Comparative analysis of flavor differences of six Chinese commercial smoked chicken

ABSTRACT This study was intended to characterize six types of commercial smoked chicken products in China by using gas chromatography-mass spectrometry (GC-MS), odor-activity values (OAVs), and sensory evaluation. Results demonstrated that a total of 89 components were identiﬁed in all samples, and 34 were considered as odor-active compounds because their OAVs were greater than one. Liaocheng Chicken that smoked with fruit tree sawdust had more phenols, which contributed to the smoky aroma. Jinshan and Goubangzi Chicken that smoked with sugar had more furans which contributed the overall odor with sweety and caramel aroma. Zhuozishan and Laoting Chicken that smoked with sugar and wood chips had similar flavor and volatile compounds. Tengqiao Chicken that smoked with sugar, tea and rice had significant difference with other chicken in smoky, bittern and caramel aroma (P< .05). The diversity of these smoked chicken flavors was mainly due to the cooking culture differences.


Introduction
Smoked meat products are typical traditional meat products. It refers to the cured or cooked meat that heated in the medium of smoke, high temperature gas or solid (Pöhlmann et al., 2013). There are many kinds of smoked meat products, such as bacon (Saldaña et al., 2018;Soladoye et al., 2017), smoked chicken, smoked ham, etc. (Jerković et al., 2007). In China, there are mainly two kinds of smoked meat products: Chinese bacon in western and southern China, smoked chicken in central and northern China. As a common smoked meat product, smoked chicken can be processed into many different taste and flavor products due to the different processing technology in ingredients and smoked materials. There are many commercial brands of smoked chicken in China, such as Liaoning Goubangzi Chicken, Shandong Liaocheng Chicken, Zhuozishan Chicken in Inner Mongolia and Laoting Chicken in Hebei Province, which have become the typical representative of the local food culture, with strong regional characteristics and histories more than 100 years. In recent years, Zhejiang Tengqiao Chicken and Inner Mongolia Jinshan Chicken are also widely welcomed by consumers. The geographical location of six smoked chicken products is shown in Figure 1.
Among the sensory attributes of smoked products, flavor has been rated as one of the most important indexes to evaluate the quality of smoked products (Marusic et al., 2016). Because of the diversity of food culture, there are some differences in flavor of the same product. Sam Al-Dalalia characterized the volatile compounds in three commercial Chinese vinegars by SPME-GC-MS and GC-O and a total of 75 volatile compounds were identified (Al-Dalali et al., 2019). Zhu studied the flavor of tomato juice from Garden Gem and Roma tomatoes with comparison to commercial tomato juice (Y. Zhu et al., 2018). Thanakorn Yimdee et al. studied the odor and taste components of 14 commercial brands of fish sauces from various Asian countries, and seventy-nine different types of volatile compounds were identified, described, and quantified by concentration and odor active value (Thanakorn & Wang, 2016). Smoked chicken with different regional characteristics present different flavor, however, there are few reports on the flavor research of the smoked chicken.
In this study, volatile compounds of the selected six top famous regional brands of smoked chicken, and the characteristic odor-active substances were analyzed by SPME-GC-MS (Zhou et al., 2016). The present research was intended to provide a useful reference for understanding consumers' taste preference in different regions and the later study of flavor regulation of smoked chicken products. Geographical classification can also be useful to help consumers select food items.

Materials and chemicals
The six selected brands of smoked chicken samples were purchased from corresponding companies in China. Goubangzi (GBZ) Chicken was provided by Goubangzi Company Among the six types of smoked chicken, GBZ Chicken was smoked with refined cane sugar, LC Chicken was smoked with fruit tree sawdust, ZZS Chicken was smoked with refined cane sugar and cypress sawdust, TQ Chicken was smoked with brown sugar, tea and rice, JS Chicken was smoked with brown sugar and cypress sawdust, LT Chicken was smoked with refined cane sugar and fruit tree sawdust. All samples were smoked for 5-8 min, except the LC Chicken, which was 24 h. All samples were at the most recent production dates and divided into two portions: one portion was used for the sensory evaluation, and another portion was frozen by liquid nitrogen. C 7~ C 30 normal alkanes standard samples were bought from Supelco (Shanghai, China). Cyclohexanone standard was obtained from Sigma-Aldrich (Castle Hill, Australia).

