Direct analysis of volatile compounds during coffee and

Direct analysis of volatile compounds during coffee and
Direct analysis of volatile compounds during
coffee and tea brewing with Proton Transfer
Reaction Time of Flight Mass Spectrometry
Cumulative Dissertation
To obtain the academic degree
doctor rerum naturalium (Dr. rer. nat.)
of the Faculty of Mathematics and Natural Sciences
of the University of Rostock
by
José Antonio Sánchez López
Born on 08.02.1982
in Castellón de la Plana (Spain)
Rostock, September 2016
urn:nbn:de:gbv:28-diss2017-0030-7
1. Reviewer:
Prof. Dr. Ralf Zimmermann
Universität Rostock
2. Reviewer:
Prof. Dr. Elke Richling
Technische Universität Kaiserslautern
Date of submission: 31.08.16
Date of defense: 17.01.17
ii
ERKLÄRUNG
Ich versichere hiermit an Eides statt, dass ich die vorliegende Arbeit selbstständig angefertigt und ohne
fremde Hilfe verfasst habe. Dazu habe ich keine außer den von mir angegebenen Hilfsmitteln und
Quellen verwendet und die den benutzten Werken inhaltlich und wörtlich entnommenen Stellen habe ich
als solche kenntlich gemacht.
Die vorliegende Dissertation wurde bisher in gleicher oder ähnlicher Form keiner anderen
Prüfungsbehörde vorgelegt und auch nicht veröffentlicht.
Rotterdam, 20 August 2016
____________________
José Antonio Sánchez López
iii
iv
CONTRIBUTION TO THE MANUSCRIPTS THAT FORM THIS CUMULATIVE THESIS José A. Sánchez‐López has been author of the following manuscripts. His contribution to each one is described below. Insight into the Time‐Resolved Extraction of Aroma Compounds during Espresso Coffee Preparation: Online Monitoring by PTR‐TOF‐MS Analytical Chemistry, Volume 86, Issue 23, 2014, Pages 11696–11704. DOI: 10.1021/ac502992k José A. Sánchez‐López designed and performed all the experiments. He also performed the data analysis and prepared the manuscript. His work to this publication accounts for approximately 90%. Extraction kinetics of coffee aroma compounds using a semi‐automatic machine: On‐line analysis by PTR‐ToF‐MS International Journal of Mass Spectrometry, Volume 401, 2016, Pages 22–30. DOI: 10.1016/j.ijms.2016.02.015 José A. Sánchez‐López was involved in the design and execution of the experiments. He performed the data analysis and prepared the manuscript. His work to this publication accounts for approximately 80%. Rapid and direct volatile compound profiling of black and green teas (camellia sinensis) from different countries with PTR‐ToF‐MS Talanta, Volume 152, 2016, Pages 45–53. DOI: 10.1016/j.talanta.2016.01.050 José A. Sánchez‐López was involved in the design and execution of the experiments. He collaborated on the data analysis and the preparation of the manuscript. His work to this publication accounts for approximately 40%. Extraction Dynamics of Tea Volatile Compounds as a Function of Brewing Temperature, Leaf Size and Water Hardness: On‐Line Analysis by PTR‐ToF‐MS. Submited to Talanta on 06.04.2016 José A. Sánchez‐López was involved in the design and execution of the experiments. He performed data analysis and prepared the manuscript. His work to this publication accounts for approximately 70%. Rotterdam, 20.08.2016 ___________________________________________ José Antonio Sánchez López v
vi
This thesis is based on work done in the time between April 2013 and March 2016 at the Institute of
Chemistry and Biotechnology of the Zürcher Hochschule für Angewandte Wissenschaften and at the
Division of Analytical Chemistry of the University of Rostock under supervision of Prof. Dr. R.
Zimmermann.
vii
viii
Acknowledgements
First of all, I would like to thank Prof. Dr. Ralf Zimmermann for giving me the opportunity to conduct
my PhD program under his supervision.
I would also like to give special thanks to my thesis advisor Prof. Dr. Chahan Yeretzian, for his
unconditional trust, allowing me to fully manage this research.
I would like to thank my colleagues at the Analytical Chemistry department of ZHAW, especially to Dr.
Marco Wellinger and Dr. Samo Smrke for their support, sharing their knowledge and the fruitful
discussions in coffee science, analytical techniques and life in general.
I would like to thanks Dr. Franco Biasioli for giving me the opportunity of perform some of my research
in his group. Dr. Biasioli and his entire group make me feel like home. I would like to thank specially Dr.
Sine Yener for working hand in hand with me, and showing me the beauty of Trentino and the
Dolomites.
I would like to acknowledge the PIMMS (Proton Ionisation Molecular Mass Spectrometry) ITN
supported by the European Commission's 7th Framework Programme under Grant Agreement Number
287382. I would like to thank to Prof. Christopher Mayhew for all his effort starting this network and
keeping it together. I am thankful for knowing all the members in this network, all the supervisors and
ESRs.
Special thanks to Michael Reber, for listening to all my complaints and providing me with coffee and beer
when needed. Your friendship helped to survive to boring Wädenswil.
And my greatest thanks to Eva Tarrasón, for not killing me when I moved to Switzerland and support
me in all aspects of life. T’estim!
ix
x
Abstract
Coffee and tea are the two most consumed beverages in the world. Reasons for consuming them are
varied and include the energy boost from their caffeine content, multiple health benefits and, above all, a
pleasant sensory experience while drinking a cup of tea or coffee. Aroma is one of the main contributors
to the sensory perception, and therefore of utmost importance for both coffee and tea. This thesis work
focuses on the development of simple, fast, sensitive and reliable methods based on PTR-TOF-MS for
the analysis of volatile organic compounds and how they are extracted into the liquid phase during the
brewing of both coffee and tea.
In a first study, an on-line method to analyze extracted volatiles directly from the coffee flow was
developed and applied to the study of single dose coffee capsules. Volatile concentration could be
followed at 1 Hz resolution during the 42 seconds that the coffee extraction last. Differences in
extraction between compounds were revealed, implying an aroma profile that changed with extraction
time. Coffee capsules could be differentiated according to their extraction profiles by unsupervised
statistical methods (Principal component analysis, PCA; and Hierarchical Cluster Analysis, HCA).
A follow up study was performed using a semi-automatic coffee machine and varying the brewing
parameters (temperature and pressure). The different brewing conditions resulted in different timeintensity profiles that could be differentiated by PCA. Furthermore, all the compounds extracted were
grouped according to their extraction profiles into 5 families by HCA and Self Organizing Three
Algorithm (SOTA), with compounds in the same family sharing their physicochemical characteristics
(mainly water solubility and volatility). An increase in either brewing pressure or temperature resulted in
an increase in compound extraction. However, differences were only significant in the second part of the
extraction (after 10-15 seconds) and more pronounced in the less polar compounds.
In the case of tea, a preliminary study was performed in order to obtain the volatile profiles of a large
number of commercial teas. In that study, an automated headspace sampling method was used in
combination with PTR-TOF-MS to allow the screening of a large amount of samples in a short time. The
aroma profiles of 63 black teas and 38 green teas were analyzed for both full leaves and the infusion
obtained after brewing the leaves. Differences between leaves and infusions were found, indicating
incomplete extraction of some compounds (e.g. terpenes) and formation of others (e.g. alcohols) during
hot water extraction. Using multivariate analysis, black and green teas were successfully discriminated.
Also the origin of the samples could be partially discriminated although with some miss-assignments,
mainly between neighboring countries.
The extraction of volatiles from tea leaves and how it is affected by several parameters (leaf size,
temperature, brewing time and water composition) was further studied. Tea aliquots were taken every 30
seconds during five minutes and the headspace was analyzed with PTR-TOF-MS coupled to an
autosampler. An increase in brewing temperature resulted in increased volatile extraction, with
differences more pronounced at longer brewing times. Reduced leaf size resulted in faster extraction, a
difference that was more significant during the first minute of brewing. On the other hand, water
mineralization had low impact on the extraction kinetics and the volatile aroma content in the cup. Using
PCA and HCA, not only the impact of brewing parameters was assessed but also different sets of
brewing conditions resulting in analogous volatile profiles (i.e. same aroma profile) were identified.
xi
The positive outcomes of this thesis support the use of PTR-TOF-MS to follow the extraction of volatile
aroma compounds during preparation of hot beverages, both on-line and off-line. The use of multivariate
methods on the dynamic data increased the applicability of the methods allowing differentiation of
samples according to its composition (e.g. coffee capsules), country of origin (e.g. tea origins) or
parameters used for preparation (e.g. temperature-pressure combinations).
xii
Table of contents
Acknowledgements .............................................................................................................................................................ix Abstract .................................................................................................................................................................................xi Table of contents ...............................................................................................................................................................xiii 1. Coffee ...........................................................................................................................................................................1 1.1. Green Coffee. .....................................................................................................................................................1 1.2. Composition of green coffee beans ................................................................................................................2 1.3. Coffee roasting. ..................................................................................................................................................3 1.4. Coffee brewing. ..................................................................................................................................................5 2. Tea ................................................................................................................................................................................9 2.1. Economic importance .......................................................................................................................................9 2.2. Tea classification ................................................................................................................................................9 2.3. Tea Aroma........................................................................................................................................................ 11 2.4. Tea brewing. ..................................................................................................................................................... 13 3. Aroma........................................................................................................................................................................ 14 3.1. Aroma analysis. .................................................................................................................................................... 15 3.2. Isolation of the volatile fraction. .................................................................................................................. 15 3.2.1. Distillation and Solvent extraction. ......................................................................................................... 15 3.2.2. Static Headspace (SHS). ............................................................................................................................ 16 3.2.3. Dynamic Headspace (DHS)...................................................................................................................... 16 3.2.4. Solid-phase micro-extraction (SPME)..................................................................................................... 16 3.3. Analysis of the volatile fraction. ................................................................................................................... 17 3.3.1. Gas chromatography (GC) ....................................................................................................................... 17 3.3.2. Gas Chromatography Mass Spectrometry (GC-MS) ............................................................................ 18 3.3.3. Gas chromatography – olfactometry (GC-O) ....................................................................................... 18 3.3.4. Direct injection mass spectrometry (DIMS) .......................................................................................... 19 3.3.4.1. Photoionization .......................................................................................................................................... 19 3.3.4.2. Atmospheric Pressure Chemical Ionization (APCI) ............................................................................. 20 3.3.4.3. Selected Ion Flow Tube (SIFT) ............................................................................................................... 20 3.3.4.4. Proton transfer reaction (PTR). ............................................................................................................... 21 3.3.4.4.1. Components of PTR-MS instrument. ................................................................................................ 21 3.3.4.4.1.1. Ion source. .......................................................................................................................................... 21 3.3.4.4.1.2. Drift tube. ........................................................................................................................................... 22 3.3.4.4.1.3. Mass spectrometer. ........................................................................................................................... 23 xiii
3.3.4.4.2. 4. Reactions in PTR-MS ............................................................................................................................ 25 Applications of direct injection mass spectrometry in food science. .............................................................. 28 4.1. Food authenticity. ........................................................................................................................................... 28 4.2. Food quality ..................................................................................................................................................... 29 4.3. Flavor generation. ........................................................................................................................................... 31 4.4. Flavor release. .................................................................................................................................................. 32 4.4.1. 5. Flavor release during consumption. .................................................................................................... 32 Summary of results .................................................................................................................................................. 33 5.1. Extraction of volatiles during espresso coffee brewing (papers 1 and 2) .............................................. 33 5.2. Volatile profiling of tea leaves and their infusions. Application in authenticity. (paper 3) ................. 37 1.1. Extraction of volatiles during tea brewing (paper 4) ................................................................................. 38 6. References................................................................................................................................................................. 40 7. Annex ........................................................................................................................................................................ 52 I. List of figures........................................................................................................................................................ 52 II. List of abbreviations ....................................................................................................................................... 54 III. List of publications for thesis work.............................................................................................................. 55 IV. List of further publications ............................................................................................................................ 55 V. Publications ...................................................................................................................................................... 56 VI. Conference contributions ............................................................................................................................ 117 a. Oral contributions ............................................................................................................................................. 117 b. Poster contributions .......................................................................................................................................... 117 VII. VIII. xiv
Outreaching publications (non-scientific journals).................................................................................. 118 Publications previous to thesis work ..................................................................................................... 118 1. Coffee
Coffee is not only one of the most consumed beverages around the world, it is also one of the most
important commodities[1]. According to the International Coffee Organization (ICO), more than 140
million of green coffee bags (60kg/bag) were produced in 56 different countries during 2013[2]. Brazil
was the biggest producer with more than 30% of the worldwide coffee production. Although some
countries consume more coffee than they produce (e.g. Philippines) this is not a general feature. Most of
the coffee produced is exported. Taking a look at the two main coffee producers in 2013, Brazil and
Vietnam, they exported 60% and 90% of their total production respectively. Although coffee
consumption is higher in non-producer developed countries, not all imported coffee is meant to be
consumed in the country. Those countries roast the imported green coffee, adding value to the product,
and re-export it as roasted coffee. Therefore, coffee has a huge economic importance not only for the
coffee producing countries but worldwide.
1.1. Green Coffee.
The coffee plant, or coffee tree, is indigenous of Ethiopia and belongs to the Rubiaceae family, genus
Coffea. Despite the existence of more than 100 species of the Coffea genus, only two of them are of
commercial importance: Coffea Arabica, commonly known as Arabica coffee and accounting for 62% of
the world coffee production; and Coffea canephora var. robusta, known as Robusta coffee [3,4]. The coffee
plant is easily damaged by frost, does not tolerate temperatures higher than 30 °C and needs a minimum
of 1500 L/m2 of rainfall per year[4]. These requirements limit the regions where coffee can be cultivated.
Hence coffee is mainly produced in the area between the Tropics of Cancer and Capricorn, a region
commonly known in the coffee industry as “The bean belt”.
The coffee plant starts to produce white and fragrant flowers after 3-4 years, which develop into the
fruits (or berries). Coffee berries need 7-11 months of maturation from flowering, changing their color
from green to dark read with ripening. The coffee berry is formed by an exocarp (skin), mesocarp
(mucilage), endocarp (husk) and the seeds (coffee beans). Each coffee berry usually contains two coffee
beans, each one wrapped with a perispem that is commonly called silver skin.
Harvesting starts when most of the cherries are ripe and it is performed by selective handpicking of the
ripe cherries or non-selective stripping of all the cherries at the same time (either manually or
mechanically). Selective hand picking results in high quality but it needs extensive work force therefore
not being feasible in big plantations. There is a general believe that non-selective stripping results in lower
quality coffee but, with post-harvesting classification of the berries, top-quality coffee can be obtained
independently of the harvesting method[3].
Post-harvesting process to separate the beans from the other parts of the coffee berry is carried out at
origin by the dry or the wet processes[3,4]. The dry process is more economical as the cherries are sun
dried (or dried using mechanical driers) during several weeks. Then the dried husks are removed leaving
the bean with the silver skin attached. Wet process is more complex and requires high amounts of water,
but it generally provides better quality coffee. Pulp is removed in a pulper prior to fermentation in a
water tank, which enzymatically breaks down the mucilage surrounding the beans allowing its separation.
After washing away the mucilage, coffee beans are dried – either at sun or at mechanical driers.
1
1.2. Composition of green coffee beans
The chemical composition of green beans presents high variability as it is generally the case with
biological materials. Variations in green coffee composition have been attributed to both agricultural
factors linked to geographical origin (soil[5,6], altitude[7,8] or other environmental conditions[8,9]) and to
the post-harvesting process[9–11]. Table 1 provides a general overview on the non-volatile constituents
of green coffee beans. Comprehensive data about the composition of green beans can be found in the
reviews of Illy (1995)[3], Flament (2002)[12] and Clarke (2012)[4].
Table 1. Main non-volatile compounds reported in raw green beans.
NITROGENOUS COMPONENTS
Alkaloids
Proteins and free amino acids
caffeine
alanine*
cystine
leucine*
serine
liberine
b-alanine
glutamic acid
Lysine
taurine
methylliberine
g-aminobutyric acid
glutamine
methionine
threonine
arginine
glycine*
ornithine
tryptophan
tyrosine*
valine*
paraxanthine
theacrine
asparagine
histidine
phenylalanine*
theobromine
aspartic acid
1-(or 3-)methylhistidine
pipecolic acid
theophylline
cystathionine
hydroxyproline
proline
trigonelline
cysteine
isoleucine*
pyroglutamic acid 5-oxoproline
CARBOHYDRATES
Monosaccharides
Oligosaccharides
xylose (a-D-)
glucose (a-D-)
rhamnose
sucrose
arabinose (a-D-)
galactose (a-D-)
fructose (a-D-)
raffinose
ribose (a-D-)
mannose (a-D-)
psicose
stachyose
LIPIDS
Free Fatty Acids
Sterols
Diterpenes
linoleic acid
stearic acid
sitosterol
cafestol
16-O-methylcafestol
palmitic acid
arachidic acid
stigmasterol
kahweol
dehydrocafestol
oleic acid
linolenic acid
campesterol
cafestal
deydrokahweol
kahweal
CHLOROGENIC ACIDS
5-caffeoylquinic acid
5-feruloylquini acid
5-p-coumaroylquinic acid
3,5(4,5 and 2,4)-dicaffeoylquinic acid
Data compiled from Clifford and Willson (1985)[13] and Flament (2002)[12]. * indicates Strecker-active amino acids
2
In general terms, green beans have weights between 100 and 200 mg with water content from 9 to 13%
in weight. Different composition has been reported for Arabica and Robusta varieties. Arabica contains
more lipids (~15%), sucrose (6-9%) and trigonelline (0.6-1.3%) than Robusta (10, 3-7 and 0.3-0.9%
respectively). On the other side, Robusta contains more caffeine (2.2-2.4%) and chlorogenic acids
(~10%) than Arabica (1.2-1.3% and ~6% respectively). For both coffee varieties, the main constituents
are carbohydrates which account for 50% of the green bean’s dry weight.
Green coffee beans have a weak aroma that resembles that of peas or green bell peppers. Although the
volatile composition of green beans is less complex than that of roasted beans, around 300 volatile
compounds have been identified to date, of which 100 are unique to green beans. As green beans are not
used for consumption, most of the work done on volatile constituents had the objective of identifying
defective beans[14–16] or contaminants[17,18].
1.3. Coffee roasting.
Roasting is a crucial process that highly impacts the quality of the final coffee product. It basically
consists on heating the green coffee beans to obtain a desired taste and aroma, with a texture that allows
grinding and further extraction of the roasted coffee beans. The main aspects of bean roasting are
highlighted in Figure 1.
Figure 1. Roasting of coffee beans – main aspects[19,20]
The coffee roasting process can be basically divided in three different steps: (i) drying of the beans, (ii)
roasting phase and (iii) cooling of the beans. The drying step is an endothermic process and happens
when the coffee bean temperature is at around 100°C; free water evaporates and the bean reduces its
moisture from around 12% to a few percent. At this time the first changes in the beans are observed,
with a color change from green to pale yellow. The chemical reactions taking place at this stage are
endothermic but when the bean moisture is lower than 6%, the temperature keeps rising up to ~150-190
°C and the chemical reactions in the bean become exothermic (pyrolytic reactions). These reactions lead
to the production of large amounts of gas that increases the pressure inside the bean cell walls, resulting
in swelling of the beans and formation of coffee aroma. The high pressure inside the beans produces the
break of some cells generating a characteristic sound known as “first crack”. If beans are roasted for longer
time, the beans become dark roasted and start to release oils to the surface after a second small popping
(“second crack”). The final step of roasting is the cooling down of the beans in order to stop all the
reactions, which can be done either with air or water quenching[21].
The changes that coffee beans experiment during roasting can be divided into physical or chemical
transformations. The main physical changes are the increase in volume (up to 100%) and reduction of the
3
density from 550-700 g·L-1 in green beans to 300-450 g·L-1 in roasted beans[19]. Furthermore, the cellular
structure of the bean is damaged with formation of microporous and a fractured structure in the cell
walls, which allows the mass transfer of compounds from the inner cavities to the surface (e.g. oil) and it
accelerates degassing (loss of CO2 and volatile compounds)[22].
The chemical changes occurred during roasting are crucial, as they produce a large number of volatile
compounds that contribute to the final aroma of roasted coffee. More than 900 compounds have been
identified in roasted coffee to date[23]. The main reactions that lead to aroma formation are summarized
in Figure 2 and include: Maillard reaction, between reducing sugars and amino acids that leads to the
formation of volatile compounds and melanoidins (colored compounds); Strecker degradation, reaction
between and amino acid and a α-dicarbonyl to form aminoketones; trigonelline thermal degradation,
forming compounds with pyridine or pyrrole rings; phenolic acid degradation, to form phenol
derivatives; oxidative degradation of lipids, resulting in aldehydes, ketones and terpenes; breakdown of
sulfur amino acids, resulting in sulfur containing volatiles; and degradation of hydroxy amino acids that
mainly produce alkylpyrazines[4,21].
Figure 2. Main volatiles produced during roasting from non-volatile precursors. Adapted from Yeretzian et al 2002[21]
Despite the large number of volatile compounds found in coffee, only a small fraction is responsible for
the characteristic aroma of roasted coffee. The aroma composition of coffee depends on several factors
like origin, post-harvesting processes and roasting; therefore, the compounds reported as important
odorants in coffee are highly determined by the specific coffee analyzed and the methods used to
4
determine the contribution of each compound to the overall coffee aroma. Sunarharum and co-workers
recently reviewed the chemical components that are responsible of Arabica coffee flavor and ended up
with a short list of the main volatiles contributing to the aroma of Arabica coffee (Table2)[24].
Table 2. Compounds reported as important odorants of roasted Arabica coffee. Adapted from Sunarharum et al.
(2014)[24].
Aldehydes
2-methylbutanal
2-methylpropanal
3-methylbutanal
E-2-nonenal
Acetaldehyde
Propanal
p-anisaldehyde
phenylacetaldehyde
Acids
2-methylbutyric acid
3-methylbutyric acid
Esters
ethyl-2-methylbutyrate
ethyl-3-methylbutyrate
Furans
Furfural
2-((methylthio)methyl)furan
2-furanemethanol acetate
2-methylfuran
5-methyl-2-furancarboxaldehyde
furfurylmethyl ether
Fufurylformate
Furfuryldisulfide
Sulfur-containing compounds
dimethyl trisulfide
bis(2-methyl-3-furfuryl)disulphide
Methional
Thiols
3-mercapto-3-methylbutylformate
2-furfurylthiol
2-methyl-3-furanthiol
3-mercapto-3-methylbutylacetate
3-methyl-2-butene-1-thiol
methanethiol
Thiophenes
3-methylthiophene
Thiazoles
2,4-dimethyl-5-ethylthiazole
Furanones
dihydro-2-methyl-3(2H)-furanone
2-ethyl-4-hydroxy-5-methyl-3(2H)-furanone
3-Hydroxy-4,5-dimethyl-2(5H)-furanone
4-Hydroxy-2,5-dimethyl-3(2H)-furanone
5-Ethyl-3-hydroxy-4-methyl-2(5H)-furanone
5-ethyl-4-hydroxy-2-methyl-3(2H)-furanone
Ketones
1-octen-3-one
2,3-hexadione
2,3-butanedione
2,3-pentanedione
3,4-dimethylcyclopentenol-1-one
4-(4'-hydroxyphenyl)-2-butanone
1-(2-furanyl)-2-butanone
Phenolic compounds
Guaiacol
4-ethylguaiacol
4-vinylguaiacol
vanillin
Pyrazines
2,3-dimethylpyrazine
2,5-dimethylpyrazine
2,3-diethyl-5-methylpyrazine
2-ethenyl-3,5-dimethylpyrazine
2-ethenyl-3-ethyl-5-methylpyrazine
2-ethyl-3,5-dimethylpyrazine
2-ethyl-3,6-dimethyl-pyrazine
2-methoxy-3,5-dimethylpyrazine
2-methoxy-3,2-methylpropylpyrazine
2-methoxy-3-isopropylpyrazine
3-ethenyl-2-ethyl-5-methylpyrazine
3-isobutyl-2-methoxypyrazine
6,7-dihydro-5-methyl-5H-cyclopentapyrazine
ethylpyrazine
Pyridines
pyridine
pyrrole
1-methyl pyrrole
Terpenes
linalool
limonene
geraniol
Norisoprenoids
E-β-damascenone
1.4. Coffee brewing.
Brewing is the last step of coffee preparation and it is crucial to obtain the desired quality in the final
beverage. It basically consists the percolation of hot water through a bed of ground roasted coffee
particles. Water not only dissolves the soluble components of coffee but also sweeps along solids and
oils present in the surface of the ground coffee particles, which will impact the final organoleptic
properties of the coffee brew.
Brewing methods can be mainly classified in: (i) decoction methods, in which the hot water is kept in
contact with the ground coffee for a considerable amount of time (e.g. boiled coffee, Turkish coffee,
Vacuum coffee, French press); (ii) infusion methods, where water flows through the ground coffee bed
5
with shorter contact times (e.g. filter coffee, Napoletana coffee); and (iii) pressure methods, where a
relative high pressure is used to create a flow through the compacted coffee bed (e.g., Moka,
espresso)[25]. The method used for coffee brewing will determine the final characteristics of the cup
including caffeine content[26–29], antioxidants[27–29], lipids[30] or volatile composition[28,29,31]. The
only coffee brewing method in the scope of this thesis is espresso brewing; thus, it will be described in
detail.
Espresso coffee can be defined as a beverage prepared on request by percolation of hot water (90 ± 5
°C) under pressure (9 ± 2 bar) through a compacted cake of roast and ground coffee beans (6 ± 1.5 g)
for a short defined time (30 ± 5 s)[25]. The extraction conditions used for brewing espresso are
responsible for the characteristics of the final beverage that make espresso unique and different from
other coffee beverages. As defined by Petracco, “espresso is a polyphasic beverage, prepared from roast and ground
coffee and water alone. It is constituted by a foam layer of small bubbles with a tiger-tail pattern, on top of an emulsion of
microscopic oil droplets in an aqueous solution of sugars, acids, protein-like material and caffeine, with dispersed gas bubbles
and colloidal solids.”[3]
The methods to brew an espresso can be grouped in three main categories: use of a bar machine, coffee
pods and single dose coffee capsules. The last two are responsible of the increase in espresso coffee
domestic consumption. Both methods allow preparing an espresso with high reproducibility just by
pressing a button, in a clean and easy way. As a drawback, the selection of coffees is limited and the
coffee machines designed for both pods and capsules do not usually allow the user to adjust the water
temperature or pressure for extraction.
The first step in coffee brewing is grinding of the roasted coffee bean. Grinding is generally considered
an independent step from brewing but, the requirement of freshly ground beans in order to get a top
quality espresso together with the high impact that grind size has in regulating the flow of water through
the coffee cake, sustain the need of considering both processes –grinding and brewing- together.
Grinding of the roasted coffee beans produces not only physical but also chemical changes that impact
the final beverage. The most obvious effect is the fragmentation of the bean in small particles (0.2-650
μm) increasing the effective surface area for extraction and destroying the cellular structure, therefore
liberating most of the trapped CO2[32]. The finer the grind, the highest the loss of CO2[33] and
consequently of volatile aroma compounds. Particles bigger than 50 μm can still contain intact cell
cavities filled with the gases generated during roasting which are responsible for the formation of
“crema” (coffee foam)[22,32]. Another physical effect of grinding is the release of coffee oil and its
deposition on the coffee particles surface. Coffee oil is extracted and found as emulsified oil droplets in
the final beverage, contributing to the mouthfeel and aroma lingering typical from espresso
beverages[34]. The increased surface exposed to oxygen upon grinding, also results in an increased
oxidation rate of coffee lipids and aroma compounds, negatively affecting the final flavor of the beverage.
Therefore, it is crucial to brew the espresso immediately after grinding the roasted coffee beans. In the
case of pods and capsules systems, the manufacturer has to provide a packaging that ensures the
freshness of the product until its consumption.
In the case of semi-automatic bar machines, the roast and ground coffee is dosed into a porta-filter
holder and tamped. Tamping consists of the application of pressure on the ground coffee in order to
eliminate the void spaces between coffee particles. The dose, particle size of the ground coffee and the
compacting of the bed determine the hydraulic resistance of the system, therefore controlling the flow of
6
water through the coffee cake[19]. Different flows imply different contact times between the coffee
particles and water, thus affecting the extraction yield of the different coffee compounds. A non-uniform
coffee bed favors the creation of water channels of high flow, reducing the extraction efficiency.
Two main mechanisms can be distinguished in the extraction of soluble substances from the coffee bed
into water: washing out of the solutes from the surface of the coffee particles and diffusion from inside
the particles to the solution. In the case of espresso extraction, where the contact time between the hot
water and the coffee particles is short, washing out can be considered the main effect[19,35] but diffusion
still plays an important role[36]. As different compounds have different water solubility and diffusion
coefficients, their extraction rates differ, varying the composition of the cup of coffee over the extraction
time[37]. Parameters that affect the water/coffee contact time, the solubility of compounds in water or
their diffusion will also affect the extraction efficiency.
Several authors have studied the effect of different extraction parameters on the composition and quality
of the final espresso beverage. The particle size of ground coffee showed an inverse correlation with the
extraction of solids, soluble compounds, caffeine and lipids[38,39]. Finer grounds resulted in higher
compaction of the coffee bed and therefore higher water/coffee contact times, consequently increasing
the extraction yields. Regarding the organoleptic characteristics of the espresso coffee, higher levels of
volatile compounds could be determined in espresso coffees prepared from fine grounds, with roasty,
fermented and woody notes; while coarse particles produced coffees that were described as burnt or
rubbery by the sensory panelists[38]. Closely related to the coffee dose is the coffee/water ratio, in other
words, the amount of water used to extract the ground coffee. Coffees prepared with high coffee/water
ratios had higher density, viscosity, surface tension and total solids content[40]. Total lipids and diterpene
content also showed a positive correlation with coffee/water ratio[39]. Additionally, higher coffee/water
ratio resulted in higher caffeine and chlorogenic acids extraction which gave bitterness and astringency to
the beverage. Although the volatile composition was not significantly affected by the coffee/water ratio,
samples with higher ratio were described as less fresh and presented a burnt, acrid and fermented flavor.
Water is the second main component of coffee. Three parameters attributed to water will affect the
extraction of espresso coffee: water composition, pressure and temperature. The water used to brew
coffee needs to be free of any impurity that can impact the organoleptic properties of the final cup[41].
Water is generally softened in order to reduce calcium and magnesium content as these cations produce
insoluble salts that precipitate, affecting the heat transfer efficiency and damaging the coffee
machines[19]. The most common way of water softening is the exchange of Ca2+ and Mg2+ ions with
sodium ions, although this replacement by Na+ was found to dramatically affect the extraction time,
increasing it up to 80%[42]. Furthermore, the binding energy of Na+ with some compounds (e.g. caffeine
or eugenol) is higher than that of water according to quantum-chemical calculations, therefore sodium
ions negatively affect the extraction of those compounds[43]. Another water component that influences
both extraction and final characteristics of coffee is the hydrogen carbonate ion. Both carbonate CO32and hydrogen carbonate HCO3- ions can be neutralized by coffee acids and produce CO2. The dissolved
carbon dioxide produced from hydrogen carbonate affects the compaction of the coffee bed, increasing
the extraction time and affecting the coffee crema. Higher CO2 content produces higher foam volume in
the cup but the foam has an undesired texture when too much CO2 is released, with big bubbles that
collapse quickly, negatively affecting foam persistence[32,44].
7
Water temperature and pressure can be easily adjusted by the barista in semi-automatic bar machines in
order to optimize coffee extraction. Increase of water temperature results in an increased extraction yield
and higher foam volume and persistence. Coffees extracted at high temperatures also show larger
amounts of caffeine, lipids and volatile compounds that impact the final flavor of the beverage[39,45,46].
Significantly higher values for sensorial descriptors have been found between coffees extracted at 90 or
100°C, including color intensity, body, texture, olfactive intensity, bitterness or astringency amongst other
attributes[46]. Extraction efficiency also shows positive correlation with pressure. Increase in total solids,
lipids or chlorogenic acids has been reported between 7 and 9 bar[47], and higher diterpenes were
extracted at 11 bar[39]; however, other compounds as caffeine were not significantly affected by pressure.
Decrease of diterpenes, total lipids and chlorogenic acids with pressures higher than 11 bar have also
been reported and attributed to a differences in compaction of the coffee bed that resulted in a reduced
flow[47]. The higher extractability of compounds with high water temperature and pressure has been
related with two different processes. The first one is the change in polarity of water with temperature and
pressure that allows medium and low polarity compounds to dissolve in water and it improves mass
transfer from the beans to water. The second effect is the higher penetrability of water into the coffee
particles due to pressure that can assist the extraction of compounds that remain inside the ground coffee
pores[48].
The composition of espresso coffee changes with time. Despite this widely known fact, most research
has been focused on the final cup and only few studies have investigated the extraction kinetics for
espresso coffee. Kinetics of extraction have been determined for compounds with health related
properties as caffeine, chlorogenic acids or acrylamide. Despite the high water solubility of these
compounds, only 70-80% of the total content in the beans is extracted during the short time in which
espresso is prepared (~25s)[19,25,49]. Figure 3 summarizes the results from Ludwig and co-workers
studying the extraction of caffeine and chlorogenic acids during espresso brewing (45 mL in 24 seconds).
The first fraction corresponds to the first 8 seconds (~17mL) and it contained around 70% of the total
amount extracted for monocaffeoylquinic acids and caffeine, and half of the dicaffeoylquinic acids[27].
After those 8 seconds, extraction was much slower for all the compounds. The same behavior has been
observed in other studies with caffeine[50], chlorogenic acids and melanoidins[51]; with the accumulate
concentration during extraction following a 2-parameter hyperbolic curve of formula y = a·x/(b+x).
Kinetics of 20 volatile aroma compounds have also been studied for a single dose coffee capsule
system[52]. By extracting capsules to different volumes (10 to 150 mL) and the subsequent analysis of the
coffee headspace, Mestdagh and co-workers found the same hyperbolic curves and an inverse correlation
between the slope at the beginning of the curve and the polarity of the compounds. The more polar the
compound the faster it was extracted from ground coffee[52]. Figure 3 shows the concentration of three
volatile compounds at different extraction volumes (from 0 to 160 mL). The different slopes at the
beginning of the curve for each compound reflect different extraction rates. Therefore, the volatile
profile in the final beverage is highly dependent on the volume (and therefore time) in which the coffee is
extracted, highlighting the importance of time resolved data in order to study coffee the brewing process.
8
Figure 3. Dynamics of espresso brewing. Left: Extraction of non-volatile compounds. Fractions were taken every 8
seconds and had volumes between 14-17mL and data is presented as % of the total amount extracted [27]. Right:
Extraction of volatile compounds. Coffees were brewed to different final volumes and the headspace concentration
determined by GC-MS[52].
In order to better describe the percolation process and the extraction of soluble compounds, Bandini and
co-workers created a computational model to simulate the motion of water across a compacted coffee
bed and the extraction of soluble compounds from the coffee particles. The results of the simulation are
shown in Figure 4. In the case of water, the number of elements leaving the system follows an initial
growth that stabilizes with time. For soluble compounds, extraction is more efficient at the beginning and
it slowly decreases with time. These results are in agreement with experimental results[53].
Figure 4. Water (left) and soluble compounds (right) exiting the ground coffee bed system in a computational model for
espresso brewing.[53]
2. Tea
2.1. Economic importance
The Food and Agriculture Organization (FAO) of the United Nations (UN) recognizes 50 tea producing
countries in the world, although the first ten (China, India, Kenya, Sri Lanka, Vietnam, Turkey, Iran,
Indonesia, Argentina and Japan) accounted for more than 90% of the total production in 2013. Tea
production has been almost doubled during the last 20 years. That increase has been attributed to an
increased consumption in the producing countries and to an increment in the exports to other countries,
mainly the United States of America, Russian Federation and United Kingdom. According to FAO data,
2 million tons of tea were exported on 2013, accounting for 5.7 billion USD (United States Dollars)[54].
2.2. Tea classification
The tea plant presents high heterogeneity on its vegetative (somatic) characteristics which has made
difficult its taxonomical classification. Two major varieties with economic importance are cultivated:
Camellia sinensis var. sinensis, which represents the majority of green teas from China and Japan, and
9
Camellia sinensis var. assamica, from which black tea is produced[55]. Owing to that heterogeneity, together
with the large differences between producing countries in terms of climate (temperature, day length,
humidity), soil, and growing, harvesting and processing practices, tea is normally commercialized with
the name of the production region (often including also the name of the plantation and the harvesting
season).
Figure 5. Tea producing countries color-coded according to 2013 FAOSTAT data.
Tea is mainly classified according to its fermentation degree: green (non-fermented), white (lightly
fermented), oolong (semi-fermented) and black tea (fermented). The tea manufacturing process is highly
variable, with differences not only between regions but within producers. Most times, producers have
specific steps in the manufacturing processes that are kept secret, as it represents the main differentiator
from competition. The main steps in tea manufacturing are: harvesting (plucking), fixing (enzyme
inactivation), withering, rolling and drying.
It all begins with the plucking of the two youngest leaves and a bud from the tea plant. For high quality
tea, plucking is done manually to avoid removal of stalks. The so-called fermentation is in fact a
combination of enzymatic reactions induced by endogenous oxidative enzymes (e.g. polyphenol oxidase
or peroxidases). Therefore, in the case of the non-fermented green teas, enzymes need to be deactivated
after harvesting and the tea leaves are then subjected to a thermal treatment, being pan firing and
steaming the most commonly used. The thermal treatment used for fixing will impact the chemical
composition and organoleptic properties of the final tea[56]. After fixing, and while still hot, the leaves
are rolled in order to break the leaf cells and liberate the inner content, developing the tea aroma. To
keep the quality of the tea until brewing, leaves are finally dried. For those teas in which fermentation is
required, leaves are withered after plucking. This process can last for few hours to several days, either at
ambient conditions in the sun or indoors in heated rooms. In the case of black tea, withering is
performed until the water content of the leaf is around 60%. Then, leaves are rolled until 80-90% of the
leaf cells are broken and they are kept at controlled conditions of temperature and humidity so the
enzymatic reactions can take place. The leaves turn darker and the aroma evolves from green to fruity as
the chemical reactions evolve. Once the fermentation is finished, the tea leaves are dried[56,57].
10
2.3. Tea Aroma
Tea quality is assessed by the color and aroma of the dry leaves together with the aroma and taste of the
infusion obtained after brewing the tea leaves with hot water. Therefore, the volatile content of teas is of
utmost importance. More than 600 volatiles have been reported in tea, and their concentrations in tea
leaves depend on the tea cultivar, environmental conditions and leaf processing. Extensive research has
been performed on the volatile content of tea since the early 90’s and the current knowledge of tea flavor
has already been reviewed[57–59].
The main pathways leading to volatile compounds in tea are depicted in figure 6. Saturated and
unsaturated C6 and C9 aldehydes and alcohols are formed from enzymatic oxidation of fatty acids,
mainly by lipoxygenase. Therefore, any change in the enzymatic activity (e.g. seasonal variations) also
alters the concentration of these compounds which are accountable for the green notes of tea aroma.
Methyl jasmonate also derives from fatty acids, although its biosynthetic pathway has not been yet
elucidated in tea plants[58]. The main volatile terpenes found in tea are geraniol, linalool and linalool
oxides. They are key compounds in the aroma of black tea, contributing to the floral aroma[60]. Other
compounds contributing to the floral and fruity aroma of tea are those derived from L-phenylalanine:
benzaldehyde, benzylalcohol, phenylethanol, phenylacethaldeyde and coumarin, the last one with a
pleasant sweet odour.
Carotenoids are also important precursors of aroma volatiles in tea. The total carotenoid content in the
tea leaves is variable but it can be up to 10% in dry weight[61]. Enzymatic and thermal degradation of
carotenoids lead to formation of important aroma compounds as ionones and damascenones[62]. Finally,
there are a large number of volatile compounds present in glycosidically-bound forms, which increases
their solubility but reduces their volatility. Enzymatic hydrolysis releases some of those volatiles during
tea processing[63], while others can be hydrolyzed during the hot water extraction of the tea leaves[60].
Figure 6. Simplified proposed biosynthetic pathways of tea volatiles. From Yang et al. 2013[58].
11
Most of the research on tea volatiles has been focused on the identification of new compounds or the
description of the volatile composition of a specific tea. Wang and co-workers determined the volatile
composition of 87 different teas (8 white, 27 green, 27 oolong, 15 black and 10 Pu-erh)[64]. Although the
volatiles analyzed were similar for all teas, their concentration varied largely. All tea types had linalool and
linalool oxides among the five compounds in highest concentration. Geraniol was also among the most
abundant compounds for all the teas with exception of pu-erh teas. As not all volatiles are relevant to the
aroma of tea, some authors have identified the key aroma compounds in tea by using flavor dilution
technique in combination with GC-Olfactometry[65,66] (Table 3).
Differences in the compounds contributing to aroma have been found not only between different tea
types but also between the tea leaves and their infusions. Most of the differences found were attributed to
the lower concentration of compounds in the beverage compared to the leaves. Interestingly, some
alcoholic volatiles like geraniol were found in higher concentration in the hot water infusions than in the
leaves, probably due to thermal hydrolysis of their glycosides[60].
Table 3. Potent odorants of green and black teas identified by flavor dilution.
Compound
(E)-2-nonenal1,2
(E)-methyl jasmonate2
(E,E)-2,4-decadienal1,2
(E,E)-2,4-heptadienal1,2
(E,E)-2,4-nonadienal1,2
(E,E)-2,4-octadienal1
(E,E,Z)-2,4,6-nonatrienal1
(E,Z)-2,6-nonadienal1,2
(E,Z)-2,6-nonadienol1
(Z)-1,5-ocatadien-3-one1,2
(Z)-3-hexenol1,2
(Z)-3-hexenal1
(Z)-4-heptenal1,2
(Z)-jasmone2
(Z)-methyl jasmonate2
1-octen-3-one1,2
2,3,5-trimethylpyrazine2
2,3-butanedione1,2
2,3-diethyl-5-methylpyrazine2
2,3-pentanedione2
2-acetyl-1-pyrroline2
2-acetyl-2-thiazoline2
2-acetylpyrazine2
2-aminoacetophenone2
2-ethyl-3,5-dimethylpyrazine2
2-ethyl-3,6-dimethylpyrazine2
2-isobutyl-3-methoxypyrazine2
2-methoxyphenol1
2-methoxy-4-vinylphenol2
2-phenylethanol1
3-ethylphenol1
3-hydroxy-4,5-dimethyl-2(5H)-furanone1,2
3-methylbutanal1,2
3-methylbutanoic acid1
1
Aroma descriptor
green
floral
fatty
fatty
fatty
fatty
oat flake like
cucumber like
cucumber like
metallic
green
green
hay like
green
floral
mushroom like
nutty
buttery
nutty
buttery
popcorn-like
popcorn-like
roasty
grape like
nutty
nutty
earthy, musty
smoky
spicy
honey like
phenolic
seasoning
malty
sweaty
Compound
3-methylnonane-2,4-dione1,2
4,5-dihydro-3(2H)-thiophene2
4,5-epoxy-(E)-2-decenal1
4-hydroxy-2,5-dimethyl-3(2H)-furanone1,2
4-mercapto-4-methyl-2-pentanone2
4-nonanolide2
4-octanolide2
5-octanolide2
bis(2-methy-3-furyl) disulfide1
butanoic acid1
coumarin2
ethyl-2-methylbutanoate1
eugenol2
geraniol1,2
geranylacetone2
guaiacol2
hexanal1
hexanoic acid1
indole2
jasmine lactone2
linalool1,2
methional1,2
methyl anthranilate1
nonanal2
octanal1,2
p-cresol2
pentanoic acid1
phenyl acetic acid1
phenylacetaldehyde1,2
trans-4,5-epoxy-(E)-2-decenal1
vanilin1,2
β-ionone1
β-damascenone1,2
γ-nonalactone1
Volatiles identified in Darjeeling Black tea by Schuh et al.(2006) [60]
Volatiles identified in Sen-cha, Kamari-cha and Longjing green teas by Kumazawa et al. (2002)[72]
2
12
Aroma descriptor
green
roasty
metallic
caramel like
meaty
sweet
sweet
sweet
meaty
sweaty
sweet
fruity
spicy
floral
floral
burnt
grassy, green
sweaty
animal like
sweet
floral
potato like
grape like
orange like
orange like
phenolic
sweaty
honey like
honey like
metallic
vanilla like
violet like
honey like
coconut like
Volatile compounds have also been used to discriminate among various teas. Using only five volatiles
(trans-2-hexenal, benzaldehyde, methyl-5-hepten-2-one, methyl salicylate and indole), it was possible to
separate non-fermented teas green teas (34 samples) from fermented ones (22 samples)[67]. In another
study, seventy-five oolong teas from five different varieties were also discriminated by their volatile
content[68]. Similar approaches based in either GC-MS or eNoses have been successfully applied for the
discrimination of teas according to their quality[69–71].
Although with some discrepancy in the discrimination efficiency, attempts of discriminating tea origin by
their volatile content have also been reported. Togari and co-workers applied pattern recognition
techniques to volatiles extracted from 44 different samples and analyzed by GC-MS. They could classify
black teas according to 3 origins, but classification was not possible for green or oolong teas[73]. Other
study reported successful classification of green teas from two different regions in China also based on
GC-MS data[74]. Baldermann and coworkers analyzed 38 samples belonging to 5 different origins. They
could discriminate teas according to fermentation (green, oolong and black teas) by cluster analysis.
Although most samples from same origin were clustered together, some samples were misclassified and
was not possible to relate origin with volatile profile[75]. Lee et al. neither found any relationship between
the country of origin and the volatiles measured in the headspace of 24 different green teas by
SPME/GC-MS[76]. All these difficulties in discriminating tea samples points to the complexity and
variability of their volatile composition.
2.4. Tea brewing.
Tea brewing consists of the immersion of tea leaves in hot water during a defined period of time.
Although this process might seem simple a priori, the optimal brewing conditions depend on the kind of
tea, culture, social environment and individual preferences. Therefore tea brewing can go from a
ceremonial preparation like the Japanese tea ceremony to pouring boiling water in a mug containing a tea
bag in UK, to exemplify two extreme situations. The main factors affecting the quality of the final tea
infusion are the water/tea ratio, leaf shape and size, water composition, water temperature and brewing
time.
As it happens in the case of coffee, work studying the kinetics of extraction of tea compounds has been
focused on caffeine and other health related compounds, mainly polyphenols. Detailed studies on the
kinetics of caffeine and theoflavins extraction have been performed by several authors. Spiro and Jago
stated that, according to the two phase and steady-state theories, the extraction kinetics for soluble
compounds in tea leaves should obey first-order kinetics (Equation 1)[77].
ln (𝐶
𝐶∞
∞ −𝐶
) = 𝐾𝑜𝑏𝑠 𝑡
Eq. 1
Where C is the concentration of the compound at time t, C∞ is the concentration at equilibrium and Kobs
is the overall rate constant.
In a series of papers, Spiro, Price and other authors have used equation 1 to fit experimental data and
describe how the kinetics and equilibria of tea infusions are affected by the leaf size[78], water
composition[79–81] or extraction temperature[82–84] amongst other factors[85–88]. Their results suggest
that the rate determining step for extraction of tea components into water is the diffusion of the
component through the leaf matrix to the surface. Therefore, the extraction rate will depend on
13
compound structure, mass, solubility, water temperature and the concentration gradient between the tea
leaf and the solution.
Leaf size had a minimum effect on the equilibrium concentrations of caffeine and theaflavin[89] but it
affected their extraction kinetics[78]. Smaller particle size resulted in larger rate constants. For full leaves,
extraction takes part mainly through the two large surfaces of the leaf, but the effect of the edges
becomes important when the particle size is reduced, increasing the surface area of the leaf, and
enhancing the extraction rate. In the case of fine ground tea leaves, caffeine was found to be extracted
fast from leaves into the infusion, with no significant differences between 0.5 to 30 minutes infusion time
[90]. The enhanced extraction of soluble compounds from fine ground leaves resulted in increased
antioxidant activity and higher concentrations of polyphenols and flavonoids in the final beverage.
However, as it often happens with flavor, those infusions were found extremely bitter and astringent,
therefore being less pleasant than infusions prepared from full leaves[91].
The dissolved mineral content of water also impacted the extraction kinetics of caffeine[80] and
theflavins[81] but had no influence on the equilibrium concentration of those compounds[79]. In the
case of caffeine, minerals can modify its solubility affecting extraction kinetics. A decrease on caffeine
extraction rate was found with NaCl, NaOAc, KCl or CaCl2 that correlates with a decrease in caffeine
solubility in the presence of those minerals. On the other hand, higher extraction rate is obtained with
Bu4N+ ions which increase caffeine solubility[80,92]. In the case of theaflavins, addition of salts slightly
increased extraction rate with no significant differences between the different salts, with the exception of
CaCl2 which highly decreased the extraction rate. This pronounced decrease in extraction with CaCl2
have also been observed for caffeine, polyphenols and the total organic carbon content of tea infusions,
and it has been related to the formation of Ca-pectin complexes. These complexes modify the cell wall
properties, affecting the diffusion of compounds, hence hampering extraction[93].
In the case of volatile compounds, research has focused on the identification of the volatiles present in
the final infusion (after brewing) and not on the extraction process itself. To the best of our knowledge,
no studies have been reported about the kinetics of volatile compounds extraction from tea leaves into
the infusion and how this extraction is affected by different brewing parameters (e.g. temperature, leaf
size or water composition).
3. Aroma.
While the biological purpose of eating food is nutrition, it also has a very important hedonic element. The
pleasure associated with food consumption is a result of the stimulation of different receptors and it
involves the five human senses. The multimodal nature of food sensation –commonly referred as flavor–
implies not only the combination of different senses but the interaction between them (crossmodality) to
form the final flavor perception[94]. Among all the attributes that result in flavor perception, aroma plays
a key role. Aroma substances are those volatile compounds that are detected at the sensory receptors
located in the olfactory epithelium, inside the nasal cavity. Odorants can reach the olfactory receptors
during inhalation through the nose (orthonasal) or via the throat (retronasal) after they are released due to
oral food processing and also after swallowing.
To date, more than 8000 volatile compounds have been identified in food products, exhibiting a broad
range of physicochemical properties. In a single foodstuff, the volatile fraction varies from few hundred
compounds in the case of simple natural products (e.g. strawberries) to more than 900 in complex
14
products (e.g. roasted coffee)[95]. Despite the large number of volatiles present in foods, only a small
fraction is responsible for the final aroma as their concentration has to be higher than their odor
threshold (OT) in order to be relevant for sensation. The OT is defined as the lowest amount of an
aroma compound needed in order for humans to recognize its odor. OT concentrations are compound
dependent and vary several orders of magnitude, from few pptv to ppmv. The OT also depends on the
medium, temperature, the route of perception (orthonasal or retronasal) or the interactions with other
aroma compounds present in the food product. Therefore, in order to allow comparison between
compounds, OT values are generally calculated in water or air by orthonasal evaluation. A common
practice to determine which compounds contribute to the final aroma of a certain food is to calculate
their odor activity values (OAV), which is the ratio between the concentration of the compound and the
OT value in that specific food[96]. Compounds with OAV>1 are present in a higher concentration than
their thresholds and will most probably contribute to the aroma.
3.1. Aroma analysis.
The large number of aroma compounds, their chemical diversity, the low concentrations at which they
can be sensorially significant and the varied physicochemical characteristics of the food matrices
complicate aroma analysis. Unfortunately, there is not a universal method to study aroma and a careful
choice of the analytical procedure has to be done depending on the objective of the study, food material
and volatiles of interest. The main purposes of volatile analysis in food are: (i) identification of the key
compounds contributing to the aroma of the product, (ii) obtaining the volatile profile of the sample, (iii)
detection of any off-odors that affect the final quality of the product, (iv) monitoring changes in time
(aroma production, release or degradation), (v) linkage of volatile content to sensory attributes and (vi)
checking the authenticity of the food product or existence of fraud related to its composition[97].
