Effects of heat treatment and β-cyclodextrin addition on soymilk flavor

Shi Xiaodi, Lv Yanchun,2, Guo Shuntang

?

Effects of heat treatment and-cyclodextrin addition on soymilk flavor

Shi Xiaodi1, Lv Yanchun1,2, Guo Shuntang1※

(1.100083,; 2.563000,)

Flavor is an essential indicator to evaluate the quality of soymilk products, but the effects of heat treatment and-cyclodextrin addition during soymilk processing on the entire flavor quality are not clear.In this study, the changes in the contents of critical soymilk flavor compounds during heat treatment (30-90 ℃) were investigated, and then the flavor profiles of soymilk with the addition of-cyclodextrin (0.25%, 0.50%, 0.75%, 1.00%) at different heating periods (before heating, at 40℃, at 60 ℃, after heating) were analyzed. Results showed that, different flavor compounds varied in the sensitivity to temperature, and as the temperature increased during heat treatment, the intensity of beany flavor tended to decrease, whereas non-beany flavor was relatively enhanced.-cyclodextrin addition could decrease the contents of critical soymilk flavor compounds, and a high concentration of-cyclodextrin (≥0.50%) added at 60 ℃ led to the most significant decrease in the contents of critical beany flavor compounds including hexanal, hexanol, 1-octen-3-ol (＜0.05), but trans-2-octenal, as the critical non-beany flavor compound, also suffered the greatest loss. Similar tendency was obtained according to the scores of beany odor, mushroom flavor, and sweet aroma by sensory evaluation, and the results of comprehensive scoring indicated that, the soymilk with the addition of 0.75%-cyclodextrin at 60℃ during heat treatment had the best flavor quality. Since-cyclodextrin is low-priced and safe, and the addition during soymilk processing is also easily practiced, it can be well applied in soymilk production for the improvement of flavor quality.

heat treatment; temperature; flavors; non-beany flavor;-cyclodextrin; sensory evaluation

0 Introduction

Flavor is a critical factor to determine the acceptable level of soymilk among customers. Western customers are very sensitive to the off-flavor in soymilk, whereas oriental consumers have a higher preference for flavor-rich soymilk products[1]; accordingly, researchers have paid increasing attention to soymilk flavor characteristics in recent years.

Soymilk flavor is composed of many volatile flavor compounds. At present, more than 70 kinds of volatile flavor compounds have been identified in soymilk[2]. According to the study of Lv et al.[3], 12 critical compounds significantly affecting soymilk flavor are confirmed, and among these compounds, hexanal, trans-2-hexenal, 1-octen-3-ol, hexanol, pentanol, acetic acid, benzaldehyde and trans, trans-2, 4-decadienal, respectively representing herbal flavor, leaf flavor, mushroom flavor, raw flavor, alcohol flavor, sour flavor, bitter almond flavor, and oil/fatty flavor, are defined as the beany flavor unacceptable to consumers; by contrast, 3-methyl-butyraldehyde, trans-2-octenal, nonanal and trans-2-nonenal, respectively representing chocolate flavor, green cucumber flavor, broccoli flavor, and fruit flavor, are defined as the non-beany flavor favored by consumers. Similar results are obtained by Poliseli-Scopel et al.[4], even though different soymilk preparation methods are adopted.

Soymilk flavor is profoundly affected by heat treatment. Yuan et al.[5]report that, if soybean is bla-nched in hot water, the content of hexanal in soymilk is significantly decreased. When soaked soybean is ground at 3, 15, 30, 55 and 80℃, hexanal has the lowest content at 80℃[6]. Furthermore, steam heating, compared with direct- fire heating, leads to lower contents of hexanal, hexanol, 1-octen-3-ol, and trans- 2-nonenal in soymilk[7].

Cyclodextrin is a common food ingredient to mask flavor. It has three types, namely,-,-,-, and-cycl-odextrin is often used in food processing. Cyclodextrin can entrap flavor compounds to form “host-guest” complex, which is considered a dynamic equilibrium with water mol-ecules as the solvent, resulting in the beneficial modification of food flavor profile by controlling the release of unpl-easant odors[8].

Beany flavor compounds are unacceptable, but some non-beany flavor compounds in soymilk are favored by consumers[3]. Many reseachers have mainly focused on how to eliminate beany flavor with specific processing methods, but the effects of these methods on non-beany flavor are rarely reported. In fact, provided that beany flavor is reduced or removed by effective measures, non-beany flavor is reduced as well, leading to a bland and insipid flavor profile of soymilk.

Therefore, in this study, the changes in the contents of critical soymilk flavor compounds during heat treatment were investigated and the flavor quality of soymilk added with-cyclodextrin was analyzed in order to explore applicable methods to improve the flavor quality of soymilk products.

1 Materials and methods

1.1 Materials

The soybean cultivar Zhonghuang 13 was purchased from Chinese Academy of Agricultural Sciences. Hexanal, hexanol, trans-2-hexenal, 1-octen-3-ol, trans-2-octenal, non-anal, trans-2-nonenal and trans, trans-2, 4-decadienal were of chromatographic grade and obtained from Sigma-Aldrich Co., Ltd. (Steinheim, Germany).-cyclodextrin was of food grade and obtained from Ingredion Co., Ltd. (Shanghai, China).

