View PDF Version
 
DOI: 10.1039/D5AY00420A (Paper) Anal. Methods, 2025, Advance Article

Rapid identification of high-viscosity rice soup quality using neutral desorption extractive electrospray ionization mass spectrometry†

Manman Qina, Leting Wangb, Kaixin Xub, Zhehao Liangb, Chao Zhonga, Yanfei Xiea, Dan Shanc, Jun Yu*c, Jiang Wang*d and Hao Fan*e
aCenter for Translational Medicine, College of Traditional Chinese Medicine, Jiangxi University of Chinese Medicine, Nanchang, 330004, China
bThe Jiangxi Province Key Laboratory for Diagnosis, Treatment, and Rehabilitation of Cancer in Chinese Medicine, Jiangxi University of Chinese Medicine, Nanchang, 330004, China
cDepartment of Cardiovascular Sciences and Centre for Metabolic Disease Research, Lewis Katz School of Medicine, Temple University, Philadelphia, PA, USA. E-mail: jun.yu@temple.edu
dAcademician Workstation, Jiangxi University of Chinese Medicine, Nanchang, 330004, China. E-mail: wj0121@yeah.net
eDepartment of Pharmacy, Jiangxi University of Chinese Medicine, Nanchang, 330004, China. E-mail: fanhao11@aliyun.com

Received 13th March 2025 , Accepted 1st July 2025

First published on 4th August 2025

Abstract

High-viscosity rice soup is not only a popular food item but also a traditional Chinese medicinal remedy. Herein, we developed a high-throughput characterization platform for high-viscosity rice soup analysis using Neutral Desorption Extractive Electrospray Ionization Mass Spectrometry (ND-EESI-MS) with untargeted metabolomics. The chemical fingerprint of high-viscosity rice soup was directly analyzed by ND-EESI-MS, with exogenous (e.g., heavy metals and pesticides) and endogenous pollutants detected via qualitative and quantitative tandem mass spectrometry. Multivariate statistical analysis was then applied to distinguish the quality of the high-viscosity rice soup effectively. The concentrations of heavy metals and pesticide residues in high-viscosity rice soup were quantified, demonstrating a strong linear relationship (R2 > 0.99) across the 0.05–2000.00 ng mL−1 concentration range. The limits of detection (LOD) ranged from 0.27 ng mL−1 to 2.18 ng mL−1, while the limits of quantification (LOQ) ranged from 0.82 ng mL−1 to 6.61 ng mL−1. Furthermore, three specific compounds—3-methyl-2-butenoic acid, salicylic acid, and rhamnose—were quantified. These compounds demonstrated excellent linearity (R2 > 0.99) within the linear range of 1.10–1500.00 ng mL−1. The LOD ranged from 0.56 to 1.30 ng mL−1, and the LOQ ranged from 1.71 to 3.95 ng mL−1. Owing to its capability for simultaneous detection of exogenous and endogenous contaminants in complex viscous matrices, this ND-EESI-MS platform demonstrated significant potential to enhance quality control in food safety surveillance and pharmaceutical manufacturing.

1. Introduction

Rice soup, also referred to as congee oil, is a concentrated layer that forms when congee undergoes prolonged simmering. This traditional food is especially popular in China and parts of Southeast Asia. Additionally, cities in southwestern China sell both rice soup and sour rice soup as snacks. In fact, beyond being a foodstuff, high-viscosity rice soup is recognized as a traditional Chinese medicinal remedy. Shixiong Wang, a renowned physician from the Qing Dynasty, regarded high-viscosity rice soup as a viable alternative to ginseng soup, citing a comparable tonic effect.1 Contemporary medical research has confirmed the presence of essential nutrients, including proteins, fats, starch, and vitamins, in rice soup.2 As a result, rice soup consumption possesses notable tonic properties. For instance, in the treatment of diarrhea in children, the concomitant administration of glucose oral rehydration solution (G-ORS) and rice soup was found to be more effective than G-ORS alone.3 Concurrently, Yi Cao et al. demonstrated that rice soup in combination with parenteral nutrition (PN) may be a safe and effective approach for treating congenital chylous ascites.4 Moreover, rice soup could be used as an adjuvant to enhance the aroma and efficacy of various Chinese medicines.5,6 The incorporation of rice soup during the preparation of curcuma could increase its viscosity, thus reducing the likelihood of fracturing during slicing.5 Xin Dong et al. observed that adding rice soup to Gastrodia elata Bl. increased the concentration of its primary active ingredients and enhanced its aroma compared with Gastrodia elata Bl. prepared with traditional steaming.6 Chunli Luo et al. compared rice soup-roasted Codonopsis pilosula and traditional rice-fried Codonopsis pilosula. The results show that the indicator component content in the rice soup-roasted sample is significantly higher than that in the rice-fried sample.7