Independently of their objective, analytical methods for aroma analysis can be divided into two different
processes: sampling or isolation of the volatile fraction and analysis of the volatile fraction.
3.2. Isolation of the volatile fraction.
The first step in aroma analysis is the separation of the volatile fraction from other compounds present in
the sample. The complexity of this step depends on the number of volatiles that are object of study, their
concentration, the complexity of the matrix and the technique that will subsequently be used for the
analysis; it can also be a rather simple process such as the withdrawal of a fraction of the headspace
above the food product, or a very elaborate process involving the combination of extraction, distillation
and concentration steps. The main methods for aroma isolation will be described, pointing out their
advantages and disadvantages.
3.2.1. Distillation and Solvent extraction.
The two simplest methods to separate aroma compounds from the food matrix are direct extraction of
the volatiles from the food matrix with an organic solvent, and distillation. Distillation can be performed
by heating the food matrix, applying vacuum or a combination of both. The high volatile compounds are
then condensed on a cooled trap. When high vacuum is used for distillation, water is also drawn into the
distillate and a further extraction with an organic solvent is needed. Nickerson and Likens developed a
method for the simultaneous distillation and extraction of volatiles with organic solvents[98].
Simultaneous distillation-extraction (SDE) is one of the most flexible aroma isolation techniques but the
use of high temperatures for distillation might result in the generation of artifacts. More recently, Engel
and co-workers developed a new distillation system called Solvent Assisted Flavor Evaporation (SAFE)
15
that reduces artifact formation[99]. The main problem of both SDE and SAFE resides in the bias
introduced by the different solubility of the compounds in the different solvents. Figure 7 shows how the
volatile recovery varies depending on the solvent used for extraction, thus resulting in completely
different volatile profiles.
3.2.2. Static Headspace (SHS).
In SHS, the sample is placed in a closed vessel, leaving some air between the sample and the lid of the
vessel. Then the sample is heated at a fixed temperature during a certain time period, until equilibrium is
reached between the sample and the air in the headspace above it. At that point, a few mL of air are
drawn for analysis. Static headspace is, a priori, the best method for aroma analysis as it directly reflects
the genuine volatile composition that arrives to the olfactive receptors (orthonasally). The main drawback
of SHS is its lack of sensitivity for aroma applications as it is difficult to detect compounds at low
concentrations in a small volume of the headspace air. Considering that several important aroma
compounds have very low threshold values and occur at very low concentrations in the headspace, the
lack of sensitivity of the SHS technique represents a real limitation in aroma applications. As it will be
discussed later, one of the most common techniques for aroma analysis is gas chromatography coupled
to mass spectrometry (GC-MS). When SHS is coupled to GC-MS, the detection limit is in the ppb range
while the concentration of most aromas in the headspace is in the ppt range, therefore allowing detection
of only the major volatile compounds[100]. In order to overcome these problems, SHS can be used with
high sensitivity instruments (e.g. GC-Chemoluminesence, PTR-MS, eNose) or the headspace can be
concentrated into a sorbent (e.g. purge and trap, SPME) in order to increase sensitivity.
3.2.3. Dynamic Headspace (DHS)
A continued flow of an inert gas is bubbled through the sample or blown above its surface in order
deplete the maximum amount of volatiles into the headspace. The gas leaving the vessel is then passed
through a trap where volatiles get retained either by condensation (cryogenic trap) or adsorption
(adsorptive polymer trap). This is the reason why DHS is commonly referred as purge-and-trap method.
Volatiles trapped are then released by fast heating and drawn by a carrier gas into the analytical
instrument (e.g. GC-MS). This method is fast, simple, and easily automated, but still has some
disadvantages. The main problem of cryogenic traps is that they will also condensate water which might
need to be separated again from the volatile fraction. The use of adsorptive polymer traps overcome the
problem of trapping water but, as the different compounds can have different adsorption affinities for
the polymeric phase, these traps alter the volatile profile of the headspace analyzed. For example, the
widely used TENAX trap has high affinity for nonpolar compounds but low for polar ones, therefore
resulting in selective enrichment of the first ones and loss of important aroma compounds.
3.2.4. Solid-phase micro-extraction (SPME)
Solid phase micro extraction was developed in the early 90s at Pawliszyn group as a solvent free sample
preparation technique[101], and it has become one of the most popular techniques to extract volatiles
from samples’ headspace and thermally desorb them into a GC. It consists of a fused-silica fiber (inert)
coated with a polymeric stationary phase incorporated into a syringe-like device. In the case of headspace
sampling, the syringe is placed inside the vessel containing the sample and the stationary phase is then
exposed to the air in the headspace. When the stationary phase is exposed, partition of the volatiles
between the headspace and the polymeric coating occurs. Once equilibrium between the fiber coating
and the headspace is reached, the fiber is removed and directly placed in a GC injector for thermal
16
desorption. This procedure can be done manually or easily automated using an autosampler. As the
principle of SPME is the partition of the analyte between fiber and headspace of the sample, fiber
coatings can be selected to have high affinity for specific compounds and a wide range of stationary
phases with different polarities is commercially available. The high surface area of the coating, together
with its high affinity for organic compounds, results in concentration of the volatiles in the fiber, thus
substantially increasing the sensitivity of headspace sampling[102]. Despite SPME being one of the most
used techniques for aroma analysis, it also has some associated problems. As it is an equilibrium
technique, the volatile profile obtained is dependent of the sampling parameters (temperature, time), the
absorption/adsorption properties of the stationary phase used and the competition of different volatiles
for the adsorption sites of the fiber[103]. Another important drawback is the stability of the fiber over
time and the differences between fibers of different batches that complicate the comparison of results
from different fibers. These disadvantages can be overcome with the use of isotopically label standards
that allow the quantification of the compounds and correct for differences in fibers performance, making
SPME a valuable tool for target analysis.
Figure 7. The recovery of volatiles depends on the isolation method and has to be chosen wisely depending on the
objective of the analysis. Figure adapted from Reineccius 2005 [104]
3.3. Analysis of the volatile fraction.
3.3.1. Gas chromatography (GC)
Modern gas chromatography is commonly attributed to the invention of Martin and James presented in
1952[105]. The contribution of GC to volatile analysis is beyond doubt. GC can be considered a
reference method for trace volatile analysis and therefore for aroma analysis. This technique basically
consists in the separation of volatiles from a mixture by partitioning between a gas phase (mobile phase)
and a solid or liquid one (stationary phase). Nowadays the stationary phase is generally a thin layer of a
viscous liquid coating the inside of a capillary column. The analyte is injected upstream of the column
17
and drawn by a carrier gas through the column. The different interactions between the volatile
compounds and the stationary phase result in separation of the compounds which elute the column at
different times (retention time). Those interactions will depend on the column coating and temperature,
thus the column is kept inside an oven to control temperature over time. At the end of the column, a
detector produces a signal each time a compound is eluting.
3.3.2. Gas Chromatography Mass Spectrometry (GC-MS)
Several detectors can be coupled to GC, each one with its own strengths and drawbacks. The selection of
the GC detector will depend on its selectivity for the analyte of interest, its sensitivity, linear range or
limit of detection. GC detectors have been reviewed by Colón and Baird[106] and only the mass
spectrometer (MS) detector will be discussed here as it is the most commonly used for aroma analysis.
The basic configuration of an MS detector consists of an electron impact (EI) ionization source, a
quadrupole mass filter, an electron multiplier and an ion counter. Standard electron ionization uses 70 eV
which is higher than the energy needed to ionize most organic compounds. Therefore, the excess of
energy results in high fragmentation of the organic molecules. Fragmentation could be considered a
negative aspect of EI but the mass spectra obtained gives structural information about the compound
and it can be used as its fingerprint. The GC retention time together with the mass spectra obtained by
EI-MS can be compared with libraries (e.g. FFNSC GC/MS library) for compound identification.
GC-MS is a reliable and well stablished technique in aroma analysis. Its main advantages are the
separation of the compounds, that can be improved by the use of multi-dimensional GC (i.e. GCxGC);
and the possibility to identify compounds by their mass spectra. But it also has some limitations. The first
one is the lack of sensitivity, which makes almost imperative the isolation and concentration of the aroma
compounds prior to analysis. A second limitation is the analysis time. The time needed for
chromatographic separation, together with the time needed for pre-concentration of the compounds,
limits the amount of samples that can be analyzed in a daily basis. This time constraint has a big relevance
when fast changes in volatile composition need to be monitored as it happens in the case of flavor
generation or release.
3.3.3. Gas chromatography – olfactometry (GC-O)
GC can also be used to detect which compounds contribute to the aroma of food. When it comes to
aroma analysis, the reference detector is indeed the nose. Considering that the human nose is more
sensitive for some odor active compounds than any other detector coupled to GC, techniques were
developed where the detector was replaced by a real human nose: GC-Olfaction (GC-O). The GC-O
approach consists in sniffing the effluent of the GC column in order to identify the aroma quality at the
respective retention times for odor active molecules[66]. Generally, the effluent is split in two, one part
going to the sniffing port and the other to a detector (e.g. MS), so the chromatographic peaks can be
associated to an odor descriptor[107]. GC-O is a qualitative technique per se, but quantitative data can be
obtained in combination with dilution analysis methods (e.g. AEDA, CHARM). Those methods consist
in the sequential dilution of the aroma extract until the odor cannot be detected. The dilution for the
lowest concentration at which a substance can still be smelled is known as Flavor Dilution (FD) factor. It
is assumed that the higher the FD of a compound in the mixture, the higher its contribution to the aroma
of the foodstuff will be[65].
18
3.3.4. Direct injection mass spectrometry (DIMS)
Analytical techniques based on direct injection mass spectrometry (DIMS) have been successfully used to
monitor volatile organic compounds in different fields, including food analysis. They overcome some of
the limitations of GC-MS methods as they allow on-line analysis of volatiles in real time, without any pretreatment and with high sensitivity. As no chromatographic separation takes place, it is crucial that the
ionization is soft (i.e. does not result in compound fragmentation) to ease the interpretation of the
recorded mass spectra. It is also important to use mass spectrometers with fast response (≤1 Hz) to
follow fast dynamic processes and with high mass resolution to discriminate between isobaric
compounds (i.e. same nominal mass). Some of the main soft ionization DIMS methods will be discussed
in this section.
Figure 10. A) Schematic mechanism of SPI and REMPI ionization (M: electronic ground state; M *: excited intermediate
state; and M+: ionized product). B) Ionization energies for some compound groups and energy of VUV light sources.
Adapted from Hanley and Zimmermann[108]
3.3.4.1.
Photoionization
Pulsed lasers can be used for gas-phase soft ionization of organic molecules with high sensitivity. The
two main photoionization methods are resonance-enhanced multi-photon ionization (REMPI) and
vacuum ultraviolet-single photon ionization (VUV-SPI). In REMPI, the energy of a single photon is
lower than the ionization potential (IP) of the molecule, therefore the absorption of at least a second
photon is needed to ionize the molecule. Although ionization is possible without excitation via an
intermediate state, the ionization is strongly enhanced when the energy of the first photon is in resonance
with an excited state of the neutral molecule (Figure 10A)[109]. Therefore, REMPI has a great selectivity
because the wavelength can be selected in order to ionize only some of the molecules of a mixture (e.g.
aromatic compounds), which can be used to discriminate between isomers[110,111]. This high selectivity
makes REMPI a valuable ionization scheme for targeted analysis but restricts its use as a universal
ionization method. By contrast, SPI, while retaining some selectivity, results in soft ionization of all
organic compounds and can be considered a universal soft ionization method. It is based in the
absorption of a single VUV photon. This photon has energy between 8 and 11 eV and most compounds
have ionization energies lower than 10 eV (Figure 10B). This will result in ionization of the molecule
without fragmentation due to the small excess of energy. At the same time, common inorganic
constituents of air have higher ionization energies and are not ionized by SPI (O2 12.06 eV, N2 15.58 eV,
CO2 13.77 eV and H2O 12.62 eV), making VUV-SPI a versatile tool to directly and selectively analyze
19
volatile organic compounds in gas samples[108]. Photoionization mass spectrometry has been mainly
applied to direct online monitoring of gas samples and some applications in food sciences will be
presented in section 5.
3.3.4.2.
Atmospheric Pressure Chemical Ionization (APCI)
Ionization of volatile organic compounds by proton transfer reaction (PTR) via the hydronium ion
(H3O+) is a universal mechanism. For compounds with higher proton affinity than that of H3O+ (Table
4) the following reaction occurs:
𝐻3 𝑂+ + 𝑅 → 𝐻2 𝑂 + 𝑅𝐻 +
Eq. 2
𝑘
The simplest method to produce hydronium ions is APCI. Ions are generated from air due to a corona
discharge (e.g. N2+) and react with the water vapor present in air to generate H3O+ and water clusters
(H20)nH+[112]. Ion-molecule reactions in the gas phase result in the ionization of the volatile compounds
by proton transfer, adduct formation or charge transfer.
Table 4. Proton affinities (PA) of some compounds.
PA
Molecule
(kJ mol-1)
O2
421
N2
494
CO2
541
O3
626
H2O
691
NH3
854
methane
544
ethene
681
benzene
750
methanol
754
Source: Hunter and Lias 1998[113]
Molecule
ethanol
phenol
acetaldehyde
propanal
acetone
2-pentanone
acetic acid
propanoic acid
acetonitrile
dimethyl sulfide
PA
(kJ mol-1)
776
817
769
786
812
833
784
797
779
831
One of the limitations of APCI is the complexity of the mass spectra due to the presence of different
ionization agents that can produce other species than the protonated molecule RH+; this complicates the
identification of the compounds based on their m/z values. With a proper design and controlling the
instrumental conditions, it is possible to obtain a mass spectra containing mainly the protonated
molecular ions[114]. APCI has been successfully used for headspace analysis[115,116], in-vivo aroma
release during consumption[117,118] and food authenticity[119].
3.3.4.3.
Selected Ion Flow Tube (SIFT)
Selected ion flow tube mass spectrometry (SIFT-MS) has been the technique of choice to study the
kinetics of ion-molecule reactions. In SIFT, reactant ions (e.g. H3O+, NO+) are selected from all the ions
formed in the ion source via a quadrupole mass filter and then injected with a carrier gas (He) into the
flow tube. Analyte volatiles are also injected via a separate inlet into the flow tube where the ionization
reaction takes place:
𝐴+ + 𝑅 → 𝐵 + + 𝐶 + 𝐷 …
𝑘
Eq. 3
Finally, the formed ions are detected by a mass spectrometer. By introducing a known concentration of
analyte gas (R) via a flow meter, the decay of the primary ion (A+) as a function of the analyte flow rate
can be measured and used to calculate the reaction rate coefficient (k). Similarly, for reactions with
20
known k, it is possible to calculate the concentration of the volatile compounds in a mixture from the
intensities of the product ions and the primary ion.[120]. A limitation of SIFT-MS is its sensitivity. Due
to the primary ion selection process, a low ion number density is found in the flow tube. However, subppb concentrations can be measured at expenses of increasing the ion counting time of the mass
detector. SIFT-MS applications for real-time quantification of VOCs have been reported in different
fields including food sciences, and they have been reviewed by Smith and Španěl[120].
3.3.4.4.
Proton transfer reaction (PTR).
Proton transfer reaction mass spectrometry (PTR-MS) was developed in the 1990s by Werner Lindinger’s
group [121–123]. PTR-MS can be considered as an evolution of SIFT-MS with the aim of increasing its
sensitivity for quantification of organic volatile compounds in complex gas mixtures. There are a few and
specific modifications in the design which drastically modify the performance of PTR-MS in comparison
to SIFT-MS. The first one is the ion source which has been replaced to ensure an almost pure source of
H3O+, eliminating the need for a mass filter. Another difference is the replacement of the flow tube by a
much shorter drift tube in which the analyte gas is delivered without the need of using helium as a carrier
gas. The kinetic energy is provided to the ions by an electric field inside the drift tube. This allows
increasing the amount of analyte gas and therefore the sensitivity of the instrument. The negative effect
of using an electric field is that the additional collision energy can result in some fragmentation. PTR-MS
has been the technique used in this thesis for the analysis of volatile compounds and therefore it will be
described in more detail.
3.3.4.4.1.
Components of PTR-MS instrument.
PTR-MS instruments consist on three basic parts: the ion source, where the hydronium ions are formed;
the drift tube, where the proton transfer reaction takes place; and the mass spectrometer, to separate ions
according to their m/z ratio. (Figure 11).
3.3.4.4.1.1.
Ion source.
The ion source is the region of the instrument where the hydronium ions, needed for the proton transfer
reaction, are produced. Although some alternative ion sources have been suggested (direct current
discharge[124], circular glow discharge[125] or radioactive[126] ion sources) most of the PTR-MS
instruments are equipped with a hollow cathode discharge ion source. Two main parts can be
differentiated in the ion source: the hollow cathode region (HC) and the source drift region (SD) (Figure
11). Water vapor is introduced in the HC region where it is ionized by electron impact and converted in
the following ions[127]:
𝑒 − + 𝐻2 𝑂 → 𝐻2 𝑂+ + 2𝑒 −
Eq. 4
𝑒 − + 𝐻2 𝑂 → 𝑂𝐻 + + 𝐻 + 2𝑒 −
Eq. 5
𝑒 − + 𝐻2 𝑂 → 𝑂+ + 2𝐻 + 2𝑒 −
Eq. 6
𝑒 − + 𝐻2 𝑂 → 𝐻 + + 𝑂𝐻 + 2𝑒 −
Eq. 7
𝑒 − + 𝐻2 𝑂 → 𝐻2 + + 𝑂 + 2𝑒 −
Eq. 8
21
To ensure a high purity (>99.5%) of hydronium ions leaving the ion source, a SD region is included. In
the SD region, the ions previously formed by electron impact undergo ion-molecule reactions that lead to
the production of H3O+ (or an ion that can be further converted into H3O+)[122]:
!! ! ! + !! !
!! ! ! + !"
Eq. 9
!" ! + !! !
!! ! ! + !
Eq. 10
!" ! + !! !
!! ! ! + !"
Eq. 11
! ! + !! !
!! ! ! + !
Eq. 12
! ! + !! !
!! ! ! + !
Eq. 13
!! ! + !! !
!! ! ! + !
Eq. 14
!! ! + !! !
!! ! ! + !!
Eq. 15
Impurity ions can be formed in the SD region due to some air entering from the drift tube region. These
ions are mainly O2+ and NO+ and, although in relative low quantities (<1%), can undergo reactions with
the analyte in the drift tube complicating the interpretation of the mass spectra.
3.3.4.4.1.2.
Drift tube.
The drift tube can be considered the main part of PTR-MS instruments as it is where the ion-molecule
reactions take place (mainly proton transfer), via collision between hydronium ions and analyte
molecules. Reactions happening in the drift tube will be briefly discussed in section 4.2.4.6.
The drift tube consists on a series of electrodes to which a voltage is applied, generating a uniform
electric field along the drift tube. The analyte gases are introduced upstream of the drift tube where they
mix with the positive ions coming from the ion source. These positive ions reach a constant velocity as a
result of two opposite forces: acceleration due to the applied electric field and collisions with the neutral
gas molecules. A portion of those ions leave the drift tube through an orifice, and are guided by a transfer
lens system into the mass spectrometer.
The drift tube can be heated up to 120 °C in order to avoid condensation of volatiles. It is generally
operated at pressure values around 2 mbar with an electric field strength (E) near 60 V·cm-1. These
parameters are commonly combined and referred to as the reduced electric field value of the drift tube
(E/N), where N is the gas number density (cm-3). E/N values are expressed in Townsends (Td, 10-17
V·cm-2) and can be easily adjusted in PTR-MS instruments in order to control the proton transfer
reaction[128]. High E/N values result in higher collision energies between H3O+ and the analyte
molecules resulting in undesired fragmentation that can complicate the interpretation of the mass spectra.
PTR-MS instruments are commonly operated at E/N values between 100 and 140 Td.
22
Figure 11. Scheme of a PTR-TOF-MS instrument
3.3.4.4.1.3.
Mass spectrometer.
In PTR-MS, volatile compounds are directly injected into the drift tube without previous separation. The
function of the mass spectrometer is to separate the ions generated in the drift tube according to their
m/z ratio. Both mass accuracy and mass resolution of the mass spectrometer are crucial in PTR-MS for
identification of compounds according to their exact mass and discrimination between near-isobaric
compounds (compounds with the same nominal mass).
Mass accuracy is defined as the difference between the mass measured in the mass spectrometer and the
theoretical mass of the compound. It is generally expressed in ppm and calculated according to:
𝑚𝑎𝑠𝑠 𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦 =
(𝑚/𝑧𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑 −𝑚/𝑧𝑡ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 )
𝑚/𝑧𝑡ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙
𝑥 106 𝑝𝑝𝑚
Eq. 16
High mass accuracy is needed to assign the elemental composition (i.e. molecular formula) to each
individual ion. As the mass of the ion increases, the number of possible molecular formulas increases as
well; therefore, a higher mass accuracy is needed for a reliable compound identification. For example, the
theoretical mass for protonated 2,3-butanedione is 87.0446. If the mass accuracy is 15 ppm, this means
that the measured mass is in the range 87.0446 ± 0.0013. In that range, only one molecular formula can
be assigned (C4H7O2), therefore the mass accuracy is good enough to determine the elemental
composition. If protonated caffeine (m/z 195.0882) is measured with the same mass accuracy, the
difference between the measured mass and the theoretical one could be up to 0.003 amu. In that range,
16 different molecular formulas can be assigned to the measured mass, implying that the accuracy is not
enough for compound identification. In some cases, previous knowledge about the compounds that
might be present in the sample can be used to tentatively identify the compound when the mass accuracy
is not high enough to get a unique molecular formula from the experimentally measured mass.
Another important parameter of the mass spectrometer is the mass resolution, which is the ability to
distinguish between two different peaks. Mass resolution is calculated as the quotient between the
measured m/z and the width of the peak (Δm) at a specific fraction of total height (usually 50%). As it
can be seen in Figure 13 (left), with high mass resolution it is possible to distinguish two different peaks:
m1 and m2. As the resolution decreases, the peak width increases to the point in which only one peak can
be distinguished. In the case of low mass resolution instruments (e.g. quadrupole), nearly-isobaric
compounds mix up in the same peak, preventing compound identification. As an example, a commercial
PTR-ToF-MS instrument was capable to distinguish between 2,3-butanedione and 3-methylbutanal in a
coffee sample (Figure 13 – center).
23
Figure 13. Left: Scheme showing the effect of different mass resolution in peak separation. Center: two nearly isobaric
compounds (2,3-butanedione and 3-methylbutanal) in a coffee sample analyzed by PTR-ToF-MS. Right: Mass peak
tentatively identified as methylpyrazine in a coffee sample analyzed by PTR-ToF-MS which might hide other compounds
as phenol. Although native molecules are drawn, the mass peaks measured in PTR-ToF-MS correspond to the
protonated molecules MH+.
Hence, in PTR-MS, high mass resolution is needed in order to separate and identify the compounds in
the sample. Three main types of mass analyzers have been used in PTR-MS instruments: quadrupole, ion
trap and time of flight (TOF) mass spectrometers.
The first PTR-MS instruments were equipped with a quadrupole filter due to its robustness, compact
size, low cost and high sensitivity and dynamic range[121,128]. Ion separation via quadrupole mass filters
is based on the stability of the trajectory of the ions in oscillating electric fields as a function of their m/z
ratio. This working principle has some advantages and disadvantages. When a quadrupole is operated in
single ion monitoring (SIM) mode (i.e. only one m/z is selected), it has a duty cycle of 100%. That means
that all ions of that specific m/z introduced in the mass spectrometer will be detected, therefore resulting
in high sensitivity. The problem arises when more than one m/z needs to be monitored or when all the
m/z in a defined range have to be measured (full scan). As only one m/z travels through the quadrupole
at a time, with all the others being excluded, the duty cycle is drastically reduced. For example, for a full
scan between 0 and 500 amu, the duty cycle will be 1/500 = 0.2%, decreasing both the sensitivity and the
scan speed. Another drawback of quadrupole filters is the low mass resolution. Although quadrupoles
can reach resolutions up to 0.1 amu, most quadrupoles have low resolution and are able to discriminate
only nominal masses[129]. The low mass resolution prevents compound identification, as near-isobaric
compounds (i.e. compounds with same nominal mass) will be detected together. The slow scan speed
and low mass resolution make quadrupoles not suitable for the online analysis of complex gas mixtures
experimenting fast changes.
Ion trap mass spectrometers have been also coupled to PTR[130–132]. Ion traps also use oscillating
electric fields but in this case they are used to store ions. Ions over the full mas range can be stored and
scanned in milliseconds, resulting in a duty cycle >90%, which can go up to >99% by increasing the
trapping time[131]. This fast response (<1s for a full spectrum) represents a huge advantage over
quadrupole filters for monitoring fast dynamic processes. Another advantage is the possibility to perform
collision induced dissociation, allowing identification of isobaric compounds[131]. The last advantage of
ion traps over quadrupoles is the possibility to use lower electric fields in order to increase sensitivity.
Low E/N values result in cluster formation, but those clusters are fragmented after isolation in the ion
trap[132]. Despite the advantages of ion traps, the number of PTR-MS equipment using this mass
spectrometer is rather low.
24
The actual tendency is to equip PTR-MS with a TOF mass spectrometer. In PTR-TOF-MS, m/z ratio is
calculated using the time that ions take to move in a field-free region from the source to the detector.
Ions coming from the drift tube enter the pulse extraction region were they are focused and accelerated
into the field-free section. As the same kinetic energy is applied to each ion, their velocities will be
inversely proportional to the square root of the mass, therefore taking different times to reach the
detector[129]. In order to increase the mass resolution, a reflectron TOF-MS is preferred to a linear one
(Figure 11). The reflectron is an electrostatic reflector that corrects for differences in the kinetic energy of
the ions leaving the extraction region. Those ions with higher kinetic energy (red path in Figure 11), will
penetrate deeper and will spend more time in the reflector. Consequently, slow and fast ions will reach
the detector at similar times[129].
The possibility of getting a full spectrum in milliseconds, the high ion transmission, high sensitivity (pptv
to ppqv range) and the high mass resolution (6000 – 10000 m/Δm) of commercial PTR-TOF-MS
instruments[133,134] justify its choice for real-time online monitoring of fast processes in food industry
(e.g aroma formation and aroma release). Figure 14 shows the mass spectra recorded for the headspace
of ground roasted coffee. The full mass spectrum was acquired with a commercial PTR-TOF-MS
instrument in one second. More than 200 mass peaks were found in the range 19-200 m/z. Thanks to the
mass accuracy of the instrument, 100 compounds were tentatively assigned to mass peaks and the mass
resolution allowed discrimination of compounds like furan (m/z 69.034) and isoprene (m/z 69.070) or
furfural (m/z 97.028) and dimethylfuran (m/z 97.065). Unfortunately, PTR-TOF-MS cannot discriminate
isobaric compounds as ethylmethylpyrazine and trimethylpyrazine (m/z 123.092). It also has problems to
detect near isobaric compounds when the intensity of one of them is high. In figure 13 (right), a mass
peak at m/z 95,060 was tentatively identified as methylpyrazine. In that coffee sample it was expected to
find also phenol at m/z 95.049, but the high intensity of the methylpyrazine mass peak made separation
impossible. In that case, higher mass resolution or previous separation of the compounds via fast-GC
approaches is needed in order to discriminate between the compounds.[135]
3.3.4.4.2.
Reactions in PTR-MS
The main reaction happening in the drift-tube of a PTR-MS instrument is non-dissociative proton
transfer and it is defined by the reaction:
𝐻3 𝑂+ + 𝑅 → 𝐻2 𝑂 + 𝑅𝐻 +
𝑘
Eq. 2
This reaction is fast, usually taking place at collisional rate, and it is thermodynamically spontaneous when
the proton affinity (PA) of the compound R is higher than the PA of hydronium ions. It enables
ionization of most organic compounds (with exemption of alkanes) while not reacting with the common
constituents of air. Although PTR is considered a soft ionization technique, some fragmentation is
observed in the final mass spectrum. Some molecules can undergo dissociation after proton transfer,
resulting in fragmentation into a charged fragment (F+) and a neutral one (N):
𝐻3 𝑂+ + 𝑅 → 𝐻2 𝑂 + (𝑅𝐻 + )∗
(𝑅𝐻 + )∗ → 𝐹 + + 𝑁
Eq. 17
Eq. 18
An example of the dissociative ionization is the dehydration of alcohols, aldehydes or carboxylic
acids[128]:
𝐻3 𝑂+ + 𝐶𝑛 𝐻2𝑛+1 𝑂𝐻 → 2𝐻2 𝑂 + 𝐶𝑛 𝐻2𝑛+1 +
Eq. 19
25
Figure 14. PTR-TOF-MS spectrum of roasted ground coffee headspace highlighting some of the main aroma compounds.
Tentatively identified compounds are drawn as neutral molecule M, although the mass peaks correspond to the protonated
compounds [MH]+.
26
Other possible reaction with the hydronium ion is association, as it happens in the case of heptane and
higher alkanes[128]:
𝐻3 𝑂+ + 𝐶𝑛 𝐻2𝑛+2 + 𝑀 → 𝐻3 𝑂 + (𝐶𝑛 𝐻2𝑛+2 ) + 𝑀
Eq. 20
Where M is a third body.
The main reactions between H3O+ and volatile organic compounds have been summarized in the book
by Ellis and Mayhew[135] and are presented in Table5.
Table 5. Main product channels for reaction of H3O+ with volatile organic compounds at 300K
Functional group
Reaction products
References
No reaction for C5 and lighter alkanes. C6 and higher species give only
association products (H3O+·M) but reaction is much slower than the
collisional limit. Cycloalkanes react at the collisional limit
Fast reaction for all alkenes except ethane.
100% MH+ for small alkenes but fragmentation important for C 7 and higher
alkenes
[136,137]
Slow (endothermic) reaction for acetylene but fast reaction for larger alkynes
dominated by MH+ production
[136–138]
Yield almost exclusively MH+
[137]
Alcohols
C3 and higher alcohols show an increasing tendency to undergo dehydration
on protonation. Dehydration is the only product channel for tertiary alcohols
[139,140]
Ethers
Main product MH+, although dissociative channels grow in importance as the
complexity of the ether increases
[141]
Aldehydes
100% production of MH+ of C3 and lower aldehydes, but increasing tendency
to eject H2O for C4 and higher species
[142]
Almost 100% production of MH+ regardless of chain length
[142]
Carboxylic acids
Dominated by MH+ formation, but accompanied by small amount of
dehydration product
[143]
Esters
MH+ formation is main channel for small esters but increasing propensity for
major ion fragmentation, particularly once the alcohol conjugate is a propyl or
higher unit.
[143,144]
Nitriles
100% production of MH+
[145]
Amines
MH+ main product, with minor channels involving alkene or H2 loss also
found for some amines
[145]
100% production of MH+
[146,147]
Alkanes
Alkenes
Alkynes
Aromatic
hydrocarbons
Ketones
Organosulfur
compounds
Organohalides
[136–138]
Most halomethanes undergo slow reaction. Heavier halomethanes undergo a [148,149]
variety of reactions, some of which include termolecular association with
H3O+. On the other hand, aromatic halides show 100% MH + production in
fast reactions.
(Source: Ellis A.M and Mayhew C.A. Proton Transfer Reaction Mass Spectrometry: Principles and Applications. 2013)
27
The ion source of PTR-MS instruments is optimized to produce high purity H3O+, but unfortunately,
the presence of unreacted water vapor from the source and from the sample leads to formation of water
clusters H3O+(H2O)n. Despite the efforts to minimize the presence of water clusters, they are frequently
present, especially when analyzing humid samples or operating at low E/N[128]. Proton affinity of water
clusters is higher than that of H2O, therefore some compounds will be ionized via proton transfer from
H3O+ but not from the clusters H3O+(H2O)n. However, ligand switching reactions can occur:
𝐻3 𝑂+ (𝐻2 𝑂) + 𝑅 → 𝐻3 𝑂+ (𝑅) + 𝐻2 𝑂
Eq. 21
These reactions lead to ions different from the expected RH+, complicating the obtained mass spectrum
in the case of complex gas mixtures and affecting quantitative analysis.
Recently, a new system has been developed that allows fast switching between H3O+, NO+, O2+, Xe+,
and Kr+ as primary ions. The system is commercially available under the name of Switchable Reagent Ion
(SRI). Some advantages are the possibility to analyze compounds that are not ionized by H3O+ and to
discriminate isobaric compounds due to different fragmentation[150,151]. The main drawback of using
other reagent ions is that the ion chemistry is not as well-known as in the case of H3O+, thus
complicating the mass spectra. Therefore these other reagents have been only used for target analysis or
fingerprinting of samples for further classification using multivariate methods[151,152].
4. Applications of direct injection mass spectrometry in food science.
4.1. Food authenticity.
The aim of food authentication is to protect the consumer from fraudulent products. The three main
authenticity issues are: (i) economic motivated adulteration of a food product, that comprise substitution
or addition of a substance in the product to increase its apparent value while decreasing the cost; (ii) false
declaration of geographical origin; and (iii) false declaration of farming regime (e.g. conventional products
labelled as organic).
Laboratories dealing with food safety and authenticity require fast and high-throughput analytical
techniques that allow fast analysis of a large number of samples. With DIMS techniques it is possible to
get a fingerprint of the volatile composition of a sample within a few minutes, and without need of
sample preparation. The classification of samples based on those fingerprints is possible even without
compound identification, which further simplifies the process. Successful application of DIMS methods
for food authentication have been reported in the literature for edible fats[153–156], cheese [157],drycured ham[158,159], honey[160], apple juice[119], strawberries[161], chocolate[162] or coffee[152,163–
165] amongst others. Only the case of coffee will be discussed here as an example.
Coffee is a natural product that is generally marketed as roasted (or roasted ground) coffee beans.
Commercially available coffees can belong to only one coffee variety (Arabica or Robusta), or can be sold
as a blend of both varieties. Robusta has a lower market value than that of Arabica coffee; thus the
incorporation of higher levels of Robusta than those specified in the label can be considered economic
motivated adulteration. Using SPI-TOF-MS, it was possible to discriminate between Arabica and
Robusta varieties by the ratio between two different terpenes: cafestol and kahweol[165]. Although the
28
study needed further statistical analysis, the potential of SPI-TOF-MS for discriminating the content of
Robusta in different blends was clear.
Coffee origin is also an issue that affects the consumer. The origin of coffee has an impact in its sensory
attributes and also in the price. Using PTR-TOF-MS coupled to an autosampler (Figure 15D) for the
high throughput analysis of coffee headspace, roasted ground coffee samples from three different origins
(Brazil, Ethiopia and Guatemala) were successfully separated by Principal Component Analysis (PCA unsupervised method) and the separation was further confirmed by Partial Least Square Regression –
Discriminant Analysis (PLS-DA – supervised method)[164]. In a follow up study, the same authors
analyzed both the roasted ground beans and the coffee brews prepared with six coffees of different
origins (Brazil, Ethiopia, Guatemala, Costa Rica, Colombia and India). In this case, different ions were
used for ionization (H3O+, NO+ and O2+). All the ions resulted in good classification of the sample, but
combination of the information obtained for all the ions via data infusion techniques increased the
efficiency of the classification[152]. Different coffee origins have also been distinguished by on-line
monitoring of the roasting process with PTR-TOF-MS [163]. The recorded time-intensity profiles during
roasting of coffee beans showed differences in flavor formation between the different origins, making
possible their discrimination.
The last issue in coffee authenticity is the false declaration of farming regime. Organic and fair trade
coffees present an added value and therefore their retail prices are higher than those of regular coffee.
Using mass spectral fingerprints and multivariate analysis it was demonstrated that it is possible to
discriminate between organic and regular coffee[171]. The study involved the headspace analysis of 110
commercially available coffees by PTR-MS. PLS-DA clearly separated organic coffees (43 samples) from
regular ones (67 samples) according to their volatile profiles.
4.2. Food quality
Food quality is of utmost importance for the consumer. First, because the consumer expects the food
product to be fresh, good looking and to have a pleasant taste when consuming it; and even more
important, because the product has to be safe for consumption. DIMS methods are precise, reliable and
allow rapid screening of samples, which allows their use as quality control techniques. Furthermore, they
are non-destructive methods which allow the long-term studies to determine the impact of storage
conditions in the food product.
Lipid oxidation in food is associated with negative aroma attributes, like rancid, and it has been studied
by DIMS in several food products. For example, the accelerated thermal oxidation of New Zealand extra
virgin oil was followed by SIFT-MS. Out of the 13 compounds analyzed, propanal, acetone and acetic
acid increased their levels with oxidation but it was not possible to correlate sensory rancidity with the
volatile profile obtained[172]. In a similar study with Italian extra virgin oils, the evolution of the volatile
profile was followed during thermal oxidation. Volatiles correlated with peroxidase activity were
identified and mainly comprise aldehydes which were associated to rancid smell[173]. Lipid oxidation
aldehydes have been also analyzed in the headspace of tomato that was left at three different
temperatures (4, 23 and 37°C) after blending. SIFT-MS measurements revealed that, while the increased
concentration of most compounds with temperature was related to the increase in volatility, three
compounds (E-2-pentenal, E-2-heptenal and E-2-octenal) were produced at much higher rates at 37°C.
Based on those results, the authors proposed a new temperature dependent lipoxygenase pathway in
29
which some unknown enzymes were activated at 37°C[174]. As a last example, SIFT-MS was compared
to SPME-GC-MS for the determination of lipid oxidation products from beef meat packed under a 20%
CO2/80% H2O atmosphere and refrigerated at 4°C. Measurements were done at 0, 2, 5, 8 and 12 days of
storage. Both techniques detected a significant increase of aldehydes with time but SIFT-MS could detect
the differences earlier than the GC-MS method[175].
Other studies based on DIMS methods for quality control include the storage stability of meat
products[176,177], fish[178], broccoli[179] or fruits[180,181] .
Figure 15. Different set ups used for volatile analysis with direct injection mass spectrometry techniques. A) Double
stripping cell used for calculation of partition coefficients in water-air systems using PTR-MS[166]. B) Simultaneous
thermal analysis coupled to SPI-MS to study single bean coffee roasting[165]. C) Direct sampling of gases produced
during coffee roasting into a SPI-MS[167]. D) Automatic headspace sampling coupled to a PTR-MS[152]. E) Artificial
chewing device coupled to PTR-MS to study aroma release during consumption[168]. F) Glass nose-piece for analysis of
in-vivo flavor release [169]. G) micro-probe used to monitor volatile formation inside coffee beans during roasting[170].
30
4.3. Flavor generation.
During food processing and cooking, chemical reactions take place, generating volatile compounds that
contribute to the final aroma of the product. Food industries demand fast analytical techniques with high
time resolution in order to study those reactions and to optimize the processing parameters leading to the
desired final flavor.
One of the most important reactions in aroma generation is the Maillard reaction. It is actually not a
single reaction, but a cascade of reactions that starts with the reaction between the carbonyl group of a
reducing sugar and the amino group of an amino acid. Maillard reaction leads to the production of
compounds that positively affect the sensory quality of food, like flavor compounds (e.g. pyrazines or
pyrroles) or colored compounds (e.g. melanoidins). Non-desired compounds are also produced by
Maillard reaction, including potential carcinogenic products as hydroxymethylfurfural and
acrylamide[182,183].
Maillard reaction has been extensively studied via online DIMS. Acrylamide formation in
asparagine/fructose model systems as function of temperature was studied by PTR-MS. High
temperatures (>150°C) significantly increased the production of acrylamide. At 170°C acrylamide levels
reached a maximum after only 2 minutes of heating, but the high temperature also favored its release,
therefore decreasing its concentration rapidly. In the case of potato slices heated at 170°C, acrylamide
maximum appeared later (~10 minutes) but the decrease was much slower, with 50% of the maximum
concentration still present after 20 minutes[184]. Formation of Maillard products has also been studied
using DIMS after thermal treatment of foods including carrot drying[185], dry sausages[175], skim milk
powder[186,187], pumpkin seeds roasting[188] or cocoa roasting[189].
Flavor generation during coffee roasting is crucial for the final quality of the product. Several reactions
lead to the formation of hundreds of compounds responsible for the roasted coffee aroma (section 1.3).
The complexity of coffee chemistry together with its economic importance explains the large amount of
research performed on coffee roasting. The first study for online monitoring of coffee roasting with
DIMS dates back to 1996[190]. Volatiles produced during the whole roasting process, from green to dark
roasted beans, were analyzed on-line with REMPI-MS from the off-gas of a simulated roaster, with a
time resolution of 1 Hz. Time intensity profiles during roasting were obtained for 4-vynilguaiacol,
guaiacol and indole. The first two were formed during roasting but, while guaiacol increased over the
whole roasting time, 4-vynilguaiacol reached a maximum and then decreased, suggesting the degradation
of the compound. Indole is a compound found in green beans, and its concentration decreased during
roasting. This first work, showed the potential of DIMS for real-time analysis of volatile formation during
coffee roasting and opened the door to several other studies using SPI-MS, REMPI-MS and PTR-MS.
Coffee roasting has been studied in small sample roasters[167,191], batch roasters[163,192] or even for
single beans[165,170]. The online analysis of volatiles in such different conditions needed different
sampling setups, some of them shown in Figure 15B, C and G. Figure 15 C shows direct sampling inside
a roaster by introducing a quartz tube into the rotating drum. A filter was introduced between the tube
and the instrument inlet to avoid solid particles entering the system, and all the system was heated at
250°C to avoid condensation of low volatile compounds and coffee oil[167]. Single bean roasting can be
performed directly in a heated vial[21] but some other interesting sampling methods have been reported
in literature. The first consists of a μ-probe that was drilled into the coffee bean (Figure 15G) and
allowed sampling of the volatiles directly in the place where they are formed (i.e. inside the bean)[170]. A
31
second one was based in the coupling of a simultaneous thermal analysis and SPI-TOF-MS (Figure 15B).
This setup allowed analysis of evolved gasses at the same time that the weight loss of the single bean was
monitored over time[165].
4.4. Flavor release.
The last application of DIMS that will be discussed is flavor release. Flavor release consists of the
liberation of volatile compounds from the food matrix so they can reach the olfactory epithelium.
Therefore, flavor release depends on the partition coefficient of the compound between air and the
matrix and it will be affected by any process that alters the matrix structure or composition (e.g. oral
processing). Although DIMS techniques have been used to determine volatile release into the headspace
in thermodynamic equilibrium (i.e. static headspace), the main advantage of DIMS over traditional
techniques (e.g. GC-MS) resides in the possibility of measuring dynamic processes in real-time. An
important application of DIMS in flavor release is the on-line monitoring of in-vivo flavor release during
food consumption.
4.4.1. Flavor release during consumption.
Release of volatiles during consumption is affected by the food matrix and how it is disrupted during oral
processing of food. Therefore, mastication, saliva flow, tongue movements or breathing patterns will
impact the release of volatiles into the mouth space and their transport to the odor receptors in the nose.
The study of flavor release during consumption is essential for understanding sensory perception of food,
and DIMS have proved to be valuable tools for that kind of analysis.
The simplest scenario of food consumption is drinking of beverages. When a liquid is placed in the
mouth, the soft palate is closed thus blocking volatiles to pass to the nasal cavity and preventing retronasal sensation. During swallowing, volatiles are released and can reach the olfactory epithelium during
subsequent breathing. Volatiles reaching the nose cavity can be analyzed by DIMS with different
sampling methods (Figure 15F)[117]. The effect of fat content in flavor release was analyzed by spiking
different liquids (water and milk with different fat content) with aroma compounds and sampling of the
volatiles in-nose with PTR-MS[193]. In all cases, the amount of compound reaching the nose was only a
small fraction of the concentration present in the liquids. As expected, lipophilic compounds were
retained to a higher extent in the samples containing fat, with differences up to 70% between water and
full fat milk. In the case of low lipophilic compounds (2,3-butanedione), no significant difference was
found between the different solutions. Fat content of the drink not only impacted the amount of volatiles
released to the nose-space but also how long those volatiles are delivered (persistence). Studies using
APCI-MS revealed that lipophilic compounds were more persistent in breath after swallowing oil/water
emulsions, and persistence was higher when the oil concentration was increased. On the contrary, oil
concentration had no effect on non-lipophilic compounds[194]. This oil dependent behavior could be
explained by the existence of emulsion residues on the oral tract that kept releasing volatiles over
time[195]. Aroma persistence in water solutions depended on the physicochemical properties of the
compounds and could be modeled by the quantitative structure property relationship approach (QSPR),
taking into account the partition coefficient, the vapor pressure and the ether linkage of the
compound[196].
For more rigid food matrices, where mastication of the product is needed, the situation is more complex.
The effect of viscosity on flavor release has been analyzed with the use of custard desserts flavored with
32
aroma compounds. Firmer custards resulted in higher concentrations of volatiles in-nose as measured by
PTR-MS. Those custards involved more mouth/tongue movements than the soft ones, which resulted in
opening of the soft palate, allowing volatiles to pass to the nasal cavity[197]. Other studies confirmed that
differences in oral processing between subjects had a higher effect on the volatiles released than that due
to differences in the texture of the food product[198]. By sampling volatiles at different points of the
nose and in the nasopharynx after eating soft custards, it was observed that only a small fraction of the
volatiles present in the nasopharynx reach the olfactory cleft (20-50%). The decrease was dependent on
the hydrophobicity of the compound and could be related to absorption into the watery mucosa of the
nose[199]. Therefore physiological parameters also need to be taken into account during in-vivo
experiments. For example, the interactions of the aroma compounds with the mouth and nose mucosa
can impact the amount of compounds that reach the olfactory receptors, affecting the sensory
perception.
In-vivo aroma analysis has also been used in combination with sensory analysis to correlate the volatile
content measured in-nose with sensory data. As compounds persist in breath even after swallowing, it is
of particularly importance to analyze both the volatile content and the sensations over time. This is the
purpose of the Temporal Dominance of Sensations (TDS) method, in which panelists rank different
attributes during a defined amount of time. By coupling TDS and on-line nose-space analysis with PTRMS, correlations have been found between volatiles and sensation for several food products including
candy[200], yogurt [201], flavored vodka[202] or coffee[203,204]. In the case of coffee, the effect of
crema (coffee foam) was studied. In the presence of crema, the roasted attribute was dominant; this
attribute, however, was not perceived in coffees without crema. One compound was also found to be
dominant in the volatile profile when crema was present, 2-methylfuran. Although this compound is
considered a marker for roasting, it does not contribute to the roasted aroma. Other compounds might
be responsible of the roasted attribute but it was not possible to assign them, pointing out the difficulty
of relating analytical and sensory data[203]. In another study, two coffees with different roasting degrees
were consumed either with or without sugar. Differentiation of coffees related to sugar was clear in the
sensory data as the dominant attribute during the first 15 seconds was sweet. On the other hand,
analytical data allowed discrimination of coffees according to their roasting degree but not to the sugar
content. Some compounds detected in the nose-space could be correlated with sensory as is the case of
methyl-pyrrole and acetyl-methyl-pyrrole with burnt notes or pyrazines with roasted flavor [204].
5. Summary of results
5.1. Extraction of volatiles during espresso coffee brewing (papers 1 and 2)
Coffee is not a food product consumed because of its nutritional value but for pleasure. As discussed in
section 1, volatiles responsible for coffee aroma are generated during coffee roasting. Figure 16 shows
PTR-TOF mass spectra of coffee at different stages: green coffee, roasted ground coffee and coffee
brew. The headspace of green coffee contains volatile compounds in low concentration (Figure 16A).
Once that coffee is roasted (Figure 16B), the volatile content increases and it is possible to find thermally
generated compounds, like pyrazine derivatives. After brewing (Figure 16C), the headspace composition
is again altered and presents a different profile of that of roasted ground coffee. The amount of volatiles
transferred from the ground coffee to the final brew depend on the brewing method and brewing
parameters (i.e. temperature, pressure), used in the preparation of the coffee brew. Aroma is a key factor
33
in the sensorial experience of drinking coffee; thus, it is quite surprising not to find more studies about
how volatile compounds are incorporated into the liquid phase during coffee brewing.
Figure 16. PTR-ToF-MS spectra of A) green coffee, B) roasted ground coffee and C) coffee brew. Tentatively identified
compounds are drawn as neutral molecule M, although the mass peaks correspond to the protonated compounds [MH]+.
34
In the case of espresso coffee, brewing is a fast process as the preparation of a full cup takes from 15s (in
the case of ristretto) to 40 seconds (for lungo). This short brewing time complicates the sampling of
fractions for off-line measurements, which results in large measurement errors. PTR-TOF-MS, as other
DIMS methods, has high time resolution and allows recording of a full mass spectra in less than a
second, allowing the online monitoring of fast processes as coffee brewing.
The first approach was to monitor volatiles directly from the coffee flow during brewing[205]. For this
study, a capsule system was chosen due to its ease of operation and the reproducibility on coffee
preparation (extracting time and extracted volume). Direct sampling from the flow in an open
atmosphere was not only possible but also highly reproducible. Figure 17 shows the time evolution of the
compound tentatively identified as pyridine during the preparation of a lungo espresso.
Figure 17. Picture of the set up for on-line monitoring of volatiles in the coffee flow (left). Normalized time intensity
profile for pyridine (blue) and the cumulative intensity (red) over the extraction of a lungo capsule. Lines are the average
of 10 capsules with ribbons representing the 95% confidence level.
Differences in extraction dynamics were found both between compounds extracted from the same coffee
and for the same compound in different coffee capsules. The differences within one capsule could be
explained due to the different physicochemical properties of the compounds, mainly polarity, and were in
agreement with the previous study of Mestadgh et al. using SPME-GC-MS methods[52]. The main
advantage of using PTR-TOF-MS was that we obtained full mass spectra with 1 Hz time resolution over
the whole brewing time (i.e. 42s) and we could tentatively identify 47 different compounds. The SPMEGC-MS method, although allowing unambiguous identification of the compounds, was based in the
analysis of different aliquots, which added experimental error to the measurements. Furthermore, it
required 2 different GC methods (with two columns of different polarity) that took between 60 and 70
minutes to analyze 20 aroma compounds in each sample, and 6 samples to have the profile of the full
extraction time.
Differences between coffee capsules could be attributed to different factors as the coffee blend used,
roasting degree, particle size or the dose of coffee in the capsule (information that was not available in
the commercial product). In order to check if the differences observed in the extraction dynamics were
enough to differentiate coffees, we performed PCA and HCA on the area under the curve for all the m/z
analyzed at different brewing times. Discrimination of capsules was possible even when only the first 15
35
seconds of extraction were considered, confirming the usefulness of the dynamic data for classification
purposes.
The only drawback of using coffee capsules is that the only variable that allowed manipulation was the
capsule type. In a follow-up experiment, the sampling set up was adapted and optimized to be used in a
semi-automatic coffee machine where both pressure and temperature could be controlled (Figure
18)[206]. Results showed that, regardless of the conditions used for extraction, volatiles could be
clustered by HCA in 5 groups according to their extraction kinetics (Figure 18 right). Compounds in the
same group shared physicochemical properties like water solubility, polarity or volatility.
One of the limitations of PTR-TOF-MS is that it is not possible to distinguish between isomers, as they
have the same molecular formula. This approach allowed us to discard some possible compound
identities as their physicochemical properties did not match the family to which the extraction profile was
assigned (e.g. the highly soluble pyridine was grouped in the family with the lowest water solubility).
