ARTICLE

pubs.acs.org/IECR

Simultaneous Extraction and Separation Process for Coffee Beans

with Supercritical CO2 and Water
Siti Machmudah,†,‡ Kiwa Kitada,§ Mitsuru Sasaki,§ Motonobu Goto,*,† Jun Munemasa,# and
Masahiro Yamagata#
†
Bioelectrics Research Center, Kumamoto University, Kurokami 2-39-1, Kumamoto 860-8555, Japan
‡
Chemical Engineering Department, Sepuluh Nopember Institute of Technology, Kampus ITS Keputih, Sukolilo, Surabaya 60111,
Indonesia
§
Graduate School of Science and Technology, Kumamoto University, Kurokami 2-39-1, Kumamoto 860-8555, Japan
#
Kobe Steel Ltd., Shinhama Arai-cho, Takasago-city, Hyogo 676-8670, Japan

ABSTRACT: The aim of this work was to develop a new process for extraction and separation of polar and nonpolar compounds
from natural materials using supercritical CO2 (SCCO2) in water. In this work, chlorogenic acid and caffeine from coffee beans were
used as model compounds of polar and nonpolar compounds, respectively. Extraction and separation were conducted in three types
of semicontinuous extractors with different sizes and flow types of CO2 and water. The effect of temperature and pressure in the
separation mode using SCCO2 in flow and water in batch mode has been reported. Recovery of caffeine in the SCCO2 phase
increased with decreasing temperature and increasing pressure. However, only the increasing pressure could promote the increasing
recovery of caffeine and chlorogenic acid in water phases. The change of height of the nonpolar recovery section on the recovery of
caffeine and chlorogenic acid in SCCO2 and water phases and on the separation factor has been studied using SCCO2 and water in
co-current and counter-current flow modes. Recovery of caffeine and separation factor of caffeine from chlorogenic acid significantly
increased with increasing height of nonpolar recovery section. The effect of temperature in counter-current type of separation on the
recovery of caffeine and chlorogenic acid in SCCO2 and water phases exhibited different trend with batch type of separation. The
counter-current type of separation was more effective compared to both batch and co-current types of separation. On the basis of the
results of this work, it is suggested to design longer extractor columns with a higher nonpolar recovery section in the counter-current
type in order to increase separation efficiency.

1. INTRODUCTION respectively. Caffeine and chlorogenic acids from coffee beans are
Supercritical fluids extraction (SFE) has immediate advan- used as nonpolar and polar model compounds, respectively.
tages over conventional extraction techniques: it is a flexible Coffee is one of the most popular beverages in the world.
process due to the possibility of adjustment of the solvent power Coffee beans contain around 20 wt % of water-soluble substances,
or selectivity of the SCF, it allows protection of polluting organic including caffeine and chlorogenic acids.2 Caffeine (1,3,7-trime-
solvents, and it eliminates the expensive cost of postprocessing of thylxanthine) is an alkaloid generally responsible for about 0.9-
the extract for solvent reduction. Even though several com- 2.5% of coffee dry matter composition.3 Even though caffeine
pounds have been examined as SFE solvents, the most popular has been widely consumed and studied for centuries, research
one is carbon dioxide (CO2). CO2 is an inert, inexpensive, easily results are inconclusive about both adverse and beneficial effects
available, odorless, tasteless, environment-friendly, and generally of caffeine to several health outcomes. Low to moderate caffeine
regarded as safe (GRAS) solvent. CO2 offers supercritical oper- intake is generally associated with improvements in alertness,
ations at relatively low pressures and at near-room temperatures. learning capacity, exercise performance, and perhaps mood.4
Recently, water has also been used as supercritical solvent for Caffeine is also often used as an additive in pain medications.5
extraction, especially in the field of food application. Water is a However, its stimulatory effects may also adversely affect sensi-
polar component that is a environmentally friendly solvent, especially tive individuals by causing tachycardia, increase of blood pres-
for non-heat-sensitive compounds.1 sure, anxiety, and insomnia.6
Supercritical CO2 (SCCO2) is a nonpolar solvent and usually Decaffeination is performed prior to the roasting process. The
used for extraction of nonpolar compounds. Extraction of polar most common and least costly caffeine extraction methods in the
compounds is conducted using SCCO2 with the addition of a coffee industry employ an organic solvent associated with the use
polar solvent, such as ethanol, methanol, and water. Separation of water/vapor prior to and after extraction. Water alone has also
and extraction of nonpolar and polar compounds simultaneously
from natural products are necessary to simplify the separation Received: June 8, 2010
process. In this work, a simultaneous extraction and separation Accepted: December 9, 2010
process was proposed using SCCO2 and water as solvent to Revised: December 8, 2010
extract nonpolar and polar compounds from natural products, Published: January 14, 2011

r 2011 American Chemical Society 2227 dx.doi.org/10.1021/ie101252w | Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