The quality of rice soup is significantly influenced by various factors, including the quality of the rice and the purity of the drinking water. Moreover, endogenous and exogenous pollutants from aged rice compromise its nutritional value. The aging of rice during storage can lead to a decline in its nutritional content.8,9 Additionally, the environmental storage conditions affect the quality of rice. Under natural storage conditions, rice moisture content fluctuates dynamically, whereas fatty acid content rises proportionally with temperature. This elevated fatty acid level accelerates mold proliferation on rice grains, potentially causing detrimental health effects.10,11 Apart from endogenous pollutants, rice is also vulnerable to exogenous pollutants. The misuse of pesticides during cultivation leads to residue accumulation in grains, thus posing significant risks to food quality and safety.12,13 Moreover, industrialization and urbanization have significantly increased heavy metal residues in soil and water sources.14,15 This suggests that rice grown in these regions may be contaminated with heavy metals. Among the 45 known heavy metals, copper, lead, arsenic, chromium, mercury, and cadmium are recognized for their relatively high biotoxicity.16–21 These heavy metals deposit readily in soil, are absorbed and accumulated by crops, and ultimately enter the human body through the food chain, posing significant health risks.22,23 Consequently, detecting these pollutants and their metabolites is essential.

Currently, analytical techniques used for pollutant detection in rice soup include high performance liquid chromatography (HPLC),24–26 gas chromatography (GC),27,28 gas chromatography–tandem mass spectrometry (GC-MS/MS)29,30 and high performance liquid chromatography–tandem mass spectrometry (HPLC-MS/MS).31–34 However, these methods require complex sample pretreatment processes. For samples with complex matrices and high viscosity (e.g., rice soup), dilution is commonly employed during pretreatment. This step, however, may compromise the precision of the analytical results. Consequently, developing a rapid in situ analytical approach for highly viscous complex samples is imperative.

Neutral Desorption Extractive Electrospray Ionization Mass Spectrometry (ND-EESI-MS) is a rapid mass spectrometry technique that enables direct analysis of viscous samples with complex matrices. The ND-EESI-MS method employs nitrogen gas to carry desorption reagents, which desorb target molecules from the samples to form an aerosol. The aerosol is subsequently sampled and ionized in the EESI source, producing ions that undergo mass spectrometric analysis. The technique has been successfully applied to analyze various viscous matrices, including honey, cheese, toothpaste, and cosmetics.35–39

This study achieved the first simultaneous detection of exogenous pollutants, endogenous pollutants, and inherent components in viscous rice soup through integration of ND-EESI-MS with untargeted metabolomics. By incorporating ethylenediaminetetraacetic acid disodium salt (EDTA-2Na) into the extraction solvent, we accomplished online heavy metal extraction, with extracted species analyzed via organic mass spectrometry. These advancements established a rapid detection method for organic and inorganic contaminants in viscous rice soup, and a high-throughput analytical platform for quality assessment of rice soup (Fig. 1a).


image file: d5ay00420a-f1.tif
Fig. 1 Study design. (a) Workflow diagram of the study design; (b) schematic diagram of the ND-EESI-MS ion source.

