Industrial Crops & Products 141 (2019) 111723

Contents lists available at ScienceDirect

Industrial Crops & Products

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

Enhanced extraction of spent coffee grounds oil using high-pressure CO2 T

plus ethanol solvents
Micheli Nolasco Araújo, Ana Queren Paladonai Leandro Azevedo, Fabiane Hamerski,
⁎
Fernando Augusto Pedersen Voll, Marcos Lúcio Corazza
Department of Chemical Engineering, Federal University of Paraná, CEP 81531-990, Curitiba, PR, Brazil

A R T I C LE I N FO A B S T R A C T

Keywords: This work reports the extraction of spent coffee grounds (SCG) oil using supercritical CO2 plus ethanol
Supercritical carbon dioxide (scCO2+EtOH) solvent in semi-batch process as an alternative approach to the use of organic solvent extraction.
CO2-expanded ethanol In addition, the results are compared to supercritical carbon dioxide (scCO2) and pressurized ethanol (EtOH).
Pressurized ethanol Soxhlet was performed and used as a standard extraction method. The extraction yield, fatty acid profile, total
Spent coffee grounds
phenolic content (TPC), antioxidant activity (AA by ABTS and DPPH methods), phenolic compounds and caf-
Biorefinery
feine content were determined, as well the overall extraction curves. The highest extraction yield of 15.9% was
obtained using scCO2+EtOH using an ethanol to SCG ratio mass of (2:1), 80 ºC, 20 MPa in 25 min of extraction
time. Major fatty acids present in the oil were linoleic (45%) and palmitic (31%). The oil obtained with
scCO2+EtOH (2:1) presented high TPC values (294.47 to 392.96 mg GAE/100 g oil), similar to those obtained
by pressurized ethanol extraction. High values of AA by ABTS and DPPH methods were also found in the oil from
scCO2+EtOH extraction. The main phenolic acids found were dihydroxybenzoic acid and caffeic acid. The
highest concentrations of dihydroxybenzoic (17.66 mg/100 g oil) and caffeic acid (9.36 mg/100 g oil) were
obtained with scCO2+EtOH (2:1) at 80 ºC and 10 MPa. The highest caffeine content (711.70 mg/100 g oil) was
obtained with scCO2+EtOH (0.5:1) at 60 ºC and 15 MPa. The results obtained in this work provided higher
extraction yield when compared to scCO2 and similar to compressed ethanol, however in a shorter extraction
time and using lesser organic solvent amounts. The results are promising and demonstrate the technical feasi-
bility of SCG oil extraction using scCO2+EtOH to obtain valuable a product from an agricultural and urban
waste biomass.

1. Introduction be interesting for food, pharmaceutical and cosmetic industry applica-

tions. However currently, the SCG are being incinerated, generating
Coffee is the most consumed worldwide beverage after water and greenhouse gases and/or allocated into landfills. Thus, SCG have been
the second largest commodity in stock exchange after petroleum investigated as a source of value-products to reduce its environmental
(Akgün et al., 2014; Andrade et al., 2012; Mata et al., 2018). According impact (Mata et al., 2018; Panusa et al., 2013).
to International Coffee Organization (ICO) approximately 9.4 million of Considering the quantity and diversity of products that can be ob-
tons were produced globally in 2018 (ICO, 2018). During the coffee tained from SCG, some authors have studied the feasibility of im-
production and processing huge quantities of waste, such as skins and plementing a biorefinery with SCG as feedstock. According to Karmee
spent coffee grounds (SCG) are generated. Around 650 kg of SCG are (2018) and Mata et al. (2017), a biorefinery based on spent coffee
generated by 1000 kg of green coffee beans processed (Karmee, 2018). grounds can be a feasible strategy by using a combination of compatible
SCG contain many organic compounds as fatty acids, amino acids, chemical, biotechnological, and thermochemical methods, where bio-
lignin, cellulose, hemicellulose, and other polysaccharides, which can diesel production (Abdullah and Bulent Koc, 2013; Al-Hamamre et al.,
be processed and add value to this waste (Campos-Vega et al., 2015). 2012; Couto et al., 2009; Döhlert et al., 2015; Kwon et al., 2013; Rocha
Besides that, it also contains bioactive compounds, particularly alka- et al., 2014), extractions of antioxidants to the pharmaceutical and food
loids and polyphenols, and caffeine (major alkaloid), tannins, flavanols, industry (Ahangari and Sargolzaei, 2013; Andrade et al., 2012;
flavones and phenolics acids. Thus, recovering these compounds might Benavides et al., 2016; Mussatto et al., 2011; Page et al., 2017; Shang

⁎
Corresponding author at: Department of Chemical Engineering, Federal University of Paraná, PO Box 19011 Polytechnic Center, Curitiba 81531-980, PR, Brazil.
E-mail address: corazza@ufpr.br (M.L. Corazza).

https://doi.org/10.1016/j.indcrop.2019.111723
Received 5 June 2019; Received in revised form 21 August 2019; Accepted 23 August 2019
Available online 31 August 2019
0926-6690/ © 2019 Elsevier B.V. All rights reserved.
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

et al., 2016; Xu et al., 2015), bioethanol (Kwon et al., 2013; Rocha Pressurized liquid extraction (PLE) is an attractive extraction tech-
et al., 2014), pellets (Zuorro and Lavecchia, 2012) and tannins (Low nique that can promote high efficient extraction in short extraction time
et al., 2015), are some examples of by-products that can be obtained using low relatively low amounts of organic liquid solvents (Oliveira
from SCG. et al., 2018). Currently, PLE has been applied to green coffee (Oliveira
The crude oil is the most economically valuable and easily re- et al., 2018), wild almond (Balvardi et al., 2015), soybean (Rodrigues
covered compound from SCG (Campos-Vega et al., 2015). For example, et al., 2017) and spent coffee grounds (Shang et al., 2016). The main
there is an increasing interest to produce biodiesel from the residual oil advantages of PLE is related to the use of high temperatures that pro-
obtained with spent coffee grounds as a feedstock, since this is a sus- motes high solubility of solutes in the solvent by increasing the diffu-
tainable practice for waste reduction and a low-cost source of fatty sional rate of the solute to the solvent and breaking the strong inter-
acids (Al-Hamamre et al., 2012). SCG have potential to biodiesel pro- action solute-matrix interactions, such as van der Waals bonds, and
duction since its oil content varies from 12 to 18.3% (Andrade et al., solute molecules and active sites in the matrix, such hydrogen and di-
2012; Couto et al., 2009) and some authors report 85.5% (Al-Hamamre pole-dipole interactions (Oliveira et al., 2018).
et al., 2012) and 98.61% (Liu et al., 2017) of converted coffee oil to A gas-expanded liquid (GXL) technique is a mixture of a compres-
biodiesel by alkali-catalyzed esterification and direct acid transester- sible gas dissolved in a liquid solvent. The CO2-expanded liquids (CXL)
ification (in situ), respectively. Assuming an oil content of 15% and 92% are the mostly common class of GXL used, due to the operational ad-
of converted coffee oil to biodiesel, the production could be approxi- vantages of the CO2 already cited (Jessop and Subramaniam, 2007).
mately 0.85 million tons of biodiesel by year. Furthermore, SCG oil The CXL is a solvent with gas and liquid characteristics, where CO2
quality can be enhanced to applications in cosmetic and pharmaceutical improves the gas solubility and mass transfer, while liquid organic
industry or use it as source of other valuable compounds as caffeine, solvents increase the solubility of liquid and solid solutes (Herrero
sterols, terpenes and tocopherols (Campos-Vega et al., 2015). Caffeine et al., 2017; Siougkrou et al., 2014). In general, CXL can be considered a
is the most studied compound from coffee due its psychoactive effects switchable solvent technique found midway between pressurized li-
and the promotion of energetic metabolism (Panusa et al., 2013). quids and compressed CO2 by increasing the amount of CO2 (Herrero
Moreover, phenolics compounds from coffee oil have been studied due et al., 2017). The CXL has been successfully applied to the extraction of
their beneficial effects to human health, as its protection against Moringa oleifera (Rodrígues et al., 2016), where the authors reached a
chronic degenerative disease (cataracts, macular degeneration, neuro- extraction yield two-fold higher using 50% of ethanol in the CXL
degenerative diseases, and diabetes mellitus) (Ballesteros et al., 2017) (ethanol and CO2) than with pure scCO2. The main advantages of CXL
and its other bioactivities properties like antioxidant, anti-bacterial, compared to scCO2 is the possibility of using lower pressure operations,
antiviral, anti-inflammatory and anti-carcinogenic activities (Barbosa and when compared to PLE the use of lesser solvent amounts
et al., 2014). (Rodrígues et al., 2016; Toda et al., 2016). However, to the best of our
Oil extraction is usually performed using organic solvents, however knowledge, a semi-batch technique using CO2-expanded solvent to re-
the environment safety rules and the increased public health risk are covery the oil from SCG has not been reported in the literature.
pushing the industry to seek for alternative extraction methods Therefore, this study aimed to evaluate the influence of
(Ahangari and Sargolzaei, 2013). Supercritical fluid extraction (SFE), scCO2+EtOH (CXL) extraction parameters and compare the semi-batch
pressurized liquid extraction (PLE), and gas-expanded liquid extraction CXL methodology with other green solvents techniques, such as pure
(GXL) have been considered as green technology and efficient methods scCO2 and pressurized ethanol, in terms of extraction yields, overall
to extract compounds from natural matrices due some advantages, such extraction curves (kinetics of extraction) and chemical profiles of the
as, elimination of large amount of organic solvents, solvent selectivity extracted oil from spent coffee grounds. Fatty acid (FA) profile, anti-
by changing the operational conditions to enhanced the recovery a class oxidant activity (AA), total phenolic content (TPC), phenolics com-
of specific compounds (Jessop and Subramaniam, 2007; Reverchon and pounds and caffeine were also evaluated in order to assess the oil
De Marco, 2006; Shang et al., 2016). qualify and its potential commercial application.
Extraction of compounds from natural matter is the most studied
application of supercritical fluids (SCF). SFE has immediately ad- 2. Materials and methods
vantages over traditional extraction methods, such as processing flex-
ibility and elimination of solvent from the extracts in a post-processing 2.1. Samples
stage (Reverchon and De Marco, 2006). SFE is generally performed
with carbon dioxide (CO2) due to its low critical pressure (74 bar) and Spent coffee ground (SCG) samples were obtained in a local coffee
temperature (32 °C), relative non-toxicity, non-flammability, avail- shop (Zaraffa Coffee Shop, Curitiba, Brazil) with initial moisture and
ability in high purity at relativity low cost and easy removal from the volatile compounds of 65.19 ± 0.89%. The raw material was dried in
extracts (Pourmortazavi and Hajimirsadeghi, 2007). In SFE, the addi- air circulation oven at 40 °C for 48 h until constant moisture and vo-
tion of a liquid solvent can improve the extraction efficiency by in- latile compounds content of 4.94 ± 0.29%. The average particle size
creasing the solubility of the solute (Pereira and Meireles, 2010). Couto was estimate with the method presented by Gomide (Gomide, 1983)
et al. (2009) performed SFE of lipids using supercritical carbon dioxide using Tyler series sieves in a vertical vibratory sieve shaker, where the
(scCO2) at different pressures (15.0, 20.0, 25.0 and 30.0 MPa) and milled material retained in each sieve was: mesh 14 (8.05 ± 0.45%),
temperatures (40, 50 e 55 °C) aiming the subsequent biodiesel pro- mesh 24 (15.86 ± 0.04%), mesh 28 (14.32 ± 0.20%), mesh 35
duction. The highest yield was 15.4% obtained at 25.0 MPa and 50 °C (14.68 ± 2.19%), mesh 48 (29.81 ± 5.46%) and mesh 80
after 3 h of extraction. Moreover, those authors (Couto et al., 2009) (16.38 ± 2.52%). These fractions were blended and packed in plastic
evaluated the addition of ethanol as co-solvent corresponding to an bags and stored at 4 °C until the use.
ethanol to CO2 mass ratio of 6.5:93.5 (w/w) at 50 °C and 20.0 MPa,
which resulted in a reduction of 60% on the extraction time and the 2.2. Determination of moisture and volatile compounds content and particle
total amount of solvent required to achieve an oil extraction yield of density
12.9% at the same condition using pure scCO2. Similar results were
reported by Andrade et al. (2012), who studied the SCG oil extraction Moisture and volatile compounds content (M) were determined by
with scCO2 and scCO2 + co-solvent, at 40, 50 and 60 °C and 10 to automatic infrared moisture analyzer method (Gehaka, model IVT 200),
30 MPa during 2.5 h. The highest yield from SFE extraction was determined in triplicate according to National Renewable Energy
10.5 ± 0.2%, while the highest among the extractions yield was Laboratory (NREL) methodology (Sluiter et al., 2008) using Eq. 1,
15 ± 2%, obtained by solid-liquid extraction with ethanol. where minitial is the SCG mass before drying (g) and mfinal is the SCG

