Introduction

Vinegar and soy sauce were important food additives which effectively improved the flavor and color quality of food products (Kim et al. 2004; Tesfaye et al. 2002). A great number of vinegar and soy sauce were consumed in daily life of Asian countries such as China and Japanese. Usually, the colorants like caramel color were added into the vinegar and soy sauce to improve the color quality (Theobald et al. 1998; Stute et al. 2002). However, the addition of caramel color in vinegar and soy sauce induced the presence of 5-hydroxymethylfurfural (HMF), which was a potential carcinogenic substance (Abraham et al. 2011) and may have a negative influence on gene. In addition, HMF may be produced during the process of vinegar and soy sauce microbiological fermentation or chemical synthesis and storage (Theobald et al. 1998). However, there were little report about HMF in soy sauce and vinegar. Therefore, it is necessary to analyze the content of HMF in vinegar and soy sauce.

Up to now, lots of methods were used to analyze HMF in honey, fruit juices, coffee, wine, and so on (Andrade et al. 2016). The analytical methodology involved high-performance liquid chromatography (HPLC) coupled with a diode array detector (DAD) (Abu-Bakar et al. 2014; Rovira et al. 1993) or mass spectrum (MS) detector (Takino et al. 2003) and gas chromatography (GC) (Gaspar and Lopes et al. 2009). Although the use of chromatography methods (like GC or HPLC) for HMF analysis was accurate and stable, the operation cost of chromatography methods was the main concern including the consumption of analytical column, solvent, and sample clean-up supply. A few years ago, the relatively economic method was developed for HMF analysis in honey sample on the basis of micellar electro kinetic capillary chromatography technique (MEKC) (Teixidó et al. 2011). The previous studies indicated that CE involved low solvent consumption and offered high separation efficiency (Acunha et al. 2016), being an alternative to HPLC in the analysis of residues in different fields (Tejada-Casado et al. 2016). Nevertheless, the separation of HMF on CE system was readily interfered by the complex matrix with high ion strength like acid and salt in food samples (such as vinegar and soy sauce). Thus, the sample pretreatment process is necessary prior to the CE analysis for high ion strength food samples, especially for vinegar and soy sauce samples.

Generally, solid phase extraction (SPE) and liquid–liquid extraction (LLE) were applied to eliminate the ionic compounds from samples. However, both of the cleanup methods were time-consuming or solvent-consuming. Especially, SPE was relatively expensive, and the low enrichment factors were obtained without solvent evaporation. In recent years, the miniaturization extraction technique has been developed as rapid sample pretreatment methods like solid phase micro-extraction (SPME) and liquid–liquid micro-extraction (LLME). Because of the limited amount of solid absorbent in expensive SPME, LLME as a promising pretreatment method was inexpensive and environmentally friendly. In addition, the diversified operation modes were developed for LLME application such as single-drop micro-extraction (SDME) that could be achieved on-line injection (ALOthman et al. 2012), dispersive liquid–liquid micro-extraction(DLLME) (Chandrasekaran et al. 2011), and hollow fiber–liquid phase micro-extraction (HF-LPME) (María Ramos Payán et al. 2009). Furthermore, kinds of assisted energy fields like vortex (Yiantzi et al. 2010), ultrasonic (Sereshti et al. 2011), and microwave (Ahmadi-Jouibari and Fattahi 2015) were attempted to be introduced into LLME for improving the extraction efficiency. Unfortunately, the organic phase from LLME was not compatible with direct CE analysis. Though a solvent evaporation after LLME was suitable to avoid the organic solvent interference in CE analysis, the operation duration including vacuum evaporation and aqueous redissolution was time-consuming and tedious. Using the aqueous phase to back extract, the organic phase after LLME could be an applicable and handy technique for CE determination HMF.

In order to detect HMF in high ionic strength samples like vinegar and soy sauce, the two-step ultrasonic assisted liquid–liquid micro-extraction was applied to reduce the ionic effecting on the analysis of HMF by capillary electrophoresis-ultraviolet (CE-UV). In this study, the optimization of experimental parameters for the second step UALLME was carried out, and the method performances were evaluated carefully.

