User manual | Manual 21392354

The determination of Arsenic in soil by ICP-OES
by
Alexander Edward Sibiri Whaley
Submitted in partial fulfilment of the requirements for the degree
MAGISTER SCIENTIAE
in the Faculty of Science University of Pretoria Pretoria January 2000 © University of Pretoria
SYNOPSIS
Arsenic has always played a major role in the environment and human life in general. From its
earliest uses, in ancient times, as a poison to its most recent use, in medicine, as an anti-leukemia
agent, this metal has fascinated mankind. This fascination has already yielded several surveys
on its toxicity, concentration, source and specie. However, some approaches have not been fully
explored. One of these is its occurrence in Phosphate bearing rocks and subsequent possible
contamination offertilizers derived from these rocks. To this end, a new variation ofa speciation
mechanism used, solvent extraction followed by ion-exchange, has been developed. Although
this method has mainly been used in connection with marine samples, in this dissertation, it has
been applied to solid samples. The levels of arsenic were then determined by Inductively
Coupled Plasma Optical Emission Spectrometry (ICP-OES). Phosphate bearing rocks,
commercial fertilizers and soils were digested in HCI and speciated out as As (III), As(V), Mono­
Methyl Arsenic acid (MMAA) and Dimethyl Arsenic acid (DMAA). The effect of different
organic phases were examined as well as two ion-exchange resins. Phosphoric acid, being an
intermediate stage in the manufacture of diverse fertilizers, was also investigated. Due to the
close relationship between arsenic and phosphorus, some steps in the method had to be reversed
in order to yield better results. The resulting samples were analysed at two wavelengths, ie
188.979nm en 193.696nm. Although the values obtained at the two wavelengths differ by
between O.Smg/kg and 1.Smg/kg, a simple t-test proves that the values are reconcilable. A micro­
concentric nebuliser was studied regarding its aerosol particle size, with a )l-LASER-particle­
analyser, and the detection limit of arsenic. The second part of this study could not be completed
due to the loss of the ICP. The following four nebulisers were compared: Micro-Concentric
Nebuliser (MCN), Meinhard Nebuliser, V-groove Nebuliser and a Cross-flow Nebuliser. Due
to their similar design, the MCN and Meinhard nebulisers have similar characteristic, the main
differences being in their operating characteristics, ie regarding particle size and background
intensity. As the inside diameter of the MCN is only O.2)lm special glassware had to be
developed to insure that the MCN does not become blocked.
SAMEVATTING
Arseen het nog a1tyd 'n be1angrike ro1 in die natuur en mens like 1ewe oor die a1gemeen
gespeel. Hierdie metaal het die mens van die begin af bet ower, van die vroegste gebruik,
in die verre verlede, as gifstof tot die hedendaagse gebruik in die mediese wereld as anti­
leukemie middel. Hierdie betowering het reeds gelei tot vele studies na die giftigheid,
konsentrasie, bronne en chemiese vorme van arseen. Sekere benaderings is egter nie ten
volle bestudeer nie. Een van die, is die voorkoms van arseen in fosfaatbevattende rots en
die gevolglike moontlike besoedeling van kunsmis wat vanuit die rotse vervaardig word.
Vir hierdie doel is 'n nuwe variasie van 'n bestaande metode om arseen in sy verskillende
chemiese vorme te verkry, gebruik; naamlik oplosmiddel-ekstraksie gevolg deur ioon­
uitrui1ing in 'n gepaste kolom.
A1hoewel die metode meestal in die ontleding van
seemonsters gebruik word, word die metode tydens
vastetoestand-monsters toegepas.
hierdie verhandeling op
Die arseenvlakke is met behulp van Induktief
Gekoppe1de Plasma Optiese Emissie Spektroskopie (IGP-OES) bepaal.
Fosfaat­
bevattende rots, bedryfskunsmis en grondmonsters is in Hel opgelos en geskei in die
volgende chemiese vorme: As(III), As(IV), monometie1arseensuur (MMAS) en
dimetie1arseensuur (DMAS). Die effek van verskillende organiese fases, sowel as twee
ioon-uitruilingsharse is ondersoek. Fosforsuur, as tussenproduk in die vervaardiging van
verskeie kunsmisstowwe, is ook ondersoek. Weens die nabye verwantskap tussen arseen
en fosfor moes verskeie stappe in die gebruikte metode omgeruil word, om beter resu1tate
te lewer. Die verkrygde monsters is by twee golflengtes, 188.979nm en 193.696nm,
ontleed.
Alhoewel daar 'n verskil van 0.5 - 1.5mglkg in die waardes is wat by die
verskillende golflengtes verkry word, toon 'n eenvoudige t-toets aan dat die waardes
versoenbaar is.
'n Mikro-konsentriese newelaar is ondersoek ten opsigte van die sproei­
deeltjiegrootte, met behulp van 'n ).!-LASER-deeltjiegrootte-ontleder en ook ten opsigte
van die bepaalbaarheid van arseen. Die deel is egter nie voltooi nie, weens verlies aan
die IGP.
Tydens die deel van die studie is vier newelaars, die Mikro-Konsentriese
Newelaar (MKN) , Meinhard Newelaar, V-Groef Newelaar en ' n Kruis-Vloei Newelaar
vergelyk. Weens hulle soortgelyke ontwerp is die eienskappe van die MKN en Meinhard
newelaars ook soortgelyk, met die grootste verskille in hul bedryfseienskappe, naamlik
ten opsigte van deeltjiegrootte en agtergrondsterkte. Spesiale glasware moes ontwikkel
word, om te voorkom dat die MKN verstop word, aangesien die binne deursnee van die
MKN slegs O.2).!ffi is.
Table of Contents
1
Introduction
1.1
ICP as an analytical method
1.2
Arsenic in the envirorunent
1.3
2
3
1.2.1
Types of arsenic found in soils and relative toxicity
3
1.2.2
Properties of arsenic useful in analysis
7
1.2.3
Soil conditions
8
1.2.3.1 Type of soil and parent rock type
8
1.2.3.2 pH and Eh of soil
8
1.2.3.3 Presence of other metals
9
1.2.3.4 Depth factors
11 1.2.3.5 Other considerations
11 1.2.4
Extraction methods
12 1.2.5
Speciation methods
l3
1.2.6
Hydride Generation
15 References
17 Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES)
18 2.1
The theory ofICP-AES
18 2.1.1
Sample introduction
20 2.1.2
Plasma formation
22 2.1.3
RF Generator
23 2.1.4
Detection of emission
24 2.2
Interferences in ICP-AES
24 2.2.1
Matrix effects
25
2.2.2
Spectral Overlap
26 2.2.2.1 Direct overlap
26 2.2.2.2 Wing overlap
27 2.2.2.3 Continuum radiation
28 Stray light
29 2.2.3
2.3
3
Ion Exchange
30 31 3.1
Introduction
31 3.2
Ion exchange resins
32 3.3
4
References
3.2.1
Anion exchange resins
33 3.2.2
Cation exchange resins
34 3.2.3
Chelating ion exchange resins
34 3.2.4
Capacity
35 3.2.5
Selectivity
35 3.2.5.1 Selectivity coefficient
35 3.2.5 .2 Factors influencing selectivity
36 3.2.6
Distribution coefficient
37 3.2.7
Kinetics of ion exchange
38 References
41 Comparison of a Micro-concentric nebulizer with a Meinhard nebulizer
42 4.1
Introduction to nebulizers
42 4.2
Theory of nebulizers
42 4.2.1
Pneumatic nebulizers
42 4.2.1.1 Concentric nebulizers
43 4.2.2
5
4.2.1.2 Cross-flow nebulizers
44
4.2.1.3 Babington-type nebulizers
45 Ultrasonic nebulizer
45 4.3
Experimental for the particle size distribution
47 4.4
Discussion of limits of detection
65 4.5
References
66 Extraction and speciation of arsenic
68 5.1
Introduction
68 5.2
Experimental 1: preparation and calibration of resins
69 5.3
Experimental 2: calibration of organic solvents
70 5.4
Experimental 3: extraction of soil samples
70 5.5
Experimental 4: inclusion ofMMAA
71 5.6
Experimental 5: extraction of phosphoric rocks and fertilizers
73 5.7
Experimental 6: new method for analysing phosphorus bearing rocks
76 6
General conclusions
7
Appendix 80 CHAPTERl INTRODUCTION 1.1
ICP as an analytical method
Inductively Coupled Plasma Atomic Emission spectrometry (ICP-AES) is a technique
used predominantly for the quantitative multi-element analysis ofmost types ofsamples,
ego biological and geological materials. It is an ideal technique due to its high dynamic
range which allows for the analysis of both major and trace elements from a single
sample (i .e. from gil to J.lgli range) .. Major elements can easily be analysed with high
accuracy and precision through internal standardisation and proper calibration
procedures. As with many other analytical techniques, trace elements, however, can be
more problematic. These problems can be related to the background enhancement and
spectral line overlap from concomitant elements. Physical interference can occur during
the nebulization process that will affect accuracy and precision. Here again, internal
standardisation can be used to rectify the sample transport variation.
Sensitive
measurements of refractory elements such as Band P are also catered for by this
technique through the high temperature of the ICP.
There are two systems available for ICP-AES, simultaneous and sequential.
Simultaneous systems have a higher sample throughput as more elements can be
analysed for a time period. This is because the wavelengths at which the elements are
analysed are preselected in the manufacturing process. A direct result of this however
is the lack of control over the analysis. Sequential systems have the advantage of
flexibility of wavelength selection. This in effect, allows for variations in analyte
-1­
concentrations and matrix types and the removal of interfering peaks by the selection of
a different wavelength, with the major drawback of much longer analysis time. As
geological materials are generally very complex in composition, several sources of
interference can generally be expected in ICP-AES. These can be eliminated by using
a high resolution spectrometer, on-line background compensation, predetermined
interference coefficients or matrix matched standards. Several methods for background
compensation exist, many mathematical, which usually involve scanning a spectral
segment on either or both sides ofpeak positions. Then, the spectrum for each interferent
is measured and removed mathematically from the analyte signal.
Any productive spectrometer should, in theory, fulfil these capabilities:
a high degree of resolution to achieve good separation from nearby
spectral lines, thereby reducing inter-element interference
exhibit high sensitivity through excellent light-gathering and least stray light
be stable, both mechanically and thermally, ensuring both the precision and
repeatability of the analysis and ensuring that the correct wavelength has been
selected.
low detection limits to measure trace elements plus major elements.
As ICP-AES is a very similar spectroscopic tool to flame and graphite furnace atomic
emission spectroscopy, the advantages and disadvantages of this technique can best be
described though a brief comparison of the three. Although atomic absorption is a
stronger phenomenon than atomic emission, where typically only 1* 10.6 photons emitted
by atoms reaches the detector, ICP-AES offers a tremendous narrowing of the detection
-2­
limits gap. Species are highly excited due to the high temperatures ofthe plasma. This
translates into extensively populated excited states yielding a simultaneous, intense
emission from many lines. Flame and graphite furnaces' emissions are much less intense
as the nonnal operating temperatures of these techniques are characteristically between
2,000 and 3,500 K. More popularly, flame and graphite furnaces operate as absorption
spectrometers, as absorption is a stronger phenomenon than emission, through the
absorption of light by excited states with low energies less than 3 eV.
The low detection limits achieved by ICP-AES are predominantly due to the large
emission signals with respect to the noise of the background of some elements (Ilgll
sometimes). These detection limits are obviously sample dependant. Therefore, they are
easily degraded by difficult matrices, increased background, and by spectral overlaps.
1.2 ARSENIC IN THE ENVIRONMENT
1.2.1 Types of arsenic found in soils and relative toxicity:
Although several fonns of arsenic can be found in soils, there are four major groups that
are important. These different fonns of arsenic can be separated into two major groups,
inorganic arsenic in the oxidation states As(III) and As(V) and organo-arsenicals, mono­
methyl (MMA)- and dimethyl-arsenic (DMA) acids. These last two arsenic fonns are
mainly found in an aquatic environment and are thought to be lOO times less toxic than
the inorganic fonns. These methylated fonns are the result of biological transfonnation
of the inorganic species [1,2].
-3­
Arsenate
AS04l' pH 8-9
HAs0 4' pH 6-7
Reduction
Arsenite
Bacteria
AsO l'
Methyl
arsonic acid
Bacteria
•
Dimethyl
Arsinic aci
Oxidation
j
+Sl'
+Fe
adsorp,
Pptn,
AsS l'
Reduction
j
FeAsO
j
Ppm.
AslS]
Figure 1: Chemical forms of arsenic and their transformations in soils [1].
The first or most ancient source of arsenic in the environment would be weathering of
the parent rock. Indeed, extensive surveys of the levels of arsenic have found that
different parent rocks led to varying levels of arsenic in soils, For example, arsenic in
soils derived from granite is lower than in those derived from basalt while the level in
soil derived from sedimentary rocks may attain a value of20 to 30 mg/Kg. Furthermore,
the amounts found in the earth's crust and shales are 1.8 and 13 mglKg respectively
which is mainly a result of its accumulation during weathering and translocation in
colloid fractions. The average concentration around the world ofAs is 6 mglKg, ranging
from 1 to 50 mglKg. Arsenic is concentrated in magmatic sulfides and iron ores. The
most important ores of arsenic are arsenic pyrites or mispickel, realgar, and orpiment.