2
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

mass after drying (g). Real density was measured using a helium the confinement time, the dynamic extraction step with a solvent flow
pycnometer (Quantachrome Ultrapyc 1200e), at the Analytical Central- rate around 2.0 mL/min was started, and the extract oil samples were
Institute of Chemistry/Unicamp, Campinas, Brazil. The apparent den- collected in test tubes at predetermined times to obtain the kinetic
sity was calculated as a direct relation between the total SCG sample curves. Samples from the extraction with scCO2+EtOH and pressurized
mass and the extractor vessel volume. ethanol were placed into an air circulation oven (IKA, Nova Ética,
minitial − mfinal model 400-2) at 70 °C until constant weight. After that, the extracts
M= × 100 mass was measured and stored in amber vials under refrigeration at
m initial (1)
approximately -4 °C. In order to obtain the extraction time and the mass
of SCG oil extracted in each run, the extraction was considered finished
2.3. Soxhlet extraction when the amount of oil extracted in the test tube was equivalent to 1%
of the total amount of extracted oil measured in the previous test tubes.
Soxhlet extractions were used as a standard method to provide a The overall extraction yields were also calculated using Eq. 2.
comparation with the extraction with pressurized solvents (CLX, SFE
and PLE). These extractions were carried out using 5 g of dried SCG and 2.5. Fatty acid profile
150 mL of solvent, where the solvents used were n-hexane (Neon,
99.5% purity), ethyl acetate (Neon, 99.5% purity) and ethanol (Neon, Fatty acids (FA) profile of the SCG oil samples was performed in a
99.8% purity). Oil extracted was dried in a rotary vacuum evaporator Shimadzu chromatograph (GC 2010 Plus), with a polar capillary
(IKA, Model RV 10 digital). After, it was dried using an air circulation column (Shimadzu, model SH-Rtx-Wax, 30 m x 0.32 mm x0.25 μm),
oven (IKA, Nova Ética, model 400-2) until constant mass at 60 °C for n- flame ionization detector (FID) at 250 °C and split injection mode (1:10)
hexane, 65 °C for ethyl acetate (EtAc) and 70 °C for ethanol. In addition, at 240 °C. The oven temperature was programmed to start at 100 °C,
the oil was stored in amber flasks and kept under refrigeration at -4 °C. and it was maintained for 5 min. Then, it was increased from 100 °C to
The extraction yields were calculated using Eq. 2. 240 °C at a rate of 4 °C/min, and maintained at 240 °C for 5 min. The
Mass of extracted oil (g ) carrier gas was helium at 32.5 cm3/min. The samples were prepared in
Yield ( %) = × 100 agreement to official method Ce 2–66 to convert the lipids into fatty
Mass of spent coffee grounds sample (g ) (2)
acid methyl esters (FAMEs) (AOCS, 1997). FAMEs were identified by
comparison with retention times of the standard mixture FAMEs (Su-
2.4. Extraction with compressed fluids pelco, MIX FAME 37, Bellefonte, PA, USA). The quantification of fatty
acid was conducted by area normalization procedure. Results were
Extraction with compressed fluids was performed in a laboratory expressed as a percentage of each individual fatty acid present in the
scale unit that has been presented in previous works (Fetzer et al., sample.
2018), in which a few modifications were performed for the PLE ex-
tractions. Basically, the experimental setup consisted of a solvent re- 2.5.1. Iodine value (IV) and saponification value (SV)
servoir (CO2 cylinder, for SFE and CXL extractions; or ethanol bottle, for The iodine value (IV) and the saponification value (SV) of the SCG
PLE extractions), a thermostatic bath (Nova Ética, Sppencer, model oil were determined using the fatty acid profile according to AOCS
521-5D) to maintain the temperature of the high pressure syringe pump methods Cd 1c-85 and Cd 3a-94. The IV is a measure of the average
(ISCO, model 500D, Lincoln, NE 68504, USA) at 10 °C, a jacketed ex- amount of unsaturated fats and oils, expressed in g of iodine absorbed
tractor vessel with internal volume of 62.4 cm3 (length 22 cm and per 100 g of sample (g I2/100 g oil) (Knothe, 2002). The IV values were
diameter of 1.9 cm) coupled to a thermostatic bath (Quimis, model calculated with Eq. 3, using the percentages (in mass basis) of the fatty
Q214 s), which is used to keep the extractor temperature constant, and acid found in SCG oil.
a micrometric needle valve o control the solvent mass flow.
Different compressed solvents (scCO2, scCO2+EtOH and pressur- IV = ( % hexadecenoic acid × 0.99760) + ( % octadecenoic acid × 0.8986)
ized EtOH) were used in this study in order to assess the solvent effect + ( % octadienoic acid × 1.810) + ( % octadecatrienoic acid × 2.735)
in the spent coffee oil recovery from SCG. Suppliers and purity of
+ ( % eicosenoic acid × 0.8175) + ( % docosenoic acid × 0.7497)
chemicals used in the extractions at high pressure were: CO2, White
(3)
Martins S.A., 99.5% purity; and ethanol, Neon, 95% and 99.8% purity.
The extractions were carried out at temperatures of 40, 60 and 80 °C The saponification value (SV) is defined as the amount of alkali
and pressures of 10, 15 and 20 MPa. For extractions with scCO2 and needed to saponify a defined sample weight. It is expressed in mg of
pressurized EtOH the extractor was loaded with approximately 17 g of potassium hydroxide per g of sample and specifies the total fatty acid
SCG. For scCO2+EtOH (CXL) extractions, the ethanol amount related to content in oil (mg KOH/g oil). The average molecular weight of fatty
the specified ethanol to spent coffee grounds mass ratio (MRES) was acids in a lipid system, as well as the number of ester bonds per gram of
mixed with approximately 17 g spent coffee grounds and then the sample, can be derived from SV (Al-Hamamre et al., 2012). The SVs
mixture (ethanol and SCG) was loaded into the extractor. The MRES were calculated with Eq. 4, where mmw is the mean molecular weight
(gethanol:gSCG) studied were 0.25:1 (4.25 g of EtOH and 17 g of SCG), (sum fractional), 56.1 is the molecular weight of KOH and 92.09 mo-
0.5:1 (8.50 g of EtOH and 17 g of SCG), 1:1 (17 g of EtOH and 17 g of lecular weight of glycerol.
SCG) and 2:1 (34 g of EtOH and 17 g of SCG). To prevent solid particles 3 × 56.1 × 1000
from entering the piping system, the extremities of the extractor were SV =
[(mmw × 3) + 92.09] − (3 × 18) (4)
coated with cotton.
All the extractions were performed in two distinct steps defined as
static and dynamic extraction stages. The solvent (scCO2, for SFE and 2.6. Total phenolic content (TPC)
CXL extractions, or EtOH, for PLE extractions) was slowly added to the
extractor until the desired pressure was reached, and the temperature Total phenolic content (TPC) in the SCG oil samples obtained by the
was previously set to the extraction setpoint. After reaching the desired different extraction methods was determined according to the
pressure, the static extraction took place. During this stage, the solid Folin–Ciocalteu method (Singleton and Rossi, 1965). Briefly, the oil
matrix was kept in contact with the solvent phase under the conditions samples were weighed, approximately 150 mg of each sample, and then
of pressure and temperature of each experiment. The static extraction 1.5 mL of methanol chromatographic grade were added. This mixture
times (confinement time) evaluated were: 10, 30, 60 and 90 min. After was vigorously shacked for 5 min and centrifuged for 10 min at

