Green Extracts from Co ﬀ ee Pulp and Their Application in the Development of Innovative Brews

: Co ﬀ ee pulp, a by-product of co ﬀ ee production, contains valuable compounds such as ca ﬀ eine and chlorogenic acid with high antiradical activity. In this study, aqueous solutions of β -cyclodextrin ( β -CD) were used as a non-conventional solvent for the extraction of targeted compounds from co ﬀ ee pulp. The parameters of β -CD concentration (C β cd ), liquid-to-solid ratio (L / S), and temperature (T) were evaluated based on the antiradical activity (A AR ) and the ca ﬀ eine content (C Caf ). The optimum operational conditions were found to be C β cd : 9.25 mg / mL, L / S: 30 mL / g and T: 80 ◦ C. The sensory proﬁles of brews prepared with co ﬀ ee and co ﬀ ee pulp with or without cyclodextrin were studied with quantitative descriptive analysis. The brew from the by-product had fruity, botanic, sweet and sourness sensory properties, and cyclodextrin was found to be able to a ﬀ ect the overall taste of the brew. The same observed in tea infusion, in cyclodextrin selective in the extraction of compounds from tea [34]. These results give a new insight in the contribution of cyclodextrin in taste differentiation.


Introduction
Agricultural production leads to the accumulation of agro-food wastes and by-products. These wastes can be reused as they are a source of bioactive compounds such as phenolics that can be extracted by several conventional and non-conventional techniques.
Coffee is one of the most consumed products in the word; world coffee production for 2018/19 was estimated to be about 170,937,000 bags (60 kg/bag) [1]. For the preparation of coffee, coffee grounds must be released from the coffee cherry. During this step, many parts of the cherry are thrown away. Coffee pulp is the main by-product; other by-products of coffee production include husks, silver skin, and spent waste coffee. It is estimated that around one ton of coffee pulp is produced during the manufacture of two tons of coffee [2]. Coffee pulp is rich in cellulose, hemicellulose, phenolic compounds, and caffeine [3]. Chlorogenic acid is the main phenolic compound in coffee pulp contributing to the sensory attributes of coffee [4,5]. Specifically, it is responsible for the bitter, sour, and astringency tastes. During roasting, chlorogenic acid turns into lactones and contributes to the bitter taste of the brew [6,7]. Caffeine is another significant compound in coffee, contributing 10-30% of the final bitter taste of the brew [8].
Solid-liquid extraction is the most common technique for the extraction of polyphenols. The solvent penetrates into the plant tissue and contributes to the dilution of the bioactive compounds [9,10]. temperatures (T, 30-80 • C), cyclodextrin (CD) concentrations (C βcd , 0-18.5 mg/mL), and liquid-to-solid ratio values (L/S, 13-47 mL/g). The experimental conditions were selected based on data from the literature. A three-factor, five-level central composite rotatable design (2 3 + star) was used. This design consisted of three groups of design points, including two-level factorial design points, axial or star points, and center points. Therefore, the three selected independent variables were studied at five different levels, coded as −α, −1, 0, 1, and +α (Table 1). The value for alpha (1.68) was chosen to fulfill the rotatability in the design. According to the central composite design matrix, a total of 20 experiments was required ( Table 2). The caffeine content (C Caf ) and the antiradical activity (A AR ) were chosen as the dependent variables.

Extraction Process
Ground coffee pulp was mixed with an aqueous solution of β-cyclodextrin of different concentrations (C βcd , 0-18.5 mg/mL), in different liquid-to-solid ratios (L/S, 13-47 mL/g), at different temperatures (T, 30-80 • C) in a stoppered glass bottle based on the experimental conditions presented in Table 2. The material was subjected to extraction under stirring at 600 rpm. In all experiments, the extracts were collected after 120 min. The time of extraction was selected based on previous experiments. Following extraction, samples were centrifuged in a bench centrifuge (Hermle Z300K, Germany) at 9000 rpm for 5 min and were separated under vacuum filtration. The clear supernatant was stored in the refrigerator until used for further analysis. Table 2. Measured and predicted values of extracts antiradical activity (A AR , µmol TRE * /g) and caffeine extraction yield (C Caf , mg/g) for the individual design points.

