Raw materials

Roasted coffee (C. arabica L.) grains produced in Nayarit, Mexico, were obtained from local markets for this study. The coffee had a roasted appearance, an attractive aroma, and a desirable visible coffee quality. Coffee samples were kept in storage in zip-lock bags while awaiting analysis to avoid the loss of volatile compounds.

Plasma generator equipment

The plasma generating equipment’s design and construction correspond to the Plasma Physics Laboratory of the National Institute for Nuclear Research in Toluca, Mexico. The system was made up of the 13.56 MHz radio frequency generator (RF) designed to directly supply a pair of stainless-steel plates in a glass DBD reactor (^{Peña-Eguiluz et al., 2010}).

Roasted coffee powder treated with cold plasma

Roasted coffee (C. arabica L.) grains were pulverized, and 0.5 g of sample was placed on plate Petri on thin layers. They had approximately 1 mm of thickness and were im- mediately treated with cold plasma energy at atmospheric pressure. Plasma was generated with commercial helium gas (Praxair, Mexico) at a flow of 1.5 L/min, and the energy was applied at different times (0, 1, 4, 8, 12, 16, 20, 24, and 30 min) and 30 W input power and an output voltage of 850 V. Before and after each plasma treatment, we determined antioxidant activity and total polyphenol content. In the case of physicochemical and sensory properties, treatment times were 0 and 30 min.

Physicochemical quality parameters in roasted coffee Surface color

The color assessment was evaluated with a Hunter Lab MINOLTA Chroma Meter CR300 colorimeter (Hunter Laboratory, México) in the reflectance mode for roasted coffee. The color was expressed as L (brightness), a (redness), and b (yellowness) values. Results were expressed as the average of four measurements. Besides, the total color difference (ΔE) was calculated as below (^{Kovačević et al., 2016}) (Eq. 1):

Where L₀, a₀, and b₀ were the untreated control values, and L^*, a^*, and b^* were the values for treated roasted coffee.

Aroma compounds evaluation

Samples subjected to a 30 min cold plasma treatment were analyzed with a gas chromatograph (GC-7890) coupled to mass spectrometry (NIST) system with ion trap mass spectrometer from Agilent Technologies Instruments operated in the split mode. A CP- Wax (60 m Χ 0.25 mm i.d.) column with a 0.25 μm film thickness was employed for chromatographic separation of aromatic compounds. Ion trap mass spectrometer was operated in the electron ionization (EI) mode. The NIST library was used to detect possible intermediates, and also our mass library was created with this program. The free volatile compounds were identified by GC/MS by means of retention time and mass spectrum. Helium at 1.6 mL/min was used as the carrier gas. Divinylbenzene / carboxen / polydimethylsiloxane (DVB/CAR/PDMS) fiber with coating thicknesses of 50 and 30 (m was used (^{Akiyama et al., 2008}).

The oven temperature program was as follows 50 ºC for 2 min (rate 2 ºC min^-1) to 220 ºC for 85 min. Hydrogen was used as the carrier gas at 100 kPa; the injector temperature and the splitless flow were set at 260 ºC and 30 mL min⁻¹, respectively, after a splitless time of 1 min. Solid-phase mi-croextraction (SPME) was carried out according to ^{Solis-Solis et al. (2007)}, with some modifications. Samples of 0.5 g of non-treated and treated roasted coffee were mixed with 1 g NaCl and 1.6 mL deionized water and subsequently sealed with PTFE-silicone septa. Sample vial equilibration was in-cubated at 40 ºC for 30 min; the DVB/CAR/PDMS fiber was then exposed to the headspace above the sample for 30 min followed by 10 min of thermal desorption of the adsorbed substances in the injector port.

Polyphenols extraction

0.5 g of roasted coffee were mixed with 20 mL of acidified methanol solution (0.8 M HCl, 50:50, v/v) and extracted by shaking for 1 h in a wrist-action (Burrel, USA) at maximum speed (room temperature). The extracts were then centrifuged (Hermle, Z32HK, Labortechnik GmbH, Wehingen, Germany) at 6000 x g for 15 min at 4 ºC. The supernatants were recovered, and the residues were reextracted using 20 mL of an acetone solution (70:30, v/v, 60 min), supernatants were collected in a flask with the combination of acidified methanol solution and acetone solution (50:50, v/v). The supernatants were considered as total soluble polyphenols (TSP) and were used to evaluate the AOX. Samples were stored at 4 ºC in the dark (^{Pérez-Jiménez et al., 2008}).

Total Soluble Polyphenols (TSP) content

The soluble polyphenols content was determined by the method of ^{Alvarez-Parilla et al. (2010)} with some modifications. An 250 (L aliquot of extracts from roasted coffee were mixed with 1000 (L Na₂ CO₃ (sodium carbonate, 7.5 w/v). After 3 min, 1250 μL of Folin-Ciocalteu’s reagent were added, and the mixture heated in a water bath for 15 min at 50ºC, which resulted in a final volume of extract sample of 270 μL. A microplate reader (Bio-Tek®, Synergy HT, Winooski, VT, USA) in a multi-mode spectrophotometric detection with 96-well plates and the Gen5 Program was used. Gallic acid was used as a standard (0.0125 - 0.2 mg/mL R² ≥ 0.9997), and results were expressed as mg of gallic acid equivalents (mg GAE/g DW), the absorbance was read at 750 nm against a blank.

