Journal of the Science of Food and Agriculture J Sci Food Agric 80:85±90 (2000)
Kinetics of coffee infusion: a comparative study
on the extraction kinetics of mineral ions and
caffeine from several types of medium roasted
coffees
Deogratius Jaganyi* and Sizwe Paulos Madlala
Department of Chemistry, University of Natal, Private Bag X01, Scottsville, Pietermaritzburg 3209, South Africa
(Rec

* Co
Sout
Cont

# 2
Abstract: A comparative study of the equilibrium concentrations, rates of infusion and intra-bean

diffusion coef®cients of caffeine, P (as H2POÿ4 ), K�, Mg2� and Mn2� in Milli-Q-Water at 80°C was

carried out. Medium roasted coffees of particle size range 1.70±2.00mm from six different countriesÐ

Special Kenya (Kenya), Santos (Brazil), Blue Mountain Java (Sumatra), Mocha (Ethiopia), Zimbabwe

(Zimbabwe) and South African Grown (South Africa)Ðwere used in the investigation. High-

performance liquid chromatography was used for the analyses of caffeine, and inductively coupled

plasma atomic emission spectrometry for the chosen elements. The equilibrium concentrations of all

the species and the trend were found to be independent of the various coffee beans. The order of the

rate of infusion was found to be K�> caffeine>P (as H2POÿ4 )>Mg2�>Mn2�. Examination of the rate

constants clearly indicated that P (as H2POÿ4 ), Mg2� and Mn2� cannot be used for identi®cation of the

coffee origin, but the values for caffeine and K� can be used. Diffusion coef®cients of caffeine and

mineral ions were calculated in two separate ways using rate constants and half-lives of infusion. These

were then compared with known diffusion coef®cients of the same species in water for determination of

hindrance factors. The hindrance factor for caffeine was found to be much smaller than the

corresponding factors recorded at 25.5°C. In general the hindrance factors in the bean were all of the

order of 10, with Mn2� being the most hindered species. This is an indication that the infusion of the

various species through the coffee bean is a hindered process. This is because of the association of

caffeine and mineral ions with other coffee solubles and the physical restraint within the bean matrix.

# 2000 Society of Chemical Industry
Keywords: coffee; caffeine; mineral ions; extraction kinetics; diffusion coef®cient
INTRODUCTION
Coffee beans (Arabica and Robusta) contain a variety

of different elements, of which potassium represents

the largest proportion. In addition, other metallic

elements found in reasonable quantity are magnesium

and calcium, while the non-metallic elements are

mainly phosphorus and sulphur. The other mineral

ions exist in trace amounts. Numerous workers1±4

have found that the mineral content of coffee beans

depends upon the country of origin, a factor which is

associated with the soil conditions. The other factor is

the method of processing the green coffee, whether

subjected to the wet or the dry method.1,5,6 Clarke and

Walker7 have shown that soaking of coffee beans

during the fermentation and washing stages causes

potassium to diffuse out, resulting in a wide range of

potassium contents in coffee beans. The mineral

constituents in coffee beans are not lost during the

roasting process, except for sulphur and phosphorus to
eived 23 December 1998; accepted 23 July 1999)

rrespondence to: Deogratius Jaganyi, Department of Chemistry, Univ
h Africa
ract/grant sponsor: University of Natal Research and Development.

000 Society of Chemical Industry. J Sci Food Agric 0022±5142/2
a small degree.1 It is also believed that minerals act as

catalysts during roasting and during the various

biochemical transformations in the plant. Arabica

cannot be distinguished from Robusta by determina-

tion of mineral ions, as indicated by Kroplien's data.8

However, recent work9 has shown that the mineral

content can be related to the origin of the green coffee

and can be used as a tool for characterising coffee

varieties.

