- Email: [email protected]
A simple and efficient method for enrichment of cocoa polyphenols from cocoa bean husks with macroporous resins following a scale-up separation
Accepted Manuscript A simple and efficient method for enrichment of cocoa polyphenols from cocoa bean husks with macroporous resins following a scale-up separation Jia-Lun Zhong, Nadeem Muhammad, Yu-Chao Gu, Wei-Dong Yan PII:
S0260-8774(18)30360-1
DOI:
10.1016/j.jfoodeng.2018.08.023
Reference:
JFOE 9373
To appear in:
Journal of Food Engineering
Received Date: 3 June 2017 Revised Date:
19 August 2018
Accepted Date: 22 August 2018
Please cite this article as: Zhong, J.-L., Muhammad, N., Gu, Y.-C., Yan, W.-D., A simple and efficient method for enrichment of cocoa polyphenols from cocoa bean husks with macroporous resins following a scale-up separation, Journal of Food Engineering (2018), doi: 10.1016/j.jfoodeng.2018.08.023. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
A simple and efficient method for enrichment of cocoa polyphenols from
2
cocoa bean husks with macroporous resins following a scale-up separation
3
Jia-Lun Zhong, Nadeem Muhammad, Yu-Chao Gu, Wei-Dong Yan1
4
Department of Chemistry, Zhejiang University, Hangzhou 310027, China
RI PT
1
5
Abstract
SC
6
The aim of this study is to develop an efficient method for the decaffeination and enrichment of the
8
polyphenols from cocoa husks extracts. LX-17 was selected for the exploration of optimal processing
9
parameters from eight kinds of macroporous resins. The adsorption kinetics and thermodynamics of
10
(-)-epicatechin (EC) were studied prior to the scale-up separation. The optimum parameters for separation
11
were as follow: 6.0 mg/mL cocoa extracts, pH 2.0, 25 °C column temperature, flow rates of adsorption
12
and desorption 1.6 BV/h (Bed Volume, the volume of the resin) and ethanol-water (20:80, 50:50, 95:5,
13
v/v) solutions in the gradient elution. In the scale-up separation, 2 kg of cocoa husks were extracted in 20
14
L 50 % ethanol solutions, separated on 3 L LX-17 resins and yielded 34.99 g cocoa polyphenols. This
15
method significantly increased the total polyphenol contents from 2.23 % to 62.87 % with a recovery
16
yield of 78.57 %.
17
Key words
18
Cocoa husk, Polyphenol, Macroporous resin, Adsorption, Separation, Scale-up
1
AC C
EP
TE D
M AN U
7
Corresponding author at: Department of Chemistry, Zhejiang University, Hangzhou 310027, China
Tel: 0086 571 87951430. Fax: 0086 571 8795189. E-mail: [email protected] ( W. D. Yan) 1
ACCEPTED MANUSCRIPT
19
Chemical compounds studied in this article:
20
(-)-Epicatechin (PubChem CID: 72276); Caffeine (PubChem CID: 2519); Theobromine (PubChem CID:
21
5429); Procyanidin B2 (PubChem CID: 122738); Procyanidin C1 (PubChem CID: 169853);
22
(PubChem CID: 702); Sodium Hydroxide (PubChem CID: 14798); Formic Acid (PubChem CID: 284);
23
N-Hexane (PubChem CID: 8058); Hydrochloric Acid (PubChem CID: 313); Acetonitrile (PubChem CID:
24
6342)
25
1. Introduction
M AN U
SC
RI PT
Ethanol
Cocoa beans (Theobroma cacao L.) have been cultivated for a long period of time as a major
27
ingredient of cocoa and chocolate (Baba et al., 2000; Andres-Lacueva et al., 2008). Cocoa beans contain
28
approximately 6-8 % polyphenols by dry weight (Grassi et al., 2008). Polyphenols are secondary plant
29
metabolites with high antioxidant properties and also have potentially beneficial effect on human health,
30
such as treatment and prevention of cancer, cardiovascular disease and other pathologies (Ignat et al.,
31
2011; Quideau et al., 2011). EU Commission approves health claim: Cocoa flavanols support a healthy
32
blood circulation (2013).
AC C
EP
TE D
26
33
The data obtained from the Food and Agriculture Organization of the United Nations Statistics
34
Division have shown that 4.58 million ton of cocoa beans were produced in 2013. Whereas, cocoa husks
35
about 5-10 % weight of cocoa beans were generated as a waste by-product of the cocoa industries. Cocoa
36
husks can be a rich source of polyphenols and its low cost as compared to cocoa beans make a better
37
choice. Three main polyphenol groups can be distinguished in cocoa: catechins, anthocyanins and
2
ACCEPTED MANUSCRIPT
procyanidins, whose common skeletal structure is a flavanoid-type structure (Wollgast and Anklam,
39
2000). The molecular structure of EC is shown in Figure 1. The crude polyphenol extracts of cocoa husks
40
always contain alkaloids, proteins, polysaccharides and other impurities. Therefore, an efficient
41
purification method is needed to enrich cocoa polyphenols.
RI PT
38
The conventional method of decaffeination and enrichment of cocoa polyphenols were carried out by
43
means of solid-liquid extraction (Chemat et al., 2011), precipitation or silica gel column chromatography.
44
However, these separation methods are inefficient, time consuming, laborious and using high contents of
45
toxic organic solvents.
M AN U
SC
42
Macroporous adsorption resins have been widely used for the separation and enrichment of
47
biochemical products. Researches like madecassoside and asiaticoside from Centella asiatica (Jia and Lu,
48
2008), rutin and quercetin from Euonymus alatus (Thunb.) Siebold (Zhao et al., 2011), flavonoids from
49
Glycyrrhiza glabra L. leaf (Dong et al., 2015) and rosavin from Rhodiola rosea (Ma et al., 2009) have
50
been reported due to the large surface area and distinct pore structure on macroporous resin which
51
increases the efficiency of purification process. In addition, the functional groups on the surface of the
52
macroporous resin also enhance the separation selectivity. Therefore, compounds can be eluted based on
53
polarity, molecular size and hydrogen-bond interactive forces. The mildness and low toxicity of these
54
adsorption and desorption interactions make this process food-friendly.
AC C
EP
TE D
46
55
Therefore, in this study the adsorption and desorption characteristics of EC and the optimal
56
separation condition for polyphenols on macroporous resins were studied. Meanwhile, the adsorption
57
kinetics and thermodynamics were also investigated. The decaffeination was achieved by gradient elution 3
ACCEPTED MANUSCRIPT
with different ratios of ethanol-water solutions. In the end, the experiments were scaled up to evaluate the
59
feasibility of the industrial process.
60
2. Experimental
61
2.1 Chemicals and materials
RI PT
58
Distilled water was purchased from Hangzhou Wahaha Group Co., Ltd while HPLC grade methanol
63
and acetonitrile were purchased from MERYER (Shanghai, China). Theobromine, caffeine and
64
(-)-epicatechin standards of analytical grade were purchased from Sigma-Aldrich.
65
phenol reagent, formic acid, n-hexane and ethanol were purchased from Sinopharm Chemical Reagent
66
Co., Ltd (Shanghai, China). The standard solutions were prepared by weighing the appropriate amounts
67
of standards and dissolving in 20 % ethanol-water at the concentration of 1.06 mg/mL. Cocoa beans
68
(Forastero) were purchased from Wuxi Aokang food Corp., Ltd (Wuxi, China). Cocoa husks were
69
purchased from Zhejiang Qili Xingguang Cocoa Products Corp., Ltd. The cocoa beans/cocoa husks were
70
milled into small pieces using liquid nitrogen.
Folin-Ciocalteu's
EP
TE D
M AN U
SC
62
Macroporous resins LX-68, LX-32, LX-17, XDA-1, XDA-8, AB-8, and DM130 were purchased
72
from Lanxiao technology Co., Ltd (Xian, China), D101 were purchased from Lansheng Co., Ltd (Xian,
73
China). The moisture contents of macroporous resins were determined. Macroporous resins were
74
pretreated by soaking in 95 % ethanol (1:5, v/v) and eluting with 5 % NaOH (1:5, v/v), 5% HCl (1:5, v/v)
75
and distilled water (1:5, v/v) before using.
76
2.2 Apparatus
AC C
71
4
ACCEPTED MANUSCRIPT
The scale-up extraction was carried out in 20L-Jacketed Pilot Plant Reactors connected with
78
20L-circulation constant temperature water bath (GYY-20L, Zhengzhou Zhuocheng instrument Co., Ltd.,
79
China). The centrifugation was performed on low speed centrifuge (SC-3610, Anhui USTC Zonkia
80
Scientific Instruments Co., Ltd., China). The pH value of the cocoa husk extracts was determined by
81
Waterproof pHScan2 Tester, Eutech Instruments Pte Ltd. Water bath was carried out in a thermostatic
82
water bath (THD-2006, Ningbo Tianheng Instrument Works Co., Ltd., China). The cocoa husk extracts
83
and the fractions eluted by macroporous resins were concentrated by a RE2000 rotary evaporator
84
(Shanghai Yarong Biochemical Co., Ltd., China). In the scale-up experiments, the fractions were
85
concentrated by a RE-5220 rotary evaporator (Shanghai Yarong Biochemical Co., Ltd., China). The
86
aqueous solutions were dried by a freeze drier (FD-1A-50, Beijing Boyikang experimental instrument Co.,
87
Ltd., China). The effluent and eluted solutions were collected by auto parts collector (BS-100N, Shanghai
88
Luxi analytical instrument Co., Ltd., China). HPLC analyses were carried out on Wufeng LC-100 HPLC
89
(Shanghai, China) coupled with the Wufeng LC-100 UV detector (Suzhou, China). The column used was
90
Phenomenex Luna C18 100A column (250 mm × 4.6 mm, 5 µm). The ultraviolet absorption of total
91
polyphenols was measured by using Photometer (UV-1800PC SPECTROPHOTOMETER, MAPADA,
92
Shanghai, China).
93
2.3 Extraction of cocoa beans/husks
AC C
EP
TE D
M AN U
SC
RI PT
77
94
The extraction procedure of cocoa beans/cocoa husks was carried out by following Langer (Langer
95
et al., 2011) scheme with slight modifications. The milled cocoa beans/cocoa husks were defatted and
96
extracted together with n-heptane (1: 5, w/v) and 50 % ethanol in a solvent ratio of 1: 10 (w/v) following
5
ACCEPTED MANUSCRIPT
it was vortexed (1 min), sonicated (50 °C, 30 min) and filtered. Solvents were added to the residue after
98
filtration and the process was repeated thrice. After that, the n-heptane layer was discarded and ethanol in
99
the lower extracted phase was removed by rotary evaporation under partial vacuum at 80 °C. The
RI PT
97
resulting yield was dry froze and stored in the fridge at -20 °C for further use.
101
2.4 HPLC analysis
SC
100
A linear gradient elution was used for the HPLC analyses from A (acetonitrile containing 20 % water)
102
to B (water containing 1 % formic acid). Gradient elution was followed as: from 11 % A to 20 % A in 20
104
min, keep 20 % A for 10 min, from 20 % A to 80 % A in 20 min, keep 80 % A for 5 min, and return to
105
11 % A in 3 min at a flow rate of 1.0 mL/min, 20 µL of the sample was injected in duplicates into the
106
column kept at 30 °C. The wavelength of the detector was 280 nm. Total polyphenols contents (TPC)
107
were determined with the Folin–Ciocalteu procedure described by Silke Elwers (Elwers et al., 2009).