been used to replace organic solvents in the process. By using

water, the caffeine content is usually reduced to 0.02-0.3%.7
Supercritical fluid, particularly carbon dioxide, is an alternative to
be applied in the decaffeination process. More recent commercial
application was the decaffeination of coffee and tea.2,8 The de-
caffeination process advantageously eliminates residual solvent.
The decaffeination process with SCCO2 can be improved by
adding water as a cosolvent.9 Soaking the raw beans in water prior
to decaffeination enhanced the rate of extraction, which in-
creased markedly with water content. Using CO2 saturated with
water also increased the rate of extraction.
Chlorogenic acids (CGA) are water-soluble phenolic compo-
nents of coffee and other plants formed by the esterification of
certain trans-cinnamic acids, such as cafeic (CA), ferulic (FA),
and p-coumaric acids (CoA), with (-)-quinic acid.10 CGA not
only contributes to coffee flavor but also may be of potential
biopharmacological importance in humans. The most studied
pharmacological activities of phenolic compounds such as CGA
have been related to their antioxidant properties because they are
thought to have positive effects on chronic degenerative diseases.11,12 Figure 1. Concept of simultaneous extraction and separation process.
Blue circle = water; O = SCCO2; red circle = polar compound; b =
In the decaffeination process with water, CGA is usually extracted nonpolar compound.
together with caffeine and their related compounds. Azevedo et
al. has reported that CGA might be extracted by SCCO2 with into the beans and as a result caffeine is dissolved in water and
polar solvents, such as ethanol and isopropyl alcohol.13 CO2 is a transferred from inside to the surface of the beans' matrix. Fur-
nonpolar solvent with low affinity for polar substances. So, the thermore, caffeine on the surface of the beans is directly dissolved
solubility of substances in supercritical CO2 decreases with the in SCCO2. The extraction is efficient if all caffeine contained in
increase in the number of polar functional groups (e.g., hydroxyl, the beans is dissolved and extracted in the water, then the caffeine
carboxyl, amino, and nitro). Thus, the solubility of CGA mole- existing in the water is expected to be transferred and dissolved in
cules is expected to be low, particularly as the molecular weight SCCO2. In addition, water may act as an entrainer that can
increases.14-16 enhance the solubility of caffeine in SCCO2.8
To extract both CGA and caffeine, and to separate CGA from CGA is one of the polar compounds contained in coffee beans,
caffeine in the decaffeination process, a simultaneous extraction which has pharmaceutical importance in humans. CGA can be
process was developed in this work. Both supercritical CO2 and decomposed in the roasting process, and as a result small amount
water were used as solvents in this process to extract caffeine of CGA remained in the coffee beans. Therefore, it is necessary to
and CGA, respectively. The CGA was expected to be isolated extract both caffeine and CGA from coffee beans before roasting
from caffeine in the water phase. The concept of the extraction process for collecting the valuable compound (in this case CGA)
process will be explained in detail in the next section. In this work, and removing the unnecessary compound (in this case caffeine).
the extraction and separation process was proposed experi- In the case of CGA, it will be recovered in the water phase due to
mentally to study the effect of extraction conditions and type its polarity and can be added again into the coffee drink or coffee
of extractors on the recovery of CGA and separation of caffeine beans after the roasting process by dispersing the CGA into the
from the coffee beans and extracted compounds in the water beans. In this extraction process, recoveries of compounds in
phase. SCCO2 and water phases are expected to contain caffeine and
CGA in high purity, respectively.
2. CONCEPT OF SIMULTANEOUS EXTRACTION PROCESS
3. MATERIALS AND METHODS
Concept of the simultaneous extraction and separation pro-
cess is shown in Figure 1, where both water and SCCO2 flow 3.1. Materials and Chemicals. In this work, green Arabica
counter-currently through the extractor or only SCCO2 flows coffee beans from Costa Rica with a moisture content of less than
while water is steady in the extractor. In this case, SCCO2 is 5% purchased from a local market were used as starting material.
dispersed in water. Target component of nonpolar compound is Caffeine (purity: 99.0%) and CGA (purity: 99.5%) standards and
directly extracted or dissolved from the surface and inside of the HPLC grade of methanol, acetonitrile, and phosphoric acid were
raw material into SCCO2. A part of nonpolar compounds is provided by Wako Chemical Ltd. CO2 with purity of 99.9% and
initially dissolved in water from the surface or inside of the raw glass beads for experiments were obtained from Uchimura Co.,
material, and then it is dissolved into SCCO2. The nonpolar Japan, and As One Co., Japan, respectively.
compounds are expected to be collected in the SCCO2 phase. On 3.2. Experimental Method. The experiments were con-
the other hand, the target component of the polar compound is ducted in three types of semicontinuous extractors made of
dissolved in water and collected in the water phase. In SCCO2 SUS316. The first type of extractor had an inside diameter of 40 mm,
extraction of natural products, water may play an important role, length of 50 mm, and volume of 60 mL. The second type of
and extraction of caffeine from coffee beans is as an example. extractor had an inside diameter of 40 mm, length of 200 mm,
Caffeine is extracted by SCCO2 from coffee beans wetted and and volume of 250 mL. The third type of extractor had an inside
swollen by water. In this case, coffee beans were saturated with diameter of 21.4 mm, length of 697 mm, and volume of 250 mL.
water prior to SCCO2 extraction. Water penetrated and diffused In this work, whole coffee beans were used as the starting
2228 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

