Characterization of odor-active compounds of various Chinese “Wuliangye” liquors by gas chromatography–olfactometry, gas chromatography–mass spectrometry and sensory evaluation

ABSTRACT

Odor-active compounds of Wuliangye liquors with markedly different aging years were identified by aroma extract dilution analysis (AEDA) and gas chromatography−olfactometry (GC–O). A total of 62 aroma compounds were identified in three Wuliangye liquors and 45 odorants were further screened out and quantified as the important odorants according to Flavor Dilution (FD) values. They were selected as specific compounds correlated to sensory attributes by the Pearson coefficient. The correlation results showed that ethyl hexanoate, ethyl 3-methylbutyrate, ethyl octanoate, ethyl 2-methylbutyrate, pentanoic acid, ethyl octanoate, furfural, 4-methylphenol, hexanoic acid, isovaleric acid, and 1,1-diethoxyethane were related to the characteristic aroma of Wuliangye liquors. It will be helpful for the improvement of characteristic aroma of Wuliangye liquors through adjusting fermentation parameters or compensating typical aroma compounds after alcoholic fermentation.

KEYWORDS:

• Odor-active compounds
• Wuliangye liquors
• GC–MS
• GC–O
• Sensory evaluation

Introduction

Chinese liquor, also called “Baijiu” or “Shaojiu” in Chinese, is one of the most ancient distillates in the world and occupies an irreplaceable position in China.^[1,2] Due to different raw materials and specialized brewing techniques, Chinese liquors from different manufacturers have significant differences in aroma type. According to the aroma characteristics, they are generally divided into the following categories: light-aroma-type, strong-aroma-type, soy sauce-aroma-type, sweet- and honey-aroma-type and miscellaneous-aroma-type liquors.^[3]

As one of the strong aroma-type liquor, “Wuliangye” produced in Yibin, Sichuan province of China, enjoys a long-lasting popularity in China and many other countries. It is fermented from grains (rice, sticky rice, sorghum, wheat, and corn) with Daqu powder usually made of pulverized wheat, as a starter.^[4] Freshly distilled liquor made from the fermented cereals has undesirable characteristics often described as “harsh,” “green,” and “raw.” These characteristics are often associated with young liquor and generally decrease or disappear through a long-aging process, ranging from months to years. Many chemical reactions such as oxidation, esterification, hydrolysis and rearrangement can occur during this aging process.^[5] As a result, a well-balanced, “matured” liquor of unique flavor and taste is developed during this process.^[6]

Gas chromatography–mass spectrometry (GC–MS) is a rapid method to identify and quantify volatile compounds, but it cannot identify odor-active compounds. Gas chromatography−olfactometry (GC–O) provides a valuable tool for investigating the pattern of odorants in terms of both their odor descriptors and activity. Many studies have been done to identify typical aroma compounds in liquors by GC–MS and GC–O. Wang et al.^[7] developed a method to characterize the volatile compounds in Chinese Daohuaxiang liquors by GC–MS and GC–O and 57 components were identified. Zheng et al.^[6] have investigated 62 compounds in the 36 raw liquors from three distilling stages (head, heart, and tail) of two typical Luzhou-flavor liquors (Fenggu-FG and Jiannanchun-JNC) by GC–MS. Zhang et al.^[8] detected 118 volatile compounds in chixiang-aroma-type liquor by headspace solid-phase microextraction (HS-SPME)–GC–MS. Zhu et al.^[9] developed a method to characterize the volatile compounds in Moutai liquor by comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry and 528 components were identified. So far, numerous studies were focused on the identification and quantitation of aroma composition but none comprehensively made the comparison of the aroma contributing compounds in Wuliangye liquor of different vintages. The objective of this paper was (a) to identify the components contributing to the aroma of liquors by liquid–liquid extraction (LLE)/GC–O; (b) to make a comparison between the liquors from different years and quantify the key aroma composition; and (c) to elucidate the relationship between sensory analysis and aroma compounds.

Materials and methods

Liquors

Three different vintages of Chinese “Wuliangye” liquors were under investigation: Wuliangye liquor, produced in the 2014s (W1, 52% ethanol by volume, 1 year), Wuliangye liquor, produced in the 2000s (W2, 50% ethanol by volume, 15 years), and Wuliangye liquor, produced in the 1985s (W3, 50% ethanol by volume, 30 years). All liquors (500 mL each) were purchased from Yibin Co. Ltd, and stored at 4 ºC until analysis.

Chemicals

Authentic standards were purchased from Sigma-Aldrich Chemical Co. Ltd. (Shanghai, China). The internal standards of 2-octanol, 2,2-dimethylpropanoic acid and octyl propanoate and a C7–C30 n-alkane mixture were also obtained from Sigma-Aldrich Chemical Co. Ltd. (St. Louis, MO). Absolute ethanol, dichloromethane, sodium chloride (NaCl), sodium bicarbonate (NaHCO₃), hydrochloric acid (HCl), and anhydrous sodium sulfate (Na₂SO4) were obtained from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China). All of them were analytical grade. Water purified through a Milli-Q purification system (Millipore, Bedford, MA) was used for all solution preparations and dilutions.