108
2.5 Static adsorption and desorption tests
109
2.5.1
EP
TE D
M AN U
103
AC C
Selection of macroporous resins
110
2.0 g of accurately weighed pretreated D101, LX-68, LX-32, LX-17, XDA-1, XDA-8, AB-8, and
111
DM130 macroporous resins were put into a 250 mL conical flask and 50 mL of 6.0 mg/ml aqueous cocoa
112
beans extracts solution (pH 2.0) was added. TPC (mg EC/ml) of the aqueous cocoa beans extracts
113
solution was analyzed before adsorption. The sealed flasks were shaken at 25 °C for 60 min. After
114
adsorption, a 1 mL sample solution was withdrawn with an injection syringe a having a long needle for
115
analysis of TPC. Then the resins were washed with 50 mL distilled water twice, filtered and the residues 6
ACCEPTED MANUSCRIPT
were desorbed with 50 mL ethanol-water (50:50, v/v) solution, stirred for 60 min at 25 °C. After
117
desorption, TPC of the desorbed solution was analyzed. A circulating water bath with a thermostat was
118
applied to maintain the temperature within ±0.01 K. Adsorption capacities were calculated using equation
119
(1).
120
Qt = V0 (C0 − Ct ) / M
121
Desorption ratios were calculated using equation (2).
122
D = 100CdVd /[V0 (C0 − Ce )]
123
2.5.2
(1)
M AN U
SC
RI PT
116
(2)
Effects of initial concentration, temperature, pH and adsorption time on the adsorption amounts
The freeze-dried cocoa beans extracts were diluted with distilled water to different concentrations
125
(0.409, 2.04, 4.09, 6.13, 8.17 mg/mL). The solutions were adsorbed and desorbed as described in section
126
2.5.1 on the selected resin and the concentrations of EC were analyzed with HPLC. The effects of
127
temperature (methods described in section 2.6.3), pH (2.0, 3.0, 4.0, 5.0, 7.0 and 9.0) and adsorption time
128
(0, 5, 10, 15, 30 and 60 min) on adsorption capacity were also studied respectively.
129
2.5.3
130
AC C
EP
TE D
124
Adsorption kinetics and thermodynamics
The results of adsorption time effect were linearly regressed using the pseudo-first and second order
131
models.
132
Pseudo-first order model
7
ACCEPTED MANUSCRIPT
133
ln(Qe − Qt ) = ln Qe − K f t
134
Pseudo-second order model
135
t 1 1 = + t 2 Qt K s Qe Qe
RI PT
(3)
(4)
The selected resin (8.0 g) was added to series of 25 mL standard solutions having cumulative
137
concentration of EC (0.018, 0.026, 0.039, 0.054, 0.087 mg/mL). Adsorption was conducted at four
138
different temperatures (28.5, 39.8, 54.1, 68.0 °C) in a 100 mL boiling flask-3-necked with a stirring
139
paddle and a mercury thermometer inside. The adsorption isotherms for EC were determined using
140
Langmuir and Freundlich equations.
141
Langmuir equation (5):
142
Qe =
143
Freundlich equation (6):
144
Qe = K F Ce
145
The enthalpy changes (∆H) were calculated by the Van’t Hoff equation:
146
ln K = −
147
2.6 Dynamic adsorption and desorption
1
TE D
(5)
EP
AC C
QmCe K L + Ce
M AN U
SC
136
n
(6)
∆H +A RT
(7)
8
ACCEPTED MANUSCRIPT
148
2.6.1
Effects of flow rates and ethanol-water ratios on dynamic adsorption and desorption
Dynamic adsorption and desorption experiments were performed on glass column (length 50 cm × Φ
150
4 cm) wet-packed with the selected resin (100 mL, 25 °C). The aqueous cocoa beans extracts solutions
151
(6.0 mg/mL, pH 2.0) were adsorbed and then eluted by ethanol-water (10:90, 20:80, 40:60, 50:50, 60:40,
152
70:30, 95:5, v/v, 1.6 BV/h) eluent and the volume of each solution was 2.5 BV. An auto parts collector
153
was used to collect eluted solutions, analyzed by HPLC and TPC was determined. Effects of flow rates
154
(1.6, 2.4, 3.2 BV/h) on dynamic adsorption and desorption were also studied to obtain dynamic
155
breakthrough curves and desorption curves.
156
2.6.2
Enrichment under optimal conditions
M AN U
SC
RI PT
149
The 100.0 g cocoa husks were extracted as described in Section 2.3. The experiments were
158
performed on glass column (length 50 cm × Φ 4 cm) wet-packed with the selected resin (300 mL, 25 °C).
159
The aqueous cocoa husks extracts solutions (6.0 mg/mL, pH 2.0) were adsorbed at the flow rate of 1.6
160
BV/h. After the adsorption, the column was eluted by ethanol-water (0:100, 20:80, 50:50, 95:5, v/v. 1.6
161
BV/h) solutions and the volume of each solution was 5 BV. Each part of the eluted solutions was
162
concentrated in a rotary evaporator at 70 °C, analyzed by HPLC and TPC was determined.
163
2.7 The scale-up experiments
AC C
EP
TE D
157
164
The extraction of 2 kg cocoa husks was carried out in a 20 L-Jacketed Pilot Plant Reactor connected
165
with 20 L-circulation constant temperature water bath. The milled cocoa husks were extracted with 50 %
166
ethanol in a solvent ratio of 1:10 (w/v). Afterward 20 L solvents were added, the agitating vane in the 9
ACCEPTED MANUSCRIPT
167
reactor stirred up at the rate of 550 rpm. And the circulation constant temperature water bath was kept at
168
85 °C (80 °C in the reactor).
To validate the stability of EC and polyphenols, three sealed conical flask filled with 50 % ethanol
170
cocoa husk extracts (50 mL) were heated to 80 °C and maintained by the circulating water bath with a
171
thermostat. Meanwhile, at the time interval of 0 h, 1 h and 4 h the solution was withdrawn and injected to
172
the HPLC system to determine the concentration of EC and TPC.
SC
RI PT
169
The extraction was repeated for three times and the extracted solutions at the end were combined.
174
The extracts were evaporated in a rotary evaporator having a 20 L-round-bottom flask to remove the
175
organic solvents at 70 °C. The glass column (length 100 cm × Φ10 cm) was filled with 3L the selected
176
resin. The aqueous cocoa husks extracts solutions (6 mg/mL, pH 2.0) were adsorbed at the flow rate of
177
1.6 BV/h, 25 °C. The elution and analysis processes are described in section 2.6.2.
178
2.8 Evaluation of the scale-up process
EP
TE D
M AN U
173
The scale-up procedure described in section 2.7 was repeated for 10 cycles on the selected LX-17
180
resin (1 L, 25 °C) to evaluate the regeneration capacity of the exhausted resin. After each adsorption and
181
desorption process, the resin was pretreated as described in section 2.1.
182 183
AC C
179
Total polyphenols produced per volume of the selected resin per day was calculated using the equation:
10
ACCEPTED MANUSCRIPT
M tp
184
Pt =
185
The volume of waste solvent produced in the production of one mass unit of polyphenols was using the
186
equation:
187
Psc =
× 24
RI PT
(8)
(9)
SC
Vs M tp
The description of the equations (1-9) can be seen in Supporting Information and the reference
189
(Dong et al., 2015; Ma et al., 2009).
190
3. Results and discussion
M AN U
188
Vr ⋅ tr
According to Langer’s research (Langer et al., 2011), (-)-epicatechin (linear regression with total
192
polyphenols, R2 = 0.96) was a reliable marker of total polyphenols. The molecular structure of
193
(-)-epicatechin is similar to other polyphenols like catechin, epigallocatechin, procyanidin oligomers and
194
procyanidin polymers. Furthermore, the adsorption and desorption characteristics of these polyphenols on
195
the macroporous resins would also be similar. Therefore, the adsorption capacity and desorption ratios of
196
EC were determined in some parts of the experiments instead of TPC.
AC C
EP
TE D
191
197
The typical chromatogram of cocoa beans extracts is shown in Figure S1. The peaks of theobromine,
198
caffeine, procyanidin B2, (-)-epicatechin and procyanidin C1 were identified by the standard solutions
199
based on their retention time on HPLC (Esatbeyoglu et al., 2011). TPC was determined at equivalents of
200
EC (mg EC/mL). A calibration curve was made with the EC standards. The linear regression equation y =
11
ACCEPTED MANUSCRIPT
0.1056x – 0.0019 was obtained with R2 = 0.9975. The linear regression of the calibration curve of EC is
202
shown in Figure S2.
203
3.1 Extraction and analysis of cocoa beans and cocoa husks
RI PT
201
To compare the constituents and contents, the concentrations of theobromine, caffeine and EC and
205
TPCs in cocoa husks and cocoa beans were determined as listed in Table S1. It shows that cocoa beans
206
have higher TPCs than cocoa husks. In addition, the concentrations of theobromine, caffeine and EC in
207
cocoa beans were higher while there were no significant differences in constituents between cocoa beans
208
and cocoa husks. It should be noted that there was some cocoa bean powder left in the cocoa husks. The
209
TPC content can be lower in coca husks without the cocoa bean powder. However, TPC content in cocoa
210
husks are acceptable regard to the prices of cocoa beans and cocoa husks. Therefore, in this study, cocoa
211
beans were used in the static and dynamic experiments and cocoa husks were used in the final scale-up
212
separations.
213
3.2 Static adsorption and desorption
214
3.2.1 Selection of macroporous resins
AC C
EP
TE D
M AN U
SC
204
215
The pretreated resins were soaked in water and filtered to remove the free water before the
216
determination of moisture. The chemical composition, physicochemical parameters, adsorption capacities
217
and desorption ratio of polyphenols on these resins are shown in Table 1. The material of the tested resins
218
is all polystyrene series. LX-32, XDA-1, XDA-8 and LX-17 are all polar resins and have shown high
219
adsorption capacities while moderate-polar resins AB-8 and DM130 displayed low adsorption capacities. 12
ACCEPTED MANUSCRIPT
220
The maximum handling capacity of cocoa husks in one run depends on the maximum adsorption capacity
221
of macroporous resins.
Also, the physical adsorption should be reversible in the desorption process. The specific surface
223
areas of XDA-1 and XDA-8 were higher than LX-17 and there were more irreversible adsorptions on the
224
resins. In this way, the desorption ratios of XDA-1 and XDA-8 were lower than LX-17 resin. Thus,
225
LX-17 was selected as the effective macroporous resin in the following experiments.
226
3.2.2 Effects of initial concentration and pH on the adsorption capacity
M AN U
SC
RI PT
222
The effects of initial concentration and pH on the adsorption capacity are shown in Figure 2A and 2B.
228
Figure 2A showed that adsorption capacity rose as the concentration increased and the highest adsorption
229
capacity was observed when the initial concentration of EC reached 0.025 mg/mL (The aqueous cocoa
230
beans extracts solution, 6.0 mg/mL). On further increase of EC concentration of 0.035 mg/mL the
231
adsorption capacity decreased. It can be attributed to decrease in the number of free active sites and more
232
chance of impurities adsorption on LX-17 resin, resulting in competition for active sites between the
233
polyphenols and the impurities (Sun et al., 2013), which led to a slight drop in adsorption capacity. Thus,
234
the initial concentration of the aqueous cocoa beans extracts should be 6.0 mg/mL or less than it for the
235
adequate adsorption of EC.