A subject that has received very little attention until

recently is the kinetics of extraction of mineral ions.10

Even for caffeine there is no comparative study to

determine whether the rate of extraction is dependent

on the coffee species, the blend of coffee or the country

of origin. This paper presents a comparative study of

the extraction of caffeine and mineral ions from coffees

from six different geographical regions. It was hoped

that the kinetic and diffusion data could be used to

distinguish the origin of each coffee bean or to shed
ersity of Natal, Private Bag X01, Scottsville, Pietermaritzburg 3209,

000/$17.50 85

http://crossmark.crossref.org/dialog/?doi=10.1002%2F%28SICI%291097-0010%2820000101%2980%3A1%3C85%3A%3AAID-JSFA489%3E3.0.CO%3B2-%23&domain=pdf&date_stamp=2000-01-05


D Jaganyi, SP Madlala

 10970010, 2000, 1, D
ow

nloaded f
some light on the differences or similarities of the

beans.
rom
 https://scijournals.onlinelibrary.w

iley.com
/doi/10.1002/(SIC

I)1097-0010(20000101)80:1<
85::A

ID
-JSFA

489>
3.0.C

O
;2-# by M

ount K
enya U

niversity, W
iley O

nline L
ibrary on [09/06/2025]. See the T

erm
s and C

onditions (https://onlinelibrary.w
iley.com

/term
s-and-conditions) on W

iley O
nline L

ibrary for rules of use; O
A

 articles are governed by the applicable C
EXPERIMENTAL
The coffee beans investigated were medium roasted

Arabica or Robusta types, namely Kenyan Special,

Sumatran Blue Mountain Java, Brazilian Santos,

Ethiopian Mocha, Zimbabwean coffee and South

African Grown (Colombo Coffee and Tea Co,

Durban). A Glen Creston mill equipped with an agate

mortar and pestle was used to grind the coffee beans,

which were later sieved through a standard set of

stainless steel Endecotts sieves using an Endecotts

sieve shaker machine. The sieved fraction between

1.70 and 2.00mm was selected for the study. All

containers used were made of plastic, and the water

was demineralised Milli-Q-Water (Milli-Pore). The

cleaning process involved using Extran detergent

(Merck), which is phosphate-free, then rinsing the

apparatus with a 0.1M HCl (BDH Aristar) solution.

This precaution was necessary because of the deter-

mination of mineral ions.

A plastic conical ¯ask containing 200ml of water

equilibrated to 80°C in a thermostatically controlled

water bath was used in all the kinetic experiments. The

lid of the ¯ask had two holes, one for the thermometer

and the second for the sampling tube made out of a

thin plastic material. On the outside of the ¯ask the

sampling tube was attached to a 2ml sabre syringe,

while on the inside the end of the tube had a hollow

plastic cone to which a Gilson ®lter (Anachem) was

inserted so as to exclude coffee beans and particulates

during sampling. It has been found11 that simply

dropping tea leaves into hot water tends to give

irreproducible results. Because of this, 4g of coffee

beans were quickly added to the 200ml of water via a

long wide-spout glass funnel. The stopwatch and the

immersible magnetic stirrer attached to the water bath

were started as soon as the coffee was added.

Samples (1ml) were withdrawn at 0.5min intervals

for the ®rst 5min, and the last equilibrium sample after

90min. The samples taken were transferred into 20ml

plastic sample bottles containing 9ml of water and

mixed thoroughly. After each sampling, the sample

tube was cleaned of any solution by attaching a clean

syringe and injecting air through it and the ®lter.

Corrections were made to the concentration of the

solubles12,13 for volume lost through sampling and

evaporation by weighing the ¯ask and the contents

before and after each run.

