Characterization of volatile compounds in Swedish yellow and gray peas: Implications for new legume‐based ingredients

There is a growing demand for alternative protein‐source ingredients from domestically cultivated pulses in Europe, including Sweden. However, the use of legumes as a food ingredient is limited by the presence of a distinct beany flavor. Mapping the volatile compounds composition in a standardized approach will aid in comparing different legume varieties and processing treatments. The composition of volatile compounds in flour from yellow and gray peas (raw and boiled) was investigated and compared. Volatile compounds were isolated by headspace solid‐phase microextraction (HS‐SPME) and analyzed using gas chromatography‐mass spectrophotometry (GC‐MS). A total of 43 volatiles were identified, consisting mostly of aldehydes, followed by alkanes, alcohols, ketones, alkenes, furans, terpenes, aromatics, and sulfur‐containing compounds. Boiling led to a marked reduction in alcohols and an increase in aldehydes. Several markers of beany flavor, such as 1‐octen‐3‐ol, 2‐pentylfuran, and 3,5‐octadien‐2‐one, were significantly decreased after boiling. The composition of volatiles collected from yellow and gray peas was comparable, but boiled yellow pea had a higher abundance of beany flavor as compared to gray pea. Gray pea is an interesting variety to be explored further as a potential alternative to the well‐known yellow pea.

over, the water content in cooked mashed pea might be highly varied depending on the raw pea seeds; thus, it will add a challenge to adjust the product formulation. Hence, it motivates additional studies to investigate the factors affecting volatile profiles in raw pea flour as well as boiled peas that have been dried and milled, to obtain flours suitable for food applications.
Headspace solid-phase microextraction (HS-SPME), in combination with gas chromatography (GC) and mass spectrophotometry (MS), is commonly used for profiling volatile compounds in various matrices (Vas & Vékey, 2004). HS-SPME is a simple, sensitive, and fast sampling technique for collecting volatile compounds from samples without using any solvents (Vas & Vékey, 2004). The method only requires a small amount of sample and is cost-effective (Vas & Vékey, 2004). A number of researchers have analyzed the composition of volatile compounds and studied the effect of processing on the volatile profile in beans using HS-SPME-GC-MS (Jiang et al., 2016;Mishra et al., 2017;Oomah et al., 2014;Oomah & Liang, 2007;Szczygiel et al., 2017). However, only a few studies have focused on peas Ma et al., 2016;Murat et al., 2013;Xu et al., 2019). Furthermore, there are no studies on the volatile compound profile of yellow and gray peas cultivated in Sweden.
Therefore, the objectives of the present study were (i) to identify and compare volatile compounds in flours from raw Swedish yellow and gray peas and (ii) to investigate the effects of processing (i.e., a combination of soaking, boiling, and drying) on the volatile compound profile of pea flours. Data on the volatile compound profiles of yellow and gray peas flours can provide the basis for development of food products with little or no beany flavor, which is of particular interest to manufacturers of products such as beverages, bakery goods, snacks, or meat analog products.

| Materials
Dried yellow pea (Pisum sativum, Clara variety) and gray pea (Pisum sativum, unknown Latvian variety), grown in Öland, Sweden, were obtained from Kalmar-Ölands Trädgårdsprodukter (KÖTP), Kalmar, Sweden. All materials were harvested in 2017. The pulses were stored packed in a cardboard box at room temperature (20 C) for approximately 22 months until processing.

| Preparation of pea flours
Flour from raw pea: Whole dried yellow and gray peas were ground using a laboratory-scale mill (Cyclotec 1093, Tecator, Sweden). Flour from boiled pea: Dried peas were soaked in tap water (1:3 w/v) at room temperature for 14 h then boiled in water (1:5 w/v) until they get soft (50 min for yellow pea and 35 min for gray pea; Ferawati et al., 2019).
The boiled peas were dried in a convection oven at 50 C for 16 h.
The dried boiled peas were ground (500 μm particle size) using the laboratory-scale mill referred to above.

| Headspace solid-phase microextraction
Volatile compounds extraction from the pea flours was performed with a modification of published extraction conditions (Oomah & Liang, 2007), using freshly milled raw or dried boiled pea seeds. The ground samples (approximately 50 g) were analyzed within 4 h after milling. For this, pea flour was transferred directly after milling to three 100-ml Erlenmeyer flasks (10 g each), which were tightly covered with aluminum foil and kept in room temperature until analysis. One sample at a time was equilibrated at 50 C for 30 min in a water bath. After that, the volatile compounds were extracted by exposing a 1-cm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) SPME fiber (Supelco, Bellefonte, PA, USA) to the headspace of the sample for 30 min. The DVB/CAR/PDMS fiber is recommended for studies of volatile compounds in food and beverages due to its broad binding capacity for a variety of volatiles (Murat, Gourrat, Jerosch, & Cayot, 2012;Oomah & Liang, 2007;Vas & Vékey, 2004;Xu et al., 2019). The extraction was done in pace with the chromatography analysis. Extracted volatile compounds were desorbed by injecting and exposing the fiber at the injection port of the GC at 250 C for 10 min. All samples were analyzed within 24 h. The SPME fiber was conditioned before use at 270 C for 1 h, as recommended by the supplier. A blank extraction was also performed daily to detect contamination. The same fiber was used for all analyses.