3
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

3000 rpm. The methanol phase was removed and used to determine the an equipment Agilent 1200 Series, with a diode array detector (DAD),
TPC. In this procedure, aliquots of 0.05 to 0.2 mL were prepared as quaternary pump, automatic injector, degassing system and a column
described above, and methanol was added to complete the volume up (Zorbax Eclipse XDB-C18 (4.6 mm diameter, 150 mm in length and
to 0.5 mL and mixed with 2.5 mL of Folin–Ciocalteu phenol reagent 5 μm particles). The elution of the compounds was performed in gra-
(diluted with distilled water to 0.2 mol/L). After 4 min, 2.0 mL of so- dient mode between the mobile phase A and B, with constant flow rate
lution of sodium carbonate (75 g/L) was added. The mixture was kept of 0.9 mL/min. Phase A was composed of the volumetric ratio of 95%
at ambient temperature and protected from light for 120 min. Subse- purified water in a Milli-Q system, 2% formic acid PA (Vetec, 85%
quently, the absorbance was measured in a spectrophotometer at purity) and 3% acetonitrile (Panreac, 99% purity). Phase B was com-
760 nm. The quantitative results were calculated using an analytical posed of 48% water, 2% formic acid and 50% acetonitrile. Both phases
curve of gallic acid and are expressed as mg of gallic acid equivalent per were previously filtered in a vacuum system, with a Nylon membrane
100 g of oil (mg AGE/100 g oil). filter with a diameter of 47 mm and a porosity of 0.45 μm.
Subsequently, the gases dissolved in the mobile phase were eliminated
2.7. Antioxidant activity (AA) in an ultrasonic bath, 154 W of power (Unique, model USC-1880 A) at
20 °C for 10 min. The gradient composition of the chromatographic run
2.7.1. ABTS method was initially 90% for phase A and 10% for phase B, ranging to 10% of A
Antioxidant activity (AA) was determined from 2,2-azino-bis-(3- in 25 min. At 27 min the composition of the mobile phase returned to
ethylbenzotiazoline-6-sulfonic acid) (ABTS) with radical scavenging. It the initial condition of 90% A and 10% B. The wavelength used for all
was performed based on the procedure described by Re et al. (Re et al., compounds analyzed was 280 nm. The compounds identification was
1999). ABTS was dissolved in water to a 7 mmol/L final concentration. performed based on comparison of retention times and absorption
This solution (5 mL) was mixed with 88 μL potassium persulfate solu- spectra between standard peaks (dihydroxybenzoic acid, caffeic acid, p-
tion (140 mmol/L) and then incubated in the dark for 16 h at room coumaric acid, ferulic acid and caffeine obtained from Sigma-Aldrich,
temperature to produce a stock solution of the radical cation (ABTS•+). St. Louis, MO, USA) and the peaks obtained on the chromatograms of
The ABTS•+ working solution was prepared by diluting the stock so- the samples. The quantification was performed by external calibration.
lution with absolute ethanol until reaching an absorbance of Results were expressed as milligrams of each individual compound per
0.700 ± 0.020 at 734 nm. For the sample analyses, aliquots of me- 100 g of oil (mg/100 g oil).
thanolic solutions prepared as previously described (procedure of TPC),
and methanol up to 100 μL was added to amber bottles and mixed with 2.9. Statistical analysis
3.9 mL of the ABTS•+ radical cation working solution (A734
nm = 0.700 ± 0.020). The resulting solution was kept in a dark en- In this work the results obtained were analyzed using the software
vironment for 6 min. After that time the absorbance was measured at Statistica 7.0® (Analytical Software, Tallahassee, FL, USA), and each
734 nm with a UV–vis spectrophotometer (model UV-1100, Pro-Ana- response was evaluated independently. Experimental data were ana-
lysis). Trolox (Sigma-Aldrich Co, St. Louis, USA) was used as an anti- lyzed for variance (ANOVA), and the average comparison considering
oxidant standard for construction of the standard curve. The results of the expanded uncertainties with 95% of confidence level calculated
AA were expressed as μmol of Trolox equivalent antioxidant capacity from triplicate results and assumed the same value for all other con-
(TEAC) per 100 g of oil (μmol TEAC/100 g oil). ditions in the same group of experiments (scCO2, PLE and scCO2 +
EtOH groups of experiments).
2.7.2. DPPH method
Antioxidant activity determination by the DPPH (2,2-diphenyl-1- 3. Results and discussion
picrylhydrazyl) method was performed according to the method pro-
posed by Kalantzakis et al. (2006) with minor modifications. Samples of 3.1. Raw material and Soxhlet extraction
SCG oil were prepared as described in the determination of TPC. Sub-
sequently, volumes of 0.05 to 0.1 mL of this methanol solution were SCG samples used in this work presented an average particle dia-
added to 3.9 mL of freshly prepared solution of DPPH in methanol meter, real density and apparent density of 0.466 ± 0.003 mm, 1.23 g/
(0.06 mmol/L) and stirred until homogenization. The resulting solution mL and 0.27 g/mL, respectively. The results of the Soxhlet extractions
was kept in a dark environment for 60 min. After that time the absor- are presented in Table 1. The overall extraction yield obtained with n-
bance was measured at 515 nm on a UV–vis spectrophotometer (model hexane was around 15%, value in agreement with the literature that
UV-1100, Pro-Analysis). The quantification was performed using a reports values within 12 to 18.3% (Ahangari and Sargolzaei, 2013;
Trolox analytical curve, and the results were expressed as μmol of Akgün et al., 2014; Andrade et al., 2012; Barbosa et al., 2014; Couto
Trolox equivalent antioxidant capacity (TEAC) per 100 g of oil (μmol et al., 2009; Melo et al., 2014). Such variation in the extraction yield
TEAC/100 g oil). might be attributed to the difference in the coffee cultivated depending
of the local, climatic conditions, harvest time and drying method re-
2.8. Phenolic compounds and caffeine lated to each region (Jenkins et al., 2014). Furthermore, the composi-
tion of the Coffea arabica and Coffea robusta (most commercialized
Phenolic compounds and the caffeine of the SCG oil samples were species) are significantly different, whence total lipids in blends formed
determined by high performance liquid chromatography (HPLC), using from these species must vary even more.

Table 1
Results for Soxhlet extraction of SCG oil with different solvents.
Run Solvent BTa (°C) Polarity indexb This work# Couto et al. (2009) Andrade et al. (2012) Ahangari and Sargolzaei (2013)
Extraction yield (%)

S1 n-Hexane 69 0.0 14.52 ± 0.52 18.3 12 ± 1 16.7 ± 0.1

S2 Ethyl acetate 71 4.4 15.04 ± 0.60 – 11.8 –
S3 Ethanol 79 5.2 15.64 ± 0.17 – 15 –

a b
BT: boiling temperature obtained from Efthymioupulos et al. (2018); Andrade et al. (2012); #Average and expanded uncertainties with 95% of confidence level
calculated from triplicate results.

4
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Results presented in Table 1 indicate a tendency to increase the temperature and pressure variation, which explain the univariant ex-
yield by increasing the solvent boiling temperature and the polarity traction yield.
index, as expected. The highest yield was 15.64 ± 0.52% obtained
with ethanol, which has the highest boiling temperature and higher
polarity among the solvents studied in this work. The high yields at 3.2.1. Overall extraction curves using scCO2
high temperatures are due to the breakdown of the cellular structure of Fig. 1 depicts the extraction curves of scCO2 extraction. The influ-
the coffee, which makes the solutes more available, thus increasing the ence of the pressure is better evidenced by comparing the results at
solubility of these compounds in the solvent (Efthymiopoulos et al., 40 °C, where the yield increased from 0.82% (obtained at 165 min ex-
2018). Besides provided the highest extraction among the organic sol- traction) to 12.19% at 20 MPa (295 min extraction). According to
vents evaluated in Soxhlet extraction, the ethanol also presents lower Akgün et al. (2014) the pressure is the most important parameter that
risks to human health and the environment when compared to n-hexane influences positively the scCO2 extraction yield, where high yield va-
and other organic solvents (Shang et al., 2016), making it desirable for lues can be obtained through increasing pressure that consequently
applications in the food, pharmaceutical and cosmetic industries. increases the density and solvation power of the CO2 (Ahangari and
Therefore, the ethanol was selected as the liquid solvent for the further Sargolzaei, 2013).
extraction presented in this work. Influence of the temperature on the extractions with scCO2 is more
complex than pressure due to the combination of two variables: density
and vapor pressure. The vapor pressure of the solutes increases with the
3.2. SCG extraction with scCO2 and pressurized ethanol (PLE) temperature, increasing the solubility in compressed CO2, however, the
density of the solvent decreases by increasing the temperature dimin-
In order to compare the scCO2 + EtOH extraction approach used in ishing the solutes solubility. The main effect depends on the magnitude
this work (CXL), which is the main subject, extractions using scCO2 and of these variables for each system (Pourmortazavi and Hajimirsadeghi,
compressed ethanol (PLE) were also performed and the results of the 2007). In Fig. 1, at pressures from 10 to 20 MPa, it is observed a ne-
extraction curves are presented in Figs. 1 and 2. All experiments of gative effect on extraction yields when the temperature increased from
scCO2 and PLE were performed using 30 min of static extraction time 40 °C to 80 °C. Couto et al. (2009) also observed a negative effect of the
(confinement period). The overall extraction curves of different ex- temperature on the extraction at pressures up to 20 MPa, and reported
traction technique will be further discussed and compared later, for that only at high pressures (25 and 30 MPa) the density of scCO2 de-
instance an analysis of main factor of each extraction is presented. To creased moderately with increasing temperature and the effect of the
compare the extraction yield obtained in scCO2 and PLE the same ex- increase in the solute vapor pressure prevailed resulting in increasing in
traction time was set for each process, 145 and 35 min for scCO2 and the overall extraction yield.
PLE, respectively, and these results are presented in Table 2 along with Extraction yield obtained in Soxhlet with n-hexane
the experimental conditions evaluated. ScCO2 results were statistically (14.52 ± 0.52%) was higher than the highest yield (12.19 ± 0.21%)
significant (p ≤ 0.05) when considering the CO2 density as the main obtained with scCO2 (40 °C and 20 MPa in 4.9 h of extraction). Couto
factor, showing as expected that the extraction yield increases as the et al. (2009) obtained similar results, reaching an extraction yield with
solvent density is increased, which was obtained at the lower tem- n-hexane in Soxhlet of 18.3% and the maximum yield in SFE was
perature and higher pressure evaluated. The supercritical fluid-based 15.4%, at pressures of 15 and 30 MPa and temperatures of 40 and 60 °C
extraction is affected by the thermodynamic conditions because at the in 3 h of extraction. In addition, the effects of pressure and temperature
critical region the fluid exhibits a liquid-like density and a much more were the same as those observed in the present study. Andrade et al.
increased solvent capacity that is pressure-dependent. For PLE, ex- (2012) also reported pressure and temperature effects as those pre-
traction yield reached high values, comparable to the results obtained sented in this study. Those obtained yields of 0.43–10.5 % of SCG oil
in Soxhlet with ethanol (Run S3, Table 1), and both temperature and using scCO2 at pressures of 10–30 MPa and temperatures of 40–60 °C
pressure evaluated were statistically non-significant (p ≥ 0.05) at 35 with 2.5 h of extraction, while the yield obtained with n-hexane in
min of extraction, which means that all conditions achieved the same Soxhlet was 12%.
extraction conditions regardless of temperature and pressure variation. Results regarding SCG extraction with scCO2 obtained in this work
It can be seen in Table 2 that the solvent density changed a little with and those presented in the literature are indicating that high-pressure

Fig. 1. Experimental overall extraction curves for spent coffee ground (SCG) oil with scCO2 at different temperature (40 °C, 60 °C and 80 °C) and pressure (10 MPa,
15 MPa and 20 MPa).

5
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Fig. 2. Experimental overall extraction curves for spent coffee grounds oil with pressurized ethanol at different temperature (40 °C, 60 °C and 80 °C) and pressure
(10 MPa, 15 MPa and 20 MPa).

conditions are need to recover most of the oil present in this raw ma- ethanol (15.64%). Similar effects of pressure and temperature on the
terial, and indeed the addition of an organic cosolvent or a combined extraction yield of green coffee oil were reported by Oliveira et al.
mixture of scCO2 plus organic solvent (as ethanol) can be a feasible (2018), where temperature (50–74 °C) and static extraction time
option, since lower pressure conditions would be requited. (3–9 min) were the only variables that had significant effect on PLE. At
a fixed pressure of 10.30 MPa, Oliveira et al. (2018) obtained extraction
3.2.2. Overall extraction curves of SCG using PLE yields of 6.60–9.78 % of green coffee oil, higher than the yield of 7.57%
The overall extraction curves for PLE extraction using ethanol under obtained with n-hexane in Soxhlet.
the same experimental conditions used in the extraction with Even though the PLE results obtained in this work are very pro-
scCO2+EtOH (Table 2) and 30 min of static extraction are presented in mising for the oil extraction from SCG, it is worth mention that this
Fig. 2. For PLE, the highest extraction yields (15.12% and 15.74%) were technique used a large amount of solvent and it separation when re-
obtained at the highest temperature of 80 °C and 10 and 20 MPa, re- quired is energy consuming operation, which is normally performed by
spectively. Similar yields for both pressures at the same temperature either distillation or evaporation to get concentrated extracts for further
were expected, since EtOH is a non-compressible fluid and its density is applications. Therefore, a process that demands lower amounts of li-
slightly affected by varying the pressure, as discussed on PLE results. In quid solvent might be a potential strategy.
Fig. 2, the temperature increase from 40 °C to 80 °C, at both fixed
pressures, increased both the extraction yield and the initial extraction 3.3. Extraction using scCO2+EtOH
rate. In PLE, high temperatures improve the contact of the solvent with
the solute, therefore the extraction is enhanced due to the increase in The yields obtained for the extractions with scCO2+EtOH at dif-
the extraction rates, the mass transfer and the solubility of the com- ferent conditions of pressure, temperature, ethanol to SCG mass ratio
pounds in the solvent, and also due to the decrease of the surface ten- (MRES) and confinement time are shown in Table 3. In addition,
sion and viscosity (Shang et al., 2016). Table 3 presents additional experiments for scCO2+EtOH extraction
Extraction yields for PLE using ethanol at 80 °C were higher than the using ethanol at 95% of purity (around azeotropic ethanol). The highest
overall yields obtained in Soxhlet with n-hexane (14.52%), and in the overall yield among all conditions performed was 15.87% (Run 19),
same order to the yields obtained with ethyl acetate (15.04%) and obtained using scCO2+EtOH with a MRES of 2:1 at 80 °C and 20 MPa in