AC C
EP
TE D
227
236
As Figure 2B shows, the resin showed the highest adsorption capacity of EC when the pH value of
237
the sample solution was 2.0. Figure 2B also shows that from pH 3.0 to pH 5.0 the adsorption capacity
238
changed little. The adsorption capacity of EC declined when the pH value was higher than 5.0. This may
13
ACCEPTED MANUSCRIPT
be due to existence of phenolic compounds as in molecule forms at low pH value. As the pH value rose,
240
polyphenols ionized and the physical interactions between polyphenols and the adsorptive sites of the
241
adsorbent declined. Similar results were found according to Xi et al. (Xi, Mu, & Sun, 2015). Thus, pH 2.0
242
was chosen as the optimum pH value of the sample solution.
243
3.2.3 Effects of adsorption time on the adsorption capacity and adsorption kinetics
SC
RI PT
239
The effect of adsorption time is shown in Figure 2C. The adsorption capacity of EC increased with
245
the adsorption time and reached to equilibrium at about 30 min. In the first 5 min, the adsorption
246
capacities increased rapidly and from 5 to 15 min, the adsorption capacity increased slowly. Later from 15
247
to 60 min the adsorption capacity changed little indicating attaining of the adsorption equilibrium.
248
Therefore, the adsorption process can be clarified into three stages. The results of adsorption time effect
249
were linearly regressed using the pseudo first order ( ln(0.725 − Qt ) = ln 0.725 − 0.1175t R2 = 0.9904) and
250
second order model (
251
pseudo-first-order model could be used to describe the initial (0 - 5 min), second (5 - 15 min) and third
252
(15 - 45 min) adsorption process with different slopes while the pseudo second order model can be used
253
to describe the whole process. Similar observations were found in previous studies (Buran et al., 2014;
254
Dong et al., 2015) and 60 min was selected as the optimal adsorption time for static tests.
255
3.2.4 Effects of temperature on the adsorption capacity and adsorption thermodynamics
TE D
M AN U
244
R2 = 0.9998) as shown in Figure S3.1-3.2. The
AC C
EP
t t 1 = + Qt 1.13 0.38 × 1.132
256
The effect of temperature on static adsorption is shown in Figure 2D, which indicates that the
257
adsorption capacity decreased as the temperature rose from 28.5 °C to 68.0 °C and it is also confirmed by 14
ACCEPTED MANUSCRIPT
Freundlich equation. The data were linearly regressed by the Langmuir equation and Freundlich equation,
259
respectively. As it is shown in Table 2 and Figure S4, the Freundlich equation appeared fitting better. The
260
constant of Freundlich equation KF decreased as the temperature increased which indicated that the
261
adsorption capacity was higher at the room temperature. Thus, 25.0 °C was selected as the column
262
temperature. According to the Van’t Hoff equation (linear regressed in Figure S5), the enthalpy change
263
(∆H) were calculated as -21.12 kJ/mol which indicated the adsorption process was exothermic and the
264
low temperature was appropriate for the adsorption process. Macroporous resins adsorb the targeted
265
constituents selectively through electrostatic force, hydrogen bonding interactions, complexation and size
266
sieving (Gao et al., 2007). The interactions declined as temperature rose due to molecular thermodynamic
267
movement, the swelling of resin and the increasing kinetic energies of the polyphenol molecules at higher
268
temperatures (Sun et al., 2013; Wu et al., 2010).
269
3.3 Dynamic adsorption and desorption
270
3.3.1
TE D
M AN U
SC
RI PT
258
EP
Effect of ethanol concentration on the desorption ratio and the elution order
Different content of ethanol-water solutions (10:90, 20:80, 40:60, 50:50, 60:40, 70:30 and 95:5, v/v)
272
were used to perform desorption tests in order to study the effect of ethanol content on the ratio of
273
desorption and the elution order of adsorbates as results are shown in Table S2 and Figure S6. The main
274
peaks of cocoa bean extracts are theobromine, caffeine, procyanidin B2, EC, procyanidin C1 and
275
procyanidin polymers as shown in Figure S6A. Most of theobromine could be eluted by water as shown
276
in Figure S6B whereas caffeine could be eluted by 20 % ethanol Figure S6D. Thus, removal of
277
theobromine and decaffeination were achieved by eluting with 20 % ethanol content and most of the
AC C
271
15
ACCEPTED MANUSCRIPT
polyphenols were still adsorbed on the macroporous resin. It should be noted that about 20 % of the
279
polyphenols were eluted in the water fraction while the resin has not reached the saturated adsorption. It
280
was probably because sugars and proteins were eluted in the water fractions. On the one hand, the result
281
determined by the Folin–Ciocalteu method would be interfered by sugars in the water fraction and a part
282
of the polyphenols would combine proteins by chemical bonds. From 40 % to 60 % ethanol fractions, EC
283
and polyphenols were enriched in these fractions as shown in Figure S6E. TPCs were all over 45 % and
284
the rest of the adsorbates were eluted in the 70 % and 95 % fraction. In this way, ethanol-water (20:80,
285
50:50, 95:5, v/v) solutions were used in the gradient elution.
286
3.3.2
M AN U
SC
RI PT
278
Effects of flow rate on dynamic adsorption and desorption
By choosing the initial concentration of EC 25 mg/L, the flow rate was investigated at 1.6 BV/h, 2.4
288
BV/h and 3.2 BV/h. It can be seen from Figure 3A that better adsorption performance of EC was
289
exhibited at the flow rate of 1.6 BV/h. The breakthrough point was defined as 1% ratio of the exit to the
290
inlet solute concentration (Jung et al., 2001). The breakthrough point of LX-17 resin at the flow rate of
291
1.6 BV/h was approximately 4500 mL. While at the flow rate of 2.4 BV/h and 3.2 BV/h, the breakthrough
292
volume was approximately 3000 mL and 2500 mL respectively. At the volume of 5600 mL the
293
concentration of EC eluted was coincident. The concentration of EC eluted rose steadily after 5600 mL at
294
the flow rate of 1.6 BV/h. In contrast, at the flow rate of 3.2 BV/h it showed as a flat curve. And the cocoa
295
extract solutions broke through at the volume of approximately 9000 mL. As section 3.2.3 indicated, EC
296
had less time to interact with the surface of the resin at higher flow rates and the adsorption process was
297
inadequate resulting in EC broke through the macroporous resin. Thus, 1.6 BV/h was selected as the best
AC C
EP
TE D
287
16
ACCEPTED MANUSCRIPT
298
dynamic adsorption flow rate. At the flow rate of 1.6 BV/h, the adsorption capacity of EC on LX-17 resin
299
was 1.27 mg/g which was coincident with the results obtained from section 3.2.2.
Better desorption performance of EC was exhibited at the flow rate of 1.6 BV/h according to Figure
301
3B. Polyphenols were totally desorbed by ethanol-water solutions (50:50, v/v) in 2 BV at the flow rate of
302
1.6 BV/h. As the flow rate increased, polyphenols were totally desorbed in 3 BV, 3.6 BV at the flow rate
303
of 2.4 BV/h, 3.2 BV/h respectively. At the lower flow rate, adsorbents were eluted in higher
304
concentrations with lower cost of ethanol-water solutions. 1.6 BV/h was selected as the proper flow rate
305
of desorption.
306
3.3.3
M AN U
SC
RI PT
300
Gradient elution tests
100.0 g cocoa husks were extracted and separated under optimal conditions on LX-17 resin in small
308
scale. The results are listed in Table S3. Although it was far from saturated adsorption, the similar results
309
were given as section 3.3.1 that about 20 % polyphenols were eluted in the water fraction. The
310
discrepancy was that 50 % of theobromine was eluted in the 20 % ethanol fraction while in section 3.3.1
311
the value was only 10 %. It is probably because the saturated adsorption of theobromine has reached and
312
most of the theobromine (90 %) broke through in the water fraction. The bed volume of LX-17 in section
313
3.3.1 was 100 mL. The test on small scale was generally satisfactory. Theobromine and caffeine were
314
eluted in 20 % ethanol fraction. In the 50 % ethanol fraction, the mass of dried residue was 1.79 g and the
315
mass of total polyphenols was 1.04 g EC with TPC 58.10 %.
316
3.4 The scale-up experiments
AC C
EP
TE D
307
17
ACCEPTED MANUSCRIPT
Before the experiments, the stability of EC and polyphenols was validated at 80 °C and the
318
concentration of EC and TPC changed slightly as shown in Table S4. In the study of Taeye (Taeye et al.,
319
2014), EC and procyanidin B2 degraded obviously at 60 and 90 °C, where these compounds were studied
320
individually. While in this study, both EC and polyphenols were extracted from cocoa husks and they
321
existed as mixtures in an equilibrium state. The degradation and epimerization had finished in natural.
322
Thus, EC and polyphenols at the extraction temperature of 80 °C were proved to be stable and the
323
extraction was sufficient at a higher temperature.
M AN U
SC
RI PT
317
The results of scale-up separation are listed in Table 3. Two kilograms of cocoa husks containing
325
44.54 g polyphenols were enriched through 3 L LX-17 resins. The results were similar to that of small
326
scale experiments. Theobromine and caffeine were eluted mainly in water fraction and 20 % ethanol
327
fraction, respectively. The mass of 50 % ethanol fraction was 55.66 g with 34.99 g polyphenols. The
328
value of TPC was 62.86 % and the recovery yield was 78.56 %. The content of caffeine by dry weight in
329
the yields was only 0.16 %. The difference with small scale tests was that much more polyphenols were
330
extracted in the scale-up separation, it might be due to high temperature that made the extraction
331
complete.
332
3.5 Evaluation the feasibility of scale-up process
AC C
EP
TE D
324
333
The regeneration capacity of the exhausted resin is an important factor in the process economy in the
334
enrichment of polyphenols. Consecutive polyphenols adsorption and desorption were repeated for 10
335
cycles with the same amount of resins and cocoa husks. At the breakthrough point, the adsorption
336
capacity was calculated every 2 cycles and then desorbed to obtain the desorption ratio. The results were 18
ACCEPTED MANUSCRIPT
shown in Table S5. The regeneration capacity of the exhausted resin performed consistently in 10 cycles
338
or more which were encouraging for production. Moreover, the yielding capacity, waste solvent volume
339
and estimated cost of cocoa polyphenols production are listed in Table 4. With Pt and the volume of resin,
340
polyphenols produced in one month could be calculated. While some of the ethanol in wasted solvents
341
can be regenerated. According to the TPC obtained in section 3.1 and recovery yield in section 3.4, the
342
exact costs of total polyphenols in cocoa husks and cocoa beans were 189.9 ¥/kg EC and 822.3 ¥/kg EC,
343
respectively. And the price of 40 % cocoa polyphenols was 200 ¥/kg, approximately. Despite of the cost
344
of ethanol and resins, cocoa husks could be used to produce cocoa polyphenols with profits.
345
4 Conclusion
M AN U
SC
RI PT
337
The decaffeination and enrichment of cocoa polyphenols from low cost cocoa husks were achieved
347
by gradient elution on the LX-17 macroporous resin. The optimal conditions for adsorption and
348
desorption were achieved by static and dynamic adsorption experiments as: cocoa extracts 6.0 mg/mL,
349
pH 2.0, column temperature 25 °C, flow rates of adsorption and desorption 1.6 BV/h and ethanol-water
350
(20:80, 50:50, 95:5, v/v) solutions were used in the gradient elution. The results of scaling-up separation
351
are satisfactory which suggested that the resin purification technology was efficient and food-friendly that
352
can be scaled up in the industry to produce polyphenol rich products with low-cost cocoa husks.
