Tobias Kluge dissertation 2008
DISSERTATION
submitted to the
Combined Faculties for the Natural Sciences and for Mathematics
of the Ruperto-Carola University of Heidelberg, Germany
for the degree of
Doctor of Natural Sciences
presented by
Diplom-Physicist Tobias Kluge
born in Heilbronn
Oral examination:
22.10.2008
Fluid inclusions in speleothems as a new
archive for the noble gas palaeothermometer
Referees:
Prof. Dr. Werner Aeschbach-Hertig
Prof. Dr. Augusto Mangini
Zusammenfassung
Fluideinschlüsse in Speläothemen stellen ein einzigartiges Archiv für Paläowässer dar. Diese
Arbeit beschäftigt sich mit der Untersuchung der Fluideinschlüsse in Speläothmen bzgl. deren
Verwendbarkeit als Paläotemperaturarchiv. Anhand der temperaturabhängigen Löslichkeit
verschiedener Edelgase (He, Ne, Ar, Kr und Xe) können Edelgastemperaturen gewonnen
werden. Dazu ist die Bestimmung von Edelgaskonzentrationen unerlässlich, welche neben
der Bestimmung absoluter Gasmengen auch die Messung der freigesetzten Wassermenge voraussetzt. Es werden zwei Möglichkeiten vorgestellt, mit denen die sehr kleinen, aus dem
Gestein extrahierten Wassermengen (≤ 1 µl) mit ausreichender Genauigkeit gemessen werden können. Darüberhinaus wurden verschiedene Verfahren zur Extraktion von Edelgasen
aus Speläothemen untersucht und in Bezug auf die angestrebte Anwendung analysiert. Als
besonders geeignet hat sich das Zerkleinern in einem Stahlzylinder unter Hochvakuum mittels
einer magnetisch bewegten Stahlkugel erwiesen. Außerdem werden die Edelgasaufbereitung
sowie das auf sehr kleine Gasmengen spezialisierte massenspektrometrische Messverfahren
erläutert.
Abschließend wird anhand von 6 Proben aus einer Wachstumsschicht des Stalagmiten BU-U
aus der Bunkerhöhle (Sauerland) aufgezeigt, dass es möglich ist, reproduzierbare und glaubwürdige Temperaturen zu berechnen. Die Messungen an weiteren Stalagmiten der Bunkerhöhle (BU-1 bzw. BU-2) ergaben darüber hinaus Temperaturen, die mit den erwarteten
klimatischen Bedingungen übereinstimmen. Typische Temperaturunsicherheiten liegen bei
diesen Probenstücken zwischen < 1 ℃ und 2 ℃. An den Stalagmiten BU-U und BU-1 wurden über die Messung von Edelgaskonzentrationen Paläotemperaturverläufe erstellt und in
Verbindung mit den stabilen Isotopendaten diskutiert. Diese exemplarischen Anwendungen
machen das große Potential der hier vorgestellten Methode deutlich.
Abstract
Fluid inclusions in speleothem constitute a unique archive for palaeo-waters. This thesis deals
with the investigation of fluid inclusions in speleothems and their possible use as a palaeotemperature archive. The main objective focuses on the calculation of noble gas temperatures,
which can be derived from the temperature-dependent solubility of the noble gases He, Ne,
Ar, Kr and Xe. An essential requirement is the determination of noble gas concentrations,
which implies measuring the absolute gas amounts as well as determining the amounts of
water released. Two ways of measuring the tiny water amounts (≤ 1 µl) extracted from
the speleothems are presented and will be discussed with regard to the required precision.
Furthermore, various techniques for the extraction of noble gases from the speleothems are
investigated and analysed in terms of the intended application. It turned out that crushing
under vacuum in a steel cylinder by milling with a magnetically movable steel ball is the
most suitable technique. Additionally, the noble gas preparation and the mass spectrometric
procedure, optimized for the measurement of tiny gas amounts, will be discussed.
Finally, it is demonstrated that it is possible to determine reliable temperatures from fluid
inclusions in speleothems and that the acquired results can be reproduced to a certain extent.
From the stalagmite BU-U (Sauerland, NW Germany) six samples from one growth period
were extracted and measured. Their results agree within the uncertainties although the
samples are not totally identical. Measurements on other stalagmites (BU-1, BU-2) from the
same cave revealed temperatures corresponding to the expected climatic conditions in the
respective growth period. Typical uncertainties for these samples range from ≤ 1 ℃ to 2 ℃
at most. From the stalagmites BU-U and BU-1 a temperature record has been established by
noble gas concentrations and will be discussed in combination with the stable isotope data.
These exemplary applications reveal the high potential of the method presented.
Contents
1 Introduction
5
2 Theory and basics
2.1 Speleothems as a climate archive . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1 Speleothem formation and types . . . . . . . . . . . . . . . . . . . . .
2.1.2 The importance of speleothems in climate research and possible applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.3 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Fluid inclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Definition, origin, properties . . . . . . . . . . . . . . . . . . . . . . . .
2.2.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Noble gases and common applications . . . . . . . . . . . . . . . . . . . . . .
2.3.1 Noble gases - occurrence and solubility . . . . . . . . . . . . . . . . . .
2.3.2 Temperature determination using noble gas concentrations . . . . . .
2.3.3 Further applications . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.4 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4 Noble gases from speleothems as a proxy . . . . . . . . . . . . . . . . . . . . .
2.4.1 Studies using noble gases in fluid inclusions of speleothems and minerals
2.4.2 Basic idea of climate reconstruction from noble gases in speleothems .
2.4.3 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.4 Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5 Diffusion of noble gases in speleothems . . . . . . . . . . . . . . . . . . . . . .
2.5.1 Theoretical background . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.2 Literature values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6 Adsorption effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7 Goals of this study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
9
9
13
15
16
16
18
19
19
21
22
23
24
24
25
26
26
28
29
31
31
35
3 Working with speleothems
3.1 Sample selection . . . . . . . . . . . . . . .
3.1.1 Optical methods . . . . . . . . . . .
3.1.2 Summary of the thin-section analysis
3.2 Water determination . . . . . . . . . . . . .
3.2.1 Water determination by weighing . .
3.2.2 Water determination by pressure . .
3.2.3 Precision and limits . . . . . . . . .
3.2.4 Summary . . . . . . . . . . . . . . .
3.3 Extraction of water and noble gases . . . .
3.3.1 Design of the extraction line . . . . .
3.3.2 Extraction using a metal crusher . .
37
37
37
42
43
44
46
48
50
51
51
54
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4
Contents
3.4
3.5
3.6
3.7
3.3.3 Extraction by crushing in a copper tube . . . . . . . .
3.3.4 Extraction by microwave heating . . . . . . . . . . . .
3.3.5 Extraction by thermal decrepitation . . . . . . . . . .
3.3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . .
Separation techniques . . . . . . . . . . . . . . . . . . . . . .
3.4.1 Stepwise crushing . . . . . . . . . . . . . . . . . . . . .
3.4.2 Stepwise heating . . . . . . . . . . . . . . . . . . . . .
3.4.3 Combined stepwise procedures . . . . . . . . . . . . .
3.4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . .
Mass spectrometry . . . . . . . . . . . . . . . . . . . . . . . .
3.5.1 Gas separation and purification . . . . . . . . . . . . .
3.5.2 Measurement sequences . . . . . . . . . . . . . . . . .
3.5.3 Calibration . . . . . . . . . . . . . . . . . . . . . . . .
3.5.4 Reproducibility and uncertainties . . . . . . . . . . . .
3.5.5 Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.6 Blank values . . . . . . . . . . . . . . . . . . . . . . .
3.5.7 Measurement automation . . . . . . . . . . . . . . . .
Mass spectrometric procedures and data evaluation . . . . . .
3.6.1 He and Ne measurement . . . . . . . . . . . . . . . . .
3.6.2 Argon measurement . . . . . . . . . . . . . . . . . . .
3.6.3 Magnet stability and implications for data evaluation
3.6.4 Data evaluation . . . . . . . . . . . . . . . . . . . . . .
Test of the measurement procedure with an artificial standard
4 Results
4.1 Cave air and dripwater measurements . . . . . . . .
4.1.1 Investigation of the cave air . . . . . . . . . .
4.1.2 Dripwater measurements . . . . . . . . . . . .
4.2 Extraction . . . . . . . . . . . . . . . . . . . . . . . .
4.2.1 Sample selection . . . . . . . . . . . . . . . .
4.2.2 Water determination and water content . . .
4.2.3 Air/water volume ratio . . . . . . . . . . . .
4.2.4 Applications . . . . . . . . . . . . . . . . . .
4.2.5 Theory of fluid inclusion origin and frequency
4.3 Separation techniques . . . . . . . . . . . . . . . . .
4.4 Noble gas fractionation and enrichment . . . . . . .
4.5 Dating via Helium . . . . . . . . . . . . . . . . . . .
4.6 Case studies . . . . . . . . . . . . . . . . . . . . . . .
4.6.1 Reproducibility and uncertainties . . . . . . .
4.6.2 Case study BU-1 . . . . . . . . . . . . . . . .
4.6.3 Case study BU-U . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
56
60
65
69
71
71
72
75
78
78
78
79
80
81
82
83
85
85
85
92
93
95
96
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
99
99
99
104
114
114
114
116
117
124
127
130
134
142
142
148
153
5 Summary and outlook
161
References
164
Appendix
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
List of figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
179
179
182
185
Chapter 1
Introduction
The ten warmest years since 1880 have occurred from 1995 to 2007, of which since 2001 every
year has been one of the ten warmest (DWD, 2008; WMO, 2008). Such an accumulation of
extraordinary mean annual air temperatures encourage the discussion about global warming
and climate change. Not only temperatures are rising significantly compared to the beginning
of the last century, but also the content of climate-relevant gases in the atmosphere. The CO2
concentration is now (2008) at about 385 ppm which is considerably above the highest values
at interglacial times (280 ppm) and of the glacial periods (180 ppm) as reconstructed from
ice cores from Antartica (Siegenthaler et al., 2005). Other greenhouse gases like methane are
as well far above the values of the last 650 000 years. Today a methane concentration of
about 1750 ppb is reached. In glacial and interglacial times the methane level was oscillating
between 400 and 770 ppb (Spahni et al., 2005). Such a strong modification of the natural
conditions is assumed to cause effects in the whole climate system. To estimate the impact
of man-made changes an immense number of climate models has been created and is used
to make projections with different greenhouse gas emission scenarios (e.g., HADCM3 - used
in IPCC TAR 2001, NCAR-CCSM and the today frequently used ECHAM5 - described by
Roeckner et al., 2003 ).
To achieve plausible results it is necessary to describe the complex climate system as precisely
as possible. Therefore, we have to know stable states, unstable transition states, interconnections between the different compartments, the magnitude of the connection and especially
the forcings of the system.
Instrumental weather records, in general, only date back to 1850 (Jones and Moberg, 2003)
and in exceptional cases back to the 18th century. The instrumental records started e.g. in
Austria in some places in the year 1767, s. Auer et al. (2001). Furthermore weather chronologies can be extracted from historical notes, from some hundred to up to some thousand years
back in the past (Brázdil et al., 2005). However, the information gets scarce the further we
go back in history. For modelling purposes the time span covered by instrumental records
is by far not sufficient and additionally it is difficult to derive, from the certainly subjective
notations, information about interconnections, stable states and transitions in the historical
time scale. Therefore it is imperatively necessary to resort to palaeoclimate data derived from
different archives as for example ice cores, marine and lacustrine sediments, stalagmites and
tree rings. The more precise this data, the more the archives cover the whole earth and the
more it extends over sufficiently large time scales, the better the past climate and especially
the climate changes together with forcings and triggers can be inferred and used for future
projections.
Although much is known about the interaction between forcings and climate change in principle, its details are mostly left in the dark. Exact timing, temporal and spatial development
6
Chapter 1. Introduction
and magnitude of such events are basically unknown. For example, most archives sustain
dating problems and often it is not possible to transfer the (stable isotope) signals into unambiguous and sufficiently precise temperature information. Speleothems can be dated rather
precisely, but often it is also difficult to assign temperatures to the measured isotope data.
In this work, supported in the context of the scientific group DAPHNE, we are trying to
introduce a new method to determine absolute temperatures using phyiscal principles based
on noble gas concentrations in liquids which will overcome the problems of stable isotope
data interpretation.
The basic idea
The idea of temperature determination via noble gas measurements (He, Ne, Ar, Kr and
Xe) on fluid inclusions (air-, water- or partially water-filled cavities) in calcite precipitates is
based on the temperature-dependent solubility. Low temperatures are correlated with a high
solubility and a corresponding high gas concentration in the water. Increasing temperatures
lead to a decreased solubility and gas concentrations. As noble gases are not affected by
chemical reactions, the noble gas concentrations in water are a direct proxy for temperature.
In the case of groundwater the temperature determination using noble gas concentration is a
well established method with precise results (uncertainty <1℃), but lacks a good temporal
resolution as well as a precise dating. In contrast, calcite precipitates in caves, as e.g. stalagmites or stalactites, summarized under the term speleothems, are precisely datable up to
500 kyr with U-Th and further with U-Pb and enable in some cases even seasonal resolution.
Due to the commonly present inclusions speleothems constitute a unique archive for palaeowater. In the fluid inclusions, there exist small water amounts which contain noble gases
corresponding to the climatic conditions during growth. Measurements of the water amount
as well as the dissolved noble gases can give information about palaeotemperatures using the
well-known solubility temperature dependence.
In the past some effort has been made to use this promising archive. So far no successful
data has been obtained with regard to this objective. However, we have been able to develop
a suitable extraction procedure for noble gases and water as well as a sufficiently precise
measurement method for the released water and the extracted noble gases, which allowed
temperature determination at selected speleothems with an uncertainty of 1 ℃ or even less.
The water is determined with a manometrical method and the noble gases are measured by
a sector-field mass spectrometer.
The major complication is the presence of air-filled inclusions. In general, they contain noble gases with the typical mixing ratio of atmospheric air delivering no information about
palaeotemperatures. Stepwise extraction procedures may enable the application of this
method also to samples with a larger air content as it separates air- and water-filled inclusions to a certain extent.
Case studies on samples from Bunker cave in North-Western Germany resulted in reproducible temperatures for a growth layer and yielded reasonable temperature differences between the Early Holocene (∆T - 4.2 ℃, 11-12 kyr BP), a warm period during the last Glaciation (∆T - 2.6 ℃, 53 kyr BP), and the Eemian (∆T + 3 ℃, 125 - 134 kyr BP) compared to
more recent values (1.3 kyr BP).
This study shows the potential of noble gas measurements on fluid inclusions for temperature
determination and presents methods to achieve the aimed resolution in temperature of better
than 1 ℃.
7
The experiments
The key to precise temperatures is a suitable extraction. A high extraction efficiency is
necessary to obtain water and gas amounts in a measurable range. In addition, low background values as well as a small air contribution to the total noble gas signal have to be
achieved. Gases from air-filled inclusions mask the temperature information of the dissolved
noble gases. An investigation of different procedures like squeezing in copper tubes, heating
with micro waves and crushing with a steel ball under vacuum, revealed that crushing in
a cylinder with a steel ball to be the best method. It offers a high efficiency and enables
to partially separate the water- and air-filled inclusions by a stepwise crushing procedure in
combination with heating. Furthermore the blank can be reduced by preheating and is in
general better controllable compared to squeezing in copper tubes.
With regard to the water determination two methods have been investigated. Highest precision in the typical scale (0.1 - 1 mg) can be achieved by a manometrical method, measuring
the water vapour pressure in adequate and calibrated volumes.
Outline
First an overview over the different forms of speleothems and their importance in climate
research is given. The fluid inclusions are of major importance in this project and are presented in the subsequent section. The second part of the first chapter focuses on noble gases,
their occurrence, their solubility as well as possible applications in the different research
fields. Especially, we discuss the constraints for precise temperature determination via noble
gas measurements. Diffusion and adsorption can influence the results and are also briefly
discussed.
In the third section of this work the main elements for successful noble gas studies on
speleothems are presented. The first part explains which samples are suitable and how they
can be selected in advance using simple inspection by eye and microscopy. The following
part investigates two ways of water determination in the mg range and their advantages as
well as their disadvantages. The main part presents several methods for gas and water extraction as well as stepwise procedures. Finally an overview on the basics of gas processing,
mass spectrometric measurement and data evaluation is given. Problems related to the mass
spectrometric measurement and the use of artificial standards are briefly discussed.
The last section presents the main results covering sample selection and extraction methods.
In a special section the results of noble gas measurements of cave air, drip water and water
from a cave pond are described. The effect of stepwise extraction for separation of waterand air-filled inclusions is discussed on exemplary data. Radiogenic He and the water content
exhibit interesting features with regard to dating, respectively paleoclimate and are also
presented in this chapter. The main part is focused in the temperature determination and
discusses several case studies. The first gives an overview over repeated measurements on a
growth layer with regard to reproducibility of temperature. The second case study presents
temperature values over a large part of the Holocene and compares them with published data.
Finally, paleotemperatures are given for the Eemian, a period during the last Glaciation and
the Early Holocene, and are compared to published studies. Furthermore, they are discussed
in combination with the stable oxygen and carbon isotopes obtained from the same stalagmite.
The chapter is closed with a short summary and a brief outlook for future research related
to this topic.
8
Chapter 1. Introduction
Chapter 2
Theory and basics
2.1
Speleothems as a climate archive
In the last decade the importance of speleothems as a climate archive has strongly increased.
Due to new techniques such as laser ablation and improved detection processes even annual
resolution can be achieved for stable isotopes and trace elements. Their use in the climate
reconstruction as well as advantages and limits will be discussed in the following sections.
2.1.1
Speleothem formation and types
In caves not only stalactites and stalagmites are abundant, there also exist flowstones, soda
straws and lots of other types of precipitated calcite (Fig. 2.2). In this chapter the different
species of speleothems together with their formation will be presented.
Speleothems are formed by precipitation of carbonates inside the cave. Rain water can absorb
large amounts of CO2 while percolating through the soil, as the CO2 partial pressure is very
high (about 0.1 atm, see White, 1976) in the soil air. Meanwhile carbonic acid is produced
according to the following balance reactions, the first one is important for low pH-values,
where the second equation is valid for pH ≥ 7 (see Buhmann and Dreybrodt, 1985):
H2 O + CO2 H2 CO3 H+ + HCO−
3
(2.1)
CO2 + OH− HCO−
3
(2.2)
The elevated content of carbonic acid in the water dissolves calcium carbonate from the host
rock and soil material. This process can be described by the so-called Plummer-WigleyParkhurst equation (Plummer et al., 1978):
CaCO3 + H+ Ca2+ + HCO−
3
(2.3)
CaCO3 + H2 CO3 Ca2+ + 2HCO−
3
(2.4)
−
CaCO3 + H2 O Ca2+ + HCO−
3 + OH
(2.5)
In the cave itself the CO2 partial pressure is higher than in ambient air, e.g. 1.5 · 10−3 atm
like in Obir Cave (Spoetl et al., 2005), but in general rarely above 3 · 10−3 atm (White, 1976)
and thus significantly lower than in the soil air. Therefore, reactions are shifted towards the
left side. The water is supersaturated in carbonate and precipitation of calcite occurs.
According to growth conditions and water composition different crystals are formed. The
most common type is calcite, other more rare forms are aragonite and dolomite. Aragonite is
10
Chapter 2. Theory and basics
Dolomite
Aragonite
Calcite
CO3
Mg
rhombohedron
Ca
Carbonate
dipyramid
Figure 2.1: In the upper row the unit cell configuration of calcite and dolomite as well as the form of an
aragonitic crystal is shown (adapted from Ford and Williams, 2007). Below, possible ideal configurations
of calcite as e.g. the rhombohedron and the dipyramidal configuration are displayed. Three others
configurations are also possible: pinacoid, prism and scalenohedron (adapted from Ford and Williams,
2007). Images of crystals which grew under laboratory conditions and in cave environment can be found
in Fig.2.3 and Fig.2.6 respectively.
F
D
E
B
C
A
Figure 2.2: Selection of common speleothem forms in a cave. A: flowstone, B: stalagmite column,
C: stalagmites, D: drapery, E: stalactites, F: soda straws. Modified from a photo of Bunnell (2006).
2.1. Speleothems as a climate archive
11
the high-pressure polymorph of CaCO3 and is supposed to be the third most common cave
mineral (the second being gypsum, White, 1976). At typical cave temperatures, its solubility
is about 11% higher than for calcite. Dolomite is an ordered structure in close relation
to calcite, but with alternating layers of magnesium and calcium (White, 1976). Actively
forming dolomite has rarely been observed. Therefore, in the following we focus on the most
common type, the calcite minerals.
Typical properties of the three carbonates are listed in Table 2.1 and the according mineral
forms are displayed in Fig. 2.1.
Table 2.1: Properties of the carbonate species calcite, aragonite and dolomite. Data from Ford and
Williams (2007).
mineral
chemical composition
Calcite
Aragonite
Dolomite
CaCO3
CaCO3
CaMg(CO3 )2
specific weight
(g/cm3 )
2.71
2.95
2.85
hardness
(Mohs scale)
3
3.5 - 4
3.5 - 4
crystal type
rhombohedral
orthorombic, dypramidal
hexagonal, rhombohedral
Speleothems can be distinguished in 3 different groups:
• dripstone and flowstone forms
– stalactites
– stalagmites
– draperies
– flowstone sheets
• erratic forms
– shields
– helictites
– botrioidal forms
– anthodites
– moonmilk
• sub-aqueous forms
– rimstone pools
– concretions
– pool deposits
– crystal linings
The most important group is the first one. Stalagmites are the most prominent speleothems
due to their simple stratigraphy. They are built on ground by water dripping on top and thus
growing up to the cave ceiling. The older parts are always below the younger layers. The
stalagmite growth is strongly correlated to temperature and the amount of carbon dioxide
available in the soil (Kaufmann, 2003) as well as the content of calcium in the dripwater
(Genty et al., 2001). In a cave in England growth rates ranging from 0.04 to 0.16 mm yr−1
are reported with a mean of 0.077 mm yr−1 , in a cave in France 0.2- 0.9 mm yr−1 and in
a artificial tunnel in Belgium of up to 1.3 mm/a (Baker et al., 1998). Genty et al. (2001)
12
Chapter 2. Theory and basics
Figure 2.3: SEM image of a calcite which grew in a channel experiment under controlled and constant
laboratory conditions. The image was made at the Faculty of Chemistry of the University of Heidelberg.
reported similar values from six caves in Europe with a maximum growth rate of 1.5 mm yr−1 .
The stalagmite shape varies due to the calcite deposition process. An increased drip rate
leads to a larger stalagmite diameter (Franke, 1965). For drip intervals ≥ 1000 s a minimum
equilibrium diameter of about 4 cm is established (Kaufmann, 2003; Curl, 1973). Other
effects like drip fall height (Gams, 1981) also influence the shape as well as temperature and
soil CO2 variations. A uniform diameter implies constant growth conditions over a long time
period. Terraces suggest periodic variations and a conical style implies a decrease in drip rate
(White, 1976). However, due to the complex interdependence of stalagmite stratigraphy and
palaeoclimate it is not possible to infer the climate information only from shape- and growth
information (Kaufmann, 2003).
The inverse form of stalagmites, which grow from the top of the cave to the bottom, are
stalactites. As sometimes the water is flowing through the stalactite core, it is difficult to
estimate the age of different growth layers. Sections cut perpendicular to the growth axis
usually show structures analog to tree rings which is due to the additional flow at the outside
of the stalactite. As the water flow channels are quite large, sand grains and other material
can be transported as well. Blockage due to this matter transport can form stalactites with
a complicated growth history. The size of stalactites is limited by weight.
Soda straws or soda stalactites are long and thin calcite tubes. The water percolates through
these tubes, and a small amount of mineral material is deposited as a ring similar to the drop
diameter. The soda straw is growing downwards and also on the outside, because water can
seep between boundary layers. If the central channel is blocked, the water is forced to flow
on the outer side and a conical stalactite may be formed.
Another group of speleothems which can be dated rather well are flowstones. They form in
places where water is flowing over cave walls. According to the pathways different shapes like
2.1. Speleothems as a climate archive
13
flowstone cascades, draperies and ”bacon-strip” forms can occur. Growth rates reported from
three sites in southeastern England range from 0.009 to 0.05 mm yr−1 (Baker and Smart,
1995), which is about one order of magnitude less than the mean of most stalagmites.
Subaqueous forms, like pool deposits may be a very well suited speleothem for inclusion
measurement, as they grew below the water surface. Therefore, few air-filled inclusions can
be expected. The water-filled inclusions may be predominant. Due to CO2 loss from the pool
surface, water can be supersaturated and lead to calcite precipitation. Growth is most rapid
near the water surface as this is where the loss of CO2 and the strongest supersaturation
occur.
Due to their shapes, the erratic forms can hardly be used for palaeoclimatic purposes. It
will be rather difficult to address representative ages to such a speleothem. However, these
special calcites show interesting forms. The helictites for instance grow in curved respectively
spiraled paths. In case of helictites the crystal growth is dominant. The crystallization follows
the fast growing mineral axis of the calcite. The rapid changes in growth directions are caused
by blockage of the drip channel through clastic material. Similar but larger as the helictites
are eccentric stalactites. Botryoidal forms are globular protuberances on the cave walls.
Anthodites are chunks of crystalline aragonite growing radially.
2.1.2
The importance of speleothems in climate research and possible applications
Speleothems constitute a unique archive for climate information on continents due to their
properties and also due to the environment in which they grow. The cave itself is protected
against direct human influence and against fast erosion, and is therefore able to conserve
information about the past climate even in urbanized or heavily polluted areas.
Speleothems offer a variety of environmental tracers like stable isotopes, trace elements and
trapped pollen grains as well as parameters such as inter-annual thickness variation of growthlaminae, growth rate changes, organic acid content and dust layers. Mostly, stable isotope
signals are used in combination with precise dating to estimate the onset and duration of
climate events, such as the transition from glacial to interglacial, Dansgaard-Oeschger oscillations or short-term climate changes with high amplitudes like the 8.2 kyr event. Even
growth rates provide information about the climate. In glacial times the growth rate slows
down and, for instance, during Marine Isotope Stage 2 no calcite precipitation is found on
the Northern Hemisphere (≥ 35◦ N). Often slower growth rates can be related to a cooler
climate (compare review of McDermott, 2004).
This data gains importance with the increasing use of global circulation models. The information from speleothems can be added to the data from ice-cores, marine and lacustrine
sediments, pollen and tree rings to test and validate the models. Well dated speleothems
exhibit a high potential in refining the chronology of high-latitude ice-cores, as many of the
stable isotope features are also found in speleothems (Spoetl and Mangini, 2002). Furthermore, they can give insight into forcing and feedback mechanism of the global climate system.
Short-term changes like the 8.2 kyr event or the Medieval Warm Period as well as major
climate changes are saved in the isotopic composition of the growth layers. As there has been
proven annual lamination in some stalagmites (Frisia et al., 2003), it is possible to derive high
resolution data from speleothems. With laser-ablation gas chromatography isotope ratio mass
spectrometry (LA-GC-IRMS) a spatial resolution of 250 µm can be achieved (McDermott
et al., 2001), which corresponds to 4 years at a mean annual growth rate of 80 µm. Even
better resolution was achieved by Frappier et al. (2002) with about 20 µm using micro-milling.
Similar resolution can be obtained by ion microprobe techniques (Kolodny et al., 2003). In
addition, this archive provides samples over time spans of thousands of years up to some
14
Chapter 2. Theory and basics
Figure 2.4: Picture of different speleothem forms. On the left side a huge stalagmite (’witchs finger’ in
the Carlsbad Cavern) can be seen. The upper picture on the right side shows a flowstone drapery, in the
middle soda straws and on the bottom some helictites have been photographed.
2.1. Speleothems as a climate archive
15
hundred thousand years. So in principle speleothems can cover the whole range of the earth’s
climate, which is governed by the glaciation - deglaciation pattern with a 100 kyr period.
In caves the temperature is typically rather constant throughout the year. The recorded
temperature signal reflects the mean annual air temperature of the cave region (McDermott,
2004; Genty et al., 2002). However, common studies focus on the precise estimation of timing
and duration of major stable isotope-defined climatic events as it is hardly possible to achieve
unambiguous temperatures due to kinetic fractionation effects and difficult relations between
the absolute values of stable isotopes and environmental parameters at a certain sample site.
As a new and independent proxy, noble gases in fluid inclusions of speleothems will be used
to determine absolute temperatures and temperature differences between different climatic
periods. Speleothems can become one of the most important climate archives as they can
provide precise dating and unambiguous temperatures using noble gas measurements. In
combination with the high-resolution stable isotope data it may be possible to reconstruct
the past climate in a complete way.
2.1.3
Limitations
The use of stable isotopes and to some extent that of trace elements is restricted as the
reason for changes in isotope data sometimes is difficult to figure out. Some solutions are
ambiguous, and only in very few cases it is possible to assign a signal change in stable isotopes
directly to a change in temperature. The change in stable isotopes could be caused by a set of
factors e.g. changes in the isotopic composition of the precipitation or changes in the isotope
fractionation during calcite deposition. δ 18 O and δD can be influenced by changes in the
moisture source, changes in the δ 18 O value of the ocean surface due to increased or decreased
continental ice volume, changes in the fraction of non-oceanic precipitation or by an increased
or decreased amount effect (McDermott, 2004). Even if vegetation changes above the cave,
a feedback in isotope data can be expected.
However, with an independent temperature calculation, using noble gases from fluid inclusions, it should become feasible to derive the different components of the total isotope signal,
to assign a certain signal variation to the temperature change and to changes in the moisture
source and the type of precipitation as well as to modifications of land cover and calcite
precipitation.
For very old stalagmites the archive is limited in dating. The uranium - thorium method
can only be used for up to 500 000 years. If the stalagmite is older than 300 kyrs or 400
kyrs, the precision in dating gets smaller or it even gets impossible to determine a reliable
age. Nervertheless, using the helium data from the noble gas measurements, a rough age
estimation may be possible. Furthermore, geochronological methods like U/Pb (Richards
et al., 1998; Polyak et al., 2008) and possibly U/He may be applied as well.
16
2.2
Chapter 2. Theory and basics
Fluid inclusions
In this section origin, composition and properties of fluid inclusions in stalagmites are explained. Furthermore, possible and actually performed applications are presented.
2.2.1
Definition, origin, properties
Fluid inclusions are microscopic cavities within a crystal, filled with gas or liquids or with a
mixture of both. Fluid inclusions are present in approximately all minerals as they formed
mostly in presence of liquids or came into contact with liquids when fractures were built or
healed (Roedder, 1984). Even in certain meteorites liquid inclusions are reported (Zolensky
et al., 1999; Bodnar and Zolensky, 2000).
Figure 2.5: Thin section of a sample from the stalagmite H12 from Oman. The dark irregular shaped
inclusions are assumed to be only air-filled. The light-reflecting, round or ellipsoidal inclusions are waterfilled.
Fluid inclusions are in general very small, about some µm and in rare cases up to 50 µm.
Generally they are ellipsoidal or sausage shaped and are frequently distributed inhomogeneously (Schwarcz et al., 1976). Zones with a high abundance of water-filled inclusions are
milky white (Genty et al., 2002), especially in the case of a coarse crystalline type. In these
samples about 109 inclusions/cm3 are assumed (Roedder, 1984). Fluid inclusions typically
constitute trace amounts up to 3 ‰ of the total speleothem sample weight (Serefiddin et al.,
2002), written as ‰ wt. In addition to the liquids in fluid inclusions, water molecules can
be trapped along grain boundaries or even in the calcite matrix and may contribute to the
released water in special extraction steps like, for example, strong heating.
Water-filled fluid inclusions in speleothems are typically generated by entrapment of water
in voids inside the crystal granules. Scheidegger et al. (2007a) reported that they found
water-filled inclusions always to be located inside the calcite crystals. Furthermore, most
2.2. Fluid inclusions
17
Figure 2.6: SEM image of a single calcite crystal, which grew in the Bunker Cave on a glass plate
within one year. Picture with courtesy by Dr. R. D. Neuser / D. Riechelmann.
of the water-filled inclusions seem to contain a gas-filled bubble. As air-filled inclusions are
different in their location (in general between the calcite crystals) and size (20-60 µm, Scheidegger et al.,2007a) it may be possible to separate the two fractions by adequate extraction
techniques (Scheidegger et al., 2007b).
The speleothem is formed due to the degassing of CO2 from the carbon saturated water
(further details: see Schwarcz (1986); Ford and Williams (1989)). Typically, the cave interior
shows a high humidity and therefore evaporation can be neglected. Thus, the enclosed water
in fluid inclusions reflects the characteristics of dripwater. As the dripwater normally shows
negligible salinity this can be assumed as well for fluid inclusions.
A considerable difference of calcite growth under laboratory conditions and natural conditions can be seen with regard to the shape the calcite establishes and the deviation from
the expected shape. In laboratory very regular crystals with a precisely defined shape are
detectable (Fig.2.3), whereas in the cave the crystals show strongly different behaviour. As
it can be seen on the SEM image of a calcite sample from Bunker Cave (s. Fig.2.6) there
is an overall crystalline feature combined with an unexpected crater type pattern, possibly
referring to micro-crystalline growth. The little holes are in general very small (some µm)
and scattered over the whole surface. Comparing their size to the typical inclusion dimension,
it gives the impression that these holes might be the preceding structure responsible for the
amount and size of the fluid inclusions.
With respect to the spatial distribution of fluid inclusions, thin section analysis is a useful
tool. In many thin sections a sequence of inclusion-rich and inclusion-free layers can be
detected. In Fig. 3.2 we can see on the upper left side a significant change in the inclusion
pattern from an area with few and large inclusions to an area dominated by bands of rather
small inclusions. A similar pattern showing a change between inclusion-rich and inclusion-free
layers can be seen in Fig. 3.3.
18
Chapter 2. Theory and basics
According to this, the reason for such patterns is an interesting question. Brook et al. (1999)
investigated two stalagmites which showed a strong layered pattern. By using Sr isotopes they
proved that inclusion-poor layers are correlated with a slow calcite precipitation and inclusionrich layers with faster precipitation. In accordance with this data, they derived a similar
correlation using the MgCO3 content of the calcite. Inclusion-poor layers are correlated with
low MgCO3 and the inclusion-rich parts with a higher MgCO3 - value. Higher MgCO3 content
is related with a faster precipitation caused by an increase in the saturation state (Goldstein
and Reynolds, 1994). Brook et al. (1999) concluded from their data, that the inclusion-rich
layers are deposited during dry seasons and the inclusion-poor layers during wet seasons.
Similarly, Genty and Quinif (1996) interpreted couplets of dark and light laminae as annual
cycles caused by differences in the water excess. Thus, it is possible to derive information
about the climatic conditions from the chronology of different layers. Furthermore, this
argument can be turned around and can be used for the search for adequate samples with a
high content of inclusions.
2.2.2
Applications
As the fluid inclusions preserve information about the environment at time of enclosure, they
are used to derive this unique data from the past.
In stratigraphy and sedimentation they are used to identify the provenance of detrital grains
in sandstone, quartzite and conglomerate; in geology for studies concerning the petrogenesis
and the tectonics, for instance to trace erosion and uplift; in astrophysics to reconstruct
extraterrestrial or possible early terrestrial processes from lunar and meteoritic samples.
Furthermore, fluid inclusions are applied in the search for oil as they can provide information
about the tectonics and pressure-temperature evolution of the oil-bearing basin. Even in
security investigations of nuclear waste and nuclear reactor sites they can provide data about
the properties of the underlying rocks, for instance, the last movement of faults at this site
(for more detail see Roedder, 1984).
In the field of climate research some attempts have been made to derive information about
precipitation and temperature. Most of the studies focused on stable isotope signals in fluid
inclusions and not on noble gases. In the last decade remarkable advances in extraction and
measurement techniques had been achieved with regard to the stable isotopes. Therefore, it
was possible to derive reliable stable isotope data from fluid inclusions (Matthews et al., 2000;
Dennis et al., 2001; Genty et al., 2002; Serefiddin et al., 2002; McGarry et al., 2004; Zhang
et al., 2008). However, it is rather difficult to extract and especially to collect the released
water entirely, if the extraction is not performed under vacuum conditions. Due to adsorption
of water molecules significant fractionation of the stable isotopes can occur. Furthermore,
the measured isotopic signal may not only reflect undisturbed dripwater, but can also be
influenced by a post-depositional isotopic exchange with the host calcite (Schwarcz et al.,
1976; Wilkinson, 2001).
Some pilot projects have been performed to assess the potential of entrapped air in fluid
inclusions in order to reconstruct palaeoclimatic conditions. Scheidegger et al. (2007b) and
Badertscher et al. (2007) investigated N2 , O2 , N2 O and Ar as well as the gases methane
and carbon dioxide using dynamic mass spectrometry. Methane and the CO2 -concentration
may be of special interest in the air-filled inclusions. As they have been closely related to
the climatic cycles, their concentration development along the growth axis may give valuable
insights into timing and evolution of the signals as well as the connection between temperature
change and variations in the methane respectively CO2 level.
2.3. Noble gases and common applications
19
With regard to noble gas measurements on fluid inclusions aiming at temperature determination few studies have been published (Böhlke and Irwin, 1992b; Stuart and Turner, 1992;
Scheidegger, 2005) or presented on conferences (Scheidegger et al., 2007b, 2006). This may
be due to the fact that nobody has successfully derived absolute noble gas temperatures with
palaeoclimatic implication from speleothem fluid inclusions so far. In this thesis, the method
for a successful noble gas analysis and temperature calculation as well as the results of the
most striking measurements are presented.
2.3
Noble gases and common applications
In this section the basics about noble gases and their applications in the research field of
palaeoclimatology will be presented. Noble gases are widely used in groundwater research in
order to investigate flow regimes and residence times on the one hand and on the other hand
to deduce information about the past climate.
2.3.1
Noble gases - occurrence and solubility
Noble gases owe their name due to their chemical inert behaviour. They build only very few
real compounds, otherwise they normally exist in atomic form. The commonly used noble
gas isotopes in palaeoclimate and dating studies are 3 He and 4 He, 20 Ne and 22 Ne, 36 Ar, 40 Ar,
Kr- and Xe-isotopes with high abundances like 84 Kr and 132 Xe and the short-lived 222 Rn.
The standard and reference for noble gas measurements is the terrestrial atmosphere. Typical
atmospheric mixing ratios, which are assumed to have stayed constant over the last million
years, are shown in Table 2.2. Water samples in equilibrium with the atmosphere yield noble
gas concentrations corresponding to these values and the solubility of each isotope. However,
in some samples different values can be found. 3 He, 4 He and 222 Rn are formed by radioactive
decay or other nuclear processes and can therefore show an excess in older samples compared
to equilibrium values. Radon itself is also radioactive and decays further; the other mentioned
noble gases are stable.
Beyond the atmospheric abundance, neon can be formed by nucleogenic reactions in the
crust or mantle (Yatsevich and Honda, 1997). The fraction of neon due to this source is
related to the radio-element and target-element concentration. The present-day production
rate for 20 Ne lies at about 1.4·10−23 ccSTP yr−1 g−1 in mantle and 1.5·10−21 ccSTP yr−1 g−1
in crustal material and in the case of 22 Ne 1.5·10−24 ccSTP yr−1 g−1 in mantle-, respectively
5.2·10−21 ccSTP yr−1 g−1 in crustal material (Leya and Wieler, 1999).
As in the case of neon, argon may have contributions beyond the atmospheric reservoir. 40 Ar
production is dominated by 40 K decay and therefore proportional to the potassium concentration. About 102 atoms 40 Ar are produced in this way per µg of K and year (Ballentine and
Burnard, 2002). 36 Ar production can be neglected compared to the background of ambient
36 Ar.
Krypton and xenon are rare noble gases in the environment and are typically dominated by
the constant atmospheric abundance. However, 84 Kr and 132 Xe can also be produced to a very
small extent by fission of 238 U. Besides this dominant process, fission of 232 Th and 235 U can
be a further production path. The present day crustal production rate for 84 Kr (2.1·10−23 ccSTP yr−1 g−1 ) is about 10 orders, for 132 Xe (5.2·10−22 ccSTP yr−1 g−1 ) about 9 orders of
magnitude smaller compared to the production rate of 4 He of 3.3·10−13 ccSTP yr−1 g−1 (g
referred to one gram of crustal material). Even compared with 40 Ar it is about 10 respectively
8 orders of magnitude smaller (Ballentine and Burnard, 2002). In the case of samples with no
extraordinary composition (e.g. high uranium contents) and a moderate age, all nucleogenic
as well as the fissiogenic contributions to the measured signals can be neglected.
20
Chapter 2. Theory and basics
Table 2.2: Atmospheric mixing ratios of noble gases (compare Ozima and Podosek (2001) or Porcelli
et al. (2002))
isotope
3
He
He
20
Ne
21
Ne
22
Ne
36
Ar
40
Ar
84
Kr
132
Xe
22
Rn
4
volume mixing ratio
7.33 10−12
5.24 10−6
1.645·10−5
4.872·10−8
1.678·10−6
3.14·10−5
9.30·10−3
6.50·10−7
2.34·10−8
∼ 6·10−20
As the atmospheric noble gas mixing ratio is constant, valuable information can be deduced
from noble gases dissolved in water, for instance from groundwater or even from fluid inclusions in speleothems. At equilibrium, the dissolved concentration (Ciwater ) of a noble gas i
gas
in water is proportional to the concentration in the gas phase (Ci ) as it is described by
Henry’s law:
gas
Ci = Hi (T, S) · Ciwater
(2.6)
The proportionality factor Hi , called Henry coefficient, inversely describes the solubility of
the gaseous element in the fluid and depends on temperature T and salinity S and is different
for each gaseous element i. Furthermore the equilibrium concentration Ci,water
of water in
eq
contact with air is dependent on the atmospheric pressure p. Therefore the atmospheric
equilibrium concentration Ci∗ is often used, which is the dissolved concentration of a gas i in
water at equilibrium with vapour saturated air at a total pressure of p0 = 1 atm:
Ci∗ =
(p0 − e) · zi
patm
|p=po
i
=
Hi (T, S)
Hi (T, S)
(2.7)
patm
is the partial pressure of the corresponding noble gas i, e the saturation vapour pressure
i
and zi the volume fraction of the gas in dry air (s. Table 2.2). The Henry coefficient Hi is not
dependent on the pressure p, but as mentioned above the equilibrium concentration Ci,water
eq
changes with pressure. The complete formula is then:
Ci,water
eq =
(p − e) · zi
(p0 − e) · zi p − e
·
=
Hi (T, S)
p0 − e
Hi (T, S)
(2.8)
The solubility strongly depends on temperature. As temperature increases the solubility
decreases. This effect is most pronounced for the heavy noble gases. A temperature change
from 0℃ to 30℃ reduces the solubility of e.g. He by only 11%, whereas it is about 63% in
case of Xe.
Normally the salinity plays a minor role as typical groundwater has a negligible salt content.
The effect of salinity on the solubility is also stronger for the heavier noble gases, but the
difference between light and heavy elements is less pronounced. He solubility is reduced by
about 0.5% per g of salt in 1 kg water, in case of xenon it is about 0.67 % per g of salt in
1 kg water.
As described in the last equation, Ci,water
is directly proportional to the ratio of the actual
eq
pressure and to p0 in case of dry air. The pressure p changes according to the barometric
2.3. Noble gases and common applications
21
Table 2.3: Noble gas concentrations of water equilibrated with atmospheric air at a pressure of 1013.25
mbar and no salinity. Data calculated using solubilities of Weiss (1971) for He and Ne, of Weiss (1970)
for Argon and of Weiss and Kyser (1978) for Kr. Xenon concentrations are calculated using the data
presented by Clever (1979). Rn - values are corresponding to an air activity concentration of 1 Bq/m3
and solubilities given by Weigel (1978).
isotope
3
He
He
20
Ne
22
Ne
21
Ne
36
Ar
40
Ar
84
Kr
132
Xe
222
Rn
4
0℃
noble gas concentration in cc/g at
10℃
20℃
30℃
6.66·10−14
4.90·10−8
2.04·10−7
2.08·10−8
6.03·10−10
1.68·10−6
4.96·10−4
7.07·10−7
5.18·10−9
9.03·10−15
6.32·10−14
4.65·10−8
1.83·10−7
1.86·10−8
5.41·10−10
1.30·10−6
3.85·10−4
5.19·10−8
3.54·10−9
6.22·10−15
6.09·10−14
4.48·10−8
1.68·10−7
1.71·10−8
4.96·10−10
1.05·10−6
3.11·10−4
3.97·10−8
2.56·10−9
4.50·10−15
5.94·10−14
4.36·10−8
1.56·10−7
1.59·10−8
4.61·10−10
8.74·10−7
2.59·10−4
3.15·10−8
1.93·10−9
3.45·10−15
height formula. At 1000 m above sea level the p/p0 ratio is 0.886 and leads to a correspondingly reduced noble gas concentration in the liquid phase.
With regard to fluid inclusions of speleothems it is assumed that the salinity is in a first order
approximation related to the dripwater. As in most cases the dripwater shows no significant
values the salinity is in general negligible. Niggemann (2000) made an intensive monitoring of
different caves in north-western Germany. Dripwater always showed electrical conductivities
in the range of groundwater. Similar values have been reported by Galdenzi and Maruoka
(2003). They measured a value of 0.2 - 0.4 g/l in case of water from the vadose zone and even
in rather rare sulfidic groundwater values did not exceed 2 g/l. In an intensive monitoring of
gypsum karst waters Krawczyk and Ford (2007) could emphasize this finding, as almost all
water samples showed salinities below 4 g/l.
However, data derived by microthermometry of fluid inclusions indicate that in some special
cases a considerable salinity may have built (Scheidegger et al., 2007b). Thus, prior to noble
gas data analysis this point has to be investigated.
2.3.2
Temperature determination using noble gas concentrations
Using the temperature dependent solubility we are able to derive temperatures from measured noble gas concentrations. The so-derived values are called noble gas temperatures
(NGTs). The noble gas concentrations are the total extracted noble gas amounts related to
the corresponding water volume.
The simplest case, water in equilibrium with the ambient air, the so-called air-equilibrated
water (AEW), only consists of the dissolved atmospheric noble gas component. Typical
noble gas concentrations of water in equilibrium with the atmosphere are shown in Table 2.3.
However, in case of groundwater an additional component, referred to as excess-air, is found
(Heaton and Vogel, 1981, 1979; Andrews and Lee, 1979). Through fluctuations of the water
table bubbles may be included and partially or totally dissolved afterwards (Holocher et al.,
2002). This causes in groundwater the additional excess-air contribution to the equilibrium
noble gas concentrations, which in general consists of a certain amount of atmospheric air. If
the bubbles are only partially dissolved an elemental and isotopic fractionation may occur.
The excess-air contribution is handled with an additional parameter, the air-water volume
ratio A (= Vair /Vwater ), respectively the Ne-excess ∆Ne. An air/water volume ratio of 0.1
22
Chapter 2. Theory and basics
indicates that the measured noble gases originated from a ten times larger water volume
compared to the air volume, e.g. 0.1 µl air and 1 µl water. If fractionation plays a role,
another parameter (F ) is included in the calculation. Nevertheless, all these complications
can be managed with sufficient precision due to adequate inverse modeling concepts developed
by Aeschbach-Hertig et al. (1999) and Ballentine and Hall (1999). By using the whole dataset
of neon, argon, krypton and xenon it is possible to derive all unknown parameters such as
temperature T , excess-air A and fractionation F .
With respect to the speleothems, the pressure parameter p can be fixed according to the
height above sea level, as the location of the cave as well as the position of the speleothem
is known. Fractionation can be excluded due to the incorporation process of water in the
speleothem. The water gets equilibrated before being closed-off in calcite. The excess-air
consists of air-bubbles inside the water inclusions, respectively of totally air-filled inclusions.
As free parameters for the inverse modelling temperature T and excess air A are remaining.
They can be well fitted in the over-determined system by using neon, argon, krypton and
xenon. The simplest model approach consists of the unfractionated excess-air model (UAmodel):
(2.9)
Ci = Ci,eq · (1 + A · Hi (T ))
In this equation, Ci describes the modeled concentration, Ci,eq the volumetric concentration
of the noble gas i in air-equilibrated water (AEW), Hi is the dimensionless Henry’s law
coefficient and A the volume ratio of enclosed air to water. This model is used in the
following for the speleothem samples to derive NGTs.
In contrastto this, groundwater studies have to take into account the possible fractionation
processes and use therefore more sophisticated models, such as the partial re-equilibration
model (PR-model, Stute et al., 1995) or the closed-system equilibration model (CE-model,
Aeschbach-Hertig et al., 2000). Fitting the neon, argon, krypton and xenon concentrations in
groundwater samples to the model by χ2 -minimization, reasonable temperature differences
between distinct climatic periods (Aeschbach-Hertig et al., 2002) with small errors, usually
well below 1 ℃ (Aeschbach-Hertig et al., 1999), can be obtained. Furthermore, Stute and
Sonntag (1992) and Stute and Schlosser (1993) could show, that the NGTs are related to
the ground temperature during recharge. As cave temperatures, like ground temperatures,
are closely related to the mean annual air temperature (Gascoyne, 1992; McDermott, 2004),
the fluid inclusions and therein dissolved noble gases provide information about the past
climate. Thus, the idea concerning temperature reconstruction via the solubility is expanded
to speleothems and discussed in the following chapters.
2.3.3
Further applications
Noble gases can be used for a widespread set of applications, from tracing of flows to dating
purposes and investigation of palaeoclimatic conditions. In the following, some exemplary
applications related to the work with speleothems are presented.
3 He, 4 He
and 222 Rn are well suited for dating purposes. 222 Rn is a short-lived radio-isotope
with a half-life of 3.82 d. It can be used to trace submarine groundwater discharge (Cable
et al., 1996; Corbett et al., 2000) and infiltrating groundwater in lakes due to the strong
concentration gradients between the two reservoirs (Kluge et al., 2007) or it can be used to
estimate the travel time of artificially recharged water in aquifers as well as the flow velocity
of the groundwater body (Hoehn and von Gunten, 1989). Because of the short half-life the
applications are limited to a maximum of two to three weeks. Afterwards the signal is lost
due to natural decay. Applied in caves radon can be a useful tool to investigate the ventilation
(Hakl et al., 1996) or to control the dripwater whether a degassing has just occurred.
2.3. Noble gases and common applications
23
4 He
can be used at least as a semi-quantitative tool for dating of old groundwater (Andrews
and Lee, 1979; Mazor and Bosch, 1992). 4 He accumulates in groundwater traveling in the
aquifer due to α decay of radioactive elements like uranium. However, an important requirement for a robust age calculation is a well-known value for the Helium flux from the aquifer
and neighbouring layers to the groundwater body.
This idea can be expanded to the age determination of speleothems. If all the helium has
stayed inside the calcite and the inclusions, the 4 He built up by radioactive decay can be
used as an internal clock. In the case of minerals this method is already used (Lippolt et al.,
1994; Wernicke and Lippolt, 1994). Some attempts for speleothem dating are discussed in
chapter 4.5.
The combination of 3 He and tritium (3 H) can be used to date young groundwater (Szabo
and Rice, 1996; Solomon et al., 1992; Schlosser et al., 1988) and to trace mixing in the ocean
(Andrié et al., 1988). The naturally occurring tritium is formed in the atmosphere by a
nuclear reaction of nitrogen atoms induced by cosmic rays. The tritium decays through βdecay into 3 He with a half-life of 12.32 years (Lucas and Unterweger, 2000). If both isotopes,
3 He and tritium are used, an age determination without knowing the tritium-input function
is possible. After a certain residence time t the tritium content is:
3
H(t) = 3 H(0) · e−λt
(2.10)
The amount of 3 He formed during this time t due to the decay of tritum is:
3
He(t) = 3 H(0) − 3 H(t) = 3 H(0)(1 − e−λt )
(2.11)
Therefore it is sufficient to measure 3 He and tritium at one certain point of time to determine
the residence time of e.g. water in the aquifer. A complication are other sources of 3 He in
the water sample. A water sample contains an excess-air component 3 Heex , an equilibrium
component 3 Heeq , a possible terrigenic component (reaction of Li with neutrons) and the
tritiogenic component (Aeschbach-Hertig, 1994). The equilibrium and the excess-air component can be estimated by measuring all noble gases. The terrigenic component can be
deduced from 4 He, which has also a radiogenic component. The tritiogenic helium 3 Hetrit is
calculated from the total 3 He (3 Hetot ) in the following way:
3
Hetrit = 3 Hetot − (3 Heeq + Rex · Lex · Neex + 4 Herad · Rrad )
(2.12)
Rex means the ratio of ecxess-air 3 He to the corresponding 4 He value. Rrad is the radiogenic
3 He-4 He ratio. L
4
ex describes the He/Ne ratio of excess air. Thus the penultimate term in
the equation describes the excess-air and the last term the terrigenic component.
With regard to caves and fluid inclusions in speleothems, it is not known a priori how long it
takes the water to percolate through the vadose zone down to the stalagmites. In some cases,
few weeks or less are necessary for the water to move from the surface to the dripping site
due to fissures and fractures. However, in deep caves and caves with quasi-saturated layers
above, considerable delay can be present. Therefore the dating via 3 H-3 He has been applied
to dripwater in an exemplary site and is discussed in chapter 4.1.2.
2.3.4
Limitations
The main objective of this thesis is to investigate the possibility of using noble gases in
speleothems as a new climate proxy. Therefore, it is suggestive to compare the available data,
its quality and the limitations in case of speleothems with the information that can be derived
from groundwater, where noble gases are originally used to determine palaeotemperatures.
24
Chapter 2. Theory and basics
The groundwater constitutes a reliable and unique palaeoclimate archive, but is nevertheless
limited because of signal dispersion, mixing between aquifers and dating constraints.
If there has been a sudden temperature change with a duration of only some 100 years this
signal will become smaller and will disappear in some thousand years due to the diffusion
of the noble gases according to the concentration gradients (Stute and Schlosser, 1993). So,
high frequency fluctuations like the 8.2 ka event (Alley et al., 1997; Alley and Ágústsdóttir,
2005) can hardly be seen in the groundwater archive. Nevertheless, substantial changes in
the climate system are very well recorded, for instance the transition from the last glaciation
to the holocene (Stute et al., 1995; Aeschbach-Hertig et al., 2002; Stute and Deák, 1989).
A strong constraint in the use of groundwater as a palaeoclimate archive is dating. Usually
14 C is used to determine the age of the water samples. Unfortunately, the 14 C content is
influenced by various constituents as, e.g. the dissolution of 14 C-active CO2 from the soil
gas, 14 C in the soil layer and dissolution of mostly 14 C-free carbon from rock and aquifer
materials (the so-called dead carbon fraction). In addition to this, chemical reactions of
water flowing through the aquifer may change the 14 C content and can furthermore influence
the 14 C value by fractionation. Determining the different constituents as well as the initial
14 C concentration is difficult, but can be managed by an additional measurement of the
δ 13 C-value (Fontes and Garnier, 1979). In general, dating is restricted to a timespan of
about 40 000 years, as the half life of 14 C is only 5730 years.
Furthermore, mixing of groundwater bodies may result in misleading data. Different proxies can, for instance, show totally different ages which complicates the data interpretation
(Mazor, 1997). With the use of chemical parameters and an intensive isotope study the end
members of mixing can still distinguished, however with rather high uncertainties.
In case of the speleothems, most of these problems are of minor importance or do not exist. Mixing and dispersion does not occur and dating is possible up to 500 000 yrs with a
rather small uncertainty when using the U-Th-disequilibrium method. Thus, noble gas measurements on fluid inclusions of speleothems would provide an ideal tool for palaeoclimate
reconstruction.
2.4
Noble gases from speleothems as a proxy
Now, the advantages of noble gas measurements in liquids, the calculation of absolute temperatures, shall be combined with the precise dating of stalagmites. In this section the idea is
explained in detail and the constraints on a feasible application are discussed. Furthermore,
published studies using noble gases in fluid inclusions are presented and interpreted with
focus on the NGT calculation.
2.4.1
Studies using noble gases in fluid inclusions of speleothems and minerals
Noble gases are of widespread use in geological applications, especially the 3 He/4 He measurements as well as the Ar measurements for dating. However, in the past (before the 1980s)
they have rarely been measured in combination with fluid inclusions.
One of the early studies refers to Kelley et al. (1986), who attempted to apply the 40 Ar-39 Ar
dating method to fluid-inclusion bearing vein quartz. Similarly, Turner and Bannon (1992)
tried to apply this method to fluid inclusions in quartz and fluorite veins. Another early study
was conducted by Stuart and Turner (1992), who investigated noble gases from fluid inclusions
in quartz, fluorite, mantle materials and also minerals that precipitated in equilibrium with
2.4. Noble gases from speleothems as a proxy
25
the atmosphere (a flowstone sample). They did not measure the amount of released water
and thus were not able to derive noble gas concentrations. However, they calculated noble
gas ratios which at least could indicate a temperature range for the equilibration of the
inclusions fluids as well as the contribution of air. They concluded this air contribution to
be an unfractionated addition of atmospheric air.