Figure 18. Sampling set up used for sampling volatiles in semi-automatic coffee machine (left). Normalized time-intensity
profiles for the 5 differentiated families. Line represents the average of all compounds in the family and the ribbon the
95% confidence interval.
Regarding the impact of brewing parameters, both an increase in pressure or temperature resulted in
higher levels of volatiles although the effect was more pronounced for the less soluble compounds, and
higher in the case of temperature changes. This result was expected as it has been reported in literature
that high temperature extraction leads to higher content of compounds like caffeine, lipids or
volatiles[39,45,46]. A similar situation occurred in the case of pressure, with some compounds increasing
their concentration when extracted at 9 bar instead of 7 bar; however, pressure did not affect other
compounds like caffeine[39,47]. Our work supported the results found in literature (which were
measured in the final cup). In addition, and since the whole time profile was obtained, it was possible to
determine when the differences between brewing parameters started being significant. In general terms,
those differences started being significant after 10 seconds of extraction, thus having a greater effect in
the preparation of long cups (i.e. lungo) rather than in short cups (i.e. ristretto).
These two studies in coffee extraction showed the suitability of PTR-TOF-MS for the online analysis of
volatiles from a liquid flow in real time, with high sensitivity and high time resolution. They provided new
information about extraction of VOCs from ground coffee to water and opened a new line of research in
36
coffee. Potential applications which can take advantage of this analytical approach include: comparison of
coffee machine performance, optimization of full automatic systems (including single dose systems),
study of other important parameters in coffee brewing (e.g. water alkalinity), or relating extraction
profiles with sensory data, amongst others.
5.2. Volatile profiling of tea leaves and their infusions. Application in authenticity.
(paper 3)
The volatile profile of tea is highly dependent on the tea variety, growing conditions, country of origin or
post-harvesting conditions. The big differences expected between teas have conditioned tea aroma
research and most studies have focused on the identification of the main contributors to aroma for a
specific tea. In order to analyze the profiles of a large amount of samples in a fast and reproducible
fashion, PTR-TOF-MS was coupled to an autosampler, allowing headspace analysis of a sample in less
than 5 minutes. This approach made possible the analysis of 101 commercial teas, both full leaves and
infusions, with replicates (909 analyses) in a short time[207].
This profiling study was particularly important for several reasons. First, because it was the first time that
the volatile profile of tea had been analyzed for an extensive amount of samples based on static
headspace techniques. Wang and co-workers determined the volatile composition of 87 different teas by
SDE-GC-MS[208]. As the volatile profile obtained from SDE depends on the solvent used for
extraction, our aim was to analyze it by SHS so it better reflects the volatile profile reaching the nose via
the orthonasal pathway. Furthermore, this study considered both leaves and infusions. Differences
between leaves and infusions had been already reported and some compounds presented higher levels on
the infusions, implying that they were formed during tea brewing[60].
Figure 19. Concentration of volatiles in the headspace of full leaves or brews. Values represent the average of 63 black
teas and 38 green teas. VOCs were grouped in different chemical families.
Figure 19 shows the total concentration of volatile compounds present in the headspace of the different
samples. Despite the differences between black and green tea, it was possible to observe a clear trend
between leaves and infusions. Tea leaves presented high concentration of aldehydes, esters/acids,
hydrocarbons and terpenes. While the brews kept similar concentrations of aldehydes, the content of
hydrocarbons and terpenes was heavily reduced. This result was expected as non-polar compounds may
not be completely extracted during tea brewing. On the other hand, alcohols and esters/acids were found
in higher concentration in the infusions. As some tea alcohols are found in the leaves as glycosides,
37
hydrolysis of the glycosidic precursors may explain why some compounds presented higher levels on the
infusions than those on the leaves[60].
The obtained profiles were subjected to statistical analysis with the aim of classifying the samples
according to (i) their tea type (black or green) and (ii) the country of origin. Green and black teas could
be distinguished by PCA and separation was confirmed with four different classification models and
cross-validation. This separation was possible using either the leaves or the infusion data with
classification errors <4% and <1% respectively. In the case of the country of origin, classification was
more difficult. The classification errors ranged between 30-50%, and most of the confusion was between
samples of neighboring countries (e.g. Korea-China). This implies that country of origin is not the best
factor to discriminate teas as political borders often split regions with similar climate, growing conditions
or processing traditions which will impact the tea aroma and could therefore be better discriminators.
1.1. Extraction of volatiles during tea brewing (paper 4)
The kinetics of extraction of non-volatile soluble compounds from tea leaves has been widely studied
[79–89]. Unfortunately, in the case of volatiles, all previous studies have focused on the final cup and not
on the dynamics of extraction. The volatile profile of the tea infusions has shown significant differences
with that of the leaves, therefore it is important to understand the effect that the brewing process has in
the volatile composition of the final infusion. Furthermore, as tea preparation is highly dependent on
culture and lifestyle, brewing parameters need to be optimized in order to get a pleasant aroma that
guarantees the acceptance of the consumer.
Figure 20. Scheme of the off-line measurement of tea brewing with PTR-TOF-MS (left). Time intensity profile for
dimethylsulfide showing the average of three replicates and the standard deviation.
Using PTR-TOF-MS the extraction of volatiles from loose leaves into the final infusion was followed
over 5 minutes extraction, at different temperatures, using waters of different mineral composition. As
tea brewing is a relatively long process, the initial temperature of water drops several degrees with time.
To ensure the monitoring of differences in extraction and not on release (due to the different
temperatures), the brewing process was followed off-line by taking fractions every 30 seconds and
analyzing them at the same temperature (Figure 20).
Using this approach, the time intensity profiles of 88 m/z (including 34 tentatively identified compounds)
were obtained for each of the brewing parameters considered (3 temperatures – 60,70 and 80 °C; 2 leaf
sizes – full and broken; and 2 waters – soft and hard). Due to the large amount of samples measured and
38
the different variables that may impact the extraction, data was subjected to different statistical analysis in
order to better interpret the results.
A simple PCA including all analyzed samples allowed differentiation of the effect of the different
variables in the volatile profile (Figure 21). HCA using the concentration at all time points for the 88 m/z
measured in each sample, allowed discrimination of all temperature and leaf size combinations with
exception of broken leaves brewed at 70 or 80°C which were clustered together. The extraction from
broken leaves at these temperatures was fast, with no significant differences in volatile levels between
broken and full leaves. It was also not possible to discriminate between the two different waters used for
extraction within one temperature-size combination.
Figure 21. PCA containing all samples analyzed in the study (3 temperatures x 2 leaf sizes x 2 waters x 10 time points x 3
replicates)
HCA was also performed on each individual measurement (10 time points x 3 temperatures x 2 waters x
2 leaf sizes x 3 replicates) revealing the brewing conditions that resulted in equivalent volatile profiles.
In order to compare how volatiles were extracted from leaves, the time intensity profiles were normalized
to the intensity at the end of the extraction for each of the measured m/z. HCA on that data resulted in
two differentiated groups that corresponded with the two different leaf sizes. This results indicate that
temperature plays a role in the total amount of compound being extracted into the infusion (higher levels
with higher temperatures), but not on the mechanism by which compounds are extracted; therefore all
temperatures had similar extraction profiles. On the other hand, when the leaf size is reduced, the rate of
extraction changes, modifying the extraction profile of the compounds.
The combination of PTR-TOF-MS data with multivariate tools proved to be a useful approach to study
the tea brewing process and how it is affected by different parameters (temperature, leaf size, brewing
time and water hardness).
39
6. References
[1]
M. Pendergrast, Coffee: second to oil?, Tea Coffee Trade J. (2009).
[2]
ICO, Coffee trade statistics. International Coffee Organization., (n.d.). http://www.ico.org/new_historical.asp
(accessed March 1, 2016).
[3]
A. Illy, R. Viani, Espresso Coffee: The Chemistry of Quality, in: Acad. Press. London, 1995.
[4]
R.J. Clarke, Coffee: Volume 1: Chemistry, Springer Science & Business Media, 2012.
[5]
K.A. Anderson, B.W. Smith, Chemical Profiling To Differentiate Geographic Growing Origins of Coffee, J.
Agric. Food Chem. 50 (2002) 2068–2075. doi:10.1021/jf011056v.
[6]
V. Krivan, P. Barth, A.F. Morales, Multielement analysis of green coffee and its possible use for the
determination of origin, Mikrochim. Acta. 110 (1993) 217–236. doi:10.1007/BF01245106.
[7]
E. Alpizar, B. Bertrand, A.S.I. du Café, Incidence of elevation on chemical composition and beverage quality of
coffee in Central America., in: ASIC 2004. 20th Int. Conf. Coffee Sci. Bangalore, India, 11-15 Oct. 2004.,
Association Scientifique Internationale du Café (ASIC), 2005: pp. 322–327.
[8]
J. Avelino, B. Barboza, J.C. Araya, C. Fonseca, F. Davrieux, B. Guyot, et al., Effects of slope exposure, altitude
and yield on coffee quality in two altitudeterroirs of Costa Rica, Orosi and Santa María de Dota, J. Sci. Food
Agric. 85 (2005) 1869–1876. doi:10.1002/jsfa.2188.
[9]
T. Joët, A. Laffargue, F. Descroix, S. Doulbeau, B. Bertrand, A. de kochko, et al., Influence of environmental
factors, wet processing and their interactions on the biochemical composition of green Arabica coffee beans,
Food Chem. 118 (2010) 693–701. doi:10.1016/j.foodchem.2009.05.048.
[10]
G. Bytof, S.-E. Knopp, P. Schieberle, I. Teutsch, D. Selmar, Influence of processing on the generation of
gamma-aminobutyric acid in green coffee beans, Eur. Food Res. Technol. 220 (2004) 245–250.
doi:10.1007/s00217-004-1033-z.
[11]
S. Knopp, G. Bytof, D. Selmar, Influence of processing on the content of sugars in green Arabica coffee beans,
Eur. Food Res. Technol. 223 (2005) 195–201. doi:10.1007/s00217-005-0172-1.
[12]
I. Flament, Y. Bessière-Thomas, Coffee Flavor Chemistry, 2002.
[13]
M.N. Clifford, Coffee: botany, biochemistry, and production of beans and beverage, 1985.
[14]
A.T. Toci, A. Farah, Volatile compounds as potential defective coffee beans’ markers., Food Chem. 108 (2008)
1133–41. doi:10.1016/j.foodchem.2007.11.064.
[15]
P.D.C. Mancha Agresti, A.S. Franca, L.S. Oliveira, R. Augusti, Discrimination between defective and nondefective Brazilian coffee beans by their volatile profile, Food Chem. 106 (2008) 787–796.
doi:10.1016/j.foodchem.2007.06.019.
[16]
J.C. Spadone, G. Takeoka, R. Liardon, Analytical investigation of Rio off-flavor in green coffee, J. Agric. Food
Chem. 38 (1990) 226–233. doi:10.1021/jf00091a050.
[17]
V. Sberveglieri, I. Concina, M. Falasconi, E. Gobbi, A. Pulvirenti, P. Fava, et al., Early detection of fungal
contamination on green coffee by a MOX sensors based Electronic Nose, in: OLFACTION Electron. NOSE
Proc. 14TH Int. Symp. OLFACTION Electron. NOSE, AIP Publishing, 2011: pp. 119–120.
doi:10.1063/1.3626328.
[18]
M. Falasconi, I. Concina, E. Gobbi, V. Sberveglieri, A. Pulvirenti, G. Sberveglieri, Electronic Nose for
Microbiological Quality Control of Food Products, Int. J. Electrochem. 2012 (2012) 1–12.
doi:10.1155/2012/715763.
[19]
A. Illy, R. Viani, Espresso Coffee: The Science of Quality, Academic Press, 2005.
[20]
R. Eggers, A. Pietsch, Technology I: roasting, Coffee Recent Dev. (n.d.).
[21]
C. Yeretzian, A. Jordan, R. Badoud, W. Lindinger, From the green bean to the cup of coffee: Investigating coffee
roasting by on-line monitoring of volatiles, Eur. Food Res. Technol. 214 (2002) 92–104. doi:10.1007/s00217001-0424-7.
[22]
S. Schenker, S. Handschin, B. Frey, R. Perren, F. Escher, Pore Structure of Coffee Beans, J. Food Sci. 65 (2000)
452–457. doi:10.1111/j.1365-2621.2000.tb16026.x.
40
[23]
Y.-F. Chu, Coffee: Emerging Health Effects and Disease Prevention, 2012.
[24]
W.B. Sunarharum, D.J. Williams, H.E. Smyth, Complexity of coffee flavor: A compositional and sensory
perspective, Food Res. Int. 62 (2014) 315–325. doi:10.1016/j.foodres.2014.02.030.
[25]
M. Petracco, Chapter 7-Technology IV. beverage preparation: brewing trends for the new millennium, CoffeeRecent Dev. (2001) 140–164.
[26]
L.N. Bell, C.R. Wetzel, A.N. Grand, Caffeine content in coffee as influenced by grinding and brewing
techniques, Food Res. Int. 29 (1996) 785–789. doi:10.1016/S0963-9969(97)00002-1.
[27]
I.A. Ludwig, L. Sanchez, B. Caemmerer, L.W. Kroh, M.P. de Peña, C. Cid, Extraction of coffee antioxidants:
Impact of brewing time and method, Food Res. Int. 48 (2012) 57–64. doi:10.1016/j.foodres.2012.02.023.
[28]
I. López-Galilea, N. Fournier, C. Cid, E. Guichard, Changes in headspace volatile concentrations of coffee brews
caused by the roasting process and the brewing procedure, J. Agric. Food Chem. 54 (2006) 8560–8566.
doi:10.1021/jf061178t.
[29]
N. Caporaso, A. Genovese, M.D. Canela, A. Civitella, R. Sacchi, Neapolitan coffee brew chemical analysis in
comparison to espresso, moka and American brews, Food Res. Int. 61 (2014) 152–160.
doi:10.1016/j.foodres.2014.01.020.
[30]
W.M.N. Ratnayake, R. Hollywood, E. O’Grady, B. Stavric, Lipid content and composition of coffee brews
prepared by different methods, Food Chem. Toxicol. 31 (1993) 263–269. doi:10.1016/0278-6915(93)90076-B.
[31]
A.N. Gloess, B. Schönbächler, B. Klopprogge, L. D`Ambrosio, K. Chatelain, A. Bongartz, et al., Comparison of
nine common coffee extraction methods: instrumental and sensory analysis, Eur. Food Res. Technol. 236 (2013)
607–627. doi:10.1007/s00217-013-1917-x.
[32]
E. Illy, L. Navarini, Neglected Food Bubbles: The Espresso Coffee Foam., Food Biophys. 6 (2011) 335–348.
doi:10.1007/s11483-011-9220-5.
[33]
E. Shimoni, T.P. Labuza, Degassing kinetics and sorption equilibrium of carbon dioxide in fresh roasted and
ground coffee, J. Food Process Eng. 23 (2000) 419–436. doi:10.1111/j.1745-4530.2000.tb00524.x.
[34]
L. Navarini, R. Capuccio, F. Suggi-Liverani, A. Illy, Espresso coffee beverage: classification of texture terms, J.
Texture Stud. 35 (2005) 525–541. doi:10.1111/j.1745-4603.2004.35504.x.
[35]
B. Zanoni, E. Pagliarini, C. Peri, Modelling the aqueous extraction of soluble substances from ground roasted
coffee, J. Sci. Food Agric. 58 (1992) 275–279. doi:10.1002/jsfa.2740580217.
[36]
M. Spiro, Modelling the aqueous extraction of soluble substances from ground roast coffee, J. Sci. Food Agric.
61 (1993) 371–372. doi:10.1002/jsfa.2740610314.
[37]
T.A. Lee, R. Kempthorne, J.K. Hardy, Compositional Changes in Brewed Coffee as a Function of Brewing
Time, J. Food Sci. 57 (1992) 1417–1419. doi:10.1111/j.1365-2621.1992.tb06872.x.
[38]
S. Andueza, M. Paz De Peña, C. Cid, Chemical and Sensorial Characteristics of Espresso Coffee As Affected by
Grinding and Torrefacto Roast, J. Agric. Food Chem. 51 (2003) 7034–7039. doi:10.1021/jf034628f.
[39]
M. Moeenfard, J.A. Silva, N. Borges, A. Santos, A. Alves, Diterpenes in espresso coffee: impact of preparation
parameters, Eur. Food Res. Technol. 240 (2014) 763–773. doi:10.1007/s00217-014-2381-y.
[40]
S. Andueza, M.A. Vila, M. Paz de Peña, C. Cid, Influence of coffee/water ratio on the final quality of espresso
coffee, J. Sci. Food Agric. 87 (2007) 586–592. doi:10.1002/jsfa.2720.
[41]
E.E. Lockhart, C.L. Tucker, M.C. Merritt, The effect of water impurities on the flavor of brewed coffee, J. Food
Sci. 20 (1955) 598–605. doi:10.1111/j.1365-2621.1955.tb16874.x.
[42]
D. Rivetti, L. Navarini, Effects of water composition and water treatment on espresso coffee percolation, Proc.
19th …. (2001).
[43]
C.H. Hendon, L. Colonna-Dashwood, M. Colonna-Dashwood, The role of dissolved cations in coffee
extraction., J. Agric. Food Chem. 62 (2014) 4947–50. doi:10.1021/jf501687c.
[44]
L. Navarini, D. Rivetti, Water quality for Espresso coffee, Food Chem. 122 (2010) 424–428.
doi:10.1016/j.foodchem.2009.04.019.
[45]
P. Masella, L. Guerrini, S. Spinelli, L. Calamai, P. Spugnoli, F. Illy, et al., A new espresso brewing method, J.
41
Food Eng. 146 (2015) 204–208. doi:10.1016/j.jfoodeng.2014.09.001.
[46]
S. Andueza, L. Maeztu, L. Pascual, C. Ibáñez, M.P. de Peña, C. Cid, Influence of extraction temperature on the
final quality of espresso coffee, J. Sci. Food Agric. 83 (2003) 240–248. doi:10.1002/jsfa.1304.
[47]
S. Andueza, L. Maeztu, B. Dean, M.P. de Peña, J. Bello, C. Cid, Influence of water pressure on the final quality
of Arabica espresso coffee. Application of multivariate analysis, J. Agric. Food Chem. 50 (2002) 7426–7431.
doi:10.1021/jf0206623.
[48]
C.C. Teo, S.N. Tan, J.W.H. Yong, C.S. Hew, E.S. Ong, Pressurized hot water extraction (PHWE)., J.
Chromatogr. A. 1217 (2010) 2484–94. doi:10.1016/j.chroma.2009.12.050.
[49]
R.C. Alves, C. Soares, S. Casal, J.O. Fernandes, M.B.P.P. Oliveira, Acrylamide in espresso coffee: Influence of
species, roast degree and brew length, Food Chem. 119 (2010) 929–934. doi:10.1016/j.foodchem.2009.07.051.
[50]
M. Petracco, Physico-chemical and structural characterization of espresso coffee brew, Proc. 13th Int. Colloq.
Coffee. (1989).
[51]
C. Bartel, M. Mesias, F.J. Morales, Investigation on the extractability of melanoidins in portioned espresso coffee,
Food Res. Int. 67 (2015) 356–365. doi:10.1016/j.foodres.2014.11.053.
[52]
F. Mestdagh, T. Davidek, M. Chaumonteuil, B. Folmer, I. Blank, The kinetics of coffee aroma extraction, Food
Res. Int. 63 (2014) 271–274. doi:10.1016/j.foodres.2014.03.011.
[53]
S. Bandini, E. Illy, C. Simone, F.S. Liverani, Cellular Automata: Research Towards Industry: ACRI’98 --Proceedings of the Third Conference on Cellular Automata for Research and Industry, Trieste, 7--9 October
1998, in: S. Bandini, R. Serra, F.S. Liverani (Eds.), Springer London, London, 1998: pp. 157–164.
doi:10.1007/978-1-4471-1281-5_15.
[54]
FAOSTAT, (n.d.). http://faostat3.fao.org/browse/Q/QC/E (accessed March 16, 2016).
[55]
B. Banerjee, Botanical classification of tea, Tea. (1992).
[56]
IV. Processing of tea, Food Rev. Int. 11 (1995) 409–434. doi:10.1080/87559129509541052.
[57]
K. Willson, M. Clifford, Tea: Cultivation to consumption, (2012).
[58]
Z. Yang, S. Baldermann, N. Watanabe, Recent studies of the volatile compounds in tea, Food Res. Int. 53 (2013)
585–599. doi:10.1016/j.foodres.2013.02.011.
[59]
VIII. Flavor of tea, Food Rev. Int. 11 (1995) 477–525. doi:10.1080/87559129509541056.
[60]
C. Schuh, P. Schieberle, P.E.S. Chieberle, Characterization of the key aroma compounds in the beverage
prepared from Darjeeling black tea: Quantitative differences between tea leaves and infusion, J. Agric. Food
Chem. 54 (2006) 916–924. doi:10.1021/jf052495n.
[61]
T. Yamanishi, Flavor of Tea, Food Rev. Int. 11 (1995) 477–525.
[62]
R. Ravichandran, Carotenoid composition, distribution and degradation to flavour volatiles during black tea
manufacture and the effect of carotenoid supplementation on tea quality and aroma, Food Chem. 78 (2002) 23–
28. doi:10.1016/S0308-8146(01)00303-X.
[63]
D. Wang, T. Yoshimura, K. Kubota, A. Kobayashi, Analysis of Glycosidically Bound Aroma Precursors in Tea
Leaves. 1. Qualitative and Quantitative Analyses of Glycosides with Aglycons as Aroma Compounds, J. Agric.
Food Chem. 48 (2000) 5411–5418. doi:10.1021/jf000443m.
[64]
K. Wang, F. Liu, Z. Liu, J. Huang, Z. Xu, Y. Li, et al., Comparison of catechins and volatile compounds among
different types of tea using high performance liquid chromatograph and gas chromatograph mass spectrometer,
Int. J. Food Sci. Technol. 46 (2011) 1406–1412. doi:10.1111/j.1365-2621.2011.02629.x.
[65]
W. Grosch, Detection of potent odorants in foods by aroma extract dilution analysis, Trends Food Sci. Technol.
4 (1993) 68–73. doi:10.1016/0924-2244(93)90187-F.
[66]
T.E. Acree, Peer Reviewed: GC/Olfactometry GC With a Sense of Smell, Anal. Chem. 69 (1997) 170A–175A.
doi:10.1021/ac971560s.
[67]
L.-F.F. Wang, J.-Y.Y. Lee, J.-O.O. Chung, J.-H.H. Baik, S. So, S.-K.K. Park, Discrimination of teas with different
degrees of fermentation by SPME–GC analysis of the characteristic volatile flavour compounds, Food Chem.
109 (2008) 196–206. doi:10.1016/j.foodchem.2007.12.054.
42
[68]
J. Lin, P. Zhang, Z. Pan, H. Xu, Y. Luo, X. Wang, Discrimination of oolong tea (Camellia sinensis) varieties
based on feature extraction and selection from aromatic profiles analysed by HS-SPME/GC-MS., Food Chem.
141 (2013) 259–65. doi:10.1016/j.foodchem.2013.02.128.
[69]
H. Yu, J. Wang, Discrimination of LongJing green-tea grade by electronic nose, Sensors Actuators B Chem. 122
(2007) 134–140. doi:10.1016/j.snb.2006.05.019.
[70]
Q. Chen, J. Zhao, Z. Chen, H. Lin, D.-A. Zhao, Discrimination of green tea quality using the electronic nose
technique and the human panel test, comparison of linear and nonlinear classification tools, Sensors Actuators B
Chem. 159 (2011) 294–300. doi:10.1016/j.snb.2011.07.009.
[71]
H. Yu, J. Wang, H. Xiao, M. Liu, Quality grade identification of green tea using the eigenvalues of PCA based on
the E-nose signals, Sensors Actuators B Chem. 140 (2009) 378–382. doi:10.1016/j.snb.2009.05.008.
[72]
K. Kumazawa, H. Masuda, Identification of potent odorants in different green tea varieties using flavor dilution
technique, J. Agric. Food Chem. 50 (2002) 5660–5663. doi:10.1021/jf020498j.
[73]
N. Togari, a Kobayashi, T. Aishima, Pattern recognition applied to gas chromatographic profiles of volatile
components in three tea categories, Food Res. Int. 28 (1995) 495–502. doi:10.1016/0963-9969(95)00029-1.
[74]
N. Ye, L. Zhang, X. Gu, Discrimination of Green Teas from Different Geographical Origins by Using HSSPME/GC–MS and Pattern Recognition Methods, Food Anal. Methods. 5 (2011) 856–860. doi:10.1007/s12161011-9319-9.
[75]
S. Baldermann, Z. Yang, T. Katsuno, V.A. Tu, N. Mase, Y. Nakamura, et al., Discrimination of Green, Oolong,
and Black Teas by GC-MS Analysis of Characteristic Volatile Flavor Compounds, Am. J. Anal. Chem. 05 (2014)
620–632. doi:10.4236/ajac.2014.59070.
[76]
J. Lee, D.H. Chambers, E. Chambers IV, K. Adhikari, Y. Yoon, Volatile aroma compounds in various brewed
green teas, Molecules. 18 (2013) 10024–10041. doi:10.3390/molecules180810024.
[77]
M. Spiro, D.S. Jago, Kinetics and equilibria of tea infusion. Part 3.—Rotating-disc experiments interpreted by a
steady-state model, J. Chem. Soc. Faraday Trans. 1 Phys. Chem. Condens. Phases. 78 (1982) 295.
doi:10.1039/f19827800295.
[78]
W.E. Price, M. Spiro, Kinetics and equilibria of tea infusion: Rates of extraction of theaflavin, caffeine and
theobromine from several whole teas and sieved fractions, J. Sci. Food Agric. 36 (1985) 1309–1314.
doi:10.1002/jsfa.2740361216.
[79]
M. Spiro, Kinetics and equilibria of tea infusion—Part 6: The effects of salts and of pH on the concentrations
and partition constants of theaflavins and caffeine in Kapchorua Pekoe fannings, Food Chem. 24 (1987) 51–61.
doi:10.1016/0308-8146(87)90083-5.
[80]
M. Spiro, W.E. Price, Kinetics and equilibria of tea infusion: Part 7—The effects of salts and of pH on the rate
of extraction of caffeine from Kapchorua Pekoe Fannings, Food Chem. 25 (1987) 49–59. doi:10.1016/03088146(87)90053-7.
[81]
M. Spiro, W.E. Price, W.M. Miller, M. Arami, Kinetics and equilibria of tea infusion: Part 8—the effects of salts
and of pH on the rate of extraction of theaflavins from black tea leaf, Food Chem. 25 (1987) 117–126.
doi:10.1016/0308-8146(87)90060-4.
[82]
W. Price, J. Spitzer, The temperature dependence of the rate of extraction of soluble constituents of black tea,
Food Chem. 46 (1993) 133–136. doi:10.1016/0308-8146(93)90025-B.
[83]
M. Spiro, D. Jaganyi, M.C. Broom, Kinetics and equilibria of tea infusion: Part 9. The rates and temperature
coefficients of caffeine extraction from green Chun Mee and black Assam Bukial teas, Food Chem. 45 (1992)
333–335. doi:10.1016/0308-8146(92)90033-X.
[84]
W.E. Price, J.C. Spitzer, The kinetics of extraction of individual flavanols and caffeine from a Japanese green tea
(Sen Cha Uji Tsuyu) as a function of temperature, Food Chem. 50 (1994) 19–23. doi:10.1016/03088146(94)90086-8.
[85]
D. Jaganyi, T. Ndlovu, Kinetics of tea infusion. Part 3: the effect of tea bag size and shape on the rate of caffeine
extraction from Ceylon orange pekoe tea, Food Chem. 75 (2001) 63–66. doi:10.1016/S0308-8146(01)00186-8.
[86]
D. Jaganyi, S. Mdletshe, Kinetics of tea infusion. Part 2: the effect of tea-bag material on the rate and
temperature dependence of caffeine extraction from black Assam tea, Food Chem. 70 (2000) 163–165.
43
doi:10.1016/S0308-8146(99)00262-9.
[87]
M. Spiro, Y.Y. Chong, D. Jaganyi, Kinetics and equilibria of tea infusion. Part 13. Further studies on tea scum:
the effects of calcium carbonate, lemon juice and sugar, Food Chem. 57 (1996) 295–298. doi:10.1016/03088146(95)00236-7.
[88]
D. Jaganyi, R.D. Price, Kinetics of tea infusion: the effect of the manufacturing process on the rate of extraction
of caffeine, Food Chem. 64 (1999) 27–31. doi:10.1016/S0308-8146(98)00101-0.
[89]
W.E. Price, M. Spiro, Kinetics and equilibria of tea infusion: Theaflavin and caffeine concentrations and partition
constants in several whole teas and sieved fractions, J. Sci. Food Agric. 36 (1985) 1303–1308.
doi:10.1002/jsfa.2740361215.
[90]
S. Suteerapataranon, J. Butsoongnern, P. Punturat, W. Jorpalit, C. Thanomsilp, Caffeine in Chiang Rai tea
infusions: Effects of tea variety, type, leaf form, and infusion conditions, Food Chem. 114 (2009) 1335–1338.
doi:10.1016/j.foodchem.2008.11.013.
[91]
S. Castiglioni, E. Damiani, P. Astolfi, P. Carloni, Influence of steeping conditions (time, temperature, and particle
size) on antioxidant properties and sensory attributes of some white and green teas, Int. J. Food Sci. Nutr. 7486
(2015) 1–7. doi:10.3109/09637486.2015.1042842.
[92]
P. Pérez-Tejeda, A. Maestre, M. Balón, J. Hidalgo, M.A. Muñoz, M. Sánchez, Setschenow coefficients for
caffeine, theophylline and theobromine in aqueous electrolyte solutions, J. Chem. Soc. Faraday Trans. 1 Phys.
Chem. Condens. Phases. 83 (1987) 1029. doi:10.1039/f19878301029.
[93]
A. Mossion, M. Potin-Gautier, S. Delerue, I. Le Hécho, P. Behra, Effect of water composition on aluminium,
calcium and organic carbon extraction in tea infusions, Food Chem. 106 (2008) 1467–1475.
doi:10.1016/j.foodchem.2007.05.098.
[94]
A. Taylor, D. Roberts, Flavor perception, (2008).
[95]
L.M. Nijssen, C.A. van Ingen-Visscher, J.J.. Donders, VCF Volatile Compounds in Food: database. Version 16.2,
Triskeleton, Zeist (The Netherlands), 2016.
[96]
S. Patton, D. Josephson, A method for determining significance of volatile flavor compounds in foods, J. Food
Sci. (1957).
[97]
G. Reineccius, A. Voilley, P. Etiévant, Choosing the correct analytical technique in aroma analysis., (2006) 81–97.
[98]
G. Nickerson, S. Likens, Gas chromatography evidence for the occurrence of hop oil components in beer, J.
Chromatogr. A. (1966).
[99]
W. Engel, W. Bahr, P. Schieberle, Solvent assisted flavour evaporation - a new and versatile technique for the
careful and direct isolation of aroma compounds from complex food matrices, Eur. Food Res. Technol. 209
(1999) 237–241. doi:10.1007/s002170050486.
[100]
J. Schaefer, Isolation and concentration from the vapour phase, Handb. Der Aroma Forchung Akad. (1981).
[101]
C.L. Arthur, J. Pawliszyn, Solid phase microextraction with thermal desorption using fused silica optical fibers,
Anal. Chem. 62 (1990) 2145–2148. doi:10.1021/ac00218a019.
[102]
J. Pawliszyn, Solid Phase Microextraction: Theory and Practice, 1997.
[103]
R. Marsili, SPME comparison studies and what they reveal, in: R. Marsili (Ed.), Flavor, Fragrance, Odor Anal.,
New York, 2002: pp. 205–227.
[104]
G. Reineccius, Flavor Chemistry and Technology, Second Edition, 2005.
[105]
A.T. James, A.J.P. Martin, Gas-liquid partition chromatography: the separation and micro-estimation of volatile
fatty acids from formic acid to dodecanoic acid, Biochem. J. 50 (1952) 679–690. doi:10.1042/bj0500679.
[106]
L.A. Colón, L.J. Baird, Detectors in Modern Gas Chromatography, in: R.L. Grob, E.F. Barry (Eds.), Mod. Pract.
Gas Chromatogr. Fourth Ed., John Wiley & Sons, Inc., 2004: pp. 275–337. doi:10.1002/0471651141.ch6.
[107]
S. Van Ruth, Methods for gas chromatography-olfactometry: a review, Biomol. Eng. (2001).
[108]
L. Hanley, R. Zimmermann, Light and molecular ions: the emergence of vacuum UV single-photon ionization in
MS., Anal. Chem. 81 (2009) 4174–82. doi:10.1021/ac8013675.
44
[109]
R. Zimmermann, Laser ionisation mass spectrometry for on-line analysis of complex gas mixtures and
combustion effluents, Anal. Bioanal. Chem. 381 (2005) 57–60. doi:10.1007/s00216-004-2886-5.
[110]
C. Weickhardt, R. Zimmermann, U. Boesl, E.W. Schlag, Laser mass spectrometry of dibenzodioxin,
dibenzofuran and two isomers of dichlorodibenzodioxins: Selective ionization, Rapid Commun. Mass Spectrom.
7 (1993) 183–185. doi:10.1002/rcm.1290070303.
[111]
C. Weickhardt, R. Zimmermann, K.W. Schramm, U. Boesl, E.W. Schlag, D.M. Lubman, Laser mass
spectrometry of the Di-, Tri- and tetrachlorobenzenes: Isomer-selective ionization and detection, Rapid
Commun. Mass Spectrom. 8 (1994) 381–384. doi:10.1002/rcm.1290080508.
[112]
W.C. Byrdwell, Atmospheric pressure chemical ionization mass spectrometry for analysis of lipids, Lipids. 36
(2001) 327–346. doi:10.1007/s11745-001-0725-5.
[113]
E.P.L. Hunter, S.G. Lias, Evaluated Gas Phase Basicities and Proton Affinities of Molecules: An Update, J. Phys.
Chem. Ref. Data. 27 (1998) 413. doi:10.1063/1.556018.
[114]
A. Taylor, R. Linforth, Food flavour technology, (2009).
[115]
A.J. Taylor, R.S.T. Linforth, Direct mass spectrometry of complex volatile and non-volatile flavour mixtures, Int.
J. Mass Spectrom. 223-224 (2003) 179–191. doi:10.1016/S1387-3806(02)00789-3.
[116]
J. Wright, F. Wulfert, J. Hort, A.J. Taylor, Effect of preparation conditions on release of selected volatiles in tea
headspace, J. Agric. Food Chem. 55 (2007) 1445–1453. doi:10.1021/jf062657v.
[117]
A. Taylor, Atmospheric pressure chemical ionisation mass spectrometry for in vivo analysis of volatile flavour
release, Food Chem. 71 (2000) 327–338. doi:10.1016/S0308-8146(00)00182-5.
[118]
M. Hodgson, R.S.T. Linforth, a. J. Taylor, Simultaneous real-time measurements of mastication, swallowing,
nasal airflow, and aroma release, J. Agric. Food Chem. 51 (2003) 5052–5057. doi:10.1021/jf030118+.
[119]
H.-H. Gan, C. Soukoulis, I. Fisk, Atmospheric pressure chemical ionisation mass spectrometry analysis linked
with chemometrics for food classification – A case study: Geographical provenance and cultivar classification of
monovarietal clarified apple juices, Food Chem. 146 (2014) 149–156. doi:10.1016/j.foodchem.2013.09.024.
[120]
D. Smith, P. Spanel, Selected ion flow tube mass spectrometry (SIFT-MS) for on-line trace gas analysis., Mass
Spectrom. Rev. 24 661–700. doi:10.1002/mas.20033.
[121]
W. Lindinger, J. Hirber, H. Paretzke, An ion/molecule-reaction mass spectrometer used for on-line trace gas
analysis, … J. Mass Spectrom. Ion …. (1993).
[122]
A. Hansel, A. Jordan, R. Holzinger, P. Prazeller, W. VOGEL, W. Lindinger, Proton-Transfer Reaction MassSpectrometry - Online Trace Gas- Analysis At the Ppb Level, Int. J. Mass Spectrom. Ion Process. 150 (1995)
609–619.
[123]
W. Lindinger, A. Hansel, Analysis of trace gases at ppb levels by proton transfer reaction mass, Plasma Sources
Sci. Technol. 6 (1997) 111–117. doi:10.1088/0963-0252/6/2/004.
[124]
S. Inomata, H. Tanimoto, N. Aoki, J. Hirokawa, Y. Sadanaga, A novel discharge source of hydronium ions for
proton transfer reaction ionization: Design, characterization, and performance, Rapid Commun. Mass Spectrom.
20 (2006) 1025–1029. doi:10.1002/rcm.2405.
[125]
D.R. Hanson, M. Koppes, A. Stoffers, R. Harsdorf, K. Edelen, Proton transfer mass spectrometry at 11 hPa
with a circular glow discharge: Sensitivities and applications, Int. J. Mass Spectrom. 282 (2009) 28–37.
doi:10.1016/j.ijms.2009.01.021.
[126]
D.R. Hanson, J. Greenberg, B.E. Henry, E. Kosciuch, Proton transfer reaction mass spectrometry at high drift
tube pressure, Int. J. Mass Spectrom. 223-224 (2003) 507–518. doi:10.1016/S1387-3806(02)00924-7.
[127]
M.M. Mann, A. Hustrulid, J.T. Tate, The ionization and dissociation of water vapor and ammonia by electron
impact, Phys. Rev. 58 (1940) 340–347. doi:10.1103/PhysRev.58.340.
[128]
R.S. Blake, P.S. Monks, A.M. Ellis, Proton-transfer reaction mass spectrometry., Chem. Rev. 109 (2009) 861–96.
doi:10.1021/cr800364q.
[129]
E. Hoffmann, Mass spectrometry, (1996).
[130]
P. Prazeller, P.T. Palmer, E. Boscaini, T. Jobson, M. Alexander, Proton transfer reaction ion trap mass
45
spectrometer., Rapid Commun. Mass Spectrom. 17 (2003) 1593–9. doi:10.1002/rcm.1088.
[131]
C. Warneke, J.A. De Gouw, E.R. Lovejoy, P.C. Murphy, W.C. Kuster, R. Fall, Development of proton-transfer
ion trap-mass spectrometry: On-line detection and identification of volatile organic compounds in air, J. Am.
Soc. Mass Spectrom. 16 (2005) 1316–1324. doi:10.1016/j.jasms.2005.03.025.
[132]
M.M.L. Steeghs, C. Sikkens, E. Crespo, S.M. Cristescu, F.J.M. Harren, Development of a proton-transfer
reaction ion trap mass spectrometer: Online detection and analysis of volatile organic compounds, Int. J. Mass
Spectrom. 262 (2007) 16–24. doi:10.1016/j.ijms.2006.09.031.
[133]
A. Jordan, S. Haidacher, G. Hanel, E. Hartungen, L. Märk, H. Seehauser, et al., A high resolution and high
sensitivity proton-transfer-reaction time-of-flight mass spectrometer (PTR-TOF-MS), Int. J. Mass Spectrom. 286
(2009) 122–128. doi:10.1016/j.ijms.2009.07.005.
[134]
P. Sulzer, E. Hartungen, G. Hanel, S. Feil, K. Winkler, P. Mutschlechner, et al., A Proton Transfer ReactionQuadrupole interface Time-Of-Flight Mass Spectrometer (PTR-QiTOF): High speed due to extreme sensitivity,
Int. J. Mass Spectrom. 368 (2014) 1–5. doi:10.1016/j.ijms.2014.05.004.
[135]
M. Fischer, S. Wohlfahrt, J. Varga, G. Matuschek, M.R. Saraji-Bozorgzad, A. Walte, et al., Evolution of Volatile
Flavor Compounds During Roasting of Nut Seeds by Thermogravimetry Coupled to Fast-Cycling Optical
Heating Gas Chromatography-Mass Spectrometry with Electron and Photoionization, Food Anal. Methods.
(2016). doi:10.1007/s12161-016-0549-8.
[136]
A.M. Ellis, C.A. Mayhew, Proton Transfer Reaction Mass Spectrometry: Principles and Applications, John Wiley
& Sons, 2013.
[137]
M. Norman, A. Hansel, A. Wisthaler, O2+ as reagent ion in the PTR-MS instrument: Detection of gas-phase
ammonia, Int. J. Mass Spectrom. 265 (2007) 382–387. doi:10.1016/j.ijms.2007.06.010.
[138]
G.H. Wannier, On the Motion of Gaseous Ions in a Strong Electric Field. I, Phys. Rev. 83 (1951) 281–289.
doi:10.1103/PhysRev.83.281.
[139]
L. Keck, U. Oeh, C. Hoeschen, Corrected equation for the concentrations in the drift tube of a proton transfer
reaction-mass spectrometer (PTR-MS), Int. J. Mass Spectrom. 264 (2007) 92–95. doi:10.1016/j.ijms.2007.03.021.
[140]
G.H. Wannier, Motion of Gaseous Ions in Strong Electric Fields, Bell Syst. Tech. J. 32 (1953) 170–254.
doi:10.1002/j.1538-7305.1953.tb01426.x.
[141]
M. Mcfarland, D.L. Albritton, F.C. Fehsenfeld, E.E. Ferguson, A.L. Schmeltekopf, Flow-drift technique for ion
mobility and ion-molecule reaction rate constant measurements. II. Positive ion reactions of N+, O+, and
[formula omitted] with O2 and O+ with N2 from thermal to [inverted lazy s]2 eV, J. Chem. Phys. 59 (1973)
6620–6628. doi:10.1063/1.1680042.
[142]
N.F. Dalleska, K. Honma, P.B. Armentrout, Stepwise solvation enthalpies of protonated water clusters: collisioninduced dissociation as an alternative to equilibrium studies, J. Am. Chem. Soc. 115 (1993) 12125–12131.
doi:10.1021/ja00078a059.
[143]
Y. Xie, R.B. Remington, H.F. Schaefer, The protonated water dimer: Extensive theoretical studies of H5O+2, J.
Chem. Phys. 101 (1994) 4878–4884. doi:10.1063/1.467409.
[144]
X. Huang, B.J. Braams, J.M. Bowman, Ab initio potential energy and dipole moment surfaces for H 5O 2+, J.
Chem. Phys. 122 (2005) 1–12. doi:10.1063/1.1834500.
[145]
A. Bekstein, M. Benhenni, M. Yousfi, Collision cross sections and swarm coefficients of water vapour ion
clusters (H2O) n H+ with n = 1, 2 and 3 in N2, O2 and air, Eur. Phys. J. D. 61 (2010) 153–159.
doi:10.1140/epjd/e2010-10376-7.
[146]
Y. Lau, S. Ikuta, P. Kebarle, Thermodynamics and kinetics of the gas-phase reactions H3O+ (H2O) n-1+
water= H3O+ (H2O) n, J. Am. Chem. …. (1982) 1462–1469. doi:10.1021/ja00370a002.
[147]
D.L. Albritton, I. Dotan, W. Lindinger, M. McFarland, J. Tellinghuisen, F.C. Fehsenfeld, Effects of ion speed
distributions in flow-drift tube studies of ion–neutral reactions, J. Chem. Phys. 66 (1977) 410.
doi:10.1063/1.433986.
[148]
L. Viehland, E. Mason, Statistical–mechanical theory of gaseous ion–molecule reactions in an electrostatic field,
J. Chem. Phys. 66 (2008) 422–434. doi:10.1063/1.433987.
46
[149]
D.A. Fairley, D.B. Milligan, C.G. Freeman, M.J. McEwan, P. ??pan??l, D. Smith, Competitive association and
charge transfer in the reactions of NO+ with some ketones: A selected ion flow drift tube study, Int. J. Mass
Spectrom. 193 (1999) 35–43. doi:10.1016/S1387-3806(99)00144-X.
[150]
J. De Gouw, C. Warneke, T. Karl, J. de Gouw, C. Warneke, T. Karl, et al., Sensitivity and specificity of
atmospheric trace gas detection by proton-transfer-reaction mass spectrometry, Int. J. Mass Spectrom. 223-224
(2003) 365–382. doi:10.1016/S1387-3806(02)00926-0.
[151]
A. Jordan, S. Haidacher, G. Hanel, E. Hartungen, J. Herbig, L. Märk, et al., An online ultra-high sensitivity
Proton-transfer-reaction mass-spectrometer combined with switchable reagent ion capability (PTR + SRI - MS),
Int. J. Mass Spectrom. 286 (2009) 32–38. doi:10.1016/j.ijms.2009.06.006.
[152]
P. Sulzer, A. Edtbauer, E. Hartungen, S. Jürschik, A. Jordan, G. Hanel, et al., From conventional protontransfer-reaction mass spectrometry (PTR-MS) to universal trace gas analysis, Int. J. Mass Spectrom. 321-322
(2012) 66–70. doi:10.1016/j.ijms.2012.05.003.
[153]
S. Yener, A. Romano, L. Cappellin, P.M. Granitto, E. Aprea, L. Navarini, et al., Tracing coffee origin by direct
injection headspace analysis with PTR/SRI-MS, Food Res. Int. 69 (2015) 235–243.
doi:10.1016/j.foodres.2014.12.046.
[154]
M. Maçatelli, W. Akkermans, A. Koot, M. Buchgraber, A. Paterson, S. van Ruth, Verification of the geographical
origin of European butters using PTR-MS, J. Food Compos. Anal. 22 (2009) 169–175.
doi:10.1016/j.jfca.2008.10.009.
[155]
S.M. van Ruth, B. Villegas, W. Akkermans, M. Rozijn, H. van der Kamp, A. Koot, Prediction of the identity of
fats and oils by their fatty acid, triacylglycerol and volatile compositions using PLS-DA, Food Chem. 118 (2010)
948–955. doi:10.1016/j.foodchem.2008.10.047.
[156]
C. Ruiz-Samblás, A. Tres, A. Koot, S.M. van Ruth, A. González-Casado, L. Cuadros-Rodríguez, Proton transfer
reaction-mass spectrometry volatile organic compound fingerprinting for monovarietal extra virgin olive oil
identification, Food Chem. 134 (2012) 589–596. doi:10.1016/j.foodchem.2012.02.135.
[157]
N. Araghipour, J. Colineau, A. Koot, W. Akkermans, J.M.M. Rojas, J. Beauchamp, et al., Geographical origin
classification of olive oils by PTR-MS, Food Chem. 108 (2008) 374–383. doi:10.1016/j.foodchem.2007.10.056.
[158]
S.A. Galle, A. Koot, C. Soukoulis, L. Cappellin, F. Biasioli, M. Alewijn, et al., Typicality and geographical origin
markers of protected origin cheese from The Netherlands revealed by PTR-MS., J. Agric. Food Chem. 59 (2011)
2554–63. doi:10.1021/jf104170r.
[159]
J.S. Del Pulgar, C. Soukoulis, F. Biasioli, L. Cappellin, C. García, F. Gasperi, et al., Rapid characterization of dry
cured ham produced following different PDOs by proton transfer reaction time of flight mass spectrometry
(PTR-ToF-MS)., Talanta. 85 (2011) 386–93. doi:10.1016/j.talanta.2011.03.077.
[160]
J. Sánchez del Pulgar, C. Soukoulis, A.I. Carrapiso, L. Cappellin, P. Granitto, E. Aprea, et al., Effect of the pig
rearing system on the final volatile profile of Iberian dry-cured ham as detected by PTR-ToF-MS., Meat Sci. 93
(2013) 420–8. doi:10.1016/j.meatsci.2012.10.003.
[161]
A. Agila, S. Barringer, Application of selected ion flow tube mass spectrometry coupled with chemometrics to
study the effect of location and botanical origin on volatile profile of unifloral American honeys., J. Food Sci. 77
(2012) C1103–8. doi:10.1111/j.1750-3841.2012.02916.x.
[162]
P. GRANITTO, F. BIASIOLI, E. APREA, D. MOTT, C. FURLANELLO, T. MARK, et al., Rapid and nondestructive identification of strawberry cultivars by direct PTR-MS headspace analysis and data mining
techniques, Sensors Actuators B Chem. 121 (2007) 379–385. doi:10.1016/j.snb.2006.03.047.
[163]
V. Acierno, S. Yener, M. Alewijn, F. Biasioli, S. van Ruth, Factors contributing to the variation in the volatile
composition of chocolate: Botanical and geographical origins of the cocoa beans, and brand-related formulation
and processing, Food Res. Int. 84 (2016) 86–95. doi:10.1016/j.foodres.2016.03.022.
[164]
A.N. Gloess, A. Vietri, F. Wieland, S. Smrke, B. Schönbächler, J.A.S. López, et al., Evidence of different flavour
formation dynamics by roasting coffee from different origins: On-line analysis with PTR-ToF-MS, Int. J. Mass
Spectrom. 365-366 (2014) 324–337. doi:10.1016/j.ijms.2014.02.010.
[165]
S. Yener, A. Romano, L. Cappellin, T.D. Märk, J. Sánchez Del Pulgar, F. Gasperi, et al., PTR-ToF-MS
characterisation of roasted coffees (C. arabica) from different geographic origins., J. Mass Spectrom. 49 (2014)
929–35. doi:10.1002/jms.3455.
47
[166]
M. Fischer, S. Wohlfahrt, J. Varga, M. Saraji-Bozorgzad, G. Matuschek, T. Denner, et al., Evolved gas analysis by
single photon ionization-mass spectrometry, J. Therm. Anal. Calorim. 116 (2014) 1461–1469.
doi:10.1007/s10973-014-3830-3.
[167]
Ö. Özdestan, S.M. van Ruth, M. Alewijn, A. Koot, A. Romano, L. Cappellin, et al., Differentiation of specialty
coffees by proton transfer reaction-mass spectrometry, Food Res. Int. 53 (2013) 433–439.
doi:10.1016/j.foodres.2013.05.013.
[168]
B.M. Davis, M.J. McEwan, Determination of olive oil oxidative status by selected ion flow tube mass
spectrometry., J. Agric. Food Chem. 55 (2007) 3334–8. doi:10.1021/jf063610c.
[169]
E. Aprea, F. Biasioli, G. Sani, C. Cantini, T.D. Märk, F. Gasperi, Proton transfer reaction-mass spectrometry
(PTR-MS) headspace analysis for rapid detection of oxidative alteration of olive oil., J. Agric. Food Chem. 54
(2006) 7635–40. doi:10.1021/jf060970r.
[170]
Y. Xu, S. Barringer, Effect of temperature on lipid-related volatile production in tomato puree., J. Agric. Food
Chem. 57 (2009) 9108–13. doi:10.1021/jf902192r.
[171]
A. Olivares, K. Dryahina, P. Spaněl, M. Flores, Rapid detection of lipid oxidation in beef muscle packed under
modified atmosphere by measuring volatile organic compounds using SIFT-MS., Food Chem. 135 (2012) 1801–
8. doi:10.1016/j.foodchem.2012.06.075.
[172]
E.S. Holm, A.P.S. Adamsen, A. Feilberg, A. Schäfer, M.M. Løkke, M.A. Petersen, Quality changes during storage
of cooked and sliced meat products measured with PTR-MS and HS-GC-MS., Meat Sci. 95 (2013) 302–10.
doi:10.1016/j.meatsci.2013.04.046.
[173]
D. Jaksch, R. Margesin, T. Mikoviny, J.D. Skalny, E. Hartungen, F. Schinner, et al., The effect of ozone
treatment on the microbial contamination of pork meat measured by detecting the emissions using PTR-MS and
by enumeration of microorganisms, Int. J. Mass Spectrom. 239 (2004) 209–214. doi:10.1016/j.ijms.2004.07.018.
[174]
B. Noseda, M.T. Islam, M. Eriksson, M. Heyndrickx, K. De Reu, H. Van Langenhove, et al., Microbiological
spoilage of vacuum and modified atmosphere packaged Vietnamese Pangasius hypophthalmus fillets., Food
Microbiol. 30 (2012) 408–19. doi:10.1016/j.fm.2011.12.025.