Materials and Methods

Chemicals and Solutions

HMF was purchased from Sigma-Aldrich (St. Louis, MO, U.S.A). Trichloromethane (CHCl3), sodium hydroxide, acetic acid, and hydrochloric acid of analytical grade were provided by Sigma-Aldrich (St. Louis, MO, USA). All vinegar and soy sauce were bought from a local supermarket in Da-Lian city of Liaoning province, PR China. Deionized water was used in the whole experiment.

HCl of 1 M and 0.1 M NaOH were prepared for CE, and 1% of acetic acid solution was used else. All solutions were filtered by inorganic membrane (0.22 μm). Stock solution of HMF (2000 mg/L) was prepared in acetic acid solution, stored at 4 °C until analysis.

UALLME Procedure

The process of two-step UALLME was carried out as follows. First, the vinegar or soy sauce was diluted in deionized water (1:4, v/v), then added 500 μL CHCl3 (accepted phase) to extract the analyte under the ultrasonic condition. In the first step of UALLME, the optimal condition was carried out as follows: the ultrasonic extractor (SCIENTZ-II D, Shenyang, China) was applied, and the ultrasonic power density was 168 W/cm2 at 12 °C (the temperature was maintained by a circulating water bath) for 4 min. In the second step of UALLME, 150-μL accepted phases containing HMF were added into acetic acid solution (aqueous phase) and extracted by the UALLME method. The aqueous phase (30 μL) could be directly analyzed using CE-UV.

CE-UV Analysis of HMF

The CESI 8000 Plus provided by BECKMAN COULTER (USA) coupled with UV was used for the separation and determination of HMF. The capillary was coated with silica-tubing (50 cm effective length, 75 μm i.d., 375 μm o.d.). The new capillary was conditioned by rinsing with 1 M HCl (5 min), followed by deionized water (2 min) and 0.1 M NaOH (10 min), then by deionized water (2 min) and 10% acetic acid (5 min). The whole process of condition was operated under 20 psi at 25 °C (forward). The sample injection time, injection press, and separation voltage were 10 s, 5 psi, and 25 kV, respectively. The detection wavelength was set at 280 nm. The capillary was rinsed by 0.1 M NaOH (5 min), water (2 min), and 10% acetic acid (5 min) under 20 psi at 25 °C. Sample was injected to the instrument with a 10-μL loop.

Statistical Analysis

Mean and standard deviation of the data were calculated for each treatment. Analysis of variance was carried out to determine any significant differences (p < 0.05) among the applied treatments by the SPSS software package (SPSS 10.0 for Windows).

Results and Discussion

Optimization of UALLLME Conditions

In this study, the two-step UALLME method was applied in eliminating the ionic compounds from vinegar and soy sauce for compatibility CE analysis. The high density organic solvent including CHCl3, CCl4, and CS2 was selected as the extractant screened through the property of solvent like density, vapor pressure, boiling point, and solubility in aqueous solution. Unlike the dispersive liquid–liquid micro-extraction method, the disperse solvent was absent under UALLME operation due to a micro-drop formation during ultrasonic cavitation (Tadeo et al. 2010). In our previous study, the excellent extraction efficiency for HMF in the first step of UALLME could be obtained by using CHCl3 as extractant (Ting-Ting et al. 2017). Meanwhile, the optimum conditions (the power density of ultrasound: 168 W/cm2, volume ratio of sample to extraction solvent: 25:1, ultrasound time: 4 min) at room temperature were applicable for extracting HMF into organic solvent.

The optimization of the back extraction of HMF into aqueous solution by UALLME method based on CHCl3 solvent was studied at 24-mg/L HMF standard solutions. The extracted organic solution (150 μL) was further extracted by the 10% acetic acid solution. As shown in Fig. 1, the extraction parameters including the aqueous phase volume (extraction volume), the power density of ultrasonic, and extraction time significantly affected the extraction efficiency. For example, the increase of back extraction solvent (10% acetic acid solution) volume was performed from 20 to 50 μL; the back-extracted HMF level was increased at 30 μL back extraction solvent but lowered at 50 μL back extraction solvent. In addition, the power density of ultrasonic was very important to obtain an excellent back-extraction of HMF. In this study, the power density of ultrasonic including 0.14, 0.17, and 0.21 W/cm2 was applied to improve the back-extraction efficiency. Obviously, the 0.17 W/cm2 of ultrasonic power density was suitable for HMF back extraction, while a higher ultrasonic power density like 0.21 W/cm2 decreased the level of HMF back extraction. The result indicated that the increase of HMF back-extraction could be induced by the increase of the miscibility between water and organic solvent by ultrasonic cavitation, while the too high intensity of ultrasound could increase the miscibility between water and organic solvent and induce the decrease of HMF back-extraction.