Aroma extraction

Liquor samples (50 mL) were diluted to 10% ethanol by volume with Milli-Q water, saturated with NaCl, spiked with 4500 μL of a mixed internal standard (2,2-dimethylpropanoic acid, octyl propanoate, and 2-octanol with a final concentration of 333.3, 333.3, and 133.3 mg/L, respectively), and then extracted thrice with 50 mL dichloromethane. The organic phase was labeled as “extract 1.” The combined extracts were further separated into acidic, water-soluble, and neutral/basic fractions according to the procedures described previously.^[10] The “extract 1” was extracted with 50 mL Milli-Q water with pH 10 by 10% NaHCO₃, then separated in a separatory funnel and saved. The aqueous phase was adjusted to pH 2 with 2 mol/L HCl, saturated with NaCl. It was extracted twice with 20 mL dichloromethane, then separated in a separatory funnel. This extract was labeled as the “acidic fraction.” Then the organic phase was separated and labeled as “extract 2.”

The ract 2 “gan” was washed twice with 10 mL Milli-Q water and the organic phase was labeled as the “neutral/basic fraction.” With alkaline, aqueous phase was further extracted twice with 20 mL dichloromethane; this extract was labeled as the “water-soluble fraction.” Afterward, each fraction was dried with anhydrous Na₂SO4 overnight and concentrated to a final volume of 200 μL under a gentle stream of nitrogen. These concentrated fractions were stored at −20°C prior to the GC–O analysis.

GC–O and GC–MS analysis

GC–O analysis was performed on a Agilent 7890A gas chromatograph equipped with a flame ionization detector (FID) and an ODP-2 Olfactory Detector Port (Gerstel, Mulheim an der Ruhr, 103 Germany). Separation was performed on an HP-INNOWAX fused-silica capillary column (60 m × 0.25 mm ID, 0.25 μm film thickness, Agilent Technologies). Hydrogen was used as the carrier gas at a constant flow rate of 2 mL/min. The column effluent was split 1:1 into FID and olfactometer. Sample (1 μL) was injected into the GC injector at splitless mode. The GC injector and detector temperatures were at 250 °C and 280 °C, respectively. The oven temperature was held at 40°m film min, ramped at 3 °C/min to 100°C, and then increased to 230°m film min, ramped at 3°C/min to 100° min. Six well-trained panelists (three females and three males, 26-year-old on average) were selected for GC–O study. The panelists were trained for 3 months in GC–O using at least 30 odor-active reference compounds in a concentration 10 times above their odor thresholds in air. During a GC run described above, a panelist placed his/her nose close to the sniffing port, and recorded the aroma descriptor and intensity value as well as retention time. The aroma descriptors were determined by an evaluation of the odor quality of reference odorants previously. Each fraction was replicated three times by each panelist.

For aroma extract dilution analysis (AEDA), a series of fourfold dilutions of each fraction were prepared using dichloromethane as a solvent.^[11,12] The FD factors were determined for the odor-active compounds in each sample. When a volatile compound was detected at least twice, this analyte was registered to be an aroma compound.

GC–MS analysis were performed on an Agilent 7890N GC with an Agilent 5975 mass selective detector (MSD) instrument operating under electron ionization (EI) mode (70 eV, ion source temperature 230°C) with the quadrupole in a scanning mode (scan range was m/z 30–450 at a scan rate of 1 scan/s). Helium (purity = 99.999%) at a constant flow rate of 1 mL/min was used as carrier gas. The column, oven, and injector temperatures were identical to those of GC–O analysis, as described above. The odorants were identified by comparing their odors, linear retention indices (LRIs), and mass spectra with those of pure standards. Several compounds were tentatively identified by comparing their retention indices from the literatures and with mass spectra in the Wiley and NIST 11 database (Agilent Technologies Inc.). The LRIs were calculated from the retention times of n-alkanes (C7−C30), according to Van den Dool and Kratz.^[13]

Quantification of aroma compounds

Quantification of the major aroma compounds was performed on the Innowax–Wax column by GC–MS. Model liquor was prepared containing standard compounds in 10% of ethanol of Milli-Q water. The standard curves of each compound were obtained at seven different concentrations which simulated the concentration ranges of various odor-active compounds in three Wuliangye liquors. Extraction of the standard volatiles for making the standard curves was the same as that of samples. The limit of detection (LOD) and limit of quantification (LOQ) were performed by establishing the minimum concentration at which the standard could be reliably detected and determined.