[175]
S. Raseetha, S.P. Heenan, I. Oey, D.J. Burritt, N. Hamid, A new strategy to assess the quality of broccoli
(Brassica oleracea L. italica) based on enzymatic changes and volatile mass ion profile using Proton Transfer
Reaction Mass Spectrometry (PTR-MS), Innov. Food Sci. Emerg. Technol. 12 (2011) 197–205.
doi:10.1016/j.ifset.2010.12.005.
[176]
E.I. Boamfa, M.M.L. Steeghs, S.M. Cristescu, F.J.M. Harren, Trace gas detection from fermentation processes in
apples; An intercomparison study between proton-transfer-reaction mass spectrometry and laser photoacoustics,
Int. J. Mass Spectrom. 239 (2004) 193–201. doi:10.1016/j.ijms.2004.07.017.
[177]
A. Boschetti, F. Biasioli, M. van Opbergen, C. Warneke, A. Jordan, R. Holzinger, et al., PTR-MS real time
monitoring of the emission of volatile organic compounds during postharvest aging of berryfruit, Postharvest
Biol. Technol. 17 (1999) 143–151. doi:10.1016/S0925-5214(99)00052-6.
[178]
T. Karl, C. Yeretzian, A. Jordan, W. Lindinger, Dynamic measurements of partition coefficients using protontransfer-reaction mass spectrometry (PTR–MS), Int. J. Mass Spectrom. 223-224 (2003) 383–395.
doi:10.1016/S1387-3806(02)00927-2.
[179]
R. Dorfner, T. Ferge, A. Kettrup, R. Zimmermann, C. Yeretzian, Real-time monitoring of 4-vinylguaiacol,
guaiacol, and phenol during coffee roasting by resonant laser ionization time-of-flight mass spectrometry, J.
Agric. Food Chem. 51 (2003) 5768–5773. doi:10.1021/jf0341767.
[180]
B. Farneti, A.A. Alarcón, S.M. Cristescu, G. Costa, F.J.M. Harren, N.T.E. Holthuysen, et al., Aroma volatile
release kinetics of tomato genotypes measured by PTR-MS following artificial chewing, Food Res. Int. 54 (2013)
1579–1588. doi:10.1016/j.foodres.2013.09.015.
[181]
C. Yeretzian, P. Pollien, C. Lindinger, S. Ali, Individualization of Flavor Preferences: Toward a Consumer-centric
and Individualized Aroma Science, Compr. Rev. Food Sci. Food Saf. 3 (2004) 152–159. doi:10.1111/j.15414337.2004.tb00066.x.
[182]
R. Hertz-Schünemann, T. Streibel, S. Ehlert, R. Zimmermann, Looking into individual coffee beans during the
roasting process: Direct micro-probe sampling on-line photo-ionisation mass spectrometric analysis of coffee
roasting gases, Anal. Bioanal. Chem. 405 (2013) 7083–7096. doi:10.1007/s00216-013-7006-y.
48
[183]
V. Somoza, Five years of research on health risks and benefits of Maillard reaction products: an update, Mol.
Nutr. Food Res. (2005).
[184]
E. Capuano, V. Fogliano, Acrylamide and 5-hydroxymethylfurfural (HMF): A review on metabolism, toxicity,
occurrence in food and mitigation strategies, LWT - Food Sci. Technol. 44 (2011) 793–810.
doi:10.1016/j.lwt.2010.11.002.
[185]
P. Pollien, C. Lindinger, C. Yeretzian, I. Blank, Proton transfer reaction mass spectrometry is a suitable tool for
monitoring on-line the formation of acrylamide in food and maillard systems, Anal. Chem. 75 (2003) 5488–5494.
doi:10.1021/ac0344586.
[186]
H. DUAN, S.A. BARRINGER, CHANGES IN FURAN AND OTHER VOLATILE COMPOUNDS IN
SLICED CARROT DURING AIR-DRYING, J. Food Process. Preserv. 36 (2012) 46–54. doi:10.1111/j.17454549.2011.00550.x.
[187]
J.A. Turner, L.R. Sivasundaram, M.-A. Ottenhof, I.A. Farhat, R.S.T. Linforth, A.J. Taylor, Monitoring Chemical
and Physical Changes during Thermal Flavor Generation, J. Agric. Food Chem. 50 (2002) 5406–5411.
doi:10.1021/jf0203803.
[188]
J.A. Turner, R.S.T. Linforth, A.J. Taylor, Real-Time Monitoring of Thermal Flavor Generation in Skim Milk
Powder Using Atmospheric Pressure Chemical Ionization Mass Spectrometry, J. Agric. Food Chem. 50 (2002)
5400–5405. doi:10.1021/jf020381v.
[189]
T. Bowman, S. Barringer, Analysis of factors affecting volatile compound formation in roasted pumpkin seeds
with selected ion flow tube-mass spectrometry (SIFT-MS) and sensory analysis., J. Food Sci. 77 (2012) C51–60.
doi:10.1111/j.1750-3841.2011.02465.x.
[190]
Y. Huang, S.A. Barringer, Monitoring of cocoa volatiles produced during roasting by selected ion flow tube-mass
spectrometry (SIFT-MS)., J. Food Sci. 76 (2011) C279–86. doi:10.1111/j.1750-3841.2010.01984.x.
[191]
R. Zimmermann, H.J. Heger, C. Yeretzian, H. Nagel, U. Boesl, Application of Laser Ionization Mass
Spectrometry for On-line Monitoring of Volatiles in the Headspace of Food Products: Roasting and Brewing of
Coffee, Rapid Commun. Mass Spectrom. 10 (1996) 1975–1979. doi:10.1002/(SICI)10970231(199612)10:15<1975::AID-RCM786>3.0.CO;2-X.
[192]
R. Dorfner, T. Ferge, C. Yeretzian, A. Kettrup, R. Zimmermann, Laser mass spectrometry as on-line sensor for
industrial process analysis: process control of coffee roasting., Anal. Chem. 76 (2004) 1386–402.
doi:10.1021/ac034758n.
[193]
F. Wieland, A.N. Gloess, M. Keller, A. Wetzel, S. Schenker, C. Yeretzian, Online monitoring of coffee roasting
by proton transfer reaction time-of-flight mass spectrometry (PTR-ToF-MS): towards a real-time process control
for a consistent roast profile., Anal. Bioanal. Chem. 402 (2012) 2531–43. doi:10.1007/s00216-011-5401-9.
[194]
D.D. Roberts, P. Pollien, N. Antille, C. Lindinger, C. Yeretzian, Comparison of nosespace, headspace, and
sensory intensity ratings for the evaluation of flavor absorption by fat., J. Agric. Food Chem. 51 (2003) 3636–42.
doi:10.1021/jf026230+.
[195]
M.D. Hodgson, J.P. Langridge, R.S.T. Linforth, a J. Taylor, Aroma release and delivery following the
consumption of beverages., J. Agric. Food Chem. 53 (2005) 1700–6. doi:10.1021/jf040316g.
[196]
S. Camacho, V. van Riel, C. de Graaf, F. van de Velde, M. Stieger, Physical and sensory characterizations of oral
coatings of oil/water emulsions., J. Agric. Food Chem. 62 (2014) 5789–95. doi:10.1021/jf5010868.
[197]
R. Linforth, A.J. Taylor, Persistence of volatile compounds in the breath after their consumption in aqueous
solutions, J. Agric. Food Chem. 48 (2000) 5419–5423. doi:10.1021/jf000488n.
[198]
S.M. van Ruth, L. de Witte, A.R. Uriarte, Volatile Flavor Analysis and Sensory Evaluation of Custard Desserts
Varying in Type and Concentration of Carboxymethyl Cellulose, J. Agric. Food Chem. 52 (2004) 8105–8110.
doi:10.1021/jf048647g.
[199]
E. Aprea, F. Biasioli, F. Gasperi, T.D. Märk, S. van Ruth, In vivo monitoring of strawberry flavour release from
model custards: effect of texture and oral processing, Flavour Fragr. J. 21 (2006) 53–58. doi:10.1002/ffj.1702.
[200]
S.M. van Ruth, J. Frasnelli, L. Carbonell, Volatile flavour retention in food technology and during consumption:
Juice and custard examples, Food Chem. 106 (2008) 1385–1392. doi:10.1016/j.foodchem.2007.08.093.
[201]
I. Déléris, A. Saint-Eve, F. Dakowski, E. Sémon, J.-L. Le Quéré, H. Guillemin, et al., The dynamics of aroma
49
release during consumption of candies of different structures, and relationship with temporal perception, Food
Chem. 127 (2011) 1615–1624. doi:10.1016/j.foodchem.2011.02.028.
[202]
J. Mesurolle, A. Saint-Eve, I. Déléris, I. Souchon, Impact of fruit piece structure in yogurts on the dynamics of
aroma release and sensory perception., Molecules. 18 (2013) 6035–56. doi:10.3390/molecules18056035.
[203]
I. Déléris, A. Saint-Eve, Y. Guo, P. Lieben, M.-L. Cypriani, N. Jacquet, et al., Impact of swallowing on the
dynamics of aroma release and perception during the consumption of alcoholic beverages., Chem. Senses. 36
(2011) 701–13. doi:10.1093/chemse/bjr038.
[204]
D. Barron, N. Pineau, W. Matthey-Doret, S. Ali, J. Sudre, J.C. Germain, et al., Impact of crema on the aroma
release and the in-mouth sensory perception of espresso coffee., Food Funct. 3 (2012) 923–30.
doi:10.1039/c2fo30046j.
[205]
M. Charles, A. Romano, S. Yener, M. Barnabà, L. Navarini, T.D. Märk, et al., Understanding flavour perception
of espresso coffee by the combination of a dynamic sensory method and in-vivo nosespace analysis, Food Res.
Int. 69 (2015) 9–20. doi:10.1016/j.foodres.2014.11.036.
[206]
J.A. Sánchez-López, R. Zimmermann, C. Yeretzian, Insight into the Time-Resolved Extraction of Aroma
Compounds during Espresso Coffee Preparation: Online Monitoring by PTR-ToF-MS, Anal. Chem. 86 (2014)
11696–11704. doi:10.1021/ac502992k.
[207]
J.A. Sánchez López, M. Wellinger, A.N. Gloess, R. Zimmermann, C. Yeretzian, Extraction kinetics of coffee
aroma compounds using a semi-automatic machine: On-line analysis by PTR-ToF-MS, Int. J. Mass Spectrom.
401 (2016) 22–30. doi:10.1016/j.ijms.2016.02.015.
[208]
S. Yener, J.A. Sánchez-López, P.M. Granitto, L. Cappellin, T.D. Märk, R. Zimmerman, et al., Rapid and direct
volatile compound profiling of black and green teas (camellia sinensis) from different countries with Ptr-ToF-ms,
Talanta. (2016). doi:10.1016/j.talanta.2016.01.050.
[209]
K. Wang, F. Liu, Z. Liu, J. Huang, Z. Xu, Y. Li, et al., Comparison of catechins and volatile compounds among
different types of tea using high performance liquid chromatograph and gas chromatograph mass spectrometer,
Int. J. Food Sci. Technol. 46 (2011) 1406–1412. doi:10.1111/j.1365-2621.2011.02629.x.
[210]
N. Khan, H. Mukhtar, Multitargeted therapy of cancer by green tea polyphenols., Cancer Lett. 269 (2008) 269–
80. doi:10.1016/j.canlet.2008.04.014.
[211]
C. Cabrera, R. Artacho, R. Giménez, Beneficial Effects of Green Tea—A Review, J. Am. Coll. Nutr. 25 (2006)
79–99. doi:10.1080/07315724.2006.10719518.
[212]
V. Crespy, G. Williamson, A Review of the Health Effects of Green Tea Catechins in In Vivo Animal Models, J.
Nutr. 134 (2004) 3431S–3440.
[213]
C. Astill, M.R. Birch, C. Dacombe, P.G. Humphrey, P.T. Martin, Factors affecting the caffeine and polyphenol
contents of black and green tea infusions, J. Agric. Food Chem. 49 (2001) 5340–5347. doi:10.1021/jf010759+.
[214]
C. Alasalvar, B. Topal, A. Serpen, B. Bahar, E. Pelvan, Flavor Characteristics of Seven Grades of Black Tea
Produced in Turkey, (2012).
[215]
M. Shimoda, H. Shigematsu, H. Shiratsuchi, Y. Osajima, Comparison of Volatile Compounds among Different
Grades of Green Tea and Their Relations to Odor Attributes, J. Agric. Food Chem. 43 (1995) 1621–1625.
doi:10.1021/jf00054a038.
[216]
H. Chen, H. Liang, J. Ding, J. Lai, Y. Huan, X. Qiao, Rapid Differentiation of Tea Products by Surface
Desorption Atmospheric Pressure Chemical Ionization Mass Spectrometry Rapid Differentiation of Tea
Products by Surface Desorption Atmospheric Pressure Chemical, (2007) 10093–10100. doi:10.1021/jf0720234.
[217]
W. Lindinger, A. Hansel, A. Jordan, On-line monitoring of volatile organic compounds at pptv levels by means
of proton-transfer-reaction mass spectrometry (PTR-MS) medical applications, food control and environmental
research, Int. J. Mass Spectrom. Ion Process. 173 (1998) 191–241. doi:10.1016/S0168-1176(97)00281-4.
[218]
L. Cappellin, F. Biasioli, A. Fabris, E. Schuhfried, C. Soukoulis, T.D. Märk, et al., Improved mass accuracy in
PTR-TOF-MS: Another step towards better compound identification in PTR-MS, Int. J. Mass Spectrom. 290
(2010) 60–63. doi:10.1016/j.ijms.2009.11.007.
[219]
L. Cappellin, F. Biasioli, P.M. Granitto, E. Schuhfried, C. Soukoulis, F. Costa, et al., On data analysis in PTRTOF-MS: From raw spectra to data mining, Sensors Actuators B Chem. 155 (2011) 183–190.
50
doi:10.1016/j.snb.2010.11.044.
[220]
R Core Team, R: A Language and Environment for Statistical Computing, (2015).
[221]
T. Kiribuchi, T. Yamanishi, Studies On the Flavor of Green Tea. IV Dimethyl Sulfide and its Precursor, Agr.
Biol. Chem. 27 (1963) 56–59.
[222]
M. Kawabata, K. Ohtsuki, H. Kokura, Y. Wakahara, Determination of Dimethylsulfide in the Head Space Vapor
of Green Tea by Gas Chromatography, Agric. Biol. Chem. (2014).
[223]
M. Liebert, U. Licht, V. Böhm, Antioxidant properties and total phenolics content of green and black tea under
different brewing conditions, (1999) 217–220.
[224]
Y. Liang, Y. Xu, Effect of extraction temperature on cream and extractability of black tea [Camellia sinensis (L.)
O. Kuntze], Int. J. Food Sci. Technol. 38 (2003) 37–45. doi:10.1046/j.1365-2621.2003.00631.x.
51
7. Annex
I.
List of figures
Figure 1. Roasting of coffee beans – main aspects[19,20] ..............................................................................
Figure 2. Main volatiles produced during roasting from non-volatile precursors. Adapted from
Yeretzian et al 2002[21] ......................................................................................................................................
Figure 3. Dynamics of espresso brewing. Data is presented as % of the total amount extracted in 24
seconds. Fractions were taken every 8 seconds and had volumes between 14-17mL. Data obtained
from Ludwig et al. 2012[27] ................................................................................................................................
Figure 4. Water (left) and soluble compounds (right) exiting the ground coffee bed system in a
computational model for espresso brewing.[53] .............................................................................................
Figure 5. Tea producing countries color-coded according to 2013 FAOSTAT data.................................
Figure 6. Simplified proposed biosynthetic pathways of tea volatiles. From Yang et al. 2013[58]. .........
Figure 7. The recovery of volatiles depends on the isolation method and has to be chosen wisely
depending on the objective of the analysis. Figure adapted from Reineccius 2005 [104] .........................
Figure 10. A) Schematic mechanism of SPI and REMPI ionization (M: electronic ground state; M*:
excited intermediate state; and M+: ionized product). B) Ionization energies for some compound
groups and energy of VUV light sources. Adapted from Hanley and Zimmermann[108] .......................
Figure 11. Scheme of a PTR-TOF-MS instrument .........................................................................................
Figure 12. Left: Scheme showing the effect of different mass resolution in peak separation. Right:
two nearly isobaric compounds (2,3-butanedione and 3-methylbutanal) in a coffee sample analyzed
by PTR-ToF-MS ...................................................................................................................................................
Figure 13. Left: Scheme showing the effect of different mass resolution in peak separation. Right:
two nearly isobaric compounds (2,3-butanedione and 3-methylbutanal) in a coffee sample analyzed
by PTR-ToF-MS ...................................................................................................................................................
Figure 14. PTR-TOF-MS spectrum of roasted ground coffee headspace highlighting some of the
main aroma compounds.
Figure 15. Different set ups used for volatile analysis with direct injection mass spectrometry
techniques. A) Double stripping cell used for calculation of partition coefficients in water-air systems
using PTR-MS[166]. B) Simultaneous thermal analysis coupled to SPI-MS to study single bean coffee
roasting[165]. C) Direct sampling of gases produced during coffee roasting into a SPI-MS[167]. D)
Automatic headspace sampling coupled to a PTR-MS[152]. E) Artificial chewing device coupled to
PTR-MS to study aroma release during consumption[168]. F) Glass nose-piece for analysis of in-vivo
flavor release [169]. G) micro-probe used to monitor volatile formation inside coffee beans during
roasting[170]. .........................................................................................................................................................
Figure 17. Picture of the set up for on-line monitoring of volatiles in the coffee flow (left).
Normalized time intensity profile for pyridine (blue) and the cumulative intensity (red) over the
extraction of a lungo capsule. Lines are the average of 10 capsules with ribbons representing the 95%
confidence level. ...................................................................................................................................................
52
Figure 18. Sampling set up used for sampling volatiles in semi-automatic coffee machine (left).
Normalized time-intensity profiles for the 5 differentiated families. Line represents the average of all
compounds in the family and the ribbon the 95% confidence interval. ......................................................
Figure 19. Concentration of volatiles in the headspace of full leaves or brews. Values represent the
average of 63 black teas and 38 green teas. VOCs were grouped in different chemical families. ............
Figure 20. Scheme of the off-line measurement of tea brewing with PTR-TOF-MS (left). Time
intensity profile for dimethylsulfide showing the average of three replicates and the standard
deviation. ................................................................................................................................................................
Figure 21. PCA containing all samples analyzed in the study (3 temperatures x 2 leaf sizes x 2 waters
x 10 time points x 3 replicates) ...........................................................................................................................
53
II.
List of abbreviations
APCI
Atmospheric Pressure Chemical Ionization
DHS
Dynamic Head Space
DIMS
Direct Injection Mass spectrometry
FAO
Food and Agriculture Organization
GC
Gas Chromatography
GC-O
Gas Chromatography Olfactometry
HC
Hollow Cathode
HCA
Hierarchical Cluster Analysis
HLC
Henry Law Constant
ICO
International Coffee Organization
MS
Mass Spectrometry
OAV
Odor Activity Value
OT
Odor Threshold
PA
Proton Affinity
PCA
Principal Component Analysis
PLS-DA
Partial Least Square Discriminant Analysis
PTR
Proton Transfer Reaction
QSPR
Quantitative Structure Property Relationship
REMPI
Resonant Enhanced Multiple Photon Ionization
SAFE
Solvent Assisted Flavor Extraction
SDE
Simultaneous Distillation Extraction
SHS
Static Head Space
SIFT
Selected Ion Flow Tube
SIM
Single Ion monitoring
SOTA
Self Organizing Three Algorithm
SPI
Single Photon Ionization
SPME
Solid Phase Micro Extraction
TDS
Temporal Dominance of Sensations
TOF
Time of Flight
UK
United Kingdom
54
UN
United Nations
USD
United States Dollar
VUV
Vacuum Ultraviolet
III.
List of publications for thesis work
Paper 1: José A. Sánchez-López, Ralf Zimmermann, Chahan Yeretzian. Insight into the Time-Resolved
Extraction of Aroma Compounds during Espresso Coffee Preparation: Online Monitoring by PTR-ToF-MS, Anal.
Chem. 86 (2014) 11696–11704.
Paper 2: José A. Sánchez López, Marco Wellinger, Alexia N. Gloess, Ralf Zimmermann, Chahan
Yeretzian. Extraction kinetics of coffee aroma compounds using a semi-automatic machine: On-line analysis by PTRToF-MS, Int. J. Mass Spectrom. 401 (2016) 22-30
Paper 3: Sine Yener, José A. Sánchez-López, Pablo M. Granitto, Luca Cappellin, Tilmann D. Märk,
Ralf Zimmerman, Günther K. Bonn, Chahan Yeretizian, Franco Biasioli. Rapid and direct volatile compound
profiling of black and green teas (camellia sinensis) from different countries with PTR-ToF-MS, Talanta 152 (2016) 4553
Manuscript 1: José A. Sánchez-López, Sine Yener, Tilmann D. Märk, Günther Bonn, Ralf Zimmerman,
Franco Biasioli , Chahan Yeretizian. Extraction Dynamics of Tea Volatile Compounds as a Function of Brewing
Temperature, Leaf Size and Water Hardness: On-Line Analysis by PTR-ToF-MS. Submitted to Talanta.
06.04.2016
IV.
List of further publications
Paper 4: José A. Sánchez-López, Aldo Ziere, Sara I.F.S. Martins, Ralf Zimmermann, Chahan Yeretzian.
Persistence of aroma volatiles in the oral and nasal cavity. Real-time monitoring of decay-rate in air exhaled through the nose
and mouth. Accepted for publication in Journal of Breath Research. 30.05.2016
55
V.
Publications
Publication 1.
Insight into the Time-Resolved Extraction of Aroma Compounds during Espresso Coffee
Preparation: Online Monitoring by PTR-TOF-MS
by
José A. Sánchez-López, Ralf Zimmermann, Chahan Yeretzian.
Analytical Chemistry
Volume 86, Issue 23, 2014, Pages 11696–11704.
DOI: 10.1021/ac502992k
José A. Sánchez-López designed and performed all the experiments. He also performed the data analysis
and prepared the manuscript. His work to this publication accounts for approximatelly 90%.
56
Article
pubs.acs.org/ac
Insight into the Time-Resolved Extraction of Aroma Compounds
during Espresso Coffee Preparation: Online Monitoring by
PTR-ToF-MS
José A. Sánchez-López,†,‡ Ralf Zimmermann,‡,§ and Chahan Yeretzian*,†
†
Zurich University of Applied Sciences, Institute of Chemistry and Biological Chemistry, 8820 Wädenswil, Switzerland
Joint Mass Spectrometry Centre, Chair of Analytical Chemistry, Institute of Chemistry, University of Rostock, D-18059 Rostock,
Germany
§
Joint Mass Spectrometry Centre, Cooperation Group Comprehensive Molecular Analytics/CMA, Helmholtz Zentrum München,
D-85764 Neuherberg, Germany
‡
ABSTRACT: Using proton-transfer-reaction time-of-flight
mass-spectrometry (PTR-ToF-MS), we investigated the
extraction dynamic of 95 ion traces in real time (time
resolution = 1 s) during espresso coffee preparation. Fiftytwo of these ions were tentatively identified. This was achieved
by online sampling of the volatile organic compounds (VOCs)
in close vicinity to the coffee flow, at the exit of the extraction
hose of the espresso machine (single serve capsules). Ten
replicates of six different single serve coffee types were
extracted to a final weight between 20−120 g, according to
the recommended cup size of the respective coffee capsule (Ristretto, Espresso, and Lungo), and analyzed. The results revealed
considerable differences in the extraction kinetics between compounds, which led to a fast evolution of the volatile profiles in the
extract flow and consequently to an evolution of the final aroma balance in the cup. Besides exploring the time-resolved
extraction dynamics of VOCs, the dynamic data also allowed the coffees types (capsules) to be distinguished from one another.
Both hierarchical cluster analysis (HCA) and principal component analysis (PCA) showed full separation between the coffees
types. The methodology developed provides a fast and simple means of studying the extraction dynamics of VOCs and
differentiating between different coffee types.
F
reaction time-of-flight mass spectrometry (PTR-ToF-MS)
allows volatile organic compounds (VOCs) to be quantified
and exhibits low ion fragmentation, high sensitivity, and high
time and mass resolution.20
In this piece of work, we will focus on the analysis of the
dynamic extraction of an espresso coffee using PTR-ToF-MS.
Coffee is a food product of great economic relevance and an
icon of western life style. The unique and highly appreciated
flavor of a cup of coffee is the final expression of a long chain of
chemical and physical transformations that link the seed to the
cup. The genetic makeup (the variety), agronomic practices,
the soil, climatic conditions, and the care given by the farmers
set the stage for the later development of the typical coffee
flavor. The flavor of unroasted coffee does not bear any
resemblance to what is considered the typical flavor of coffee.
Roasting generates around 1000 VOCs, although less than
50 might be relevant to the aroma of roasted coffee.21 Roasting
is the most important step for the formation of the coffee
aroma, and hence it is also one of the most thoroughly studied
ood flavor is a highly complex phenomenon and the
strategies and technologies used to elucidate perceived
flavors are becoming increasingly sophisticated, requiring
multidisciplinary approaches.1 With more than 7000 flavor
compounds reported in food products to date,2 the unequivocal
identification and quantification of these compounds is a crucial
step in flavor analysis. The ability to separate compounds by gas
chromatography (GC), to identify them by comparison with
mass spectral reference libraries and to quantify them using
standard compounds makes GC/MS an indispensable technique for flavor scientists. Coupled with olfactory techniques
such as GC-olfactometry (GC/O), these approaches allow sensory relevant compounds to be elucidated, and their relative
contributions to the flavor of the food product to be
estimated.3,4 While GC/MS is highly suitable for identifying
and quantifying flavor-active compounds, it performs less well
when it comes to monitoring the temporal evolution of fast
dynamic processes and needs to be complemented with other
analytical techniques when processes such as flavor generation5−11 or in vivo release12−17 need to be monitored. This has
led to the introduction of new analytical technologies capable of
monitoring volatiles in real-time, including electronic sensors,18
and direct injection mass spectrometry.19 Among the various
techniques for direct injection techniques, proton transfer
© 2014 American Chemical Society
Received: August 11, 2014
Accepted: November 5, 2014
Published: November 5, 2014
11696
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
57
Analytical Chemistry
Article
Table 1. Characterization of the Coffee Capsules
blend/origins
particle size
capsule
type
arabica
robusta
RF
EA
EC
EI
LC
LF
Central America + Africa
Central + South America
Central + South America
Central America + India
Central+South America+Asia
Central+South America
India
India
India
Asia
powder weight
(g)
roast degree
(Pt)
±
±
±
±
±
±
69
77
88
67
98
73
6.01
5.95
6.24
6.00
6.30
6.04
0.17
0.11
0.09
0.15
0.02
0.07
processing steps.5−7 However, equally important to the final
flavor profile in the cup is the extraction of ground coffee with
water. The extraction technique and conditions used for coffee
preparation strongly influence the flavor profile in the cup22
and is often the only parameter that can be influenced by the
consumer at home.
Several studies have investigated how the extraction of flavor
compounds is affected by the brewing technique,22,23 temperature,24−26 pressure,26,27 water composition,28,29 and cup
length.30,31 In all of these studies, measurements were carried
out on the final extract, but there is a lack of information on how
the above-mentioned parameters affect the kinetics of extraction.
Few quantitative studies have been published to date on the timeresolved extraction of volatile coffee compounds. By using
different volumes of water in the extraction process or taking
fractions over the whole extraction time/volume, some authors
have published findings on the extraction process of acrylamide,
caffeine, and antioxidants.30,31 To the best of our knowledge, only
the recently published work by Mestdagh et al. has reported data
on the kinetics of extraction for selected aroma compounds, using
solid phase micro extraction (SPME)-GC/MS.32
The approach taken here examines whether it is possible to
measure VOC release from the coffee flow at the exit of the
extraction hose using PTR-ToF-MS. We make the assumption
that each compound in the liquid extract is partitioned in the
gas phase, so that the gas phase concentration of VOCs at the
exit of the hose is proportional to the liquid concentration, with
the Henry’s Law Constant (HLC) being the proportionality
constant.33,34 Hence the time-evolution in the gas-phase mimics
the extract concentration. An analytical approach that is based
on online sampling of the volatiles released from the coffee flow
was developed and tested for real-time monitoring of the
extraction of volatile aroma compounds from single serve coffee
capsules.
d3,2 (μm)
±
±
±
±
±
±
55 ± 18
47 ± 8
57 ± 1
48 ± 5
49 ± 6
56 ± 10
345
330
346
331
343
341
53
5
3
32
20
58
extraction time
(s)
14.2
28.9
22.0
23.5
41.1
42.0
±
±
±
±
±
±
0.6
0.8
0.0
0.8
0.5
0.4
extracted weight
(g)
20.06
47.21
62.10
49.81
111.64
117.80
±
±
±
±
±
±
1.37
1.57
1.05
1.40
2.81
2.27
machine (Delica, Birsfelden, Switzerland). These were operated
according to the factory settings to pump three different
volumes of water in an unrestricted mode (no capsule in the
brewing unit): 40 mL for Ristretto, 72 mL for Espresso and
131 mL for Lungo. Depending on the coffee inside each
capsule type, the actual weight of the extract in the cup (final
column in Table 1) showed significant variations, but was very
stable for repetitions of the same type. The total time for
extraction of the cup and its final weight were measured
(Table 1). Note that the expression “espresso” can have two
meanings. Either it describes the general fact that coffee was
prepared using a pressurized brewing/extraction process and
may refer to different extracted volumes (Ristretto, Espresso,
and Lungo) or it designates the volume of the extract (here,
50 mL), the context clarifies the meaning. Just before extraction
of each capsule, 110 mL of water was passed through the circuit
to remove possible residues from the previous extraction and to
preheat the circuit. Both for cleaning and extraction, tap water
was mixed with filtered water (PURITY 600 Quell ST, BRITA
Professional, Taunusstein, Germany) to adjust the extraction
water to Alkalinity 4 dH° ± 1° dH, Hardness 6 dH° ± 1° dH
(German water hardness).
Sampling Set Up. VOCs released from the coffee flow
were measured with the set up shown in Figure 1. Coffee was
extracted over an ice-cold water bath to ensure that interference
from volatiles from the collected extract were eliminated. The
sampling lance was positioned 0.5 cm from the coffee flow and
coupled to the inlet of the PTR-ToF-MS. Using a custom built
gas dilution system, adapted from Wellinger et al.,35 we diluted
the sampled VOCs 7.5-fold to avoid condensation of VOCs on
the tubing and to adjust their concentration to within the dynamic
working range of the mass spectrometer. The dilution gas was
dry compressed air containing 2-isobutyl-3-methylpyrazine as
a standard for mass range calibration. All the sampling and
dilution lines were heated to 90 °C and all flows were controlled
by mass flow controllers (Bronkhorst, Ruurlo, The Netherlands)
and verified using a bubble meter.
PTR-ToF-MS. A commercial PTR-ToF-MS 8000 instrument
(Ionicon Analytik GmbH, Innsbruck, Austria) was used. The
diluted sample was introduced with a flow of 200 sccm into
the drift tube, which was operated at 2.2 mbar, 70 °C and
600 V drift voltage. PTR-ToF-MS data were recorded by
TOFDAQ v.183 data acquisition software (Tofwerk AG, Thun,
Switzerland). Mass spectra were recorded in the mass-to-charge
(m/z) range of 0−205 with one mass-spectrum recorded per
second. Mass axis calibration was performed on [H318O]+,
[C3H7O]+, and [C813CH15N2]+.
Data Processing. A PTR-TOF DATA Analyzer software v4.1736 was used for data analysis. Duty cycle corrected
signals were normalized to 106 H3O+ primary ions. During
extraction, fluctuations in the flow (mL/s) and the foam
■
MATERIALS AND METHODS
Coffee. Six commercial Delizio coffee capsule types (Delica,
Birsfelden, Switzerland) were selected: Ristretto Forte (RF),
Espresso Intenso (EI), Espresso Alba (EA), Espresso Classico
(EC), Lungo Fortissimo (LF), and Lungo Crema (LC). All the
capsules of a given type were from the same production batch.
To ensure reproducibility among the same types of capsules
and to compare different types of capsules, the coffee powder in
the capsules was characterized according to (i) total weight in
the capsule; (ii) roasting degree, measured with a Colorette 3B
instrument (Probat, Emmerich am Rhein, Germany); and (iii)
particle size distribution, measured with a Mastersizer 2000
(Malvern Instruments, Worcestershire, UK). The results are
summarized in Table 1.
Coffee Preparation. Ten different capsules of each coffee
type were extracted using a Delizio Compact Automatic coffee
11697
58
d4,3 (μm)
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
Analytical Chemistry
Article
differences were calculated using ANOVA and Tukey’s test (p <
0.05); these numbers are provided in (Table 2). Furthermore,
to examine differences in the total amount of extracted VOCs
the area under the curve of the time-intensity profiles was also
calculated (numerical integration in discrete time-intervals of
one second, corresponding to the integration time window
during data collection) and subjected to statistical analysis.
Statistical Analysis. Three different areas under the timeintensity profiles for each of the 95 mass traces were calculated:
(i) the area under the time intensity curve from t = 0 to 15 s;
(ii) the area from t = 0 s to the end of the extraction (total
extraction time, which depends on capsule type); and (iii) the
area calculated under point ii, normalized/divided by the
amount (in grams) of the extracted coffee.
These three sets of 60 samples (10 replicates of 6 different
capsule types) with 95 different variables were subsequently
subjected to statistical analysis. Hierarchical Cluster Analysis
(HCA) was performed by Ward’s minimum variance method
using half-squared Euclidean distances. Principal Component
Analysis (PCA) was performed on mean-centered scaled data.
All analysis and graphs were performed with packages and
scripts in R (R foundation for statistical computing, Vienna,
Austria).
Figure 1. Set up for sampling VOCs from the coffee flow. Volatiles
were introduced into the dilution lancet by a flow created with a
vacuum pump and were then diluted 7.5 fold using dried compressed
air containing a standard for mass calibration.
■
RESULTS AND DISCUSSION
The time-intensity profiles show different extraction dynamics
for the VOCs analyzed (Figure 2A). The time at which the
maximum intensity was reached ranged from 2 to 24 s,
although for 95% of the compounds it was reached in less than
10 s. Once the maximum had been achieved, the intensity fell at
different rates, depending on the compound. This decrease of
intensity provides information on how the compounds are
extracted. A fast decrease implies that the compound is
extracted over a relatively short time period while a slow
decrease implies that the compound is extracted over a longer
period. Using t1/2 as a measure of the intensity decrease, we
observe a large variability between the different VOCs,
encompassing a range of 3 to 25 s for t1/2. A few compounds
did not fall below 50% of the maximum intensity by the end of
the extraction and hence their t1/2 could not be determined.
Although the extraction of some compounds was relatively
slow, 70% of them reached t1/2 in less than 10 s and showed
intensities lower than 20% of the maximum by the time that the
coffee had been prepared (∼24 s).
Plotting the integrated intensity of the time-intensity curves
for each VOC, we obtained the cumulative concentration of the
(different bubble sizes) were observed. To correct for small differences in the absolute intensity and allow for a better
comparison between capsules, the intensity of the VOCs was
normalized to the maximum intensity of the m/z 69.035 ion
trace, before averaging for replicates.
Mass Peaks Selection. Ten replicates for each of the six
coffee capsule types (RF, EA, EC, EI, LC, and LF) were
analyzed with the set up described in the Sampling Set Up
section. Around 300 mass peaks were found in the m/z range
recorded, although the exact number was dependent on the
capsule type. Only peaks that changed over time and that were
present in all samples were included in the subsequent data
analysis, yielding a list of 95 ion traces. Out of these, 52 were
tentatively identified, based on the literature and were reduced
to 47 after removing fragments and isotopologues.37−39
Each m/z time-intensity profile was characterized using the
following parameters: (i) the time at maximum intensity (tmax),
(ii) the time elapsed between the maximum intensity and the
drop to half of the maximum intensity (t1/2) and significant
Figure 2. Time intensity profiles in the LC capsule showing differences in extraction. (A) Data normalized to the maximum intensity of four m/z.
Integration of the area under the curve at each time point as a percentage of the total area at the end of the extraction for (B) the four selected m/z
and (C) for all peaks considered. Shaded ribbons show the 95% confidence interval. Colors in panel C represent different peaks.
11698
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
59
60
theoretical
(m/z)
31.018
33.033
45.033
47.013
55.054
57.033
57.07
59.049
61.028
63.026
68.049
69.033
70.04
71.049
72.044
73.065
75.044
80.049
82.065
83.049
85.065
87.044
87.08
89.06
95.06
97.028
99.044
101.06
103.075
107.06
109.076
110.06
111.044
113.06
115.075
117.055
measured
(m/z)
31.018
33.033
45.034
47.013
55.055
57.035
57.071
59.05
61.029
63.027
68.051
69.035
70.039
71.051
72.046
73.066
75.045
80.052
82.068
83.052
85.067
87.046
87.083
89.062
95.06
97.032
99.045
101.062
103.078
107.056
109.079
11699
110.06
111.045
113.063
115.078
117.052
C5H9O3+
C6H11O2+
C6H7O2+
C6H9O2+
C6H8NO+
C6H7N2+
C6H9N2+
C3H6NO+
C4H9O+
C3H7O2+
C5H6N+
C5H8N+
C5H7O+
C5H9O+
C4H7O2+
C5H11O+
C4H9O2+
C5H7N2+
C5H5O2+
C5H7O2+
C5H9O2+
C5H11O2+
CH3O+
CH5O+
C2H5O+
CH3O2+
C4H7+
C3H5O+
C4H9+
C3H7O+
C2H5O2+
C2H7S+
C4H6N+
C4H5O+
C2H4N3+
C4H7O+
sum
formula
tentative identification
formaldehyde
methanol
acetaldehyde
formic acid
1,3-butadiene
2-propenal, prop-1-en-1-one
1-butene
acetone, propanal
acetic acid
dimethyl sulfide
pyrrole
furan
triazole
methyl-propenal, 3-buten-2one
acrylamide
methyl propanal
propanoic acid, ethyl acetate
pyridine
methyl pyrrole
methyl furan
methylbutenal
2,3-butanedione
methylbutanal
methylpropanoate
methylpyrazine
furfural
furfuryl alcohol
pentanedione
hidroxypentanone, methyl
butanoic acid
ethenylpyrazine
dimethylpyrazine,
ethylpyrazine
acetylpyrrole, methylpyrrolyl ketone
acetylfuran
methylfurfuryl alcohol,
dimethylfuranone
4-methyltetrahydro-2Hpyran-2-one
2-oxopropyl acetate, acetol
acetate
1a
1ab
1a
1b
1b
1a
1a
1b
1bc
1a
1b
1a
1b
1a
1b
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
5 ± 1n
3 ± 1n
4 ± 1n
3 ± 1o
4 ± 1n
7 ± 1b
7 ± 1b
6 ± 1bc
7 ± 1b
6 ± 1a
1mn
1o
1o
1n
1no
1o
1n
1no
1o
1no
1n
1no
1n
1no
1n
2 ± 1o
4 ± 1n
5
5
4
5
3
3
4
3
3
4
4
4
4
4
4
3 ± 1b
7 ± 1bc
4
4
4
7
7
7
6
5
7
5
6
6
6
6
6
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
1op
1o
1op
1m
1no
1no
1n
1o
1o
2mo
1n
1n
1n
1no
t1/2 (s)
4
4
4
8
5
4
4
4
4
4
4
4
3
1
1a
1a
1a
1a
1ab
0b
1ab
1a
1a
2a
1a
1a
1b
1a
±
±
±
±
±
±
±
±
±
±
±
±
±
±
3
3
4
4
4
5
7
3
3
6
6
6
6
6
tmax (s)
RF
0a
0a
1a
1a
2b
1ab
3b
1a
1a
2a
2a
1a
2ab
2a
6 ± 2a
7 ± 3b
7 ± 2b
8 ± 2c
9 ± 1c
9 ± 2c
8 ± 2c
6 ± 2b
6 ± 2b
4 ± 1a
9 ± 1c
9 ± 2c
8 ± 2a
6 ± 2a
5 ± 1ab
10 ± 1d
5 ± 1a
6 ± 2ab
6 ± 2a
6 ± 2ab
6 ± 2a
5 ± 1ab
±
±
±
±
±
±
±
±
±
±
±
±
±
±
tmax (s)
3
3
4
4
6
5
8
3
4
6
6
5
6
6
EA
7 ± 2m
9 ± 3mn
8 ± 2mo
8 ± 2m
8 ± 2mo
12 ± 2n
8 ± 2mo
5 ± 2mn
5 ± 2mo
5 ± 1mo
9 ± 2mo
7 ± 2mn
14 ± 2m
7 ± 2m
4 ± 1o
5 ± 2no
6 ± 1mo
7 ± 2m
6 ± 2mo
6 ± 2m
6 ± 2mo
6 ± 1o
6 ± 1mp
4 ± 1o
5 ± 1mp
11 ± 2n
6 ± 2mo
5 ± 1mo
14 ± 3m
4 ± 1o
4 ± 1o
9 ± 2n
6 ± 2mn
7 ± 1m
6 ± 2m
6 ± 2mo
t1/2 (s)
Table 2. List of Mass Peaks, Assigned Sum Formulae, and Tentative Compound Identificationa
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
1a
1a
1a
1a
1a
3a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
4 ± 1a
4 ± 1a
5 ± 1a
4 ± 1a
5 ± 1a
6 ± 1a
5 ± 1a
4
4
4
5
4
8
4
4
4
4
4
4
4
4
4
±
±
±
±
±
±
±
±
±
±
±
±
±
±
tmax (s)
3
3
4
4
4
4
5
4
4
4
4
4
4
4
EC
t1/2 (s)
8 ± 1m
12 ± 2m
9 ± 1m
9 ± 1m
9 ± 1m
13 ± 2mn
9 ± 1m
7 ± 1m
7 ± 1m
6 ± 1m
10 ± 1m
10 ± 1m
14 ± 3m
7 ± 1m
6 ± 1m
10 ± 2m
7 ± 1m
7 ± 1m
7 ± 1m
8 ± 1m
7 ± 1m
8 ± 1m
6 ± 1m
6 ± 1m
7 ± 1m
9 ± 1mn
7 ± 1m
7 ± 1m
15 ± 2m
6 ± 1m
6 ± 1m
8 ± 1mn
7 ± 1m
7 ± 1m
7 ± 1m
7 ± 1m
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
0a
1ab
1a
1a
1ab
2a
1a
1ab
1ab
1a
1ab
1a
1ab
1a
1ab
1a
1a
1a
1a
1ab
1ab
1a
1a
1a
1a
1a
1a
1ab
1a
5 ± 1a
5 ± 1ab
5 ± 1ab
5 ± 1ab
6 ± 1ab
5 ± 1a
6 ± 1ab
4
5
4
6
5
6
5
5
5
5
5
5
5
5
5
±
±
±
±
±
±
±
±
±
±
±
±
±
±
tmax (s)
4
4
4
4
5
5
5
4
4
5
5
5
5
5
EI
1no
1no
1n
2m
1n
1n
2n
1n
1n
2o
1n
1n
1n
1n
3 ± 1n
4 ± 2no
3 ± 2n
3 ± 1n
4 ± 2n
7 ± 2m
4 ± 2no
3 ± 1n
2 ± 1n
2 ± 1n
5 ± 2no
5 ± 2mn
14 ± 2m
2 ± 2n
2 ± 1n
8 ± 2mn
3 ± 1n
3 ± 1n
2 ± 1n
2 ± 1n
2 ± 1n
3 ± 1n
±
±
±
±
±
±
±
±
±
±
±
±
±
±
t1/2 (s)
3
2
2
7
3
2
7
2
2
3
3
3
2
1
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
1a
1ab
1a
1a
1a
1a
1a
1ab
1abc
1a
1ab
1a
1ab
1a
1ab
1a
1a
1a
1a
1ab
1ab
1a
1a
1a
1a
1a
1a
1ab
1a
5 ± 1a
5 ± 1ab
5 ± 1ab
5 ± 1ab
6 ± 1ab
6 ± 1a
5 ± 1ab
5
5
4
5
5
6
5
4
6
5
5
5
5
5
5
±
±
±
±
±
±
±
±
±
±
±
±
±
±
tmax (s)
4
4
4
4
4
5
6
4
4
4
5
5
5
5
LC
t1/2 (s)
3 ± 1n
7 ± 2mo
4 ± 1no
5 ± 1n
5 ± 1n
9 ± 1mn
5 ± 1no
4 ± 1n
3 ± 1no
2 ± 1n
6 ± 1no
2 ± 1o
10 ± 2n
3 ± 1n
2 ± 1n
8 ± 2mn
3 ± 1n
4 ± 1n
3 ± 1n
4 ± 1n
3 ± 1n
4 ± 1n
3 ± 1n
2 ± 1n
3 ± 1no
3 ± 1o
4 ± 1n
2 ± 1n
13 ± 2m
2 ± 1n
2 ± 1n
6 ± 2mno
3 ± 1n
3 ± 1n
3 ± 1n
1 ± 1n
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
1a
1ab
1a
1ab
1ab
2a
1a
1ab
2c
1a
1ab
1a
1ab
1a
1ab
1a
1a
1a
0a
1ab
1ab
1ab
1a
1a
1a
1a
1a
1ab
1a
6 ± 1a
6 ± 1ab
5 ± 1ab
6 ± 1ab
6 ± 1ab
6 ± 2a
6 ± 1ab
4
5
4
6
6
7
5
4
7
5
5
5
5
5
5
±
±
±
±
±
±
±
±
±
±
±
±
±
±
tmax (s)
4
4
4
4
4
5
6
4
4
5
5
5
5
5
LF
1no
1n
1no
1m
1no
1n
2n
1n
1n
2mno
1n
1n
1n
1no
2 ± 1n
6 ± 1no
4 ± 1n
4 ± 1n
5 ± 1n
7 ± 2m
6 ± 1no
4 ± 1n
3 ± 1no
3 ± 1n
6 ± 1no
5 ± 1no
13 ± 2mn
4 ± 1n
2 ± 1n
5 ± 2no
3 ± 1n
4 ± 1n
3 ± 1no
4 ± 1n
3 ± 1no
3 ± 1n
±
±
±
±
±
±
±
±
±
±
±
±
±
±
t1/2 (s)
3
2
3
7
5
2
8
2
2
5
3
4
3
1
Analytical Chemistry
Article
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
17 ± 3a
16 ± 2m
5 ± 1ab
11 ± 4ab
7 ± 1a
15 ± 3a
3 ± 2n
1ab
1a
1ab
1ab
±
±
±
±
6
6
5
6
2m
2no
1n
2no
±
±
±
±
9
7
2
6
21 ± 1bc
5 ± 1ab
17 ± 5c
6 ± 2a
20 ± 2b
9 ± 1m
1a
1a
1ab
1a
±
±
±
±
6
5
4
6
14 ± 3m
12 ± 2m
8 ± 1m
12 ± 2m
tmax (s)
11 ± 6ab
7 ± 2a
15 ± 3m
13 ± 2n
t1/2 (s)
17 ± 3a
17 ± 3m
4 ± 1a
11 ± 4ab
6 ± 1a
15 ± 4a
8 ± 2mo
1a
1a
1a
1a
±
±
±
±
6
5
4
5
13 ± 2m
11 ± 2mo
4 ± 2n
9 ± 2mo
tmax (s)
7 ± 3a
7 ± 2a
14 ± 3m
14 ± 2m
t1/2 (s)
24 ± 3c
6 ± 2n
7 ± 2b
15 ± 5bc
10 ± 3b
22 ± 3b
4 ± 1n
2c
2b
1b
2c
±
±
±
±
9
9
5
9
4
4
3
4
1n
1n
1n
1n
±
±
±
±
±
±
±
±
tmax (s)
12 ± 3b
10 ± 2b
t1/2 (s)
5 ± 1n
5 ± 1n
tmax (s)
8 ± 1ab
8 ± 1a
10 ± 1d
6 ± 1ab
11 ± 2a
7 ± 2a
9 ± 1c
1b
1a
1a
1bc
7
7
3
7
165.091
165.1
C10H13O2+
131.07
135.092
137.071
149.107
131.074
135.096
137.067
149.113
C6H11O3+
C8H11N2+
C7H9N2O+
C9H13N2+
124.076
125.06
127.039
127.075
124.083
125.065
127.038
127.08
C7H10NO+
C7H9O2+
C6H7O3+
C7H11O2+
tentative identification
ethenylmethyl-pyrazine
ethyl-methylpyrazine,
trimethylpyrazine
acetylmethylpyrrole
guaiacol, methylbenzenediol
maltol, methylfuroate
ethylbenzenediol,
ethylcyclopentanedione
ethyl acetoacetate
ethylvinylcyclopentapyrazine
methyl-pyrazinylehtanone
dihydro-dimethyl
cyclopentapyrazine
allylguaiacol
C7H9N2+
C7H11N2+
121.076
123.092
121.075
123.096
For each coffee variety the time at maximum intensity (tmax) and time elapsed until the intensity drops to half of maximum intensity (t1/2) is shown. Data followed by different letters (a−d for tmax and
m−p for t1/2) are significantly different within the capsule tipes according to ANOVA (p < 0.05).
22 ± 2o
19 ± 4ab
22 ± 3o
6 ± 1ab
11 ± 3ab
7 ± 2a
17 ± 3ab
6 ± 2no
14 ± 3m
17 ± 2m
13 ± 4m
1ab
1a
1ab
2ab
±
±
±
±
6
7
4
6
2m
2no
2m
2no
±
±
±
±
9
8
9
7
tmax (s)
8 ± 2ab
7 ± 1a
12 ± 3m
9 ± 2m
t1/2 (s)
tmax (s)
t1/2 (s)
11 ± 3m
11 ± 3m
7 ± 2a
6 ± 1a
LC
EI
EC
EA
RF
sum
formula
theoretical
(m/z)
measured
(m/z)
Table 2. continued
VOC released from the flow at each time-point (Figure 2B).
The slope of these curves reflects the extraction rate.
Normalizing this data to the total amount of compound
extracted (intensity at the end of the extraction time was set to
100%), it is possible to compare the extraction behavior/rate of
the different compounds within a coffee capsule. For each
compound, and as a function of time, these curves represent the
extracted fraction with respect to the total amount in the final
cup. We can observe that, as a consequence of the different
extraction behavior of the different compounds over time, the
VOC profiles and ratios of aroma compounds in the samples
differ at each time point.
Extraction of single serve capsules is similar to espresso
extraction, where hot water at high pressure passes through the
ground coffee bed and results in an extract containing dissolved
compounds, suspended solid particles and emulsified oil and
foam. The high pressure at which the water is pumped through
the coffee makes espresso extraction much faster than other
coffee brew techniques (e.g., compared to filter coffee
extraction by gravitational force). A simple visual inspection
of the coffee flow out of an espresso machine, shows that the
color of the extract becomes progressively lighter with
extraction time. This indicates that most of the colored
compounds are extracted at the beginning of the extraction, in
the first few seconds (first few milliliters). The same happens
with the VOCs, although it is expected that VOC extraction is
even faster and occurs more quickly than the colored, higher
molecular weight compounds. Extraction of VOCs mostly
occurs at the very beginning of the espresso extraction,
resulting in an intense signal at the start of the time intensity
profile, which is expressed as a steep slope on the integrated
curve. Our results agree with those of Mestdagh et al.,32 who
extracted Nespresso coffee capsules stepwise with increasing
volumes of extracts, from 10 mL up to 150 mL, and quantified
20 flavor active VOCs using GC-MS and isotopically labeled
standards. Despite the variance associated with the use of
different capsules for each volume point and the low time
resolution (six points for 150 mL), they were able to describe
the kinetics of extraction for 20 compounds and found some
correlation between the polarity of the compound and
extraction efficiency: more polar compounds were extracted
faster. The same behavior was observed by Ludwig et al.30
for nonvolatile compounds, such as caffeine, 3-, 4-, and
5-caffeoylquinic acids. They found that 70% of these
compounds were extracted in the first 8 s, while only 50% of
the total 3,4-,3,5- and 4,5-dicaffeoylquinic acids were extracted
in the same 8 s time window, showing slower rates during the
whole process of making an espresso coffee. Diccaffeoylquinic
acids are less polar than monocaffeolyquinic acids and have
stronger chemical interactions with melanoidins, due to
potential esterification. This suggests that not only polarity
but also possible interactions with other polymers present in
the coffee powder modulate the rate of extraction of the
different compounds.