The arsenic levels in soil enriched in these ores are often higher than in normal soil. The
arsenic content of soil may be closely related to the underlying bedrock if the parent
materials have not been mixed or redistributed by pedogenetic processes, wind, or
glaciation [1].
-4­
The second source (and least important) is from geological upheavals such as volcanoes where the average arsenic soil content is about 20 mglKg while outside these localities the soil content is closer to 2 mg/Kg [1]. Table 1: Movement of Arsenic in the environment [I]. From:
Land
Atmosphere
Oceans
Sediments
To:
Approx.
Amount
( *108 g/year)
Oceans
3,000
Atmosphere
1,000
Biota
300
Oceans
2,000
Land
1,000
Sediments
2,500
Biota
1,300
Dissolved
1,000
Land
2,400
Mining, smelting
500
Terrestial biota
Land
300
Volcanoes
Land
54
Sediments
40
Atmosphere
3
The third and obviously most recent source of arsenic is anthropogenic. This category
can be further broken down into two. These being, direct contamination or rather
purpose driven and indirect contamination or incidental [1]. From table 1 it is possible
to see that although arsenic has been used by humans, such use is not the predominant
cause of movement of arsenic in the environment. Indeed, anthropogenic arsenic
pollution tends to occur in localised areas though in some places such as Bangladesh, the
ground water can become polluted due to such activities. Table 2 highlights the major
-5­
uses of arsenic by humans, which can be seen to encompass most aspects of modem
industrial production.
In the fifties and sixties, arsenic compounds were used as fungicides, pesticides
and insecticides (some of which are still in use today). This use is still reflected
in the levels of arsenic found in some areas today that can be greater than 100
mg!Kg, especially in orchards.
The other source, incidental contamination comes from mines, industry and uses
of arsenic containing compounds (fertilizers for example) which have not been
properly examined. In Japan, the arsenic content ofsoils was 1475 mg/Kg at 300
yards and 11.15 mg!Kg at nine miles from the chimney of the Obuasi goldmine.
Since the 1920's the recorded world production of Arsenic Trioxide (White Arsenic) has been between 30,000 and 65,000 metric tons with a maximum output of63,939 in 1970. As recently as 1990, 47,632 metric tons were still being produced despite the well-
recorded dangers to human health and the environment [1]. Table 2: Uses of Arsenic in the economy [1]. Sector
Uses
Agriculture
Pesticides, insecticides, defoliants, wood preservatives, debarking trees, soil
sterilant
Livestock
Feed additives, disease ~revention (swine dysentery, heartwom infection),
cattle and sheep dips, a gaecides
Medicine*
Antisyphilic drugs, treatment of trypanosomiasis, amebiasis, sleeping
sickness
Electronics
Solar cells, ,&toelectronic devices, semiconductor applications, lightemitting dio es (digital watches)
Industry
Glassware, electrophotogra~hY, catalysts, pyrotechnics, antifouling paints,
dyes and soaps, ceramics p annaceutical substances
Metallurgy
Alloys (automotive body solder and radiators), battery plates (hardening
agents)
* Arsenic still used for medicinal purposes in some developing countries
-6­
1.2.2 Properties of arsenic useful in analysis
Although arsenic has several properties that are useful in specific analytical techniques,
some properties (e.g. thermodynamic, ease of redox) are useful in all. Foremost among
these are the differing chemical properties ofthe various oxidation states. For example,
arsenite, As(III) can be easily reduced to arsine, which has a low boiling point (-55°C)
and therefore can easily be distilled from complex sample matrices. This, coupled with
the fact that arsenate, As (V), needs to be reduced prior to hydride generation gives us
a powerful tool for speciation. The methylated forms ofarsenic can also directly produce
arsines (under reducing conditions) which have widely separated boiling points, thus
allowing for further speciation through distillation.
Thermodynamic considerations also playa major role in analysis. By addition of simple
oxidizing or reducing agents such as KI or KIO), arsenite and arsenate can easily be
interconverted although arsenate is the more thermodynamically stable form. This means
that Arsenite can be converted to arsine along with the methylated arsenic, leaving
arsenate behind.
Other properties that playa major role in the speciation ofarsenic are the easy conversion
of arsenite into arsenic trichloride by treatment with strong hydrochloric acid. This
arsenic trichloride, being covalent in nature is soluble in organic solvents while arsenate,
which doesn't react with HCl, is not extracted by an organic solvent. Furthermore,
arsenite can form complexes with sulfur compounds which gives another route for
-7­
speciation. Arsenate on the other hand can be studied with colorimetry by complexing
with molybdic acid and reduced to give a characteristic blue colour [1].
1.2.3 Soil conditions:
1.2.3.1 Type of soil and parent rock type
As mentioned earlier, the principal factors influencing the concentration of elements in
soils are the parent rock and human activities. Studies have shown that arsenic contents
are positively correlated to the level ofclay and negatively correlated to the level of sand
in the soil. Therefore, two main comparison points must be taken into account: the
parent rock and soil texture, especially as typically, the arsenic content is related to the
concentration of organic material [1].
1.2.2.2 pH and Eh of soil
Although these values do not affect the concentration of arsenic found in soils as such,
they do influence the relative abundance of the various species of As. As(III) is much
more toxic and soluble than As(V) and more mobile. Studies ofthe relationship between
solubility and Eh and pH have shown that soluble arsenic increased significantly with
diminishing Eh and increasing pH. Furthermore, arsenic adsorption is related to the pH
level, with a decrease in sorption of 46.5 Ilg As/g soil with a pH change from 5.2 to 9.4.
The quantity of adsorbed arsenic by soil was maximum at pH 6 to 8 for arsenite and at
pH 4 for arsenic acid [1].
-8­
1.2.3.3 Presence of other metals
In a mineral rich region, the levels of arsenic may be much higher as As is associated
with several minerals. Of particular interest are iron mines as arsenic is often found as
arsenopyrite and in copper containing compounds. During the formation ofsedimentary
rocks, arsenic is carried down by precipitation of iron hydroxides and sulfides.
Therefore, iron deposits and sedimentary iron ores are rich in arsenic. Because arsenic
is very similar in chemistry to P but also can have more oxidation states, it can insert
itself into many different lattices, such as S compounds. Hydrous oxides of AI, Fe, and
Mn affect arsenic surface reactions [1]:
(M=metal)
M-OH + H 20
p;
M-OH/ + OR
or
M-OH/ + H2As0 4- p; MHAs0 4 + H 20
Addition ofAs(III) as HAs0 2to untreated or Mn0 2-coated sediments result in oxidation
to As(V) or adsorption onto the surfaces ofthe oxides. The oxidation ofAs(III) by Mn02
proceeds as:
HAs0 2 + Mn02
-+
(Mn0 2}HAs02 + H20
H)As04
-+
(Mn02}HAs02
-+
H)As0 4 + MnO
H 2As04- + H+
H 2As0 4- -+ HAsOt + H+
-9­
If the Mn0 2 has much iron oxide coating, large amounts of arsenic are adsorbed; pure
iron oxide is an even more effective adsorbent. Furthermore, iron oxides adsorb arsenic
even more strongly than do manganese oxides. The ability of amorphous iron oxides to
adsorb arsenic strongly is related to their loose and highly hydrated form, allowing other
hydrated ions to diffuse freely throughout the structure without being restricted to
external surface sites, as in more crystalline solids. As(V) is less strongly adsorbed than
As(III).
Anions when present, especially phosphate (H 2P04-) can effectively compete with arsenic
for adsorption sites, particularly aluminum and iron oxide surfaces. About 60% of the
adsorbed As(V) and 70% of the adsorbed As(III) were displaced by H 2P04- in a solution
of 10- 6 M phosphate. Other anions were less effective in displacing arsenic, with the
order of effectiveness decreasing for As(V) ofH2P04 -> H 2As04- > sot> C0 32- and for
As(III) of H 2P0 4- > H 3As0 3 > F->
sot> CO/.
Phosphate substantially suppressed
arsenic adsorption, but this varied from soil to soil [1,2,4].
The role of calcium in arsenic fixation is less pronounced than the role of aluminium or
iron as calcium arsenate is more soluble than aluminium or iron arsenate. However,
plant growth on soils containing toxic levels of arsenic does not improve with liming,
since lime does not reduce the availability of arsenic by formation of calcium arsenate.
Arsenic forms sparingly soluble compounds with Ca2+ with a log K value of -18.17.
-10­
1.2.3.4 Depth factors
The quantity of arsenic which is found in soils depends on many factors. For example,
arsenic is not very mobile. It tends to remain in one place unless displaced by other
chemicals such as phosphates. However, when considering fields that are or have been
in use several factors must be considered. The first one being fertilization, as many
chemical fertilisers are phosphate rich. Also, if the field is ploughed then the depth to
which it is ploughed is important as ploughing causes the soil to be mixed. Therefore the
results over this depth should be similar and may not be true values when compared with
unused soil.
Another reason that the depth of sampling is important is that recent laboratory studies
have shown that arsenic deriving from Lead Arsenate can be displaced by phosphate
introduced by using fertiliser. This would then mean that the level of As in the top soil
would be less than the expected value and that the level in the subsoil would be greater
than the expected value. While this means that the phytotoxicity and human health
hazards associated with exposure to high As concentrations in topsoil are generally
reduced, this can lead to a contamination of shallow water tables which obviously caries
its own risks [1,5].
1.2.3.5 Other considerations
When studying the water soluble part of As, the values may be lower during the rains.
This value could be higher in the summer for the first 5 cm due to atmospheric
deposition. Also the frequency of fertilising should be considered and the time elapsed
since the last fertilising. The more fertilised the field, the higher the value may be (if the
fertiliser is the source of Arsenic). Other factors are obviously the use of pesticides,
-11­
herbicides or fungicides and previous use of the soil. Fields which were orchards in the
50's and 60's might have much higher levels of As [4].
Plant availability: Not all As in the soil is available to plants. This is an important factor
as As has similar characteristics to P and therefore can be easily taken up by plants.
However, most plants cannot tolerate a high level of As and die, should their intake be
too great. However, while feeding some animals may take in As directly from the soil.
The age and health of the plant are important (ie, must not compare apple trees and
wheat) [1].
1.2.4 Extraction methods:
There are several methods of extraction possible. The particular one used depends on
what information is required. The two main methods used can be classified as follows:
sequential leaching or total content. The sequential leaching procedure, highlighted in
table 3, is based on speciating the As into its various natural phases, these being: water
soluble, exchangeable, carbonate, easily reducible, moderately reducible, organic matter
and sulphide and the residual mineral. Further subdividing these fractions into the
different species ofArsenic (III, V, MMA, DMA) may prove to yield too small quantities
of As to measure without concentration. The total content procedure, which is the
favourite among the articles, involves "digesting" the soil in HN0 3 or Hel.
-12­
Table 3: Fractional extraction of Arsenic from soils [1].
Extraction Solution
Arsenic fraction
Water soluble
H 20 (distilled)
Exchangeable
1 molll NH4 acetate, pH 7, 1:20 (sample: solution),
20°C.
Carbonate
1 molll Na acetate, pH 5, 1:20, 20°C
Easily reducible
0.1 molll hydroxylamine hydrochloride, 0.1 molll
HNO), pH 2,1:100
Moderately reducible
0.2 molll NH4 oxalate, 0.2 molll oxalic acid, pH 3,
1:100
Organic matter and sulfide
H20 2 (0.30 gig), HN03, pH 2, 85°C, NH4 acetate
(1 molll)
Residual mineral
HFIHCI0 4
1.2.5 Speciation methods:
As mentioned earlier, arsenic has several useful properties which can be of great
assistance during speciation. The easiest method (used mainly when the two inorganic
forms of As (III, V) need to be speciated), is a simple conversion of As (III) into AsCI)
by simple addition of strong HCI (5 - 9 M). This covalent compound can then be
extracted into an organic phase such as benzene [1]. In environmental samples, however,
the most used method may be hydride generation. Although only arsenite is reduced to
arsine by sodium borohydride at pH 4 to 9, all four forms of arsenic can be reduced at a
lower pH. The arsines produced can be collected by a cold trap and speciated by
distillation.
A more efficient way to achieve separation would be to use gas
chromatography at this stage with a careful control of the experimental conditions to
avoid a rearrangement of the arsines.
The importance of ion exchange in the study of environmental samples cannot be
stressed enough. Used properly, this technique can serve as an online pre-concentration
-l3­
system of ultra trace elements and also a reproducible speciation technique, which is not
always the case with Hydride Generation. Ion exchange methods can provide high
enrichment factors, lower detection limits by separating the metals of interest from the
matrix and less interferences from other metals or oxidation states by speciating the
various metals and oxidation states. The on-line ion exchange techniques can also offer,
in contrast with manual pre-concentration techniques, the advantages in the reduction of
both analysis time and required sample volumes, not to mention a high sample
throughput. In addition, by working with a closed system, the risk ofcontamination from
open laboratory environment is reduced [3].