1.2 Preparation of soymilk and addition of-cyclodextrin

Preparation of soymilk was conducted according to the procedures described by Lv et al.[3]. The soybeans (100 g) were initially rinsed and soaked in 300 mL of distilled water for 12 h at 4℃. The soaked soybeans were then drained and ground with 700 mL of distilled water at 25 ℃ using a blender (JYL-350A, Shandong Joyoung Household Elec-trical Appliances Co., Ltd. China) for 3 min at the speed of 18 000 r/min. The resulting slurry was ?ltered through a defatted cotton sheet, and the filtrate was designated as raw soymilk.

Raw soymilk (20 mL) was respectively added into 7 glass bottles with caps, and then they were heated in a water bath with continuous stirring. Once the temperature reached 30, 40, 50, 60, 70, 80 and 90 ℃, they were successively taken out and put into an ice bath.

Soymilk samples with the addition of-cyclodextrin (0.25%, 0.50%, 0.75%, 1.00%) at different heating periods (before heating, at 40℃, at 60 ℃, after heating) were prepared as follows. 1) Control: raw soymilk (50 mL) was heated in a water bath and maintained above 95℃ for 5 min, and then immediately cooled in an ice bath. 2) Before heating: raw soymilk (50 mL) was added with-cyclodextrin and mixed well, and then heated and cooled in the same procedures of the control. 3) At 40 ℃ (or at 60 ℃): raw soymilk (50 mL) was heated in a water bath, and when the temperature reached 40 ℃ (or at 60 ℃), it was added with-cyclodextrin and mixed well, and then kept on heating above 95℃ for 5 min and cooled in an ice bath. 4) After heating: raw soymilk (50 mL) was heated and cooled in the same procedures of the control, and then added with-cyclodextrin and mixed well.

1.3 Headspace solid-phase microextraction (HS-SPME) and gas chromatography (GC)

HS-SPME and GC were conducted according to the pro-cedures described by Shi et al.[9]. The SPME fiber (50/ 30m divinyl benzene/carboxen/polydimethylsiloxane; Stable Flex, Supelco Co., Bellefonte, Pa., U.S.A.) was utilized for flavor extraction. Before each use, the fiber was conditioned at 270℃ for 1 h. The soymilk sample (5 mL) was placed in a 15 mL clear glass vial (Supelco Co.) with a small magnetic stirring bar and sealed with a lid containing Teflon- coated rubber septum. Internal standard 2-methyl-3-hept-anone (1 mg/L of soymilk) and 1.5 mg of NaCl were added. The sample in the vial was heated at 50℃ for 30 min in a water bath and extracted for 30 min using the fiber. Sub-sequently, the fiber was injected into GC column (Shimadzu 2014A, Kyoto, Japan) in splitless mode for analysis. A DB-WAX capillary column (30 m × 0.25 mm i.d., 0.25m film thi-c-kness; J&W Scientific) was used. The oven temp-erature was initially set to 50℃ for 5 min; programmed at 3℃/min to 160℃, which was held for 3 min; and sub-s-e-qu-ently progr-ammed at 10℃/min to 230℃, which was held for 10 min.

1.4 Determination of the standard curves

The standard curves of hexanal, trans-2-hexenal, 1-octen-3-ol, trans-2-octenal, hexanol, nonanal, trans-2- nonenal and trans, trans-2, 4-decadienal were established using 2% fat cow’s milk as the matrix[1]. Different con-centrations of the standard compounds with added internal standard 2-methyl-3-heptanone (1 mg/L of the solution) were prepared. The conditions of extraction and analysis were similar to those of HS-SPME and GC. The results were showed in Table 1.

Table 1 Standard curves of flavor compounds

Note:is the peak area of one flavor compound;is the concentration of one flavor compound and the unit is mg×L-1.

1.5 Sensory evaluation

Sensory evaluation was conducted according to the pro-cedures described by Murray and Delahunty[10]and Shi et al.[9]. A total of 10 trained panelists (6 female, 4 male, 22 to 33 years of age) from China Agricultural University were invited for the study. Each panelist had more than 100 h of previous experience in the sensory evaluation of food products, received an additional training on soymilk, and also agreed to the definition and specific attributes of soy-milk flavor. The descriptions of soymilk flavor were listed in Table 2. Paper ballots with rating scales were used, and the scoring of 0 to 5 points respectively represented “none” to “high” intensity of a perceptive flavor. Samples were taken out from refrigeration 1 h prior to evaluation and shaken well. Samples (20 mL) were dispensed in 30 mL portion cups labeled randomly with 3-digit codes, covered, and kept at room temperature. The panelists independently evaluated the samples according to the intensity of the specific flavor sensed, and were given a 3 min rest between individual samples, followed by an additional 15 min rest between each replication.

Table 2 Descriptions of soymilk flavor

1.6 Statistical analysis

All experiments were conducted in triplicate, and data were expressed as means ± SDs. All graphs were performed using MS Excel 2007 and OriginPro 8.5 softwares, and ANOVA was conducted using SPSS 17.0 software. Sign-if-icant differences among variables were determined by Dun-can’s multiple range test (=0.05).

Odor activity value (OAV) was the ratio of the conc-entration of one flavor compound and the corresponding sensory threshold. Therefore, the higher the OAV of a cer-tain flavor compound is, the greater its contribution is to the whole flavor profile. OAV sum of critical beany flavor com-pounds was calculated by the sum of the OAV of hexanal, trans-2-hexenal, 1-octen-3-ol, hexanol and trans, trans-2, 4-decadienal, while OAV sum of critical non-beany flavor compounds was the sum of the OAV of trans-2- octenal, nonanal and trans-2-nonenal.