353
Acknowledgements
354
This project was financially supported by Sky Herb Co., Ltd (Huzhou, China).
355
References
AC C
EP
TE D
346
19
ACCEPTED MANUSCRIPT
Andres-Lacueva, C., Monagas, M., Khan, N., Izquierdo-Pulido, M., Urpi-Sarda, M., Permanyer, J., &
357
Lamuela-Raventos, R. M. (2008). Flavanol and flavonol contents of cocoa powder products: Influence of
358
the manufacturing process. Journal of Agricultural and Food Chemistry, 56, 3111–3117.
359
Baba, S., Osakabe, N., Natsume, M., Yasuda, A., Takizawa, T., Nakamura, T., & Terao, J. (2000). Cocoa
360
powder enhances the level of anti-oxidative activity in rat plasma. British Journal of Nutrition, 84,
361
673–680.
362
Buran, T. J., Sandhu, A. K., Li, Z., Rock, C. R., W, W. H., & Gu, L. W. (2014). Adsorption/desorption
363
characteristics and separation of anthocyanins and polyphenols from blueberries using macroporous
364
adsorbent resins. Journal of Food Engineering, 128, 167–173.
365
Chemat, F., & Zill-e-Huma and Khan, M. K. (2011). Applications of ultrasound in food technology:
366
Processing, preservation and extraction. Ultrasonics Sonochemistry,18, 813-835.
367
Dong, Y., Zhao, M., Sun-Waterhouse, D., Zhuang, M., Chen, H., Feng, M., & Lin, L. (2015). Absorption
368
and desorption behaviour of the flavonoids from Glycyrrhiza glabra L. leaf on macroporous adsorption
369
resins. Food Chemistry, 168, 538–545.
370
Duran, C., Ozdes, D., Gundogdu, A., & Senturk, H. B. (2011). Kinetics and isotherm analysis of basic
371
dyes adsorption onto almond shell (Prunus dulcis) as a low cost adsorbent. Journal of Chemical and
372
Engineering Data, 56, 2136–2147.
373
Elwers, S., Zambrano, A., Rohsius, C., & Lieberei, R. (2009). Differences between the content of
374
phenolic compounds in Criollo, Forastero and Trinitario cocoa seed (Theobroma cacao L.). European
375
Food Research and Technology, 229, 937–948.
376
Esatbeyoglu, T., Wray, V., & Winterhalter, P. (2011). Isolation of dimeric, trimeric, tetrameric and
AC C
EP
TE D
M AN U
SC
RI PT
356
20
ACCEPTED MANUSCRIPT
377
pentameric procyanidins from unroasted cocoa beans (Theobroma cacao L.) using countercurrent
378
chromatography. Food Chemistry, 179, 278–289.
379
Gao, M., Huang, W.
380
breviscapus(vant.) Hand. Mazz. By macroporous resins. Journal of Chromatography B, 858, 22–26.
381
Grassi, D., Desideri, G., Necozione, S., Lippi, C., Casale, R., Properzi, G., & Ferri, C. (2008). Blood
382
pressure is reduced and insulin sensitivity increased in glucose-intolerant, hypertensive subjects after 15
383
days of consuming high-polyphenol dark chocolate. Journal of Nutrition, 138, 1671–1676.
384
Ignat, I., Volf, I., & Popa, V. I. (2011). A critical review of methods for characterization of polyphenolic
385
compounds in fruits and vegetables. Food Chemistry, 126, 1821-1835.
386
Jia, G. T., & Lu, X. Y. (2008). Enrichment and purification of madecassoside and asiaticoside from
387
Centella asiatica extracts with macroporous resins. Journal of Chromatography A, 1193, 136–141.
388
Jung, M.W., Ahn, K.H., Lee, Y., Kim, K.P., Paeng, I.R., Rhee, J.S., … Paeng, K.J. (2001). Evaluation on
389
the adsorption capabilities of new chemically modified polymeric adsorbents with protoporphyrin IX.
390
Journal of Chromatography A, 917, 87–93.
391
Langer, S., Marshall, L. J., Andrea, J. D., & Morgan, R. A. (2011). Flavanols and Methylxanthines in
392
Commercially Available DarkChocolate: A Study of the Correlation with Nonfat Cocoa Solids. Journal
393
of Agricultural and Food Chemistry, 59, 8435–8441.
394
Ma, C., Tao, G., Lou, J. Z., Wang, H., Gu, X., Hu, L., & Yin, M. (2014). Preparative separation and
395
purification of rosavin in Rhodiola rosea by macroporous adsorption resins. Separation and Purification
396
Technology, 69, 22–28.
397
Quideau, S., Deffieux, D., Douat-Casassus, C., & Pouysegu, L. (2011). Plant polyphenols: Chemical
AC C
EP
TE D
M AN U
SC
RI PT
& Liu, C. Z. (2007). Separation of scutellarin from crude extracts of Erigeron
21
ACCEPTED MANUSCRIPT
properties, biological activities, and synthesis. Angewandte Chemie International Edition, 50, 586-621.
399
Sun, L., Guo, Y., Fu, C., Li, J., & Li, Z. (2013). Simultaneous separation and purification of total
400
polyphenols, chlorogenic acid and phlorizin from thinned young apples. Food Chemistry, 136,
401
1022–1029.
402
Taeye, C., Cibaka, M. K., Jerkovic, V., & Collin, S. (2014). Degradation of (−)-Epicatechin and
403
Procyanidin B2 in Aqueous and Lipidic Model Systems. First Evidence of “Chemical” Flavan-3-ol
404
Oligomers in Processed Cocoa. Journal of Agricultural and Food Chemistry, 62, 9002−9016.
405
Wollgast, & Anklam. (2000). Review on polyphenols in Theobroma cacao: Changes in composition
406
during the manufacture of chocolate and methodology for identification and quantification. Food
407
Research International, 33, 423-447.
408
Wu, Z., Li, H., Ming, J., & Zhao, G. (2010).Optimization of adsorption of tea polyphenols into oat
409
b-glucan using response surface methodology. Journal of Agricultural and Food Chemistry, 59, 378–385.
410
Xi, L. S., Mu, T. S., & Sun, H. N. (2015). Preparative purification of polyphenols from sweet
411
potato(Ipomoea batatas L.) leaves by AB-8 macroporous resins. Food Chemistry, 172, 166–174.
412
Zhao, Z., Dong, L., Wu, Y., & Lin, F. (2011). Preliminary separation and purification of rutin and
413
quercetin from Euonymus alatus (Thunb.) Siebold extracts by macroporous resins. Food and Bioproducts
414
Processing, 89, 266–272.
SC
M AN U
TE D
EP
AC C
415
RI PT
398
22
ACCEPTED MANUSCRIPT Table 1. Chemical composition, physicochemical parameters, adsorption capacities and desorption ratio on different resins. Resin
Particle size
(mm)
area (m2/g)
(nm)
Moisture
Adsorption
(%)
(mg EC/g)
Desorption ratio (%)
Nonpolar
Polystyrene series
0.37 - 1.25
600 - 650
8.5 – 9.0
59.1 ± 4.2
36.9 ± 3.9
54.8 ± 2.2
LX-68
Nonpolar
Polystyrene series
0.37 - 1.25
1000 - 1050
7.0 – 7.5
58.2 ± 3.0
38.8 ± 4.1
43.6 ± 4.1
AB-8
Moderate-polar
Polystyrene series
0.37 - 1.25
1150 - 1200
12 - 16
58.3 ± 1.2
31.1 ± 1.0
68.5 ± 0.2
DM130
Moderate-polar
Polystyrene series
0.37 - 1.25
450 - 500
LX-32
Polar
Polystyrene series
0.37 - 1.25
650 - 700
XDA-1
Polar
Polystyrene series
0.37 - 1.25
XDA-8
Polar
Polystyrene series
LX-17
Polar
Polystyrene series
10 - 11
46.0 ± 0.9
20.2 ± 1.0
79.5 ± 0.1
-a
65.3 ± 2.0
40.5 ± 3.7
39.7 ± 1.5
1000 - 1100
7.0 – 7.5
48.1 ± 1.4
45.9 ± 2.2
32.2 ± 0.7
0.37 - 1.25
1050 - 1100
-a
53.6 ± 0.4
50.3 ± 0.9
52.4 ± 0.2
0.37 - 1.25
400 - 450
-a
55.2 ± 0.7
48.2 ± 1.8
87.7 ± 0.4
TE D
EP
The data was not collected.
SC
D101
AC C
419
a
Pore size
capacity
types
418
Specific surface
Material
RI PT
Polarity
M AN U
416 417
23
ACCEPTED MANUSCRIPT Table 2. Langmuir and Freundlich equation Temperatures (°C)
Langmuir equation
R2
Freundlich equation
R2
0.9977
Qe =
0.80Ce −4.91 + Ce
0.7663
Qe = 135.78Ce0.981
39.8
Qe =
2.78Ce 23.72 + Ce
0.5473
Qe = 117.65Ce1.028
0.9902
54.1
Qe =
8.87Ce 175.12 + Ce
0.1162
Qe = 70.13Ce1.067
0.9928
68.0
Qe =
−0.49Ce −28.43 + Ce
0.5359
RI PT
28.5
SC
420
M AN U
Qe = 53.77Ce1.133
421
AC C
EP
TE D
422
24
0.9997
ACCEPTED MANUSCRIPT Table 3. The results of the scale-up separations extracted residue Cocoa husks (kg) theobromine (g) (g) 490.77 ± 10.05 12.43 ± 0.48 2.00 ± 0.01 417.53 ± 8.76 9.99 ± 0.62 Water fraction 20% ethanol 9.67 ± 0.99 3.04 ± 0.81 fraction 50% ethanol 55.66 ± 2.63 0.28 ± 0.05 fraction 95% ethanol 4.23 ± 0.35 fraction
caffeine (g)
EC (g)
polyphenols (g EC)
1.74 ± 0.19
0.69 ± 0.04
44.54 ± 3.77
0.47 ± 0.08
-
10.87 ± 0.50
1.02 ± 0.07
-
0.47 ± 0.07
0.09 ± 0.01
-
0.63 ± 0.08
34.99 ± 1.21
0.05 ± 0.01
0.21 ± 0.03
SC
424
RI PT
423
AC C
EP
TE D
M AN U
425
25
ACCEPTED MANUSCRIPT
427
Table 4. The yielding capacity, waste solvent volume and estimated cost of cocoa polyphenols production Pt Typology (g EC/(L·day)) value
Cocoa husk
Cocoa bean
(¥/kg)
(¥/kg)
2
25
Labor cost
Energy cost
Psc (L/g EC)
96.4
0.42
(¥/man-month)
(¥/month)
5000
15000
AC C
EP
TE D
M AN U
SC
428
RI PT
426
26
Figure 1. The chemical structures ofACCEPTED (-)-epicatechin. MANUSCRIPT Figure 2. Factors affect the adsorption characteristics of LX-17 resin (A) Effects of initial concentration on the adsorption amounts, (B) effects of pH on the adsorption amounts, (C) effects of adsorption time on the adsorption amounts, (D) effects of temperature on the adsorption amounts.
RI PT
Figure 3. Dynamic breakthrough curve (A) and desorption curve (B) of EC packed with LX-17 resin
AC C
EP
TE D
M AN U
SC
at different flow rates.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights
AC C
EP
TE D
M AN U
SC
RI PT
Polyphenols from cocoa husks can be enriched with macroporous resins. The content of total polyphenols increased from 1.34% to 57.9%. Adsorption kinetics and thermodynamics are studied respectively. Caffeine and polyphenols were enriched in two fractions using LX-17 resin. The scaling-up separation was found satisfactory.