material, while CGA and caffeine were used as model com- 3.4. Statistical Analysis. To determine the interaction of
pounds of polar and nonpolar compounds, respectively. SCCO2 extraction variables and their effect on the components recovery,
and water were used as solvents to extract caffeine and CGA, an analysis of variance (ANOVA) was carried out according to
respectively. Three types of extraction and separation were the Excel Statistics 2004 program. The significance level was
used. They are extraction and separation with SCCO2 and water stated at 95%, with a p-value 0.05.
in flow and batch, respectively, and extraction and separation
with both SCCO2 and water in flow co-currently and counter- 4. RESULTS AND DISCUSSION
currently, respectively. Schematic diagrams of each type of sepa-
ration apparatus are shown in Figure 2a-c. The apparatus in- On the basis of the Soxhlet extraction, maximum extractable
cludes a chiller (cooling unit CLU-33, Iwaki Asahi Techno Glass, caffeine and CGA from green coffee beans were 32.39 ( 0.2
Japan), two pumps (syringe pump model 260D, ISCO, USA, for and 54.94 ( 0.3 mg/g of coffee beans, respectively. Collected
CO2, and Intelligent Prep. pump PU-2086 Plus, Jasco, Japan, for extract in the SCCO2 and water phase were a white solid and dark
water), a heating chamber (ST-110, ESPEC Corp., Japan), an brownish-green solution, respectively.
extractor (Taiatsu, Japan), back pressure regulators (SCF-Bpg, 4.1. Extraction and Separation Using SCCO2 in Flow and
Jasco, Japan and AKICO, Japan), collection vials, and a wet gas Water in Batch Modes. In the first type of extraction and
meter (Sinagawa Co., Japan). separation, the effect of temperature, pressure, and water mass on
In the first type of extraction and separation, glass beads and the recovery of caffeine and CGA in SCCO2 and water phases
coffee beans were put in the extractor and soaked with water, was investigated to determine of optimum condition for the sepa-
while bubbling SCCO2 was flowed and exhausted from the top of ration process. Recovery of caffeine or CGA was defined as the
the extractor. Extract in the CO2 phase was collected every hour weight of caffeine or CGA extracted divided by the weight of
for 6 h, while extract in the water phase was collected at the end of caffeine or CGA extracted by Soxhlet extraction, respectively.
the extraction process. In the co-current type of extraction and The effect of extraction conditions on the recovery of caffeine
separation, SCCO2 was flowed from the top and bubbled at the and CGA in SCCO2 and water phases was investigated. Recovery
bottom of the extractor while water was flowed from the middle of caffeine and CGA in SCCO2 and water phases at various
of the extractor. Extract in the CO2 and water phase was collected extraction conditions is shown in Table 1. The effect of tempera-
at the top and bottom of the extractor every hour for 6 h, ture on the recovery of compounds in SCCO2 and water phases
respectively. In the counter-current type, SCCO2 and water were was studied at a maximum extractor pressure of 25 MPa, CO2
flowed from the bottom and top of the extractor, respectively. flow rate of 3 mL/min, and water volume of 30 mL with 4 g of raw
Extract in the CO2 and water phase was collected every hour for coffee beans. As expected, the extracted compound in SCCO2
6 h from the top and bottom of the extractor, respectively. Sepa- contained 100% purity of caffeine. Even though in some experi-
ration and extraction were carried out at a constant water flow mental conditions SCCO2 that came out from the extractor was
rate (1 mL/min) and various temperatures of 40-60 °C, pres- saturated in water, the extract did not contain CGA. In this case,
sures of 15-30 MPa, ratios of coffee and water mass of 1/30-4/30, water may act as an entrainer to increase the solubility of caffeine
height of glass beads of 0-40 cm, and CO2 flow rates (at the in SCCO2. On the other hand, the extracted compound in water
pump) of 3-5 mL/min. In the batch, co-current and counter- contained both caffeine and CGA. On the basis of the HPLC
current, the amount of whole coffee beans with an average dia- chromatogram, the contents of caffeine and CGA in the water
meter of 0.8 cm used in the extraction was 4, 12, and 25 g, respec- phase varied from 20 to 30 and from 35 to 40%, respectively,
tively. The extracts from the CO2 and water phases were weighed depending on the extraction condition. It can be noticed that
immediately after the collection and stored in a refrigerator for both caffeine and CGA are well soluble in water (2.71 g/100 mL
a day. and 25 mg/mL at 25 °C, respectively17). On the other hand, caf-
Coffee beans were also extracted by Soxhlet extraction with feine is also soluble in SCCO2 with a solubility of 0.138 g/100 mL
methanol as the solvent to determine the maximum amount of at 40 °C and 30 MPa,18 while CGA is almost not soluble in
extractable caffeine and CGA. Methanol with 150 mL of volume SCCO2.14-16 Thus, a high concentration of caffeine was found
was used to extract caffeine and chlorogenic acid from 7 g of both in the SCCO2 and water phases. However, the caffeine
coffee beans. Extract was then analyzed by HPLC to determine concentration was higher in the water phase compared to the
the caffeine and CGA content. SCCO2 phase. As shown in Table 1, recovery of caffeine in
3.3. Analytical Method. Caffeine and CGA extracted from SCCO2 significantly decreased with increasing temperature due
CO2 and water phases were analyzed by using a high perfor- to the decrease in SCCO2 density, resulting in decreasing mass
mance liquid chromatograph LC-10AD gradient system, equipped transfer and solubility of caffeine in SCCO2. The temperature
with diode array detector SPD-M10A (Shimadzu, Japan). interaction with the caffeine recovery was also confirmed by
Ten microliters of extract dissolved in methanol was injected by ANOVA with a significance level of more than 95%. This result is
SIL-10AF autosampler (Shimadzu, Japan) and separated with a in agreement with other reports19,20 that the solubility of caffeine
STR ODS II column (5 μm; 4.6
250 mm; Shinwa Chemical in SCCO2 increased with decreasing temperature at pressures
Industries, Ltd., Japan) at 40 °C. The mobile phase consisted of lower than 30 MPa. However, the increasing temperature almost
eluent A (10 mM phosphoric acid) and eluent B (acetonitrile). had no effect significantly on the composition of caffeine and
Separation of caffeine and CGA was achieved by the follow- CGA in the water phase. It might be because the increasing
ing gradient procedure: 10% of B for 5 min; a linear gradient from extractor temperature could not increase the temperature of the
10 to 70% of B within 10 min; 70% of B for 3 min; 10% of B for water, and as a result water properties and composition of extract
13 min, at a flow rate of 1.0 mL/min. The absorption spectra of did not change significantly.
caffeine and CGA were displayed between 190 and 800 nm. The effect of pressure on the recovery of compounds in
Peaks were measured at a wavelength of 270 and 325 nm to SCCO2 and water phases was investigated at a temperature of
facilitate the detection of caffeine and CGA, respectively. 40 °C, CO2 flow rate of 3 mL/min, and water volume of 30 mL
2229 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

Figure 2. Schematic diagram of simultaneous extraction and separation apparatus with SCCO2 and water as solvent. (a) Batch type; (b) co-current
type; (c) counter-current type.

with 4 g of raw coffee beans. In Table 1, increasing pressure due to the increasing SCCO2 density. The increasing pressure
promoted increasing recovery of caffeine in the SCCO2 phase also caused easy penetration of SCCO2 into the coffee beans,
2230 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