Due to the several advantages of ion exchange, several variations ofthe technique have
been developed. For example, the simplest method is to have an anion exchanger and
a cation exchanger in separate tubes and to run the arsenic containing solution through
both tubes. The various forms of As would be retained by the tubes and can be released
with acid or base at specific molarity. This method has one major flaw in that DMAA
is defined as the portion of As being eluted from the cation exchanger with a base. This
is nonspecific and could cause confusion when dealing with As samples coming from
marine animals that appear to contain large amounts of basic As. These compounds may
exhibit very similar behaviour to that ofDMAA under these conditions [5].
Another method of ion exchange is to pack the column with successive layers of anion
and cation exchangers. The solution can then be run through the column and the various
forms of As should be separated through their different retention times. The usual order
of elution is: As(III), MMA, As(V), DMAA. The main problem with this method is that
-14­
the columns will contain much packing and therefore would be subject to a large amount
of back-pressure. This can be overcome by using HPLC-type pumps and tube designed
to withstand these conditions.
1.2.6 Hydride Generation
Hydride generation has been extensively researched in conjunction with arsenic studies.
The main problems associated with arsenic studies have often been the speciation of the
various oxidation states of arsenic plus the removal of interferences due to other metals.
Indeed, the different toxicity of the inorganic forms alone has been a cause for concern
and therefore the need to determine the exact composition of the arsenic in the
environment has arisen. For example, As (III) is considered the most toxic form of
arsenic while arsenobetaine and arsenocholine are tolerated by living organisms and are
commonly found in crustaceans and mollusks such as lobster and shrimp. Although
several techniques have been developed over the years to speciate As (III) and As (V)
such as solvent extraction, these do not necessarily remove all the interferences from
other metals. Occasionally, the use ofthe wrong solvent and/or pH can even lead to the
co-precipitation of the arsenic and the interferent. Other problems are the high risk of
sample contamination arising from these techniques that are very labour intensive and
a low replication value (high chance of human errors and/or loss of sample due to
transferring the sample from one container to another several times). Another problem
is the very low concentration levels of the As species in the environment that therefore
require the use of highly sensitive techniques.
-15­
Many hydride generators can be directly coupled to an ICP-AES system. This step alone
reduces the risk of sample loss and contamination. Furthermore, the conditions under
which the hydride generation occurs (temperature, pH, ... ) can be carefully selected and
controlled thus effecting a good speciation of the arsenic compounds and also isolation
of the arsenic from the matrices. This method also can be partially automated thus
allowing for a higher sample throughput [1,3].
-16­
1.3 References
1. Advances in Environmental Science and Technology, Arsenic in the
environment, Part 1: cycling and characterization, Vol. 26, Ed: Jerome O.
Nriagu, Wiley & Sons, New-York, 1994
2. B. Weltz and M . Melcher, Influence of the Valency State of Arsenic on the
Degree of Signal Depression Caused by Copper, Iron and Nickel, Analyst, May
1984, Vol. 109, pp 573-575
3. D. Wickstmm, W. Lund and R. Bye, Determination of Arsenic and
Tellurium by Hydride Generation Atomic Spectrometry: Minimizing
Interferences from Nickel, Cobalt and Copper by Using an Alkaline Sample
Solution, Analyst, Nov. 1995, Vol 120, pp 2695-2698
4. F. 1. Peryea and R. Kammereck, Phosphate-enhanced movement oflead
arsenate contaminated topsoil and through uncontaminated subsoil, Water, air
and soil pollution, 1997, Vol 93, pp 243-254
5. G. Bombach, A. Pierra and W. Klemm, Arsenic in contaminated soil and
river sediment, Fresenius 1. Anal. Chem., 1994, pp49-53
-17­
CHAPTER 2
INDUCTIVELY COUPLED PLASMA ATOMIC EMISSION SPECTROSCOPY (ICP-AES) 2.1 The theory of ICP-AES
ICP-AES is based, as its name indicates, on the principle of emission. Simply put, excited atoms
and ions emit radiation of characteristic wavelength when they return to a ground state.
Therefore, by analysing the intensity of the wavelengths present in a spectrum, it is possible to
obtain a quantitative analysis ofa sample while identifying these wavelengths yields a qualitative
analysis.
To this end, an ICP-AES system must comprise of at least a source of excitation and a detector.
Figure 1 shows a typical arrangement for an ICP-AES which typically consists of:
A sample introduction system The ICP torch and gas supplies A radio-frequency generator An optical spectrometer Detectors and other electronics Computerised instrument control, data collection and analysis -18­
Figure 1: schematic representation of an ICP-AES [1].
RFsupp~
Monochromator
Transfer opffcs
Photo­
multiplier
signal
processing
electronics
pressure
gauge
@
spray chamber
perlstatllc pump
'V
Brain
-19­
AI supp~
ressure regulated
2.1.1 Sample introduction
Although most samples used in ICP-AES are in solution form, there are possibilities for the
introduction of gas, slurry or solid samples. Samples are most commonly acid digested which
minimises and in some cases eliminates interferences due to the sample matrix. In the case of
an organic matrix, this procedure has the advantage of improving the detectability of the analyte
elements as the energy that would normally be used to destroy the matrix can then be used to
excite the analyte. In the case of slurries or solutions, nebulizers are the most common means
of sample introduction. The samples is brought to the tip of the nebulizer through the aid of a
peristaltic pump where it is converted to an aerosol by Ar gas. The lighter part of the aerosol is
then transported to the ICP with the larger droplets removed in the spray chamber to afford an
almost homogeneous distribution of analyte elements.
Figure 2: Process of sample introduction and analyte emission [1].
0>
Ion
t
Excitation Ionization
> Photon Emissio
t
Excitation Dissociation
o
Atom
Molecule
Deso:vau"/
Aerosol
. ... . ..... . .
(])
t
Particle
Vaporization
0
Introduction of
Solid Sample
•• •• •
• •
•
Photon Emissio
••
Dissolution
<
-20­
Solid
Sample
These nebulizers can be subdivided into two classes: pneumatic and non-pneumatic. A
pneumatic nebulizer relies on compressed gas to generate the aerosol. These nebulizers are most
commonly concentric, cross-flow and V-groove. These are further described in chapter 4.
Pneumatic nebulizers have a tend to generate a finer aerosol, i.e. on average smaller droplets,
with an increase of the nebulizing gas flow rate [2]. However, they do have a major drawback.
Their sample solution capillaries tend to be very narrow, in some cases under 0.5 microns, which
tend to be clogged by solutions with high salt contents or high percentages of suspended
particulate.
The other category ofnebulizers, non-pneumatic nebulizers such as ultrasonic nebulizers, operate
on a variant of this process. The main component of an ultrasonic nebulizer is a piezoelectric
transducer oscillating at ultrasonic frequencies. The energy generated by this transducer is
applied to the sample which breaks down into small particles. The longitudinal wave propagated
by the crystal produces a pressure that breaks the surface of the liquid-air interface into an
aerosol. The efficiency of aerosol generation has been shown to be much greater than that of
other nebulizers, in some cases, up to ten times better. Several advantages can immediately be
deduced from this. More aerosol is produced and this is independent of the gas flow rate.
Therefore, more analyte can be transported to the Iep torch at much lower gas flow rates [3].
This translates into longer residence times in the plasma and better sensitivity and detection
limits. However, these advantages are lost with complicated matrices and other interferences
such as spectral interferences and background shifts, can be enhanced as well. This method of
nebulization can also result in greater memory effect [4].
-21­
,
I 1l\.t::.W '- "l1:.1
1.0 I ~
-:OS ~ ~ "to Y.
2.1.2 Plasma formation
Plasma formation occurs in a fused-silica torch. The torch consists of three concentric tubes,
most often made of quartz, through which streams of Ar and aerosol flow. A plasma is a gas in
which a significant number of atoms or molecules are ionised. Therefore, a magnetic field can
be made to interact with the plasma. One of these interactions, the inductive coupling of a time­
varying magnetic field with the plasma forms the basis ofICP-AES. The inductive coupling is
analogous to the inductive heating ofa metal cylinder [5]. Argon gas is ionised by a spark in the
presence of the radio frequency (RF) field of the induction coil. As a result, some electrons in
the spark gain enough energy to undergo inelastic collisions with Ar atoms. This collision may
result in a transfer of energy from the electron to the Ar atom which ionises, releasing another
electron which can in tum transfer energy from the coil to the gas. The magnetic field is
generated by the high frequency currents flowing in the induction coil. The lines of force ofthe
magnetic field are axially oriented inside the torch and follow elliptical paths outside the coil [I].
The electrons and positive Ar ions formed are both accelerated by the high frequency field of
the coil. However, due to their far lesser mass, the electrons are accelerated to much higher
velocities than the Ar ions, thus creating a domination in energy transfer by electrons. A steady­
state plasma is obviously produced when the rate at which electrons are released by ionising
collisions equals the rate at which they are lost by recombination processes. Ion-electron
recombination in tum, emit light, producing a continuous spectrum corresponding to the
distribution of ion kinetic energies in the plasma [I].
As the plasma thus achieved can obtain a gas temperature of between 5,000 to 10,000 K in the
region of maximum eddy current flow, some form of thermal isolation must be provided to
-22­
protect the quartz torch. Reed's vortex stabilisation technique uses a tangential flow of Ar in the
plasma. The flow of Ar cools the inside walls of the outermost quartz tubes and centres the
plasma in the tube [6]. The position of the RF tube which generates the plasma is such that the
plasma itself is anchored at the exit end ofthe concentric tube arrangement. The innermost tube
is used to introduce the aerosol with Ar carrier gas to the centre of the plasma.
Figure 3: Zones in the ICP [1].
--+-----~
Tail Flame
Normal Analytical Zone (NAZ
Induction Zone
Initial Radiation Zone (IRZ)
~
Pre-heating Zone (PHZ)
2.1.3 RF Generator
The RF generator therefore is required to supply an alternating current to the induction coil
which is used to create and maintain the ICP's plasma. This current must have a selectable
frequency as different samples call for various plasma states. There are again two main types
of generators, "free-running" and "crystal controlled". The "crystal controlled" generator
maintains a constant frequency regardless of the plasma impedance due to the presence of a
piezoelectric crystal. This crystal usually operates in the low milliwatt output range as it cannot
carry large currents. The high frequency modulation of the RF generator results in both high
precision of analysis and enhanced signal to background ratio. By contrast, the "free-running"
-23­
generator allows the frequency of the oscillating current to vary according to the impedance of
the plasma. The stability of the RF generator in both cases is ensured by water cooling [1].
2.1.4 Detection of emission
A photo multiplier tube (PMT) in the spectrometer is used to convert the light intensity into an
electrical signal that can be quantified and therefore related to the concentration of the analyte
in solution. As there are two main types of spectrometers available, two different set-ups are
possible here. In the case of a simultaneous scanning ICP, an array of detectors is used to
measure a number of lines at the same time, while a sequential ICP measures one spectral line
after another by using a moving grating, giving an unrestricted choice of wavelengths. As
discussed in chapter 1, simultaneous systems have a higher sample throughput as more elements
can be analysed per time period. This is due to the fact that the wavelengths at which the
elements are analysed are pre-selected in the manufacturing process. A direct result of this
however is the lack of control over the analysis. Sequential systems have the advantage of
flexibility ofwavelength selection. This in effect, allows for variations in analyte concentrations
and matrix types and the removal of interfering peaks by the selection of a different wavelength,
with the major drawback of much longer analysis time [1].
2.2 Interferences in ICP-AES
There are several sources of error in ICP-AES. These can mainly be attributed to plasma
affecting factors such as matrix effects, wing overlaps and chemical interferences. Matrix effects
are usually less in ICP-AES than in flame or graphite furnaces. This is due to the environment
of the plasma. Incomplete atomisation, especially of refractory species such as oxides, has
plagued flame and graphite furnaces. Water and oxygen atom traces also cause incomplete
-24­
atomisation. In the plasma however, two factors minimise this type of matrix effect: an "inert"
environment, as well as, high temperatures. Although the plasma is made up of mainly Ar gas,
it is not totally inert. Water vapour from the sample and aerosol can produce a certain level of
oxygen atoms within this plasma. the concentration of the oxygen atoms is similar to that in
flames, approximately 2
* 10 16 cm· 3 [1].
2.2.1 Matrix Effects
Matrix effects however cannot be fully discounted. The very type of mineral acids or organic
solvent used can play an important role in the emission signal. For instance, CI from HCI has
been shown [1] to weaken the signal from several elements such as Fe. While this can be in part
compensated by the use ofmatched standards and blanks, this phenomenon cannot be completely
removed. Organic solvents also playa major role in matrix effects [7]. As organic solvents are
quite often made of large, complicated molecules, part of the energy normally used to excite
atoms and ions is diverted to break down these molecules. This in effect results in a colder
plasma and an entirely different emission spectrum. Organic solvents and very viscous solvents
such as phosphoric acid, can also cause the plasma to extinguish itself due to high changes in
electrical and thermal properties [8, 9]. Further problems are related to the vaporisation of the
solvent. Not all solvents have the same viscosity which can result in different transport rates and
nebulization rates as in the case of phosphoric acid. In many cases, excessive solvent loading
of volatile solvents can be a problem. This can be circumvented by slower aspiration rates for
readily nebulizable solvents. This can be shown by benzene, a very volatile solvent, where the
maximum attainable aspiration rate is as low as 0.1 to 0.2 ml/min for an analytically useful
plasma [8, 9].