The calculation for the comprehensive scoring system of sensory evaluation was as follows: the sum of the scores of beany flavor, including beany odor and mushroom flavor, was subtracted by the score of non-beany flavor, including sweet aroma. Accordingly, a lower score of beany flavor and a hig-her score of non-beany flavor would result in the lower com-prehensive score, corresponding to the better flavor quality.

2 Results and analysis

2.1 Changes in critical flavor compounds during heat treatment

In order to investigate the effect of heat treatment on soymilk flavor, five critical beany flavor compounds (hex-anal, hexanol, 1-octen-3-ol, trans-2-hexenal, trans, trans-2, 4-decadienal) and three critical non-beany flavor compounds (nonanal, trans-2-octenal, trans-2-nonenal) were selected in this study, and the changes in the contents of these flavor compounds during heat treatment of soymilk were an-alyzed.

2.1.1 Changes in critical beany flavor compounds dur-ing heat treatment

Hexanal had the maximum content (3.68 mg/L) at 30℃. As the heating temperature increased, the content of hexanal gradually decreased and reached a minimum of 2.27 mg/L at 60℃. Subsequently, only a slight increase was observed, and it decreased by 21.5% at 90℃ compared with that at 30℃ (Fig. 1). Hexanal, which represents herbal flavor, is known to have a low threshold of 0.0004 mg/kg and a high conc-entration in soymilk, indicating that it contributs the most to the beany flavor and is the main beany flavor compound in soymilk[11]. Moreover, the formation of hexanal is mainly induced by lipoxygenase 2, which is most active at 20- 30℃[12-13]; thus, hexanal had the highest content at 30℃. However, lipoxygenase 2 was gradually inactivated during soymilk heating.

Fig.1 Content changes of critical beany flavor compounds during heat treatment

Hexanol also represents herbal flavor and paint flavor, and its threshold is 0.5 mg/kg, which is much higher than that of hexanal. When soymilk was heated to 40℃, the content of hexanol rapidly increased to 1.6 times as high as that at 30℃, followed by a slow decrease with a further increase in temperature; however, the content at 90℃ was still 35.1% higher than that at 30℃ (Fig. 1). Hexanol is reported to be derived from 13-hydroperoxide, which is generated by the degradation of unsaturated fatty acids via lipoxygenase 1[14]. lipoxygenase 1 is most active at app-roximately 50℃[7]; thus, a significant amount of 13-hydr-operoxide was generated and further degraded into hexanol at this temperature. For this reason, the maximum content of hexanol was obtained at 40-50℃. Furthermore, hexanol might be partly converted into hexanal[15], and its content actually had a tendency to decrease above 50℃.

1-octen-3-ol is endowed with mushroom flavor, wh-ich is undesirable to western consumers who consider it as the main source of soymilk off-flavor[16]; however, for oriental consumers, its flavor is relatively acceptable[3]. The thre-shold of this compound is 0.001 mg/kg. Throughout the heat treatment, the content of 1-octen-3-ol presented a bell- shaped variation (Fig. 1), and its content at 90℃ was still significantly higher than that at 30℃. The formation pat-hway of 1-octen-3-ol has been a longstanding con-troversial issue; some researchers believe that 1-octen-3-ol is formed with high flavor dilution factor during soybean soaking[17], whereas others speculate that it is derived from 10-hyd-rop-eroxide, which is formed by some special oxidation process, rather than catalyzed by lipoxygenase[16]. Frankel et al.[18]further infer that 10-hydroperoxide is the product of linoleic acid via autoxidative degradation.

Trans-2-hexenal imparts leaf flavor to soymilk, and its threshold is 0.027 mg/kg[3]. The formation of trans-2-hexe-nal results from the decomposition of 13-hydroperoxide by hydroperoxide lyase (HPL), and 13-hydroperoxide is origin-ated from linolenic acid[3]. As shown in Fig. 1, the maximum content of trans-2-hexenal was obtained at 40℃, and then the content quickly decreased until 50℃. Subsequently, the content of trans-2-hexenal remained stable; nevertheless, its content at 90℃was 18.9% lower than that at 30℃.

Trans, trans-2, 4-decadienal primarily contributes the oxidized oil flavor, of which the threshold is 0.000 07 mg/kg, and 9-hydroxy linoleic acid is reported as its precursor[18]. The content of trans, trans-2, 4-decadienal had no remarkable change during the entire heat treatment (Fig. 1), and only a slow decline was observed above 70℃. According to the report of Kobayashi et al.[16], soymilk prepared by soybean cultivars lacking lipoxygenase has much lower contents of this compound than normal soymilk, indicating that lipox-ygenase plays an important role in the formation of trans, trans-2, 4-decadienal.

Furthermore, the capability of protein binding with flavor compounds by hydrophobic interaction changes because of the thermal denaturation of soymilk protein[19]. Fat-soluble flavor compounds can also be absorbed and dis-solved by fat[20]. Therefore, the interaction of protein or fat in soymilk with flavor compounds could also reduce the release concentrations of these compounds or change the ratio of each flavor compound in the entire flavor profile, thus affecting soymilk flavor[21].

