Coffee residues processing

A commercial supplier (CAFFENIO^®) in Hermosillo, Mexico, donated coffee silverskin flakes from dark Coffea arabica L. To obtain the powder, coffee silverskin flakes were pulverized at 20 mesh of particle size (hammer mill, Pulvex 200, DF, Mexico). Both materials were packed under vacuum (vacuum sealer, Food Saver^®, FLA., USA) and stored at room temperature (25 °C) until analysis.

Physicochemical characterization Measurement of pH

The pH of coffee residues and the commercial standard was measured after homogenization with distilled water at a 1:10 ratio with a potentiometer (pH211, Hanna Instruments Inc., RI, USA) with automatic temperature control (^{AOAC, 2020}).

Color measurement

The color measurement was performed using a spectrophotometer (CM-508d, Konica Minolta Inc., TYO, Japan). The registered values were lightness (L*, ranging from 0-black to 100-white), redness (a*, takes positive values for reddish colors and negative values for greenish colors), yellowness (b*, takes positive for yellowish colors and negative for bluish colors), Chroma (C*), and hue angle (h*). The coffee residues were placed in 3.5 cm diameter Petri dishes, and ten measurements were performed on the samples surface (^{Robertson et al., 1977}).

Techno-functional characterization

The techno-functional properties of coffee residues and commercial standard were measured as previously described with slight modifications (^{Haque et al., 2020}; ^{López-Marcos et al., 2015}).

Water-holding capacity (WHC)

Samples (0.1 g) were homogenized with 1.5 mL of distilled water at 10,000 rpm for 1 min, and kept at 25 °C for 30 min. Subsequently, the test tubes were centrifuged at 15,000 x g for 20 min at 4 °C. The supernatant was decanted, and the test tubes with the sediments were weighed. The WHC (%) was calculated as [(Final weigh of the tube with the sample - Initial weigh of the tube with the sample) / (Weigh of the sample)] x 100.

Oil-holding capacity (OHC)

Samples (0.1 g) were homogenized with 1.5 mL of commercial corn oil at 10,000 rpm for 1 min, and kept at 25 °C for 30 min. Subsequently, the test tubes were centrifuged at 15,000 x g for 20 min at 4 °C. The supernatant was decanted, and the test tubes with the sediments were weighed. The OHC (%) was calculated as [(Final weigh of the tube with the sample - Initial weigh of the tube with the sample) / (Weigh of the sample)] x 100.

Swelling capacity (SC)

Samples (0.5 g) were homogenized with 5 mL of distilled water at 10,000 rpm for 1 min, using a graduated tube, and kept at 25 °C for 1 h. The initial and final volume occupied by the samples was measured, and the increase in sample volume after incubation indicates SC.

Foaming capacity (FC)

Samples (0.5 g) were homogenized with 10 mL of distilled water at 5,000 rpm for 5 min at 4 °C (Ultraturrax T25, IKA, Stau, Germany). The foam produced within 30 sec, shows FC.

Foaming stability (FS)

The volume of the foam initially produced was measured within 30 sec, and test tubes were incubated at 25 °C for 30 min. Then, the volume of the foam was measured, and a reduction in the volume showed the absence of FS.

Emulsion capacity (EC)

Samples (0.5 g) were homogenized with 10 mL of distilled water at 8,000 rpm for 1 min at 4 °C. The resultant solution was homogenized with 10 mL of commercial corn oil, under the same conditions. A two-phase suspension formation shows EC.

Emulsion stability (ES)

The previously prepared emulsion was heated for 30 min at 80 °C, cooled at 25 °C, and centrifuged at 12,000 x g for 5 min at 4 °C. If the first emulsion is not broken, it indicates ES.

Gelling capacity (GC)

The coffee residues were homogenized with distilled water to obtain several suspensions at 0, 2.5, 5, 10, 15, and 20 % (w/v) and boiled for 1 h. Subsequently, test tubes were cooled for 1 h at 0 °C. The tubes were inverted to see gel formation.