| Gas chromatography-mass spectrometry (GC-MS) analysis
The volatile compounds from the samples were analyzed using GC-  (Mishra et al., 2017). The injection port was maintained at 250 C in splitless mode and fitted with a straight, narrow bore liner (0.75 mm I.D). The split vent was set at 80 ml/min at 1 min, and gas saver mode was 20 ml/min after 2 min.
Helium was used as the carrier gas in constant flow mode at 1 ml/min.
The GC oven temperature was programmed as follows: an initial temperature of 35 C was held for 5 min, then increased to 200 C at 4 C/min, followed by a 10 C/min increase to 280 C, which was held for 1 min.
Reference compounds and alkane series were analyzed by injecting 1 μl sample using an autosampler (Agilent 7683). The concentration of each compound was approximately 50 ng/ml in dichloromethane. All GC conditions were identical, except for the liner that was changed to a standard liner (4 mm I.D). The fiber used for the analysis of pea flour headspace was never in contact with the reference compounds in order to avoid contamination.
The mass spectrophotometer (MS) was operated in electron ionization mode at 70 eV. The ionization source temperature was set at 230 C and quadrupole at 150 C. The mass spectrometer scanned masses from m/z 35 to 500 and data collection started at 2.5 min after injection.
The volatile compounds were identified by comparison with a mass spectra library (Wiley 275/NIST 05; ≥80% match quality), reference compounds, and retention index (RI) based on a series of C7-C18 n-alkanes using the equation of Van Den Dool and Kratz (1963). Peaks with a total peak area ≥0.25 × 10 6 counts and a signal to noise ratio (S/N) ≥ 20 were selected. The level of each volatile compound was reported as the average of total area counts from three analyses.

| Statistical analysis
The total peak area was expressed as the mean of area counts ± standard deviation (n = 3). Significant differences in the content of each volatile compound identified in flours from raw versus boiled yellow pea, raw versus boiled gray pea, and boiled yellow versus gray pea were determined using Student's t test. The level of significance was set to 0.05. All statistical analyses were performed using GraphPad Prism 7.

| Sample analysis
A typical HS-SPME-GC-MS chromatogram contains a few large peaks and a high number of peaks that are close to the detection limit of the instrument. The present study aimed to identify as many peaks as possible in order to create a "volatile signature" of the different pea flours. Several small peaks below the limit of detection (  The risk of contamination from the environment, for example, from packaging materials, is an essential aspect to consider in HS-SPME-GC-MS studies. Plastic materials were found to contribute to contaminating peaks (data not shown), which is in agreement with others who found several volatile compounds from plastic packaging in the plastic-wrapped cheese (Panseri, Chiesa, Zecconi, Soncini, & De Noni, 2014). Thus, all plastics were excluded in the analytical procedure. Glassware was heat-treated at 105 C, and blank extractions from empty flasks were performed every day to exclude the possibility of contamination. Ground samples (10-g pea flour) were incubated at 50 C for 30 min to ensure that volatiles in the samples were released to the headspace, and then the SPME fiber was inserted for 30 min. Previous studies have shown that an extraction time of 30 min is enough to reach equilibrium Mishra et al., 2017). The same SPME fiber was used throughout the study to avoid differences due to manufacturing variations of the fiber.
It is important to note that the abundance of volatile compounds in the headspace is expressed in peak area counts in this
Different capital letters indicate a significant difference for each volatile compound between flour from boiled yellow and gray peas (t test, p < 0.05). The results agree with a previous HS-SPME-GC-MS study on volatiles from raw peas stored at room temperature in which aldehydes were identified as the most abundant group, followed by hydrocarbons and alcohols . These authors also reported that the storage at 4 C decreases the total peak area of volatile compounds and changes the composition in the headspace. As a result, alcohols and ketones were found to be the main compounds instead of aldehydes, which were explained by a lower lipid oxidation (Azarnia, Boye, Warkentin, Malcolmson, Sabik, et al., 2011). Others (Ma et al., 2016;Murat et al., 2012) have also reported that alcohols were the most abundant group of volatile compounds in raw yellow pea, but the storage time or temperature was not specified in these studies.
Hexanal, nonanal, octane, 1-octen-3-ol, and 2-pentylfuran were the most abundant volatile compounds in the present study (Table 1). These five compounds accounted for 66%-72% of the total peak area in the raw flour samples. Three out of five main volatile compounds found in the present study (hexanal, 1-octen-3-ol, and 2-pentylfuran) were among the markers for a beany flavor. Additionally, 3,5-octadien-2-one was observed in low abundance in raw yellow and gray peas. The aroma of odor active compounds such as 1-octen-3-ol is described as earthy, green, oily, and fungal, whereas furan-2-pentyl is described as having a green, earthy, beany, and vegetable aroma (Acree & Arn, 2004;The Good Scent Company, 2018). The aroma of 3,5-octadien-2-one is