Sensory evaluation
Sensory evaluation has always been an important indicator in flavor research, and an effective method to detect the quality of smoked meat products. Twelve healthy, nonsmoking, trained panelists (9 females and 3 males between the ages of 18-22) from the College of Food Science and Engineering, Bohai University, were recruited in this study. Firstly, the panelists were asked to write down sensory descriptors for the chicken samples while smelling. Then, the descriptors were collected and four aroma attributes were finally determined by discussion, which were smoky, caramel aroma, bittern aroma, and meaty aroma (Dandan et al., 2019). After that, the panelists were asked to score the intensity of each sensory descriptor on a five-point scale from 1 to 5, which indicate very weak, weak, moderate, strong, very strong, respectively. Based on the specified sensory evaluation criteria, the panelists received several trainings before the experiment. After the evaluation of each sample, the panelists had 5 min for break. Within 1 h prior to the sensory experiment, the panelists were not allowed to eat or drink anything.

Headspace solid -phase microextraction -gas chromatography -mass spectrometry (SPME/GC -MS) analysis
According to the reference to extract the volatile compounds by using SPME method (Chen et al., 2018). The minced chicken sample (4.5 ± 0.20 g) was added into a 20 mL headspace vial. Then, 7 μL of internal standard (cyclohexanone) was injected, and the vial was sealed with PTFE-silicon stopper (Agilent). Before the extraction, the vial was placed under the condition of 50 °C constant temperature water bath for 10 min. Subsequently, the volatile compounds were extracted at 50°C for 45 min using a 75 μm SPME fiber (Carboxen/polydimethylsiloxane; Supelco, Bellefonte). After the extraction, the loaded SPME fiber was quickly inserted into the injection port of the GC-MS to desorb at 250°C for 4 min.
In this study, GC-MS (7890B GC System, 5977A MSD) were used and samples were analyzed on HP-5 MS capillary column (30 m × 250 μm×0.25 μm, Agilent Technologies, Santa Clara, CA). Helium (purity of 99.99%) was used as the carrier gas at a constant flow rate of 1.0 mL/min in splitless injection mode. The injector temperature was 250°C. The temperature program was as follows: the oven temperature was kept at 40°C for 3 min, then increased to 70°C at a rate of 3°C/min and increased to 180°C at 5°C/min. The temperature was then increased to 230°C at a rate of 10°C/min, and finally held at 230°C for 5 min. The mass spectral detection conditions were as follows: the temperature of the MS source was set at 230°C, MS fragmentation was detected in electron-impact (EI) mode (ionization energy of 70 eV) with an acquisition range from 30 to 550 m/z in full-scan mode.

Identification and quantitation of aroma compounds
Volatile compounds in smoked chicken sample were identified by matching the retention index (RI) and library spectra (MS). The comparison with the mass-spectrometry library was based on the NIST 2.0 mass-spectrometry database. RIs were determined using a series of standard n-alkanes (C 7 -C 30 ) under the same GC-MS-detection conditions.
Taking cyclohexanone as internal standard, the aroma components in smoked chicken were analyzed semi-quanti tatively. In addition, 7 μL of cyclohexanone was added into the sample and analyzed by GC-MS. The concentration of volatile compounds was calculated according to the ratio of peak area to cyclohexanone concentration.

Odor-activity values
To evaluate the contributions of aroma compounds to smoked chicken, the odor-activity values (OAVs) were calculated, which is the ratios of the concentration to the perception threshold. The perception-threshold value of each compound was determined as described in the literature (Czerny et al., 2008;Greger & Schieberle, 2007;Kerler & Grosch, 1997;Wu & Liou, 1992;J. Zhu et al., 2016). Aroma compounds with OAVs greater than 1 may have a major contribution to smoked chicken, whereas compounds with OAVs less than 1 indicate a minor contribution (Schieberle & Hofmann, 2011).

Statistical analysis
All experimental results were repeated in triplicate. All statistical analyses of volatile compounds from six types of smoked chicken were analyzed by SPSS 19.0 software (IBM Corporation). Results are expressed as the means ± standard deviations from three measurements. The significant differences (P < .05) were analyzed by the Duncan test. The PLSR analysis was processed by XLSTAT 2016, where x variables were the sensory attributes and y variables were the released volatile compounds with OAVs greater than one.

Sensory evaluation
To visually classify the sensory attributes of different kinds of smoked chicken, a heatmap was generated. As shown in the heatmap (Figure 2), color coding was graded on the basis of the scale from blue to red with the relative intensity increasing from low (blue) to high (red).
LC Chicken sample had the highest smoky aroma score and was significantly different with other chicken samples (P < .05). This may because LC Chicken was smoked for 24 h, which was the longest smoking time of the six chicken samples. Because of the long smoking time, more smoke enters the chicken and creating stronger smoky aroma of the sample. JS Chicken sample had the strongest caramel aroma. The meaty aroma was mainly produced in the process of chicken marinating. Six brands of chicken all went through the process of marinating, so the meaty aroma scores were relevant higher. The meaty aroma of GBZ sample was significantly stronger (P < .05), which may be related to the high fat content of GBZ Chicken. GBZ Chicken sample also had the strongest bittern flavor, which was significantly different from other chicken samples (P < .05). Compared   with other chicken samples, the smoky, bittern and caramel aroma of TQ Chicken was significantly weaker (P < .05), which may be caused by the tea and rice added in the smoky material.