Böhlke and Irwin (1992a) introduced a system for high resolution studies on fluid inclusions
using tiny sample amounts and a laser microprobe system. Based on microstandards and
synthetic inclusions they demonstrated the possibility for precise and reproducible measurements in the range of these very small samples (down to 10−11 l). The new system was
applied to high-salinity hydrothermal fluid inclusions from the lead-fluorite-barite deposits at
Hansenburg (New Mexico). The combined measurements on noble gases as well as on halogens yielded an equilibration temperature of (20 ± 5)℃ for these samples (Böhlke and Irwin,
1992b). They used the obtained noble gas and halogen data primarily to derive conclusions
about the origin of hydrothermal fluids and dissolved salts.
Burnard et al. (1994) used CO2 , He and Ar isotopes in fluid inclusions of a dunite nodule to
distinguish different inclusion types and to infer some information about the mantle history.
Norman and Musgrave (1994) used N2 , Ar and He to distinguish the source of the inclusion
fluids in different hydrothermal systems. Stuart et al. (1995) also made a study on hydrothermal fluids. They used helium and argon isotopes to constrain mantle and crustal components
in the fluids and to derive information about crust and mantle behaviour of the Pacific plate
in the late Cretaceous. Newman et al. (1996) attempted to use the N2 , Ar, He, CH4 and
CO2 values, gained from fluid inclusions of different calcites, to understand the processes and
environmental conditions which control the calcite precipitation.
Scarsi (2000) performed a detailed study on olivine and phenocrysts focused on the crushing
technique under vacuum. He detected the fluid inclusions to be opened in the first steps and
a second source to be responsible for gases released in further steps. The second source may
be related to the mineral matrix or to secondary inclusions.
Recently, some groups started investigations concentrating on fluid inclusions of speleothems.
Scheidegger (2005) investigated different ways to separate noble gases from air- and waterfilled inclusions based on noble gas ratios: Rice (2004) built an extraction system for speleothems using crushing in a steel cylinder and Träumner (2005) tested several stalagmites with
regard to noble gases from fluid inclusions. All samples showed ratios close to atmospheric
values. All of these first attempts to use noble gases in fluid inclusions as a temperature
proxy lack the water determination which rules out the possibility of meaningful temperature
calculation.
2.4.2
Basic idea of climate reconstruction from noble gases in speleothems
The solubility of noble gases in water is, amongst other effects, dependent on temperature.
Therefore, the noble gas concentrations in water that has not been in contact with the
atmosphere after the close-off, can be used to derive the temperature at time of enclosure, if
other effects like salinity and pressure are known. Fluid inclusions are small residuals of the
water dripping on the speleothems. As they are growing, the fluids are embedded between
grains, inside grains or in the calcite lattice. The cave’s temperature at this time is crucial
for the noble gas concentrations in the trapped fluid parcel volume. After including the fluid
in the calcite a change in noble gas composition is unlikely as the diffusion is extremely small
(s. chapter 2.5). Only the helium concentration can be modified by radioactive decay or by
diffusive loss at high temperatures.
Hence, long-term temperature changes in the environment, for instance during the transition
from a cold period, like the last glaciation to a warmer period, like the Holocene, should be well
26
Chapter 2. Theory and basics
conserved in the noble gases from fluid inclusions. Even changes with a smaller duration (for
example Dansgaard-Oeschger events) should be detectable as diffusion is strongly suppressed
or negligible. So a climate signal remains in the fluid inclusion of different layers.
2.4.3
Objective
The main objective of this work is the calculation of absolute temperatures from speleothems
with the highest possible resolution. Special extraction techniques have been developed so
that even stalagmites with an inappropriately high amount of air-filled inclusions in an untreated case can be used for this purpose. Altogether, palaeotemperatures with an uncertainty
of about 1 ℃ should be achieved. By comparison with other proxies like stable isotopes and
trace elements, information about precipitation and other climate parameters will be derived.
Similarly, the NGT data may be used to disentangle the different effects provoking changes
in the stable isotope values of the calcite.
2.4.4
Constraints
The determination of minute quantities of water from the inclusion fluids may be difficult. In a
simple approach it is also possible to determine temperatures with noble gas ratios only, which
can already be obtained by noble gas measurements without water amount determination.
However, these temperatures yield uncertainties of at least 2 ℃ (Aeschbach-Hertig and Kluge,
2006). This is too large for high precision climate studies and rejects the use of noble gas
ratios only. Thus, we are forced to determine also the according amount of water for the
calculation of noble gas concentrations. The achievable uncertainties for the temperature
calculation based on noble gas concentrations can be derived from Fig. 2.7. This plot was
generated by synthetic data varying the air/water volume ratio as well as the analytical error
originating from the measurement process. The temperature uncertainty was calculated
by the application of the inverse modeling method (Aeschbach-Hertig et al., 1999) on the
synthetic data.
At first, measurement precision has to be better than 3 % to achieve temperature uncertainties
lower than 1 ℃. This requires a clean extraction line and very low or at least reproducible
and stable background values. Secondly, it is necessary to reduce the air/water volume ratio
to below 0.1 in the case of a 2% analytical error. At higher values temperature uncertainty
is too large for useful interpretation. A simple estimation based on e.g. 1‰ error in the
air-water volume ratio A already shows this effect. Assuming an A of 1, we get about
1.83·10−13 ccSTP Ne from 1 µl of water and 1.65·10−11 ccSTP Ne from 1 µl of atmospheric
air. The total signal consists of 1.67·10−11 ccSTP Ne. If we would subtract the air-derived
part to obtain the equilibrium component, a 1 ‰ error in A would result in an uncertainty
of 9 % in the case of the remaining Ne equilibrium fraction. Larger A values provoke even
higher uncertainties and show that this is a crucial value for the aimed precision.
Furthermore, the cave air composition should be atmospheric and not changing with time. If
the noble gas mixing ratio would be different between the caves, representative air samples
would be necessary beyond the speleothem samples. Moreover, in this case it is not sure
whether the noble gas composition is constant over time. However, there are few reasons for
such situations as noble gases (except helium) have no significant sources over the investigated
time scales of some thousand to 100 000 years. To investigate this aspect, cave air samples
have been taken and measured (see chapter 4.1.1: cave air measurements).
Another important point for this technique is the retention of noble gases inside the fluid
inclusions. The temperature signal would be modified if considerable diffusion occured. High
diffusion constants would also influence sampling and the preparation in the laboratory.
2.4. Noble gases from speleothems as a proxy
27
5
5%
3%
Temperature error [˚C]
4
2%
1%
3
2
1
0
1E-3
0.01
0.1
1
10
A = V air / V water
Figure 2.7: Maximum precision achievable at a certain analytical error and various air/water volume
ratios. Fitting parameters are temperature and excess air. The pressure is known from the location of
the speleothem inside the cave, salinity is assumed to be zero and also fractionation is supposed to be
negligible.
Calculations and a discussion are summarized in chapter 2.5 .
In the case of water samples excess air and fractionation play an important role. As waterfilled inclusions sometimes also contain some bubbles, it is possible that fractionation has
to be considered in the calculation. However, considering growth history of speleothems
this can almost be ruled out. The process of embedding liquid drops in calcite takes a long
time (typical growth of a stalagmite is about some tens of µm per year), so that a very
good equilibrium with the surrounding atmosphere is achieved. Additional air bubbles only
possess a composition according to the cave atmosphere. As the pressure inside the fluid
inclusions does not increase essentially, the bubble stays with an unchanged mixing ratio and
no fractionation is supposed to occur. Even if fractionation between water and gas occurs,
no separation takes place. As we extract the total noble gas amount, including the fractions
from bubbles as well as from water, both parts are measured. The total gas amount itself
is only influenced by conditions during growth, in particular the temperature governing the
solubility. Changes after the close-off do no alter the total noble gas amount.
Furthermore, the precision will be limited by the sample size and water content. If we assume
a speleothem with a water content of 1 ‰ wt and negligible air-filled inclusions we will get
1 µl water from 1 g speleothem. If the speleothem has grown at a temperature of 10 ℃
4.64 · 10−11 cc He, 2.02 · 10−10 cc Ne, 3.85 · 10−7 cc Ar, 9.10 · 10−11 cc Kr and 1.32 · 10−11 cc
Xe are contained in this sample. This corresponds to about 1013 atoms in case of 40 Ar and
9.5·107 atoms in the case of 132 Xe. To achieve a signal with an uncertainty of less than 5 %
at least some 107 atoms should be available in a sample for each noble gas isotope. The
minimum number depends on the ionization efficiency and the background for each isotope.
The limiting case is 132 Xe, which is the less frequent noble gas isotope measured with our
mass spectrometer. Thus the sample size is limited to 0.1 g at a water content of 1 ‰ wt.
28
Chapter 2. Theory and basics
However, speleothems with an advantageous air/water volume ratio sometimes only contain
0.05 ‰ wt water. In this case the minimum sample size is 2 g.
Böhlke and Irwin (1992a) showed that it is possible to measure noble gas abundances and
isotopic compositions by laser microprobe noble gas mass spectrometry down to 10−11 l of
inclusion fluids. In fact they did not measure the water amount of the water-filled inclusions
opened by laser. However, they were able to assign a temperature to fluid inclusions samples
from hydrothermal brines with a mean value of 20 ± 5 ℃ using the ratio of 84 Kr to 36 Ar vs
Cl to 36 Ar (Böhlke and Irwin, 1992b).
An important practical problem is the abundance of air-filled or at least partially air-filled inclusions (Kluge and Aeschbach-Hertig, 2007; Scheidegger et al., 2006). They often contribute
much more noble gases than the water and mask noble gas signals from the water-filled inclusions. This will therefore limit the NGT determination to a small number of speleothems
if a simple one-step extraction is used. In the case of our method all the noble gases are measured together and separated afterwards via inverse modelling. As the air-filled inclusions
are supposed to contain only unfractionated atmospheric noble gases (Stuart and Turner,
1992; Böhlke and Irwin, 1992b) it is feasible to model the total gas amount with regard to
the interesting water-referring signal. In contrast to the results of Stuart and Turner and
Böhlke and Irwin we observed fractionation or enrichment patterns in some special samples
(see chapter 4.4). In this case a more sophisticated procedure and data analysis is necessary.
2.5
Diffusion of noble gases in speleothems
Is the calcite tight for noble gases at typical cave temperatures? This is one of the crucial
questions concerning this method. If the diffusion coefficient of the noble gases in the calcite
is too high, not only for helium, but also for heavier noble gases, the method can not be
applied to speleothems, because gases and the temperature signals will be affected severely
in the same way.
Copeland et al. (2007) investigated the diffusion of He in calcite and dolomite. Their results
indicate that the speleothems are closed for He and therefore also for the heavier noble
gases at typical cave temperatures. Furthermore, our own measurements affirm these results.
Some pieces of one growth layer of the stalagmite BU-U(we) (chapter 4.6.3) were measured.
From the mean surplus-He an age was calculated via α decays in the decay chain using the
measured uranium and thorium content. The age obtained by fitting the time parameter to
the measured radiogenic He concentration is about 16 500 yrs. Unfortunately, the uncertainty
is rather high and in the order of some thousand years, as the extraction was not 100 %
efficient in the copper tubes. However, the uranium-thorium dating, which resulted in a time
span from 10 500 to 12 100 yrs, is not very far from this rough estimation (s. chapter 3.7).
To substantiate this result and to prove the diffusion behaviour of the different noble gases,
heating experiments have been performed.
Furthermore, the investigation of diffusion gains importance with regard to background control and reduction. In geological applications the extraction line and the extraction device
including the sample is heated some days up to ≥ 100 ℃ to remove the superficially absorbed
and adsorbed gases (baking overnight at 150-200 ℃: Yamamoto et al., 2004; baking overnight
at 150 ℃: Matsubara et al., 1988 and Podosek et al., 1980). In the case of the calcite it was
not known to which temperature the preheating could be extended without loosing the gases
from inside the calcite and from inside the fluid inclusions. The heating experiments should
also deliver information about this topic.
2.5. Diffusion of noble gases in speleothems
2.5.1
29
Theoretical background
Usually the diffusion coefficient is written according to the Arrhenius law, where T0 is the
reference temperature in Kelvin, D0 the diffusion coefficient at temperature T0 and E the
activation energy:
−E
(2.13)
D = D0 · exp
kT
The diffusive
flux of gases is proportional to the diffusion constant, driven by the concentration
∂C
gradient ∂x and can be written according to Fick’s law in one dimension:
J = −D ·
∂C
∂x
(2.14)
Using the continuity equation we get Fick’s second law:
∂
∂C
∂C
=
D
∂t
∂x
∂x
(2.15)
If the diffusion constant does not depend on the position parameter x, the equation can be
simplified to the following form:
∂2C
∂C
=D 2
(2.16)
∂t
∂x
The one-dimensional solution for an initial pulse with a fixed total tracer amount is given by
C0
x2
C(x, t) = √
· exp −
4Dt
4πDt
(2.17)
C0 refers to the initial concentration at x = 0. For calculating concrete numbers the diffusion
length xd is a helpful parameter. xd corresponds to the variance of the Gaussian distribution
(compare equation 2.17 and Fig. 2.8 ).
x2d = 2 · D · t
(2.18)
If the speleothem sample is pumped inside the vacuum chamber, the boundary conditions
are different. Outside the speleothem the gas concentration is strongly reduced due to the
pumping process and compared to atmospheric values inside air-filled inclusion by about 10
orders of magnitude smaller. Therefore, the value outside can be fixed to 0 for a first order
approximation. The solution of the diffusion equation is in this case given by one branch of
the error function for the concentration inside the speleothem and 0 outside:
C(x, t) =
C0 · erfc
0
x−x1
√
4Dt
for x ≥ x1
for x ≤ x1
(2.19)
C0 is the initial concentration. x1 is the point where the concentration is fixed to 0. The
temporal evolution is displayed in an exemplary case in Fig.2.9 for one side, and the assumption that this part is not influenced by the diffusion from the other side. A detailed
discussion of solutions for the diffusion equation can be found for example in Barrer (1941)
and Crank (1994). A similar problem occurs in the case of ice cores. Noble gas measurements
on ice core samples also require pumping in a vacuum chamber prior to the measurement to
avoid contamination by ambient air. The description of the diffusive behaviour leads in one
dimension also to eq.(2.19). The three-dimensional solution is discussed e.g. by Friedrich
(2003) or Gölzhäuser (2008).
30
Chapter 2. Theory and basics
1.10
C/C o
1.05
1.00
0.95
-8
-6
-4
-2
0
2
4
6
8
x
Figure 2.8: Evolution of an initial pulse (continuous line) in one dimension at three different points in
time (dotted and dashed line). The signal was caused for example by cooling and therefore more noble
gases are stored in the water-filled inclusions. With time these higher values are reduced by diffusion and
the concentrations in nearby points are slightly increased.
1.2
relative concentration
1.0
0.8
0.6
100 a
1 000 a
10 000 a
200 000 a
500 000 a
0.4
0.2
0.0
-2
0
2
4
6
8
x (µm)
Figure 2.9: Diffusion from a calcite cube in one dimension. The concentration outside the calcite is set
to 0. As diffusion coefficient the literature value of 2.4·10−25 m2 /s at 30 ℃ (Copeland et al., 2007) was
used for calculation.
2.6. Adsorption effects
2.5.2
31
Literature values
So far, little systematic research on noble gas diffusion in calcite was performed. Therefore,
only few data is available in literature. For a rough estimation of magnitude, diffusion
coefficients from other minerals can be used. Musset (1968) investigated the diffusion of
argon in feldspar, sanidine, microcline and other minerals. In all cases the extrapolation of
the diffusion coefficient at room temperature led to a value for argon in the order of 10−24 to
10−26 m2 /s. Assuming a mean value of 10−25 m2 /s to be a good estimation for speleothem
calcite, the time for diffusion by 1 µm would be 160 000 years. In 10 000 years the diffusion
would broaden the argon signal by 250 nm. As the annual growth of a stalgmite is in the
order of some µm even an annual signal should be preserved very well in the case of a holocene
stalagmite or a calcite grown during the last glaciation.
A recent publication confirms this data. Copeland et al. (2007) investigated the diffusion
of helium in calcite. Step heating experiments led to a closure temperature of 70 ± 10℃
at a cooling rate of 10℃/Myr. They get a weighted average of the activation energy for
the treated samples of 29.3 ± 0.7 kcal/mol. Based on the presented data, a mean value for
the diffusion constant D of 2.4·10−25 m2 /s at 30℃ respectively 3.5·10−19 m2 /s at 150℃ can
be calculated. At 30℃ this would implicate a helium diffusion of about 1 µm in ≈ 70 000
years. As the cave temperature is normally below this value, even higher timescales for the
same diffusion length can be assumed. In consideration of the mean annual growth rates of
stalagmites, the diffusive loss of significant multi-annual events becomes even more unlikely.
With regard to the preparation of the samples at e.g. 150℃ the timescale for diffusion of
1 µm is about 400 hours. About 50 % of the speleothem sample stays larger than 200 µm
after being crushed 60 times with the steel ball. Therefore, the diffusive loss of noble gases
during pumping of the samples, prior to the measurements, can be neglected as additionally
the pumping time normally does not exceed 24 hours.
Heating measurements on a cube of the H12-stalagmite revealed no significant diffusive loss or
fractionation after a pumping period of more than 12 days with temperatures of up to 250℃
(s. chapter 2.3.5). A subsequent crushing step of the heated and for a long time pumped
sample yielded a well measurable water content as well as Ne and Ar ratios (9.92 ± 0.08,
295 ± 3) close to atmospheric air and noble gas concentrations in correspondence with airequilibrated water. Thus the diffusive loss of noble gases during pumping and mild heating
can be neglected.
Helium is the lightest noble gas with the highest diffusion constant, the heavier noble gases
are better retained in the calcite with respect to diffusion. Therefore it can be stated recapitulatory that a diffusive loss of noble gases is unlikely to occur in the cave or after sampling
and that pumping with mild heating ≤ 150 ℃ does only marginally affect the noble gases.
2.6
Adsorption effects
Gas molecules can be ”bound” to solid surfaces by adsorption. Adsorption consists of two
different types, phyisical adsorption by Van-der-Waals forces and chemical adsorption by
orbital overlap or charge transfer. The physical adsorption is a long-range bonding, but
relatively weak with a bond energy of less than 50 kJ/mol. The chemical adsorption has
a bond energy of 50 kJ/mol up to 500 kJ/mol, but is only acting on short distances. The
chemisorption is surface specific, whereas the physisorption takes place on any surfaces at
adequate low temperatures. Physisorption can generate multi-layers of adsorbed atoms, which
can be described by the BET-isotherm. In case of chemisorption mono-layers are expected,
which can be described by the Langmuir isotherm in the adsorption equilibrium state.
32
Chapter 2. Theory and basics
5
BET isotherm
Langmuir isotherm
Freundlich isotherm
coverage v/v mono
4
3
2
1
0
0.0
0.2
0.4
0.6
0.8
1.0
P (atm)
Figure 2.10: Freundlich, Langmuir and BET isothermes. The isotherms describe how well certain
molecules or atoms are bond to the adsorbing solid. The isotherm displays the surface concentration
of the adsorbed material in dependence of the equilibrium concentration (liquid), respectively the gas
pressure or the volume ratio.
The isotherm is a function which connects the number of adsorbed atoms or molecules on
the adsorbent with the pressure in the gas phase. A first empirical approach has been made
by Küster (1894) and Freundlich (1907):
1
x
=k·Pn
m
(2.20)
x describes the adsorbed quantity, m the adsorbent quantity, P refers to the pressure. k and
n are empirical constants for a given temperature.
Langmuir expanded this approach to a larger range including high pressures and low temperatures (Langmuir, 1916, 1918):
αP
(2.21)
Θ=
1 + αP
Θ describes the percent coverage of the surface. α is the Langmuir adsorption constant,
which increases with the strength of adsorption and decreases with temperature. P is the
gas pressure.
The Langmuir isotherm is restricted to the case of monolayers. Uniform surface, no interaction between the adsorbed molecules and the same mechanism for adsorption of all molecules
and atoms are assumed. For the more general case Brunauer et al. (1938) developed a modified formula (BET-isotherm):
1
x(C − 1)
x
=
+
v(1 − x)
vmono C
vmono C
(2.22)
where x is ratio of the equilibrium to the saturation pressure of the adsorbate at a given
temperature. v is the adsorbate volume at standard conditions (STP), C is an empirical
constant at the given temperature. vmono is the volume in STP necessary to form a monolayer.
The BET-isotherm describes the physisorption, whereas the Langmuir formula is better suited
for the chemisorption.
2.6. Adsorption effects
33
von Antropoff (1954) investigated the adsorption of N2 and Ar at different pressures and
temperatures on charcoal. He found a strong dependence on the concentration in the gas
phase. The adsorbed Ar amount is about 3 orders of magnitude smaller at room temperature compared to -76 ℃, whereas the effect of increased temperature gets smaller at higher
temperature values. The adsorption of noble gases was also investigated by Gvozdev and
Tovbin (1997). They found a strong decrease of the adsorbed gas at a decreasing pressure in
agreement with the behaviour of the Freundlich, respectively the Langmuir isotherm at low
pressures. Munakata et al. (2003) measured the adsorption of Kr and Xe on different materials (charcoal, molecular sieves, mordenite) with the objective to remove radioactive noble
gases from the exhaust air of nuclear power plants and reprocessing units by adsorption. The
adsorbed amount is strongly dependent on the adsorbent material and shows differences up
to two orders of magnitude. Ag mordenite yielded the highest ability for adsorption and a
special molecular sieve the lowest ability amongst the investigated materials. At a Xe partial pressure of 10 Pa 10−6 mol Xe per g of adsorbate are adsorbed on the molecular sieve,
whereas it is about 10−4 mol Xe in the case of the Ag mordenite. For Kr the values are
between 10−6 and 10−7 mol Kr per g in the case of this two materials. These values indicate
that a considerable amount of noble gases can be adsorbed on certain materials. Typically
the pressure in the preparation and extraction line is considerably lower (8-9 orders of magnitude) and therefore a smaller contribution from adsorbed noble gases, released from the walls
during extraction, can be expected. However, regarding calcite growth, a gas addition with
preferential adsorption of heavy noble gases may occur in correspondance with the higher
ability of certain materials to adsorb the heavy gases.
For very low range noble gas measurements (total sample gas amount: Ar ≤ 10−6 ccSTP,
Xe ≤ 10−10 ccSTP) a clean line and an adequate sample preparation are necessary. As the
signals are very low, even adsorption of atmospheric gases on the surfaces of the line, the
crushing device and the sample after venting of the system play a role.
Measurements have revealed a long pumping time for the heavier noble gases and show
substantial values of adsorbed Xe remaining before thermal decrepitation (s. Marx, 2008).
Thus, it can be assumed that Xe has a higher coefficient of adsorption on solid surfaces
compared to the light noble gases. A similar conclusion can be extracted from the studies of
Podosek et al. (1980) and Matsubara et al. (1988). They revealed a progressive enrichment of
the heavier noble gases and especially Xe in sediments. Torgersen et al. (2002) detected noble
gases to physi-adsorb rapidly on natural rocks. They also suggest a quick modification of this
weak physisorption into a strong chemisorption mode. This finding is in good agreement with
our laboratory experience, as we need temperatures remarkably above room temperature to
release the adsorbed component efficiently. E.g. at 120℃ a peak was found for all noble gases
likely caused by desorption from the copper tube walls (Fig. 3.38). The physisorption yields
a bond-energy of up to 50 kJ/mol which corresponds to about 0.5 eV per atom or molecule.
For chemisorption the bond energy per atom or molecule is even up to 5 eV. The thermal
energy is about 0.025 eV at 20 ℃ and therefore too low to release the adsorbed components.
The sorption isotherms show a dependence of the adsorbed component on the partial pressure,
respectively the concentration. Thus, the two possibilities - low partial pressure due to
pumping and high temperatures through heating - are combined for reducing the adsorption.
In order to remove adsorbed noble gases from the crusher surface, it has been heated at least
overnight (minimum 8 hours) at temperatures between 70 ℃ in the case of the simple steel
cylinder and at 150 ℃ with the new crusher system including the external sample pocket in
a modified valve. The pumping procedure is orientated to the preparation process used in
geology. E.g., Matsubara et al. (1988) baked the samples at 150 ℃ overnight.
34
Chapter 2. Theory and basics
With regard to the noble gases in the sample itself, the adsorption on the calcite surface
during the speleothem growth may play a role for the final noble gas concentration. As the
calcite surface is typically not absolutely plane (s. Fig.2.6) a large surface for adsorption
exists. In the case of samples with a very small number of water- or air-filled inclusions the
adsorbed noble gas components may alter the noble gas concentrations towards higher values,
indicating apparent cooler temperatures like in the case of the CG-stalagmite (water content:
≈ 0.03 wt‰, low gas amount, fitted temperature about 5 to 10 ℃ below expected values)
or the two Spannangel samples (SPA 12, water content: 0.08 wt‰; SPA 52, water content:
0.1 wt‰, fitted temperature in both cases below 0℃). This effect is amplified by the smaller
mobility of the heavy noble gases Kr and Xe, which are more likely adsorbed and less easily
removed from the calcite surface. Thus, the water content of the speleothem samples as well
as the total gas amount can indicate the influence of adsorption. If there is a large number
of inclusions, a possibly adsorbed component will be of minor importance.
To get an impression of theoretically adsorbed molecules or atoms on a typical surface we
made a simple estimation. On a surface of 5 mm x 5 mm, which corresponds to typical
dimensions of the measured pieces, about 5.4·107 Xe atoms could be placed in one monolayer
assuming a mean atomic radius of 1 Å, atmospheric air and typical mixing ratios. This is a
rather large number compared to the Xe content in an inclusion-poor stalagmite. In 0.1 µl of
water, equilibrated at 10℃, about 3.5·107 Xe atoms are abundant. Actually, polar molecules
like H2 O will be adsorbed preferentially and hamper the adsorption of other atoms and
gases. In detailed experiments Marx (2008) found crushed calcite to adsorb water vapour very
efficiently. The too low NGTs obtained from the CG as well as the Spannagel samples may at
least partially be explained by an underestimated water amount due to the adsorption of water
on the crushed calcite. Underestimated water amounts or an enhanced noble gas content lead
in both cases to higher noble gas concentrations indicating apparent low temperatures. In
most samples we found no indication of increased noble gas content for the heavier isotopes.
This may be an important hint that adsorption of noble gases during growth plays only a
minor role, at least for the majority of the samples.
2.7. Goals of this study
2.7
35
Goals of this study
The main objective of this study is to establish the determination of noble gas temperatures
from speleothem fluid inclusions.
To achieve this goal different points have to be considered and examined accurately:
• sample selection: for reconstruction of the palaeoclimate only well datable samples
are useful. Speleothem types with adequate properties for noble gas measurement and
also dating purposes have to be figured out. A procedure for a simple sample selection
has to be developed.
• sample preparation: to achieve maximum time resolution, samples have to be sufficiently thin, but with regard to the extraction line and the crusher they should possess
certain dimensions, e.g. a not too big cross sectional area. An optimum sample preparation should be figured out.
• sample treatment prior to extraction: this is one of the most critical points as the
results of the measurement can strongly depend on the preparation and pumping, which
influences the blanks in the same way. A procedure has to be developed, which reduces
the blank by pumping of the sample and the whole extraction line in high vacuum at
either moderate heating or at room temperature.
• development of a noble gas extraction line for speleothems. The extraction line should
be built with regard to the lowest blanks possible and a convenient line control via pressure gauges to monitor the extraction process or detect possible leaks. The line should
possess connections which provide the possibility for a high sample throughput and
special line parts for water determination including cold fingers, pressure gauges and
expansion volumes. For maintaining a high vacuum (≈ 10−8 mbar), an adequate pumping line including turbo-molecular and scroll pumps for cleaning should be available.
• development of a system for water determination in the low µl-range. Speleothem
samples rarely contain more than 1 µl of water in 1 g of carbonate. For NGT calculation
these low water amounts have to be determined with sufficient precision of better than
3 % to achieve an overall uncertainty in the range of 3 %.
• development of an appropriate extraction method. In general all of the noble gases from
opened water-filled inclusions should be released. Additionally an extraction procedure
to separate the noble gases from the air-filled part should be developed. The extraction
has to yield a low air/water volume ratio. As a constraint for the procedure, the
extracted water amount should normally be above 0.1 µl, so that the noble gases can
be measured with sufficient precision.
• development of an adequate noble gas preparation procedure. The separation of heavy
and light noble gases as well as the purification process should be designed in such
way that the largest part of the released noble gases are being transferred to the mass
spectrometer. By this way very small samples can be measured also.
• development of a special speleothem measurement procedure. Speleothem samples
can not be measured like water samples as they deliver about 3 orders of magnitude less
noble gases. In some cases the ion multiplier has to be used instead of the faraday cup
for measurements in the mass spectrometer. Furthermore, the ion source tuning has
to be changed to achieve maximum efficiency and the actual measurement procedure
has to be adapted concerning peak centering, magnet precycling and ion source tuning
changes to achieve a good signal stability and mass resolution.
36
Chapter 2. Theory and basics
• development of a measurement procedure for diluted calibration standards. For calculation of absolute noble gas amounts, measured signals from speleothems have to
be compared with a known standard. A diluted standard has to be prepared which
yields gas amounts comparable to the samples. Furthermore, scripts for the automatic
calibration measurement have to be established. Using automation it is possible to run
the measurements 24 hours and 7 days a week.
• development of a sophisticated background control. As the sample signal is rather
low, blank contributions are of critical importance. The relevant background fraction
has to be figured out and scripts for automatic measurements have to be established.
Furthermore, scripts for contribution control of double ionized atoms to the measured
signals have to be written. This is of importance especially in case of 20 Ne due to the
possible double ionisation of 40 Ar.
• development of an appropriate data evaluation. In one measurement run several hundred of single noble gas measurements are performed. Data evaluation by hand would
take some weeks and would be quite error-prone. Instead, an automatic evaluation
routine would be less time consuming and in general less error charged.
• optimization of calibration. The calibration measurements should show reproducible
values, low scattering, as little as possible non-linear behaviour with regard to gas
amounts and small uncertainties for all noble gases (at best below 1 %).
• test of the obtained noble gas data - reasonable results? At first, tests samples
from a certain site and from known climatic periods should deliver reasonable values.
Holocene samples, for instance from a cave with a present-day cave temperature of
about 10 ℃ should at least give NGTs between 7 ℃ and 13 ℃. Permanent shifts
above or below this range would indicate systematic problems either in extraction and
measurement or in data evaluation.
• test of data reproducibility: to establish this new method for palaeoclimate reconstruction reliable data have to be produced. A speleothem sample, which could be
measured with low overall uncertainties (including sample, line and preparation blanks,
calibration and water determination uncertainties) in the range of few percent and an
acceptable air-water volume ratio below 0.1 should reproduce within 1 ℃.
• comparison with stable isotopes. Assuming all the preceding points had been achieved
and reliable as well as precise data can be derived, measurements along a stalagmite
growth axis can be performed. In this case the NGTs should be compared to the stable
isotope data. As the temperature is known from the noble gas concentrations, it should
be attempted to infer information about precipitation or soil cover changes from the
temperature corrected isotope data.
• development of a measurement procedure for routine work. In the final phase a
larger number of samples should be measured in a short time. To reach this goal an
adequate procedure has to be developed. According to the sample property a typical
extraction method has to be used and the extraction line has to be prepared for a high
sample throughput. This can be achieved with a set of three or more crushers, as for
example one is pumped via turbo-pump connection, one used for sample crushing and
the third is in the measurement stage at the same time.
Chapter 3
Working with speleothems
In this chapter the basics about deriving noble gas data from speleothems will be presented.
At first, the sample selection will be discussed. This is of special interest as the noble
gas composition, especially the fraction from air-filled, respectively water-filled inclusions is
dependent on the type and structure of the stalagmite. Investigation of thin sections play a
major role with regard to this task.
Noble gas temperatures with low uncertainties can only be obtained from noble gas concentrations. Therefore, it is necessary to determine the water amount which corresponds to the
measured gas signal. In the following section a method is presented and discussed of how to
determine the tiny water amounts with sufficient precision. Additionally, different extraction
techniques have been investigated with regard to the efficiency and the gas background. High
air/water volume ratios complicate the precise temperature determination. Using stepwise
procedures may yield better conditions and is discussed briefly. Finally, the subtle noble gas
measurements with the mass spectrometer and the sophisticated data analysis is presented.
3.1
Sample selection
An ideal sample for the calculation of noble gas temperatures has to accomplish several constraints. Primarily, the ratio of noble gases from air-filled inclusions to the noble gases from
water-filled inclusions has to be sufficiently small (air/water volume ratio ≤ 0.3). Secondly,
the sample should provide enough water and noble gases in order to be measured with the
necessary precision. The total uncertainty including water determination and noble gas signal
uncertainty should be smaller than 3 %. In this section the properties of the measured samples are discussed using microscopic analysis of thin sections. Based on this data a conclusion
can be drawn which samples can be assumed to be useful for NGT measurements with the
above described properties.
3.1.1
Optical methods
A good impression of the speleothem structure can be obtained from thin sections. They offer
the possibility to connect the noble gas signals and other properties like the air/water-volume
ratio and the water content to the structure. Furthermore, they give insight into distribution,
size and type of inclusions.
The preparation was successfully realized in the laboratories of the Institute of Mineralogy.
Several steps have to be taken in order to get a good thin section. Small parts (≤ 3 cm x
5 cm x 0.3 cm) are cut of the speleothem, embedded to a grinding plate with cold-embedding
material like Epofix or Caldofix and polished with SiC-paper in a way that the sides are totally
plane without any grooves. Subsequently the speleothem piece is attached to a microscope
38
Chapter 3. Working with speleothems
BC
O
A
SP
BU
BI
G
C
H
MA
Figure 3.1: Investigated speleothems are from different parts of the Earth. Caves, from which samples
have been taken are marked with a flag. BC Bear Cave, BU Bunker Cave, CG Dos Anas Cave, H Hoti
Cave, MA Marcelo Arevalo Cave, OBI Obir Cave, SPA Spannagel Cave.
slide, grinded roughly down to 500 - 100 µm and lapped afterwards with SiC powder with
grain size F150, F400 or F600 to a thickness of about 35 µm. To achieve a high reflecting
surface and to remove deformations caused by the fine-grinding step, the sample is diamondor oxide-polished.
Afterwards, the microscopic analysis was done in the Institute of Geology and Paleontology
with support of Dr. Glasmacher using a special automated transmitting- or reflecting-light
microscope.
Inclusions may be larger than 30 µm and can be destroyed during the thin-section preparation.
Therefore, usually thick sections are prepared to avoid the damage of larger inclusions and
the loss of a three-dimensional perspective. For sedimentary rocks a thickness of 40 to 60 µm
is sufficient (Goldstein and Reynolds, 1994). In the case of speleothems up to 150 µm may be
appropriate. However, all the data presented here refers to thin-sections with a thickness of
35 µm. The smaller inclusions (<20 µm) are assumed not to be affected by the preparation
and thus the thin-section can give a rough overview of the inclusion distribution and its
general dimensions.
A set of samples with significantly different noble gas signals has been investigated using
microscopic inspection of thin sections. The samples were taken from caves with different
climatic conditions (Fig.3.1):
• OBI5, stalagmite from the Obir Cave in Austria (Spoetl et al., 2005). The cave is
situated at about 1100 m above sea level with a mean annual air temperature of about
6 ℃.
3.1. Sample selection
39
200µm
200µm
20µm
Figure 3.2: Thin sections of stalagmites H12 and OBI5. The upper pictures are displayed with the
same magnification. The upper left picture is a thin section of H12, the right of OBI5. The scale bar is in
both cases 200µm. The lower picture illustrates the fluid inclusions of H12 with a higher magnification the scale bar corresponds to 20 µm.
• H12, stalagmite from the Hoti Cave in Oman (Neff, 2001). The cave is located 950 m
above sea level. In this cave temperature ranges between 23 ℃ and 26 ℃. The investigated stalagmite H12 reflects arid to semiarid climatological conditions of the Holocene
and is younger than 6000 years. The about 80 cm long H12 stalagmite shows layering
and in general a grey apperance with highly porous zones and denser milky white parts.
• flowstone H8Z, extremely pronounced layered speleothem from the Hoti Cave in Oman.
Small brownish up to black layers are alternating with thicker as well as brighter layers.
• CG, 72 cm tall whitish stalagmite from Cuba from the Dos Anas cave in the MajaguasCantera cave system (Pajon et al., 2006). This sample has grown at 90 - 100 m above sea
level in a tropical environment with typical precipitation values of 1600 - 1800 mm/yr
and a mean cave air temperature between 21 and 22 ℃. The cave chamber from which
the sample was taken is at 1.2 - 1.3 km distance from the entrance area.
• BU-U, stalagmite from the Bunker Cave in Sauerland (NW Germany). The cave is
located 180 m above sea level. The recent mean annual air temperature is about
9.5 ℃ (1960-1990). The region is characterized by a temperate climate with regular
precipitation. The investigated thin section from the about 10 cm long stalagmite
originate from a milky white Early Holocene growth layer.
40
Chapter 3. Working with speleothems
1mm
100 µm
200 µm
20 µm
Figure 3.3: Thin sections of the flowstone H8Z from Hoti cave (Oman) using different magnifications.
The thin-section of stalagmite H12 (Fig. 3.2) shows a microsparitic fabric without a clear caxis orientation under polarized light. Furthermore, the thin section is characterized by many
small and also some large inclusions. They are distributed with a certain pattern over the
whole thin-section, which may reflect seasonality as found e.g. by Genty and Quinif (1996).
Two denser lines of fluid inclusions are visible on the right side. On the left side grains and
large inclusions can be seen. The large inclusions with irregular forms are air-filled. Waterfilled inclusions are rarely larger than 20 µm (s. Fig. 3.2, lower picture). Due to the huge
amount of inclusions this speleothem has a a considerable noble gas signal. However, the
large air-filled inclusions contribute overwhelmingly to the signal. The change in structure
and also the alternation of zones with a higher fluid inclusion density with areas exhibiting
few inclusions, may be due to the frequent and strong changes in precipitation and growth
conditions in the Hoti Cave, which was marginally influenced by the Asian monsoon (Neff
et al., 2001; Burns et al., 2001).
This speculation can be verified by the following samples from the Obir Cave and also from
Bunker cave. OBI5 from Obir Cave (right side of Fig. 3.2) shows large columnar crystals,
which are mostly orientated in the same direction. In contrast to H12, we found randomly
distributed inclusions which were significantly smaller than in the case of H12. Even using
a higher magnification, no grains are visible, but a layered structure became visible. As the
dimension and amount of inclusions is obviously smaller, the noble gas signal is also strongly
reduced compared to H12. However, the ratio of noble gases from the air-filled inclusions to
noble gases from water-filled inclusions is even worse. The uniform distribution of inclusions
in the shown thin section may be related to more constant growth conditions as the cave is
situated in the west-wind zone, where precipitation is provided over the whole year.
The flowstones are formed by a water flow over substrate. As there is normally a constant
water film above the speleothem, the growth conditions are rather constant. Therefore, a
similar appearance as in case of OBI5 should be expected. However, Figure 3.3 shows a
3.1. Sample selection
41
200µm
20 µm
Figure 3.4: Thin sections of the stalagmite CG from Cuba.
100 µm
Figure 3.5: BU-U thin section. Most of the inclusions are very small, round and randomly distributed
over the displayed photo.
pronounced layered structure with large grains and a high amount of inclusions in certain
areas. Similarly, the distinct layers can be seen with polarized light. Bands with dendritic
fabric alternate with layers of columnar crystals. The flowstone originates from the same cave
system as the stalagmite H12 and is similarly influenced by the monsoon. Dry periods deliver
less water and can therefore cause a change in the saturation state of dripwater and influence
the calcite development. Otherwise, a very fast dripping site with low or no supersaturation
may provoke a stop in growth or even calcite dissolution. These changes in growth conditions
generate also different amounts of inclusions. In large grains few inclusions are visible whereas
at the layer boundary and between small grains a huge number can be detected.
The age-model of the stalagmite from Cuba shows that it was a fast and constantly growing
speleothem (0.5 mm per year). As in the case of OBI5, the constant conditions led to
randomly distributed inclusions and furthermore to a hardly layered structure without visible
smaller grains (≤ 0.5 mm). Similarly to the thin-section OBI5 we can see large crystals, but
with an erratic shape and irregularly interspersed small crystals. In the case of the highest
42
Chapter 3. Working with speleothems
Figure 3.6: BU-U thin section on a closer view. Most of the inclusions are water-filled. They have a
round form and appear white in the translucent light image (see places marked by arrows), if the distance
of the lenses to the object is adequate. This can be seen in the right photo. The left photo shows the
same detail, but with different distance object-lenses. Scale bar 20 µm.
magnification only few inclusions, which are at least partially water-filled (light reflecting
points), and no purely air-filled inclusion (should be dark) are visible (Fig. 3.4). Furthermore,
the inclusions are extremely small ( 1µm) compared to the inclusions shown in the other
thin sections. The noble gas signal and fortunately also the air/water volume ratio was
accordingly small. Thus, most of the noble gases originate from the water-filled inclusions.
With regard to the Bunker Cave, a very low air/water volume ratio A was measured in
stalagmite BU-U. A mean value, calculated from noble gas fitting, was about 0.035. In
Figure 3.5 a thin-section is taken through the microscope at a low magnification. There, a
large number of generally very small inclusions can be seen. Larger inclusions may have been
destroyed during the thin-section preparation. A closer look on the photograph reveals a
structure which reminds of a gravel deposition in river valley. In general, very large columnar
crystals (from some mm up to several cm) with a uniform orientation are prevailing.
In another thin-section of BU-U (Fig. 3.6), recorded at higher magnification, the size of the
water-filled inclusions can be estimated. Compared to the scale bar with a total length of
20 µm they, at best, have a diameter of 1 µm. Another interesting point is the change in
behaviour if the focus setting is varied a bit. If the inclusion is totally air-filled, no reflection
can be expected; whereas the water-filled inclusions can show a white glow due to the optical
behaviour of a water drop. Varying the focus settings, the water-filled inclusion on the shown
pictures can be detected (indicated by arrows). Comparing the two pictures of Fig. 3.6
only the two largest inclusions (marked by circles) do not show reflections and are therefore
assumed to be only filled with air. Recapitulatory, few air-filled inclusions with moderate
diameters (≤ 10 µm) and a large number of water-filled inclusions with a small diameter
(≤ 1 µm) have been found.
3.1.2
Summary of the thin-section analysis
Samples with a high amount of inclusions often show a grain-type feature as e.g. the Oman
samples. In contrast to these samples, the milky white speleothems provide a smaller total
amount of inclusions and also less large (air-filled) inclusions. An example is the BU-U
stalagmite or the CG speleothem. Both show a milky white appearance with large columnar
crystals and are similar in the inclusion distribution. Very few large inclusions can be found,
3.2. Water determination
43
whereas the mostly water-filled and in general tiny inclusions (≤ 1 µm) are distributed
randomly. A milky colour can mostly be related to an advantageous air/water volume ratio,
as e.g. demonstrated by the CG and the BU-U stalagmite. In contrast, the brownish or grey
appearance of stalagmite H12 or the samples from Chile (MA), is associated with a huge
number of inclusions, which are dominated by a high contribution of extremely large air-filled
inclusions. Interestingly, both H12 and MA are affected by monsoon or high monsoon-like
precipitation.
Another interesting parameter is the water content, which is determined through crushing. If
a structure with large columnar crystals is prevailing, then the total water amount decreases
in the way the speleothem gets more translucent. A very translucent sinter-piece with large
crystals yields e.g. a low water content, whereas the more milky soda straw yields a water
content about one order of magnitude higher.
Noble gas studies should focus on inclusion-rich speleothems which are dominated by waterfilled inclusions. They can be selected primarily by their colour, which should be in general
milky white and not too translucent, and additionally by thin sections, which should show
few large air-filled inclusions and a great number of small (water-filled) inclusions.
3.2
Water determination
Temperature calculation with uncertainties <2 ℃ is not possible using only the noble gas
ratio. The only way towards precise palaeotemperatures is given by the noble gas concentrations. To determine these concentrations it is necessary to know the water amount from which
the noble gases originate. Moreover, an interesting issue is the survey of special techniques
like stepwise heating and stepwise crushing in consideration of therein opened water-filled
inclusions. Similarly, the released water is an important parameter for the different methods
and can give information about the extraction efficiencies.
Dennis et al. (2001) and McDermott et al. (2006) say, that water amounts are typically in the
range of 0.6 - 1.9 ‰wt, but in this study we found out to vary them considerably more (0.001
to 4 ‰wt). Therefore a precise measurement down to the submicrolitre range is necessary
for speleothem samples with a typical weight of about 1 gram. To solve this problem, two
methods have been developed. First, a relatively precise balance can be used to determine
the amount of extracted water, and second, the relationship between water vapour pressure
and water amount, derived from a special calibration curve is useful. In the microlitre range a
measurement with an uncertainty of 2 - 3% can be achieved by both methods. The objective
was to reduce the uncertainty to less than 1% in the microlitre range (≈ mg), to 1 - 2 % at
0.1 µl (100 µg) and in maximum to 10 % in the range from 0.1 to 0.01 µl.
To determine the released water from speleothems two different methods have been used.
One with acceptable precision is weighing of water in the mg range. The disadvantage is
the high time consumption (at least 1 hour per sample). A practicable method, especially
for small samples, is the water determination by pressure. There are only 10 - 15 minutes
necessary for one sample.
Beyond these two methods, the absorption of light may be used to determine the water
amount. Perhaps this method can be used in the range of extremely low water amounts,
because the laser can be tuned to a special H2 O-absorption line. However, this method still
awaits its testing. FTIR spectroscopy has just been used to control the presence of liquid
water in crushed stalagmite samples (McDermott et al., 2006), but not with the objective to
determine water amounts.
44
Chapter 3. Working with speleothems
Figure 3.7: High precision balance (Mettler Toledo) with a precision of 2 µg. The sample is put on the
small roundish dish inside a chamber which is protected from air motion by sliding windows.
3.2.1
Water determination by weighing
The weighing tests have been performed using samples prepared at the extraction line of Rice
(2004). To test the efficiency and the effects of different handling, artificial samples have been
built. Glass-capillaries (Hirschmann Laborgeräte, 12.5 cm length, 0.3 mm inner diameter,
0.03 µl reading accuracy, see Fig. 3.8) have been filled with tap water. The water is sucked
into the glass tube by the capillary effect or the use of a pipette. Afterwards the two ends of
the glass tubes are flame-sealed. Then the filled capillaries are opened in a copper tube by
crushing under high vacuum conditions (before crushing ≤ 1·10−7 mbar). The released water
is frozen into a copper cold finger (Fig. 3.8) using dry ice. After 30 minutes the cold finger is
closed by squeezing. The closed copper tube part and the rest of the copper tube as well as
the gaskets are weighed using a special balance (Mettler Toledo, maximum weight: 19.99 g,
reading accuracy: 2 µg). For comparison the empty copper tube as well as the gaskets are
measured before the experiment. Moreover a buoyancy-correction is performed. The weight
is reduced by 1000 µg, because of the lower pressure inside the tube. The weight difference
of the empty copper tube and the gaskets to the weight after freezing of the water, including
the corrections, leads to the amount of water captured in the cold finger.
The application of the weight method requires a very attentive handling. One careless movement can influence the results substantially. Several points have been tested in this context
and should be taken into account in the weight calculation (s. Fig. 3.10).
Rubbing the copper tube several times with a paper results in a large reduction of the weight.
In contrast to this, there is no deviation detectable if the copper tube is slightly rubbed once
with a sand paper. One finger-touch increases the total weight because of the adsorbed grease
or other particles. Hence, the sample can only be touched carefully with tweezers. Squeezing
in order to close the water-containing part of the tube leads to a weight reduction in the order
of (105 ± 10) µg. Because this effect can not be avoided, it is included in the correction of the
weight calculation. Carrying the copper tube parts with tweezers is difficult and sometimes
provokes them to fall off to the ground. However, falling tests did not reveal a deviation of
the expected value.
In consideration of all the effects that have to be taken into account (buoyancy correction due
to lower pressure inside the squeezed copper part, mass-changes by different handling, drifts
of the high-precision balance), it is fairly time-consuming to reach the maximum precision.
Additionally, the uncertainty can not be reduced in the necessary order of magnitude as the
correction uncertainty alone is about 50 µg. This would cause a total uncertainty of at least
50 % in the case of 0.1 µl of water, which is by far too high for a meaningful NGT calculation.
3.2. Water determination
45
Figure 3.8: Copper tube freezing finger attached with gaskets. Only the lowermost part of the finger is
cooled by dry ice for 30 minutes to freeze out the water. The copper tube is squeezed directly below the
attachment point by a special pincer to enclose the water. Then the valve is closed and the two copper
tube parts are weighed. Right picture: glass capillary filled with a known water amount for calibration
purposes. Between two black rings 1 µl of water is contained.
100
efficency (%)
80
60
40
20
1. lower value due to leakage of the copper tube
0
0
1
2
3
4
5
test no.
Figure 3.9: Result of five efficiency tests. The water amount measured by weighing is compared to
the water input before the extraction (recovery in % of the input value). The first value is lower due to
incomplete freezing and a leak in the copper tube. The other four measurements show that the water
can be well collected and detected with an overall recovery of (99.0 ± 1.3) % (first value not included).
Therefore we can assume that the water recovery of this method is complete.
46
Chapter 3. Working with speleothems
500
deviation caused by different handling
400
300
200
weight (µg)
100
0
-100
-200
-300
1: rubbing with paper
2: touch with fingers
3: use of sand paper
4: falling test
5: squeezing
-400
-500
-600
1
2
3
4
5
test no.
Figure 3.10: Effects of different handling on the determined weight. The deviations are given in µg with
statistical uncertainties resulting from the weight determination (no.1 - no.4) respectively from repeated
measurements (no.5).
3.2.2
Water determination by pressure
Motivated by the large effort applying the weight method, some other ways have been considered to solve the problem of a precise water determination. Due to the rather high precision
which can be achieved with special pressure gauges (Keller Lex1, precision 0.01 % full scale,
range 0 to 3000 mbar and Pfeiffer CMR 263, 2 ‰ accuracy, range 10−3 to 11 mbar, both
are capacitance pressure sensors), we tried to determine the water amount by the pressure of
the water vapour in certain line volumes. Unless the water does not condensate (saturation
water vapour pressure at different temperatures: s. Table 3.1) the pressure can be used to
calculate the corresponding water amount. 1 mg of water corresponds to 1.25 ccSTP of water
vapour, whereas it only contains dissolved atmospheric gases on the order of 10−5 ccSTP.
Thus, the pressure in the different volumes is almost totally produced by the water vapour.
Knowing the volume, the determined pressure can be transferred into a gas volume using
the equation for ideal gases or the Van-der-Waals gas, and finally be transferred into a water
amount. However, we did not follow this way of water determination.
Another possibility is given by the use of calibration curves. These curves are prepared using
known water amounts sealed in glass capillaries. The capillary consists of a fixed total volume,
wherein certain water amounts are filled. The remaining volume contains air. The sealed
glass capillaries are inserted into a copper tube, where they are pumped until a sufficient low
pressure is reached. Then the glass capillaries are opened by squeezing of the surrounding
copper tube. The released water leads to a certain pressure, which is noted manually.
The results of the first attempt using the calibrated pressure curve at the line of Rice (2004)
can be seen in Figure 3.11. For pressure determination a Keller Lex1 gauge has been used.
Until a pressure of about 60 mbars the relation between pressure and water amount remained
quite linear. The condensation point was about 60 mbars, because the line was heated to
3.2. Water determination
47
Table 3.1: Saturation vapour pressure of water respectively ice at different temperatures.
temperature (o C)
pice (mbar)
pwater (mbar)
-100
-78
0
20
21
22
23
24
36
37
40
60
100
−5
3.6·10−5
1.62·10−3
6.1121
23.39
24.87
26.44
28.10
29.85
59.5
62.8
73.8
199.3
1013.25
1.4 ·10
7.5·10−4
6.1115
-
50
45
40
pressure (mbar)
35
30
25
first test 6.4.06
second test 10.4.06
Linear Fit of Data1_C
20
15
10
5
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
water amount (µl)
Figure 3.11: Calibration curve prepared at the line constructed by Rice (2004). Two test series showed
a good linear relationship up to 60 mbar at a temperature of 36 ℃. This calibration curve was used to
determine the water amount in the first crushing and heating experiments
48
Chapter 3. Working with speleothems
about 36 ℃. Further heating to shift the condensation point to higher values is not possible,
because this would damage the electronics of the pressure gauge. A second series of tests
reproduced the results of the first tests with regard to the linearity. A linear fit of the data
led to the following result:
p(v) = 0.49(±0.6) + 11.78(±0.25)v
(3.1)
The pressure p is given in mbar as function of the water amount v in µl. If the fitting error
is taken into account, a stalagmite sample with a total amount of opened water inclusions of
about 4 µl has an uncertainty of 2.4 % according to this calibration-curve.
The extraction line has been rebuilt and different expansion volumes have been attached
with the purpose to achieve higher precision in the case of different sample amounts. In
the final stage, the calibration curves were prepared using three different expansion volumes
to detect very small water amounts in the range of 0.1 µl as well as larger water amounts
up to 1.5 µl with a typical precision of 2 - 3 %. The glass capillaries had been prepared
in the same way as explained above. After squeezing of the capillaries, the water is frozen
to the cold finger for 20 min. Subsequent to line pumping the freezing finger is warmed
up to room temperature, which is held quite constant by air-conditioning to (23 ± 1)℃.
Then the water vapour is expanded to the first volume (detailed description see Fig. 3.15),
where the pressure is determined 5 min after expansion by the capacitance pressure sensor
CMR 263. Subsequently, the water vapour is expanded to the next volume where the pressure
is determined again after 5 min.
The results are displayed in Fig. 3.12 for all three volumes. For speleothem samples all
three volumes are used and an average is calculated from these three measurements. The
uncertainty is calculated using simple error propagation from measurement errors and fitting
uncertainty. Therefore a typical sample with a water amount of 0.5 µl yields an uncertainty
of about 2 %, a water sample of 0.1 µl an uncertainty of 6 %. The error increases to about
10 % if the released water amount is less than 0.06 µl.
3.2.3
Precision and limits
• Weighing method
The relative error of water determination by weighing is in the range of 2 % in case
of 4 µl-samples. As typical stalagmite samples (0.5 - 1 g) provide only 0.1 to 1 µl of
water, the relative error is significantly higher. The uncertainty is dominated by the
uncertainty of the correction terms and, therefore, difficult to reduce. The weight itself
can be determined with a higher precision. The overall absolute uncertainty including
the correction terms can hardly be reduced below 50 µg which results in an uncertainty
of at least 50 % if only 0.1 µl water is released.
• Manometric method
In this case the precision is limited by the uncertainty of the calibration curve and the
precision of the pressure gauge.
The main uncertainty derives from scattering in the calibration curve. One part belongs
to the slope error and the other part to the error of the intersection point at the y-axis.
These errors can be reduced if the number of calibration samples is increased. The
relative slope error is, in the case of the first attempt (Fig. 3.11) 2 %, and in the
final state for the expansion volume 2 about 1 % (Fig. 3.12). Thus, the uncertainty
due to the calibration curve can be reduced to about 1 %. The uncertainty of the
3.2. Water determination
12
49
volume 1
10
8
Equation: y = A + B*x
Weighting: y
No weighting
Chi^2/DoF
= 0.01907
R^2
= 0.99883
A
-0.32153
±0.12242
B
26.62673
±0.45642
6
4
2
water vapour pressure [mbar]
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
12
expansion volume 1
10
8
Equation: y = A + B*x
Weighting: y
No weighting
Chi^2/DoF
= 0.04554
R^2
= 0.99641
A
-0.16674
±0.11926
B
10.29021
±0.20602
6
4
2
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
6
expansion volume 2
5
4
3
2
1
Equation: y = A + B*x
Weighting: y
No weighting
Chi^2/DoF
= 0.0068
R^2
= 0.9984
A
-0.10065
±0.03501
B
3.14394
±0.0336
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
water amount [µl]
Figure 3.12: Water determination at the extraction line using different expansion volumes (increasing
from top to bottom). The calibration curve was derived via artificial standard samples. The volumes
are not heated during the water determination, but kept at room temperature (about 23.0 ℃), which
is stabilized by an air-conditioning system. In each case a linear relation is obvious. The fitting results
are given in the small box at the right side. These calibration curves were used to determine the water
amount of all speleothem samples measured beginning with measurements in December 2007.