Independent Variables Responses
Design Point C βcd (mg/mL)

Determination of Antiradical Activity (A AR )
The antiradical activity was determined according to a previously described protocol [21]. Briefly, an aliquot of 0.025 mL of extract was added to 0.975 mL DPPH • solution (100 µM in MeOH) and the absorbance at 515 nm was read at 0 and 30 min. Trolox ™ equivalents (mM TRE) were determined from a linear regression, after plotting %∆A 515 of known solutions of Trolox ™ against concentration, where: (1) Results were expressed as µmol TRE per g of coffee pulp weight.

Determination of Caffeine Extraction Yield
Caffeine quantitation was based on a previously described protocol [22]. Briefly, 12 mL of extract was mixed with an equal amount of dichloromethane. Then, using a separatory funnel, the caffeine was extracted using dichloromethane. The caffeine extraction yield (C Caf ) was expressed as mg caffeine per g of coffee pulp.

Determination of Total Polyphenol Yield (Y TP )
The total phenolic content of the brews was determined according to a protocol [21] using the Folin-Ciocalteu methodology. Yield in total polyphenols (Y TP ) was expressed as mg gallic acid equivalents (GAE) per g of coffee pulp weight.

Determination of Caffeine and Chlorogenic Acid Content
The quantification of caffeine and chlorogenic acid in brews was performed by ultra-high-performance liquid chromatography (UHPLC) using an ECOM spol. s r.o., Czech Republic system (model ECS05). The system is comprised of a quaternary gradient pump (ECP2010H), a gradient box with degasser (ECB2004), a column heating/cooling oven (ECO 2080), an autosampler (ECOM Alias) and diode array detector (ECDA2800 UV-VIS PDA Detector). A Phenomenex ® reversed-phase column (Synergi ™ Max-RP 80 Å; 4 µm particle size, 150 × 4.6 mm) was used at 25 • C. The sample injection volume was 10 µL. Chromatographic analysis was performed using a gradient of Milli-Q water with 0.1% formic acid (solvent A) and methanol with 0.1% formic acid (solvent B), at constant flow rate of 0.5 mL/min. The gradient program was as follows: solvent A was decreased from 70% to 55% after 5 min; followed by another decrease to 35% until 15 min; while it was finally reduced to 10% at 18 min. Then, solvent A was maintained at 10% for 2 min and returned to initial conditions (70% solvent A). Detection was accomplished with the diode array detector and chromatograms were recorded at 276 nm for caffeine and 330 nm for chlorogenic acid. Identification of caffeine and chlorogenic acid was performed by comparing the retention time and the UV-Vis spectra with those of reference standards, while quantification was established with the aid of calibration curves (Equations (2) and (3), respectively). All chromatographic data were analyzed using Clarity Chromatography Software v8.2 (DataApex Ltd.).

Preparation of Brews
A set of four samples was prepared for the sensory evaluation. The applied experimental conditions are displayed in Table 3 for design point 13, as these conditions were closest to the conditions of the cold brewing technique. All brews were prepared by using a liquid-to-solid ratio of 30 mL/g, an extraction time of 2 h, and a temperature of 30 • C. Two of the four samples were from Arabica coffee seeds (AQC) while the other two samples (AQCW) were made by using coffee pulp as the main extractable raw material. In each subset of samples, β-cyclodextrin was dissolved in water at a concentration of 9.25 mg/mL in one of the two samples prior to the start of the coffee extraction (AQC/CD and AQCW/CD). Detailed sample formulation is presented in Table 3. Subsequently, the extracts were collected by filtration and stored in a refrigerator until the sensory session. For each panelist, a quantity of 20 mL was served at room temperature in a plastic cup with a lid. The samples were randomly presented in four hourly sessions on three consecutive days. AQC: aqueous extract from coffee seeds, AQC/CD: β-cyclodextrin extract from coffee seeds, AQCW: aqueous extract from coffee pulp, AQCW/CD: β-cyclodextrin extract from coffee pulp.

