Modulated Interactions between the Electrochemical Label and Liquor Compounds: Differentiating between Aging Times of Typical Chinese Liquors

Ferrocenylmethanol, ethanol, acetic acid, propionic acid, n-butyric acid, n-hexylic acid, n-octanoic acid, glyoxal, ethyl acetate, and potassium chloride were purchased from Aladdin (Shanghai, China). All chemicals were used as received. All solutions were freshly prepared by ultrapure water, with a resistivity of 18.2 MΩcm⁻¹.

The polycrystalline gold electrode (1 mm in radius) was polished with 1.0, 0.3 and 0.05 μm alumina in particle size successively to achieve a mirror-like surface, rinsed with ethanol and ultrapure water. Then the polished electrode was electrochemically cleaned by cycling the potential from −0.2 to +1.6 V with a scan rate of 50 mV · s⁻¹ in 0.1 M H₂SO₄. Cyclic voltammetry (CV) was carried out in a conventional three-electrode cell at room temperature near 298 K, with a commercial electrochemical workstation (CHI660C). All the electrodes were purchased from CH Instruments, Inc. (Shanghai, China). Before recording of CV, FcCH₂OH and KCl were added to the liquor samples to achieve a final concentration of 0.5 mM and 0.1 M respectively. The oxygen was removed from the sample solutions by passing of N₂ gas not less than 10 min before the recording of the cyclic voltammogram. Three samples are taken for each CV test (three parallel measurements). An Abbe refractometer (HT512ATC, Hengan Electronic Technology CO. LTD, Shandong, China) was used to determine the alcohol content of the liquors. The conformation of FcCH₂OH and C₂H₅OH was optimized by calculations using density functional theory (DFT). Details about the DFT calculation can be found in our previous work.²⁰

The typical Chinese liquor samples were collected from the manufacturer Xiangjiao Group Ltd, Shaoyang, China. Six samples with different biological aging times (1, 2, 3, 4, 5, 10 years) were collected. All liquor samples used in this study were directly collected from storage containers without any additive. The constituent analysis of liquor samples above mentioned was primarily performed by gas chromatography (Agilent Technologies Co. Ltd), as shown in Table SI in supporting information (available online at stacks.iop.org/JES/167/027517/mmedia). GC chromatograms of the liquor samples with different aging times are shown in supporting information (Figs. S1-S6).

Classification of the aging time of the liquor samples is first performed by the electronic tongue system, which is the traditional technology for food analysis. The scheme of the electronic tongue is shown in Fig. 1a. As seen, the sensor array of the electronic tongue is composed of a working electrode (six inert electrodes: platinum, gold, tungsten, titanium, nickel and silver electrode), an Ag/AgCl electrode reference electrode, and a platinum counter electrode for standard three-electrode systems. The electronic tongue system is based on the multi-frequency pulse voltammetric scanning method (shown in Fig. 1b) and uses the multiple working electrodes to optimize the combination of sensor arrays. The data analysis methods of the system include principal component analysis (PCA), discriminant function analysis (DFA), linear discriminant analysis (LDA) and partial least squares regression analysis (PLS).

Zoom In Zoom Out Reset image size

As shown in Fig. 1c, the score plot of PC1 and PC2 derived from PCA analysis did not bring interest in the sample separation (other combinations of PCs were also examined and showed similar results). The LDA method was further performed, shown in Fig. 1d, and a general separation between samples with different aging times was observed. As can be seen, the electronic tongue system was effective to differentiate samples from six aging times, however, the compound difference within the samples is not clear. Electrochemistry-based methods emerge as an alternative strategy for a quick and economic analysis of food constituent when we consider some of the characteristics of conventional methods, like large number of samples, use of expensive reagents unfriendly to the environment and expensive equipment.^21–23 Therefore, an electrochemistry-based method was designed to reveal the certain compound differences within the samples of various aging time, shown in the following experiments.

