Journal of Food Engineering 200 (2017) 13e21

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Selective decaffeination of tea extracts by montmorillonite

Takashi Shiono*, Kenichiro Yamamoto, Yuko Yotsumoto, Junichiro Kawai, Nanami Imada,
Jumpei Hioki, Hiroyuki Naganuma, Takahiro Eguchi, Makoto Kurihara, Aruto Yoshida,
Keiji Deuchi
Research Laboratories for Beverage Technologies, Research & Development Division, Kirin Company, Ltd., 1-17-1, Namamugi, Tsurumi-ku, Yokohama,
Kanagawa 230-8628, Japan

a r t i c l e i n f o a b s t r a c t

Article history: We evaluated the decaffeination of green tea extract via montmorillonite (MMT) to develop a selective
Received 19 August 2016 and versatile decaffeination technology. MMT is a clay mineral used worldwide in the food industry.
Received in revised form MMT adsorbed caffeine with no signicant binding to catechins, whereas activated carbon (AC) bound
16 December 2016
caffeine and catechins simultaneously. This suggests that MMT has a higher selectivity for caffeine
Accepted 22 December 2016
adsorption than AC. The taste of the extract decaffeinated by MMT was not signicantly altered. MMT
Available online 23 December 2016
stably adsorbed caffeine at 5e35
C and pH 6e8; however, prolonged incubation caused an undesired
elution of Fe ions, which decreases the lightness of green tea beverages. MMT exhibited similar caffeine
Chemical compounds studied in this article:
Caffeine (PubChem CID: 2519)
adsorption properties with various types of tea extracts and a simple caffeine solution. Overall, our
Epigallocatechin (Pubchem CID: 72277) ndings suggest that MMT is a useful adsorbent for the decaffeination of a range of tea beverages.
Epicatechin (Pubchem CID: 72276) 2016 Elsevier Ltd. All rights reserved.
Epigallocatechin gallate (Pubchem CID:
65064)
Epicatechin gallate (Pubchem CID: 107905)

Keywords:
Activated carbon
Caffeine
Decaffeination
Langmuir isotherm
Montmorillonite
Tea

1. Introduction (Yanagimoto et al., 2003), anti-microbial (Hamilton-Miller, 1995),

Tea is one of the most widely consumed beverages in the world.
It is prepared by pouring hot water over cured leaves of Camellia
sinensis L. The major types of tea depend upon the manufacturing
process, but generally include green tea, oolong tea, and black tea.
Related tea products include tea leaves, packaged beverages and
instant tea. The production volume of these tea products world-
wide was 5 million tons in 2013 (Chang, 2015).
Tea contains polyphenols, including catechins, theaavins, and/
or thearubigin; these polyphenols have attracted attention in
recent years because of their positive effects on health (Khokhar
and Magnusdottir, 2002; Perva-Uzunali c et al., 2006). Tea poly-
phenols exhibit numerous health benets, such as anti-oxidative
* Corresponding author.
E-mail address: Takashi_Shiono@kirin.co.jp (T. Shiono).
anti-allergic (Suzuki et al., 2000; Yoshino et al., 2010), hypoglyce-
mic (Matsumoto et al., 1993; Toyoda-Ono et al., 2007) and anti-
obesity effects (Ueda and Ashida, 2012; Uchiyama et al., 2011).
Caffeine is an important compound that accounts for about 2%e
5% of constituents in the tea leaf (Naik and Nagalakshmi, 1997). It
has a bitter taste and exerts both positive and negative physiolog-
ical effects. For example, positive effects of caffeine include
improvement of drowsiness, cognitive decline, and fatigue; these
effects are facilitated via stimulation of the central nervous system
(Lorist and Tops, 2003; Arab et al., 2013). Some of negative effects of
caffeine include sleep disorder (Hindmarch et al., 2000), elevation
of blood pressure (Nurminen et al., 1999) and risk of fetal growth
disorder (Fernandes et al., 1998). These effects of caffeine vary
among individuals according to differences in their caffeine meta-
bolic rate, which can be affected by age, pregnancy, and smoking
habits (Von Borstel, 1983). Therefore, limited tea intake might be
recommended, depending on the situation.