2. Materials and methods

2.1. Chemicals and materials

Rice was purchased from local supermarkets (Nanchang, Jiangxi Province, China). The pesticides (carbendazim, molinate, pirimiphos-methyl and cartap) were bought from Aladdin (Shanghai, China). Heavy metal standard solutions (copper, lead and cadmium) were bought from the National Center for Analysis and Detection of Nonferrous Metals and Electronic Materials. Chromium trichloride was bought from Aladdin (Shanghai, China). Methanol, ethanol and acetic acid (all HPLC-grade) were purchased from ROE Scientific Inc. (Newark, DE, USA). Ethylenediaminetetraacetic acid disodium salt (EDTA-2Na) (purity ≥ 99%) was bought from Rhawn (Shanghai, China). 3-Methyl-2-butenoic acid (purity ≥ 98%), salicylic acid (purity ≥ 98%) and rhamnose (purity ≥ 98%) were bought from Sichuan Weikeqi Biological Technology Co., Ltd. (Chengdu, Sichuan Province, China). Ultrapure water (resistivity: 18.2 MΩ cm) was prepared using a Hetai Smart-DUVF water purifier (Shanghai, China).

A micro-injection pump was obtained from Baoding Lange Constant Flow Pump Co., Ltd. (Baoding, Hebei Province, China). An ultrasonic cleaner (Model KQ3200DE) was procured from Kunshan Ultrasonic Instrument Co., Ltd. (Kunshan, Jiangsu Province, China). An Electronic balance (Model YHM-10001) was obtained from Huizhou Yingheng Electronic Technology Co., Ltd. (Huizhou, Guangdong Province, China). An analytical balance (Model BSA124S) was obtained from Sartorius Scientific Instruments Co., Ltd. (Beijing, China). A smart-DUVF ultrapure water system was provided by Shanghai Hetai Instrument Co., Ltd. (Shanghai, China). A linear ion trap mass spectrometer (Model LTQ-XL) was from Thermo Fischer Scientific Co., Ltd. (San Jose, CA, USA). The mass spectrometry data were collected using Thermo Xcalibur Roadmap 2.0 software. Body-vision microscopes were purchased from Leica (Germany).

2.2. Preparation of samples and standard solutions

2.2.1 Preparation of high-viscosity rice soup samples. Rice (50.00 g) was rinsed once and placed in a cooking pot. Deionized water (1000 mL) was added, and the mixture was boiled at 100 °C for 20 min. After cooling to room temperature until a surface oil layer formed, the soup was filtered to remove solids. Aliquots (50 mL) were stored in centrifuge tubes at −20 °C for further use.
2.2.2 Preparation of aged rice soup samples. The rice samples were separated into four portions of approximately 100 g each and placed within a gauze bag. A simulated accelerated aging test was conducted in a constant temperature and humidity chamber, which was set to replicate the climatic conditions of a typical high-temperature and high-humidity region in China (40 °C, 80% relative humidity).40 Once the parameters had been established, the samples were transferred to the constant temperature and humidity chamber and sampled at seven-day intervals (0, 7, 14, 21 and 28 days) to assess the degree of aging. The application of different aging times was employed to indicate the degree of aging (Table 1).
Table 1 Rice soup sample information
No. Product name Specification (mL) Color Preparation method
1 Rice soup 5 Ivory-white 40 °C, RH 80% (7 days, 14 days, 21 days, 28 days)
2 Rice soup 5 Ivory-white Carbendazim standard solution (50.00–2000.00 ng mL−1)
Molinate standard solution (10.00–1000.00 ng mL−1)
Pirimiphos-methyl standard solution (10.00–1500.00 ng mL−1)
Cartap standard solution (5.00–1000.00 ng mL−1)
3 Rice soup 5 Ivory-white Copper standard solution (10.00–1500.00 ng mL−1)
Lead standard solution (0.50–100.00 ng mL−1)
Chromium standard solution (5.00–500.00 ng mL−1)
Cadmium standard solution (0.05–5.00 ng mL−1)