The typical Chinese liquors were characterized using the electrochemical label (FcCH₂OH) with clean and simple design: A gold disk electrode (1 mm in radius) was used as the working electrode, a platinum wire as the counter electrode and an Ag/AgCl (sat. KCl) electrode as the reference electrode for standard triple electrode system. CV of liquors in the absence of FcCH₂OH, and CV of liquors (from six aging times, 1, 2, 3, 4, 5 and 10 year) containing 0.5 mM FcCH₂OH and 0.1 M KCl are shown in Fig. 2a. As seen, a pair of well-shaped redox peaks are observed for both the CVs in the aqueous solution which contains 0.5 mM FcCH₂OH. However, in the absence of FcCH₂OH, no redox peaks are presented in the same potential window of the CV (dotted line in Fig. 2a). Therefore, it illustrates that the well-shaped redox peaks observed in the CV are attributed to the redox process of the electrochemical label, FcCH₂OH. Moreover, the evident changes for the CVs are found, as shown in Fig. 2b, the anodic peak potential (E_pa) is decreasing along with the increase of the aging time while the cathodic peak potential (E_pc) keeps a constant. Generally, the peak potential of FcCH₂OH, which is a stable redox species, will not change in a certain solution environment. However, the obvious potential shifts indicate that the redox process of the electrochemical label FcCH₂OH,¹⁹ is closely related to the intermolecular interactions between FcCH₂OH and the diverse compounds in the liquors. The shifts in redox potential, of course, is not particularly novel or surprising observation. Certainly, the effect of the intermolecular interactions (e.g., hydrogen bonding) on redox potentials has been recognized, particularly in the case of DNA²⁴ and proteins.²⁵ In other words, the liquors from various aging times provide a diverse solution environment for FcCH₂OH, thus the redox potential of FcCH₂OH may become an indicator of the aging time of the liquors. In order to check the memory effect²⁶ of this system, the first, second and tenth cycle of the CV curves are shown in Fig. S7. As seen, the CV curves for the first, second and tenth cycle are almost the same (except the curve of first cycle is slightly different due to the double-layer charging current). Thus, it can be suggested that the potential shifts for the liquor samples are due to the intermolecular interactions between FcCH₂OH and the diverse compounds in the liquors rather than memory effect of this system. Thus, the CV curves were collected with three cycles every time, and the CV curve of second cycle was used for the subsequent measurements.

