Green Extraction Methods for Active Compounds from Food Waste—Cocoa Bean Shell

This is the first report on the extraction of cocoa bean shell (CBS) using deep eutectic solvents (DESs). Screening results with 16 different choline chloride-based DESs showed how choline chloride:oxalic acid DES was the most suitable solvent for the extraction of the bioactive compounds from CBS and that concentrations varied greatly depending on the used solvent. The DES extraction was compared to the DESs coupled with microwave extraction (MAE), and the yields of the extracted compounds were higher for DES/MAE. For theobromine, the obtained yields for DES extraction were 2.145–4.682 mg/g, and for caffeine, were 0.681–1.524 mg/g, whereas for DES/MAE, the same compounds were obtained in 2.502–5.004 mg/g and 0.778–1.599 mg/g. Antioxidant activity was also determined, using DPPH method, obtaining 24.027–74.805% activity for DES extraction and 11.751–55.444% for DES/MAE. Water content significantly influenced the extraction of targeted active compounds from CBS, whereas extraction time and temperature did not show statistically significant influence. The extraction temperature only influenced antioxidant activity. The study demonstrated how extraction using DES and microwaves could be of a great importance in the future trends of green chemistry for the production of CBS extracts rich in bioactive compounds.


Introduction
To minimize food waste during processing of raw materials, a new age tendency is to transform it into valuable products with certain nutritional value. Cocoa bean shell (CBS) is just one of these food industry by-products, which contains a number of bioactive compounds that can be extracted and used for different purposes. In the production of chocolate and its products, the CBS is removed from the cocoa bean cotyledons, during the pre-roasting or after the roasting process. In the fermentation, roasting, and alkalization process, some bioactive compounds, like theobromine and phenols, can migrate from cocoa bean to CBS [1,2]. Thus, enriched CBS could potentially be an interesting ingredient in other production processes [1].
The chemical composition of CBS mostly depends on the origin of Theobroma cacao L. and its processing. Okyama et al. [2] in their review study mentioned Africa, America, Asia, and Oceania as the major parts in the world for the production of cocoa beans.
To address the environmentally friendly extraction protocols, we have employed DESs for the extraction of different constituents of CBS, as well as the combination of DESs and MAE extraction techniques on the same matrix. In distinction from the existing papers, the present research provides novelty with respect to (a) detail extraction of cocoa bean shell (CBS) using deep eutectic solvents (DESs) enhanced by microwave-assisted extraction (MAE); (b) High Performance Liquid Chromatography (HPLC) analysis of methylxantines and polyphenols in obtained extracts; and (c) optimization of the main extraction parameters using response surface methodology (RSM), based on the antioxidant capacity and the concentration of each individually selected active compound.

Material
Cocoa bean shell (CBS) material was obtained from Kandit d.o.o. Chocolate Factory, Osijek, Croatia, in the summer of 2017. Before it was delivered from the factory, CBS was primarily separated from the cotyledon after the roasting process which was done at 135 • C for 55 min. The countries of geographical origin of CBS were Ghana, Cameroon, Nigeria, New Guinea and Ivory Coast and this is known as West Africa blend.

Preparation of Deep Eutectic Solvents (DES)
DESs were prepared according to our previous work [35]. Choline chloride as a hydrogen bond acceptor (HBA) was mixed with different hydrogen bond donors (HBD), as indicated in the Table 1. The mixture was stirred and heated at 80 • C until a clear liquid was formed and used in further extraction experiments as such.

Extraction of Bioactive Compounds from CBS
Prior to the extraction, CBS was grounded on the standard laboratory mill. The extraction of CBS was performed by weighting 50 mg of the grinded CBS in the centrifuge tube vial and adding 1 mL of the solvent (10% or 50% H 2 O/DES mixture). The extraction was performed at 50 • C during 60 min and by stirring. Afterwards, the extracts were centrifuged and decanted. Screening analysis was performed in sixteen different choline chloride-based DESs, where choline chloride was combined with different hydrogen bond donors according to the Table 1. Furthermore, when the best DES was chosen for this type of extraction, 17 runs, including five replicates, were performed according to the Box-Behnken design by RSM (Table 2) [35].