Besides the differences in the extraction dynamics between
the VOCs for each coffee, differences in individual compounds
for different capsule types were also apparent. Figure 3 shows
the integrated time-intensity profiles of two selected compounds as an example for this observation, methylbutanal and
pyridine. Both the slopes and the final intensities are different
for each capsule type with only pyridine exhibiting the same
profile for EC and LC.
a
4 ± 1n
23 ± 3o
17 ± 2m
10 ± 4mo
2mn
2no
2n
2no
±
±
±
±
9
7
3
7
t1/2 (s)
Article
LF
13 ± 3m
9 ± 2m
Analytical Chemistry
11700
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
61
Analytical Chemistry
Article
Figure 3. Integrated intensity over time for pyridine and methylbutanal for the six coffee types analyzed. The two upper graphs show the
accumulative intensity over the whole extraction time. The lower graphs show the dynamics of extraction during the first 15 s: the accumulated
intensity for each coffee at 15 s is considered 100%. Shaded ribbons represent the 95% confidence interval.
Espresso extraction is affected by two different sets of
parameters, those related to the water, such as temperature,
pressure, and mineralization contentand those related to the
coffee bedsuch as dose, particle size distribution, compressing force, blend, and roast. In this case, parameters regarding
the water flowing through the coffee were kept constant. Thus,
observed differences can be linked to differences in the coffee
powder inside the capsules. The particle size distribution and
the amount of coffee in the samples were measured for all
capsules; they turned out to be very similar for all capsules and
capsule types. Hence, this could not account for any of the
observed differences. The coffees extracted were blends from
different origins and species/varieties and were roasted to
different roasting degrees, which leads to the formation of
different amounts and profiles of VOCs, depending on the
blends and the roasting condition. Consequently, it is expected
that the differences observed between the various types of
capsules, were related to different initial concentrations of
compounds in the coffee powder. To corroborate this
hypothesis, we checked how the compounds were extracted
in the first 15 s (∼30 mL), before the compounds are exhausted
in the coffee powder. To allow for the comparison between
capsules of different concentrations (different blends and roast
degrees), we normalized the compound extracted to its
accumulated value after 15 s (Figure 3). We can clearly
observe that the slopes for the different coffee types are very
similar for both methylbutanal and pyridine. Nevertheless,
some differences can be observed, such as for methylbutanal in
RF. In this case, methylbutanal reached the plateau before 15 s,
suggesting that most of the methylbutanal in the powder had
already been extracted by that time. By normalizing the
integrated intensities at 10 s (instead of 15 s), the slope of RF
becomes the same as for the other capsule types (data not
shown), showing the same rate of extraction for all the capsules,
relative to the total amount of extracted VOCs in that time
window. These data suggest that, when comparing the
extraction kinetics of different coffee types for selected VOCs,
these extraction kinetics are essentially identical for all capsule
types, as long as the VOC has not been depleted from the roast
and ground coffee bed.
The dynamic time-intensity data discussed above provide
insights into the extraction rates of the different compounds in
the coffee capsules. However, besides exploring the extraction
dynamics, the data were also used to distinguish between coffee
types, by means of statistical analysis. Three approaches were
used: (i) integration up to 15 s, (ii) integration over the whole
extraction time, and (iii) integration over the whole extraction
time, but normalized to the amount of coffee extracted
(division by the total weight of the final cup). For each
approach, HCA and PCA were performed.
Fifteen Seconds. The shortest to prepare coffee included
in this study was RF (30 mL) with a total extraction time of
11701
62
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
Analytical Chemistry
Article
Figure 4. Hierarchical clustering and score plots for the first three dimensions of PCA of the six capsule varieties using the integrated area at
15 s (A, B, C), full extraction time (D, E, F), and full time corrected by weight of extracted coffee (G, H, I).
15 s. We selected this specific time window for the first
comparison between capsules since all the capsule types were
extracted for at least 15 s, allowing direct comparison of the
time-intensity profiles. HCA (Figure 4A) showed good
separation for four out of the six capsules (RF, EA, EC, and
LC); each of these four clusters exclusively contain the ten
repetitions for each capsule type. Only EI and LF could not be
separated into individual clusters. PCA provided similar
information. The first principal component (66.4% of the
total variance) could only separate RF from the rest, but better
separation was obtained for the second and third components.
The PCA showed that, except for EI and LF, all the other
capsule types could be separated on the plots for the first three
principle components. It also showed that EC and LC are close
to each other on the plots for the three first components of
the PCA.
Since all the capsules were extracted for at least for 15 s,
differences between types of capsules in HCA and PCA are
indicative of differences between the coffees (i.e., coffee
varieties, blend, roasting degree) used to manufacture each of
the capsule types. Our results suggest that the coffees used for
EI and FL are similar and therefore appear close on the PCA
plots. A similar situation is observed for EC and LC. Checking
the capsule characteristics, it was possible to see that both
EI and FL contained more coffee powder (around 6.25 g) than
the others (around 6.00 g). Furthermore, together with RF,
they had the darkest roasting degrees of those included in this
study. EC and LC showed the lightest roasting of all the coffees.
Roasting is one of the key factors that affect the final coffee
aroma profile and is most probably responsible for the observed
clustering, although it is not the only factor at play. The blend
used for each coffee also impacts the compounds formed during
roasting, as the aroma precursors differ. Since the origins of the
coffees used for each blend are only known based on the
manufacturer’s general descriptions, observed similarities
cannot be related to the specific composition of the blends.
Full Time Extraction. When the time-intensity profiles are
integrated over the full extraction time (which varies between
the different coffee types), HCA is able to separate all replicates
of each coffee type into six individual clusters (Figure 4D).
PCA analysis also shows total separation of the six capsule types
on the plots for the first three principle components (Figures 4E
and 4F). Integration over the full extraction time allowed the
EI and LF coffees to be separated, which was not possible when
integrating over the first 15 s.
By integrating the whole area under the time intensity
profiles, it was possible to obtain a value proportional to the
total amount of the extracted compound in the cup. The
extraction of a Lungo takes approximately 20 s longer than for
an Espresso, and during that extra time some compounds are
still in the process of being extracted, resulting in better
separation in the HCA and PCA plots. Although the main
driving force for separation is the difference in extraction
time, compounds with identical extraction times can also be
11702
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
63
Analytical Chemistry
Article
Notes
separated from each other. These results indicate that, when
extracted according to manufacturer recommendations, the
amount and ratio of the VOCs in the final product is different
for all six different capsule types.
Final Concentration. Full extraction−time integration
reflects the total amount of each compound extracted, but
does not account for the dilution factor due to different cup
volumes. Hence it is a measure of the total amount in the
cup, but does not reflect the volatile profiles above the cup
(the headspace). As shown, the majority of the compounds
are extracted during the first seconds of the coffee extraction
process. As extraction evolves, the remaining amounts of the
compounds in the coffee bed decrease, and their concentrations
in the extract decrease as well. To get data that is closer to the
concentration in the final cup (and to the HS), the results for
the amount of coffee extracted were normalized. One of the
advantages of PTR-MS is that the signal is proportional to the
measured concentration,40 and therefore, the data could be
easily corrected for dilution by dividing the total amount of
compound extracted (integrated area over the total time of
extraction) by the weight of the final coffee extract. This result
reflects the concentration of each compound in the final coffee
and is comparable to headspace measurements for the final cup.
Both HCA and PCA showed good clustering for all the
capsules when the data are corrected for dilution. Therefore, we
can conclude that the aroma profile of the extracts and
consequently the HS profiles are clearly different for the six
capsule types investigated here.
The authors declare no competing financial interest.
■
■
CONCLUSIONS
We have presented a novel, high time-resolution methodology
for monitoring the extraction dynamics of espresso coffee and
applied it to six different capsules types. The results presented
in this work show the suitability of PTR-ToF-MS for
monitoring changes in the volatile composition of a liquid
flow in an open atmosphere. Online analysis of coffee
extraction revealed the kinetics of extraction for different
VOCs and highlighted the differences between commercial
coffee capsules over the whole extraction time. The presented
method overcomes the problems of previous GC-based
approaches: (i) it increases temporal information, from a few
data points over the whole extraction time to a one second
resolution, (ii) it reduces sources of variability, as the timeevolution of each VOC is monitored online in a single
extraction process and is not a combination of multiple
different extracts. The simplicity, high sensitivity and time
resolution of the method makes it a perfect approach for
investigating the impact of different parameters that affect
extraction dynamics of flavor compounds. On the basis of such
data, the process can be fine-tuned in order to achieve the
desired aroma balance in the final cup.
The methodology also allows the user to differentiate
between coffee types, by applying HCA and PCA on the
cumulated intensities of VOCs over specific time windows.
AUTHOR INFORMATION
Corresponding Author
*E-mail: chahan.yeretzian@zhaw.ch.
Author Contributions
All authors have given approval to the final version of the
manuscript
11703
64
REFERENCES
(1) Blank, I.; Wust, M.; Yeretzian, C. J. Agric. Food Chem. 2009, 57,
9857−9859.
(2) Chemistry and Technology of Flavors and Fragrances; Blackwell
Publishing Ltd.: Oxford, U.K., 2004.
(3) Semmelroch, P.; Grosch, W. LWTFood Sci. Technol. 1995, 28,
310−313.
(4) Chin, S. T.; Eyres, G. T.; Marriott, P. J. J. Chromatogr. A 2011,
1218, 7487−7498.
(5) Dorfner, R.; Ferge, T.; Yeretzian, C.; Kettrup, A.; Zimmermann,
R. Anal. Chem. 2004, 76, 1386−1402.
(6) Gloess, A. N.; Vietri, A.; Wieland, F.; Smrke, S.; Schönbächler, B.;
López, J. A. S.; Petrozzi, S.; Bongers, S.; Koziorowski, T.; Yeretzian, C.
Int. J. Mass Spectrom. 2014, 365−366, 324−337.
(7) Wieland, F.; Gloess, A. N.; Keller, M.; Wetzel, A.; Schenker, S.;
Yeretzian, C. Chimia 2012, 66, 443.
(8) Blank, I.; Devaud, S.; Matthey-Doret, W.; Pollien, P.; Robert, F.;
Yeretzian, C. In Proceedings of the 10th Weurman Flavour Research
Symposium; INRA−Institut National de la Recherche Agronomique:
Paris, 2002.
(9) Yeretzian, C.; Jordan, A.; Badoud, R.; Lindinger, W. Eur. Food
Res. Technol. 2002, 214, 92−104.
(10) Dorfner, R.; Ferge, T.; Kettrup, A.; Zimmermann, R.; Yeretzian,
C. J. Agric. Food Chem. 2003, 51, 5768−5773.
(11) Pollien, P.; Lindinger, C.; Yeretzian, C.; Blank, I. Anal. Chem.
2003, 75, 5488−5494.
(12) Roberts, D. D.; Pollien, P.; Antille, N.; Lindinger, C.; Yeretzian,
C. J. Agric. Food Chem. 2003, 51, 3636−3642.
(13) Romano, A.; Cappellin, L.; Ting, V.; Aprea, E.; Navarini, L.;
Gasperi, F.; Biasioli, F. Int. J. Mass Spectrom. 2014, 365−366, 20−27.
(14) Mayr, D.; Maerk, T.; Lindinger, W.; Brevard, H.; Yeretzian, C.
Int. J. Mass Spectrom. 2003, 223−224, 743−756.
(15) Roberts, D. D.; Pollien, P.; Lindinger, C.; Yeretzian, C. In
Handbook of Flavor Characterization: Sensory Analysis, Chemistry, and
Physiology; Deibler, K., Ed.; Dekker: New York, 2003; pp 151−162.
(16) Yeretzian, C.; Pollien, P.; Lindinger, C.; Ali, S. Compr. Rev. Food
Sci. Food Saf. 2004, 3, 152−159.
(17) German, J. B.; Yeretzian, C.; Tolstoguzov, V. B. In Flavours:
Chemistry, Technology and Resources; Berger, R. G., Ed.; Springer
Verlag: Berlin, Germany, 2007; pp 25−41.
(18) Gutiérrez, J.; Horrillo, M. C. Talanta 2014, 124, 95−105.
(19) Biasioli, F.; Yeretzian, C.; Märk, T. D.; Dewulf, J.; van
Langenhove, H. TrAC, Trends Anal. Chem. 2011, 30, 1003−1017.
(20) Jordan, A.; Haidacher, S.; Hanel, G.; Hartungen, E.; Maerk, L.;
Seehauser, H.; Schottkowsky, R.; Sulzer, P.; Maerk, T. D. Int. J. Mass
Spectrom. 2009, 286, 122−128.
(21) Czerny, M.; Mayer, F.; Grosch, W. J. Agric. Food Chem. 1999,
47, 695−699.
(22) Gloess, A. N.; Schönbächler, B.; Klopprogge, B.; D’Ambrosio,
L.; Chatelain, K.; Bongartz, A.; Strittmatter, A.; Rast, M.; Yeretzian, C.
Eur. Food Res. Technol. 2013, 236, 607−627.
(23) Parenti, A.; Guerrini, L.; Masella, P.; Spinelli, S.; Calamai, L.;
Spugnoli, P. J. Food Eng. 2014, 121, 112−117.
(24) Andueza, S.; Maeztu, L.; Pascual, L.; Ibanez, C.; de Pena, M. P.;
Cid, C. J. Sci. Food Agric. 2003, 83, 240−248.
(25) Albanese, D.; Di Matteo, M.; Poiana, M.; Spagnamusso, S. Food
Res. Int. 2009, 42, 727−732.
■
■
ACKNOWLEDGMENTS
The research leading to these results has received funding
through the PIMMS ITN, which is supported by the European
Commission’s seventh Framework Programme under grant
agreement number 287382. The authors also thank Dr. Marco
Wellinger and Dr. Alexia Glöss for fruitful discussions.
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
Analytical Chemistry
Article
(26) Caprioli, G.; Cortese, M.; Cristalli, G.; Maggi, F.; Odello, L.;
Ricciutelli, M.; Sagratini, G.; Sirocchi, V.; Tomassoni, G.; Vittori, S.
Food Chem. 2012, 135, 1127−1133.
(27) Andueza, S.; Maeztu, L.; Dean, B.; de Peña, M. P.; Bello, J.; Cid,
C. J. Agric. Food Chem. 2002, 50, 7426−7431.
(28) Navarini, L.; Rivetti, D. Food Chem. 2010, 122, 424−428.
(29) Hendon, C. H.; Colonna-Dashwood, L.; Colonna-Dashwood,
M. J. Agric. Food Chem. 2014, 62, 4947−4950.
(30) Ludwig, I. A.; Sanchez, L.; Caemmerer, B.; Kroh, L. W.; Paz de
Peña, M.; Cid, C. Food Res. Int. 2012, 48, 57−64.
(31) Alves, R. C.; Soares, C.; Casal, S.; Fernandes, J. O.; Oliveira, M.
B. P. P. Food Chem. 2010, 119, 929−934.
(32) Mestdagh, F.; Davidek, T.; Chaumonteuil, M.; Folmer, B.;
Blank, I. Food Res. Int. 2014, 63, 271−274.
(33) Staudinger, J.; Roberts, P. V. Crit. Rev. Environ. Sci. Technol.
1996, 26, 205−297.
(34) Pollien, P.; Jordan, A.; Lindinger, W.; Yeretzian, C. Int. J. Mass
Spectrom. 2003, 228, 69−80.
(35) Wellinger, M.; Biollaz, S.; Wochele, J. r.; Ludwig, C. Energy Fuels
2011, 25, 4163−4171.
(36) Müller, M.; Mikoviny, T.; Jud, W.; D’Anna, B.; Wisthaler, A.
Chemom. Intell. Lab. Syst. 2013, 127, 158−165.
(37) Flament, I. Coffee Flavor Chemistry; John Wiley & Sons, Ltd.:
Chichester, U.K., 2002.
(38) Chin, S.-T.; Eyres, G. T.; Marriott, P. J. J. Chromatogr. A 2011,
1218, 7487−7498.
(39) Yeretzian, C.; Jordan, A.; Lindinger, W. Int. J. Mass Spectrom.
2003, 223−224, 115−139.
(40) Lindinger, W.; Hansel, A.; Jordan, A. Chem. Soc. Rev. 1998, 27,
347−354.
11704
dx.doi.org/10.1021/ac502992k | Anal. Chem. 2014, 86, 11696−11704
65
Publication 2.
Extraction kinetics of coffee aroma compounds using a semi-automatic machine:
On-line analysis by PTR-ToF-MS
by
José A. Sánchez-López, Marco Wellinger, Alexia N. Gloess, Ralf Zimmermann, Chahan Yeretzian.
International Journal of Mass Spectrometry
Volume 401, 2016, Pages 22–30.
DOI: 10.1016/j.ijms.2016.02.015
José A. Sánchez-López was involved in the design and execution of the experiments. He performed the
data analysis and prepared the manuscript. His work to this publication accounts for approximatelly 80%.
66
International Journal of Mass Spectrometry 401 (2016) 22–30
Contents lists available at ScienceDirect
International Journal of Mass Spectrometry
journal homepage: www.elsevier.com/locate/ijms
Extraction kinetics of coffee aroma compounds using a
semi-automatic machine: On-line analysis by PTR-ToF-MS
José A. Sánchez López a,b , Marco Wellinger a , Alexia N. Gloess a , Ralf Zimmermann b,c ,
Chahan Yeretzian d,∗
a
Zurich University of Applied Sciences, Institute of Chemistry, 8820 Wädenswil, Switzerland
Joint Mass Spectrometry Centre, Chair of Analytical Chemistry, Institute of Chemistry, University of Rostock, D-18059 Rostock, Germany
c
Joint Mass Spectrometry Centre, Cooperation Group Comprehensive Molecular Analytics/CMA, Helmholtz Zentrum München, D-85764 Neuherberg,
Germany
d
Zurich University of Applied Sciences, Institute of Chemistry and Biotechnology, 8820 Wädenswil, Switzerland
b
a r t i c l e
i n f o
Article history:
Received 2 November 2015
Received in revised form 16 February 2016
Accepted 22 February 2016
Available online 2 March 2016
Keywords:
Coffee
Extraction
On-line
PTR-MS
Aroma
Coffee machine
a b s t r a c t
The hot-water extraction process used to make an espresso coffee is affected by a large number of factors.
A proper understanding of how these factors impact the profile of the final cup is important to the quality
of an espresso coffee. This work examines the effect of water temperature and pressure on the extraction
kinetics of volatile organic compounds (VOCs) in coffee. This was achieved by on-line monitoring of the
volatiles directly from the coffee flow, using proton-transfer-reaction time-of-flight mass-spectrometry
(PTR-ToF-MS). Using hierarchical cluster analysis (HCA), tentatively identified compounds were grouped
into 5 families according to their time–intensity profiles. VOCs grouped into each family had similar
physicochemical properties while polarity was found to be one of the main forces driving VOC extraction
kinetics. The effect of pressure was studied by extracting espresso coffees at 7, 9 and 11 bar. A pressure of
11 bar resulted in an increased extraction of volatiles over the entire extraction time (25 s). To study the
effect of temperature, espresso coffees were extracted at 82, 92 and 96 ◦ C. An increase in temperature
produced a significant increase in the extraction of VOCs, especially during the last part of the extraction.
The effect of temperature on extractability was more pronounced for the less polar compounds.
© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Coffee is one of the most widely consumed beverages [1]. The
beverage is made from coffee beans that are first harvested and
processed, then roasted and ground before finally being extracted.
Each and every single transformation step, from the seed to the
cup, must be mastered and performed with great care in order to
deliver the best quality in the cup [2–9]. Here, we focused on the last
and crucial transformation step, the extraction – more specifically,
espresso coffee extraction using a semi-automatic coffee machine.
During extraction, soluble compounds are dissolved and,
depending on the extraction technique, non-soluble compounds
are washed away with the extraction water, ending up in the extract
as dissolved or suspended solids [10–15]. Many different extraction
techniques have been introduced over the past centuries, which
vary according to factors such as geography, culture and social context, as well as personal preferences; these different factors can
∗ Corresponding author. Tel.: +41 0589345526.
E-mail address: chahan.yeretzian@zhaw.ch (C. Yeretzian).
result in vastly different flavor profiles in the extract. Of all coffee
brewing methods, espresso brewing is among the most popular
techniques.
Starting with whole roasted coffee beans, the preparation of the
“perfect” espresso is as much a science as an art. It is the result of the
interplay between several parameters that must be carefully controlled. These parameters include the particle size distribution of
the ground coffee, the water-to-coffee ratio, the final volume of the
brew in the cup and the temperature and pressure of the extracting water. An espresso is defined as a 25–35 ml beverage prepared
from 7 to 9 g of coffee, through which clean, 92–95 ◦ C water has
been forced at 9–10 atmospheres of pressure, and where the grind
of the coffee is such that the brewing ‘flow’ time is approximately
20–30 s [11].
An increase in the extraction temperature, for example, leads to
higher quantities of non-volatiles (i.e. total solids, caffeine, lipids) as
well as higher quantities of some volatiles, such as pyrazines, in the
final cup [16–18]. This may result in over-extraction and a coffee
cup with negative flavor notes, such as woody, burnt or acrid flavors. Increasing pressure up to 11 bar also resulted in coffees with
higher odor intensity and lower consumer acceptance than coffees
http://dx.doi.org/10.1016/j.ijms.2016.02.015
1387-3806/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.
0/).
67
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
extracted at 7 or 9 bar [19]. In these previous studies, the authors
focused on the composition and sensorial attributes of the final
cup in order to assess the impact of water temperature and pressure on espresso preparation. In a complementary line of research,
some scientists have been exploring the extraction kinetics of the
volatiles that contribute to the aroma of coffee. Two main methods have been previously used to determine the extraction kinetics
of coffee volatiles: off-line analysis of fractions using GC–MS [20]
and on-line analysis of the volatiles released by the coffee flow
using Proton Transfer Reaction Time of Flight Mass Spectrometry
(PTR-ToF-MS) [21,22]. In both cases, the methods were applied to
single-serve capsule systems in which both temperature and pressure were kept constant. The objective of the research presented
here was to investigate the effect of temperature and pressure on
the extraction kinetics of coffee aroma compounds by applying
on-line analysis by PTR-ToF-MS. We focused on 46 VOCs in particular and explored the link between extraction kinetics and their
physicochemical properties.
2. Materials and methods
2.1. Coffee extraction
The yellow bourbon variety of Coffea Arabica L. from Mogiana,
Brazil (Roaster: Kaffeepur, Switzerland, “Yellow Sun”), roasted to a
medium roast degree of 95 Pt (Colorette 3b, Probat, Germany), was
used for all of the extraction experiments. The coffee was frozen
two weeks after roasting and defrosted 12 h prior to the experiments to ensure a constant and equal freshness of the coffee for
all of the experiments. Less than one month elapsed from roasting
to extraction. The beans were ground using a Compak K10 grinder
(Barcelona, Spain) using position 47 on a scale from 0 (fine ground,
Turkish coffee) to 60 (coarse ground, French Press pot). 18 g of
the ground coffee were weighed into a double porta-filter basket,
tapered by hand and extracted for 25 s using a semi-automatic coffee machine (Dalla Corte Mini, Dalla Corte, Italy). The water used for
the extractions was commercially available Volvic mineral water
(total mineralization 130 mg/L; HCO3 − : 71 mg/L; SO4 2− : 8.1 mg/L;
Na+ : 11.6 mg/L; Ca2+ : 11.5 mg/L; Mg2+ : 8 mg/L). Extractions were
performed with five different combinations of water pressure and
temperature, one within the recommendations provided by the
Specialty Coffee Association of America (center point conditions)
and the others with values that exceed or were lower than recommended: (9 bar/92 ◦ C: center point; 7 bar/92 ◦ C; 11 bar/92 ◦ C;
9 bar/82 ◦ C; and 9 bar/96 ◦ C). Although all the coffees were prepared
by an experienced barista, variations resulting from the manual
preparation process were expected. To reduce this variability, we
performed 8 extractions for each set of conditions and selected the
5 replicates for which the final weight of the extract was closest to
30 g, ending up with coffee weights in the range of 31.5 g ± 2 g.
2.2. Sampling setup
Volatiles were sampled using a previously used setup [22], with
certain modifications (Fig. 1). The extracted coffee flowed into a
custom built system that was heated to 96 ◦ C to avoid condensation.
Volatiles were drawn out using a vacuum pump and diluted 10fold with dry compressed air to reduce their concentration to the
dynamic range of the PTR-ToF-MS being used.
2.3. PTR-ToF-MS
A commercial PTR-ToF-MS 8000 instrument (Ionicon Analytik
GmbH, Innsbruck, Austria) was used for all measurements. The
diluted sample was introduced via a 90 ◦ C heated sampling line into
68
23
Fig. 1. Setup used for the online monitoring of volatiles during coffee extraction.
the drift tube operated at 2.3 mbar, 90 ◦ C and 600 V drift tube voltage, resulting in an E/N value (electric field strength/gas number
density) of 140 Townsend (Td, 1 Td = 10−17 cm2 /V s). PTR-ToF-MS
data were recorded by TOFDAQ v.183 data acquisition software
(Tofwerk AG, Thun, Switzerland). Mass spectra were recorded in
the mass-to-charge (m/z) range of 0–300 with one mass-spectrum
recorded every 2 s.
2.4. Data processing
Dead time correction, mass calibration, peak extraction and
integration were performed using PTR-TOF DATA Analyzer software (v4.17) [23]. Duty cycle corrected signals were normalized to
106 H3 O+ primary ions and the concentration in parts per billion by
volume (ppbv) was estimated using 2 × 10−9 cm3 s−1 as a reaction
rate constant coefficient [24].
More than 500 mass peaks were detected in the range
0–300 m/z. Ions not related to the sample (O2 + , NO+ and water
clusters) were eliminated, the background was subtracted and a
concentration threshold of 1 ppb was set for further peak selection.
This resulted in a reduction to 120 mass peaks that were present
in all of the samples. From these peaks, 46 compounds (Table 1)
were tentatively identified by comparing them to the literature
[25,26].
2.5. Statistical analysis
The areas under the time–intensity curves were calculated for
the 120 mass traces of each of the five replicates of the five different extraction conditions. Principal Component Analysis (PCA)
was performed using mean centered and scaled areas. Analysis of Variance (ANOVA) was applied to assess the effect of the
different extraction variables on the total area of the selected
compounds using Tukey’s Honest Significant Difference (HSD)
post hoc test (p < 0.01). In order to identify compounds with
similar dynamic behavior, first the time–intensity profiles of all
the 120 m/z were normalized to their maximum intensity before
performing self-organizing tree algorithm (SOTA). Subsequently
the same SOTA analysis was performed on only the tentatively
identified 46 VOCs. In addition the normalized time–intensity profiles of the 46 tentatively identified compounds were subjected
to Hierarchical Cluster Analysis (HCA) using Ward’s minimum
variance method and half-squared Euclidean distances. All analyses were performed and all graphs were created using existing
packages (clValid, multcomp, and ggplot2) and scripts developed
in R [27].
24
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
Table 1
List of tentatively identified mass peaks, assigned sum formula and physic-chemical properties. Compounds are grouped in families according to the results obtained from
Hierarchical Cluster Analysis (HCA).
Compound
number
Measured
m/z
Theoretical
m/z
Sum formula
Tentative
identification
Boiling
point/◦ C
log Kaw
Vapor pressure/
KPa at 25 ◦ C
Water
solubility/g L−1
log Kow
Formaldehyde
Methanol
Acetaldehyde
Acetone
Propanal
Acetic acid
Propanoic acid
Ethyl acetate
2,3-Butanedione
Butyrolactone
−19
65
20
56
48
118
141
77
88
204
−4.861
−3.730
−2.564
−2.790
−2.523
−5.388
−4.740
−2.261
−3.265
−5.667
518
16.9
120
30.9
42.3
2.09
0.471
12.4
7.57
0.060
400
1000
1000
1000
306
1000
1000
80
200
1000
0.35
−0.77
−0.34
−0.24
0.59
−0.17
0.33
0.73
−1.34
−0.64
Formic acid
2-Propenal
2-Methylpropenal
3-Buten-2-one
Methylbutenal
Methylpropanoate
Pentanedione
!-Valerolactone
101
53
68
81
117
80
138
192
−5.166
−2.302
−2.023
−2.721
−2.444
−2.148
−4.017
−2.255
7.01
36.5
20.7
12.2
2.39
11.2
0.495
0.073
1000
212
50
60.63
25
62.4
166
93.81
−0.54
−0.01
0.74
0.41
1.15
0.84
0.4
0.11
C4 H7 +
C2 H7 S+
C4 H9 O+
1,3-Butadiene
Dimethyl sulfide
Butyraldehyde
2-Methylpropanal
Butanone
−4
37
75
65
80
0.478
−1.182
−2.328
−2.133
−2.633
Pyrrole
Furan
Triazole
Methylpyrrole
Methylpyrazine
Furfural
Furfuryl alcohol
Hydroxypentanone
1-Methyl-2-butanoic acid
Dimethylpyrazine
Ethylpyrazine
Acetylfuran
5-Methylfurfural
5-Methylfurfuryl alcohol
4-Methyltetrahydro-2Hpyran-2-one
2-Oxopropyl acetate
Ethyl acetoacetate
130
31
203
112
135
162
171
179
177
156
153
175
187
191
213
−3.133
−0.656
−4.212
−3.388
−4.046
−3.861
−5.493
−3.255
−4.468
−3.838
−3.999
−3.398
−3.218
−5.010
−2.131
1.11
80.0
0.080
2.85
1.29
0.295
0.081
0.033
0.065
0.365
0.476
0.126
0.091
0.017
0.025
171
181
−4.467
−4.309
0.199
0.104
115
94
145
167
220
188
169
171
205
241
265
225
201
−3.347
−2.187
−0.949
−4.121
−6.171
−3.762
−3.715
−3.795
−4.309
−8.580
−8.456
−5.081
−3.948
2.77
1.39
0.853
0.240
0.004
0.072
0.081
0.193
0.014
4.6 × 10−4
1.0 × 10−4
0.017
0.019
1000
10
0.310
21.38
17.59
7.284
1.903
15.21
18.7
16.48
5.52
104.8
2.416
0.65
1.23
2.95
0.84
0.93
1.33
1.53
0.95
1.32
1.58
2.07
−0.05
1.83
209
247
292
237
−4.718
−6.085
−3.195
−3.985
0.022
3.1 × 10−4
0.011
0.004
3.716
3.327
0.568
0.557
1.6
1.65
2.5
2.51
264
−5.706
3.1 × 10−4
0.305
2.73
−1
65
0.978
−0.615
0.2
3
2.4
1.85
Family A
1
2
3
4
31.019
33.034
45.034
59.047
31.018
33.033
45.033
59.049
CH3 O+
CH5 O+
C2 H5 O+
C3 H7 O+
5
6
61.029
75.045
61.028
75.044
C2 H5 O2 +
C3 H7 O2 +
7
87.046
87.044
C4 H7 O2 +
Family B
8
9
10
47.014
57.036
71.051
47.013
57.033
71.049
CH3 O2 +
C3 H5 O+
C4 H7 O+
85.066
89.061
101.061
85.065
89.060
101.060
Family C
14
15
16
55.057
63.03
73.064
55.054
63.026
73.065
Family D
17
18
19
20
21
22
23
24
68.051
69.036
70.041
82.064
95.062
97.031
99.046
103.077
68.049
69.033
70.040
82.065
95.060
97.028
99.044
103.075
C4 H6 N+
C4 H5 O+
C2 H4 N3 +
C5 H8 N+
C5 H7 N2 +
C5 H5 O2 +
C5 H7 O2 +
C5 H11 O2 +
25
109.078
109.076
C6 H9 N2 +
26
111.046
111.044
C6 H7 O2 +
27
28
113.062
115.075
113.060
115.075
C6 H9 O2 +
C6 H11 O2 +
29
30
117.057
131.073
117.055
131.070
C5 H9 O3 +
C6 H11 O3 +
Family E
31
32
33
34
35
36
37
80.052
87.082
105.068
107.06
110.064
121.075
123.093
80.049
87.080
105.070
107.060
110.060
121.076
123.092
C5 H6 N+
C5 H11 O+
C8 H9 +
C6 H7 N2 +
C6 H8 NO+
C7 H9 N2 +
C7 H11 N2 +
38
125.063
125.060
C7 H9 O2 +
39
127.075
127.075
C7 H11 O2 +
40
135.091
135.092
C8 H11 N2 +
41
138.087
138.091
C8 H12 NO+
+
11
12
13
C5 H9 O+
C4 H9 O2 +
C5 H9 O2 +
42
43
148.077
149.112
148.076
149.107
C9 H10 NO
C9 H13 N2 +
44
165.093
165.091
C10 H13 O2 +
Pyridine
Methylbutanal
Vinylbenzene
Ethenylpyrazine
2-Acetylpyrrole
Ethenylmethyl-pyrazine
2-Ethyl-5-methylpyrazine
Trimethylpyrazine
Guaiacol
Methylbenzenediol
Ethylbenzenediol
Ethylcyclopentanedione
5-Methyl-6,7-dihydro-5Hcyclopentapyrazine
2-Acetyl-1-ethylpyrrole
3-Acetyl-2,4-dimethylpyrrole
1-Furfurylpyrrole
Dihydro-dimethyl
cyclopentapyrazine
Allylguaiacol
Separate compounds
57.072
45
83.051
46
57.070
83.049
C4 H9 +
C5 H7 O+
1-Butene
Methylfuran
281
66.9
14.8
23.1
12.1
300
20.8
0.735
22
71
89
223
1.99
0.92
0.88
0.74
0.29
45
10
240
10
80
77
1000
48.868
45
38.16
28.41
39.1
29.11
49.18
32.19
0.75
1.34
−0.29
1.21
0.21
0.41
0.28
0.2
1.18
0.54
0.69
0.52
0.67
0.99
0.6
151.9
110
−0.19
0.25
69
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
3. Results and discussion
0.50
families and a sixth one containing only 49 profiles corresponding
to mass traces that did not fit with the other families (Supplementary material 1). This corroborates the robustness of the five main
families of volatile, and of their corresponding classification according to different extraction dynamics during espresso brewing.
Fig. 2 shows the time–intensity profiles of the center point
experiment for each of the five families. Time zero corresponds to
the moment at which the pump of the coffee machine started to
run, however, coffee did not start to flow out of the portafilter until
5–6 s later. After 25 s, the pump was stopped and 1 s later the coffee flow also stopped. This figure shows the distinguishing features
found between the different families.
Family A was characterized by a fast rise in intensity 6 s after the
coffee machine started to pump water, reaching maximum intensity at 10 s. From that point until the end of the extraction, this
family exhibited the fastest decrease in intensity of all of the families, with final values lower than 40% of the maximum intensity.
Independent of the extraction conditions, the same seven compounds were always clustered into this family.
Family B’s profile was similar to that of A during the first 10 s,
after which and in contrast to compounds of family A the intensity
did not change (i.e. decrease) significantly, resulting in a plateau
of maximum intensity until second 16. At this point it started to
decrease, before finally reaching 60–70% of the maximum of the
intensity at the end of the extraction process.
Family C exhibited a local maximum in intensity (visible as a
peak shoulder) at 4 s. This shoulder was also present in the other
families, although much less pronounced. From that point on, the
intensity rose rapidly until it reached a maximum at 10 s, after
which the profile was almost identical to family B, with a plateau
until 16 s and a subsequent decrease. This high similarity in profiles, in particular for the second half of the extraction time, meant
that families B and C were clustered together for the extraction at
96 ◦ C. Despite not being separated by HCA at the highest extraction temperature, the profiles for family C could still be visually
differentiated by the characteristic shoulder at 4 s.
In Family D, the intensity started to rise after 6 s and reached
a maximum at between 14 and 16 s. After that, the intensity
decreased to 70–80% of the maximum. For extraction at 7 bar, three
compounds from this family were grouped together with those in
family B. As will be discussed in Section 3.2, extraction at 7 bar
showed a slower increase in intensity and lower maximum intensities than the center point experiment (9 bar/92 ◦ C) for compounds
belonging to family B, making the profiles of this family more similar to those of D and hence affecting the classification into families.
Family E showed the slowest increase in intensity from 6 s
to a maximum at 20 s. The intensity decreased slightly from
the maximum until the end of the extraction, with values that
were approximately 80–95% of the maximum intensity. When the
extraction was performed at the highest temperature, the characteristic decrease in intensity during the last part of the extraction
was not observed for some compounds in family D. In fact the intensity increased until the end of the extraction for some compounds
in family E (Fig. 5). This effect resulted in similar profiles to family
D, meaning that four compounds from family E were grouped with
those of family D for the 96 ◦ C extraction.
0.25
3.2. Classification of families
3.1. Dynamics of extraction
During the 25 s coffee extraction, the intensity over time of
the volatile compounds in the coffee was monitored and analyzed in detail for the 46 tentatively identified compounds. The
focus was put on these 46 compounds, in contrast to the complete list of 120 compounds, as this allowed to link the identified
VOC to their physical properties. Hierarchical Clustering Analysis
(HCA) was firstly applied to the normalized time–intensity profiles of the center point experiment (9 bar/92 ◦ C) resulting in the
classification of 44 compounds into five different families that
shared similar time–intensity profiles (Table 1 and Fig. 2). Two
compounds did not fit into any of the five main families: 1-butene
(C4 H9 + ) and methylfuran (C5 H7 O+ ). In the case of 1-butene, the
intensity showed an irregular profile during extraction, with high
standard deviation between the replicates. The m/z attributed
to 1-butene (57.073) has also been reported as an alcohol fragment [28]. This suggests m/z 57.073 does not correspond only
to 1-butene but rather to a superposition of ion intensities from
additional compounds and/or fragments with the same mass-tocharge ratio, which interfered with the 1-butene signal. In the
case of methylfuran, the signal rose to its maximum value at 4–6 s
and then remained constant until the end of the extraction process. This behavior was not observed for any other compound. It
should also be noted that the time–intensity profiles recorded for
methylfuran did not show any significant differences for any of the
extraction conditions analyzed.
HCA was subsequently applied to the time–intensity profiles
of all the extraction conditions together, to check if the VOC
families observed for the center point were independent of the
extraction conditions. Essentially the same five families were
reproduced when considering all conditions, although a few compounds clustered differently for the lowest pressure and the
highest temperature conditions, relative to the center pint extraction: (i) for extraction at 7 bar, three compounds from family
D (methylpyrazine, furfural and acetol acetate) were classified
as members of family B; (ii) for extraction at 96 ◦ C, all three
compounds from family C (butadiene, dimethylsulfide and methylpropanal) appeared in family B, and four compounds from family
E (ethenyl pyrazine, pyridine, acetylpyrrole and ethylbencenediol)
appeared in family D.
Clustering was performed only on the 46 tentatively identified
compounds. In order to confirm that those compounds are representative of all the measured m/z, we used the Self Organizing Tree
Algorithm (SOTA) on a data set composed of the 3000 intensity
profiles recorded (5 brewing conditions × 5 replicates × 120 m/z).
Six clusters were obtained of which five were identical to the main
normalized intensity
1.00
0.75
0.00
0
5
10
15
time/s
20
25
Fig. 2. Normalized time–intensity profiles of the different families of compounds
extracted at 92 ◦ C and 9 bar. Lines represent the mean and the error bars represent the standard deviation of all the compounds in family A (!), B (!), C ("),
D (#), E (").
70
25
The extraction of aroma compounds from the coffee beans by
water is mainly driven by polarity [20]. For the 46 compounds that
were tentatively identified, values for log Kow (partition coefficient
between octanol and water), log Kaw (partition coefficient between
air and water), water solubility, boiling point and vapor pressure
are provided in Table 1. Apart from a few exceptions, water solubility decreases and log Kow increases as one moves from family A
26
7
11 bar
B
22
0.3
A
29
4 45
10
1
0.1
19
0.0
7 bar
−5
0
5
17
40
−0.2
−4
96°C
−10
46
10
15
PC1 75.9%
12
13
15
16
18
9
−0.1
0
center
−15
5
PC2 7.6%
2
82°C
−2
PC2 7.6%
0.2
4
6
0.4
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
83
11
23
27
44 20
37
242 30
28 24 35
25
31
39
26
32
14
34
33
38 6 36
43
41
21
0.09
0.10
0.11
0.12
0.13
0.14
0.15
0.16
PC1 75.9%
Fig. 3. Score-plot (A) and loading-plot (B) of the first two components detected in the PCA, performed on the area under the time–intensity profiles for the 120 m/z. Data
points on the score plot represent the five different extractions performed for each set of conditions. Numbers on the loading plot correspond to the compound list in Table 1.
Non-identified compounds have been omitted from the loading plot for the purposes of clarity. A plot showing the loadings of all 120 m/z can be found in the supplementary
materials.
to E, indicating that the most polar compounds belong to family A
and the least polar ones to family E. Compounds in family A – with
water solubility of between 80 and 1000 g L−1 and log Kow between
1.34 and 0.73 – were extracted quickly, within the first seconds of
extraction, and their concentration levels decreased significantly at
the later stage of the extraction. Polarity and water solubility of the
compounds in family B are slightly lower than in family A, meaning that the intensity also increased quickly at the beginning, but
their extraction lasted longer. In family C, the compounds have a
lower water solubility (0.7–223 g L−1 ) and lower polarity (log Kow
0.29–1.99) than those of families A and B. This would imply that
these compounds are extracted more slowly from the coffee bed.
However, their intensity did increase rapidly during the first 6 s.
This fast increase in signal can be attributed to the high volatility
of the compounds in this family (vapor pressure: 12–281 KPa and
log Kaw : −2.64 to 0.48) that favored their release to the gas phase.
This could also explain the characteristic shoulder at 4 s, time at
which the coffee had not started to flow, but the coffee powder
had already been wetted by hot water and, as a consequence, compounds were released to the gas phase. Families D and E contain
the least polar, water soluble and volatile compounds. Therefore
compound transfer from the ground coffee particles to water was
slow, with an important fraction of these compounds still being
extracted after 25 s.
Grouping of dynamic data into families can also be used to
improve compound identification. Generally, tentative identification of compounds using PTR-ToF-MS and other direct mass
spectrometry techniques is performed by assigning a sum formula
to the measured mass and comparing it with compounds previously reported in the literature. However, this can potentially
lead to errors in compound assignment. When clustering all the
compounds according to their dynamic behavior, it is expected
that compounds in the same group will have similar physicochemical characteristics and those which differ might have been
miss-identified. In Table 1, some compounds can be singled out
as possibly having been miss-identified. For example, compound
number 6 (C3 H7 O2 + ) has been reported in the coffee literature
as both propanoic acid and ethyl acetate. The polarity and water
solubility of ethyl acetate are lower than those of the other compounds in family A, suggesting that the measured compound was
most probably propanoic acid. Another potential miss-assignment
is compound 31 (C5 H6 N+ ), which was reported as pyridine and clustered in family E. However, the physicochemical characteristics of
pyridine are closer to those of family A or B than to those of family E, indicating we may have detected a fragment of a less polar
compound containing a pyridine ring instead of pyridine itself.
3.3. Factors affecting extraction
To evaluate the impact of temperature and pressure on
the extraction, a PCA was performed using the total area
under the time–intensity curves of the selected 120 mass traces
(Fig. 3). The first two principal components explained 82.5% of the
total variability in the data, and the graphical representation of
the scores for these two components allowed differentiation into
five different groups corresponding to the different conditions used
for extraction (Fig. 3A). More information can be drawn from the
loading plot (Fig. 3B). Loadings for the 46 tentatively identified compounds were all positive for PC1, indicating an increase in total area
under the curve for higher temperatures and pressures. For PC2,
18 tentatively identified compounds had positive scores and are
related to high pressure extraction. All but four of the compounds
(18, 19, 22, 29) belong to the most polar groups A, B or C. Compounds
with negative scores for PC2 belonged to families D and E, together
with compounds number 2, 6, 12, 13 and 14 from the other families. Negative PC2 scores are related to an increased area under the
curve for higher extraction temperatures, indicating that increases
in temperature had a greater effect on the less polar compounds.
To obtain additional information on how pressure and temperature affected the extraction, the respective time–intensity profiles
and the time evolution of the area under the curve were compared
for one compound from each family (Fig. 4 for pressure and Fig. 5
for temperature effect). As discussed in the Section 3.1, the grouping of the different time–intensity profiles hardly varied between
the different extraction parameters and so we, therefore, assumed
that one compound would be representative of the whole family.
3.3.1. Effect of pressure
The time–intensity profiles of family A showed no significant
differences based on extraction pressure for the representative
compounds (Fig. 4). For C5 H9 O+ and C4 H9 O+ , representatives of
families B and C respectively, extraction at 7 bar resulted in lower
intensities, but only during the middle phase of the extraction
(8–16 s). The highest effect of pressure on the time–intensity profiles was observed for families D and E, where there were no
differences between 7 and 9 bar. However, extraction at 11 bar
resulted in significantly higher intensities during the last 10 s of
extraction. Since the differences in intensities at each point in time
were small, but could accumulate and have high impact in the final
cup, we also calculated and plotted the total area under the curve
up to each point in time (second column Fig. 4). Boxplots showed
no significant differences between 7 and 9 bar for any of the families, but extraction at 11 bar always resulted in significantly higher
71
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
27
Fig. 4. Time–intensity profiles of compounds chosen as single representatives of each family during coffee extraction at three different pressures. Points represent the mean
and the error bars represent the standard deviation of the replicates. Boxplots represent the area under the curve at that point in time and the insert in the left corner is a
magnification of the last point (26 s), with letters representing measurements that differ significantly for the different extraction parameters (Tukey’s test, p < 0.01).
72
28
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
Fig. 5. Time–intensity profiles of one representative compound of each family during coffee extraction at three different temperatures. Points represent the mean and the
error bars represent the standard deviation of the replicates. Boxplots represent the area under the curve at that point in time and the insert in the left corner is a magnification
of the last point (26 s) with letters representing measurements that differ significantly for the different extraction parameters (Tukey’s test, p < 0.01).
73
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
concentrations over the entire extraction time, compared to the
extraction at 7 bar.
Pressure is the driving force required to produce a flow of water
through the compacted coffee cake, assisting the extraction of
compounds trapped inside the coffee particle structure, and also
helping to transfer solid particles and oil droplets from the ground
beans to the cup [15]. Some authors have reported an increase
in chlorogenic acids, lipids, coffee oil, diterpenes and aroma compounds when increasing pressure from 7 to 9 bar [19,29] although
no differences in caffeine or total solids were observed. A further pressure increase to 11 bar had either no effect or produced
a decrease in the above compounds. The negative effect on extraction efficiency of high pressures has been attributed to a decrease
in flow [19,29]. In our case, the extraction at 11 bar did not significantly change the average flow of the coffee. This might be the
reason why the extraction efficiency of aroma compounds did not
decrease and extraction at 11 bar resulted in the highest intensity
of volatile compounds. This general increase in concentrations of
volatile compounds at 11 bar also correlates with the highest ranking for odor intensity reported by Andueza et al. when extracting
at this pressure [19].
3.3.2. Effect of temperature
Increases in temperature resulted in an increase in the measured intensity of VOCs, as shown on the time–intensity profiles
(Fig. 5). This was especially visible in the second half of the extraction (t > 14 s); however, the effect was different for each compound
family.
For family A, no significant difference was observed between
extractions at 92 and 96 ◦ C, but the extraction at 82 ◦ C resulted
in lower intensities. Boxplots of the evolution of the area under
the curve showed differences at the two extreme temperatures (82 ◦ C and 96 ◦ C) after 20 s of extraction, with all three
temperatures resulting in significantly different areas under
the curve at the end of the extraction. Family C displayed
similar behavior, but there was only a statistical difference
between 82 ◦ C and 96 ◦ C at the end of the extraction. For
family B, differences were evident on the time–intensity profiles after 16 s, although only in the last 2 s of the extraction
was the area under the curve significantly different for all three
temperatures.
The greatest effect of temperature was observed for families D and E. Significant differences in the area under the curve
for the two extreme temperatures appeared earlier than for the
other families (starting at 16 and 12 s for family D and E, respectively), and increased with increasing extraction time. In the case of
dimethylpyrazine, the representative of family D, we observed that
at 96 ◦ C the intensity reached a maximum at 16 s and then remained
constant until the end of the extraction, while for the other
extraction temperatures the intensity started to decrease once the
maximum had been reached. In the case of the representative
compound of family E (furfurylpyrrole), the increase of extraction efficiency at 96 ◦ C was even more evident. The time–intensity
curve increased until the end of the extraction, suggesting that
the maximum had not been reached within the extraction
time of 25 s.
In general, an increase in water temperature results in higher
water solubility for some compounds. The use of water at high
temperatures for brewing espressos has been related to increases
in extraction yield, caffeine, diterpenes, coffee oil and lipids
[15–17,29]. The more efficient extraction of coffee oil and lipids
at higher extraction temperatures may, in turn, favor the extraction of more lipophilic compounds. This could explain the greater
effect of temperature on the extraction of volatiles from families D
and E, which contain lower polarity compounds.
74
29
4. Conclusions
On-line PTR-ToF-MS analysis of volatile coffee compounds
released from the coffee flow during extraction has revealed itself
to be a very powerful approach for studying the kinetics of coffee
aroma extraction for various pressure and temperature parameters
using a semi-automatic coffee machine. The time–intensity profiles
showed large differences in the extraction kinetics between different volatile compounds and allowed compounds to be grouped
into five families with similar physicochemical characteristics. It
was shown that the polarity of the volatile compounds was the
main driving force for their extraction. Extraction profiles of the
aroma compounds changed with the different brewing parameters
used: increases in both pressure and temperature resulted in higher
extraction of VOCs, with the least polar compounds being the most
affected, mainly impacting the aroma balance in the last stage of
the extraction.
Acknowledgments
The research leading to these results has received funding
through the PIMMS ITN, which is supported by the European Commission’s seventh Framework Programme under grant agreement
number 287382. We also would like to thank the Specialty Coffee
Association of Europe for financial support.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ijms.2016.02.015.
References
[1] M. Pendergrast, Coffee second only to oil? Is coffee really the second largest
commodity? Tea Coffee Trade J. April (2009).
[2] A.N. Gloess, A. Vietri, F. Wieland, S. Smrke, B. Schönbächler, J.A. Sánchez-López,
S. Petrozzi, S. Bongers, T. Koziorowski, C. Yeretzian, Evidence of different flavour
formation dynamics by roasting coffee from different origins: on-line analysis
with PTR-ToF-MS, Int. J. Mass Spectrom. 365–366 (2014) 324–337.
[3] A.N. Gloss, B. Schonbachler, M. Rast, L. Deuber, C. Yeretzian, Freshness indices
of roasted coffee: monitoring the loss of freshness for single serve capsules and
roasted whole beans in different packaging, Chimia 68 (3) (2014) 179–182.