Table 2
Experimental conditions and results for SCG oil extraction using scCO2 (SFE) and pressurized ethanol (PLE).
a
Run Solvent T (°C) P (MPa) ρ solvent (g/mL) m0, solvent b (g) msolvent c (g) Dynamic extraction time (min) Extraction yield (%)

SFE
1 scCO2 40 10 0.628 51.27 303.98 145 0.69
2 scCO2 80 10 0.222 23.01 174.20 145 0.11
3 scCO2 40 20 0.840 64.40 578.38 145 6.25
4 scCO2 80 20 0.593 48.83 284.70 145 1.67
5 scCO2 60 15 0.604 50.20 314.51 145 0.82 ± 0.21$

PLE
6 EtOH 40 10 0.780 72.77 87.54 35 13.74
7 EtOH 80 10 0.742 70.30 55.94 35 15.12
8 EtOH 40 20 0.787 78.03 72.51 35 13.72
9 EtOH 80 20 0.750 75.49 88.92 35 15.29
10 EtOH 60 15 0.743 72.77 87.54 35 14.25 ± 0.39$

a
CO2 and ethanol densities obtained from the NIST database (NIST, 2018) and (Green and Perry, 2007) calculated at pressure and temperature conditions of the
experiments (T and P); b Mass of solvent (CO2 or ethanol) used in the static extraction; c Mass of solvent (CO2 or ethanol) used in the dynamic extraction; $ Average
and expanded uncertainties (with 95% of confidence level) calculated for SFE and PLE extraction yields in triplicate experiments.

6
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Table 3
Experimental conditions and results for SCG oil extraction using scCO2+EtOH (CXL).
Run CXL (MRES) T (°C) P (MPa) IPECa m0, EtOH b (g) m0, CO2 c (g) msolvent d (g) CTe (min) Extraction time (min) Extraction yield (%)
(wt%)

Experimental design with ethanol 99.8%

11 scCO2+EtOH (0.5:1) 40 10 11.89 6.93 51.33 18.,71 30 10 10.35
12 scCO2+EtOH (0.5:1) 80 10 21.96 7.31 25.98 138.50 30 75 0.46
13 scCO2+EtOH (0.5:1) 40 20 10.23 6.92 60.72 146.95 30 75 12.85
14 scCO2+EtOH (0.5:1) 80 20 13.25 7.31 47.85 108.00 30 55 5.59
15 scCO2+EtOH (0.5:1) 60 15 13.72 7.53 47.35 28.63 30 15 9.13 ± 0.34f

16 scCO2+EtOH (2:1) 40 10 57.22 33.43 24.99 46.04 30 25 13.42

17 scCO2+EtOH (2:1) 80 10 61.54 33.43 20.89 19.38 30 10 12.62
18 scCO2+EtOH (2:1) 40 20 55.99 34.00 26.72 88.15 30 45 13.82
19 scCO2+EtOH (2:1) 80 20 65.78 32.26 16.78 49.03 30 25 15.87
20 scCO2+EtOH (2:1) 60 15 59.95 33.76 22.55 47.71 30 25 14.14 ± 0.16f

Additional experiments with ethanol 99.8%

21 scCO2+EtOH (0.25:1) 60 15 8.28 4.41 48.84 66.77 30 35# 2.62
22 scCO2+EtOH (1:1) 60 15 28.10 16.32 41.75 28.62 30 15 12.46
23 scCO2+EtOH (0.5:1)* 40 10 14.65 16.13 93.99 41.49 30 22.5 13.53

Additional experiments with 95%

24 scCO2+EtOH (1:1) 60 15 27.29 15.63 41.64 38.58 10 20 11.76
25 scCO2+EtOH (1:1) 60 15 21.95 13.29 47.26 38.78 30 20 11.54 ± 0.55f
26 scCO2+EtOH (1:1) 60 15 27.53 15.80 41.60 38.69 60 20 11.00
27 scCO2+EtOH (1:1) 60 15 26.31 15.85 44.40 28.88 90 15 11.55
28 scCO2+EtOH (0.25:1) 60 15 6.11 3.37 51.75 29.55 30 15 0.99
29 scCO2+EtOH (0.5:1) 60 15 12.35 6.92 49.13 24.25 30 12.5 7.40
30 scCO2+EtOH (2:1) 60 15 60.02 33.63 22.4 38.38 30 20 13.61

a
Initial ethanol mass fraction (wt%) in the scCO2+EtOH mixture in the CXL extraction; b Mass of ethanol used in the static extraction; c Mass of CO2 used in the static
extraction; d Mass of solvent (CO2 or ethanol) used in the dynamic extraction; e CT: confinement time; f Average and expanded uncertainties (with 95% of confidence
level) values in triplicate experiments; # Extraction did not reach the theoretical yield of extractable solute; * Sequential extraction.

25 min extraction. Compared to the extraction using scCO2 (Run 4) at ethanol addition above MRES of 0.5:1 increased on average 90% the
same condition (80 °C and 20 MPa), this result represents an increase of extraction yields when compare to extraction with pure scCO2 showing
14.20 p.p. in the extraction yield and 83% reduction in the extraction that the ethanol addition contributed to enhance the extraction effi-
time. In addition, the mass of CO2 used during the static extraction ciency.
decreased from 48.83 to 16.78 g of CO2 and the consumption of solvent For extractions with MRES of 2:1 (Runs 16–20), where IPEC values
during the dynamic step (solvent) was 284.70 to 49.03 g of CO2. ranged from 57.22 to 65.78 wt% ethanol in CXL mixture. The extraction
Results obtained with scCO2+EtOH (2:1), Run 19 (80 °C and yields at condition with high amount of ethanol the system behaved
20 MPa), were better in terms of ethanol consumption compared to the likely PLE system reaching yields within 12.97 to 15.97%, while pres-
pressurized ethanol extraction from Run 9 (80 °C and 20 MPa). surized ethanol the extraction yields were from 13.92 to 15.74 %.
Although the overall extraction yields were similar, the comparison However, it is worth to mention that smaller amounts of ethanol were
between the extraction with pressurized ethanol and scCO2+EtOH used in the extractions with scCO2+EtOH (2:1) (32.26 to 33.7 g of
(2:1) shows that the ethanol consumption during the static extraction ethanol) than in extractions with pressurized ethanol (150.54 to
(m 0, EtOH ) decreased from 75.49 to 32.26 g and during the dynamics 164.41 g of ethanol). The scCO2+EtOH (2:1) extraction performance
extraction stage the CO2 (Run 19) was lower than the mass used in the was similar to the extraction with pressurized ethanol, where the
pressurized ethanol extraction (Run 9). A disadvantage of the extraction highest yields were obtained by increasing the temperature to a fixed
with scCO2+EtOH is the use of two solvents, but the CO2 at ambient pressure, except for the extraction at 80 °C and 10 MPa.
pressure becomes gas and it is easily separated from the extracts, and In Table 3 can be observed that for extractions using scCO2+EtOH,
the ethanol separation might be unnecessary if the SCG oil will be used the initial percentage of ethanol in the semi-batch CXL mixture (IPEC)
for biodiesel production, for example, since ethanol is used in the varied for the same MRES. The addition of liquids into the extractor
transesterification reaction (Döhlert et al., 2015). However, if the sol- vessel can be performed in two ways: as a mixed fluid in the pumping
vent removal is required, ethanol separation from the extracts obtained system using a second pump and a mixing chamber; or adding the li-
with pressurized ethanol will be more expensive than from the ex- quid direct into the sample before the extraction (Pourmortazavi and
traction with scCO2+EtOH because the ethanol volume is higher in the Hajimirsadeghi, 2007). In this work, ethanol was added as a static co-
first case. solvent, being mixed with the matrix and rapidly added to the ex-
In general, the results in Table 3 indicates the increase in extraction tractor, and after that the CO2 was added until reach the pressure fixed
yields obtained in the semi-batch CXL process with the increase in the for the extraction. As CO2 density depends on temperature and pressure
ethanol to spent coffee grounds mass ratio (MRES). The highest yield the CO2 amount inject varied according to the extraction conditions,
variation related to MRES of increase from 0.5:1 to 2:1 was 12.16 p.p., even though the MRES was constant. Among the advantages of using a
obtained at condition of 80 °C and 10 MPa. For the extractions with system with the static method, stand out the simplicity to operate the
scCO2+EtOH, the effects of pressure and temperature on the yield extraction equipment and energy saving related to the use of only one
showed a significant variation with the MRES variation. pump for solvent flow and pressure control during the dynamic ex-
In extractions using MRES of 0.5:1 (Runs 11–15) the initial per- traction.
centage of ethanol in scCO2+EtOH mixture (named IPEC) ranged from In order to study the influence of the confinement time (static ex-
10.23 to 21.96 wt% ethanol, indicating a predominant presence of traction) on the extraction yields, experiments were performed at 60 °C,
scCO2 in the extraction vessel. As observed for the scCO2 extractions, 15 MPa, MRES 1:1 and using ethanol 95% purity (5% water, wt%) at
low temperature and high pressure resulted in higher yields. However, 10, 30, 60 and 90 min for the static extraction period (Run 24–27). As

7
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Fig. 3. Experimental overall extraction curves for spent coffee grounds oil with scCO2+EtOH at 60 °C, 15 MPa and different temperature ethanol to SCG mass ratio
(MRES = 0.25:1, 0.5:1, 1:1 and 2:1).

the extraction yields obtained with different confinement times were In Fig. 3, increasing the MRES values in CXL extractions resulted in
similar, 30 min was adopted as the confinement time for extractions to higher overall extraction yields. Extraction with scCO2+EtOH (2:1)
provide a longer contact time of solvents and matrix during the static showed a yield of 14.14%, a result superior to the yield of 9.13% ob-
extraction stage. tained with scCO2+EtOH (0.5:1). However, the amount of ethanol used
Fig. 3 depicts the overall CXL extraction curves with scCO2+EtOH, in the MRES of 0.5:1 was 4 times lower than that used in the extraction
at different MRES (0.25:1, 0.5:1, 1:1 and 2:1), using 99.8% purity with 2:1 and the extraction time was 40% lower, which implies in lesser
ethanol, 60 °C and 15 MPa, which can be compared with the overall solvent during the dynamic extraction. In Fig. 3, increasing MRES from
extraction curves with scCO2 and pressurized ethanol (PLE) at same 0.25:1 to 1:1 also increased the initial extraction rate. However, the
experimental condition. From Fig. 4 it can be observed that the addition initial extraction rate at MRES of 2:1 is lower than the initial extraction
of ethanol increased the extraction yield from 0.98% (55 min extrac- rate at 1:1. For the MRES conditions of 2:1, higher mass transfer re-
tion) using scCO2 (Run 5) to 2.62% (35 min extraction) using scCO2+EtOH sistance caused by the ethanol excess related to the solids in extraction
(0.25:1) (Run 21). Thus, the extraction yield using scCO2+EtOH with bed probably affected the CO2 diffusion into the solvent + solids phase
8.28% of ethanol in the CXL mixture (IPEC) was 2.6 times higher than leading to lower extraction rates. Similar effects of increasing the per-
the extraction yield with scCO2 at same condition. The highest yield in centage of ethanol in the CXL mixture (IPEC) were observed by
scCO2+EtOH extraction was 14.14% obtained with a MRES of 2:1, Rodrígues et al. (2016). Those authors observed that at 50 °C and 7 MPa
which is around the same extraction yield obtained with pressurized using 50% ethanol (lower percentage of ethanol studied) the yield of
ethanol (14.25%). However, it worth mentioning again that the ex- Moringa oleifera extract was 2 times higher than the yield for the SFE
traction with CXL has the advantage that lower amounts of ethanol extraction with scCO2 at same experimental condition in 200 min of
were used when compared to PLE. extraction. Analyzing the percentage of ethanol at a fixed extraction

Fig. 4. Experimental overall extraction curves for spent coffee grounds oil with scCO2+EtOH at MRES of 0.5:1, different temperature (40 °C, 60 °C and 80 °C) and
pressure (10 MPa, 15 MPa and 20 MPa).