Reverse phase high-pressure liquid chromatography

(HPLC) was used for the analyses of caffeine. The

instrument employed was a Thermo Separation

Products ®tted with a Spectra System UV 3000

(Scanning) detector which was set at 275nm. The

column used was a Waters Nova-Pak1 C18 60AÊ 4mM

which was calibrated with known concentrations of

caffeine (Fluka). The mobile phase used was similar to

that of Jaganyi and Price.14 The peak integration was
86
performed automatically by the PC1000 software

which operated the HPLC. The mineral ions were

analysed using inductively coupled plasma atomic

emission spectroscopy (ICP-AES). The instrument

was a Varian Liberty AXISO Turbo. The ICP-AES

was calibrated against known concentrations of the

various ions, and the instrument was set to accept

calibration graphs with a correlation coef®cient of

0.995 and over.
RESULTS AND DISCUSSION
The variation in caffeine and mineral ion concentra-

tion C with time t followed ®rst-order behaviour

according to the equation

ln
C1

C1 ÿ C

� �
� kobst � a �1�

which is predicted by the steady state theory of

extraction.15 Here C and C? are the concentrations

of the soluble components at any time t and at

equilibrium, while kobs and a represent the ®rst-order

rate constant and the intercept respectively. Plots of

the ln function against time were indeed linear with

small intercepts for all the coffees.

The mean rate constants obtained from the slopes of

the graphs of eqn (1) and the half-life values t1/2 of

infusion determined from

t1=2 � �ln 2ÿ a�=kobs �2�
are listed in Table 1. Each result is based on at least

three independent experiments whose results were

found to be reproducible. This is indicated by the

standard deviations given in parentheses in Table 1.

These were calculated using MINITAB statistical

software.

Kinetic data
Statistical analysis (one-way analysis of variance) was

performed on the equilibrium concentrations as well

as the rate constants so as to determine which of the

results were similar or signi®cantly different. The

analysis calculates P-factors at two levels: if the value is

lower than 0.01, it means that there is a signi®cant

difference among the means; the opposite is true if the

factor is greater than 0.05.

The analyses of the present results were based on

individual 95% con®dence intervals. The caffeine

equilibrium concentrations fell into two groups: Kenya

and Brazil with a P-factor of 0.28; and Sumatra,

Zimbabwe, Ethiopia and South Africa with a P-factor

of 0.31. Combining the two groups clearly indicated

that there was a signi®cant difference between them

with a P-factor of zero. The concentrations of

phosphorus and magnesium showed no signi®cant

difference, having P-factors of 0.08 and 0.04 respec-

tively. The concentration of manganese from Ethio-

pian beans was very different from all the other beans

investigated which had a P-factor of 0.21. The
J Sci Food Agric 80:85±90 (2000)

reative C
om

m
ons L

icense



Table 1. Mean kinetic data for the infusion of caffeine and mineral ions from different types of coffee beans (1.70–2.00mm) into Milli-Q-Water at 80°C

Type of coffee (country) Species C?(ppm) C?(mM) kobs (10ÿ3sÿ1) Mean intercept a t1/2 (s)