2.2.3 Preparation of standard solutions for characteristic components. Stock standard solutions of 3-methyl-2-butenoic acid (110.00 μg mL−1), salicylic acid (130.00 μg mL−1) and rhamnose (150.00 μg mL−1) were prepared by dissolving 1.1 mg, 1.3 mg and 1.5 mg of the respective compounds in separate 10 mL volumetric flasks with methanol. Working standard solutions (1.00–2000.00 ng mL−1) were prepared by serial dilution of the stock solutions with rice soup (Table S1†). All solutions were stored at 4 °C. Detailed procedures are provided in the ESI.†

To evaluate ionization suppression or enhancement effects induced by the rice soup matrix, the aforementioned stock solutions (110.00, 130.00, and 150.00 μg mL−1) were serially diluted with methanol to achieve concentrations matching those of the working standard solutions (1.00–2000.00 ng mL−1). These solutions were similarly stored at 4 °C.

2.2.4 Preparation of standard solutions for pesticides. 2 mL of 10.00 μg mL−1 carbendazim, 100 μL of 1.00 mg mL−1 molinate, 1.5 mL of 10.00 μg mL−1 pirimiphos-methyl and 1 mL of 10.00 μg mL−1 cartap, respectively, were diluted with rice soup using the gradient dilution method to obtain working solutions with concentrations ranging from 5.00–2000.00 ng mL−1 (Table S1†). For matrix effect assessment, the same concentration range was prepared by diluting the same stock solutions with methanol. All solutions were stored at 4 °C prior to analysis, with complete preparation details documented in the ESI.†
2.2.5 Preparation of standard solutions for heavy metals. Single-element stock solutions (10.00 μg mL−1) of copper, lead, chromium and cadmium were separately prepared by diluting 100 μL of 1000.00 μg mL−1 standard solutions to 10 mL with rice soup in volumetric flasks. Serial dilution with rice soup generated working solutions spanning 0.05–1500.00 ng mL−1 for each metal (Table S1†). The same concentration ranges were prepared in methanol from stock solutions for matrix effect evaluation. All solutions were stored at 4 °C prior to analysis, with complete procedures detailed in the ESI.†

2.3. ND-EESI-MS analysis

The ND-EESI experiment was conducted using a homemade neutral desorption electrospray ion source. As shown in Fig. 1b, this device consisted of a container filled with neutral desorption solution, a container holding the sample, a spray pipeline and an electrospray channel. When in use, nitrogen gas was first introduced and passed through the neutral desorption solution and the sample solution sequentially via the pipeline, and was then sprayed out through the spray pipeline to undergo extraction ionization with the charged reagent sprayed out from the electrospray section. Moreover, the mass scanning range was m/z 50–800; ionization voltage, 4 kV; ion transfer tube temperature, 200 °C; capillary voltage, 1 V; lens voltage, 10 V; extraction reagent, methanol; flow rate, 10 μL min−1; parent ion isolation width, 1.0–2.0 u; collision energy, 15–40%; activation value Q, 0.25; collision time, 100 ms; and other detection parameters were automatically optimized by an LTQ-Tune system. Background subtraction was a crucial step in mass spectrometric data analysis that allowed for the identification and removal of ions originating from the background or sample matrix. This technique can be applied to full-scan datasets without making any assumptions about the behavior or properties of the ions of interest.41 Under the same conditions, the mass spectrometry signal without adding samples was considered the background signal. It should be noted that all mass spectrometry results must undergo background signal subtraction.

2.4. Data analysis

This study employed ND-EESI-MS and untargeted metabolomics analysis to investigate the impact of exogenous factors such as heavy metals and pesticides, as well as endogenous substances (e.g., aging degree), on rice soup quality. ND-EESI-MS enabled quantitative analysis of characteristic ion generated by aging, heavy metal pollution, and pesticide pollution, while four heavy metals and four pesticides were also quantitatively analyzed.