DESs Extraction Coupled with MAE
The extraction of bioactive compounds from Theobroma cacao L. CBS, a byproduct in the production of cocoa and its products, by DES extraction coupled with microwave-assisted extraction (MAE) was performed using Milestone flexiWAVE (Milestone Srl, Sorisole (BG), Italy) microwave system, equipped with rotating carousel with 15 positions for PTFE high-pressure vessels. The microwave power was set to 600-800 W. A 500 mg of CBS was weighted in the high-pressure vessel, followed by the addition of 10 mL of the solvent (10% or 50% H 2 O/DES mixture). All extraction experiments were performed at the same extraction conditions as the DES extraction with stirring (Table 3).

HPLC Analysis
Identification and quantification of bioactive compounds in CBS extracts was done according to modified Alves de Oliveira Nascimento et al. [36] study. After the extraction, and prior to HPLC analysis, extracts were diluted 1:10 with water and filtered through 0.2 µm polytetrafluoroethylene (PTFE) syringe filter. The measurement was done on the High-Performance Liquid Chromatograph (HPLC) (Infinity 1260 Agilent Technologies, Santa Clara, USA), which contained an autosampler G7129A, quaternary pump G7111B 1260, and diode array detector (DAD) G7117C 1260 DAD HS. A column used was Zorbax C 18 150 mm x 4.6 mm x 5µm, which was thermostated on 30 • C. The wavelength used for the measurement was 276 nm while the injection volume was set to 20 µL. A gradient mobile phase was used, starting by 1% formic acid and acetonitrile (95:5) at the beginning, changing to (80:20) till 9 min, and returning to (95:5) till 13 min. The flow of the mobile phase was set to 1 mL/min, and the analysis was performed in triplicate.

DPPH Scavenging Activity
Antioxidant activity of the obtained CBS extracts was determined using DPPH radical. The solution mix for the measurement was prepared by adding 1.2 mL of a CBS extract (10 mg/mL) and 0.5 mL of 0.2 mM DPPH solution freshly prepared in methanol.
After 30 min of incubation in the dark and at ambient temperature, the apsorbance (A) was measured at 517 nm on the UV-Visible Spectrophotometer (Termo Spectronic, Cambridge, Great Britain) and DPPH scavenging activity was determined according to the Equation (1): where A DPPH was the absorbance of the control (instead of the sample, methanol was added); A b was the absorbance of the sample, where instead of DPPH methanol was added and A s was the absorbance of the sample mixed with DPPH radical solution [37]. All measurements were performed in triplicate.

Statistical Analysis
The commercial Design-Expert ® software (ver. 9, Stat-Ease Inc., Minneapolis, MN, USA) was used for the statistical analysis of obtained experimental data. To estimate the quality of the obtained models, the analysis of variance (ANOVA) was used. The test of the statistically significant difference was based on the total error criteria with the level of confidence of 95.0%. The response plots were generated using the same software for better understanding the correlation of independent and response variables.