Antioxidant activity

Phytochemical content

The quantitative phytochemical analysis of the coffee residues and the commercial standard was also conducted. In brief, the compounds were extracted using an aqueous solution as solvent (1:10, solid-solvent ratio) by ultrasound-assisted method at 42 KHz for 30 min at 25 °C (Bransonic 3800, Ultrasonics Corp., Jeju, Korea). The resultant solution was filtered (Whatman No. 4 filter paper), concentrated under reduced pressure at 60 °C (rotary evaporator Yamato RE-301BW, Tokyo, Japan) and dried (Freeze dryer Yamato DC401, Tokyo, Japan). The obtained extracts were stored at −20 °C in the dark until analysis.

Total phenolic content (TPC) was determined by the Folin-Ciocalteu method (^{Ainsworth and Gillespie, 2007}). An aliquot of each extract (20 µL, 5 mg/mL) was homogenized with 80 µL of distilled water, 60 µL of Na ₂CO₃ (7 %, w/v), and 40 µL of Folin- Ciocalteu’s reagent (0.25 N). The resultant solution was mixed with 80 µL of distilled water and incubated for 1 h at 25 °C, in the dark. The absorbance was measured at 750 nm in a spectrophotometer (Multiskan FC UV-Vis, Thermo Scientific, Tokyo, Japan), and results were expressed as mg equivalents of gallic acid per g of dried extract (mg GAE/g).

Total flavonoids content (TFC) was evaluated by the aluminum chloride-complex formation method (^{Popova et al., 2004}). An aliquot of each extract (20 µL, 5 mg/mL) was homogenized with 130 µL of methanol and 20 µL of AlCl₃ (5 %, w/v). The resultant mixture was incubated for 30 min at 25 °C, in the dark. The absorbance was measured at 415 nm, and results were expressed as mg equivalents of quercetin per g (mg QE/g).

Caffeoylquinic acid content (CAC) was also determined (^{Griffiths et al., 1992}). An aliquot of each extract (100 µL, 5 mg/mL) was homogenized with 200 µL of urea (0.17 M), 200 µL of glacial acetic acid (0.1 M) and 500 µL of distilled water. The resultant solution was mixed with 500 μL of NaNO2 (0.14 M) and 500 μL of NaOH (1 M), centrifuged at 2,250 x g for 10 min at 4 °C (Sorvall ST18R, Thermo Fisher Scientific, Waltham, USA). The absorbance was measured at 510 nm, and results were expressed as mg equivalents of chlorogenic acid per g (mg CAE/g).

Free-radical scavenging activity

The free-radical scavenging activity (FRSA) was determined by the DPPH test (^{Molyneux, 2004}). An aliquot of each extract (100 µL, 5 mg/mL) were homogenized with 100 µL of DPPH• solution (300 µmol), and incubated for 30 min at 25 °C, in the dark. Ascorbic acid (25 µg/mL) was used as a positive control. The absorbance was measured at 520 nm, and results expressed as FRSA (%) = [(Absorbance of DPPH• solution at 0 min) - (Absorbance of DPPH• solution + extract at 1 h) / [(Absorbance of DPPH• solution at 0 min)] × 100.

Radical cation scavenging activity

The radical cation scavenging activity (RCSA) was determined by the ABTS•+ test (^{Re et al., 1999}). Prior to obtain the radical cation, equal parts of ABTS solution (7 mM) and potassium persulfate (2.45 mM) were homogenized and incubated for 16 h at 25 °C, in the dark. Afterward, the formed radical was diluted with an ethanol solution to obtain an absorbance of 0.8, and 20 µL of each extract (5 mg/ mL) was mixed with 180 µL of the radical. Ascorbic acid (25 µg/mL) was used as a positive control. The absorbance was measured at 734 nm, and results were expressed as RCSA (%) = [(Absorbance of ABTS•+ solution at 0 min) - (Absorbance of ABTS•+ solution + extract at 10 min) / [(Absorbance of ABTS•+ solution at 0 min)] × 100.