Sensory analysis

The sensory analysis of the three liquors was carried out in a sensory laboratory set in line with ISO 8589 (2007).^[14] The sensory analysis was conducted according to previous studies^[15–17] and the ISO 8589 standard procedure with minor modifications. Screening tests chose 24 examples of odoriferous substances (e.g., ethyl butyrate, ethyl hexanoate, 2-n-butylfuran, and so on) that could be used for training; the assessor performed the assessment of the odor by sniffing the smelling strip, waving it gently a few centimeters from his/her nose. The strip should under no circumstances touch the nose, a mustache, or the skin. Once a decision had been made, the assessor discarded the strip and replied to the questions on the answer form. The assessor then went on to examine the next substance. At last, the test supervisor interpreted the results and eliminated those assessors who had made repeated errors. A sensory panel consisted of 10 trained members passed screening tests according to the ISO standard. Ten panelists had discussed aroma attributes of samples through three preliminary sessions (each spent 2 h), until everyone agreed to use them as the attributes. Seven sensory attributes including cellar, grain, aging, floral, sweet, fruity, and caramel were defined and descripted in Table 4). Wuliangye liquor samples (20 mL) were poured into a glass cup at 20 ºC and presented in random order. The sensory panel smelled the different liquors, noted the specific perceived attributes, and rated the intensity of each sensory attribute on a 10-point scale, in which 0 indicated that the descriptor was not perceived and 9 represented very strong attribute intensity. The average scores based on the scores given by 10 panelists (each repeated three times) were provided as evaluation results.

Table 1. Aroma compounds identified in W1, W2, and W3 by GC–O.

Code	Compound identified	Odor descriptors	RIcal^a	Fraction^b	FD factors
					W1	W2	W3
1	Ethyl acetate	Pineapple	909	N + A	64	16	16
2	1,1-Diethoxyethane	Fruit, cream	888	N	16	256	256
3	3-Methylbutanal	Malt	909	N	1	16	16
4	Ethyl isobutyrate	Sweet	963	N	256	16	16
5	2-Pentanone	Fruit	981	N	1
6	2-Butanol	Wine	1003	N + A + W	1	1	4
7	Ethyl butyrate	Apple	1006	N	1024	1024	4096
8	Ethyl 2-methylbutyrate	Apple	1018	N		64	64
9	Ethyl 3-methylbutyrate	Fruit	1030	N		256	1024
10	unknown	Fruit	1042	N + W	16	16	256
11	Isobutyl alcohol	Wine, solvent, bitter	1057	N + A + W	1	4	4
12	2-Pentanol	Green	1071	N + A + W	16	16	16
13	Ethyl pentanoate	Sweet, fruit	1076	N	1024	1024	1024
14	1-Butanol	Medicine, Alcoholic	1085	N + A + W	64	64	64
15	3-Methylbutanol	Malty, burnt	1141	N + A + W	256	256	256
16	1-Pentanol	Balsamic	1192	A + W	1	4	64
17	Ethyl hexanoate	Fruit	1202	N	1024	1024	4096
18	2-n-Butylfuran	Roasted, baked	1234	N		64
19	Hexyl acetate	Fruit, herb	1239	N	4
20	2-Methylbutyl 2-methylbutyrate	Fruit	1260	N	16
21	2-Heptanol	Mushroom	1270	N + W	1	4
22	1,1,3-Triethoxypropane	Fruit, vegetal	1271	N	16
23	Ethyl heptanoate	Fruit	1289	N	4	64	64
24	Propyl hexanoate	Fruit	1285	N	64
25	2-Methylpropyl hexanoate	Fresh	1289	N	16
26	2,6-Dimethylpyrazine	Roasted nut	1292	N			1
27	Ethyl lactate	Fruit	1295	N + W	1024	256	256
28	1-Hexanol	Resin, flower, green	1315	N + W	16	16	64
29	2,3,5-Trimethylpyrazine	Roast, potato	1365	N			64
30	Butyl hexanoate	Fruit	1370	N	256
31	unknown	Flower	1383	N	16
32	Ethyl octanoate	Fruit	1406	N	1024	1024	1024
33	Acetic acid	Sour, vinegar	1404	A	16	256	4
34	Isoamyl hexanoate	Fruit, flower	1417	N	1
35	Furfural	Almond, sweet	1426	N + A		256	16
36	2-Ethyl-1-hexanol	Fose, green	1443	N	16
37	Benzaldehyde	Almond, burnt Sugar	1495	N + A	16	4	4
38	Propionic acid	Pungent, Rancid, soy	1494	A	16	64	4
39	Isobutyric acid	Rancid, butter, cheese	1522	A	16		1
40	5-Methyl-2-furancarboxaldehyde	Almond, caramel	1546	N + A	4	16	1
41	Hexyl hexanoate	Apple peel, peach	1568	N	256	16	16
42	Butyric acid	Rancid, cheese, sweat	1589	N + A + W	1024	1024	1024
43	2-Acetyl-5-methylfuran	Roasted	1582	N	4
44	Ethyl decanoate	Grape, fruit	1593	N	16	16	16
45	2-Phenylacetaldehyde	Floral, rose	1614	N	16	1	1
46	Furfuryl alcohol	Burnt	1622	N + A + W	16	64	16
47	Isovaleric acid	Sweat, acid, Rancid	1632	N + A + W	16	256	64
48	Diethyl succinate	Wine, fruit	1643	N		4
49	Pentanoic acid	Sweat	1705	N + A + W	16	64	64
50	Ethyl phenylacetate	Fruit, sweet	1766	N	64	16	16
51	Isocaproic acid	Sweat, sour	1772	A	16	16
52	Hexanoic acid	Sweat	1820	A	1024	256	4
53	Furfuryl hexanoate	Caramel, fruity	1848	N	1
54	2-Phenylethanol	Honey, spice, rose, lilac	1914	N + A + W	256	256	256
55	Heptanoic acid	Unpleasant	1941	A	16	64	16
56	Phenol	Phenol	1997	N + A	64	64	64
57	Octanoic acid	Sweat, cheese	2073	A	4	4	4
58	4-Methylphenol	phenol, Smoke	2079	N + W		4	16
59	unknown	Smoke	2211	A		16
60	Ethyl hexadecanoate	Wax	2266	N + A		1	1
61	Decanoic acid	Rancid, fat	2263	A		4
62	Ethyl 2-hydroxy-3-phenylpropanoate^c	Smoke	2295	N	4