Volatile composition of six smoked chicken
As shown in Table 1

OAVs of the odor-active compounds
The importance of aroma compounds in smoked chicken depends on not only the amounts of the compounds but also their OAVs and contribution rates. By calculating the OAV, the contribution of each flavor component to the smoked chicken was analyzed, so as to determine the main flavor components of the smoked chicken. The 34 aroma compounds with OAVs greater than 1 are shown in Table 2. The production of hydrocarbons (except terpenes) is closely related to lipid oxidation and thermal decomposition, which mainly comes from the break of alkoxy radical of fatty acid. D-limonene has a pleasant lemon aroma and contributed greatly to the overall flavor of six smoked chickens (OAVs > 1). The concentration of D-limonene in LC Chicken sample was significantly higher than other chicken samples (P < .05). D-limonene in smoked chicken may come from spices added during marinating, such as star anise, dried orange peel and pepper. Previous study showed that D-limonene accounts for 75.39% of the major constituents in the volatile oil of tangerine peel. P-Xylene, p-Cymene and γ-Terpinene were all detected in 6 smoked chickens, of which p-Xylene contributed the most significant to LC Chicken sample (OAV = 2.81, P < .05). p-Cymene had a great contribution to the flavor of LT and LC chicken samples, and mainly exists in cumin. γ-Terpinene played an important role in LC Chicken and mainly existing in spicy pepper and nutmeg. β-Ocimene has the aroma of grass and is mainly found in prickly ash. It had a great contribution to the flavor of GBZ, LT and TQ Chicken. Caryophyllene mainly comes from cinnamon and fragrant leaves, which had an important contribution to the flavor of LT and LC Chicken. In addition, styrene was only detected in GBZ sample with OAV greater than 1. To sum up, most of the hydrocarbons come from spices added in the stewing process of smoked    chicken. These spices, together with each other, contributed the unique aroma of smoked chicken. Alcohols mainly come from oxidative degradation of unsaturated lipids (Pham et al., 2008). Due to the higher threshold value of saturated fatty alcohol, the contribution of saturated fatty alcohol to the overall flavor is smaller, while the threshold value of unsaturated alcohol is relatively lower, and the contribution to the flavor is larger (Drumm & Spanier, 1991). 1-octene-3-ol were the most abundant alcohol in the six types of smoked chickens, and significantly contributes higher to LC Chicken (P < .05). It is formed by autooxidation of linoleic acid or other polyunsaturated fatty acids and mainly shows mushroom and earthy aroma (Lorenzo & Domínguez, 2014). The sensory threshold of 1-octene-3-ol was low and OAV was in the range of 103.-09-941.5, which gave smoked chicken meaty aroma and enhanced the fatty aroma (Kawai & Sakaguchi, 1996;Wilson & Katz, 1972). 2-ethyl-1-hexanol was detected in ZZS and LC Chicken, which contributed significantly higher to the aroma of LC Chicken (P < .05). Therefore, it can be concluded that longer smoking time can create more alcohols. Linalool mainly came from the spice such as cardamom and ginger added in the stewing process of smoked chicken, which contributes a lot to the overall aroma of GBZ, LT, JS and TQ Chicken with OAVs range from 9.67 to 238.29 and especially contributed significantly higher in GBZ Chicken sample (P < .05). The low odor threshold of linalool indicates that it contributes with a strong flowery and citrus aroma and it is present in higher amounts in sugar-smoked chicken (Larsen & Poll, 1992).
Aldehydes were the most abundant groups of compounds in all investigated types of smoked chicken and play an important role to the overall flavor due to its low odor thresholds. Aldehydes are mainly generated from two pathways: lipid oxidation and Strecker degradation of amino acids (Yang et al., 2017). It has been reported that hexanal, octanal, nonanal, (E, E) −2,4-nondialdehyde and other aldehydes are the main aroma compounds in boiled chicken (Kerler & Grosch, 1997). The most abundant aldehydes were pentanal, hexanal, heptanal, benzaldehyde, nonanal and decanal in six smoked chicken samples. Hexanal was the most abundant of the identified smoked chicken volatile compounds and is generally considered as a good indicator of the oxidation level. Previous studies have shown that hexanal could be generated from n-6 polyunsaturated fatty acids when heated at 37°C (Frankel et al., 1989). Pentanal and hexanal had great contribution to the aroma of GBZ Chicken, while nonanal, decanal and heptanal contributed the most to the flavor of LC Chicken. This result may be related to the high fat content of smoked chicken. When heated, the lipid is hydrolyzed to form free fatty acids. Then both saturated and unsaturated fatty acids are decomposed into hydroperoxides and react to form aldehydes (Pham et al., 2008). Besides linear saturated fatty aldehyde, alkenal is also the characteristic aroma of chicken fat when heated. (E)-2-Octenal contributed significantly to GBZ and JS Chicken (P < .05), (Z)-2-Heptenal contributed significantly to the aroma of GBZ Chicken (P < .05). Aldehydes mainly came from lipid oxidation degradation, which mainly endows smoked chicken with rich meaty aroma.
Ketones are formed by oxidation and degradation of fat, and may also come from Maillard reaction, mainly with fragrance, cream and fruit aroma. As shown in Table 2, Table 3. Contribution rates (%) of aroma compounds with OAVs contributing to the top ten.
Tabla 3. Tasas de contribución (%) de compuestos de aroma con OAV contribuyendo a los diez primeros compuestos.  acetophenone and 2,3-Octanedione were detected in 6 smoked chickens with OAVs greater than 1. The OAVs of 2,3-Octanedione ranged from 164.4 to 689.58 and significantly contributed the most to the aroma of JS Chicken (P < .05). Acetophenone has a certain contribution to the aroma of LC, ZZS and JS Chicken and contributes significant higher to LC Chicken (P < .05). Ketones are mainly produced in the process of chicken cooking at high temperature, and the threshold is much higher than the aldehydes of its isomers, so the contribution to the overall aroma of smoked chicken is relatively smaller. Phenols contributed to the overall aroma with smoky and toasty aroma, mainly produced by the decomposition of lignin in smoke materials (Guillén & Manzanos, 1999). In the process of smoking, lignin was first decomposed into 4-vinyl guaiacol, and then formed into the homologues of guaiacol, such as 4-methyl guaiacol, 4-ethyl guaiacol, etc. (Yu & Sun, 2005). As can be seen from Table 2, eugenol was detected in all smoked chickens, except for TQ Chicken, which may due to that the smoke materials of TQ Chicken did not contain wood chips. The OAVs were between 5.81 and 404.28, and eugenol may come from spices such as cloves added during the stewed process and contribute greatly to the flavor of smoked chicken. In addition, guaiacol, 4-methyl guaiacol, 4-ethyl guaiacol and phenol were detected only in LC sample, with the OAVs of 1220.07, 50.58, 7.31 and 0.72, respectively. Among them, guaiacol contributed the most to the overall aroma of LC Chicken. This may because that LC Chicken was smoked with fruit tree sawdust for 24 h and then lignin in smoke materials produced more phenols. 4-methyl-guaiacol has pungent and medicinal aroma. These phenols are the characteristic flavor substances of smoked products, which were also the unique volatile compound of LC Chicken sample because LC Chicken was smoked by wood sawdust.