Special Kenya Caffeine 269.8 (�21.5) 1.39 2.71 (�0.10) 0.11 215

(Kenya) K� 365.8 (�10.3) 9.36 3.50 (�0.09) 0.14 176

P (as H2POÿ4 ) 25.1 (�3.8) 0.81 1.89 (�0.16) 0.14 293

Mg2� 28.6 (�0.9) 1.18 1.16 (�0.10) 0.14 477

Mn2� 0.39 (�0.02) 0.0071 0.98 (�0.08) 0.09 615

Santos Caffeine 253.0 (�9.5) 1.30 3.59 (�0.14) 0.08 171

(Brazil) K� 318.2 (�13.7) 8.14 3.95 (�0.12) 0.11 135

P (as H2POÿ4 ) 25.9 (�2.6) 0.84 2.42 (�0.18) 0.13 233

Mg2� 24.3 (�3.1) 1.00 1.55 (�0.14) 0.12 370

Mn2� 0.34 (�0.06) 0.0062 0.98 (�0.09) 0.16 544

Blue Mountain Java Caffeine 292.2 (�16.4) 1.50 3.37 (�0.12) 0.11 173

(Sumatra) K� 376.0 (�17.8) 9.62 4.23 (�0.13) 0.06 150

P (as H2POÿ4 ) 25.2 (�1.6) 0.81 2.13 (�0.15) 0.13 264

Mg2� 27.4 (�1.2) 1.13 1.25 (�0.12) 0.16 427

Mn2� 0.36 (�0.04) 0.0066 0.93 (�0.09) 0.20 530

Zimbabwe Caffeine 295.7 (�12.9) 1.52 3.42 (�0.13) 0.14 162

(Zimbabwe) K� 190.6 (�10.5) 4.87 5.66 (�0.20) 0.25 78

P (as H2POÿ4 ) 22.3 (�1.2) 0.72 2.23 (�0.21) 0.17 235

Mg2� 25.5 (�1.0) 1.05 1.24 (�0.09) 0.91 406

Mn2� 0.42 (�0.02) 0.0076 0.95 (�0.07) 0.13 593

Mocha Caffeine 306.9 (�8.1) 1.58 2.72 (�0.12) 0.12 211

(Ethiopia) K� 345.4 (�12.9) 8.83 3.08 (�0.10) 0.13 183

P (as H2POÿ4 ) 27.5 (�1.5) 0.89 1.96 (�0.19) 0.13 287

Mg2� 27.4 (�1.5) 1.13 1.34 (�0.12) 0.09 450

Mn2� 0.27 (�0.01) 0.0049 1.04 (�0.10) 0.14 532

South African Grown Caffeine 287.6 (�9.2) 1.48 3.06 (�0.11) 0.13 184

(South Africa) K� 300.2 (�14.6) 7.68 4.48 (�0.15) 0.13 126

P (as H2POÿ4 ) 22.6 (�0.4) 0.73 1.87 (�0.13) 0.16 285

Mg2� 23.5 (�2.4) 0.97 1.14 (�0.08) 0.16 468

Mn2� 0.38 (�0.03) 0.0069 0.75 (�0.05) 0.14 738

Kinetics of coffee infusion

 10970010, 2000, 1, D
ow

nloaded from
 https://scijournals.onlinelibrary.w

iley.com
/doi/10.1002/(SIC

I)1097-0010(20000101)80:1<
85::A

ID
-JSFA

489>
3.0.C

O
;2-# by M

ount K
enya U

niversity, W
iley O

nline L
ibrary on [09/06/2025]. See the T

erm
s and C

onditions (https://onlinelibrary.w
iley.com

/term
s-and-conditions) on W

iley O
nline L

ibrary for rules of use; O
A

 articles are governed by the applicable C
equilibrium concentrations which were found to be

dependent on the origin of the coffee were those of

potassium whose P-factor was zero.

Comparing the equilibrium concentrations shown

in Table 1, the amount of potassium leached from the

beans was highest, followed by caffeine. The values for

phosphorus and magnesium were next with similar

magnitude, while manganese was least. This trend was

the same in all the coffees investigated, apart from the

potassium concentration of the Zimbabwean coffee

whose value (190.6ppm) was less than that of caffeine.

It was also 49% smaller than the highest equilibrium

concentration of potassium. The difference between

the potassium and the caffeine concentration of the

South African Grown coffee was only 4% as compared

with � 24% in all the other coffees, with the exception

of Mocha with 11%. This difference in equilibrium

concentration is also observed between the Special

Kenya reported here and Kenyan A.10 The reasons for

this difference may be twofold. First is that the total

potassium content in the coffees investigated was

naturally very low. Second, and the most probable

explanation as to the low level of potassium, is the

handling of the green coffee, whether it underwent the

wet or the dry process.1,5,16

Inspection of the results in Table 1 shows that the

order of infusion is K> caffeine> P (as
J Sci Food Agric 80:85±90 (2000)
H2POÿ4 )>Mg2�>Mn2�. This trend was found to