2.1.2 Changes in critical non-beany flavor compounds during heat treatment

As one of the most essential non-beany flavor compo-unds, nonanal is characterized by broccoli flavor and orange flavor, and the sensory threshold is 0.001 mg/kg[3]. Nonanal is derived from 9-hydroperoxide decomposed by HPL, and 9-hydroperoxide is the product of linoleic acid and linolenic acid degraded by lipoxygenase 1[22], which is regarded as the key enzyme in the formation of 9- hydroperoxide. During the heat treatment of raw soy-milk, lipoxygenase 1 was act-ivated, and nonanal was formed. Nonanal had the maximum content at 60℃. Subsequently, the content slightly decr-eased and tended to stablize above 70℃. When the soymilk was heated to 90℃, its content was still 1.6 times as high as that at 30℃ (Fig. 2).

Trans-2-octenal, as the main source of green cucumber flavor, of which the threshold is 0.003 mg/kg, is from 11-hydroperoxide, which is formed by linolenic acid via lipoxygenase 1[23]. Given that lipoxygenase 1 has the hig-hest activity at 50℃, the maximum content of trans-2-octenal was reached at 50℃ (Fig. 2). Furthermore, octanal, which is reported to be the degradation product of 11-hyd-roperoxide, can be converted into octenal during heat treatment[24].

Trans-2-nonenal has fruit flavor and cooked carrot flavor[25], and the threshold is 0.000 5 mg/kg. No significant change in the content of trans-2-nonenal occurred during the soymilk heating (Fig. 2), indicating that this flavor com-pound is thermally stable, which might be attributed to the reinforcement of the binding between whey protein and trans-2-nonenal as the temperature increases[26]. However, the formation mechanism of trans-2-nonenal has not been explicitly confirmed. The content of trans-2-nonenal in soymilk prepared by soybean cultivars lacking lipoxygenase is reported to be lower than that in normal soymilk[7]; ther-efore, lipoxygenase is possibly involved in the for-mation of this compound. Some researchers believe that it might be formed by non-enzymatic reaction during heat treatment[25]. According to Frankel et al.[18], trans-2- nonenal is derived from 9-/10-hydroperoxide, which is originated from linolic acid by autoxidation and photo-oxidation. Therefore, the oxidation of unsaturated fatty acids in soymilk still could be inferred to occur even if lipoxy-genase was inactivated by heating as long as sufficient oxygen or light was provided, resulting in the stable content of trans-2-nonenal.

2.1.3 Changes in OAV sums of critical flavor comp-ounds during heat treatment

As shown in Fig. 3, the maximum OAV sum of beany flavor compounds was reached at 50℃, and then a rapid decrease was observed, followed by a tendency to slow decrease above 70℃. The sum at 90℃ was 36.6% lower than that at 30℃. This result indicated that the critical beany flavor compounds in soymilk were largely generated below 60℃, but the intensity of these compounds sign-ificantly decreased as the temperature rose. The OAV sum of non-beany flavor compounds gradually increased during heat treatment, reached the maximum value at 60℃, slightly decreased at 70℃, and finally remained stable. The sum at 90℃ was 43.6 % higher than that at 30℃. The-refore, an increasing trend in the intensity of non-beany flavor compounds during the soymilk heating demonstrated that the increase of temperature was beneficial to the release of these compounds. All these results above revealed that significant fluctuation of soymilk flavor compounds occurred at the range of 30-60℃ during the heat treatment, particularly 40 and 60℃, which could be considered as the jump points. In addition, when soymilk was sufficiently heated, the beany flavor was remarkably reduced, whereas the non-beany flavor was relatively inc-r-eased, and the ratio of good flavor in the whole flavor profile was enhanced, making soymilk flavor more acc-ep-table.

2.2 Effect of-cyclodextrin addition on critical soymilk flavor compounds

Since-cyclodextrin is able to entrap flavor compounds, it might change the flavor quality of soymilk. Accordingly, different concentrations (0.25%, 0.50%, 0.75%, 1.0%) of-cyclodextrin were respectively added into soymilk at four different heating periods, namely, before heating, at 40 or 60℃, and after heating. The effects of-cyclodextrinon addition on three critical beany flavor compounds (hexanal, hexanol, 1-octen-3-ol) and one critical non-beany flavor compound (trans-2-octenal) were analyzed.

Note: Critical beany flavor compounds included hexanal, trans-2-hexenal, 1-octen-3-ol, hexanol and trans, trans-2, 4-decadienal, and critical non-beany flavor compounds included nonanal, trans-2-octenal and trans-2-nonenal.

As shown in Fig. 4a, the addition of-cyclodextrin could effectively decrease hexanal content in soymilk. When 0.25%-cyclodextrin was added into soymilk before heating, at 40 or 60 ℃, hexanal content decreased by approximately 25.8% compared with the control; as the concentration of-cyclodextrin increased, only the slight decrease occurred when it was added before heating or at 40 ℃, while the further decrease was achieved by increasing the conc-entration of-cyclodextrin at 60 ℃, and the hexanal content decreased byapproximately 55% compared with the control. A similar effect was observed when-cyclodextrin was added after heating, but there seemed no significant diff-erence in hexa-nal content with the increasing concentration of-cyclod-extrin at this period (＞0.05).

For hexanol (Fig. 4b), the addition of 0.25%-cyclo-dextrin before heating, at 40 or 60 ℃ had little effect on decreasing the content of this compound (＞0.05), but the addition conducted after heating could significantly decrease its content to 0.26 mg/L, which was 36.7% lower than the control, and the similar effects were obtained when ≥0.50%-cyclodextrin was added into soymilk before heating or at 40 ℃. The addition of 0.50%-cyclodextrin at 60 ℃ or after heating led to the further decrease in hexanol content (about 0.13 mg/L), but further increasing the concentration of-cyclodextrin at these two periods could only help slightly decrease the content of this beany flavor compound.