Zoom In Zoom Out Reset image size

As the major constituent in liquors is alcohol (C₂H₅OH), the influence of alcohol content on the redox potential of FcCH₂OH was firstly investigated. Six artificial model liquors are obtained by mixing water and alcohol, the content of alcohol is 0%, 12%, 24%, 48%, 64%, and 72% respectively. CV of model liquors containing 0.5 mM FcCH₂OH and 0.1 M KCl are shown in Fig. 3a. As seen, all the CVs show one couple of quasi-reversible redox peaks, however, along with the increase of alcohol content, the redox peaks are shifting to the considerably positive potential. The extent of potential shift, described by Diane K. Smith et al.,²⁷ who observed a selective recognition of urea and amide derivatives, has been evaluated by electrochemical means: It is realized that the extent of potential shift is related with the strength of the intermolecular interactions. As shown in Fig. 3b, the E_pa and E_pc were plotted vs volume ratio of C₂H₅OH, the obvious increase in E_pa and E_pc were observed. Moreover, the extent of potential shift for E_pa and E_pc was nearly uniform, which suggests that the strength of the intermolecular interactions between C₂H₅OH and the reduced form (FcCH₂OH) is nearly equal to that between C₂H₅OH and the oxidized form (Fc⁺CH₂OH). The matrix effect in case of using the electrochemical label was further examined with freshly prepared liquor samples. As shown in Fig. S8, along with the increase of alcohol content, E_pa and E_pc are shifting to the positive potential. This phenomenon is similar to that observed for the artificial model liquors, suggesting that the CV peak potentials are correlated to the alcohol content in the liquors. In consideration of the intermolecular interactions, we further optimized the conformation of FcCH₂OH and C₂H₅OH by computational methods²⁰ (Calculations using density functional theory). As seen in Fig. 3c, an H-bonding interaction between the O atom of C₂H₅OH and the H atom of FcCH₂OH was observed, which was similar to the previous reports.²⁰ Julian Vrbancich et al. have demonstrated that the positive charge of the oxidized form (Fc⁺CH₂OH) locates at the iron atom of the ferrocene ring,²⁸ thus it can be suggested that the net positive charge on iron atom will not affect the H-bonding between the H atom of Fc⁺CH₂OH and the O atom of C₂H₅OH, due to the large distance between the H-bond and the positively charged iron. Therefore, it is reasonable to suggest that the H-bonding interaction between C₂H₅OH and the reduced form (FcCH₂OH) is nearly equal to that between C₂H₅OH and the oxidized form (Fc⁺CH₂OH). The redox process of FcCH₂OH in the model liquor is summarized in Fig. 3d: The presence of hydrogen bonding leads to the formation of the complex (FcCH₂OH·C₂H₅OH or Fc⁺CH₂OH·C₂H₅OH ), a more stable redox state than FcCH₂OH/Fc⁺CH₂OH, resulting in large positive shifts in the redox potential.^29,30 The formal potential, E_1/2 = 1/2(E_pa+E_pc), which is a good indicator of the energy barrier of redox reaction,³¹ are plotted vs volume ratio of C₂H₅OH. As shown in the inset of Fig. 3b, a positive correlation between E_1/2 and volume ratio of C₂H₅OH is observed, which is consistent with the proposed redox process shown in Fig. 3d.

Zoom In Zoom Out Reset image size

Now we are aware that alcohol content difference in the liquors affects both E_pa and E_pc of the CV with FcCH₂OH redox species. In addition, we have measured the alcohol content of the liquors from different aging times by an Abbe refractometer and the result shows that the alcohol content of the liquors is within the range from 63.0% to 64.5%, in volume ratio. No obvious alcohol content difference is observed for the liquors. Thus we confirm that the negative shift in E_pa, along with the increase of the aging times of the liquors (as shown in Fig. 2b), is not due to the effect caused by alcohol content difference.

As we previously reported, the redox species FcCH₂OH/Fc⁺CH₂OH benefits the formation of hydrogen bonding^18,32,33 or electrostatic attraction³⁴ because of the evident distinction of electric charge between the reduced form (FcCH₂OH) and the oxidized form (Fc⁺CH₂OH). We next investigated the influence of the common liquor compounds on the redox potential of FcCH₂OH. The model liquor is obtained by mixing water and alcohol, and the final alcohol content is 64% in order to keep the same alcohol content with the real liquor samples. CV of model liquors containing 0.5 mM FcCH₂OH and different liquor compounds are shown in Fig. 4a. As seen, addition of 50 mM (100 equiv) n-butyric acid, 50 mM propionic acid and 50 mM acetic acid in model liquors leads to the positive potential shift of E_pa in 10 ± 3 mV, 38 ± 4 mV and 43 ± 6 mV respectively, while E_pc almost remains a constant. To understand this phenomenon, we have compared the potential shifts caused by addition of other liquor compounds, the effect of these compounds on the redox potential of FcCH₂OH is outlined in Table I. As seen, addition of 50 mM of the carboxylic acids, including C2, C3, C4, C5, and C6, leads to the potential shift of E_pa, and the shift extent is positively related to the ionization degree of these acids. The results suggest that the intermolecular interactions between the carboxylic acid and FcCH₂OH have a significant impact on the redox potential of FcCH₂OH. Further, we compared the potential shifts by addition of carboxylate and carboxylic acid. As seen, addition of 2 equiv of acetic acid do not cause potential shift while addition of 2 equiv of sodium acetate leads to obvious potential shift (+40 ± 5 mV). This allows us to understand the mechanisms that lead to potential shift of E_pa, as represented in Fig. 4b: The negatively charged carboxylate in the liquors, selectively interacts with the positively charged Fc⁺CH₂OH (formation ion-pairs), resulting in large positive shifts of E_pa. It should be noted that, as the negatively charged carboxylate does not interact with the uncharged FcCH₂OH, addition of carboxylate does not cause the shift of E_pc. As a control group, CV of model liquors containing 0.5 mM FcCH₂OH and uncharged compounds (glyoxal and ethyl acetate) were also performed. As shown in Table I, addition of uncharged compounds does not cause obvious shift in E_pc or E_pc which verified the proposed ion-pair mechanism. As some other negatively charged ions may be present in liquor as additives, the interactions between the electrochemical label and these additives were taken into account. Sodium citrate and sodium malate are commonly known as sour salts and are mainly used as food additives, usually for flavors or as preservatives.³⁵ As seen, the potential shifts for sodium citrate (2 equiv) and sodium malate (2 equiv) are +38 ± 4 mV, +42 ± 5 mV respectively (shown in Table I). The shifts are nearly equal to that for sodium acetate. Besides, addition of the mixture of sodium acetate (1 equiv) and sodium citrate (1 equiv) leads to obvious potential shift (+40 ± 4 mV). The results suggest that the potential shift of the electrochemical label is correlated with the total content of the negatively charged ions present in the liquor.