be true for all the coffees investigated. Since a kinetic

study of caffeine infusion from Kenyan coffee has been

reported in the literature, the rate constant obtained

with the Special Kenyan beans will be examined. The

kobs of 2.71� 10ÿ3sÿ1 obtained here is comparable

with the value of 2.67� 10ÿ3sÿ1 reported by Jaganyi et
al 10 for Kenyan A. Spiro and Selwood15 (1984) have

reported a value of 9.0� 10ÿ3sÿ1 for particle size

0.85±1.18mm at 84.1°C. Since the rate of infusion

varies inversely with the square of the particle

radius15,17 and ln kobs is inversely related to tempera-

ture in the Arrhenius equation, the reported value is

equivalent to a rate constant of 2.68� 10ÿ3sÿ1 using

the present experimental conditions of particle size

and temperature. The agreement with the results given

in Table 1 is very good considering the fact that food

products do vary from year to year.

The statistical analysis of rate constants for different

species revealed that the rate constants of P (as

H2POÿ4 ) were not signi®cantly different (P = 0.12).

In the case of Mg2� the value for the Brazilian coffee

was signi®cantly different from the rest which had a P-

factor of 0.21. The analysis also showed that the values

for Mn2� were similar for all the coffees (P = 0.61),

with the exception of the South African Grown.

Because of the similarities observed in the rate
87

reative C
om

m
ons L

icense



Table 2. Effective diffusion coefficients and hindrance factors of caffeine and mineral ions for various types of coffee beans (1.70–2.00mm) at 80°C

Type of coffee Species

Dbean (from kobs)

(10ÿ11m2sÿ1)

Dbean (from t1/2)

(10ÿ11m2sÿ1)

Mean Dbean

(10ÿ11m2sÿ1)

Daq

(10ÿ9m2sÿ1) HF

Special Kenya Caffeine 19.3 12.2 15.8 2.2 14

(Arabica) K� 22.5 14.9 18.7 5.0 27

P (as H2POÿ4 ) 13.5 8.9 11.2 2.6 23

Mg2� 8.3 5.5 6.9 2.0 29

Mn2� 7.0 4.3 5.7 2.1 37

Santos Caffeine 25.6 15.3 20.5 2.2 11

(Arabica) K� 30.6 19.2 24.9 5.0 20

P (as H2POÿ4 ) 17.3 11.2 14.3 2.6 18

Mg2� 11.1 7.1 9.1 2.0 22

Mn2� 7.0 4.8 5.9 2.1 36

Blue Mountain Java Caffeine 24.0 15.1 19.6 2.2 11

(Arabica) K� 30.2 17.4 23.8 5.0 21

P (as H2POÿ4 ) 15.2 9.9 12.6 2.6 21

Mg2� 8.9 6.1 7.5 2.0 27

Mn2� 6.6 4.9 5.8 2.1 36

Zimbabwe Caffeine 24.4 16.2 20.3 2.2 11

(Med-Arabica) K� 40.4 33.5 37.0 5.0 20

P (as H2POÿ4 ) 15.9 11.1 13.5 2.6 19

Mg2� 8.8 6.4 7.6 2.0 26

Mn2� 6.8 4.4 5.6 2.1 38

Mocha Caffeine 19.4 12.4 15.9 2.2 14

(Robusta) K� 22.0 14.3 18.2 5.0 27

P (as H2POÿ4 ) 14.0 9.1 11.6 2.6 22

Mg2� 9.6 5.8 7.7 2.0 26

Mn2� 7.4 4.9 6.2 2.1 34

South African Grown Caffeine 21.8 14.2 18.0 2.2 12

(Arabica) K� 31.9 20.8 26.4 5.0 19

P (as H2POÿ4 ) 13.3 9.2 11.3 2.6 23

Mg2� 8.1 5.6 6.9 2.0 29

Mn2� 5.3 3.5 4.4 2.1 48

D Jaganyi, SP Madlala

 10970010, 2000, 1, D
ow

nloaded from
 https://scijournals.onlinelibrary.w

iley.com
/doi/10.1002/(SIC

I)1097-0010(20000101)80:1<
85::A

ID
-JSFA

489>
3.0.C

O
;2-# by M

ount K
enya U

niversity, W
iley O

nline L
ibrary on [09/06/2025]. See the T

erm
s and C

onditions (https://onlinelibrary.w
iley.com

/term
s-and-conditions) on W

iley O
nline L

ibrary for rules of use; O
A

 articles a
constants of these three species, they cannot be used

for identi®cation of the coffee origin. However, the

values of caffeine can be used for grouping the coffees.