AOX activity determination

This assay was based on the ability of different substances to scavenge 2,2´-Azinobis-3-ethyl benzothiazo-line-6-sulfonic acid (ABTS) radical cation. The method pro- posed by ^{Re et al. (1999)} with some modifications was used. The radical cation was prepared by mixing a 7 mM stock solution with 2.42 mM potassium persulfate and kept in the dark at room temperature for 14 h. The ABTS solution was adjusted with phosphate buffer at an absorbance of 0.70 (±0.02). Trolox (6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid) was used as a standard and methanol as a blank. Extract samples of 20 (L were read on a microplate reader (Biotek, Synergy HT, Winooski, VT, USA) with 300 (L capacity, and 280 (L of the ABTS radical was added. Then, it was incubated at 37 ºC in the dark and read for 6 min at 734 nm; a calibration curve was prepared using an aqueous solution of Trolox as a standard. We reported the results as mmol TE per 100 g sample DW.

Sensory evaluation

The roasted coffee (untreated control and 30 min cold plasma treated) was prepared by adding the coffee powder (1:4 w/v) and sugar 5 % (w/v) in boiling water. Prepared coffee was evaluated for sensory attributes by 40 untrained judges. The experiment was divided into two phases: (1) triangular test was performed to compare between three samples simultaneously, two untreated samples and one treated with 30 min cold plasma, without having to specify the sensory characteristic(s) that differ; (2) hedonic test was based on the evaluation of color, odor, taste and overall acceptability parameters. A 9-point hedonic scale (1: dislike extremely to 9: like extremely) was used to analyze the sensory score corresponding to each tasted sample (^{Kumar et al., 2012}).

The sensory results were analyzed by a two-factor analysis (panelist and product) with an ANOVA test and differences were established by the LSD test.

Statistical Analysis

A unifactorial design where the independent factor was the treatment time, was employed. Statistical analysis was performed using SAS for Windows V9. Analysis of variance (ANOVA) and multiple comparison procedures (Least Significant Difference-LSD) test were conducted to determine whether there were significant differences (p(0.05) among treatments for physicochemical properties, sensory analysis, and antioxidant compounds.

Color evaluation in cold plasma treated roasted coffee

Hunter color parameters for roasted coffee, both treated with cold plasma and untreated, are shown in Table 1. No differences in the surface measured parameters were observed (p > 0.05), and the value of (ΔE of 0.78 was obtained. In conditions of ideal vision, values of (ΔE ≤ 1 represent a minimum difference of color, hardly perceptible for the human eye (^{Vervoort et al., 2012}).

The impact of cold plasma on color depends on critical factors associated with the product such as type (cut, solid, liquid), plasma treatment conditions (input voltage, time, power, working gas), and storage conditions (^{Pankaj et al., 2018}).

Aroma compounds determination

The aroma compounds detected in the untreated roasted coffee and cold plasma treated samples were not statically significant (p > 0.05). A total of 9 volatile compounds were identified in both samples (Figure 1).

Figure 1 GC-MS chromatograms of aromatic compounds detected in untreated roasted coffee and cold plasma samples treated for 30 min.
Figura 1. Cromatogramas GC-MS de compuestos aromáticos detectados en café tostado sin tratar y tratado con plasma frío durante 30 min.  

This suggests that cold plasma treatment did not significantly affect the covalent bonds responsible for aroma, compounds’ integrity and flavor characteristics (^{Bressanello et al., 2017}). However, the chemistry of cold plasma collision reactions generates N_xO_y, O₃ (ozone), peroxy-radicals in a discharge produced O atom and attacks molecular O₂ by a three-body reaction to yield ozone induced oxidation in foods (^{Gavahian et al., 2018}). Also, oxygen species (ROS) can be formed in nature under UV-light influence generating photo-oxidation of lipids (^{Vandamme et al., 2015}). Some studies have demonstrated that, especially fatty foods, can interact with ROS (OH and ¹O₂) and be linked by double bonds (^{Van Durme and Vandamme, 2016}). Therefore, the primary oxidation products (non-volatile components) and secondary oxidation products (volatile components) contribute to off-flavors, for example, vinyl ketones, trans, cis-alkadienals, alkanals, hydrocarbons present oils and fats (^{Kerrihard et al., 2015}).

^{Vandamme et al. (2015)}, performed a qualitative and semi-quantitative identification of volatile oxidation compounds from by-products in vegetable oil samples treated with DBD plasma jet (Ar / 0.6 % O₂), generating a wide range of highly reactive oxidative species as atomic oxygen, hydroxyl radicals and singlet oxygen, while maintaining ambient temperatures. ^{Alves et al. (2017)}, showed processed by ozone and plasma (oxygen gas) for a long period (1 to 3 min), and obtained oxidation of limonene, y-terpinene and linalool, causing off-flavor and volatile composition on orange juice.