GBZ
Furans belong to heterocyclic compounds, which can be formed by sugar cracking reaction and Maillard reaction (Fiddler et al., 1966). 2-pentylfuran has plant fragrance and contributed substantially to smoked chicken because of its low odor threshold value. It can be formed by lipid degradation and sometimes regarded as a marker of lipid oxidation (Ho et al., 1978). Among the six smoked chickens, the OAV of 2-pentylfuran in TQ Chicken is significantly higher (P < .05) and its presence could play an important role in the overall flavor. Furfural is mainly generated by polysaccharide containing hexose and pentose fragments under heating conditions (Frankel, 1983) and significantly contributed the most to JS Chicken (P < .05). 5-methyl-2-furaldehyde was mainly produced by cellulose pyrolysis and contributed to smoked chicken with sweet and caramel aroma (Schlotzhauer et al., 1982). Previous studies have shown that furfural, 5-methyl-2-furaldehyde were the main cracking products of sugars. The smoke adsorbed and permeated into the smoked chicken, giving the chicken caramel and baking aroma. Only a small content of furans was detected in LC sample, which may be due to the fact that wood fumigation can also produce furfural, 5-methyl-2-furaldehyde and other furans. This result was consistent with previous research which indicated that the main products of sugar fumigation were furfural, 5-methyl-2-furaldehyde and wood fumigation were guaiacol, 4-methyl guaiacol, 4-ethyl guaiacol and furfural etc. (Fay & Brevard, 2005).
In addition, acids, esters, ethers and other components were also detected in smoked chicken, which also have contribution to the overall aroma of smoked chicken.  Acetic acid may be produced along with the formation of furan in the process of glycolysis, and its odor threshold is 22,000 ng/g, which is too high to has little contribution to the overall aroma of smoked chicken. Acetic acid, butyl ester has a pleasant fruit aroma, which only detected in LC and ZZS Chicken (OAV > 1). Estragole and anethole were all detected in six smoked chickens. These ethers mainly came from the spices added in the stewing process of smoked chicken, which brought strong bittern aroma to the smoked chicken. Previous studies showed that 2,6-diethyl-pyrazine is the main compound produced by tea fumigation. Therefore, 2,6-diethyl-pyrazine was only detected in TQ sample which smoke material contained a small amount of tea.