Quantitative Descriptive Analysis
Nine trained panelists from the Department of Food Science and Technology (Aristotle University of Thessaloniki, Greece), who had already participated in several trained panel studies for other food products and with at least 1 year of experience in sensory evaluation, developed a consensus vocabulary of 11 descriptors using a Quantitative Descriptive Analysis approach. The attributes were categorized under the modalities of aroma and taste and reference standards were used when required. For a better explanation of the developed attributes, the reference standards are presented in Table 4. The terms botanic, fruity, earthy, and roasted were used to describe the aroma characteristics, whereas the words bitter, sweet, sourness, sour-roasted, botanic, earthy, and astringency were used to characterize the taste of the samples. Table 4. Developed attributes and references standards in the sensory evaluation.

Modality Descriptor Definition
Odor Botanic Characteristic odor associated with typical dried black tea notes Fruity Overall odor associated with floral, sweet, ripe fruits and characteristic odor of coffee pulp Earthy Odor associated with bread and wet soil Roasted Characteristic odor of over-roasted hazelnuts (220 • C/10 minutes) Taste Bitter The fundamental sensation associated with caffeic acid Sweet The fundamental sensation associated with sucrose Sourness Taste associated with sour/fermented-like aromatics Sour-roasted like Characteristic acidic, sharp and pungent taste associated with excessively roasted coffee beans Botanic Characteristic taste of black tea infusion Earthy Characteristic taste associated with bread crust Tactile sensation Astringent A dry penetrating sensation in the nasal cavity Attributes were scored using unstructured line scales (0-100) and panelists were seated in booths under appropriate environmental conditions (in accordance with International Organization for Standardization recommendation (ISO 8589:2010). Panelists individually rated samples in duplicate on two separate examinations. Samples were presented monadically according to a balanced design, labeled with arbitrary three-digit codes in opaque white plastic cups and maintained at ambient temperature before serving for evaluation. Bottled water and unsalted crackers were provided as palate cleanser between samples.

Statistical Analysis
The response surface methodology (RSM) data were analyzed using the statistical software MINITAB (release 13.32). To identify the significance of the effects and interactions between them, analysis of variance (ANOVA) was performed for each p-value less than 0.05, which was considered to be statistically significant. Regression analysis was used to fit a full second order polynomial to the data of the response variables. To evaluate the goodness of fit of each model, two criteria were used: the coefficient of determination, R 2 , which is the relative variance explained by the model with respect to the total variance. Sensory data were analyzed using the SENPAQ software (Qi Statistics, Ruscombe, UK). A two-way model analysis of variance (ANOVA) was performed; panelists were treated as random effects and samples and replicates as fixed effects. Multiple pairwise comparisons were conducted by Tukey's method and a significant difference was determined at an alpha risk of 5% (p ≤ 0.05). All determinations were carried out in triplicate. The values obtained were averaged. Figure 1 presents the effect of the process variables on the responses A AR and C Caf . In the case of A AR , it can be observed that an increase of C βcd generally leads to increased values of A AR . Similar studies on plant extracts rich in phenolics showed that the extract antiradical activity increased with the concentration of β-cyclodextrin in the solution [23,24], since the existence of β-CD can lead to inclusion complex formation with phenolic compounds. According to [9], the protection of phenolic compounds with β-CD can contribute to increased values of A AR . However, the β-CD concentration does not seem to affect the concentration of caffeine in the extract and it can be observed that the highest C Caf was obtained in the absence of cyclodextrin. The inclusion complex formation of caffeine has already been studied [25] and the usage of cyclodextrin has been proposed for the decaffeination process [26]. In complicated extracts, it is difficult to figure out which specific compounds form complexes [13]. It seems that the presence of cyclodextrin preferably boosted the extraction of more hydrophobic polyphenol molecules than caffeine [27]. The liquid-to-solid ratio (L/S) has a great impact on the examined responses. The extraction yield increased with increasing ratio. A similar trend was reported by many researchers [12,28,29]. A higher ratio results in a larger concentration gradient during diffusion from the solid into the solution and in excessive swelling of the plant material, increasing the contact surface area between the material and the solvent [30].