The statistical analysis for caffeine showed that the rate

constants fell into three groups: Kenya and Ethiopia;

Brazil, Sumatra and Zimbabwe; with South Africa

lying in between. This grouping was con®rmed by

performing a cluster analysis of the rate constants and

the half-life values. The rate constants of K� produced

a P-factor of zero and showed no clustering, an

indication that the individual values were signi®cantly

different. Therefore, even though the Brazilian,

Sumatran and Zimbabwean coffee beans have been

grouped together in the caffeine analysis, they can be

differentiated by using the rate constants for K� which

are different. This is also true for the Kenyan and

Ethiopian coffees. It can be argued that the rate

constants are likely to change from one year's crop to

the next, but if one looks at the rate constant for

caffeine from Kenyan coffee which has been studied

for many years,10,15 it is clear that the change is very

small.
re governed by the applicable C
Diffusion coefficients and hindrance factors
The diffusion coef®cient was calculated in two

different ways. The ®rst calculation made use of the
88
steady state theory equation

kobs � 12Dbean=r
2 �3�

where r and Dbean are the radius of the bean particle

and the diffusion coef®cient of the species within the

bean respectively. The second value was calculated

from the half-life values shown in Table 1 by applying

the following equation,18 which is a solution of Fick's

second law of diffusion:19

1=t1=2 � 32:7Dbean=r
2 �4�

The values obtained from Fick's law were lower than

those obtained using the steady state equation. An

average value Mean Dbean was used in calculating the

hindrance factor HF from the equation

HF � Daq=Dbean �5�

where Daq is the diffusion coef®cient of the species in

aqueous medium. Price et al 20 have determined the

aqueous value for caffeine at 80°C. Spiro and Lam21

have listed in their Table 4 the values for K�, Mg2�

and P (as H2POÿ4 ) assuming that the major orthophos-

phate ion analysed as phosphorus was H2POÿ4 , since

the average pH of the coffee liquor was 4.5. To

determine the limiting tracer diffusion coef®cient for
J Sci Food Agric 80:85±90 (2000)

reative C
om

m
ons L

icense



Kinetics of coffee infusion

 10970010, 2000, 1, D
ow

nloaded from
 https://scijournals.onlinelibrary.w

iley.com
/doi/10.1002/(SIC

I)1097-0010(20000101)80:1<
85::A

ID
-JSFA

489>
3.0.C

O
;2-# by M

ount K
enya U

niversity, W
iley O

nline L
ibrary on [09/06/2025]. See the T

erm
s and C

onditions (https://onlinelibrary.w
iley.com

/term
s-and-conditions) on W

iley O
nline L

ibrary for rules of use; O
A

 articles are governed by the applicable C
Mn2� at 80°C, the Nernst equation

Daq � RT�o=jZjF2 �6�

was used. The symbols R, T, lo, Z and F in eqn (6)

represent gas constant, temperature in kelvin, limiting

equivalent conductance, ionic charge number and

Faraday constant respectively. Since the limiting

equivalent conductance for Mn2� is only available at

25°C,22 the Walden rule

�o� � constant �7�
was used to determine the value of lo for Mn2�. The

values for viscosity (Z) of water were taken from Ref

23. The diffusion coef®cients and the resulting

hindrance factors HF are listed in Table 2.