Dominant aroma compounds in smoked chicken
Through the calculation and analysis of OAVs of six kinds of smoked chicken, the top ten compounds of OAV value of each kind of smoked chicken were selected. The results are shown in Table 3. To intuitively evaluate the sensory contribution of each volatile compound, contribution rates were calculated. Contribution rates of the aroma compounds were calculated by dividing the OAV of each compound by the total OAV of all compounds. The contribution rates of contributing top 10 aroma compounds of each smoked chicken are summarized in Table 3, which shows 10 aroma compounds totaling a contribution of more than 95%, including aldehydes, alcohols and furans, among which decanal (contribution rate of 19.99-37.82%), nonanal (contribution rate of 16.31-26.56%) and hexanal (contribution rate of 9.69--32.04%) were the most decisive compounds. Except the nonanal, hexanal and decanal, the 2,3-octadione, 1-octene-3-ol and 2-pentylfuran are also the common flavor substances of six kinds of smoked chicken, which are the main flavor substances of smoked chicken. The above compounds mainly came from the lipid oxidation in the process of chicken stewing, bringing strong meaty aroma to smoked chicken. In addition to the above common compounds, fennel, coriander, artemisinin (from spices) and D-limonene contributing the smoked chicken with bittern aroma; furfural (GBZ Chicken) and guaiacol (LC Chicken) give smoked chicken a unique smoky aroma. To sum up, the flavor compounds in smoked chicken mainly come from lipid oxidation, spices added in the process of brine and the substances produced by smoking.

Relationship between dominant aroma compounds and sensory attributes
The correlations between the sensory results and released volatile compounds with OAV greater than 1 were analyzed by PLSR. PLS analysis is suitable to find out the potent aroma compounds that contributing to the sensory attributes of different brands of smoked chicken. As shown in Figure 3, most of the variables are loaded around the circle (r 2 =100%, r 2 represents the degree of interpretation).
According to Dimension 1 (Figure 3), six brands of smoked chicken samples was divided into 2 groups. Among the six brands of smoked chicken, JS, GBZ and LC Chicken were located on the positive axis, while ZZS, LT and TQ Chicken were located on the negative axis. In addition, LC Chicken showed strong smoky characteristic, while JS and GBZ Chicken showed stronger caramel flavor than other chicken samples. According to Dimension 2, JS and GBZ Chicken lied in the positive side of Dimension 2, while LC Chicken lied in the negative side. This might due to that LC Chicken was smoked by fruit tree sawdust for 24 h, but JS and GBZ Chicken were smoked by sugar, therefore, they had similar aroma components and sensory attributes. The correlation matrix shows that the smoky aroma was mainly correlated with phenols, such as guaiacol, 4-ethyl-guaiacol, eugenol etc. Caramel aroma is correlated with furfural, 5-methyl-2-furancarboxaldehyde, benzaldehyde etc. Aldehydes are substances that come from the meat itself during the heating process, thus the meaty aroma attribute is closer to decanal and hexanal etc.

Conclusion
In this experiment, 89 volatile compounds were identified and quantified in six brands of smoked chicken samples using SPME-GC-MS. These compounds can be divided into ten categories, including hydrocarbons, alcohols, aldehydes, ketones, phenols, furans, acids, esters, ethers and nitrogencontaining compounds. The key odor-active volatiles of all evaluated samples were determined as nonaldehyde, hexanal, decanal, 2-pentylfuran, 1-octene-3-ol and 2,3-octanedione due to their relatively higher OAVs compared with other compounds. Moreover, through PLSR analysis, smoked chicken of six brands could be classified into four separate groups (GBZ-JS, ZZS-LT, LC, TQ). The volatile compounds and sensory differences were mainly caused by smoked materials. This study is useful for enterprises to understand consumers' preference of smoked product.