8
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Fig. 5. Phase diagram for CO2(1) + ethanol(2) at 40 °C (black

lines), 60 °C (red lines) and 80 °C (blue lines). Continuous and
dashed lines are the saturated liquid and vapor phases, re-
spectively. Diagram calculated using the Peng-Robinson EoS
with the binary interaction parameters (k12 = 7.8332 × 10−2
and l12 = -3.0999 × 10−2) as presented by Araújo et al.
(2012).

time, Rodrígues et al. (2016) found that increasing the ethanol fraction experimental conditions (0.222 g/mL). Thus, the CXL solvent phase
from 50 to 60% resulted in increased yield. While the increase of 60 to formed by CO2 and ethanol had insufficient solvating power to solu-
70% ethanol decreased the extraction yield. Such behavior may be an bilize the spent coffee ground oil. Therefore, the results for this con-
indicative of the decrease in the initial extraction rate at high ethanol dition (80 °C and 10 MPa) with scCO2+EtOH and MRES of 0.5:1
concentrations observed in the present work, which requires a longer reached around the same values of the extraction yield (Run 2) and
extraction time to obtain higher extraction yields. initial extraction rate obtained with pure scCO2 (Fig. 1).
Fig. 4 shows the overall CXL extraction curves using scCO2+EtOH The overall curves for scCO2+EtOH extraction with MRES of 2:1,
at MRES of 0.5:1 with 30 min of confinement time. At this mass ratio, presented in Fig. 6, show a different behavior when compared to the
the extraction behavior was different depending on the experimental curves in Fig. 4. For all conditions it is observed that after the initial
condition of temperature and pressure. For extractions of Run 11 (40 °C extraction rate step the oil recovery was abruptly interrupted and the
and 10 MPa), Run 14 (80 °C and 20 MPa) and Run 15 (60 °C and extraction curves presented different behavior at different temperatures
15 MPa), in Fig. 5, after the initial extraction rate step the oil recovery (40 and 80 °C). At 40 °C the yields were around the value comparing the
was suddenly interrupted. This indicates that the first minutes of the results at 10 and 20 MPa, like that observed in the extraction with
extraction are controlled by the interaction between ethanol and the pressurized ethanol under the same conditions of temperature and
solute molecules, and then the oil recovery is abruptly reduced when pressure. While at 80 °C a decrease is observed by decreasing the
the ethanol concentration in the extractor vessel decreases significantly pressure. This occurs because during the dynamic extraction at 80 °C
(Fetzer et al., 2018). For the extraction at 40 °C and 20 MPa (Run 12) and 10 MPa the CXL mixture goes through the VLE region during the
the three periods of supercritical extraction (CER, FER and DCR) extraction path and the solvation power is not enough to achieve the
(Pereira and Meireles, 2010) were observed. At this condition, high CO2 yield obtained at 80 °C and 20 MPa, as observed in the extraction with
density allows the extraction of SCG oil even after the ethanol con- pressurized ethanol under the same conditions. However, even the
centration has decreased. In the extraction at 80 °C and 10 MPa (Run phase diagram presented in Fig. 5 is indicating phase separation at
13), the lowest yield among the experimental conditions was obtained 80 °C and 10 MPa, the ethanol excess in the extractor vessel resulted in
during a single extraction step with a constant and low extraction rate. oil recovery of 86.91%.
The different behavior among the extraction curves in Fig. 4 can be Regarding the extraction curves using scCO2+EtOH (2:1) presented
explained by the extraction path within the phase diagram of the in Fig. 6, it is observed that lower pressure (10 MPa) resulted in higher
CO2+ethanol mixture, as shown in Fig. 5. During the dynamic ex- initial extraction rates. In addition, the highest extraction rate at 2:1
traction, the pressure is constant and the CO2 concentration in the ex- RMEB were obtained at 80 °C, whereas for the extraction using
tractor vessel increases with the extraction time, because the ethanol is scCO2+EtOH (0.5:1) the highest initial rates were obtained at 40 °C.
also “extracted” from the vessel by the scCO2 carrying the solutes (ex- This is indicating that in the extraction with scCO2+EtOH (2:1) the
tracts) along with them. At 10 MPa and 80 °C, as the CO2 concentration effects related to the solubility of compounds in ethanol are more sig-
inside the extractor increases, the CXL mixture tends to the vapor-liquid nificant, while for the extraction using scCO2+EtOH at 0.5:1 (Fig. 4)
(VLE) region. For the other experimental conditions, at the beginning of the effect of increasing density and solvation power of CO2 by de-
the extraction the CXL mixture is in a single phase, above the VLE re- creasing temperature are more important.
gion, and even the scCO2 switches the ethanol the systems remain in the Fig. 7 depicts a comparison of overall extraction curves obtained at
one-phase region. As the extraction time increases, the CO2 molar different process conditions. For extraction using ethanol along with
fraction increases and the CXL mixture tends to the supercritical region. scCO2 the initial extraction rate is similar even at different process
However, the solvent phase is always above the saturation line of the condition (temperature and pressure), showing that the extraction is
VLE. essentially controlled by the ethanol presence and the interactions be-
The lowest yield for the MRES of 0.5:1 was obtained at 80 °C and tween the organic solvent and solutes in the raw material. However,
10 MPa. In this condition CO2 has the lowest density among the overall extraction yields reached different equilibrium values. The

9
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Fig. 6. Experimental overall extraction curves for spent coffee grounds oil with scCO2+EtOH at MRES of 2:1, different temperature (40 °C, 60 °C and 80 °C) and
pressure (10 MPa, 15 MPa and 20 MPa).

ethanol addition of 10.23% in the CXL mixture (MRES = 0.5:1) in- (Table 3). For extraction using a MRES of 2:1, the extraction yield in-
creased the overall extraction yield from 2.21% to 12.37% in 55 min of creased about 0.53 p.p. comparing the results with ethanol 95% (Run
extraction. At MRES of 2:1, the highest extraction yield of 15.87% (in 30) to ethanol 99.8% (Run 20). These results indicate the viability of
25 min of extraction) was obtained at 80 °C and 20 MPa, the same the use of a lower ethanol purity (for example azeotropic ethanol) in the
condition where the highest extraction yield of 15.74% was obtained extraction with scCO2+EtOH in terms of keeping high extraction
using pressurized ethanol, but in a extraction time of 55 min. However, yields.
as it can be noted, the constant initial extraction rate (CER step) for
scCO2+EtOH (2:1) was higher than for PLE extraction with pure
ethanol. 3.3.1. Sequential extraction using scCO2+EtOH
The influence of the ethanol purity of the solvent (5% water in A sequential extraction was also carried out to evaluate the possi-
ethanol) on the extraction yield was evaluated at 60 °C and 15 MPa, bility of enhancing the oil recovery by performing a multistep extrac-
using MRES of 0.25:1, 0.5:1, 1:1 and 2:1 and comparing ethanol 95% tion. The experimental condition selected for this test was 0.5:1
and 99.8% purity. These results are shown in Fig. 8. The yields obtained (MRES), 40 °C and 10 MPa (Run 10), because at such condition the
with ethanol 99.8% were higher than those obtained with 95% ethanol extraction yield was 10.34% (around 2/3 of the total extraction cap-
for all MRES evaluated. However, the yields with ethanol 95% tend to ability), but with the highest initial extraction rate and consequently
approximate to the results obtained with ethanol 99.8% by increasing lower extraction time (10 min), and temperature, pressure and lower
the MRES. Increasing the ethanol purity from 95% to 99.8%, for ex- solvent consumption during dynamic extraction (18.71 g CO2).
periments performed with a MRES of 0.5:1, resulted in an extraction Additionally, as shown in Fig. 7, the conditions at MRES of 2:1 also
yield increase of 7.40% (Run 29) to 9.13% (Run 15), about 1.73 p.p. presented a high initial extraction rate, such as the extraction at 80 °C
and 10 MPa (Run 17), however these extractions used 33.43 g of

Fig. 7. Extraction curves with highest yields for the different solvents scCO2 (40 °C and 20 MPa), scCO2+EtOH at MRES od 0.5:1 (40 °C and 20 MPa), scCO2+EtOH at
MRES of 2:1 (80 °C and 20 MPa) and pressurized ethanol (80 °C and 20 MPa).

10
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Fig. 8. Experimental overall extraction curves for spent coffee grounds oil using scCO2+EtOH at 60 °C, 15 MPa and different MRES (0.25:1, 0.5:1, 1:1 and 2:1), using
ethanol with 99.8% of purity and ethanol with 95% of purity.

ethanol during the static extraction, while Run 10 used 6.93 g of 10 MPa was obtained in the extractions using the lowest MRES (0.5:1).
ethanol. As observed in Table 3, the CO2 mass used during the static steps for
Fig. 9 shows the overall extraction curves at 40 °C and 10 MPa for the sequential extraction was the largest amount among all conditions.
the sequential extraction and its comparison with others single step Due to limitations of the experimental setup it was needed to stop the
extractions with each solvent (scCO2, scCO2+EtOH (0.5:1 and 2:1) and extraction, release the CO2 at the end of the CER step to add another
pressurized ethanol). In the sequential extraction, with the accom- charge of ethanol, followed by a new pressurization, which resulted in
plishment of a second extraction step, the yield of 10.21% obtained in greater mass of CO2 used. However, this drawback can be easily solved
10 min of extraction was increased to 13.53% with second extraction by performing the ethanol addition using a high-pressure pump. Thus,
step of 12.5 min. As shown in Fig. 10, such enhance obtained using the the results presented in this work suggest that the extraction yield of
multistep extraction approached the overall extraction yield to the SCG oil can be improved with the addition of ethanol using sequential
maximum obtained in this study (single extraction using an ethanol to extraction steps of extraction.
SCG of 2:1) but using half of the solvent amount. In addition, extraction
yield in the sequential approach was higher than the yield obtained for 3.4. Fatty acids profile
the best experimental condition at MRES of 0.5:1 (40 °C and 20 MPa).
It can be notice from Fig. 9 that the increase in the ethanol amount Table 4 shows the fatty acid profile for the SCG oil samples obtained
added into the extraction vessel resulted in lower initial rates of ex- at the highest extraction yields of each solvent (SFE, CXL, PLE and
traction. This is due to the increase in mass transfer resistance caused Soxhlet). The major fatty acids found in the SCG oil were linoleic
by the ethanol excess inside the extractor vessel in the beginning of the (ranged from 44.05 to 45.79 %) and palmitic (29.70–33.04 %), fol-
extraction process. Thus, as suggested by the results obtained in this lowed by oleic (8.43–8.89 %) and stearic (7.69–8.79 %). Other fatty
study, the highest initial extraction rate at a fixed condition of 40 °C and acids found were cis-11-eicosenoic, linolenic and tricosanoic. Similar

Fig. 9. Experimental extraction curves for each solvent (scCO2, scCO2+EtOH (MRES = 0.5:1 and 2:1) and pressurized ethanol) at 40 °C and 10 MPa, for sequential
extraction at 40 °C and 10 MPa and for scCO2 at 40 °C and 20 MPa.