50
Chapter 3. Working with speleothems
80
Pressure (mbar)
60
40
20
gas pressure
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
water (µl)
Figure 3.13: The relation between pressure and water amounts stays linear until 55 - 60 mbar, when the
line is heated to about 35 ℃. Then the water condensates at the coldest part of the line and no further
pressure rise can be detected. The method is limited by this effect and therefore adequate expansion
volumes have to be chosen to prevent condensation. In the final state the line is held at about 23 ℃
which corresponds to a saturation vapour pressure of 27.2 mbar.
pressure gauge (0.1 mbar using the Keller Lex1 and 0.005 mbar using the CMR 263
gauge) plays a minor role, if the line volume is adapted for smaller water amounts.
For instance, a sample with 0.1 µl water produces a pressure of about 2.2 mbar in the
case of volume 1 used in Fig. 3.12. Then, the relative error due to the CMR pressure
gauge is 2 ‰. However, measuring smaller water amounts is limited as it is difficult to
prepare standard water amounts smaller than 0.04 µl for calibration . Precision and
calibrated range could potentially be improved by precise volume determination (can
be calibrated to ± 0.3 %) and calculation of the water amount using the ideal gas laws
(corrections due to Van-der-Waals behaviour of water are small).
3.2.4
Summary
In most cases the released water amount is in the range of submicroliters (0.1 - 0.7 µl) and
can be even less than 0.1 µl for some special techniques like stepwise extraction procedures or
samples with an extremely low water content (≤ 0.1 ‰wt). The uncertainty is at least about
50 µg with regard to the weighing and, therefore, results in a total uncertainty larger than
50 % in the case of 0.1 µl water. However, using the pressure-based water determination via
calibration curves, the total uncertainty is about 6 % even at 0.1 µl. For NGT calculation the
total uncertainty should not exceed 5 % to achieve meaningful results. Therefore, the water
determination via weighing has to be rejected. The noble gas concentrations are derived
using the water vapour pressure as the uncertainty is in most cases below 5%.
3.3. Extraction of water and noble gases
3.3
51
Extraction of water and noble gases
In this section different extraction methods like crushing in copper tubes, milling with a
steel ball, thermal decrepitation and extraction by microwave heating will be introduced
and discussed. Of special interest is the separation of gases from the air- and water-filled
inclusions, which can only be implemented with adequate extraction techniques.
3.3.1
Design of the extraction line
The extraction line built by Rice (2004), which was used in a modified version at the beginning, is displayed in Fig. 3.14. In the basic version, a vessel with degassed water was
attached instead of the pressure gauge. It was intended to provide a constant gas flow from
the stalagmite crusher to the sample container. The sample container was therefore cooled
with liquid nitrogen to establish the gas flow from the crusher to the gas container. However,
this did not work very well as it resulted to be difficult to prepare degassed water in the necessary quality. The vessel was filled with about 250 ml of water, which needs to be degassed
to more than 99.99995 % to prevent a major influence on the speleothem results. Even in the
case of this efficiency a significant noble gas amount will derive from the degassed water as it
is about 500 000 times more than the water typically released from speleothems. Therefore
this design was rejected.
In the modified version a freezing finger and a capacitance pressure gauge (Keller Lex1) were
attached instead of the water vessel. The speleothems were crushed either in a steel cylinder
respectively in a copper tube, which had been evacuated. Before crushing the hand valve 4 is
closed, valve 1,2 and 3 are opened to freeze the water at the lowest end of the left side. Then
valve 1 is closed and the line is heated to determine the water amount using the water vapour
pressure. Afterwards the gases are expanded to the whole line including the sample container,
which is closed off after a certain time. The gas-filled sample container is attached to the
preparation line of the mass spectrometer for noble gas measurements. However, using this
procedure not all noble gases can be transferred to the mass spectrometer due to the volume
splitting at the extraction line, which cuts off about 50 % of the total gas amount. This setup
was used for some samples in the first test runs (”run August” and ”run November”).
At the same time a simplified extraction was implemented directly at the mass spectrometer
preparation line, which is also shown in Figure 3.14. At first, the sample is crushed in a
copper tube and then the water vapour pressure is determined in the unheated line using a
capacitance pressure gauge (Keller Lex1). Afterwards the gases are frozen totally into the
cryogenic traps without any splitting. This setup was used for most samples in the first two
measurement runs.
The final extraction line (shown in Fig. 3.15) is designed to enable the extraction of at least
one sample per day including 24 hours pumping time. Therefore the crushing devices (steel
cylinder as well as copper tubes and the cracking device for glass ampuls) can be attached at
the extraction system at three different inlet points. One sample can be pumped by the scroll
pump, the second sample can be pumped with the turbo-molecular pump and the third can
be extracted and its gases being frozen to the traps at the same time. If a shorter pumping
time is used (e.g. 12 hours), it is even possible to measure two samples a day.
The pre-vacuum of the extraction line is established by a scroll pump (Varian SHO 1001
UNIV) and is controlled by a pirani pressure gauge (Pfeiffer compact pirani TPR 261). To
avoid longer operation times of the scroll pump, an expansion volume ( ≈ 1 l) is used to
52
Chapter 3. Working with speleothems
Extraction modified stalagmite line
Pirani
2
4
Scroll/
Turbopump
3
1
Pressure
Gauge
Sample container
(Gas)
Crusher
(Stalagmite)
Extraction on line
Gas Traps
Pressure
Gauge
Water Trap
Scroll
Crusher
Sample
Figure 3.14: Design of the two extraction lines, used in the first measurement runs. The upper drawing
presents the external extraction at the modified line of Rice (2004), the lower one shows the extraction at
the preparation line of the mass spectrometer. In each case the sample is crushed (or heated) and then
the water vapour pressure is controlled by the pressure gauge. The gases are entirely transferred to the
mass spectrometer in the case of the direct extraction. If the sample is prepared at the line of Rice, a
volume splitting was necessary. The gases have been expanded to the sample container with a volume of
about 100 cm3 , so that the main part is in the sample container. Nevertheless, about 47 % are lost due
to this splitting.
3.3. Extraction of water and noble gases
53
1 Expansion Vol1
2 Expansion Vol 2
3 Pressure Gauge CMR263
4 Cold finger
5 Pirani Crusher
6 Crusher A
7 Crusher B
8 Crusher C
9 Penning Turbo
10 Turbomolecular pump
11 Scroll pump
PrepLine
MM5400
1
cracker
3
2
4
5
9
6
7
8
10
11
Figure 3.15: Design of the extraction line used in November 2007 and in subsequent measurements, e.g.
run Gamma and run Delta. For higher throughput and longer pumping times the number of crushers
has been increased (no. 6-8) and additional pump lines have been established. Number 1 and 2 indicates
the expansion volumes for water determination. The water is frozen into the cold finger (no.4) and then
expanded to the line between 3 and 4 in the first step. Subsequently, the expansion volumes are attached.
preserve a good pressure (10−3 - 1 mbar) for at least 2 days. If the prevacuum is worse
or if samples have to be pumped, the scroll is switched on manually. There are two inlets
designated for pumping with the scroll.
The high vacuum is established using a turbo-pump (Pfeiffer TMU 071 P) at 1500 Hz,
which is operating all the time and which can be used via two different pathways for sample
pumping. The pressure is controlled above the turbo pump using a compact cold cathode
gauge (Pfeiffer IKR 250) with a range of 2·10−9 to 1·10−2 mbar. Typically, the line pressure is
in the 10−8 mbar range. Additionally, the line pressure is controlled by a pirani gauge (Pfeiffer
compact pirani TPR 261) with an operating range down to 5·10−4 mbar to separately control
the extraction of the samples. Another pressure gauge (Pfeiffer compact capacitance gauge
CMR 263) with a higher precision (2 ‰) is used for the water determination via water vapour
pressure.
The water released during the extraction is frozen for 20 minutes into the cold finger (no.4,
see Fig. 3.15) using a dry-ice alcohol slush. Afterwards, the cold finger is closed off from the
line by a hand valve and the gases in the line are transferred to the gas traps. The gas loss
due to the splitting is in the low permil range as the volume of the freezing finger is about
2 ml and the rest of the line including the crusher more than 750 ml. The water vapour
pressure is measured using different expansion volumes (≈ 100 ml and ≈ 275 ml) to prevent
condensation and to achieve maximum precision by ideal utilization of the working range
(1 · 10−3 - 11 mbar) of the pressure gauge.
Before the manometric water determination, the gases are transferred to the noble gas preparation and separation line (for detailed description see Friedrich, 2007). The first trap, a naked
steel croygenic trap, is held at 25 K so that all gases except He and Ne are condensed. After
20 min the connection to the second cryogenic trap, held at about 10 K and filled with activated charcoal, is opened for 20 min in order to condense He and Ne there. Subsequently, the
54
Chapter 3. Working with speleothems
VCR-connection
copper vacuum ring
modified
hand valve
preheated
area
magnet
sample
crusher old type
crusher - modified
Figure 3.16: Extraction by crushing in a steel cylinder. On the left side the original crusher is depicted.
The modified version is displayed on the right. As the sample is located for pumping in the storage space
offered by the modified hand valve, the lower part of the crusher can be baked out without affecting the
sample.
gases are released from the traps in discrete steps and admitted to a GV 5400 noble gas mass
spectrometer for analysis. In the first step the charcoal trap is heated to 45 K to release the
He fraction. Next, the trap is set to 90 K to release Ne. Finally the naked steel trap is heated
to 130 K to release the heavy noble gases (Ar, Kr, Xe) as well as nitrogen and most other
gases that may be present in the sample. The reactive gases in this fraction are removed
by a heated SAES GP50 W2F getter and a cold SAES AP10N getter, before the purified
Ar-Kr-Xe mixture is admitted to the mass spectrometer. The cold getter is also used for the
He and Ne fraction and for the rather pure calibration gases (so-called fastcals), which are
measured to correct short-term drifts of the mass spectrometer.
3.3.2
Extraction using a metal crusher
The noble gases can be extracted from the stalagmites by crushing or heating. A well-known
setup is the use of a stainless steel cylinder, which is common in geological investigations.
The sample is put inside, pumped until a high vacuum is reached and then the inclusions can
be opened by crushing with a steel ball (Dennis et al., 2001; Rozanski and Dulinski, 1987;
Schwarcz et al., 1976). First test series with a similar extraction system showed unacceptably
high background values (Rice, 2004). Therefore, this method was not used at the beginning
of this project. It turned out that the high background was caused by small leaks or problems
with the vacuum fittings and not by the crusher itself or the crushing procedure.
In a second attempt, the extraction in a steel cylinder by crushing with a steel ball was
investigated more closely. Detailed information can be found in the thesis of Marx (2008). In
the following, the setup as well as the main results are presented briefly. A schematic drawing
(Fig 3.16) shows the two types of crushers. The first one is a simple 33 cm long steel cylinder
3.3. Extraction of water and noble gases
55
50
weight-%
40
30
1
2
3
4
5
<63 µm
63 µm-200 µm
200 µm-630 µm
630 µm-2,0 mm
> 2,0 mm
copper tube
20
10
0
1
2
3
4
5
50
mean crusher
weight-%
40
30
20
10
0
1
2
3
4
5
Figure 3.17: Grain size distribution of crushing in the steel cylinder and copper tubes. The copper
tube plot refers to speleothems wich are squeezed one time inside the tube by a vice. The distribution
was determined from a BU-U sample and shows a typical pattern for squeezing in copper tubes. The
crusher distribution was determined from several BU-U pieces by calculation of a mean. The stalagmite
pieces have been hit 60 times with the steel ball.
with an inner diameter of 26 mm containing a 23 mm steel ball. The cylinder is attached by
a copper vacuum ring to a flange, which is welded to a VCR-fitting.
The second crusher type is similar to the basic version, but additionally offers a place for
sample storage during bake-out of the crusher at high vacuum, preventing the speleothem to
heat to the same extent. This special place for sample storage is constructed from a modified
hand valve with an inner diameter of 14 mm and a length of 10 mm. The sample is placed
in the attachment by tweezers. Due to this storage the cylinder can be baked out at higher
temperature (150 - 250 ℃) without affecting the sample. Subsequent to the cleaning and
pumping process, the steel ball, which is also baked out, is moved above the sample inlet
connection, so that the sample can be thrown down to the bottom of the cylinder by moving
the valve. Afterwards the sample is crushed in high-vacuum by the magnetically moveable
ball.
According to the number of strokes a corresponding grain size distribution was achieved. The
fraction of fine grains (≤ 200 µm) is about 40 % wt in the case of 60 hits and even larger than
85 % wt if the sample was hit 400 times. Compared to crushing in copper tubes (see Fig. 3.17)
a higher fraction of small crystals can be obtained. Using copper tubes, the fraction of fine
grains (≤ 200 µm) consists only of about 15 % wt. Furthermore, the extraction is better
reproducible with regard to the grain size distribution as well as the background contribution.
The grain size distribution is of special interest regarding the separation of noble gases from
air and water-filled inclusions as well as for the extraction efficiency. Water-filled inclusions
56
Chapter 3. Working with speleothems
are supposed to be smaller than 10 µm and mainly even smaller than 1 µm, whereas the airfilled inclusions are in general larger. Crushing with the steel ball enables stepwise separation
and extraction techniques, which are discussed in section 3.4.
A further advantage of the modified steel crusher is the possibility of preheating the system.
Especially for thermal decrepitation as well as for simple crushing the blank values can be
reduced with this method. To investigate the effect of preheating, the steel crusher was
vented for 20 minutes with atmospheric air, then pumped for 20 minutes with the scroll
pump and afterwards 4 hours with the turbo pump at different temperatures. Subsequent to
this pumping procedure the hot-blank was determined. For thermal decrepitation the whole
system, except the crusher, is heated to 100 ℃ and the crusher itself to about 150 ℃ for 2
hours. Three steps have been tested, preheating at 70 ℃, which is common in the case of the
original steel cylinder, preheating at 250 ℃ and 350 ℃.
With little preheating (70 ℃) the hotblank signal (2h extraction at 150 ℃) on the multiplier
consists of 90 counts per second (cps) in Xe and 340 cps in Kr, whereas it is about 30 cps for
Xe and 160 cps for Kr if the crusher is preheated at 350 ℃ using the same pumping time.
Only with this reduced hotblank values reasonable results can be obtained using thermal
decrepitation.
3.3.3
Extraction by crushing in a copper tube
A second possibility for noble gas extraction is the use of collapsible materials like copper for
crushing (Roedder et al., 1964; Harmon et al., 1979; Scheidegger, 2005). The sample is cut
into elongated or cubic pieces and afterwards put into the copper tube (8 mm inner diameter,
about 30 cm length), which was welded on one side and attached to the the extraction line
by a UHV-tight connection (Swagelok tube fitting) on the other end. The tube with the
sample is pumped until a good pressure is reached (about 2-5·10−8 mbar). Then the part
of the copper tube where the sample is located is squeezed with a vice. To avoid disruption
of the copper, it is never squeezed at the same parts. Bottlenecks have been prepared with
pliers to enable a multi-stage extraction. The sample is filled into the upper part, squeezed
once, then the first narrow part is opened with special pliers, so that the crushed pieces fall
down to the next stage and can afterwards be squeezed again (s. Fig. 3.19). This procedure
is repeated several times.
The extraction of noble gases from speleothems by crushing in copper tubes is a rather
simple method, but it is difficult to obtain a stable and well-defined background. As the
pressure of the vice on the copper tube and the stalagmite samples is not strictly reproducible,
the background values for the noble gases as well as other parameters like the grain size
distribution are variable.
Empty copper tubes have been squeezed several times to examine the relationship between
blank signal and squeezing steps (s. Fig. 3.20). All noble gases show a clear trend. Towards a
higher number of steps the amount of released atoms from the copper increases in an almost
exponential way. Therefore it was attempted to reduce the squeezing steps to a minimum
number, that means mostly one or two squeezing steps.
It has to be noted, that the displayed results consist of a summary of different runs, in which
different methods have been used. Additionally, the results are influenced by blank problems
due to the not UHV-tight pressure gauge in the very first measurement. Therefore the
results are not strictly comparable. However, the points which are located on the exponential
curve were measured in one run. The rest is displayed to give an impression of the signal
scattering (see results at two squeezing steps), which depends on the number of steps and
3.3. Extraction of water and noble gases
57
Figure 3.18: Extraction using copper tubes. The picture in the middle shows how an elongated
speleothem piece is inserted in the one-side welded tube. On the left side the vice is displayed, which
was used to squeeze the tubes. On the right, a copper tube after squeezing is shown. In the lower left
picture pieces of the stalagmite H12 are displayed, which were extracted using copper tubes.
UHV connection
presqueezed
part
welded end
Figure 3.19: Squeezing technique used in the case of copper tubes. The sample (black square) is inserted
in the upper part of the precrimped tube, where it is squeezed for the first time, then the first crimped
point is opened by special pliers. The crushed sample pieces fall down on the second stage, where the
procedure is repeated.
58
Chapter 3. Working with speleothems
11 000
29 000
40 000
1200
Ne-20
800
signal (V)
signal (cps)
1000
600
400
200
-2
2.5x10
-2
2.0x10
-2
1.5x10
-2
1.0x10
-2
5.0x10
-3
0.07
Ar-40
0.0
0
0
1
2
3
4
0
5
1
2
3
4
5
2
3
4
5
80
100
Kr-84
Xe-132
70
80
60
signal (cps)
signal (cps)
3.0x10
60
40
50
40
30
20
20
10
0
0
0
1
2
3
squeezing steps
4
5
0
1
squeezing steps
Figure 3.20: Copper tube squeezing blanks. The tests have been carried out with empty copper tubes
using the multi-stage extraction explained above. The tests at 0, 1, 3 and 5 squeezing steps refer to one
measurements series (filled black circles), whereas the test at 2 squeezing steps (open circles) have been
determined earlier using different methods and setup.
also on the measurement run due to changed gas preparation and sensitivities. For the blanks
shown at two squeezing steps, Xe yields about 53 cps with an uncertainty of 12 cps. These
measurements belong to the older test series, which were influenced by the leaking pressure
gauge. Prior to the last runs the copper tubes were preheated to about 200 ℃ to reduce the
background and scattering. By preheating the copper tubes and replacing the pressure gauge
a background of 15 cps with an uncertainty of 8 cps was obtained and used for Xe correction of
the two-squeezing steps extraction. As the samples mostly yield signals below 100 cps for Xe
(except in the last two runs, Delta and Gamma), this background contribution and especially
the uncertainty is not acceptable with regard to the aimed precision. An overview of the
influence of the crushing uncertainty on the sample signal precision is given in Table 3.2.
For separation of noble gases from air- and water-filled inclusions different techniques are
suggested to achieve this objective. One idea is based on the heating of the crushed sample.
However, heating can release noble gases not only from the sample itself, but also from the
copper material. Therefore some heating measurements of empty copper tubes have been
performed.
Empty copper tubes were attached and pumped for at least 8 hours and up to 48 hours,
before a blank of the cold tube was performed. Afterwards, measurements of the copper tube
at 100℃ and 200℃ have been made. Precedent to the 200℃ - blank the tube was pumped
at 100℃ for more than 3 hours. All measurements in one run have been performed with the
same copper tube and without venting the system in between. This procedure was repeated
twice to check the representativeness of the results.
The measurements revealed significant background values (s. Fig. 3.21). The 132 Xe blank,
e.g. rises from signals below 100 cps up to 400 - 600 counts during heating at 100℃ and
drops back to significantly better values after the 100℃ measurement and the subsequent
3.3. Extraction of water and noble gases
600
59
Xe-132
500
signal (cps)
400
300
200
100
0
cold / heated / heated
100˚C
200˚C
cold /
heated / heated
100˚C
200˚C
Figure 3.21: Blank of empty copper tubes determined at different temperatures. The copper tubes have
not been exposed to the atmosphere between the heating steps, but instead pumped at room temperature
and 100℃ respectively before the start of the next measurement.
pumping step at the same temperature. Even if the temperature is higher in the last heating
measurement, better blank values can be achieved. This behaviour is similar for all noble
gases, but is most pronounced in the case of the heavier ones.
This problem may be due to the adsorption of noble gases in the copper tube or even to
higher noble gas concentrations in the copper material compared to the normal atmospheric
composition, caused by the production and purification process of the copper tubes. The
extremely high xenon-blanks at 100℃ may not be fully assigned to the adsorption. Otherwise
longer pumping should be able to remove these gases. In the second test, the copper tube
was pumped 4 times longer (48 hours) than the first copper tube, but nevertheless the signal
during heating was comparable. However, this test may also show that it is necessary to use
higher temperatures to enable the desorption of Xe from the surfaces. Heating under vacuum
can remove a certain part of the noble gases and prepare the system in a more defined way.
Table 3.2: Corrections used for the last copper tube measurements (June 07, run Beta). In the row
’crushing once’ the mean blank with the estimated uncertainty based on the preceding blank measurements is displayed. The blank values refers to the interpolation from step 5 to zero of the squeezing
measurements. The sample ’uncertainty due to the blank’ refers only to the crushing uncertainty.
sample type
crushing once (mean)
typical sample signal
uncertainty due to blank
4
He(cps)
38 ± 4
1500 - 6000
1%
20
Ne (cps)
410 ± 40
7000 -20000
1 %
40
Ar (V)
Kr (cps)
−3
(6.9 ± 3)·10
0.4
≤ 1%
16 ± 8
400-1000
≤2%
132
Xe (cps)
15 ± 8
40 - 90
≤ 20 %
The effect of preheating has not been tested with a sample inside the tube, because diffusive
loss of noble gases has to be suspected. Thus, the copper tubes have been preheated under
vacuum conditions, cooled down, opened and then the speleothem piece was inserted. It can
be expected that a certain amount of gases gets adsorbed during the filling. So we pumped
the tube including the sample again during mild heating (70 ℃ ). As we do not know and are
not able to determine the blank in this case exactly, we can only use the values of the empty
tubes as a first-oder estimation. In the case of hot sample extractions the blanks turned out
to be by far to imprecise to achieve meaningful results. Due to this background problems the
use of copper tubes was rejected.
60
Chapter 3. Working with speleothems
50 mm
attachement for
pumping with
o-ring connection
capillary for
flame sealing
100 mm
glass vial
outer diameter: 10 mm
inner diameter: 8.2 mm
sample pieces
Figure 3.22: Glass vial for microwave extraction of noble gases. The sample pieces are inserted before
the capillary is built by the workshop. Then the glass vial is pumped using an o-ring connection on the
upper part. The pumped sample is closed off by flame sealing.
3.3.4
Extraction by microwave heating
Noble gases from water-filled inclusions can also be extracted by heating stalagmite samples
in the vacuum system. A sophisticated method for opening of the water-filled inclusions is
microwave heating. The sample is crushed into small parts, put into a glass vial and then
pumped for a certain time. When a sufficient pressure (≈ 10−8 mbar) is reached, the vial
is flame sealed and afterwards put into a common microwave oven. The microwave-treated
sample is opened at the mass spectrometer in a special o-ring-free device (explained below).
Due to the microwave heating especially the (small) water-filled inclusions are supposed to
be opened, because of an increase in pressure. The air-filled inclusions are assumed to remain
closed.
In Figure 3.22 the glass vial is displayed. It is about 10 cm long and has an inner diameter of
8.2 mm in the lower part and is made of a low helium permeability glass. Due to a capillary
at the top part it is not possible to fill in the sample directly. It has to be put inside the glass
vial before the capillary is manufactured.
The vial is attached at the extraction system via an o-ring connection for pumping. The upper
part is divided into two sections by the small capillary area, which is needed for sealing. Using
a small gas blowpipe the capillary is warmed up first and then pulled down slowly with a
circular motion, when the glass is glowing reddish. After the sample is pumped and sealed, it
is put into a common microwave oven (Lunik 340) and heated typically for 5 min at 800 W.
Unfortunately, the samples were not only influenced by the microwaves itself but also heated
thermally by the hot rotary disc of the oven. To prevent this effect in later experiments the
sample was put on ceramics. In some cases condensation of water could be seen at the sharp
end of the capillary (Fig. 3.24). In the case of a flowstone with a slightly brownish colour and
a set of red stripes, a gas discharge with a whitish or light blue flash has been seen during
microwave treatment.
3.3. Extraction of water and noble gases
61
Figure 3.23: Glass vials for microwave extraction. On the left side a vial including a speleothem sample
is displayed, which is attached to the preparation system using an o-ring connection. On the right side
two pumped and by flame-sealing closed vials as well as one original glass vial are shown.
Figure 3.24: The effect of (microwave) heating on a speleothem sample, which consists of fine powder,
is obvious on the sharp end of the glass vial. In the lower part of the picture condensation of water
vapour inside the capillary part of the glass vial can be seen (black circle). The water vapour is produced
by the cracked water-filled inclusions in the milled stalagmite parts.
62
Chapter 3. Working with speleothems
Figure 3.25: System for extraction of microwave treated glass vials. The vial can be inserted totally
in the cracker system. Thus, contamination by leaking atmospheric air, as it is present in the case of an
o-ring connection, can be prevented. The hand valve is used for moving of a plunger, which cracks the
capillary and releases the gases from the vial. Below the cracker an opened sample is displayed.
The flame-sealed and microwave-heated samples are then opened in a special extraction
system (Fig. 3.25). A steel tube with a diameter of about 12 mm was provided with typical
VCR - fittings at each end. The glass vial can be put totally inside the tube, closed by a
fitting (blind flange) and is pumped afterwards. The glass is broken by a modified hand
valve at the capillary part, which was scored before. The released gases are frozen into the
cryogenic traps for further preparation.
To control the background, empty glass vials have been prepared and treated the same way as
real samples. The measurements show that the blank signal depends on the pressure during
flame sealing (Fig. 3.26). For 40 Ar an almost linear trend is visible, which is comparable
to the trend for the other noble gases. To achieve a lower background the glass and the
capillary is preheated three times. During the twisting off this results in a lower pressure
of 3 ·10−8 mbar to 2 ·10−7 mbar compared to values in the 10−6 mbar range without this
procedure. The glass background values counted in the first microwave run are comparable
to the extraction line blank, except at a pressure above 1 · 10−6 mbar during flame sealing.
Typical line blanks have been for instance 0.3 to 0.6 V in case of 40 Ar in the first microwave
run. This rather high value was reached due to the connection of the glass via an o-ring to
the noble gas measurement line. In the following run the o-ring connection was replaced by
the above explained special extraction system, which enables the opening of the vial totally
included in a vacuum chamber.
The comparison of an empty and microwave-treated glass vial with the lower line blank in
the second run, using the modified setup with preheating prior to flame sealing, showed the
two signals to be comparable. Therefore, it was assumed that the microwave treatment does
Table 3.3: Mean line blank and sample signals for the two microwave test series.
sample type
1.
1.
2.
2.
run,
run,
run,
run,
line blank
samples
line blank
samples
4
He(cps)
5
3·10
3·105 to 3·106
3 ·102
3
1·10 to 2·105
20
Ne (cps)
6
7 ·104 to 2
2
7 ·102 to 1
·10
·106
·102
·106
4
40
Ar (V)
−1
3.5 ·10
0.6 to 7.7
1 ·10−3
−3
7 ·10 to 1.3
132
Xe (cps)
65
80 - 1400
2.5
1.4 - 1100
3.3. Extraction of water and noble gases
1.2
63
Ar-40
signal (V)
1.0
0.8
0.6
0.4
0.2
0.0
0.0
-7
-6
-6
1.0x10
1.5x10
2.0x10
5.0x10
pressure at flame sealing (mbar)
-6
2.5x10
-6
Figure 3.26: 40 Ar blank values in the case of the microwave extraction. The blanks have been determined by the measurement of empty glass vials, which have been prepared and processed in the same
way as the speleothem sample. An almost linear relation between pressure during flame sealing and the
mass spectrometric noble gas signal can be detected. A similar trend has been observed for the other
noble gases.
not release additional noble gases from the glass. Furthermore, the background of the line
could be reduced considerably due to the new extraction system, for instance down to the
range of 8 ·10−4 to 2 ·10−3 V for 40 Ar respectively 1 to 5 cps for 132 Xe.
Table 3.4: Samples measured in the second microwave run. All samples are from not previously
measured material. They were chopped into coarse grains before they were filled into the glass vials.
’crush’ indicates the samples from which the noble gases were extracted by simple crushing. They are
mentioned for comparison and refer to the mean of all crushing measurements.
sample
H12-I
H12-II
H12-crush
CG-I
CG-II
CG-crush
flowstone
flowstone-crush
released water (µl)
not detectable
1.14
0.2 - 1.4
not detectable
1.6 ·10−3
−2
1.7 ·10 to 1.6 ·10−1
1.5 ·10−1
3.4 - 5.1 ·10−1
water content (wt‰)
0.58 ± 0.02
2.6
6 ·10−3 ± 6 · 10−3
0.005 up to 0.07
0.119 ± 0.001
0.20 - 0.29
air/water volume ratio A
2.9
1.8
2.2
0.35
0.91
1.1 - 7.7
In the first microwave extraction test four samples have been investigated. All vials have
been filled with pieces of the H12 stalagmite from the Hoti Cave. To test the extraction
efficiency, the stalagmites pieces were prepared in different ways. Two parts were crushed
to coarse grains prior to the pumping and microwave heating. One sample consisted of fine
powder from already measured samples and the fourth sample of a simple cube. During
flame sealing rather high pressure values occurred (mostly 2 − 3 · 10−6 mbar, one sample up
to 1 · 10−5 mbar). All glass vials including the speleothem samples were heated for 5 min at
800 W in the microwave oven, then connected with an o-ring to the mass spectrometer line
and opened after the line was pumped sufficiently (final pressure 5 · 10−8 - 1 · 10−7 mbar).
The noble gas measurements did not yield the results we hoped for. Although it was not
possible to determine the released water amount, the three-isotope-plots (Kr/Ar vs. Ne/Ar or
Chapter 3. Working with speleothems
Kr/Ar
64
2.4x10
-4
2.0x10
-4
1.6x10
-4
1.2x10
-4
8.0x10
-5
crushing/heating
air
asw
microwave samples
1.0x10
-3
2.0x10
-3
3.0x10
-3
4.0x10
-3
Ne/Ar
Figure 3.27: Kr-Ar-Ne plot of samples measured in the first microwave test run. For comparison
samples from H12 and MA (Chile) are displayed, from which the noble gases have been extracted by
crushing. They are indicated by black circles in the figure. The microwave sample in the lower left edge
is displayed without error bar. The errors extend over the whole plot.
Kr/Ar vs. Xe/Ar, s. Fig. 3.27) provide some information about the air/water volume ratio.
Unfortunately, the samples extracted with microwaves did not differ from pieces extracted
by normal crushing methods. The air/water ratio is similar and not significantly better as
has been assumed before.
The microwave measurement with the lowest excess-air belongs to the stalagmite powder.
Although this powder consists of the remaining part of preceding measurements, an obvious
water amount (Fig. 3.24) and the according gases could be extracted. This sample resulted
in a slightly improved air/water ratio compared to normally crushed samples.
The difference between the coarse grains and the uncrushed cube was not significant. The
coarse grains, which have been measured in a previous run, were totally away from the rest
(lowest red star on the left side) and therefore showed that most of the water and also the
noble gases have been extracted in the previous run and few water and gases could be released
with the microwave treatment. Furthermore, the error was considerably elevated in this case
due to the comparatively high o-ring-blank as the vacuum cracker was not usable yet.
Concluding from the first microwave tests, this method does not yield significantly lower
air/water volume ratios, but shows that, at least in the case of fine powder, a certain improvement may be possible compared to the crushing. Furthermore, the measurements indicate that it is feasible to open tiny inclusions (≤ 1 µm) by microwave heating. The already
measured powder, which was crushed and pumped under vacuum in a previous test series,
released again water and noble gases due to the microwaves.
In the second run five samples from different stalagmites were investigated using the microwave method. One from a flowstone from the Hoti Cave in Oman, two pieces from the
H12 stalagmite and two parts from the stalagmite CG from Cuba. The speleothem samples
were crushed to coarse grains before they have been put into the vials. The glass vials were
prepared using the modified flame sealing process with preheating of the capillary and were
3.3. Extraction of water and noble gases
65
CG micro
2.5x10
-4
aew 20˚C
H12
CG
flowstone
other
aew 0˚C
-4
1.5x10
-4
Kr/Ar
2.0x10
CG
micro
flowstone
micro
air
1.0x10
-4
0.0
H12
micro
H12 micro
2.0x10
-3
4.0x10
-3
Ne/Ar
Figure 3.28: Kr-Ar-Ne plot of samples measured in the second microwave run. The samples scattering
in the region between AEW and air belong to measurements of pieces from the stalagmite CG extracted
by simple crushing and are given for comparison. The microwave sample ”CG micro” in the middle of
the plot is displayed without error bars. The errors extend over the whole plot.
opened totally included in the vacuum system (Fig. 3.25). Therefore, the blanks could be
reduced substantially. The blank signals were up to three orders of magnitude lower in this
test series compared to the first microwave run (Table 3.3).
The second run confirmed the results from the first microwave test series. The samples extracted by microwaves showed in general a worse air/water-volume ratio compared to similar
speleothem pieces crushed within copper tubes. If the released water is referred to the sample
amount in g, the content of water-filled fluid inclusions can be calculated. In general this
value of microwave extraction is significantly below the numbers of crushing. As the extraction via copper tube is not 100% efficient, even higher amounts of water will be released in
case of milling with a steel ball and therefore the extraction by microwave heating appears
especially poor. However, in the second microwave run the speleothems have not been milled
to powder, which might yield in general better values.
The uncertainties are much higher in the case of microwave extraction compared to simple
crushing as the total amount of noble gas and water is in general rather small. Therefore, it
will be inappropriate to use this method for calculation of noble gas concentrations and subsequent fitting. Microwave treatment of milled stalagmite powder may yield some potential.
However, it requires an extremely low background and a precise water determination at the
lower limit of the method.
3.3.5
Extraction by thermal decrepitation
A further possibility to extract noble gases from speleothems is a simple heating process as
already investigated by Yonge (1982), Matthews et al. (2000) and Scheidegger (2005). The
sample is heated under vacuum to a certain temperature, either crushed before to a smaller
fraction or uased as a whole cube. Due to heating, the pressure in the water-filled inclusions
will increase and crack the calcite. Thereby the noble gases are released. Matthews et al.
66
Chapter 3. Working with speleothems
0.45
H12 200˚C
H12 100˚C
CG heat
CG crush
0.40
0.35
0.30
A
A
0.25
1
0.20
0.15
0.10
100
120
1
at 20
15 mi
0 n
˚C
80
2
at 0 m
20 in
0
˚C
60
2
at 0 m
20 in
0
˚C
40
extraction time (min)
cr
us
hi
ng
20
on
ly
0.05
Figure 3.29: Effect of the heating duration on the air/water-volume ratio. On the left side H12 samples
are displayed, which were heated between 20 and 120 minutes. The two H12 samples extracted at 200℃
with a duration of 20 and 30 minutes were crushed to small pieces prior to the measurement. The rest
was extracted subsequent to a crushing step under vacuum. The first CG data point at 20 min and 200℃
belongs to a combined extraction of crushing and heating. The second point at 200℃ reflects the result
of a single heating step after two crushing extractions under vacuum.
(2000) investigated thermal decrepitation at 900℃, whereas Scheidegger (2005) limited the
heating steps to about 600℃ to avoid decomposition of the calcite.
Different test series have been performed to investigate the potential of thermal decrepitation.
For the first measurements small pieces of the stalagmite H12 have been milled into grains
with diameters below 2 mm. Subsequently, they were inserted into the copper tubes for
pumping and thermal decrepitation. The hot extraction was carried out 20 min respectively
30 min at 200℃. One measurement consisted of a combination of crushing under vacuum
and a subsequent heating step, which was performed 20 min at 200℃. In the beginning of
2007 a series of heating tests have been performed, in which the stepwise crushing extraction followed by the 20 min lasting thermal decrepitation at 200℃ was investigated. Every
sample was squeezed up several times, analysed and pumped after the initial crushing steps.
Subsequently, the speleothems were heated for 20 min to 200℃ for gas release. The effect of
this treatment is displayed in plot 3.29.
In the case of the H12 stalagmite, a heating of 20 minutes at 100℃, respectively 200℃ did
not yield better A values compared to the crushing data. However, a significantly improved
result was obtained for the CG stalagmite. In general, the effect of reducing A by thermal
decrepitation is most pronounced at longer heating times. An exposure time of 20 minutes of
the crushing devices to the heat coils does not seem not to be sufficient to heat up the whole
setup. With regard to the temperature a lower A value was obtained for the H12 sample
at 200℃. For pieces from H12 the released water amount was larger in case of the higher
temperatures as well as for longer heating times. 30 minutes heating at 200℃ released 0.13
and 0.048 µl/g, whereas at the 120 minutes heating step 0.14 (100℃) and 0.21 µl/g (200℃)
were released. A similar trend was found for the CG pieces with 0.001 µl/g at 20 min and
200℃ and 0.042 µl/g at 120 minutes heating with 150℃. However, longer heating periods
and higher temperatures bear the problem of increased noble gas desorption from the crusher
walls as well as the diffusive release of noble gases from the calcite matrix. In geological
3.3. Extraction of water and noble gases
67
A - H12
H12 Crush
A
10
1
50
100
150
200
250
temperature (˚C)
Figure 3.30: Stepwise heating experiment on one H12 sample. The first data point refers to a different
sample extracted by crushing only. The subsequent data points have been obtained by heating of one
sample at different temperatures, but using always the same extraction time of 120 minutes.
applications heating periods for minerals range from 20 to 30 minutes at high temperatures
of e.g. 200 - 1500 ℃ (Podosek et al., 1980; Stuart et al., 1995).
In ”run Beta” (2007) the effect of different temperatures on the extraction efficiency as well as
the air/water volume ratio was tested. One sample was inserted into a copper tube, cleaned
from superficially adsorbed noble gases by pumping and extracted in subsequent heating
steps at various temperatures. In each case the extraction time was fixed to 120 minutes. A
clear trend towards a lower air/water volume ratio can be observed at higher temperatures
(Fig. 3.30). The best values are obtained between 140 and 200℃ and are about 0.4 to 0.5.
At 250℃ a small increase can be observed, which may be due to noble gases released from
the copper tube walls or the calcite matrix. The amount of water was easily measurable as
it ranged between 0.15 and 0.5 µl (except at the first step at 50℃).
Heating may influence the noble gas concentration by diffusion. Therefore, deviations from
noble gas ratios corresponding to air-equilibrated water and as well deviations from the
excess-air line in the two-isotope-plots should be observable. In Fig. 3.31 all samples which
have been extracted by thermal decrepitation are displayed. Most of them are lying on the
the excess-air line of water equilibrated at about 30℃.
The pattern is more complex concerning the Xe and Ne concentrations. In Fig. 3.32 the same
samples as in the Kr-Ar plot 3.31 are displayed. However, some deviations from the expected
line can be observed. Most samples are located on a parallel below the excess-air line at
30℃, which indicates an excess in Ne. Points above the expected excess-air range belong to
samples from the Cuban stalagmite CG and to the two samples from H12, which have been
measured in the very first runs. A location above the excess-air line indicates surplus-Xe.
This excess in Xe can be explained for the H12 samples by the release of adsorbed Xe from
the copper tube walls. A correction for this effect was not possible. The Xe-excess may be
explained in the same way for the CG samples, but it may be influenced additionally by a
matrix component, which is enriched in Xe.
Despite the rather high and uncertain hotblanks of the copper tubes, the deviations from
the expected concentrations can not always be explained by systematic errors or technical
problems. E.g. the Ne-excess in case of the H12 samples occurs also using the crushing
extraction. Thus, it is not related to the extraction procedure, but rather to the calcite
Chapter 3. Working with speleothems
Kr (ccSTP/g)
68
2.5x10
-6
2.0x10
-6
1.5x10
-6
1.0x10
-6
5.0x10
-7
0.0
0.000
0.005
0.010
0.015
0.020
Ar (ccSTP/g)
Figure 3.31: Kr and Ar noble gas concentrations of all samples extracted by thermal decrepitation.
The values are given in ccSTP per g of water. The red points indicate air-equilibrated water. The upper
dashed line refers to water equilibrated at 0℃ with addition of various air amounts. The lower line
represents the excess-air line corresponding to water equilibrated at 30℃.
and its inclusions. This assumption can be approved by the examination of the Ne- and the
Ar-isotope ratios. They are displayed in two small boxes in the same plot. In the case of
the 40 Ar/36 Ar -ratio most of the data is scattering very close to the value of atmospheric air.
The deviations can be explained by the measurement uncertainty of the noble gases, which
is below 1 % for 40 Ar and some % for 36 Ar. Only one measurement exceeds this range with
an ratio of about 4800. This value belongs to a heating measurement at 50 ℃ with a very
low amount of released water, but a relatively high noble gas signal. The gases are supposed
to be mainly desorbed from the walls as well as from the calcite surface.
The neon ratios scatter in the range of the atmospheric ratio of 9.80. The deviations are
in general lower than 2 %, which corresponds to the measurement uncertainty of 20 Ne and
22 Ne. Taking into account the Ne- and Ar-isotope ratios, no fractionation due to the thermal
decrepitation at temperatures below 250℃ and 2 hours duration occurs. Therefore, it is
feasible to use thermal decrepitation as an extraction method or in a stepwise procedure
without alternating or fractionating the noble gas concentrations of the fluid inclusions.
Measurements of other samples confirm the ability of reducing the air/water volume ratio by
a heating step. A set of samples is displayed in a three-isotope plot (Fig. 3.33) including the
extraction by crushing. The crushing results of two samples (Spa12, Spa 52b) from Spannagel
Cave (Vollweiler et al., 2006) and a sinter plate from Bunker Cave scatter around the air point
in the same area as the H12 samples. Subsequent heating improves the air/water volume ratio,
which is indicated by a position closer to the equilibrated water points. In principle, a fitting
of the noble gas concentrations with regard to temperature determination becomes possible
using this stepwise procedure, as it gets easier to subtract the air contribution from the
total signal. However, the large disadvantage, so far, was the strong increase in uncertainty,
which made a meaningful fitting of the data impossible. This occurs due to a significant
background during heating which often exceeds 10 % of the sample signal. Additionally, the
water amount decreases similarly to the noble gas signal. Therefore, the uncertainty of water
3.3. Extraction of water and noble gases
8.0x10
69
-7
H12
AEW
6.0x10
-7
-7
Ar/
4.0x10
40
Ar
305
36
Xe(ccSTP/g)
310
300
295
AEW at 30 ˚C
290
10.3
10.2
10.1
Ne
10.0
20
Ne/
22
2.0x10
-7
9.9
9.8
9.6
0.0
0.0
5.0x10
-5
1.0x10
-4
air
9.7
800 1000
1540
1560
meaurement no.
1.5x10
-4
1580
1600
2.0x10
-4
Ne (ccsTP/g)
Figure 3.32: Xe and Ne noble gas concentrations of all samples extracted by thermal decrepitation.
The values are given in ccSTP per g of water. The blue points indicate air-equilibrated water. The upper
dashed line refers to water equilibrated at 0 ℃ with addition of various air amounts. The lower line
represents the excess-air line corresponding to water equilibrated at 30 ℃. In the small-sized inserts the
corresponding isotope ratios of Ne and Ar are displayed.
determination also increases to higher values, which makes it difficult, in combination with
the elevated uncertainties in the noble gas measurements, to push the total uncertainty of the
temperature determination below 10℃. The most difficult step is related to the background
control. It is not possible to perform an ideal and representative hotblank measurement
with the real sample. A measurement without the sample is not practicable as well, as the
crusher has to be opened to insert the speleothem sample. In this case noble gases will be
adsorbed on the surfaces and the measured blank will not reflect the real conditions. A
hotblank measurement including the unheated sample inside the vacuum system bears the
disadvantage of outgassing from the sample, which results in an overestimation of the actual
hotblank. The blank values are discussed in detail in chapter 3.5.6.
However, combined methods using a sophisticated procedure based on low background values
may help to overcome these problems.
3.3.6
Summary
Four different methods have been tested for the extraction of noble gases from fluid inclusions
in speleothems. The most exotic procedure using microwave heating did not keep the promise
of preferential and especially efficient opening of water-filled inclusions. It may be of interest
using very fine powder and a more powerful microwave oven, but so far no convincing results
have been achieved.
Squeezing of speleothems in copper tubes showed some success, but difficulties with the background as well as problems to reproduce extraction excluded this method for advanced studies.
Thermal heating in copper tubes releases too much noble gases from the copper walls and
therefore increases the total uncertainty due to difficult hotblank corrections. Furthermore,
it is not possible to perform preheating of the sample or to measure a hotblank including
the sample. Reproducible extraction is laborious or even impossible, as the pressure of the
Chapter 3. Working with speleothems
Kr/Ar
70
2.6x10
-4
2.4x10
-4
2.2x10
-4
BU-U
heat
2.0x10
-4
1.8x10
-4
1.6x10
-4
1.4x10
-4
1.2x10
-4
BU-U
heat
Soda
BU-U
Cu-tube
Spa12
heat
BU-U
Crushing
H12-heat
CG
sinter
heat
Spa52
heat
H12-crush
H12-heat
-5
-5
-5
-5
-5
-5
1.0x10 1.5x10 2.0x10 2.5x10 3.0x10 3.5x10 4.0x10
-5
Xe/Ar
Figure 3.33: Three-isotope plot of samples measured in a speleothem run in 2007 (run Beta). The
crushed samples of Spa12, Spa 52b from the Spannagel Cave and a sinter piece from Bunker Cave are
scattering around the atmospheric value (red square). Subsequent heating led to a reduced air/water
volume ratio indicated by a position more close to the air-saturated water points (indicated by blue sqares
- left point 25℃, right point 0℃).
vice on the sample during squeezing can not be controlled. The crushing steps are limited
to a rather small number (≤ 5) to prevent the formation of leaks. Another disadvantage is
the varying and low extraction efficiency. The achieved grain size distribution shows a high
variation with a small fraction of small grains (< 200 µm).
Thermal decrepitation seems to be very efficient as a tool for the reduction of the air/watervolume ratio and can release considerable water amounts from small grains, whereas further
crushing would not yield such results. However, extraction by thermal decrepitation is very
laborious as the background signal is rather high. Without a sophisticated preheating process,
too much desorbed gases will alter the measured values. The hotblank signal is high compared
to the extraction signal and therefore the total uncertainty is in general situated above the
aimed values. Another problem is the sometimes low water amount.
The most efficient method for noble gas extraction turned out to be crushing inside a steel
cylinder by a magnetically movable steel ball. The blank values are sufficiently low, in both
the empty crusher and the moving steel ball inside the crusher with and without calcite. The
steel cylinder enables a stepwise crushing procedure and leads to a reproducible extraction
efficiency. In general, the grain size distribution yields a large fraction (≥ 40 % wt) of very
small grains (≤ 200 µm) and is thus suitable for combined extraction procedures, which are
specialized on different inclusion dimensions. The crushing inside the steel cylinder as well
as a comparison with the copper tube squeezing and the achieved grain size distributions are
discussed by Marx (2008).
3.4. Separation techniques
71
2.2
BU-U
2.0
1.0
Sinter Bu
1.8
0.8
1.4
pressure mbar
water pressure (mbar)
1.6
1.2
1.0
0.8
0.6
0.4
0.6
0.4
0.2
0.2
0.0
0
10
20
30
40
50
60
0.0
0
10
20
30
40
50
60
number of strokes
3.0
BU2
water vapour pressure
(mbar)
2.5
2.0
1.5
1.0
0.5
0.0
0
10
20
30
40
50
60
number of strokes
Figure 3.34: Pressure increase due to the water vapour during crushing of different samples from Bunker
Cave. Most of the water is released in the first crushing steps. The red curve is a fit with an exponential
growth model.
3.4
Separation techniques
Many speleothems do not exhibit a favorable air/water volume ratio A and are mostly dominated by air-filled inclusions. In such cases no meaningful temperature can be calculated.
Stepwise procedures are assumed to reduce A (Scheidegger, 2005) by separation of noble
gases from air- and water-filled inclusions and thus this technique may enable the calculation
of temperatures with sufficiently low uncertainties.
3.4.1
Stepwise crushing
One of the crucial points, the separation of noble gases from air- and water-filled inclusions,
may be solved by a combination of several techniques or by stepwise procedures. We first
focus on the stepwise crushing and stepwise heating and then discuss the combined technique.
To investigate the extraction by crushing with a steel ball, the pressure in the extraction
line was recorded for a number of crushing steps. The pressure refers mainly to the released
water from the fluid inclusions. 0.1 µl of water corresponds to 0.125 ccSTP of water vapour,
whereas it contains only on the order of 10−6 ccSTP of dissolved atmospheric gases.
Three samples from Bunker Cave have been extracted by crushing in a steel cylinder in
the same way. With regard to the water vapour pressure, they show a similar behaviour,
which can be described by a strong pressure increase in the first steps and only a small
rise in pressure or even stagnation at more than 30 - 50 hits (Fig. 3.34). As the pressure
increase during crushing indicates that the main part of water is released in the first steps,
it is not feasible to mill the speleothem to fine powder first, to pump the gases and to take
an additional extraction step afterwards (by crushing or heating) to release the gases from
the smaller water-filled inclusions. However, by applying reasonable steps, as for instance
5 hits followed by 15 and finally 30 additional hits, stepwise crushing can lead to a reduced
air/water volume ratio and easily measurable water amounts.
72
Chapter 3. Working with speleothems
1.6x10
-6
1.4x10
-6
1.2x10
-6
1.0x10
-6
8.0x10
-7
6.0x10
-7
4.0x10
-7
2.0x10
-7
1
H12
aew
Kr (ccSTP/g)
3
5
10
100
200
+heat
0.0
0.0
2.0x10
-3
4.0x10
-3
6.0x10
-3
8.0x10
-3
1.0x10
-2
1.2x10
-2
Ar (ccsTP/g)
Figure 3.35: Effect of stepwise crushing on a sample from H12. The first crushing step resulted in
the most distant values from air-equilibrated water. In each crushing step the air contribution could be
reduced. All steps were performed by crushing (total number of strokes indicated in the figure), except
the last step, which refers to a combination of 100 additional strokes and heating at 150 ℃ .
Test measurements with a piece of the H12 stalagmite showed that each crushing step leads
to a reduction of A, which starts at about 1.1 at hitting the sample once (Fig. 3.35). In the
last step A was reduced by one order of magnitude to about 0.13. The released water amount
was about 0.2 µl in the first step and 0.18, 0.08, 0.12, 0.29 and 0.21 µl in the following steps.
A disadvantage of the stepwise procedure may be the preferential opening of inclusions of
a certain type. During the first hits (≤ 10) the larger inclusions (Ø ≥ 1 µm) are supposed
to be opened, whereas in the subsequent steps smaller and mainly water-filled inclusions are
cracked. This can lead to an alteration of the noble gas signals (Fig. 3.36). In run Delta
a sample from MA and H12 was extracted by stepwise crushing. Hitting the speleothems
few times under vacuum resulted in each case in an extreme offset in Ar. This offset became
smaller in the subsequent steps and finally ended in the expected range of air-equilibrated
water with addition of atmospheric air. Additionally, a systematic offset for Xe and Kr was
found for high crushing numbers. This may be due to diffusion of the lighter noble gases
from the very small water-filled inclusions (Ø 1 µm) in the precedent steps, but can also
be influenced by noble gases from the calcite matrix. This effects needs to be investigated in
the future.
3.4.2
Stepwise heating
A further possibility for the separation of noble gases according to their provenance is a
heating process as already stated by Scheidegger (2005) and Scheidegger et al. (2006). The
sample is heated to a certain temperature, either crushed before to a smaller fraction or
used as a whole cube. Stepwise heating can reduce A by more than one order of magnitude.
The results of a stepwise heating run is displayed in Fig. 3.30. A stalagmite sample from
H12 was inserted as a whole uncrushed cube and pumped for about 1 day. The thermal
decrepitation in each step lasted 120 min. The temperature was varied between 50℃ and at
maximum 250℃. Between the thermal decrepitation steps, two blanks for background control
have been measured. Alternatively, the time was used for pumping the copper tube including
the sample up to several days.
3.4. Separation techniques
2.0x10
73
-5
MA
H12
aew
aew+air
1.5x10
5
-5
Kr (ccSTP/g)
20
1.0x10
60
-5
Ar offset
160
3
5.0x10
200
-6
Ar offset
30
0.0
0.0
10
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Ar (ccSTP/g)
Figure 3.36: Ar offset measured during stepwise crushing procedures (numbers refer to the total of
applied strokes). The first crushing steps of two samples, one from H12 and one from MA2, led to
extremely high Ar values, which are far from an atmospheric contribution. In subsequent steps the noble
gas concentration tended towards air-equilibrated water with addition of atmospheric air.
Low temperature extraction yields high A values comparable to simple crushing. In each of
the subsequent steps A was reduced, except at 120℃ and 250℃. Between 150℃ and 200℃ an
optimum is achieved. At 50℃ a measurable water amount (≈ 0.05 µl) was released from the
calcite (Fig. 3.37). The stalagmite was pumped prior to the extraction for about 1 day and
reached a pressure of about 1·10−8 mbar. Despite this sample preparation a certain amount
of water seems to rest inside the surface-near pore space. A relatively high water amount
is released at 70℃ and at 160℃, respectively 250℃. The first maximum may be due to the
opening of larger surface-near inclusions. From 50℃ to 160℃ a trend towards increasing
water release persists. At 180℃ and 200℃ only a small number of water-filled inclusions
are opened additionally. Therefore we focused in the subsequent measurement runs on the
extraction at 150℃ which integrates over all the temperature steps from 50℃ to 150℃ and
liberates a significant fraction of the inclusion water. At higher temperatures the additional
water gain is small and furthermore the contribution of the blanks becomes significant.
The water release at 50℃ and 70℃ is an important discovery with regard to the sample
preparation. As only 70℃ is sufficient to release a rather high percentage of the included
water, it should be avoided to preheat the sample at too high temperatures. However, a
certain part of this water fraction may be related to adsorption and has to be investigated
further.
Investigating of the gas concentrations shows a similar pattern for all noble gases (Fig. 3.38).
A peak at 120℃ is followed by a minimum between 150℃ and 200℃ and a slight increase
at 250℃. The high concentrations at 120℃ are most likely due to the release of noble gases
from the copper tube walls (compare hotblanks of copper tubes, chapter 3.3.3). The increase
at 250℃ may be an artefact of the background correction, as a hotblank at 200℃ was used
for the blank correction. Higher temperatures may enable an increased desorption from the
walls.
74
Chapter 3. Working with speleothems
1,47
0.5
water amount [µl]
"Xe-signal" / water amount [1/µl]
10
crushed
0.4
0.3
0.2
0.1
0.0
50
100
150
200
250
T [˚C]
1
crushed
50
100
150
200
250
T [˚C]
1E-4
Kr(ccSTP/g)
Ne(ccSTP/g)
Figure 3.37: Xe signal per water amount for a stepwise heated sample (H12, 1.14g). The result of a
final crushing step after the stepwise heating procedure is given for comparison.
1.2x10
-5
1.0x10
-5
8.0x10
-6
6.0x10
-6
4.0x10
-6
2.0x10
-6
1E-5
0.0
8.0x10
7.0x10
-7
6.0x10
-7
5.0x10
-7
4.0x10
-7
3.0x10
-7
2.0x10
-7
1.0x10
-7
100
150
200
250
50
100
50
100
150
200
250
150
200
250
0.10
0.08
Ar (ccSTP/g)
Xe (ccSTP/g)
50
-7
0.06
0.04
0.02
0.0
50
100
150
T (˚C)
200
250
0.00
T (˚C)
Figure 3.38: Noble gas concentrations versus extraction temperature for a stepwise heated sample (H12,
1.14g). All data are background corrected with a hotblank at 100 ℃, respectively 200 ℃. A mimimum
is found between 150 ℃ and 200 ℃ indicating low contribution of noble gases from air-filled inclusions.
3.4. Separation techniques
Kr (ccSTP/g)
5.0x10
75
-6
4.0x10
-6
3.0x10
-6
H12 3step
aew
H12 stepheat
H12 Crush
H12 100 ˚C
2.0x10
-6
1.0x10
-6
H12 200 ˚C
0.0
0.00
0.01
0.02
0.03
0.04
Ar(ccSTP/g)
Figure 3.39: Kr versus Ar noble gas concentration for a stepwise heated sample (H12, 1.14g). All
data are background corrected with a hotblank at 100 ℃, respectively 200 ℃. All values plot around the
expected results of air-equilibrated water and a certain addition of air.