As regards the HF values, Table 2 shows that Mn2�

is the most hindered species in all the coffees. Its

diffusion coef®cient is approximately three to four

times smaller than that of caffeine, which is the least

hindered of the species investigated. The hydrody-

namic radii in water for Mn2�, Mg2� and caffeine are

very similar as indicated by their Daq values, but their

HF values are very different. The value of Mg2� is

approximately half-way between the other two. Even

though some of the HF values for a particular species

differ for different coffees, their magnitudes remain

very similar.

When the results in Tables 1 and 2 are considered

together, it is clear that the diffusion of caffeine and

mineral ions within the coffee bean particle is a

hindered process. The comparison of the HF values

for caffeine, Mg2� and Mn2� clearly indicates that one

of the causes of slower diffusion is the association

between the species and insoluble components of the

bean. Manganese, being a transition metal, has the

ability to form more complexes than magnesium ion.

Another explanation for the hindrance process is

complex formation between the dissolved species

and other solubles to create a more bulky and

slower-moving entity. One of the compounds sus-

pected to fall into this group is the caffeine±potassium

chlorogenate molecular complex15,18,24,25 which has

been isolated as a 1:1 complex between caffeine and

potassium chlorogenate. Spiro and Chong26 have

demonstrated that the slow infusion of caffeine is not

due to its self-association as was originally suspected.

The equilibrium concentration in Table 1 shows that

the ratio of caffeine to potassium is mostly 1:6. This

big difference may be an indication that potassium

does not only exist as a caffeine±potassium chlorogen-

ate complex in the coffee bean, a fact that can only be

proven by determining the total amounts of the three

species in the bean. Since aqueous K� ions rarely

associate with other species and potassium salts in

water are completely dissociated, it can be concluded

that in whatever form potassium exists in the bean,

during infusion it ionises and the free potassium anion

diffuses out of the bean as a single species much faster

than the organic or the inorganic counterparts. This
J Sci Food Agric 80:85±90 (2000)
idea is clearly supported by the differences in rate

constant and hindrance factor between potassium and

caffeine. In addition to the above explanation, one

factor which is believed to be the main cause of slow

diffusion for all the species is the physical restraint

within the bean matrix. This forces the solubles to

follow a diffusion path which is tortuous in nature.
ACKNOWLEDGEMENTS
The authors thank University of Natal Research and

Development for funding this work, and Professor M

Spiro for his critical comments and suggestions.
REFERENCES
1 Ferreira LAB, Fragoso MA, Peratta MF, Silvo MCC and

Rebeto MC, Mineral constituents in coffees of Angola. 5th

Colloq Sci Int CafeÂ, Paris, pp 51±62 (1971).

2 Rof® JA, Corte dos Santos A, Mexia JT, Busson F and Maigrot

M, Green and roasted Angola coffees. Chemical studies. 5th

Colloq Sci Int CafeÂ, Paris, pp 179±200 (1971).

3 Tserevitnov OB, Padaryan EM, Andreeva EV and Buneeva PI,

Atomic-absorption spectroscopy in studying the mineral

composition of green coffee. Vop Pitan 30:85±90 (1972).

4 Quijano Rico M and Spettel B, Determination of the content of

various elements in samples of different varieties of coffee. 7th

Colloq Sci Int CafeÂ, Paris, pp 165±173 (1975).

5 Clarke RJ and Walker LJ, Potassium and other mineral contents

of green, roasted and instant coffees. J Sci Food Agric 25:1389±

1404 (1974).

6 Clarke RJ and Macrae R, Coffee: Volume 1 Chemistry, Elsevier

Applied Science, London, pp 42±82 (1985).

7 Clarke RJ and Walker LJ, The interrelation of potassium

contents of green, roasted and instant coffees. 7th Colloq Sci

Int CafeÂ, Paris, pp 159±163 (1975).

8 Kroplien U, Green and Roasted Coffee Tests. Gordian, Hamburg

(1961).

9 Martin MJ, Pablos F and Gonzalez AG, Characterization of

green coffee varieties according to their metal content. Anal

Chim Acta 358:177±183 (1998).