^aThe linear retention index (LRI) calculated of unknown compounds on a HP-INNOWAX capillary column.

^bThe odorants were detected in the following fractions: A – acidic fraction, N – neutral/basic fraction or W – water-soluble fraction. Each fraction was analyzed by GC–O.

^cTentatively identified. This compound was identified by comparison with the retention index from the literatures and mass spectrum comparison using Wiley or Nist library.

Table 2. Result of recovery (%) and RSD (%) (n = 6) in W1, W2, and W3 for the 13 volatile compounds.

Code	Compound	W1		W2		W3
		Recovery	RSD	Recovery	RSD	Recovery	RSD
1	Ethyl acetate	73.2	8.2	79.7	9.2	83.0	8.5
2	Ethyl butyrate	89.6	8.3	92.1	9.7	92.9	6.8
3	1-Butanol	99.3	10.6	90.7	9.8	98.8	7.3
4	Ethyl hexanoate	91.2	10.2	102.6	11.3	100.4	7.1
5	2-n-Butylfuran	–	–	101.0	9.0	–	–
6	Ethyl heptanoate	98.9	8.8	102.2	10.2	95.6	5.7
7	Ethyl lactate	104.0	9.5	99.7	10.3	99.5	7.9
8	2,3,5-Trimethylpyrazine	–	–	–	–	100.7	5.2
9	Butyric acid	95.5	9.4	90.5	9.0	93.9	8.9
10	Hexyl hexanoate	87.5	6.8	95.8	9.1	89.1	8.2
11	Ethyl phenylacetate	87.0	5.2	88.6	7.9	93.7	8.2
12	Hexanoic acid	79.9	8.6	96.9	7.4	87.3	6.4
13	4-Methylphenol	–	–	75.4	9.0	80.3	9.5

In some cases, recovery was not calculated because concentration data of some compounds were not available in the liquors.

Table 3. Concentrations of the volatile components.