Fig. 4c exhibited the content change of 1-octen-3-ol. When 0.25%-cyclodextrin was added into soymilk before heating or at 40 ℃, no significant difference in 1-octen-3-ol content was observed compared with the control; however, when the addition occurred at 60 ℃ and after heating, the content of this compound was 0.067 mg/L and 0.080 mg/L, which was 32.3% and 19.2% lower than the control (＜0.05). The addition of 0.5%-cyclodextrin at the four periods all could result in the further decrease of 1- octen-3-ol content at different levels, and the most rem-arkable decrease occurred at 60 ℃, since the content of 1-octen-3-ol decreased by 65.7% compared with the control, but further increase of-cyclodextrin concentration could hardly decrease the con-tent of this compound.

Note: Control was the soymilk without β-cyclodextrin addition during heat treatment.

As to trans-2-octenal, it could be seen from Fig. 4d that, when 0.25%-cyclodextrin was added before heating, at 40 or 60 ℃, significant decrease in trans-2-octenal content was observed, compared with the control; the addition of 0.50%-cyclodextrin at these three periods more significantly decreased the content of this compound, especially when the addition was conducted at 60 ℃. At 60 ℃, the trans-2- oct-enal content decreased by 71.2% compared with the control, and the higher conc-entration of-cyclodextrin (0.75% or 1%) at this period had little effect on the further decrease in trans-2-octenal content. When-cyclodextrin was added into soymilk after heating, the content of trans- 2-octenal was approximately 29.7% lower than the control, and no sig-nificant variation in the content of this compound was obs-erved with the increase of-cyclodextrin conc-e-ntration.

It has been reported by a lot of researches that,-cyc-lodextrin is a cyclic carbohydrate composed of seven glu-cose units connected by-1, 4 glycosidic bonds, which is a cage-like structure, and the primary or secondary hydroxyl groups of glucose are at the outer surface, making the exterior of-cyclodextrin hydrophilic, whereas the skeletal carbons and glycosidic oxygen atoms are oriented in the interior, imparting a hydrophobic character to its cavity[27-28]. Possessing this specific structure,-cyclodextrin has the notable ability to form inclusion complexes (host-guest complexes) with a wide range of compounds[29], especially for flavor compounds, including straight or branched chain aldehydes, ketones, alcohols, aromatics, etc., since the lip-ophilic cavity of-cyclodextrin provides a microe-nvir-on-ment into which these compounds can enter, and no covalent bonds are broken or formed during the formation of incl-usion complex[8]. The main driving force of complex form-ation is the release of enthalpy-rich water molecules from the cavity, which tend to be displaced by more hydrophobic molecules (like many flavor compounds) pres-ent in the sol-ution to attain the decrease of-cyclodextrin ring strain, resulting in a more stable lower energy state[8]. Furthermore, temperature has important influence on cyclodextrin co-m-plexes and a relatively high temperature might promote the solubility of cyclodextrin and accelerate the formation of complexes[8]. These may help explain why-cyclodextrin ad-ded into soymilk could significantly decrease the contents of critical flavor compounds, and the addition conducted at 60 ℃ had a more significant effect than the other three per-i-ods. In addition, although it was expected that higher conc-entrations of-cyclodextrin in soymilk would create greater opportunities for complex for-mation with soymilk fla-v-or compounds[30], the decreasing tendency in the con-tents of these flavor compounds with the increasing con-c-entrations of-cyclodextrin were not linear, and it may be related with the limited efficacy of-cyclodextrin to form complexes with flavor compounds[31], since it is a one-site host, i.e., its empty cavity can accom-modate only one guest compound, and other components in soymilk, like fatty acids or pho-sp-holipids, are also the possible competitive guests[27].

Meanwhile, the binding of host-cyclodextrin with guest compounds is not fixed or permanent, but rather a dynamic equilibrium, and the binding strength depends on several key factors. The first is dimensional fitness, which greatly depends on the sizes of host and guest or certain key functional groups within the guest. If one compound has the wrong size, it will not fit properly into the-cyclodextrin cavity. The second is thermodynamic driving force among host cyclodextrin, guest compounds and solvent (like water molecules), and an energetic driving force that pulls the guest into the host is the guarantee for the formation of the inclusion complexes. The third is specific local interactions between surface atoms of host cavity and guest compounds[8]. Therefore, those complexes with stronger binding strength between host and guest are possibly more stable in the system. It is estimated that, the-cyclodextrin cavity can accommodate about five methylene groups of a hydrocarbon chain[27], so it can be easily speculated that, C6 compounds, rather than C8 compounds, are more compatible for the cavity and preferentially entrapped by-cyclodextrin, and the complexes formed are also more stable. This might be the reason why the contents of hexanal and hexanol were more significantly decreased with the addition of-cycl-odextrin, compared with those of 1-octen-3-ol and trans-2- octenal, especially when the addition occurred after heating.

2.3 Effect of-cyclodextrin addition on sensory evalu-ation of soymilk flavor

Each flavor compound has a corresponding sensory threshold, which is the lowest concentration of this com-pound that can be perceived. A certain flavor compound with a relatively low concentration in soymilk can still be clearly perceived because of its low sensory threshold. The-refore, the sensory evaluations of three typical flavor indices, including beany odor, mushroom flavor, and sweet aroma, were conducted to investigate the actual effect of-cyc-lodextrin addition on soymilk flavor (Table 3).