Neon shows the most pronounced behaviour and exhibits an additional peak at 70 ℃, which
is not found in case of the other noble gases. This can be explained by a high fraction of
gases from air-filled inclusions.
Heating may affect the noble gas pattern as the lighter noble gases show a higher diffusivity.
To investigate this effect noble gas concentrations are plotted against each other in Fig. 3.39
and Fig. 3.40. In case of the stepwise heating of an uncrushed sample (black points) no
deviation from the expected values of air-equilibrated water with addition of atmospheric air
can be detected for Kr vs. Ar. For a sample that was crushed in a first step, then heated
to 100℃ and finally to 200℃ (open squares) also no deviation was found. Similar results
were detected for Xe vs. Ne. If we take into account that the H12-samples generally yield a
small Ne-offset of about 20 %, the stepwise heating data plots on the expected line (with one
exception: thermal decrepitation at 70℃, indicated by an arrow). The three-step extraction
using crushing and two heating steps led to the same results. Significant diffusive loss of
noble gases can be ruled out, as there is no obvious excess in Kr and Xe compared to the
lighter noble gases, even after a set of 11 heating steps at different temperatures on the same
sample. Thus, in principle a stepwise heating procedure can be used to reduce sequentially
the air-contribution to the total signal.
3.4.3
Combined stepwise procedures
One of the crucial points, the separation of noble gases from air- and water-filled inclusions,
may be solved by a combination of several techniques. A combination of crushing with heating
turned out to be most promising.
Chapter 3. Working with speleothems
Xe (ccSTP/g)
76
8.0x10
-7
7.0x10
-7
6.0x10
-7
5.0x10
-7
4.0x10
-7
3.0x10
-7
2.0x10
-7
1.0x10
-7
H12 3step
aew
H12 stepheat
Crush
100˚C
-4
200˚C
0.0
0.0
-8
(3.1x10 ,2.7x10 )
5.0x10
-5
1.0x10
-4
1.5x10
-4
2.0x10
-4
Ne (ccSTP/g)
Kr(ccSTP/g)
Figure 3.40: Xe versus Ne noble gas concentrations for a stepwise heated sample (H12, 1.14g). All data
are background corrected with a hotblank at 100 ℃, respectively 200 ℃. The systematic shift towards
higher Ne-concentrations is due to an intrinsic Ne-surplus of the H12-stalagmite sample.
8.0x10
-7
6.0x10
-7
H12
CG
BU-U
BU-2
aew
5
10
100
3
4.0x10
-7
2.0x10
-7
0.0
0.0
160
200 +heat
+heat
135 35
+heat
A=0.2
200
+heat
A=0.1
2.0x10
-3
A=0.5
60
100
A=0.3
4.0x10
-3
6.0x10
-3
Ar (ccSTP/g)
Figure 3.41: Kr versus Ar concentrations for samples extracted by a combined procedure of crushing in
the steel cylinder and subsequent heating. The first step always consists of simple crushing, the following
steps either of crushing, crushing and heating or a simple heating step. Numbers refer to the total of
applied strokes. For comparison the limit for certain A is given.
3.4. Separation techniques
77
An analysis of the microwave heating pointed out that this method is not useful with regard
to the separation of air- and water-filled inclusions in a single step. However, it was not
explicitly tested in a combined stepwise procedure and is therefore not discussed further. We
will now focus on crushing in the steel cylinder and thermal heating.
Taking into account the findings with regard to the different extraction methods, a standard
stepwise procedure was developed. First, the modified crushing cylinder is heated for at least
8 hours at 150℃ to reduce the background . The sample is included at the same time, but
exposed to lower temperatures below 70℃. After the initial pumping phase a first crushing
step is applied to mainly open the air-filled inclusions. This is attempted by hitting the
sample typically 5 times with the steel ball. The gases from this step may be pumped away
as they mostly contain noble gases from the air-filled part. We also measure this fraction as
we are interested in the radiogenic He. In the subsequent step the principal water fraction is
released by crushing the sample 60 times. A higher number of strokes would be associated
with a linearly increasing background, but almost no additional water released. As the largest
air-filled inclusions are opened in the first crushing step the air/water volume ratio is reduced
in the second crushing phase. Another crushing step may be applied if the first two yield
high water values (≥ 0.5 µl). Otherwise a final heating step is applied to further reduce A.
The sample is heated for 60 minutes, respectively 120 minutes at 150℃. Meanwhile, the
water is frozen into a cold finger. As mostly the intra-granular inclusions are remaining after
the crushing, this mainly water-filled fraction will contribute to the signal. Details of this
extraction procedure as well as some results are discussed by Marx (2008). Here we give a
short summary of the data.
In Fig. 3.41 some results of samples from the stalagmites H12, CG, BU-U and BU-2 are
summarized. H12 was investigated in a 6-step procedure. In the first five steps there was
crushing only. The last step consisted of a combination of 100 additional strokes and heating
of 2h at 150℃. The CG stalagmite was extracted in six steps, whereof three are displayed.
The first data point refers to the extraction by hitting 3 times with the steel ball, followed by
the 100 strokes extraction and a combined extraction of 100 additional strokes with 2 hours
heating at 150℃ in the last step. The BU-2 sample was crushed 35 times in the first step
and 100 times more in the second step, which was combined with 2 hours heating at 150 ℃.
A similar procedure was applied to the BU-U sample, however with 60 hits in the first step.
From the Holocene H12 sample only the last four steps are displayed in Fig. 3.41, whereof
the results of the higher number of steps always refers to data points increasingly closer
to the noble gas concentrations of air-saturated water. The displayed H12 values are all
plotting on a line referring to one temperature of air-equilibrated water. Fitting the noble gas
concentrations of the last H12-measurement leads to a value of 24 ± 4 ℃, which corresponds
well with the present-day cave temperature.
The CG data shows a larger scatter and higher uncertainties, which are mostly due to the large
uncertainties of water determination in consequence of the extremely small water amounts.
However, the result of the last step, which is the CG-point with the lowest Ar concentration,
shows a strongly reduced air-contribution and therefore allows a temperature calculation with
an uncertainty of 1.8 ℃. Similarly, the BU-U and the BU-2 samples show a decreased contribution of unfractionated air to the noble gas concentration. In both cases the concentrations
of the different extraction steps are shifted parallel to the excess-air line.
In all cases the excess-air contribution could be reduced strongly, so that the last step resulted
in values which can be used for meaningful fitting. The calculated temperatures, 4.9 ± 0.6 ℃
for the 51 000 yr old BU-2, 5.6 ± 1.8 ℃ for an early Eemian BU-U piece and 24 ± 4 ℃
78
Chapter 3. Working with speleothems
for the Oman-stalagmite H12, show that it is possible to obtain reasonable temperatures
with acceptable uncertainties even from speleothems or speleothem sections with unfavorable
properties.
Unfortunately, the combined stepwise procedure was not successful in all cases. It is always
possible to reduce A by about 1 order of magnitude, which works fine for the H12 stalagmite,
but which is otherwise not sufficient if the initial A values are above 5 (like in the case of
the MA-speleothems). Furthermore, the low water amount in some steps is challenging (CG
0.15 µl at the heating step, 0.08 µl after crushing the H12 sample 5 times, 0.11 µl during
the heating step of BU-U). Thus, further development is necessary balancing the demand of
air-reduction and precise water and noble gas determination.
3.4.4
Summary
The temperature determination gets feasible, if noble gas amounts and the corresponding
water can be measured sufficiently precise and if the contribution of air is sufficiently small. To
accomplish these constraints, a combined stepwise procedure is most suitable. The air/water
volume ratio can be reduced in each step and, furthermore, the water and gas amounts are
still rather precisely measurable in the presented three-step procedure. In general a reduction
of A in the order of one magnitude can be achieved. Tests on pieces from H12, BU-U and BU2
showed the efficiency of the stepwise procedure and yielded reasonable temperature results
with acceptable uncertainties.
However, in the case of samples with a very unfavourable A (≥ 1) the stepwise procedure has
to be extended by additional crushing and heating steps. Unfortunately, not each time useful
data will be obtained as water as well as gas amounts get increasingly smaller. There, more
work has to be done and a modified procedure has to be found in order to achieve meaningful
results.
3.5
Mass spectrometry
This section is focused on the technical point of view with regard to the preparation, separation and measurement of the noble gases. The general procedure is orientated at current
methods developed for water and gas samples in a common range (about 1 ccSTP of ambient
air) as it is described by Friedrich (2007). The preparation procedure itself was developed
according to the system of Beyerle et al. (2000). A special calibration standard enables the
determination of absolute gas amounts. The reproducibility of the standard gives a measure for typical uncertainties. Finally, the sensitivity is investigated and blank values of our
extraction and measurement line are given in comparison with literature values.
3.5.1
Gas separation and purification
The sample gases are frozen into two cold traps after being released from the speleothem.
The first trap, a bare steel volume, is cooled to 25 K, which enables the freezing of the
heavy noble gases (Ar, Kr, Xe). After 20 minutes the second trap, a charcoal-trap held at
10 K, is opened to condensate Ne and retain the He. This two-step freezing procedure needs
altogether about 40 minutes to quantitatively freeze the noble gases.
Subsequently, they are released from the traps and admitted to different line volumes for
purification and splitting, if necessary. At the beginning, the charcoal trap is heated to 42 K
to release the He, but not the Ne. From this fraction a small aliquot is taken for a control
measurement in a quadrupol mass spectrometer. According to this result a splitting of the
total He amount is performed in order to obtain a He signal in a well measurable range.
3.5. Mass spectrometry
79
Before the gas inlet into the sector-field mass spectrometer (GV 5400), the He is purified by
a cold SAES AP 10N getter for 120 s and again 120 s during the inlet. To release the Ne
from the charcoal the trap temperature is raised to 90 K. The rest of the procedures is the
same as for He, except the purification time, which is increased to 180 s before the gas inlet
into the mass spectrometer and to 150 s during the inlet.
The heavy noble gases are prepared in a different way. They are released altogether from the
bare steel trap at 130 K. In this case not only Ar, Kr and Xe are admitted to the system, but
also N2 (boiling point: 77 K), O2 (90 K) and other gases like methane (111 K). Therefore,
purification becomes an important point. In a first step, the whole gas amount is admitted
to a hot SAES GP50 W2F getter for 10 minutes to remove the reactive gases. Subsequently,
the remaining fraction is processed to the cold SAES AP 10N getter for further purification.
This step lasts 240 s before the inlet into the mass spectrometer and 210 s during the inlet.
In contrast to water samples no splitting is applied; all the gas from the naked steel trap is
used.
3.5.2
Measurement sequences
This section presents the measurement procedure for a typical speleothem sample. Furthermore the changes and differences with regard to water samples are explained.
A measurement run consists of sample, calibration and blank measurements. Each of these
types is subdivided into the sample gas measurement and an according fast-calibration measurement (”fastcal”), which is performed to correct short-time electronics or ion source
changes. According to the gas release from the cryogenic traps, He, Ne and Ar-Kr-Xe are
measured subsequently. Thus, a sample, calibration or blank measurement consists of a
sample-He, fastcal He, sample Ne, fastcal Ne, sample Ar-Kr-Xe and fastcal Ar-Kr-Xe measurement.
The water samples that are usually processed on the preparation line have a weight of about
20 g, which is 20 000 times more than the water extracted from 1 g of stalagmite with a
water content of 0.1 %wt. Due to the air-inclusions of the speleothems, the signals of the
light noble gases (He, Ne) are actually only 5 000 to 10 000 times smaller. In the case of
the combined Ar-Kr-Xe measurement the difference between the signals of the speleothems
and water samples is even much smaller, because the corresponding fraction from the water
samples is split by taking a small aliquot of about 0.5 cm3 from a 2000 cm3 expansion volume
(Friedrich, 2007). This procedure was developed by Beyerle et al. (2000) in order to measure
Ar on a Faraday cup and Kr-Xe on an ion multiplier in a single measurement. For stalagmite
samples, the expansion volume is bypassed, leading to comparable gas amounts in the mass
spectrometer as in the case of water samples.
The signals of the isotopes 4 He, 20 Ne, 22 Ne, although much smaller than for water samples,
where they are measured on a Faraday detector, can still easily be measured on an ionmultiplier (MasCom SEV MC-217). Typical count rates in the case of speleothems are 103
to 104 cps for 4 He, 104 to 105 cps for 20 Ne, and 103 to 104 cps for 22 Ne. 36 Ar and 40 Ar are
measured on the Faraday cup with a 1011 Ω resistor with typical signals of 10−3 to 10−2 V for
36 Ar and 0.5 to 10 V for 40 Ar. 84 Kr and 132 Xe are measured by the multiplier with typical
count rates of 103 to 104 respectively 102 to 103 cps. In order to achieve maximum precision
and to reduce variations due to the ion source, the source tuning is normally not changed
during the whole speleothem measurement run. This provides maxiumum stability of the ion
source.
80
Chapter 3. Working with speleothems
The fast calibrations, consisting of pure He, Ne or Ar-Kr-Xe volumes, are normally performed
after the according sample and calibration measurement. Therewith we want to obtain a
background as low as possible for sample measurements. As the mass spectrometer is pumped
only 6 minutes after each measurement, a small fastcal gas amount may rest in the system
after the pumping time.
The He fast calibration is measured before the sample, because enough time is available while
the gases are frozen into the traps for separation purposes. As He is the noble gas which can
be pumped the fastest, the pumping time between fast calibration and sample is assumed to
be long enough (≥ 5 min). Blank measurements show that the influence of the residual gas
of the He fastcal is negligible.
The fast calibrations of He and Ne for speleothem measurements as well as for the diluted
standards are prepared in a special way. They are also diluted to values in the order of typical
sample or diluted standard values by using a splitting volume. The fastcal for Ar, Kr and Xe
is prepared as for water samples and the undiluted standard. It is taken from a volume with
a rather pure Ar, Kr and Xe mixture and admitted to the mass spectrometer after an initial
purification step.
The mass-spectrometric noble gas measurement starts with the gas inlet into the spectrometer. At first one or more peak centerings are performed to determine the precise peak
position. Then, baseline measurements are performed at 0.5 atomic mass units difference
to the isotope peak. Subsequently, the signal is integrated alternately at the isotope masses
determined during the peak centering. The measurement value corresponds to the baselinecorrected intercept value, which is the back-calculated signal at the time of gas inlet into the
spectrometer.
3.5.3
Calibration
The noble gas measurement is based on a comparison of the sample signal with a standard
signal. The standard signal refers to a gas volume with a known composition. For normal
water and gas samples a standard was prepared with ambient air at precisely known conditions. From this standard volume certain pipettes (0.2148 cc and 0.9989 cc) are taken to
calibrate the signals. In the whole measurement procedure the standards are handled in the
same way as the samples to achieve maximum precision and comparability.
With regard to speleothem samples the gas amounts are strongly reduced, typically to about
10 000 times smaller than for a 20 ml groundwater sample. Therefore, the normal standard
can not be used for calibration of the speleothem signals. For speleothem samples we prepared
a diluted standard by using one big pipette (0.9989 cc) of the normal air standard and
expanding it under precisely known conditions into a volume of 6.37 l. In this case we were
able to generate a calibration curve for the nonlinearity correction which expands over the
range of typical speleothem samples. The gas amounts in the case of one small pipette of the
diluted standard are shown in Table 3.5.
Processing one big pipette (about 1cc) of the diluted standard produces about 14 000 cps
for 4 He, 107 000 cps in the case of 20 Ne and 9 800 cps for 22 Ne. The Faraday measurement
of Ar yields about 2.75·10−2 V for 36 Ar and 8.2 V for 40 Ar. The heavier noble gases are
less abundant and therefore generate a smaller number of counts, of about 650 cps in the
case of 132 Xe and 16 000 cps for 84 Kr. The count rates as well as the measured voltage are
taken from run Delta (spring 2008). The values are varying slightly between different runs
(some %) according to ion source tuning or gas preparation.
3.5. Mass spectrometry
81
Table 3.5: Noble gas amounts in one small pipette (0.2148 cc) of the diluted standard before dilution
by taking aliquots.
isotope
4
He
Ne
22
Ne
21
Ne
36
Ar
40
Ar
84
Kr
132
Xe
20
gas amount in ccSTP
1.5606·10−10
4.9001·10−10
5.0001·10−11
1.4500·10−12
9.3760·10−10
2.7706·10−7
1.9352·10−11
1.0545·10−13
From the calibration data a curve for nonlinearity corrections can be determined. Calibration
signals are measured in different volume sizes, from one small (0.2148 cc) to up to 4 times a
big pipette (4 x 0.9989 cc). The fast-calibration corrected sample signal is then transferred
into absolute gas amounts using the nonlinearity curve and subsequently into a noble gas
concentration through dividing by the water amount.
3.5.4
Reproducibility and uncertainties
A measure for typical uncertainties is given by the reproducibility of the diluted standard.
All investigated noble gas isotopes (except the rare 21 Ne) show uncertainties below 2 % or
about 2% (132 Xe). Typical values are summarized in Table 3.6. The reproducibility of the
diluted standard was calculated by taking the mean of the 1-σ deviations from the fit function
accounting for the non-linearity. The shown data refer to run Beta and are representative for
the diluted standard. So far, all speleothem runs confirmed the rather high reproducibility
of the calibration data. Furthermore, it was possible to improve the results for He and Ne.
Table 3.6: Typical calibration reproducibilities for all measured noble gas isotopes in the case of the
diluted standard and the normal standard. 21 Ne has a rather high uncertainty, but is not used for noble
gas temperature calculation. The reproducibility of the diluted standard was calculated by taking the
1 σ deviations from the fit function accounting for the non-linearity. The values for the normal standard
are taken from Friedrich (2007).
isotope
4
He
Ne
22
Ne
21
Ne
36
Ar
40
Ar
84
Kr
132
Xe
20
typical uncertainties
diluted standard normal standard
1.6 %
1.3 %
1.5 %
5%
1.7%
≤ 1%
≤ 1%
2%
0.8 ± 0.7 %
0.4 ± 0.2 %
0.4 ± 0.3 %
0.8 ± 0.4 %
0.5 ± 0.4 %
0.9 ± 0.7 %
1.4 ± 0.7 %
To obtain noble gas temperatures with an uncertainty below 1℃ either the air contribution
has to be lower than A = 0.01 at a total uncertainty of 5% or, if we also want to measure
samples with a higher air contribution, the total uncertainty has to be reduced strongly. If
it is intended to calculate noble gas temperatures for a speleothem sample with A = 0.7 the
temperature uncertainty can only be reduced to 2 ℃, even if the total analytical error is about
1 %. In consideration of the calibration uncertainties given in Table 3.6 it is totally unrealistic
to reduce the total analytical error to a value below 2 %, including the sample measurement
82
Chapter 3. Working with speleothems
uncertainty, the uncertainty of blank correction, of calibration and the uncertainty of water
determination. Under really ideal conditions a total analytical error between 2 and 3 % may
be possible. In this case the limit for A is about 0.1, if a temperature uncertainty below
1 ℃ should be achieved (compare Fig. 2.7). The reproducibility of the diluted standard is
similar to the normal undiluted calibrations. In the case of Ne some improvements have been
achieved by a modified measurement procedure, which reduces the uncertainty to about 1 %
for 20 Ne and 22 Ne. However, it is unlikely that the analytical error can be reduced further.
The reproducibility of the diluted standard shows that it is possible to measure small gas
amounts (e.g., 1 small pipette of the diluted standard corresponds to 3.37 ·10−5 ccSTP of air)
with a rather high precision (≤ 2%). This implies that noble gas temperature calculations
with an uncertainty below 1 ℃ are possible for adequate samples with a low air/water-volume
ratio.
3.5.5
Sensitivity
Palaeoclimate studies to investigate rapid changes and short-time events require a temporal
resolution as high as possible. Therefore, the sample size has to be reduced and consequently
less gas is available. The limit in sample size is given by the counting statistics and the
maximum tolerable uncertainty. A high mass spectrometric sensitivity can push down the
lower limit of measurable samples.
The sensitivity of two runs in 2007, run Beta and the subsequent run Gamma, is summarized in Table 3.7. In run Gamma the ion source tuning was modified for the He and Ne
measurements. With regard to the heavy noble gases a splitting step was removed, so that
the sensitivity was strongly increased. The numbers for typical water samples at the same
system and from a high precision laser microprobe noble gas mass spectrometric system are
given for comparison (data from Böhlke and Irwin, 1992a).
Table 3.7: Mass spectrometer sensitivity. Run Beta and Run Gamma refers to a speleothem measurement series, the water samples are measured using a standard procedure as described by Friedrich (2007),
microprobe refers to a highly sensitive system of Böhlke and Irwin (1992a).
reference
run Beta
run Gamma Ar tuning
run Gamma He tuning
water samples HD
microprobe MS
sensitivity in 103 atoms per cps
He
Ne
Ar
Kr
1 600
1 600
77
150
-
665
665
65
920
-
482
69
69
1 850 000
107
2 300
280
280
3 920 000
146.6
Xe
4 000
440
440
6 580 000
156.2
Measurements using He tuning turned out to be of no use, because of high background values
in the case of He. A line blank, including only the pathway of the calibrations to the mass
spectrometer, yields already about 70 % of the total He signal caused by one small pipette.
Fluctuations in the blank values can therefore provoke unacceptably high uncertainties for
calibrations as well as for sample signals. In the case of Ne the blank contribution to the
signal of one small calibration is about 10 %, which is lower, but still too high to achieve
a confidently low uncertainty at lower sample count rates. Furthermore, the Ne signal is
strongly influenced by twofold ionized Ar atoms. The Ar value increases permanently after
closing the pumping valve due to degassing from the spectrometer walls, which is an effect
caused by the memory of larger water samples. Test series showed that 19.7 % of the Ar get
double ionized in the case of He tuning. If the ion source is changed to Ar tuning, only about
3.5. Mass spectrometry
83
1 % of the He-signal and about 3% of the Ne signal of one small calibration are due to the
line blank. Thus, in the subsequent measurements and in run Gamma the ion source was set
to Ar tuning.
0.1 µl of water-filled fluid inclusions at 10 ℃ contain about 3.5·10−13 ccSTP 123 Xe, which
corresponds to about 35·106 atoms of Xe. Thus, we can expect about 80 counts per second,
which is well measurable. Assuming a water content of 0.2 ‰ would mean that a sample size
down to 0.5 g is adequate, if the setting of run Gamma is used.
3.5.6
Blank values
Due to the small gas amounts the background control becomes very important. Several types
of blanks are determined depending on the extraction and measurement procedure.
The calibration data are corrected by a line blank which excludes the extraction part. Prior
to each sample a blank of the extraction line including the pumped sample was performed.
Sample crushing measurements are corrected with this value. Furthermore, crushing blanks
of the empty copper tubes respectively crushing cells have been measured to determine the
difference between empty untreated crushing devices and the background of the extraction
process. Additionally a hotblank was determined to investigate the influence of the background on the signal in the case of thermal decrepitation. Moving of the steel ball also
produces a certain background. Its value has been examined in an experiment using ”degassed” stalagmite powder (prepared from inclusion-poor calcite by crushing several 100 times
and heating >150 ℃ during pumping). Similarly, a blank is produced by squeezing of empty
copper tubes (chapter 3.3.3).
In Table 3.8 a summary of different measurements is given. Unfortunately, the types are
not directly comparable as they have been measured using different settings. However, it is
clearly visible that the simple crushing mechanism does have little influence on the signals in
case of the steel cylinder. Squeezing of an empty copper tube releases in contrast much more
Xe compared to the unsqueezed tube. Heating creates in all cases strongly elevated values in
the first heating step, which are extremely pronounced using copper tubes.
A comparison with the laser microprobe noble gas spectrometer of Böhlke and Irwin (1992a)
reveals that is is possible to achieve lower blanks applying an adequate preparation. Our
system yields blanks up to one order of magnitude higher.
A comparison of our hotblank values shows, that it is unrealistic to achieve useful results with
the copper tube extraction. The hotblank obtained in the steel cylinder during 2 h heating
at 150 ℃ is in general significantly lower, especially for the heavy noble gases. Systematic
experiments with varying pumping periods and heating temperatures revealed an exponentially decreasing background for increasing pumping times and temperatures (Marx, 2008).
Compared to the blank values achieved in other labotatories our values are rather good for
the light noble gases, but show some potential for improvement in the case of the heavy noble
gases (Table 3.9). The results of the ”cold” steel crusher at 70 ℃ constitute a lower limit
for the achievable blank. As this value is up to three orders smaller than the heating blank
at 150 ℃ it is possible to reduce the hotblank by special treatments prior to the measurement. So far, preheating was performed at medium temperatures of about 150 ℃ and rather
short time periods of < 24 h. Increasing the temperature and prolongating pumping time
reduces the hotblank values (Marx, 2008). This is an essential requirement for the further
developments in stepwise extraction methods as well as the sophisticated water collection
and determination.
84
Chapter 3. Working with speleothems
Table 3.8: Typical blank values of different background measurements. The microprobe system blank
refers to a highly sensitive system of Böhlke and Irwin (1992a) and is given for comparison. He, Ne, Kr
and Xe blanks are given in counts per second. Ar is given in Volt except the microprobe blank (cps).
The values of the copper tube and steel cylinder measurements are hardly comparable as they have been
determined in different runs with different settings.
type
He
copper tube
copper tube squeezing (empty)
copper tube heating (empty, 100 ℃)
steel cylinder (empty, 70 ℃)
steel cylinder (empty, 150 ℃)
steel cylinder crushing
(empty, 300 strokes)
steel cylinder crushing
(calcite, 100 strokes)
steel cylinder crushing
(calcite, 300 strokes)
microprobe system blank
20
40
Ne
84
Ar
−3
132
Kr
Xe
150
38
100-200
730
1100
410
1400
1500
2500
4·10
6.9·10−3
0.2
0.01
0.06
12
16
600
20
130
1
15
600
2
11
300
3300
0.01
20
1.3
6110
1030
0.03
80
8
17700
1600
0.08
220
24
-
-
1900 - 3600
0.3 - 0.4
0.13-0.18
Table 3.9: Typical blank values of different background measurements in ccSTP. ’hotblank Tokio’
refers to a system for geological applications (e.g., Podosek et al.1980) as well as the ’hotblank Kobe’
(Matsubara et al., 1988), which was determined at 1500 ℃. ’hotblank Heidelberg’ has been determined
at 800 ℃, respectively at 1700 ℃ and refer to data published by Trieloff et al. (2001). ’hotblank Houston’
(Copeland et al., 2007) was derived from several long duration (≥ 16 h) hotblank measurement at 400 ℃.
’blank manchester’ refers to extraction blanks described by Turner and Bannon (1992). Values are given
in ccSTP.
type
copper tube heating
(empty, 100 ℃)
steel cylinder heating
(empty, 70 ℃)
steel cylinder heating
(empty, 150 ℃)
steel cylinder crushing
(calcite, 60 strokes)
hotblank Tokio
hotblank Kobe
hotblank Heidelberg
hotblank Houston
blank Manchester
4
He
(10−10 )
20
40
Ne
−11
(10
)
Ar
(10−8 )
84
132
Kr
−12
(10
)
Xe
(10−13 )
≈ 0.5
≈1
≈ 10
≈ 200
0.01
0.08 - 0.8
0.001 - 0.01
0.001 - 0.01
0.001 - 0.01
0.7
4
3
4
2
1.7
0.5
0.31
0.29
1
1.5
8.9
210 - 230
0.3 - 1.2
-
50
0.66
-
30
0.1 - 0.01
20
0.45
0.01-0.07
-
30
1.5
-
-
3.6. Mass spectrometric procedures and data evaluation
3.5.7
85
Measurement automation
After manually extracting the noble gases from the speleothems, the whole gas separation
and measurement process is run automatically. Software and control tools (Prepline 5400)
developed by R. Friedrich and K. Träumner are used. For each type of sample (water, gas
or speleothem samples) a special script can be written, which corresponds to the specific
requirements. The scripts can open and close the valves, control the temperature of the
cryo gas traps for subsequent gas separation, check the gas amount in a quadrupol before
the inlet into the mass spectrometer as well as start and stop the measurement in the mass
spectrometer. Furthermore, the details of peak centering, the scanning sequences as well as
the time-scales of the measurement can be determined by the individual scripts.
We developed three different speleothem scripts for background, sample and calibration.
Apart from the extraction process, the gas separation and purification as well as the measurement in the mass spectrometer is completely identical for the three sample types. In
the case of the background measurement only the line part between pipettes and traps is
included, the extraction part itself is closed off. These values are used for the correction of
the calibration and are named with B for automatic data evalution.
The calibrations use the same line parts as the blanks. In this case the extraction part is
also closed off. These measurements are marked with C for the data processing with the
evaluation program Calc5400.
The sample measurements are split into two parts, although the same scripts are used for
both. Before each sample measurement a blank is made for the line, including the extraction
part and the sample. Subsequent to the blank measurement the sample is processed using
the same script. It is marked with an S with regard to automatic evaluation.
3.6
Mass spectrometric procedures and data evaluation
In this section, the measurement of the different noble gases is discussed in detail. Due to the
low signals and the strong non-linear behaviour of the electron-multiplier special routines have
been developed for the He as well as the Ne measurement. The magnet-stability is the major
problem for the measurement of the heavy noble gases. Due to hysteresis the peak moves
after jumps to other masses and therefore provokes to fail the peak maximum. Different blank
contributions require the development of a sophisticated data evaluation which is presented
at the end of this section.
3.6.1
He and Ne measurement
The signal interpretation was most difficult for He and Ne. The signals are about 10 000
times smaller as for typical water measurements. In contrast to this, the heavy noble gases
yield signals, which are comparable because a big splitting volume for water samples is not
used for stalagmites.
The measured signals include several parts: the real signal due to the sample, a memory
contribution due to the Ne memory of the mass spectrometer and a contribution of double
ionized 40 Ar in the case of 20 Ne (Fig. 3.42 c, d). For He the measured signal is composed of
the sample signal, but furthermore influenced by a contamination peak (Friedrich, 2007).
As the Ne signal is rather large in case of water samples (some V, corresponding to some
100·106 cps), a remarkable Ne memory has built up in the mass spectrometer during several
years of use. If the pumping of the mass spectrometer is stopped, the neon atoms are released
from the walls and are leading to an increasing signal for all Ne isotopes (measured: 20 Ne,
86
Chapter 3. Working with speleothems
He
a) He-tuning
b) Ar-tuning
Ne
c) signal rise
d) He-tuning
e) Ar-tuning
Figure 3.42: Signal forms of He and Ne using different tuning parameters. He tuning yields a better
ionisation efficiency but furthermore produces multiple ionisation, which affects the peaks (a,d). a) and
c) show the increasing signal due to the He, respectively Ne memory of the mass spectrometer. b)d) and
e) displays the peaks for typical sample gas amounts.
21 Ne, 22 Ne).
This background contribution is assumed to be constant in time, that means for
the same time the same number of atoms are outgassing. 20 Ne to 22 Ne ratios (s. Table 3.12),
which are not deviating extremely from atmospheric values even in the case of line blanks,
indicate that this is the dominating part of the signal background.
The contribution of double ionized 40 Ar was calculated by special measurement series. A
pure heavy noble gas composition, containing only Ar, Kr and Xe isotopes, was investigated.
Mass 20, 22 and 40 were recorded in each measurement using Ar tuning. The experiment
was repeated 10 times to achieve higher statistical significance (s. Table 3.10).
In none of the measurements a significant 22 Ne signal was visible, however a well measurable
20 Ne value. As the 20 Ne signal was about 3.9 ·10−2 V, we would expect a signal of about
Table 3.10: Measurement series for the check of double ionized 40 Ar isotopes which influences the 20 Ne
results. A pure heavy noble gas composition containing only Ar, Kr and Xe isotopes was used for this
experiment. The measurement was performed using Ar tuning.
measurement no.
10434
10435
10436
10437
10438
10439
10440
10441
10442
10443
Ar-40 (V)
2.3086
2.3428
2.3471
2.3521
3.3507
2.3505
2.3514
2.3510
2.3496
2.3480
isotope signals
mass 20 (V) mass 22 (V)
3.8446
3.9305
3.9344
3.9149
3.9454
3.9735
3.9687
3.9554
3.9568
3.9131
·10−2
·10−2
·10−2
·10−2
·10−2
·10−2
·10−2
·10−2
·10−2
·10−2
negative
negative
negative
negative
negative
negative
negative
negative
negative
negative
mass 20/40 ratio
(in %)
1.665
1.678
1.676
1.666
1.678
1.690
1.688
1.682
1.684
1.667
3.6. Mass spectrometric procedures and data evaluation
87
Table 3.11: Measurement series for the check of double ionized 40 Ar isotopes with He tuning. A pure
heavy noble gas composition containing only Ar, Kr and Xe isotopes was used for this experiment.
measurement no.
12956
12957
12958
12959
12960
12961
Ar-40 (V)
8.5891
8.4722
8.3729
8.3596
8.3156
8.3046
isotope signals
mass 20 (V) mass 22 (V)
1.7329
1.6737
1.6511
1.6332
1.6223
1.6109
negative
negative
negative
negative
negative
negative
mass 20/40 ratio
(in %)
20.17
19.76
19.72
19.54
19.51
19.40
3.9 ·10−3 V for 22 Ne. This signal is easily measurable with the Faraday cup, which exhibits
a lower limit of about 2 ·10−4 V. Thus, it can be assumed that there is no Ne in the gas.
Otherwise 10 % of the 20 Ne values should be visible at mass 22. Therefore, the signal at
mass 20 was caused by double ionized 40 Ar. Taking into consideration all 10 measurements
(1.677 ± 0.009)% of the argon get ionized twice.
The experiment was repeated with He tuning, which yields a higher electron energy for
ionisation. Instead of about 62 eV the ion-source produces electrons with about 71 eV. This
difference is sufficient to increase the amount of double ionized 40 Ar significantly to about
(19.7 ± 0.3)% (Table 3.11). We tested the double-ionisation as we wanted to measure Ne
and also He in the He tuning. This would result in a higher ionisation efficiency and enables
the 20 Ne detection with the Faraday Cup. We intended to avoid the use of the multiplier
for high count rates above some 10 000 cps as the sensitivity increases very strongly in the
first minutes at high count rates. However, the high ionisation efficiency of the He tuning
provokes also the double ionisation of memory- 40 Ar and a background at mass 4, e.g. a line
blank yields 80 000 cps for 4 He, whereas about 200 000 cps have been counted for a Fastcal
at the same isotope. The line blank accounts for 50% of the sample signal in the case of 1
small pipette of the undiluted standard and still about 9 % in the case of 2 big pipettes of
the undiluted standard. The contribution is smaller for Ne, about 10 % in the case of 1 small
pipette and 2 % in the case of 2 big pipettes. However, the varying background leads to large
uncertainties in the final gas amounts. Therefore, we rejected the measurement with the He
tuning. Additionally, the signals become more stable because the tuning of the ion-source is
not changed during the whole measurement series.
A neon measurement takes about 27 minutes if 20 Ne, 22 Ne and 21 Ne are measured. The
mass 40 value subsequent to the neon measurement was about 2.9 ·105 cps (s. Table 3.13) after
30 minutes and is rather constant. Therefore only about 4400 cps in the neon measurement
are due to the 40 Ar using the Ar tuning for 27 min. In the case of the line blanks the total
signal is about 1.5 ·104 cps at the end of the measurement. This means that the contribution
of the double ionized argon to the total signal is at most 35 %. Most samples have some
100 000 cps and more, so that the 40 Ar contribution is in the low % range and can be
corrected according to the subsequent argon measurement. Actually, this is rather difficult
and charged with high uncertainty, as we need to extrapolate the background data to the
beginning of the sample measurement. However, the double ionized 40 Ar is a constant offset
for all Ne measurements and thus no correction for the sample as well as the calibration was
performed. Furthermore, if the concentration is above 4·105 cps the argon contribution to
the neon signal is below 1 % and can therefore be neglected.
Analysis of the 20 Ne/22 Ne ratio shows, that the double-ionized 40 Ar leads to an increased
ratio in the raw data (Table 3.12). 22 Ne is affected neither by the multiple ionized argon nor
by the low number of atoms at mass 44. Therefore, the 20 Ne/22 Ne ratio is above 10, but
88
Chapter 3. Working with speleothems
Table 3.12: Neon isotope ratios in the case of different calibrations, calculated from raw data. The
size of the calibrations is comparable to that of the speleothems. The neon ratios are constant and not
dependent on the size, except the (line)blank. For comparison the 20 Ne/22 Ne ratio of blank-corrected
sample data (mean of 50 measurements) is given.
measurement
line blank
Cal 0.2 cc
Cal 0.4 cc
Cal 0.6 cc
Cal 0.8 cc
Cal 1.0 cc
Cal 2.0 cc
Cal 3.0 cc
Cal 4.0 cc
theory atmosphere
mean samples run Delta
Ne-20/Ne-22
ratio
Ne-20/Ne-21
Ne-21/Ne-22
11.20
10.38
10.34
10.32
10.26
10.27
10.30
10.30
10.33
9.80
9.74
375
376
358
369
362
365
359
360
367
338
-
33.7
35.6
35.1
35.4
34.8
34.9
34.9
34.9
34.9
34.5
-
Table 3.13: Ar and C02 background for different sample sizes. This values have been determined
subsequent to the sample-Ne measurement.
measurement
line blank
Cal 0.2 cc
Cal 0.4 cc
Cal 0.6 cc
Cal 0.8 cc
Cal 1.0 cc
Cal 2.0 cc
Cal 3.0 cc
Cal 4.0 cc
Ar-40 (105 cps)
CO2 (cps)
2.07
2.51
2.50
2.69
3.00
3.17
2.81
2.20
2.40
204
207
214
213
212
214
211
207
205
relatively constant over the whole range of the calibrations. Similarly, the 20 Ne/21 Ne ratio is
elevated, as 21 Ne is not affected by multifold ionized atoms in contrast to 20 Ne. Furthermore,
the 21 Ne/22 Ne ratio is very close to the expected values, even in the case of the line blank
and the uncorrected calibration data. Blank correction and comparison of the sample signals
with calibration signals yield the expected ratios. The mean calculated from about 50 single
measurements in run Delta resulted in a ratio of 9.74 and shows the correctness of the data
evaluation including the background correction and the calibration procedure.
In addition to the sample value, the 40 Ar background and the Ne memory, the signal is
influenced by attrition due to the ion source. As the double ionized Ar plays a minor role for
the signal evolution of 20 Ne by using Ar tuning we neglect this contribution in the following.
The signal change of 20 Ne can be described by the following differential equation:
dN
(t) = RNe − z · N (t)
dt
(3.2)
RNe describes the constant outgassing of Ne from the walls and - z · N (t) is the attrition of
Ne by the ion source. For the factor z it is assumed, that the attrition is proportional to the
3.6. Mass spectrometric procedures and data evaluation
89
amount of Ne in the system. A solution according to Bronstein et al. (2000) is:
RNe
· (1 − exp(−zt))
z
R
R
N0 − Ne · exp(−zt) + Ne
=
z
z
N (t) = N0 · exp(−zt) +
(3.3)
(3.4)
(N0 − RzNe ) · exp(−zt) is the attrition term with (N0 − RzNe ) negative for most cases (N0 ≤ 106
cps), RzNe gives the end member by balancing of attrition and outgassing, N0 is the signal
due to 20 Ne amount extracted from the sample.
The typical signal evolution in time is plotted in Figure 3.43. The evolution is strongly
dependent on the initial neon concentration N0 . If this value is higher than the limit RNe /z
≈ 106 cps then only an exponential decrease can be seen. The turning point between increase
and attrition was found by analysis of all measured data sets. Up to some 100 000 cps a
linear increasing trend is dominating, between 105 and 106 cps an inverse exponential decay is
visible and at 1.0 ·106 cps the signal is constant. Above this point the values are exponentially
decreasing.
For typical stalagmite samples the count rate is about some 100 000 counts or even lower.
Then the difference (N0 − RzNe ) is very large and the signal is strongly increasing for the first
time after the gas inlet. Therefore the equation can be simplified by series expansion for a
short time scale after the gas inlet:
R
R
N0 − Ne exp(−zt) + Ne
z
z
RNe
+ N0 −
z
RNe
+ N0 −
z
RNe
=
· (1 − (zt) − (zt)2 /2 − ...)(3.5)
z
RNe
RNe
− zt N0 −
+ ... (3.6)
=
z
z
N
= N0 + RNe · 1 − 0 · z t + ...
(3.7)
RNe
N0
NNe
·z is a very small number for small gas amounts (≤ 105 cps) as RNe is about 106 cps and z
about 10−2 s−1 . Therefore, we expect a linearly increasing count rate for a small gas amount
according to the following equation derived by simplification from the above formulas:
N (t) = N0 + RNe · t
(3.8)
Comparison with measurements confirm the theoretically derived signal evolution for small
sample gas amounts (see for example Figure 3.44).
If the signal is larger (105 to 106 cps), no longer a linear trend can be expected, because the
difference to the equilibrium value gets smaller. The slope decreases very fast and an inverse
exponential decay curve is found, as it was deduced theoretically. Several measurements have
been recorded by their peak centre values at the beginning. These plots (see Fig. 3.45) show
this behaviour quite well. According to this result the fitting routines have been adapted to
this data. Evidently, a linear fit does lead to wrong data, only the exponential curve gives
reasonable results.
Recapitulatory it can be said, that up to 100 000 cps the neon data is fitted linearly. Between
1·105 and 3 ·105 cps the last cycles showing the exponentially decreasing slope are deleted
and the rest is fitted linearly. Above 300 000 cps the data is fitted according to the inverse
exponential decay curve.
90
Chapter 3. Working with speleothems
RNe /z
signal
N0 big
~(N 0- R Ne /z) exp(-zt)
N0 small
time
Figure 3.43: Time dependent evolution of the
increase by the memory effect.
20
Ne signal due to attrition by the ion source and the
16000
14000
10474
Linear Fit of Data1_A
12000
cps
10000
8000
6000
4000
0
200
400
600
800
1000
1200
1400
1600
time (s)
Figure 3.44: Ne signal for a sample with a low count rate. The trend is totally linear, as the outgassing
of Ne due to the memory effect is strongly dominating.
3.6. Mass spectrometric procedures and data evaluation
690000
91
10544
exp decay fit
linear fit
680000
670000
cps
660000
650000
640000
630000
620000
0
200
400
600
800
1000
1200
1400
1600
time (s)
Figure 3.45: Ne signal for a sample with a high count rate. The trend is non-linear, as the signal is
influenced not only by outgassing but also by attrition from the ion source. A linear fit leads to much
too high intercept values, whereas the exponential fit can reproduce the peak center values and therefore
will give a reasonable intercept value.
The He and Ne measurements on the multiplier are furthermore complicated by its sensitivity
change. In the case of relatively high count rates (above some 10 000 cps) the sensitivity
increases rapidly if the ion-beam constantly hits the multiplier (Fig. 3.46, right diagrams).
After a certain time the sensitivity for both He and Ne stabilises. For He this takes about
100 to 120 s. To solve the problem concerning the multiplier sensitivity change, we extended
the measurement cycle of He to about 10.5 min. For small gas amounts the trend is totally
linear with a correlation coefficient R of 0.995 (Fig. 3.46, lower left diagram).
If the count rates are in a higher level and thereby influencing sensitivity, this has to be taken
into account. The first data points can not be used for fitting due to this effect. We rejected
mostly the first 10 measurement points and performed a linear fit through the remaining
points. Using this technique the intercept value of e.g. the He measurement in Fig. 3.46
(lower right diagram) can be determined with an uncertainty of 2 ‰. The maximum deviation
of the 2 big pipettes calibration from the linear trend is 2.7 % for the very first measurement
readback.
In the case of 20 Ne much larger effects are obvious. Already one small pipette of the diluted
standard is in the same signal range than He in the case of 2 big pipettes. However, for Ne the
sensitivity increase is most pronounced above 100 000 cps. In contrast, one small pipette leads
to an almost totally linear increase of the count rate (Fig. 3.46, upper left diagram). At some
100 000 cps the signal is nearly completely dominated by the multiplier sensitivity leading to
an increase of about 3.7 % in few seconds (Fig. 3.46, upper right diagram). Thus, it is not
useful to integrate over the 20 Ne-peak for a longer time, but rather changing between 20 Ne
and 22 Ne. After each peak jump the same, or at least a similar effect of increasing sensitivity
occurs. Using a minimum of 3 peak jumps a linear fits gets possible. The intercept value can
be determined with an uncertainty of less than 5 ‰ using this technique.
In all examples shown in Fig. 3.46 the signal increase due to the memory effect is visible. The
interplay with the attrition is not as important as displayed in Fig. 3.45 due to the reduction
of the measurement time and is hardly visible in the upper right diagram of Fig. 3.46.
92
Chapter 3. Working with speleothems
36000
20
228000
Ne
20
Ne
224000
cps
32000
220000
28000
24000
216000
212000
200
300
400
500
600
700
200
800
300
400
500
600
700
800
4000
30000
He
3800
29000
3600
cps
He
29500
28500
3400
28000
3200
27500
200
300
400
500
600
700
200
time after inlet (s)
300
400
500
600
700
time after inlet (s)
Figure 3.46: Ne and He signal trends for different count rates. The upper figures refer to 20 Ne, the
lower ones to 4 He. On the left side, samples with low gas amounts and on the right side measurement
results with higher signals are displayed. The linear fit in the lower right figure was performed without
the first 10 data points.
In summary, the precise determination of absolute Ne and He gas amounts is possible using a long He-measurement and a Ne-measurement consisting of at least three peak jumps.
An extension of the Ne measurement with one or two additional peak jumps and reduced
measurement iterations on the same peak may further reduce the intercept uncertainties.
However, for long measurement times the effect of attrition in combination with the memory
effect has to be taken into account.
3.6.2
Argon measurement
In general, the argon signal is significantly above the memory values (0.005 V) during the
measurement cycles for calibrations (one small pipette: ≈ 1.6 V) as well as for samples. As
the equilibrium value between attrition and outgassing for t → ∞ is approximately 0.01 V,
the outgassing of argon due to the memory does not play a role and is neglected in the
following. However, the attrition has to be taken into account. The argon amount changes
by:
dN
= −z · N
(3.9)
dt
Integration leads to the following time dependent signal:
N (t) = N0 · exp(−zt)
(3.10)
where N0 is the argon amount at inlet time and the z the attrition factor.
In Figure 3.47 a typical 40 Ar signal evolution is shown. The exponential fit has a significantly
lower uncertainty and fits better to the signal evolution. The linear fit is an approximation
which leads to an underestimation of the intercept value of about 1 % compared to the
exponential result. As the exponential behaviour is physically based and leading to a better
fit, this routine was used in run Beta. In general, the linear approximation can also be
3.6. Mass spectrometric procedures and data evaluation
93
2.30
2.25
signal (V)
2.20
raw data
linear fit
exponential fit
2.15
2.10
2.05
2.00
200
400
600
800
1000
1200
1400
time (s)
Figure 3.47: 40 Ar signal during a typical measurement. Ar shows exponentially decreasing values due
to attrition. Therefore the linear fit underestimates the gas at inlet time by ≈ 0.7 %.
Table 3.14:
40
Ar/36 Ar ratios. The calibration data are not line blank corrected.
measurement
40
Ar/36 Ar
samples run Delta
Cal 0.2 cc
Cal 0.8 cc
theory
296 ± 4
299 ± 5
299.2± 0.8
295.5
no. of samples
28
17
12
-
used if the same fit is applied to all measurements. Therefore, we evaluated our data in the
subsequent runs with a linear fit. The investigation of sample 40 Ar/36 Ar ratios proves the
argon measurements and data evaluation to be accurate (Table 3.14). The sample ratios agree
within the uncertainty with the theoretically expected values. The ratios of the calibrations
are independent of the gas amount, but more precise for the larger calibrations due to less
scattering.
3.6.3
Magnet stability and implications for data evaluation
The behaviour of the magnet is a further complication in the measurement procedure as
well as in the data evaluation. As the peaks are quite sharp, the multiplier measurement
depends strongly on the precision of the deflection magnet. If the magnet setting for the
peak measurement changes a bit, the multiplier will be off the peak. As in all measurements
(except for He) more than one isotope is recorded, the magnet has to jump between different
masses.
Magnetic fields are known to be influenced by the behaviour of the material which is used to
increase the magnitude of the magnetic field. The ferromagnetic material between the copper
coils leads to hysteresis and therefore the return point after a peak jump is not exactly the
same. The magnet is provided with a Hall probe to achieve the desired magnetic field.
However, in the first measurement cycles drifts are clearly visible despite the Hall probe
94
Chapter 3. Working with speleothems
a) Ne
c) Xe
b) Ar-40
d) Xe - 2. Time
e) Xe- 3. time
Figure 3.48: The five screen prints show typical examples for the effects of hysteresis on the data.
Evidently, a drift in the peak center position occurs (indicated by vertical black lines). The peak center
moves to lower masses in subsequent scans. a) refers to 20 Ne after a jump from mass 4, b) to 40 Ar after
a jump from mass 4 and c) d) e) to 132 Xe after a jump from mass 40.
measurements (Fig. 3.48). Peak scans show drifts for all masses, which are most pronounced
for the heavier gases after a large peak jump (Ne: mass 4 to mass 20, Xe: mass 40 to
mass 132). The deviation between the first peak scan and the last recorded cycles in Fig. 3.48
are 0.0076 amu for 20 Ne, 0.0081 amu for 40 Ar, and 0.054 amu for 132 Xe after the first jump
from mass 40 to 132. In the second and third step at Xe the drift is reduced with a value
of 0.039 amu and 0.026 amu, respectively. As it is obvious in Fig. 3.48 the peaks stabilise
at a little bit lower mass compared to the first peak scan. Thus, the measurement occurs in
general at too high masses and provokes the risk of wrong readback values. For Ne and Ar
the drift is not important as the peaks are normally rather wide. The peaks of Ne and Ar
in Fig. 3.48 refer to the outgassing due to the memory. However, the Xe values in the same
figure refer to a fastcal AKX and correspond to typical signal levels. Such large drifts can
lead to measurements at wrong positions and even missing of the peak.
To stabilise the magnet several techniques have been developed, e.g a precycling is performed
before the measurement and additionally a second peak center is done after the baseline
measurements. Despite these procedures the effect of hysteresis can not be avoided totally.
Consequently, we attempted to prevent peak jumps. A typical measurement procedure for
the heavy noble gases Ar, Kr and Xe is composed of a long Xe measurement (5 min) followed
3.6. Mass spectrometric procedures and data evaluation
95
by a Kr measurement of about 1 min. The procedure starts with Xe to get the intercept
value as precise as possible. Subsequent to the combined 36 Ar and 40 Ar measurement, the
cycle starts again with Xe and is repeated once. Finally, an additional Xe measurement
is performed to raise the precision of the curve and the intercept value. This technique is
very sensitive to the order of investigated masses. Changes in the sequence provoke that the
peaks of Kr and Xe are not found any more. Furthermore, the magnet has to be stabilized
also between one measurement of the same element (especially 132 Xe) even if no peak jump
occurs. Without precycling, 132 Xe was lost after the third or fourth recording period in the
first cycle. Therefore, the magnet current is changed slightly after each single measurement
to mimic adequate small peak jumps to antagonize hysteresis effects. Due to this method
a rather stable signal could be achieved with low variations for Ar and Kr and acceptable
values for Xe. However, the procedure has to be optimized further with regard to the 132 Xe
measurement as in some cases the Xe signal is not as stable as expected.
3.6.4
Data evaluation
The raw data is evaluated by WinCalc, which is used for discarding of outliers and fitting
the intercept values. For further data processing always the fitted intercept value including
the resulting uncertainty is used. Subsequently, the data is evaluated automatically with the
program Calc5400, which was developed by Friedrich (2007).
Firstly, the data is fastcal-corrected to remove sensitivity changes. The raw data from the
calibration SC , blank SB and sample measurement SS is divided by the dilution-corrected
fastcal signal SF∗ :
SC
SF∗
S
SEB = B
SF∗
S
SES = S∗
SF
SEC =
(3.11)
(3.12)
(3.13)
(3.14)
Then, the fastcal-corrected blank values are subtracted from the calibration as well as from
the sample data. Afterwards, the background corrected calibration values SEC - SEB are
dilution corrected, as the gas amount in the standard containers are reduced every time by
taking a calibration pipette:
SE − SEB
∗
SEC-B
= C
(3.15)
ni
i=1,2 di
di dilution factor for pipette i, ni total number of portions taken from pipette i.
The sample values are converted into absolute gas amounts by comparison with a fit function, which accounts for the nonlinearity of the fastcal-, background and dilution-corrected
∗
. Due to non-linearity effects a curve has to be established by calicalibration values SEC−B
bration measurements of different sizes, which cover the range of typical speleothem samples.
Therefore, the so-called inverse sensitivity Sensinv
c is calculated:
Sensinv
c =
Vc
∗
SEC-B
(3.16)
Vc gas amount of one pipette of the diluted standard before taking aliquots, representing the
initial value.
96
Chapter 3. Working with speleothems
The inverse sensitivity Sensinv
c , which represents a gas amount per signal, is fitted with
a suitable function f(SEC−B ). Absolute gas amounts MS for the samples, as well as for
the different extraction blanks (extraction line blank Mextr , crushing blank Mcrush ) can be
calculated from the signals using this function f(SEC−B ):
MS = SES-B · f (SEC-B )|SES-B
(3.17)
SES−B represents the fastcal- and background-corrected sample signal SES - SEB . The value
of the fitting function is taken at the position SES−B .
The program Calc5400 executes the presented steps and delivers absolute gas amounts for
the speleothem samples as well as the extraction blanks. The absolute sample gas amount
MS is corrected by the extraction blank gas amount Mextr , and depending on the extraction
procedure also by other factors, like the contribution of the crushing itself (Mcrush ). For
calculation of noble gas concentrations c the resulting absolute gas amount is divided by the
released water amount mw :
c=
MS − (Mextr + Mcrush + ...)
mw
(3.18)
The uncertainty is calculated by using the error propagation of the different contributions.
The most important part is generally the uncertainty of the water determination, which is in
the range of 2 to 3%. For small samples also the uncertainty of the noble gas measurement,
especially in the case of Xe, gets significant.
The data evaluation for speleothems is discussed in detail by Marx (2008).
3.7
Test of the measurement procedure with an artificial standard
To test extraction, measurement and data evaluation, an artificial standard with a known
noble gas content would be ideal. At first glance the glass capillaries, used for calibration of
the water amount - gas pressure curve, could provide the desired properties. A well known
amount of equilibrated water can be inserted and additionally the excess-air contribution can
be adjusted by the melting-controlled closing.
Seven samples have been prepared (Table 3.15) with an appropriate water and air content
to cover a certain range of real speleothem samples. The water was taken in glass capillaries
from a drum with equilibrated water at about 24 ℃, respectively from a small bowl which
was filled before with tap water at 25.5 ℃. The glass capillaries were prepared as for the
water vapour-pressure curves and were measured in the same way as speleothem samples.
In the three isotope plot (Fig. 3.49) all samples are approximately located on a line, even
though the water sources have been different. In general, they show a much too high noble
gas content for all gases. This can not be assigned to the fact that tap water was used, which
was probably not perfectly equilibrated with ambient air. Otherwise the samples prepared
with equilibrated water should show a different behaviour, which is not the case.
In general, He, Kr, and Xe are strongly enriched, whereas Ne and Ar are depleted relative
to the expected values (Table 3.16). The absolute offset respectively deficit decreases with
the air amount in the capillary. The highest absolute offset was found in the sample with
only 1.129 µl water and 6.52·10−3 ccSTP air, whereas the capillary with 6.13 µl water and
1.81·10−3 ccSTP air shows the lowest value. Therefore, it is unlikely that the deviations from
the expected data are caused by changes in the solubility as for example by negative pressure
3.7. Test of the measurement procedure with an artificial standard
97
Table 3.15: Artificial standard samples.The first 5 samples have been prepared with tap water from a
small bowl, the last two with equilibrated water from a large ton. Extraction ’at line’ means the sample
is prepared directly at the line of the mass spectrometer, ’external’ stands for extraction in a separate
line (Fig. 3.14), from where only the gases are transfered to the measurement line with a special sample
container after the extraction.
no.
water amount (µl)
air volume (10−3 cc)
274
275
276
281
282
330
331
1.42
4.06
2.48
1.13
3.26
6.13
6.19
6.52
3.87
5.45
6.81
4.71
1.81
1.74
extraction
at line
at line
at line
at line
at line
external
external
Table 3.16: Absolute offset (in ccSTP/g) of the artificial samples compared to the expected values
according to the inserted water and the additional air amout. Sample no.331 shows a deviation from the
overall trend, which can be assigned to the fact, that the sample container for the gas transfer was not
pumped properly before. No. 281 was affected by splitting problems.
no.
water amount (µl)
281
274
276
282
275
330
331
1.129
1.420
2.484
3.258
4.064
6.129
6.194
He (10−6 )
Ne (10−5 )
Ar (10−3 )
Kr (10−6 )
Xe (10−7 )
9.65
5.57
3.73
3.64
-1.29
1.58
-1.88
-1.39
-0.63
-0.55
-0.30
-0.22
-0.30
4.15
-10.28
-4.02
-2.99
-1.73
-0.69
-1.11
6.13
2.68
1.67
1.49
0.83
0.22
0.93
6.44
3.39
2.02
1.73
1.02
0.32
0.94
due to the capillary effect (e.g. Mercury et al., 2004). The dependency of the offset on the
enclosed air volume implies a major effect of the air. The reason for this result could not be
figured out. It is possible that the flame sealing causes a fractionation and removes especially
Ne and Ar, or that the blowpipe gas may be a source with a non-atmospheric gas composition
containing a large fraction of heavy noble gases.