This also ensures smaller primary droplets and increases the evaporation
efficiency, allowing for a reduction in the nebulizer pressure [10-13].
-25­
2.2.2 Spectral overlap Spectral overlap, one of the biggest problems of ICP-AES, is a direct result of the high temperatures required to minimise vaporisation-type matrix effects and to maximise emission [14]. The usefulness ofICP-AES is directly derived from its ability to excite atoms to high energy levels thus, providing high signal-to-background (SIB) and signal-to-noise (SIN) ratios for many spectral lines. As many species are injected into the plasma which emit profusely, a high level of spectral interference can be expected. In the case of major elements this can be often discounted however, in the case oftrace or weakly emitting elements, wavelength selection is necessarily dependant on other spectral features near that wavelength. Wavelength tables [15] have long been established with an average of294 spectral lines being emitted per element with more than 100,000 wavelength lines being recorded in the 200-1000 nm range. This abundance of wavelengths can cause two types of spectral line overlap: Direct spectral overlap and Wing overlap. 2.2.2.1 Direct overlap Direct spectral overlap occurs when two or more elements emit light at the same wavelength. In this case, no amount of increased resolution would achieve a separation of the two lines. Figure 4: Direct Overlap [1, 16] Wavelen
-26­
2.2.2.2 Wing overlap
Wing overlap occurs through Doppler broadening due to the high temperature ofthe instrument.
In this case, several steps can be taken. Another line could be selected or better resolution could
be sought. Inter-element correction is not recommended as small changes in the profile setting
ofthe spectrometer can seriously alter the degree ofoverlap. Another source ofwing broadening
is resonance broadening. In this case, the width of the line profile can be said to broaden in
proportion to the number of collisions between excited and ground state atoms . For example,
the Ca I 393 .37 nm and Ca II 396.85 run lines, even at a concentration of 1 mg/ml can cause a
non-linear elevation of the background as far as 10 nm away from the line centre. Therefore, a
Ca matrix would severely interfere in the determination of Al at either the 394.4 or 396.2 run
lines.
Figure 5: Wing Overlap [1, 16]
-27­
2.2.2.3 Continuum radiation
Another source of error in an rcp spectrum is continuum radiation. This radiation occurs from
many sources such as electrons, Ar, and matrix species, both atomic and molecular As many
operating parameters can affect the profile of the continuum radiation, several levels of
background overlap can occur, as described in figure 6.
Simple changes in operational
parameters such as the Ar gas flow rate can severely alter the background radiation. A change
in the matrix or recombination of ions such as Mg with free electrons can cause radiation over
a wide range. These do not necessarily occur in the ground state and therefore cannot always be
quantified. By using blanks, much ofthis background radiation can be mathematically removed.
However, it is often necessary to use a dynamic background correction method in order to
separate out this phenomenon.
Figure 6: Continuum Radiation [1, 16]
-------­ ------------­----------­ ---------------- (iii)
-----­ ---------------------------­ -------------- (ii)
- -------------------------------------­ --------- (i)
Wavelen
A
-28­
2.2.3 Stray light
Stray light should always be considered when using a spectrometer. Stray light is defined as
light which reaches the detector unintentionally. There are several sources, mainly occurring
from defects in the grating, for stray light such as line "ghosts" from periodic errors in the grating
spacing; satellites, near scatter, or "grass" resulting for regular or irregular grating defects; or far
scatter which occurs several bandpass away from an affected line, generated by a localised
roughness in the groove shape. As with all man-made instruments, dispersive devices are not
perfectly made, which results in instrumental beam aberrations. This problem is carried over to
other monochromator devices such as slits, slit mechanisms or simply optical misalignment.
Particles in the optical path, degradation ofanti-reflection coatings on optics can also cause stray
light. Several methods to minimise stray light interferences have been devised. These include
solar blind PMT which will not respond to radiation above 350 nm, double monochromators, as
well as, correction coefficients and linear scanning across the analyte line to provide dynamic
correction.
-29­
2.3
References
1. M. Selby, B. Sturman and 1.B. Willis, Analytical methods for the Liberty spectrometer,
Varian Australia Pty. Ltd., 1991, pp 1-19 .
2. L. Ebdon and M.R. Cave, Analyst, 1982,106, ppl72.
3. K.W. Olson, WJJr. Haas and V.A. Fassel, Anal. Chern., 1977,49, pp 632.
4. R.H. Scott, V.A. Fassel, R.N/ Kniseley and D.E. Nixan, Anal. Chern., 1974,46, pp 75.
5. V.A. Fassel and R.N. Knisseley, Anal. Chern., 1974,46, pp 1155A.
6. V.A. Fassel, "Electrical Plasma Spectroscopy" XVI Colloquium Spectroscopicum
Internationale, Adam Hilger, London, 1973.
7. 1. Farino, 1.R. Miller, D.D. Smith, R.F. Browner, Anal. Chern., 1987,59, pp 2303.
8. L.M. Faires and T.M. Niemczyk, Appl. Spectrosc., 1983,37, pp 553 .
9. M.R. Tripkovik and I.D. Holclajtner-Antunovic, 1. Anal. At. Spectrom., 1993,8, pp
349.
10. D.D. Nygaard, R.G. Schreicher and 1.1. Sotera, Appl. Spectrosc., 1986, 40, pp 1074­
1075.
11. R.I. Botto, Spectrochim. Acta, 1987, 42(B), pp 181-199.
12. F.1.MJ. Maessen, G. Kreunig and J. Balke, Spectrochim. Acta, 1986, 41(B), pp 3-25.
13. T. Brotheron, B. Barnes, N. Vela and J. Caruso,1. Anal. At. Spectrom., 1987,2, pp
389-396.
14. L. Ebdon, E.H. Evans and N.W. Barnett, 1. Anal. At. Spectrom., 1989,4, pp 505-508
15. R.C. Ng, H. Kaiser and B. Meddings, Spectrochimica Acta, 1985, 40(B), pp 63-72.
16. M. Thompson and J.D. Walsh, Handbook ofInductively Coupled Plasma Spectrometry,
Second Edition, Blackie and Son, 1989.
-30­
CHAPTER 3
ION EXCHANGE
3.1 Introduction
Ion exchange is a fairly new technique in chemistry. This natural phenomenon has come
to the fore of research in the past 50 years as indicated by scientific pUblications [1].
This technique, while on the surface fairly simple, has become a very useful tool for
speciation and extraction. Over the last half century this method has been extensively
developed, mainly in the purification and demineralisation of water. One of the earliest
records of ion exchange comes from Aristotle who found that sea water loses some ofits
salt content by filtration through certain types of sand [1,2]. This phenomenon was then
largely left alone until recently when the first synthetic, industrial ion exchanger was
developed by Horm and Rumpler [1] . This development, among others, sparked research
into new resins that had superior chemical stability at both low and high pH values, and
better properties than the inorganic ion exchangers used at the time. In 1935 Adams and
Holmes were able to synthesise the first organic ion exchange resins [1].
The principle of ion exchange as mentioned before is fairly simple. Ion exchange is the
reversible exchange of ions between a solid matrix such as the ion exchanger and a
solution. Ion exchanger must therefore be insoluble, solid material with ionogenic
groups. Ionogenic groups are charged centres to which anions or cations can bond.
Through an electrolyte solution, these counter ions can then be stoichiometrically
exchanged for equivalent amounts ofsimilarly charged ions [1,3]. This phenomenon has
proven to be highly useful in the recovery of metals from industrial wastes and the
-31­
separation of rare earth metals. Within a chemical laboratory context, ion exchangers
have proven to be highly useful in a variety of sample preparation and manipulation
procedures.
These include the removal of interferents from samples, and pre­
concentration of trace levels of metals as well as the separation or speciation of
equivalent species from samples. In this work, the main use of ion exchangers was the
last one, namely the speciation of As and less important, the removal of other ions
present in the sample.
3.2 Ion Exchange Resins
Although several different substances exhibit some ability to act as ion exchangers,
synthetic organic resins have become the major players in this field . This prominence
is largely due to the ability of synthetic chemists to generate fairly ion specific resins
[1,3]. These resins consist of a 3-dimensional network of organic polymers such as
styrene-divinyl benzene, to which ionic groups such as NR3 + are chemically bonded. In
essence these groups are the ionogenic groups ofthe ion exchanger. The major drawback
of these groups is the hydrophilic character which the polymers adopt which then allows
them to dissolve in aqueous solutions. To prevent this, a cross linking compound such
as divinyl benzene is introduced during the polymerisation of the resins in order to
interconnect the hydrocarbon chains. Through this simple chemical process, the
polymer's molecular weight and mechanical stability can be dramatically increased while
its solubility decreases. This high level of cross linking however does have two side
effects which may not always be beneficial. Primarily, the cross linking will prevent the
resin from extensive swelling in solutions. This in tum inhibits the mobility of counter
-32­
ions and therefore lowers the exchange rates. From the name of the resin, ego Dowex
1X8 resin, we can tell the percentage of cross linking (8% in this case).
Synthetic resins as mentioned before have become the preferred ion exchange medium.
This is due to the fact that the ion exchange behaviour ofthe resin is a direct result ofthe
number and nature of the fixed ionogenic groups. The number of groups fixes the
capacity of the resin while the nature of these groups affects the exchange equilibria.
These resins tend to be pH dependent as the functional groups are often acidic or basic
in nature. This means that they will tend to be more readily ionized at certain pKa (or
pKb). For instance, when the pH is lower than the p~ of an acid functional group as is
the case for NR3-type functional groups, then the ionogenic groups will not be ionized
and no ion exchange will be possible. It is therefore critical that the proper conditions
are used to allow the maximum potential of these resins [1].
The resins are often divided into three major categories, based on their ionogenic groups.
These are: Anion, Cation, and Chelating ion exchange resins.
3.2.1 Anion Exchange Resins
When the fixed ionogenic groups have positive charges (or are basic groups), the resin
is classified as an Anion Exchange Resin. These tend to have quaternary ammonium
functional groups as their ionogenic centres [3]. By varying the degree of substitution
of these terminal amine groups, it is possible to engineer strong, medium and weak base
anion exchange resins with highly different pKb.
Strong anion exchange resins contain highly substituted ionogenic groups such as
-33­
trimethylamine, which ionizes at pH values from 1 to 15 [2].
These resins will
deteriorate at temperatures above 60°C, but will withstand most common solvents as well
as certain oxidising agents. Weak anion exchange resins are typically lower substituted
ammines or ammonia, ego NH 2R+ and NH3+which are only protonated below a pH value
below 9. The third class of anion exchange resins, medium anion exchange resins, falls
between these two and contain both strong and weak base functional groups and will not
therefore withstand high pH values.
3.2.2 Cation Exchange Resins
In this case, the counter ions are bonded to negatively charged ionogenic groups in the
resin. These are the opposites to anion exchange resins in that the functional groups are
acidic in nature with varying strength. There are also three main classes of Cation
Exchange Resins, strong, medium and weak, based on their PI<... values.
Sulphonic functional groups (S03-) fonn the basis ofthe most common strong acid cation
exchange resins. These are typically bonded to benzene rings within the organic polymer
matrix. The PI<... of these groups is approximately equal to one and therefore they will
be ionized by the presence ofa strong acid [2] . Weaker acid groups such as PO(OH)2 or
OPO(OH)2 are the typical ionogenic groups ofmedium cation exchange resins. As these
groups are not found dissociated at low pH values, the minimum pH at which these are
useful is 5. Weak cation exchange resins are made up of carboxylic acid functional
groups as a rule. These are very pH dependent and are used at a pH ofbetween 6 and 14.
3.2.3 Chelating Ion Exchange Resins
This class of resins are probably the most selective. The functional group is, in this case,
a chelating agent which will interact with a limited number or metal ions [3]. Therefore,
by carefully selecting the chelating agent, the resin's affinity for certain metal ions can
-34­
easily be designed while the size, charge and other physical characteristics of the metal
ions and resin will be of secondary importance. By careful pH control, the selectivity of
this resin can be further enhanced. Trace element concentration and separation ofcertain
elements is enhanced in this method as the metals are chemically bonded to the resin.
However, as the exchange rates are determined by either secondary chemical reactions
or particle diffusion [2], the kinetics of the exchange reactions are unfavourable.
3.2.4 Capacity
Each resin has a specific capacity which is best defined as the number of ionogenic
groups contained by 1 g of the dry resin [1,3,4]. Resins with higher capacity will
therefore accommodate more ions though the trade off is more difficult elution and
therefore will require higher eluent concentrations to effect complete elution. Low
capacity resins are generally preferred in ion chromatography as the separation of ions
should be achieved quickly and with low eluent concentrations.
3.2.5 Selectivity
3.2.5.1 Selectivity coefficient
Ion exchange, like many other reversible chemical reactions can be expressed as a
general equilibrium reaction:
(1)
Where a and b represent the molar quantities of exchanged ions A and B respectively.