Zoom In Zoom Out Reset image size

Liquor is an alcoholic beverage consisting of hundreds of components that can have an impact on its quality and flavor. As far as the chemistry of liquor is concerned, there is a close link between liquor aging time and liquor flavor,³⁶ it is important to understand how liquor aging time affect liquor chemistry. The constituent analysis of liquor samples with different aging times (1, 2, 3, 4, 5, 10 years) was performed by gas chromatography, as shown in Table SI in supporting information. As seen, the main constituents of the liquor samples are similar: saturated ethanols, short chain carboxylic acids and esters. The chromatography characterization enables us to confirm the main constituents in the liquors; however, it is not likely to provide further details in regard to the discrimination of aged liquor.

As mentioned above, the negatively charged carboxylate in the liquors, selectively interacts with the positively charged Fc⁺CH₂OH (formation ion-pairs), resulting in large potential shifts of E_pa. Thus, we further investigated the correlation between redox potentials and the carboxylic acid contents in the liquors. The total content of C2, C3, and C4 was plotted vs the E_pa in the CV of liquors from different aging times, as shown in Fig. 5 (in black dotted line). As seen, a positive correlation between E_pa and the total carboxylic acid contents (C2 + C3 + C4) is observed. Other combinations of the carboxylic acid contents were also examined, but they did not bring additional interest in the correlation between E_pa and the total carboxylic acid contents. For instance, the total content of C2, C3, C4, C5, and C6 was plotted vs E_pa, as shown in Fig. 5 (in red dotted line), the total content towards a random variation.

Zoom In Zoom Out Reset image size

However, the result is reasonable as the short chain compounds tend to become long chain compounds in the liquor fermentation process: As shown in Table SI, the content of short chain carboxylic acids (e.g., acetic acid) in 1 year samples (135.7 mg/100 ml) is significantly higher than that in 10 year samples (50.2 mg/100 ml). On the contrary, the content of the relatively long chain carboxylic acids (e.g., n-hexylic acid) in 1 year samples (58.5 mg/100 ml) is significantly lower than that in 10 year samples (120.6 mg/100 ml). In other words, the result illuminates the certain compound differences within the samples from various aging time and unveils the electrochemistry behind the aged liquor: the potential shift in E_pa, that observed in the cyclic voltammograms (shown in Fig. 2b), can be attributed to the decrease in the total content of the relatively short chain carboxylic acids (e.g., C2 + C3 + C4) in the liquor fermentation process. And, clearly, an alternative approach for differentiating the aging time of typical Chinese liquor with the electrochemical label, FcCH₂OH, has been developed in this study.