Code	Components	Concentration (g/L)
		W1	W2	W3
1	Ethyl acetate	0.42 ± 0.01^a	0.01 ± 0.00	0.09 ± 0.02
2	1,1-Diethoxyethane	0.82 ± 0.04	4.49 ± 0.23	1.99 ± 1.59
3	3-Methylbutanal	0.30 ± 0.01	3.99 ± 0.12	2.24 ± 0.45
4	Ethyl isobutyrate	0.80 ± 0.03	0.27 ± 0.08	0.35 ± 0.12
5	Ethyl butyrate	3.55 ± 0.47	3.61 ± 0.60	5.25 ± 3.23
6	Ethyl 2-methylbutyrate	n.d.	0.55 ± 0.01	0.69 ± 0.14
7	Ethyl 3-methylbutyrate	n.d.	1.08 ± 0.08	2.34 ± 0.94
8	2-Pentanol	0.10 ± 0.01	0.05 ± 0.00	0.07 ± 0.04
9	Ethyl pentanoate	1.77 ± 0.47	7.69 ± 0.35	5.81 ± 3.16
10	1-Butanol	4.37 ± 0.68	2.15 ± 0.08	2.19 ± 2.32
11	3-Methylbutanol	2.68 ± 0.87	1.92 ± 1.39	1.57 ± 4.56
12	1-Pentanol	0.01 + 0.00	0.02 + 0.00	0.12 ± 0.01
13	Ethyl hexanoate	10.41 ± 8.08	26.07 ± 13.00	34.36 ± 38.71
14	2-n-Butylfuran	n.d.	0.01 ± 0.00	n.d.
15	2-Methylbutyl 2-methylbutyrate	0.01 ± 0.00	n.d.	n.d.
16	1,1,3-Triethoxypropane	0.01 ± 0.00	n.d.	n.d.
17	Ethyl heptanoate	0.24 ± 1.09	2.06 ± 1.68	10.98 ± 2.07
18	Propyl hexanoate	0.02 ± 0.00	n.d.	n.d.
19	2-Methylpropyl hexanoate	0.01 ± 0.00	n.d.	n.d.
20	Ethyl lactate	7.76 ± 0.17	3.57 ± 0.04	3.42 ± 1.37
21	1-Hexanol	1.31 ± 0.01	1.43 ± 0.01	3.64 ± 0.36
22	2,3,5-Trimethylpyrazine	n.d.	n.d.	0.02 ± 0.00
23	Butyl hexanoate	0.02 ± 0.00	n.d.	n.d.
24	Ethyl octanoate	6.62 ± 0.67	9.85 ± 0.20	7.43 ± 0.19
25	Acetic acid	0.58 ± 0.02	3.87 ± 0.19	0.33 ± 0.32
26	Furfural	n.d.	5.65 ± 0.17	1.98 ± 0.90
27	2-Ethyl-1-hexanol	0.01 ± 0.00	n.d.	n.d.
28	Benzaldehyde	0.18 ± 0.00	0.06 ± 0.00	0.08 ± 0.02
29	Propionic acid	0.34 ± 0.00	1.18 ± 0.02	0.12 ± 0.02
30	Isobutyric acid	0.02 ± 0.00	n.d.	0.01 ± 0.00
31	5-Methyl-2-furancarboxaldehyde	0.60 ± 0.01	1.29 ± 0.05	0.33 ± 0.03
32	Hexyl hexanoate	3.87 ± 0.16	0.50 ± 0.03	0.25 ± 0.23
33	Butyric acid	3.15 ± 0.36	1.33 ± 0.11	1.07 ± 0.54
34	Ethyl decanoate	0.70 ± 0.01	0.66 ± 0.01	0.11 ± 0.02
35	2-Phenylacetaldehyde	0.44 ± 0.01	0.10 ± 0.00	0.01 ± 0.07
36	Furfuryl alcohol	1.77 ± 0.02	0.66 ± 0.05	0.45 ± 0.13
37	Isovaleric acid	0.23 ± 0.00	6.95 ± 0.23	2.11 ± 0.04
38	Pentanoic acid	0.11 ± 0.00	0.24 ± 0.00	0.32 ± 0.07
39	Ethyl phenylacetate	2.89 ± 0.12	0.54 ± 0.03	0.88 ± 0.08
40	Isocaproic acid	0.01 ± 0.00	0.01 ± 0.00	n.d.
41	Hexanoic acid	8.42 ± 0.01	6.20 ± 0.12	2.29 ± 0.86
42	2-Phenylethanol	2.82 ± 0.06	1.74 ± 0.07	1.94 ± 0.28
43	Heptanoic acid	3.05 ± 0.52	7.57 ± 0.38	2.74 ± 1.55
44	Phenol	0.18 ± 0.01	0.07 ± 0.00	0.18 ± 0.02
45	4-Methylphenol	n.d.	0.01 ± 0.00	0.01 ± 0.00

^aMean standard deviation (average of triplicate). n.d.: not detected.

Table 4. The mean scores of the seven attributes for the three Chinese liquors in sensory analysis and description of sensory attributes.

Sensory attribute	Mean score			Description
	W1	W2	W3
Aging	2.67 ± 0.82^c	4.31 ± 0.36^b	7.33 ± 0.86^a	The mature aroma characteristics after long-term storage of liquor by solid natural formation
Fruity	7.00 ± 0.49^a	5.76 ± 0.83^b	5.13 ± 0.49^b	Exhilarating scent and smelling strongly of varieties of ripe fruits
Cellar	3.63 ± 0.33^b	7.61 ± 0.22^a	8.81 ± 0.9^c	The aroma characteristics of ethyl hexanoate
Floral	6.75 ± 0.78^a	4.62 ± 0.61^b	3.93 ± 0.47^c	The perfume of fresh flowers
Sweet	3.25 ± 0.43^c	4.80 ± 0.66^b	6.91 ± 0.67^a	The aroma of honeys and fruities
Caramel	2.30 ± 0.51^c	4.85 ± 0.5^b	6.21 ± 0.58^a	The aroma of baking cereals and grains and caramel
Grain	4.15 ± 0.85^c	8.60 ± 1.1^a	5.24 ± 0.46^b	The steamed food aroma from the fermentation and distillation of Sorghum, rice, wheat and other grains

^a,b,cDifferent letters of the same row on the upright side mean significantly differences (p < 0.05) using Duncan’s multiple comparison test (n = 30: 10 panelists with 3 replications).

Table 5. Correlation observed between the volatile compounds and sensory attributes.