Extraction Yield
The positive effect of extraction temperature on AAR and CCaf is presented in Figure 1. It is widely recognized that temperature enhances mass transfer by improving the extraction rate; a fact that can The liquid-to-solid ratio (L/S) has a great impact on the examined responses. The extraction yield increased with increasing ratio. A similar trend was reported by many researchers [12,28,29]. A higher ratio results in a larger concentration gradient during diffusion from the solid into the solution and in excessive swelling of the plant material, increasing the contact surface area between the material and the solvent [30].
The positive effect of extraction temperature on A AR and C Caf is presented in Figure 1. It is widely recognized that temperature enhances mass transfer by improving the extraction rate; a fact that can be attributed to the effect of temperature on the vapor pressure, surface tension, and viscosity of the solvent. Moreover, the increase in the yield may be associated with the increased ease with which solvent diffuses into cells and the enhancement of desorption and solubility at high temperatures [31]. Previous studies have reported a decrease of extract antiradical activity with temperature due to the possible degradation of polyphenolic compounds [32].
The measured and the predicted values of the responses are analytically depicted in Table 2. The extraction conditions of 9.25 mg/mL (C βcd ), 30 mL/g (L/S), and 80 • C (T) are the optimum conditions. Under these conditions, the predicted values of antiradical activity and caffeine content were 30.56 µmol TRE/g dry matter and 4.733 mg/g, respectively, whereas the observed experimental values were 30.324 µmol TRE/g dry matter and 4.761 mg/g, respectively.