[4] R. Hertz-Schönemann, R. Dorfner, C. Yeretzian, T. Streibel, R. Zimmermann,
On-line process monitoring of coffee roasting by resonant laser ionisation timeof-flight mass spectrometry: bridging the gap from industrial batch roasting to
flavour formation inside an individual coffee bean, J. Mass Spectrom. 48 (12)
(2013) 1253–1265.
[5] R. Hertz-Schönemann, T. Streibel, S. Ehlert, R. Zimmermann, Looking into individual coffee beans during the roasting process: direct micro-probe sampling
on-line photo-ionisation mass spectrometric analysis of coffee roasting gases,
Anal. Bioanal. Chem. 405 (22) (2013) 7083–7096.
[6] C. Lindinger, D. Labbe, P. Pollien, A. Rytz, M.A. Juillerat, C. Yeretzian, I. Blank,
When machine tastes coffee: instrumental approach to predict the sensory
profile of espresso coffee, Anal. Chem. 80 (5) (2008) 1574–1581.
[7] S. Smrke, I. Kroslakova, A.N. Gloess, C. Yeretzian, Differentiation of degrees of
ripeness of Catuai and Tipica green coffee by chromatographical and statistical
techniques, Food Chem. 174 (2015) 637–642.
[8] F. Wieland, A.N. Gloess, M. Keller, A. Wetzel, S. Schenker, C. Yeretzian, Online
monitoring of coffee roasting by proton transfer reaction time-of-flight mass
spectrometry (PTR-ToF-MS): towards a real-time process control for a consistent roast profile, Anal. Bioanal. Chem. 402 (8) (2012) 2531–2543.
[9] C. Yeretzian, A. Jordan, R. Badoud, W. Lindinger, From the green bean to the cup
of coffee: investigating coffee roasting by on-line monitoring of volatiles, Eur.
Food Res. Technol. 214 (2) (2002) 92–104.
[10] G. Caprioli, M. Cortese, G. Sagratini, S. Vittori, The influence of different types
of preparation (espresso and brew) on coffee aroma and main bioactive constituents, Int. J. Food Sci. Nutr. 66 (5) (2015) 505–513.
[11] A. Gloess, B. Schönbächler, B. Klopprogge, L. D‘Ambrosio, K. Chatelain, A. Bongartz, A. Strittmatter, M. Rast, C. Yeretzian, Comparison of nine common coffee
extraction methods: instrumental and sensory analysis, Eur. Food Res. Technol.
236 (4) (2013) 607–627.
[12] A.N. Gloess, B. Schönbächler, B. Klopprogge, L. D‘Ambrosio, K. Chatelain, A. Bongartz, A. Strittmatter, M. Rast, C. Yeretzian, Comparison of nine common coffee
extraction methods: instrumental and sensory analysis, Eur. Food Res. Technol.
236 (4) (2013) 607–627.
30
J.A. Sánchez López et al. / International Journal of Mass Spectrometry 401 (2016) 22–30
[13] F. Mayer, M. Czerny, W. Grosch, Sensory study of the character impact aroma
compounds of a coffee beverage, Eur. Food Res. Technol. 211 (4) (2000)
272–276.
[14] M. Petracco, Technology IV: beverage preparation: brewing trends for the new
millennium, in: R.J. Clarke, O.G. Vitzthum (Eds.), Coffee: Recent Developments,
Blackwell Science, London, 2001.
[15] M. Petracco, Technology IV: beverage preparation: brewing trends for the new
millennium, in: Coffee, Blackwell Science Ltd., 2008, pp. 140–164.
[16] D. Albanese, M. Di Matteo, M. Poiana, S. Spagnamusso, Espresso coffee (EC) by
POD: study of thermal profile during extraction process and influence of water
temperature on chemical–physical and sensorial properties, Food Res. Int. 42
(5-6) (2009) 727–732.
[17] S. Andueza, L. Maeztu, L. Pascual, C. Ibanez, M. Paz de Peña, C. Cid, Influence of
extraction temperature on the final quality of espresso coffee, J. Sci. Food Agric.
83 (3) (2003) 240–248.
[18] S. Andueza, M.A. Vila, M. Paz de Peña, C. Cid, Influence of coffee/water
ratio on the final quality of espresso coffee, J. Sci. Food Agric. 87 (4) (2007)
586–592.
[19] S. Andueza, L. Maeztu, B. Dean, M.P. de Peña, J. Bello, C. Cid, Influence of water
pressure on the final quality of arabica espresso coffee. Application of multivariate analysis, J. Agric. Food Chem. 50 (25) (2002) 7426–7431.
[20] F. Mestdagh, T. Davidek, M. Chaumonteuil, B. Folmer, I. Blank, The kinetics of
coffee aroma extraction, Food Res. Int. 63 (2014) 271–274.
[21] J.A. Sánchez-López, R. Zimmermann, C. Yeretzian, Extraction-kinetics of
volatile coffee aroma compounds by on-line PTR-ToF-MS, in: A.J.M. Taylor,
S. Donald (Eds.), Weurman Flavour Research Symposium, Context, Queens’
College Cambridge, UK, 2014, pp. 147–150.
[22] J.A. Sánchez-López, R. Zimmermann, C. Yeretzian, Insight into the timeresolved extraction of aroma compounds during espresso coffee preparation:
online monitoring by PTR-ToF-MS, Anal. Chem. 86 (23) (2014) 11696–11704.
[23] M. Müller, T. Mikoviny, W. Jud, B. D’Anna, A. Wisthaler, A new software tool
for the analysis of high resolution PTR-TOF mass spectra, Chemom. Intell. Lab.
Syst. 127 (2013) 158–165.
[24] W. Lindinger, A. Hansel, A. Jordan, On-line monitoring of volatile organic compounds at pptv levels by means of proton-transfer-reaction mass spectrometry
(PTR-MS): medical applications, food control and environmental research, Int.
J. Mass Spectrom. 173 (1998) 191–241.
[25] I. Flament, Coffee Flavor Chemistry, John Wiley & Sons, Ltd., 2002.
[26] C. Yeretzian, A. Jordan, W. Lindinger, Analysing the headspace of coffee by
proton-transfer-reaction mass-spectrometry, Int. J. Mass Spectrom. 223–224
(January) (2003) 115–139.
[27] R-Development-Core-Team, A language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria, 2008.
[28] K. Buhr, S. van Ruth, C. Delahunty, Analysis of volatile flavour compounds by
proton transfer reaction-mass spectrometry: fragmentation patterns and discrimination between isobaric and isomeric compounds, Int. J. Mass Spectrom.
221 (1) (2002) 1–7.
[29] M. Moeenfard, J.A. Silva, N. Borges, A. Santos, A. Alves, Diterpenes in espresso
coffee: impact of preparation parameters, Eur. Food Res. Technol. 240 (2015)
763–773.
75
Publication 3.
Rapid and direct volatile compound profiling of black and green teas (camellia sinensis)
from different countries with PTR-ToF-MS
by
Sine Yener, José A. Sánchez-López, Pablo M. Granitto, Luca Cappellin, Tilmann D. Märk, Ralf
Zimmerman, Günther K. Bonn, Chahan Yeretizian, Franco Biasioli
Talanta
Volume 152, 2016, Pages 45–53.
DOI: 10.1016/j.talanta.2016.01.050
José A. Sánchez-López was involved in the design and execution of the experiments. He collaborated on
the data analysis and the preparation of the manuscript. His work to this publication accounts for
approximatelly 40%.
76
Talanta 152 (2016) 45–53
Contents lists available at ScienceDirect
Talanta
journal homepage: www.elsevier.com/locate/talanta
Rapid and direct volatile compound profiling of black and green teas
(Camellia sinensis) from different countries with PTR-ToF-MS
Sine Yener a,b, José A. Sánchez-López c,d, Pablo M. Granitto e, Luca Cappellin a,
Tilmann D. Märk g, Ralf Zimmermann d,f, Günther K. Bonn b,h, Chahan Yeretzian c,
Franco Biasioli a
a
Department of Food Quality and Nutrition, Research and Innovation Centre, Fondazione Edmund Mach, 38010 San Michele all’Adige, Italy
Institute of Analytical Chemistry and Radiochemistry, Leopold-Franzens University of Innsbruck, CCB- Center of Chemistry and Biomedicine, Innrain 80-82,
6020 Innsbruck, Austria
c
Zurich University of Applied Sciences (ZHAW), Institute of Chemistry and Biological Chemistry, 8820 Wädenswil, Switzerland
d
Joint Mass Spectrometry Centre, Chair of Analytical Chemistry, Institute of Chemistry, University of Rostock, d-18059 Rostock, Germany
e
CIFASIS, French Argentine International Center for Information and Systems Sciences, UNR–CONICET, Bv 27 de Febrero 210 Bis, 2000 Rosario, Argentina
f
Joint Mass Spectrometry Centre, Cooperation Group Comprehensive Molecular Analytics/CMA, Helmholtz Zentrum München, d-85764 Neuherberg,
Germany
g
Institute of Ion Physics and Applied Physics, Leopold-Franzens University of Innsbruck, Technikerstr. 25/3, 6020 Innsbruck, Austria
h
Austrian Drug Screening Institute, Innrain 66a, 6020 Innsbruck, Austria
b
art ic l e i nf o
a b s t r a c t
Article history:
Received 16 November 2015
Received in revised form
18 January 2016
Accepted 25 January 2016
Available online 26 January 2016
Volatile profiles of 63 black and 38 green teas from different countries were analysed with Proton
Transfer Reaction-Time of Flight-Mass Spectrometry (PTR-ToF-MS) both for tea leaves and tea infusion.
The headspace volatile fingerprints were collected and the tea classes and geographical origins were
tracked with pattern recognition techniques. The high mass resolution achieved by ToF mass analyser
provided determination of sum formula and tentative identifications of the mass peaks. The results
provided successful separation of the black and green teas based on their headspace volatile emissions
both from the dry tea leaves and their infusions. The volatile fingerprints were then used to build different classification models for discrimination of black and green teas according to their geographical
origins. Two different cross validation methods were applied and their effectiveness for origin discrimination was discussed. The classification models showed a separation of black and green teas according to geographical origins the errors being mostly between neighbouring countries.
& 2016 Elsevier B.V. All rights reserved.
Keywords:
Tea aroma
Tea leaf
Tea infusion
Volatile profiling
Headspace volatile fingerprinting
Geographic origin classification
1. Introduction
In tea production, the leaves of the tea plant Camellia sinensis
are used as the same starting material but the differences in the
processing techniques result in a wide range of characteristic teas
with distinct sensory properties. According to the way of processing, teas are usually classified into three big groups based on their
fermentation degrees: non-fermented (green and white), semifermented (oolong) and fully fermented (black tea including puerh tea) [1]. There are several tea producing countries in the world.
The main five tea producing countries are China, India, Kenya, Sri
Lanka and Turkey [2]. Each country has different regions with their
own climate and tea processing methods which characterize colour, appearance and flavour of the final product. For this reason,
most tea products are marketed with the indication of the production region for product authentication and valorization.
Aroma compounds play an important role for consumer preferences and perception of tea. Starting with the fresh tea leaves,
which have a greenish and unripe odour, the characteristic tea aroma
is developed during tea leaves processing. The most investigated
volatile compounds (VOCs) in tea mainly consist of non-terpenoid
and terpenoid components; the former are products of fatty acid
degradation and provide the fresh green flavour, the latter are mostly
monoterpene alcohols which give a floral sweet aroma [3,4].
Various studies have been conducted in the field of tea aroma
research as recently reviewed by Yang et al. (2013) [5]. In short, gas
chromatography-mass spectrometry (GC–MS) is generally used as
a reference method in order to identify and quantify VOCs. The
odour characteristics of volatiles have been detected with aroma
dilution and GC-olfactometry; and recently electronic nose techniques have been used for fast analysis of tea aroma. These
methods have allowed analysing the volatile profiles of teas at
http://dx.doi.org/10.1016/j.talanta.2016.01.050
0039-9140/& 2016 Elsevier B.V. All rights reserved.
77
46
S. Yener et al. / Talanta 152 (2016) 45–53
different fermentation degrees and also to classify green, black and
oolong teas according to their geographical origins [6–11]. Among
them, GC–MS has turned out to be the most accurate and effective
method for identification, separation and quantification of volatile
compounds; however it requires capturing volatiles by various
extraction methods which are generally time consuming and their
efficiency depends on the characteristics and limitations of the
analytical approach (e.g. the absorption and desorption of volatiles
from a specific material in the case of SPME) [12].
To link sensory perception of tea with instrumental data, direct
and non-destructive instrumental analysis of volatiles can be
considered to be the most appropriate approach because it provides a direct estimation of the VOCs released of from tea and that
reach the human olfactory system. In this regard, proton transfer
reaction-mass spectrometry, PTR-MS, provides an efficient approach as a direct injection, soft chemical ionization method for
the analysis of VOCs at trace levels. The direct injection method
requires no sample pre-treatment which allows real-time monitoring of VOCs [13, 14] without making any changes in the volatile
composition of samples. The technique uses H3O þ ions for protonation of VOCs with proton affinities higher than that of water
which can be further analysed by a quadruple or a time-of-flight
(ToF) mass analyser [15]. ToF mass analysers provide high sensitivity that leads to detection of volatiles at ppt levels and high
mass resolution which allows, in most cases, the identification of
the sum formula of the observed peaks [13].
PTR-MS allows collecting the overall mass spectral fingerprints
of the samples which can be further processed with advanced data
analysis tools for successful discrimination and classification of the
food products [16]. To the best of our knowledge, neither a study
has been conducted on the analysis of volatile compounds emitted
from various tea types by PTR-MS nor was this method applied for
discrimination of teas from different geographical origins.
With this study, we aim to apply PTR-ToF-MS, for the first time,
for aroma profiling of black and green tea samples, both leaves and
brew, from different countries and to investigate the possibility of
origin tracing on the basis of their geographical origins with the
aid of chemometric tools.
2. Materials and methods
2.1. Tea samples
In total, 101 commercially available pure tea samples, without
addition of flavouring agents, from 16 different countries (Table 1)
Table 1
Distribution of tea samples according to tea types and countries of origin.
Country (code)
Argentina (ARG)
China (CHI)
India (IND)
Indonesia (INDO)
Japan (JAP)
Kenya (KEN)
Korea (KOR)
Nepal (NEP)
Rwanda (RWA)
Sri Lanka (SRI)
Tanzania (TNZ)
Turkey (TUR)
Vietnam (VIE)
Zimbabwe (ZIM)
(Total)
78
Tea types
Black
Green
1
13
25
3
–
2
1
4
–
8
1
3
1
1
63
–
15
4
–
9
–
4
1
1
1
1
–
2
–
38
were purchased from the market; 63 black teas and 38 green teas.
The samples were stored in their original bags at room temperature
before analysis. Trademarks and producers were kept confidential
but the commercial names, origins and other characteristics of the
tea samples are provided in Supplementary material S1.
2.2. Analysis of tea volatiles by PTR-ToF-MS
The volatile compounds of dry tea leaves and their infusions
were analysed with PTR-ToF-MS by direct injection headspace
analysis without destructing the original samples. For the analysis
of dry tea leaves, 500 mg tea leaves were weighted into 22-ml
glass vials (Supelco, Bellefonte, PA) and 3 replicates were prepared
for each tea sample. Tea brewing was performed by applying a
3 min fixed infusion time for all tea samples. Deionized hot water
(25 ml, 85 °C) was used for brewing of tea leaves (400 mg) in
40 ml amber vials (Supelco, Bellefonte, PA). The liquid infusion was
taken right after brewing by a micropipette and 2 ml of aliquots
were transferred into 22-ml glass vials. Each tea sample was
brewed 3 times and each brew was analysed in duplicate.
The headspace measurements were performed by using a
commercial PTR-ToF-MS 8000 instrument (Ionicon Analytik
GmbH, Innsbruck, Austria). The instrumental conditions in the
drift tube were set as following: drift voltage 550 V, drift temperature 110 °C, drift pressure 2.33 mbar affording an E/N value
(electric field strength/gas number density) of 140 Townsend (Td,
1 Td ¼10 # 17 V cm2). All the vials containing samples and blank
vials (air for tea leaves and hot water for tea brews) were incubated at 37 °C for 30 min before headspace analysis. The headspace mixture was directly injected into PTR-MS drift tube with a
flow rate of 40 sscm via a PEEK tube at 110 °C. Sample injection
was performed with a multipurpose autosampler (Gerstel GmbH,
Mulheim am Ruhr, Germany). A different sample was analysed
every 5 min. Each sample was measured for 30 s, at an acquisition
rate of one spectrum per second. The measurement order was
randomized while measuring the volatile emissions of tea leaves
and tea brews.
2.3. Data processing and analysis
2.3.1. Treatment of mass spectrometric data
Data processing of ToF spectra included dead time correction,
internal calibration and peak extraction steps performed according
to a procedure described elsewhere [17] to reach a mass accuracy
(Z0.001 Th) which is sufficient for sum formula determination.
The baseline of the mass spectra was removed after averaging the
whole measurement and peak detection and peak area extraction
was performed by using modified Gaussian to fit the data [18].
Whenever a peak was detected, the volatile concentrations were
calculated directly via the amount of detected ions in ppbv (part
per billion by volume) levels according to the formulas described
by Lindinger et al. [13] by assuming a constant reaction rate
coefficient (kR ¼ 2 $ 10 # 9 cm3/s). For H3O þ as a primary ion, this
introduces a systematic error for the absolute concentration for
each compound that is in most cases below 30% and can be accounted for if the actual rate constant is available [19].
2.3.2. Selection of mass peaks
The direct injection headspace analysis of tea (leaves and infusion) samples resulted in identifying 455 mass peaks in the
range 15–300 m/z. After eliminating the interfering ions (O2 þ ,
NO þ and water clusters) and their isotopologues, 438 mass peaks
remained for further analysis. The signals belonging to blank vials
were subtracted from the whole data set (air from tea leaf emissions, water from infusion emissions). A concentration threshold
of 0.1 ppb was set for further reduction of noise in the data
47
S. Yener et al. / Talanta 152 (2016) 45–53
matrices. After this step 257 mass peaks ! 303 (i.e. 101 samples,
three biological replicates) data points were left to build the matrix containing tea leaf emissions; 162 mass peaks ! 606 (i.e. 101
samples, three infusions, two analytical replicates) data points
were left for tea infusion data matrix. These final data matrices
were used for univariate and multivariate data analysis methods.
After mass peak selection and extraction, tentative peak identification was performed by using an in-house library developed
by the authors where the peak annotations were done automatically with the scripts developed under R programming language [20].
2.3.3. Statistical analyses
The significant differences between tea types were calculated
using ANOVA (99% confidence level) and the pairwise comparison
was performed with Tukey’s test to highlight these differences
with letter annotations.
As a first step, the final data matrices were subjected to principal component analysis (PCA). Secondly, Random Forests (RF),
Penalized Discriminant Analysis (PDA), Support Vector Machines
(SVM) and Discriminant Partial Least Squares (dPLS) classification
methods were applied for sample discrimination [21] and their
classification power was compared.
Two types of validation methods were tested for each classification method: a simple 6-fold cross validation and Leave-GroupOut (LGO) cross-validation. The six-fold cross-validation was performed by randomly dividing the whole data set into 6 folds. One
of the folds was removed at each time and used as a test set where
the rest of the data (the train set) was used to build the discriminant method and predict the origins of samples. Using this
cross validation method, analytical or biological replicates of the
same tea sample can be at the same time in both the train and test
sets. With the highly flexible classification methods used in this
work, this can easily leads to overfitting the data and to produce
biased estimates of classification errors. This effect was verified in
preliminary experiments (not shown) and the method was discarded. In the case of the more elaborated LGO cross-validation,
the analytical and biological replicates of each tea sample were
considered as a group when discriminating tea types and geographical origin. Each time, one group was removed from the full
dataset and used as a test set. Mean classification errors and
confusion matrices were used to evaluate the performance of each
classification method. All the multivariate data analyses were
performed by using the scripts and packages developed under R
programming language [20].
3. Results and discussion
3.1. Volatile profiling of black and green teas and discrimination
based on tea type
One-way ANOVA of the mass peaks extracted in black and green
tea headspace, showed 135 mass peaks significantly different
(po0.01 with Bonferroni correction) between emissions of black and
green tea leaves and 125 mass peaks between their infusions. Among
the mass peaks extracted, 62 of them were tentatively identified as
one or more volatile compounds based on their presence in dry tea
leaves and brews reported in literature. The details of the tentatively
identified mass peaks are shown in Table 2 with their average concentrations in black and green tea leaves and infusions.
The leaves of different tea types showed greater volatile
emissions as compared with infusions. Various terpenes and their
fragments dominated the volatile emission of tea leaves, followed
by esters/acids and aldehydes/ketones. In particular, green tea
leaves emitted more terpenes and sulphur compounds than black
teas. The most abundant volatile compounds in the headspace of
green tea infusions were sulphur compounds, aldehydes/ketones
and terpenes. The headspace of black tea infusions contained aldehydes/ketones the highest; sulphur compounds, terpenes and
alcohols were other most abundant chemical groups.
Some distinct differences and similarities can be pointed out
between black and green teas: the most abundant sulphur compound detected in both tea infusions was tentatively identified as
dimethyl sulphide. It has been reported that this sulphur compound improves the flavour of green teas harvested in spring [3].
The information about the season when the green teas were
picked was not available for all the tea samples but for some of the
black teas. Interestingly, we observed that the black teas that had
the highest dimethyl sulphide contents were indeed picked during
spring (e.g. sample no 102, 110, and 116 in Supplementary file 1).
We observed that the percentage of total monoterpenes and
their fragments in the headspace of black tea infusions ( "20%)
was higher than the amount emitted from green tea infusion
( "12%). Terpenes, especially monoterpenes, are responsible for
the characteristic floral odour of tea [22]. Important aroma compounds derived from breakdown of carotenoids during black tea
processing like linalool, geraniol, linalool oxide and ionone [3]
were also higher in the headspace of black teas and their infusions
than in green teas. Most of the monoterpenes and derived compounds were significantly lost during tea brewing; in particular
linalool oxide (m/z 171.133) in green tea infusions.
Vanillin was previously reported to be one of the compounds of
highest flavour dilution factor (FD) in black tea infusion [23]. In our
study, the peak corresponding to vanillin was negligible in green
tea infusions, but clearly observable in black teas with little effect
of brewing.
When PCA was performed, the first three principal components
provided a good separation of black teas from green teas based on
the volatile emissions both from dry leaves and infusions (Fig. 1a
and b). The first PCs explain 53.2 and 54.7% variances for the dry
tea leaves and infusions, respectively. This reflects the high variance between black and green tea volatile emissions as well as
within each tea type (black or green) depending on the different
production methods and origins. The release mechanisms of volatiles might be influenced by matrix characteristics (i.e. leave
shape and size) as teas can be produced in various shapes. For
example, the green teas can be shaped like needle, twisted, flat,
round, compressed shape or even as ground powder as a results of
fixing and drying methods. Besides, leaf disruption also occurs in
cutting and rolling steps of black tea production that leads to
grading of black teas according to leaf size [22].
Fig. 2a and b show score plots of the first two PCs of tea leaves and
tea infusions (loadings of the first two components of tea leaves and
infusions are provided in Supplementary file S2). According to these
Fig. 2a and b, some black tea samples with broken leaves (sample
numbers 1, 10, 30, 43, 112, 146-148) were closely located and separated
from others. These samples were characterized by the mass peaks at
m/z 59.049, 85.065, 97.065, 99.081, 111.081, 113.096, 115.074, 115.112,
139.113, 141.127 and 143.144 which were mostly attributed to aldehydes and ketones; mass peaks at m/z 101.096 and 87.080 to alcohols
and mass peak at m/z 169.126 to geranic acid in the headspace of dry
tea leaves. In addition, mass peaks; m/z 71.049 (butenal), 77.058
(propandiol), 129.099 (hexenyl formate), 127.112 (methylheptenone)
had high loadings in the headspace of tea infusions with broken leaf
shape. Broken and smashed tea may release more catechins than
firmly pressed tea leaves and they may undergo heavier oxidation
[24]. Broken leaves also provide a larger surface area during fermentation favouring enzymatic (i.e. glycosidases, fatty acid hydroperoxide
lyase) activity for production of volatile aldehydes [25]. These findings
indicate the importance of leaf shape on volatile compound generation
and their extraction during the infusion process.
79
80
Theoretical
mass
33.033
45.034
47.049
49.011
59.049
61.028
63.026
69.034
69.070
71.049
73.065
75.044
79.054
81.070
83.086
85.065
87.044
87.080
91.058
93.037
93.070
95.016
95.049
95.086
96.081
97.028
97.065
99.080
101.096
103.075
105.037
105.070
107.049
107.086
109.065
109.101
111.044
Measured
mass
33.0336
45.0333
47.0491
49.0110
59.0488
61.0280
63.0260
69.0333
69.0697
71.0489
73.0646
75.0438
79.0536
81.0697
83.0854
85.0646
87.0431
87.0802
91.0559
93.0365
93.0698
95.0173
95.0478
95.0854
96.0814
97.0282
97.0647
99.0803
101.0960
103.0755
105.0343
105.0689
107.0488
107.0855
109.0658
109.1013
111.0466
þ
þ
C7H8OH þ
C8H12H þ
C6H6O2H þ
C7H6OH
C8H10H þ
C8H8H þ
C5H4O2H þ
C6H8OH þ
C6H10OH þ
C6H12OH þ
C5H10O2H þ
C4H8OSH þ
C6H9NH þ
C6H6OH
C7H10H þ
þ
C2H6O2SH
C7H8H
þ
C3H8OSH þ
C6H8H þ
C6H10H þ
C5H8OH þ
C4H6O2H þ
C5H10OH þ
C4H10SH þ
C4H4OH þ
C5H8H þ
C4H6OH þ
C4H8OH þ
C3H6O2H þ
C6H6H þ
Heterocyclic
compounds
Aldehydes
Aldehydes/Furans
Aldehydes
Alcohols
Esters and acids
Sulphur
compounds
Aromatic
hydrocarbons
Aldehydes
Aromatic
hydrocarbons
Phenols
Hydrocarbons
Furans
Alcohols
Aldehydes
Alcohols
Sulphur
compounds
Aldeydes/ketones
Esters and acids
Sulphur
compounds
Furans
Terpene fragment
Aldehydes
Aldehydes
Esters and acids
Aromatic
hydrocarbons
Terpene fragment
Terpene fragment
Aldehydes/Ketones
Ketones
Alcohols
Sulphur
compounds
Sulphur
compounds
Aromatic
hydrocarbons
Sulphur
compounds
Phenols
Terpenes
CH4OH þ
C2H4OH þ
C2H5OH þ
CH4SH þ
C3H6OH þ
C2H4O2H þ
C2H6SH þ
Chemical class
Sum formula
6.5 7 4.8
877 121
3.2 7 1.9
257 29
3.8 7 3.5
337 15
58 7 43
Benzaldehyde
Xylene/ethylbenzene
Benzyl alcohol (cresol)
Cyclooctadiene
Acetyl furan
137 5
137 12
197 23
197 16
247 28
127 9
2.0 7 0.8
Styrene/ethylbenzene/vinylbenzene
Furfural
Hexadienal/ethylfuran
Hexenal/methylpentenone
Hexenol
Methylbutanoic acid
Methional
3.8 7 2.1
64 761
2.4 7 0.9
Dimethyl sulfone
(methylsulfonylmethane)
Phenol
Methylcyclohexadiene (α-terpinene
fragment)
Ethylpyrrole
7.1 78.2
1.3 7 0.7
557 88
1.8 7 1.3
21 717
2.2 7 1.6
137 9
437 37
1.4 7 0.8
2.9 7 1.9
197 24
157 9
107 13
117 11
0.4 7 0.4
5.9 7 4.4
347 34
5.6 7 2.6
Toluene
423 7 461
607 70
37 737
6.4 7 4.6
627 157
4.4 7 2.7
4.7 7 6.1
220 7 217
12 78
1957 565
427 42
267 22
265 7 784
390 7 848
12 712
9555 7 3826
265 7 260
1107 360
0.5 7 0.3
Green tea leaves
8.6 7 3.7
612 7905
1757 166
327 33
197 16
557 37
7.9 7 3.6
7.0 7 7.4
183 7145
197 17
206 7 143
967 103
407 18
340 7 479
804 7 605
107 8
11,756 7 3992
804 7 608
1387 152
0.9 7 0.5
Black tea leaves
Cyclohexadiene (Terpene fragment)
Cyclohexene (Terpene fragment)
Pentenal/pentenone
Butanedione
Pentenol
Diethylsulphide/butanethiol
(fragment)
Methylsulfanylethanol
Furan fragment
Isoprene
Butenal
Methylpropanal
Propionic acid
Benzene
Propanal/acetone
Acetic acid
Dimethylsulfide
Methanol
Acetaldehyde
Ethanol
Methanetiol
Tentative identification
Average concentration7standart deviation (ppbv)
170 7 82
416 7 158
6 7 14
0.4 7 0.5
Black tea infusion
1.8 7 2.7
25 7 26
21 7 22
13 7 11
1.1 7 1.4
0.2 7 0.2
1.0 7 0.7
9.4 7 5.9
9 7 16
0.5 7 0.3
4.0 7 3.5
0.3 7 0.7
o 0.001
0.797
0.009
o 0.001
0.414
o 0.001
o 0.001
o 0.001
0.002
o 0.001
0.154
o 0.001
0.3 7 0.4
4.2 7 3.2
o 0.001
0.051
2.1 7 1.8
0.2 7 0.3
o 0.001
0.144
5.9 7 7.5
1.7 7 1.2
o 0.001
0.014
29 7 24
71 7 63
12 7 9
2.7 7 3.1
165 7 89
0.9 7 0.9
0.035
o 0.001
0.229
o 0.001
0.549
o 0.001
0.004 7.2 7 7.8
0.067 145 7 76
o 0.001 4.6 7 4.0
0.811 292 7 159
o 0.001 9.7 7 6.1
o 0.001 5.3 7 7.3
0.2 7 0.2
2.3 7 2.2
0.2 7 0.4
1.8 71.8
87 15
0.1 7 0.3
0.4 7 0.7
8.3 7 8.6
4.6 7 2.8
3.7 7 4.5
0.7 7 1.1
n.d.
2.5 7 1.6
0.4 7 0.4
2.6 7 2.2
n.d.
2.2 7 2.3
0.4 7 0.6
97 12
197 23
7.5 77.6
1.3 70.7
237 14
0.3 7 0.5
2.0 7 3.0
357 21
2.2 7 2.3
437 26
4.5 7 2.6
4.1 7 7.7
927 85
6.34 7 17
2757 299
120 763
1087 58
117 18
1.2 71.5
Green tea infusion
Average concentration 7 standart deviation (ppbv)
0.293 115 7 99
o 0.001 2.3 7 4.8
0.191 264 7 289
o 0.001
o 0.001
0.345
o 0.001
p-Value
[28]
[28]
[28]
[28]
Reference
[28]
[28]
[24,29,32]
[29]
[24,29]
[33]
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
[32,33]
[24,32]
[24,29]
[32,33]
[23]
[30]
o0.001 [24,29,33]
o0.001 [29]
0.028 [24,33]
o0.001 [24,29,33,35]
0.398 [24,35]
o0.001 [9]
o0.001
o0.001
o0.001
o0.001
0.002
o0.001
0.002 [32]
0.188 [34]
o0.001 [33]
o0.001 n.a.
o0.001 [28,29,32]
o0.001 n.a.
[30]
n.a.
[31]
[29]
[28]
[29]
o0.001
o0.001
o0.001
o0.001
o0.001
0.043
0.002 [28]
o0.001 [29]
0.675 [28]
o0.001
o0.001
o0.001
o0.001
p-Value
Table 2
The average concentrations (ppb) of tentatively identified mass peaks in the headspace of black and green tea leaves and infusions. Peaks were selected on the basis of one-way ANOVA and the relative p-values are listed in the right
columns.
48
S. Yener et al. / Talanta 152 (2016) 45–53
123.117
127.112
129.091
129.127
131.107
135.117
136.112
137.133
139.112
141.127
143.143
151.112
153.055
153.127
155.143
171.138
193.159
195.088
123.1170
127.1116
129.0901
129.1276
131.1069
135.1170
136.1212
137.1321
139.1124
141.1271
143.1435
151.1114
153.0550
153.1272
155.1430
171.1332
193.1587
195.0879
C10H16H þ
C9H14OH þ
C9H16OH þ
C9H18OH þ
C10H14OH þ
C8H8O3H þ
C10H16OH þ
C10H18OH þ
C10H18O2H þ
C13H20OH þ
C8H10N4O2H þ
C9H13NH þ
C9H14H þ
C8H14OH þ
C7H12O2H þ
C8H16OH þ
C7H14O2H þ
C10H14H þ
C7H10OH þ
C7H12OH þ
C6H10O2H þ
C7H14OH þ
C7H4O2H þ
C8H8OH þ
C9H12H þ
n.a.: Not available, n.d.: Not detected.
111.080
113.096
115.074
115.112
121.028
121.065
121.101
111.0805
113.0960
115.0738
115.1119
121.0291
121.0648
121.1004
Aldehydes
Aldehydes
Ketones
Ketones
Terpenes
Aldehydes
Aromatic
hydrocarbons
Terpenes
Ketones
Esters and acids
Ketones
Esters and acids
Aromatic
hydrocarbons
Heterocyclic
compounds
Terpenes
Aldehydes
Aldehydes
Ketones/Aldehydes
Terpenes
Aldehydes
Aldehydes
Alcohols
Terpenes
Terpenes
Ketones
Various monoterpenes
Nonadienal
Nonenal
Nonanone/nonanal
Carvacrol/safranal
Vanillin, methyl salicylate
Decadienal
Linalool/geraniol
Linalool oxide
Β-ionone
Caffeine
Butyl-pyridine/ethyl-propylpyridine
Santene
Octenone/methylheptenone
Hexenyl formate
Octanone/Dimethylcyclohexanol
Heptanoic acid/hexyl formate
Methylpropylbenzene
Heptadienal
Heptenal
Caprolactone
Heptanone
cyclohexadienone (fragment)
Methylbenzaldehyde-coumaran
Methylethylbenzene
368 7 548
27 716
3.2 7 1.5
0.9 7 0.9
1.0 7 0.4
6.17 4.0
297 26
2.1 7 1.5
2.9 7 2.6
0.4 7 0.2
n.d.
3.7 7 4.9
5.6 7 2.3
97 11
1.7 7 1.1
2.9 7 2.7
9.9 7 9.4
127 14
177 20
3.3 7 2.5
2.5 7 1.7
107 10
0.8 7 0.4
2.5 7 1.2
8.7 7 6.9
252 7 277
197 12
2.5 7 1.2
0.6 7 0.4
0.9 7 0.5
n.d
4.4 7 2.9
1.17 1.5
0.2 7 0.1
0.3 7 0.2
0.2 7 0.1
2.3 7 2.3
5.9 7 2.3
7.3 7 6.2
0.8 7 1.1
2.1 71.8
8.2 7 11.0
6.2 7 4.9
207 24
2.6 7 2.2
1.2 7 1.0
5.5 7 5.9
n.d.
1.17 0.6
7.6 7 6.1
0.033
o 0.001
o 0.001
0.002
0.365
o 0.001
o 0.001
o 0.001
o 0.001
0.141
o 0.001
0.005
0.242
0.030
o 0.001
0.002
0.138
o 0.001
0.213
0.010
o 0.001
o 0.001
o 0.001
o 0.001
0.134
13 7 14
2.8 7 1.5
0.9 7 0.6
0.8 7 0.8
0.3 7 0.13
4.2 7 3.4
2.9 7 1.9
0.6 7 0.4
0.2 7 0.2
0.2 7 0.2
0.2 7 0.1
0.2 7 0.2
1.2 7 0.5
2.6 7 2.3
0.4 7 0.3
1.2 7 1.1
2.3 7 2.1
1.0 7 1.0
7.3 7 7.2
1.6 7 1.3
0.2 7 0.1
3.6 7 2.9
0.2 7 0.3
2.3 7 1.3
0.8 7 0.4
5.6 7 7.5
1.3 70.9
0.5 7 0.3
0.3 7 0.3
0.2 7 0.1
n.d.
0.6 7 0.5
0.2 7 0.3
n.d.
0.2 7 0.2
n.d.
n.d.
0.7 7 0.4
1.2 71.1
n.d.
0.4 7 0.3
1.17 1.7
0.4 7 0.4
6.4 7 7.4
0.7 7 0.7
0.1 7 0.1
1.2 71.1
n.d.
0.5 7 0.6
0.5 7 0.5
n.a.
[24,29]
[32]
[32,33]
[24,32,33]
[29]
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
0.626
o0.001
[29,32,33,36]
[29,33,36]
[29,36]
[24,29]
[24,33,36]
[33,36,37]
[36]
[29,33,35,36]
[24,29,36]
[29,33,35,36]
[29]
o0.001 n.a.
[24,36]
[24,29,32]
[29]
[28]
[24,29,33]
[24,32,35]
[29]
0.120
o0.001
o0.001
o0.001
o0.001
o0.001
o0.001
S. Yener et al. / Talanta 152 (2016) 45–53
49
81
50
S. Yener et al. / Talanta 152 (2016) 45–53
10
20
10
0
10
−10
−20
20
10
0 −10 −20 −30
0
−10
)
−20
%
.0
−30
6
3
−40
1(
−50
C
P
PC 3 (9.9 %)
PC 3 (6.6 %)
PC 2 (17.2 %)
30
5
0
30
20
10
−5
0
−10
−10
15
10
5
0
−5
−10
−20
PC
1
2
2.
(4
%
)
PC 2 (12.5 %)
Fig. 1. 3D PCA score plots of black and green tea leaves (a) and tea infusions (b). Black and green colours represent black and green teas, respectively. (For interpretation of
the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 2. 2D PCA score plots of black and green tea leaves and infusions. Black and green colours represent black and green teas, respectively. Due to the good repeatability of
the analytical replicates, PCA was built via averaging the replicates. This improved the visualization of each sample. The numbers on the points indicate the sample codes
given in Supplementary file S1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
PCA showed a relatively good separation of black and green teas by
using three principal components (Fig. 1a–b, a–b). To be able to assess
the performance of discrimination, we applied 4 classification models
by using two different cross validation methods for discrimination of
black and green teas. According to LGO cross validation the average
errors for classification black and green tea leaves were 2.6%, 3.9%, 1.3%
and 3.6%; the average errors for classification black and green tea infusions were 0.6%, 0.2%, 0.0% and 0.5% obtained by RF, PDA, SVM and
dPLS classification models, respectively. In general, all the classification
techniques showed very good classification efficiency with an average
error rate less than 4.0% for differentiating black and green tea volatile
profiles emitted from leaves and infusions. In all cases, the classification errors were lower for tea infusions than tea leaves.
3.2. Geographical origin discrimination with supervised classification methods
The results described above highlighted significant differences
between black and green tea aroma profiles and successful
82
separation of large number of tea samples according to tea type.
However it would more relevant to demonstrate that the volatile
composition of tea might be related to its geographical origin, as
well. For this reason, we applied supervised classification methods
on the black and green tea volatile profiles in order to differentiate
them according to their origins.
To get a more representative data set for classification studies, we
selected origins (countries) represented by at least 4 different teas.
Black teas from China, India, Sri Lanka and Nepal (50 samples) were
included for classification of black teas; China, India, Japan and Korea
(32 samples) were selected for classification of green teas. Each classification algorithm ended up with an average classification error and a
confusion matrix where the original tea origins were compared with
the origins assigned by the classification method. The classification
methods were applied on normalized volatile concentrations with LGO
cross validation tests. The normalized concentrations were obtained by
normalizing each mass spectrum to unit area as described in [26].
The classification performances obtained by using emissions of
the tea leaves and tea infusions were similar and they provided
51
S. Yener et al. / Talanta 152 (2016) 45–53
Table 3
Confusion matrices showing the origin separation of black and green tea infusions for leave-group-out cross validation obtained by random forests (RF), penalized discriminant analysis (PDA), support vector machine (SVM) and discriminant partial least squares (dPLS) classification models.
Classification method
Black tea
Average error rate (%)
Chi
Ind
Sri
Nep
Chi
Ind
Sri
Nep
58
9
0
12
19
146
24
35
0
0
0
0
1
1
0
1
32.4
Chi
Ind
Sri
Nep
55
2
0
0
11
108
19
24
4
30
5
0
8
16
0
24
Chi
Ind
Sri
Nep
57
4
0
1
16
105
18
26
1
22
6
0
Chi
Ind
Sri
Nep
44
2
0
0
12
101
19
27
7
29
5
0
Green tea
Average error rate (%)
Chi
Ind
Jap
Kor
Chi
Ind
Jap
Kor
74
15
10
16
6
0
3
0
4
9
42
8
0
0
5
0
39.6
36.5
Chi
Ind
Jap
Kor
69
5
17
5
2
5
0
0
10
14
28
16
3
0
15
3
45.3
4
35
0
21
37.5
Chi
Ind
Jap
Kor
62
7
17
8
5
14
0
0
14
3
31
12
3
0
12
4
42.2
15
24
0
21
43.3
Chi
Ind
Jap
Kor
67
7
17
6
0
4
0
0
12
13
24
14
5
0
19
4
48.4
RF
PDA
SVM
dPLS
relatively good separations which were between 30-50%. Due to
the fact that, the tea infusions are the final consumed products, in
the following discussion, we will focus on the classification models
tested for black and green tea infusions.
Table 3 shows confusion matrices and the performances of the
classification models applied on black and green tea infusions.
Among the 50 black teas from 4 countries tested, the lowest prediction error was around 32% obtained by RF; the same method
also provided the highest classification performance for green tea
infusions prepared by 32 samples from 4 countries. Among the
black teas, teas from Sri Lanka were classified with lowest errors
followed by India, China and Nepal. In the case of green teas,
Chinese teas had the lowest error followed by Japan, India and
Korea. The confused tea samples were mostly from the neighbouring countries. For instance, Korean green teas were confused
with Chinese and Japanese green teas but not with Indian green
teas with RF method. This finding is not surprising because political borders are not likely to affect tea quality while climate,
growing conditions, picking method and processing traditions
[3,22,27] are the key factors for differentiating tea classes and their
characteristics. Unfortunately we were not able to find better
geographical indications for many samples.
Similar cases have been reported in literature with various
classification performances when different tea samples were discriminated according to geographical origin based on their volatile
profiles. Togari, Kobayashi and Aishima (1995) [10] performed the
first study on the geographical origin determination of different
tea categories based on their volatile profiles. The study included
GC–MS analysis of 44 tea samples where tea volatiles were extracted by simultaneous dilution and extraction (SDE) method by
mixing the tea samples with hot water. Black teas from India (8),
Sri Lanka (4) and Japan (1) were successfully classified by supervised pattern recognition techniques, however neither oolong
(China (10) and Taiwan (4)) nor green teas (15, from different regions of Japan) could be classified according to origin. Kovács et al.,
[8] applied electronic nose technology with electronic tongue and
sensory assessment for geographical origin discrimination of five
Sri Lankan teas. When electronic tongue responses of tea infusion
headspace was treated with linear discriminant analysis, 100%
correct classification was obtained for middle and low elevation
regions (n ¼3) however two samples from high elevation showed
overlapping. Ye, Zhang and Gu [11] analysed volatile profiles of 23
green tea samples produced in two different regions of China with
SPME/GC–MS via extracting the volatiles from tea powder. They
could classify the production areas of tea samples. Lee et al. [9]
analysed 24 green tea samples from 8 different countries (China
(7), India (1), Japan (6), Kenya (2), Korea (4), Sri Lanka (2), Tanzania
(1) and Vietnam (1)) with GC-SPME method nevertheless no relationship has been found between country of origin and aroma
where specific information about the samples other than origin
was not known for the tested tea samples. In another study, 38 tea
samples from China (2 oolong, 2 green, 3 black), Japan (5 green,
3 black, 2 oolong), India (5 black), Sri-Lanka (5 black), and Chinese
Taipei (6 oolong, 2 black) were analysed by GC–MS and they were
classified according to their origins by clustering methods [6].
Lastly, four varieties of oolong teas were analysed by olfaction and
gustation sensing systems, the samples were classified according
to producing regions by using the information each sensing system provided. When all information was merged with data fusion
techniques, the discrimination power increased compared to individual classification performances suggesting the possibility to
use these systems with multivariate methods for discriminating
and classifying tea samples [7].
When our results and the literature were considered together,
different tea types from various countries can be discriminated to
some extent according to geographical origin based on their volatile emissions from dry tea leaves or tea infusions. Moreover
fermented tea products are better classified than non-fermented
and semi-fermented teas which was also observed from our results when we compare the classification efficiencies of black and
green teas.
When black tea infusions from India (Assam (9) and Darjeeling
(12)), China (Anhui (3) and Yunnan (7)), Sri Lanka (all country) and
Nepal (all country) were classified according to tea producing
83
52
S. Yener et al. / Talanta 152 (2016) 45–53
regions a significant improvement on the classification performance was observed providing 15% average error rate (confusion
matrices not shown). The results indicated 4 classes: China Anhui
(class 1); Sri Lanka and India Assam (class 2) Nepal and Darjeeling
(class 3) and China Yunnan (class 4) by showing the geographically
close regions in the same group.
Overall, these findings point out that the regions might be
better differentiators instead of the country and the regions closely
located to each other share more similar properties and they are
likely to create a group. Besides, there might be other factors affecting the volatile composition of different types of tea in addition to geographical location such as the age of the tea plant,
plucking (fine or coarse), plucking season, tea processing, packaging of tea, conditions during storage and storage time, which
should be taken into consideration.
4. Conclusions
In this study, for the first time, the volatile profiles of black and
green teas from 12 different geographical origins were analysed by
PTR-ToF-MS. The volatile compounds of a large sample set (101
samples, with replicates, both leaves and infusions) were analysed
by direct injection of the headspace without altering the original
tea components and destructing the original sample. The high
mass resolution and sensitivity achieved by the mass analyser
enabled annotation of sum formulas to the detected mass peaks.
Tentative identifications lead defining important aroma compounds in black and green tea volatile emissions and pointed out
the differences among them.
Black and green teas were correctly classified by the volatile
compounds emitted from tea leaves and their infusions independent from their geographical origins. Classification models
were built to predict the geographical origins of black and green
teas. Results provided a good separation of tea origins; however
countries geographically close to each other were most likely to be
confused. Preliminary analysis indicated that a better discrimination of tea samples might have been achieved if teas were classified according to production region rather than just country of
origin. This was not feasible here, since information about production region was available only for a limited number of samples.
Our results showed that PTR-ToF-MS fingerprints combined
with multivariate statistical techniques provided successful evaluation of tea products. Considering the very promising results
obtained so far, in discriminating for processing and country, it
seems highly warranted to collect significantly more detailed information about the individual tea samples, for future studies. This
may include e.g. information on production region, producer,
harvesting season, post-harvest treatment and age of the product.
It may also be significant to investigate the effect of tea leave
shape and infusion conditions. Finally, it is also important to direct
our interest towards the consumer, by analysing the volatile
compounds release from the nosespace and analysed by PTR-ToFMS, when a tea product is being consumed, and conducting sensory profiling as well. Combining such a large spectrum of different data sets might currently seem to be a veritably challenging
task; we believe this will need to be approached in steps towards a
more complete understanding of the factors affecting tea aroma
profiles.
Acknowledgements
This research has been funded by PIMMS (Proton Ionisation
Molecular Mass Spectrometry) ITN (287382), which is supported
84
by the European Commission's 7th Framework Programme under
Grant Agreement number 287382.
Appendix A. Supplementary material
Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.bios.2014.05.063.
References
[1] L. Zhang, Z. Zeng, C. Zhao, H. Kong, X. Lu, G. Xu, A comparative study of volatile
components in green, oolong and black teas by using comprehensive twodimensional gas chromatography-time-of-flight mass spectrometry and multivariate data analysis, J. Chromatogr. A 1313 (2013) 245–252.
[2] FAOSTAT, Available from: 〈http://faostat3.fao.org/browse/Q/*/E〉, (accessed on
02.11.15).
[3] Z.M. Chen, H.F. Wang, X.Q. You, N. Xu, in: Y.S. Zhen (Ed.), Tea Bioactivity and
Therapeutic Potential, Taylor & Francis, New York, USA, 2005, pp. 89–120.
[4] P. Pripdeevech, S. Wongpornchai, in: V.R. Preedy (Ed.), Tea in Health and
Disease Prevention, Elsevier, USA, 2013, pp. 307–322.
[5] Z. Yang, S. Baldermann, N. Watanabe, Recent studies of the volatile compounds
in tea, Food Res. Int. 53 (2013) 585–599.
[6] S. Baldermann, Z. Yang, T. Katsuno, V.A. Tu, N. Mase, Y. Nakamura,
N. Watanabe, Discrimination of green, oolong, and black teas by GC–MS
analysis of characteristic volatile flavor compounds, Am. J. Anal. Chem. 05
(2014) 620–632.
[7] Q. Chen, C. Sun, Q. Ouyang, Y. Wang, A. Liu, H. Li, J. Zhao, Classification of
different varieties of Oolong tea using novel artificial sensing tools and data
fusion, LWT-Food Sci. Technol. 60 (2015) 781–787.
[8] Z. Kovács, I. Dalmadi, L. Lukács, L. Sipos, K. Szántai-Kőhegyi, Z. Kókai, A. Fekete,
Geographical origin identification of pure Sri Lanka tea infusions with electronic nose, electronic tongue and sensory profile analysis, J. Chemom. 24
(2010) 121–130.
[9] J. Lee, D.H. Chambers, E. Chambers, K. Adhikari, Y. Yoon, Volatile aroma
compounds in various brewed green teas, Molecules 18 (2013) 10024–10041.
[10] N. Togari, A. Kobayashi, T. Aishima, Pattern recognition applied to gas chromatographic profiles of volatile components in three tea categories, Food Res.
Int. 28 (1995) 495–502.
[11] N. Ye, L. Zhang, X. Gu, Discrimination of green teas from different geographical
origins by using HS-SPME/GC–MS and pattern recognition, Methods Food
Anal. Methods 5 (2011) 856–860.
[12] G. Reineccius, Flavor analysis, in: G. Reineccius (Ed.), Flavor Chemistry and
Technology, second edition,CRC Press, FL, USA, 2005, pp. 33–72.
[13] W. Lindinger, A. Hansel, A. Jordan, On-line monitoring of volatile organic
compounds at pptv levels by means of proton-transfer-reaction mass spectrometry (PTR-MS) medical applications, food control and environmental research, Int. J. Mass Spectrom. Ion Process. 173 (1998) 191–241.
[14] R.S. Blake, P.S. Monks, A.M. Ellis, Proton-transfer reaction mass spectrometry,
Chem. Rev. 109 (2009) 861–896.
[15] R.S. Blake, C. Whyte, C.O. Hughes, A.M. Ellis, P.S. Monks, Demonstration of
proton-transfer reaction time-of-flight mass spectrometry for real-time analysis of trace volatile organic compounds, Anal. Chem. 76 (2004) 3841–3845.
[16] F. Biasioli, F. Gasperi, C. Yeretzian, T.D. Märk, PTR-MS monitoring of VOCs and
BVOCs in food science and technology, TrAC Trends Anal. Chem. 30 (2011)
968–977.
[17] L. Cappellin, F. Biasioli, A. Fabris, E. Schuhfried, C. Soukoulis, T.D. Mark, F. Gasperi,
Improved mass accuracy in PTR-TOF-MS: another step towards better compound
identification in PTR-MS, Int. J. Mass Spectrom. 290 (2010) 60–63.
[18] L. Cappellin, F. Biasioli, P.M. Granitto, E. Schuhfried, C. Soukoulis, F. Costa, T.
D. Mark, F. Gasperi, On data analysis in PTR-TOF-MS: From raw spectra to data
mining, Sens. Actuator B-Chem. 155 (2011) 183–190.