S0260-8774(18)30360-1
DOI:
10.1016/j.jfoodeng.2018.08.023
Reference:
JFOE 9373
To appear in:
Journal of Food Engineering
Received Date: 3 June 2017 Revised Date:
19 August 2018
Accepted Date: 22 August 2018
Please cite this article as: Zhong, J.-L., Muhammad, N., Gu, Y.-C., Yan, W.-D., A simple and efficient method for enrichment of cocoa polyphenols from cocoa bean husks with macroporous resins following a scale-up separation, Journal of Food Engineering (2018), doi: 10.1016/j.jfoodeng.2018.08.023. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
A simple and efficient method for enrichment of cocoa polyphenols from
2
cocoa bean husks with macroporous resins following a scale-up separation
3
Jia-Lun Zhong, Nadeem Muhammad, Yu-Chao Gu, Wei-Dong Yan1
4
Department of Chemistry, Zhejiang University, Hangzhou 310027, China
RI PT
1
5
Abstract
SC
6
The aim of this study is to develop an efficient method for the decaffeination and enrichment of the
8
polyphenols from cocoa husks extracts. LX-17 was selected for the exploration of optimal processing
9
parameters from eight kinds of macroporous resins. The adsorption kinetics and thermodynamics of
10
(-)-epicatechin (EC) were studied prior to the scale-up separation. The optimum parameters for separation
11
were as follow: 6.0 mg/mL cocoa extracts, pH 2.0, 25 °C column temperature, flow rates of adsorption
12
and desorption 1.6 BV/h (Bed Volume, the volume of the resin) and ethanol-water (20:80, 50:50, 95:5,
13
v/v) solutions in the gradient elution. In the scale-up separation, 2 kg of cocoa husks were extracted in 20
14
L 50 % ethanol solutions, separated on 3 L LX-17 resins and yielded 34.99 g cocoa polyphenols. This
15
method significantly increased the total polyphenol contents from 2.23 % to 62.87 % with a recovery
16
yield of 78.57 %.
17
Key words
18
Cocoa husk, Polyphenol, Macroporous resin, Adsorption, Separation, Scale-up
1
AC C
EP
TE D
M AN U
7
Corresponding author at: Department of Chemistry, Zhejiang University, Hangzhou 310027, China
Tel: 0086 571 87951430. Fax: 0086 571 8795189. E-mail: [email protected] ( W. D. Yan) 1
ACCEPTED MANUSCRIPT
19
Chemical compounds studied in this article:
20
(-)-Epicatechin (PubChem CID: 72276); Caffeine (PubChem CID: 2519); Theobromine (PubChem CID:
21
5429); Procyanidin B2 (PubChem CID: 122738); Procyanidin C1 (PubChem CID: 169853);
22
(PubChem CID: 702); Sodium Hydroxide (PubChem CID: 14798); Formic Acid (PubChem CID: 284);
23
N-Hexane (PubChem CID: 8058); Hydrochloric Acid (PubChem CID: 313); Acetonitrile (PubChem CID:
24
6342)
25
1. Introduction
M AN U
SC
RI PT
Ethanol
Cocoa beans (Theobroma cacao L.) have been cultivated for a long period of time as a major
27
ingredient of cocoa and chocolate (Baba et al., 2000; Andres-Lacueva et al., 2008). Cocoa beans contain
28
approximately 6-8 % polyphenols by dry weight (Grassi et al., 2008). Polyphenols are secondary plant
29
metabolites with high antioxidant properties and also have potentially beneficial effect on human health,
30
such as treatment and prevention of cancer, cardiovascular disease and other pathologies (Ignat et al.,
31
2011; Quideau et al., 2011). EU Commission approves health claim: Cocoa flavanols support a healthy
32
blood circulation (2013).
AC C
EP
TE D
26
33
The data obtained from the Food and Agriculture Organization of the United Nations Statistics
34
Division have shown that 4.58 million ton of cocoa beans were produced in 2013. Whereas, cocoa husks
35
about 5-10 % weight of cocoa beans were generated as a waste by-product of the cocoa industries. Cocoa
36
husks can be a rich source of polyphenols and its low cost as compared to cocoa beans make a better
37
choice. Three main polyphenol groups can be distinguished in cocoa: catechins, anthocyanins and
2
ACCEPTED MANUSCRIPT
procyanidins, whose common skeletal structure is a flavanoid-type structure (Wollgast and Anklam,
39
2000). The molecular structure of EC is shown in Figure 1. The crude polyphenol extracts of cocoa husks
40
always contain alkaloids, proteins, polysaccharides and other impurities. Therefore, an efficient
41
purification method is needed to enrich cocoa polyphenols.
RI PT
38
The conventional method of decaffeination and enrichment of cocoa polyphenols were carried out by
43
means of solid-liquid extraction (Chemat et al., 2011), precipitation or silica gel column chromatography.
44
However, these separation methods are inefficient, time consuming, laborious and using high contents of
45
toxic organic solvents.
M AN U
SC
42
Macroporous adsorption resins have been widely used for the separation and enrichment of
47
biochemical products. Researches like madecassoside and asiaticoside from Centella asiatica (Jia and Lu,
48
2008), rutin and quercetin from Euonymus alatus (Thunb.) Siebold (Zhao et al., 2011), flavonoids from
49
Glycyrrhiza glabra L. leaf (Dong et al., 2015) and rosavin from Rhodiola rosea (Ma et al., 2009) have
50
been reported due to the large surface area and distinct pore structure on macroporous resin which
51
increases the efficiency of purification process. In addition, the functional groups on the surface of the
52
macroporous resin also enhance the separation selectivity. Therefore, compounds can be eluted based on
53
polarity, molecular size and hydrogen-bond interactive forces. The mildness and low toxicity of these
54
adsorption and desorption interactions make this process food-friendly.
AC C
EP
TE D
46
55
Therefore, in this study the adsorption and desorption characteristics of EC and the optimal
56
separation condition for polyphenols on macroporous resins were studied. Meanwhile, the adsorption
57
kinetics and thermodynamics were also investigated. The decaffeination was achieved by gradient elution 3
ACCEPTED MANUSCRIPT
with different ratios of ethanol-water solutions. In the end, the experiments were scaled up to evaluate the
59
feasibility of the industrial process.
60
2. Experimental
61
2.1 Chemicals and materials
RI PT
58
Distilled water was purchased from Hangzhou Wahaha Group Co., Ltd while HPLC grade methanol
63
and acetonitrile were purchased from MERYER (Shanghai, China). Theobromine, caffeine and
64
(-)-epicatechin standards of analytical grade were purchased from Sigma-Aldrich.
65
phenol reagent, formic acid, n-hexane and ethanol were purchased from Sinopharm Chemical Reagent
66
Co., Ltd (Shanghai, China). The standard solutions were prepared by weighing the appropriate amounts
67
of standards and dissolving in 20 % ethanol-water at the concentration of 1.06 mg/mL. Cocoa beans
68
(Forastero) were purchased from Wuxi Aokang food Corp., Ltd (Wuxi, China). Cocoa husks were
69
purchased from Zhejiang Qili Xingguang Cocoa Products Corp., Ltd. The cocoa beans/cocoa husks were
70
milled into small pieces using liquid nitrogen.
Folin-Ciocalteu's
EP
TE D
M AN U
SC
62
Macroporous resins LX-68, LX-32, LX-17, XDA-1, XDA-8, AB-8, and DM130 were purchased
72
from Lanxiao technology Co., Ltd (Xian, China), D101 were purchased from Lansheng Co., Ltd (Xian,
73
China). The moisture contents of macroporous resins were determined. Macroporous resins were
74
pretreated by soaking in 95 % ethanol (1:5, v/v) and eluting with 5 % NaOH (1:5, v/v), 5% HCl (1:5, v/v)
75
and distilled water (1:5, v/v) before using.
76
2.2 Apparatus
AC C
71
4
ACCEPTED MANUSCRIPT
The scale-up extraction was carried out in 20L-Jacketed Pilot Plant Reactors connected with
78
20L-circulation constant temperature water bath (GYY-20L, Zhengzhou Zhuocheng instrument Co., Ltd.,
79
China). The centrifugation was performed on low speed centrifuge (SC-3610, Anhui USTC Zonkia
80
Scientific Instruments Co., Ltd., China). The pH value of the cocoa husk extracts was determined by
81
Waterproof pHScan2 Tester, Eutech Instruments Pte Ltd. Water bath was carried out in a thermostatic
82
water bath (THD-2006, Ningbo Tianheng Instrument Works Co., Ltd., China). The cocoa husk extracts
83
and the fractions eluted by macroporous resins were concentrated by a RE2000 rotary evaporator
84
(Shanghai Yarong Biochemical Co., Ltd., China). In the scale-up experiments, the fractions were
85
concentrated by a RE-5220 rotary evaporator (Shanghai Yarong Biochemical Co., Ltd., China). The
86
aqueous solutions were dried by a freeze drier (FD-1A-50, Beijing Boyikang experimental instrument Co.,
87
Ltd., China). The effluent and eluted solutions were collected by auto parts collector (BS-100N, Shanghai
88
Luxi analytical instrument Co., Ltd., China). HPLC analyses were carried out on Wufeng LC-100 HPLC
89
(Shanghai, China) coupled with the Wufeng LC-100 UV detector (Suzhou, China). The column used was
90
Phenomenex Luna C18 100A column (250 mm × 4.6 mm, 5 µm). The ultraviolet absorption of total
91
polyphenols was measured by using Photometer (UV-1800PC SPECTROPHOTOMETER, MAPADA,
92
Shanghai, China).
93
2.3 Extraction of cocoa beans/husks
AC C
EP
TE D
M AN U
SC
RI PT
77
94
The extraction procedure of cocoa beans/cocoa husks was carried out by following Langer (Langer
95
et al., 2011) scheme with slight modifications. The milled cocoa beans/cocoa husks were defatted and
96
extracted together with n-heptane (1: 5, w/v) and 50 % ethanol in a solvent ratio of 1: 10 (w/v) following
5
ACCEPTED MANUSCRIPT
it was vortexed (1 min), sonicated (50 °C, 30 min) and filtered. Solvents were added to the residue after
98
filtration and the process was repeated thrice. After that, the n-heptane layer was discarded and ethanol in
99
the lower extracted phase was removed by rotary evaporation under partial vacuum at 80 °C. The
RI PT
97
resulting yield was dry froze and stored in the fridge at -20 °C for further use.
101
2.4 HPLC analysis
SC
100
A linear gradient elution was used for the HPLC analyses from A (acetonitrile containing 20 % water)
102
to B (water containing 1 % formic acid). Gradient elution was followed as: from 11 % A to 20 % A in 20
104
min, keep 20 % A for 10 min, from 20 % A to 80 % A in 20 min, keep 80 % A for 5 min, and return to
105
11 % A in 3 min at a flow rate of 1.0 mL/min, 20 µL of the sample was injected in duplicates into the
106
column kept at 30 °C. The wavelength of the detector was 280 nm. Total polyphenols contents (TPC)
107
were determined with the Folin–Ciocalteu procedure described by Silke Elwers (Elwers et al., 2009).