Untargeted metabolomics profiling of rice soup samples was performed using MetaboAnalyst (https://www.metaboanalyst.ca). Raw mass spectrometry data (.xls format) underwent a standardized preprocessing pipeline to ensure analytical rigor. First, ion signals from 25–35 mass spectral files were aligned based on retention time and mass-to-charge ratio (m/z), followed by median-based intensity averaging across samples to generate a normalized data matrix with m/z as the independent variable and signal intensity as the dependent variable. Subsequent data refinement involved a multi-step approach to enhance data quality, beginning with the exclusion of features exhibiting over 50% missingness to prioritize variables with sufficient coverage. For the remaining features, missing values were imputed using one-fifth of the minimum positive value for each respective variable, a strategy designed to mitigate zero-inflation artifacts while preserving the structural integrity of the data. Finally, low-variance features were systematically removed by applying an interquartile range (IQR) threshold of <0.5, effectively eliminating background noise while retaining biologically or analytically relevant variation for downstream analyses. Systemic batch effects were mitigated via total intensity normalization (TIC) or probabilistic quotient normalization (PQN). Multivariate analyses including principal component analysis (PCA), orthogonal partial least squares-discriminant analysis (OPLS-DA) and volcano plots were performed on the preprocessed dataset. Statistical metrics (p-value from t-tests/ANOVA, variable importance in projection [VIP], and fold change [FC]) were computed to identify differentially abundant metabolites. Stringent thresholds were applied: statistical significance (p < 0.05, two-tailed), model discriminative power (VIP > 1), and biological relevance (FC > 2 or <0.5). Parameter selections (e.g., IQR cutoff and imputation strategy) were justified by their capacity to preserve low-abundance metabolites, attenuate instrumental noise, align with OPLS-DA classification criteria, and adhere to metabolomic community standards for biologically meaningful changes. All workflows were executed using MetaboAnalyst's default settings, with critical parameters documented in exportable JSON configuration files to ensure full reproducibility.

2.5. Method validation

2.5.1 Linear range of the standard curve and sensitivity. Using blank rice soup as the sample, mixed pesticide standard stock solutions of different concentrations were added and fully vortex-mixed to prepare five or six mixed standard samples. Analysis was performed using ND-EESI-MS according to the method discussed in Section 2.3. Secondary fragment ions of four pesticides (carbendazim, molinate, pirimiphos-methyl and cartap) were selected as quantitative ions for analysis, and the secondary ion signal intensity of each pesticide was recorded. To account for the influence of the background pesticide residues in blank rice soup samples on detection, the signal intensity of each concentration in the standard curve was subtracted from the background signal intensity of the original sample. The quantitative curve was plotted with the concentration of each pesticide as the abscissa and the average of three mass spectrometry signal intensities of the target ion as the ordinate. The linear relationship of the standard curve was evaluated by the correlation coefficient (R2). Four heavy metals (copper, lead, chromium and cadmium) were also analyzed according to the above standard curve procedure.

In mass spectrometry, the limit of detection (LOD) and the limit of quantification (LOQ) are key parameters defining sensitivity. The formulae for calculation are expressed as follows:42

LOD = 3.3σ/S, LOQ = 10σ/S
where σ denotes the standard deviation of baseline noise derived from nine blank measurements, and S represents the slope of the calibration curve.

2.5.2 Repeatability and accuracy. To test the reliability of the above standard curve, a spiked recovery experiment was conducted. For the four pesticides (50.00, 100.00, and 500.00 ng mL−1 carbendazim; 10.00, 50.00, and 100.00 ng mL−1 molinate; 10.00, 50.00, and 100.00 ng mL−1 pirimiphos-methyl; 5.00, 10.00, and 50.00 ng mL−1 cartap) and heavy metals (10.00, 50.00, and 100.00 ng mL−1 copper; 0.50, 5.00, and 10.00 ng mL−1 lead; 5.00, 10.00, and 50.00 ng mL−1 chromium; 0.05, 0.10, and 0.50 ng mL−1 cadmium), three different spiking levels were used. Each spiking level was analyzed three times, and the spiked recovery rate and relative standard deviation (RSD) for each analyte at different concentrations were calculated.
2.5.3 Stability test. Using blank rice soup samples kept at room temperature, sampling and measurement were performed at 0, 2, 4, 8, 18, and 24 h. Detection was carried out using ND-EESI-MS according to the parameters described in Section 2.3. The mass spectrometry detection range was set to m/z 50–1000. The total ion signal intensity of the rice soup sample was recorded, and the RSD of the six total ion signal intensities over 24 h was calculated according to the formula.
2.5.4 Matrix effect test. The sample contains multiple substances such as lactose, inorganic salts, phospholipids, etc., which may affect the ionization of the target in the ion source. Whether ionization of the target is suppressed or enhanced, it will affect the accuracy of quantifying the target. Therefore, it is necessary to investigate the matrix effect of rice soup to ensure the accuracy of the detection method. For different matrix samples, representative blank matrix samples were selected to investigate the matrix effect. Using a blank matrix to prepare standard working solutions and plotting working curves and comparing these with working curves plotted using pure aqueous solutions at the same concentration points, the ratio of the slopes of the two curves was calculated (matrix effect = slope of matrix standard curve/slope of solvent standard curve × 100%). A ratio greater than 100% indicates a matrix enhancement effect, and a ratio less than 100% indicates a matrix suppression effect, which is not conducive to target ionization.