Subscripts rand s represent the phase ofthe ion, r for the resin phase while s stands for
solution. The equilibrium constant for the reaction can then be represented as:
-35­
[Ar]G *[Bs]b
(2)
[As]G*[Br]b
Where [Ar] and [BJ are the concentrations of ions A and B in the resin and solution
phases respectively. Normally a and b are expressed in terms ofmmol/l for solutions
and mmol I g for resins when a and b are not equal. [5] From the affinity of ion
exchangers for certain species, it is possible to deduce that the equilibrium constant will
not equal one. From this, it is possible to infer a selectivity coefficient indicating the
specific affinity of a resin for an ion. Therefore, should the selectivity constant K/ be
greater than one, then the resin can be said to have a higher affinity for ion A than for ion
B. This selectivity coefficient can also be used to determine whether ions would be
useful as eluents ego ions (A) with high K/ values are often good eluents.
3.2.5.2 Factors influencing selectivity
Many factors can influence the selectivity of a resin. Some of these have been referred
to previously such as the nature ofthe resin as well as the type and concentration ofthe
ions in solution. The exchange of ions on the resin typically involves the formation of
chemical bonds between ions in the solution and the ionogenic groups of the resin. The
affinity of the resin for certain ions is directly related to the interaction between the ions
and the ionic groups or the matrix. The stronger this interaction when these bonds are
formed, the higher the affinity of the resin.
At the same time, ion exchange can also be due to purely electrostatic interactions instead
of chemical bond formation.
As electrostatic interactions can be proven to be
-36­
proportional to the charge of the ion while being inversely proportional to the distance
between the charges, ion exchange resins will have a preference for ions of higher
valencies as well as ions with smaller, solvated equivalent volumes and a greater
polarizability. Due to this effect, the ions tend to be localized in the neighbourhood of
the fixed ionic groups.
Selectivity can be decreased through secondary reactions such as complex formation.
However, should cations form anionic complexes with ligands in solution, anion
exchangers tend to prefer the anion which will form the stronger complex or the complex
with the greater average ligand number [3].
By decreasing either the temperature or the solution concentration as well as increasing
the degree of cross linking, the selectivity ofa specific resin may be increased [1].
3.2.6 Distribution coefficient
The distribution coefficient, like KB A is a measure ofthe affinity ofa resin for a particular
solute ion A and is defined as below:
(3)
As previously, [Ar] and [As] are the concentrations of the exchanged ion A in the resin
(mmol/g) and in the solution (mmolll) respectively. Therefore, a large distribution
coefficient would signify a greater affinity of the ion for the resin than for the solution.
This value, D g , however is specific for the ion exchange resin and the eluent conditions
used. This value also tends to increase with increasing atomic weight. [3]
-37­
3.2.7 Kinetics of Ion Exchange
An exchange resin's usefulness in either separation or concentration processes is often
determined by the rate at which ion exchange occurs. As the ion exchange process is
basically a diffusion process, the following equation can be used to represent the overall
process:
A+B",,"A
+B s
s
r -''I
(4)
The simpleness ofthis equation belies the fact that the process often involves five distinct
reactions occurring simultaneously. Three further steps can be used to describe this
reaction based on the electroneutrality principle, which require that the flux of ions A and
B are equal: [5]
(i) The diffusion of ions A from the solution to the surface of the resin
particle occurs while ion B diffuses from the resin surface into the
external solution.
(ii) Ion A diffuses from the resin surface to the ion exchange position in the
particle while ion B diffuses from the ion exchange position to the surface
of the resin.
(iii) Ion B and ion A are exchanged at the exchange position.
From any of these reactions, the ion exchange rate can be extrapolated [1,5]. Generally,
reaction (iii) is not considered to be the rate-determining reaction as reactions of free
dissociated ions in aqueous solutions are very fast [5]. In the case of chelating ion
exchange resins, this rule does not usually apply as complexing reactions are very
slow[ I] and therefore, the chemical exchange of the ions will then be the ratedetermining reaction.
-38­
This in effect leaves us with two reactions, (i) and (ii) from which the ion exchange rate
can be determined. These two reactions are diffusion processes. In reaction (i), the
diffusion rates of ions A and B will be primarily determined by the degree to which the
solution is agitated. Under normal condition, the solution is well agitated in both batch
and column processes. Hydrodynamic factors, however, indicate that a thin film of
solution which directly surrounds the resin particle will be static, eg no agitation will
occur in this film . The name of this film, the Nemst film or the Nemst diffusion layer
(8) is a tribute to Nemst who developed the concept. This thin film then can be
considered to be the limiting factor for the diffusion of the ions A and B. As these ions
are well mixed in the bulk of the solution, the rate of diffusion ofthe ions depends on the
film diffusion rate. [1]
Ion exchange reSInS with low degrees of cross linking, small particle sizes, and
insufficiently mixed, dilute solutions are especially affected by film diffusion. The
overall reaction rate can be increased as follows in these cases:[5]
o The Nemst film may be decreased through increased agitation of the solution in
batch processes, or increased flow rates in column processes.
o The concentration of exchangeable species can be increased.
o Smaller resin particles will result in an increased surface area of the ion exchange
resIn.
o Higher temperature should increase the diffusion rates.
-39­
Should reaction (ii) be the rate determining step, then particle diffusion is involved. This
means that the diffusion of the ions within the particle becomes more important[5].
Generally, this indicates that the resins used have a high degree of cross linking and are
made up of large resin particles. The diffusion rate can be increased through the use of
smaller resin particles, higher reaction temperatures, higher capacity resins, as well as
through the use of a resin with a lower degree of cross linking which will result in an
increase in the porosity of the resin particles [2,5].
When solutions have a concentration of 0.001 molll and lower then, film diffusion is the
rate determining step.
For solutions with concentrations near 0.01 molll, particle
diffusion starts to playa bigger role and at 0.1 molll or higher, particle diffusion becomes
the rate determining step.
As increasing the reaction temperature and using smaller particles will increase the
reaction rates of both film and particle diffusion, the overall ion exchange reaction rate
can be increased without first determining which of these two processes is rate
determining.
-40­
3.3 References
1. F. Helfferich, Ion Exchange, McGraw-Hill Book Company, New York, 1962 2. D.T. Gerde, J.S. Fritz, Ion Chromatography, Second Editon, Heidelberg, 1987 3. M. Marhol, Ion Exchangers in Analytical Chemistry: Their Application to Inorganic Analytical Chemistry, Volume XIV, Elsevier Scientific Publishing Company, Amsterdam, 1982 4. F.e. Smith, R.e. Chang, The Practice ofIon Chromatography, Wiley and Sons, 1983 5. O. Samuelson, Ion Exchange Separations in Analytical Chemistry, Wiley and Sons, 1963 -41­
Chapter 4
Comparison of a Micro-concentric nebulizer with a Meinhard nebulizer
4.1 Introduction to nebulizers
As with many emission and absorption techniques, ICP-AES relies heavily on the sample
introduction efficiency. Most analytical samples do not conform in size, concentration,
viscosity or even phase. The sheer array of sample types makes it difficult to standardize
sample introduction methods. However, some specialised equipment has been developed
for the ICP-AES to facilitate this process. In the case of solutions, the sample introduction
ofchoice is a nebulizer. These involve several categories such as, pneumatic nebulizers and
ultrasonic nebulizers.
Research into nebulizer design and performance, the effects ofspray-chamber design, factors
influencing droplet size, formation, and distribution, and methods for the characterization
of nebulizers has intensified, as demonstrated by the increased volume ofpapers published.
4.2 Theory of nebulizers
4.2.1 Pneumatic nebulizers
The most widely used solution sample introduction method is Pneumatic nebulization. The
efficient generation of aerosol with a small average droplet size requires a high gas velocity
in the nebulizer. To generate a flowrate of aerosol carrier gas of about 1 lImin, a nebulizer
must have very small orifices, typically 200 ~m for the gas and the liquid flows. This means
that the manufacture and alignment of these nebulizers can be quite difficult. The main
disadvantages of these nebulizers is the ease with which they can be blocked.
-42­
Fine
suspended matter present in the solution or even fibres introduced when the tubing is cleaned
before each sample introduction can result in the blockage of these orifices. Should the
solution analysed have a high salt content, then the salt tends to precipitate out at the tips of
the nebulizer, leading to partial or complete blockage. This effect, the "salting out" effect
occurs most commonly around the annular gas aperture of concentric nebulizers. By using
a humidified argon gas flow (a 'gas-wetter'), this problem can be alleviated somewhat or
alternatively, the tip of the nebulizer can be washed between sample aspirations (a 'tip­
washer').[ 1]
The most common pneumatic nebulizers are:[l]
•
concentric nebulizer, e.g. the Meinhard nebulizer
•
cross-flow nebulizer
•
Babington-type nebulizer
•
glass-frit nebulizer
•
grid nebulizer
•
jet-impact nebulizer
4.2.1.1 Concentric nebulizers
The concentric design offers a greater mechanical stability than the adjustable cross-flow
designs at the cost of more blockage. Both the concentric and cross-flow designs make use
ofthe Venturi effect, using the reduced pressure resulting from a fast moving gasjet to cause
the solution to be drawn into this gas jet and to be broken into droplets of various sizes.
Because of the Venturi effect, concentric and cross-flow nebulizers are generally self­
feeding, removing the need to pump the sample. However, in the case of samples with
-43­
differing viscosities, the feed rates will differ and a peristaltic pump should be used to
minimize these differences.[2]
After primary nebulization, a spray chamber is used to filter out the larger droplets, and the
finer droplets pass on into the excitation source.
Some designs use an impact band
immediately in front of the jet for the secondary nebulization of the larger droplets.
The finer details of these nebulizers such as the alignment of the gas and solution tubes and
overall geometry depends on the manufacturers. Some designs have a fixed geometry while
others allow the alignment ofthe gas and solution tubes to be adjusted by the user. Several
designs are currently in use though the Meinhard type C nebulizer is reported to be more
tolerant of high-solids sample solutions than types A and B.
4.2.1.2 Cross-flow nebulizers
Though the cross-flow design is much more tolerant of solutions with high salt contents than
the concentric design, both systems are subject to periodic blockages by stary particulate
matter, and to salting out, which is often caused by the reduction in temperature that
accompanies the Venturi effect. Problems associated with salting out tend to occur in
solutions in which the total salt content is greater than 1 percent. The cross-flow nebulizer,
as its name indicates, relies on the gas flow and the sample solution flow coming at 90
degrees to each other. The gas flow is typically delivered at a horizontal position with the
solution coming vertically up a capillary. The fast moving aerosol gas jet causes the droplets
formed at the top ofthe solution capillary to from a spray which can then be passed through
a spray chamber.[l]
-44­
Here again, the overall design can vary as the capillaries can either be fixed in position by
the manufacturer to allow greater stability and reproducibili ty or, they can be adjusted by the
user to allow for better nebulization of different solutions.
4.2.1.3 Babington-type nebulizers
Due to the design of these nebulizers, this particular model is commonly used for solutions
with high salt content, such as 10 percent sodium or more, and slurries of solid powders.
The original design involved the solution being poured over a hollow sphere with a hole in
it through which the nebulizing gas issued in a jet. Since then the design has undergone
many modifications. The modem Babington-type nebulizer has the sample solution or slurry
trickling down along a 'V' -shaped groove. The sample is fed into the V-groove by a
peristaltic pump. A gas jet which issues from a capillary hole in the middle of this groove,
disrupts the solution flow and causes nebulization. This design has the major advantage that
it is virtually unblockable in routine analytical work. Another advantage is that Babington­
type nebulizers which match the performance of commercial nebulizers can be constructed
in the laboratory from simple materials. [1]
4.2.2 Ultrasonic nebulizers
Ultrasonic nebulizers are somewhat less common than pneumatic nebulizers even though
their sensitivity has been reported as being up to 4 times higher. This is higher sensitivity
is mainly due to the proportion of small particles (less than 10 J.!m) generated in the aerosol
which is higher than that produced by pneumatic nebulizers. The sample solution is pumped
onto a vibrating crystal transducer (1 to 10 MHz) which generates the aerosol.[3]
-45­
One of the major drawbacks of this type of nebulizer is that often a desolvation step is
required to reduce the water loading in the aerosol and therefore keep the plasma ignited.
This means that a desolvation apparatus (heater and condenser) has to be used which
increases the path followed by the aerosol and increases the analysis time as well as the
memory effects. The increased memory effects in turn require longer wash-out times which
again lengthens the analysis time. Ultrasonic nebulizers also suffer from poor long-term
stability, and their performance is sensitive to small changes in the operating parameters.
However there are several benefits to using this type of nebulizer.