Table 1. Recovery of Caffeine in SCCO2 and Recovery of Caffeine and CGA in the Water Phase at Various Extraction Conditions
for Batch Type

temperature (°C)a pressure (MPa)b ratio of coffee and water mass (mc/mw)c

accumulation time
recovery (%) (min) 40 50 60 15 20 25 1/30 2/30 4/30

caffeine in SCCO2 60 2.55 ( 0.04 1.50 ( 0.02 1.72 ( 0.02 1.18 ( 0.02 0.72 ( 0.01 2.55 ( 0.04 3.67 ( 0.04 2.13 ( 0.03 2.55 ( 0.04
120 4.67 ( 0.05 3.30 ( 0.03 3.28 ( 0.04 2.19 ( 0.02 2.38 ( 0.04 4.67 ( 0.05 6.62 ( 0.07 5.34 ( 0.06 4.67 ( 0.05
180 6.30 ( 0.05 5.31 ( 0.06 4.64 ( 0.05 3.02 ( 0.04 3.61 ( 0.04 6.30 ( 0.05 9.42 ( 0.10 7.33 ( 0.07 6.30 ( 0.05
240 7.64 ( 0.06 6.66 ( 0.07 6.02 ( 0.06 3.79 ( 0.04 4.73 ( 0.06 7.64 ( 0.06 11.6 ( 0.12 9.70 ( 0.11 7.64 ( 0.06
300 8.69 ( 0.09 8.01 ( 0.08 7.53 ( 0.08 5.63 ( 0.08 5.83 ( 0.09 8.69 ( 0.09 13.6 ( 0.14 11.2 ( 0.11 8.69 ( 0.09
360 9.44 ( 0.12 9.34 ( 0.11 8.55 ( 0.10 5.10 ( 0.13 6.84 ( 0.11 9.44 ( 0.12 15.2 ( 0.14 12.4 ( 0.10 9.44 ( 0.12
caffeine in H2O 360 45.0 ( 0.21 44.2 ( 0.20 50.7 ( 0.20 27.4 ( 0.20 39.7 ( 0.23 45.0 ( 0.21 52.1 ( 0.20 40.1 ( 0.20 45.0 ( 0.21
CGA in H2O 360 51.3 ( 0.25 47.7 ( 0.22 56.6 ( 0.30 26.9 ( 0.19 44.8 ( 0.20 51.3 ( 0.25 79.3 ( 0.31 55.0 ( 0.30 51.3 ( 0.25
a
P = 25 MPa, mc/mw = 4/30, CO2 flow rate = 3 mL/min. b T = 40 °C, mc/mw = 4/30, CO2 flow rate = 3 mL/min. c T = 40 °C, P = 25 MPa, CO2 flow
rate = 3 mL/min.

thus increasing the mass transfer of caffeine and resulting in in- the SCCO2 phase was also still low for 6 h extraction; however,
creasing solubility of caffeine in SCCO2.17 The increasing pres- the recovery tends to increase with increasing extraction time. It
sure also promoted increasing recovery of both caffeine and indicated that a longer extraction time may increase the recovery
CGA in the water phase. High pressure of CO2 in water caused and separate the caffeine from the coffee beans.
easy penetration of both CO2 and water into the coffee beans to 4.2. Co-current Extraction and Separation using SCCO2
extract both caffeine and CGA. In addition, the increasing pres- and Water in Flow Modes. In section 4.1, extraction and
sure also causes the increasing water properties related to the separation using water in batch mode, a larger amount of water
mass transfer rate of solute to the solvent, such as density, ion yielded high caffeine recovery and separation efficiency. Further-
product, and dielectric constant.21 Moreover, a part of caffeine more, it is necessary to design continuous flow extractor column
dissolved in water was transferred to the SCCO2 phase that with longer residence time of water to collect caffeine prior to the
resulted in increasing recovery of caffeine in the SCCO2 phase. extraction with SCCO2. For this purpose, glass beads were
Separation of caffeine from CGA in the water phase seems to installed in the bottom of the extractor to accommodate water,
be higher with increasing pressure. This result indicated that which has contact with the coffee beans and extracted a part of
caffeine might be separated from CGA by increasing the pressure the caffeine and/or CGA. Furthermore, the installation of glass
of the process. beads was called nonpolar recovery section. The higher nonpolar
To understand the effect of water used in the process, the recovery section indicates the larger amount of water loaded into
ratio of coffee and water mass (mc/mw) was changed to a tem- the extractor. The height of the nonpolar recovery section was
perature of 40 °C and a pressure of 25 MPa. The effect of ratio of changed to study the effect of nonpolar recovery section height
coffee and water mass used for extraction on the recovery of on the recovery of caffeine and CGA in SCCO2 and water phases,
caffeine and CGA in SCCO2 and water phases is shown Table 1. and on the separation factor. The separation factor was defined as
In the SCCO2 phase, the increasing water mass caused an the weight of caffeine extracted in the SCCO2 phase divided by
increased caffeine recovery. It indicated that a larger amount of the weight of caffeine in SCCO2 and water phases.
water in the extractor caused decreasing mass transfer resistance Table 2 shows the effect of nonpolar recovery section height
of caffeine into both SCCO2 and water, thus increasing mass on the recovery of caffeine and CGA in SCCO2 and water phases.
transfer of caffeine from water to SCCO2, and as a result an in- The effect of nonpolar recovery section was investigated using
creasing recovery of caffeine in SCCO2. A larger amount of water 12 g of coffee beans and various glass beads weights from 0 to
might also swell the coffee beans matrix and favor penetration of 168 g depending on the height of the nonpolar recovery section.
SCCO2 inside the matrix, leading to enhancement of the solute As shown in Table 2, initially the increase in height of the
diffusion out of the plant tissue. A similar enhancement in extrac- nonpolar recovery section from 0 to 5 cm had no effect on the
tion recovery with an addition of water was observed in the recovery of caffeine; however, at a longer extraction time the
caffeine and epigallocatechin gallate extraction from green tea22 recovery of caffeine significantly increased with increasing non-
and in bioactive compounds extraction from Helianthus annuus polar recovery section. Furthermore, higher nonpolar recovery
L.23 In the water phase, recovery of CGA dramatically increased section could improve caffeine recovery, especially from 5 to
with increasing water mass, while recovery of caffeine slightly 7.5 cm. However, the increasing nonpolar recovery section
increased with increasing water mass, and as a result the increas- height from 7.5 to 10 cm slightly increased caffeine recovery at
ing separation of caffeine from CGA. It can be explained that the initial extraction time, and then dramatically increased at
water is a polar solvent that induces changes in the structure of longer extraction times. It showed that the nonpolar recovery
the cellular matrix via intracrystalline and osmotic swelling and section is necessary to enhance extraction efficiency in the
breaks solute-matrix bindings by competing with polar interac- SCCO2 phase for longer extraction time. The recovery of caffeine
tions between the matrix and the compounds, in this case CGA, and CGA in the water phase is also displayed in Table 2. As ex-
to be extracted.24 pected, higher nonpolar recovery section could enhance recovery
For all conditions, recovery of caffeine in the SCCO2 phase of both caffeine and CGA because a larger amount of water
was much lower than that in the water phase. Caffeine recovery in was available for transferring both compounds from the beans.
2231 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