10 Jaganyi D, Vanmare J and Clark T, Kinetics of coffee infusion:

kinetic study of mineral ion extraction from Kenyan Arabica

coffee. J Sci Food Agric 79:323±326 (1999).

11 Spiro M and Siddique S, Kinetics and equilibria of tea infusion.

Part 1: Analysis and partition constants of thea¯avins,

thearubigins and caffeine in Koomsong Broken Pekoe. J Sci

Food Agric 32:1027±1032 (1981).

12 Spiro M and Jago D, Kinetics and equilibria of tea infusion. Part

3: Rotating disc experiments interpreted by a steady-state

model. J Chem Soc, Faraday Trans 1 78:295±305 (1982).

13 Jaganyi D, Aspects of tea brewing. PhD Dissertation, University of

London (1992).

14 Jaganyi D and Price RD, Kinetics of tea infusion: the effect of the

manufacturing process on the rate of extraction of caffeine.

Food Chem 64:27±31 (1998).

15 Spiro M and Selwood RM, The kinetics and mechanism of

caffeine infusion from coffee: the effect of particle size. J Sci

Food Agric 35:915±924 (1984).

16 Menchu JF and Ibarra E, The chemical composition and the

quality of Guatemala coffee. 3rd Colloq Sci Int CafeÂ, Paris, pp

146±154 (1967).

17 Spiro M, The rate of caffeine infusion from Kenyan Arabica

coffee beans. 12th Colloq Sci Int CafeÂ, Paris, pp 260±264

(1988).

18 Spiro M, Toumi R and Kandiah M, The kinetics and mechanism

of caffeine infusion from coffee: the hindrance factor in intra-

bean diffusion. J Sci Food Agric 46:349±356 (1989).
89

reative C
om

m
ons L

icense



D Jaganyi, SP Madlala

 10970010, 2000, 1, D
ow

nloaded from
 https://scijournals.onlinelibrary.w

iley.com
19 Crank J, The Mathematics of Diffusion, Oxford University Press,

Oxford (1975).

20 Price NE, Trickett KA and Harris KR, Association of caffeine in

aqueous solution: effects on caffeine intradiffusion. J Chem

Soc, Faraday Trans 1 85:3281±3288 (1989).

21 Spiro M and Lam P-LL, Kinetics and equilibria of tea infusion.

Part 12: Equilibrium and kinetic study of mineral ion extrac-

tion from black Assam Bukial and green Chun Mee teas. Food

Chem 54:393±396 (1995).

22 CRC Handbook of Chemistry and Physics, 72nd edn, CRC Press,

Boca Raton, FL, pp 5-96±5-97 (1991).
90
23 Robinson RA and Stokes RH, Electrolyte Solutions, 2nd edn,

Butterworths, London, pp 317±457 (1959).

24 Haslam E, Begley MJ, Martin R, Lilley TH, Falshaw CP and

Magnolato D, The caffeine±potassium chlorogenate molecular

complex. Phytochemistry 26:273±279 (1987).

25 Horman I and Viani R, The nature and conformation of the

caffeine±chlorogenate complex of coffee. J Food Sci 37:925±

927 (1972).

26 Spiro M and Chong YY, The kinetics and mechanism of caffeine

infusion from coffee: the temperature variation of the

hindrance factor. J Sci Food Agric 74:416±420 (1997).
J Sci Food Agric 80:85±90 (2000)

/doi/10.1002/(SIC
I)1097-0010(20000101)80:1<

85::A
ID

-JSFA
489>

3.0.C
O

;2-# by M
ount K

enya U
niversity, W

iley O
nline L

ibrary on [09/06/2025]. See the T
erm

s and C
onditions (https://onlinelibrary.w

iley.com
/term

s-and-conditions) on W
iley O

nline L
ibrary for rules of use; O

A
 articles are governed by the applicable C

reative C
om

m
ons L

icense