11
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Fig. 10. Antioxidant activity (AA) by the total phenolic compounds content (TPC) for spent coffee grounds oil.

results of fatty acid profile for SCG oil were found by Akün et al. (2014), Table 5
using n-hexane, dichloromethane e scCO2 (linoleic 47.77–49.01 %, Iodine values (IV) and saponification value (SV) for spent coffee grounds (SCG)
palmitic 24.47–27.70 %, oleic 9.00–10.39 %, stearic 7.41–8.25 %), and oil samples.
Melo et al. (2014), using n-hexane (linoleic 44.67%, palmitic 37.37%, Extraction Solvent Condition Iodine Saponification
oleic 8.31%, stearic 7.07%). The fatty acid profile of the SCG oil is value value
similar to cottonseed oil, where the major fatty acid normally found is (g I2/ (mg KOH/g oil)
linoleic (46.7–58.2 %), followed by palmitic (21.4–26, 4%), oleic 100 g oil)

(14.7–21.7 %) and stearic (2.1–3.3 %) (Knothe, 2002). CXL scCO2+EtOH(0.5:1) 40 °C/10 MPa 90.34 197.12
Table 5 presents the iodine value (IV) for the SCG samples de- scCO2+EtOH(0.5:1) 40 °C/20 MPa 92.51 196.76
termined by using the fatty acid profile presented in Table 4 The IV scCO2+EtOH(0.5:1) 60 °C/15 MPa 91.12 196.98
results for the different extractions were similar, varying from 90.34 to scCO2+EtOH(2:1) 80 °C/10 MPa 92.26 196.73
scCO2+EtOH(2:1) 40 °C/20 MPa 92.52 196.77
93.05 g I2/100 g oil. These values are close to that observed by
scCO2+EtOH(2:1) 80 °C/20 MPa 91.37 196.27
Najdanovic-Visak et al. (2017) of 83.7 g I2/100 g SCG oil extracted with
SFE scCO2 40 °C/20 MPa 91.58 196.49
n-hexane in Soxhlet. Oils with low iodine content such as olive oil
(80.3 g I2/100 g oil) are of great interest for use in human diets and in PLE EtOH 80 °C/10 MPa 92.56 196.63
EtOH 40 °C/20 MPa 93.05 196.13
food and cosmetic and pharmaceutical industries. The IV values of the
EtOH 80 °C/20 MPa 91.23 196.56
SCG oil obtained in this study were lower than IV values observed for
other edible oils, such as sunflower oil (102.02 g I2/100 g oil), peanut Soxhlet n-Hexane – 92.59 194.75
EtAC – 92.44 197.03
oil (119.19 g I2/100 g oil) (Bozdogan Konuskan et al., 2019) and corn Ethanol – 92.22 196.90
oil with (130.6 g I2/100 g oil) (El-Hadad and Tikhomirova, 2018). In

Table 4
Fatty acid profile for spent coffee grounds (SCG) oil samples.
Extraction (Solvent) Palmitic acid Stearic acid Oleic Linoleic acid Gama- linolenic Cis-11-eicosenoic Docosadienoic acid Tricosanoic acid
Condition (C16:0) (C18:0) acid (C18:2) acid (C18:3) (%) acid (C20:1) (C22:2) (C23:0)
(C18:1)

CXL (scCO2+EtOH)
40 °C/10 MPa/0.5:1 31.53 8.79 8.88 44.05 1.52 3.22 0.95 1.06
40 °C/20 MPa/0.5:1 31.42 8.30 8.79 45.38 1.49 3.03 0.77 0.82
60 °C/15 MPa/0.5:1 31.69 8.52 8.62 44.68 1.60 3.06 0.89 0.94
80 °C/10 MPa/2:1 31.41 8.42 8.57 45.35 1.38 3.03 0.99 0.85
40 °C/20 MPa/2:1 31.36 8.41 8.61 45.48 1.41 3.01 0.87 0.85
80 °C/20 MPa/2:1 31.08 8.26 8.53 44.92 1.38 2.94 2.05 0.84

SFE (scCO2)
40 °C/20 MPa 33.04 7.69 8.43 45.38 1.43 2.28 1.21 0.54

PLE (EtOH)
80 °C/10 MPa 30.43 8.46 8.66 45.39 1.68 3.21 1.22 0.95
40 °C/20 MPa 29.70 8.15 8.53 45.79 1.67 3.06 2.18 0.92
80 °C/20 MPa 30.86 8.47 8.89 44.57 1.72 3.14 1.44 0.91

Soxhlet
n-hexane 31.31 8.60 8.65 45.43 1.36 3.17 0.57 0.91
EtAc 31.44 8.34 8.78 45.34 1.67 3.04 0.49 0.90
Ethanol 31.55 8.47 8.59 45.29 1.40 3.09 0.73 0.88

Standard uncertainties estimated for the analysis were considered < 1%.

12
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Table 6
Total phenolic content (TPC) and antioxidant activity (AA) by ABTS and DPPH methods for spent coffee grounds oil.
Extraction Solvent Experimental condition TPC AA ABTS AA DPPH
mg GAE/100 g oil μmol TEAC/100 g oil

CXL
scCO2+EtOH(0.5:1) 40 °C/10 MPa 269 4610 3148
scCO2+EtOH(0.5:1) 40 °C/20 MPa 272 3308 2840
scCO2+EtOH(0.5:1) 60 °C/15 MPa 411 ± 79# 4143 ± 115# 3215 ± 383#
scCO2+EtOH(2:1) 80 °C/10 MPa 393 4061 2868
scCO2+EtOH(2:1) 40 °C/20 MPa 295 3440 2691
scCO2+EtOH(2:1) 80 °C/20 MPa 348 3629 2815
SFE
scCO2 40 °C/20 MPa 64 397 345
PLE
EtOH 80 °C/10 MPa 406 ± 40# 4559 ± 760# 3534 ± 420#
EtOH 40 °C/20 MPa 267 4451 2841
EtOH 80 °C/20 MPa 351 3905 3199
Soxhlet
n-Hexane – 66 676 642
EtAc – 177 2443 1618
Ethanol – 330 3502 2606

#
Expanded uncertainties estimated with 95% of confidence level considering triplicate experiments.

addition, the IV values for SCG oil obtained in this work are within the regulation established by the EU no. 432/2012, in which the minimum
European standard for biodiesel production, in which the maximum of 25 mg GAE/100 g oil is established (Khmelinskii et al., 2019). In
value of 120 g I2/100 g oil is established (Ramos et al., 2009). addition, the samples of SCG oil showed TPC higher than the values of
Saponification values (SV), in Table 5, presented similar values for 13.8–70.0 mg GAE/100 g oil found in virgin olive oil by Khmelinski
all experimental conditions and the solvents used in this study, ranging et al. (2018).
between 194.75 and 197.12 mg KOH/g oil. Among the solvents ana-
lyzed, the lowest SV value was obtained with n-hexane in Soxhlet
3.6. Antioxidant by ABTS and DPPH methods
(194.75 mg KOH/g oil). Values of 173.9 mg KOH/g oil and 183.0 mg
KOH/g of SCG oil were obtained by Al-Hamare et al. (2012) using n-
Table 6 shows the results for antioxidant activity (AA) for the SCG
hexane and ethanol in Soxhlet, respectively. However, SV values of SCG
oil samples with the highest yields for each extraction (SFE, CXL, PLE
oil obtained in this work were close to the SV for the corn oil (190.3 mg
and Soxhlet), performed according to ABTS and DPPH analyses. In the
KOH/g oil, El-Hadad e Tikhomirova, 2018).
Soxhlet extraction, an increase in AA values was observed with the
increase in the solvent polarity, thus, the highest AA was obtained with
3.5. Total Phenolic Content (TPC) ethanol (3501.80 μmol TEAC/100 g oil). The highest AA performed by
the ABTS method was 4609.53 μmol TEAC/100 g oil, obtained with
Total phenolic content (TPC), presented in Table 6, ranged from scCO2+EtOH (0.5:1) at 40 °C and 10 MPa. The highest AA by the DPPH
63.90 to 411.10 ± 18.45 mg GAE/100 g oil. The highest TPC value of method was 3534.16 μmol TEAC/100 g oil, obtained with pressurized
411.10 ± 18.45 mg GAE/100 g oil was obtained for the extraction ethanol at 80 °C and 10 MPa. Difference in the content of compounds
with scCO2+EtOH at 60 °C, 15 MPa and MRES of 0.5:1. In the Soxhlet responsible for the antioxidant characteristics detected by the DPPH
extraction, TPC values increased through increasing solvent polarity, and ABTS method in each sample can generate results like above, where
hence the highest TPC value was obtained with ethanol (329.94 mg the highest AA were obtained at different samples. Hence, the necessity
GAE/100 g oil) and the lowest value was obtained with n-hexane of performing more than one AA analysis method is evidenced.
(66.12 mg GAE/100 g oil). For the extractions with scCO2+EtOH and pressurized ethanol, in
From the results presented in Table 6, at 40 °C and 20 MPa, the both methods of analysis (ABTS and DPPH), Table 6 shows that AA is
lowest TPC value among the solvents (scCO2+EtOH, scCO2 and pres- improved by increasing temperature and decreasing pressure, a beha-
surized ethanol) was 63.90 mg GAE/100 g oil obtained with scCO2. vior similar to that observed for TPC. The relationship between the
Adding ethanol in the extractor vessel with scCO2 at an ethanol to SCG values of AA and TPC are shown in Fig. 10, where a linear correlation
mass ratio (MRES) 0.5:1 the TPC increased from 63.90 to 272.43 mg between these variables with R² of 0.76 for the ABTS and R² method of
GAE/100 g oil, due to the ethanol polarity and consequent high affinity 0.86 for DPPH methods are observed. Such a relationship between TPC
for the analytes. At same condition (40 °C and 20 MPa), with MRES and AA is expected since the antioxidant activity of the SCG oil is at-
increasing from 0.5:1 to 2:1 in extractions using scCO2+EtOH, the TPC tributed to the phenolic compounds present in the oil. The antioxidant
increased from 272.43 to 294.47 mg GAE/100 g oil. Overall, the results activity of phenolic compounds increases the resistance of the oil to
indicate that increasing the amount of ethanol in the extractor vessel oxidation and increases its range of applications (Khmelinskii et al.,
increases the recovery of phenolic compounds. Thus, TPC values ob- 2019).
tained with scCO2+EtOH at MRES = 2:1 (294.47 to 392.96 mg GAE/ In general, it can be observed that at same experimental condition
100 g oil) were high and similar to those observed in pressurized (40 °C and 20 MPa), in Table 6, the addition of ethanol to scCO2
ethanol extraction (267.12 to 406. 12 mg GAE/100 g oil), however, in (scCO2+EtOH) enhanced the AA (ABTS and DPPH). Also, through in-
the extractions with scCO2+EtOH (2:1) smaller amounts of ethanol creasing the amount of ethanol in the extractor increased the AA by the
were used. In addition, it can be observed in Table 6 that the extractions ABTS method, and similar results were found for AA by the DPPH
with scCO2+EtOH and pressurized ethanol showed similar behavior, in method. The use of scCO2+EtOH resulted in AA ABTS values of
which high temperatures combined with low pressures produced high 3308.30–4609.53 μmol TEAC/100 g oil and AA DPPH of
TPC values for the SCG oil. 2691.25–3214.76 μmol TEAC/100 g oil, similar values to those ob-
TPC values for SCG oil (63.90–411.10 ± 18.45 mg GAE/100 g oil) tained with pressurized ethanol for AA ABTS (3904.57–4559.42 μmol
are classified as having human health benefits according to the TEAC/100 g oil) and AA DPPH (3199.03–3534.16 μmol TEAC/100 g