Table 2. Effect of Nonpolar Recovery Section Height on the Recovery of Caffeine in SCCO2 and Recovery of Caffeine and CGA in
Water Phase for Co-current Type at 60 °C and 25 MPa

height of nonpolar recovery section (cm)

accumulation recovery (%) time (min) 0 5 7.5 10

caffeine in SCCO2 60 0.02 ( 0.00 0.00 ( 0.00 0.17 ( 0.01 0.15 ( 0.01
120 0.06 ( 0.01 0.04 ( 0.01 0.22 ( 0.03 0.25 ( 0.03
180 0.12 ( 0.02 0.12 ( 0.01 0.36 ( 0.05 0.35 ( 0.04
240 0.17 ( 0.03 0.17 ( 0.02 0.54 ( 0.06 0.62 ( 0.06
300 0.24 ( 0.04 0.22 ( 0.03 0.60 ( 0.05 0.84 ( 0.06
360 0.30 ( 0.05 0.54 ( 0.05 0.66 ( 0.05 0.92 ( 0.09
caffeine in H2O 60 3.49 ( 0.08 7.04 ( 0.10 5.05 ( 0.20 6.79 ( 0.25
120 7.64 ( 0.15 11.1 ( 0.31 9.67 ( 0.25 12.0 ( 0.25
180 11.4 ( 0.50 15.2 ( 0.50 13.2 ( 0.60 16.5 ( 0.70
240 14.6 ( 0.56 18.3 ( 0.42 16.1 ( 0.45 20.0 ( 0.50
300 17.4 ( 0.45 20.7 ( 0.60 18.4 ( 0.40 22.6 ( 0.60
360 19.6 ( 0.41 22.5 ( 0.62 20.2 ( 0.42 24.31 ( 0.60
CGA in H2O 60 2.84 ( 0.05 6.70 ( 0.10 5.94 ( 0.15 7.29 ( 0.25
120 7.42 ( 0.12 12.2 ( 0.50 12.5 ( 0.30 14.7 ( 0.50
180 12.7 ( 0.50 18.9 ( 0.43 18.4 ( 0.40 22.7 ( 0.60
240 17.9 ( 0.46 25.0 ( 0.75 23.5 ( 0.45 29.9 ( 0.70
300 22.6 ( 0.65 29.9 ( 0.72 28.0 ( 0.50 35.8 ( 0.65
360 26.8 ( 0.50 34.2 ( 0.96 31.2 ( 0.52 40.6 ( 0.75

the result the amount of caffeine dissolved in the water was large
and easy to be extracted by SCCO2. Furthermore, it is suggested
to apply a longer extractor column to increase the separation
factor. In addition, the observation of a small difference in the
separation factor might be caused by the irregularity of the water
flow rate during the process. In this case, it is necessary to design
level control in the system to control the flow rate.
For all experimental results using co-current extraction type,
caffeine recovery in the SCCO2 phase was lower than that in
batch type. It might be caused by water stored in the extractor. In
co-current extraction type, fresh water flowed continuously to
replace water containing extract collected every 60 min. While in
batch type, a certain amount of water was used for one extraction
process and coffee beans was soaked in the water during the
extraction. Thus, the content of extracted compounds in water
increased with progressing time and longer contacting time of
water with SCCO2 to dissolve extracted caffeine from the water.
Figure 3. Effect of nonpolar recovery section height on the separation Furthermore, it resulted in a higher caffeine recovery in the batch
factor for co-current type at 60 °C and 25 MPa. type.
4.3. Counter-Current Extraction and Separation Using
Moreover, a larger amount of water might decrease mass transfer SCCO2 and Water in Flow Modes. In this type of extraction
resistance and enhance the mass transfer rate due to the influence and separation, a longer extractor column was used in order to
of both intraparticle diffusion in water-soaked beans and external improve extraction and separation efficiency. Effect of tempera-
mass transfer.9 Because the migration of caffeine from water into ture, height of nonpolar recovery section and CO2 flow rate on
SCCO2 either increased with increasing nonpolar recovery the recovery of caffeine and CGA in SCCO2 and water phases,
section height, selectivity of caffeine to CGA in the water phase and separation factor was investigated. Maximum pressure used
might increase for higher nonpolar recovery section. for the extraction was 30 MPa based on the maximum pressure of
The effect of nonpolar recovery section height on the separa- the extractor.
tion factor is shown in Figure 3. High separation factor indicates Table 3 shows the effect of temperature on the recovery of
high separation efficiency of caffeine from coffee beans and the caffeine and CGA in SCCO2 and water phases at 3 mL/min of
water phase. Even though a small difference in the separation CO2 flow rate, 1 mL/min of water flow rate, and 0 cm of non-
factor was observed, the separation factor has a tendency to be polar recovery section height. Recovery of caffeine both in SCCO2
increased with increasing height of nonpolar recovery section. It and water phases, and recovery of CGA in water phase increased
can be explained that at a higher nonpolar recovery section, the with an increase in temperature. The increasing temperature might
amount of water accommodated in the extractor was large, and as cause increasing temperature of water and vapor pressure of both
2232 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