Fig. 1
figure 1

Optimum condition of the second extraction step: extraction volume (the experiment parameters of extraction volume were 20, 30, and 50 μL, the organic accepted phase volume was 150 μL, the power density of ultrasonic was 0.17 W/cm2, the ultrasound time was 2 min), the power density of ultrasonic (the parameters of ultrasonic power density were 0.14, 0.17, and 0.21 W/cm2, the organic accepted phase volume was 150 μL, the extraction volume was 30 μL, the ultrasound time was 2 min), the ultrasound time (the parameters of ultrasound time were 1, 2, 3, 4, and 5 min, the organic accepted phase volume was 150 μL, the power density of ultrasonic was 0.17 W/cm2, the extraction volume was 30 μL), each experiment was replicated three times

Full size image

In addition, the extraction time was an important factor on HMF extraction into aqueous solution. For example, the increase of extraction time from 1 to 4 min induced a significant improvement of HMF back extraction. However, a prolonged ultrasonic treatment induced the decrease of HMF extraction. The phenomenon could be explained that the long time ultrasonic treatment could induce the strong emulsification between water and organic solvent, which had a bad effect on the phase separation and induced the extraction efficiency decrease.

CE-UV Analysis of HMF

CE involves the high separation efficiency, being an optional substitute for liquid chromatography (LC) in separation of analytes with small sample volume. Therefore, this study established the two-step UALLME coupled with CE-UV method (simple, convenient) to detect the HMF in vinegar and soy sauce.

The HMF separation was performed on capillary tube and detected by the characteristic UV absorption at 280 nm; the quantification and identification were realized by the CE-UV. The results are shown in Fig. 2; HMF was eluted out at 13.5 min. All peaks were confirmed by comparing the retention times to the standard solution. In addition, the UALLME method significantly improved the method selectivity. As shown in Fig. 2, the samples including vinegar and soy sauce without two-step UALLME pretreatment were directly injected into CE system. Obviously, the 5-HMF peak was totally interfered by the impurities in soy sauce (Fig. 2a) and vinegar (Fig. 2b). Thus, the HMF in the unclean-up sample was not suitable to be analyzed by the CE-UV. Moreover, no HMF peak appeared in the extracts containing organic solvent, which was also not applicable for direct analysis of HMF in CE-UV (Fig. 2c). As a result, the organic solvent should be removed before CE-UV analysis. In this study, the second step UALLME was applied to back extract the HMF into aqueous solution in order to remove the organic solvent. Besides, compared with the previous method (Bignardi et al. 2014), the mass spectrometer was coupled to CE system for improving the selectivity of HMF analysis. However, the two-step UALLME method in this study was very effective to improve the method selectivity in the vinegar and soy sauce. The results are shown in Fig. 3. Clearly, the HMF peaks in vinegar (D) or soy sauce (B) were symmetrical and totally separated from the impurities. To confirm the matrix effect in vinegar and soy sauce on the separation of HMF, the spiked HMF standards were added into vinegar (Fig. 3c), soy sauce (Fig. 3a) during the process of two-step UALLME treatment. As shown in Fig. 3, nearly no shoulder peaks appeared, and the increased peak areas are listed as recoveries in Table 2. Therefore, the two-step UALLME method was important to improve the method selectivity.

Fig. 2
figure 2

Samples directly analyzed by capillary electrophoresis-ultraviolet (CE-UV). a Soy sauce filtered by organic membrane (0.22 μm). b Vinegar filtered by organic membrane (0.22 μm). c Vinegar extracted into organic accepted phase by the first step ultrasonic assisted liquid–liquid micro-extraction. d 5-Hydroxymethylfurfural (HMF) standard (8 mg/L)

Full size image
Fig. 3
figure 3

Samples with two-step ultrasonic assisted liquid–liquid micro-extraction pretreatment analyzed by CE-UV. a Soy sauce spiked with HMF standard (8 mg/L). b Soy sauce. c Vinegar spiked with HMF standard (8 mg/L). d Vinegar

Full size image

Method Validation

The validation parameters involving linearity, correlation coefficient (R 2), limit of detection (LOD), limit of quantitation (LOQ), and recovery were tested under the optimum condition of the two-step UALLME method.