http://dx.doi.org/10.1016/j.jfoodeng.2016.12.015
0260-8774/ 2016 Elsevier Ltd. All rights reserved.
14 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21
Abbreviations
MMT montmorillonite
AC activated carbon
EGC epigallocatechin
EC epicatechin
EGCg epigallocatechin gallate
ECg epicatechin gallate
Alternatively, consumers may choose decaffeinated tea. To date,
two approaches have been used for the decaffeination of tea: one is
caffeine removal from tea leaves during the processing of the raw
materials; and the other is caffeine elimination after tea extraction.
The former approach extracts caffeine using hot water (Vuong et al.,
2013) or supercritical carbon dioxide (Huang et al., 2007; Park et al.,
2007; Tang et al., 2010). In some countries, organic solvents such as
ethyl acetate, dichloromethane, and dimethyl ether are also used
(Kanda et al., 2013). The latter approach eliminates caffeine from
tea extract using an adsorbent such as activated carbon (AC),
lignocellulose, poly(acrylamide-co-ethylene glycol dimethylacry-
late), or porous polymeric resin (Sakanaka, 2003; Ye et al., 2007; Ye
et al., 2009; Lu et al., 2010; Sevillano et al., 2014). None of the above
methods are completely satisfactory for consumers because, in
addition to caffeine, they reduce other substances such as catechins
in tea (Lee et al., 2007, 2009).
Montmorillonite (MMT) is a layered clay mineral belonging to
the smectite group and a main component of naturally mined
bentonite. In the food industry, it is used as a processing aid for the
clarication of wine and juice, as well as the bleaching of oils
(Gonza lez-Pradas et al., 1993; Tajchakavit et al., 2001; Yildirim,
2011). MMT particles comprise approximately 1 nm thick plate
crystals; they are composed of a basic three-layer structure of silica
tetrahedral sheets of Si4O6(OH)4 that sandwich a central alumina
octahedral sheet of Al(OH)6 (Bergaya et al., 2006). The Al in the
alumina central layer partially undergoes isomorphous substitution
with Mg and Fe, causing the plate crystal to carry a negative charge;
therefore, positively charged ions are present between the layers
for replenishment. MMT swells with water and has a capacity for
cation exchange; it is capable of adsorbing organic cationic sub-
stances, e.g., methylene blue and alkylammonium ions, as well as
inorganic cations such as Cs and Rb (Hurel et al., 2009; Sarma
et al., 2011; Okada et al., 2014). In addition, MMT adsorbs non-
ionic organic compounds such as 3-aminotriazole and benzimid-
azole fungicides (Russell et al., 1968; Aharonson and Kafka, 1975).
Last year, we reported that MMT and saponite, both belonging to
the smectite group, adsorbed caffeine in aqueous solutions.
Furthermore, caffeine adsorption was improved by modifying MMT
with benzylammonium (Okada et al., 2015). Similarly, Maral et al.
reported that saponite that had been organically modied with 3-
aminopropyltriethoxysilane adsorbed caffeine (Maral et al., 2015).
To date, however, there has been no research involving a detailed
examination of the caffeine adsorption properties of these adsor-
bents in tea extracts.
In this study, we evaluated the utility of MMT for caffeine
adsorption in various tea beverages with the aim of developing a
selective and versatile decaffeination technology.
2. Materials and methods
2.1. Materials and reagents
Adsorbents MMT (MIZULITE) and AC (FP-3) were obtained from
Mizusawa Industrial Chemicals, Ltd. (Tokyo, Japan) and Osaka Gas
Chemicals, Co., Ltd. (Osaka, Japan), respectively. Caffeine and cate-
chins (epigallocatechin, EGC; epicatechin, EC; epigallocatechin
gallate, EGCg; epicatechin gallate, ECg) were purchased from Wako
Pure Chemical Industries, Ltd. (Osaka, Japan) and Funakoshi, Co.
Ltd. (Tokyo, Japan), respectively. Various tea leaves, green tea
(Sencha in Shizuoka, Japan), oolong tea (Tie-Guan-Yin in China and
Taiwan, Dong-Ding-Oolong in China), and black tea (Dimbula in Sri
Lanka and Darjeeling in India) were purchased from specialized tea
shops in Japan.
2.2. Adsorption test of caffeine and catechins
The adsorption test of caffeine and catechins was carried out
with green tea extract, which was prepared as follows: 100 g of tea
leaves was added to 1000 mL of hot water at 80
C and extracted for
8 min. Either 160e2000 mg of MMT or 32e200 mg of AC was added
to 40 mL of the diluted green tea extract, whose nal caffeine
concentration was 3.6 mmol/L. After incubation at 25
C for
120 min, the suspension was centrifuged at 1920g for 10 min, and
the resulting supernatant was ltered through a membrane lter
(0.2 mm, PTFE). The concentrations of caffeine and catechins (EGC,
EC, EGCg and ECg) were measured by HPLC (LC-2000Plus, JASCO,
Corp., Tokyo, Japan). The HPLC conditions were: CAPCELL PAK C18
UG120 (4.6 mm I.D. x 150 mm, 3 mm, Shiseido Co., Ltd., Tokyo,
Japan), column temperature 40
C, ow rate 0.9 mL/min, mobile
phase A water/acetonitrile/phosphoric acid (1000/25/1.0, v),
mobile phase B water/methanol/acetonitrile/phosphoric acid
(600/300/15/0.6, v), mobile phase C methanol/acetonitrile/
phosphoric acid (800/200/1.0, v), step gradient: 0e3 min 94% A
and 6% B, 3e8 min from 94% A and 6% B to 50% A and 50% B,
8e22 min from 50% A and 50% B to 0% A and 100% B,
22e24 min from 100% B to 0% B and 100% C, after 5 min 100% C,
detector UV (caffeine, 275 nm; catechins, 230 nm). The taste quality
of green tea extracts before and after decaffeination by MMT or AC
was compared by ve tea experts; these experts were qualied
either as Japanese Tea Instructors or had been trained in sensory
evaluation for more than two years.
For the adsorption test of a simple caffeine solution,
160e2000 mg of MMT or 16e800 mg of AC was added to 40 mL of a
3.6 mmol/L caffeine solution, and the residual caffeine concentra-
tion was measured by HPLC as described above.
2.3. Kinetics of caffeine adsorption
Either 800 mg of MMT or 120 mg of AC was added to 40 mL of
the caffeine solution (5.2 mmol/L of caffeine) or green tea extract
(5.2 mmol/L of caffeine) and incubated from 2 to 240 min at 25
C.
The residual caffeine concentration was measured by HPLC as
described above. The amount of caffeine adsorbed per weight of
adsorbent at t (min) contact time (Qt, mmol/g) was calculated by
the equation:
Qt V  C0  Ct =M (1)
where V (L) is the volume of green tea extract; C0 (mmol/L) is the
initial caffeine concentration in green tea extract before adsorption;
Ct (mmol/L) is the caffeine concentration in green tea extract at t
contact time; and M (g) is the dry weight of adsorbent.
Subsequently, we tested whether the caffeine adsorption pro-
cess coincides with a pseudo-rst order kinetic model or a pseudo-
second order kinetic model. The linear form of the pseudo-rst
order model (Dang et al., 2009) is given by the equation:
 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21
lnQe  Qt lnQe  K1  t (2)
where Qe (mmol/g) and Qt are the amounts of caffeine adsorbed per
weight of adsorbent at equilibrium and t contact time, respec-
tively; and K1 (1/min) is the adsorption rate constant.
The linear form of the pseudo-second order model (Ho and
McKay, 1999) is given by the equation:
.  water temperature was altered to 95
C. The caffeine concentra-
t=Qt 1 K2  Qe2 t=Qe
where Qe and Qt are the amounts of caffeine adsorbed per weight of
adsorbent at equilibrium and t contact time, respectively; and K2
(g/mmol/min) is the adsorption rate constant.
2.4. Analysis of caffeine adsorption isotherm
By using the results of the caffeine adsorption test, the mecha-
nism of caffeine adsorption was also evaluated using Freundlich
and Langmuir models. The linear form of the Freundlich adsorption