Table 3. Recovery of Caffeine in SCCO2 and Recovery of Caffeine and CGA in the Water Phase at Various Extraction Conditions
for Counter-Current Type

temperature (oC)a nonpolar recovery section height (cm)b CO2 flow rate (mL/min)c

accumulation time
recovery (%) (min) 40 60 80 0 20 30 40 3 5

caffeine in SCCO2 60 0.12 ( 0.01 0.55 ( 0.03 0.70 ( 0.02 0.55 ( 0.03 0.65 ( 0.02 0.47 ( 0.02 0.66 ( 0.03 0.28 ( 0.02 0.27 ( 0.02
120 0.34 ( 0.03 1.12 ( 0.07 1.42 ( 0.10 1.12 ( 0.07 1.13 ( 0.05 1.31 ( 0.04 1.74 ( 0.04 0.54 ( 0.04 0.70 ( 0.05
180 0.57 ( 0.05 1.83 ( 0.07 2.22 ( 0.09 1.83 ( 0.07 1.96 ( 0.06 1.96 ( 0.04 2.84 ( 0.09 0.80 ( 0.06 1.07 ( 0.07
240 0.91 ( 0.10 2.32 ( 0.10 2.85 ( 0.10 2.32 ( 0.10 2.66 ( 0.08 2.89 ( 0.06 3.47 ( 0.08 1.25 ( 0.08 1.54 ( 0.09
300 1.07 ( 0.09 3.15 ( 0.15 3.42 ( 0.14 3.15 ( 0.15 3.28 ( 0.75 3.48 ( 0.09 3.95 ( 0.10 1.64 ( 0.05 2.00 ( 0.08
360 1.16 ( 0.07 3.61 ( 0.09 3.50 ( 0.10 3.61 ( 0.09 3.84 ( 0.08 3.74 ( 0.10 4.25 ( 0.13 1.85 ( 0.07 2.41 ( 0.09
caffeine in H2O 60 1.59 ( 0.08 5.10 ( 0.10 7.69 ( 0.20 5.10 ( 0.10 3.52 ( 0.10 2.64 ( 0.05 2.61 ( 0.06 2.55 ( 0.05 1.99 ( 0.07
120 4.04 ( 0.15 11.8 ( 0.35 16.0 ( 0.70 11.8 ( 0.35 8.94 ( 0.12 7.64 ( 0.12 5.91 ( 0.13 9.14 ( 0.07 7.25 ( 0.09
180 6.27 ( 0.25 18.4 ( 0.50 20.7 ( 0.55 18.4 ( 0.50 14.4 ( 0.32 12.8 ( 0.25 10.6 ( 0.20 14.9 ( 0.10 13.4 ( 0.11
240 8.29 ( 0.32 22.2 ( 0.55 23.1 ( 0.60 22.2 ( 0.55 18.0 ( 0.45 16.4 ( 0.30 14.5 ( 0.40 19.0 ( 0.25 17.9 ( 0.20
300 10.1 ( 0.25 22.7 ( 0.60 24.3 ( 0.50 22.7 ( 0.60 21.3 ( 0.50 19.0 ( 0.40 16.8 ( 0.40 21.8 ( 0.50 20.9 ( 0.50
360 11.8 ( 0.30 24.3 ( 0.40 24.9 ( 0.45 24.3 ( 0.40 22.8 ( 0.60 20.6 ( 0.55 18.4 ( 0.50 23.6 ( 0.50 22.7 ( 0.45
CGA in H2O 60 1.54 ( 0.09 6.69 ( 0.10 12.7 ( 0.50 6.69 ( 0.10 5.41 ( 0.15 4.67 ( 0.12 4.06 ( 0.07 4.06 ( 0.07 2.74 ( 0.05
120 4.32 ( 0.11 18.1 ( 0.60 29.8 ( 0.85 18.1 ( 0.60 16.4 ( 0.40 15.6 ( 0.30 13.6 ( 0.25 16.0 ( 0.10 12.5 ( 0.12
180 7.32 ( 0.35 30.8 ( 0.90 41.7 ( 0.92 30.8 ( 0.90 28.2 ( 0.80 26.7 ( 0.90 25.4 ( 0.88 28.0 ( 0.50 25.1 ( 0.50
240 10.6 ( 0.50 39.7 ( 0.85 49.4 ( 0.90 39.7 ( 0.85 37.1 ( 0.90 36.2 ( 1.05 34.5 ( 0.90 37.0 ( 1.2 35.9 ( 1.50
300 13.7 ( 0.45 46.2 ( 0.80 54.9 ( 1.05 46.2 ( 0.80 43.2 ( 1.25 43.4 ( 1.20 42.2 ( 1.10 43.5 ( 1.30 43.9 ( 1.60
360 16.8 ( 0.50 51.7 ( 1.70 57.9 ( 1.75 51.7 ( 1.70 47.8 ( 1.50 48.9 ( 1.20 47.7 ( 1.10 48.8 ( 1.50 49.7 ( 1.20
a
P = 30 MPa, CO2 flow rate = 5 mL/min, H2O flow rate = 1 mL/min, nonpolar recovery section height = 0 cm. b T = 60 °C, P = 30 MPa, CO2 flow rate =
5 mL/min, H2O flow rate = 1 mL/min. c T = 60 °C, P = 30 MPa, H2O flow rate = 1 mL/min, nonpolar recovery section height = 30 cm.

