Comparative Profiling of Hot and Cold Brew Coffee Flavor Using Chromatographic and Sensory Approaches

Coffee brewing is a complex process from roasted coffee bean to beverage, playing an important role in coffee flavor quality. In this study, the effects of hot and cold brewing on the flavor profile of coffee were comprehensively investigated on the basis of chromatographic and sensory approaches. By applying gas chromatography–mass spectrometry and odor activity value calculation, most pyrazines showed higher contribution to the aroma profile of cold brew coffee over hot brew coffee. Using liquid chromatography, 18 differential non-volatiles were identified, most of which possessed lower levels in cold brew coffee than hot brew coffee. The sensory evaluation found higher fruitiness and lower bitterness and astringent notes in cold brew coffee than hot brew coffee, which was attributed by linalool, furfural acetate, and quercetin-3-O-(6″-O-p-coumaroyl) galactoside. This work suggested coffee brewing significantly affected its flavor profile and sensory properties.


Introduction
Coffee (Coffea sp.) is the second most globally traded commodity after petroleum [1] and accounts for broad market and high economic values. It is estimated that the worldwide production of coffee increased to 9.9 million tons in 2019 [2]. Coffee is popular among consumers due to its pleasant flavor profile. The flavor quality of coffee is principally determined by chemical components including volatile compounds, which contribute to the aroma properties, and non-volatile compounds to the taste. These flavor compounds can be significantly affected by many complex factors, such as processing method, roasting degree, and brewing method, which all consequently affect the consumer choice [3].
In recent years, the brewing process, as the final step closest to coffee consumption, is increasingly attracting the interest of food manufacturers. Extraction parameters play key roles in coffee sensory quality and flavor perception, such as the ground coffee particle size, extraction time, pressure, and temperature [4]. On the basis of the brewing temperature [5], it can be divided into classical hot brew coffee and novel cold brew coffee (≤25 • C). As reported by most sensory studies, cold brew coffee has been considered as fruity, floral, and sweet [5,6], while hot brew coffee possesses signature roasted and smoky-like smells [7].
Generally, the high temperature increases the kinetic energy of water molecules, which favor the leaching out of chemical compounds from the coffee powders. Meanwhile, the high temperature can affect the release of volatile compounds in coffee, thereby altering the sensory perception [8]. With advanced analytical techniques, chromatography-mass spectrometry was applied to compare the differences of physiochemical compounds between hot and cold brew coffee. In a comparative study on the volatiles in cold and hot brew Arabica coffee from Colombia with the medium roasting degree [9], it was reported that cold brew coffee with immersion extraction had a higher abundance in total furans and pyrazines than hot brew coffee by the French press method. As for non-volatiles, our previous study reported that the contents of physiologically functional norhorman and harman were richer in hot brew coffee with pour-over and boiled methods than cold brew coffee using liquid chromatography quadrupole time-of-flight mass spectrometry [10]. Interestingly, the hot brew coffee exhibited significantly higher bitterness and astringency over cold brew coffee [11], but no significant difference was detected in the quantities of a few common compounds associated with coffee bitterness, such as total caffeoylquinic acids, caffeine, and trigonelline [11,12]. Yet, a significant difference was frequently reported in previous sensory studies [4,5], especially in the flavor perception, between hot and cold brew coffee. Until now, very limited knowledge is available on the compounds that are responsible for the differential sensory profile between hot and cold brew coffee. Moreover, the coffee preparation in most studies referred to homemade process without standardization after coffee extraction, which would be of little help in understanding the coffee flavor for industrial processing. Therefore, a comprehensive analysis is needed.
The objective of this study was to compare the aroma-active compounds in hot and cold brew coffee by gas chromatography-mass spectrometry (GC-MS) and odor activity value calculation, to differentiate non-volatile compounds including organic acid, sugar, and phenolic compounds using the liquid chromatography with pulsed amperometric or mass detector and multivariate statistical analysis. Lastly, the sensory triangle test and quantitative descriptive analysis (QDA) of hot and cold brew coffee were conducted to tentatively figure out potential correlations between differential flavor compounds and sensory attributes of hot and cold brew coffee prepared on the basis of industrial processing.