With or without
desolvation, ultrasonic nebulizers can give detection limits that are an order of magnitude
or more lower than those normally obtained by ICP-AES. With a desolvation system, Taylor
and Floyd have reported detection limits that were 5 to 10 times lower for 31 elements while
Fassel and Bear report that on a continuous-flow nebulizer with a desolvation system the
detection power was improved by five- to fifty-fold.[l]
The principal cause for improvement of detection limits and signal to noise ratio is in most
cases an improvement of the droplet size generated by the nebulizer. Smaller droplets lead
to lower background emission with, at the same time a slight drop in signal intensity due to
lower analyte levels. Further improvements can be derived from an overall raising of the
plasma temperature as the plasma solvent loading is reduced. This means that less of the
energy contained within the plasma is diverted to breakdown the matrix. Not only does the
droplet size play an important role in the signal to noise ratio, but the exact size distribution
has a major impact too . With smaller droplets comes usually lower droplet weights. This
translates into more of the analyte being transported, through the spray chamber where the
-46­
heavier droplets are removed, to the plasma. In other words, as the spray chamber tends to
let a range of droplets through, the ideal nebulizer would have a Gaussian particle size
distribution to match the spray chamber's operating range. Should a nebulizer produce only
some low mass or diameter droplets, with the majority being heavy droplets, that nebulizer
would be wasting most of the solution. Ideal nebulizers would therefore be those which
allow the introduction of a "dry" analyte as opposed to those used currently which depend
on the creation of an aerosol [I].
Nukiyama and Tanasawa [4] developed an equation describing the particle size distribution
for concentric nebulizers:
(5)
where p is the density (g/cmJ), a is the surface tension (dynes/cm), J..l is the coefficient of
viscosity (dynes/cm\ c is the relative velocity between the gas and the liquid (cg-c,)(rn!s),
and Q, and Qg are the volume flows ofthe liquid and gas, respectively. By playing with these
parameters, one at a time, optimal operating parameters can be found for each nebulizer.
4.3 Experimental for the particle size distribution:
The ).l-Laser Particle Analyser ().l-LPA) was mounted on a stable base (see appendix). A
holder for the nebulizer was placed on a screw next to a ruler which allows us to determine
the distance from the nebulizer tip to the cell of the ).l-LP A, see diagram. Several sleeves
were designed to allow the various nebulizers to fit into the holder. A 0.5 mm i.d. tube was
used across the peristaltic pump with a 0.4 ).lm i.d. tube making the connection to the
-47­
nebulizer. In the case ofthe Micro-Concentric Nebulizer (MCN) the size ofthis tubing was
of particular importance as the inner diameter ofthe MCN is 0.2
)lm
and therefore the MCN
can get easily blocked by particulates in the solution. Special glassware was also designed
(see appendix) for use with the MCN to minimize contamination and filter out particles
larger than 0.45
)lm.
The pump speed was varied from 200 rpm to 900 rpm, the Ar pressure
from 40 kPa to 200 kPa and the distance from the nebulizer tip to the )l-LPA from -1.0 cm
(nebulizer tip inside the cell) to 2.0 cm. Four nebulizers were considered: MCN, Meinhard
type 'C', V-Groove and Cross-Flow. The effect ofa surfactant with the MCN and Meinhard
nebulizers was also explored. The resulting data was plotted as number of particles vs
particle diameter size. This allowed us to explore the effect of changing each parameter of
the Nukiyama and Tanasawa equation in turn.
Results and Discussion:
Most of the graphs appear to have two common features. These are the initial peak ca. 0.45
)lm
and a secondary peak starting near 1 )lm. For this reason most nebulizers operate with
a spray chamber which removes the larger droplets from the aerosol. The Cross-flow and
V -Groove nebulizers seem to follow the same trend although in most cases the initial peak
may have started well below the detection limit of the )l-LPA, as can be seen from the
decreasing shoulder at the beginning of the graph. The V-Groove's secondary peak is also
found at a much lower diameter range, between 0.5
)lm
and 1.0
)lm.
When comparing graphs 1-4, it can be seen that the particle count under these conditions is
gr~atest
for the MCN and smallest for the V-Groove. This can be attributed to the fact that
the flow rate delivered by the tubing is optimal for the MCN and probably insufficient for
-48­
Graph 1: MeN at different positions 2500r-----------------------------------------------------------------------~
- - + - - MCN at 0 em ...
MCN at - 1.0 em *-MCN at - 1.5 em - -x - MCN at + 1.0 em .... ·111··· ··
2000 ---- - . - ---
.
..
'
1500
,
.....
t:
::l
0
U
~
u
.
y....--­
.'
1000
tC'C
./
.,.­
x'"
_.m
./
/'
a.
/'
.,.. . 19"' ­
/'
~.'
.... ;._....>t---­
500
,,.-. . .: --.~-.:;.--'"
: ..:.. ."'- . ..- ...- . - -=- . - .
0.5
-500
-.-~/
._ . 1I:'t"'
. --. - ....
----~~~.~--~~~
~
:.~.--~~--~--~:_--_;--~~---+----~----"
01
-.~
_ ·on
1
-
1.5
2
2.5
3
3.5
4
4.5
L ______________________________--'
Particle Size (microns)
Graph 2: Meinhardt at different positions 2000~------------------------------------------------------------------------------------
.m--- .....
- - ... - - Meinhardt at 0 em
- -m-- ' Meinhardt at - 1.0 em
Meinhardt at - 1.S em
~ Meinhardt at + 1.0 em
/
--* -1S00
/
/
"
,/' , '••-----.; ~ -"'--e..- -­
'
. '-. ....: : -,
-.
',,- 'W
N .'
>
,
,./ i
/
/
/
o
I
/
/
I
U
Q)
/
.I
u
/
tC\l
/
SOO
I
r
/
I
o I'
_ , __
~
,/" .•..
~.-. -- --_-.0!1;:~':.-.-.-,'
­...'-. -- ........ - -- - - - -.- - - - -- ...... -- ........ - -- -­
• -_.
O.S
-SOO
/
!
t:
::l
a.
/
/
.... 1000
/
•___
.,,_._~ --w-
/
I
I
1.S
2
2.S
3
3.S
4
.. -- .•
4.S
~----------------------------------------------------------------------------------------~
PArticle Size (microns)
Graph 3: V-Groove at different positions 40r-------------------------------------------------~
,
35
....... V-Groove at 0 em -m
V-Groove at + 1.0 em
30
25
-g
t: 20 () Q) u
tra 15 c..
•
10
5
o
I
."" .. - . , ~~-~
..........
0.5 1
S
1.5
~
2
~. ------.£:.1
I-
2.5
3
m
--------m-----~~e~------_:~:_-----3.5
4
~
4.5
-5 ~--------------------------------------------------------------------------------------------------~
Particle Size (microns)
Graph 4: Cross-Flow at different positions 250~--------------------------------------------------------------------------------------------~
- -+- . Cross-Flow at 0 em -s--- Cross-Flow at + 1.0 em 200
/
/
""
...
- . -+- . - - . - . -.
"
"".­
..... 150
c:
/
:l
0
j¥'
U
"" ""
./
Cl)
u
./
t
ra
0..
""
,/
100
"
((
t'
50
o
I
o
......... ..,... . _-­
0.5
1.5
2
2.5
Particle Size (microns)
~
3.5
4
4.5
5
Graph 5: MeN at Ar 100 kPa =
~-------------------------------------------------------------------------------------------.
800
700 I...... .
,
500
t:
::::l
o
U
~
~.
. ......J: .........
... . ~a- .. ~ .. :.:: : Ii:::... - .
600
-t­
...
,
,
400!
" /
. ·m
./
./
,-
,-
~,
, , , 'U"
,
~ j
300
.. - .. - .. ". ..
. /
jl,
ell
u
,,
,I:
r;r
~ .......... ~ , '"
a.
, ,­
.....
I
200 .
. ...... Pump speed = 200
-
100 -.
o
1
Pump speed = 300 ----.- Pump speed = 500 ;:'
0.5
-100
-iJ -
1.5
2
2.5
3
3.5
4
4.5
L ____________________~____________'
Particle Size (microns)
Graph 6: MeN at Ar = 140 kPa 1800 ~----------------------------------------------------------------------------------------,
1600
- ... - Pump speed = 500
- - .. - - Pump speed = 200
--m--- Pump Speed = 300 1400 1200
..... - -­
.... 1000
t:::
.....
::l
o
U
Q.)
--+
..K
800~-
."
u
t
C'il
a..
600
J
"
-.­
;'i,
400
I;
:-.:¥Jr-
200
01
1-":---'-= :--:,'- -~- ----- -I- - -- - - --' -1- - - - - - - - - -j;- -. - - - - - __ j;
­
--------::~
. ';;('
~~
0.5
1.5
2
3
2.5
3,5
4
4.5
-200 ~------------------------------------------------------------------------------------------~
Particle Size (microns)
!:II
the V -Groove and the Cross-flow. By using a Il-LPA which could go lower than the 0.3 11m
limit imposed by the current instrument, this premise could be fully explored.
Increasing the pump speed at lower pressures such as 100 kPa changes the position of the
initial peak. Higher flow rates cause greater particle diameters as can be seen most easily
on graph 5. However, at the higher pressures this effect is no longer evident. Graphs 4 and
6 illustrate the relationship between the Ar pressure and the flow rate. At lower pressures,
lower flow rates are required in order to maintain the initial peak while at higher pressures,
the higher flow rates generate larger initial peaks.
The MCN nebulization process is most efficient at approximately 100 kPa Ar for a pump
speed of 500 rpm (graph 7) while the Meinhard nebulizer appears to have its highest
efficiency at 200 kPa for the same pump speed (graph 4). The V-Groove and Cross-Flow
nebulizers were disappointing in this respect as they did not produce any large level of
particles.
When the % RSD is compared for all these nebulizers, the trend indicated from graphs 8-12
is that the MCN produces the most even spray while the Cross-flow has the largest standard
deviation in aerosol size. The Meinhard nebulizer produces in most cases a very constant
spray although its %RSD is slightly larger than the MCN. The inhomogeniety ofthe sprays,
especially for the MCN can be attributed to the pump. Indeed when the spray is examined
closely, it is possible to see it pulsating at the speed of the peristaltic pump. While this
phenomenon is more difficult to observe with the Meinhard nebulizer, the trend still exists.
It was not possible to determine this trend visually for the Cross-flow or V-Groove
nebulizers.
-55­
Graph 7: Meinhard at pump 500 RPM =
.
­
••
.'
,.
•
. .... -Ar
=200 kPa
- -m - Ar=140kPa
. ·x.
x"
/
/
x'
/
/
'.
0.3
0.35
0.4
m.
/!;;I--
/"...
JIl'
,/
••
.
':>:':. .
1:s-Ar = 100 kPa
.... x ... Ar = 60 kPa
,
--m-- •.. _
'
-II!t". • - .• " ...D- .... -
-m.. -- --
-l'J
·x ..
·····x·· '. ...• .. •x-. _.... - ·.x· ...... . · ·x
..
,/
,: Jf --­ ..•'
\
/
,/
0.45
'.\ .
0.5
/
0.6
0.8
"
/
~.'
1
1.5
Particle Size (microns)
2
2.5
3
3.5
4
4.5
Graph 8: MeN at Ar = 60 kPa 1600
1400
- .. - Pump speed = 200
f
~
Pump speed = 300
- - ... - - Pump speed = 500
1200
,.
1000
....c:
::l
0
\
800
()
(l)
Co)
'-2
C'il
a..
600
400
200
o
."- ..
.. ..................
------
1.5
.. ..................
35
2
-200
Particle Size (microns)
Graph 9: MeN at Ar 200 kPa =
2500r---------------------~-----------------------------------------------------
--m---- Pump speed = 300 -Pump speed = 500 - +- - Pump speed = 200
--* 2000 ",
,.
...--- --
",
---.
1500
.....
/"
c:
:::l
0
U
~
,.
/
/
1000--
II
u
.;'
tttl
.•... __ . ___ .. i··
/
a.
.;'
",
,. ..
500
.*­
'"
_ -~:_~ *' :~_ -~' -,..-_--~- _--_ -'
-00- '
oI
1*,114 . -.. - "
0.5
-500
1.5
2
2,5
3
3.5
4
4.5
~--------------------------------------------------------------------------------------------~
Particle Size (microns)
Graph 10: Meinhardt at Ar 200 kPa =
600~----------------------------------------------------------------------------------------.
500 -­
- .. - Pump speed = 200 --m-- Pump speed = 300 Pump speed = 500
--* --
_ -1----- ----}"---------I
400
t::J
I
300
o
u
"
"
..!!:!
~
0..
.'
,'..
u
't
.
,
200­
,
•
•
_
_o.~
_
• _ 3r:
x­
0
,
-
0
-
-
-
-
:K:_
x--
0' - '
'
-
--~
../
......... ",'"
.
..
100
04
T±±~~---4--------~--------+--------4---------r--------+--------+---------r------~
1.5
-100
2
2.5
3
3.5
4
4.5
-'----------------------~--------~
Particle Size (microns)
Graph 11: V-Groove at Ar = 200 kPa 400
=200
=300
- - ... - - Pump speed = 500
- -+- - Pump speed
I!l!
300 -
Pump speed
200
-
100
c:
::l
0
U
Q)
0
u
05
t(tI
1.l5
25
a..
-100
-200
-300
-400
Particle Size (microns)
35
4
45
Graph 12: Cross-Flow at Ar = 200 250~1------------------------------------------------------------~
- -+- - Pump speed = 200
200
~
--*
Pump speed = 300
- -Pump speed = 500
. L//--­
~ . "" 'I
150
" -",
... 100
:J
0
(J
50
u
t!1l
t
c..
oj----$, , s I Itl.