caffeine in SCCO2 and resulted in higher separation of caffeine

from CGA in the water phase.
As resulted in the co-current type, at a higher nonpolar re-
covery section, higher separation efficiency was obtained. In the
counter-current type, nonpolar recovery section height was
changed to obtain a higher separation efficiency. The nonpolar
recovery section height was changed from 0 to 40 cm by installa-
tion of glass beads in the bottom of the extractor. The effect of
nonpolar recovery section height on the recovery of caffeine in
the SCCO2 phase and the recovery of caffeine and CGA in water
phase are shown in Table 3. Recovery of caffeine in the SCCO2
phase significantly increased with increasing height of nonpolar
recovery section. As expected, the increasing nonpolar recovery
section height could increase caffeine recovery because caffeine
dissolved in water was easily transferred to the SCCO2 phase. On
the contrary, caffeine recovery in the water phase significantly
decreased with increasing nonpolar recovery section height due
Figure 4. Effect of temperature on the separation factor for counter- to migration of caffeine from water to SCCO2 phase. The same
current type at 30 MPa, 3 mL/min of CO2 flow rate, 1 mL/min of water trend either could be observed for CGA recovery; however, CGA
flow rate, and 30 cm of nonpolar recovery section. recovery did not significantly change by changing the nonpolar
recovery section height from 20 to 40 cm. It might be because the
caffeine and CGA that resulted in increasing recovery of caffeine in coffee beans were not soaked well with water on the glass beads
SCCO2 and both components in water. Moreover, high tempera- installed at the bottom of extractor. As a comparison with the co-
ture might break the plant cells of the coffee beans that caused easy current type, recovery of caffeine in SCCO2 increased more than
penetration of SCCO2 and water into the cells. Because high con- five fold, while recovery of caffeine in the water phase decreased
centration of caffeine was available in water, caffeine was easily about 25% at the highest nonpolar recovery section. It indicates
transferred into the SCCO2 phase. that the counter-current type is more effective than the co-
Evaluation of temperature effect on the separation factor is current type for separation of caffeine from CGA. The effect of
shown in Figure 4. The separation factor slightly increased with nonpolar recovery section height on the separation factor is
increasing temperature; however, the highest separation factor shown in Figure 5. Separation factor of the process dramatically
was obtained at 60 °C for 6 h of extraction time. As explained increased with the increasing nonpolar recovery section height,
before, the increasing temperature caused increasing dissolved especially from 30 to 40 cm. It is suggested that higher nonpolar
2233 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235
Industrial & Engineering Chemistry Research ARTICLE

5. CONCLUSION
Extraction and separation of caffeine and CGA from raw coffee
beans using SCCO2 and water were studied. Three types of
extraction and separation modes were used in this work. Extrac-
tion and separation were conducted in three types of extractor
columns with different sizes and flow types of CO2 and water.
The effect of temperature and pressure in the separation mode
using SCCO2 in flow and water in batch mode was reported.
Recovery of caffeine in the SCCO2 phase increased with decreas-
ing temperature and increasing pressure. However, only the
increasing pressure could promote the increasing recovery of
caffeine and CGA in the water phases. The change of height of
the nonpolar recovery section on the recovery of caffeine and
CGA in the SCCO2 and water phases and on the separation
factor was studied using SCCO2 and water in co-current and
counter-current flow modes. Recovery of caffeine and separation
factor of caffeine from CGA significantly increased with increas-
Figure 5. Effect of nonpolar recovery section height on separation ing height of the recovery section. The effect of temperature in
factor for counter-current type at 60 °C and 30 MPa.
counter-current type of separation on the recovery of caffeine
and CGA in SCCO2 and water phases exhibited different trends
with the batch type of separation. Even though the recovery of
caffeine in SCCO2 and separation efficiency obtained in this
work was very low (less than 10%), the proposed simultaneous
extraction and separation process can be developed further to
obtain maximum efficiency. In addition, it is necessary to employ
another substance as a nonpolar model compound. On the basis
of the results of this work, it is suggested to design longer
extractor columns with higher nonpolar recovery sections in
counter-current type in order to increase the separation effi-
ciency. Moreover, it is better to employ substances from natural
materials with high polarity and nonpolarity in this process.

’ AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Phone: þ81-96-342-3664.
Fax: þ81-96-342-3665.
Figure 6. Effect of CO2 flow rate on the separation factor for counter-
current type at 60 °C and 30 MPa. ’ ACKNOWLEDGMENT
This work was supported by Kumamoto University Global
recovery section resulted in higher separation efficiency of caf- COE Program “Global Initiative Center for Pulsed Power Engi-
feine from the coffee beans. Moreover, higher nonpolar recovery neering” and Japan Society for the Promotion of Science (JSPS).
section caused a higher concentration of CGA to remain in the
coffee beans. The separation factor of counter-current type was ’ REFERENCES
5-fold higher than the co-current type due to the longer con- (1) Mukhopadhyay, M. Natural Extracts Using Supercritical Carbon
tacting time between SCCO2 and water in the higher nonpolar Dioxide; CRC Press LLC: Boca Raton, FL, 2000.
recovery section and longer extractor column. (2) Lack, E.; Seidlitz, H. Decaffeination Commercial Scale of Coffee
In order to obtain a higher separation efficiency, the CO2 flow and Tea using Supercritical CO2. In Extraction of Natural Products Using
rate was changed from 3 to 5 mL/min. As shown in Table 3, Nearcritical Solvents; King, M. B., Bott, T. R.; Blackie Academic &
caffeine recovery in the SCCO2 phase significantly increased with Professional: Birmingham, 1993; pp 101-140.
an increase in the CO2 flow rate, especially at the end of extrac- (3) Ky, C.; Louarn, J.; Dussert, S.; Guyot, B.; Hamon, S.; Noirot, M.
tion. It can be explained that the increasing CO2 flow rate caused Caffeine, Trigonelline, Chlorogenic Acids and Sucrose Diversity in Wild
an increasing number of CO2 molecules to penetrate and diffuse Coffee arabica L. and C. canephora P. Accessions. Food Chem. 2001, 75, 223.
into beans to extract caffeine. On the other hand, the change of (4) Flores, G. B.; Andrade, F.; Lima, D. R. Can Coffee Help Fighting
the Drug Problem? Preliminary Results of a Brazilian Youth Drug Study.
CO2 flow rate had no significant effect on the recovery of both
Acta Pharmacol. Sin. 2000, 21, 1059.
caffeine and CGA in the water phase. It indicated that the number (5) De Paulis, T.; Commers, P.; Farah, A.; Zhao, J.; McDonald,
of CO2 molecule did not effect the interaction between caffeine M. P.; Galici, R.; Martin, P. R. 4-Caffeoyl-1,5-Quinide in Roasted Coffee
and CO2 molecules. Because caffeine recovery in the SCCO2 Inhibits [3H] Naloxone Binding and Reverses Antinociceptive Effects of
phase increased, the separation factor of the process increased Morphine in Mice. Psychopharmacology. 2004, 176, 146.
(Figure 6), which indicated the increasing separation efficiency of (6) Ramalakshmi, K.; Bhagavan, B. Caffeine in Coffee: Its Removal.
caffeine from the coffee beans. Why and How?. Crit. Rev. Food Sci. Nutr. 1999, 39, 441.