isotherm is given by the equation:
lnQe 1=n  lnCe lnKF (4)
where Qe (mmol/g) is the amount of caffeine adsorbed per weight
of adsorbent at equilibrium; Ce (mmol/L) is the equilibrium caffeine
concentration; and n and KF are the Freundlich constants.
The linear form of the Langmuir adsorption isotherm is given by
the equation:
Qe KL  Qm  Qe =1 KL  Ce (5)
where Qe is the amount of caffeine adsorbed per weight of adsor-
bent at equilibrium; Qm (mmol/g) is the maximum adsorption ca-
pacity of caffeine; Ce is the equilibrium caffeine concentration; and
KL (L/mmol) is the Langmuir constant.
2.5. Analysis of caffeine adsorption conditions
To determine the effects of pH, temperature, and time, on
caffeine adsorption, 8.0 g of MMT was added to 400 mL of diluted
green tea extract. The contact pH of the green tea extract was
changed from 4 to 8 by using a McIlvaine buffer prepared from
0.2 M Na2HPO4 and 0.1 M citric acid (McIlvaine, 1921). The incu-
bation was carried out at 25
C for 60 min. After centrifugation and
ltration of the suspension as described above, ascorbic acid or
NaHCO3 was added to the decaffeinated green tea extract to adjust
the pH to 6.6. Alternatively, the contact temperature was raised
from 5 to 35
C at pH 5.8 for 60 min. The contact time was altered
from 0.5 to 60 min at pH 5.8 and 25
C. After centrifugation and
ltration of the suspension as described above, 400 mg of ascorbic
acid was added to 400 g of the decaffeinated tea extract. Subse-
quently, the pH of the extract was readjusted to 6.6 by the addition
of NaHCO3. Finally, 1000 g of the decaffeinated green tea beverage
was made by adding water. The concentrations of caffeine and Fe
ion were measured by HPLC and ICP-AES (iCAP6500Duo, Thermo
Fisher Scientic, K. K., Yokohama, Japan), respectively. Lightness (L*
value) was measured based on a Hunter Lab color space using a
CM3500-d spectrophotometer equipped with Spectra Magic soft-
ware (Konica Minolta, Inc., Tokyo, Japan) as previously reported
(Chugh et al., 2014).
To assess the appearance of the beverage, treatment of the green
tea extract with MMT was carried out at 5, 20, or 35
C for 60 min. If
necessary, the pH of the extract was adjusted to 7.3 with NaHCO3. In
parallel, FeSO4 was added to the green tea extract at a nal
15
concentration of 1e50 mmol/L for color reference.
2.6. Caffeine adsorption properties in various types of tea
The caffeine adsorption properties of MMT were analyzed in
various types of tea. Extraction of oolong tea and black tea was
carried out under the conditions used for green tea except that the
(3) tions of these tea extracts were between 7.3 and 10.3 mmol/L, and
diluted extracts ranging from 0.7 to 6.2 mmol/L were prepared. For
the adsorption test, 400 mg of MMT was added to 40 mL of each
diluted extract. From the results, Ce and Qe were calculated to draw
adsorption isotherms, and then Freundlich parameters (n and KF)
and Langmuir parameters (Qm and KL) were calculated using
equations (4) and (5) described above. The data were evaluated via
Pearson correlation using SPSS Statistics software (IBM Japan, Ltd.,
Tokyo, Japan) to determine the relationship between Ce and Ce/Qe
among tea extracts.
3. Results and discussion
3.1. Comparison of caffeine adsorption in green tea extract
Our preliminary experiment suggested that MMT and AC were
potential candidates for decaffeination; these two candidates are
among approximately 100 adsorbents that are permitted for use in
food manufacturing in Japan. We tested the ability of these two
adsorbents to adsorb caffeine from a green tea extract and
compared the proportion of remaining catechins along with the
caffeine removal rate (Fig. 1). In the case of MMT, the caffeine
removal rate reached 96.5%, while the proportion of remaining
catechins was not affected greatly. A maximum of 6.7% of the total
catechins was adsorbed. By contrast, during AC treatment, the
proportion of remaining catechins gradually decreased as the
caffeine removal rate increased. At a 96.5% caffeine removal rate, AC
removed 54.0% of total catechins from the green tea extract.
HPLC analyses of caffeine and catechins (EGC, EC, EGCg, and ECg)
in the MMT and AC samples are shown in Figs. 2 and 3. To remove
96.5% of the caffeine from a green tea extract containing 3.3 mmol/L
of caffeine (Fig. 3a), the minimal addition rates of MMT and AC were
5.0 and 0.5 wt%, respectively. In the MMT-treated extract, no
marked decrease in the peak areas of catechins was observed
(Fig. 2a and b), and the percentages of EGC, EC, EGCg, and ECg
remaining were 96.8%, 97.1%, 86.6%, and 81.1%, respectively
(Fig. 3b). Although the amounts of catechins bound to MMT were
Fig. 1. Effect of decaffeination treatment on the percentage of caffeine removed and
the percentage of catechins remaining. Closed circle, MMT; open circle, AC.
16 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21
Fig. 2. HPLC chromatograms of green tea extract before and after decaffeination by
MMT or AC. (a,b) Before and after decaffeination by MMT, respectively; (c) after
decaffeination by AC. 1, EGC; 2, caffeine; 3, EGCg; 4, EC; 5, ECg.
Fig. 3. Concentration of caffeine and catechins in green tea extract before and after
decaffeination by MMT or AC. (a) Caffeine; (b) catechins.
marginal, the results suggest that a gallate group might enhance
the binding of catechins to MMT. By contrast, the peak areas of all of
the catechins were signicantly decreased alongside decaffeination
with AC (Fig. 2a and c), and the percentage of EGC, EC, EGCg, and
ECg remaining were 41.6%, 35.4%, 58.3%, and 57.8%, respectively
(Fig. 3b). These data suggested that AC adsorbed all four kinds of
catechins, and bound relatively more non-gallate-type catechins
than gallate-type catechins.
MMT treatment increased the ratio of total catechins to caffeine
in the green tea extract from 1.88 to 50.4, whereas AC treatment
showed a limited increase to 25.0; this value is comparable to a
previous study in which AC treatment showed a ratio of 23.6 at a
caffeine removal rate of 90% (Ye et al., 2007). Likewise, the ratio of
the amount of caffeine to EGCg adsorbed was 13.8 for MMT, as
compared with only 4.4 for AC. The value for MMT was also better
than the value for supercritical carbon dioxide extraction of 2.57
(Kim et al., 2008). Taken together, these ndings indicate that MMT
preferentially adsorbed caffeine but not catechins in green tea
extract, and the selectivity of MMT was superior to both AC and
carbon dioxide extraction. In accordance with these alterations in
composition, the taste of the tea extract decaffeinated by MMT
changed very little, while those decaffeinated by AC varied
markedly.
3.2. Kinetics of caffeine adsorption
The kinetics of caffeine adsorption in a caffeine solution and
green tea extract are shown in Fig. 4. Because the standard de-
viations were too small to visualize, the raw data have been
included as Supplementary Fig. S1 and Table S1. MMT quickly
adsorbed more than 99% of caffeine within 5 min from the caffeine
solution, and within 20 min from the green tea extract (Fig. 4a).
However, the adsorption of caffeine by AC in the caffeine solution
and green tea extract differed signicantly (Fig. 4b). Although AC
adsorbed caffeine within 20 min in the caffeine solution, caffeine
adsorption was much slower in the green tea extract and took more
than 240 min to complete.
Subsequently, we tried to t these experimental data to the
Fig. 4. Caffeine adsorption kinetics curves of MMT and AC. (a) MMT; (b) AC. Open
circle, caffeine solution; closed circle, green tea. The bars indicate standard deviation
(n 3).
 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21 17