108
2.5 Static adsorption and desorption tests
109
2.5.1
EP
TE D
M AN U
103
AC C
Selection of macroporous resins
110
2.0 g of accurately weighed pretreated D101, LX-68, LX-32, LX-17, XDA-1, XDA-8, AB-8, and
111
DM130 macroporous resins were put into a 250 mL conical flask and 50 mL of 6.0 mg/ml aqueous cocoa
112
beans extracts solution (pH 2.0) was added. TPC (mg EC/ml) of the aqueous cocoa beans extracts
113
solution was analyzed before adsorption. The sealed flasks were shaken at 25 °C for 60 min. After
114
adsorption, a 1 mL sample solution was withdrawn with an injection syringe a having a long needle for
115
analysis of TPC. Then the resins were washed with 50 mL distilled water twice, filtered and the residues 6
ACCEPTED MANUSCRIPT
were desorbed with 50 mL ethanol-water (50:50, v/v) solution, stirred for 60 min at 25 °C. After
117
desorption, TPC of the desorbed solution was analyzed. A circulating water bath with a thermostat was
118
applied to maintain the temperature within ±0.01 K. Adsorption capacities were calculated using equation
119
(1).
120
Qt = V0 (C0 − Ct ) / M
121
Desorption ratios were calculated using equation (2).
122
D = 100CdVd /[V0 (C0 − Ce )]
123
2.5.2
(1)
M AN U
SC
RI PT
116
(2)
Effects of initial concentration, temperature, pH and adsorption time on the adsorption amounts
The freeze-dried cocoa beans extracts were diluted with distilled water to different concentrations
125
(0.409, 2.04, 4.09, 6.13, 8.17 mg/mL). The solutions were adsorbed and desorbed as described in section
126
2.5.1 on the selected resin and the concentrations of EC were analyzed with HPLC. The effects of
127
temperature (methods described in section 2.6.3), pH (2.0, 3.0, 4.0, 5.0, 7.0 and 9.0) and adsorption time
128
(0, 5, 10, 15, 30 and 60 min) on adsorption capacity were also studied respectively.
129
2.5.3
130
AC C
EP
TE D
124
Adsorption kinetics and thermodynamics
The results of adsorption time effect were linearly regressed using the pseudo-first and second order
131
models.
132
Pseudo-first order model
7
ACCEPTED MANUSCRIPT
133
ln(Qe − Qt ) = ln Qe − K f t
134
Pseudo-second order model
135
t 1 1 = + t 2 Qt K s Qe Qe
RI PT
(3)
(4)
The selected resin (8.0 g) was added to series of 25 mL standard solutions having cumulative
137
concentration of EC (0.018, 0.026, 0.039, 0.054, 0.087 mg/mL). Adsorption was conducted at four
138
different temperatures (28.5, 39.8, 54.1, 68.0 °C) in a 100 mL boiling flask-3-necked with a stirring
139
paddle and a mercury thermometer inside. The adsorption isotherms for EC were determined using
140
Langmuir and Freundlich equations.
141
Langmuir equation (5):
142
Qe =
143
Freundlich equation (6):
144
Qe = K F Ce
145
The enthalpy changes (∆H) were calculated by the Van’t Hoff equation:
146
ln K = −
147
2.6 Dynamic adsorption and desorption
1
TE D
(5)
EP
AC C
QmCe K L + Ce
M AN U
SC
136
n
(6)
∆H +A RT
(7)
8
ACCEPTED MANUSCRIPT
148
2.6.1
Effects of flow rates and ethanol-water ratios on dynamic adsorption and desorption
Dynamic adsorption and desorption experiments were performed on glass column (length 50 cm × Φ
150
4 cm) wet-packed with the selected resin (100 mL, 25 °C). The aqueous cocoa beans extracts solutions
151
(6.0 mg/mL, pH 2.0) were adsorbed and then eluted by ethanol-water (10:90, 20:80, 40:60, 50:50, 60:40,
152
70:30, 95:5, v/v, 1.6 BV/h) eluent and the volume of each solution was 2.5 BV. An auto parts collector
153
was used to collect eluted solutions, analyzed by HPLC and TPC was determined. Effects of flow rates
154
(1.6, 2.4, 3.2 BV/h) on dynamic adsorption and desorption were also studied to obtain dynamic
155
breakthrough curves and desorption curves.
156
2.6.2
Enrichment under optimal conditions
M AN U
SC
RI PT
149
The 100.0 g cocoa husks were extracted as described in Section 2.3. The experiments were
158
performed on glass column (length 50 cm × Φ 4 cm) wet-packed with the selected resin (300 mL, 25 °C).
159
The aqueous cocoa husks extracts solutions (6.0 mg/mL, pH 2.0) were adsorbed at the flow rate of 1.6
160
BV/h. After the adsorption, the column was eluted by ethanol-water (0:100, 20:80, 50:50, 95:5, v/v. 1.6
161
BV/h) solutions and the volume of each solution was 5 BV. Each part of the eluted solutions was
162
concentrated in a rotary evaporator at 70 °C, analyzed by HPLC and TPC was determined.
163
2.7 The scale-up experiments
AC C
EP
TE D
157
164
The extraction of 2 kg cocoa husks was carried out in a 20 L-Jacketed Pilot Plant Reactor connected
165
with 20 L-circulation constant temperature water bath. The milled cocoa husks were extracted with 50 %
166
ethanol in a solvent ratio of 1:10 (w/v). Afterward 20 L solvents were added, the agitating vane in the 9
ACCEPTED MANUSCRIPT
167
reactor stirred up at the rate of 550 rpm. And the circulation constant temperature water bath was kept at
168
85 °C (80 °C in the reactor).
To validate the stability of EC and polyphenols, three sealed conical flask filled with 50 % ethanol
170
cocoa husk extracts (50 mL) were heated to 80 °C and maintained by the circulating water bath with a
171
thermostat. Meanwhile, at the time interval of 0 h, 1 h and 4 h the solution was withdrawn and injected to
172
the HPLC system to determine the concentration of EC and TPC.
SC
RI PT
169
The extraction was repeated for three times and the extracted solutions at the end were combined.
174
The extracts were evaporated in a rotary evaporator having a 20 L-round-bottom flask to remove the
175
organic solvents at 70 °C. The glass column (length 100 cm × Φ10 cm) was filled with 3L the selected
176
resin. The aqueous cocoa husks extracts solutions (6 mg/mL, pH 2.0) were adsorbed at the flow rate of
177
1.6 BV/h, 25 °C. The elution and analysis processes are described in section 2.6.2.
178
2.8 Evaluation of the scale-up process
EP
TE D
M AN U
173
The scale-up procedure described in section 2.7 was repeated for 10 cycles on the selected LX-17
180
resin (1 L, 25 °C) to evaluate the regeneration capacity of the exhausted resin. After each adsorption and
181
desorption process, the resin was pretreated as described in section 2.1.
182 183
AC C
179
Total polyphenols produced per volume of the selected resin per day was calculated using the equation:
10
ACCEPTED MANUSCRIPT
M tp
184
Pt =
185
The volume of waste solvent produced in the production of one mass unit of polyphenols was using the
186
equation:
187
Psc =
× 24
RI PT
(8)
(9)
SC
Vs M tp
The description of the equations (1-9) can be seen in Supporting Information and the reference
189
(Dong et al., 2015; Ma et al., 2009).
190
3. Results and discussion
M AN U
188
Vr ⋅ tr
According to Langer’s research (Langer et al., 2011), (-)-epicatechin (linear regression with total
192
polyphenols, R2 = 0.96) was a reliable marker of total polyphenols. The molecular structure of
193
(-)-epicatechin is similar to other polyphenols like catechin, epigallocatechin, procyanidin oligomers and
194
procyanidin polymers. Furthermore, the adsorption and desorption characteristics of these polyphenols on
195
the macroporous resins would also be similar. Therefore, the adsorption capacity and desorption ratios of
196
EC were determined in some parts of the experiments instead of TPC.
AC C
EP
TE D
191
197
The typical chromatogram of cocoa beans extracts is shown in Figure S1. The peaks of theobromine,
198
caffeine, procyanidin B2, (-)-epicatechin and procyanidin C1 were identified by the standard solutions
199
based on their retention time on HPLC (Esatbeyoglu et al., 2011). TPC was determined at equivalents of
200
EC (mg EC/mL). A calibration curve was made with the EC standards. The linear regression equation y =
11
ACCEPTED MANUSCRIPT
0.1056x – 0.0019 was obtained with R2 = 0.9975. The linear regression of the calibration curve of EC is
202
shown in Figure S2.
203
3.1 Extraction and analysis of cocoa beans and cocoa husks
RI PT
201
To compare the constituents and contents, the concentrations of theobromine, caffeine and EC and
205
TPCs in cocoa husks and cocoa beans were determined as listed in Table S1. It shows that cocoa beans
206
have higher TPCs than cocoa husks. In addition, the concentrations of theobromine, caffeine and EC in
207
cocoa beans were higher while there were no significant differences in constituents between cocoa beans
208
and cocoa husks. It should be noted that there was some cocoa bean powder left in the cocoa husks. The
209
TPC content can be lower in coca husks without the cocoa bean powder. However, TPC content in cocoa
210
husks are acceptable regard to the prices of cocoa beans and cocoa husks. Therefore, in this study, cocoa
211
beans were used in the static and dynamic experiments and cocoa husks were used in the final scale-up
212
separations.
213
3.2 Static adsorption and desorption
214
3.2.1 Selection of macroporous resins
AC C
EP
TE D
M AN U
SC
204
215
The pretreated resins were soaked in water and filtered to remove the free water before the
216
determination of moisture. The chemical composition, physicochemical parameters, adsorption capacities
217
and desorption ratio of polyphenols on these resins are shown in Table 1. The material of the tested resins
218
is all polystyrene series. LX-32, XDA-1, XDA-8 and LX-17 are all polar resins and have shown high
219
adsorption capacities while moderate-polar resins AB-8 and DM130 displayed low adsorption capacities. 12
ACCEPTED MANUSCRIPT
220
The maximum handling capacity of cocoa husks in one run depends on the maximum adsorption capacity
221
of macroporous resins.
Also, the physical adsorption should be reversible in the desorption process. The specific surface
223
areas of XDA-1 and XDA-8 were higher than LX-17 and there were more irreversible adsorptions on the
224
resins. In this way, the desorption ratios of XDA-1 and XDA-8 were lower than LX-17 resin. Thus,
225
LX-17 was selected as the effective macroporous resin in the following experiments.
226
3.2.2 Effects of initial concentration and pH on the adsorption capacity
M AN U
SC
RI PT
222
The effects of initial concentration and pH on the adsorption capacity are shown in Figure 2A and 2B.
228
Figure 2A showed that adsorption capacity rose as the concentration increased and the highest adsorption
229
capacity was observed when the initial concentration of EC reached 0.025 mg/mL (The aqueous cocoa
230
beans extracts solution, 6.0 mg/mL). On further increase of EC concentration of 0.035 mg/mL the
231
adsorption capacity decreased. It can be attributed to decrease in the number of free active sites and more
232
chance of impurities adsorption on LX-17 resin, resulting in competition for active sites between the
233
polyphenols and the impurities (Sun et al., 2013), which led to a slight drop in adsorption capacity. Thus,
234
the initial concentration of the aqueous cocoa beans extracts should be 6.0 mg/mL or less than it for the
235
adequate adsorption of EC.