2234 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235

Industrial & Engineering Chemistry Research ARTICLE

(7) Balyaya, K. J.; Clifford, M. N. Individual Chlorogenic Acids and

Caffeine Contents in Commercial Grades of Wet and Dry Processed
Indian Green Robusta Coffee. J. Food Sci. Technol. 1995, 32, 104.
(8) McHugh, M.; Krukonis, V. Supercritical Fluid Extraction Princi-
ples and Practice; Butterworth-Heinemann: Boston, 1994.
(9) Peker, H.; Srinivasan, M. P.; Smith, J. M.; McCoy, B. J. Caffeine
Extraction Rates from Coffee Beans with Supercritical Carbon Dioxide.
AIChE J. 1992, 38, 761.
(10) Farah, A.; de Paulis, T.; Moreira, D. P.; Trugo, L. C.; Martin,
P. R. Chlorogenic Acids and Lactones in Regular and Water-Decaffei-
nated Arabica Coffees. J. Agric. Food Chem. 2006, 54, 374.
(11) Daglia, M.; Papetti, A.; Gregotti, C.; Berte, F.; Gazzani, G. In
Vitro Antioxidant and Ex Vivo Protective Activities of Green and
Roasted Coffee. J. Agric. Food Chem. 2000, 48, 1449.
(12) Del Castilho, M. D.; Ames, J. M.; Gordon, M. H. Effect of
Roasting on the Antioxidant Activity of Coffee Brews. J. Agric. Food
Chem. 2002, 50, 3698.
(13) de Azevedo, A. B. A.; Mazzafera, P.; Mohamed, R. S.; de Melo,
S. A. B. V.; Kieckbusch, T. G. Extraction of Caffeine, Chlorogenic Acids
and Lipids from Green Coffee Beans using Supercritical Carbon Dioxide
and Co-solvents. Braz. J. Chem. Eng. 2008, 25, 543.
(14) Clifford, M. N. Chlorogenic Acids. In Coffee: Chemistry; Clarke,
R. J., Macrae, R.; Elsevier Applied Science Publishers Ltd.: New York,
1985; pp 89-103.
(15) Brunner, G. Gas Extraction an Introduction to Fundamentals of
Supercritical Fluids and the Applications to Separation Processes; Steinkopff
Darmstadt Springer: New York, 1994.
(16) Taylor, L. Supercritical Fluid Extraction; John Wiley & Sons Inc.:
New York, 1996.
(17) Farah, A.; Monteiro, M.; Donangelo, C. M.; Lafay, S. Chloro-
genic Acids from Green Coffee Extract are Highly Bioavailable in
Humans. J. Nutr. 2008, 138, 2309.
(18) Burgos-Solorzano, G. I.; Brennecke, J. F.; Stadtherr, M. A.
Solubility Measurements and Modeling of Molecules of Biological and
Pharmaceutical Interest with Supercritical CO2. Fluid Phase Equilib.
2004, 220, 57.
(19) Li, S.; Varadarajan, G. S.; Hartland, S. Solubilities of Theobro-
mine and Caffeine in Supercritical Carbon Dioxide: Correlation with
Density-Based Models. Fluid Phase Equilib. 1991, 68, 263.
(20) Saldana, M. D. A.; Mohamed, R. S.; Baer, M. G.; Mazzafera, P. J.
Extraction of Purine Alkaloids from Mate (Ilex paraguariensis) using
Supercritical CO2. J. Agric. Food Chem. 1999, 47, 3804.
(21) Kitada, K.; Machmudah, S.; Sasaki, M.; Goto, M.; Nakashima,
Y.; Kumamoto, S.; Hasegawa, T. Antioxidant and Antibacterial Activity
of Nutraceutical Compounds from Chlorella vulgaris Extracted in
Hydrothermal Condition. Sep. Sci. Technol. 2009, 44, 1228.
(22) Kim, W.-J.; Kim, J.-D.; Kim, J.; Oh, S.-G.; Lee, Y.-W. Selective
Caffeine Removal from Green Tea using Supercritical Carbon Dioxide
Extraction. J. Food Eng. 2008, 89, 303.
(23) Casas, L.; Mantell, C.; Rodriguez, M.; Torres, A.; Macias, F. A.;
de la Ossa, E. M. Effect of the Addition of Cosolvent on the Supercritical
Fluid Extraction of Bioactive Compounds from Helianthus annuus L.
J. Supercrit. Fluids. 2007, 41, 43.
(24) Erkucuk, A.; Akgun, I. H.; Yesil-Celiktas, O. Supercritical CO2
Extraction of Glycosides from Stevia rebaudiana Leaves: Identification
and Optimization. J. Supercrit. Fluids. 2009, 51, 29.

2235 dx.doi.org/10.1021/ie101252w |Ind. Eng. Chem. Res. 2011, 50, 2227–2235