Reducing power ability

The reducing power ability (RPA) was measured by the Prussian-blue test (^{Berker et al., 2010}). An aliquot of each extract (100 µL, 5 mg/mL) was homogenized with 300 µL of phosphate buffer (0.2 M, pH 6.6) and 300 µL of potassium ferricyanide (1 %, w/v), and incubated in a water bath for 20 min at 50 °C. The resultant solution was mixed with 300 µL of trichloroacetic acid (10 %, w/v) and centrifuged at 4,200 x g for 10 min at 4 °C. Thereafter, 500 µL of the supernatant was mixed with 100 µL of distilled water and 250 µL of ferric chloride (0.1 %, w/v). Ascorbic acid (25 µg/mL) was used as a positive control. The absorbance was measured at 700 nm, and results expressed as absorbance increase at the same wavelength.

Ferric-reducing antioxidant power

The ferric-reducing antioxidant power (FRAP) test was also determined (^{Benzie and Strain, 1999}). An aliquot of each extract (20 µL, 5 mg/mL) was homogenized with 180 µL of FRAP solution [10:1:1, 300 mM buffer sodium acetate in glacial acetic acid at pH 3.6 and 10 mM 4,4,6-tripyridyl-S-triazine (TPZ) in 40 nM HCl and 20 mM FeCl ₃] and incubated for 8 min at 25 °C, in the dark. Ascorbic acid (25 µg/mL) was used as a positive control. The absorbance was measured at 595 nm, and results were expressed as mg of Fe2+ equivalents per g (mg Fe2+/g).

Lipid oxidation in a meat system

The lipid oxidation (LOX) was measured by the TBARS test (^{Kim et al., 2016}). An aliquot of each extract (1 mL, 5 mg/ mL) were homogenized with 10 mL of an aqueous pork meat extract (10 %, w/v), and incubated at 37 °C for 8 h in a water bath. After, 0.5 mL of the mixture were mixed with 1 mL of TCA (10 %, w/v) and 1.5 mL TBA solution (0.02 M), placed in a water bath for 20 min at 97 °C, and then cooled. The absorbance was measured at 531 nm, and results were expressed as mg of malondialdehyde per kg of meat (mg MDA/kg).

Statistical analysis

Results obtained from physicochemical, WHO and OHC, antioxidant, phytochemical content, antiradical, reducing power, and lipid oxidation inhibition tests, were expressed as mean ± standard deviation. These data from the three independent trials were subjected to one-way analysis of variance (ANOVA), while Tukey tests were conducted for means comparison at p < 0.05. In addition, results of qualitative phytochemical screening, the rest of the test of techno-functional properties and antioxidant activity were expressed as follows: (+), present slight functionality; (++), present moderate functionality; (+++), present high functionality; (-), absent.

Physicochemical properties

Table 1 reports the physicochemical properties of the coffee residues compared to the commercial standard. The results show that coffee silverskin powder and flake showed lower pH values than texturized soy (p < 0.05). Also, results of color parameters write down that coffee silverskin powder and flakes showed lower L*, b*, C* and h* values than texturized soy, as well as the highest a* values (p < 0.05), while coffee silverskin powder showed the lowest L* values (p < 0.05). In addition, the color data also includes information related to the RGB and Hex color of soybeans and the silverskin of coffee in flakes and flour, which according to these the color name is light brown, roman coffee or brown bear, and deep coffee or bole, respectively.

A recent work proposed coffee silverskin powder derived from a mixture of five Arabicas and five Robusta varieties as a potential food-safe ingredient, thus, measuring physicochemical parameters like pH (5.34) and color (L* = 20.8, a* = 4.27, b* = 13.4, C* = 14.2). They concluded that pH values could be associated with phenolic components. At the same time, the color depends on the interaction of constitutive components (polyphenols, melanoidins, among others), protein degradation, and Maillard reactions by the roasting process (^{Martuscelli et al., 2021b}). Another study showed that a reduction of particle size in coffee beans increases the sensory acidity in the beverage, which denotes in the extractable compounds (^{Getaneh et al., 2020}).