AC C
EP
TE D
227
236
As Figure 2B shows, the resin showed the highest adsorption capacity of EC when the pH value of
237
the sample solution was 2.0. Figure 2B also shows that from pH 3.0 to pH 5.0 the adsorption capacity
238
changed little. The adsorption capacity of EC declined when the pH value was higher than 5.0. This may
13
ACCEPTED MANUSCRIPT
be due to existence of phenolic compounds as in molecule forms at low pH value. As the pH value rose,
240
polyphenols ionized and the physical interactions between polyphenols and the adsorptive sites of the
241
adsorbent declined. Similar results were found according to Xi et al. (Xi, Mu, & Sun, 2015). Thus, pH 2.0
242
was chosen as the optimum pH value of the sample solution.
243
3.2.3 Effects of adsorption time on the adsorption capacity and adsorption kinetics
SC
RI PT
239
The effect of adsorption time is shown in Figure 2C. The adsorption capacity of EC increased with
245
the adsorption time and reached to equilibrium at about 30 min. In the first 5 min, the adsorption
246
capacities increased rapidly and from 5 to 15 min, the adsorption capacity increased slowly. Later from 15
247
to 60 min the adsorption capacity changed little indicating attaining of the adsorption equilibrium.
248
Therefore, the adsorption process can be clarified into three stages. The results of adsorption time effect
249
were linearly regressed using the pseudo first order ( ln(0.725 − Qt ) = ln 0.725 − 0.1175t R2 = 0.9904) and
250
second order model (
251
pseudo-first-order model could be used to describe the initial (0 - 5 min), second (5 - 15 min) and third
252
(15 - 45 min) adsorption process with different slopes while the pseudo second order model can be used
253
to describe the whole process. Similar observations were found in previous studies (Buran et al., 2014;
254
Dong et al., 2015) and 60 min was selected as the optimal adsorption time for static tests.
255
3.2.4 Effects of temperature on the adsorption capacity and adsorption thermodynamics
TE D
M AN U
244
R2 = 0.9998) as shown in Figure S3.1-3.2. The
AC C
EP
t t 1 = + Qt 1.13 0.38 × 1.132
256
The effect of temperature on static adsorption is shown in Figure 2D, which indicates that the
257
adsorption capacity decreased as the temperature rose from 28.5 °C to 68.0 °C and it is also confirmed by 14
ACCEPTED MANUSCRIPT
Freundlich equation. The data were linearly regressed by the Langmuir equation and Freundlich equation,
259
respectively. As it is shown in Table 2 and Figure S4, the Freundlich equation appeared fitting better. The
260
constant of Freundlich equation KF decreased as the temperature increased which indicated that the
261
adsorption capacity was higher at the room temperature. Thus, 25.0 °C was selected as the column
262
temperature. According to the Van’t Hoff equation (linear regressed in Figure S5), the enthalpy change
263
(∆H) were calculated as -21.12 kJ/mol which indicated the adsorption process was exothermic and the
264
low temperature was appropriate for the adsorption process. Macroporous resins adsorb the targeted
265
constituents selectively through electrostatic force, hydrogen bonding interactions, complexation and size
266
sieving (Gao et al., 2007). The interactions declined as temperature rose due to molecular thermodynamic
267
movement, the swelling of resin and the increasing kinetic energies of the polyphenol molecules at higher
268
temperatures (Sun et al., 2013; Wu et al., 2010).
269
3.3 Dynamic adsorption and desorption
270
3.3.1
TE D
M AN U
SC
RI PT
258
EP
Effect of ethanol concentration on the desorption ratio and the elution order
Different content of ethanol-water solutions (10:90, 20:80, 40:60, 50:50, 60:40, 70:30 and 95:5, v/v)
272
were used to perform desorption tests in order to study the effect of ethanol content on the ratio of
273
desorption and the elution order of adsorbates as results are shown in Table S2 and Figure S6. The main
274
peaks of cocoa bean extracts are theobromine, caffeine, procyanidin B2, EC, procyanidin C1 and
275
procyanidin polymers as shown in Figure S6A. Most of theobromine could be eluted by water as shown
276
in Figure S6B whereas caffeine could be eluted by 20 % ethanol Figure S6D. Thus, removal of
277
theobromine and decaffeination were achieved by eluting with 20 % ethanol content and most of the
AC C
271
15
ACCEPTED MANUSCRIPT
polyphenols were still adsorbed on the macroporous resin. It should be noted that about 20 % of the
279
polyphenols were eluted in the water fraction while the resin has not reached the saturated adsorption. It
280
was probably because sugars and proteins were eluted in the water fractions. On the one hand, the result
281
determined by the Folin–Ciocalteu method would be interfered by sugars in the water fraction and a part
282
of the polyphenols would combine proteins by chemical bonds. From 40 % to 60 % ethanol fractions, EC
283
and polyphenols were enriched in these fractions as shown in Figure S6E. TPCs were all over 45 % and
284
the rest of the adsorbates were eluted in the 70 % and 95 % fraction. In this way, ethanol-water (20:80,
285
50:50, 95:5, v/v) solutions were used in the gradient elution.
286
3.3.2
M AN U
SC
RI PT
278
Effects of flow rate on dynamic adsorption and desorption
By choosing the initial concentration of EC 25 mg/L, the flow rate was investigated at 1.6 BV/h, 2.4
288
BV/h and 3.2 BV/h. It can be seen from Figure 3A that better adsorption performance of EC was
289
exhibited at the flow rate of 1.6 BV/h. The breakthrough point was defined as 1% ratio of the exit to the
290
inlet solute concentration (Jung et al., 2001). The breakthrough point of LX-17 resin at the flow rate of
291
1.6 BV/h was approximately 4500 mL. While at the flow rate of 2.4 BV/h and 3.2 BV/h, the breakthrough
292
volume was approximately 3000 mL and 2500 mL respectively. At the volume of 5600 mL the
293
concentration of EC eluted was coincident. The concentration of EC eluted rose steadily after 5600 mL at
294
the flow rate of 1.6 BV/h. In contrast, at the flow rate of 3.2 BV/h it showed as a flat curve. And the cocoa
295
extract solutions broke through at the volume of approximately 9000 mL. As section 3.2.3 indicated, EC
296
had less time to interact with the surface of the resin at higher flow rates and the adsorption process was
297
inadequate resulting in EC broke through the macroporous resin. Thus, 1.6 BV/h was selected as the best
AC C
EP
TE D
287
16
ACCEPTED MANUSCRIPT
298
dynamic adsorption flow rate. At the flow rate of 1.6 BV/h, the adsorption capacity of EC on LX-17 resin
299
was 1.27 mg/g which was coincident with the results obtained from section 3.2.2.
Better desorption performance of EC was exhibited at the flow rate of 1.6 BV/h according to Figure
301
3B. Polyphenols were totally desorbed by ethanol-water solutions (50:50, v/v) in 2 BV at the flow rate of
302
1.6 BV/h. As the flow rate increased, polyphenols were totally desorbed in 3 BV, 3.6 BV at the flow rate
303
of 2.4 BV/h, 3.2 BV/h respectively. At the lower flow rate, adsorbents were eluted in higher
304
concentrations with lower cost of ethanol-water solutions. 1.6 BV/h was selected as the proper flow rate
305
of desorption.
306
3.3.3
M AN U
SC
RI PT
300
Gradient elution tests
100.0 g cocoa husks were extracted and separated under optimal conditions on LX-17 resin in small
308
scale. The results are listed in Table S3. Although it was far from saturated adsorption, the similar results
309
were given as section 3.3.1 that about 20 % polyphenols were eluted in the water fraction. The
310
discrepancy was that 50 % of theobromine was eluted in the 20 % ethanol fraction while in section 3.3.1
311
the value was only 10 %. It is probably because the saturated adsorption of theobromine has reached and
312
most of the theobromine (90 %) broke through in the water fraction. The bed volume of LX-17 in section
313
3.3.1 was 100 mL. The test on small scale was generally satisfactory. Theobromine and caffeine were
314
eluted in 20 % ethanol fraction. In the 50 % ethanol fraction, the mass of dried residue was 1.79 g and the
315
mass of total polyphenols was 1.04 g EC with TPC 58.10 %.
316
3.4 The scale-up experiments
AC C
EP
TE D
307
17
ACCEPTED MANUSCRIPT
Before the experiments, the stability of EC and polyphenols was validated at 80 °C and the
318
concentration of EC and TPC changed slightly as shown in Table S4. In the study of Taeye (Taeye et al.,
319
2014), EC and procyanidin B2 degraded obviously at 60 and 90 °C, where these compounds were studied
320
individually. While in this study, both EC and polyphenols were extracted from cocoa husks and they
321
existed as mixtures in an equilibrium state. The degradation and epimerization had finished in natural.
322
Thus, EC and polyphenols at the extraction temperature of 80 °C were proved to be stable and the
323
extraction was sufficient at a higher temperature.
M AN U
SC
RI PT
317
The results of scale-up separation are listed in Table 3. Two kilograms of cocoa husks containing
325
44.54 g polyphenols were enriched through 3 L LX-17 resins. The results were similar to that of small
326
scale experiments. Theobromine and caffeine were eluted mainly in water fraction and 20 % ethanol
327
fraction, respectively. The mass of 50 % ethanol fraction was 55.66 g with 34.99 g polyphenols. The
328
value of TPC was 62.86 % and the recovery yield was 78.56 %. The content of caffeine by dry weight in
329
the yields was only 0.16 %. The difference with small scale tests was that much more polyphenols were
330
extracted in the scale-up separation, it might be due to high temperature that made the extraction
331
complete.
332
3.5 Evaluation the feasibility of scale-up process
AC C
EP
TE D
324
333
The regeneration capacity of the exhausted resin is an important factor in the process economy in the
334
enrichment of polyphenols. Consecutive polyphenols adsorption and desorption were repeated for 10
335
cycles with the same amount of resins and cocoa husks. At the breakthrough point, the adsorption
336
capacity was calculated every 2 cycles and then desorbed to obtain the desorption ratio. The results were 18
ACCEPTED MANUSCRIPT
shown in Table S5. The regeneration capacity of the exhausted resin performed consistently in 10 cycles
338
or more which were encouraging for production. Moreover, the yielding capacity, waste solvent volume
339
and estimated cost of cocoa polyphenols production are listed in Table 4. With Pt and the volume of resin,
340
polyphenols produced in one month could be calculated. While some of the ethanol in wasted solvents
341
can be regenerated. According to the TPC obtained in section 3.1 and recovery yield in section 3.4, the
342
exact costs of total polyphenols in cocoa husks and cocoa beans were 189.9 ¥/kg EC and 822.3 ¥/kg EC,
343
respectively. And the price of 40 % cocoa polyphenols was 200 ¥/kg, approximately. Despite of the cost
344
of ethanol and resins, cocoa husks could be used to produce cocoa polyphenols with profits.
345
4 Conclusion
M AN U
SC
RI PT
337
The decaffeination and enrichment of cocoa polyphenols from low cost cocoa husks were achieved
347
by gradient elution on the LX-17 macroporous resin. The optimal conditions for adsorption and
348
desorption were achieved by static and dynamic adsorption experiments as: cocoa extracts 6.0 mg/mL,
349
pH 2.0, column temperature 25 °C, flow rates of adsorption and desorption 1.6 BV/h and ethanol-water
350
(20:80, 50:50, 95:5, v/v) solutions were used in the gradient elution. The results of scaling-up separation
351
are satisfactory which suggested that the resin purification technology was efficient and food-friendly that
352
can be scaled up in the industry to produce polyphenol rich products with low-cost cocoa husks.
353
Acknowledgements
354
This project was financially supported by Sky Herb Co., Ltd (Huzhou, China).