DES Extraction of Bioactive Compounds from CBS
This research is the first detailed report on the extraction of CBS using DESs. In our previously study [25], only a short insight into DES extraction was given, where only one DES was described, namely, choline chloride:oxalic acid (0, 25, and 50% of water and 60, 180, and 360 min). There we indicated that the time of 60 min is sufficient for the extraction of bioactive compounds from plant material (theobromine 3.700 mg/g and 3.644 mg/g, and caffeine 0.507 mg/g and 0.536 mg/g with 25% and 50% H 2 O), whereas the energy consumption during heating is also reduced in shorter time. Furthermore, some of the DES possess a high viscosity at room temperature, which can be reduced by the increased temperature and the addition of water. Both thermolabile compounds and DES can be degraded at high temperatures, so the temperature of 50 • C was applied. Hereby, in this paper, we investigated a larger number of DESs, followed by the optimization of the process parameters.
In general, DESs can be formed from various HBAs and HBDs, and their combination, as well as the ratio, affects their physical and chemical properties greatly. Due to the different chemical and physical properties, they can exhibit a different extraction potential of various bioactive compounds. Therefore, an extraction of selected compounds in this study was performed in 16 different DESs. Our research was conducted utilizing choline chloride based DESs, where choline chloride was combined with different HBDs, as indicated in the Table 1. A screening was performed with the addition of water to each DESs (10% and 50% v/v) for 60 min, at temperature of 50 • C.
A screening experiment revealed that the extracted amount of bioactive compounds varied greatly depending on the type of the solvent, as expected. Gallic acid was extracted only with ChCl:AA (0.009 and 0.016 mg/g), ChCl:EG (0.008 and 0.002 mg/g) and ChCl:Glu (0.009 mg/g). The highest content of theobromine was extracted with ChCl:BD (3.639 mg/g), ChCl:OA (3.605 mg/g), and ChCl:U (3.613 mg/g) with 50% of water, whereas the lowest amount of theobromine was extracted with ChCl:OA (0.620 mg/g) with 10% of water. Catechin was extracted in the highest amount with ChCl:Fru (0.095 mg/g) with 50% of added water, whereas ChCl:MA, ChCl:Lev, ChCl:CA, ChCl:Mal, ChCl:LA, ChCl:OA, and ChCl:TA yielded no catechin. Caffeine was found in the highest amount when extracted with ChCl:OA (0.909 mg/g) with 50 % of water and ChCl:Lev (0.916 mg/g) with 10 % of water, and in the lowest amount in ChCl:OA (0.156 mg/g) with 10 % of water. Caffeic acid was the most abundant in the ChCl:Sor (0.04 mg/g) with 50% of water and in the lowest concentration with ChCl:OA (0 mg/g) with 10% of water. The amount of epicatechin was the highest in ChCl:Mal (0.114 mg/g) with 50% of water, whereas the lowest was in ChCl:OA (0 mg/kg) with 10% of water. Futhermore, the best antioxidant activity, expressed as DPPH scavenging activity, was obtained with ChCl:OA, 66.307% with 50% of water and 64.322% with 10% of water. When compared to the ultrasound assisted water extraction of CBS, performed in our previous research [25], obtaining theobromine yield of 4.44-5.61 mg/g and caffeine content of 0.537-0.591 mg/g, we can conclude that DESs possess a comparable efficacy for the extraction of selected compounds.