Code	Components	Aging	Fruity	Cellar	Floral	Sweet	Caramel	Grain
1	Ethyl acetate	−0.649	0.873	−0.924	0.917	−0.708	−0.868	−0.803
2	1,1-Diethoxyethane	0.146	−0.481	0.579	−0.566	0.226	0.472	0.997
3	3-Methylbutanal	0.376	−0.675	0.756	−0.745	0.450	0.667	0.950
4	Ethyl isobutyrate	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
5	Ethyl butyrate	0.928	−0.740	0.658	−0.670	0.895	0.747	−0.309
6	Ethyl 2-methylbutyrate	0.844	−0.978	0.995	−0.994	0.884	0.976	0.593
7	Ethyl 3-methylbutyrate	0.992	−0.974	0.941	−0.946	0.999	0.976	0.192
8	2-Pentanol	−0.415	0.705	−0.783	0.772	−0.487	−0.698	−0.936
9	Ethyl pentanoate	0.532	−0.794	0.859	−0.850	0.599	0.787	0.881
10	1-Butanol	−0.973	0.993	−0.972	0.976	−0.988	−0.994	−0.298
11	3-Methylbutanol	0.577	−0.825	0.885	−0.877	0.641	0.819	0.854
12	1-Pentanol	0.954	−0.790	0.714	−0.725	0.926	0.796	−0.235
13	Ethyl hexanoate	0.998	−0.955	0.915	−0.921	1.000	0.958	0.123
14	2-n-Butylfuran	−0.169	−0.185	0.297	−0.282	−0.088	0.175	0.972
15	2-Methylbutyl 2-methylbutyrate	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
16	1,1,3-Triethoxypropane	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
17	Ethyl heptanoate	−0.656	0.352	−0.241	0.257	−0.592	−0.362	0.709
18	Propyl hexanoate	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
19	2-Methylpropyl hexanoate	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
20	Ethyl lactate	−0.788	0.953	−0.982	0.978	−0.836	−0.950	−0.667
21	1-Hexanol	0.953	−0.789	0.712	−0.724	0.926	0.795	−0.237
22	2,3,5-Trimethylpyrazine	0.938	−0.759	0.678	−0.690	0.907	0.765	−0.283
23	Butyl hexanoate	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
24	Ethyl octanoate	−0.906	0.997	−0.999	1.000	−0.938	−0.996	−0.483
25	Acetic acid	−0.230	−0.123	0.237	−0.221	−0.150	0.113	0.955
26	Furfural	0.183	−0.513	0.609	−0.596	0.262	0.505	0.993
27	2-Ethyl-1-hexanol	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
28	Benzaldehyde	−0.680	0.893	−0.939	0.933	−0.737	−0.888	−0.778
29	Propionic acid	−0.362	0.015	0.101	−0.084	−0.285	−0.025	0.905
30	Isobutyric acid	−0.554	0.809	−0.872	0.864	−0.620	−0.803	−0.868
31	5-Methyl-2-furancarboxaldehyde	−0.434	0.093	0.023	−0.006	−0.359	−0.103	0.870
32	Hexyl hexanoate	−0.808	0.962	−0.987	0.985	−0.853	−0.960	−0.643
33	Butyric acid	−0.836	0.975	−0.994	0.992	−0.878	−0.973	−0.604
34	Ethyl decanoate	−0.958	0.797	−0.722	0.733	−0.931	−0.804	0.223
35	2-Phenylacetaldehyde	−0.769	0.944	−0.976	0.972	−0.819	−0.940	−0.689
36	Furfuryl alcohol	−0.857	0.983	−0.998	0.996	−0.896	−0.981	−0.572
37	Isovaleric acid	0.106	−0.445	0.546	−0.532	0.186	0.436	0.999
38	Pentanoic acid	0.948	−0.999	0.989	−0.991	0.971	1.000	0.381
39	Ethyl phenylacetate	−0.678	0.891	−0.938	0.932	−0.735	−0.887	−0.780
40	Isocaproic acid	−0.938	0.759	−0.678	0.690	−0.907	−0.765	0.283
41	Hexanoic acid	0.957	−0.998	0.984	−0.987	0.978	0.999	0.353
42	2-Phenylethanol	−0.651	0.875	−0.925	0.918	−0.711	−0.870	−0.801
43	Heptanoic acid	0.120	0.233	−0.344	0.329	0.039	−0.223	−0.982
44	Phenol	0.207	0.146	−0.260	0.244	0.127	−0.136	−0.962
45	4-Methylphenol	0.947	−1.000	0.989	−0.992	0.970	1.000	0.383

The larger the positive number, the stronger the positive correlation but negative correlation was the opposite.

Statistical analysis

The sensory analysis was evaluated by analysis of variance (ANOVA) using SAS V8 (SAS Institute Inc., Cary, NC, USA). ANOVA with Duncan’s multiple comparison tests were performed to determine whether there were differences among individual sample for sensory attributes. The differences were considered to be significant at p < 0.05. The correlation analysis was employed to show the correlations between odor-active compounds and sensory attributes. The data were operated using XLSTAT version 7.5 (Addinsoft, New York, NY, USA).