Although a clear trend from atmospheric air towards air-saturated water can be detected
with an increasing amount of water in the glass capillaries, the data is far away from the
expected noble gas concentrations. Therefore, it is not possible to use them as an artificial
standard for measurement, extraction and the data evaluation process. However, the overall
measurement uncertainties of typically about 1% show the general possibility of high precision
measurements in the low-level noble gas mass spectrometry.
Modification of the flame sealing may enable the use of capillaries as artificial standards.
After the water is filled in and moved in the middle of the glass capillary, the ends are closed
with an appropriate material (e.g, rubber). In this case contact with the blowpipe gas can be
avoided during flame sealing. It may also be possible to create small steel volumes, which are
closed by squeezing and opened in the same way as copper tube diffusion samplers. Further
tests should show the applicability of these ideas. The development of an artificial standard
was not pursuited further as the data from the diluted calibration standard suggest a reliable
measurement and data evaluation procedure.
98
Chapter 3. Working with speleothems
3.0x10
-4
330
2.5x10
-4
Kr/Ar
282
2.0x10
-4
275
glascaps
air
asw
glascaps old line
expected values
276
*
1.5x10
274
281
-4
* splitting problems
1.0x10
-5
1.5x10
-5
2.0x10
-5
2.5x10
-5
3.0x10
-5
3.5x10
-5
Xe/Ar
Figure 3.49: Results of the artificial samples in the three isotope plot. On the lower end of the two
excess-air lines the theoretical values of the glass samples are plotted. The measurement no. is given for
referencing.
A similar idea was used by Böhlke and Irwin (1992a). They applied air-filled capillaries with
a special technique. One end was sealed, cooled at room temperature, then the other end was
sealed while immersed partially in water. Shorter segments have been produced by halving
the remaining parts. With this procedure they achieved results which are scattering around
the expected line of various amounts of unfractionated air. Although the number of counts
was very small in the case of 132 Xe and 84 Kr the uncertainty is impressively low. As an
additional micro standard they used a synthetic basalt glass with a known noble gas content,
which yields values similar to air. However, the variability and uncertainty in this case was
higher. Furthermore, they prepared synthetic fluid inclusions of a special optical quartz to
demonstrate that it is possible to measure noble gases in microscopic samples of trapped
fluids.
Recently precipitated calcite from the last decades may also be used as a standard sample
as well as artificially produced calcite precipitates. Speleothems can be grown in special
experiments under very well known controlled conditions. Thus, this samples may provide
the highest potential for control and calibration measurements. However, they have not been
tested with regard to NGTs and unfortunately it is not known, if they provide the required
properties as e.g. low air content and sufficiently high water amount. Laboratory experiments
may furthermore be an important source to fully understand the conditions which lead to
high water content and which also control the air/water volume ratio.
Chapter 4
Results
4.1
Cave air and dripwater measurements
The first section is focused on the cave and its environment. Deriving temperatures from
noble gas concentrations by inverse modelling requires the knowledge of the noble gas mixing
ratios in the air. In a first approach we can assume the noble gas composition of the cave
air to be simply atmospheric. However, in some cases, as e.g. in badly ventilated caves,
deviations may occur. To put this important factor on a solid scientific basis, we collected
air and dripwater samples in three caves, which were investigated in the framework of the
DFG-group Daphne and from which the most important speleothems were taken.
4.1.1
Investigation of the cave air
Noble gas temperatures are calculated from noble gas concentrations by inverse modeling
using the temperature dependence of the equilibrium concentration (as well as taking into
account the addition of air and fractionation, s. Aeschbach-Hertig et al., 2000). The equilibrium values depend at a certain temperature on the mixing ratio in the air. With regard
to the application of the method on speleothems, the composition of the cave air has to be
tested prior to the data evaluation and interpretation. If the cave is scarcely ventilated, the
noble gas mixing ratio could be changed by terrigenic or radiogenic production of some elements. To prove or revise this assumption we took air samples in two caves in the Sauerland
(51◦ 22’ N, 7◦ 40’ E).
One site, Bunker Cave, is monitored in the framework of the DFG research group DAPHNE.
Most of the later discussed speleothem samples, which could be measured with a simple
technique and which resulted in reasonable data, derive from this cave. It is situated in the
north-western part of Germany near Bochum (20 km). The surrounding geology is dominated
by Devonian Massive Limestones. Above the cave there are 10 m to 20 m of rock with few
soil (about 1 m), which is covered by small trees and scrubs. The cave air sampling points
are marked in Fig. 4.1 with white stars.
Altogether five samples have been taken at two different dates. In January 2007 two cleaned
and evacuated 100 ml stainless steel cylinders were filled inside the cave at the indicated
sampling points by opening the vacuum-valve for some minutes. Additionally, two small
copper tubes were used for collecting the cave air at the same locations. They were closed
airtight by special pliers. In November 2007, again cave air samples were taken at the same
places by using an evacuated steel cylinder. It was placed on the ground and opened for 10
to 15 minutes.
Additionally, the radon values have been monitored during the sampling in January 2007 to
100
Chapter 4. Results
Figure 4.1: Map of the Bunker cave, modified from Grebe (1993). Sampling points are indicated by
white stars. The entrance is located at the lowerleft part of the red rectangle. Subsequent a very narrow
part is following, which disables the air flow.
control the ventilation and to estimate the influence of the inflowing air. A radon monitor
(RAD 7, Durridge Company) was carried through the cave for continuous activity measurements (Fig. 4.2). The first data point indicates the activity concentration outside the cave
(≤ 40 Bq/m3 ). Passing a long narrow part (about 50 cm x 50 cm), the radon activity increases rapidly to a steady value of about 1800 Bq/m3 . Going back to the entrance area
after sampling in the inner cave chambers, the activity concentration drops after passing the
crawling passage. In the entrance area a lower, but compared to the outside air significantly
elevated Rn-level of 330 ± 30 Bq/m3 is found. As the differences between the radon concentration outside and inside the cave are large, even small mixing events with atmospheric
air or small inflow can be traced (Hakl et al., 1999, 1997, 1996). However, it is difficult to
determine Rn-fluxes or calculate the amount of inflowing air. The activity concentration
in a cave chamber Ac is determined by an advective component Fad through fractures and
passages with cross-section Sf , the diffusive flux Fdiff from the rock surface S and the decay
of radon in the chamber volume. In the case of steady state conditions this can be used for
quantitative calculations:
(Fad · Sf + Fdiff · S) − λRn · V · Ac = 0
(4.1)
The advective component consist of the influx Qin · Ain and the outflow Qout · Ac . The
ouflowing air masses are equal to the inflowing volume due to mass conservation. Thus, Fad
= Q · (Ain − Ac ) is negative and reduces the activity concentration in the cave chamber.
Two parameters are unknown, the advective volume flux Q and the diffusive flux. The
last component may be estimated from a badly ventilated cave chamber by neglecting the
advective term. Subsequently, the advection can be calculated for other cave parts in a first
order approximation by assuming the diffusive flux to be constant allover. In the case of
the Bunker Cave this calculation was not possible, as no measurement in highly unventilated
parts has been performed. Moreover, Rn measurements at different places inside the cave
can show the signal development in the case of fresh air intrusion. This can also be used for
estimation of underground airflow velocities and chamber volumes (Hakl, 1997).
4.1. Cave air and dripwater measurements
101
activity concentration (Bq/m³)
2000
1500
inner cave part
sampling points
1000
entrance area
500
outside
0
10.5
11.0
11.5
12.0
12.5
13.0
13.5
time
Figure 4.2: Radon activity concentration in different parts of the Bunker cave. The activity is plotted
against the time, as the measurement was continous. The investigation started at quarter to 11 in front
of the cave entrance. Then the monitor was carried through the cave. The high values have been detected
after a narrow part, where the activity had increased rapidly.
The noble gas air samples have been taken in the part of the cave, where the highest radon
activity was measured, so that a small contribution of ambient air can be assumed. According
to the sample type, a slightly different measurement procedure was used. The gas from
sample containers was measured by taking pipettes of about 0.41 cc. The pressure as well
as the temperature were recorded for each pipette. The copper tube samples are inserted
completely in a cracker system where they are opened under vacuum. In this case, all the gas
is transfered to the mass spectrometer and the temperature, humidity as well as the pressure
at the sampling point is used for calculation of the noble gas mixing ratio in the cave air. The
measurement results are compared to a set of laboratory air measurements using calibrated
pipette volumes of atmospheric air. The lab air pipettes reproduce very well, e.g. in run 33
(January 2007) the standard deviation is ≤ 0.4% for 4 He, 20 Ne, 22 Ne, 40 Ar and 84 Kr, and
≤ 1 % for 36 Ar and 132 Xe. The standard deviation for the cave air sample container C,
measured in the same run, is comparable. The standard deviation is about 2 % for repeated
measurements of laboratory air with copper tube samples.
The noble gas measurements of the air samples in the copper tubes show no systematic
deviation of the noble gas mixing ratios from the atmospheric values (Table 4.1). The sample
containers taken in January show a slightly elevated noble gas concentration, except for
He. The sample container C, filled with cave air in November, is closer to atmospheric
values, except for He. Calculating the mean of all Bunker Cave measurements, the noble
gas concentrations are marginally above the atmospheric values. In case of Xe its in the
range of the measurement uncertainty. The standard deviation of the mean values indicates
that at least Xe, Kr and Ar are comparable to atmospheric values. Only Ne is 2σ above
the air concentration. He seems to be a special case. All samples from the first campaign
in January did not show an excess in He. However, the value of the sample-container C is
significantly above atmospheric values, which could be confirmed by a duplicate measurement.
The elevated He concentration can be caused by a lower degree of ventilation at the time of
sampling. Unfortunately, no radon concentrations have been measured at this time.
102
Chapter 4. Results
Table 4.1: Deviation of the noble gas mixing ratios of the Bunker cave air samples compared to atmospheric standard values prepared by lab air pipettes. The precision of the noble gas measurement is
about 1 % for all noble gases except xenon, where it is about 2-3%. SD mean refers to the standard
deviation of the mean values. Dripwater refers to the mean of four dripwater measurements and indicates
the difference to the model expectation. It is given for comparison.
sample
sampling date
Coppertube sample 1
Coppertube sample 2
Sample ContainerA-I
Sample ContainerA-II
Sample ContainerB-I
Sample ContainerB-II
Sample ContainerC-I
Sample ContainerC-II
mean all
SD mean
dripwater (mean)
23.1.07
23.1.07
23.1.07
23.1.07
23.1.07
23.1.07
28.11.07
28.11.07
23.1./24.4.07
deviations from atmospheric ngc in %
∆4 He ∆20 Ne ∆40 Ar ∆84 Kr ∆132 Xe
-1.1
2.0
0.8
0.7
1.4
-2.3
0.7
-0.5
-1.5
-2.1
-0.3
3.4
2.0
2.9
6.6
-0.0
2.8
2.2
3.7
3.9
0.2
3.1
2.3
4.2
1.5
-0.4
2.2
2.2
2.8
5.1
16.9
1.7
1.7
0.9
-1.7
14.9
1.7
2.1
0.8
1.4
3.5
2.2
1.6
1.8
2.0
4.6
1.0
1.2
2.2
3.2
-1.1
-0.5
-1.6
1.1
0.6
Additionally, water samples have been taken from dripwater inside the cave. Fitting of the
obtained noble gas concentrations did not show systematic deviations from the model, which
is based on typical atmospheric mixing ratios for the equilibrium concentration. Significant
deviations of the cave air from atmospheric mixing ratios would result in systematic shifts
in the noble gas concentrations of water equilibrated with the cave atmosphere, as e.g. of
dripwater which before had been in contact with the cave air.
If we disregard the He values, we can conclude from the Ne, Ar, Kr and Xe results of the
cave air measurements and the dripwater concentrations that the air inside the Bunker Cave
has most likely atmospheric composition. Water included in speleothems from this cave
should therefore possess typical noble gas concentrations of water in equilibrium with the
atmosphere.
For comparison and as a test of the representativeness of the Bunker Cave data, a second
cave was investigated. The second site is the adjacent B7 cave (Niggemann, 2000), which
is a larger system (5100 m) with about 50 m of rock covering above the galleries where air
samples have been taken. These galleries are also characterized by a stable cave climate,
which indicates low ventilation. In the beginning of 2008, samples were taken at 4 different
places and by using two methods, filling of evacuated steel cylinders and enclosing of cave air
in copper tubes by squeezing. The sampling points are indicated in Fig. 4.3.
Unfortunately, the measurement of the largest part of the copper tube samples failed. Possibly
due to the difficult transport through the cave small leaks had been generated, which made the
measurement impossible. Furthermore, problems during sample opening and gas extraction
lead to the loss of additional samples. However, one copper tube sample in combination
with the measurement of equilibrated water from a cave pond and several dripwater samples
(chapter 4.1.2) provide reliable information about the atmosphere in different cave parts.
The Ne, Ar, Kr and Xe mixing ratios of the copper tube sample can not be distinguished
from the ambient air (Table 4.2). Similarly, the water samples show negligible deviations
from the model indicating that the dripwater as well as the the water in the cave pond was
equilibrated with air possessing typical atmospheric noble gas mixing ratios. Calculating the
mean of the deviation from the atmospheric mixing ratio for the copper tube sample as well
as the water samples, no significant values can be found for Ne, Ar and Kr. Xe seems to
have a small uncertainty, but typical measurement errors are rather 1 to 2 % and thus the
4.1. Cave air and dripwater measurements
103
KH
6 c.t.
1 container
BBH
1 c.t.
Teich
2 c.t.
1 container
Traverse
2 c.t.
Entrance
Figure 4.3: Map of the B7 cave, modified from Grebe (1998). Sampling points are indicated by arrows.
c.t. refers to air enclosed in copper tubes, container indicates sampling in 100 ml steel cylinders.
Table 4.2: Deviation of the noble gas mixing ratios of the B7 cave air samples compared to atmospheric
standard values. WS refers to water samples from the cave pond, respectively from corresponding stalactite drip sites. In these cases the numbers describe the deviation from the model. SD refers to the
standard deviation of the mean values of water and air samples.
sample
Container Teich-I
Container KH-I
C.t. Teich
WS Teich
C.t. Traverse
C.t. BBH
WS BBH I
WS BBH II
C.t. KH-I
mean B7 cave
S.D B7 cave
sampling date
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
2.2.08
deviations from atmospheric ngc in %
∆4 He ∆20 Ne ∆40 Ar ∆84 Kr ∆132 Xe
not yet measured
not yet measured
2.3
-0.0
-0.3
-0.6
1.6
2.2
-1.0
0.0
1.7
1.0
failed
failed
2.5
0.0
0.0
0.5
1.3
3.0
0.0
0.0
0.9
0.8
failed
2.5
-0.3
-0.1
0.6
1.2
0.4
0.5
0.2
1.0
0.4
104
Chapter 4. Results
Xe can hardly be distinguished from expected atmospheric values. However, He has a lower
uncertainty and perhaps shows a certain enrichment of He in this deeper and less ventilated
cave.
Several measurements of the cave air in the shallow Bunker cave and the deeper B7 cave
using different methods could prove the cave air to be atmospheric with regard to Ne, Ar,
Kr and Xe. Thus, the NGT calculation is well based on atmospheric concentrations for the
two investigated caves and is not influenced by strange cave atmospheres. This result is
additionally confirmed by the measurement of water from a large cave pond. The deviations
from the expected noble gas concentrations in equilibrium with atmospheric air are negligible.
Furthermore, dripwater measurements in the B7 cave as well as the Bunker cave yielded
similar results. However, He concentrations may be elevated, as found in the copper tube air
and the water samples from the B7 cave. Fortunately, this does not affect the temperature
calculation, but can be of importance for dating by radiogenic He.
4.1.2
Dripwater measurements
Karstified sites and especially caves constitute a unique environment as they contain calcite
precipitates which provide interesting insights into climate history. Stalagmite trace element
concentrations and also stable isotope profiles can be used to determine climatological events
on multi-annual, annual or even seasonal scale (Mattey et al., 2008). However, to combine
isotope signals with seasonal or multi-annual forcings, it is important to understand the drip
hydrology (Baldini et al., 2006) and to know how long the water needs to travel through
the overlying structure. The question of the residence time is in particular interesting for
the detection of strong short time scale events as severe droughts and extremely warm or
cold seasons. A shift of several years will be significant in interpretation. Furthermore,
information about the percolation time may be useful for water resources management.
Simple observation of the drip intervals and the drip behaviour may be misleading. Even
if an immediate response of the drip rate to precipitation changes were detected, as found
by Baker et al. (1997), the water may need significantly longer from the rain event at the
surface to the dripping point. Immediate response can be caused by changes in the height
of the water column in fissures above the drip site or the water content of the unsaturated
zone. Furthermore, threshold controlled flow can be present, which can be established by
e.g. an overflow of a cave pond above the dripping site during rain events with sufficient
precipitation. Moreover, a direct connection of the dripping site with cracks (fractured flow)
is also possible.
A better solution is a multi-tracer approach to determine the time span of water staying in
the rock cover and the overlying soil. For young waters an age estimation can be derived
by combined 3 He and tritium measurements (see chapter 2.3.3). 3 He is built up in the
flow path due to the decay of the tritium contained in the precipitation. An important
constraint for this gas-tracer dating technique is the closing-off from the atmosphere. If the
water flowing through the underground outgases, like in cracks or cavities, then only the
equilibrium component of 3 He or at least a reduced fraction will present in the water sample.
A certain amount of the surplus helium, including the radiogenic component, will be lost.
Beyond this, a possible excess air component has to be subtracted from the 3 He-signal, which
is due to the inclusion of small bubbles in the water. This correction is performed using Ne,
Ar, Kr and Xe in an inverse modelling routine (”Noble”, Peeters et al.,2003) by applying the
CE Model. As the whole data-set of noble gases is measured, also a noble gas temperature
is calculated, which can be used for fitting control.
4.1. Cave air and dripwater measurements
105
Figure 4.4: Maps from the three investigated caves. On the top Bunker Cave, below Dechen Cave and
on the bottom B7 - cave. The sampling points are marked with red stars. The pictures are adapted from
Grebe (1993, 1998) and E. Hammerschmidt.
106
Chapter 4. Results
outflow from
the stalactite
plastic hoses
funnel
valve
copper tube
Figure 4.5: Dripwater sampling in the cave. In the picture it is shown how sampling under field
conditions looks like. In the scheme below the typical setup, consisting of a copper tube, plastic hoses
and a funnel, is illustrated.
In addition to the 3 H-3 He approach, the stable oxygen isotopes as well as the tritium may
be used for dating purposes. From the stable isotope data and possibly also from tritium
values an age estimation for very young waters can be derived by comparing the seasonally
influenced precipitation data. Tritium shows a pronounced seasonal signal with low values
in winter and significantly higher values in summer (Fig.4.6) and varies in the Bunker Cave
region between 4 TU in winter and 14 TU in summer (Fig.4.7), whereas the δ 18 O values are
oscillating between -6 and -12‰ VSMOW in this region. Thus, the precipitation is imprinted
by a clear seasonal signal, which can be used for an age estimation.
Dripwater sampling technique
Similar to groundwater noble gas measurements, we have taken copper tube samples from
drip sites in three adjacent caves. As the available water amount is significantly smaller than
for groundwater wells, we reduced the sample amount to about 6 ml and used a modified
sampling technique (Fig. 4.5). Hoses can not be attached to the soda straws or other dripping
sites, because they mostly stop dripping after attaching them. Thus, we are forced to hold
the hoses and a small funnel closely below the dripping point. The water is collected until the
level in the funnel is above the dripping point to prevent or at least to reduce instantaneous
degassing. Subsequently, the water flows slowly through the attached plastic hoses, which
are smaller (inner diameter 4.8 mm) than in the case of groundwater sampling to reduce
the necessary water amount. The copper tubes are closed by clamps after flushing several
times combined with knocking for removal of attached air-bubbles. Only stalactites, which
are dripping from inside, have been used. If the water flows on the outside of the speleothem
the degassing can not be prevented and therefore too small gas ages have to be expected.
4.1. Cave air and dripwater measurements
107
Site description
In the vicinity of Bochum in North-Western Germany, two shallow and one deeper cave have
been investigated. The two shallow caves, Bunker Cave and Dechen Cave, bear about 10
to 20 m soil or rock cover. The sampling points in the adjacent deeper cave, B7 Cave, are
covered by about 50 m of rock and soil. The three caves are located 180 m asl. Today,
Bunker Cave shows a mean temperature of about 10.5 ℃, which is comparable to the values
in the two other caves. The caves are described in more detail by Niggemann (2000) and
Hammerschmidt et al. (1995). Most investigated drip sites are seasonal drips with a medium
discharge (≤ 15 ml/min). The drip site B7 BBH is a rather constant drip site with a relatively
high discharge (≥ 50ml/min) and thus belongs to the seepage flow regime. Samples have been
taken from the three caves in three different sample campaigns, from Dechen cave at 23.1.2007
and 24.4.2007, from Bunker cave at 23.1.2007 and 24.4.2007 and from B7-Cave at 25.4.2007
and 2.2.2008. Sampling places are marked in the maps with white stars (Fig. 4.4).
Results of the first sampling in January 2007
In January 2007 we made the first attempt to sample dripwater. As we have not been
confronted before with the sampling conditions in a cave and were not used to the sampling
of dripping stalactites, the sampling was rather difficult and challenging. During sampling a
certain contact with the atmosphere could not be prevented. Altogether the sampling time
for one copper tube was about 10 - 15 min corresponding to the drip rate. Therefore some
outgassing could have occurred, which may have influenced the age calculation and led to
wrong ages because of the gas loss. Additionally, small air bubbles could have been included
due to the sampling. Two samples from the campaign in January show a rather high excessair value (see Table 4.3). In this case it is likely that this is caused by inclusion of small
bubbles.
The fitting results show acceptable χ2 -values together with very low temperature uncertainties. The calculated temperatures are in the range of the mean cave values. The cave temperature is about 11 ℃ in Dechen Cave and 10.5 ℃ in Bunker Cave. The fitting results indicates
that the water has equilibrated in the soil at the mean annual soil temperature. The calculated temperature from Bunker Cave is a little bit elevated, perhaps due to re-equilibration
during the longer lasting sampling.
The tritum results are quite low, but fitting well to the tritium values in the winter precipitation (compare Fig. 4.6 and Fig. 4.7). As the samples have been taken in January, they could
therefore reflect winter precipitation of the current year, but the 3 He excess requires a longer
travel time through the soil. At the time of the first sample run the mass spectrometric measurement of 3 He was not perfect, so it can not be excluded that measurement problems have
caused deviations in the signals and therefore influenced the resulting ages. Furthermore, the
uncertainty of the 3 He measurement was high, so that the age error had risen similarly to
significant values. Therefore, it was not possible to derive a definitive conclusion from the
measured data which motivated a second sampling.
Results of the second sampling in April - 24.4 - 25.4.07
To check the results of the first campaign in January, the same drip sites have been sampled
for noble gases and tritium. In Bunker cave the drip site was dry, so we were forced to use
another place. Additionally, three samples have been taken in the much deeper and larger
cave B7. Furthermore water was filled into 40 ml glass bottles for radon analysis in the case
of sufficiently fast dripping stalactites.
108
Chapter 4. Results
Table 4.3: Data for dripwater samples. Age, T, χ2 and ∆Ne are obtained by fitting T and excess-air to
the noble gas concentrations using the CE-model (Aeschbach-Hertig et al., 1999, 2000). Rn and tritium
values are determined from additional samples. If no age is given, a 3 He deficit was existent. The noble
gas results have been corrected by an extraction dependent offset. The abbreviations are BH Bunker
Cave, DH Dechen Cave and B7 for B7 Cave. TS, BBH or names indicate the dripping site. Duplicate
samples are numbered by I or II. B7 Teich is a reference sample of a cave pond.
sample
Rn (Bq/l)
BH TS4 (23.1)
BH TS1-I (24.4)
BH TS1-II (24.4)
BH TS7 (24.4)
DH Grufthalle(23.1)
DH Grufthalle(24.4)
DH Kristallg.(23.1)
DH Kristallg.(24.4)
B7 BBH-I (25.4)
B7 BBH-II (25.4)
B7 BBH-I (2.2.08)
B7 BBH-II (2.2.08)
B7 Kerzenhalle(25.4)
B7 Teich (2.2.08)
3.3 ± 0.9
3.3 ± 1.6
5.9 ± 1.5
7.6 ± 1.7
17.5 ± 2.2
17.5 ± 2.2
-
6.3
8.4
8.4
7.4
7.4
8.6
5.0
6.7
6.0
6.0
7.2
7.2
8.4
TU
± 1.1
± 0.9
± 0.9
± 1.0
± 1.0
± 1.0
± 1.0
± 0.9
± 0.9
± 0.9
± 1.0
± 1.0
± 1.0
results of fitting with noble 90
Age (years) ∆Ne (%)
T (℃)
13 ± 14
220 12.3 ± 0.4
0.9 ± 2.0
0
9.3 ± 0.3
2.5 ± 1.8
0
9.4 ± 0.3
5.8 ± 1.9
0 12.4 ± 0.2
0 10.8 ± 0.3
4.8 ± 1.7
0 10.0 ± 0.3
0.9 ± 5.2
71 10.2 ± 0.3
0 10.1 ± 0.3
2.7 ± 2.6
9.0 10.0 ± 0.3
2.3 ± 2.7
9.1 10.3 ± 0.3
8.6
9.7 ± 0.1
2.5 ± 5.9
10.2
9.7 ± 0.1
0.1 ± 2.6
0 14.6 ± 0.3
1.9 ± 3.9
0
9.6 ± 0.1
χ2
0.2
15.1
4.5
8.1
3.6
11.2
1.1
9.4
2.4
3.8
0.2
0.3
58
17.3
The noble gas data were analysed with noble 90 and are also displayed in Table 4.3. The χ2
looks slightly worse, but this is mainly an effect of the relatively small analytical errors. The
maximum deviation from the models is in case of the B7 samples BBH-I and BBH-II smaller
than 1.4 %. The rest of the samples show deviations from the models of 3 % at maximum.
Only the sample B7-Kerzenhalle is really a badly fitted sample, Ne is 6% lower than the
model and Xe 4 % higher. Therefore this result can hardly be used further. The elevated
temperature indicates that this sample has re-equilibrated during sampling, which has taken
about 15 min or even longer and additionally, there have been few drips per minute. As
there is a degassing pattern visible (He most depleted and Xe most enriched compared to the
models, other noble gases accordingly between), gas loss during the time of sampling might
have occurred.
A second slowly dripping site - BH-TS7 has a slightly elevated temperature. In this case,
sampling took also 15 minutes, so a loss of noble gases can not be excluded. However,
there is no pronounced degassing pattern visible, the 3 He values are quite high. As the 3 He
determination via mass spectrometer was not completely stable, the resulting age has to be
handled with care.
The two duplicate samples TS1-I and TS1-II from Bunker cave are reproducing quite well.
Their mean temperature is 9.35 ℃, which is close to the mean annual air temperature in the
cave area. They did not show any excess air. Therefore, it can be assumed that the water
did equilibrate within the soil in the unsaturated zone. The calculated age is charged with a
high uncertainty. It is not possible to derive an age with the 3 H - 3 He method in this case.
However, a glance on the monthly collected tritium data from rain and dripwater (Fig. 4.7)
shows that the 3 H values may follow the typical annual cycle with some delay indicating a
residence time of about 6 months. Unfortunately, the measurement precision for the 3 H of
the dripwater is not high enough and therefore interpretation is difficult. Higher resolution
is required to better constrain the tritium results. Assuming the 3 H value from February
to be actually the tritium peak in the dripwater, the trend is similar to the stable isotopes
(Fig. 4.8).
4.1. Cave air and dripwater measurements
109
22
20
18
16
14
TU
12
10
8
6
4
2
0
1.1.2004
1.1.2003
1.1.2005
1.1.2006
Figure 4.6: Tritium measurement of monthly collected rain water from Hof (Bavaria) from 1.1.2003 to
1.1.2006, showing a typical annual cycle. High values are found in the summer rain and low activities in
winter precipitation. Samples have been provided by the German Weather Service and were measured
in the Heidelberg tritium laboratory.
10
16
tritium cave
tritium rain
14
9
10
8
TUcave
TU rain
12
8
7
6
4
6
-1
0
1
2
3
4
5
month
Figure 4.7: Seasonal tritium cycle at the Bunker Cave site. Samples were taken of rain and dripwater
on a monthly basis and were measured in the Heidelberg tritium laboratory. Month 1 refers to January
2008.
Chapter 4. Results
200
-6.0
d18 O (‰ VSMOW)
-6.5
-7.0
300
400
500
600
TS1
TS2
TS3
TS4
TS5
TS6
TS7
TS8
soil water
rain water
700
-6
-8
-7.5
-10
-8.0
d18 O rain (‰ VSMOW)
110
-12
-8.5
200
300
400
500
600
700
days (starting 1.1.2006)
Figure 4.8: Stable isotope data from rain and dripwater at/in Bunker cave. TS1 to TS8 refers to
dripping places inside the cave. Soil water and dripwater is plotted in the same diagramm (scale on the
left side). The rainwater (taken from the cave vicinity) refers to the numbers on the right. Take notice
of the different scales. The stable isotope data has been provided by C. Spötl from the University of
Innsbruck.
The results from Dechen Cave are more complex. The sample Kristallg. from 23.1.2007 has
a high excess air value, which indicates problems during sampling by inclusion of bubbles.
The calculated age seems to be a little bit high. All other samples from Dechen Cave have no
excess air and similar fitted temperatures with a mean of 10.3℃, which is about 0.5℃ higher
than the mean annual air temperature. The temperature may be elevated because Dechen
Cave is a public cave, which is illuminated by lots of spotlights. Temperature measurements
of the cave air showed a mean value of 10.6℃ (Pflitsch et al., 2000). Again the determination
of the age is difficult with the 3 H-3 He method, because of the high uncertainties in the 3 Hemeasurements and the young ages. The radon samples show activity concentrations which
are some orders of magnitude higher than it would be expected in the case of equilibrium.
7.6 kBq/m3 in water at 11℃ would require an activity concentration of 22.5 kBq/m3 in
air. Test measurements in the less ventilated Bunker Cave (Fig. 4.2) showed a maximum
activity concentration in the air of 1.8 kBq/m3 , which corresponds to about 0.6 kBq/m3 in
equilibrated water at 10℃. Thus the water stays at least some weeks in the unsaturated zone
with suppressed gas exchange. Due to similar soil cover and geological settings as in the case
of the Bunker Cave and similar 3 H-3 He results, the residence time can be estimated to at
least 6 months.
B7 Cave is a large cave with a considerable soil cover above the sampling points (50 to 60 m).
Three samples have been taken, duplicate samples from a fast dripping stalactite and one
from a very slowly dripping soda straw. As mentioned before, the fitting was bad for this
sample, which can likely be due to a degassing. The duplicate samples BBH-I and BBHII reproduce very well with similar temperatures, excess-air values and even ages. As the
sampling had taken place under ideal conditions the excess air was not caused by bubbles,
but is rather due to the water itself and therefore reflects typical excess air. As all parameters
are reliable, a mean residence time derived by the 3 H-3 He method for this dripwater is about
2.5 years. However, the uncertainty of a single measurement is rather high (2.6 and 2.7 yr).
4.1. Cave air and dripwater measurements
111
8
TS1-II TS1-I
TS7
Grufth.
24.4
Kristallg.
23.1
Tritium-He3-age (yr)
6
BBH
08
07
II
I II
Kerzenhalle
4
2
0
Bunker Cave
Dechen Cave
B7 Cave
Figure 4.9: Summarized result of the 3 H-3 He-measurements. Two samples of Dechen Cave and one
sample from B7 Cave yield a small deficit in 3 He, one sample from Bunker Cave a high uncertainty and
are therefore not displayed.
Results of the third sampling in February 2008
Again we took samples from the deeper B7 chambers for validation of the already obtained
results, three from the fast dripping stalactite in chamber BBH and some samples from a
pond (Teich) for comparison. One sample from the cave pond and two samples from the drip
site were measured. The χ2 is very well for the two dripwater samples and somewhat worse
for the cave pond water, but this being mainly an effect of the relatively small analytical
errors. The maximum deviation of Ne, Ar, Kr and Xe from the model is in this case smaller
than 1.7 %. The two dripwater samples show similar excess-air values of 8.6 % and 10.2 %
compared to the previous measurements from April 2007. As expected, the cave pond water
possesses no excess-air component and should therefore be water equilibrated with the cave
air. The fitted NGTs of all three measured samples are very similar at about 9.7 ℃. For
BBH-I (2.2.2008) no age could be calculated as the sample yielded a small deficit in 3 He.
The second sample from this site resulted in an age of about 2.5 yrs, which corresponds well
to the data obtained in April 2007. The apparent age of the single measurement from the cave
pond should be interpreted carefully as the measurement uncertainty of 3 He was quite high.
The measurement of the remaining duplicate samples can decide the question of tritiogenic
3 He in the cave pond water in future using an improved 3 He measurement procedure.
Discussion
Only six of the 13 copper tube samples of dripwater show a measurable excess-air value. The
high values of two samples (BH TS4, DH Kristallg.) taken in the first campaign indicate
inclusions of some bubbles. The rest, except multiple samples from a deep B7 cave part (B7
BBH), show no excess-air. This may be explained by the hydrological situation as there is no
typical aquifer with a saturated zone. At some points a perched layer with quasi-saturated
conditions may be present, which can be an explanation for the excess-air values from the
B7 samples (BBH samples from 25.4.2007 and 2.2.2008).
112
Chapter 4. Results
Radon measurements in the Bunker Cave showed an air activity concentration of about
1 800 Bq/m3 corresponding to an equilibrium value in water of 630 Bq/m3 at 10 ℃. Spot
tests of the dripping water revealed significantly higher activity concentrations. This implies
that the copper tube samples do not represent re-equilibrated water at least (correct sampling
provided), but furthermore preserve a large fraction of the original gas composition of the
dripwater.
All noble gas samples from the second campaign (24.4.07 and 25.4.07) contain no radiogenic
4 He, but rather show a very small deficit. However, most have a tritiogenic 3 He component
(except DH Kristallg., DH Grufthalle and B7 BBH-I from 2.2.2008). Therefore it was possible
to calculate 3 H-3 He ages from the noble gas measurements and the corresponding tritium
value. For B7 Kerzenhalle no tritium data was available, thus the same value as for B7
BBH-I was assumed.
The results for the Bunker Cave (BH) are scattering over a large scale. Two samples overlap
with young water, whereas the other two indicate ages above 1 year up to 8 years. Two of the
Dechen Cave samples could not be evaluated due to the 3 He-deficit. One sample yields a high
uncertainty and one an age between 3 and 7 years. The results from the B7-Cave are more
uniform. The multiple samples from the dripsite BBH are reproducing well and are resulting
in a mean age of 2.5 ± 0.2 years, which is reasonable if we take into account the 50 m thick
overlaying structure of rock and soil. The sample B7 Kerzenhalle agrees with these results
within the uncertainty, even though the sample seems to be partially re-equilibrated at least.
Unfortunately, the measurement uncertainty of 3 He was relatively high during the first measurements. This could be reduced in the second campaign. However, the uncertainty has to
be further reduced to achieve reliable and useful data. For constraining the results, the stable
oxygen isotopes can be used for young waters. Precipitation values are varying strongly on
a seasonal scale (s. Fig. 4.8). Isotope values of the dripwater can be compared to the seasonality of the rain water. Shifts between the two water types can be used for the estimation
of the travel time. In the case of the Bunker Cave a shift of 50 days can be found between
rain water and soil water and an additional shift of 50 to 100 days between soil water and
dripwater. Altogether, the dripwater is shifted by about 100 to 150 days compared to rain
water, which implies a travel time of about 4 to 5 months. This result is reasonable for the
shallow Bunker Cave, but only more precise 3 H-3 He data can decide, if the water needs 4 to
5 months or rather 1 year and 4 months. Niggemann (2000) found only small influence of the
seasonal rain isotopic signal on the dripwater in the B7 cave due to attenuation of the signal.
Tritium can be used as an additional tracer for the water percolation time through the aquifer
for shallow caves or short time scales. The tritium values in the rain water show a similar or
even more pronounced seasonality than the stable oxygen isotopes (Fig. 4.6). A comparison
of dripwater with rain water, as e.g. displayed in Fig. 4.7 can give an estimate for the
timescales. Assuming the highest tritium value in the dripwater (in February) to be the
seasonal peak, a similar shift between rain and dripwater than for the stable oxygen isotopes
occurs. Due to the signal attenuation the resolution for the tritium has to be improved to
<0.5 TU to achieve unambiguous results.
We tried to date the dripwater to gain important information with regard to the palaeoclimatic interpretation of speleothem tracer data. Using 3 H-3 He it was possible to derive gas
ages for dripwater from sites in three different caves. Theses gas ages have to be a lower
limit for the percolation time, as we can not expect conditions like in a normal aquifer. Degassing may occur in the mostly unsaturated parts of the aquifer. However, excess-air found
in the dripwater of a deeper cave part indicates a perched aquifer, which may provide similar
conditions as in normal groundwater sampling. Radon samples from some drip sites showed
a large excess in comparison with the cave air activity and proved, that the investigated
4.1. Cave air and dripwater measurements
113
parts are not in equilibrium with the cave atmosphere and, therefore, can be used for gas-age
estimation. As a further complication, a diffusive exchange of certain tracers with the rock
matrix is possible (Cook et al., 2005). E.g. He may be released from the quasi-immobile
water in the matrix and therefore can provoke an apparent higher concentration. Inversely,
a retention of, for instance, the stable isotopes or tritium can occur and influence the age
estimation. Therefore, a multitracer approach seems to be most suited for age determination
as diverse tracers are affected in a different extent by the interaction with the porous rock
parts. With regard to the dripwater dating, the seasonal signals of oxygen stable isotopes
and tritium can also be useful for constraining the noble gas results. They are less affected
by degassing, but are rapidly smoothed. In the shallow Bunker Cave the seasonal signal was
found to be present in the dripwater with a shift of 100 to 150 days compared to the rain. In
the deeper B7 cave it was not possible to relate the signals of drip and rain water.
Combining all results of 3 H-3 He, stable isotopes and tritium, a residence of 4 - 6 months for
Bunker Cave and Dechen Cave, and about 2.5 years for the deeper B7 cave seems to be most
likely.
In summary, dating of water from karstified aquifers seems to be feasible by the use of 3 H3 He, as e.g. shown by Yamada et al. (2008). However, the application has to be limited due
to some constraints:
• highly fractured and karstified sites enable the equilibration of the percolating water
with ambient air in fractures. Thus, a too low age has to be expected. In this case stable
isotopes or tritium may be better suited. A perched aquifer as found most likely above
the drip site BBH in the B7 Cave provides conditions for determination of reliable ages.
Shallow caves are thus commonly not well suited for the method, whereas deeper caves
may exhibit perched aquifers. However, only spot checks focused on the excess-air and
the tritiogenic 3 He can decide about the applicability.
• degassing during sampling should be minimized. This implies the improvement of
sampling with regard to the gas exchange. The sampling should not take longer than
five minutes, otherwise the degassing will make the results senseless with regard to age
determination. Thus, the drip rate should be high enough. Furthermore, it should be
aimed to apply the method also to slower dripping sites, which are also of importance
for the palaeoclimatic investigations.
• the sample size has to be adequate so that the 3 He can be measured with sufficient
precision in the range of months to few years. If the uncertainty of the 3 He measurement
is too high, the error of the age calculation will make the results irrelevant. 6 ml water
samples seem to be a good trade-off between sampling time and sampling size.
• the water flow should not be too high, rather reduced by a thick soil cover above the
cave. Residence times in the order of some weeks or months could hardly be detected
by 3 H-3 He. However, tritium data and a comparison with the annual tritium cycle can
help to estimate an age even for residence times of weeks.
114
4.2
4.2.1
Chapter 4. Results
Extraction
Sample selection
One measurement of a speleothem sample takes at least one day or more for preparation,
pumping and cleaning. It consists of 4 hours blank measurement, up to three times 4 hourlasting sample measurements due to the stepwise procedures and a certain time for data
evaluation. If the sample is not adequate for our purpose, namely calculation of noble gas
temperatures, all time spent and the effort were useless. Therefore, it is very helpful to select
in advance the best and most promising samples.
A porous stalagmite (as e.g. MA2 with mean A of 10.6) with a large amount of holes can be
excluded in advance from the measurement. Due to the large air inclusions the noble gases
from this parts will make it impossible to derive the noble gas fraction from the water-filled
inclusions with the needed precision. In contrast, compact and whitish spelothems with no
visible holes and fractures show a very low air fraction. E.g. the whitish stalagmite CG from
Cuba with large columnar crystals has a mean A of 0.9, and the the milky-whitish stalagmite
BU-U from Bunker Cave even 0.16. Thus, in a first step, samples can be chosen due to their
obvious structure and colour. Porous ones should be excluded and the whitish, compact
speleothems prepared for the next step.
Using thin sections it is possible to check the distribution of water and air-filled inclusions.
Furthermore the homogeneity of the material and the dimension of the different inclusions can
be observed. At first, surface effects due to the preparation process should be distinguished
from real structure. Using transmitted and reflected light, this differentiation is possible
and the main focus can be on the different inclusion types. Water-filled inclusions can be
identified by light reflection or bubbles inside. In general they are roundish and relatively
small, but more irregular shapes can not be excluded. Air-filled inclusions are in general
larger, more diffuse and irregular in their shape. Mostly, they appear between calcite grains,
whereas water-filled inclusions are often inside the crystals (Scheidegger et al., 2007a).
Again, very large air-filled inclusions are a good argument to exclude the according sample
from further processing. However, if very few and additionally small water-filled inclusions are
present, it is also a challenge to achieve good results. The CG stalagmite shows a low fraction
of air-derived noble gases, but as it also yields few and only µm-sized water-filled inclusions,
the released water amount is in general too small to determine meaningful temperatures.
Similarly, we found a low A-value in the case of a translucent sinter piece with very large
columnar crystals. The water content was larger (about 0.2 ‰wt), but not comparably large
as in the case of the milky whitish stalagmites BU-U and BU-1 from the Bunker cave.
The most promising samples with a small fraction of air-filled inclusions and a large water
amount seem to be speleothems with a whitish colour and columnar crystals.
4.2.2
Water determination and water content
A reproducible and precise water measurement in the µl-range could be established using the
water vapour pressure. The comparison with a calibration curve, prepared by using known
water quantities, enables determination of the water amount from the resulting pressure.
The preparation of the calibration curve as well as the determination of the fluid inclusions
follow the same procedure. After the crushing, respectively the release of a known water
amount from calibrated capillaries, the line is heated to about 100 ℃, whereas the crusher
itself is heated to about 70 ℃. The released water is frozen into a dry-ice cooled cold-finger for
20 minutes. Subsequently, it is closed off from the system and warmed to room-temperature,
which is stabilized by an air-conditioning system to (23 ± 0.5)℃. After at least waiting
40 minutes, the water vapour is admitted to different volumes (s. Fig. 3.12) for pressure
4.2. Extraction
115
Table 4.4: Uncertainty of the water determination in the case of different quantities. error (cal-...)
refers to the uncertainty, which is due to the fitting error of the calibration in the according volume.
∗
refers not to the line, but to the first expansion volume.+ corresponds to the first expansion volume,
”Error averaging” gives an impression of the typical deviations of the determined water amounts in the
different volumes.
water amount (µl)
3 µl
2 µl
1 µl
0.5 µl
0.1 µl
0.03 µl
used volume
Exp.vol 2
Exp.vol 2
Exp.vol 1/2
Exp.vol 1/2
Exp.vol 1/2/Line
Exp.vol 1/Line
error (cal-vol 2)
1.1 %
1.2 %
1.6 %
2.5 %
10%
32%+
error (cal-line)
2.3%∗
3.0%∗
5%
16 %
error (averaging)
1.0 %
1.2 %
3.4 %
20 %
determination with a compact capacitance pressure gauge (CMR 263, Pfeiffer). In each
volume it remains for five minutes until the pressure is recorded.
The overall error for different water amounts is displayed in Table 4.4. If we use all the
available expansion volumes and average the results, even water amounts down to 0.1 µl
can be determined with an error of ≈ 3%. The water amount calculated from the pressure
in the different volumes is scattering much less than would be expected from the fitting
uncertainties. We use an average value and the corresponding uncertainty which takes into
account the typical scattering. This scattering is about 1 % for water amounts larger than
0.5 µl. However, the uncertainty increases rapidly for smaller samples. Similarly, the scatter
between different volumes rises drastically and reaches 20 % at 0.03 µl. Thus, the lower limit
of water determination with total uncertainties below 10 % is at about 0.05 µl.
An interesting property of the speleothems is their water content, which is discussed more
detailed in chapter 4.2.4. As the samples are always weighed before the extraction, the water
content in ‰ of the total speleothem weight can be calculated by dividing the released water
amount by the calcite mass. A summary is given in Fig. 4.10.
A piece of the BU-1 stalagmite yields the highest value with a water content of 6.6 ‰wt.
The lowest value was measured in the case of a CG-sample and was about 0.0012 ‰wt,
which is 6 000 times less. Despite this very large difference most of the samples vary between
0.1 and 1 ‰wt. Furthermore, the water content seems to be characteristic for the various
speleothems. The milky whitish BU-1 has a mean value of 2.88 ‰wt, the whitish, but more
crystalline BU-U 0.44 ‰wt. The mean values of two soda straws and also from a sample of
BU-2 are in a similar range. They are also crystalline with a whitish colour. H12 shows a
highly layered structure consisting of an alternation of broad milky white with small brownish
layers. The thin-section analysis revealed a high number of large water- (and also air-) filled
inclusions, which is reflected in the high water content. MA2 and the flowstone as well as
the samples from the Spannagel cave and the sinter from Bunker cave range between at 0.1
to 0.2 ‰wt. They show either a more translucent crystalline structure (sinter, Spa) or an
alternation of relatively dark layers with few whitish parts (flowstone). The stalagmite with
the lowest water content shows a milky colour. However, the cut samples look very crystalline
and translucent. On thin-sections, a large number of inclusions could be found, however, only
very small and µm-sized. This may be related to the growth conditions in the cave. The
sample was taken from a chamber, which is at least 1.3 km distant from any entrance and
with some 100 metres of rock cover above (J.M. Pajón, personal communication). Therefore,
the growth was presumably less influenced by temperature and drip interval changes and also
less affected by organics in the dripwater. This may cause the growth of ideal crystals and
prevent the development of large inclusions.
116
Chapter 4. Results
mean H12
2.62
mean BU-1
2.88
BU-2
()
mean BU-U
0.44
water content (‰ wt)
1
mean flowstone
0.22
mean MA2
0.21
sinter
0.1
Spa
H12
CG
MA2
OBI5
others
BU-U
BU-1
Soda
mean CG
0.036
mean OBI5
0.047
0.01
1E-3
0
10
20
30
40
50
60
70
80
90
100
no.
Figure 4.10: Water content in ‰ of the total speleothem weight. The water content has been calculated
by summarizing the results of the crushing and heating steps.
4.2.3
Air/water volume ratio
For the determination of temperatures from noble gas concentrations in fluid inclusions it is
necessary to subtract the air-derived part from the total noble gas signal as it does not contain
any temperature information. The contribution of noble gases from the air-filled inclusions
is one of the most decisive points. If their amount is too high, i.e. the air/water volume ratio
is above 0.3 (Fig. 2.7), no meaningful results can be achieved. About 100 times more He and
between 5 to 10 times more Xe is contained in air compared to the same volume of water. If
we have an uncertainty of 1 % in the magnitude of the air-fraction the uncertainty gets much
larger for the remaining water part due to the subtraction process.
If the noble gases are extracted by a simple crushing method the results are strongly dependent on the stalagmite type and structure. The investigated samples cover a wide range of
some magnitudes (Fig. 4.11). In the worst case, the speleothems MA2 and MA1 from Chile,
an A of 42 in maximum and a mean of 10.6 was reached. In this case the air inclusions are
dominating. The H12-Oman samples have a mean of 1.8 and scatter between 0.4 and 7.8,
which is clearly lower than for the Chilean stalagmite. Already half of the mean of the H12
stalagmites is achieved with samples from Cuba. There a mean value of 0.9 with a scatter
between 0.05 and 4.1 was measured. For the BU-U stalagmite from the Bunker cave even
lower values have been achieved. The mean is 0.16 with a scatter between 0.026 and 0.44
in maximum. The lowest values have been measured for a soda straw from the same cave.
It reveals an air/water-volume ratio of 0.002, which is comparable to a typical groundwater
sample. This provides ideal conditions for the NGT calculation. Therefore, at least some
samples from the stalagmite CG and of the speleothems from Bunker cave are in a suitable range and theoretically enable meaningful NGT calculation using a simple extraction
procedure instead of a stepwise method.
4.2. Extraction
117
10
A
1
H12
CG
MA
Oman others
BU-U
Bunker cave
others
soda straw
0.1
0.01
no.
Figure 4.11: Air/water volume ratio for different stalagmite samples on a logarithmic scale. The results
show a large scatter, even for different pieces from the same stalagmite. Nevertheless, clustering of results
from certain stalagmites is visible. Only the Cuban stalagmite CG and the speleothems from Bunker
Cave (BU-U, BU-1, BU-2, soda straw) theoretically enable meaningful noble gas temperature calculation
using a simple extraction procedure instead of a stepwise method.
4.2.4
Applications
One interesting application beyond noble gas measurements on fluid inclusions may be the
quantification of the entrapped water and its spatial distribution along the growth axis.
Brook et al. (1999) approved a relationship of inclusion concentrations with stalagmite growth
and precipitation. They found inclusion-reach layers in the case of fast calcite precipitation
and inclusion-poor calcite in the case of slow growth. Genty and Quinif (1996) interpreted
couplets of dark and light laminae as annual cycles caused by differences in the water excess.
They suggested the compact type of columnar crystals to be an indicator for higher drip
rate. Frisia et al. (2000) investigated fabrics in speleothems in Alpine and Irish caves and
found relations to the growth conditions. Columnar fabrics grow during wet phases, which
implicates the speleothems to be continuously wet. Furthermore they grow in the case of
low supersaturation, from fluids at near-equilibrium conditions, which also can be related
to wetter conditions. Dentritic fabrics are related to high supersaturation and periodic dry
conditions. Fairchild et al. (2007) also stated the occurrence of couplets to be related to
changes in the dripwater composition and flow rate. The milky white part of these couplets
is supposed to contain a large number of water-filled inclusions. In the following, we will
discuss the relation of speleothem parts with a high water content with precipitation and
growth conditions based on our own data.
In the case of the BU-1 stalagmite a highly variable water content was found. The values
fluctuate between 0.6 and 6.6 ‰wt. An investigation of the calcite structures shows in the
most measured parts columnar crystals, only at two places dentritic fabrics. The sample
pieces yield a highly different amount of inclusions, less in the dentritic part and a rather
large number in the sections with columnar crystals. This has to be related to the growth
conditions, such as temperature, precipitation, CO2 -level in soil and cave, driprate and water
film on the speleothem. In the case of BU-1 temperature changes can be ruled out for
explanation, as the mean annual air temperature has not changed significantly in this region
118
Chapter 4. Results
A
A
B
C
B C
D E F G
DE FG
Figure 4.12: Water content of BU-1 in ‰ of the total speleothem weight, measured along the growth
axis (upper panel). The water content has been calculated by summarizing the results of the crushing
and heating steps. For comparison the δ 18 O-values of the Comnispa-record (Vollweiler et al., 2006) are
displayed in the lower panel. The BU-1 results (characters) are allocated to the the Comnispa record
using the ages obtained by U/Th measurements along the BU-1 growth axis.
4.2. Extraction
119
A
B
C
D
BC
D EFG
E FG
BP
mm/a
A
Figure 4.13: Water content of BU-1 in ‰ of the total speleothem weight, measured along the growth
axis (upper panel). The water content has been calculated by summarizing the results of the crushing and
heating steps. For comparison the reconstructed precipitation of a Scandinavian Pollen record (Seppä and
Birks, 2001) is displayed in the lower panel. The BU-1 results (characters) are allocated to the the pollen
record using the ages obtained by U/Th measurements along the BU-1 growth axis.
120
Chapter 4. Results
0.45
MA2 stalagmite
0.40
water content (‰ wt)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
depth (cm)
age (ka)
0
5
0.60
0.96
10
1.6
2.1
15
2.45
3.10
20
4.09
4.66
25
5.0
5.50
Figure 4.14: Water content in ‰ of the total speleothem weight, measured along the growth axis of
the Chile stalagmite MA2. The water content has been calculated by summarizing the results of the
crushing and heating steps. Note that the age scale is not linear.
over the last 8 kyr (Davis et al., 2003). Therefore, the water film on the growing stalagmite
may have played the dominating role.
In Fig. 4.12 the water content is compared to the δ 18 O-values of the Comnispa record.
This record reflects changes in precipitation and temperature. Despite the high-frequency
fluctuations it is obvious that the BU-1 water content is related to the δ 18 O-values. Periods
with more depleted δ 18 O correspond to an extremely low water content, whereas high water
content is observed in periods with less negative δ 18 O . Higher lake levels during phases with
less negative δ 18 O (Vollweiler et al., 2006) suggest this to be related to a larger amount of
precipitation.
A comparison with reconstructed data from a pollen record, derived from lake sediments in
northern Finland, shows a certain agreement. Although the amplitudes are different, the
general trend is the same. Periods of high precipitation correspond to a high water content,
lower precipitation to a reduced amount of water-filled inclusions. Consistently, the pollen
record, Comnispa δ 18 O as well as the water content in BU-1, show, that the periods around
5 kyr B.P. and 7.5 to 9 kyr B.P. are affected by a higher amount of precipitation. In contrast
the period around 6 to 7 kyr B.P. is most likely a period of reduced humidity in large parts
of Europe. In summary, the agreement of the BU-1 water content with precipitation-related
proxies indicates that the water content may be a useful tool for a qualitative estimation of
palaeo-humidity and precipitation.
As an interesting application we investigated the water content of MA2. Along the growth
axis 9 samples have been taken and measured. With exception of two extremely small pieces
(0.148 g and 0.146 g) the error in the water content is small enough to derive significant
trends. The temperature during the growth period of the stalagmite is assumed to be rather
stable. However, the precipitation may have changed considerably, as the cave is located at
the southern margin of the Westerlies in Chile.
4.2. Extraction
121
-3.5
-4.0
-4.5
d18 O
-5.0
-5.5
-6.0
-6.5
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
water content (‰)
Figure 4.15: δ 18 O values against water content of the same pieces from the stalagmite MA2. The red
line represents a linear fit of the 8 data points with a correlation coefficient R of -0.91.
Figure 4.16: 100*Mg/Ca against water content of the same pieces from the stalagmite MA2. The red
line represents a linear fit of the 6 most reliable data points with an R of -0.997, the two measurements
with large uncertainties have been discarded for fitting, but agree with the fitted line taking into account
the errors.
0.45
0.40
100*Mg/Ca
0.35
0.30
0.25
0.20
0.00
0.05
0.10
0.15
0.20
0.25
0.30
water content (‰ wt)
0.35
0.40
0.45
122
Chapter 4. Results
0.50
0.45
0.40
water content (‰ wt)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
10
20
30
40
50
60
lithic concentration (in %)
Figure 4.17: Comparison of the MA2 water content with an ENSO related marine sediment core. The
sediment core data are taken from (Rein, 2007).