Given all the data, and the fact that some compounds are not always desirable in the obtained extracts, we have decided to optimize the extraction conditions of selected compounds in ChCl:OA DES. This DES was chosen because the influence of the water addition on the amount of bioactive compounds was the highest, thus being the most suitable for tuning the parameters for extraction of desired compounds. Therefore, both CBS extracts, with either high or low theobromine or caffeine content, can be obtained with the same solvent, only varying the water content and based on this parameter one can tune the desirable properties of the selected DES. Furthermore, the capability of the solvent to extract different undesirable compounds from the material, offers an opportunity for Foods 2020, 9, 140 7 of 14 these solvents to be used in the pretreatment process of the material prior to its further utilization for different purposes.
The optimization of the extraction parameters was performed in a combination of 17 different experiments according to BBD, varying the temperature, extraction time, and the amount of water. Furthermore, for comparison, all experiments were performed by stirring ( Table 2) and under microwaves (Table 3), applying the same DES and process parameters.
Above all, the results represented in Tables 2 and 3 showed how DES extraction coupled with MAE gave slightly higher concentrations of selected bioactive compounds in CBS extracts. Some compounds that were found in traces after DES extraction, gave better yields after using DES/MAE, like gallic acid. These results were expected, as it is well established that the application of microwaves in many chemical processes, including extraction, can enhance the yield, shorten the reaction time, and reduce the overall energy consumption. This is well explained by the thermal and non-thermal effects of the microwaves explained in detail by de la Hoz, Diaz-Ortiz, and Moreno [38]. Furthermore, by observing the obtained concentrations of those bioactive compounds (Tables 2 and 3) it is noticeable how different parameters variously influence on the amount of every specific compound. The highest theobromine and caffeine concentrations were found by applying 60 • C and 30% of water during 10 min of extraction time (5.004 mg/g and 1.599 mg/g), as well as the highest DPPH scavenging activity (55.444%). There are many data in the literature describing the amount of available theobromine in the CBS. Adamafio [39] stated in his review how theobromine content in CBS can vary from 5-21 mg/g according to different authors. Arlorio et al. [40] determined the amount of theobromine in cocoa hulls, which was 12.9 ± 1.8 mg/g d.w. , whereas Sotelo and Alvarez [41] performed a water extraction and determined that the amount of theobromine in CBS was 0.223 mg/g of dry sample in fermented samples and 0.174 mg/g in not fermented samples, while caffeine was 0.056 mg/g in fermented, and 0.051 mg/g in non-fermented samples. Prabhakaran Nair [42] claims that the amount of theobromine in CBS is 1.3 % and caffeine 0.1 %. Okiyama et al. [20] performed a pressurized liquid extraction with absolute ethanol and found that the amount of theobromine was 9.89 ± 0.09 mg/g, epicatechin 3.5 ± 0.1 mg/g, caffeine 1.44 ± 0.01 mg/g, and catechin 0.178 ± 0.003 mg/g. The highest content of caffeic acid in our research, was found by applying 30 • C and 50% of water content, during 10 min of extraction time (0.137 mg/g), whereas catechin and epicatechin were 0.065 and 0.107 mg/g, respectively.