In contrast with our study, a decrease was reported for pH values of coffee parchment residue (C. arabica) by effect of the milling procedure, which was associated with the bioavailability of certain phytochemicals such as phenolic acids (^{Benitez et al., 2019}). It is important to know pH values of natural ingredients because this allows knowing the type of food matrix into which they could be incorporated without affecting their functional properties (^{López-Marcos et al., 2015}; ^{Benitez et al., 2019}). In addition, color also plays a key role in the perception, acquisition, and consumption of food. Therefore, knowing and measuring this sensory attribute is crucial for chefs, food bloggers, packaging designers, industrial food processors, among others (^{Spence, 2018}). In agreement, it has been proven that a reduction in particle size x time does not affect a* values (^{Mugabi, 2021}). While a reduction in particle size of various powders rich in fiber reduce L* values, which show that samples became darker (^{Liu et al., 2016}).

Techno-functional properties

According to the ^{Codex Alimentarius (1995)}, food additives are employed for different purposes, such as sweeteners, flavor enhancers, and colorants, as well as pH regulators and bind food components. In this context, Table 2 reports the techno-functional properties of coffee residues compared to the commercial standard. The results say that coffee silverskin flakes showed the highest WHC and OHC compared to coffee silverskin powder and texturized soy. The WHC of coffee residues depends on the polysaccharides content in coffee residues. A reduction of these components during processing could affect WHC due to a lack of site interactions for water retention. This parameter is desirable in meat products like sausages because these residues can absorb water without dissolving proteins. While OHC of coffee residues depends on the surface availability of hydrophobic non-polar chains, such as dietary components or amino acids, which bind the hydrocarbon side chain of oil. This parameter is related to the physical entrapment of oil and food stabilization, which could function as a flavoring retainer (^{Benitez et al., 2019}).

Further, all samples exert slight SC, FC, and FS, without effect of milling process. While no EC, ES, and GC were shown in coffee silverskin powder, they were the highest in texturized soy. In agreement, it has been reported that a mixture of coffee silverskin powder (C. arabica and C. robusta) exerts WHC and OHC (^{Martuscelli et al., 2021b}). Also, previous work revealed that coffee silverskin and spent coffee grounds powders (a mixture of C. arabica and C. robusta) showed approximately 50 % of WHC and OHC. In disagree with our study, both residues showed EC and ES (^{Ballesteros et al., 2014}). Additionally, it has been reported a decrease of WHC values of coffee parchment residue (C. arabica) by effect of the milling procedure. In contrast with our study, OHC values were unaffected (^{Benitez et al., 2019}).

WHC and OHC are important parameters that express the potential of an ingredient to retain water and oil in a food matrix (^{López-Marcos et al., 2015}; ^{Martuscelli et al., 2021b}). SC is the capacity of an ingredient to increase the surface volume in the presence of excess water (^{Benitez et al., 2019}), while FC and FS express the capacity of an ingredient to increase the interfacial area and the effectiveness to maintain the foam stable after the time (^{Haque et al., 2020}). EC express the capacity of an ingredient to form a homogeneous dispersion of two immiscible liquids or emulsion, while ES describes the effectiveness of an ingredient in maintaining a stable emulsion after the thermal effect (^{Ballesteros et al., 2014}; ^{López-Marcos et al., 2015}; ^{Martuscelli et al., 2021b}).

Antioxidant properties

Moreover, phytochemicals (from the Greek word Phyto, meaning plant) are bioactive chemical compounds found in different amounts in plant parts (fruits, seeds, flowers, leaves, stems, and roots), which depends on plant species, growing, and processing conditions. Also, it is well known that phytochemicals improve the protection to plant cells against environmental hazards like pollution, stress, UV-exposure, and pathogenic damage. Although these compounds are not essential nutrients for human body, it has been reported that they exert several biological functions of interest for the food industry (^{Saxena et al., 2013}). Table 3 reports the phytochemical content of the coffee residues compared to the commercial standard. The results of the qualitative phytochemical profile showed that phenols and caffeoylquinic acid were highly present in coffee silverskin powder, while these compounds were not detected in texturized soy. Flavonoids were moderately present in both coffee residues, and coffee silverskin flakes showed moderately presence of alkaloids. In addition, terpenoids, steroids, and saponins were not detected in all samples. Regarding the quantitative profile, both coffee residues showed high (p < 0.05) TPC and TFC values (>200 mg GAE and 60 mg QE/g, respectively) than texturized soy, while coffee silverskin powder showed high CAC (approx. 80 mg CAE/g) than coffee silverskin flakes and texturized soy (p < 0.05).