Materials and Chemicals
The same commercial batch of medium-dark roasted Arabica coffee beans (Pacific, Shanghai, China) blended from Costa Rica, Colombia, and Indonesia were selected for coffee brewing, which were purchased from the local market in January 2020 in Beijing city. A total of 2.0 kg of coffee beans were fully mixed right before coffee preparation.
High-performance liquid chromatography (HPLC) grade of acetonitrile and methanol were purchased from Thermo-Fisher (Waltham, MA, USA). Formic acid, quinic acid, shikimic acid, malic acid, citric acid, tartaric acid, and oxalic acid were of HPLC grade and purchased from Sigma-Aldrich (St. Louis, MO, USA). HPLC grade of glucose, sucrose, fructose, maltose, arabinose, and galactose were from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). The 3-octanone standard was from TCI Chemical Industry (Tokyo, Japan). Other reagents used were of analytical grade and bought from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China).

Coffee Preparation
Coffee beans were grounded by the semi-automatic grinder (E10, HERO, Beijing, China) and sieved to obtain 30-mesh to 60-mesh coffee powders for hot and cold brew coffee preparation. On the basis of previous studies with some modifications [13,14], a coffee machine (NC-F400, Panasonic, Tokyo, Japan) was used for hot brewing with a waterto-powder ratio of 10:1 (w/w). Briefly, 60 g of coffee powders were placed on a coffee filter paper (No. 4, Melitta, Germany) after 600 g of purified water was added into the machine. The water was heated to brew coffee with the outlet coffee extract temperature at 80 ± 3 • C, with the brewing time of approximately 6 min. For cold brew coffee, 60 g of coffee powders were placed into a 300-mesh filter bag and let soak in 600 g of purified drinkable water in a glass jar at room temperature (20 ± 3 • C) for 8 h, with the jar lid closed. Additional 120 g purified drinkable water was applied to rinse coffee powders and the rinsed liquid was collected into the jar and mixed with previous obtained cold brew coffee. The liquid in the jar was filtered through a coffee filter paper before sending for stock. Referring to industrial processing, hot and cold brew coffee were diluted to standardize to 2% total dissolved solids (TDS) for sensory and subsequential instrumental analysis, which was consistent with previous studies [6,12]. Hot and cold brew coffee preparation was performed in three replicates for each independent experiment.

Extraction of Volatile Compounds
Hot and cold brew coffee samples were sent to GC-MS to determine the volatile compounds. The extraction of volatile compounds in coffee was conducted using solid phase microextraction (SPME) as described by Pan et al. [15], with slight modifications. Coffee (4 mL) was transferred into a 20 mL headspace bottle containing 2.5 g NaCl and 50 µL of 3-octanone (8.2 mg/L in methanol, as internal standard). The bottle was sealed by PTFE-silicone septum and equilibrated at 60 • C for 10 min with agitation. Next, a 50/30 µm divinylbenzene/carboxenTM/polydimethylsiloxane SPME fiber was exposed to the headspace of the coffee for 45 min at the same temperature without stirring. Finally, the fiber was obtained and introduced into the gas chromatography injector at 250 • C for 5 min.

Gas Chromatography-Mass Spectrometry
GC-MS analysis was carried out according to the method described by Heo et al. [14] using an Agilent 7890 gas chromatography system (Agilent Technologies, Santa Clara, CA, USA) equipped with an Agilent 5975C series mass spectrometer. The volatile compounds were isolated with HP-INNOWAX (30 m × 0.25 mm i.d. × 0.25 µm; Agilent Technologies) fused silica capillary columns. The carrier gas was helium at a rate of 1 mL/min constant flow. The oven temperature was held at 40 • C for 10 min, ramped at the rate of 8 • C/min to 180 • C, followed by ramping to 280 • C at the rate of 10 • C/min for 10 min. Mass spectrometry was carried out at an electron impact mode of 70 eV with a scan range of m/z 33-550. The volatile compounds were identified on the basis of the matching of the mass spectra with those in the standard NIST10 database. The quantification of volatile compounds was performed using peak areas normalized with 3-octanone added to each sample as an internal standard.

Identification of Aroma-Active Compounds
Odor activity value (OAV) was the ratio of concentration to odor threshold [16]. In most cases, the OAV value > 1 means the compound contributes to the overall coffee aroma profile.