-50
.. . . .'
1.
.. '
I
.'
2
· .'.. ''.
·..'.t
.'
t:
~
-".
I
........... ..
..
1--.--.----'1--.--.-... -1
" "
2.5
3
3.5
4
4.5
-100
-150
~------------------------------~
Particle Size (microns)
Graph 13: Effect of Surfactant: MeN at Ar
=30 kPa 300 ,-------------------------------------------------------------------------------------------,
.. + .. Deionized Water
--m-- One drop Surfactant
2S0
- .. - Two drops Surfactant
E
- ... - Three Drops Surfactant
200
....
§ 1S0
o
'. ()
Q)
. ~
u
't
C\l
Cl..
100 -.­
~
... .
A~
so
t
r{,\ ---.
--_'. _- ".....~~~~
U"~~'--_
'*~~~~:~
n~" . " . .-~~.:---_:.~
~
'-<~F.,Z
t t"":,&
oI
It
O.S
1.S
2
2.S
-so
Particle Size (microns)
3
3.S
4
_ _-I• • _
4.S
~
__
Graph 14: Effect of Surfactant: MeN at Ar = 140 kPa 500 ,----------------------------------------------------------------------------------------------,
.-
- - .. - - Deionized Water
450
400
. - -- -..
-m- One Drop Surfactant
- ~ - Two Drops Surfactant
-)E- - Three Drops Surfactant
4- 1-
¥
;'
./ /
;'
.......
".
-
,
-
~
~
....
"\ ,
'l
/
0
~
f
•
,~
t
C'tl
,,
~ ....
f
250
••
'\
i-
U
f;
.....
..........
..... ~
I
~
c.. 200
... \,
/
::l
-
'\
/
-c: 300
-.
........... ~
/~
350
----
-.
-. -... - . -- - - - -.
---~--:&... ..........
-,A
150
..
100 -­
"
-.
50
-.- -)(.. -- -- -- -* -- -- -- -* -- -- -- -* --
-- -- -x
0
0
0.5
1.5
2
2.5
Particle Size (microns)
3
3.5
4
4.5
5
Graph 15: Effect of Surfactant: MeN at Ar = 200 kPa 2500
~-----------------------------------------------------------------------------------------.
.. + .. Deionized Water -----m- One Drop Surfactant ..
- .. - Two Drops Surfactant - .*" . Three Drops Surfactant
2000
-
.-
.......
,
c: •
1500
::l 0
U
a.l U
t
~
0..
1000
•
+
500
..
•
­
___ - -A---:-.- ...:._ ..... __ _
oI
o
~~;..:.~:,...--.~-.;.-.-.-.;.-.-.-.~.
4;'-
0.5
. - --
1.5
.~.
·tt·-·-·-·~·-·-·-~
- - ... - - - - - .... - - - - - .....
2
2.5
Particle Size (microns)
3
3.5
-----A------A
4
4.5
5
As expected, the use of a surfactant caused an increase in the smaller particles as can be seen
in graphs 13-15. At the lower pressures, for the MeN, it is possible to see a third peak
ending at our limit of detection, 0.3 )lm while the peak above 1 )lm decreases dramatically
in size. At 200 kPa, however, no peak is clearly in evidence over the entire range (there is
a remnant ofthe secondary peak). Presumably, at these levels, the solution is nebulized into
particles smaller than 0.3
)lm.
By varying the pressure of the argon gas while keeping the same solution flow rate, it is
possible to see that the initial peak shifts to the left, ie. towards a smaller droplet diameter.
The secondary peak does not seem to have any definite pattern relating it to the argon
pressure for either the MeN or Meinhard nebulizers. In the case of the MeN, the optimal
operating argon pressure for a pump speed of200 rpm seems to be 100 kPa. When the data
for the Meinhard is considered, this optimum shifts to a pressure of 200 kPa for a pump
speed of 300 rpm.
4.4 Discussion of limits of detection:
Although an attempt was made at determining the limits of detection for the MeN, the Iep
was recalled by SMM Instruments, from whom it was on loan, and this data could not be
confirmed. The preliminary tests did show that the MeN had a great impact in lowering the
background emission, by a factor of 5 in some cases. Unfortunately, this improvement was
countered by a similar decrease in the analyte's peak height. In the case ofa weakly emitting
analyte such as Sulphur, this meant that there was no apparent improvement on the signal
-65­
to noise ratio while, for arsenic the improvement could be seen. The major gain from the
MCN over the Meinhardt nebulizer was the level of solution required to run an analysis:
I ml as opposed to 5 ml in our case. This means that the MCN should come into its own
when only low volumes of solution are available, as is often the case in environmental
samples.
-66­
4.5 References
1. A. Montaser and D. W. Golightly, Inductively Coupled Plasmas in Analytical Atomic
Spectrometry, VCH Publishers, 1987, New-York.
2. R. N. Kniseley, H. Amenson,
c.c. Butler, and V.A. Fassel, App!. Spectrosc., 28,
1974,285-286.
3. K.W. Olson, W. 1. Haas Jr, and V. A. Fassel, Anal. Chern, 49,1977,632-637
4. S. Nukiyamaand Y. Tanasawa, Trans. Soc. Mech. Eng. (Japan), 4,5,6 (1938-1940),
E. Hope (transl.), Defence Research Board, Department ofNational Defence, Ottawa,
Canada, 1950.
-67­
Chapter 5 Extraction and speciation of Arsenic 5.1 Introduction:
As mentioned previously, there are several methods for the extraction and speciation of arsenic.
These would be too numerous to fully explore within this dissertation. Therefore, only one method
was followed. Most arsenic studies involving soil samples follow the hydride generation method.
In this case however, the method chosen was that of solvent speciation followed by sequential resin
speciation, which is usually used for marine samples. The reason for this was two fold. Primarily,
the availability of reagents and equipment and secondarily, the availability of expertise. Ion
exchange resins were readily available within the research group as well as suitable solvents for the
solvent extraction step. At the same time, a few people within the Department of Chemistry,
University of Pretoria, had already dealt with ion exchange, specifically with these resins which
gave us an advantage when using them. Hydride generation on the other hand had not yet been
performed at this university and would have required the purchasing or manufacture of specialized
glassware. Also, hydride generation is a less sensitive method and uses more reagents. A the same
time, other researchers tried to analyze the level of arsenic in soils by working out the total
concentration of arsenic by reducing As (V) to As (III) and subtracting from this the level of As
(III) found previously. This method, however, proved to be highly inefficient and unreliable.
-68­
5.2 Experimental 1: Ion exchange
Initially, 0.922g of Dowex lx8 anion exchanger was slurry loaded in a tube (blue/orange code).
This gave us a resin length of 15cm. This resin was conditioned as described below. Various
concentrations of As (V) were prepared from a 1000 mgIL stock solution. The resin was washed
prior to the run with 50 ml 1 M Hel followed by 50 ml de-ionized water. The "sample" was run
through and then stripped from the resin with 100 ml of 1 M Hel. The resin was then washed again
in preparation for the next run.
Results and discussion:
From graph 1, it is possible to see the effect of Hel on the Arsenic loaded resin. Although it might
be argued that 80 ml might have been sufficient to completely remove the arsenic from the resin, it
was decided to use a larger volume, partially for ease of preparation and also to ensure that all the
arsenic was removed. This is due to the fact that not all the resins which were prepared had the
same mass, a result ofthe difficulty ofieproducibly slurry-loading these resins. In effect, this larger
volume yielded a greater margin of error in the weighing of subsequent resins. Should a larger
amount of resin be used, there would still be enough Hel to strip it completely clean of arsenic.
The resin could hold approximately 13 mg of As, as can be seen from graph 2. As the research
progressed, it became apparent that a new resin would have to be used after about 50-60 runs as the
percent recovery of As dropped. This could be due to the acid used destroying the resin or, from the
resin binding sites becoming irreversibly bonded with As or other metals over time.
-69­
5.3 Experimental 2: solvent extraction
A stock solution of As (III) was prepared in I M HCl. Individual standards of As in 100 ml of 11
M HCI were prepared from this stock solution. These were then extracted with 4x 25 ml of an
organic solvent. These portions were combined and extracted with 2x 50 ml of de-ionized water.
Results and discussion:
The purpose of this section was to convert the As (III) in its inorganic form to AsCI) which would
favor the non-polar organic phase. As was expected the more non-polar solvents were better at
extracting the AsCI). From graph 3, we can see that there is little difference between benzene and
cyclohexane. However, hexane did not do as well. Toluene was also a pretty good extractor and has
the advantage of being less harmful than benzene. In the end, benzene was selected as it was the
purest, most readily available solvent in the stores at the time.
5.4 Experimental 3: soil samples, extraction of As(III) and As(V)
Soil samples weighing 5 g were spiked with both As (III) and As (V) and left to air-dry.
These
were subsequently milled and mixed with 100 ml ofHCl at different concentrations. The sample
was then made up to 11 M HCl and solvent extracted with 100 ml of benzene. The organic phase
was then back extracted with 100 ml de-ionized water. The aqueous phase was neutralized back to
a pH of 7 with NaOH and filtered with a Whatman 4 filter. It was then passed through the resin
using the method discussed above.
-70­
Results and discussion:
From these tests, it soon became apparent that the ideal conditions not involving a microwave
digestion step would use 11 M HCl. The soil was visually better digested and at the same time, the
As (III) was converted to AsCl 3 without the need for additional steps prior to the solvent extraction.
This meant less wet chemistry and therefore better chances at achieving accurate results. From the
table below, it is possible to see that the recovery rates were very good, almost all falling within a
range of95-100%.
TABLE 1: Recovery of As (III) and As (V) from spiked soil samples.
Sample No
1
2
3
4
5
Weight of soil (g)
5
5
5
5
5
Spike As (V) (g)
10.111
10.111
10.111
10.111
Recovery (g)
% Recovered
Spike As (Ill) (g)
9.604
94.99
9.809
97.01
9.780
96.73
10.111
9.675
95.69
10.001
10.001
10.001
10.001
Recovery (g)
10.036
10.001
9.847
9.778
9.933
9.843
% Recovered
100.35
98.46
97.77
99.32
98.42
9.966
98 .57
5.5 Experimental 4: soil samples, extraction of M1\1AA
Soil samples (5 g) were spiked with As (III) and As (V) as well as MMAA. These were acid
digested with 11 M HCI, filtered and then extracted with 100 ml Benzene. The organic layer was
back extracted with 100 ml de-ionized water while the aqueous layer was neutralized with NaOH,
filtered with a Whatrnan 4 filter and then passed through the anion exchange resin. The resin was
then first eluted with acetic acid / sodium acetate mixture (1: 1, 50 ml) followed by 100 ml I M
HCl.
-71­
Results and discussion:
.MMAA is readily stripped from the Dowex resin with an acetic Acid / sodium acetate mixture
while As (V) is not. Therefore, it is possible to elute them separately by choosing the acids
carefully. This method in effect allows us to speciate all the various type ofAs commonly found in
soils: As (III), As (V), MMAA and DMAA. DMAA has a more basic character and therefore will
be retained only by the cation exchange resin used in series with the anion exchanger. The eluent
tested at the end of this sequence showed no detectable trace of As, even after being evaporated
down to 10 ml (10 % of initial volume). The recovery rates of this experiment are shown below in
table 3. As DMAA could not be obtained commercially at the time, it was not possible to test the
accuracy of the recovery process.
Table 2: Recovery values of As for spiked soil samples.
Sample No
Weight of soil (g)
Spike As (V) (g)
Recovery (g)
% Recovered
Spike As (III) (g)
Recovery (g)
% Recovered
Spike MMAA (g)
Recovery (g)
% Recovered
1
5.003
10.009
9.644
96.35
9.998
9.374
93.76
1.006
1.003
99.70
2
5.122
10.009
9.770
97.61
9.998
9.640
96.42
l.006
.995
98.90
3
5.056
5.004
5.016
100.23
4.999
4.784
95.69
0.503
0.468
93.04
-72­
4
5.021
l.001
1.015
1Ol.42
0.999
0.984
98.50
0.503
0.504
100.20
5
5.035
6.005
5.983
99.63
5.998
5.857
97.65
0.503
0.751
99.47
5.6 Experimental 5: phosphoric rock and fertilizers
Phosphoric rock samples were obtained from Palaborwa, as well as Phosphoric Acid samples from
Indian Oceans Fertilizers. Fertilizer samples were also obtained from commercial sources as well as
from a farm. These were spiked with both As (III) and As (V) and then, were all acid digested in 11
M HCl, filtered and solvent extracted with benzene. The organic phase was then back extracted into
de-ionized water while the aqueous phase was neutralized with NaOH. The neutral aqueous phase
was then filtered twice, initially with a Whatman 42 filter and then with a Whatman 4 filter. The
filtrate was then passed through the anion exchange resin followed by the cation exchange resin.
The arsenic fractions were then released from the resin as described previously.
Results and discussion:
These samples caused quite a few problems. Primarily, phosphate rock can very efficiently be
digested by HCl. Indeed, the reaction tended to be violent should too much HCl be added at once.