As shown in Table 1, the correlation coefficient of standard curves in vinegar and soy sauce matrix was 0.9918 and 0.9942, respectively. The calibration standards were prepared at concentrations of 0.50, 1.00, 2.00, 4.00, 8.00, 16.00, 24.00, and 32.00 mg/L to generate the calibration curves. In addition, the LOD and LOQ were calculated as the signal to noise ratio of 3 and 10, to evaluate the performance of CE-UV instruments. The values of LOD and LOQ in both vinegar and soy sauce sample were 0.03 and 0.10 mg/L, respectively. Significantly, the LOD and LOQ in this study were comparable with the previous method such as CE-MS (LOD = 0.03 mg/L and LOQ = 0.10 mg/L) (Bignardi et al. 2014). In addition, the method in this study was more sensitive for HMF determination with respect to the micellar electrokinetic capillary chromatography (MEKC) methodology (LOD = 0.09 mg/L and LOQ = 0.30 mg/L) (Rizelio et al. 2012). Actually, the two-step UALLME method for HMF pre-concentration could be obtained in this study.

Table 1 Calibration range, regression equation, limit of detection (LOD), and limit of quantification (LOQ) for 5-hydroxymethylfurfural(HMF)
Full size table

In addition, the addition of 8.00, 16.00, and 24.00 mg/L HMF standards into the vinegar and soy sauce sample was carried out to evaluate the accuracy calculated by recovery, while the relative standard deviation (%RSD, n = 3) was tested to evaluate the stability of method. As shown in Table 2, the observed RSD % was ranging from 0.53 to 3.74, which indicated that the method stability was satisfactory for HMF analysis in vinegar and soy sauce sample. The recovery values of HMF, calculated as the ratio of the tested level to spiked level, were ranging from 91.24 to 109.39%. The results indicated that the tested HMF contents in samples by CE-UV after UALLME pretreatment were reliable and available.

Table 2 Recoveries with the relative standard deviations (%RSD) of 5-hydroxymethlfurfural (HMF) in vinegar and soy sauce
Full size table

Analysis of HMF in Vinegar and Soy Sauce

Although HMF was presented in vinegar and soy sauce, there was little literature available up to now. Moreover, the complex matrix in vinegar and soy sauce such as colored components and salts could interfere with the analyte separation and detection (Fig. 2). By the two-step UALLME method, the influence factors could be mostly eliminated to improve the accuracy and flexibility, and the analysis could be practiced in a stabilized condition. In this study, the content of HMF was measured under the optimum conditions. As shown in Table 3, the levels of HMF in vinegar and soy sauce depended on the different brands. HMF could not be detected in some brands of vinegar such as vinegar 1 and vinegar 6. The levels of HMF in vinegars without caramel color (vinegars 3 and 4, 22.92 and 17.54 mg/L, respectively.) were higher than those in vinegars with addition of caramel color (vinegars 2 and 5, 7.66 and 9.01 mg/L). The results suggested that HMF in vinegar could be formed during the process of fermentation (Kowalski et al. 2013) and the addition of caramel colorants induced the increase of HMF in vinegar (Hewala et al. 1993). In addition, the contents of HMF in soy sauce 2 (produced by solid-state fermentation) were 17.68 mg/L, which was lower than that of HMF in soy sauces containing caramel color(soy sauces 1 and 3, 4.29 and 5.56 mg/L, respectively). In sum, the addition of caramel color could be an important factor for the increase of HMF in vinegar or soy sauce (Eixidó et al. 2011). Thus, the two-step UALLME pretreatment was applicable for the HMF determination in CE-UV in food samples with high ion strength.

Table 3 Concentration of 5-hydroxymethylfurfural (HMF) in vinegar and soy sauce
Full size table

Conclusion

The HMF level in vinegar and soy sauce after two-step UALLME pretreatment could be analyzed by CE-UV. Moreover, excellent linearity, recovery, and repeatability were obtained in the method. Therefore, two-step UALLME coupled with CE-UV was an alternative option for accurate, rapid, and effective detection of HMF in vinegar and soy sauce.