Table 3 Effect of β-cyclodextrin during heat treatment on sensory evaluation of soymilk flavor

Note: Any two means in the same line followed by different letters were significantly different (＜0.05).

For beany odor, the addition of 0.25%-cyclodextrin after heating could markedly decrease the score of this odor, which was 35.1% lower than the control, while the addition at the other three periods seemed to have little effect on decreasing this undesirable odor. As the concentration of-cyclodextrin increased, the score of beany odor tended to decrease, which depended greatly on the adding period of-cyclodextrin, and the addition of 0.75% or 1.00%-cyc-lodextrin occurred at 60 ℃ or after heating was proved to be best, since the score of beany odor decreased by app-ro-ximately 54.8% compared with the control.

According to the results of mushroom flavor, when 0.25%-cyclodextrin was added into soymilk before heating or at 40℃, no significant difference in the score of this flavor was observed compared with the control (＞0.05), however, when the addition occurred at 60 ℃ or after hea-t-ing, the score decreased by 45.3% (＜0.05). Similarly,- cyclodextrin with the higher concentration added at the four periods could also weaken this flavor, especially when ≥0.50%-cyclodextrin was added at 60℃, the score of mus-hroom flavor was 73.8% lower than the control (＜0.05).

As to sweet aroma, the addition of-cyclodextrin into soymilk didn’t cause significant difference in the score of this aroma (＞0.05), except that, when 0.50% or 0.75%-cyclodextrin was added at 60℃, the sweet aroma suffered a slight loss, compared with the control.

Given that beany and non-beany flavors coexisted in soymilk, a comprehensive scoring system was applied to better characterize the effect of-cyclodextrin addition on the soymilk flavor profile. As shown in Table 4, compared with the control,-cyclodextrin, with the higher conc-entration (≥0.50%) added at 60 ℃ or after heating, could effectively decrease the comprehensive score of soymilk flavor, thus improving the entire flavor quality of soymilk. The lowest score was achieved when 0.75%-cyclodextrin was added at 60℃ during the soymilk heating, therefore, soymilk prepared under this condition was considered to have the best flavor quality.

Table 4 Comprehensive score of sensory evaluation of soymilk flavor with addition of β-cyclodextrin during heat treatment

2.4 Analysis for application of-cyclodextrin to soymilk production

Since-cyclodextrin has been widely used as a food ingredient for flavor modification in recent years[8], it can be also well applied to soymilk production and has a good prospect of industrialization.

First, it is low-priced, and 1 kg-cyclodextrin costs about 0.8 yuan. If 0.5%-1.0%-cyclodextrin is used, it can be calculated that, 5-10 kg-cyclodextrin is needed in the production of 1 t soymilk, which only costs 4-8 yuan. On the other hand, since the addition of-cyclodextrin can greatly improve the entire flavor quality of soymilk, it is conducive to promoting the sales and increasing the profits of soymilk products.

Second, it is safe and accessible. Many toxicity studies have demonstrated that,-cyclodextrin is practically non- toxic due to lack of absorption from the gastrointestinal tract[32]. In addition, although-cyclodextrin is derived from starch and can be partly hydrolyzed by human-amylase, the low degree of hydrolysis is not sufficient to cause significant changes in sweetness[30], and it will not affect the taste quality of soymilk products.

Third, it is easily practiced. According to the results in this study, it is better for-cyclodextrin to be added at 60 ℃ or after heating. Accordingly,-cyclodextrin can be dis-s-olved in the batching tank, which is commonly used for dissolving or mixing ingredients like sugars in the pro-d-uc-tion line of soymilk, and then separately added into soy-milk at 60 ℃ during heat treatment. It can also be mixed with the other ingredients in the tank and added after the heat treatment of soymilk. Therefore, it can be seen that no extra equ-ipment is needed to realize the addition of-cyc-lod-extrin.

3 Conclusion

Heat treatment during soymilk processing has an essential impact on the entire flavor profile of soymilk. Acc-ording to the content changes and OAV sums of critical flavor compounds, although different flavor compounds in soymilk varied in terms of sensitivity to temperature, the intensity of critical beany flavor had a tendency to decrease during heat treatment, while the increase of temperature was beneficial to the release of non-beany flavor compounds; moreover, significant fluctuation in the contents of these fla-vor compounds occurred at the range of 30-60℃.-cyc-l-o-dextrin with a higher concentration (≥0.50%) added at 60 ℃ led to the significant decrease in the contents of critical soy-milk flavor compounds. According to the sensory evaluation, when 0.75%-cyclodextrin was added at 60℃ during heat treatment, soymilk presented the best flavor quality with the lowest comprehensive score of 1.83. Alth-ough the effects of heat treatment and-cyclodextrin addition on soymilk flavor have been investigated, the cha-nging mechanisms of critical flavor compounds during the soymilk heating remain unclear. Thus, the releated mech-anisms will be clarified in our future work, which is aimed at providing a scientific basis on how to eliminate beany flavor and retain non-beany flavor, thus funda-m-en-tally improving the flavor quality of soymilk.

[1] Yuan Shaohong, Chang S K C. Selected odor compounds in soymilk as affected by chemical composition and lipoxy-ge-n-ases in five soybean materials[J]. J Agric Food Chem, 2007,

55(2): 426-431.