13
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

Table 7
Profile of phenolic acids and caffeine for spent coffee grounds oil extracted from different methods.
Solvent (MRES) Experimental condition Caffeine Dihydroxybenzoic acid Caffeic acid p-Coumaric acid Ferulic acid
(mg/100 g oil) (mg/100 g oil)

CXL
scCO2+EtOH(0.5:1) 40 °C/10 MPa 421.10 5.92 4.68 2.59 1.41
scCO2+EtOH(0.5:1) 40 °C/20 MPa 439.15 5.87 3.97 2.10 1.74
scCO2+EtOH(0.5:1) 60 °C/15 MPa 711.70 ± 45.17# 8.81 ± 0.46# 5.77 ± 2.08# 2.33 ± 0.33# 1.76 ± 0.09#
scCO2+EtOH(2:1) 80 °C/10 MPa 593.33 17.66 9.36 2.18 1.54
scCO2+EtOH(2:1) 40 °C/20 MPa 404.58 7.38 5.47 2.61 1.35
scCO2+EtOH(2:1) 80 °C/20 MPa 612.71 16.80 9.02 1.60 1.65
SFE
scCO2 40 °C/20 MPa 64.05 ND ND ND ND
PLE
EtOH 80 °C/10 MPa 451.05 ± 49.10# 13.53 ± 1.42# 9.08 ± 1.55# 3.44 ± 0.82# 1.49 ± 0.34#
EtOH 40 °C/20 MPa 196.36 3.85 2.87 1.25 0.93
EtOH 80 °C/20 MPa 393.51 11.2 7.69 2.75 1.34
Soxhlet
n-Hexane – 23.06 ND ND ND ND
AcEt – 24.94 4.8 2.34 0.24 1.31
Ethanol – 550.23 16.94 7.13 2.29 2.68

#
ND: not detected. Average and expanded uncertainties estimated with 95% of confidence level in triplicate experiments.

oil), but with lower amount of solvent and extraction time. Thus, the higher values than those observed in this work. On the other hand,
scCO2+EtOH extraction is presented as an interesting alternative Shang et al. (2016) obtained caffeine content of 30–130 mg/100 g SCG
aiming the spent coffee grounds oil extraction with high antioxidant using PLE extraction with ethanol, values similar to those obtained in
activities. the extractions of the present study (3–97 mg/ g SCG). These variations
may be related to the differences between the coffee blends used in each
study, since coffee has different caffeine contents depending on the
3.7. Phenolics compounds and caffeine
species (arabica or robust) (Dorsey and Jones, 2017).
The phenolic compounds and caffeine identified in the samples of
SCG oil with the highest overall extraction yield form the different 4. Conclusions
process (SFE, CXL, PLE and Soxhlet) are presented in Table 7. The
phenolic acids were detected in all samples except those obtained with In this work, the spent coffee grounds (SCG) oil extraction with
scCO2 and n-hexane (non-polar solvents). The main phenolic acid found scCO2+EtOH as solvent was studied and compared to the extractions
was dihydroxybenzoic acid, followed by the caffeic acid. The highest with supercritical carbon dioxide (scCO2), pressurized ethanol (EtOH)
concentrations of dihydroxybenzoic and caffeic acid were 17.66 mg/ and to the conventional Soxhlet technique. The results demonstrated
100 g oil and 9.36 mg/100 g oil, respectively, obtained with the technical feasibility of SCG oil extraction using scCO2+EtOH,
scCO2+EtOH (2:1) at 80 °C and 10 MPa. In the Soxhlet extraction, an where the addition of ethanol significantly enhanced the extraction
increase in phenolic acid content was observed with increasing polarity recovery. Among of the process parameters (temperature, pressure,
and boiling temperature of the solvent. For the extractions using mass ratio of ethanol to SCG, confinement time and ethanol purity), the
scCO2+EtOH and pressurized ethanol, it is observed in Table 7 that mass ratio of ethanol to SCG (MRES) was the most influent in the ex-
high temperatures and low pressures favor the recovery of phenolic traction yield and chemical profile of the oil sample obtained. At MRES
acids. In the extraction with scCO2+EtOH the increase of the MRES of 0.5:1 the highest overall extraction yield was 12.85% obtained at
resulted in an increase in the phenolic acid content. Also, the values of lower temperature (40 °C) and higher pressure (20 MPa), same condi-
phenolic acid content obtained with scCO2+EtOH (2:1) were higher tion where the highest result for the SFE extraction was obtained. The
than those obtained with pressurized ethanol. highest extraction yield obtained among the overall solvents was
The caffeine content in the SCG oil samples, in Table 7, ranged from 15.87% (in 25 min of extraction) obtained with scCO2+EtOH at MRES
23.06 to 711.70 mg/100 g oil. The highest caffeine content (711.70 mg/ 2:1, at highest temperature (80 °C) and pressure (20 MPa). The use of
100 g oil) was obtained with scCO2+EtOH (0.5:1) at 60 °C and 15 MPa. scCO2+EtOH allowed the extraction of oil with high yield from 10 to
Addition of ethanol to scCO2 (40 °C and 20 MPa) raised the caffeine 20 MPa, pressures where low yields and high times of extraction were
content from 64.05 to 439.15 mg/100 g oil. In the CXL extractions with observed in the SFE with scCO2. The yields of the CXL extraction were
scCO2+EtOH, the increase in temperature and pressure resulted in comparable to the yields obtained in the PLE but using lower amount of
higher caffeine contents. While in the PLE extraction with pressurized ethanol and the extraction time.
ethanol, higher contents of caffeine were obtained at higher tempera- The fatty acid profiles of the oils obtained by the different extraction
ture and lower pressure. The highest average caffeine content was ob- techniques (CXL, SFE, PLE and Soxhlet) were similar. The main fatty
tained with scCO2+EtOH (2:1). However, at 40 °C and 20 MPa, it was acids found in coffee grounds were linoleic (44.05 - 15.79%) and pal-
observed that the increase in MRES resulted in a decrease in the caf- mitic (29.70–33.04 %). The IV for the SCG oil (90.34–93.05 g I2/100 g
feine content. oil) were lower than the IV observed for other edible oils (sunflower oil,
Caffeine is considered the most consumed stimulant for the central peanut oil and corn oil) and were within the European standard for
nervous system, whether as coffee or tea drink, soft drinks or choco- biodiesel production.
lates. In addition, it is widely used as a supporting or agent in phar- The results of total phenolic content (TPC) and antioxidant activity
maceutical formulations. The consumption of moderate doses has sti- (AA) for the SCG oil samples obtained with scCO2+EtOH (2:1) were
mulating effect and reduces fatigue, without causing damages (Andrade similar to those obtained with pressurized ethanol, which among the
et al., 2012). The caffeine content obtained by Andrade et al. (2012) in overall solvents had the highest values. The extractions with
Soxhlet with n-hexane was 210 mg/100 g oil and in the extraction with scCO2+EtOH and pressurized ethanol also showed similar behavior, in
scCO2 values between 2720 and 4130 mg/100 g oil were obtained, which high temperatures combined with low pressures produced high