Result and discussion

Identification of odor-active constituents

AEDA is a GC–O method commonly used for the ranking and identification of characteristic aroma compounds^[18–20] In this study, the extract was serially diluted in a 1:4 ratio. FD factors and odor descriptors are listed in Table 1. Following this procedure, a total of 62 aroma compounds were detected in the Wuliangye liquors of different ages. These volatiles included 23 esters, 11 organic acid, 10 alcohols, 5 aldehydes, 5 furans, 2 phenols, 2 pyrazines, and 1 ketone. In comparison, the liquors of 1, 15, and 30 years had 51, 46, and 39 volatile components, respectively.

On the basis of the FD values detected on a Innowax–Wax column, potentially most important alcohols were 2-pentanol, 1-butanol, 2-phenylethanol, 1-pentanol, 3-methylbutanol, 2-ethyl-1-hexanol, furfuryl alcohol, and 1-hexanol (FD ≧ 16). Among them, 2-ethyl-1-hexanol was only detected in 1-year Wuliangye liquor and long-chained alcohols were probably formed in the fermentation of liquor.^[21] In addition, 3-methylbutanol having high FD value in three liquors (FD ≧ 256) was detected, which gave intense malty and burnt aromas. Esters seemed to be the most important aroma compounds in Wuliangye liquors. Among them, the highest FD factor of 4096 was determined in 30-year Wuliangye liquor for ethyl butyrate and ethyl hexanoate, which were previously regarded as the characteristic aromas of Wuliangye liquor.^[1] Ethyl 3-methylbutyrate, ethyl pentanoate and ethyl octanoate had the highest FD values (FD ≧ 1024) in both young and aged liquors. Several aldehydes were detected in three liquors, including 3-methylbutanal, benzaldehyde, and 2-phenylacetaldehyde which elicited intense malt, almond, and floral aromas, respectively. Furthermore, Noguerol-Pato et al.^[22] found that the progressively increased concentration of benzaldehyde might indicate a deterioration of the raw materials. Aldehydes can be converted into other compounds during the liquor-aging process.^[23,24] In addition, 1,1-diethoxyethane had the relative high FD value (FD ≧ 256) in aged liquors which have also been found as important aroma components in freshly distilled Calvados^[25] and white wine ^[26] They were formed from reaction of alcohols and aldehydes in the presence of excess alcohols. Phenols could be potentially important aroma compounds to Wuliangye liquors. 4-Methylphenol was not detected in 1-year liquor and had the highest FD value (FD ≧ 16) in 30-year liquor. Phenolic compounds were probably derived from lignin degradation of raw materials.^[27] Butyric acid (FD ≧ 1024) was detected in three liquors whose aroma contribution was also be confirmed in Godello white wines.^[28] Most of the acids in the liquors were produced by microbial fermentation.^[29,30] Two pyrazine compounds, 2,6-dimethylpyrazine (FD ≧ 1) and 2,3,5- trimethylpyrazine (FD ≧ 64), were only detected in 30-year Wuliangye liquor. These pyrazines can be formed through both nonenzymatic and enzymatic pathways.^[31,32]

Quantitation of odor-active compounds

To gain a deeper insight into the aroma of three Wuliangye liquors, a total of 45 odorants with FD value of ≥16 among three liquors were quantified. The concentrations of the volatiles were calculated on the basis of the standard curves obtained from their total ion chromatograms (TICs). Furthermore, a recovery study was performed to investigate the recovery of 13 compounds. The recoveries ranged from 73.2% to 103.9% with relative standard deviations (RSDs) below 12% for all compounds are revealed in Table 2. The result demonstrated that this method was reliable to quantify aroma compounds.

Among these odorants, ethyl hexanoate, ethyl butyrate, ethyl pentanoate, ethyl octanoate, and ethyl lactate, which had been reported as the most abundant odorants in sauce-aroma-type liquor,^[33] appeared with the highest concentrations as expected. These esters were mainly the markers of fermentative aroma^[22] and contributed to the pleasant fruity aroma of liquors. Ethyl hexanoate showed the highest concentrations in three Wuliangye liquors and all of them were >10.0 g/L. Esters were mostly formed through esterification of alcohols with fatty acids during the fermentation and aging processes. Fan et al.^[34] reported that Daqu had both high hydrolase and esterase activities. The esterases could be very active during the fermentation process and catalyze ester synthesis.