[2] Kaneko S, Kumazawa K, Nishimura O. Studies on the key aroma compounds in soy milk made from three different soybean cultivars[J]. J Agric Food Chem, 2011, 59(22): 12204-12209.

[3] Lv Yanchun, Song Huanlu, Li Xin, et al. Influence of blanching and grinding process with hot water on beany and non-beany flavor in soymilk[J]. J Food Sci, 2011, 76(1): S20-S25.

[4] Poliseli-Scopel F H, Gallardo-Chacón J J, Juan B, et al. Characterisation of volatile profile in soymilk treated by ultra high pressure homogenization[J]. Food Chem, 2013, 141(3): 2541-2548.

[5] Yuan Shaohong, Chang S K C, Liu Zhisheng, et al. Elimination of trypsin inhibitor activity and beany flavor in soy milk by consecutive blanching and ultrahigh-temperature (UHT) processing[J].J Agric Food Chem, 2008, 56(17): 7957-7963.

[6] Mizutani T, Hashimoto H. Effect of grinding temperature on hydroperoxide and off-flavor contents during soymilk manufacturing process[J]. J Food Sci, 2004, 69(3): 112-116.

[7] Yuan Shaohong, Chang S K C. Selected odor compounds in cooked soymilk as affected by soybean materials and direct steam injection[J]. J Food Sci, 2007, 72(7): S481-S486.

[8] Del Valle E M M. Cyclodextrins and their uses: A review [J]. Process Biochem, 2004, 39(9): 1033-1046.

[9] Shi Xiaodi, Li Jingyan, Wang Shuming, et al. Flavor characteristic analysis of soymilk prepared by different soybean cultivars and establishment of evaluation method of soybean cultivars suitable for soymilk processing[J]. Food Chem, 2015, 185: 422-429.

[10] Murray J M, Delahunty C M. Selection of standards to reference terms in a cheddar-type cheese flavor language[J]. J Sens Stud, 2000, 15(2): 179-199.

[11] Poliseli-Scopel F H, Hernández-Herrero M, Guamis B, et al. Sterilization and aseptic packaging of soymilk treated by ultra high pressure homogenization[J]. Innov Food Sci Emerg, 2014, 22: 81-88.

[12] Mellor N, Bligh F, Chandler I, et al. Reduction of off-flavor generation in soybean homogenates: A mathematical model [J]. J Food Sci, 2010, 75(7): 131-138.

[13] Yang Aijun, Smyth H, Chaliha M, et al. Sensory quality of soymilk and tofu from soybeans lacking lipoxygenases[J]. Food Sci Nutr, 2016, 4(2): 207-215.

[14] Ties P, Barringer S. Influence of lipid content and lipoxygenase on flavor volatiles in the tomato peel and flesh [J]. J Food Sci, 2012, 77(7): 830-837.

[15] Prese?ki A V, Vasi?-Ra?ki ?. Mathematical modelling of the dehydrogenase catalyzed hexanol oxidation with coenzyme regeneration by NADH oxidase[J]. Process Biochem, 2009, 44(1): 54-61.

[16] Kobayashi A, Tsuda Y, Hirata N, et al. Aroma constituents of soybean [glycine max (L.) Merril] milk lacking lipoxygenase isozymes [J]. J Agric Food Chem, 1995, 43(9): 2449-2452.

[17] Boué S M, Shih B Y, Carter-Wientjes C H, et al. Identification of volatile compounds in soybean at various developmental stages using solid phase microextraction[J]. J Agric Food Chem, 2003, 51(17): 4873-4876.

[18] Frankel E N, Neff W E, Selke E. Analysis of autoxidized fats by gas chromatography-mass spectrometry: VII. Volatile thermal decomposition products of pure hydroperoxides from autoxidized and photosensitized oxidized methyl oleate, linoleate and linolenate[J]. Lipids, 1981, 16(5): 279-285.

[19] Wang Kun, Arntfield S D. Binding of carbonyl flavours to canola, pea and wheat proteins using GC/MS approach [J].Food Chem, 2014, 157: 364-372.

[20] Mattes R D. Accumulating evidence supports a taste component for free fatty acids in humans[J]. Physiol Behav, 2011, 104(4): 624-631.

[21] Achouri A, Boye J I, Zamani Y. Changes in soymilk quality as a function of composition and storage[J]. J Food Quality, 2007, 30(5): 731-744.

[22] Kiralan M, Kiralan S S. Changes in Volatile Compounds of Black Cumin Oil and Hazelnut Oil by Microwave Heating Process[J]. J Am Oil Chem Soc, 2015, 92(10): 1445-1450.

[23] Clapp C H, McKown J, Xu Hui, et al. The action of soybean lipoxygenase-1 on 12-iodo-cis-9-octadecenoic acid: The importance of C (11)-H bond breaking[J]. Biochem, 2000, 39(10): 2603-2611.

[24] Yettella R R, Castrodale C, Proctor A. Oxidative stability of conjugated linoleic acid rich soy oil[J]. J Am Oil Chem Soc, 2012, 89(4): 685-693.

[25] Zhang Yan, Guo Shuntang, Liu Zhisheng, et al. O?-flavor related volatiles in soymilk as a?ected by soybean variety, grinding, and heat-processing methods[J]. J Agric Food Chem, 2012, 60: 7457-7462.