The main parameters involved in by-products formation are food matrix, voltage, treatment time, oxygen concentration, plasma configuration, and also water and carrier gas concentration. From those parameters, high lipid content in food matrices is critical for off-flavor formation during cold plasma treatment. The coffee beans have a low lipid content, and thus, no formation of off-flavor molecules was observed.

Total Soluble Phenols (TSP) content and AOX in roasted coffee treated with cold plasma

TSP content in roasted coffee showed significant (p < 0.05) variations during the cold plasma treatment, decreasing from 67.74 ( 2.31 GAE/g to 59.12 ( 3.92 GAE/g after 30 min of treatment (Table 2). A decrease of 12 % in total phenolic concentration (TPS) after cold plasma treatment, compared with control.

The plasma technology generates reactive oxygen species (ROS) such as hydroxyl radicals, atomic oxygen, and singlet oxygen and ozone (^{Brandenburg et al., 2007}). The TSP can capture radicals and are susceptible to ozone attack (^{Stalter et al., 2011}), explaining the reduction in TSP through the time of plasma application. According to these authors, a phenolic compounds reduction was observed, probably due to oxidation caused by active particles and radicals, depending on the plasma exposure time (^{Sarangapani et al., 2017}). Moreover, molecular ozone acts on the aromatic com- pounds favoring the formation of hydroxylated and quinone compounds, because the formation of aliphatic compounds is originated from the rupture of the aromatic ring (^{Perez et al., 2002}).

^{Almeida et al. (2015)}, showed that the total phenolic content was affected (p<0.05) only after 60 s under indirect plasma exposure on orange juice. ^{Solís et al. (2013)}, observed a reduction of 38 ( in total polyphenol content in chamo- mile samples with 10 min of plasma energy at 750 volts, and of 33 ( in cinnamon samples treated at 650 volts.

Regarding the AOX in roasted coffee, this increased 14 % after 30 min of treatment with cold plasma (p < 0.05) (Table 2). The influence of plasma species and phenolic derivatives (hydroxyl radicals, peroxyl radicals), atomic or singlet oxygen caused reaction, both direct and indirectly (^{Brandenburg et al., 2007}). Therefore, depending on the chemical structure of each of the polyphenols, the antioxidant properties might vary. These chemical conditions influence the ability of polyphenols to scavenge free radicals, changing their biological activity (^{Vajragupta et al., 2000}). The UV-B radiation during cold plasma, plays a crucial role in the individual compounds of the polar fraction due to penetration or favored biosynthesis (^{Ramazzina et al., 2015}; ^{Grzegorzewski et al., 2010}), which facilitate the increment of antioxidants (others than polyphenols) in roasted coffee.

^{Solís et al. (2013)} reported an increase of antioxidant activity in cinnamon samples at 750 V (21.4 (), 850 V (12.2 (), and 650 V (14.4 (), during 10 min of treatment. Also, ^{Trouillas et al. (2006}) observed an increase in the antioxidant properties of plasma-treated basmati rice.

The variable effect of cold plasma on functional food components may be due to differences in the food matrices, plasma equipment configuration, and processing parameters, particularly the gas used. Accordingly, further work is needed to clarify the reaction chemistry between plasma RS and antioxidants in food products. For example the elucidate the molecule fragments resulted from the erosion and UV radiations produced during cold plasma treatments to gain further insight into the relationship between antioxidant activity and polyphenols content in treated coffee samples.

Sensory evaluation

Sensory quality to evaluate triangle test was determined after taking into consideration sensory scores given by untrained judges. The results showed no significant difference between treated and untreated samples (p < 0.05). Thus, cold plasma treated samples were found to be acceptable in terms of sensory attributes.

Figure 2 shows the results of a hedonic test of sample C201 (treatment cold plasma) and sample C685 (untreated). The judges evaluated “like moderately” for both samples.

Figure 2 Sensory evaluation of roasted coffee (cold plasma treated and untreated) on a 9-point hedonic scale.
Figura 2. Evaluación sensorial de café tostado (tratado con plasma frío y sin tratar) con una escala hedónica de 9 puntos.  

Study by ^{Basaran et al. (2008)} report no significant difference on the sensory analysis, of nut samples treated with cold plasma LPCP using air gases or sulfur hexafluoride (SF₆) (appearance, color, odor, texture).

Another study of cold plasma showed that high lipid such as alkanes, alkenes, aldehydes, alcohols, ketones, and acids produce unpleasant flavors, which are sensory described as fishy, metallic, rancid, and oxidized (^{Kim et al., 2013}; ^{Jayasena et al., 2015}). The factor of relative humidity on fresh fruits and vegetables generates “sour” or “chemical” odor associated with reactive nitrogen species (nitrogen monoxide, nitrogen dioxide and oxygen) and peroxynitrite (^{Schnabel et al., 2015}; ^{Von-Woedtke et al., 2012}). Furthermore, it should be mentioned that coffee bean has a low content of a_w (10 () and lipids (16 () (^{Alvarado and Puerta, 2002}). Hence, the probability of generating new compounds with sensory impact was reduced, and no differences in sensory attributes were observed.