14
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

TPC and AA values for the spent coffee grounds oil. Crops Prod. 119, 49–56. https://doi.org/10.1016/j.indcrop.2018.04.008.
The main phenolic acids found were dihydroxybenzoic acid and El-Hadad, S.S., Tikhomirova, N.A., 2018. Physicochemical properties and oxidative sta-
bility of butter oil supplemented with corn oil and dihydroquercetin. J. Food Process.
caffeic acid. In addition, the highest concentrations of phenolic acids Preserv. 42, 1–7. https://doi.org/10.1111/jfpp.13765.
and caffeine on average were obtained with scCO2+EtOH (2:1). For the Fetzer, D.L., Cruz, P.N., Hamerski, F., Corazza, M.L., 2018. Extraction of baru (Dipteryx
extractions using scCO2+EtOH and pressurized ethanol the recovery of alata vogel) seed oil using compressed solvents technology. J. Supercrit. Fluids 137,
23–33. https://doi.org/10.1016/j.supflu.2018.03.004.
phenolic acids was favored by high temperatures and low pressures. Gomide, R., 1983. Operações com sistemas sólidos granulares. Operações unitárias 1,
In general, the spent coffee grounds oil obtained scCO2+EtOH (2:1) 27–30.
presented high yield, total phenolic content, antioxidant activity and Green, D.W., Perry, R.H., 2007. Perry’s Chemical Engineers’ Handbook, 8th ed. McGraw-
Hill, New York.
caffeine content, demonstrating the feasibility of the extraction process Herrero, M., Mendiola, J.A., Ibáñez, E., Turner, C., Wang, J., 2017. Gas expanded liquids
using CO2-expanded ethanol (CXL) to obtain oil from spent coffee and switchable solvents. Curr. Opin. Green Sustain. Chem. 5, 24–30. https://doi.org/
grounds. 10.1016/j.cogsc.2017.03.008.
ICO, 2018. Dados históricos. [WWW Document]. Int. COFFEE Organ.
Jenkins, R.W., Stageman, N.E., Fortune, C.M., Chuck, C.J., 2014. Effect of the type of
Acknowledgments bean, processing, and geographical location on the biodiesel produced from waste
coffee grounds. Energy Fuels 28, 1166–1174. https://doi.org/10.1021/ef4022976.
The authors thank the CNPq (Grants 406737/2013-4 and 305393/ Jessop, P.G., Subramaniam, B., 2007. Gas-expanded liquids. Chem. Rev. 107, 2666–2694.
https://doi.org/10.1021/cr040199o.
2016-2) and Fundação Araucária for financial support and scholarships. Kalantzakis, G., Blekas, G., Pegklidou, K., Boskou, D., 2006. Stability and radical-
This work was also financed in part by the Coordenação de scavenging activity of heated olive oil and other vegetable oils. Eur. J. Lipid Sci.
Aperfeiçoamento de Pessoal de Nível Superior – Brazil (CAPES) – Technol. 108, 329–335. https://doi.org/10.1002/ejlt.200500314.
Karmee, S.K., 2018. A spent coffee grounds based biorefinery for the production of bio-
Finance Code 001. fuels, biopolymers, antioxidants and biocomposites. Waste Manag. 72, 240–254.
https://doi.org/10.1016/j.wasman.2017.10.042.
References Khmelinskii, I., Pasqualone, A., Caponio, F., Squeo, G., Sikorska, E., Paradiso, V.M.,
Summo, C., 2019. Evaluation of total phenolic content in virgin olive oil using
fluorescence excitation–emission spectroscopy coupled with chemometrics. J. Sci.
Abdullah, M., Bulent Koc, A., 2013. Oil removal from waste coffee grounds using two- Food Agric. 99, 2513–2520. https://doi.org/10.1002/jsfa.9461.
phase solvent extraction enhanced with ultrasonication. Renew. Energy 50, 965–970. Knothe, G., 2002. Structure indices in FA chemistry. How relevant is the iodine value? J.
https://doi.org/10.1016/j.renene.2012.08.073. Am. Oil Chem. Soc. 79, 847–854. https://doi.org/10.1007/s11746-002-0569-4.
Ahangari, B., Sargolzaei, J., 2013. Extraction of lipids from spent coffee grounds using Kwon, E.E., Yi, H., Jeon, Y.J., 2013. Sequential co-production of biodiesel and bioethanol
organic solvents and supercritical carbon dioxide. J. Food Process. Preserv. 37, with spent coffee grounds. Bioresour. Technol. 136, 475–480. https://doi.org/10.
1014–1021. https://doi.org/10.1111/j.1745-4549.2012.00757.x. 1016/j.biortech.2013.03.052.
Akgün, N.A., Bulut, H., Kikic, I., Solinas, D., 2014. Extraction behavior of lipids obtained Liu, Y., Tu, Q., Knothe, G., Lu, M., 2017. Direct transesterification of spent coffee grounds
from spent coffee grounds using supercritical carbon dioxide. Chem. Eng. Technol. for biodiesel production. Fuel 199, 157–161. https://doi.org/10.1016/j.fuel.2017.02.
37, 1975–1981. https://doi.org/10.1002/ceat.201400237. 094.
Al-Hamamre, Z., Foerster, S., Hartmann, F., Kröger, M., Kaltschmitt, M., 2012. Oil ex- Low, J.H., Rahman, W.A.W.A., Jamaluddin, J., 2015. The influence of extraction para-
tracted from spent coffee grounds as a renewable source for fatty acid methyl ester meters on spent coffee grounds as a renewable tannin resource. J. Clean. Prod. 101,
manufacturing. Fuel 96, 70–76. https://doi.org/10.1016/j.fuel.2012.01.023. 222–228. https://doi.org/10.1016/j.jclepro.2015.03.094.
Andrade, K.S., Gonçalvez, R.T., Maraschin, M., Ribeiro-Do-Valle, R.M., Martínez, J., Mata, T.M., Martins, A.A., Caetano, N.S., 2018. Bio-refinery approach for spent coffee
Ferreira, S.R.S., 2012. Supercritical fluid extraction from spent coffee grounds and grounds valorization. Bioresour. Technol. 247, 1077–1084. https://doi.org/10.1016/
coffee husks: antioxidant activity and effect of operational variables on extract J.BIORTECH.2017.09.106.
composition. Talanta 88, 544–552. https://doi.org/10.1016/j.talanta.2011.11.031. Melo, M.M.R. de, Barbosa, H.M.A., Passos, C.P., Silva, C.M., 2014. Supercritical fluid
AOCS, 1997. American Oil Chemist’s Society. extraction of spent coffee grounds: measurement of extraction curves, oil character-
Araújo, O.A.S., Silva, F.R., Ramos, L.P., Lenzi, M.K., Ndiaye, P.M., Corazza, M.L., 2012. ization and economic analysis. J. Supercrit. Fluids 86, 150–159. https://doi.org/10.
Phase behaviour measurements for the system (carbon dioxide+biodiesel+ethanol) 1016/j.supflu.2013.12.016.
at high pressures. J. Chem. Thermodyn. 47, 412–419. https://doi.org/10.1016/j.jct. Mussatto, S.I., Ballesteros, L.F., Martins, S., Teixeira, J.A., 2011. Extraction of antioxidant
2011.11.029. phenolic compounds from spent coffee grounds. Sep. Purif. Technol. 83, 173–179.
Ballesteros, L.F., Ramirez, M.J., Orrego, C.E., Teixeira, J.A., Mussatto, S.I., 2017. https://doi.org/10.1016/j.seppur.2011.09.036.
Encapsulation of antioxidant phenolic compounds extracted from spent coffee Najdanovic-Visak, V., Yee, F., Lee, -Lam, Tavares, M.T., Armstrong, A., 2017. Kinetics of
grounds by freeze-drying and spray-drying using different coating materials. Food extraction and in situ transesterification of oils from spent coffee grounds. J. Environ.
Chem. 237, 623–631. https://doi.org/10.1016/j.foodchem.2017.05.142. Chem. Eng. 5, 2611–2616. https://doi.org/10.1016/j.jece.2017.04.041.
Balvardi, M., Mendiola, J.A., Castro-Gómez, P., Fontecha, J., Rezaei, K., Ibáñez, E., 2015. NIST, 2018. Isothermal Properties for Carbon Dioxide. [WWW Document]. .
Development of pressurized extraction processes for oil recovery from wild almond Oliveira, N.A. de, Santiago, H.P.C., Fukumasu, H., Oliveira, A.L. de, 2018. Green coffee
(Amygdalus scoparia). J. Am. Oil Chem. Soc. 92, 1503–1511. https://doi.org/10. extracts rich in diterpenes e process optimization of pressurized liquid extraction
1007/s11746-015-2708-2. using ethanol as solvent. J. Food Eng. 224, 148–155. https://doi.org/10.1016/j.
Barbosa, H.M.A., de Melo, M.M.R., Coimbra, M.A., Passos, C.P., Silva, C.M., 2014. jfoodeng.2017.12.021.
Optimization of the supercritical fluid coextraction of oil and diterpenes from spent Page, J.C., Arruda, N.P., Freitas, S.P., 2017. Crude ethanolic extract from spent coffee
coffee grounds using experimental design and response surface methodology. J. grounds: volatile and functional properties. Waste Manag. 69, 463–469. https://doi.
Supercrit. Fluids 85, 165–172. https://doi.org/10.1016/j.supflu.2013.11.011. org/10.1016/j.wasman.2017.08.043.
Benavides, A.H., Dorado, D., Camargo, A.D.P.S., 2016. Study of the fatty acid profile and Panusa, A., Zuorro, A., Lavecchia, R., Marrosu, G., Petrucci, R., 2013. Recovery of natural
the aroma composition of oil obtained from roasted Colombian coffee beans by su- antioxidants from spent coffee grounds. J. Agric. Food Chem. 61, 4162–4168.
percritical fluid extraction. J. Supercrit. Fluids 113, 44–52. https://doi.org/10.1016/ https://doi.org/10.1021/jf4005719.
j.supflu.2016.03.008. Pereira, C.G., Meireles, M.A.A., 2010. Supercritical fluid extraction of bioactive com-
Bozdogan Konuskan, D., Arslan, M., Oksuz, A., 2019. Physicochemical properties of cold pounds: fundamentals, applications and economic perspectives. Food Bioprocess
pressed sunflower, peanut, rapeseed, mustard and olive oils grown in the Eastern Technol. 3, 340–372. https://doi.org/10.1007/s11947-009-0263-2.
Mediterranean region. Saudi J. Biol. Sci. 26, 340–344. https://doi.org/10.1016/j. Pourmortazavi, S.M., Hajimirsadeghi, S.S., 2007. Supercritical fluid extraction in plant
sjbs.2018.04.005. essential and volatile oil analysis. J. Chromatogr. A 1163, 2–24. https://doi.org/10.
Campos-Vega, R., Loarca-Pi, G., Vergara-Casta, H.A., Dave Oomah, B., 2015. Spent coffee 1016/j.chroma.2007.06.021.
grounds: a review on current research and future prospects. Trends Food Sci. Ramos, M.J., Fernández, C.M., Casas, A., Rodríguez, L., Pérez, Á., 2009. Bioresource
Technol. 45, 24–36. https://doi.org/10.1016/j.tifs.2015.04.012. Technology Influence of fatty acid composition of raw materials on biodiesel prop-
Couto, R.M., Fernandes, J., da Silva, M.D.R.G., Simões, P.C., 2009. Supercritical fluid erties. Bioresour. Technol. 100, 261–268. https://doi.org/10.1016/j.biortech.2008.
extraction of lipids from spent coffee grounds. J. Supercrit. Fluids 51, 159–166. 06.039.
https://doi.org/10.1016/j.supflu.2009.09.009. Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yand, M., Catherine, R., 1999.
Döhlert, P., Weidauer, M., Enthaler, S., 2015. Spent coffee ground as source for hydro- Antioxidant activity applying an improved abts radical. Free Radic. Biol. Med. 26,
carbon fuels. J. Energy Chem. 25, 146–152. https://doi.org/10.1016/j.jechem.2015. 1231–1237. https://doi.org/10.1016/S0891-5849(98)00315-3.
11.012. Reverchon, E., De Marco, I., 2006. Supercritical fluid extraction and fractionation of
Dorsey, B.M., Jones, M.A., 2017. Healthy components of coffee processing by-products. natural matter. J. Supercrit. Fluids 38, 146–166. https://doi.org/10.1016/j.supflu.
In: Galanakis, C.M. (Ed.), Handbook of Coffee Processing By-Products: Sustainable 2006.03.020.
Applications. Academic Press, Cambridge, pp. 27–62. https://doi.org/10.1016/B978- Rocha, M.V.P., de Matos, L.J.B.L., Lima, L.P. de, Figueiredo, P.M. da S., Lucena, I.L.,
0-12-811290-8.00002-5. Fernandes, F.A.N., Gonçalves, L.R.B., 2014. Ultrasound-assisted production of bio-
Efthymiopoulos, I., Hellier, P., Ladommatos, N., Russo-Profili, A., Eveleigh, A., Aliev, A., diesel and ethanol from spent coffee grounds. Bioresour. Technol. 167, 343–348.
Kay, A., Mills-Lamptey, B., 2018. Influence of solvent selection and extraction tem- https://doi.org/10.1016/j.biortech.2014.06.032.
perature on yield and composition of lipids extracted from spent coffee grounds. Ind. Rodrigues, G., de, M., Cardozo-Filho, L., Silva, Cda, 2017. Pressurized liquid extraction of

15
M.N. Araújo, et al. Industrial Crops & Products 141 (2019) 111723

oil from soybean seeds. Can. J. Chem. Eng. 95, 2383–2389. https://doi.org/10.1002/ 10.1016/j.ces.2013.12.025.
cjce.22922. Sluiter, A., Hames, B., Hyman, D., Payne, C., Ruiz, R., Scarlata, C., Sluiter, J., Templeton,
Rodrígues, P.C., Mendiola, J.A., Quirantes-Piné, P.R., Ibáñez, E., Segura, C.A., 2016. D., Nrel, J.W., 2008. Determination of total solids in biomass and total dissolved
Green downstream processing using supercritical carbon dioxide, CO 2 -expanded solids in liquid process samples. Natl. Renew. Energy Lab. 9. https://www.nrel.gov/
ethanol and pressurized hot water extractions for recovering bioactive compounds docs/gen/fy08/42621.pdf.
from Moringa oleifera leaves. J. Supercrit. Fluids 116, 90–100. https://doi.org/10. Toda, T.A., Sawada, M.M., Rodrigues, C.E.C., 2016. Kinetics of soybean oil extraction
1016/j.supflu.2016.05.009. using ethanol as solvent: experimental data and modeling. Food Bioprod. Process. 98,
Shang, Y.F., Xu, J.L., Lee, W.J., Um, B.H., 2016. Antioxidative polyphenolics obtained 1–10. https://doi.org/10.1016/j.fbp.2015.12.003.
from spent coffee grounds by pressurized liquid extraction. S. Afr. J. Bot. 109, 75–80. Xu, H., Wang, W., Liu, X., Yuan, F., Gao, Y., 2015. Antioxidative phenolics obtained from
https://doi.org/10.1016/j.sajb.2016.12.011. spent coffee grounds (Coffea arabica L.) by subcritical water extraction. Ind. Crops
Singleton, V.L., Rossi, J.A., 1965. Colorimetry of total phenolics with phosphomolybdic- Prod. 76, 946–954. https://doi.org/10.1016/j.indcrop.2015.07.054.
phosphotungstic acid reagents. Am. J. Enol. Vitic. 16, 144–158. Zuorro, A., Lavecchia, R., 2012. Spent coffee grounds as a valuable source of phenolic
Siougkrou, E., Galindo, A., Adjiman, C.S., 2014. On the optimal design of gas-expanded compounds and bioenergy. J. Clean. Prod. 34, 49–56. https://doi.org/10.1016/j.
liquids based on process performance. Chem. Eng. Sci. 115, 19–30. https://doi.org/ jclepro.2011.12.003.

16