| Effects of processing on the volatile compound composition of pea flours
Flour from boiled yellow pea had a significantly higher peak area of aldehydes than flour from raw yellow pea (p = 0.006; Table 1). Others have reported a similar increase in aldehydes after cooking of green bean and faba bean (Jiang et al., 2016;Rodriguez-Bernaldo De Quiros et al., 2000), due to the enzymatic and nonenzymatic oxidation processes promoted by heating (Jiang et al., 2016). The total peak area of volatiles in boiled yellow pea was higher than seen in flours from raw pea or boiled gray pea. The seemingly higher abundance of volatiles in boiled yellow pea might be related to the longer boiling time of yellow peas (50 min) as compared with gray pea (35 min). Moreover, there was an apparent decrease in the number of volatile compounds detected in flour from boiled yellow pea compared with raw pea flour.
As found for yellow pea, there was a notable increase in the peak area of aldehydes and a lower abundance of alcohols in flour from boiled gray pea (Table 1). Moreover, only 32 volatile compounds were detected in flour from boiled gray pea. Volatile compounds in flour from boiled gray pea consisted of aldehydes (75% of total peak area), alkanes (15%), furans (4%), alcohols (2%), ketones (1%), and others (2%).
Hexanal was the main compound found in flours from boiled yellow and gray peas (Table 1) and was detected at significantly higher levels (p = 0.0003) in boiled peas than in raw pea flours F I G U R E 1 Volatile compound composition of flours from raw and boiled yellow and gray peas. Bars represent the mean of total area counts from triplicate analyses ( Figure 2). A higher peak area of hexanal in flours from boiled peas might not necessarily be associated with a stronger beany flavor because hexanal lacks beany character by itself, but it is an indicator of heat treatment. We also observed a significant reduction in the peak area of 1-octen-3-ol and furan-2-pentyl in flours from boiled peas compared with raw flours (Figure 2) and did not detect any 3,5-octadien-2-one in flours from boiled peas (Figure 2). A reduction in the 1-octen-3-ol, furan-2-pentyl, and 3,5-octadien-2-one levels after processing is in agreement with previous findings (Mishra et al., 2017). Lower abundance of these compounds in flours from boiled peas suggests lower beany flavor concentration compared with raw pea flours.
The levels, in terms of peak area counts, of the three out of four beany marker compounds (i.e., hexanal, 1-octen-3-ol, and 3,5-octadien-2-one) were higher in flour from boiled yellow pea than in flour from boiled gray pea. The exception was 2-pentylfuran, which showed significantly (p < 0.0001) higher abundance in grey pea than in yellow pea after processing ( Figure 2). Overall, the results indicated that flour from boiled yellow pea has higher abundance in beany flavor causing compounds than flour from boiled grey pea. Therefore, the two types of flours from yellow and gray peas could be used in different food applications, depending on the desired characteristics of the final products. Moreover, a previous study has shown that gray pea had a higher content of resistant starch and folate than yellow pea (Ferawati et al., 2019). Thus, gray pea is an attractive commodity for further exploration due to a higher content of nutrients and a lower level of beany flavor compounds than the well-known yellow pea.
We believe that data from this study will contribute to the general knowledge on volatile profiles of different raw and processed legumes. Our study confirms that raw peas stored at room temperature have a volatile profile dominated by aldehydes. Also, the information on the length of storage of the pea seeds will be useful to compare findings from different studies. The abundance of aldehydes increased after heat treatment, but all other beany flavor compounds decreased. The data on the presence of beany flavor compounds (i.e., hexanal, 1-octen-3-ol, 2-pentylfuran, and 3,5-octadien-2-one) might be useful in, for example, comparing the quality of different harvests, varieties, and effects of storage and treatments. Information on the volatile compound profile of flours from different varieties of pea will be interesting to food industries seeking to increase the utilization of pea in food products. Based on our findings, HS-SPME-GCMS is a F I G U R E 2 Effect of boiling yellow and gray peas on potential beany flavor compounds: (a) hexanal, (b) 2-pentylfuran, (c) 1-octen-3-ol, and (d) 3,5-octadien-2-one. Bars represent the mean of area counts ± SD (n = 3). Asterisks indicate significant differences between samples (t test, * = p < 0.01; ** = p < 0.001)