It has been evidenced that soy is an important source of nutrients like proteins, essential amino acids, and carbohydrates, although oligosaccharides, stachyose, and raffinose, trypsin inhibitors, lectins, gliadin, phytates, mycotoxins, and saponins have been considered antinutritional substances in soy, which are inactivated by adequate thermal processing (^{Banaszkiewicz, 2011}). In agreement, a preliminary phytochemical screening showed the presence of alkaloids and flavonoids in an ethanol extract obtained from coffee silverskin powder (C. arabica). In contrast, saponins and terpenoids were also identified in this extract (^{Al-Yousef et al., 2017}). In addition, a quantitative screening demonstrated the presence of phenolic, flavonoids, and caffeoylquinic acid in extracts obtained from coffee residues (C. arabica), including coffee silverskin and spent coffee grounds (^{Zengin et al., 2020}). Although, phytochemicals are present in diverse types of waste from the coffee processing industry (^{Ballesteros et al., 2014}; ^{Zengin et al., 2020}), it has been also demonstrated that a reduction of the particle size improves chemical compound recovery. For example, an increase in carbohydrates, phenolic, and flavonoid content of coffee parchment residue (C. arabica) was reported after the milling procedure (^{Benitez et al., 2019}).

The ^{Codex Alimentarius (1995)} also indicate that food additives are used with preservative purposes, for example, to function as an antimicrobial agent and improve antioxidant effect against the oxidation process. In this regard, Table 4 shows the antioxidant properties of the coffee residues compared with a commercial antioxidant standard (ascorbic acid, Asc), and the results indicate that Asc showed higher FRSA (> 90 % of inhibition) than coffee silverskin powder and flakes (approx. 40 % of inhibition for both), while, Asc and coffee silverskin powder showed higher RCSA activity (approx. 90 % of inhibition for both) than coffee silverskin flakes (< 85 % of inhibition) (p < 0.05), which indicate a positive effect of milling procedure. Concerning reducing power, RPA and FRAP values were presented in the order Asc > coffee silverskin powder > coffee silverskin flakes (p < 0.05). In addition, coffee silverskin powder showed the lowest TBARS values (< 0.1 mg MDA/kg), which indicates a positive effect of the milling procedure.

In agreement, the antiradical (FRSA and RCSA) and the reducing power activity (FRAP) of extracts obtained from coffee residues (C. arabica) like silverskin and spent coffee grounds has been demonstrated (^{Zengin et al., 2020}). A previous study indicates that coffee silverskin powders and spent coffee grounds (mixture of C. arabica and C. robusta) exert antiradical FRSA and FRAP potential (^{Ballesteros et al., 2014}). Additionally, an increase of RCSA of coffee parchment residue (C. arabica) was observed after the milling procedure (^{Benitez et al., 2019}). The FRSA (DPPH^• assay) are methods that involve the hydrogen atom transfer as reaction mechanisms [free radical (X^•) + Antioxidant compound (OH) → Neutralized radical (XH + O^•)], while the RCSA (ABTS^•+ assay), RPA (Prussian-blue) and FRAP methods involve the electron transfer mechanisms [radical cation (X^•+) + Antioxidant com-pound (OH) → Neutralized radical (XH + O^•+)] (^{Berker et al., 2010}; ^{Echegaray et al., 2021}). In addition, it has been reported that TBARS method could involve both mechanisms (^{Velázquez et al., 2021}).