Organic Acid and Sugar Profile Analysis
Coffee samples of 5 mL were centrifuged at 9000× g for 5 min, and the supernatant was collected for chromatographic analysis. According to previous reports [17,18] with some modification, the supernatant was passed through a 0.22 µm filter membrane before being injected into a Diane ICS-3000 ion chromatography coupled with a pulsed amperometric detector for analysis.
Acid chromatographic separation was achieved using the ion chromatography column (Dionex IonPacTM AS11-HC, 4 × 250 mm), with 20 µL injection volume and 0.60 mL/min flow rate, at 65 • C. Sulfuric acid solution (0.005 mol/L) was set to the eluent. External Standard solutions were applied for qualitative and quantitative analysis, which were quinic acid, shikimic acid, malic acid, tartaric acid, oxalic acid, and citric acid.

Ultra-Performance Liquid Chromatography-Tandem Mass Spectrometry (UPLC-MS/MS)
The extract was vortexed for 5 min, 500 µL coffee was added to 500 µL of 70% methanol solution, and then the extract was vortexed for 15 min. Following centrifugation at 12,000× g for 3 min, the extract was filtered through a 0.22 µm filter (SCAA-104; ANPEL, Shanghai, China) before analysis by UPLC-MS/MS.
Linear ion trap and triple quadrupole scans were performed using a triple quadrupolelinear ion trap (QTRAP) mass spectrometer (AB4500 QTRAP UPLC/MS/MS System). Sequence analyses were carried out in positive ion mode (ion spray voltage of 5500 V) and negative ion mode (ion spray voltage of −4500 V) with an ion source temperature of 550 • C. The pressures of the ion source gases I and II and the curtain gas were set at 345, 414, and 172 kPa, respectively. Instrument tuning and mass calibration were performed with 10 and 100 µmol/L polypropylene glycol solutions in triple quadrupole and linear ion trap modes, respectively. The data were collected at a scan range of m/z 100−1200.
The triple quadrupole scans were performed as multiple reaction monitoring (MRM) mode with the pressure of the collision gas (nitrogen) set to medium. The declustering potential and collision energy were selected for the individual MRM transitions and were then further optimized. A specific set of MRM transitions were monitored for each period according to the compounds eluted within this period.

Color Analysis
Colors of the hot and cold brew coffee were measured using a spectrophotometer (Hunterlab ColorQuest XE) at the transmission mode, D65 illuminant, 0.375 inch observation aperture, and 10 • observation angle. A cuvette with a light path of 5 mm was applied in this study. The L*a*b* values of hot and cold brew coffee samples were determined after calibration. L* represents lightness, and a* and b* are chromaticity indexes. The total color difference (∆E) of the two coffee was calculated via

Sensory Evaluation
All sensory evaluation was conducted at room temperature. Coffee samples were brought out from the refrigerator and let set at room temperature until 20 ± 3 • C for panel tasting.

Sensory Triangle Test
A total of 48 university students were recruited (age 19-22, 18 males and 30 females) for the sensory triangle test. According to ISO [20], three randomly coded hot and cold brew coffee samples were presented to the sensory panelist, two of which were the same and the other was different. The sensory panel was asked to select the sample different from the others. The test was carried out under red illuminant in order to avoid the interference of the coffee color difference.

Sensory Quantitative Descriptive Analysis
The descriptive sensory panel consisted of 7 university students (age 19-22, 2 males and 5 females) with regular coffee consumption at least 3 years. All panelists previously participated in at least 3 rounds of training on the identification and differentiation of the basic sensory attributes in coffee. To train panelists to understand sensory attributes of coffee, description vocabulary and their intensity scoring standards of coffee sensory evaluation were as shown in Table S1 [5,21]. After training, intensities of sensory attributes in hot and cold brew coffee were evaluated by the panel. Consistent with other sensory quantitative descriptive analysis [22], the panel was asked to evaluate coffee visual appearance and gave a score for color under the natural lighting condition, then took a deep smell of the coffee to evaluate the intensity of aroma before drinking it to evaluate its taste. The sensory description vocabulary of coffee included dark-colored, nutty aroma, coffee aroma, fruity aroma, sweetness, sourness, bitterness, and astringency, with the intensity score ranging 1-5 points.

Data Processing and Multivariate Statistical Analysis
Data were reported as mean ± standard deviation. Student's t-tests were performed using SPSS Statistics 26 (IBM, Armonk, NY, USA) at a 95% significant level.
The LC-MS/MS raw data signals were processed by Analyst 1.6.3 software (AB Sciex, Framingham, MA, USA). Principal component analysis (PCA), hierarchical cluster analysis (HCA), and orthogonal partial least-squares discriminant analysis (OPLS-DA) were carried out by the R software package (www.r-project.org) to visualize the differences between hot and cold brew coffee. Heatmap analysis and cluster analysis of non-volatiles were performed using R on the basis of their signal abundances in hot and cold coffee.