Alcohols were the second largest group. 3-Methylbutanol, 2-phenylethanol, and 1-butanol were the most abundant alcohols with the high concentration in all liquors. 3-Methylbutanol reached the highest concentration (2.68 g/L) in 1-year Wuliangye liquor. Noguerol-Pato et al.^[35] found that 3-methylbutanol was quantitatively (249 mg/L) the most abundant higher alcohol in Mencía wines, reminiscent of “alcohol, fusel” odor. 2-Phenylethanol also reached the highest concentration (2.82 g/L) in 1-year Wuliangye liquor. Fatty acid can contribute to a balanced aroma in liquors by hindering hydrolysis of their esters.^[36] Hexanoic acid which may be negative and undesirable aroma exhibited the highest concentration of 8.42 g/L in 1-year Wuliangye liquor, above 20 mg/L^[37,38] . Fifteen-year Wuliangye liquor had the highest concentration of 5-methyl-2-furancarboxaldehyde (1.29 g/L), 1,1-diethoxyethane (4.49 g/L), and furfural (5.65 g/L). One-year Wuliangye liquor had the highest concentration of benzaldehyde (0.18 g/L) and 2-phenylacetaldehyde (0.44 g/L). The group of volatile phenols only included phenol and 4-methylphenol. Thirty-year Wuliangye liquor possessed the highest amounts of them, 0.01 g/L and 0.18 g/L, respectively. 2,3,5-Trimethylpyrazine was only quantified (0.02 g/L) as the pyrazine compound in the 30-year Wuliangye liquor, which could be related to the Maillard reaction between saccharide and amino residues or the ambient temperature reaction of microbial metabolites in the solid-state fermentation in Chinese liquor.^[39]

Sensory evaluation

The sensory characteristics of three Wuliangye liquors with different vintages were evaluated by 10 trained panelists. As shown in Fig. 1, liquors aroma was described as cellar, grain, aging, floral, sweet, fruity, and caramel attributes. ANOVA analysis demonstrated that samples showed a significant difference in all attributes. The most discriminative terms were cellar, grain, aging, floral, sweet, and caramel (p < 0.001), followed by fruity (p < 0.05). Duncan’s multiple comparison test results (Table 4) revealed that the six attributes (cellar, grain, aging, floral, sweet, and caramel) with different superscripts for each sample seemed to well explain their aroma characteristics. However, fruity attribute had a little difference.

Figure 1. Graph of the mean sensory scores of three Wuliangye liquors studied. * and *** indicate the significance levels at p < 0.05 and p < 0.001, respectively.

As shown in Fig. 1 and Table 4, W1 (1 year) presented the highest fruity and floral aroma, but it showed the lowest score for another five attributes. W2 (15 year) had almost the same score in fruity, cellar, and floral, a higher score in grain aroma than others. W3 (30 year) had the least fruity and floral aroma, but it had the strongest aging, cellar, sweet, and caramel aroma.

Correlation analysis

Correlation analysis could be manipulated to further illustrate the relationships between the odor-active compounds and sensory attributes. Positive correlations were found between “aging”,“sweet” attribute and ethyl hexanoate (r = 0.998), ethyl 3-methylbutyrate (r = 0.992); “fruity” attribute and ethyl octanoate (r = 0.997). Ethyl octanoate and ethyl hexanoate had been reported as the characteristic aromas of liquors and showed an intense fruity odor. “cellar”attribute and ethyl 2-methylbutyrate (r = 0.995); “floral” attribute and ethyl octanoate (r = 0.999). “caramel” attribute and 4-methylphenol (r = 0.999), hexanoic acid (r = 0.999);“grain” attribute and isovaleric acid (r = 0.999), 1,1-diethoxyethane (r = 0.997), and furfural (r = 0.993) (Table 5). It was reported that the aroma of 1,1-diethoxyethane was not ignored and furfural, 4-methylphenol and hexanoic acid contributed to the almond, phenol and sweet aromas of liquors, respectively.^[4] They could also contribute to the overall aroma. On the other hand, negative correlations were found. Liang et al.^[40] indicated that positive as well as negative correlations suggest that the perception of an aromatic note was influenced not only by the presence of a few compounds whose aroma forms the attribute but also by the presence of other odorants that affect negatively in the perception of such aromatic attribute.

Conclusion

In summary, the potentially important aroma compounds were analyzed by GC–O and GC–MS. A total of 62 aroma-active compounds were identified in three Chinese Wuliangye liquors. The important aroma compounds were ethyl butyrate, ethyl 3-methylbutyrate, ethyl pentanoate, ethyl hexanoate, ethyl lactate, ethyl octanoate, furfural, hexyl hexanoate, hexanoic acid, 2-phenylethanol and 1,1-diethoxyethane that had high FD values. Among them, 45 odor-active compounds with FD values ≧ 16 were further screened out and quantified as the important odorants. Sensory evaluation showed the young liquor had high fruity and floral aroma, whereas the aging and cellar aroma in the old liquors were significant. The correlations between sensory analysis and odor-active compounds were elucidated. A better understanding of this knowledge will be helpful for the improvement of characteristic aroma of Wuliangye liquors through adjusting fermentation parameters or compensating typical aroma compounds after alcoholic fermentation.

Funding

The research was supported by National Key Research and Development Program Nanotechnology Specific Project [grant number 2016YFA0200304], National Natural Science Foundation of China [grant number 2147614090], and Shanghai Engineering Technology Research Center of Fragrance and Flavor [grant number 15DZ2280100].

Additional information

Funding

The research was supported by National Key Research and Development Program Nanotechnology Specific Project [grant number 2016YFA0200304], National Natural Science Foundation of China [grant number 2147614090], and Shanghai Engineering Technology Research Center of Fragrance and Flavor [grant number 15DZ2280100].