Sensory Analysis
A consensus vocabulary of eleven attributes was developed to describe the samples, under the modalities of odor, taste, and mouthfeel, while using reference standards where required. Similar sensory descriptors were generated by other trained panels to characterize brewed coffee beverages [33]. All sensory descriptors differ significantly between samples (p < 0.05, Table 5). To the best of the authors' knowledge, this is the first time that coffee pulp has been used scientifically for cold brewed coffee drink preparation and the developed lexicon is unprecedented. Cold brewing is an innovative technique and several researchers have worked on its sensory evaluation. Generally, cold brews are characterized by a sweeter and sour taste as compared to the product of the classical technique. Hence, the purpose of the sensory profiling was to investigate the sensory quality of an alternative coffee drink/beverage formulation and provide a measure of the alterations in coffee drink descriptors occurring when the raw material used is changed.
The sensory profiles of the different brews are presented in Figure 2. Brews prepared from coffee grounds (AQC and AQC/CD) were compared to the brews obtained from coffee pulp (AQCW and AQCW/CD). The panelists scored the latter samples as significantly more botanic and fruitier in odor and more sweet, sour, and botanic in taste. The earthy odor and taste intensities were significantly higher for the coffee bean brew samples as compared to the other two samples. In addition, the AQC and AQC/CD samples had significantly higher scores in bitter and sour-roasted taste than their counterparts. The results of this study show that the sensory characteristics of the cold brew coffee bean samples resembled the usual organoleptic properties of coffee samples prepared with the hot brewing technique. Attributes such as earthy, roasted, and fruity aromas, bitter and sour tastes, and astringency had been previously used to evaluate hot brewed coffee [33]; however, the perceived intensities are not similar. On the contrary, the sensory attributes of the AQC and AQC/CD samples possessed similar characteristics to tea beverages, since they were characterized by high intensities of botanic and fruity aromas, and sweet, sour, and botanic taste sensations. abc: different letters represent statistically significant differences, p < 0.05 (Tukey's method). **: O for odor, T for taste, MF for mouthfeel. ***: AQC: aqueous extract from coffee seeds, AQC/CD: β-cyclodextrin extract from coffee seeds, AQCW: aqueous extract from coffee pulp, AQCW/CD: β-cyclodextrin extract from coffee pulp.
The sensory profiles of the different brews are presented in Figure 2. Brews prepared from coffee grounds (AQC and AQC/CD) were compared to the brews obtained from coffee pulp (AQCW and AQCW/CD). The panelists scored the latter samples as significantly more botanic and fruitier in odor and more sweet, sour, and botanic in taste. The earthy odor and taste intensities were significantly higher for the coffee bean brew samples as compared to the other two samples. In addition, the AQC and AQC/CD samples had significantly higher scores in bitter and sour-roasted taste than their counterparts. The results of this study show that the sensory characteristics of the cold brew coffee bean samples resembled the usual organoleptic properties of coffee samples prepared with the hot brewing technique. Attributes such as earthy, roasted, and fruity aromas, bitter and sour tastes, and astringency had been previously used to evaluate hot brewed coffee [33]; however, the perceived intensities are not similar. On the contrary, the sensory attributes of the AQC and AQC/CD samples possessed similar characteristics to tea beverages, since they were characterized by high intensities of botanic and fruity aromas, and sweet, sour, and botanic taste sensations. As far as the effect of β-CD is concerned, it was found that the AQC sample had a less bitter taste intensity as compared to the AQC/CD sample. Based on the fact that the content of chlorogenic acid and caffeine was the same in both brews (Table 6), it seems that the selective inclusion formation of the extracted ingredients affects the sensory characteristics of the brews. The same behavior has been observed in tea infusion, in which cyclodextrin appeared selective in the extraction of different compounds from tea [34]. These results give a new insight in the contribution of cyclodextrin in taste differentiation. Cyclodextrin has been used for taste masking in different products. In bitter gourd As far as the effect of β-CD is concerned, it was found that the AQC sample had a less bitter taste intensity as compared to the AQC/CD sample. Based on the fact that the content of chlorogenic acid and caffeine was the same in both brews (Table 6), it seems that the selective inclusion formation of the extracted ingredients affects the sensory characteristics of the brews. The same behavior has been observed in tea infusion, in which cyclodextrin appeared selective in the extraction of different compounds from tea [34]. These results give a new insight in the contribution of cyclodextrin in taste differentiation. Cyclodextrin has been used for taste masking in different products. In bitter gourd juice, cyclodextrin contributes to the reduction of the bitter taste [35] and this effect was associated with the fact that the molecules responsible for bitter sensations, caffeine and chlorogenic acid, partially form stable inclusion complexes with β-CD [36], resulting in masking of the bitter sensation by more than 30%. The mechanism of complex formation between β-CD and chlorogenic and caffeic acids in aqueous solutions was previously investigated, suggesting that catechol hydroxyl groups are trapped inside the β-CD cavity [17]. Another effect of β-CD addition is related to the intensity of astringency. More specifically, the coffee pulp brewed sample with β-CD was characterized by a less astringent mouthfeel as compared to the other coffee bean brewed sample. β-CD can be considered as a slightly sweet substance, showing a recognition threshold for sweetness in water around 0.5% [36]; hence, its sweetness cannot be ignored. This study showed two results concerning sweetness: there was no perceived variation in sweet taste between the coffee bean brewed samples, whereas sweetness was more pronounced in AQCW/CD than in the sample prepared by brewing coffee pulp with water. This observation could be attributed to the phenomenon of mixture intensity suppression; in the case of coffee brew samples, intensity was at such high levels that it suppressed the perception of sweet taste, whereas the coffee pulp brewed samples scored very low in bitterness, "allowing" the perception of different intensities in sweetness.
Earlier studies utilized coffee by-products as a source of antioxidants and dietary fibers in bakery products, such as biscuits [37] and muffins [38], investigating consumer acceptance and the sensory quality of the novel products. There is currently very little research published on the development of functional beverages from coffee waste. Coffee silver skin was brewed and used as an ingredient in a beverage aiming to reduce body fat accumulation [39]. The study concluded that the acceptance level of the beverages made with coffee silver skin was about 95%. However, the effect of coffee by-product use on the sensory profile of the functional beverages was not investigated. To the best of our knowledge, the current study is the first systematic approach in determining variations in sensory attributes of cold brew coffees made from coffee seeds or coffee waste with or without β-CD addition. The brewed coffee waste samples demonstrated sensory properties similar to beverages based on tea brew, with highly perceived intensities of fruity and botanic aroma characteristics and sweet, sour, and botanic taste characteristics. This novel beverage prepared by utilizing a coffee by-product could be widely accepted for its sensory quality and consumed by non-coffee drinkers as well.

Conclusions
The present study showed, for the first time, that aqueous solutions of β-cyclodextrin could be very effective solvent systems regarding the extraction of phytochemicals from coffee pulp. The experimental design based on a response surface methodology optimized the conditions, enabling the production of extracts with enhanced antioxidant and caffeine content. The applied extraction method with cyclodextrin resulted in innovative brews from coffee and coffee by-products with modified sensory attributes.