[19] L. Cappellin, T. Karl, M. Probst, O. Ismailova, P.M. Winkler, C. Soukoulis,
E. Aprea, T.D. Märk, F. Gasperi, F. Biasioli, On quantitative determination of
volatile organic compound concentrations using proton transfer reaction
time-of-flight mass spectrometry, Environ. Sci. Technol. 46 (2012) 2283–2290.
[20] R. Core Team, in: R Foundation for Statistical Computing, 2014.
[21] F. Biasioli, P.M. Granitto, E. Aprea, D. Mott, C. Furlanello, T.D. Mark, F. Gasperi,
Rapid and non-destructive identification of strawberry cultivars by direct PTRMS headspace analysis and data mining techniques, Sens. Actuators B Chem.
121 (2007) 379–385.
[22] Z.M. Chen, N. Xu, in: Y.S. Zhen (Ed.), Tea Bioactivity and Therapeutic Potential,
Taylor & Francis, New York, USA, 2005, pp. 35–56.
[23] C. Schuh, P. Schieberle, Characterization of the key aroma compounds in the
beverage prepared from darjeeling black tea: quantitative differences between
tea leaves and infusion, J. Agric. Food Chem. 54 (2006) 916–924.
[24] L.F. Wang, J.Y. Lee, J.O. Chung, J.H. Baik, S. So, S.K. Park, Discrimination of teas
with different degrees of fermentation by SPME–GC analysis of the characteristic volatile flavour compounds, Food Chem. 109 (2008) 196–206.
[25] M.E. Harbowy, D.A. Balentine, Tea chemistry, Crit. Rev. Plant Sci. 16 (1997)
415–480.
53
S. Yener et al. / Talanta 152 (2016) 45–53
[26] P.M. Granitto, F. Biasioli, E. Aprea, D. Mott, C. Furlanello, T.D. Maerk, F. Gasperi,
Rapid and non-destructive identification of strawberry cultivars by direct PTRMS headspace analysis and data mining techniques, Sens. Actuators B 121
(2007) 379–385.
[27] M. Johns, The Evaluation of Volatile Quality Factors in Black Tea : A Report for
the Rural Industries Research and Development Corporation, in: RIRDC Publication, 1440-6845; no. 98/46, Rural Industries Research and Development
Corporation, Barton, A.C.T, 1998, pp. 1–58.
[28] T. Yamanishi, The aroma of various teas, in: G. Charalambous, G.E. Inglett
(Eds.), Flavor of Foods and Beverages, Chemistry and Technology Academic
Press, New York, USA, 1978, pp. 305–328.
[29] C. Alasalvar, B. Topal, A. Serpen, B. Bahar, E. Pelvan, V. Gokmen, Flavor characteristics of seven grades of black tea produced in Turkey, J. Agric. Food
Chem. 60 (2012) 6323–6332.
[30] K. Kumazawa, H. Masuda, Identification of potent odorants in japanese green
tea (Sen-cha), J. Agric. Food Chem. 47 (1999) 5169–5172.
[31] M. Kawakami, T. Shibamato, The volatile constituents of piled tea: Toyama
Kurocha, Agric. Bioi. Chem., 55, (1991) 1839–1847.
[32] W. Renold, R. Näf-Müller, U. Keller, B. Wilhalm, G. Ohloff, An investigation of
the tea aroma Part I. New volatile black tea constituents, Helv. Chim. Acta 57
(1974) 1301–1308.
[33] P. Pripdeevech, T. Machan, Fingerprint of volatile flavour constituents and
antioxidant activities of teas from Thailand, Food Chem. 125 (2011) 797–802.
[34] T. Yamanishi, M. Kosuge, Y. Tokitomo, R. Maeda, Flavor constituents of pouchong tea and a comparison of the aroma pattern with jasmine tea, Agric. Biol.
Chem. 44 (1980) 2139–2142.
[35] E.S. Kim, Y.R. Liang, J. Jin, Q.F. Sun, J.L. Lu, Y.Y. Du, C. Lin, Impact of heating on
chemical compositions of green tea liquor, Food Chem. 103 (2007) 1263–1267.
[36] P. Schieberle, C. Schuh, Aroma compounds in black tea powders of different
origins-changes induced by preparation of the infusion, in: Wender L.
P. Bredie, Mikael Agerlin Petersen (Eds.), Flavour Science: Recent Advances
and Trends, 2006, pp. 151–156.
[37] H. Chi-Tang, Z. Xi, L. Shiming, Tea aroma formation, Food Sci. Hum. Wellness 4
(2015) 9–27.
85
Publication 4.
Extraction Dynamics of Tea Volatile Compounds as a Function of Brewing Temperature, Leaf Size
and Water Hardness: On-Line Analysis by PTR-ToF-MS.
by
José A. Sánchez-López, Sine Yener, Tilmann D. Märk, Günther Bonn, Ralf Zimmerman, Franco
Biasioli , Chahan Yeretizian.
Submited to Talanta
on 06.04.2016
José A. Sánchez-López was involved in the design and execution of the experiments. He performed data
analysis and prepared the manuscript. His work to this publication accounts for approximatelly 70%.
86
Title: Extraction Dynamics of Tea Volatile Compounds as a Function of Brewing Temperature,
Leaf Size and Water Hardness: On-Line Analysis by PTR-ToF-MS
Authors
José A. Sánchez-López1,2, Sine Yener3,4, Tilmann D. Märk5, Günther Bonn4, Ralf Zimmermann2,6, Franco
Biasioli3 ,Chahan Yeretzian1
Affiliations
1Zurich
University of Applied Sciences (ZHAW), Institute of Chemistry and Biotechnology, 8820
Wädenswil, Switzerland
2Joint
Mass Spectrometry Centre, Chair of Analytical Chemistry, Institute of Chemistry, University of
Rostock, D-18059 Rostock, Germany
3Department
of Food Quality and Nutrition, Research and Innovation Centre, Fondazione Edmund
Mach, San Michele all’Adige, Italy
4Leopold
Franzens University Innsbruck, Institute of Analytical Chemistry & Radiochemistry, Innsbruck,
Austria
5Leopold-Franzens
University of Innsbruck, Institute of Ion Physics and Applied Physics, Innsbruck,
Austria
6Joint
Mass Spectrometry Centre, Cooperation Group Comprehensive Molecular Analytics/CMA,
Helmholtz Zentrum München, D-85764 Neuherberg, Germany
Highlights
Extraction of volatiles into tea infusion was analyzed by PTR-TOF-MS
The impact of temperature, leaf size, water hardness and brewing time was determined
Samples with similar volatile profiles were identified and grouped
Grouping of profiles revealed differences in extraction dynamics between conditions
87
Abstract
Changes in the volatile profile of tea infusion during brewing were determined by analyzing the
headspace of aliquots taken every 30 seconds with PTR-ToF-MS (Proton Transfer Reaction Time-ofFlight Mass Spectrometry) coupled with Principal Component Analysis (PCA) and Hierarchical Cluster
Analysis (HCA). The effect of three different brewing temperatures (60, 70 and 80 °C), two leaf sizes
(broken and full leaves) and two water mineralizations (soft and hard) on the concentration of volatile
compounds in the head-space of tea was studied. An increase in brewing temperature resulted in
increased volatile content, with the differences on extraction efficiency becoming more pronounced at
longer brewing times. Leaf size had also a big impact on the extraction of volatile compounds, but mainly
during the beginning of the brewing. Water mineralization had low impact in the volatile content.
Furthermore, samples prepared with different combinations of brewing parameters but resulting in
analogous volatile profiles could be identified using HCA
Key words
PTR-MS, tea aroma, volatile extraction, tea infusion
1
INTRODUCTION
Tea, one of the world most popular beverages in the world, is obtained from the hot water infusion of
Camellia sinensis leaves. Tea leaves are classified in three main classes according to the way they have been
processed: green tea (unfermented), oolong tea (semifermented) and black tea (fermented). Green tea is
mainly produced and consumed in Asia, however its sensory characteristics and health promoting
properties [209–211] have contributed to its growing popularity and increasing consumption in the
Western world as well.
Green tea brewing is considered an art in some cultures and recommendations on brewing temperature,
brewing time, water-to-tea ratio or the number of repeated infusions (re-steeping) depend on the tea
variety, region and individual preferences. Extensive research has been performed on the effects of
brewing parameters on extraction kinetics of health-related green tea compounds such as flavanols,
flavonols, catechines or caffeine [77,79,81,84,212]. In the case of volatile aroma compounds, previous
studies have focused on the analysis of either the dry leaves or the final infusion with the aim of
characterizing the aroma profile of different tea varieties[58,72,76,213,214], discrimination of tea
according to variety[215], grade[69] , fermentation [67] or origin[207], and on the impact of some
parameters such as the temperature on the final volatile composition of the infusion[116]. To the best of
our knowledge, no work has been performed on the extraction dynamics of tea volatiles and how
different brewing parameters can influence the volatile profile of the tea infusion over the brewing time.
In this work, we investigated the effect of three different parameters - water temperature, leaf size and
water mineralization - on the extraction profiles of green tea volatiles by analyzing the headspace of
aliquots taken over the brewing time. Volatile compound concentration in the head-space of tea samples
was measured by Proton Transfer Reaction Time of Flight Mass Spectrometry (PTR-ToF-MS). The use
of this fast and highly sensitive direct injection mass spectrometry technique allowed the direct sampling
88
and monitoring of the headspace of tea infusions without the need for any pre-concentration step, which
could otherwise change the volatile composition of the samples[128,216]. PTR-ToF-MS allows not only
the recording of the overall mass spectral fingerprint of each sample, but, due to its high mass resolution,
also the assignment of a sum formula to the detected peaks and the tentative identification of
compounds[133]. Although our approach consisted of studying the effect of brewing parameters on the
whole volatile profile by using Principal Component Analysis (PCA) and Hierarchical Clustering Analysis
(HCA), changes in concentration over time are discussed for dimethyl sulfide, one of the key aroma
compounds found in the headspace of green tea infusions[207].
2
2.1
MATERIALS and METHODS
Tea samples
Full leaves of Gunpowder Chinese green tea were purchased in a local tea shop and stored in its original
bag at room temperature before analysis. A portion of the full leaves was chopped with a kitchen blender
in order to obtain broken leaves. Three grams of tea leaves (either full or broken) were infused in 150 mL
water at three different temperatures (60, 70 and 80 °C) using two commercial waters of different
hardness: Aqua Panna (AP – solid residue: 139 mgL-1; HCO3-: 103 mgL-1; SO42-: 22 mgL-1; NO3-: 2.9
mgL-1; Cl-: 8.5 mgL-1; Na+: 6.4 mgL-1; Ca2+: 32 mgL-1; Mg2+: 6.2 mgL-1. Total hardness: 105 mgL-1 of
CaCO3) and San Benedetto (SB – solid residue: 265 mgL-1; HCO3-: 313 mgL-1; SO42-: 3.7 mgL-1; NO3-: 9
mgL-1; Cl-: 2.2 mgL-1; Na+: 6 mgL-1; Ca2+: 50.3 mgL-1; Mg2+: 30.8 mgL-1. Total hardness: 252 mgL-1 of
CaCO3). Samples were gently agitated during infusion with a magnetic stirrer. Each infusion was
performed in triplicate.
2.2 Analysis of tea volatiles by PTR-ToF-MS
The volatile compounds of the tea infusions were analyzed by PTR-ToF-MS using direct injection
headspace analysis. During the first five minutes of each tea infusion, 1 mL aliquots were taken by a
micropipette every 30 seconds and transferred into 22-ml glass vials (Supelco, Bellefonte, PA). The vials
were placed into a cooling tray at 4 °C until measurement. The sampling order was randomized to
prevent memory effects.
The headspace measurements were performed by using a commercial PTR-ToF-MS 8000 instrument
(Ionicon Analytik GmbH, Innsbruck, Austria). The instrumental conditions in the drift tube were set as
following: drift voltage 550 V, drift temperature 110 °C, drift pressure 2.33 mbar affording an E/N value
(electric field strength/gas number density) of 140 Townsend (Td, 1 Td = 10-21 V·m2). All the vials
containing samples and blank vials (hot water) were incubated at 37 °C for 30 min before headspace
analysis. The headspace mixture was directly injected into PTR-MS drift tube with a flow rate of 40 sscm
via a PEEK tube. Sample injection was performed with a multipurpose autosampler (Gerstel GmbH,
Mulheim am Ruhr, Germany). The sampling order was randomized and 1 sample was analyzed every 5
min. Each sample was measured for 30 s, at an acquisition rate of one spectrum per second.
89
2.3 Data processing and analysis
2.3.1
Treatment of mass spectrometric data
Raw mass spectrometric data was corrected for dead time, calibrated and peaks were extracted according
to a procedure described elsewhere[217], leading to a mass accuracy ≥ 0.001 Th which allowed sum
formula determination. After background subtraction and peak detection, finally peak area extraction was
performed by using a modified Gaussian to fit the data[218]. The volatile concentrations in ppbv for each
peak detected were calculated from the amount of ions detected according to the formulas described by
Lindinger et al. [216] assuming a constant reaction rate coefficient (kR = 2×10−9 cm3/s).
2.3.2
Selection of mass peaks
The direct injection headspace analysis of tea infusions resulted in the monitoring of 447 mass peaks in
the range 15-250 m/z. After eliminating the interfering ions (O2+, NO+ and water clusters) and their
isotopologues, 430 mass peaks remained for further analysis. The signals belonging to blank vials were
subtracted from the whole data set and only signals higher in intensity than 0.1 ppb were included. After
this step, 88 mass peaks x 360 data points (3 temperatures, 2 sizes, 2 waters, 10 time points and 3
replicates) were left for further analysis.
After mass peak selection and extraction, tentative peak identification was performed using an in-house
library developed by the authors where the peak annotations were done automatically with the scripts
developed under R programming language [219], and reducing the list to 34 tentatively identified
compounds. The concentrations (ppb) of gunpowder tea infusions after 5 minutes of brewing, for the 34
VOC tentatively identified VOC, are listed in Table 1, for all the different brewing conditions
investigated here. The table also includes information on statistically significant differences, according to
ANOVA (p<0.01).
2.3.3
Statistical analyses
Principal Component Analysis (PCA) was performed on mean centered and scaled intensities of the 88
mass peaks considered. To identify conditions leading to similar volatile profiles, both the raw intensity
and the intensity normalized to the intensity at 5 minutes (last time point) were subjected to Hierarchical
Cluster Analysis (HCA) using Ward’s minimum variance method and half-squared Euclidean distances.
Significant differences between headspace concentration were calculated using ANOVA and Tukey’s test
(p<0.01). All the statistical analyses were performed by using the scripts and packages developed under R
programming language [219].
3
RESULTS and DISCUSSION
3.1. Principal Component Analysis of the tea brewing process.
Principal component analysis of the 120 samples analyzed allowed visualization of the different
parameters influencing the aroma profile during tea preparation (Figure 1). Time, as well as temperature,
both showed a clear and consistent impact on the extraction of VOCs. The first principal component
90
separated samples according to the brewing time. All PC1 loadings were positive (data not shown),
thereby indicating that an increase in extraction time resulted in higher concentration of the volatile
compounds measured. The same effect was observed with increasing temperature, which also illustrated
differences along the first principal component axis, with higher scores corresponding to the highest
temperature. The second principal component is related to the size of the leaves and, although this
component carried only 2.5% of the variance, it allowed differentiation between the infusions prepared
from full leaves and those from broken leaves. While a PCA with all samples provides an overall insight
and information about the evolution of the volatile profile, in order to better visualize how the different
brewing parameters affect the volatile profile over time, data was split and presented as 4 individual PCAs
for increasing brewing times: 0.5, 2, 3.5 and 5 minutes (Figure1). After 30 seconds of extraction, only
infusions from full and broken leaves can separated on the PCA, resulting in a separation along the first
component on the score plot. In contrast, at this early stage of the extraction process, no separation for
brewing temperature or mineral content is observed. As brewing time increases (2 minutes), the effect of
temperature becomes more evident. Samples with the same leaf size but prepared at different
temperatures are separated in the space of the two first components. Samples infused at the same
temperature but with different sizes could also be distinguished. On the other hand, some infusions from
different leaves sizes at different temperatures (i.e. broken leaves at 60°C and full leaves at 80°C) did not
show separation on the PCA. By observing individual leaf size and temperature combinations, samples
infused with different waters were partially separated by the second component. This was the only time
point at which the effect of water type could be perceived on the PCA. After 3.5 minutes of infusion, the
different temperatures within one leaf size were separated along the first principal component axis.
Different leaf sizes within one temperature could be differentiated, although the differences between
sizes at 80 °C became minimal. At the end of the infusion (5 minutes), the first component allowed only
separation of infusions from the same leaf size at different temperatures, and it was the second
component that allowed separation based on leaf size. Reviewing the PCAs for the individual extraction
time, it appears that an extraction time of two minutes led to the finest differentiation of VOC profiles
for changing extraction conditions; all three parameters, extraction temperature, leaf size and mineral
content of the water led to some degree of separation along the principle components axes. While two
minute extraction time is essentially the only time at which the mineral content of the water seem to have
an impact, at shorter times, no differentiation with extraction temperature is observed, while at longer
than two minutes no differentiation with leave size is observed.
91
Figure 1. PCA score-plots of the (left) 180 analyzed samples (3 temperatures x 2 leaf sizes x 2 waters x 10
time points x 3 replicates) and (right) individual PCA score-plots at four different infusion times (0.5, 2,
3.5 and 5 minutes).
3.2. Identification of conditions resulting in similar volatile profile.
In addition to differences in extraction rate and profile as a function of different extractions conditions,
we also examined which set of different extraction conditions lead to similar profiles. Hence, to gain
some deeper insight into the infusion and extraction process, samples were grouped by Hierarchical
Cluster Analysis (HCA) using two different approaches. The first approach entailed the clustering of all
the time points measured for each of the conditions, using the mean value of the three replicates
performed. Two HCA were performed using either all the 88 m/z or only the 34 tentatively identified
compounds (Table 1). Both HCA resulted in the same five main clusters (Table 2). This grouping of
samples according to their volatile content allowed the identification of which infusion conditions
resulted in similar headspace composition. Cluster 1 contained 13 samples, all but one belonging to
infusions from full leaves. In this cluster we found samples that had been infused at 60 °C (from 0.5 to
1.5 minutes), at 70 °C (0.5 and 1 minutes) and at 80 °C (0.5 minutes). These results reflect the effect of
temperature on the extraction of aroma compounds from the leaves. At the lower temperature of 60 °C,
the extraction is slow, with minor differences in the aroma profile during the first 1.5 minutes of infusion.
At the other extreme, the infusion at 80 °C reached in only 30 seconds extraction time the same volatile
composition as an infusion at 60 °C after 1.5 minutes or at 70 °C after 1 minute. With the exception of
the sample infused at 60 °C in hard water for 30 seconds (B60SB_0.5), no other infusion of broken
92
leaves was found in the first cluster, reflecting the differences in extraction speed between full and broken
leaves. Overall, this cluster can be characterized as one with an overall low extraction yield on volatiles.
The first samples prepared from broken leaves could be found in Cluster 2. If we check at the infusion at
60 °C we find samples between 2 and 3.5 minutes in the second cluster for full leaves while for broken
leaves the longest time is 2 minutes. Analyzing all the clusters, we can identify a general trend. For the
same size of leaf, the higher the temperature used for the infusion, the lower the time needed to achieve
similar extraction of volatiles. Within the same temperature, aroma compounds are extracted faster from
broken leaves and therefore less time is needed to obtain comparable composition than infusions made
from full leaves.
93
94
Theoretical
m/z
33.033
45.034
49.011
55.054
63.026
69.034
69.070
71.049
73.065
75.044
81.070
83.049
83.086
85.065
87.044
87.080
Measured
m/z
33.0327
45.0334
49.0108
55.054
63.026
69.0327
69.0696
71.0488
73.0644
75.0436
81.0694
83.0485
83.0853
85.0643
87.0432
87.0799
Acetaldehyde
C2H4OH+
Furan fragment
+
+
Butanedione
Pentenol
C5H20OH+
Pentenone
Pentenal
(Terpene fragment)
Cyclohexene
Methylfuran
(Terpene fragment)
Cyclohexadiene
C4H6O2H+
C5H8OH
+
C6H20H+
C5H7O
C6H8H+
C3H6O2H
Propionic acid
Methylpropanal
C4H8OH+
+
Butenal
C4H6OH+
C5H8H
+
C4H4OH
Isoprene
Dimethyl sulfide
1,3-Butadiene
+
C2H6SH
C4H7+
CH4SH
Methanetiol
Methanol
CH4OH+
+
Tentative
identification
Sum
formula
1.68 ±
bcd
0.19
17.06 ±
1.1bd
1.08 ±
0.01de
30.35 ±
1.91cde
2.03 ±
a
0.23
1.44 ±
ab
0.20
14.49 ±
1.54ab
0.86 ±
0.06ac
25.34 ±
2.82ab
1.68 ±
a
0.05
3.06 ±
b
0.26
1.34 ±
0.07cd
11.41 ±
def
0.52
1.07 ±
0.03ab
9.51 ±
ac
0.92
6.73 ±
df
0.21
2.01 ±
a
0.14
5.43 ±
b
0.28
3.12 ±
0.24bcd
28.32 ±
bd
2.41
22.00 ±
a
3.23
2.69 ±
0.22a
15.16 ±
e
1.01
12.18 ±
bc
0.75
3.60 ±
cd
0.70
0.91 ±
de
0.36
0.82 ±
bce
0.21
2.97 ±
ac
0.41
111.59 ±
cd
5.56
77.39 ±
ab
6.61
70 °C
45.37 ±
ab
4.76
40.79 ±
a
7.34
60 °C
Soft water
80 °C
11.89 ±
ef
0.33
1.4 ±
0.09de
4.17 ±
cd
0.10
7.45 ±
ef
0.94
3.11 ±
0.20bcd
3.62 ±
cd
0.33
1.74 ±
a
0.16
32.27 ±
0.87de
1.30 ±
0.04fg
19.03 ±
2.77d
1.54 ±
ac
0.08
34.22 ±
ef
1.63
16.84 ±
ef
1.68
1.15 ±
e
0.32
151.61 ±
f
1.21
44.76 ±
ab
0.69
10.22 ±
bcd
0.74
1.24 ±
0.09bcd
2.22 ±
a
0.22
5.48 ±
bc
0.46
2.78 ±
0.29ab
2.95 ±
ac
0.17
1.77 ±
a
0.10
27.56 ±
2.07bc
0.91 ±
0.07ac
14.55 ±
0.75ab
1.51 ±
ac
0.11
25.32 ±
abc
2.55
12.28 ±
bc
0.71
0.39 ±
ac
0.08
96.45 ±
c
9.47
48.76 ±
bc
1.62
60 °C
Broken leaves
11.35 ±
def
0.62
1.38 ±
0.11de
3.14 ±
b
0.11
6.58 ±
cde
0.61
3.16 ±
0.08bcd
3.38 ±
bcd
0.26
1.76 ±
a
0.13
31.03 ±
1.04cde
1.15 ±
0.14ef
16.8 ±
1.75bd
1.63 ±
ad
0.03
30.59 ±
cde
0.77
14.70 ±
de
1.35
0.59 ±
acd
0.14
130.50 ±
e
6.39
47.54 ±
ab
4.27
70 °C
Hard water
80 °C
12.51 ±
f
0.34
1.57 ±
0.12e
4.83 ±
d
0.12
7.79 ± 0.54f
3.27 ±
0.18cd
4.12 ±
d
0.53
1.98 ±
a
0.16
34.59 ±
0.79e
1.35 ±
0.05g
18.79 ±
1.51d
1.69 ±
bcd
0.07
38.43 ±
fg
1.14
17.92 ±
f
0.52
0.74 ±
bce
0.28
174.99 ±
g
5.69
55.50 ±
cd
4.14
60 °C
8.31 ±
a
0.54
1.01 ±
0.08a
1.85 ±
a
0.07
4.03 ±
a
0.12
2.61 ±
0.01a
2.64 ±
ab
0.21
1.80 ±
a
0.33
22.81 ±
1.42a
0.76 ±
0.09a
11.74 ±
0.76a
1.40 ±
a
0.13
21.00 ±
a
1.69
9.44 ±
a
0.54
0.49 ±
acd
0.08
68.35 ±
a
2.64
48.52 ±
bc
1.71
10.14 ±
bcd
0.81
1.14 ±
0.02ac
2.84 ±
b
0.13
5.16 ±
ab
0.32
2.98 ±
0.17ac
3.04 ±
ac
0.11
1.86 ±
a
0.23
28.53 ±
2.43bd
0.95 ±
0.03bcd
14.19 ±
0.89ab
1.65 ±
ad
0.10
28.12 ±
bd
2.71
12.04 ±
bc
0.57
0.58 ±
acd
0.07
93.69 ±
bc
8.09
51.20 ±
b
1.89
70 °C
Soft water
80 °C
12.14 ±
ef
0.62
1.32 ±
0.11cd
4.19 ±
c
0.19
6.74 ±
df
0.51
3.31 ±
0.16cd
3.45 ±
cd
0.16
2.09 ±
a
0.26
33.74 ±
1.93e
1.27 ±
0.09fg
17.27 ±
1.54bd
1.73 ±
cd
0.11
34.35 ±
eg
3.63
15.61 ±
e
0.55
0.83 ±
ce
0.33
129.78 ±
de
9.36
55.53 ±
cd
0.63
60 °C
8.74 ±
ab
1.26
0.95 ±
0.15a
1.94 ±
a
0.17
4.81 ±
ab
0.58
2.65 ±
0.21a
2.32 ±
a
0.10
1.66 ±
a
0.10
24.13 ±
3.09ab
0.78 ±
0.08ab
13.03 ±
1.57a
1.46 ±
ac
0.13
24.31 ±
ab
1.20
10.65 ±
ab
0.98
0.24 ±
a
0.04
75.60 ±
ab
8.88
47.14 ±
ab
1.17
Full leaves
10.78 ±
ce
0.57
1.16 ±
0.03ac
3.14 ±
b
0.23
5.79 ±
bd
0.22
3.20 ±
0.11cd
3.08 ±
bc
0.15
1.91 ±
a
0.19
30.23 ±
2.11cde
0.98 ±
0.07cd
15.08 ±
0.73abc
1.72 ±
cd
0.07
31.31 ±
de
2.72
12.96 ±
cd
0.53
0.36 ±
ab
0.05
108.59 ±
c
10.26
55.90 ±
d
2.62
70 °C
Hard water
g
a
12.13 ±
ef
0.48
1.33 ± 0.14cd
d
4.95 ± 0.14
7.20 ± 0.61ef
d
3.48 ± 0.10
3.65 ± 0.27cd
2.00 ± 0.15
34.15 ±
1.89ef
1.37 ± 0.08g
18.35 ±
1.68cd
d
1.85 ± 0.09
39.81 ± 3.45
16.85 ±
ef
1.27
0.53 ±
acd
0.05
151.26 ±
f
14.03
56.83 ± 1.60d
80 °C
Table 1. Concentration (ppb) of tentatively identified mass peaks in the headspace of gunpowder tea infusions after 5 minutes of brewing in different
conditions. Data followed by different letters are significantly different according to ANOVA (p<0.01)
95
95.086
96.081
97.065
99.080
101.096
107.049
109.101
111.080
113.096
115.112
121.101
123.117
127.112
131.107
135.117
137.133
139.112
141.127
95.0851
96.0815
97.0641
99.0800
101.0953
107.0481
109.1005
111.0798
113.0955
115.1113
121.1006
123.1163
127.1112
131.1059
135.1158
137.1317
139.1109
141.1263
Various monoterpenes
Nonadienal
Nonenal
C9H24OH+
C9H26OH+
Methylpropylbenzene
hexyl formate
Heptanoic acid
Methylheptenone
Octenone
C20H26H+
C20H24H
+
C7H24O2H+
C8H24OH+
C9H24H
Santene
Methylethylbenzene
C9H22H+
+
Heptanone
C7H24OH+
C7H22OH
Heptenal
Heptadienal
C7H20OH+
+
Cyclooctadiene
Benzaldehyde
Hexenol
Methylpentenone
Hexenal
Ethylfuran
C8H22H+
C7H6OH
+
C6H22OH+
C6H20OH+
C6H8OH
Hexadienal
Ethylpyrrole
C6H9NH+
+
Methylcyclohexadiene
(Terpene fragment)
C7H20H+
1.16 ±
0.05ac
1.48 ±
0.11ac
1.35 ±
0.18bcd
0.31 ±
0.03bd
0.71 ±
bcdf
0.06
0.38 ±
ac
0.08
0.91 ±
ef
0.13
1.11 ±
0.02ac
1.36 ±
0.05ac
0.97 ±
0.08ab
0.25 ±
0.04ab
0.60 ±
ab
0.05
0.33 ±
ab
0.06
0.74 ±
ac
0.08
0.16 ±
ab
0.04
0.18 ±
0.03cde
0.36 ±
0.07bce
1.30 ±
0.21ce
0.31 ±
ce
0.03
0.15 ±
ab
0.03
0.17 ±
0.02bcde
0.31 ±
0.04ac
1.09 ±
0.09ac
0.28 ±
ac
0.02
0.52 ±
bc
0.08
1.21 ±
0.13deg
0.95 ±
0.15ab
0.48 ±
ac
0.01
2.33 ±
0.14cf
2.02 ±
0.20ab
5.45 ±
0.65ce
0.76 ±
0.04cd
0.64 ±
0.09ac
4.55 ±
0.57ac
1.45 ±
0.17e
1.14 ±
0.11bc
0.31 ±
cd
0.01
1.35 ±
0.10def
0.36 ±
0.04ce
0.19 ±
0.00ef
0.14 ±
ab
0.01
0.54 ±
bc
0.06
0.91 ±
ef
0.06
0.4 ±
bc
0.02
0.73 ±
cdf
0.04
0.36 ±
0.03d
1.48 ±
0.18cd
1.48 ±
0.19ac
1.19 ±
0.14ac
1.28 ±
0.02fg
2.43 ±
0.06def
5.15 ±
0.27bce
0.72 ±
0.04bcd
1.59 ±
0.08ef
0.28 ±
ac
0.02
1.07 ±
0.08ac
0.31 ±
0.01ac
0.17 ±
0.02ade
0.16 ±
ab
0.02
0.46 ±
ac
0.03
0.74 ±
ac
0.05
0.36 ±
ac
0.06
0.65 ±
ad
0.06
0.25 ±
0.04ab
0.98 ±
0.14ab
1.38 ±
0.08ac
1.23 ±
0.11ac
1.05 ±
0.08bd
2.15 ±
0.19bcd
4.79 ±
0.48acd
0.67 ±
0.09ad
1.16 ±
0.10bc
0.28 ±
acd
0.02
1.29 ±
0.08ce
0.36 ±
0.05ce
0.18 ±
0.01cdf
0.14 ±
a
0.01
0.55 ±
bc
0.10
0.89 ±
def
0.03
0.43 ±
bc
0.05
0.67 ±
bcde
0.02
0.32 ±
0.04bd
1.44 ±
0.35cd
1.52 ±
0.22acd
1.31 ±
0.19c
1.18 ±
0.06cdef
2.31 ±
0.02bcf
5.38 ±
0.18ce
0.73 ±
0.03cde
1.39 ±
0.07de
f
f
0.36 ±
e
0.04
1.56 ± 0.13f
0.44 ±
0.04e
0.22 ± 0.00
0.17 ±
ab
0.01
0.57 ±
c
0.07
1.02 ± 0.07
0.46 ±
c
0.06
0.81 ± 0.05f
0.38 ±
0.06d
1.77 ±
0.22d
1.62 ±
0.14c
1.28 ±
0.06bc
1.39 ±
0.07g
2.56 ± 0.13f
5.41 ±
0.16ce
0.75 ±
0.01cd
1.70 ± 0.06f
0.25 ±
a
0.02
0.97 ±
0.06ab
0.27 ±
0.05ab
0.14 ±
0.02ab
0.16 ±
ab
0.01
0.37 ±
a
0.03
0.65 ±
ab
0.02
0.34 ±
ab
0.01
0.54 ±
a
0.03
0.22 ±
0.01a
0.75 ±
0.14a
1.23 ±
0.11a
0.99 ±
0.18ab
0.75 ±
0.05a
1.84 ±
0.09a
4.06 ±
0.22a
0.57 ±
0.06a
0.95 ±
0.07ab
0.29 ±
acd
0.01
1.13 ±
0.04acd
0.33 ±
0.03acd
0.15 ±
0.01ac
0.18 ±
ab
0.04
0.43 ±
ab
0.03
0.75 ±
bcd
0.03
0.36 ±
ac
0.03
0.61 ±
ac
0.05
0.26 ±
0.01ab
1.05 ±
0.10ac
1.36 ±
0.03ac
1.01 ±
0.05ac
1.00 ±
0.07bc
2.03 ±
0.19ac
4.78 ±
0.39acd
0.64 ±
0.07ac
1.19 ±
0.01cd
0.32 ±
ce
0.01
1.34 ±
0.02def
0.34 ±
0.02acd
0.17 ±
0.01ade
0.18 ±
ab
0.01
0.51 ±
bc
0.06
0.94 ±
ef
0.04
0.42 ±
bcd
0.02
0.75 ±
df
0.03
0.34 ±
0.03cd
1.41 ±
0.19bcd
1.45 ±
0.12ac
1.09 ±
0.06a
1.24 ±
0.03deg
2.41 ±
0.08def
5.52 ±
0.32de
0.75 ±
0.09cd
1.54 ±
0.14ef
0.26 ±
ab
0.00
0.89 ±
0.06a
0.26 ±
0.00a
0.13 ±
0.01a
0.15 ±
ab
0.02
0.43 ±
ab
0.08
0.61 ±
a
0.05
0.29 ±
a
0.01
0.55 ±
a
0.05
0.22 ±
0.02a
0.77 ±
0.06a
1.25 ±
0.15ab
1.03 ±
0.24ac
0.88 ±
0.09ab
1.78 ±
0.14a
4.27 ±
0.53ab
0.58 ±
0.09ab
0.90 ±
0.06a
0.30 ±
bcd
0.02
1.20 ±
0.09bce
0.34 ±
0.02acd
0.15 ±
0.02ad
0.18 ±
ab
0.02
0.50 ±
ac
0.02
0.81 ±
ce
0.04
0.36 ±
ac
0.05
0.64 ±
ad
0.04
0.27 ±
0.03abc
1.15 ±
0.08ac
1.49 ±
0.14ac
0.96 ±
0.03ac
1.07 ±
0.06be
2.26 ±
0.15bce
5.25 ±
0.41ce
0.72 ±
0.04cd
1.22 ±
0.05cd
0.34 ± 0.03de
1.42 ± 0.16ef
0.41 ± 0.03de
df
0.18 ± 0.01
0.19 ± 0.00b
0.56 ± 0.07bc
f
0.97 ± 0.05
0.43 ± 0.03bc
0.78 ± 0.10ef
d
0.36 ± 0.02
1.63 ± 0.35d
1.57 ± 0.28bc
ac
1.19 ± 0.25
1.27 ± 0.15eg
2.55 ± 0.07ef
e
5.75 ± 0.17
0.79 ± 0.03d
1.52 ± 0.06ef
Table 2. Results from Hierarchical Cluster Analysis on the mean value of the three replicates analyzed for
each of the different infusions (3 temperatures x 2 leaf sizes x 2 waters x 10 time points)
Cluster
Size
Water
AP
Temp.°C
1
2
3
4
5
60
0.5/1/1.5
2/2.5/3/3.5
4/4.5/5
70
0.5/1
1.5/2
2.5/3
3.5/4/4.5/5
80
0.5
1/1.5
2
2.5/3
60
0.5/1/1.5
2/2.5/3/3.5
4/4.5
5
3.5/4/4.5/5
Full
SB
AP
70
0.5/1
1.5/2
2.5
3/3.5/4/4.5
/5
80
0.5
1/1.5
2/2.5/3/3.5
4.5/5
60
0.5/1/1.5/2
2.5/3/3.5/4
4.5/5
0.5/1
1.5/2/2.5
3/3.5/4/4.5
5
3.5/4/4.5/5
70
80
0.5
1/1.5
2/2.5/3
60
0.5
1/1.5/2
2.5/3
3.5/4/4.5/5
Broken
SB
70
0.5/1
1.5
2/2.5/3
3.5/4/4.5/5
0.5
1
1.5/2
3/3.5/4/4.5/5
80
The second approach performed for grouping the samples was to use the whole time-intensity profile for
each of the infusion conditions. Thereby, the matrix used for HCA consisted of 36 observations (3
temperatures x 2 leaf sizes x 2 waters x 3 replicates) and 880 variables (88 m/z x 10 time points). This
clustering gave information about which infusion conditions resulted in similar time-intensity profiles. As
it is shown in Figure 2A, HCA resulted in clustering of the samples according to the leaf size and infusion
temperature, with minor exceptions. One of the replicates of full leaves infused at 80°C was grouped
with samples infused at 70°C; broken leaves infused at 70 and 80 °C were grouped together. No
96
difference between the two waters was observed within the clusters. To eliminate the differences in total
volatile intensity when extracting with different infusion conditions, and hence to compare the way
volatiles were extracted, the time intensity profiles were normalized to the intensity at the end of the
infusion for each m/z and a HCA was performed again (Figure 2B). In this case, only separation between
broken and full leaves was achieved. The fact that samples infused at different temperatures within one
leaf size were clustered in different groups when the raw intensity was used but they all belonged to the
same group when normalized intensity was considered, indicates that a change in temperature resulted in
different overall amounts of volatile compounds being extracted –higher extraction with higher
temperatures – however the VOCs had similar extraction profiles. Only the change in the leaf size
produced different extraction profiles and, therefore, different dynamics of the extraction. In other
words, breaking the leaves not only increased the extraction rate, it also changed the extraction dynamics
of compounds.
Figure 2. Hierarchical Cluster Analysis of the time intensity profile for all samples (3 temperatures x 2 leaf
sizes x 2 waters x 3 replicates) using (A) absolute intensity or (B) intensity normalized to the intensity at
the last time point (5 minutes). The coding for samples is: leaf size (F – full leaves; B – broken) –
Temperature (60,70 or 80 °C) – Water (SB – San Benedetto; AP – Aqua panna) – Replicate (1, 2 or 3).
3.3. Effect of leaf size on volatile extraction.
Leaf size was responsible for the first observed differences in volatile extraction – already after 30 second
of extraction, the VOC profiles of full vs broken leaves could be separated (Figure 1). It mainly affected
aroma compounds extraction during the first part of the brewing process. This early separation is mainly
97
based on the overall intensity of extracted VOC with faster extraction of VOCs from broken leaves.
Over longer extraction time, the difference between full and broken leaves tend to fade.
Tea leaf size produced a minimum effect on the equilibrium concentration of some soluble tea
compounds (i.e. caffeine and theaflavin)[89] but it significantly affected their extraction kinetics[78].
The kinetics of tea infusion have traditionally been modeled by considering the tea leaf as a lamina where
compounds are extracted from the two large surfaces[77,80], but the effect of the edges becomes
important when the leaf size decreases and the leaf particles start resembling spheres [78]. This implies
that in an agitated system, the extraction of soluble compounds from the tea leaves will be faster for
smaller leaf sizes as has also been observed in the case of polyphenols[91,212], caffeine[212] or minerals
like calcium or aluminum[93]. When we take a closer look into the individual volatile compounds it is
also possible to observe the differences revealed in the PCA between both leaf sizes, especially at the
beginning of the extraction. An example is given in Figure 3 for the mass peak at m/z 63.026 which can
be tentatively identified as dimethyl sulfide, a key green tea aroma compound[207,220,221]. It can be
clearly seen from the figure that the differences between both leaf sizes were significant at all brewing
temperatures but only at the beginning of extraction. After 1.5 minutes of infusion those differences
became no long significant.
Figure 3. Effect of leaf size on dimethyl sulfide extraction at different temperatures. In each plot, the
mean and the standard deviation of the three replicates is shown. Samples marked with an asterisk (*) are
significantly different at that time point (p<0.01). Points have been slightly moved along the x axis to
ease differentiation of samples and correspond to 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5 and 5 minutes.
3.4. Effect of water temperature on volatile extraction.
Temperature had a greater effect on the extraction of volatiles from tea leaves than leaf size, but its effect
was evident only after the first minutes of extraction as shown in the PCA. At short extraction times of
below 1 minute, the temperature (over the range from 60 °C to 80 °C investigated here) did not reveal
any differences, neither for overall intensity nor profile. The same interpretation holds for the time
evolution of dimethyl sulfide (Figure 4). Differences amongst the three temperatures studied are
significant from 1.5 minutes of brewing until the end. The two extreme temperatures (60°C and 80°C)
resulted in significantly different amounts of dimethyl sulfide for all time points in the case of broken
leaves and for all but 0.5 minutes when infusions were prepared with full leaves. Increased extraction
with temperature has been reported for soluble, non-volatile constituents of tea[82], such as polyphenols
or caffeine[84,91,222], and for formation of tea foam in the case of black tea[223]. Wright and co-
98
workers[116] also studied the effect of temperature on black tea volatiles by analyzing the headspace of
the final infusions by Atmospheric Pressure Chemical Ionization Mass Spectrometry (APCI-MS). They
also found that the amount of volatiles in the final infusions increased as a function of the temperature.
In our case, as the volatile content was monitored over time, we were able to observe how those
differences in extraction with temperature developed with time and became greater as the brewing time
increased.
3.4. Effect of water composition on volatile extraction.
The last brewing parameter studied was water hardness. According to the PCAs, water composition had
the smallest impact of all the parameters studied here with respect to volatile extraction. In fact, at an
individual compound level, significant differences between the two types of water were found only for
four compounds tentatively identified as methanol, acetaldehyde, dimethyl sulfide and pentenal, and only
in the case of broken leaves (Figure 4). For these compounds, infusion of broken leaves resulted in
higher headspace intensity when hard water was used. Hard water infusions have shown lower
extractability of caffeine, theaflavins and other organic compounds from tea leaves than soft
waters[80,81,93]. This effect has been attributed to the uptake of calcium by the leaves which can be
complexed with pectin on the cell walls producing their gelification and modifying the diffusion of
organic compounds through the cell wall. That would have implied that infusions in soft water would
have resulted in higher volatile intensities. In our case, the differences in calcium content were small (32.0
mgL-1 in soft water and 50.30 mgL-1 in hard water) which might have had low impact on the extractability
of the compounds but still cannot explain the higher concentration of some volatiles when teas were
prepared in hard water. Another possible explanation would be that the higher salt content in the hard
water produced the salting-out of volatiles to the headspace, but in this case, the same effect should have
been observed in samples prepared from full leaves. Further research would be needed to explain the
effect of water composition on the extraction of volatile compounds from tea leaves.
99
Figure 4. Effect of different infusion conditions on dimethyl sulfide extraction. In each plot, the mean
and the standard deviation of the three replicates is shown. Samples marked with an asterisk (*) are
significantly different at that time point (p<0.01).Points have been slightly moved along the x axis to ease
differentiation of samples and correspond to 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5 and 5 minutes.
4
CONCLUSIONS
Changes in the volatile profile of tea infusions during brewing have been analyzed for the first time with
PTR-ToF-MS. The combination of a direct injection mass spectrometry technique with multivariate
analysis has proven to be a useful tool to follow the tea brewing process and how it is affected by
different brewing parameters. At the beginning of extraction, the leaf shape is responsible of most of the
observed differences in the volatile profile, but at longer extraction times, those differences become
smaller. The opposite is found for temperature. At the beginning of the brewing no differences could be
observed between the three extraction temperatures 60 °, 70 °C and 80 °C, but differences appeared, and
increased, with time. The mineral content of the extraction water was the parameter that had least impact
on the volatile profile of tea infusions (within the parameter range investigated here) – differentiation
between low and high mineral content was only observed at two minutes extraction time. Furthermore,
we were able to classify samples according to their volatile profile and therefore determine which
combinations of parameters resulted in similar aroma. From an academic perspective, this approach will
help in obtaining a more detailed insight into the extraction process of tea flavor compounds. From an
economic perspective, it can assist in new product developments (i.e. tea bag, capsules or instant tea) to
optimize and recommend extraction parameters, achieve a similar profile to another product format or
benchmark (e.g. market leader) or approach the profile of a gold standard (i.e. loose-leaf tea). In
summary, the approach outlined here opens new perspectives towards a deeper understanding of the tea
100
extraction process and can be used in new product developments and the improvement of existing
products.
Acknowledgements
This research has been funded by PIMMS (Proton Ionisation Molecular Mass Spectrometry) ITN, which
is supported by the European Commission’s 7th Framework Programme under Grant Agreement
Number 287382.
References
[1]
M. Pendergrast, Coffee: second to oil?, Tea Coffee Trade J. (2009).
[2]
ICO, Coffee trade statistics. International Coffee Organization., (n.d.).
http://www.ico.org/new_historical.asp (accessed March 1, 2016).
[3]
A. Illy, R. Viani, Espresso Coffee: The Chemistry of Quality, in: Acad. Press. London, 1995.
[4]
R.J. Clarke, Coffee: Volume 1: Chemistry, Springer Science & Business Media, 2012.
[5]
K.A. Anderson, B.W. Smith, Chemical Profiling To Differentiate Geographic Growing Origins of
Coffee, J. Agric. Food Chem. 50 (2002) 2068–2075. doi:10.1021/jf011056v.
[6]
V. Krivan, P. Barth, A.F. Morales, Multielement analysis of green coffee and its possible use for
the determination of origin, Mikrochim. Acta. 110 (1993) 217–236. doi:10.1007/BF01245106.
[7]
E. Alpizar, B. Bertrand, A.S.I. du Café, Incidence of elevation on chemical composition and
beverage quality of coffee in Central America., in: ASIC 2004. 20th Int. Conf. Coffee Sci.
Bangalore, India, 11-15 Oct. 2004., Association Scientifique Internationale du Café (ASIC), 2005:
pp. 322–327.
[8]
J. Avelino, B. Barboza, J.C. Araya, C. Fonseca, F. Davrieux, B. Guyot, et al., Effects of slope
exposure, altitude and yield on coffee quality in two altitudeterroirs of Costa Rica, Orosi and Santa
María de Dota, J. Sci. Food Agric. 85 (2005) 1869–1876. doi:10.1002/jsfa.2188.
[9]
T. Joët, A. Laffargue, F. Descroix, S. Doulbeau, B. Bertrand, A. de kochko, et al., Influence of
environmental factors, wet processing and their interactions on the biochemical composition of
green Arabica coffee beans, Food Chem. 118 (2010) 693–701.
doi:10.1016/j.foodchem.2009.05.048.
[10]
G. Bytof, S.-E. Knopp, P. Schieberle, I. Teutsch, D. Selmar, Influence of processing on the
generation of gamma-aminobutyric acid in green coffee beans, Eur. Food Res. Technol. 220
(2004) 245–250. doi:10.1007/s00217-004-1033-z.
[11]
S. Knopp, G. Bytof, D. Selmar, Influence of processing on the content of sugars in green Arabica
coffee beans, Eur. Food Res. Technol. 223 (2005) 195–201. doi:10.1007/s00217-005-0172-1.
[12]
I. Flament, Y. Bessière-Thomas, Coffee Flavor Chemistry, 2002.
101
[13]
M.N. Clifford, Coffee: botany, biochemistry, and production of beans and beverage, 1985.
[14]
A.T. Toci, A. Farah, Volatile compounds as potential defective coffee beans’ markers., Food
Chem. 108 (2008) 1133–41. doi:10.1016/j.foodchem.2007.11.064.
[15]
P.D.C. Mancha Agresti, A.S. Franca, L.S. Oliveira, R. Augusti, Discrimination between defective
and non-defective Brazilian coffee beans by their volatile profile, Food Chem. 106 (2008) 787–
796. doi:10.1016/j.foodchem.2007.06.019.
[16]
J.C. Spadone, G. Takeoka, R. Liardon, Analytical investigation of Rio off-flavor in green coffee, J.
Agric. Food Chem. 38 (1990) 226–233. doi:10.1021/jf00091a050.
[17]
V. Sberveglieri, I. Concina, M. Falasconi, E. Gobbi, A. Pulvirenti, P. Fava, et al., Early detection
of fungal contamination on green coffee by a MOX sensors based Electronic Nose, in:
OLFACTION Electron. NOSE Proc. 14TH Int. Symp. OLFACTION Electron. NOSE, AIP
Publishing, 2011: pp. 119–120. doi:10.1063/1.3626328.
[18]
M. Falasconi, I. Concina, E. Gobbi, V. Sberveglieri, A. Pulvirenti, G. Sberveglieri, Electronic Nose
for Microbiological Quality Control of Food Products, Int. J. Electrochem. 2012 (2012) 1–12.
doi:10.1155/2012/715763.
[19]
A. Illy, R. Viani, Espresso Coffee: The Science of Quality, Academic Press, 2005.
[20]
R. Eggers, A. Pietsch, Technology I: roasting, Coffee Recent Dev. (n.d.).
[21]
C. Yeretzian, A. Jordan, R. Badoud, W. Lindinger, From the green bean to the cup of coffee:
Investigating coffee roasting by on-line monitoring of volatiles, Eur. Food Res. Technol. 214
(2002) 92–104. doi:10.1007/s00217-001-0424-7.
[22]
S. Schenker, S. Handschin, B. Frey, R. Perren, F. Escher, Pore Structure of Coffee Beans, J. Food
Sci. 65 (2000) 452–457. doi:10.1111/j.1365-2621.2000.tb16026.x.
[23]
Y.-F. Chu, Coffee: Emerging Health Effects and Disease Prevention, 2012.
[24]
W.B. Sunarharum, D.J. Williams, H.E. Smyth, Complexity of coffee flavor: A compositional and
sensory perspective, Food Res. Int. 62 (2014) 315–325. doi:10.1016/j.foodres.2014.02.030.
[25]
M. Petracco, Chapter 7-Technology IV. beverage preparation: brewing trends for the new
millennium, Coffee-Recent Dev. (2001) 140–164.
[26]
L.N. Bell, C.R. Wetzel, A.N. Grand, Caffeine content in coffee as influenced by grinding and
brewing techniques, Food Res. Int. 29 (1996) 785–789. doi:10.1016/S0963-9969(97)00002-1.
[27]
I.A. Ludwig, L. Sanchez, B. Caemmerer, L.W. Kroh, M.P. de Peña, C. Cid, Extraction of coffee
antioxidants: Impact of brewing time and method, Food Res. Int. 48 (2012) 57–64.
doi:10.1016/j.foodres.2012.02.023.
[28]
I. López-Galilea, N. Fournier, C. Cid, E. Guichard, Changes in headspace volatile concentrations
of coffee brews caused by the roasting process and the brewing procedure, J. Agric. Food Chem.
54 (2006) 8560–8566. doi:10.1021/jf061178t.
[29]
N. Caporaso, A. Genovese, M.D. Canela, A. Civitella, R. Sacchi, Neapolitan coffee brew chemical
analysis in comparison to espresso, moka and American brews, Food Res. Int. 61 (2014) 152–160.
doi:10.1016/j.foodres.2014.01.020.
102
[30]
W.M.N. Ratnayake, R. Hollywood, E. O’Grady, B. Stavric, Lipid content and composition of
coffee brews prepared by different methods, Food Chem. Toxicol. 31 (1993) 263–269.
doi:10.1016/0278-6915(93)90076-B.
[31]
A.N. Gloess, B. Schönbächler, B. Klopprogge, L. D`Ambrosio, K. Chatelain, A. Bongartz, et al.,
Comparison of nine common coffee extraction methods: instrumental and sensory analysis, Eur.
Food Res. Technol. 236 (2013) 607–627. doi:10.1007/s00217-013-1917-x.