Response Surface Analysis and Process Optimization
To investigate the influence of the process parameters on the extraction yield and antioxidant activity of selected compounds a response surface analysis was performed.
As gallic acid, theobromine, catechin, caffeine, and caffeic acid were obtained in higher yields applying DES with microwaves, the ANOVA was performed considering the DES/MAE of methylxantines, catechins, and antioxidant activity. The results are summarized in Table 4, Table 5, Table 6 and Table 7. For theobromine, only effect of water content in DES was significant (p < 0.05) in first order linear and quadratic effect (X 3 and X 3 2 ). The model F-value of 5.77 implies the model is significant, and a p-value of 0.0153 indicates that model terms are significant (Table 4).   Catechin shows the same trend as theobromine, where only the linear effect of the water content (p = 0.0302) as well as the quadratic effect of the temperature (p = 0.0059) had a significant influence ( Table 5). The model F-value of 3.80 implies the model is significant. There is only a 4.61% chance that an F-value this large could occur due to noise. p-values less than 0.05 indicate model terms are significant like in this case. The Lack of Fit F-value of 2.20 implies the Lack of Fit is not significant relative to the pure error. Nonsignificant lack of fit (p = 0.2306) is good and implies the good model fitting.
The model F-value of 10.01 implies the model for caffeine is significant (Table 6). There is only a 0.31% chance that an F-value this large could occur due to noise. The model p-value is less than 0.05, indicating that model is statistically significant. In this case, the linear term of water content (X 3 ) as well as quadratic term (X 3 2 ) are significant model terms for caffeine content. The Lack of Fit F-value of 2.41 implies the Lack of Fit is not significant relative to the pure error. There is a 20.71% chance that a Lack of Fit F-value this large could occur due to noise. Nonsignificant lack of fit in the case of caffeine is desirable. ANOVA for antioxidant activity (Table 7) in CBS extracts show that model is significant (p = 0.0211) and that linear term of temperature (X 1 ) as well as quadratic term of temperature (X 1 2 ) and time (X 2 2 ) influenced significantly of investigated response. In this case, water content had no significant influence. The lack of fit F-value of 1.52 implies the lack of fit is not significant relative to the pure error. Summarizing all the obtained results, it can be concluded that water content had the strongest influence on all investigated active compounds in CBS extracts (p-value is <0.05), whereas the extraction temperature only influenced an antioxidant activity.
Three-dimensional plots are given for theobromine in Figure 1, for caffeine in Figure 2, and for catechin in Figure 3. For two most abundant methylxanthines in CBS three dimensional plots ( Figure 1 and Table 2) showed very similar shapes. The plots show that by increasing water content, concentration of those two compounds significantly increase. As theobromine has a limited solubility in water [23], two opposite effects of water probably influenced its extractability. It could be assumed that the higher content of water decreases theobromine extractability due to its lower solubility, whereas at the same time, it causes a lower viscosity of the solvent, thus increasing its diffusion from the CBS. In addition, when water is added in higher amounts to the DES, it decreases a hydrogen bond formation between DES and the bioactive compound, thus decreasing its extractability [43]. Opposite to that, temperature and extraction time had no significant influence on methylxantine concentrations in obtained CBS extracts. All three parameters (water content, time, and temperature) significantly influenced on the DPPH scavenging activity (Figure 4). The increased DPPH activity can be attributed to the higher amount of antioxidants in the extracts. Bajkacz et al. [44] stated that the increase in water content (more than 10%) in DES increased the flavonoid recovery from food samples. According to them, the best yields were obtained with 30% water content, while the further increase up to 75% decreased the flavonoid content.
obtained CBS extracts. All three parameters (water content, time, and temperature) significantly influenced on the DPPH scavenging activity (Figure 4). The increased DPPH activity can be attributed to the higher amount of antioxidants in the extracts. Bajkacz et al. [44] stated that the increase in water content (more than 10%) in DES increased the flavonoid recovery from food samples. According to them, the best yields were obtained with 30% water content, while the further increase up to 75% decreased the flavonoid content.   obtained CBS extracts. All three parameters (water content, time, and temperature) significantly influenced on the DPPH scavenging activity (Figure 4). The increased DPPH activity can be attributed to the higher amount of antioxidants in the extracts. Bajkacz et al. [44] stated that the increase in water content (more than 10%) in DES increased the flavonoid recovery from food samples. According to them, the best yields were obtained with 30% water content, while the further increase up to 75% decreased the flavonoid content.       The equations, in terms of coded factors (Table 8), can be used to make predictions about the response for given levels of each factor. By default, the high levels of the factors are coded as +1 and the low levels are coded as −1. The coded equation is useful for identifying the relative impact of the factors by comparing the factor coefficients.  The equations, in terms of coded factors (Table 8), can be used to make predictions about the response for given levels of each factor. By default, the high levels of the factors are coded as +1 and the low levels are coded as −1. The coded equation is useful for identifying the relative impact of the factors by comparing the factor coefficients. Table 8. Polynomial equations calculated after implementation of BBD (in terms of coded factors).

Conclusions
The combination of extraction with DESs and MAE of CBS for the purpose of isolation targeted bioactive compounds proved to be highly efficient. The results showed how only the water content in DES significantly influenced the extraction of methylxantines (theobromine and caffeine) from CBS, whereas the linear effect of the water content as well as the quadratic effect of the temperature had a significant influence on catechin. On the DPPH scavenging activity, linear and quadratic terms of temperature as well as quadratic term of time have statistically significantly influence. The study proves how these novel green extraction methods could definitely be included in the future trends of supplying and producing enriched extracts especially from food by-products like CBS which accumulation is increasingly problem in today's world. Obtained extracts rich in various bioactive compounds could further be used as a raw material in food, pharmaceutical, and chemical industries.