[26] Ku?hn J, Considine T, Singh H. Binding of flavor compounds and whey protein isolate as affected by heat and high pressure treatments[J]. J Agric Food Chem, 2008, 56(21): 10218-10224.

[27] Damodaran S, Arora A. Off-flavor precursors in soy protein isolate and novel strategies for their removal[J]. Annu Rev Food Sci T, 2013, 4: 327-346.

[28] Arora A, Damodaran S. Competitive binding of off-flavor compounds with soy protein and-cyclodextrin in a ternary system: a model study[J]. J Am Oil Chem Soc, 2010, 87(6): 673-679.

[29] Arora A, Damodaran S.-Cyclodextrin-mediated removal of soy phospholipids from the air–water interface[J]. J Am Oil Chem Soc, 2011, 88(2): 213-222.

[30] Suratman L L I, Jeon I J, Schmidt K A. Ability of cyclodextrins to entrap volatile beany flavor compounds in soymilk[J]. J Food Sci, 2004, 69(2): 109-113.

[31] Arora A, Damodaran S. Removal of soy protein-bound phospholipids by a combination of sonication,-cyclodextrin, and phospholipase A2treatments[J]. Food Chem, 2011, 127(3): 1007-1013.

[32] Delogu G, Juliano C C A, Usai M. Thymus catharinae Camarda essential oil: β-cyclodextrin inclusion complexes, evaluation of antimicrobial activity[J]. Nat prod res, 2016, 30(18): 2049-2057.

熱處理及-環(huán)糊精添加量對豆乳風(fēng)味的影響

施小迪1，呂艷春1,2，郭順堂1※

（1. 中國農(nóng)業(yè)大學(xué)食品科學(xué)與營養(yǎng)工程學(xué)院，北京 100083；2. 遵義醫(yī)學(xué)院公共衛(wèi)生學(xué)院，遵義563000）

風(fēng)味是評價(jià)豆乳產(chǎn)品品質(zhì)的重要指標(biāo)，而在豆乳生產(chǎn)中，熱處理及-環(huán)糊精的添加對豆乳風(fēng)味品質(zhì)的影響則未可知。因此，該研究首先分析了豆乳加熱過程中（30-90 ℃范圍內(nèi)）關(guān)鍵性風(fēng)味物質(zhì)的含量變化，并比較了在不同加熱階段（加熱前、加熱至40 ℃時(shí)、加熱至60 ℃時(shí)、加熱后）添加不同質(zhì)量分?jǐn)?shù)-環(huán)糊精（0.25%、0.50%、0.75%、1.00%）所獲得的豆乳的風(fēng)味品質(zhì)。研究結(jié)果顯示，豆乳中的關(guān)鍵性風(fēng)味物質(zhì)對溫度的敏感度因風(fēng)味物質(zhì)的種類不同而異，隨著加熱過程中溫度的升高，豆腥味的強(qiáng)度呈現(xiàn)顯著降低的趨勢，而非豆腥味的強(qiáng)度則相對增強(qiáng)。添加-環(huán)糊精能夠顯著降低關(guān)鍵性風(fēng)味物質(zhì)在豆乳體系中的濃度，當(dāng)添加量為≥0.50%的-環(huán)糊精在豆乳加熱至60 ℃階段添加時(shí)，所得豆乳中己醛、己醇、1-辛烯-3-醇等關(guān)鍵性豆腥味成分含量的降低幅度最為顯著(＜0.05)，但同時(shí)反-2-辛烯醛這種非豆腥味成分的含量也有所損失。豆乳豆腥味、蘑菇味和甜香味的感官評價(jià)結(jié)果也呈現(xiàn)類似的趨勢，另外，根據(jù)感官評價(jià)的綜合得分，當(dāng)添加量為0.75%-環(huán)糊精在豆乳加熱至60 ℃時(shí)添加時(shí)，所得豆乳的整體風(fēng)味品質(zhì)最佳。由于-環(huán)糊精價(jià)格低廉、且安全性高，其在豆乳加工過程中的添加又較易實(shí)現(xiàn)，因而在豆乳生產(chǎn)中具有良好的應(yīng)用價(jià)值。

熱處理；溫度；風(fēng)味；非豆腥味；-環(huán)糊精；感官評價(jià)

10.11975/j.issn.1002-6819.2017.08.039

TS214.2 Document code: A Article ID: 1002-6819(2017)-08-0293-08

2016-09-27 Revised date: 2017-04-02

National Major Research Plan of China (grant 2016YFD0400402).

Shi Xiaodi, female, Jiangyin, Jiangsu, ph.D candidate, food science of College of Food Science and Nutritional Engineering in China Agricultural University. Email: s11070783@163.com.

Guo Shuntang, male, Haerbin, Heilongjiang, deputy secretary of College of Food Science and Nutritional Engineering in China Agricultural University, ph.D in food science. Email: shuntang@cau.edu.cn.

Shi Xiaodi, Lv Yanchun, Guo Shuntang. Effects of heat treatment and-cyclodextrin addition on soymilk flavor[J].Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(8): 293－300. (in English with Chinese abstract) doi：10.11975/j.issn.1002-6819.2017.08.039 http: //www.tcsae.org

施小迪，呂艷春，郭順堂. 熱處理及-環(huán)糊精添加量對豆乳風(fēng)味的影響[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2017，33(8)：293－300. doi：10.11975/j.issn.1002-6819.2017.08.039 http: //www.tcsae.org