Volatile Compound Profile
As shown in Table 1, a total of 40 volatile compounds were identified, including 12 pyrazines, 9 phenols, 7 furans, 4 aldehydes, 3 ketones, and others. To further evaluate the potency of aroma compounds to the flavor profile, OAV was determined, as shown in Table 1. As shown in Figure 1, concentrations of 13 volatiles in coffee exceeded their respective threshold value, and thus these compounds with OAV > 1 have high chances of being the major aroma contributors to coffee. Among them, 11 aroma-active compounds have been identified in previous research on coffee brewed with French press, coffee machine, and in ready-to-drink coffee [23,24].
Pyrazines and furans were reported as the largest shares of volatiles in coffee [25], mainly formed by the Maillard reaction during coffee bean roasting [3]. As key contributors to coffee aroma, OAVs of total pyrazines accounted for over 60% of 13 aroma-active compounds, whose contribution in cold brew coffee was almost as twice as that of in hot brew coffee (Table 1), among which, 2-methylpyrazine was only detected in cold brew coffee, requiring further study to be clarified. Although abundant in coffee, furans made limited contributions to coffee aroma, due to relatively high recognition thresholds of furans [3]. Only the furfuryl acetate was identified as aroma-active furan (Figure 1), with higher OAV of 4.23 in cold brew coffee than in hot brew coffee. Generally, the high temperature increased the saturated vapor pressure, thereby adding the losses of volatile compounds [26]. It can be used to explain higher contents of pyrazine and furan compounds in cold brew coffee, which was consistent with previous studies [9]. Moreover, 4-ethylguaiacol, 3-ethylphenol, 4-vinylguaiacol, and guaiacol, as common phenolic volatiles identified as aroma-active compounds of coffee, were produced from the degradation of chlorogenic acids during green coffee bean roasting [3]. Different phenolic volatile change tends were observed between hot and cold brew coffee, which could be explained in brewing methods could affect concentrations of these compounds through both the extraction temperature and time.
Among aroma-active compounds, linalool and methyl salicylate were only detected in cold brew coffee, and the content of N-methylpyrrole-2-carboxaldehyde in cold brew coffee was significantly higher than hot brew (Table 1). During the fermentation process of green coffee beans, linalool and methyl salicylate can be produced by various microorganisms [27], which can still keep in detectable concentrations after coffee bean roasting [28]. As an endogenous component in coffee, however, N-methylpyrrole-2-carboxaldehyde was potentially hazardous, which needs to be avoided during coffee preparation [29].  Among aroma-active compounds, linalool and methyl salicylate were only detected in cold brew coffee, and the content of N-methylpyrrole-2-carboxaldehyde in cold brew coffee was significantly higher than hot brew (Table 1). During the fermentation process of green coffee beans, linalool and methyl salicylate can be produced by various microorganisms [27], which can still keep in detectable concentrations after coffee bean roasting [28]. As an endogenous component in coffee, however, N-methylpyrrole-2-carboxaldehyde was potentially hazardous, which needs to be avoided during coffee preparation [29]. Table 1. Volatile compounds identified in hot and cold brew coffee using head space solid phase microextraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS).