The water content suggests a general trend of increasing precipitation beginning from 5 000
or 4 500 yr BP towards present day values (Fig. 4.14). Unfortunately, no pollen data or
sediment core from this region with sufficient information about the precipitation is available
to constrain this hypothesis. However, we can compare the water content with the δ 18 O value
and the Mg/Ca ratio of the same stalagmite (Fig. 4.15 and Fig. 4.16). The location of the
samples for water content measurements are somewhat uncertain (± 0.5 cm), especially in the
deepest section (25 cm from top). However, this does not affect the general trends of δ 18 O and
Mg/Ca against the water content. Both are correlated inversely with the water content. An
increasing water content corresponds to a more depleted δ 18 O value and a decreasing Mg/Ca
ratio. If we interpretate the more depleted values in the δ 18 O as a result of the amount effect
due to increased precipitation, the higher water content indicates a higher rainfall amount.
A low Mg/Ca ratio corresponds to a higher recharge into the karstic aquifer, as the calcite
precipitation in the unsaturated zone above the cave decreases during high recharge. A high
Mg/Ca ratio is an indication for low recharge (Cruz et al., 2007). A high water content in the
stalagmite is related to a low Mg/Ca ratio and vice versa. As in the case of the stable oxygen
isotopes a high water content is related to an increased precipitation. This corresponds to
our interpretation from the Bunker Cave stalagmites.
Interestingly, a certain inverse correlation with the precipitation at the northern margin of the
southern Westerlies can be found. The precipitation record of Heusser et al. (1981), derived
from pollen contained in lake sediments at 42◦ S and 73◦ W at least partially shows an anticorrelation to the suggested precipitation pattern from southern Chile. Both precipitation
records can be reconciled by a south- respectively northward shift of the Westerlies. A
northward shift would reduce the precipitation in South Chile and increase the rain fall in
the northern part and vice versa.
If we compare the water content with a laminated marine sediment core from Peru (Rein,
2007), which is a proxy for ENSO triggered flood events, a certain correlation can be found
(Fig. 4.17). The water content of the MA2 stalagmite is related to the concentration of
lithic grains in the marine sediment using the age model established by U/Th dating. High
concentrations of lithic grains are an indicator for strong El Niño events, leading to heavy
4.2. Extraction
123
11.7 ka
11.5 ka 11.8 ka
53 ka
128 ka 130 ka
12 ka
50 ka
60 ka
110 ka
130 ka
ice core depth (m)
Figure 4.18: Water content in ‰ of the total speleothem weight, measured along the growth axis of
BU-U (upper panel). The water content has been calculated by summarizing the results of the crushing
and heating steps. U-Th-dated pieces are indicated by the according age, the uncertainty is about 3 to
4 %. For comparison the δ D values of the Vostok ice-core are displayed (Petit et al., 1999) in the lower
panel. Corresponding parts are marked in colour.
precipitation and floods in Peru. If we assume high water content in the stalagmite to be
related with high precipitation, then the periods with stronger El Niño events are connected
to decreased precipitation in southern Chile. Vice versa, low El Niño activity is related to
stronger precipitation in the southern Chile region. This may be induced by changed circulation patterns due to the ENSO phenomenon (Kilian et al., 2008). Furthermore, sediment
cores from a Bolivian lake show a pronounced dry period between 6.2 and 2.3 kyr BP and
than an increase in precipitation towards present day values (Abbott et al., 2000). This may
indicate general drier periods during the mid Holocene as Baker et al. (2001) found also
corresponding dry periods in tropical South America. Clement et al. (2000) claim largely
arid coastal regions in the mid-Holocene to be caused by a relatively cool eastern equatorial
Pacific. However, the relationships between the different forcings as well as the influences
on the circulation patterns are highly controversial. Thus, it is not possible to constrain the
interpretation of the water content of MA2 definitively. Based on the common interpretation
of the Mg/Ca ratios, and the δ 18 O signal in terms of the amount effect, the water content is
closely related to precipitation. A likely interpretation of speleothem sections with increasing
water content suggest increasing recharge and precipitation.
124
Chapter 4. Results
The stalagmite BU-U expands over some periods of the last 130 kyr in its top 75 mm. We
measured 13 pieces along the growth axis. The calculated water content varies in the order
of one magnitude from about 0.10 to 0.88 ‰wt. The first five data points belong to the
early Holocene (10.5 to 12.0 kyr), which corresponds to the fastest growth period of the
whole stalagmite (top 4 cm). The mean water content of the first three samples is about
0.31 ‰wt. The water content of the two subsequent pieces, which were extracted from a
more translucent layer, is low and about 0.1 ‰wt. The next growth period of about 2 cm
corresponds to a warmer interstadial (MIS 3, D/O-events 14-17, compare Dansgaard et al.,
1993) at about 53 kyr BP. The interstadial at 53 kyr is reflected by a higher value in the
water content and a denser calcite. This corresponds also to the fact, that other stalagmites
(BU-2) show rapid growth during this phase, whereas no growth is found in other periods
during the Glaciation. The next part with low water content is a rather translucent layer
of about 8 mm. Unfortunately, no U-Th-dating is available for this period. Subsequent to
the MIS3 period in BU-U a dark brownish layer with macroscopic detritus particles (clay
minerals) is visible, which indicates a hiatus. This period may refer to MIS 4 and MIS 5a-d
from 60 to 120 kyr BP. The following larger growth layer refers to MIS 5e and is milky white
with one small brownish layer. The uppermost sample dated in this larger growth layer was
found to be about 128 kyr old. Below the dark layer of MIS4 the water content rises rapidly
to 0.67 ‰wt and peaks finally at 130 kyr BP with 0.88 ‰wt. The older samples then show a
decline in the water content. The layer from MIS 5e is followed by an extremely dark brown
layer (undated).
A comparison of the water content with the Vostok ice-core (Petit et al., 1999) shows an
astonishing agreement of the general trend. The warm periods with less depletion in δD
compares with the periods with a high water content in the stalagmite. In contrast, colder
phases in the ice core with strong depletion in δD correspond to slowly growing parts with
low water content. The most striking feature is the good replication of parts of the Eemian
peak in the water content. In the case of BU-U the water content seems to be closely
related to temperature. However, warmer periods may additionally be connected with a
more humid climate and thus the primary factor influencing the water content is most likely
the precipitation. In consideration of the BU-1 and especially the MA2 results, the water
content seems to be related to precipitation. In the following, we discuss the empirically
found correlation on a more theoretical approach.
4.2.5
Theory of fluid inclusion origin and frequency
In the precedent section some examples of water content curves along the growth axis have
been shown. They reveal correlations with precipitation and possibly also with temperature.
To use this interesting information in a less speculative way, it is necessary to know the origin
of this feature.
Lamination with an alternation between inclusion-poor and inclusion-rich layers are a common feature (Mattey et al., 2008; Frisia et al., 2003; Brook et al., 1999). However, theories about the origin are scarce (Frisia et al., 2003; Genty and Quinif, 1996; Kendall and
Broughton, 1978). Kendall and Broughton (1978) suggest changes in the environmental conditions, especially in the drip rate, to be responsible for the layering. Similarly, Fairchild et al.
(2007) and Genty and Quinif (1996) interpreted couplets of dark and light laminae as annual
or multi-annual cycles caused by differences in the water excess or the dripwater composition.
By review of the literature and our own investigation and experiences, a hypothesis of fluid
inclusion origin was developed (s. schematic drawing on page 125).
Three main environmental conditions can influence the amount of inclusions: precipitation,
temperature and the CO2 -level of soil and cave. High soil-CO2 values lead to a high amount
4.2. Extraction
125
Precipitation
Temperature
CO2-Level
soil and cave
CO2supersaturation
drip rate
water film
thickness
thick
growth
rate
gas bubbles
thickholes stay
water-filled
impurities
highproduce
cavities
lateral
coalescence
partial
enhanced
many
water-filled
inclusions
thin
evaporationdominated
complete
few
water-filled
inclusions
Figure 4.19: Factors influencing the amount of water-filled inclusions.
126
Chapter 4. Results
of H2 CO3 (see equation 2.1), which in turn increases dissolution of calcium carbonate from
the soil and the host rock. The CO2 -level in the cave influences the precipitation of calcite.
A rather low cave-CO2 -value induces a higher supersaturation of the dripwater, which favors
the calcite precipitation and thus speeds up the growth. A higher growth rate leads to an
elevated amount of partial lateral coalescence. In the case of partial lateral coalescence the
elongated imperfections are fluid inclusions and may be water-filled (Kendall and Broughton,
1978).
The growth-rate is also influenced by temperature. Higher temperatures lead to changes in
the chemical reactions and reduce the solubility of CO2 . A faster CO2 degassing in turn
accelerates the calcite precipitation. Therefore, the faster growing speleothem is expected to
yield more imperfections due to the partial lateral coalescence of the single crystals resulting
in a larger number of fluid inclusions.
The growth rate can also be changed by the precipitation, which modulates the dripwater
supply. A constant drip site, which does not dry out over the year, increases the growth,
compared to a seasonally dripping site. In contrast, slower growth favours complete lateral
coalescence, which results in inclusion-poor calcite.
Precipitation is supposed to have the largest influence on fluid-inclusion occurrence. The
precipitation governs the drip rate, which directly controls the crystal growth by the variation
in the lateral coalesence, and changes also the water film thickness as well as the amount of
impurities washed into the cave. Kendall and Broughton (1978) also speculate the water
film thickness to have an important influence on the crystal growth. Low drip rate and a
thin water film leads to enhanced drying of the speleothem surface. Therefore, the cavities,
which are occurring less often due to a lower growth rate and complete lateral coalescence,
are most likely not water-filled. In contrast, a thicker water film enables the cavities to be
water-filled all the time and thus impedes the appearance of only air-filled inclusions. By
splashing, tiny air bubbles may form, which can adhere on the growing speleothem surface and
disturb the growth (Kendall and Broughton, 1978). Taylor and Chafetz (2004) also detected
holes presumably provoked by adhesion of bubbles. The occurring spherical cavities may
be (partially) water-filled later on. Another reason for an elevated number of small waterfilled inclusions could be impurities in the dripwater. Strong precipitation and enhanced
water flow in the karstic aquifer can transport organics as well as tiny minerals. Deposited
on the speleothem surface they can influence the growth and impede or slow down crystal
development at the affected parts. Possibly, they are the explanation for skeleton-like features
found in the Bunker Cave (Fig. 2.6). In contrast, laboratory experiments with similar growth
conditions, but clean dripwater resulted in perfect crystals (Fig.2.3). The impurity derived
cavities may be water-filled and overgrown. This hypothesis can be verified by the occurrence
of a high number of very small inclusions in the µm-range (BU-U stalagmite, s. Fig. 3.5) ,
which fits to the size and distribution of holes in the SEM image from Bunker Cave (Fig. 2.6).
The argumentation can be turned around for the palaeoclimatic interpretation of the speleothem water content. Inclusion-poor layers are related to complete lateral coalescence or a
thinner water film. This may be related to a smaller growth and also to a reduced drip rate.
Finally, a reduced precipitation or lower temperatures can account for this feature.
To disentangle the different effects, temperature estimates from independent proxies or from
NGT-determination on speleothems as well as the growth rates, derived from an age-model,
established by U-Th dating or 14 C-measurements, can be used. Furthermore, thin-sections
can give an insight into the crystal growth history and help to decide if lateral coalescence
or other effects caused the existence of the water-filled fluid inclusions.
4.3. Separation techniques
127
The measurements on BU-U expand over more than 130 kyr. Therefore, temperature as well
as precipitation changes may be responsible for the observed water content variations. In
this case it is not possible to separate the effects. Most likely a warmer climate is related
to elevated precipitation. Therefore the inclusion-rich parts are related to a rather fast
growth, whereas the inclusion-poor layers, reflecting the cold glacial periods, hardly exhibit
any growth (some mm in 30-40 kyr).
The Holocene stalagmite BU-1 is not supposed to reflect significant temperature changes
(compare Davis et al., 2003). However, we revealed high variations in the water content.
In retrospect to the presented theory this can be explained by precipitation effects, whereas
the larger water film thickness, the higher drip rate and possibly the elevated content of
washed out organics have played the major role. Elongated ellipsoidal inclusions as indicators
for partial coalescence have been found in nearly all sections of the stalagmite, however
in different magnitudes. Despite the difficulties to calculate a growth rate from the age
stratigraphy, parts of higher water content coincide mostly with a higher growth rate. In
turn, this reflects the water supply and finally the precipitation, as under drier conditions the
stalagmite surface dries out and the growth slows down. Similarly, Genty and Quinif (1996)
detected a high correlation of laminae thickness with water excess and thus precipitation.
The variation in the water content for the MA2 stalagmite are smaller, but significant. As
again temperature changes are not supposed to be that large in the last 6 kyr, the changes
in water content have to reflect variations in the precipitation pattern. Correlations with
the stable isotopes suggest an increasing water content to be related to an also increased
precipitation. Although the cave is located in a super-humid climate (3 500 - 5 000 mm
per year) changes in precipitation are influencing the content of water-filled inclusions, most
likely by the increased input of detritus or potentially by bubble influenced cavity formation. The latter factor can be of importance especially for this stalagmite as we found in
general an extremely high amount of air-filled inclusions. The amount of air-filled inclusions
decreased with a higher water content perhaps due to enhanced water flow removing the
bubbles partially. The growth rate did not change substantially over the whole stalagmite
growth period.
In summary, the comparison of the water content in BU-U, BU-1 and MA-2 with other proxy
data from either the same stalagmites or from sediment and ice cores has shown that this
value can deliver precious information about the past precipitation regime. A higher water
content is correlated in all investigated samples with a higher precipitation. This can be
explained by theoretical considerations about the growth conditions.
4.3
Separation techniques
Noble gases from air-filled inclusions can contribute significantly to the total signal. E.g., in
air about 100 times more Ne is abundant compared to the same volume of water. As the
temperature information is only contained in the water-derived component, it is necessary
to subtract the noble gases of the air-filled inclusions from the total signal. In the case of a
significant amount of air-filled inclusions (A >0.3), even a small uncertainty in the air-water
volume ratio can cause a large error in the remaining noble gas amount after subtraction. This
part is used for temperature determination and thus can only deliver meaningful temperatures
if the air contribution is sufficiently small (A <0.3).
Unfortunately, only a small number of stalagmites fulfil this criterion in a simple extraction
procedure consisting of a single step (compare Fig. 4.11). Stepwise procedures may overcome
this problem by subsequent opening of different inclusions. The idea is based on findings from
128
Chapter 4. Results
200 µm
Figure 4.20: Thin-section of a piece from the H12 stalagmite. Between the calcite crystals large airfilled inclusions are visible. The inclusions within the crystals are significantly smaller and are supposed
to be mostly water-filled.
thin section analysis (Fig. 4.20). Large air-filled inclusions are located between the calcite
crystals. Inside the grains the inclusions are in general smaller (some µm in diameter) and
predominantly water-filled. Badertscher (2007) conducted intensive studies on fluid inclusion
characterization. She found the large inclusions between the crystals to be always air-filled
and water-filled inclusions to appear only inside the crystals.
Using this knowledge we developed special stepwise procedures which consist of crushing,
heating or a combination of both extraction methods. The first crushing predominantly
opens the air-filled parts, whereas the subsequent steps increasingly extract the gases from
water-filled inclusions. The results of stepwise crushing experiments are displayed in Fig.4.21.
In all cases A is reduced in subsequent steps; if crushing only is applied a linear trend on
the double-logarithmic scale is visible. Crushing up to 200 times can reduce A by a factor
of 3. A further heating step enables additional reduction of the air contribution. In summary,
the air-water volume ratio can be reduced by about 1 order of magnitude by application of
several crushing steps in combination with a final heating step.
These results indicate that the first crushing opens actually the largest (air-filled) inclusions.
Studies at ETH Zurich confirm this finding. They achieved also good results with a stepwise procedure consisting of a precrushing step followed by a main-extraction using thermal
decrepitation (Scheidegger et al., 2008).
The stepwise procedure may have influences on the obtained noble gas pattern. Mechanical
crushing may produce micro-fractures, which enable exchange with the atmosphere. In this
case it is supposed that especially the light and mobile noble gases He and Ne would be
affected. However, in some cases we found e.g. an excess in Xe in the first crushing step
of samples from BU-1 (Fig. 4.22). This is in contradiction to the expectations and can
not be explained by the extraction method. The sample is crushed few times with the
steel ball and all the gases from this step are transferred to the cryo traps and further to
the mass spectrometer. If micro-fractures had arisen from the crushing, a clear Ne-excess
should be visible. On the other hand, the treatment prior to the extraction can have an
influence. Due to the sampling of the stalagmite, subsequent transport and the cutting of
4.3. Separation techniques
A [V air /V water ]
10
129
H12
MA
CG
1
0.1
1
10
100
number of strokes
Figure 4.21: Development of the air-water volume ratio A using stepwise extraction procedures. A is
plotted against the total number of strokes. The values at 200 strokes are achieved by a combination
with heating at 150 ℃. The data point of MA2 at the highest stroke number refers to a pure heating
step.
small pieces, small micro-fractures may emerge. Possibly they enable the exchange of the
inter-granular inclusions with the atmosphere and a certain re-equilibration. The heavy noble
gases as Xe are suppressed in the exchange due to their larger atomic diameter (2.18·10−12 m,
He: 1.28·10−12 m) and diminished kinetic behaviour. Furthermore the Ne may get lost in the
initial pumping step, whereas the heavier noble gases stay mainly inside the calcite. This may
be an explanation for the deviations in the Xe-Ne plot towards Xe-excess. In the subsequent
crushing steps only smaller inclusions inside the calcite crystals are opened which are hardly
affected by the precedent extraction. Therefore, the noble gas concentrations are within the
expected range and follow the line of decreasing air-addition towards equilibrated water at a
certain temperature for all subsequent extraction steps beside the first crushing. This can be
observed in the shown examples of BU-1 for the 60 hits step towards the combined 110 hits and
heating step (Fig. 4.22). In stepwise procedures generally the noble gas concentrations evolve
along the line of air-equilibrated water with a constantly decreasing addition of unfractionated
air in higher extraction steps (Fig. 3.41). In the very first steps the mentioned deviation
may occur. Thus, the stepwise crushing under vacuum may help to overcome not only the
problem of high air addition, but also to reduce and minimize influences of micro-fractures.
The Xe excess may also be explained by another reason. Adsorption is more effective for
heavier gases. Thus, the Xe excess in the first step can also be due to removal of superficially
adsorbed Xe. As we observed also Ar excess in the first step this assumption alone is not able
to explain the noble gas pattern completely. A more detailed discussion about the deviations
from expected values is given in the following subsection.
Another important point concerns the released water amount. For meaningful temperature
determination from noble gas concentrations a water amount with small errors (<5 %) is
essential. This can not be achieved with a liberated water amount <0.1 µl. Measurements
on a very inclusion-rich stalagmite (BU-1) showed a decreasing trend in the released water
130
Chapter 4. Results
1.0x10
8.0x10
3.0x10
-8
2.5x10
-8
2.0x10
-8
1.5x10
-8
1.0x10
-8
-7
0 ˚C
-8
5 ˚C
Xe (ccSTP/g)
10 ˚C
6.0x10
-8
0.0
4.0x10
-8
2.0x10
-8
0.0
0.0
5.0x10
-7
1.0x10
-6
1.5x10
-6
2.0x10
-6
2.5x10
-6
5 hits
60 hits
110 hits & heat
aew
2.0x10
-6
4.0x10
-6
6.0x10
-6
8.0x10
-6
1.0x10
-5
1.2x10
-5
Ne (ccSTP/g)
Figure 4.22: Reduction of the air-addition by stepwise methods, shown exemplary for the BU-1 stalagmite. Each sample is first crushed 5 times, than 60 times and finally 110 times in combination with 2
hours heating at 150 ℃. The insert shows the result for the last step. The heating measurements are not
hotblank-corrected.
amount per step. The strongest decline can be found from 60 to 100 hits. Thus, with the
first 60 strokes most inclusions, at least the larger ones are opened. The remaining µmsized inclusions can not provide such large amounts, even if they are more abundant than
the big inclusions. A stalagmite with medium high water content (H12) did not show a
significant trend. The amount of released water scatters mostly between 0.1 and 0.3 µl in
one step, which is in an acceptable range for water determination. A different behaviour was
found for the CG stalagmite. There, we only have few and in addition very small inclusions.
They are opened primarily after the stalagmite is milled to fine powder and especially by
thermal decrepitation. Thus, the optimum procedure depends on the stalagmite type and its
inclusions.
In summary, the stepwise procedure enables the reduction of the air contribution to the total
noble gas signal. A moderate number of strokes combined with a heating step yield in general
enough water at the end of the stepwise extraction for precise water determination and finally
the calculation of temperatures from noble gas concentrations.
4.4
Noble gas fractionation and enrichment
As mentioned in the precedent section, some speleothems show deviations from the expected
atmospheric mixing ratio. Xe can be enriched in the first crushing step (Fig. 4.22), which may
be caused either by the differential pumping due to micro-fractures (Ne most affected) or the
release of superficially adsorbed Xe. Furthermore, extremely high values of Ar are sometimes
found in the first crushing step (Fig. 4.24). This can neither be explained by adsorption
(the deviation should tend towards Kr) nor by differential pumping (Ar should be depleted).
4.4. Noble gas fractionation and enrichment
131
H12-runGamma
H12-runDelta
BU-1
CG
* heating only
+ additional heating
3.5
3.0
2.5
2.0
1.5
1.0
released water [µl]
0.5
*
0.4
0.3
+
+
+
+
0.2
0.1
+
*
+
*
0.0
1
10
100
number of strokes
Figure 4.23: Water amounts released during the stepwise extraction. The samples at 110 and 200
strokes are additionally heated for 2h at 150 ℃.
Even a combination of both effects can hardly account for the observed signal. Turner and
Bannon (1992) also found excess Ar in quartz and fluorite samples and suggested dissolution
of adsorbed atmospheric gases from sediments through which the fluids flow as a possible
source. Furthermore, they explained a high 40 Ar concentration by a possible emplacement of
sill. However, in the case of the speleothems the Ar-excess remains an open question. To get
a hint with regard to the origin of this deviation, the main gas components (O2 , N2 ) as well
as the major trace gases (CO2 , CH4 ) should be recorded as well.
It is important to mention, that the deviations from the excess-air line are only found in
the first step or steps. At higher step numbers little or no deviations from the expected
line are found, as it can clearly be seen in Fig. 4.24 and Fig. 4.22. Only the 5 hit-points
are distant from the line, all 60-hit points as well as the 110-hit points are inside the area
of air-equilibrated water with a different contribution of atmospheric air. An interesting
result is in both cases the very good correspondence of the heating results with regard to the
temperature. They refer to a Holocene stalagmite (<8 kyr) and should therefore not show
large differences in temperature. The results of the heating step can affirm this expectation.
A systematical analysis of all measured data can provide a better insight into unexpected
effects. At first, the measurements with bad fits are rejected (χ2 >10) as they may have been
influenced by technical problems during extraction and measurements. Then, the residuals
between measurement and model are calculated and summarized according to each stalagmite
(see Table 4.5). The major part of the investigated stalagmites show no significant deviation
from the expected concentration for any noble gas. E.g. the deviation of the BU-1 samples
from the model is in general smaller than 1%. However, the H12 stalagmite from Oman
clearly stands out. Using Ne for temperature calculation always delivered bad results and
strange temperatures. After detecting the Ne excess to be not a systematical measurement
problem, but a property of the stalagmite, Ne was not used further for fitting. Subsequently,
Chapter 4. Results
Kr(ccSTP/g)
132
2.0x10
-6
1.5x10
-6
5 hits
60 hits
heat
aew
excess-air
line
1.0x10
5.0x10
2.5x10
-7
2.0x10
-7
1.5x10
-7
1.0x10
-7
5.0x10
-8
heat
aew
-6
-7
0.0
0.00
5.0x10
0.02
0.04
0.06
0.08
-4
1.0x10
0.10
-3
0.12
1.5x10
-3
0.14
Ar (ccSTP/g)
Figure 4.24: Deviations from expected signals at BU-1. The first extraction step resulted in high Ar
concentrations. None of the subsequent steps did provide similar deviations. The insert shows the results
for the last step of each sample. The heating measurements are not hotblank-corrected.
the temperature values got more reasonable and in general better fitting results have been
achieved. Ar, Kr and Xe show no significant deviation from the models, whereas the Ne
concentration is more than 2 σ away from air-equilibrated water with atmospheric air.
It is difficult to explain this unexpected effect. An explanation for the Ne excess may be an
enrichment of noble gases in some relatively undisturbed and badly ventilated caves, perhaps
in some way related to the findings of Podosek et al. (1980), who discovered enrichment
patterns towards heavier noble gases and also an enrichment of Ne in sedimentary rocks. Assuming a certain ventilation of the cave chambers it is at least in shallow caves unreasonable,
but can not be excluded in deeper caves with long and narrow passages. Our own measurements confirm that in the shallow Bunker Cave no significant deviation from the atmospheric
mixing ratio exists. Air samples from deep and less ventilated chambers in the B7 cave show
also no significant deviation from the atmospheric mixing ratios of Ne, Ar, Kr and Xe. Thus,
it is less likely that the Ne excess is caused by an excess in the cave atmosphere.
Another reason may be a diffusive isotope exchange of the inclusion fluids with the noble
gases from the host calcite and sedimentary particles in it, as known from the stable isotopes
Wilkinson (2001). A cross section of the stalagmite H12 along the growth axis (Badertscher,
2007; Neff, 2001) reveals a pronounced layering consisting of brownish, red and white layers.
The white layers are related to a larger amount of water-filled inclusions, whereas the the
brownish and especially the red layers may be due to depositions of sedimentary mineral
grains. The deposition of extrinsic material in the H12 stalagmite is also manifested in the
high detritus values ( mean c(232 T h) ≈ 14,1 ng/g, Neff, 2001).
However, the Ne excess seems not to be dependent on the air-water volume ratio or on the
extraction procedure (Fig.4.25). The deviations from the expected model values do not decrease or increase with the fraction of noble gases from the air-filled parts. Only 5 out of 30
samples exceed the 1 σ deviation from the mean Ne excess, but show no systematics. Furthermore, even stepwise measurements and microwave extraction yield a similar Ne excess
compared to model values. Thus, this Ne excess is not generated by extraction or measurement procedures, but has to be an intrinsic property of the stalagmite. Interestingly, other
studies also revealed Ne excess for samples from the same region (Scheidegger et al., 2008).
4.4. Noble gas fractionation and enrichment
133
Table 4.5: Noble gas excesses in different samples. The given numbers refer to the residuals between
measurement and model (UA-model, fit-parameter: T , A). ”Number of samples” indicates how many
measurements have been used for the calculation of the mean. The uncertainty refers to the 1σ-deviation
of all used measurements. The displayed samples have been selected by the χ2 -values. Results with bad
χ2 (>10) have been rejected. H12 was fitted without Ne.
deviations from model (in %)
Ne
Ar
Kr
sample
BU-U
BU-1
BU-2
CG
H12
MA2
1.9
0.5
1.8
-6.7
22.3
2.9
±
±
±
±
±
±
2.0
1.4
2.0
2.3
8.7
7.8
-1.5
-0.6
-3.1
4.9
0.5
-0.8
±
±
±
±
±
±
1.3
1.7
3.2
3.8
2.1
3.2
-0.1
-0.2
-0.4
6.6
-0.9
-0.4
±
±
±
±
±
±
1.1
0.8
2.1
4.1
3.2
3.3
Xe
0.9
0.6
2.4
-3.2
0.3
-0.2
±
±
±
±
±
±
number of samples
1.3
1.4
3.9
2.2
1.1
2.4
16
7
2
4
30
6
50
45
40
excess Ne (%)
35
30
25
20
15
10
5
0
0
1000
2000
3000
4000
A [-] x 1000
Figure 4.25: Excess-Ne of the H12 samples plotted against the air-water volume ratio A. Here, excessNe does not refer to air addition, but to the deviation of the measured values from model expectation
based on fits to Ar, Kr and Xe. The mean Ne excess is given in Table 4.5. Three outliers (highest
Ne excess) refer to early measurements and may therefore influenced by extraction and measurement
problems.
134
Chapter 4. Results
100
modern calcite
10
4
He (10 -11 ccsTP / g calcite )
1000
IFQ
1
10
-2
10
-1
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10
7
age (ka)
Figure 4.26: He-content measured by Stuart and Turner (1992). He is given in 10−11 ccSTP per g of
calcite. IFQ is given for comparison and refers to an inclusion-free quartz with unknown age. Modern
calcite refers to the samples BB, YB and Carl-1 from the same paper. The data includes He from airand water-filled inclusions and also the radiogenic fraction.
The CG stalagmite from Cuba also shows a significant deviation in the Ne values. However,
this result is not as robust as in the case of H12 as it is based on only 4 fitted samples.
Additional measurements may substantiate this first result.
In summary, most speleothems yield a noble gas concentration which corresponds to the
theory. The total noble gas concentration obtained from a calcite sample is composed of
a component from air-equilibrated water and unfractionated atmospheric air. In rare cases
systematic deviations can occur (H12, Oman). However, we found only systematic shifts
for Ne so far. Consequently, also for these samples reasonable temperature calculation with
acceptable errors has been possible by using the unaffected Ar, Kr and Xe concentrations.
4.5
Dating via Helium
Stalagmites are usually dated using the uranium-thorium method (Richards and Dorale,
2003). In the case of old stalagmites (≥ 300 000 yr) it gets difficult to date precisely using
the U/Th disequilibrium. The results can be charged with high uncertainty or it will be even
impossible to date if the speleothem is older than 500 000 years. For such stalagmites U/Pb
dating can yield an age (Richards et al., 1998; Walker et al., 2006; Hellstrom et al., 2008;
Polyak et al., 2008). Additionally, helium may help to estimate a reasonable age, if no U/Pb
data are available.
First indications that He in speleothems can be useful for age estimation can be drawn from
the data of Stuart and Turner (1992). Even though they claim to find very low He contents, a
remarkable difference between modern calcite and rather old precipitates (some 100s of Myr)
can be observed (Fig. 4.26). They did not perform data fitting to disentangle the different
contributions to the signal. Additionally, the 395 Myr Rhynie chert showed lower He content
4.5. Dating via Helium
135
than the younger triassic authigenic quartz minerals. Furthermore, the sample preparation
may have affected the results due to baking out of the extraction set at 200 ℃ for 36 hours.
However, an increasing He content can be observed for increasing age.
The recent study of Scheidegger (2005) stated that radiogenic He must be abundant due to
high He/Ne ratios above air or equilibrated water values. A comparison of radiogenic He
contents from a stalagmite of the Qunf-Cave with radiogenic He and age data from the H12stalagmite yields reasonable ages for the Qunf stalagmite. She concluded from this results
that U/Th-He dating should principally be possible.
With respect to our own experiments, the investigation of U/Th-He age determination was
motivated by missing U/Th data. The first stalagmite with reliable noble gas concentrations
was dated some time after the noble gas measurements. To immediately compare the calculated noble gas temperature with an expected value a rough age estimation was necessary
and was performed using the helium data.
Helium accumulates in the stalagmite due to the radioactive decays in the 238 U, 232 Th and
235 U decay chains (Table 4.6). As it can be assumed that the calcite is closed for the noble
gases at typical temperatures (≤ 30 ℃) and timescales (Copeland et al., 2007), the over
time accumulated helium will be a measure of the age. Furthermore, we assume that the
air-filled inclusions contain a gas composition equal to the atmospheric mixing ratio and the
water-filled inclusions a noble gas component which is equilibrated at cave temperature. The
radiogenic helium component Crad can therefore be calculated by subtracting the excess-air
Cair and the equilibrium component Ceq from the total gas amount Ctot :
Crad = Ctot − (Ceq + Cair )
(4.2)
The equilibrium component Ceq can be determined using the known amount of released water
and the equilibration temperature from data fitting. The excess-air component is derived by
inverse modelling from all noble gas data except He.
The radiogenic component is built up inside the calcite matrix and is, therefore, dependent
on the uranium CU238 , CU234 , CU235 and the thorium concentration CTh232 , given in µg U
(Th) per g of calcite. To establish a first-order approximation, we assume that initially there
are no daughter elements to the mentioned isotopes. In general this assumption will not limit
the calculation as the half-lives of the daughter elements are in general very low. Thus a
secular equilibrium is established in the centennial scale.
However, exceptions are 226 Ra with a half-life of 1602 a, 231 Pa with a half live of 32 760 a and
230 Th with a half-life of about 75 400 a. With regard to old samples (some 10 000 years) the
initial Ra will not play a significant role in the case that Ra is not strongly concentrated at
the beginning. Gainon et al. (2007) reported that radium is adsorbed efficiently due to iron
or manganese-hydroxies under oxidizing conditions. Therefore, the typical radium activity
concentration in groundwater and springs is rather low. E.g. in Austria > 90% of more
than 380 wells are below 20 Bq/m3 (BMLFUW, 2002). Fanale and Kulp (1961) reported a
Ra content in a calcite sample to be 3.6·10−15 g Ra per g of calcite. Based on the activity
concentration in groundwater of <20 Bq/m3 , which is also valid for the cave dripwater, the
radium content would be smaller than 5.4·10−16 g per g water. Compared to the result of
Fanale and Kulp (1961) no extreme enrichment of radium in the precipitated calcite occurs.
If we assume 3.6·10−15 g Ra per g of calcite and an age of 10 000 years, we obtain about
5.7·107 α’s from the decay of this initial Ra. This is low compared to the results of 238 U.
Using an 238 U content of 1 ppm we obtain about 3.9·109 α-particles alone from the decay of
238 U in the same timescale. The contribution of initial Ra can be of importance in the case
of low U content (<0.1 ppm) and ages below 10 kyr (Fig. 4.27). However, this consideration
is based on the measurements of Fanale and Kulp (1961). Different Ra contents due to e.g.
136
Chapter 4. Results
10
10
9
10
8
10
7
10
6
10
5
4
He-nuclei
10
-15
Ra226 + daughters: 3.6 10
g/g
U-238: 0.1ppm
U-238: 1ppm
U-238: 0.1ppm + U-234: 0.01 ng/g
10
100
1000
10000
years
Figure 4.27: Comparison of α-particles produced by 238 U and 226 Ra. The red line refers to the inital
Ra concentration of 3.6·10−15 g/ g calcite. After about 8 kyr no further contribution of the initial Ra to
the total number of the α-particles occurs, as most of the Ra is already decayed.
detritus may change these values. In the case of older speleothems the contribution gets less
significant with age as the initial Ra is almost completely decayed after 8 kyr.
The 231 Pa in the 235 U decay chain should be low. The estimated oceanic abundance is about
5·10−17 g per g water (TJlab, 2008). Furthermore, 231 Pa measurements on carbonates (corals)
showed that no initial Pa-231 is abundant in the samples (Edwards et al., 1997). Thus, initial
231 Pa is neglected in the following.
The initial 230 Th in the calcite is in general rather low as Th is hardly soluble and thus
not contained in the dripwater. However, it is particle reactive and can be included in
the calcite with detritus such as clay minerals. The initial 230 Th is calculated using the
detritus correction based on a fix relation between 232 Th and detritical 238 U (Wedepohl,
1995). Assuming decay equilibrium between the detritical components, the initial 230 Th can
be inferred:
λ238
N232 λ238
N230 = N238, det ·
·
=
= N232 · 4.44 · 10−6
(4.3)
λ230
3.8 λ230
For calculation of the total number of α particles the 235 U decay chain is of minor importance
because of the low abundance of 235 U:
CU235 = 0.0072527 · CU238
(4.4)
The content of 238 U and 232 Th can be assumed to be constant over the investigated time scale
of less than one million years. However, for the other relatively long lived isotopes (234 U,
230 Th) the initial concentration and the change over time has to be calculated. In the 238 U
decay chain a total of 8 α particles, in the 232 Th decay chain 6 α’s and in the 235 U decay
chain 7 α’s are produced.
4.5. Dating via Helium
137
Table 4.6: Thorium, uranium-actinium and uranium-radium decay series. Plotted are the main decay
channels. In summary 8 α’s are produced in the U-238 series, 7 in the U-235 and 6 in the thorium decay
series.
nuclide
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Po-212
Ti-208
Pb-208
decay
thorium
α
β
β
α
α
α
α
β
α, β
α
β
-
half life
nuclide
1.405·1010 yr
5.75 yr
6.15 h
1.91 yr
3.66 d
55.6 s
0.145 s
10.64 h
60.5 min
10−7 s
3.1 min
-
U-238
Th-234
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Bi-214
Ti-210
Pb-210
Bi-210
Po-210
Pb-206
decay
U-238
α
β
β
α
α
α
α
α
β
β
α
β
β
β
β
α
-
half life
nuclide
4.47·109 yr
24.1 d
6.7 h
2.46·105 yr
7.54·104 yr
1602 yr
3.82 d
3.05 min
26.8 min
19.9 min
0.164 µs
19.9 min
1.3 min
22.3 yr
5.01 d
138 d
-
U-235
Th-231
Pa-231
Ac-227
Th-227
Rn-223
Fr-223
Ra-223
Rn-219
Bi-215
Po-215
Pb-211
Bi-211
Ti-207
Pb-207
decay
U-235
α
β
α
β
α
β
β
α
α
β
α
β
α
β
-
half life
7.03·108 yr
25.5 h
32760 yr
21.7 yr
18.7 d
23.2 min
22 min
11.43
3.96 s
7.6 min
1.78 ms
36.1 min
2.14 min
4.77 min
-
To correctly assess the amount of He generated inside the speleothem, the calculation for the
decays in the 238 U decay chain has to be performed in a stepwise order as shown by Farley
et al. (2002).
4
He
=
+
+
+
+
+
t
(4.5)
6 · CTh232 · 1 − e− λ232
t
CU238 · 1 − e− λ238
(4.6)
t
7 · 0.007257 · CU238 · 1 − e− λ235
(4.7)
t
t
CU238 · 1 − e− λ238 + CU234 · 1 − e− λ234
(4.8)
t
t
t
7 (CU238 · 1 − e− λ238 + CU234 ) · 1 − e− λ234 + CTh230 · 1 − e− λ230
(4.9)
t
5 · CRa226,i · 1 − e− λ226
(4.10)
Equation (4.5) states the alpha decay from 232 Th within the time t, equation (4.6) the decay of
238 U, (4.7) of 235 U and equation (4.8) the decay of 234 U including the fraction of 238 U which
decays in the same time to 234 U. Equation (4.9) summarizes all α’s from the subsequent
isotopes and the initial 230 Th (which should be zero due to the chemical behaviour, but can
yield values above zero due to detritus contribution) including the Th amount which was
built up in time t due to the decay of 238 U and 234 U. Equation (4.10) is the term including
the contribution of initial 226 Ra.
This theory was applied to six speleothem samples from the Bunker cave stalagmite BU-U.
In this case all noble gases have been measured, so that the calculation of the three He
components (equilibrium, air and radiogenic component) was possible. Additionally, the
released water amount and the sample weight was measured. Results are listed below in
Table 4.7. The isotope content of U and Th was determined in the context of the U-Thdating. The radiogenic He was calculated from all He released in the crushing and the
subsequent heating step using equation (4.2). This value was then converted to a radiogenic
138
Chapter 4. Results
Table 4.7: Exemplary calculation of the expected radiogenic He for the mean of six samples of a growth
layer of BU-U. The radiogenic He is calculated by summarizing the α particles of all decay chains. Herad
indicates the mean amount of the measured radiogenic He per gram of calcite. α’s are summarized from
all expected decays within the measured U/Th-age of 10 800 years and converted into gas amounts per
gram of calcite.
parameter
CU238
CU234
CTh232
CTh230
CTh230,initial
Herad
U/Th-He age
U/Th age
α’s (theoretical)
value
108.79 ng/g
0.00919 ng/g
0.46 ng/g
0.000263 ng/g
2.02·10−6 ng/g
3.0 ± 2.2 · 10 −10 ccSTP/g
16 500 yr (+7700 yr, -10100 yr)
10 500 - 12 100 yr
2.15· 10 −10 ccSTP/g
He amount per gram of calcite and afterwards the mean of the six investigated pieces was
determined (s. Table 4.7).
From the known isotope content and the age (10 800 a) the α particles are calculated. From
the 232 Th-decay 7.93·106 , from 235 U 3.09·108 , from 238 U 9.58·108 , from 234 U 1.49·109 αparticles are produced including the additional contribution of already decayed 238 U. The
subsequent daughter decays to the stable 206 Pb account for additional 3.03 ·109 αs. In total,
we get 5.79·109 He nuclei, which corresponds to 2.15·10−10 ccSTP He. With consideration
of the dating error of the stalagmite sample of 300 years, the radiogenic He value is located
between 2.08 ·10−10 and 2.48·10−10 ccSTP. In this calculation the uncertainties of the isotopic
content of the calcite are not included, but should result in minor deviations.
A comparison of the He amount derived by decay calculation and the mean determined by
noble gas measurement show an agreement within the uncertainty. However, the scatter of
the measurement data is large. The mean value is about 3.00 ·10−10 ccSTP/gCarb , whereas
the standard deviation is 2.2 · 10 −10 ccSTP/g. The scatter of the measured values can at
least partially be attributed to the extraction process. Not only different methods have been
used, but also small deviations in the extraction procedure have occurred which influenced the
efficiency. This can be verified by a closer look to the obtained grain size distribution (Marx,
2008). A larger fraction of big grains can explain a lower amount of extracted He and vice
versa. Additionally, not all samples yield the same ages, even if they are extracted nearby.
The corresponding growth layer expands over the period from 10 500 yr BP to 12 100 yr BP.
The U/Th-He age was determined by inverse modelling. The age is calculated according
to the known U and Th content by using time as fitting parameter and by comparing the
measured radiogenic He amount with the mathematically derived α decays.
The results indicate that the measured amount of radiogenic helium is comparable to the
theoretically expected values. Turning around this statement an age estimation via inverse
modelling using the known U and Th content may be possible. Although we have to keep in
mind the not totally reproducible and uniform extraction process, at least the accumulation
of radiogenic helium may be used to give a rough estimation of the speleothem age.
In the case of very old samples (≥ 1 Myr) the helium - uranium dating had already been
used, not yet for speleothems, but for instance for corals (Bender, 1973), apatite (Wolf et al.,
1996) or other minerals (s. summary of Farley, 2002). An important requirement is that the
He is retained quantitatively in the calcite, which can be assumed with regard to the diffusion
data recently given by Copeland et al. (2007).
4.5. Dating via Helium
139
1E-7
radiogenic He (ccSTP/g
carb
)
run Beta BU-U
run Beta
run Gamma
run Delta
1E-8
1E-9
1E-10
1k
10k
100k
1M
10M
U-Th age (a)
U-Pb age (a)
Figure 4.28: Compilation of radiogenic He data for stalagmites with different ages. They refer to
different measurement runs. The radiogenic He per g of calcite is plotted against U-Th ages, respectivley
U-Pb ages. The sample with the highest radiogenic He value has not been dated so far by U-Pb and is
only arranged according to the trend.
During the duration of the project not only stalagmites from different locations have been
measured, but also speleothems covering a certain age scale. A summary of these samples is
given in Fig.4.28. There, the absolute amount of radiogenic He is normalized to the sample
weight and given per g of calcite. The youngest sample is dated by U/Th to about 1 300 years,
the oldest one dated by U/Pb to about 2.2 Myr. The different measurement runs are displayed
in distinct colours. Obviously there is a trend of increasing radiogenic He with age. Despite
some deviations in the case of samples extracted in run Gamma, a rather linear relation
between age and radiogenic He can be inferred from this diagram. Furthermore, in all runs
a similar linear trend can be found. However, the total amount of released gases depends on
the extraction efficiency. Compared to Run Delta and Run Gamma, Run Beta seems to yield
a generally lower efficiency. The other two runs are comparable. It is important to mention
that the scatter in the obtained radiogenic He is remarkable and can reach up to one order
of magnitude, even though samples with similar ages are used.
The differences in the magnitude of the measured radiogenic He can already be used to deduce
a reasonable age interval for some samples. Based on the data shown in Fig. 4.28 a linear
fit was created. The outliers of run Gamma have been rejected for this purpose. Using the
remaining 27 data points a fit with a correlation coefficient R of about 0.82 was obtained
(Fig. 4.29).
The relationship was used for age determination of the sample BC1-4 from Northern Canada.
It refers to a stalagmite from the Bear Cave in the Bear Cave Mountains in the Yukon
Territories (66°25’N, 139°20’W). Today the mean annual air temperature is about -8.9 ℃.
The measurement on a piece of BC1-4 revealed an extremely high amount of radiogenic
He, which is about one order of magnitude above all other samples measured so far. The
adjustment of the age to the fit using the radiogenic He results in a mean age of about
140
Chapter 4. Results
BC1-4
stalagmite results
linear fit
radiogenic He (ccSTP/g
carb
)
1E-7
1E-8
1E-9
1E-10
1E-11
1k
10k
100k
1M
10M
U/Th , U/Pb age
Figure 4.29: Correlation between U/Th age and radiogenic He. The data displayed in Fig.4.28 have
been used for fitting a linear relationship. The sample BCI-4 was not included for fitting and is adjusted
to the linear fit.
1E-6
He per g calcite
and per µg U
Linear Fit
rad He (ccSTP/(g*µg))
1E-7
1E-8
1E-9
1E-10
1k
10k
100k
1M
10M
U/Th-age
Figure 4.30: Correlation between U/Th ages and normalised radiogenic He. The radiogenic He is
referred to 1 g calcite and an uranium content of 1 µg U per g calcite. Again the highest value is only
adjusted to the fit. The dashed line indicates extrapolation of the linear relationship.
4.5. Dating via Helium
141
60 Myr. Assuming the uncertainty to be given by the typical scatter of the other speleothem
measurements, BC1-4 could be between 15 Myr in the minimum case and maximum 250 Myr
old. Information about growth conditions and U/Th data can help to constrain the result
from the radiogenic He. U/Th measurements revealed the sample to be out of the dating
range, i.e. older than 500 kyr. Taking into account that today no speleothem growth can be
observed due to the low mean annual air temperature, the stalagmite has most likely grown
in a warmer period. Even 5 ℃ above the present-day temperature would not be sufficient
to enable stalagmite growth. Thus, the stalagmite can not originate from the Pleistocene,
but from the Miocene, which was the last warmer period. The mean age of about 60 Myr,
derived from radiogenic He, suggests the stalagmite to have grown in the Paleocene, which
was a significantly warmer period, especially during the Paleocene/Eocene transition (55 Myr
ago), (Norris and Röhl, 1999). Another very warm period was the Cretaceous, which is also
a reasonable period for the investigated speleothem. During this time the sea level was
higher (Haq et al., 1988) and the tropical as well as the polar sea temperatures had been
elevated. The tropic sea ranged between 27 and 32 ℃ and the polar sea above 0 ℃ up to
15 ℃ (Barron, 1983). The lack of cold-water fauna as well as of any permanent ice and the
occurrence of a large palaeo flora at the coast of Northern Alaska (420 species) additionally
confirm warm conditions in the far north. Thus, it would not be unreasonable that the
investigated speleothem sample has grown during this period, which was one of the warmest
in Earth’s history.
The age of the BC1-4 sample is additionally constrained by the use of normalized He data.
The amount of accumulated He is strongly dependent on the uranium and to a smaller extent
on the thorium content. A summary of the normalized data is given in Fig.4.30, where the
radiogenic He is calculated per g calcite and µg of 238 U. Fitting the data by a linear function
leads to a similar relation as in the case of the non-normalized He content. The correlation
coefficient R is not that good (0.61) because no data have been discarded and the total sample
number was smaller (n=21). For some samples the U content was not known and thus they
could not be used for fitting the normalised data. However, the radiogenic He fits to a similar
age as inferred from Fig. 4.29, but with a higher mean of about 130 Myr. This result would
again correspond to the warm Cretaceous period.
Thus, the age estimation with radiogenic He is not unreasonable, but has first to be constrained by U/Th or U/Pb dating. Additionally, a uniform and reproducible extraction with
constant efficiencies has to be developed to reduce uncertainties and to finally enable more
precise quantitative results. Furthermore, the theoretical calculations have to be arranged
with the experimental data. Fitting the liberated radiogenic He of BC1-4 to the theoretically
expected α - decays during time, we would obtain an age of about 1.5 Myr. This is one to up
to two orders of magnitude less compared to the age inferred from Fig. 4.29 and Fig. 4.30
and may be due to a low extraction efficiency. Preliminary data obtained from U/Pb dating
of the sample BC1-4 resulted in an age of ∼ 5 Myr (unpublished, personal communication
Nicholas Utting), which suggests the dating attempts with U/Th-He to be in a similar order
of magnitude.
In summary, our data, the exemplary application as well as the low diffusivity of He are
encouraging and suggest this technique to be possibly useful as a dating tool for very old
speleothems.
142
Chapter 4. Results
Figure 4.31: Halves of the stalagmite BU-U. The left side was used for U/Th dating, the missing
part between the two halves was used for isotope, trace-element and noble gas measurements. The red
rectangle refers to the measurement of six subsamples from the according growth layer. The brownish
layers indicate hiatuses. Below the first hiatus the stalagmite has grown during MIS 3 and then in the
Eemian. The layers below the subsequent extremely dark parts have not been dated and investigated.
4.6
Case studies
In the first subsection a set of samples is presented, which demonstrates the general possibility for noble gas temperature determination via fluid inclusions in speleothems. Reproducible
values from one growth layer covering the Early Holocene (11 - 12 kyr BP) show the robustness of the data. Furthermore, a comparison with a pollen study approves the determined
temperature difference of 4 ℃ between the Early Holocene and the Early Middle Ages to be
a reasonable result.
Encouraged by these first useful data we tried to establish a temperature record along the
growth axis of the same stalagmite (BU-U). The results are presented briefly in the following
subsection and discussed in comparison with independent studies. The determined temperature differences between Holocene, Glacial values (mainly Marine Isotope Stage 3 - MIS3)
and the Eemian are in agreement with expectations and reveal the Eemian maximum to be
possibly warmer than present-day climate. Similarly, a preliminary temperature record was
established for a completely Holocene stalagmite (BU-1) from the same cave. Based on several samples along the growth axis hardly a significant temperature variation can be found
in the last 8 kyr in central Europe.
4.6.1
Reproducibility and uncertainties
The first indication for successful temperature determination from speleothem fluid inclusions
by noble gas measurements came from one small piece from the stalagmite BU-U (Fig.4.31),
which was measured in spring 2007. To substantiate this encouraging results, six pieces
from the according growth layer of this stalagmite and a small soda straw were prepared
(Table 4.8). Additionally, a small piece was cut out of the younger stalagmite BU-1. All
of these samples derive from the Bunker Cave in Northwest Germany (Sauerland). BU-U
exhibits very large columnar crystals (up to a length of some cm, Fig. 4.32) in the uppermost
growth layer belonging to the Early Holocene.
4.6. Case studies
143
Figure 4.32: The uppermost part of BU-U on a closer view. The three bluish pictures are photos taken
from microscope analysis using polarized light. Unfortunately, no exact scale bar was available, but the
displayed details shows columnar crystals in the order of one cm. These crystals are already visible by
eye as it can be seen in the lower right picture. The growth axis tends there to the lower right edge.
Table 4.8: Total sample weight, released water amount, apparent gravimetric water content, and applied
extraction method for different Bunker Cave speleothems
Sample
BU-Ua
BU-U0
BU-U1
BU-U2
BU-U3
BU-U4
BU-1
Soda Straw
weight (g)
2.748
1.084
1.001
1.069
1.064
1.428
1.424
1.140
water amount (µl)
1.31 ± 0.04
0.159 ± 0.006
0.516 ± 0.007
0.49 ± 0.01
0.411 ± 0.005
0.765 ± 0.007
0.899 ± 0.015
0.669 ± 0.006
water content % wt
0.48
0.15
0.52
0.46
0.39
0.54
0.63
0.59
crushing device
copper tube
copper tube
steel cylinder
copper tube
steel cylinder
steel cylinder
steel cylinder
steel cylinder
From the six BU-U sub-samples, three were extracted by squeezing in copper tubes and three
by crushing in the steel cylinder accompanied by mild heating at 50 ℃. The soda straw was
cut into small pieces, which were crushed in the steel cylinder. Similarly, the BU-1 sample was
extracted by crushing in the steel cylinder. The further noble gas purification, separation and
measurement procedure as well as the water determination follows the sequences described
above and is presented also in detail by Kluge et al. (2008). Most samples had a total weight
of about 1 g and yielded a well-measurable water amount of around 0.5 µl (≈ 0.5 mg). The
water content was thus quite uniform at nearly 0.5 % wt (Table 4.8). The clearly lower value
of sample BU-U0 is due to incomplete crushing in a copper tube. When the tube was opened
after the measurement, a comparatively large fraction of coarse grains was found for this
sample.
The speleothem pieces of BU-U and the soda straw from Bunker Cave produced relatively
high noble gas signals, which together with the quite precisely determined water amounts
yield well-defined noble gas concentrations (Table 4.9). With the exception of sample BU-U0,
144
Chapter 4. Results
which yielded small gas amounts because of incomplete extraction, and the Xe measurement
in the case of sample BU-U2, the relative uncertainties of the concentrations are about 3 %
or less. Comparatively large errors occurred for small Xe amounts from samples crushed in
copper tubes due to a significant blank correction in this case.
Table 4.9: Apparent dissolved noble gas concentrations in ccSTP/g , i.e. total measured gas amounts
(in ccSTP) per released water amount (in g), for different Bunker Cave speleothems.
Sample
BU-Ua
BU-U0
BU-U1
BU-U2
BU-U3
BU-U4
BU-1
Soda Straw
He (10−7 )
6.56 ± 0.50
5.54 ± 0.26
14.3 ± 0.4
6.84 ± 0.21
10.5 ± 0.3
6.05 ± 0.14
5.55 ± 0.12
1.46 ± 0.04
Ne (10−7 )
10.1 ± 0.4
11.0 ± 1.1
9.68 ± 0.26
8.67 ± 0.28
7.88 ± 0.22
7.41 ± 0.19
17.1 ± 0.31
2.40 ± 0.07
Ar (10−4 )
8.30 ± 0.25
8.56 ± 0.39
8.27 ± 0.11
7.54 ± 0.17
7.16 ± 0.08
6.86 ± 0.06
11.3 ± 0.19
4.28 ± 0.04
Kr (10−7 )
1.57 ± 0.05
1.65 ± 0.08
1.63 ± 0.03
1.50 ± 0.04
1.44 ± 0.02
1.39 ± 0.02
1.87 ± 0.04
1.01 ± 0.01
Xe (10−8 )
2.06 ± 0.68
2.57 ± 0.07
2.12 ± 0.08
2.16 ± 0.24
2.00 ± 0.07
1.92 ± 0.03
2.17 ± 0.04
1.60 ± 0.06
In a plot comparing two measured noble gas concentrations, e.g. Kr vs. Ar (Fig. 4.33)
or Xe vs. Ne (Fig. 4.34), the relative contribution of noble gases from air-filled inclusions
can be estimated. The open circles indicated in these figures represent concentrations of
AEW at temperatures between 0 and 30 ℃ (from top to bottom). The lines leading from
these points to higher concentrations represent the effect of air addition. The more distant
a sample plots from the AEW points, the more noble gases are contributed by air-filled
inclusions. Projecting a sample along an excess air line onto the curve connecting the AEW
points yields its equilibration temperature based on the two plotted noble gases. Fig. 4.33
and Fig. 4.34 show that all the samples scatter in the range between 0 and 5 ℃, except the
soda straw, which lies around 6 ℃ and the BU-1 piece at about 7 ℃.
The final NGTs obtained by the fitting procedure (Table 4.10) slightly differ from the values
indicated by Figures 4.33 and 4.34, as they take the information from all four atmospheric
noble gases into account. The optimal model parameters T and A are obtained by minimising
the sum of the error-weighted squared deviations between measured and modeled concentrations, denoted as χ2 . The goodness of fit can be assessed by the χ2 -value, which for 2 degrees
of freedom (4 measured concentrations and 2 free parameters) has an expectation value of 2
and a probability of 0.99 to be smaller than 9.2. Only one of the 8 samples yields a slightly
larger χ2 (BU-U4 with χ2 = 9.3, Table 4.10), the others are well described by the simple UA
model of equation (2.9) (chapter 2.3.2) within the limits of the experimental errors. This result confirms the assumption that unfractionated air is present in the stalagmites in addition
to an AEW component.
A surprising result obtained from the soda straw sample is the extremely low fraction of noble
gases derived from air-filled inclusions. An air/water volume ratio A of (1.80 ± 0.39)·10−3
was determined, which provides optimal conditions for the NGT calculation. Therefore, the
temperature uncertainty for this sample is only about ± 0.4 ℃, showing that is possible to
achieve high precision NGTs at least from selected speleothems.