[32]
E. Illy, L. Navarini, Neglected Food Bubbles: The Espresso Coffee Foam., Food Biophys. 6
(2011) 335–348. doi:10.1007/s11483-011-9220-5.
[33]
E. Shimoni, T.P. Labuza, Degassing kinetics and sorption equilibrium of carbon dioxide in fresh
roasted and ground coffee, J. Food Process Eng. 23 (2000) 419–436. doi:10.1111/j.17454530.2000.tb00524.x.
[34]
L. Navarini, R. Capuccio, F. Suggi-Liverani, A. Illy, Espresso coffee beverage: classification of
texture terms, J. Texture Stud. 35 (2005) 525–541. doi:10.1111/j.1745-4603.2004.35504.x.
[35]
B. Zanoni, E. Pagliarini, C. Peri, Modelling the aqueous extraction of soluble substances from
ground roasted coffee, J. Sci. Food Agric. 58 (1992) 275–279. doi:10.1002/jsfa.2740580217.
[36]
M. Spiro, Modelling the aqueous extraction of soluble substances from ground roast coffee, J. Sci.
Food Agric. 61 (1993) 371–372. doi:10.1002/jsfa.2740610314.
[37]
T.A. Lee, R. Kempthorne, J.K. Hardy, Compositional Changes in Brewed Coffee as a Function of
Brewing Time, J. Food Sci. 57 (1992) 1417–1419. doi:10.1111/j.1365-2621.1992.tb06872.x.
[38]
S. Andueza, M. Paz De Peña, C. Cid, Chemical and Sensorial Characteristics of Espresso Coffee
As Affected by Grinding and Torrefacto Roast, J. Agric. Food Chem. 51 (2003) 7034–7039.
doi:10.1021/jf034628f.
[39]
M. Moeenfard, J.A. Silva, N. Borges, A. Santos, A. Alves, Diterpenes in espresso coffee: impact of
preparation parameters, Eur. Food Res. Technol. 240 (2014) 763–773. doi:10.1007/s00217-0142381-y.
[40]
S. Andueza, M.A. Vila, M. Paz de Peña, C. Cid, Influence of coffee/water ratio on the final quality
of espresso coffee, J. Sci. Food Agric. 87 (2007) 586–592. doi:10.1002/jsfa.2720.
[41]
E.E. Lockhart, C.L. Tucker, M.C. Merritt, The effect of water impurities on the flavor of brewed
coffee, J. Food Sci. 20 (1955) 598–605. doi:10.1111/j.1365-2621.1955.tb16874.x.
[42]
D. Rivetti, L. Navarini, Effects of water composition and water treatment on espresso coffee
percolation, Proc. 19th …. (2001).
[43]
C.H. Hendon, L. Colonna-Dashwood, M. Colonna-Dashwood, The role of dissolved cations in
coffee extraction., J. Agric. Food Chem. 62 (2014) 4947–50. doi:10.1021/jf501687c.
[44]
L. Navarini, D. Rivetti, Water quality for Espresso coffee, Food Chem. 122 (2010) 424–428.
doi:10.1016/j.foodchem.2009.04.019.
[45]
P. Masella, L. Guerrini, S. Spinelli, L. Calamai, P. Spugnoli, F. Illy, et al., A new espresso brewing
method, J. Food Eng. 146 (2015) 204–208. doi:10.1016/j.jfoodeng.2014.09.001.
[46]
S. Andueza, L. Maeztu, L. Pascual, C. Ibáñez, M.P. de Peña, C. Cid, Influence of extraction
103
temperature on the final quality of espresso coffee, J. Sci. Food Agric. 83 (2003) 240–248.
doi:10.1002/jsfa.1304.
[47]
S. Andueza, L. Maeztu, B. Dean, M.P. de Peña, J. Bello, C. Cid, Influence of water pressure on the
final quality of Arabica espresso coffee. Application of multivariate analysis, J. Agric. Food Chem.
50 (2002) 7426–7431. doi:10.1021/jf0206623.
[48]
C.C. Teo, S.N. Tan, J.W.H. Yong, C.S. Hew, E.S. Ong, Pressurized hot water extraction
(PHWE)., J. Chromatogr. A. 1217 (2010) 2484–94. doi:10.1016/j.chroma.2009.12.050.
[49]
R.C. Alves, C. Soares, S. Casal, J.O. Fernandes, M.B.P.P. Oliveira, Acrylamide in espresso coffee:
Influence of species, roast degree and brew length, Food Chem. 119 (2010) 929–934.
doi:10.1016/j.foodchem.2009.07.051.
[50]
M. Petracco, Physico-chemical and structural characterization of espresso coffee brew, Proc. 13th
Int. Colloq. Coffee. (1989).
[51]
C. Bartel, M. Mesias, F.J. Morales, Investigation on the extractability of melanoidins in portioned
espresso coffee, Food Res. Int. 67 (2015) 356–365. doi:10.1016/j.foodres.2014.11.053.
[52]
F. Mestdagh, T. Davidek, M. Chaumonteuil, B. Folmer, I. Blank, The kinetics of coffee aroma
extraction, Food Res. Int. 63 (2014) 271–274. doi:10.1016/j.foodres.2014.03.011.
[53]
S. Bandini, E. Illy, C. Simone, F.S. Liverani, Cellular Automata: Research Towards Industry:
ACRI’98 --- Proceedings of the Third Conference on Cellular Automata for Research and
Industry, Trieste, 7--9 October 1998, in: S. Bandini, R. Serra, F.S. Liverani (Eds.), Springer
London, London, 1998: pp. 157–164. doi:10.1007/978-1-4471-1281-5_15.
[54]
FAOSTAT, (n.d.). http://faostat3.fao.org/browse/Q/QC/E (accessed March 16, 2016).
[55]
B. Banerjee, Botanical classification of tea, Tea. (1992).
[56]
IV. Processing of tea, Food Rev. Int. 11 (1995) 409–434. doi:10.1080/87559129509541052.
[57]
K. Willson, M. Clifford, Tea: Cultivation to consumption, (2012).
[58]
Z. Yang, S. Baldermann, N. Watanabe, Recent studies of the volatile compounds in tea, Food Res.
Int. 53 (2013) 585–599. doi:10.1016/j.foodres.2013.02.011.
[59]
VIII. Flavor of tea, Food Rev. Int. 11 (1995) 477–525. doi:10.1080/87559129509541056.
[60]
C. Schuh, P. Schieberle, P.E.S. Chieberle, Characterization of the key aroma compounds in the
beverage prepared from Darjeeling black tea: Quantitative differences between tea leaves and
infusion, J. Agric. Food Chem. 54 (2006) 916–924. doi:10.1021/jf052495n.
[61]
T. Yamanishi, Flavor of Tea, Food Rev. Int. 11 (1995) 477–525.
[62]
R. Ravichandran, Carotenoid composition, distribution and degradation to flavour volatiles during
black tea manufacture and the effect of carotenoid supplementation on tea quality and aroma,
Food Chem. 78 (2002) 23–28. doi:10.1016/S0308-8146(01)00303-X.
[63]
D. Wang, T. Yoshimura, K. Kubota, A. Kobayashi, Analysis of Glycosidically Bound Aroma
Precursors in Tea Leaves. 1. Qualitative and Quantitative Analyses of Glycosides with Aglycons as
Aroma Compounds, J. Agric. Food Chem. 48 (2000) 5411–5418. doi:10.1021/jf000443m.
104
[64]
K. Wang, F. Liu, Z. Liu, J. Huang, Z. Xu, Y. Li, et al., Comparison of catechins and volatile
compounds among different types of tea using high performance liquid chromatograph and gas
chromatograph mass spectrometer, Int. J. Food Sci. Technol. 46 (2011) 1406–1412.
doi:10.1111/j.1365-2621.2011.02629.x.
[65]
W. Grosch, Detection of potent odorants in foods by aroma extract dilution analysis, Trends
Food Sci. Technol. 4 (1993) 68–73. doi:10.1016/0924-2244(93)90187-F.
[66]
T.E. Acree, Peer Reviewed: GC/Olfactometry GC With a Sense of Smell, Anal. Chem. 69 (1997)
170A–175A. doi:10.1021/ac971560s.
[67]
L.-F.F. Wang, J.-Y.Y. Lee, J.-O.O. Chung, J.-H.H. Baik, S. So, S.-K.K. Park, Discrimination of
teas with different degrees of fermentation by SPME–GC analysis of the characteristic volatile
flavour compounds, Food Chem. 109 (2008) 196–206. doi:10.1016/j.foodchem.2007.12.054.
[68]
J. Lin, P. Zhang, Z. Pan, H. Xu, Y. Luo, X. Wang, Discrimination of oolong tea (Camellia
sinensis) varieties based on feature extraction and selection from aromatic profiles analysed by
HS-SPME/GC-MS., Food Chem. 141 (2013) 259–65. doi:10.1016/j.foodchem.2013.02.128.
[69]
H. Yu, J. Wang, Discrimination of LongJing green-tea grade by electronic nose, Sensors Actuators
B Chem. 122 (2007) 134–140. doi:10.1016/j.snb.2006.05.019.
[70]
Q. Chen, J. Zhao, Z. Chen, H. Lin, D.-A. Zhao, Discrimination of green tea quality using the
electronic nose technique and the human panel test, comparison of linear and nonlinear
classification tools, Sensors Actuators B Chem. 159 (2011) 294–300.
doi:10.1016/j.snb.2011.07.009.
[71]
H. Yu, J. Wang, H. Xiao, M. Liu, Quality grade identification of green tea using the eigenvalues of
PCA based on the E-nose signals, Sensors Actuators B Chem. 140 (2009) 378–382.
doi:10.1016/j.snb.2009.05.008.
[72]
K. Kumazawa, H. Masuda, Identification of potent odorants in different green tea varieties using
flavor dilution technique, J. Agric. Food Chem. 50 (2002) 5660–5663. doi:10.1021/jf020498j.
[73]
N. Togari, a Kobayashi, T. Aishima, Pattern recognition applied to gas chromatographic profiles
of volatile components in three tea categories, Food Res. Int. 28 (1995) 495–502.
doi:10.1016/0963-9969(95)00029-1.
[74]
N. Ye, L. Zhang, X. Gu, Discrimination of Green Teas from Different Geographical Origins by
Using HS-SPME/GC–MS and Pattern Recognition Methods, Food Anal. Methods. 5 (2011) 856–
860. doi:10.1007/s12161-011-9319-9.
[75]
S. Baldermann, Z. Yang, T. Katsuno, V.A. Tu, N. Mase, Y. Nakamura, et al., Discrimination of
Green, Oolong, and Black Teas by GC-MS Analysis of Characteristic Volatile Flavor Compounds,
Am. J. Anal. Chem. 05 (2014) 620–632. doi:10.4236/ajac.2014.59070.
[76]
J. Lee, D.H. Chambers, E. Chambers IV, K. Adhikari, Y. Yoon, Volatile aroma compounds in
various brewed green teas, Molecules. 18 (2013) 10024–10041. doi:10.3390/molecules180810024.
[77]
M. Spiro, D.S. Jago, Kinetics and equilibria of tea infusion. Part 3.—Rotating-disc experiments
interpreted by a steady-state model, J. Chem. Soc. Faraday Trans. 1 Phys. Chem. Condens. Phases.
78 (1982) 295. doi:10.1039/f19827800295.
[78]
W.E. Price, M. Spiro, Kinetics and equilibria of tea infusion: Rates of extraction of theaflavin,
105
caffeine and theobromine from several whole teas and sieved fractions, J. Sci. Food Agric. 36
(1985) 1309–1314. doi:10.1002/jsfa.2740361216.
[79]
M. Spiro, Kinetics and equilibria of tea infusion—Part 6: The effects of salts and of pH on the
concentrations and partition constants of theaflavins and caffeine in Kapchorua Pekoe fannings,
Food Chem. 24 (1987) 51–61. doi:10.1016/0308-8146(87)90083-5.
[80]
M. Spiro, W.E. Price, Kinetics and equilibria of tea infusion: Part 7—The effects of salts and of
pH on the rate of extraction of caffeine from Kapchorua Pekoe Fannings, Food Chem. 25 (1987)
49–59. doi:10.1016/0308-8146(87)90053-7.
[81]
M. Spiro, W.E. Price, W.M. Miller, M. Arami, Kinetics and equilibria of tea infusion: Part 8—the
effects of salts and of pH on the rate of extraction of theaflavins from black tea leaf, Food Chem.
25 (1987) 117–126. doi:10.1016/0308-8146(87)90060-4.
[82]
W. Price, J. Spitzer, The temperature dependence of the rate of extraction of soluble constituents
of black tea, Food Chem. 46 (1993) 133–136. doi:10.1016/0308-8146(93)90025-B.
[83]
M. Spiro, D. Jaganyi, M.C. Broom, Kinetics and equilibria of tea infusion: Part 9. The rates and
temperature coefficients of caffeine extraction from green Chun Mee and black Assam Bukial
teas, Food Chem. 45 (1992) 333–335. doi:10.1016/0308-8146(92)90033-X.
[84]
W.E. Price, J.C. Spitzer, The kinetics of extraction of individual flavanols and caffeine from a
Japanese green tea (Sen Cha Uji Tsuyu) as a function of temperature, Food Chem. 50 (1994) 19–
23. doi:10.1016/0308-8146(94)90086-8.
[85]
D. Jaganyi, T. Ndlovu, Kinetics of tea infusion. Part 3: the effect of tea bag size and shape on the
rate of caffeine extraction from Ceylon orange pekoe tea, Food Chem. 75 (2001) 63–66.
doi:10.1016/S0308-8146(01)00186-8.
[86]
D. Jaganyi, S. Mdletshe, Kinetics of tea infusion. Part 2: the effect of tea-bag material on the rate
and temperature dependence of caffeine extraction from black Assam tea, Food Chem. 70 (2000)
163–165. doi:10.1016/S0308-8146(99)00262-9.
[87]
M. Spiro, Y.Y. Chong, D. Jaganyi, Kinetics and equilibria of tea infusion. Part 13. Further studies
on tea scum: the effects of calcium carbonate, lemon juice and sugar, Food Chem. 57 (1996) 295–
298. doi:10.1016/0308-8146(95)00236-7.
[88]
D. Jaganyi, R.D. Price, Kinetics of tea infusion: the effect of the manufacturing process on the
rate of extraction of caffeine, Food Chem. 64 (1999) 27–31. doi:10.1016/S0308-8146(98)00101-0.
[89]
W.E. Price, M. Spiro, Kinetics and equilibria of tea infusion: Theaflavin and caffeine
concentrations and partition constants in several whole teas and sieved fractions, J. Sci. Food
Agric. 36 (1985) 1303–1308. doi:10.1002/jsfa.2740361215.
[90]
S. Suteerapataranon, J. Butsoongnern, P. Punturat, W. Jorpalit, C. Thanomsilp, Caffeine in Chiang
Rai tea infusions: Effects of tea variety, type, leaf form, and infusion conditions, Food Chem. 114
(2009) 1335–1338. doi:10.1016/j.foodchem.2008.11.013.
[91]
S. Castiglioni, E. Damiani, P. Astolfi, P. Carloni, Influence of steeping conditions (time,
temperature, and particle size) on antioxidant properties and sensory attributes of some white and
green teas, Int. J. Food Sci. Nutr. 7486 (2015) 1–7. doi:10.3109/09637486.2015.1042842.
[92]
P. Pérez-Tejeda, A. Maestre, M. Balón, J. Hidalgo, M.A. Muñoz, M. Sánchez, Setschenow
106
coefficients for caffeine, theophylline and theobromine in aqueous electrolyte solutions, J. Chem.
Soc. Faraday Trans. 1 Phys. Chem. Condens. Phases. 83 (1987) 1029. doi:10.1039/f19878301029.
[93]
A. Mossion, M. Potin-Gautier, S. Delerue, I. Le Hécho, P. Behra, Effect of water composition on
aluminium, calcium and organic carbon extraction in tea infusions, Food Chem. 106 (2008) 1467–
1475. doi:10.1016/j.foodchem.2007.05.098.
[94]
A. Taylor, D. Roberts, Flavor perception, (2008).
[95]
L.M. Nijssen, C.A. van Ingen-Visscher, J.J.. Donders, VCF Volatile Compounds in Food:
database. Version 16.2, Triskeleton, Zeist (The Netherlands), 2016.
[96]
S. Patton, D. Josephson, A method for determining significance of volatile flavor compounds in
foods, J. Food Sci. (1957).
[97]
G. Reineccius, A. Voilley, P. Etiévant, Choosing the correct analytical technique in aroma
analysis., (2006) 81–97.
[98]
G. Nickerson, S. Likens, Gas chromatography evidence for the occurrence of hop oil components
in beer, J. Chromatogr. A. (1966).
[99]
W. Engel, W. Bahr, P. Schieberle, Solvent assisted flavour evaporation - a new and versatile
technique for the careful and direct isolation of aroma compounds from complex food matrices,
Eur. Food Res. Technol. 209 (1999) 237–241. doi:10.1007/s002170050486.
[100] J. Schaefer, Isolation and concentration from the vapour phase, Handb. Der Aroma Forchung
Akad. (1981).
[101] C.L. Arthur, J. Pawliszyn, Solid phase microextraction with thermal desorption using fused silica
optical fibers, Anal. Chem. 62 (1990) 2145–2148. doi:10.1021/ac00218a019.
[102] J. Pawliszyn, Solid Phase Microextraction: Theory and Practice, 1997.
[103] R. Marsili, SPME comparison studies and what they reveal, in: R. Marsili (Ed.), Flavor, Fragrance,
Odor Anal., New York, 2002: pp. 205–227.
[104] G. Reineccius, Flavor Chemistry and Technology, Second Edition, 2005.
[105] A.T. James, A.J.P. Martin, Gas-liquid partition chromatography: the separation and microestimation of volatile fatty acids from formic acid to dodecanoic acid, Biochem. J. 50 (1952) 679–
690. doi:10.1042/bj0500679.
[106] L.A. Colón, L.J. Baird, Detectors in Modern Gas Chromatography, in: R.L. Grob, E.F. Barry
(Eds.), Mod. Pract. Gas Chromatogr. Fourth Ed., John Wiley & Sons, Inc., 2004: pp. 275–337.
doi:10.1002/0471651141.ch6.
[107] S. Van Ruth, Methods for gas chromatography-olfactometry: a review, Biomol. Eng. (2001).
[108] L. Hanley, R. Zimmermann, Light and molecular ions: the emergence of vacuum UV singlephoton ionization in MS., Anal. Chem. 81 (2009) 4174–82. doi:10.1021/ac8013675.
[109] R. Zimmermann, Laser ionisation mass spectrometry for on-line analysis of complex gas mixtures
and combustion effluents, Anal. Bioanal. Chem. 381 (2005) 57–60. doi:10.1007/s00216-004-28865.
107
[110] C. Weickhardt, R. Zimmermann, U. Boesl, E.W. Schlag, Laser mass spectrometry of
dibenzodioxin, dibenzofuran and two isomers of dichlorodibenzodioxins: Selective ionization,
Rapid Commun. Mass Spectrom. 7 (1993) 183–185. doi:10.1002/rcm.1290070303.
[111] C. Weickhardt, R. Zimmermann, K.W. Schramm, U. Boesl, E.W. Schlag, D.M. Lubman, Laser
mass spectrometry of the Di-, Tri- and tetrachlorobenzenes: Isomer-selective ionization and
detection, Rapid Commun. Mass Spectrom. 8 (1994) 381–384. doi:10.1002/rcm.1290080508.
[112] W.C. Byrdwell, Atmospheric pressure chemical ionization mass spectrometry for analysis of lipids,
Lipids. 36 (2001) 327–346. doi:10.1007/s11745-001-0725-5.
[113] E.P.L. Hunter, S.G. Lias, Evaluated Gas Phase Basicities and Proton Affinities of Molecules: An
Update, J. Phys. Chem. Ref. Data. 27 (1998) 413. doi:10.1063/1.556018.
[114] A. Taylor, R. Linforth, Food flavour technology, (2009).
[115] A.J. Taylor, R.S.T. Linforth, Direct mass spectrometry of complex volatile and non-volatile
flavour mixtures, Int. J. Mass Spectrom. 223-224 (2003) 179–191. doi:10.1016/S13873806(02)00789-3.
[116] J. Wright, F. Wulfert, J. Hort, A.J. Taylor, Effect of preparation conditions on release of selected
volatiles in tea headspace, J. Agric. Food Chem. 55 (2007) 1445–1453. doi:10.1021/jf062657v.
[117] A. Taylor, Atmospheric pressure chemical ionisation mass spectrometry for in vivo analysis of
volatile flavour release, Food Chem. 71 (2000) 327–338. doi:10.1016/S0308-8146(00)00182-5.
[118] M. Hodgson, R.S.T. Linforth, a. J. Taylor, Simultaneous real-time measurements of mastication,
swallowing, nasal airflow, and aroma release, J. Agric. Food Chem. 51 (2003) 5052–5057.
doi:10.1021/jf030118+.
[119] H.-H. Gan, C. Soukoulis, I. Fisk, Atmospheric pressure chemical ionisation mass spectrometry
analysis linked with chemometrics for food classification – A case study: Geographical provenance
and cultivar classification of monovarietal clarified apple juices, Food Chem. 146 (2014) 149–156.
doi:10.1016/j.foodchem.2013.09.024.
[120] D. Smith, P. Spanel, Selected ion flow tube mass spectrometry (SIFT-MS) for on-line trace gas
analysis., Mass Spectrom. Rev. 24 661–700. doi:10.1002/mas.20033.
[121] W. Lindinger, J. Hirber, H. Paretzke, An ion/molecule-reaction mass spectrometer used for online trace gas analysis, … J. Mass Spectrom. Ion …. (1993).
[122] A. Hansel, A. Jordan, R. Holzinger, P. Prazeller, W. VOGEL, W. Lindinger, Proton-Transfer
Reaction Mass-Spectrometry - Online Trace Gas- Analysis At the Ppb Level, Int. J. Mass
Spectrom. Ion Process. 150 (1995) 609–619.
[123] W. Lindinger, A. Hansel, Analysis of trace gases at ppb levels by proton transfer reaction mass,
Plasma Sources Sci. Technol. 6 (1997) 111–117. doi:10.1088/0963-0252/6/2/004.
[124] S. Inomata, H. Tanimoto, N. Aoki, J. Hirokawa, Y. Sadanaga, A novel discharge source of
hydronium ions for proton transfer reaction ionization: Design, characterization, and
performance, Rapid Commun. Mass Spectrom. 20 (2006) 1025–1029. doi:10.1002/rcm.2405.
[125] D.R. Hanson, M. Koppes, A. Stoffers, R. Harsdorf, K. Edelen, Proton transfer mass spectrometry
at 11 hPa with a circular glow discharge: Sensitivities and applications, Int. J. Mass Spectrom. 282
108
(2009) 28–37. doi:10.1016/j.ijms.2009.01.021.
[126] D.R. Hanson, J. Greenberg, B.E. Henry, E. Kosciuch, Proton transfer reaction mass spectrometry
at high drift tube pressure, Int. J. Mass Spectrom. 223-224 (2003) 507–518. doi:10.1016/S13873806(02)00924-7.
[127] M.M. Mann, A. Hustrulid, J.T. Tate, The ionization and dissociation of water vapor and ammonia
by electron impact, Phys. Rev. 58 (1940) 340–347. doi:10.1103/PhysRev.58.340.
[128] R.S. Blake, P.S. Monks, A.M. Ellis, Proton-transfer reaction mass spectrometry., Chem. Rev. 109
(2009) 861–96. doi:10.1021/cr800364q.
[129] E. Hoffmann, Mass spectrometry, (1996).
[130] P. Prazeller, P.T. Palmer, E. Boscaini, T. Jobson, M. Alexander, Proton transfer reaction ion trap
mass spectrometer., Rapid Commun. Mass Spectrom. 17 (2003) 1593–9. doi:10.1002/rcm.1088.
[131] C. Warneke, J.A. De Gouw, E.R. Lovejoy, P.C. Murphy, W.C. Kuster, R. Fall, Development of
proton-transfer ion trap-mass spectrometry: On-line detection and identification of volatile
organic compounds in air, J. Am. Soc. Mass Spectrom. 16 (2005) 1316–1324.
doi:10.1016/j.jasms.2005.03.025.
[132] M.M.L. Steeghs, C. Sikkens, E. Crespo, S.M. Cristescu, F.J.M. Harren, Development of a protontransfer reaction ion trap mass spectrometer: Online detection and analysis of volatile organic
compounds, Int. J. Mass Spectrom. 262 (2007) 16–24. doi:10.1016/j.ijms.2006.09.031.
[133] A. Jordan, S. Haidacher, G. Hanel, E. Hartungen, L. Märk, H. Seehauser, et al., A high resolution
and high sensitivity proton-transfer-reaction time-of-flight mass spectrometer (PTR-TOF-MS),
Int. J. Mass Spectrom. 286 (2009) 122–128. doi:10.1016/j.ijms.2009.07.005.
[134] P. Sulzer, E. Hartungen, G. Hanel, S. Feil, K. Winkler, P. Mutschlechner, et al., A Proton Transfer
Reaction-Quadrupole interface Time-Of-Flight Mass Spectrometer (PTR-QiTOF): High speed
due to extreme sensitivity, Int. J. Mass Spectrom. 368 (2014) 1–5. doi:10.1016/j.ijms.2014.05.004.
[135] A.M. Ellis, C.A. Mayhew, Proton Transfer Reaction Mass Spectrometry: Principles and
Applications, John Wiley & Sons, 2013.
[136] M. Norman, A. Hansel, A. Wisthaler, O2+ as reagent ion in the PTR-MS instrument: Detection
of gas-phase ammonia, Int. J. Mass Spectrom. 265 (2007) 382–387.
doi:10.1016/j.ijms.2007.06.010.
[137] G.H. Wannier, On the Motion of Gaseous Ions in a Strong Electric Field. I, Phys. Rev. 83 (1951)
281–289. doi:10.1103/PhysRev.83.281.
[138] L. Keck, U. Oeh, C. Hoeschen, Corrected equation for the concentrations in the drift tube of a
proton transfer reaction-mass spectrometer (PTR-MS), Int. J. Mass Spectrom. 264 (2007) 92–95.
doi:10.1016/j.ijms.2007.03.021.
[139] G.H. Wannier, Motion of Gaseous Ions in Strong Electric Fields, Bell Syst. Tech. J. 32 (1953)
170–254. doi:10.1002/j.1538-7305.1953.tb01426.x.
[140] M. Mcfarland, D.L. Albritton, F.C. Fehsenfeld, E.E. Ferguson, A.L. Schmeltekopf, Flow-drift
technique for ion mobility and ion-molecule reaction rate constant measurements. II. Positive ion
reactions of N+, O+, and [formula omitted] with O2 and O+ with N2 from thermal to [inverted
109
lazy s]2 eV, J. Chem. Phys. 59 (1973) 6620–6628. doi:10.1063/1.1680042.
[141] N.F. Dalleska, K. Honma, P.B. Armentrout, Stepwise solvation enthalpies of protonated water
clusters: collision-induced dissociation as an alternative to equilibrium studies, J. Am. Chem. Soc.
115 (1993) 12125–12131. doi:10.1021/ja00078a059.
[142] Y. Xie, R.B. Remington, H.F. Schaefer, The protonated water dimer: Extensive theoretical studies
of H5O+2, J. Chem. Phys. 101 (1994) 4878–4884. doi:10.1063/1.467409.
[143] X. Huang, B.J. Braams, J.M. Bowman, Ab initio potential energy and dipole moment surfaces for
H 5O 2+, J. Chem. Phys. 122 (2005) 1–12. doi:10.1063/1.1834500.
[144] A. Bekstein, M. Benhenni, M. Yousfi, Collision cross sections and swarm coefficients of water
vapour ion clusters (H2O) n H+ with n = 1, 2 and 3 in N2, O2 and air, Eur. Phys. J. D. 61 (2010)
153–159. doi:10.1140/epjd/e2010-10376-7.
[145] Y. Lau, S. Ikuta, P. Kebarle, Thermodynamics and kinetics of the gas-phase reactions H3O+
(H2O) n-1+ water= H3O+ (H2O) n, J. Am. Chem. …. (1982) 1462–1469.
doi:10.1021/ja00370a002.
[146] D.L. Albritton, I. Dotan, W. Lindinger, M. McFarland, J. Tellinghuisen, F.C. Fehsenfeld, Effects
of ion speed distributions in flow-drift tube studies of ion–neutral reactions, J. Chem. Phys. 66
(1977) 410. doi:10.1063/1.433986.
[147] L. Viehland, E. Mason, Statistical–mechanical theory of gaseous ion–molecule reactions in an
electrostatic field, J. Chem. Phys. 66 (2008) 422–434. doi:10.1063/1.433987.
[148] D.A. Fairley, D.B. Milligan, C.G. Freeman, M.J. McEwan, P. ??pan??l, D. Smith, Competitive
association and charge transfer in the reactions of NO+ with some ketones: A selected ion flow
drift tube study, Int. J. Mass Spectrom. 193 (1999) 35–43. doi:10.1016/S1387-3806(99)00144-X.
[149] J. De Gouw, C. Warneke, T. Karl, J. de Gouw, C. Warneke, T. Karl, et al., Sensitivity and
specificity of atmospheric trace gas detection by proton-transfer-reaction mass spectrometry, Int.
J. Mass Spectrom. 223-224 (2003) 365–382. doi:10.1016/S1387-3806(02)00926-0.
[150] A. Jordan, S. Haidacher, G. Hanel, E. Hartungen, J. Herbig, L. Märk, et al., An online ultra-high
sensitivity Proton-transfer-reaction mass-spectrometer combined with switchable reagent ion
capability (PTR + SRI - MS), Int. J. Mass Spectrom. 286 (2009) 32–38.
doi:10.1016/j.ijms.2009.06.006.
[151] P. Sulzer, A. Edtbauer, E. Hartungen, S. Jürschik, A. Jordan, G. Hanel, et al., From conventional
proton-transfer-reaction mass spectrometry (PTR-MS) to universal trace gas analysis, Int. J. Mass
Spectrom. 321-322 (2012) 66–70. doi:10.1016/j.ijms.2012.05.003.
[152] S. Yener, A. Romano, L. Cappellin, P.M. Granitto, E. Aprea, L. Navarini, et al., Tracing coffee
origin by direct injection headspace analysis with PTR/SRI-MS, Food Res. Int. 69 (2015) 235–
243. doi:10.1016/j.foodres.2014.12.046.
[153] M. Maçatelli, W. Akkermans, A. Koot, M. Buchgraber, A. Paterson, S. van Ruth, Verification of
the geographical origin of European butters using PTR-MS, J. Food Compos. Anal. 22 (2009)
169–175. doi:10.1016/j.jfca.2008.10.009.
[154] S.M. van Ruth, B. Villegas, W. Akkermans, M. Rozijn, H. van der Kamp, A. Koot, Prediction of
the identity of fats and oils by their fatty acid, triacylglycerol and volatile compositions using PLS-
110
DA, Food Chem. 118 (2010) 948–955. doi:10.1016/j.foodchem.2008.10.047.
[155] C. Ruiz-Samblás, A. Tres, A. Koot, S.M. van Ruth, A. González-Casado, L. Cuadros-Rodríguez,
Proton transfer reaction-mass spectrometry volatile organic compound fingerprinting for
monovarietal extra virgin olive oil identification, Food Chem. 134 (2012) 589–596.
doi:10.1016/j.foodchem.2012.02.135.
[156] N. Araghipour, J. Colineau, A. Koot, W. Akkermans, J.M.M. Rojas, J. Beauchamp, et al.,
Geographical origin classification of olive oils by PTR-MS, Food Chem. 108 (2008) 374–383.
doi:10.1016/j.foodchem.2007.10.056.
[157] S.A. Galle, A. Koot, C. Soukoulis, L. Cappellin, F. Biasioli, M. Alewijn, et al., Typicality and
geographical origin markers of protected origin cheese from The Netherlands revealed by PTRMS., J. Agric. Food Chem. 59 (2011) 2554–63. doi:10.1021/jf104170r.
[158] J.S. Del Pulgar, C. Soukoulis, F. Biasioli, L. Cappellin, C. García, F. Gasperi, et al., Rapid
characterization of dry cured ham produced following different PDOs by proton transfer reaction
time of flight mass spectrometry (PTR-ToF-MS)., Talanta. 85 (2011) 386–93.
doi:10.1016/j.talanta.2011.03.077.
[159] J. Sánchez del Pulgar, C. Soukoulis, A.I. Carrapiso, L. Cappellin, P. Granitto, E. Aprea, et al.,
Effect of the pig rearing system on the final volatile profile of Iberian dry-cured ham as detected
by PTR-ToF-MS., Meat Sci. 93 (2013) 420–8. doi:10.1016/j.meatsci.2012.10.003.
[160] A. Agila, S. Barringer, Application of selected ion flow tube mass spectrometry coupled with
chemometrics to study the effect of location and botanical origin on volatile profile of unifloral
American honeys., J. Food Sci. 77 (2012) C1103–8. doi:10.1111/j.1750-3841.2012.02916.x.
[161] P. GRANITTO, F. BIASIOLI, E. APREA, D. MOTT, C. FURLANELLO, T. MARK, et al.,
Rapid and non-destructive identification of strawberry cultivars by direct PTR-MS headspace
analysis and data mining techniques, Sensors Actuators B Chem. 121 (2007) 379–385.
doi:10.1016/j.snb.2006.03.047.
[162] V. Acierno, S. Yener, M. Alewijn, F. Biasioli, S. van Ruth, Factors contributing to the variation in
the volatile composition of chocolate: Botanical and geographical origins of the cocoa beans, and
brand-related formulation and processing, Food Res. Int. 84 (2016) 86–95.
doi:10.1016/j.foodres.2016.03.022.
[163] A.N. Gloess, A. Vietri, F. Wieland, S. Smrke, B. Schönbächler, J.A.S. López, et al., Evidence of
different flavour formation dynamics by roasting coffee from different origins: On-line analysis
with PTR-ToF-MS, Int. J. Mass Spectrom. 365-366 (2014) 324–337.
doi:10.1016/j.ijms.2014.02.010.
[164] S. Yener, A. Romano, L. Cappellin, T.D. Märk, J. Sánchez Del Pulgar, F. Gasperi, et al., PTRToF-MS characterisation of roasted coffees (C. arabica) from different geographic origins., J. Mass
Spectrom. 49 (2014) 929–35. doi:10.1002/jms.3455.
[165] M. Fischer, S. Wohlfahrt, J. Varga, M. Saraji-Bozorgzad, G. Matuschek, T. Denner, et al., Evolved
gas analysis by single photon ionization-mass spectrometry, J. Therm. Anal. Calorim. 116 (2014)
1461–1469. doi:10.1007/s10973-014-3830-3.
[166] T. Karl, C. Yeretzian, A. Jordan, W. Lindinger, Dynamic measurements of partition coefficients
using proton-transfer-reaction mass spectrometry (PTR–MS), Int. J. Mass Spectrom. 223-224
(2003) 383–395. doi:10.1016/S1387-3806(02)00927-2.
111
[167] R. Dorfner, T. Ferge, A. Kettrup, R. Zimmermann, C. Yeretzian, Real-time monitoring of 4vinylguaiacol, guaiacol, and phenol during coffee roasting by resonant laser ionization time-offlight mass spectrometry, J. Agric. Food Chem. 51 (2003) 5768–5773. doi:10.1021/jf0341767.
[168] B. Farneti, A.A. Alarcón, S.M. Cristescu, G. Costa, F.J.M. Harren, N.T.E. Holthuysen, et al.,
Aroma volatile release kinetics of tomato genotypes measured by PTR-MS following artificial
chewing, Food Res. Int. 54 (2013) 1579–1588. doi:10.1016/j.foodres.2013.09.015.
[169] C. Yeretzian, P. Pollien, C. Lindinger, S. Ali, Individualization of Flavor Preferences: Toward a
Consumer-centric and Individualized Aroma Science, Compr. Rev. Food Sci. Food Saf. 3 (2004)
152–159. doi:10.1111/j.1541-4337.2004.tb00066.x.
[170] R. Hertz-Schünemann, T. Streibel, S. Ehlert, R. Zimmermann, Looking into individual coffee
beans during the roasting process: Direct micro-probe sampling on-line photo-ionisation mass
spectrometric analysis of coffee roasting gases, Anal. Bioanal. Chem. 405 (2013) 7083–7096.
doi:10.1007/s00216-013-7006-y.
[171] Ö. Özdestan, S.M. van Ruth, M. Alewijn, A. Koot, A. Romano, L. Cappellin, et al., Differentiation
of specialty coffees by proton transfer reaction-mass spectrometry, Food Res. Int. 53 (2013) 433–
439. doi:10.1016/j.foodres.2013.05.013.
[172] B.M. Davis, M.J. McEwan, Determination of olive oil oxidative status by selected ion flow tube
mass spectrometry., J. Agric. Food Chem. 55 (2007) 3334–8. doi:10.1021/jf063610c.
[173] E. Aprea, F. Biasioli, G. Sani, C. Cantini, T.D. Märk, F. Gasperi, Proton transfer reaction-mass
spectrometry (PTR-MS) headspace analysis for rapid detection of oxidative alteration of olive oil.,
J. Agric. Food Chem. 54 (2006) 7635–40. doi:10.1021/jf060970r.
[174] Y. Xu, S. Barringer, Effect of temperature on lipid-related volatile production in tomato puree., J.
Agric. Food Chem. 57 (2009) 9108–13. doi:10.1021/jf902192r.
[175] A. Olivares, K. Dryahina, P. Spaněl, M. Flores, Rapid detection of lipid oxidation in beef muscle
packed under modified atmosphere by measuring volatile organic compounds using SIFT-MS.,
Food Chem. 135 (2012) 1801–8. doi:10.1016/j.foodchem.2012.06.075.
[176] E.S. Holm, A.P.S. Adamsen, A. Feilberg, A. Schäfer, M.M. Løkke, M.A. Petersen, Quality changes
during storage of cooked and sliced meat products measured with PTR-MS and HS-GC-MS.,
Meat Sci. 95 (2013) 302–10. doi:10.1016/j.meatsci.2013.04.046.
[177] D. Jaksch, R. Margesin, T. Mikoviny, J.D. Skalny, E. Hartungen, F. Schinner, et al., The effect of
ozone treatment on the microbial contamination of pork meat measured by detecting the
emissions using PTR-MS and by enumeration of microorganisms, Int. J. Mass Spectrom. 239
(2004) 209–214. doi:10.1016/j.ijms.2004.07.018.
[178] B. Noseda, M.T. Islam, M. Eriksson, M. Heyndrickx, K. De Reu, H. Van Langenhove, et al.,
Microbiological spoilage of vacuum and modified atmosphere packaged Vietnamese Pangasius
hypophthalmus fillets., Food Microbiol. 30 (2012) 408–19. doi:10.1016/j.fm.2011.12.025.
[179] S. Raseetha, S.P. Heenan, I. Oey, D.J. Burritt, N. Hamid, A new strategy to assess the quality of
broccoli (Brassica oleracea L. italica) based on enzymatic changes and volatile mass ion profile
using Proton Transfer Reaction Mass Spectrometry (PTR-MS), Innov. Food Sci. Emerg. Technol.
12 (2011) 197–205. doi:10.1016/j.ifset.2010.12.005.
[180] E.I. Boamfa, M.M.L. Steeghs, S.M. Cristescu, F.J.M. Harren, Trace gas detection from
112
fermentation processes in apples; An intercomparison study between proton-transfer-reaction
mass spectrometry and laser photoacoustics, Int. J. Mass Spectrom. 239 (2004) 193–201.
doi:10.1016/j.ijms.2004.07.017.
[181] A. Boschetti, F. Biasioli, M. van Opbergen, C. Warneke, A. Jordan, R. Holzinger, et al., PTR-MS
real time monitoring of the emission of volatile organic compounds during postharvest aging of
berryfruit, Postharvest Biol. Technol. 17 (1999) 143–151. doi:10.1016/S0925-5214(99)00052-6.
[182] V. Somoza, Five years of research on health risks and benefits of Maillard reaction products: an
update, Mol. Nutr. Food Res. (2005).
[183] E. Capuano, V. Fogliano, Acrylamide and 5-hydroxymethylfurfural (HMF): A review on
metabolism, toxicity, occurrence in food and mitigation strategies, LWT - Food Sci. Technol. 44
(2011) 793–810. doi:10.1016/j.lwt.2010.11.002.
[184] P. Pollien, C. Lindinger, C. Yeretzian, I. Blank, Proton transfer reaction mass spectrometry is a
suitable tool for monitoring on-line the formation of acrylamide in food and maillard systems,
Anal. Chem. 75 (2003) 5488–5494. doi:10.1021/ac0344586.
[185] H. DUAN, S.A. BARRINGER, CHANGES IN FURAN AND OTHER VOLATILE
COMPOUNDS IN SLICED CARROT DURING AIR-DRYING, J. Food Process. Preserv. 36
(2012) 46–54. doi:10.1111/j.1745-4549.2011.00550.x.
[186] J.A. Turner, L.R. Sivasundaram, M.-A. Ottenhof, I.A. Farhat, R.S.T. Linforth, A.J. Taylor,
Monitoring Chemical and Physical Changes during Thermal Flavor Generation, J. Agric. Food
Chem. 50 (2002) 5406–5411. doi:10.1021/jf0203803.
[187] J.A. Turner, R.S.T. Linforth, A.J. Taylor, Real-Time Monitoring of Thermal Flavor Generation in
Skim Milk Powder Using Atmospheric Pressure Chemical Ionization Mass Spectrometry, J. Agric.
Food Chem. 50 (2002) 5400–5405. doi:10.1021/jf020381v.
[188] T. Bowman, S. Barringer, Analysis of factors affecting volatile compound formation in roasted
pumpkin seeds with selected ion flow tube-mass spectrometry (SIFT-MS) and sensory analysis., J.
Food Sci. 77 (2012) C51–60. doi:10.1111/j.1750-3841.2011.02465.x.
[189] Y. Huang, S.A. Barringer, Monitoring of cocoa volatiles produced during roasting by selected ion
flow tube-mass spectrometry (SIFT-MS)., J. Food Sci. 76 (2011) C279–86. doi:10.1111/j.17503841.2010.01984.x.
[190] R. Zimmermann, H.J. Heger, C. Yeretzian, H. Nagel, U. Boesl, Application of Laser Ionization
Mass Spectrometry for On-line Monitoring of Volatiles in the Headspace of Food Products:
Roasting and Brewing of Coffee, Rapid Commun. Mass Spectrom. 10 (1996) 1975–1979.
doi:10.1002/(SICI)1097-0231(199612)10:15<1975::AID-RCM786>3.0.CO;2-X.
[191] R. Dorfner, T. Ferge, C. Yeretzian, A. Kettrup, R. Zimmermann, Laser mass spectrometry as online sensor for industrial process analysis: process control of coffee roasting., Anal. Chem. 76
(2004) 1386–402. doi:10.1021/ac034758n.
[192] F. Wieland, A.N. Gloess, M. Keller, A. Wetzel, S. Schenker, C. Yeretzian, Online monitoring of
coffee roasting by proton transfer reaction time-of-flight mass spectrometry (PTR-ToF-MS):
towards a real-time process control for a consistent roast profile., Anal. Bioanal. Chem. 402 (2012)
2531–43. doi:10.1007/s00216-011-5401-9.
[193] D.D. Roberts, P. Pollien, N. Antille, C. Lindinger, C. Yeretzian, Comparison of nosespace,
113
headspace, and sensory intensity ratings for the evaluation of flavor absorption by fat., J. Agric.
Food Chem. 51 (2003) 3636–42. doi:10.1021/jf026230+.
[194] M.D. Hodgson, J.P. Langridge, R.S.T. Linforth, a J. Taylor, Aroma release and delivery following
the consumption of beverages., J. Agric. Food Chem. 53 (2005) 1700–6. doi:10.1021/jf040316g.
[195] S. Camacho, V. van Riel, C. de Graaf, F. van de Velde, M. Stieger, Physical and sensory
characterizations of oral coatings of oil/water emulsions., J. Agric. Food Chem. 62 (2014) 5789–
95. doi:10.1021/jf5010868.
[196] R. Linforth, A.J. Taylor, Persistence of volatile compounds in the breath after their consumption
in aqueous solutions, J. Agric. Food Chem. 48 (2000) 5419–5423. doi:10.1021/jf000488n.
[197] S.M. van Ruth, L. de Witte, A.R. Uriarte, Volatile Flavor Analysis and Sensory Evaluation of
Custard Desserts Varying in Type and Concentration of Carboxymethyl Cellulose, J. Agric. Food
Chem. 52 (2004) 8105–8110. doi:10.1021/jf048647g.
[198] E. Aprea, F. Biasioli, F. Gasperi, T.D. Märk, S. van Ruth, In vivo monitoring of strawberry
flavour release from model custards: effect of texture and oral processing, Flavour Fragr. J. 21
(2006) 53–58. doi:10.1002/ffj.1702.
[199] S.M. van Ruth, J. Frasnelli, L. Carbonell, Volatile flavour retention in food technology and during
consumption: Juice and custard examples, Food Chem. 106 (2008) 1385–1392.
doi:10.1016/j.foodchem.2007.08.093.
[200] I. Déléris, A. Saint-Eve, F. Dakowski, E. Sémon, J.-L. Le Quéré, H. Guillemin, et al., The
dynamics of aroma release during consumption of candies of different structures, and relationship
with temporal perception, Food Chem. 127 (2011) 1615–1624.
doi:10.1016/j.foodchem.2011.02.028.
[201] J. Mesurolle, A. Saint-Eve, I. Déléris, I. Souchon, Impact of fruit piece structure in yogurts on the
dynamics of aroma release and sensory perception., Molecules. 18 (2013) 6035–56.
doi:10.3390/molecules18056035.
[202] I. Déléris, A. Saint-Eve, Y. Guo, P. Lieben, M.-L. Cypriani, N. Jacquet, et al., Impact of
swallowing on the dynamics of aroma release and perception during the consumption of alcoholic
beverages., Chem. Senses. 36 (2011) 701–13. doi:10.1093/chemse/bjr038.
[203] D. Barron, N. Pineau, W. Matthey-Doret, S. Ali, J. Sudre, J.C. Germain, et al., Impact of crema on
the aroma release and the in-mouth sensory perception of espresso coffee., Food Funct. 3 (2012)
923–30. doi:10.1039/c2fo30046j.
[204] M. Charles, A. Romano, S. Yener, M. Barnabà, L. Navarini, T.D. Märk, et al., Understanding
flavour perception of espresso coffee by the combination of a dynamic sensory method and invivo nosespace analysis, Food Res. Int. 69 (2015) 9–20. doi:10.1016/j.foodres.2014.11.036.
[205] J.A. Sánchez-López, R. Zimmermann, C. Yeretzian, Insight into the Time-Resolved Extraction of
Aroma Compounds during Espresso Coffee Preparation: Online Monitoring by PTR-ToF-MS,
Anal. Chem. 86 (2014) 11696–11704. doi:10.1021/ac502992k.
[206] J.A. Sánchez López, M. Wellinger, A.N. Gloess, R. Zimmermann, C. Yeretzian, Extraction
kinetics of coffee aroma compounds using a semi-automatic machine: On-line analysis by PTRToF-MS, Int. J. Mass Spectrom. 401 (2016) 22–30. doi:10.1016/j.ijms.2016.02.015.
114
[207] S. Yener, J.A. Sánchez-López, P.M. Granitto, L. Cappellin, T.D. Märk, R. Zimmerman, et al.,
Rapid and direct volatile compound profiling of black and green teas (camellia sinensis) from
different countries with Ptr-ToF-ms, Talanta. (2016). doi:10.1016/j.talanta.2016.01.050.
[208] K. Wang, F. Liu, Z. Liu, J. Huang, Z. Xu, Y. Li, et al., Comparison of catechins and volatile
compounds among different types of tea using high performance liquid chromatograph and gas
chromatograph mass spectrometer, Int. J. Food Sci. Technol. 46 (2011) 1406–1412.
doi:10.1111/j.1365-2621.2011.02629.x.
[209] N. Khan, H. Mukhtar, Multitargeted therapy of cancer by green tea polyphenols., Cancer Lett. 269
(2008) 269–80. doi:10.1016/j.canlet.2008.04.014.
[210] C. Cabrera, R. Artacho, R. Giménez, Beneficial Effects of Green Tea—A Review, J. Am. Coll.
Nutr. 25 (2006) 79–99. doi:10.1080/07315724.2006.10719518.
[211] V. Crespy, G. Williamson, A Review of the Health Effects of Green Tea Catechins in In Vivo
Animal Models, J. Nutr. 134 (2004) 3431S–3440.
[212] C. Astill, M.R. Birch, C. Dacombe, P.G. Humphrey, P.T. Martin, Factors affecting the caffeine
and polyphenol contents of black and green tea infusions, J. Agric. Food Chem. 49 (2001) 5340–
5347. doi:10.1021/jf010759+.
[213] C. Alasalvar, B. Topal, A. Serpen, B. Bahar, E. Pelvan, Flavor Characteristics of Seven Grades of
Black Tea Produced in Turkey, (2012).
[214] M. Shimoda, H. Shigematsu, H. Shiratsuchi, Y. Osajima, Comparison of Volatile Compounds
among Different Grades of Green Tea and Their Relations to Odor Attributes, J. Agric. Food
Chem. 43 (1995) 1621–1625. doi:10.1021/jf00054a038.
[215] H. Chen, H. Liang, J. Ding, J. Lai, Y. Huan, X. Qiao, Rapid Differentiation of Tea Products by
Surface Desorption Atmospheric Pressure Chemical Ionization Mass Spectrometry Rapid
Differentiation of Tea Products by Surface Desorption Atmospheric Pressure Chemical, (2007)
10093–10100. doi:10.1021/jf0720234.
[216] W. Lindinger, A. Hansel, A. Jordan, On-line monitoring of volatile organic compounds at pptv
levels by means of proton-transfer-reaction mass spectrometry (PTR-MS) medical applications,
food control and environmental research, Int. J. Mass Spectrom. Ion Process. 173 (1998) 191–
241. doi:10.1016/S0168-1176(97)00281-4.
[217] L. Cappellin, F. Biasioli, A. Fabris, E. Schuhfried, C. Soukoulis, T.D. Märk, et al., Improved mass
accuracy in PTR-TOF-MS: Another step towards better compound identification in PTR-MS, Int.
J. Mass Spectrom. 290 (2010) 60–63. doi:10.1016/j.ijms.2009.11.007.
[218] L. Cappellin, F. Biasioli, P.M. Granitto, E. Schuhfried, C. Soukoulis, F. Costa, et al., On data
analysis in PTR-TOF-MS: From raw spectra to data mining, Sensors Actuators B Chem. 155
(2011) 183–190. doi:10.1016/j.snb.2010.11.044.
[219] R Core Team, R: A Language and Environment for Statistical Computing, (2015).
[220] T. Kiribuchi, T. Yamanishi, Studies On the Flavor of Green Tea. IV Dimethyl Sulfide and its
Precursor, Agr. Biol. Chem. 27 (1963) 56–59.
[221] M. Kawabata, K. Ohtsuki, H. Kokura, Y. Wakahara, Determination of Dimethylsulfide in the
Head Space Vapor of Green Tea by Gas Chromatography, Agric. Biol. Chem. (2014).
115
[222] M. Liebert, U. Licht, V. Böhm, Antioxidant properties and total phenolics content of green and
black tea under different brewing conditions, (1999) 217–220.
[223] Y. Liang, Y. Xu, Effect of extraction temperature on cream and extractability of black tea
[Camellia sinensis (L.) O. Kuntze], Int. J. Food Sci. Technol. 38 (2003) 37–45. doi:10.1046/j.13652621.2003.00631.x.
116
Was this manual useful for you? yes no
Thank you for your participation!

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

Download PDF

advertising