Differential Non-Volatiles Identified in Hot and Cold Brew Coffee
Sugars and organic acids were basic taste compounds in coffee. As shown in Table 2, sucrose had the largest share of the sugar content, with 15.37 ± 0.86 µg/mL and 14.26 ± 0.72 µg/mL in hot and cold brew coffee, respectively. Consistent with other research, sucrose accounted for the majority of the coffee sweet non-volatiles in coffee beans [41]. Chen and Ho [42] reported that monosaccharides were involved in the Maillard reaction for flavor generation during coffee bean roasting, which might lead to much lower contents of fructose and glucose in coffee. As for the organic acid profile in coffee, no significant difference was observed in most organic acids (Table 2). In a word, a few differences were detected in the sugar and organic acid profiles between hot and cold brew coffee. Moreover, Casas et al. [43] reported that levels of overall amino acids dropped to about 1% of green coffee beans after roasting, which made limited contribution to non-volatile profile of coffee. Values are given as means ± standard deviation (n = 3 independent measurements). * in the same row means significant difference at p < 0.05. Therefore, UPLC-MS/MS was applied to clarify the difference of phenolic acids, alkaloids, and flavonoids between hot and cold brew coffee. Our results showed a total of 322 non-volatiles were identified in hot and cold brew coffee. As shown in Figure S1, hot and cold brew coffee could be distinguished. Moreover, the mix samples (quality controls, QC) were clustered in the center of the PCA score plot, which indicated that the instrument exhibited high stability during data acquisition. Differential non-volatiles were identified by a variable importance in projection score of ≥ 1 and a fold change score of ≥ 2 or ≤ 0.5, which are visually shown in Figure 2A and Table 3. Hot and cold brew coffee were discriminated by 18 differential non-volatiles, including 8 phenolic acids, 5 terpenoids, 3 alkaloids, 1 flavonoid, and 1 coumarin. As shown in Figure 2B, 16 differential non-volatiles showed higher abundance in hot brew coffee, which could be explained in that higher brewing temperature flavored the kinetic energy of water molecules. Due to increased mobility and higher physical forces, non-volatiles were leached out more efficiently from the coffee bed [8].
were found to be extracted after 5 min of hot filter brewing. As for cold brew coffee, concentrations of 3-caffeoylquinic acid and caffeine were described to be stable after 400 min [13]. Therefore, it may indicate that in our study, with either high-temperature-short-time or low-temperature-long-time brewing methods, the extraction of caffeine, trigonelline, and caffeoylquinic acids in hot and cold brew coffee might have reached equilibrium due to relatively high solubility in water.
(A) (B) Figure 2. Volcano plot (A) and hierarchical cluster analysis (B) of differential non-volatiles identified between hot and cold brew coffee. (A) Red dots represent differential compounds of which the abundance was upregulated; green dots represent differential compounds of which the abundance was downregulated; and gray dots represent compounds with insignificant differences. If the abscissas had large absolute values, the FC values were also observed to be large. (B) Red represents a relatively high content of the compound, whereas green represents a relatively low content. HB: hot brew coffee; CB: cold brew coffee.

Figure 2.
Volcano plot (A) and hierarchical cluster analysis (B) of differential non-volatiles identified between hot and cold brew coffee. (A) Red dots represent differential compounds of which the abundance was upregulated; green dots represent differential compounds of which the abundance was downregulated; and gray dots represent compounds with insignificant differences. If the abscissas had large absolute values, the FC values were also observed to be large. (B) Red represents a relatively high content of the compound, whereas green represents a relatively low content. HB: hot brew coffee; CB: cold brew coffee. Table 3. Differential non-volatiles identified in hot and cold brew coffee using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) with the VIP 1 ≥ 1 and the fold change score of ≥ 2 or ≤ 0.5. As essential bioactive compounds in coffee [3], the abundance of 3-caffeoylquinic acid, 4-caffeoylquinic acid, 5-caffeoylquinic acid, caffeine, and trigonelline was shown to be comparable in hot and cold brew coffee (Figure 2A), which was in agreement with previous studies [13,16]. Cordoba et al. [4] reported that over 80% of soluble materials were found to be extracted after 5 min of hot filter brewing. As for cold brew coffee, concentrations of 3caffeoylquinic acid and caffeine were described to be stable after 400 min [13]. Therefore, it may indicate that in our study, with either high-temperature-short-time or low-temperaturelong-time brewing methods, the extraction of caffeine, trigonelline, and caffeoylquinic acids in hot and cold brew coffee might have reached equilibrium due to relatively high solubility in water.

Color Analysis
The L*a*b* values of hot brew coffee were L* = 29.76 ± 1.70, a* = 23.21 ± 0.61, and b* = 38.57 ± 2.43, while the color of cold brew coffee was L* = 38.79 ± 0.78, a* = 19.14 ± 0.91, and b* = 50.19 ± 3.95. The total color difference value ∆E = 15.27 suggested that the visual appearance of hot and cold brew coffee prepared from the identical coffee beans could be easily distinguished, which was consistent with our sensory quantitative descriptive study (Table 4). To be more specific, hot brew coffee presented darker lightness and redder tone, while cold brew coffee appeared lighter and yellower, which was also observed in medium-roasted coffee brews originating from El Salvador [45]. However, different trends in color changes may be shown in brews from different coffee bean origins, despite similar roasting levels. Sensory intensity values are given as means ± standard deviation (n = 7 sensory panelists). * in the same row means significant difference at p < 0.05.
The brown color of coffee is related to melanoidins, a class of high-molecular-weight products originating from the Maillard reaction [46]. Melanoidin compounds are more available for extraction with higher temperatures, likely due to the increase in the temperaturedependent solubility [12]. Therefore, hot brewing might make it easier for melanoidin extraction, which contributes to the darker color.

Sensory Evaluation and Differential Flavor Markers
In the sensory triangle test, a total of 25 panelists correctly identified the coffee sample that was different from others, which was greater than the critical value of 23 (α = 0.05), indicating the significant sensory difference between hot and cold brew coffee could be perceived. To further clarify differential sensory attributes between hot and cold brew coffee, results of the coffee sensory QDA are shown in Table 4.
Hot brew coffee appeared darker and tasted bitterer and more astringent over cold brew coffee (Table 4), consistent with prevoius studies [11]. Phenolic acids, alkaloids, and flavonoids were reported as main contributors to the bitterness and astringency of coffee [3]. Dicaffeoylshikimic acid, 2-methoxycinnamic acid, and thihydroxycinnamoylquinic acid, as identified phenolic acids in this study ( Figure 2B), might be used to explain the difference of bitter and astringent attributes between hot and cold brew coffee. Moreover, as one of the flavonol-3-glycosides, quercetin-3-O-(6"-O-p-coumaroyl) galactoside was identified to be the differential non-volatile compound with a higher abundance found in hot brew coffee. Flavonol-3-glycosides were reported to induce a velvety and mouth-coating sensation with low thresholds, which contributed to astringency in tea and red wine [47,48], although without bitterness on their own, flavonol-3-glycosides were found to amply the bitter taste of caffeine in tea [47]. Therefore, further study is needed to confirm the contribution of quercetin-3-O-(6"-O-p-coumaroyl) galactoside to the bitter and astringent tastes in coffee ( Figure 2B). In addition to non-volatile compounds, 4-ethylphenol and 4-ethylguaiacol were reported to increase the intensity and persistence of astringency of the flavanol solutions in the wine system [49]. Therefore, volatile contents might also need to be considered when concerning the bitter and astringent perception of coffee (Table 1).
As shown in Table 4, cold brew coffee showed higher fruitiness and sweetness intensity. Coniferaldehyde, as an identified key non-volatile in hot brew coffee, was reported to attenuate the fruity perception in red wine, even with concentrations lower than the threshold [50]. It implied that the fruity sensation in hot brew coffee might be decreased by the presence of coniferaldehy ( Figure 2B). Sweet and fruity perception of coffee was not only determined by non-volatile but also volatile compounds in foods. Linalool as an aroma-active compound with sweet and fruity notes [51], showing a higher level in cold brew coffee compared with hot brew coffee (Table 1). Interestingly, Barba et al. [52] have reported that linalool in fruit could significantly improve the sweet flavor perception of fruit juice. Hence, a similar result could be found in the presence of furfuryl acetate with the sweet smell in coffee [24]. Moreover, Yu et al. [53] reported that benzaldehyde and furfural increased the intensity of sweet aroma perception in Huangjiu even with OAVs < 1. Thus, higher contents of benzaldehyde, furfural, linalool, and furfuryl acetate might also enhance the sweet and fruity perception in cold brew coffee (Table 1). However, contributions of these differential flavor compounds would need further verification on the basis of more extensive investigations on larger scale of hot and cold brew coffee production, involving more general coffee consumers.

Conclusions
In conclusion, the effects of hot and cold brewing on the flavor profile of coffee appeared significantly different, with 13 aroma-active compounds and 18 differential nonvolatiles identified using chromatography-mass spectrometry and multivariate statistical analysis. Hot brew coffee was more closely associated with differential non-volatiles, while cold brew coffee was more associated with aroma-active compounds. A higher abundance of quercetin-3-O-(6"-O-p-coumaroyl) galactoside might be correlated with stronger bitterness and astringency in hot brew coffee, whereas higher OAVs of aromaactive furfural acetate and linalool led to the fruitier and sweeter flavor in cold brew coffee. These findings provide an insightful understanding of the coffee flavor control based on the sensory perception due to brewing methods during coffee industrial processing.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/foods11192968/s1, Figure S1: Principal component analysis (PCA) of the overall non-volatiles in hot and cold brew coffee using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). HB: hot brew coffee; CB: cold brew coffee; QC: quality controls. Table S1: Hot and cold brew coffee sensory attributes and their 5-point scale reference for panel training.