As already observed in Figs. 4.33 and 4.34, the NGTs of all samples from the BU-U stalagmite
scatter around the same temperature, whereas the soda straw and BU-1 yield significantly
higher values. Even if we do not have completely identical samples and at least small deviations have to be expected, the temperatures of the BU-U samples are astonishingly similar
and reproduce within their uncertainty. It is important to mention that the shown results
derive from different measurement runs with different ion-source tuning, peak shapes and
measurement methods as well as from a different extraction procedure (in copper tubes or
4.6. Case studies
145
2.0x10
-7
1.8x10
-7
1.6x10
-7
1.4x10
-7
BU-1
1
BU-U
0
a
2
Kr
4
1.2x10
-7
1.0x10
-7
8.0x10
-8
6.0x10
-8
3
0˚C
SD
10˚C
20˚C
-8
4.0x10
-4
2.0x10
4.0x10
-4
6.0x10
-4
8.0x10
-4
1.0x10
-3
1.2x10
-3
1.4x10
-3
Ar
Figure 4.33: Noble gas concentrations in ccSTP per gram of water from fluid inclusions, Kr vs Ar.
The open circles on the left hand side represent water equilibrated with atmospheric air (AEW) at
temperatures from 0 ℃ (uppermost point) to 30 ℃ (lowest point). Samples are shown with error bars,
SD: soda straw, the numbers for BU-U indicate the corresponding subsamples. The dotted lines indicate
the effect of addition of air to the equilibrated water. As the air component is relatively small, backextrapolation along such lines yields a well-defined equilibration temperature for each pair of measured
noble gas concentrations.
in the steel cylinder) and with different preparation and pumping procedures. Furthermore,
measurements of pieces from the same layer in a subsequent run yielded also temperatures
in agreement with the older results, even if the extraction procedure was changed (different
number of strokes and steps), the preparation was done more sophisticated (baking out of
the new steel crusher at 150 ℃ overnight) and the measurement scripts and procedures have
been replaced or written completely new. In between also the filament of the ion-source was
changed. This surprising reproducibility underlines the robustness of the obtained results
for all subsamples from this growth period and gives a certain confidence in the obtained
results. However, to assess whether the NGTs and the temperature differences are plausible,
a comparison with present-day temperatures as well as with results of other palaeoclimate
studies is necessary.
The BU-U sub-samples give a mean temperature of (2.9 ± 0.7) ℃, about 7 ℃ cooler than the
modern cave and air temperatures. The recent mean annual air temperature in the region
of Bunker Cave is about 9.5 ℃ (1961 - 1990 mean of the station Hagen-Fley of the German
weather service DWD). The cave temperature is measured since 2006 in the framework of a
monitoring program and shows a mean of about 10.5 ℃. The single soda straw sample from
the Bunker Cave yields a NGT of 6.4 ± 0.4 ℃, about halfway between the cold temperatures
indicated by the stalagmite samples and the warm modern cave temperature. The young
BU-1 sample resulted in a similar temperature of 7.1 ± 0.8 ℃.
To assess whether the derived NGTs and temperature differences are plausible, the age of the
samples has to be considered. U/Th measurements on the same growth layer of the stalagmite BU-U led to an age of 10 800 - 11 700 years, which indicates that the measured BU-U
parts are from the end of the Younger Dryas or the beginning of the Holocene (Preboreal).
Because the whole soda straw was needed for noble gas measurements, there was no material
146
Chapter 4. Results
3.5x10
-8
3.0x10
-8
2.5x10
-8
2.0x10
-8
BU-U
0
Xe
BU-1
0˚C
4 3
1.5x10
-8
a
2 1
SD
10˚C
1.0x10
-8
5.0x10
-9
20˚C
30˚C
0.0
5.0x10
-7
1.0x10
-6
1.5x10
-6
2.0x10
-6
Ne
Figure 4.34: Noble gas concentrations in ccSTP per gram of water from fluid inclusions: Xe vs Ne.
The open circles on the left hand side represent water equilibrated with atmospheric air (AEW) at
temperatures from 0 ℃ (uppermost point) to 30 ℃ (lowest point). Back-extrapolation along the excessair lines yield a well-defined equilibration temperature for each pair of measured noble gas concentrations.
left to determine a U/Th age. Therefore, we considered the use of radiogenic He to estimate
its age (Table 4.10). As discussed in the precedent section, the feasibility of using He for
quantitative dating of stalagmites certainly needs further investigation. However it seems
possible to derive some qualitative statements. The mean concentration of radiogenic He per
weight of carbonate is around 2.6·10−10 cm3 STP g−1
carb in the case of the stalagmite BU-U.
The radiogenic He content of the six samples scatters considerably around this mean value,
which can be related to the extraction efficiency. Since the soda straw sample was extracted
by relatively efficient milling in the steel cylinder, its low radiogenic He concentration of
5.3·10−11 cm3 STP g−1
carb indicates a younger age compared to the stalagmite if we assume a
similar U content. The 1 300 yr old BU-1 sample contains 5.1 ·10−11 cm3 STP g−1
carb radiogenic He, which is very similar to the soda straw result. Using the empirically found relation
between radiogenic He per g of calcite and the speleothem age, the soda straw is most likely
of Holocene origin with an age between some 100 years and at most 3 kyr. Assuming the
same U and Th content than for the BU-U stalagmites an age of about 4 400 yr is obtained
based on the equations (4.5) - (4.10).
Several studies applying the noble gas thermometer in groundwater indicate that the last
glacial maximum in central and northern Europe was at least 5 ℃ and up to 9 ℃ colder
than the Holocene (Andrews and Lee, 1979; Stute and Deák, 1989; Beyerle et al., 1998;
Loosli et al., 2001). 7 ℃ colder than present temperatures are therefore plausible for the
Pleistocene including the Younger Dryas, but may seem rather low for the Preboreal. On
the other hand, Kloppmann et al. (1998) reported up to 8 ℃ cooler than present NGTs in
groundwater of the Paris Basin that was dated in the Early Holocene. Thus, the substantially
cooler than present NGTs from the BU-U stalagmite are not unreasonable. Additionally, as
we only have few dating points for this growth layer, the samples may originate from the
latest part of the Younger Dryas. Taylor et al. (1997) reported the transition time span to
be from 11 700 to 10 200 B.P. using the δ 18 O from the GSIP2 ice core.
4.6. Case studies
147
Table 4.10: Fitting results for the Bunker Cave speleothems. The age is determined for BU-U and BU1
by the U/Th method. Noble gas (equilibration) temperature (NGT), air/water volume ratio (A), and
total radiogenic He amount were derived by inverse modeling using the program ”noble” (Peeters et al.,
2003). Pressure was fixed according to the altitude of the cave of 180 m asl and zero salinity was assumed.
χ2 is the sum of the error-weighted squared deviations between model and data (Aeschbach-Hertig et al.,
1999).
Sample
BU-Ua
BU-U0
BU-U1
BU-U2
BU-U3
BU-U4
BU1
Soda Straw
10
10
10
10
10
10
Age (years)
800 - 11 700
800 - 11 700
800 - 11 700
800 - 11 700
800 - 11 700
800 - 11 700
1300
holocene
NGT (℃)
3.2 ± 0.9
2.4 ± 2.9
1.7 ± 0.7
3.6 ± 1.1
3.3 ± 0.7
3.5 ± 0.9
7.1 ± 0.8
6.4 ± 0.4
A(-)
0.043
0.045
0.040
0.035
0.030
0.026
0.081
0.0018
χ2 (-)
1.1
1.2
2.6
2.3
5.2
9.3
4.9
4.8
radiogenic He (10−10 ccSTP)
5.04
0.43
6.09
2.22
3.47
3.19
0.72
0.61
As the soda straw sample is assumed to be considerably younger than the BU-U stalagmite
samples, it makes sense that it yields a significantly higher NGT. Furthermore, its value is
in agreement with the result of BU-1. It seems to indicate a significant warming for the Late
Holocene relative to the Early Holocene represented by the older stalagmite samples. Yet,
compared to the present day cave temperature, the NGTs of the soda straw and the BU1
sample seem rather low.
Comparisons between NGTs and modern cave or air temperatures should be interpreted with
some caution. The NGTs derived from speleothem fluid inclusions are expected to correspond
to cave temperatures, but we have not yet been able to demonstrate this with samples from
the period of instrumental temperature records. Even if we had a recent stalagmite sample,
only mean annual air temperatures would be available for comparison, which probably differ
slightly from cave temperatures. For similar reasons, noble gas studies in groundwater usually
rely on temperature differences between NGTs from different periods and compare NGTs of
modern samples with air or soil temperatures only to support the validity of the method
(Andrews and Lee, 1979; Stute et al., 1992, 1995).
The reconstruction of European temperatures based on pollen data from over 500 sites of
Davis et al. (2003) indicates that major warming in western central Europe took place around
11 kyr BP. As the BU-U samples may be older than 11 kyr, their substantially cooler than
present NGTs are not unreasonable.
The NGT of the sample from BU-1 with an estimated U/Th age of 1 300 ± 300 years is
more than 2℃ lower than the modern air temperature. Indeed, the period around 700 AD
probably was relatively cool. The multi-proxy temperature reconstruction for the northern
hemisphere by Moberg et al. (2005) indicates 0.2 to 0.4 ℃ lower temperatures for this period
compared to 1961 - 1990. However, both regionally and episodically, temperatures may have
been substantially cooler. E.g., Holzhauser et al. (2005) report a major glacier advance in
the Alps during the time of 500 to 700 AD. Nevertheless, the NGT of the sample from
BU-1 appears rather low. The absolute accuracy of NGTs from fluid inclusions remains to
be tested with modern speleothem samples and direct comparison with cave temperatures.
Small offsets between NGT and cave temperature, e.g. due to unknown effects influencing
the noble gas composition in speleothems, or even to systematic analytical errors as e.g. an
underestimation of the released water amount, can presently not be ruled out completely.
Marx (2008) revealed crushed calcite to be a very efficient adsorbent for water vapour, which
may have reduced the detected water amount.
The relationship between cave and air temperature also needs to be studied for each cave.
148
Chapter 4. Results
9
8
7
temperature (˚C)
6
(4.2 +- 1.1) ˚C
5
4
3
2
1
0
0
2000
9000
10000
11000
12000
yr B.P.
Figure 4.35: Temperature differences for Bunker Cave in western central Europe between the Early
Holocene and Early Middle Ages as derived from noble gases in fluid inclusions. The BU-U samples are
arranged in the time interval between 10 800 and 11 700 yr BP according to their radiogenic He content.
Yet, the differences between NGTs obtained from individual samples from the same cave
should be less affected by such complications. Hence we consider the difference of (4.2 ±
1.1) ℃ between the 1 300 yr old BU-1 sample and the ≈ 11 kyr old BU-U samples to be the
most reliable palaeoclimatic result of this part of the study (Table 4.35). This result is in
good agreement with the temperature reconstruction of Davis et al. (2003), which indicates
about 4 ℃ cooler temperatures before 11 kyr compared to recent values in western central
Europe, where our site is located.
The most surprising result from this set of measurements is the highly reproducible temperature value of a growth layer from the stalagmite BU-U. Another indication for reasonable
temperature determination via noble gas measurements on fluid inclusions is the good agreement of the obtained temperature difference with a high resolution pollen study. Furthermore,
the low uncertainty for a soda straw sample of only about ± 0.4 ℃ shows that it is possible to
achieve high precision NGTs at least from selected speleothems. The precision of the NGTs
calculated from the BU-U stalagmite are mostly close to ± 1 ℃, which is easily sufficient to
resolve glacial-interglacial temperature changes and should, if a sufficient number of samples
is available, allow the detection of major temperature shifts within the Holocene. In the following we tried to apply the method on different stalagmites to obtain temperature records
along the growth axis.
4.6.2
Case study BU-1
In a first case study we focused on the stalagmite BU-1, which also originates from the Bunker
Cave. Due to its generally milky white appearance, we supposed a high water content and a
comparably low air contribution. A first test piece did approve this assumption and yielded
a precise temperature (s. precedent chapter). Therefore, we tried to measure more points on
this Holocene stalagmite and investigated it with regard to temperature changes during the
growth period.
4.6. Case studies
149
Bu1
1.3
4.7
4.9
6.8
7.0
7.3
7.7
Figure 4.36: Half of BU-1. Locations
of sampling for NGT determination are
marked with a red circle and anoted
with an age in kyr.
BU-1 is a relatively large (67 cm), but rather thin stalagmite (≈ 5 - 7 cm) with a candlestick type shape. Most of the speleothem consists of large columnar crystals. In some
periods, which are supposed to have been considerably drier, also dentritic fabric occurs.
The stalagmite has grown between 1 kyr and 8 kyr BP with an Eemian part at the bottom
(darker layers, s. Fig. 4.36). One hiatus is found between 1.8 and 4 kyr, which is located
in the very thin top part of the stalagmite. We extracted 7 samples, one above the hiatus,
two in the period of very fast growth at about 5 kyr, and 4 samples between 7 and 8 kyrs,
and measured them in two different runs. All samples are extracted by crushing in the steel
cylinder using a stepwise procedure consisting of some crushing steps and an optional heating
step. Details about the extraction procedure, sample weight and water content are given in
Table 4.11.
An interesting feature is the high water content of almost all pieces from BU1. Furthermore
the large differences in the released amount of water as well as the water content raise the
question about the origin of these deviations. A working hypothesis is discussed in chapter 4.2.5. We will now focus on the temperature information obtained from noble gases in
fluid inclusions. The fitting results are summarized in Table 4.12.
Using a stepwise procedure we have been able to reduce the air contribution in each step to
values of A of about 0.1 or significantly less (Fig 4.22). This low A was obtained in the heating
measurements. Fitting the noble gas concentrations of the heating extraction measurements
resulted in a quite good χ2 of mostly below 1. For 5 of the 7 samples the temperature
uncertainty is about 1.5 ℃ or better. However, this data is not hotblank corrected. It was
not possible to measure the hotblank during the measurement run. Thus, we performed a
set of totally 5 blank measurements at 150 ℃ using different steel crushers. Unfortunately,
the blank values are scattering strongly by about a factor of 5. Furthermore, the pumping
time varied in the case of the speleothem samples between at least 8 hours and 2 or 3 days
on week-ends. Therefore, the afterwards measured heating-blanks with a pumping time of
generally less than 1 day cannot precisely represent the values in the case of the real samples.
150
Chapter 4. Results
Table 4.11: Total sample weight, released water amount, apparent gravimetric water content, and
applied extraction method for the BU-1 stalagmite.
Sample
BU1-5
BU1-22
BU1-29
BU1-46
BU1-49
BU-51
BU-55
weight (g)
1.424
0.690
0.561
0.514
1.012
0.885
1.151
water amount (µl)
0.899 ± 0.015
2.351 ± 0.023
2.352 ± 0.023
0.590 ± 0.011
0.987 ± 0.014
2.794 ± 0.026
7.603 ± 0.052
water content % wt
0.63
3.40
4.19
1.15
0.98
3.16
6.61
extraction steps
3 strokes
5, 60 strokes, heat
5, 60 strokes, heat
5, 60 strokes, heat
5, 60 strokes, heat
5, 60 strokes, heat
5, 20 , 60 strokes, heat
Table 4.12: Fitting results for the BU-1 stalagmite. The age is determined from an age model established
by the U/Th method. Noble gas (equilibration) temperature (NGT) and air/water volume ratio (A) were
derived by inverse modeling using the program ”noble” (Peeters et al., 2003). Pressure was fixed according
to the altitude of the cave of 180 m asl and zero salinity was assumed. χ2 is the sum of the error-weighted
squared deviations between model and data (Aeschbach-Hertig et al., 1999). The first NGT-, A- and χ2 value refers to the not hotblank-corrected heating extraction. The second NGT (NGT-C) was obtained
by combinations of the according extraction steps and application of the blank corrections. The mean
value refers to the arithmetic mean.
Sample
BU1-5
BU1-22
BU1-29
BU1-46
BU1-49
BU1-51
BU1-55
mean
Age (years)
1 300 ± 300
4 700 ± 300
4 900 ± 300
6 800 ± 600
7 000 ± 500
7 300 ± 200
7 700 ± 300
holocene
NGT (℃)
7.1 ± 1.1
8.6 ± 1.5
4.1 ± 3.9
6.4 ± 2.9
7.7 ± 1.6
7.5 ± 0.8
6.9 ± 1.5
A(-)
0.036
0.054
0.107
0.061
0.017
0.024
0.050
χ2 (-)
0.3
0.6
0.1
4.6
0.4
0.3
1.1
NGT-C (℃)
7.1 ± 0.8
3.8 ± 1.1
7.3 ± 1.4
9.8 ± 5.6
7.9 ± 1.6
9.0 ± 1.1
5.9 ± 0.8
7.3 ± 2.0
used steps
3
60, heat
60, heat
60, heat
all
60, heat
60, heat
-
χ2 (-)
4.9
1.5
1.7
0.9
0.3
0.7
0.3
1.5
The most reliable results with low temperature uncertainty and good χ2 stem from the
heating measurements of BU1-22, BU1-51 and BU1-55. They are very close or equal to the
temperature obtained in the very first measurement of a BU1-piece, even if they represent
different periods within the Holocene. Another NGT with good χ2 refers to BU1-49. Putting
together the noble gases of all extraction steps of BU1-49 resulted in a temperature value of
7.9 ± 1.6 ℃. Calculating the arithmetic mean of these four samples and BU1-5 results in a
temperature of 7.5 ± 0.4 ℃. This is the most reliable result of this case study as it refers to
data points with best fitting results and lowest uncertainties.
Using the not-hotblank corrected data is justified by the similarity of their values. They correspond to the corrected data within their uncertainty (except BU1-22, Table 4.12, Fig. 4.37).
Similarly, the mean values calculated from both data sets are astonishingly consistent. However, in most cases the result of the combined fitting, including the 60 strokes and heating
measurement, is slightly worse than the χ2 of the hot extraction. This may be mainly due
to the applied blank correction in the case of the combined fitting. The hotblank correction
seems to be a little bit to strong, as in the case of tiny samples the resulting gas amounts
become negative. Furthermore, the uncertainty of the correction is quite large, which leads
to increased temperature uncertainties for the combined data set.
Although some more measurement have to be done in future on this stalagmite, first interpretations are possible using the most reliable data points (BU1-5, BU1-22, BU1-49, BU-51,
BU1-55). They result in an arithmetic mean value of 7.5 ± 0.4 ℃ and exclude temperature
variations larger than 2 ℃ within the last 8 kyr of the Holocene, at least in the investigated
periods. Ojala et al. (2008) found minor temperature variations for the high latitudes in Eu-
4.6. Case studies
151
16
combined values, hotblank corrected
raw data, not hotblank corrected
14
12
T (˚C)
10
mean:
7.3 +- 2.0 ˚C
8
mean:
6.9 +- 1.5 ˚C
6
4
2
0
0
10
20
30
40
50
60
depth (cm)
Figure 4.37: NGT from BU-1 samples. Red points refer to the results of the heating measurements (not
hotblank corrected), the black points to the fitting results of combined steps. The dotted line indicates
the mean values for each case.
rope during the last 8 kyr with variations of less than 1 ℃ from the mean annual temperature
in this period. Nevertheless, systematic trends are detected for summer temperatures as well
as for mean annual values from pollen- and varve-based reconstructions. This findings may
be only of regional importance. However, Viau et al. (2006) report on mean July temperature
anomalies in Northern America on the order of only 0.2 ℃ in the last 8 kyr based on more
than 750 fossil pollen records. The Holocene optimum July temperatures are found between
6 kyr and 3 kyr BP. Similarly, Ojala et al. (2008) stated the warmest period in Northern
Europe to be between 8 kyr and 4 kyr BP. Unfortunately, we can until now not resolve such
small temperature variations of some tenth of degrees, but all the mentioned studies suggest
temperature variations to be <1 ℃ during the Holocene growth period of BU-1. This is
also in agreement with the interpretation of the δ 18 O values from two Greenland ice cores.
Grootes et al. (1993) interpreted the small variations in the δ 18 O signals (1-2‰) to be an
indication for a remarkably stable climate during the Holocene. On a regional scale Davis
et al. (2003) investigated the temperature anomalies during the Holocene throughout Europe
and found especially in the region of the Bunker Cave very small changes in the last 7 kyr
(± 0.3 ℃). From 7 to 8 kyr BP they suggest a temperature decrease of about 1 ℃. Comparing the results of these studies with our own measurements, a rather good correlation can be
observed. The most reliable NGTs suggest the temperature to have been stable during the
investigated period. Nevertheless, variations in the range of ± 1 ℃ can not be excluded due
to the relatively large uncertainties of the single measurements.
Particularly interesting is a closer view on the stable isotope data and the comparison with
NGTs as well as with the water content as supposed precipitation proxy. The stable carbon
and oxygen isotope curves of BU-1 are displayed in Fig. 4.38. Apart from high frequency
fluctuations in the BU-1 stable isotope curve also a longterm trend can be found. One
maximum is located at the top at ≈ 1 kyr BP, the second at 6 kyr BP (∼ 40 cm from top)
and the last at 8 kyr BP (∼ 60 cm from top). A similar trend is obvious in the δ 13 C values.
152
Chapter 4. Results
-7
-5.0
-8
-5.5
d13C
d18O
-9
-6.0
-10
-6.5
-11
d180
25 point FFT smoothing
d13C
25 point FFT smoothing
-7.0
0
100
200
300
400
500
600
-12
700
depth (mm)
Figure 4.38: Stable oxygen and carbon isotope data of BU-1. Based on the single measurements a
curve has been fitted applying 25 points FFT smoothing. Each 300 µm isotopes have been measured.
The data have been measured in the framework of the DAPHNE-project and were provided by C. Spötl.
The major question is related to the origin of these fluctuations. In δ 18 O the variations
are small and in maximum only about 1 ‰. If we transferred this into temperature, as
it has been done e.g. by Friedman and O’Neil (1977)(-0.24 ‰ per ℃), it would indicate
a temperature change of about 4 ℃ from maximum (5 kyr) to minimum (1.3 kyr). The
most reliable NGT results, belonging to maxima as well as to minima periods of δ 18 O, can
certainly exclude such large temperature variation. The stable isotope signals are influenced
by different effects (s. e.g McDermott (2004)) and are therefore difficult to interpret without
additional information. δ 18 O in the calcite is influenced by changes in the precipitation source
and amount and may additionally be altered by different kinds of fractionation during calcite
precipitation (equilibrium vs. kinetic fractionation). Furthermore, the δ 18 O of the rain is
temperature dependent and can change also the oxygen isotopy of the calcite. However, the
temperature dependence can be site-specific and may have changed in the past. Therefore it
cannot be used to determine unambiguous temperatures only with the isotope information.
However, δ 18 O shows in some cases an interesting link to precipitation (Niggemann et al.,
2003; Fleitmann and Burns, 2004; Cruz et al., 2005; Wang et al., 2008). As the temperature
variations were expected to be small during the growth period of BU-1 it is not unreasonable,
that the isotope signal of BU-1 was strongly influenced by precipitation. The periods with
a more positive δ 18 O also may be warmer periods, but coincide well with wetter phases
reconstructed from pollen (Seppä and Birks, 2001). Similarly, the interpretation of the water
content as a proxy for precipitation confirms this assumption (Fig. 4.13). A low water
content is found for the more negative δ 18 O and significantly higher values in the case of the
less depleted oxygen. Even if this interpretation is not well constrained at the older stalagmite
part, the general agreement encourages the additional use of the speleothem water content.
The less depleted δ 18 O is possibly caused by a change in the atmospheric circulation pattern,
delivering precipitation not only from a different source but also in a higher quantity.
4.6. Case studies
153
Eem
15 14 13 12 11 10
MIS3
E. Holocene
9 8 6 5
4
3 2 1
Figure 4.39: Pieces cut out from the growth-axis of BU-U. A darker brownish layer separates the Early
Holocene from Marine Isotope Stage 3 (MIS3) and another dark layer MIS3 from the Eemian. The pieces
are numbered for referencing. The total length from sample no.1 to no.15 is about 77 mm.
Combining all results can provide a meaningful interpretation of the climate of the last
8 kyr in Northwestern Germany. Most likely, the temperature variations have been smaller
than 2 ℃. Thus the enhanced signal variations in δ 18 O may be provoked by changes in the
precipitation pattern and the atmospheric circulation. This examples show the importance
of reasonable temperature calculations via noble gases in fluid inclusions. Using the NGTs it
is possible to constrain the hypothesis or to exclude some possibilities, as e.g. in the case of
BU-1, temperature changes by 4 ℃ within the Holocene.
4.6.3
Case study BU-U
Noble gas measurements on BU-1 revealed rather stable temperatures during certain periods
of the Holocene. The determined NGTs agree with the expectations. As the assumed variations are quite small within the Holocene, it is rather difficult to confirm the small changes
with NGTs. However, the difference between Holocene and Glacial values should be easily
detectable. Therefore, we made a set of measurements on the stalagmite BU-U, which cover
the Early Holocene, Marine Isotope Stage 3 and also the Eemian (Fig. 4.31). The stalagmite
was cut into two halves, one for U/Th dating and one for stable isotope, trace element and
noble gas measurements. The latter have all been measured on one small slice (5 mm x
10 mm) cut out near the growth axis (Fig. 4.39). In summary, we extracted 15 samples,
seven from the Early Holocene, two from the period of Marine Isotope Stage 3, and six samples from the Eemian and measured them in two different runs. All samples were extracted
by crushing in the steel cylinder using a stepwise procedure consisting of some crushing steps
and an optional heating step. Details about the extraction procedure, sample weight and
water content are given in Table 4.13.
We will now focus on the temperature information gained from the noble gas measurements.
A summary of temperature values and the air-water volume ratios is given in Table 4.14. In
some cases the temperature uncertainty is quite high, which is due to background correction
and also due to tiny samples (mostly < 1 g, down to 0.4 g) as well as influenced by medium
air/water volume ratios. However, about half of the samples yield well defined temperatures
with an uncertainty below 2 ℃. In general, the model fits quite well to the measured noble
gas concentrations and therefore the χ2 is about 1 or even better. Only regarding one
measurement the obtained temperature is unlikely, as the χ2 is about 25 and indicates a bad
fit. The background correction for this sample was very strong for some noble gases and
perhaps led to an over-correction. Except of one speleothem piece, all samples resulted in
realistic temperatures above 0 ℃. To assess whether the results are reasonable, we have to
take into account the age of the growth periods. To get an impression of the temperature
development along the growth axis, the fitted temperatures are summarized in Fig. 4.40.
154
Chapter 4. Results
Table 4.13: Total sample weight, released water amount, apparent gravimetric water content, and
applied extraction method for the BU-U stalagmite
Sample
BUU-2
BUU-3
BUU-4
BUU-5
BUU-6
BUU-8
BUU-9
BUU-10
BUU-11
BUU-12
BUU-13
BUU-14
BUU-15
weight (g)
1.401
0.776
1.750
0.587
0.795
0.885
1.331
1.123
1.134
0.619
0.680
0.727
0.397
water amount (µl)
0.381 ± 0.008
0.290 ± 0.008
0.509 ± 0.010
0.064 ± 0.006
0.080 ± 0.004
0.218 ± 0.013
0.141 ± 0.008
0.748 ± 0.012
0.759 ± 0.012
0.435 ± 0.009
0.507 ± 0.011
0.442 ± 0.008
0.162 ± 0.008
water content % wt
0.27
0.37
0.29
0.11
0.10
0.25
0.11
0.67
0.67
0.70
0.75
0.61
0.41
extraction steps
50 strokes, 100 + heat
60 strokes, 100 + heat
7, 60, 160 strokes
60 strokes, 100 + heat
60 strokes, 100 + heat
5, 60 strokes, heat
60 strokes, 100 + heat
7, 60, 160 strokes
7, 60, 160 strokes
60 strokes, 100 + heat
5, 60 strokes, 50 + heat
60 strokes, 100 + heat
60 strokes, 100 + heat
Table 4.14: Fitting results for the BU-U stalagmite. The age was determined by the U/Th method
on several points within the different growth layers. Noble gas (equilibration) temperature (NGT) and
air/water volume ratio (A) were derived by inverse modeling using the program ”noble” (Peeters et al.,
2003). Pressure was fixed according to the altitude of the cave of 180 m asl and zero salinity was assumed.
χ2 is the sum of the error-weighted squared deviations between model and data (Aeschbach-Hertig et al.,
1999). NGTs were obtained by combinations of the according extraction steps.
Sample
BUU-2
BUU-3
BUU-4
BUU-5
BUU-6
BUU-8
BUU-9
BUU-10
BUU-11
BUU-12
BUU-13
BUU-14
BUU-15
Age (years)
10 800 - 11
10 800 - 11
10 800 - 11
10 800 - 11
10 800 - 11
52 900 ± 1
52 900 ± 1
125 000 - 134
125 000 - 134
125 000 - 134
125 000 - 134
125 000 - 134
125 000 - 134
700
700
700
700
700
300
300
000
000
000
000
000
000
A(-)
NGT (℃)
used steps
χ2 (-)
0.106
0.131
0.076
0.272
0.204
0.102
0.227
0.236
0.255
0.0
0.667
0.281
0.293
1.0 ± 0.9
4.2 ± 1.3
3.2 ± 1.4
0.0 ± 6.8
<0℃
4.0 ± 4.1
0.9 ± 3.8
6.0 ± 2.3
11.1 ± 3.0
5.0 ± 1.7
8.7 ± 5.3
5.6 ± 1.8
2.5 ± 4.0
all
strokes
strokes
strokes
strokes
all
strokes
strokes
strokes
heat
all
all
strokes
0.2
0.0
0.7
0.4
0.2
0.1
0.5
1.4
1.5
25.1
0.3
5.7
0.3
60
60, 160
60
60
60
60, 160
60, 160
60
4.6. Case studies
155
12
10
NGT (˚C)
8
6
4
2
0
< 0˚C
MIS3
Early Holocene
Eem
-2
0
10
20
30
40
depth (mm)
50
60
70
80
isotope-axis
Figure 4.40: NGTs derived from the growth axis of BU-U. In the early Holocene part samples are
added from the first measurement campaign focused on this single layer (chapter 4.6.1).
The Early Holocene is characterized by a mean temperature of about 3 ℃. Only BUU-2
and BUU-6 are significantly lower. BUU-5 also shows a low temperature, but can not be
distinguished from the mean due to the large uncertainty. Taking into account the previous measurements on this growth layer (s. chapter 4.6.1) the arithmetic mean value is
(2.9 ± 1.0)℃. For this calculation the NGT below 0 ℃ (BU-6) as well as BU-5 were discarded. The temperatures from MIS3 can not be distinguished from values from the Early
Holocene due to the rather large uncertainties. This is also reflected in the mean value of
(2.5 ± 2.2)℃. The Eemian differs totally from the other two periods. There we did not
find constant temperatures, but rather a temperature evolution with its highest value at
∼ 130 kyr BP.
Until now we have not been able to demonstrate that the determined temperatures are exactly
identical with the expected temperatures during growth, as we have not yet examined recent
calcite precipitates. The mean annual air temperature in the region around the Bunker
Cave was 9.5 ℃ (1960 to 1990). Today, a temperature of about 10.5 ℃is measured in the
cave. Several measurements on the stalagmite BU-1 resulted in a mean value of about 7 ℃.
Many studies suggest only minor temperature variations during the growth of BU-1 (the last
8 kyr). Thus, our relatively low NGTs from BU-1 compared to the cave temperature indicate
a systematic deviation from the expected values. Recent tests lead to the assumption that
incomplete water collection due to adsorption on the crushed stalagmite is responsible for
this deviation. However, as we did not substantially change the extraction process a similar
systematic offset should occur for all samples. Thus, the temperature differences between the
growth periods of BU-U are the most reliable results.
As a Holocene reference temperature we use the NGT of (7.5 ± 0.4)℃ (chapter 4.6.2). Assuming that the 1960-1990 period to be representative for the Holocene, the systematic deviation
of the speleothem NGT to annual air temperatures is (2.0 ± 0.6)℃. With regard to the
156
Chapter 4. Results
palaeotemperatures from the Bunker Cave speleothems, the Early Holocene is (4.6 ± 1.1) ℃
cooler than present day values, that means at about (4.9 ± 1.1) ℃. This is in good correspondence with pollen data from Central Europe (Davis et al., 2003). Using the two data points
from Marine Isotope Stage 3, this period was (5.0 ± 2.3)℃ cooler than today, resulting in a
mean annual air temperature of (4.5 ± 2.5) ℃. Due to few samples from this part and their
unfavorable properties, the temperature uncertainty is quite high and thus the mean can not
be distinguished from the Early Holocene values. Another stalagmite from the same cave,
BU-2, has a larger growth layer corresponding to MIS3. From this layer (about 51 kyr old)
we measured one piece, resulting in a NGT of (4.85 ± 0.57)℃. This is warmer than the NGT
mean of the two samples from BU-U. As the MIS3 layer of BU-U was very small, the single
values may have been affected by a mixture with other layers. However, taking into account
the large uncertainty of 2.2 ℃, the BU-2 value is in agreement with the BU-U mean (2.5 ℃).
Swann et al. (2005) investigated diatoms, C/N ratios and organic carbon isotope ratios in
lake Baikal and detected that the period between 54 and 51.5 kyr BP to had relatively warm
interstadial conditions. This corresponds to the interstadial 14 (GIS 14), detected also in
a high alpine stalagmite, which was dated to last between 54.2 and 51 kyr BP (Spoetl and
Mangini, 2002). They state the GIS 14 to have been the longest D/O warm phase during
MIS3. As the stalagmite originates from a high alpine cave (≈ 2165 m asl) with an interior present-day cave temperature of 2.3 ℃, temperatures during GIS 14 should not have
been significantly below contemporary values, at least not significantly below 0 ℃, to enable
speleothem growth. Cabioch and Ayliffe (2001) detected relatively high sea levels at about
50 kyr BP from coral sample measurements. They estimate the sea level to have been only
30 - 60 m below contemporary values, in contrast to 110 to 130 m below present-day values
in the last glacial maximum. Similar features have also been found in the stable oxygen
isotope record of Greenland ice cores (Johnsen et al., 2001). Converting the ice-core data
into temperature the GIS 14 is about 7 ℃ cooler than Holocene values. The difference to
the last glacial maximum is about 20 ℃. In Central Europe the last glacial maximum was
5 - 9 ℃ cooler (Andrews and Lee, 1979; Stute and Deák, 1989; Beyerle et al., 1998; Loosli
et al., 2001). Using a mean value of 7 ℃ and assuming a similar relation for the GIS 14, the
downscaling of the Greenland temperature differences suggested the GIS 14 temperature to
be 2.5 ℃ cooler compared to present-day values in Central Europe. This value corresponds to
the mean of the two BU-U pieces from the GIS14 taking into account the large uncertainty.
However, it would fit perfectly to the measurement on a single piece of the BU-2 stalagmite,
which suggests a temperature of (2.6 ± 0.8)℃ below present-day values. Again, the stalagmite NGTs show the possibility to better constrain the continental climate and especially the
temperature. Despite the rather few measurements with relatively large uncertainties, the
NGT range of the GIS14 can be fixed to 4.1 - 5.3 ℃ calculating the weighted mean of the
two values from BU-U and the measurement of BU-2. Taking into account the systematic
offset between NGT and mean annual air temperature, the mean air temperature value for
GIS 14 in the Bunker cave region is (7.2 ± 0.8) ℃.
Another very interesting palaeoclimatic epoch is the Eemian period, which is also contained
in stalagmite BU-U. There we found considerably higher temperatures compared to GIS 14
and the Early Holocene (Fig.4.40). The temperature values of the three periods contained in
BU-U follow the expected temperature trend during the last Glaciation known e.g. from ice
core data (Petit et al., 1999; Kawamura et al., 2007). Assuming the two uppermost points to
be representative for the Eemian maxium the temperature is (10.5 ± 2.6) ℃ calculated via
a weighted mean of NGTs. Unfortunately, so far we only have few NGTs to better constrain
this value indicating a ≈ 3 ℃ warmer Eemian period (absolute value: 12.5 ℃) compared to
contemporary values. Kukla et al. (2002) stated the Eemian to have been at least as warm as
the present climate, based on coral, deep-sea sediment, ice-core, speleothem and pollen data.
4.6. Case studies
157
-4
d18O
-5
-6
-7
d18O
linear fit
-8
-5
0
5
10
15
temperature (˚C)
Figure 4.41: NGTs from the growth axis of BU-U compared to ageraged δ 18 O-values. The δ 18 O was
calculated by averaging the isotope values in the range of the corresponding NGT sample. The uncertainty
in δ 18 O was determined by variation of the interval in consideration of the uncertainty in the location of
the NGT sample (typically 2 mm).
16
-5.5
-4
d18O
-6.0
14
-6.5
12
-7.0
-5
-7.5
0
2
4
6
8
10
12
14
10
8
-6
6
4
-7
NGT [˚C]
d18 O [‰]
NGT (˚C)
2
-8
0
18
d O
NGT
-2
-9
0
20
40
60
80
Distance from top [mm]
Figure 4.42: NGTs from the growth axis of BU-U compared to the continuous δ 18 O record. In the
insert the correlation between the NGT and δ 18 O in the Eemian period is displayed.
158
Chapter 4. Results
-4.5
-5.0
III
-5.5
II
d13C
-6.0
-6.5
-7.0
I
-7.5
-8.0
-2
0
2
4
6
8
10
12
14
temperature (˚C)
Figure 4.43: NGTs from the growth axis of BU-U compared to δ 13 C-values. The δ 13 C was calculated
by averaging the isotope values in the range of the corresponding NGT sample. The uncertainty in δ 13 C
was determined by variation of the interval in consideration of the uncertainty in the location of the NGT
sample (typically 2 mm). The data points of section I refer to the Early Holocene, II to the Eemian and
III to GIS 14, respectively the beginning of the Eemian.
For the western Central Europe Zagwijn (1996) stated the winter maximum to be 3 ℃ and
the summer maximum to be 2 ℃ above present day values in this region. This fits very well to
our speleothem measurements and the determined NGTs. Furthermore, the interpretation of
the pollen data with regard to precipitation correspond also to the interpretation of the water
content in BU-U. Zagwijn (1996) claims the beginning of the Eemian to have been relatively
dry with increasing precipitation and later a generally rather oceanic climate with higher
precipitation than in the Holocene. A nicely corresponding trend can be drawn from the
water content, including significantly higher precipitation in the Eemian maximum compared
to early Holocene values (Fig. 4.18). Klotz et al. (2003) detected from pollen reconstruction
in the northern alpine foreland also a more oceanic climate with warmer and more humid
periods in the Eemian. Similarly, Aalbersberg and Litt (1998) inferred from pollen data that
the summer temperatures were several degrees higher in Southern England than present-day
values. Felis et al. (2004) used coral records in combination with a coupled atmosphereocean circulation model to infer the influence of the North Atlantic Oscillation during the
last interglacial period. The climate model produced for the Bunker Cave region a winter
temperature with little deviation from modern values, but up to 1.5 ℃ warmer summer values.
In summary, a warmer Eemian (in comparison with Holocene temperatures) as derived from
NGTs of the Bunker Cave speleothem BU-U is reasonable and shows similar trends like other
climate proxies.
Using the temperature information from the noble gas measurements on the fluid inclusions,
we are now able to better interpret the stable isotope curves (Fig. 4.44). Comparing the
NGTs with the δ 18 O-values of the same periods does not reveal any significant correlation
with temperature in the Early Holocene and GIS 14 (Fig. 4.41). The two results from GIS 14
are slightly separated from the rest (two uppermost points). Focusing only on the Eemian
period a correlation between δ 18 O and temperature is visible (Fig. 4.42). Higher temperatures
4.6. Case studies
159
0
-5
-2
-6
-4
-7
-6
-8
-8
d18O
d13C
-4
-10
-9
0
20
40
60
80
depth (mm)
Figure 4.44: δ 13 C and δ 18 O values along the growth axis of BU-U. The stable isotopes have been
provided by Dr. C. Spötl.
refer to a less depleted oxygen value, whereas lower temperatures are correlated with a more
negative δ 18 O value. Calculating a mean δ 18 O value for the periods covered by the NGT
samples and comparing them with the temperature values suggest a high correlation for this
epoch with a correlation coefficient of R=0.97. The slope of this relation is 0.22 ± 0.03 ‰
per ℃. A clearly distinct pattern can be found in the case of the δ 13 C-values. Although, again
there is no simple relationship with temperature, the points are arranged according to the
corresponding growth period: I Early Holocene, II Eemian and III intermediate points from
GIS 14 and the beginning of the Eemian. It looks like a step function with two different stable
states. Between 4.2 and 4.5 ℃ the system jumps between the two states. The Early Holocene
is characterized by a mean value of -(7.75 ± 0.06)‰ and the Eemian by -(6.36 ± 0.08)‰. The
reason for such pronounced differences is most likely a change of vegetation above the cave.
The δ 18 O-values seems, in some periods, to correlate very well with temperature (Fig. 4.42),
but can furthermore show a certain agreement with the trend in the water content. In
summary, combined studies with determination of NGTs from fluid inclusions and additional
stable isotope measurements can help to establish reasonable climate reconstructions from
speleothem data.
160
Chapter 4. Results
Chapter 5
Summary and outlook
Summary
Before summarizing the results of the main objective, temperature determination via noble
gas concentrations, the major advances in the related topics will be discussed.
As a first step, we investigated different methods to select suitable samples in advance. Using
microscopic analysis, information about the inclusion size and distribution as well as about
the fabric can be obtained. We found speleothems with the best properties to be of milky
white appearance. This type of speleothem yields a high number of water-filled inclusions
and rather few air-filled ones.
For the extraction of water and noble gases from the speleothems we tested several methods:
squeezing in copper tubes, milling in a steel crusher and microwave treatment in evacuated
vials. The microwave based extraction is not yet investigated in all details, but the first tests
did not fulfil the expectations. Squeezing in copper tubes is not as efficient as crushing in a
steel cylinder and, furthermore, not completely reproducible. The best results can be obtained
using the steel crusher with a magnetically movable steel ball. This setup enables also stepwise
extraction, which is of high importance for a large group of speleothems. Adapted to the
extraction in the steel cylinder we constructed a special extraction line, including a setup
for ultra-high vacuum preparation, pressure recording in different ranges (10−9 -10 mbar),
to control the sample preparation, noble gas extraction and the water determination via
pressure, and three inlets with parallel pumping lines for higher sample throughput (at least
one sample per day). Based on the results of the methodological tests, we extracted the
samples in the steel crusher using different steps, e.g. 5 strokes followed by 60 strokes and an
optional heating step. The water liberated from the sample was then frozen into a cold finger
and determined via measurement of the water vapour pressure in different volumes. The
noble gases were measured in a static sector-field mass spectrometer by comparison with a
diluted air standard. The achievable precision was 2 % in the case of the water determination,
1 % for the He, Ne, Ar and Kr measurement and 2% for the Xe measurement.
We applied the described extraction method to a set of speleothems from Bunker Cave, but
also to stalagmites from Chile, Canada, Cuba, Oman and Austria. The most precise results
have been achieved for the speleothems from the Bunker Cave. Multiple measurements on one
growth layer of BU-U yielded temperatures (mean 2.9 ℃) agreeing within their uncertainties.
Furthermore, the typical uncertainty was only about 1 ℃. The NGT of an undated soda straw
from the Bunker Cave even was only charged with an uncertainty of 0.4 ℃ due to the very low
air contribution to the noble gas signal. Several pieces from the younger Holocene stalagmite
BU-1 resulted in coinciding noble gas temperatures of about 7 ℃. From the stalagmite BU-2
the measurement of a single piece resulted in a reasonable noble gas temperature of about
162
Chapter 5. Summary and outlook
5 ℃ for the according growth period (GIS 14). The NGTs obtained along the growth axis
of BU-U revealed an interesting pattern, which is in agreement with the trend of the polar
ice cores as well as marine and continental recores (pollen), and shows the high potential
of this method. However, most of the other investigated speleothems yielded far too high
contributions of noble gases from air-filled inclusions. They mask the temperature signal and
make it difficult or impossible to obtain reasonable temperatures from a simple extraction
method. Large air-filled inclusions are located between the grain boundaries. Therefore, it
is possible to reduce the fraction of air-derived noble gases by stepwise crushing.Mainly, the
inter-granular inclusions were opened in the first steps, in the subsequent steps the smaller
water-filled inclusions were increasingly affected. Applying this technique to a speleothem
from Oman (H12) with a high air-water volume ratio using simple extraction, a reasonable
temperature (24 ℃) in agreement with the cave temperature (23 ℃ - 26 ℃) and an uncertainty
of 4 ℃ was achieved. Further developments suggest the possibility to extend the method to
samples with unfavourable properties and to further reduce the uncertainty.
The water content seems to reflect a certain correlation with precipitation and may deliver
in some cases information about the palaeo-humidity. Based on published literature and
own measurements a working hypothesis is presented for this correlation, namely that higher
drip rates and faster stalagmite growth lead to more water-filled inclusions. Additionally,
the radiogenic He can help to constrain the age estimates for very old speleothems. The low
diffusivity (2.4·10−25 m2 /s at 30 ℃) supposes that quantitative extraction and measurement
of He may enable the U/Th-He dating for calcite. Preliminary data showed an encouraging
correlation between age and radiogenic He content.
In summary, a large set of successful gas extractions and measurements on speleothems
proved the possibility of NGT determination using fluid inclusions in speleothems. All temperature differences obtained from different stalagmites from the Bunker Cave agreed with
temperature changes stated in other studies. Not only the changes from the glaciation to interstadials correspond to the expected temperatures, but also the NGTs within the Holocene.
They constrain the expected climatic stability in the last 8 kyr and give an upper limit for
temperature changes of ± 2 ℃ from the value of the 1960-1990 reference period. Even though
this project was started only three years ago, already a large set of palaeoclimatically interesting conclusions can be drawn, also some beside the NGT information as e.g. of the water
content. Improving this methods will certainly help to deliver new insights into the past
climate, especially in combination with the stable isotope and trace element data.
Outlook
In the next measurement run tests with recent calcite precipitates or artificial speleothems
should be performed. They could help to investigate the absolute accuracy of the NGTs from
fluid inclusions in speleothems. Subsequently, a high resolution record may be established
on a stalagmite or a set of speleothems covering the Holocene as well as glacial periods. The
combination with stable isotopes and trace elements will not only help to disentangle different
effects, which cause the changes in the stable isotope values, but furthermore deliver a more
complete and also more precise picture of the past climate.
Some effort has to be made to further reduce the background in case of the hot extractions.
However, the combination of this step with precedent crushing is supposed to be most promising. Optimization of pumping times as well as baking out of the crushing cell may help to
diminish the disturbing background signal.
163
First analysis of the water content suppose this value to be a proxy for palaeo-humidity.
Intensive measurement of other stalagmites has to be performed to verify this assumption.
To establish the radiogenic He as a dating tool, the extraction procedure has to be modified
towards a complete release of all noble gases. This may be achieved by a strong heating step
at temperatures above 600 ℃.
This work intended to investigate the feasibility of NGT determination on speleothem fluid
inclusions. Summarizing all results shows that NGT calculation is possible and can yield low
uncertainties down to values like in groundwater studies. This conclusion should encourage
further work in this exciting field which, in all probability, will deliver important data for
palaeoclimate studies.
164
Chapter 5. Summary and outlook
Bibliography
Aalbersberg, G. and Litt, T.: Multiproxy climate reconstructions for the Eemian and Early
Weichselian, Journal of Quarternary Science, 13(5), 367–390, 1998.
Abbott, W., Wolfe, B. B., Aravena, R., Wolfe, A. P., and Seltzer, G. O.: Holocene hydrological reconstructions from stable isotopes and paleolimnology, Cordillera Real, Bolivia,
Quarternary Science Reviews, 19, 1801–1820, 2000.
Aeschbach-Hertig, W.: Helium and Tritium as Tracers for Physical Processes in Lakes, Ph.D.
thesis, Swiss Federal Institute of Technology, Zurich, 1994.
Aeschbach-Hertig, W. and Kluge, T.: On the feasibility of applying the noble gas thermometer
to fluid inclusions in speleothems, Workshop Daphne, Innsbruck, abstract, 2006.
Aeschbach-Hertig, W., Peeters, F., Beyerle, U., and Kipfer, R.: Interpretation of dissolved
atmospheric noble gases in natural waters, Water Resources Research, 35(9), 2779–2792,
1999.
Aeschbach-Hertig, W., Peeters, F., Beyerle, U., and Kipfer, R.: Palaeotemperature reconstruction from noble gases in ground water taking into account equilibration with entrapped
air, Nature, 405, 1040–1044, 2000.
Aeschbach-Hertig, W., Stute, M., Clark, J. F., Reuter, R. F., and Schlosser, P.: A paleotemperature record derived from dissolved noble gases in groundwater of the Aquia Aquifer
(Maryland, USA), Geochimica et Cosmochimica Acta, 66(5), 797–817, 2002.
Alley, R. and Ágústsdóttir, A.: The 8k event: cause and consequences of a major Holocene
abrupt climate change, Quaternary Science Reviewsk, 24(10-11), 1123–1149, 2005.
Alley, R., Mayewski, P., Sowers, T., Stuiver, M., Taylor, K., and Clark, P.: Holocene climatic
instability; a prominent, widespread event 8200 yr ago, Geology, 25(6), 483–486, 1997.
Andrews, J. N. and Lee, D. J.: Inert gases in groundwater from the Bunter Sandstone of
England as indicators of age and paleoclimatic trends, Journal of Hydrology, 41, 233–252,
1979.
Andrié, C., Jean-Baptiste, P., and Merlivat, L.: Tritium and helium 3 in the northeastern
Atlantic ocean during the 1983 Topogulf cruise, Journal of Geophysical Research , 93(C10),
12 511–12 524, 1988.
Auer, A., Böhm, R., and Schöner, W.: Austrian long-term climate 1767 to 2000: Multiple instrumental climate time series from Central Europe, Österreichische Beiträge zu
Meteorologie und Geophysik, Heft 25, Publ. 397, Zentralanstalt für Meteorologie und Geodynamik, Vienna, Austria, 2001.
165
166
Bibliography
Badertscher, S. V.: Charakterisierung von Einschlüssen in Stalagmiten zur Bestimmung der
Paläotemperatur, diploma thesis, Swiss Federal Institute of Technology, Zürich, 2007.
Badertscher, S. V., Scheidegger, Y., Leuenberger, M., Nyfeler, P., Fleitmann, D., Wieler, R.,
and Kipfer, R.: Trace gas content in air inclusions in speleothems as new paleoenvironmental archive?, Geophysical Research Abstracts, 9, 06 252, 2007.
Baker, A. and Smart, P. L.: Recent flowstone growth rates: Field measurements in comparison to theoretical predictions, Chemical Geology, 122, 121–128, 1995.
Baker, A., Barnes, W. L., and Smart, P. L.: Variations in the discharge and organic matter
content of stalagmite drip waters in Lower Cave, Bristol, Hydrol.Process, 11, 1541–1555,
1997.
Baker, A., Genty, D., Dreybrodt, W., Barnes, W. L., Mockler, N. J., and Grapes, J.: Testing
theoretically predicted stalagmite growth rate with recent annually laminated samples:
Implications for past stalagmite deposition, Geochimica et Cosmochimica Acta, 62(3), 393–
404, 1998.
Baker, P. A., Seltzer, G. O., Fritz, S. C., Dunbar, R. B., Grove, M. J., Tapia, P. M., Cross,
S. L., Rowe, H. D., and Broda, J. P.: The History of South American Tropical Precipitation
for the Past 25,000 Years, Science, 291, 640–643, 2001.
Baldini, J. U. L., McDermott, F., and Fairchild, I. J.: Spatial variability in cave drip water
hydrochemistry: Implications for stalagmite paleoclimate records, Chemical Geology, 235,
390–404, 2006.
Ballentine, C. J. and Burnard, P. G.: Production, Release and Transport of Noble Gases
in the Continental Crust, In: Porcelli, D., C.J. Ballentine, R. Wieler(Eds) Noble gases in
geochemistry and cosmochemistry, vol. 47, pp. 481–538, 2002.
Ballentine, C. J. and Hall, C.: Determining paleotemperature and other variables by using
an error-weighted, nonlinear inversion of noble gas concentrations in water, Geochimica et
Cosmochimica Acta, 63(16), 2315–2336, 1999.
Barrer, R. M., ed.: Diffusion in and through solids , Cambridge University Press, 1941.
Barron, E. J.: A Warm, Equable Cretaceous: the Nature of the Problem, Earth-Science
Reviews, 19, 305 338, 1983.
Bender, M. L.: Helium-uranium dating of corals, Geochimica et Cosmochimica Acta, 37,
1229–1247, 1973.
Beyerle, U., Purtschert, R., Aeschbach-Hertig, W., Imboden, D., Loosli, H. H., Wieler, R.,
and Kipfer, R.: Climate and groundwater recharge during the last glaciation in an icecovered region, Science, 282, 731–734, 1998.
Beyerle, U., Aeschbach-Hertig, W., Imboden, D. M., Baur, H., Graf, T., and Kipfer, R.: A
mass spectrometric system for the analysis of noble gases and tritium from water samples,
Environmental Science and Technololgy, 34, 2042–2050, 2000.
BMLFUW: Bundesministerium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft:
Jahresbericht 2002 für die Wassergüte in Österreich, Wien, 2002.
Bodnar, R. J. and Zolensky, M. E.: Fluid inclusions in meteorites. Are they useful, and why
are they so hard to find?, Meteoritics and Planetary Science, 35, A29–A30, 2000.
Bibliography
167
Böhlke, J. K. and Irwin, J. J.: Laser microprobe analyses of noble gas isotopes and halogens in
fluid inclusions: Analyses of microstandards and synthetic inclusions in quartz, Geochimica
et Cosmochimica Acta, 56, 187–201, 1992a.
Böhlke, J. K. and Irwin, J. J.: Brine history indicated by argon, krypton, chlorine, bromine,
and iodine analyses of fluid inclusions from the Mississippi Valley type lead-fluorite-barite
deposits at Hansonburg, New Mexico, Earth and Planetary Science Letters, 110, 51–66,
1992b.
Brázdil, R., Pfister, C., Wanner, H., Von Storch, H., and Luterbacher, J.: Historical Climatology In Europe, The State Of The Art, Climate Change, 70(3), 363–340, 2005.
Bronstein, I. N., Semendjajew, K. A., Musiol, G., and Mühlig, H., eds.: Taschenbuch der
Mathematik , vol. 5, Verlag Harri Deutsch, 2000.
Brook, G. A., Rafter, M. A., Railsback, L. B., Sheen, S., and Lundberg, J.: A high-resolution
proxy record of rainfall and ENSO since AD 1550 from layering in stalagmites from Anjohibe Cave, Madagascar, The Holocene, 9(6), 695–705, 1999.
Brunauer, S., Emmet, B. H., and Teller, E.: Adsorption of gases in multimolecular layers,
Jounal of the American Chemical Society, 60(2), 309–319, 1938.
Buhmann, D. and Dreybrodt, W.: The kinetics of calcite dissolution and precipitation in
geologically relevant situations of karst areas - 1.open systems, Chemical Geology, 48, 189–
211, 1985.
Bunnell, D.: http://en.wikipedia.org/wiki/Image:Labeled speleothems.jpg, 2006.
Burnard, P. G., Stuart, F., and Turner, G.: C-He-Ar variations within a dunite nodule as a
function of fluid inclusion morphology, Earth and Planetary Science Letters, 128, 243–258,
1994.
Burns, S. J., Fleitmann, D., Matter, A., Neff, U., and Mangini, A.: Speleothem evidence from
Oman for continental pluvial events during interglacial periods., Geology, 29(7), 623–626,
2001.
Cabioch, G. and Ayliffe, L. K.: Raised Coral Terraces at Malakula, Vanuatu, Southwest
Pacific, Indicate High Sea Level During Marine Isotope Stage 3, Quarternary Research, 56,
357–365, 2001.
Cable, J. E., Bugna, G. C., Burnett, W. C., and Chanton, J. P.: Application of 222 Rn
and CH4 for assessment of groundwater discharge to the coastal ocean, Limnology and
Oceanography, 41, 1347–1353, 1996.
Clement, A. C., Seager, R., and Cane, M.: Suppression of El Niño during the Mid-Holocene
by Changes in the Earths Orbit, Paleoceanography, 15(6), 731–737, 2000.
Clever, H. L., ed.: Krypton, xenon and radon - gas solubilities. Solubility data series, vol. 2,
Pergamon Press, Oxford, 1979.
Cook, P. G., Love, A. J., Robinson, N., and Simmons, C. T.: Groundwater ages in fractured
rock aquifers, Journal of Hydrology, 308, 284 – 301, 2005.
Copeland, P., Watson, E. B., Urizar, S. C., Patterson, D., and Lapen, T. J.: Alpha thermochronology of carbonates, Geochimica et Cosmochimica Acta, 71, 4488 – 4511, 2007.