Table 1
Kinetic parameters of caffeine adsorption by MMT and AC.

MMT AC

Caf. solution Green tea Caf. solution Green tea

Exp. Qe (mmol/g) 0.226 0.0003 0.193 0.0002 1.747 0.0012 1.176 0.0114
Pseudo-rst order
Cal. Qe (mmol/g) 0.0015 0.0001 0.0047 0.0012 0.0392 0.0021 0.429 0.0548
K1 (1/min) 0.0086 0.0031 0.0188 0.0100 0.0222 0.0029 0.0170 0.0040
r2 0.683e0.779 0.714e0.847 0.581e0.692 0.878e0.973
Pseudo-second order
Cal. Qe (mmol/g) 0.225 0.0002 0.193 0.0002 1.748 0.0008 1.189 0.0075
K2 (g/mmol/min) 40.72 10.43 20.09 1.971 2.482 0.0126 0.144 0.0064
r2 1.000 1.000 1.000 0.999

Values are presented as the mean standard deviation of triplicate measurements.

was much closer to the calculated Qe of the pseudo-second order

model than that of the pseudo-rst order model (Table 1). The
values of the correlation coefcients were all extremely high
(r2 > 0.999). Therefore, the adsorption of caffeine onto MMT and AC
is considered to follow pseudo-second order kinetics. While the
rate constants of the pseudo-second order model (K2) of MMT in
the caffeine solution and green tea extract were of the same order,
those of AC showed a signicant difference of one order of
magnitude. These results conrm the higher selectivity of MMT
compared with AC in the decaffeination of green tea extract.

3.3. Caffeine adsorption isotherms

Fig. 5a shows the caffeine adsorption isotherms of MMT in the

Fig. 5. Caffeine adsorption isotherms of MMT and AC. (a) MMT; (b) AC. Open circle,
caffeine solution; closed circle, green tea.
equations of pseudo-rst order and pseudo-second order models.
Regardless of the type of adsorbent or solution, the experimental Qe
caffeine solution and green tea extract. The standard deviations are
documented in Table S2 and Fig. S2. As reported previously (Okada
et al., 2015), the data from the caffeine solution converged to a
certain Qe value with an increase in Ce; these data coincided with
the Freundlich adsorption model and the Langmuir adsorption
model, showing a high correlation coefcient (r2 > 0.951 and 0.992,
respectively) (Table 2). The isotherm in the green tea extract was
similar to that in the caffeine solution (Fig. 5a); it also tted well to
the Freundlich adsorption model and the Langmuir adsorption
model, with a high correlation coefcient (r2 > 0.960 and 0.995,
respectively) (Table 2). Moreover, both showed slightly higher
correlation coefcients in the Langmuir adsorption model than in
the Freundlich adsorption model. Although Qm was almost the
same in green tea as in the caffeine solution, KL was slightly lower in
green tea than in the caffeine solution (Table 2). These observations
suggest that the adsorption of caffeine by MMT was not signi-
cantly affected by other substances in the green tea extract.
By contrast, the caffeine adsorption isotherms of AC differed
signicantly between the caffeine solution and green tea extract
(Fig. 5b). Moreover, both showed slightly higher correlation co-
efcients in the Freundlich adsorption model (caffeine solution,

Table 2
Isotherm parameters of caffeine adsorption by MMT and AC.

MMT AC

Caf. solution Green tea Caf. solution Green tea

Freunclich model
1/n 0.373 0.0007 0.491 0.0048 0.0885 0.0001 0.176 0.0019
KF 0.248 0.0001 0.179 0.0004 2.167 0.0030 1.008 0.0124
r2 0.951e0.955 0.960e0.963 0.973e0.977 0.889e0.927
Langmuir model
Qm (mmol/g) 0.346 0.0015 0.327 0.0035 2.334 0.0089 1.240 0.0301
KL (L/mmol) 3.326 0.0470 1.463 0.0347 18.59 1.252 5.647 0.4509
r2 0.992e0.993 0.995 0.878e0.899 0.825e0.898

Values are presented as the mean standard deviation of triplicate measurements.

18 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21

r2 > 0.973; green tea extract, r2 > 0.889) than in the Langmuir
adsorption model (Table 2). Because the Freundlich parameters (n
and KF) and Langmuir parameters (Qm and KL) differed greatly be-
tween the green tea extract and caffeine solution (Table 2), we
posited that the adsorption of caffeine by AC might be reduced by
two potential factors. The rst factor might be competitive inhibi-
tion resulting from the occupation of the caffeine adsorption site
with catechins or other substances in green tea. In support of this, it
was previously reported that AC adsorbed catechins according to
the Langmuir model (Gogoi et al., 2010). The second factor might be
a decrease in binding energy caused by substances around the
caffeine molecules in green tea. In support of this, the caffeine
adsorption of AC is known to be inuenced by ethanol
concentration (Ye et al., 2007). Regardless of the mechanism, our
observations indicated that the adsorption of caffeine by MMT is
more selective and stable than that by AC.
3.4. Effect of decaffeination conditions
Next, we examined the effect of contact pH, temperature, and
incubation time on the decaffeination process with MMT. Note that
the pH of the sample was readjusted to approximately 6.6 after
each treatment. The amount of adsorbed caffeine per weight (Qe) of
MMT was essentially the same, regardless of the contact pH, tem-
perature, and incubation time, although a slight increase in Qe was
observed below pH 5 (Fig. 6a, c, and e). This result was consistent

Fig. 6. Effect of MMT decaffeination condition on amount of adsorbed caffeine (Qe), Fe elution, and L* value of decaffeinated green tea beverages. (a) Contact pH vs Qe; (b) contact pH
vs Fe elution and L* value; (c) contact temperature vs Qe; (d) contact temperature vs Fe elution and L* value; (e) contact time vs Qe; (f) contact time vs Fe elution and L* value. Closed
circle, adsorbed caffeine; open square, Fe concentration; closed square, L* value.
 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21
with a previous study showing that the caffeine adsorption ca-
pacity of Ca-MMT in a caffeine solution was stable within pH 4e10
(Lailach et al., 1968).
During these tests, we noticed that some samples exhibited a
slight change in color properties, such as increased dullness; as a
result, we examined the L* value and Fe concentration of the ex-
tracts in detail. As the contact pH dropped, the L* value decreased
and the Fe concentration conversely increased (Fig. 6b). Similarly, a
higher contact temperature and longer contact time lowered the L*
value with a converse increase in Fe concentration (Fig. 6d and f). It
was apparent that these changes in L* value and Fe concentration
occurred independently of the adsorption of caffeine by MMT,
which remained stable (Fig. 6). More specically, these ndings
clearly indicated that the increase in Fe concentration was not due
to a substitution effect caused by caffeine adsorption to MMT.
The appearance of decaffeinated green tea beverages prepared
under different conditions of contact pH and temperature with
MMT are shown in Fig. 7a. For a contact pH of both 5.8 and 7.3, the
lightness of the beverages decreased as the contact temperature
increased. Similarly, for a contact temperature of both 5
C and
20
C, the lightness of the beverages decreased as the contact pH
dropped. These changes in appearance were extensively supported
by the decrease in lightness. At 35
C, by contrast, the lightness
increased with the lower contact pH. This phenomenon was likely
due to the co-occurrence of dulling and browning, as discussed
below.
We considered that the decrease in the lightness of the decaf-
feinated green tea beverages was caused by the formation of a
complex consisting of catechins and Fe ions eluted from MMT. As
shown in Fig. 6b, d, and f, a decrease in L* value was inversely
correlated with an increase in Fe concentration, suggesting that the
eluted Fe ions caused the decrease in L* value. Many decades ago, it
was reported that Fe ions make a complex with catechins in green
tea (Tsujimura, 1934; Oshima, 1936). Indeed, the addition of FeSO4
to the green tea beverage led to a dull color, and the dullness
depended on the amount of Fe ions (Fig. 7b). It is also known that Fe
Fig. 7. Effect of MMT decaffeination condition and Fe ion addition on appearance of
green tea beverages. (a) Appearance of decaffeinated green tea beverages prepared
under different conditions of MMT decaffeination; (b) appearance of green tea bev-
erages with Fe ion added.
19
ions are eluted from clay minerals such as MMT (Catarino et al.,
2008), and that the dissolution of Fe ions from saponite is accel-
erated by acid treatment and high temperature (Vicente Rodriguez
et al., 1994). These features are in accordance with our results
indicating that the concentration of Fe ions eluted from MMT
increased with lower contact pH and higher contact temperature
(Fig. 6b and d).
The liquid color of the green tea beverage prepared at pH 7.3 and
35
C turned reddish brown (Fig. 7a). This can be potentially
explained by the oxidation of catechins at a mildly alkaline pH and
high temperature. It has been reported that the color of catechins
changes to brown at pH 6 or higher (Komatsu et al., 1993) and that
iron catalyzes the oxidation of catechins, a process accompanied by
browning (Oszmianski et al., 1996).
Because the appearance of a green tea beverage is a critical
factor reecting its qualities, it is essential to avoid deterioration of
the liquid color during the decaffeination process. Thus, to prevent
the elution of Fe ions and the resulting color change, we propose
using a lower contact temperature, shorter time, and neutral pH in
the decaffeination of green tea with MMT. A low temperature
seems to be especially effective because caffeine adsorption can be
stably performed without the dissolution of Fe ions and accom-
panying color change.
3.5. Effect of tea type on caffeine adsorption properties
The caffeine adsorption isotherms and Langmuir plots of MMT
Fig. 8. Effect of tea type on the caffeine adsorption properties of MMT. (a) Caffeine
adsorption isotherms of MMT in various tea extracts; (b) Langmuir plots. Closed circle,
green tea (sencha from Japan); closed square, oolong tea (Tie Guan Yin from China);
open square, oolong tea (Tie Guan Yin fromTaiwan); gray square, oolong tea (Dongding
from China); closed triangle, black tea (Dimbula from Sri Lanka); open triangle, black
tea (Darjeeling from India).
20 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21
in a green tea extract, three oolong tea extracts, and two black tea
extracts are shown in Fig. 8a and b. The data demonstrated that the
isotherms were nearly identical, regardless of the fermentation
level or the locality.
Although experimental Ce and calculated Qe were satisfactorily
tted to both the Freundlich equation and the Langmuir equation,
the latter model consistently gave a better correlation coefcient
(data not shown). We assessed the relationship between Ce and Ce/
Qe using the Pearson correlation. The result showed a positive and
signicant relationship (P < 0.001) (r 0.999), suggesting that the
caffeine adsorption properties of MMT were more or less consistent
regardless of the type of tea.
It is known that the content of caffeine and polyphenols in green
tea and black tea varies widely, depending on both fermentation
level and locality (Astill et al., 2001). Nonetheless, our observations
suggest that MMT selectively adsorbed sufcient amounts of
caffeine in various types of tea extract without any signicant
interference from other components such as polyphenols. There-
fore, it seems likely that MMT can be widely applied to the decaf-
feination of various tea beverages.
4. Conclusions
In this study, we found that MMT selectively adsorbed caffeine
but not catechins in green tea extracts, and its selectivity was much
higher than that of AC. Furthermore, the taste of the decaffeinated
tea extract was minimally changed by treatment with MMT. MMT
quickly adsorbed caffeine regardless of the temperature or pH
examined and could be applied similarly to various types of tea
such as green tea, oolong tea, and black tea. These ndings will be
extremely useful for producing high-quality decaffeinated tea
beverages.
Funding
This research did not receive any specic grant from funding
agencies in the public, commercial, or not-for prot sectors.
Acknowledgements
We thank Dr. Okada, Professor of Shinshu University, and
Hideyuki Wakabayashi, Takehisa Nishida, Ryohei Tsuji, members of
our research groups, for valuable discussions. We also gratefully
acknowledge the technical assistance of Noriko Kumata.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jfoodeng.2016.12.015.
References
Aharonson, N., Kafka, U., 1975. Adsorption of benzimidazole fungicides on mont-
morillonite and kaolinite clay surfaces. J. Agric. Food Chem. 23, 434e437.
Arab, L., Khan, F., Lam, H., 2013. Epidemiologic evidence of a relationship between
tea, coffee, or caffeine consumption and cognitive decline. Am. Soc. Nutr. Adv.
Nutr. 4, 115e122.
Astill, C., Birch, M.R., Dacombe, C., Humphrey, P.G., Martin, P.T., 2001. Factors
affecting the caffeine and polyphenol contents of black and green tea infusions.
J. Agric. Food Chem. 49, 5340e5347.
Bergaya, F., Theng, B.K.G., Lagaly, G., 2006. Handbook of Clay Science, Ch. 2. Elsevier,
Amsterdam, pp. 42e45.
Catarino, S., Madeira, M., Monteiro, F., Rocha, F., Curvelo-Garcia, A.S., De Sousa, R.B.,
2008. Effect of bentonite characteristics on the elemental composition of wine.
J. Agric. Food Chem. 56, 158e165.
Chang, K., 2015. World Tea Production and Trade: Current and Future Development.
Food and Agriculture Organization of the United Nations, Rome.
Chugh, A., Khanal, D., Walkling-Ribeiro, M., Corredig, M., Duizer, L., Grifths, W.,
2014. Change in color and volatile composition of skim milk processed with
pulsed electric eld and microltration treatments or heat pasteurization.
Foods 3, 250e268.
Dang, V.B.H., Doan, H.D., Dang-Vu, T., Lohi, A., 2009. Equilibrium and kinetics of
biosorption of cadmium (II) and copper (II) ions by wheat straw. Bioresour.
Technol. 100, 211e219.
Fernandes, O., Sabharwal, M., Smiley, T., Pastuszak, A., Koren, G., Einarson, T., 1998.
Moderate to heavy caffeine consumption during pregnancy and relationship to
spontaneous abortion and abnormal fetal growth: a meta-analysis. Reprod.
Toxicol. 12, 435e444.
Gogoi, P., Dutta, N.N., Rao, P.G., 2010. Adsorption of catechin from aqueous solutions
on polymeric resins and activated carbon. Indian J. Chem. Technol. 17, 337e345.
Gonz alez-Pradas, E., Villafranca-Sa nchez, M., Gallego-Campo, A., 1993. Inuence of
the physical-chemistry properties of an acid-activated bentonite in the
bleaching of olive oil. J. Chem. Technol. Biotechnol. 57, 213e216.
Hamilton-Miller, J.M.T., 1995. Antimicrobial properties of tea (Camellia sinensis L.).
Antimicrob. Agents Chemother. 39, 2375e2377.
Hindmarch, I., Rigney, U., Stanley, N., Quinlan, P., Rycroft, J., Lane, J., 2000.
A naturalistic investigation of the effects of day-long consumption of tea, coffee
and water on alertness, sleep onset and sleep quality. Psychopharmacology 149,
203e216.
Ho, Y.S., McKay, G., 1999. Pseudo-second order model for sorption processes. Pro-
cess Biochem. 34, 451e465.
Huang, K.J., Wu, J.J., Chiu, Y.H., Lai, C.Y., Chang, C.M.J., 2007. Designed polar
cosolvent-modied supercritical CO2 removing caffeine from and retaining
catechins in green tea powder using response surface methodology. J. Agric.
Food Chem. 55, 9014e9020.
Hurel, C., Marmier, N., Bourg, A.C.M., Fromage, F., 2009. Sorption of Cs and Rb on
puried and crude MX-80 bentonite in various electrolytes. J. Radioanal. Nucl.
Chem. 279, 113e119.
Kanda, H., Li, P., Makino, H., 2013. Production of decaffeinated green tea leaves using
liqueed dimethyl ether. Food Bioprod. Process. 91, 376e380.
Khokhar, S., Magnusdottir, S.G.M., 2002. Total phenol, catechin, and caffeine con-
tents of teas commonly consumed in the United Kingdom. J. Agric. Food Chem.
50, 565e570.
Kim, W.J., Kim, J.D., Kim, J., Oh, S.G., Lee, Y.W., 2008. Selective caffeine removal from
green tea using supercritical carbon dioxide extraction. J. Food Eng. 89,
303e309.
Komatsu, Y., Suematsu, S., Hisanobu, Y., Saigo, H., Matsuda, R., Hara, K., 1993. Effects
of pH and temperature on reaction kinetics of catechins in green tea infusion.
Biosci. Biotechnol. Biochem. 57, 907e910.
Lailach, G.E., Thompson, T.D., Brindley, G.W., 1968. Absorption of pyrimidines, pu-
rines, and nucleosides by Co-, Ni-, Cu-, and Fe(III)-montmorillonite (clay-
organic studies XII). Clays Clay Miner. 16, 285e293.
Lee, S., Park, M.K., Kim, K.H., Kim, Y.S., 2007. Effect of supercritical carbon dioxide
decaffeination on volatile components of green teas. J. Food Sci. 72, 497e502.
Lee, S.M., Lee, H.S., Kim, K.H., Kim, K.O., 2009. Sensory characteristics and consumer