355
References
AC C
EP
TE D
346
19
ACCEPTED MANUSCRIPT
Andres-Lacueva, C., Monagas, M., Khan, N., Izquierdo-Pulido, M., Urpi-Sarda, M., Permanyer, J., &
357
Lamuela-Raventos, R. M. (2008). Flavanol and flavonol contents of cocoa powder products: Influence of
358
the manufacturing process. Journal of Agricultural and Food Chemistry, 56, 3111–3117.
359
Baba, S., Osakabe, N., Natsume, M., Yasuda, A., Takizawa, T., Nakamura, T., & Terao, J. (2000). Cocoa
360
powder enhances the level of anti-oxidative activity in rat plasma. British Journal of Nutrition, 84,
361
673–680.
362
Buran, T. J., Sandhu, A. K., Li, Z., Rock, C. R., W, W. H., & Gu, L. W. (2014). Adsorption/desorption
363
characteristics and separation of anthocyanins and polyphenols from blueberries using macroporous
364
adsorbent resins. Journal of Food Engineering, 128, 167–173.
365
Chemat, F., & Zill-e-Huma and Khan, M. K. (2011). Applications of ultrasound in food technology:
366
Processing, preservation and extraction. Ultrasonics Sonochemistry,18, 813-835.
367
Dong, Y., Zhao, M., Sun-Waterhouse, D., Zhuang, M., Chen, H., Feng, M., & Lin, L. (2015). Absorption
368
and desorption behaviour of the flavonoids from Glycyrrhiza glabra L. leaf on macroporous adsorption
369
resins. Food Chemistry, 168, 538–545.
370
Duran, C., Ozdes, D., Gundogdu, A., & Senturk, H. B. (2011). Kinetics and isotherm analysis of basic
371
dyes adsorption onto almond shell (Prunus dulcis) as a low cost adsorbent. Journal of Chemical and
372
Engineering Data, 56, 2136–2147.
373
Elwers, S., Zambrano, A., Rohsius, C., & Lieberei, R. (2009). Differences between the content of
374
phenolic compounds in Criollo, Forastero and Trinitario cocoa seed (Theobroma cacao L.). European
375
Food Research and Technology, 229, 937–948.
376
Esatbeyoglu, T., Wray, V., & Winterhalter, P. (2011). Isolation of dimeric, trimeric, tetrameric and
AC C
EP
TE D
M AN U
SC
RI PT
356
20
ACCEPTED MANUSCRIPT
377
pentameric procyanidins from unroasted cocoa beans (Theobroma cacao L.) using countercurrent
378
chromatography. Food Chemistry, 179, 278–289.
379
Gao, M., Huang, W.
380
breviscapus(vant.) Hand. Mazz. By macroporous resins. Journal of Chromatography B, 858, 22–26.
381
Grassi, D., Desideri, G., Necozione, S., Lippi, C., Casale, R., Properzi, G., & Ferri, C. (2008). Blood
382
pressure is reduced and insulin sensitivity increased in glucose-intolerant, hypertensive subjects after 15
383
days of consuming high-polyphenol dark chocolate. Journal of Nutrition, 138, 1671–1676.
384
Ignat, I., Volf, I., & Popa, V. I. (2011). A critical review of methods for characterization of polyphenolic
385
compounds in fruits and vegetables. Food Chemistry, 126, 1821-1835.
386
Jia, G. T., & Lu, X. Y. (2008). Enrichment and purification of madecassoside and asiaticoside from
387
Centella asiatica extracts with macroporous resins. Journal of Chromatography A, 1193, 136–141.
388
Jung, M.W., Ahn, K.H., Lee, Y., Kim, K.P., Paeng, I.R., Rhee, J.S., … Paeng, K.J. (2001). Evaluation on
389
the adsorption capabilities of new chemically modified polymeric adsorbents with protoporphyrin IX.
390
Journal of Chromatography A, 917, 87–93.
391
Langer, S., Marshall, L. J., Andrea, J. D., & Morgan, R. A. (2011). Flavanols and Methylxanthines in
392
Commercially Available DarkChocolate: A Study of the Correlation with Nonfat Cocoa Solids. Journal
393
of Agricultural and Food Chemistry, 59, 8435–8441.
394
Ma, C., Tao, G., Lou, J. Z., Wang, H., Gu, X., Hu, L., & Yin, M. (2014). Preparative separation and
395
purification of rosavin in Rhodiola rosea by macroporous adsorption resins. Separation and Purification
396
Technology, 69, 22–28.
397
Quideau, S., Deffieux, D., Douat-Casassus, C., & Pouysegu, L. (2011). Plant polyphenols: Chemical
AC C
EP
TE D
M AN U
SC
RI PT
& Liu, C. Z. (2007). Separation of scutellarin from crude extracts of Erigeron
21
ACCEPTED MANUSCRIPT
properties, biological activities, and synthesis. Angewandte Chemie International Edition, 50, 586-621.
399
Sun, L., Guo, Y., Fu, C., Li, J., & Li, Z. (2013). Simultaneous separation and purification of total
400
polyphenols, chlorogenic acid and phlorizin from thinned young apples. Food Chemistry, 136,
401
1022–1029.
402
Taeye, C., Cibaka, M. K., Jerkovic, V., & Collin, S. (2014). Degradation of (−)-Epicatechin and
403
Procyanidin B2 in Aqueous and Lipidic Model Systems. First Evidence of “Chemical” Flavan-3-ol
404
Oligomers in Processed Cocoa. Journal of Agricultural and Food Chemistry, 62, 9002−9016.
405
Wollgast, & Anklam. (2000). Review on polyphenols in Theobroma cacao: Changes in composition
406
during the manufacture of chocolate and methodology for identification and quantification. Food
407
Research International, 33, 423-447.
408
Wu, Z., Li, H., Ming, J., & Zhao, G. (2010).Optimization of adsorption of tea polyphenols into oat
409
b-glucan using response surface methodology. Journal of Agricultural and Food Chemistry, 59, 378–385.
410
Xi, L. S., Mu, T. S., & Sun, H. N. (2015). Preparative purification of polyphenols from sweet
411
potato(Ipomoea batatas L.) leaves by AB-8 macroporous resins. Food Chemistry, 172, 166–174.
412
Zhao, Z., Dong, L., Wu, Y., & Lin, F. (2011). Preliminary separation and purification of rutin and
413
quercetin from Euonymus alatus (Thunb.) Siebold extracts by macroporous resins. Food and Bioproducts
414
Processing, 89, 266–272.
SC
M AN U
TE D
EP
AC C
415
RI PT
398
22
ACCEPTED MANUSCRIPT Table 1. Chemical composition, physicochemical parameters, adsorption capacities and desorption ratio on different resins. Resin
Particle size
(mm)
area (m2/g)
(nm)
Moisture
Adsorption
(%)
(mg EC/g)
Desorption ratio (%)
Nonpolar
Polystyrene series
0.37 - 1.25
600 - 650
8.5 – 9.0
59.1 ± 4.2
36.9 ± 3.9
54.8 ± 2.2
LX-68
Nonpolar
Polystyrene series
0.37 - 1.25
1000 - 1050
7.0 – 7.5
58.2 ± 3.0
38.8 ± 4.1
43.6 ± 4.1
AB-8
Moderate-polar
Polystyrene series
0.37 - 1.25
1150 - 1200
12 - 16
58.3 ± 1.2
31.1 ± 1.0
68.5 ± 0.2
DM130
Moderate-polar
Polystyrene series
0.37 - 1.25
450 - 500
LX-32
Polar
Polystyrene series
0.37 - 1.25
650 - 700
XDA-1
Polar
Polystyrene series
0.37 - 1.25
XDA-8
Polar
Polystyrene series
LX-17
Polar
Polystyrene series
10 - 11
46.0 ± 0.9
20.2 ± 1.0
79.5 ± 0.1
-a
65.3 ± 2.0
40.5 ± 3.7
39.7 ± 1.5
1000 - 1100
7.0 – 7.5
48.1 ± 1.4
45.9 ± 2.2
32.2 ± 0.7
0.37 - 1.25
1050 - 1100
-a
53.6 ± 0.4
50.3 ± 0.9
52.4 ± 0.2
0.37 - 1.25
400 - 450
-a
55.2 ± 0.7
48.2 ± 1.8
87.7 ± 0.4
TE D
EP
The data was not collected.
SC
D101
AC C
419
a
Pore size
capacity
types
418
Specific surface
Material
RI PT
Polarity
M AN U
416 417
23
ACCEPTED MANUSCRIPT Table 2. Langmuir and Freundlich equation Temperatures (°C)
Langmuir equation
R2
Freundlich equation
R2
0.9977
Qe =
0.80Ce −4.91 + Ce
0.7663
Qe = 135.78Ce0.981
39.8
Qe =
2.78Ce 23.72 + Ce
0.5473
Qe = 117.65Ce1.028
0.9902
54.1
Qe =
8.87Ce 175.12 + Ce
0.1162
Qe = 70.13Ce1.067
0.9928
68.0
Qe =
−0.49Ce −28.43 + Ce
0.5359
RI PT
28.5
SC
420
M AN U
Qe = 53.77Ce1.133
421
AC C
EP
TE D
422
24
0.9997
ACCEPTED MANUSCRIPT Table 3. The results of the scale-up separations extracted residue Cocoa husks (kg) theobromine (g) (g) 490.77 ± 10.05 12.43 ± 0.48 2.00 ± 0.01 417.53 ± 8.76 9.99 ± 0.62 Water fraction 20% ethanol 9.67 ± 0.99 3.04 ± 0.81 fraction 50% ethanol 55.66 ± 2.63 0.28 ± 0.05 fraction 95% ethanol 4.23 ± 0.35 fraction
caffeine (g)
EC (g)
polyphenols (g EC)
1.74 ± 0.19
0.69 ± 0.04
44.54 ± 3.77
0.47 ± 0.08
-
10.87 ± 0.50
1.02 ± 0.07
-
0.47 ± 0.07
0.09 ± 0.01
-
0.63 ± 0.08
34.99 ± 1.21
0.05 ± 0.01
0.21 ± 0.03
SC
424
RI PT
423
AC C
EP
TE D
M AN U
425
25
ACCEPTED MANUSCRIPT
427
Table 4. The yielding capacity, waste solvent volume and estimated cost of cocoa polyphenols production Pt Typology (g EC/(L·day)) value
Cocoa husk
Cocoa bean
(¥/kg)
(¥/kg)
2
25
Labor cost
Energy cost
Psc (L/g EC)
96.4
0.42
(¥/man-month)
(¥/month)
5000
15000
AC C
EP
TE D
M AN U
SC
428
RI PT
426
26
Figure 1. The chemical structures ofACCEPTED (-)-epicatechin. MANUSCRIPT Figure 2. Factors affect the adsorption characteristics of LX-17 resin (A) Effects of initial concentration on the adsorption amounts, (B) effects of pH on the adsorption amounts, (C) effects of adsorption time on the adsorption amounts, (D) effects of temperature on the adsorption amounts.
RI PT
Figure 3. Dynamic breakthrough curve (A) and desorption curve (B) of EC packed with LX-17 resin
AC C
EP
TE D
M AN U
SC
at different flow rates.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights
AC C
EP
TE D
M AN U
SC
RI PT
Polyphenols from cocoa husks can be enriched with macroporous resins. The content of total polyphenols increased from 1.34% to 57.9%. Adsorption kinetics and thermodynamics are studied respectively. Caffeine and polyphenols were enriched in two fractions using LX-17 resin. The scaling-up separation was found satisfactory.