After 2 hours, on average, most of the rock would be dissolved, leaving behind a tiny amount of
very fine powder residue which could not be weighed. Unfortunately, when these solutions were
neutralized, a thick white precipitate (presumably sodium hypophosphite) tended to be formed at a
pH of 6.5 - 6. This was a particularly troublesome event as the finer filter paper tended to get
clogged while the larger filter paper did not remove this precipitate sufficiently well. One method
by which this was solved was to do a double filtration.
Early ICP-AES tests of this precipitate, run by re-dissolving the precipitate in HCI showed that up
to 8 % of the As (V) could be found within it, while the precipitate appeared to be largely
composed of phosphorus. This indicates a strong possibility of a phosphate based salt. As it was
thought that this phenomenon was due to the close relationship between P and As, and, as
phosphate tends from an insoluble precipitate with OR, the neutralizing base was then changed to
ammonia. This caused a lower level of precipitation, as well as, only 3-5 % of the Arsenic (V)
being co-precipitated (table 3).
-73­
Using Celite as an alternative filter proved to be more effective, allowing for a "hot" filtration, as
vacuum filtration is much faster than gravitational filtration .. When the Celite bed was dissolved,
the level of As (Y) found dropped to 1-3.5 %. However, the Celite could only be used once due to
contamination from the co-precipitated As (Y). For this reason, as well as the cheaper cost and
greater availability of filter paper, this filtration method was abandoned.
Table 3: Results from the digestion of Phosphoric rock.
Sample
No:
Weight:
(g)
I
5.023
2
5.001
3
5.012
4
4.999
5
5.008
6
5.004
7
5.015
8
5.045
9
4.970
10
5.002
II
4.998
12
4.978
13
5.053
14
5.021
15
5.006
spike mglkg
As (III)
As (V)
10.009
10.023
10.009
10.023
10.009
10.023
10.009
10.023
10.009
10.023
9.899
10.023
9.899
10.023
9.899
10.023
9.899
10.023
9.993
10.012
9.993
10.012
9.993
10.012
9.993
10.012
9.993
10.0 12
9.993
JO.012
recovery
mglkg
As (III)
As (V)
9.951
9.543
9.837
9.653
9.976
9.756
9.776
9.231
9.863
9.495
10.043
9.879
9.843
9.609
9.558
9.767
9.735
9.746
9.966
9.556
9.249
9.657
9.721
9.787
9.922
9.549
10.146
9.614
9.885
9.638
-74­
% Recovery
As (III)
As (v)
99.42
95.21
98.28
96.31
99.67
97.34
97.67
92.11
98.54
94.73
101.45
98.56
99.43
95.87
96.56
97.45
98 .34
97.24
99.73
95.45
92.56
96.45
97.28
97.75
99.29
95.38
101.53
96.02
98.92
96.26
Filtration
method
% As in residue
Base
used
Filter paper
4.56
Filter paper
4.01
Filter paper
3.02
Filter paper
7.68
Filter paper
5.02
Celite
1.00
Celite
3.42
Celite
2.97
Celite
2.31
Celite
3.34
Filter paper
4.23
Filter paper
2.96
Filter paper
4.39
Filter paper
3.64
Filter paper
126
NaOH
NaOH
NaOH
NaOH
NaOH
NH4+
NH4+
NH4+
NH+
4
NH4+
NH4+
NH4+
NH4+
NH/
NH4+
Table 4: Results from fertiliser spikes.
spike mglkg
recovery, mglkg
% Recovery
No
As (III)
As (III)
As (III)
1
5.006
As (V)
9.997
As (V)
9.732
As (V)
97.35
5.009
10.003
9.997
9.568
9.795
95.65
97.98
5.011
10.003
9.997
9.727
9.926
97.24
99.29
4.899
10.003
9.997
9.759
9.972
97.56
99.75
10.003
9.997
9.856
9.854
98.53
98 .57
10.003
9.796
97.93
Sample
Weight: (g)
I
2
3
4
5
5.00
From Hydride Generation experiments conducted by the mining companies which provided us
with some phosphate rock samples, the levels of As (III) (as As 2 0 3 ) for two types of rocks, A
and B, are known to be l3 mg/Kg and 15 mg/Kg respectively. There was a good correlation
between these two techniques as the values obtained for these two rocks in this experiment
were 12.4 mglKg for rock A and 15.6 mglKg for rock B . Unfortunately, the levels of As (V)
were not determined by Hydride Generation and therefore cannot be compared.
-75­
5.7 Experimental 6: new method for analysing phosphorus bearing rocks
Phosphorus bearing rock samples were spiked with As (III), As (V) and MMAA, then they
were digested in 11 M HCl. These solutions were solvent extracted with benzene as described
above and then the aqueous phases were neutralized. The neutral solutions were then re­
acidified to I M acetic acid, using a sodium acetate buffer. The solutions were then passed
through the resin, and the arsenic was then released from the resin as described above.
Results and discussion:
Due to the difficulties in dealing with the white precipitate, a new method had to be developed.
The easiest was to reverse the order of resin speciation. In theory, only the DMAA should be
the only fonn of As to be retained by the cation exchange resin while the As (V) and MMAA
will be retained by the anion exchange resin. Therefore, if the solution were to be prepared
with acetic acid, under the right conditions, the resin should not retain the MMAA. As the As
(V) is more tightly bound to the resin and will be only displaced by the HCl, it stood to reason
that with this method, the As (III) would be found in the benzene, the As (V) would remain in
the resin while the MMAA would be in the eluent. This could have been further tested by
passing the solution through a cation exchange resin which would retain the DMAA, leaving
the MMAA in the solution. This method raised the overall recovery of As by 3% on average
which may not seem great though it did simplify the overall process by removing the hot
filtration stage when using Celite and the double filtration stage when using filter paper.
-76­
"/­
..
"
'
0
0
~
<,'
I
I
I
I
I
I
U
0::
~
0
LO
r-­ .....-..
.....J
~
~
~
~
E
'--'
'--"
en
C
...... --<
~
0
t
~
~
~
~
0
0
C
0..
Q)
.. >­
0
0
~
..=
~
~
~
0
()
Q) 0
H
LO ~
~
c
0
LO N
I
I
o
o
~
o
co
o(()
o
N
o
I
Graph 2: Comparison of As recovery
by two anion exchange resins
105 -­
.­c
C/1
<l)
I-<
>-.
I
I
T
I
100
--- 1.- - - -
95
-------'---- -
----+--I~m_+-·----- ----··
___l -
_
- - - -iI­ ___. _ ____._
I
I
1___
L
I
,.D
C/1
--< 90
I
~
0
- l----------T
.
C 85
<l)
>­
0
C,)
<l)
I-<
!
i
I
I
I
-l
~
---·T-·-----~
--:-- - --- ­ -- - 1 - ----1
I
80
!
- - -- !
-----~______1---_l
I
0~
I
75
70
.
I
- -'- - - - -- - -- + - - - ­
~
I
-1
0
I
i
._-,- -
- -- -+-- - -- - - ---'-- - - -- - - -- Li - - - -
10
20
30
40
mg As (V) used
--Y-
Amberlite lRA-400
--i::::::J--
DOWEX lx8
50
60
Graph 3: Comparison of % recovery
of As by different organic phases
100
''-~.~------------------------------------------~
/:,\
__ -0, '"
........... __
~,
L.);'~_. ,~
, .___
--
',--~. -0- - - _ -6- -
(:)-=-------..---=:::.;6 - -
95
~
-~
.. --
85
- f
/~---.--
80
--1----
75
-1 90
_
-
-0 --1''- - - --6- -=-=~--6:-=.-::::::::::.~
,
- ~.---
-- .- . . -----.---- -------- -----"
-, o(.) Cl) .....
~
---- ,
-tY-
­ --0
-~, /Q::'-~-=--;--.---
/
-
. -­
-
':o~"~-~~----------------------
•
=_:. .:_-----_·__·_--_ ·_--_·_-_.-/_·_---_·------1
•
4.5 ----)(
x
--- --~-------· . ----- ·-..--·----..- --
I
,
I
I
I
I
I
I
6.3
7.4
--. _. Cyclohexane
- 6, -
10
13.87 16.02
mg As (III) used
Benzene
- --D -
Hexane
20
30
50
----*- Toluene
·---I
Chapter 6 General Conclusions The driving force of this study was the need for a study of viable methods of analysing
arsenic in soils, fertilisers and phosphorus bearing rocks through the use of Inductively
Coupled Plasma - Optical Emission Spectroscopy (ICP-OES). This technique has come into
its own in recent years due to its speed, sensitivity and multi-element capabilities, as an
analytical method for the determination of many elements in a wide variety of matrices. Its
benefits have rendered it the mainstay of many analytical laboratories in the environmental,
health and general industrial areas.
As a counterpoint, most analytical techniques, especially spectroscopy related methods, still
can have difficulties with the accuracy of the measurements. These techniques need to be
both accurate and precise. While several procedures such as standard addition have been
developed to overcome the matrix effects, the application ofspectroscopic techniques can still
be complicated and there is certainly room for improvement both in accuracy and
standardization. To this end new methods of sample introduction are constantly developed.
In chapter 4, we looked at one of these improvements, the Micro-Concentric Nebulizer.
While this nebulizer does provide some definite advantages in decreasing the plasma loading
and generating a finer aerosol, some inherent drawbacks are inevitable. The gain in lowering
-80­
by the
background radiation is somewhat
benefit to
intensity. There is however a
10
nebulizer. Its smaller capillaries allow for
it ideal
solution volumes required
studies involving low
quantities. The
nebulizer's ideal operating conditions were explored, showing that
aerosol's physical
in a typical
thus
on the gas flow rates and
properties were indeed
flow rates.
The internal dimensions ofthe Micro-Concentric Nebulizer required some special glassware
solutions. This
to prevent blockages due to suspended particles in
was developed with ease of use
possible solution quantities in mind.
the glassware was optimized for low quantities,
dynamic range (up to 5 or 6 orders
111'''.'-''-'
for 0.5 ml to 8 ml.
magnitude)
high
as well as
powerful software and reproducible wavelength changes have made the ICP-OES
favoured technique for multi-element analysis. The technique
automation, making
technique has
l".UIJllvU
analysis of large
popularity
U"-"'uvv,
areas
lends itself well to
sample relatively easy. Thus
as environmental, hospital
laboratories.
ease of automation and popularity for environmental analyses prompted
of the study,
mainly
analysis of arsenic in
rocks and
second
Although
was
as a detector, with most ofthe wet chemistry having already been done, a similar
requiring only the overall levels of arsenic would
-81
simple to implement.
major
advantage of the ICP-OES in this study was the relative freedom it allowed from matrix
effects. This meant that the different extraction methods could more easily be compared. Both
the lines used were free from interference, though this was mainly due to the sample
preparation stages.
Ion exchange chromatography was shown to be an effective way of speciating the major
forms of arsenic in combination with solvent extraction. By reversing the order of ion
exchange, the method could be simplified with a slight gain in recovery levels. Different
filtration methods were shown to be effective. At the same time, the effect of the bases used
to neutralize the solutions was explored. These played a role in the precipitation of
phosphorus which hampered the analysis in some cases. By the simple act of changing the
base to one in which phosphate salts were more soluble, the level of precipitation dropped
visibly. At the same time, the level of Arsenic being co-precipitated dropped as seen in
chapter 5, table 3.
Different organic phases were shown to have different extraction capabilities towards As (III)
in the form ofAsCI) which was to be expected based on their polarity. On average, the more
polar the organic phase, the less well it extracted the AsCI).
After optimization ofthe system, standard addition experiments were done, yielding excellent
recovery rates. From the experiments carried out, a reasonable level ofprecision and accuracy
was possible with this system as can be seen from chapter 5, table 4. However, the sample
-82­
preparation steps tended to be tedious and often took up to five hours due to the filtration
steps and pumping solvents through the ion exchange tubing. However in cases where more
than one sample needs to be run, several samples could be run through the ion exchange step
simultaneously, byusing more advanced pumps, thus reducing the overall sample preparation
time.
In conclusion, it is clear that ICP-OES is a strong contender in the determination of arsenic
in the environment though some more work should be done in the speciation and extraction
of arsenic. Currently, hydride generation, with all its drawbacks may still be the most
efficient method to use. Unfortunately, the ICP was lost before such premises could be fully
explored.
-83­
Appendix A: Current page: Glassware developed to prevent MCN clogging. Next page: Liquid particle analyser setup. Glass double ended lid.
Fits into sample holder and
the glass'
Plastic container holding
0.45 ).lm whatman filter paper.
Syringe can be affixed to
other end to allow introduction
Offi~es of sample.
Glass inlet which fits
into sample holder.
Thin tube fits into the
glass inlet, thus enabling
the sampling of many
solutions by moving
the glass inlet from
sample holder to sample
holder. A glass cap has
been designed along
similar lines for the
sample holders not
currently used.
Sample holder
.W
SpecifiC'nolder for nebulizer
o
LPA
Universal holder
Nebulizer fits in specific holder
which then fits into universal holder
Wood support
Support
By turning the screw for positioning, the nebulizer
can be placed at a specific distance from the optical cell.

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