acceptability of decaffeinated green teas. J. Food Sci. 74, 135e141.
Lorist, M.M., Tops, M., 2003. Caffeine, fatigue, and cognition. Brain Cognit. 53,
82e94.
Lu, J.L., Wu, M.Y., Yang, X.L., Dong, Z.B., Ye, J.H., Borthakur, D., Sun, Q.L., Liang, Y.R.,
2010. Decaffeination of tea extracts by using poly(acrylamide-co-ethylene gly-
col dimethylacrylate) as adsorbent. J. Food Eng. 97, 555e562.
McIlvaine, T.C., 1921. A buffer solution for colorimetric comparison. J. Biol. Chem. 49,
183e186.
Maral, L., De Faria, E.H., Nassar, E.J., Trujillano, R., Martin, N., Vicente, M.A., Rives, V.,
Gil, A., Korili, S.A., Ciuf, K.J., 2015. Organically modied saponites: SAXS study
of swelling and application in caffeine removal. ACS Appl. Mater. Interfaces 7,
10853e10862.
Matsumoto, N., Ishigaki, F., Ishigaki, A., Iwashina, H., Hara, Y., 1993. Reduction of
blood glucose levels by tea catechin. Biosci. Biotechnol. Biochem. 57, 525e527.
Naik, J.P., Nagalakshmi, S., 1997. Determination of caffeine in tea products by an
improved high-performance liquid chromatography method. J. Agric. Food
Chem. 45, 3973e3975.
Nurminen, M.L., Niittynen, L., Korpela, R., Vapaatalo, H., 1999. Coffee, caffeine and
blood pressure: a critical review. Eur. J. Clin. Nutr. 53, 831e839.
Okada, T., Seki, Y., Ogawa, M., 2014. Designed nanostructures of clay for controlled
adsorption of organic compounds. J. Nanosci. Nanotechnol. 14, 2121e2134.
Okada, T., Oguchi, J., Yamamoto, K., Shiono, T., Fujita, M., Iiyama, T., 2015. Organo-
clays in water cause expansion that facilitates caffeine adsorption. Langmuir 31,
180e187.
Oshima, Y., 1936. Chemical studies on the tannin substance of formosan tea-leaves.
Bull. Agric. Chem. Soc. Jpn. 12, 103e114.
Oszmianski, J., Cheynier, V., Moutounet, M., 1996. Iron-catalyzed oxidation of ()-
catechin in model systems. J. Agric. Food Chem. 44, 1712e1715.
Park, H.S., Choi, H.K., Lee, S.J., Park, K.W., Choi, S.G., Kim, K.H., 2007. Effect of mass
transfer on the removal of caffeine from green tea by supercritical carbon di-
oxide. J. Supercrit. Fluids 42, 205e211.
Perva-Uzunali 
c, A., Skerget,  Weinreich, B., Otto, F., Grner, S., 2006.
M., Knez, Z.,
Extraction of active ingredients from green tea (Camellia sinensis): extraction
efciency of major catechins and caffeine. Food Chem. 96, 597e605.
Russell, J.D., Cruz, M.I., White, J.L., 1968. Adsorption of 3-aminotriazole by mont-
morillonite. J. Agric. Food Chem. 16, 21e24.
Sakanaka, S., 2003. A novel convenient process to obtain a raw decaffeinated tea
polyphenol fraction using a lignocellulose column. J. Agric. Food Chem. 51,
3140e3143.
 T. Shiono et al. / Journal of Food Engineering 200 (2017) 13e21
Sarma, G.K., Sengupta, S., Bhattacharyya, K.G., 2011. Methylene blue adsorption on
natural and modied clays. Sep. Sci. Technol. 46, 1602e1614.
Sevillano, D.M., Van Der Wielen, L.A.M., Hooshyar, N., Ottens, M., 2014. Resin se-
lection for the separation of caffeine from green tea catechins. Food Bioprod.
Process. 92, 192e198.
Suzuki, M., Yoshino, K., Maeda-Yamamoto, M., Miyase, T., Sano, M., 2000. Inhibitory
effects of tea catechins and O-methylated derivatives of (-)-epigallocatechin-3-
O-gallate on mouse type IV allergy. J. Agric. Food Chem. 48, 5649e5653.
Tajchakavit, S., Boye, J.I., Couture, R., 2001. Effect of processing on post-bottling haze
formation in apple juice. Food Res. Int. 34, 415e424.
Tang, W.Q., Li, D.C., Lv, Y.X., Jiang, J.G., 2010. Extraction and removal of caffeine from
green tea by ultrasonic-enhanced supercritical uid. J. Food Sci. 75, 363e368.
Toyoda-Ono, Y., Yoshimura, M., Nakai, M., Fukui, Y., Asami, S., Shibata, H., Kiso, Y.,
Ikeda, I., 2007. Suppression of postprandial hypertriglyceridemia in rats and
mice by oolong tea polymerized polyphenols. Biosci. Biotechnol. Biochem. 71,
971e976.
Tsujimura, M., 1934. Isolation of a new catechin, tea catechin II or gallo-catechin
from green tea. Bull. Agric. Chem. Soc. Jpn. 10, 140e147.
Ueda, M., Ashida, H., 2012. Green tea prevents obesity by increasing expression of
insulin-like growth factor binding protein-1 in adipose tissue of high-fat diet-
fed mice. J. Agric. Food Chem. 60, 8917e8923.
Uchiyama, S., Taniguchi, Y., Saka, A., Yoshida, A., Yajima, H., 2011. Prevention of diet-
induced obesity by dietary black tea polyphenols extract in vitro and in vivo.
21
Nutrition 27, 287e292.
Vicente Rodriguez, M.A., Suarez Barrios, M., Lopez Gonzalez, J.D., Banares
Munoz, M.A., 1994. Acid activation of a ferrous saponite (grifthite): physico-
chemical characterization and surface area of the products obtained. Clays
Clay Miner. 42, 724e730.
Von Borstel, R.W., 1983. Biological effects of caffeine. Metab. Food Technol. 37,
40e43.
Vuong, Q.V., Golding, J.B., Nguyen, M.H., Roach, P.D., 2013. Preparation of decaf-
feinated and high caffeine powders from green tea. Powder Technol. 233,
169e175.
Yanagimoto, K., Ochi, H., Lee, K.G., Shibamoto, T., 2003. Antioxidative activities of
volatile extracts from green tea, oolong tea, and black tea. J. Agric. Food Chem.
51, 7396e7401.
Ye, J.H., Liang, Y.R., Jin, J., Liang, H.L., Du, Y.Y., Lu, J.L., Ye, Q., Lin, C., 2007. Preparation
of partially decaffeinated instant green tea. J. Agric. Food Chem. 55, 3498e3502.
Ye, J.H., Jin, J., Liang, H.L., Lu, J.L., Du, Y.Y., Zheng, X.Q., Liang, Y.R., 2009. Using tea
stalk lignocellulose as an adsorbent for separating decaffeinated tea catechins.
Bioresour. Technol. 100, 622e628.
Yildirim, H.K., 2011. Effects of ning agents on antioxidant capacity of red wines.
J. Inst. Brew. 117, 55e60.
Yoshino, K., Yamazaki, K., Sano, M., 2010. Preventive effects of black tea theaavins
against mouse type IV allergy. J. Sci. Food Agric. 90, 1983e1987.