168
Bibliography
Corbett, R., Dillon, K., Burnett, W. C., and Chanton, J.: Estimating the groundwater contribution into Florida Bay via natural tracers, Rn-222 and CH-4, Limnology and Oceanography, 45, 1546 – 1557, 2000.
Crank, J., ed.: The mathematics of diffusion, vol. 2, Clarendon Pr., Oxford, 1994.
Cruz, F. W., Burns, S. J., Karmann, I., Sharp, W. D., Vuille, M., Cardoso, A. O., Ferrari,
J. A., Silva Dias, P. L., and Viana, O.: Insolation-driven changes in atmospheric circulation
over the past 116,000 years in subtropical Brazil, Nature, 434, 63 – 66, 2005.
Cruz, F. W., Burns, S. J., Jercinovic, M., Karmann, I., Sharp, W. D., and Vuille, M.:
Evidence of rainfall variations in Southern Brasil from trace element ratios (Mg/Ca and
Sr/Ca) in a Late Pleistocene stalagmite, Geochimica et Cosmochimca Acta, 71, 2250 –
2263, 2007.
Curl, R. L.: Minimum diameter stalagmites, Bulletin National Speleological Society of America (Nat. Speleol. Soc. Bull), 35, 1–9, 1973.
Dansgaard, W., Johnson, S. J., Clausen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer,
C. U., Hvidberg, C. S., Steffensen, J. P., Svelnbrörnsdottir, A. E., Jouzel, J., and Bond,
G.: Evidence for general instability of past climate from a 250 kyr ice-core record, Nature,
364, 218–220, 1993.
Davis, B. A. S., Brewer, S., Stevenson, A. C., Guiot, J., and contributors, D.: The temperature of Europe during the Holocene reconstructed from pollen data, Quaternary Science
Reviews, 22, 1701–1716, 2003.
Dennis, P. F., Rowe, P. J., and Atkinson, T. C.: The recovery and isotopic measurement
of water from fluid inclusions in speleothems, Geochimica et Cosmochimica Acta, 65(5),
871–884, 2001.
DWD: Deutscher Wetterdienst, webpages: s. www.dwd.de, 2008.
Edwards, R. L., Cheng, H., Murrell, M. T., and Goldstein, S. J.: Protactinium-231 Dating
of Carbonates by Thermal Ionization Mass Spectrometry: Implications for Quaternary
Climate Change, Science, 276, 782–786, 1997.
Fairchild, I., Frisia, S., A., B., and Tooth, A.: Speleothems, In: Nash, D.J. and McLaren,
S.J.(Eds): Geochemical Sediments and Landscapes, Blackwells, Oxford, vol. 47, 2007.
Fanale, F. and Kulp, J. L.: Helium in Limestone and Marble, The American Mineralogist,
46, 155–167, 1961.
Farley, K. A.: (U-Th)/He Dating: Techniques, Calibrations, and Applications, In: Porcelli, D., C.J. Ballentine, R. Wieler(Eds) Noble gases in geochemistry and cosmochemistry,
vol. 47, 2002.
Farley, K. A., Kohn, B. P., and Pillans, B.: The effect of secular disequilibrium on (U-Th)/He
systematics and dating of Quarternary volcanic zircon and apatite, Earth and Planetary
Science Letters , 201, 117–125, 2002.
Felis, T., Lohmann, G., Kuhnert, H., Lorenz, S. J., Scholz, D., Pätzold, J., Al-Rousan, A.,
and Al-Moghrabi, S. M.: Increased seasonality in Middle East temperatures during the
last interglacial period, Nature, 429, 164–168, 2004.
Bibliography
169
Fleitmann, D. and Burns, S. J.: Palaeoclimatic interpretation of high-resolution oxygen
isotope profiles derived from annually laminated speleothems from Southern Oman, Quarternary Science Reviews , 23, 935–945, 2004.
Fontes, J. C. and Garnier, J. M.: Determination of the initial 14 C activity of the total
dissolved carbon: A review of the existing models and a new approach, Water Resources
Research , 15, 399, 1979.
Ford, D. and Williams, P., eds.: Karst Hydrogeology and Geomorphology, vol. 1, John Wiley
and Sons, Chichester, 2007.
Ford, D. and Williams, P. W., eds.: Karst Geomorphology and Hydrology, Chapman and
Hall, London, 1989.
Franke, H. W.: The theory behind stalagmite shapes, Studies in Speleology, 1, 89–95, 1965.
Frappier, A., Sahagian, D., Gonzales, L. A., and Carpenter, S. J.: El Niño events recorded
by stalagmite carbon isotopes, Science, 298, 565, 2002.
Freundlich, H.: Über die Adsorption von Lösungen, Zeitschrift für physikalische Chemie,
Abteilung A, 54, 385–457, 1907.
Friedman, I. and O’Neil, J. R.: Data of Geochemistry, vol. 47, chap. KK, pp. 1–12, USGS
Professional paper 440-KK, US Government Priniting Office, Washington D.C., 1977.
Friedrich, R.: Helium in polaren Eisschilden, diploma thesis, University of Heidelberg, 2003.
Friedrich, R.:
Grundwassercharakterisierung mit Umwelttracern:
Erkundung des
Grundwassers der Odenwald-Region sowie Implemetierung eines neuen EdelgasMassenspektrometersystems, Ph.D. thesis, University of Heidelberg, 2007.
Frisia, S., Borsato, A., Fairchild, I. J., and McDermott, F.: Calcite fabrics, growth mechanisms, and environment of formation in speleothems from the Italian Alps and southwestern
Ireland, Journal of Sedimentary Research, 70, 1183–1196, 2000.
Frisia, S., Borsato, A., Preto, N., and McDermott, F.: Late Holocene annual growth in three
Alpine stalagmites records the influence of solar activity and the North Atlantic Oscillation
on winter climate, Earth and Planetary Science Letters, 216, 411–424, 2003.
Gainon, F., Surbeck, H., and Zwahlen, F.: Natural radionuclides in gorundwater as pollutants
and as useful tracers, in: Proceedings of the 12th Symposium on Water Rock Interaction,
31 July - 5 August 2007, kunming, China, edited by Bullen, T. and Wang, Y., publisher:
Taylor and Francis Ltd, 2007.
Galdenzi, S. and Maruoka, T.: Gypsum deposits in the Frasassi Caves, Central Italy, Journal
of Cave and Karst Studies, 65(2), 111–125, 2003.
Gams, J.: Contribution to morphometrics of stalagmites, Proceedings of the 8th International
Congress of Speleology, Bowling Green, Kentucky, 8, 276–278, 1981.
Gascoyne, M.: Paleoclimate determination from cave calcite deposits, Quarternary Science
Reviews, 11, 609–632, 1992.
Genty, D. and Quinif, Y.: Annually laminated sequences in the internal structure of some
Belgian stalagmites-importance for paleoclimatology., Journal of Sedimentary Research,
66, 275–288, 1996.
170
Bibliography
Genty, D., Baker, A., and Vokal, B.: Intra- and inter-annual growth rate of modern stalagmites, Chemical Geology, 176, 191–212, 2001.
Genty, D., Plagnes, V., Causse, C., Cattani, O., Stievenard, M., Falourd, S., Blamart, D.,
Ouahdi, R., and Van-Exter, S.: Fossil water in large stalagmite voids as a tool for palaeoprecipitation reconstruction and palaeotemperature calculation, Chemical Geology, 184,
83–95, 2002.
Goldstein, R. H. and Reynolds, J. T., eds.: Systematics of fluid inclusions in diagenetic
minerals, SEPM Short Course 31, 1994.
Gölzhäuser, T.: Laborexperimente zur diffusiven Ausgasung von Helium aus polarem Eis,
diploma thesis, University of Heidelberg, 2008.
Grebe, W.: Die Bunkerhöhle in Iserlohn-Letmathe (Sauerland), Mitteilung des Verbands der
deutschen Höhlen- und Karstforscher, München, 39(2), 22–23, 1993.
Grebe, W.: Zur aktuellen Ausdehnung der B7-Höhle, Speläologisches Jahrbuch - Verein für
Höhlenkunde in Westfalen, Iserlohn , 14, 21–22, 1998.
Grootes, P. M., Stuiver, M., White, J. M. C., Johnsen, S., and Jouzel, J.: Comparison
of oxygen Isotope record from the GSIP2 and GRIP Greenland Ice cores, Nature, 366,
552–554, 1993.
Gvozdev, V. V. and Tovbin, Y. K.: New method for calculation the adsorption of noble gases
on amorphous surfaces, Russian Chemical Bulletin, 46(6), 1060–1069, 1997.
Hakl, J.: Application of radon-222, as a natural tracer in environmental studies, Ph.D. thesis,
Lajos Kossuth University Debrecen, 1997.
Hakl, J., Csige, I., Hunyadi, I., Várhegyi, A., and Géczy, G.: Radon transport in fractured
porous media - experimental study in caves, Environment International , 22, Suppl.1,
S433–S437, 1996.
Hakl, J., Hunyadi, I., Csige, I., Lénárt, L., and Várhegyi, A.: Radon transport phenomena
studied in karst caves - international experiences on radon levels and exposures, Radiation
Measurements, 28 (1-6), 675–684, 1997.
Hakl, J., Hunyadi, I., Csige, I., Géczy, G., and Bolner-Takács, K.: Site-specific radon regimes
of a cave system, Il Nuovo Cimento , 22 C (3-4), 471–474, 1999.
Hammerschmidt, E., Niggemann, S., Grebe, W., Oelze, R., Brix, M. R., and Richter, D.:
Höhlen in Iserlohn, Schriften zur Karst- und Höhlenkunde in Westfalen , 1, 1–154, 1995.
Haq, B. U., Hardenbol, J., and Vail, P. R.: Mesozoic and Cenozoic chronostratigraphy and
cycles of sea-level change, In: Wilgus, C. K., Hastings, B. S., Kendall, C. G. S. C., Posamentier, H. W., Ross, C. A., van Wagoner, J. C. (Eds) Sea-level changes: an integrated
approach SEPM Spec. Publ., 42, 72–108, 1988.
Harmon, R. S., Schwarcz, H. P., and O’Neil, J. R.: D/H ratios in speleothem fluid inclusions:
a guide to variations in the isotopic composition of meteoric precipitation?, Earth and
Planetary Science Letters, 42, 254–266, 1979.
Heaton, T. H. E. and Vogel, J. C.: Gas concentrations and ages of groundwaters in Beaufort
Group sediments, South Africa, Water SA, 5 (4), 160–170, 1979.
Bibliography
171
Heaton, T. H. E. and Vogel, J. C.: ”Excess-Air” in groundwater, Journal of Hydrology, 50,
201–216, 1981.
Hellstrom, J., Woodhead, J., Drysdale, R., Couchoud, I., and Zanchetta, G.: U-Th, U-Pb and
U-U dating of the Corchia Cave speleothems: a continuous 1 Ma chronology, Geophysical
Research Abstracts, 10, EGU2008–A–11 591, 2008.
Heusser, C. J., Streeter, S. S., and Stuiver, M.: Temperature and precipitation record in
southern Chile extended to ≈ 43’000 yr ago, Nature, 294, 65–67, 1981.
Hoehn, E. and von Gunten, H. R.: Radon in groundwater - a tool to assess infiltration from
surface waters to aquifers, Water Resources Research, 25(8), 1795–1803, 1989.
Holocher, J., Peeters, F., Aeschbach-Hertig, W., Hofer, M., Brennwald, M., Kinzelbach,
W., and Kipfer, R.: Experimental investigations on the formation of excess air in quasisaturated porous media, Geochimica et Cosmochimica Acta, 66(23), 4103–4117, 2002.
Holzhauser, H., Magny, M., and Zumbühl, H. J.: Glacier and lake-level variations in westcentral Europe over the last 3500 years, The Holocene, 15(6), 789–801, 2005.
Johnsen, S., Dahl-Jensen, D., Gundestrup, N., Steffensen, P., Clausen, H. B., Miller, H.,
Asson-Delmotte, V., Sveinbjörnsdottir, A., and White, J.: Oxygen isotope and palaeotemperature records from six Greenland ice-core stations: Camp Century, Dye-3, GRIP,
GISP2, Renland and NorthGRIP, Journal of Quarternary Science, 16 (4), 299–307, 2001.
Jones, P. D. and Moberg, A.: Hemispheric and Large-Scale Surface Air Temperature Variations: An Extensive Revision and an Update to 2001, Journal of Climate, 16, 206–223,
2003.
Kaufmann, G.: Stalagmite growth and palaeo-climate: the numerical perspective, Earth and
Planetary Science Letters, 214, 251–266, 2003.
Kawamura, K., Parrenin, F., Lisiecki, L., Uemura, R., Vimeux, F., Severinghaus, J. P.,
Hutterli, M. A., Nakazawa, T., Aoki, S. Jouzel, J., Raymo, M. E., Matsumoto, K., Nakata,
H., Motoyama, H., Fujita, S., Goto-Azuma, K., Fujii, Y., and Watanabe, O.: Northern
Hemisphere forcing of climatic cycles in Antarctica over the past 360,000 years, Nature,
448, 912–917, 2007.
Kelley, S., Turner, G., Butterfield, A., and Shepherd, T.: The source and significance of
argon isotopes in fluid inclusions from areas of mineralization, Earth and Planetary Science
Letters, 79, 303–318, 1986.
Kendall, A. and Broughton, P.: Origin of fabrics in speleothems composed of columnar calcite
crystals, Journal of Sedinmentary Petrology, 48(2), 519–538, 1978.
Kilian, R., Schimpf, D., Mangini, A., Spötl, C., Kronz, A., and Arz, H.: High resolution
precipitation and wind record from the Southern Hemisphere Westerly Zone, based on
speleothem proxies., Geophysical Research Abstracts, 10, EGU2008–A–10 551, 2008.
Kloppmann, W., Dever, L., and Edmunds, W. M.: Residence time of Chalk groundwaters in
the Paris Basin and the North German Basin: a geochemical approach, Applied Geochemistry, 13, 593–606, 1998.
Klotz, W., Guiot, J., and Mosbrugger, V.: Continental European Eemian and early Wrmian
climate evolution: comparing signals using different quantitative reconstruction approaches
based on pollen, Global and Planetary Change, 36 (4), 277–294, 2003.
172
Bibliography
Kluge, T. and Aeschbach-Hertig, W.: Noble gas measurement on fluid inclusions in stalagmites, In: Lippmann-Pipke, J. and Aeschbach-Hertig, W.(Eds) 4th Mini Conference
on Noble Gases in the Hydrosphere and in Natural Gas Reservoirs - abstract volume, doi
10.2312/GFZ.mga.050, GFZ Potsdam, Germany, 2007.
Kluge, T., Ilmberger, J., von Rohden, C., and Aeschbach-Hertig, W.: Tracing and quantifying
groundwater inflow into lakes using radon-222, Hydrology and Earth System Sciences, 11,
1621–1631, 2007.
Kluge, T., Marx, T., Scholz, D., Niggemann, S., Mangini, A., and Aeschbach-Hertig, W.: A
new tool for palaeoclimate reconstruction: Noble gas temperatures from fluid inclusions in
speleothems, Earth and Planetary Science Letters, 269, 407–414, 2008.
Kolodny, Y., Bar-Matthews, M., Ayalon, A., and McKeegan, K. D.: A high spatial resolution
δ 18 O profile of a speleothem using an ion microprobe, Chemical Geology, 197, 21–28, 2003.
Krawczyk, W. E. and Ford, D. C.: Correlating specific conductivity with total hardness in
gypsum karst waters., Earth Surface Processes and Landforms , 32, 612–620, 2007.
Kukla, G. J., Bender, M. L., de Beaulieu, J., Bond, G., Broecker, W. S., Cleveringa, P., Gavin,
J. E., Herbert, T. D., Imbrie, J., Jouzel, J., Keigwin, L. D., Knudsen, K., McManus, J. F.,
Merkt, J., Muhs, D. R., Müller, H., Poore, R. Z., Porter, S. C., Seret, G., Shackleton, N. J.,
Turner, C., Tzedakis, P. C., and Winograd, I. J.: Last interglacial climates, Quarternary
Research, 58 (1), 2–13, 2002.
Küster, F.: Beiträge zur Molekulargewichtsbestimmung an ”festen” Lösungen, Zeitschrift für
physikalische Chemie, 13, 445–458, 1894.
Langmuir, I.: The constitution and fundamental properties of solids and liquids. Part I.
Solids, Journal of the American Chemical Society, 38(11), 2221–2295, 1916.
Langmuir, I.: The adsorption of gases on plane surfaces of glass, mica and platiunum, Journal
of the American Chemical Society, 40(9), 1361–1403, 1918.
Leya, I. and Wieler, R.: Nucleogenic production of Ne isotopes in Earth’s crust and upper
mantle induced by alpha particles from the decay of U and Th, Journal of Geophysical
Research, 104, 15 439–15 450, 1999.
Lippolt, H. J., Leitz, M., Wernicke, R. S., and Hagedorn, B.: (Uranium + thorium)/helium
dating of apatite: experience with samples from different geochemical environments.,
Chemical Geology (Isotope Geoscience Section) , 112, 179–191, 1994.
Loosli, H., Aeschbach-Hertig, W., Barbecot, F., Blaser, P., Darling, W., Dever, L., Edmunds,
W., Kipfer, R., Purtschert, R., and Walraevens, K.: Isotopic methods and their hydrogeochemical context in the investigation of palaeowaters, In: Edmunds, W.M., Milne, C.J.
(Eds.) Palaeowaters in Coastal Europe: evolution of groundwater since the late Pleistocene,
vol. 189, pp. 193–212, Geol. Soc. Special Publication, Geological Society of London, London, 2001.
Lucas, L. and Unterweger, M. P.: Comprehensive Review and Critical Evaluation of the HalfLife of Tritium, ournal of Research of the National Institute of Standards and Technology,
105(4), 541–549, 2000.
Bibliography
173
Marx, T.: Weiterentwicklung des Mess- und Extraktionssystems zur Bestimmung von Edelgastemperaturen aus Speläothemen, diploma thesis, Institute of Environmental Physics,
University of Heidelberg, 2008.
Matsubara, K., Matsuda, J., Nagao, K., Kita, I., and Taguchi, S.: Xe in amorphous silica: A
new thermometer in geothermal systems, Geophysical Research Letters, 15, 657–660, 1988.
Mattey, D., Lowry, D., Duffet, J., Hodge, E., and Frisia, S.: A 53 year seasonally resolved
oxygen and carbon isotope record from a modern Gibraltar speleothem: reconstructed
drip water and relationship to local precipitation, Earth and Planetary Science Letters,
269, 80–95, 2008.
Matthews, A., Ayalon, A., and Bar-Matthews, M.: D/H ratios of fluid inclusions in Soreq
Cave (Israel) speleothem as a guide to the eastern Mediterranean Meteoric Line relationship
in the last 120 ky, Chemical Geology , 166, 183–191, 2000.
Mazor, E., ed.: Chemical and isotopic groundwater hydrology - The applies approach, vol. 2,
Dekker, New York, 1997.
Mazor, E. and Bosch, A.: Helium as a semi-quantitative tool for groundwater dating in the
range of 104 to 108 years., Consultants meeting on isotopes of noble gases as tracers in
environmental studies, IAEA, Vienna, 305 p. , 1992.
McDermott, F.: Palaeo-climate reconstruction from stable isotope variations in speleothems:
a review, Quaternary Sciene Reviews , 23, 2004.
McDermott, F., Mattey, D., and Hawkesworth, C.: Centennial-scale Holocene climate variability revealed by a high-reolution speleothem 18 O record from S.W. Ireland, Science, 294,
1328–1331, 2001.
McDermott, F., Schwarcz, H., and Rowe, P. J.: Isotopes in Speleothems, In: M.J. Leng (Ed.)
Isotopes in paleaoenvironmental research, Springer Verlag, Dordrecht, 2006.
McGarry, S., Bar-Matthews, M., Maathews, A., Vaks, A., and Ayalon, A.: Constraints on
hydrological amd paleotemperature variations in the eastern Mediterranean region in the
last 140 ka given by the δD values of speleothem fluid inclusions, Quarternary Science
Reviews, 23, 901–918, 2004.
Mercury, L., Pini, D. L., and Zeyen, H.: The effect of the negative pressure of capillary water
on the atmospheric noble gas solubility in ground water on palaeotemperature reconstruction, Earth and Planetary Science Letters, 223, 147–161, 2004.
Moberg, A., Sonechkin, D., Holmgren, K., Datsenko, N. M., and Karlén, W.: Highly variable Northern Hemisphere temperatures reconstructed from low- and high-resolution proxy
data, Nature, 433, 613–617, 2005.
Munakata, K., Kanjo, S., Yamatsuki, S., Koga, A., and Ianovski, D.: Adsorption of noble
gases on silver-mordenite, Journal of Nuclear Science and Technology, 40(9), 695–697,
2003.
Musset, A. E.: Diffusion Measurements and the Potassium-Argon Method of Dating, Geophys. J. R. astr. Soc. , 18, 257–303, 1968.
Neff, U.: Massenspektrometrische Th/U-Datierung von Höhlensintern aus dem Oman: Klimaarchive des asiatischen Monsuns, Ph.D. thesis, University of Heidelberg, 2001.
174
Bibliography
Neff, U., Burns, S. J., Mangini, A., Mudelsee, M., Fleitmann, D., and Matter, A.: Strong
coherence between solar variability and the monsoon in Oman between 9 and 6 kyr ago,
Nature, 411, 290–293, 2001.
Newman, B. D., Norman, D. I., Gundimenda, M., and Levy, S. S.: Understanding the
genesis of nonmarine calcite deposits through quadrupole mass spectrometric analysis of
fluid inclusion gases, Chemical Geology, 132, 205–213, 1996.
Niggemann, S.: Climate related research on late postglacial stalagmites from massive limestone caves in the Sauerland, Bochumer Geologische und Geotechnische Arbeiten, 55, 5–129,
2000.
Niggemann, S., Mangini, A., Mudelsee, M., Richter, D. K., and Wurth, G.: Sub-Milankovitch
climatic cycles in Holocene stalagmites from Sauerland, Germany, Earth and Planetary
Science Letters , 216, 539–547, 2003.
Norman, D. I. and Musgrave, J. A.: N2 -Ar-He compositions in fluid inclusions: Indicators of
fluid source, Geochimica et Cosmochimica Acta, 58, 1119–1131, 1994.
Norris, R. D. and Röhl, U.: Carbon cycling and chronology of climatewarming during the
Palaeocene/Eocene transition, Nature, 401, 775–778, 1999.
Ojala, A. E. K., Alenius, T., Seppä, H., and Giesecke, T.: Integrated varve and pollen-based
temperature reconstruction from Finland: evidence for Holocene seasonal temperature
patterns at high latitudes, The Holocene, 18 (4), 529–538, 2008.
Ozima, M. and Podosek, F. A., eds.: Noble gas geochemistry , Cambridge university press,
2nd edn., 2001.
Pajon, J. M., Curtis, J., Tudhope, S., Metcalfe, S., Brenner, M., Guilderson, T., Chilcott, C.,
Grimm, E., and Hernández, I.: Isotope records from a stalagmite from Dos Anas cave in
Pinar del Rio Province, Cuba. Palaeoclimatic implications, conference contribution, NURT2006. V International Symposium on nuclear and related techniques. Havana, Cuba, April
3-7., 16 pp., 2006.
Peeters, F., Beyerle, U., Aeschbach-Hertig, W., Holocher, J., Brennwald, M. S., and Kipfer,
R.: Improving noble gas based paleoclimate reconstruction and groundwater dating using
20 Ne/22 Ne ratios, Geochimica et Cosmochimica Acta, 67, 587–600, 2003.
Petit, J. R., Jouzel, J., Raynaud, D., Barkov, N., Barnola, J.-M., Basile, I., Bender, M.,
Chappellaz, J., Davisk, M., Delaygue, G., Delmotte, M., Kotlyakov, V. M., Legrand, M.,
Lipenkov, V., Lorius, C., Pépin, L., Ritz, C., Saltzmank, E., and Stievenard, M.: Climate
and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica,
Nature, 399, 429–436, 1999.
Pflitsch, A., Piasecki, J., and Niggeman, S.: Untersuchungen zum Einflu von Touristen auf
das Höhlenklima in der Dechenhöhle/Iserlohn, Mitteilungen des Verbands d.dt. Höhlenund Karstforschung, 16, 45–48, 2000.
Plummer, L. N., Wigley, T. M., and Parkhurst, D. L.: The kinetics of calcite dissolution in
CO2 -water systems at 5 ℃ to 60 ℃ and 0.0 to 1.0 atm CO2 , American Journal of Science,
278, 179–216, 1978.
Podosek, F. A., Honda, M., and Ozima, M.: Sedimentary noble gases, Geochimica et Cosmochimica Acta, 44, 1875–1884, 1980.
Bibliography
175
Polyak, V., Hill, C., and Asmerom, Y.: Age and Evolution of the Grand Canyon Revealed
by U-Pb Dating of Water Table Type Speleothems, Science, 319, 1377–1380, 2008.
Porcelli, D., Ballentine, C. J., and R.Wieler: An overview of noble gas geochemistry and cosmochemistry, In: Porcelli, D., C.J. Ballentine, R. Wieler(Eds) Noble gases in geochemistry
and cosmochemistry, vol. 47, pp. 1–17, 2002.
Rein, B.: How do the 1982/83 and 1997/98 El Niños rank in a geological record from Peru?,
Quarternary International, 161, 56–66, 2007.
Rice, S. B.: The Development of a Method for the Extraction and Measurement of Noble
Gases from Fluid Inclusions in Samples of Calcium Carbonate, master thesis, Institute of
Environmental Physics, University of Heidelberg, 2004.
Richards, D. A. and Dorale, J. A.: U-series chronology and environmental applications of
speleothems, In: Bourdon, B., Henderson, G.M., Lundstrom, C.C., Turner, S.P. (Eds)
Uranium-Series Geochimistry, vol. 52, Reviews in Mineralogy, pp. 407–460, 2003.
Richards, D. A., Bottrell, S. H., Cliff, R. A., Ströhle, K., and Rowe, P. J.: U-Pb dating of a
speleothem of quarternary age, Geochimica et Cosmochimica Acta, 62 (23/24), 3683–3688,
1998.
Roeckner, E., Buml, G., Bonaventura, L. Brokopf, R., Esch, M., Giorgetta, M., Hagemann,
S., Kirchner, I., Kornblueh, L., Manzini, E., Rhodin, A., Schlese, U., Schultzweida, U.,
and Tompkins, A.: The atmospheric general circulation model ECHAM5. Part I: Model
description, report, Max-Planck-Inst. F. Meteor., 2003.
Roedder, E.: Introduction to fluid inclusions, In: Roedder, E.(ed) Reviews in mineralogy Fluid inclusions, vol. 12, pp. 1–10, Mineralogical Society of Americy, Blacksburg, 1984.
Roedder, E., Ingram, B., and Hall, W. E.: Fluid inclusions. Extraction and quantitative
analysis of inclusions in the milligram range, Economic Geology and the Bulletin of the
Society of Economic Geologists, 58(3), 353–74, 1964.
Rozanski, K. and Dulinski, M.: Deuterium content of European paleowaters as inferred
from isotopic composition of fluid inclusions trapped in carbonate cave deposits, Isotope
Techniques in Water Resources Development, IAEA, Vienna, 565–578, 1987.
Scarsi, P.: Fractional extraction of helium by crushing of olivine and clinopyrexene phenocrysts: Effects on the 3 He/4 He measured ratio, Geochimica et Cosmochimica Acta,
64(21), 3751–3762, 2000.
Scheidegger, Y.: Stalagmiten als mögliche Klimaarchive, diploma thesis, Swiss Federal Institute of Technology, Zurch, 2005.
Scheidegger, Y., Brennwald, M., Heber, V., Wieler, R., and Kipfer, R.: A first convincing step
towards the application of dissolved noble gases in speleothem fluid inclusions to determine
local meteorological conditions in caves, Geophysical Research Abstracts, 8, EGU06–A–
04 832, 2006.
Scheidegger, Y., Badertscher, S. V., Driesner, T., Wieler, R., Heber, V., and Kipfer, R.:
Microscopial speleothem calcite investigations proofing the existence of two different types
of fluid inclusions, Geophysical Research Abstracts , 9, EGU–2007–A–06 374, 2007a.
176
Bibliography
Scheidegger, Y., Kipfer, R., R., W., Badertscher, S., and Leuenberger, M.: Noble gases in
fluid inclusions in speleothems., abstract, Goldschmidt Conference, Köln , 2007b.
Scheidegger, Y., Kluge, T., Kipfer, R., and Aeschbach-Hertig, W.: Paleotemperature reconstruction using noble gas concentrations in speleothem fluid inclusions, Pages Newsletter,
submitted, 2008.
Schlosser, P., Stute, M., Dörr., H., Sonntag, C., and Münnich, K. O.: Tritium/3He-dating of
shallow groundwater, Earth and Planetary Science Letters, 89, 353–362, 1988.
Schwarcz, H. P.: Geochronology and isotope geochemistry of speleothems, In: Fritz, P. and
Fontes, J. Ch.(Eds) Handbook of Environmental Isotope Geochemistry, pp. 271–303, Elsevier, Amsterdam, 1986.
Schwarcz, H. P., Harmon, S., Thompson, P., and Ford, D. C.: Stable isotope studies of fluid
inclusions of speleothems and their paleoclimatic significance, Geochimica et Cosmochimica
Acta , 40, 657–665, 1976.
Seppä, H. and Birks, H. J. B.: July mean temperature and annual precipitation trends during
the Holocene in the Fennoscandian tree-line area: pollen-based climate reconstructions, The
Holocene, 11(5), 527–539, 2001.
Serefiddin, F., Schwarcz, H. P., and Ford, D. C.: Palaeotemperature reconstruction using
isotopic variations in speleothem fluid inclusion water, Geochimica et Cosmochimca Acta,
66, 697, 2002.
Siegenthaler, U., Stocker, T. F., Monnin, E., Lüthi, D., Schwander, J., Stauffer, B., Raynaud,
D., Barnola, J., Fischer, H., Masson-Delmotte, V., and Jouzel, J.: Stable Carbon Cycle
Climate Relationship During the Late Pleistocene, Science, 1313 - 1317, 697, 2005.
Solomon, D. K., Poreda, R. J., Schiff, S. L., and Cherry, J. A.: Tritium and helium3 as
groundwater age tracers in the Borden aquifer, Water Resources Research, 28(3), 741–755,
1992.
Spahni, R., Chappellaz, J., Stocker, T. F., Loulergue, L., Hausammann, G., Kawamura,
K., Flückiger, J., Schwander, J., Raynaud, D., Masson-Delmotte, V., and Jouzel, J.: Atmospheric Methane and Nitrous Oxide of the Late Pleistocene from Antarctic Ice Cores,
Science, 310, 1317–1321, 2005.
Spoetl, C. and Mangini, A.: Stalagmite from the Austrian Alps reveals Dansgaard Oeschger
events during isotope stage 3: Implications for the absolute chronology of Greenland ice
cores, Earth and Planetary Science Letters, 203, 507–518, 2002.
Spoetl, C., Fairchild, I. J., and Tooth, A. F.: Cave air control on dripwater geochemistry,
Obir Caves (Austria): Implications for speleothem deposition in dynamically ventilated
caves, Geochimica et Cosmochimica Acta, 69(10), 2451–2468, 2005.
Stuart, F. M. and Turner, G.: The abundance and isotopic composition of the noble gases in
ancient fluids, Chemical Geology (Isotope Geoscience Section), 101, 97–109, 1992.
Stuart, F. M., Burnard, P. G., Taylor, R. P., and Turner, G.: Resolving mantle and crustal
contributions to ancient hydrothermal fluids: He-Ar-isotopes in fluid inclusions form Dae
Hwa W-Mo mineralisation, South Korea, Geochimica et Cosmochimica Acta, 59 (22), 4663–
4673, 1995.
Bibliography
177
Stute, M. and Deák, J.: Environmental isotope study (14C, 13C, 18O, D, noble gases) on deep
groundwater circulation systems in Hungary with reference to paleoclimate, Radiocarbon,
31, 902–918, 1989.
Stute, M. and Schlosser, P.: Principles and applications of the noble gas paleothermometer,
Climate Change in Continental Isotopic Records, 78, 89–100, 1993.
Stute, M. and Sonntag, C.: Paleotemperatures derived from noble gases dissolved in groundwater and in relation to soil temperature., Isotopes of noble gases as tracers in environmental studies, Vienna, IAEA, 111–122, 1992.
Stute, M., Schlosser, P., Clark, J. F., and Broecker, W. S.: Paleotemperatures in the Southwestern United States derived from noble gases in ground water, Science, 256, 1000–1003,
1992.
Stute, M., Forster, M., Frischkorn, H., Serejo, A., Clark, J. F., Schlosser, P., Broecker, W. S.,
and Bonani, G.: Cooling of Tropical Brazil (5 ℃) during the Last Glacial Maximum,
Science, 269, 379–383, 1995.
Swann, G. E. A., Mackay, A. W., Leng, M. J., and Dernory, F.: Climatic change in Central
Asia during MIS 3/2: a case study using biological responses from Lake Baikal, Global and
Planetary Change, 46, 1-4, 235–253, 2005.
Szabo, Z. and Rice, D. E.: Age dating of shallow groundwater with chlorofluorocarbons, tritium/helium 3, and flow path analysis, southern New Jersey coastal plain, Water Resources
Research, 32(4), 1032–1038, 1996.
Taylor, K. C., Mayewski, P. A., Alley, R. B., Brook, E. J., Gow, A. J., Grootes, P. M., Meese,
D. A., Saltzman, E. S., Severinghaus, J. P., Twickler, M. S., White, J. W. C., Whitlow,
S., and Zielinski, G. A.: The Holocene Younger Dryas Transition recorded at Summit,
Greenland, Science, 278, 825–827, 1997.
Taylor, P. M. and Chafetz, H. S.: Floating rafts of calcite crystals in cave pool, Central
Texas, U.S.A.: crystal habit vs. saturation state, Journal of Sedimentary Research, 74(3),
328–341, 2004.
TJlab: Thomas Jefferson National Accelerator Facility - Office of Science Education,
http://education.jlab.org/itselemental/ele091.html, 2008.
Torgersen, T., Kennedy, B. M., and van Soest, M. C.: Diffusive Separation of Noble Gases by
adsorption and throat constrictions: Explaining Noble Gas patterns in Sedimentary Rocks,
American Geophysical Union, Fall Meeting 2002, pp. abstract #V52B–1291, 2002.
Träumner, K.: Inbetriebnahme, Tests und erste Anwendungen einer neuen Aufbereitungslinie
zur massenspektrometrischen Messung von Edelgasen aus Grundwasser- und Stalagmitproben, diploma thesis, University of Heidelberg, 2005.
Trieloff, M., Kunz, J., and Allègre, C. J.: Noble gas systematics of the Réunion mantle plume
source and the origin of primordial noble gases in Earth’s mantle, Earth and Planetary
Science Letters, 200, 297–313, 2001.
Turner, G. and Bannon, M. P.: Argon isotope geochemistry of inclusion fluids from graniteassociated mineral veins in southwest and northeast England, Geochimica et Cosmochimica
Acta, 56, 227–243, 1992.
178
Bibliography
Viau, A. E., Gajewski, K., Sawada, M. C., and Fines, P.: Millennial-scale temperature
variations in North America during the Holocene, Journal of Geophysical Research, 111,
2006.
Vollweiler, N., Scholz, D., Mühlinghaus, C., Mangini, A., and Spötl, C.: A precisely dated
climate record for the last 9 kyr from three high alpine stalagmites, Spannagel Cave,
Austria, Geophysical Research Letters, 33, 2006.
von Antropoff, A.: Untersuchungen über die Adsorption von Gasen von kleinsten bis zu
höchsten Drücken, Kolloid-Zeitschrift, 137(2/3), 105–108, 1954.
Walker, J., Cliff, R. A., and Latham, A. G.: U-Pb Isotopic Age of the StW 573 Hominid
from Sterkfontein, South Africa, Science, 314, 1592–1594, 2006.
Wang, Y., Cheng, H., Edwards, R. L., Kong, X., Shao, X., Chen, S., Wu, J., Jiang, X., Wang,
X., and An, Z.: Millenial- and orbital-scale changes in the East Asian monsoon over the
past 224,000 years, Nature, 451, 1090–1093, 2008.
Wedepohl, K. H.: The composition of the continental crust, Geochimica et Cosmochimica
Acta, 59(7), 1217–1232, 1995.
Weigel, F.: Radon, Chemiker-Zeitung, 102, 287–299, 1978.
Weiss, R. F.: The solubility of nitrogen, oxygen and argon in water and seawater, Deep-Sea
Research, 17, 721–735, 1970.
Weiss, R. F.: Solubility of helium and neon in water and seawater, Journal of Chemical and
Engineering Data, 16(2), 235–241, 1971.
Weiss, R. F. and Kyser, T. K.: Solubility of krypton in water and seawater, Journal of
Chemical and Engineering Data, 23(1), 69–72, 1978.
Wernicke, R. S. and Lippolt, H. J.: Dating of vein Specularite using internal (U+Th)/4He
isochrons, Geophysical Research Letters, 21(5), 345–348, 1994.
White, W.: Cave minerals and speleothems In: Ford, T.D. and Cullingford, C.H. (Eds) The
science of speleology, pp. 267–327, Academic Press, London, 1976.
Wilkinson, J. J.: Fluids in hydrothermal ore deposits, Lithos, 55, 229–272, 2001.
WMO: World Meteorological Organization, webpages: s. www.wmo.ch, 2008.
Wolf, R. A., Farley, K. A., and Silver, L. T.: Helium diffusion and low-temperature thermochronometry of apatite, Geochimica et Cosmochimca Acta , 60, 4231–4240, 1996.
Yamada, M., Ohsawa, S., Matsuoka, H., Watanabe, Y., Brahmantyo, B., Maryunani, K. A.,
Tagami, T., Kitaoka, K., Takemura, K., and Yoden, S.: Derivation of travel time of limestone cave drip water using tritium/helium 3 dating method, Geopyhsical Research Letters,
35, 2008.
Yamamoto, J., Kaneoka, I., Nakai, S., Kagi, H., Prikjod’ko, V. S., and Arai, S.: Evidence
for subduction-related components in the subcontinental mantle from low 3 He/4 He and
40 Ar/36 Ar in mantle xenoliths from far eastern Russia, Chemical Geology, 207, 237–259,
2004.
Bibliography
179
Yatsevich, I. and Honda, M.: Production of nucleogenic neon in Earth from natural radioactive decay, Journal of Geophysical Research, 102, 10 291–10 298, 1997.
Yonge, C. J.: Stable isotope studies of water extracted from speleothems, Ph.D. thesis,
McMaster University, Hamilton, Ontario, Canada, 1982.
Zagwijn, W. H.: An analysis of Eemian cliamte in western and central Europe, Quarternary
Science Reviews, 15, 451–469, 1996.
Zhang, R., Schwarcz, H. P., Ford, D. C., Schroeder, F. S., and Beddows, P. A.: An absolute
paleotemperature record from 10 to 6 ka inferred from fluid inclusion D/H ratios of a
stalagmite from Vancouver Island, British Columbia, Canada, Geochimica et Cosmochimica
Acta, 72(4), 1024–1026, 2008.
Zolensky, M. E., Bodnar, R. J., Bogard, D. D., Garrison, D. H., Gibson, E. K., Nyquist, L. E.,
Reese, Y., Shih, C.-Y., and Wiesmann, H.: Asteroidal water within fluid inclusion-bearing
halite in an H5 chondrite, Science, 285, 1377–1379, 1999.
180
Bibliography
Abbreviations
abbreviation
A
AEW
amu
asl
ccSTP
cps
D/O-events
GIS
MIS3
NGT
SEM
TU
UHV
‰ wt
description
air/water volume ratio
air-equilibrated water
atomic mass unit
above sea level
gas amount in cm3 at standard conditions
(temperature T = 273.15 K, pressure p = 1 atm)
counts per second, refers to the measurement of the electron multiplier
Dansgaard-Oeschger events
relatively warm phases during the glacial periods
Greenland Interstadials
Marine Isotope Stage 3
noble gas temperature
scanning electron microscope
tritium units
1 TU corresponds to 1 atom of 3 H in 1018 atoms of 1 H
ultra-high vacuum
fraction calculated with reference to the total weight
1 ‰ wt water content implies a total of 1 mg water in a calcite sample of 1 g
181
182
Bibliography
List of Figures
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
Crystal structure of carbonate species and possible ideal configurations
Selection of common speleothem forms . . . . . . . . . . . . . . . . . .
SEM of a calcite crystal grown under lab conditions . . . . . . . . . .
Pictures from different speleothem types . . . . . . . . . . . . . . . . .
Thin section of a sample from the stalagmite H12 . . . . . . . . . . . .
SEM image of a single calcite crystal . . . . . . . . . . . . . . . . . . .
NGT uncertainty at given analytical error and air/water volume ratio
Molecular diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Diffusion with fixed boundary values . . . . . . . . . . . . . . . . . . .
Freundlich, Langmuir and BET isothermes . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10
10
12
14
16
17
27
30
30
32
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
3.15
3.16
3.17
3.18
3.19
3.20
3.21
3.22
3.23
3.24
3.25
3.26
3.27
3.28
Location of caves, from which samples have been taken . . . . . . . . . .
Thin sections of the stalagmites H12 and OBI5 . . . . . . . . . . . . . .
Thin sections of the flowstone H8Z from Hoti cave . . . . . . . . . . . .
Thin sections of the stalagmite CG from Cuba . . . . . . . . . . . . . .
BU-U thin section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BU-U thin section on a closer view. . . . . . . . . . . . . . . . . . . . . .
High precision balance . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Artificial water samples . . . . . . . . . . . . . . . . . . . . . . . . . . .
Water determination and water collection efficiency . . . . . . . . . . . .
Effects of different handling on the determined weight. . . . . . . . . . .
Calibration curve at the extraction line in first extension stage . . . . .
Pressure-water calibration curve in certain volumes . . . . . . . . . . . .
Limits of water determination by vapour pressure . . . . . . . . . . . . .
Extraction line, first versions . . . . . . . . . . . . . . . . . . . . . . . .
Extraction line, final state . . . . . . . . . . . . . . . . . . . . . . . . . .
Extraction by crushing in a steel cylinder . . . . . . . . . . . . . . . . .
Grain size distribution of crushing in the steel cylinder and copper tubes
Extraction in copper tubes . . . . . . . . . . . . . . . . . . . . . . . . .
Squeezing technique copper tubes . . . . . . . . . . . . . . . . . . . . . .
Copper tube squeezing blank . . . . . . . . . . . . . . . . . . . . . . . .
Heating blanks for copper tubes . . . . . . . . . . . . . . . . . . . . . . .
Extraction by microwave treatment . . . . . . . . . . . . . . . . . . . . .
Vials used for the microwave extraction . . . . . . . . . . . . . . . . . .
Effect of microwave heating - water released by this method . . . . . . .
O-ring free System for extraction of microwave treated glass vials . . . .
Ar-40 microwave blanks . . . . . . . . . . . . . . . . . . . . . . . . . . .
Three-isotope plot of the first microwave run . . . . . . . . . . . . . . .
Three-isotope plot of the second microwave run . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
38
39
40
41
41
42
44
45
45
46
47
49
50
52
53
54
55
57
57
58
59
60
61
61
62
63
64
65
183
184
List of Figures
3.29
3.30
3.31
3.32
3.33
3.34
3.35
3.36
3.37
3.38
3.39
3.40
3.41
3.42
3.43
3.44
3.45
3.46
3.47
3.48
3.49
Effect of the heating duration on the air/water-volume ratio. . . .
Stepwise heating experiment applied on a H12 sample . . . . . . .
Kr vs. Ar noble gas concentrations of heated samples . . . . . . .
Xe vs. Ne noble gas concentrations of heated samples . . . . . . .
Influence of heating on different samples in run Beta . . . . . . . .
Pressure increase due to the water vapour during crushing . . . . .
Stepwise crushing experiment, conducted with a sample from H12
Ar offset, measured during a stepwise crushing procedure . . . . .
Xe signal per water amount in the case of stepwise heating . . . .
Noble gas concentrations for a stepwise heated sample . . . . . . .
Kr-Ar noble gas concentrations for a stepwise heated sample . . .
Xe-Ne noble gas concentrations for a stepwise heated sample . . .
Combined procedure of crushing in steel cylinder and heating . . .
Signal forms of He and Ne using different tuning parameters . . . .
Evolution of the neon signal due to different contributions . . . . .
Neon signal for a sample with a low count rate . . . . . . . . . . .
Neon signal for a sample with a high count rate . . . . . . . . . . .
Ne and He signal trends for different count rates . . . . . . . . . .
40 Ar signal in the mass spectrometer . . . . . . . . . . . . . . . . .
Magnet stability and impact on data evaluation . . . . . . . . . . .
Results of artificial sample measurements . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
66
67
68
69
70
71
72
73
74
74
75
76
76
86
90
90
91
92
93
94
98
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
Map of the Bunker cave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Radon air activity concentrations in the Bunker cave . . . . . . . . . . . . . .
Map of the B7 cave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Map of the sampling points for dripwater . . . . . . . . . . . . . . . . . . . .
Dripwater sampling technique . . . . . . . . . . . . . . . . . . . . . . . . . . .
Recent tritium values in Germany (Hof) . . . . . . . . . . . . . . . . . . . . .
Seasonal tritium cycle at the Bunker Cave site . . . . . . . . . . . . . . . . .
Stable isotope data from rain and dripwater at/in Bunker cave . . . . . . . .
Summarized result of the 3 H-3 He-measurements. . . . . . . . . . . . . . . . .
Water content of different stalagmites . . . . . . . . . . . . . . . . . . . . . .
Air/water volume ratio from different stalagmite samples using a single step
procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Water content along the growth axis of BU-1 . . . . . . . . . . . . . . . . . .
Water content along the growth axis of BU-1, compared to a pollen record . .
Water content of the Chile stalagmite MA2 . . . . . . . . . . . . . . . . . . .
Comparison of the MA2 water content with the δ 18 0 values of the same stalagmite. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Comparison of the MA2 water content with the Mg/Ca value measured at the
same stalgmite. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Comparison of the MA2 water content with an ENSO related marine sediment
core. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Water content along the growth axis of BU-U . . . . . . . . . . . . . . . . . .
Factors influencing the amount of water-filled inclusions . . . . . . . . . . . .
Occurrence of water- and air-filled inclusions . . . . . . . . . . . . . . . . . .
Development of the air-water volume ratio using stepwise extraction . . . . .
Reduction of the air-addition by stepwise methods at BU-1 . . . . . . . . . .
Liberated water during the stepwise extraction . . . . . . . . . . . . . . . . .
Deviations from expected signals at BU-1 . . . . . . . . . . . . . . . . . . . .
100
101
103
105
106
109
109
110
111
116
4.12
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.20
4.21
4.22
4.23
4.24
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
117
118
119
120
121
121
122
123
125
128
129
130
131
132
List of Figures
4.25
4.26
4.27
4.28
4.29
4.30
4.31
4.32
4.33
4.34
4.35
4.36
4.37
4.38
4.39
4.40
4.41
4.42
4.43
4.44
Excess-Ne in the H12 samples plotted against A . . . . . . . . . . . . . . .
He-content measured by Stuart and Turner (1992) . . . . . . . . . . . . . .
α-particles produced by 238 U and 226 Ra. . . . . . . . . . . . . . . . . . . . .
Compilation of radiogenic He data for stalagmites with different ages. . . .
Correlation between U/Th age and radiogenic He. . . . . . . . . . . . . . .
Correlation between U/Th ages and normalised radiogenic He . . . . . . . .
Stalagmite BU-U . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Microscopic analysis of young BU-U part using polarized light . . . . . . . .
Kr-Ar noble gas concentrations for the Early Holocene BU-U samples . . .
Xe-Ne noble gas concentrations for the Early Holocene BU-U samples . . .
Temperature difference between the Early Holocene and Early Middle Ages
Measurement half of BU-1, sample locations . . . . . . . . . . . . . . . . . .
NGT for samples from BU-1 . . . . . . . . . . . . . . . . . . . . . . . . . . .
Stable isotope results for BU-1 . . . . . . . . . . . . . . . . . . . . . . . . .
measured pieces of BU-U . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NGTs derived from the growth axis of BU-U . . . . . . . . . . . . . . . . .
NGTs from the growth axis of BU-U compared to averaged δ 18 O-values . .
NGTs from the growth axis of BU-U compared to continuous δ 18 O values .
NGTs from the growth axis of BU-U compared to δ 13 C-values . . . . . . . .
δ 13 C and δ 18 O values along the growth axis of BU-U . . . . . . . . . . . . .
185
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
133
134
136
139
140
140
142
143
145
146
148
149
151
152
153
155
157
157
158
159
186
List of Figures
List of Tables
2.1
2.2
2.3
Properties of different carbonate species . . . . . . . . . . . . . . . . . . . . .
Atmospheric mixing ratios of noble gases . . . . . . . . . . . . . . . . . . . . .
Noble gas concentrations of equilibrated water . . . . . . . . . . . . . . . . .
11
20
21
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
3.15
3.16
Saturation vapour pressure of water respectively ice at different temperatures.
Blank of squeezing a copper tube once . . . . . . . . . . . . . . . . . . . . . .
Mean line blank and sample signals for the two microwave test series. . . . .
Samples measured in the second microwave run . . . . . . . . . . . . . . . . .
Diluted standard - gas amounts . . . . . . . . . . . . . . . . . . . . . . . . . .
Calibration reproducibilities of the diluted and the undiluted standard . . . .
Mass spectrometer sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . .
Typical blank values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Typical blank values in ccSTP . . . . . . . . . . . . . . . . . . . . . . . . . .
Double times ionized 40 Ar in Ar tuning . . . . . . . . . . . . . . . . . . . . .
Double times ionized 40 Ar in He tuning . . . . . . . . . . . . . . . . . . . . .
Neon isotope ratios in the case of different calibrations . . . . . . . . . . . . .
Ar and CO2 background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
40 Ar/36 Ar ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Artificial standard samples . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Artificial standard samples - offset . . . . . . . . . . . . . . . . . . . . . . . .
47
59
62
63
81
81
82
84
84
86
87
88
88
93
97
97
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
4.14
Cave air mixing ratios in Bunker Cave - deviation from the atmosphere
Cave air mixing ratios in B7 Cave - deviation from the atmosphere . . .
Dating of dripwater - results . . . . . . . . . . . . . . . . . . . . . . . . .
Uncertainty of the water determination . . . . . . . . . . . . . . . . . .
Noble gas excess in different samples . . . . . . . . . . . . . . . . . . . .
The three natural decay chains . . . . . . . . . . . . . . . . . . . . . . .
Exemplary calculation of the expected radiogenic He . . . . . . . . . . .
BU-U, BU-1 and soda straw - samples . . . . . . . . . . . . . . . . . . .
BU-U, BU-1 and soda straw - noble gas data . . . . . . . . . . . . . . .
BU-U, BU1 and soda straw - results . . . . . . . . . . . . . . . . . . . .
BU-1 sample data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BU-1 results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BU-U sample data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BU-U results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
187
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
102
103
108
115
133
137
138
143
144
147
150
150
154
154
Acknowledgments
Prof. Dr. Werner Aeschbach-Hertig und Prof. Dr. Augusto Mangini möchte ich für das
Ermöglichen der interessanten Arbeit und der vielseitigen Unterstützung danken.
Im besonderem Maße möchte ich meinem Betreuer Prof. Dr. Werner Aeschbach-Hertig
danken. Er hatte für Fragen immer ein offenes Ohr, bei Unsicherheiten weitergeholfen und
in begeisternder Weise in das wissenschaftliche Arbeiten und in die Welt der Forschung
eingeführt.
Prof. Dr. Augusto Mangini möchte ich dafür danken, dass er das Zweitgutachten der Arbeit
übernahm und darüberhinaus bei fachlichen Fragen immer hilfreich zur Verfügung stand.
Den Mitgliedern der Arbeitsgruppe Aquatische Physik, sowie der ”Speläogruppe” unter Prof.
Dr. Augusto Mangini danke ich für die schöne Zeit am Institut, für die sehr produktive und
angenehme Zusammenarbeit und für das Gefühl jeden Tag gern ins Institut zu kommen.
Dr. Ronny Friedrich, Dr. Laszlo Palscu und Martin Wieser danke ich ganz herzlich für ihre
Unterstützung bei der Bedienung des Massenspektrometers und der zugehörigen Software.
Darüberhinaus danke ich Martin für seine Bereitschaft an allen Expeditionen teilzunehmen
und mich dabei in selbstloser Weise zu unterstützen, ob auf dem Willersinnweiher mit löchrigem Schlauchboot oder in Iserlohn bei 10-stündigen Touren durch schlammige Höhlen.
Im speziellen danke ich Thomas Marx für die hervorragende Zusammenarbeit. Nachdem ich
ihn für das Thema interessieren konnte, hat er dem ganzen Projekt einen starken Impuls
gegeben. Durch konstruktive, aber auch kontroverse Diskussionen und eine Reihe arbeitsintensiver Messungen sowie einiger interessanter Ideen konnten das Extraktions- und Messverfahren optimiert und eine Reihe erfolgreicher Messungen durchgeführt werden.
Dr. Johann Ilmberger danke ich für seine guten Ideen, seine engagierten Beiträge bei Diskussionen und dafür, dass er bei Fragen und Problemen immer ein offenes Ohr hatte.
Danken möchte ich auch Gerhard Zimmek, der unermüdlich für das Funktionieren der Technik
des Massenspektrometers gesorgt und mir darüberhinaus sehr viel Arbeit mit der Bestellung
von Ersatzteilen und neuen Komponenten abgenommen hat.
Dr. Ullrich Glasmacher danke ich für seine Unterstützung und die Bereitschaft seine Mikroskope für die Dünnschliffuntersuchungen an Stalagmiten zur Verfügung zu stellen. Darüberhinaus möchte ich H.P. Meyer, sowie J. Fillauer für die Dünnschliffherstellung danken.
Einen besonderen Dank möchte ich an Dr. Stefan Niggemann, sowie Rasmus Dreyer und weiteren Mitgliedern des Höhlenvereins Speläogruppe Letmathe richten. Ohne ihre Unterstützung
wären die interessanten Probenahmen in der Dechen-, Bunker- und vor allem der B7-Höhle
nicht möglich gewesen.
Dana Riechelmann danke ich für die gute Zusammenarbeit. Durch das Zuverfügungstellen
von Messdaten und die Hilfe bei der Interpretation von Dünschliffen und Kristallformen habe
ich einen besseren Einblick in Entstehung und Verteilung von Fluid Inclusions bekommen.
Danken möchte ich allen, die bei der Interpretation einzelner Stalagmitdaten geholfen haben:
Claudia Fensterer für das Beschaffen der Monitoring- und Höhlendaten zum Stalagmiten CG,
Daniel Schimpf für das Bereitstellen der Isotopendaten, Monitoringwerte und Diskussionen
bzgl. der MA-Stalagmiten aus Chile, Prof. Dr. Christoph Spötl für das Messen der Isotope
am Stalagmiten BU-U und das Zuverfügungstellen einiger Proben aus der Spannagelhöhle,
René Eichstädter und Dr. Dennis Scholz für Datierungen an diversen Stalagmiten und Dr.
Stefan Niggemann für die Stalagmitprobe BU-U, die die ersten glaubwürdigen Edelgastemperaturen lieferte.
Für die hilfreichen Diskussionen und dafür, dass sie bei Fragen jederzeit ansprechbar waren,
möchte ich mich ganz besonders bei Dr. Christoph von Rohden, Dr. Dennis Scholz und Dr.
Andrea Schröder-Ritzrau bedanken. Danken möchte ich auch Daniela Polag für ihre Hilfe
bei der Literatursuche und Fragen zur Kalzitherstellung.
Für das Korrekturlesen möchte ich mich bei Dr. Christoph von Rohden, Martin Wieser, Dr.
Dennis Scholz und vor allem bei Helga Rietz für die hilfreichen Bemerkungen und Vorschläge
bedanken.
Einen besonderen Dank möchte ich an diejenigen richten, die mich motiviert und auch
in schwierigen Phasen unterstützt haben, insbesondere Andreas Meyer und Jörg Peschek.
Danken möchte ich auch meinen ehemaligen Mitbewohnern aus dem 8.Stock. Sie haben mich
motiviert nach der Diplomarbeit weiterzumachen. Außerdem haben sie mir durch die vielen
gemeinsamen Aktivitäten geholfen in Phasen der Frustration den Mut nicht zu verlieren.
Im besonderen danken möchte ich auch meinen Eltern, Großeltern und meinem Bruder, die
mich während des ganzen Studiums und darüber hinaus in jeder Hinsicht unterstützt haben.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement