Optimization of Adsorption and Desorption Time in the Extraction of Volatile Compounds in Brewed Java Arabica Coffee Using the HS-SPME/GC-MS Technique

Received: 7th August 2021 Revised: 18th February 2022 Accepted: 26th February 2022 Online: 28th February 2022


Introduction
One of the plantation products with a high-value sale is coffee. The value of coffee exports from 2016 amounted to US$ 1,008,549, an increase in 2017 reached US$ 1,187,157 [1]. The increase in coffee export results is one of the advantages for society who have a livelihood as coffee farmers. This increase is because the coffee plantation business mostly comes from society's plantations. One of the most commonly cultivated coffees by the community is Arabica coffee, which has a distinctive taste and special market [2]. Arabica coffee is often used to improve the taste and aroma of instant coffee products [3]. One of the Arabica coffees that have become worldwide from Indonesia is Java Arabica which has a characteristic thin aroma of spices.
One of the qualities of coffee is determined by its aroma. Aroma is one factor that affects the sensory perception of food originating from the food product itself [4]. Compounds that determine coffee aroma can be identified by Gas Chromatography-Mass Spectrometry (GC-MS) instrument [5]. Volatile compounds, such as 2,5-dimethyl pyrazine; 2,6-dimethyl pyrazine; acetic acid; 2-acetyl furan; and 1-furfuryl pyrrole are present in coffee [6]. Ribeiro et al. [7] succeeded in detecting coffee volatile compounds, such as dihydro-2-methyl 3(2H) furanone; methyl pyrazine; N-furfuryl pyrrole; 2furfuryl alcohol; 5-methyl furfural; and 2-furfuryl acetate. These compounds give different aroma effects on coffee, such as 2,5-dimethyl pyrazine compounds, 2,6dimethyl pyrazine, and methyl pyrazine gives a nutty aroma [5]. In contrast, the 2-furan alcohol compound affects the burning, sweet, and caramel aroma [8]. These studies used the Headspace Solid Phase Microextraction (HS-SPME) technique to extract volatile compounds from brewed coffee. SPME is one of the best sampling methods in extracting and concentrating volatile compounds simultaneously from a sample [9] without using a solvent in the extraction process. The process of extracting volatile compounds in coffee using the SPME method includes compound absorption/adsorption and compound release (desorption) on SPME fiber adsorbents. The adsorption and desorption processes are essential in analyzing volatile compounds in brewed coffee. Therefore, the length of time required for both processes greatly determines the concentration and number of compounds that can be analyzed using GC-MS [10,11]. Yang et al. [5] and Giacalone et al. [6] investigated the volatile compounds from brewed coffee using HS-SPME with 5 minutes of adsorption time and a desorption time of 15 minutes. While the research was conducted by Ribeiro et al. [7] has 20 minutes of adsorption time and a fast desorption time. The difference in the length of time for adsorption and desorption can also be seen in the research of Bressanello et al. [12], where the required adsorption time was 40 minutes and the desorption time was 5 minutes. Based on previous studies, the optimal length of time for adsorption and desorption in coffee analysis has not been studied; therefore, the purpose of this research is to acquire the optimal adsorption and desorption time when analyzing the volatile compounds in brewed Arabica coffee using HS-SPME/GC-MS.

Experimental Section
This research was conducted in the laboratory of the Ma Chung Research Center for Photosynthetic Pigments (MRCPP), Ma Chung University, Malang. The study focused on obtaining the optimal adsorption and desorption time in analyzing volatile compounds from arabica coffee using the HS-SPME/GC-MS method. Volatile compounds from detected arabica coffee will be selected based on the peak area larger than 2%. Identification of volatile compounds based on the mass spectrum compared to the library Wiley Registry of Mass Spectral data 9th Edition with a similarity percentage above 90%. In addition, the volatile compounds were compared with references from reference libraries [13,14], the Mass Bank of North America (MoNA) database, and the National Institute of Standards and Technology (NIST). The flow chart of the research to be carried out can be seen in Figure 1.

Equipment and Materials
The main raw materials used were Java Arabica coffee beans, namely Banjarnegara coffee from Central Java for optimizing adsorption time, and Malang coffee from East Java for optimizing desorption time with a medium roast. The solvent used in this study was distilled water with chromatography-grade from Merck (Darmstadt, Germany). The tool used to grind coffee beans was the Philips Coffee Grinder HR2185 (China). The tools used in coffee extraction were the Labinco L32 Basic Hotplate and magnetic stirrer (Breda, The Netherlands) and the Hermle Z 160 M centrifuge (Wehingen, Germany). The volatile compounds were analyzed by SPME Fiber Assembly 50/30 m DVB/CAR/PDMS with Manual Holder Supelco (Bellefonte, USA) and GC-MS Shimadzu QP2010 Plus (Kyoto, Japan) with RTX-WAX column (30 m × 0.25 mm). × 0.50 m) (Bellefonte, USA).

Volatile Compounds Analysis Using GC-MS
Analysis of volatile compounds was conducted based on the modified method of Yang et al. [5] and Ribeiro et al. [7]. Ten grams of coffee beans were ground into powder using a grinder for 1 minute and used directly without a sifting process. After that, 0.1 grams of coffee powder was put into the HS-SPME vial, added 10 ml of boiling water, and then tightly closed with a septum. The coffee solution was homogenized using a magnetic stirrer and incubated at 70℃. After 10 minutes, the vial containing the sample was immersed in a glass beaker filled with tap water for 10 minutes to lower the temperature. The schematic stages of the analysis of coffee volatile compounds can be seen in Figure 1. After the temperature of brewed coffee dropped to ±25℃, the adsorption process was carried out on the sample (number 10), then continued with the desorption process on the GC-MS injector (number 11).
GC-MS Shimadzu QP2010 Plus was set with the following parameters: injection temperature (235℃), ion source temperature (200℃), interface temperature (235℃), splitless mode, pressure (103 kPa), and flow rate of Helium gas is 9 mL/min. The oven column temperature was set at an initial temperature of 50℃ for 15 minutes, increased with a heating rate of 3℃/min to 220℃, then expanded with a heating rate of 8℃/minute to 240℃ held for 5 minutes. The column used was RTX WAX. Volatile compounds were detected by the scan mode with a range of 35 to 500 m/z and a 1,666 Da/sec scanning rate.

Results and Discussion
Adsorption and desorption processes were crucial in volatile compounds analysis from coffee. The amount of adsorbed volatile compounds can be increased by optimizing extraction time or, in other words, adsorption time [15]. The optimizing also applied to the desorption time, which affected the high yield [16]. Adsorption of volatile compounds from coffee using SPME was carried out on coffee grounds that have been brewed with hot water; thus, the types of volatile compounds identified by GC-MS correspond to the volatile compounds felt by coffee connoisseurs. minutes desorption using the HS-SPME/GC-MS technique. Information on peak numbers is shown in Table 1 The separation results of volatile compounds from 2 types of Java Arabica coffee using GC-MS with 20 minutes of adsorption time and 5 minutes of desorption time are shown in Figure 2. Fourteen volatile compounds ( Figure  3) have a peak area percentage of more than 2% from the optimization results adsorption and desorption times in coffee, where 5 of these compounds were found in the results of the two optimizations ( Table 1). Identification of volatile compounds was also carried out by analyzing molecular ions and fragment ions compared with reference libraries and the fragmentation pattern of the structure of the compounds.
An illustration of its identification was exhibited at the second peak, pyridine (C5H5N), which gave a sour and roasted aroma [17]. Pyridine, with an exact mass of 79.0422 Dalton (Da), eluted at a retention time (tR) 3.  Based on research conducted by Ribeiro et al. [7], analysis of volatile compounds from Arabica coffee was carried out using SPME with 20 minutes of adsorption time, followed by a desorption process of volatile compounds in a GC injector afterward. Therefore, the adsorption time was optimized in the 20-60 minutes range, with 5 minutes of desorption time in this study. The optimization of the adsorption time was carried out to determine whether the longer the adsorption time caused the content of the identified volatile compounds to increase or not. The optimization results of adsorption time based on the peak area value and the percentage of the peak area of the seven volatile compounds detected in Java Arabica coffee are served in Table 2.   The detected volatile compounds had the same amount and type at the adsorption time of 20 to 60 minutes. There were similarities in the variation of the optimization time, where 2-furfural, 2-furfuryl alcohol, and 5-methyl furfural were the volatile compounds with the highest percentage in Java Arabica coffee, where this result was similar to the study of Zakidou et al. [18]. Research from Caporaso et al. [17] also mentioned that 2furfuryl alcohol was the most abundant compound in arabica coffee.

Retention time (minutes)
The results ( Table 2) also show that the longer the adsorption time, the greater the peak area of the compound for the adsorption time from 20 minutes to 50 minutes. The reason is that at the adsorption time of 20 minutes, the ability of the SPME adsorbent to bind compounds is still immense. Meanwhile, at the adsorption time of 50 minutes, the SPME adsorbent could maximally bind the volatile compounds. Therefore, at the adsorption time of 60 minutes, the content of volatile compounds detected was less than the adsorption time of 50 minutes because the volatile compounds were no longer able to be absorbed by the SPME adsorbent and were likely to be rereleased. These compounds were 2-furfural (8), 2-acetyl furan (9), 2-furfuryl acetate (11), 5-methyl furfural (12), and 2-furfuryl alcohol (14). While the volatile compounds 3(2H)-furanone, dihydro-2-methyl-(4) had a larger peak area with 20 minutes of adsorption time, in contrast, the largest peak area of acetic acid (7) was at 60 minutes. Meanwhile, volatile compounds had the largest area for 50 minutes the adsorption time; thus, the most optimal adsorption time was 50 minutes. These results were similar to the optimization of the extraction time of volatile compounds in Arabica coffee using static headspace gas chromatography (SH-GC), where all compounds were detected within 30 minutes, while many compounds had a higher content (in %) at 45 minutes of extraction process [19].  Table 2 and Table 3 In contrast to the adsorption time, which only affects the amount of volatile compounds content, the difference in desorption time also affects the number of compounds identified. Desorption times of 5, 15, 30, 45 minutes were chosen to determine the best time for the volatile compounds to be released from the SPME adsorbent and detected by GC-MS. The use of 5 minutes of desorption time refers to Bressanello et al. [12], Mondello et al. [20], and Zakidou et al. [18], as well as standard coffee cupping, where the aroma is tested at 4 minutes after coffee being brewed with hot water. In addition, Shimadzu's application notes on analyzing coffee volatile compounds using SPME desorption for 2 minutes [21]. Yang et al. [5] and Laukaleja and Kruma [22] used 15 minutes of desorption time to analyze volatile compounds from discrepancies in the coffee bean roasting process. In addition to the two desorption times, two long periods of 30 and 45 minutes were used in testing the optimization of desorption time.
The results of the desorption time optimization of the volatile compounds of Java Arabica coffee are illustrated in Figure 4 (B) and Table 3. Fourteen volatile compounds were detected at 5 minutes of desorption time and eight compounds at 15, 30, and 45 minutes with the percentage of peak area above 2%. Three volatile compounds, such as pyridine, furfuryl formate, and 2furfuryl alcohol, have 10-19% of peak area percentage, the largest compared to other compounds, where other compounds only have <10% peak area. Four volatile compounds (peaks 1, 2, 3, and 5) that appeared at less than 10 minutes retention time were not detected at desorption times of 15, 30, and 45 minutes. This result indicated that the longer the desorption time caused some compounds to be wasted into the sewer in the GC-MS, such as the compound 1H-Pyrrole, 1-methyl-(1); pyridine (2); limonene (3); methyl pyrazine (5). Nonetheless, at the peaks of 6, 8, 9, and 13, the highest area was at 15 minutes desorption. While at 5 minutes desorption, the peaks of 10 and 14 were the most significant area. The highest area at peaks 11 and 12 was at 45 minutes. These results indicated that a rapid desorption time allowed more volatile compounds to be detected. Long desorption time served to obtain the full content of certain volatile compounds. However, the analysis of volatile compounds in coffee required complete data. Therefore, the most optimal desorption time was 5 minutes because it could detect more compounds, and there were two volatile compounds with the most significant area compared to other desorption times. The optimizing result of 5 minutes desorption time in this study was in accordance with the information on the desorption time in the research regarding the identification of volatile compounds in coffee [12,20] for 5 minutes and application notes Shimadzu [21] for 2 minutes.

Conclusion
Coffee volatile compounds were extracted using the HS-SPME technique, which included an adsorption process for compound extraction and desorption to release compounds into GC-MS. The optimization results of adsorption and desorption time detected 14 volatile compounds in Java Arabica coffee with a peak area percentage of more than 2%. Seven of 14 compounds were found at each variation of the adsorption time. It could be done for 20 minutes of adsorption to detect coffee volatile compounds. However, the most optimal adsorption time was 50 minutes because five of volatile compounds, namely 2-furfural (456.8 × 10 5 ), 2-acetyl furan (52.0 × 10 5 ), 2-furfuryl acetate (102.5 × 10 5 ), 5-methyl furfural (183.6 × 10 5 ), 2-furfuryl alcohol (227.7 × 10 5 ) had the most prominent peak area. In contrast, the desorption time affected the number of detected volatile compounds. The optimal desorption time in this study was 5 minutes which detected 12 compounds, while at the other desorption time, only eight compounds were detected. This optimal adsorption and desorption time is vital in analyzing volatile compounds in coffee because it determines the concentration and number of compounds that can be analyzed using GC-MS effectively.