3. Results and discussion

3.1. Optimization of experimental conditions in the positive ion mode

In order to optimize the ionization efficiency of ND-EESI-MS/MS, the electrospray extractant, the temperature of the ion-transport pipe, ionization voltage, the flow rate of electrospray extractant, the angle between two spray channels and the neutral desorption agent were optimized. In positive ion mode, the total signal intensity was chosen to represent the detection efficiency of the rice soup sample.
3.1.1 Optimization of the electrospray extractant. After preliminary screening, methanol/water (1[thin space (1/6-em)]:[thin space (1/6-em)]1, v/v) was found to have a better extraction effect on the sample compared with methanol or water alone. To further improve the protonation strength of the sample, different proportions of acetic acid (0–30%) were added to methanol/water solution. The experimental results showed that the total MS signal intensity of the rice soup sample was the highest when the ratio of methanol/water/acetic acid was 47.5[thin space (1/6-em)]:[thin space (1/6-em)]47.5[thin space (1/6-em)]:[thin space (1/6-em)]5 (v/v/v), as shown in Fig. S1a.† To retain as much sample information as possible, methanol/water/acetic acid (47.5[thin space (1/6-em)]:[thin space (1/6-em)]47.5[thin space (1/6-em)]:[thin space (1/6-em)]5, v/v/v) was selected for the next parameter optimization.
3.1.2 Optimization of the temperature of the ion-transport pipe. The temperature of the ion transfer tube exerts a pronounced influence on ion desorption efficiency, thereby modulating signal intensity. Since a temperature of 50 °C is close to the initial operating temperature of the instrument and excessive temperatures (>300 °C) may damage the instrument coil,42 the temperature was systematically optimized within the 50–260 °C range (Fig. S1b†). Notably, when the heating capillary temperature of the LTQ instrument was increased from 50 °C to 200 °C, the total MS signal intensity of the rice soup sample exhibited a progressive enhancement. Beyond this threshold, a gradual decline in total MS signal intensity was observed. This phenomenon is presumably attributed to enhanced desolvation at elevated temperatures. However, temperatures exceeding 200 °C may induce thermal dissociation of starch and trace proteins or amino acids in rice soup during ion transfer to the mass spectrometer, ultimately leading to signal attenuation. Based on these findings, the optimal ion transfer tube temperature was determined to be 200 °C.
3.1.3 Optimization of ionization voltage. This experiment investigated the impact of total MS signal intensity within the range of 0–7 kV, as depicted in Fig. S1c.† A significant increase in MS signal intensity was observed when the spray voltage ranged from 0 to 4 kV, with a subsequent flattening at 4–7 kV. It is possible that higher ionization voltage led to corona discharge at the nozzle, which resulted in no further improvement in ion formation efficiency. Therefore, an ionization voltage of 4 kV was chosen.
3.1.4 Optimization of the flow rate of the electrospray extractant. The impact of the electrospray solvent flow rate ranging from 1 μL min−1 to 9 μL min−1 on the MS signal intensity of the sample was investigated, and the results are depicted in Fig. S1d.† It was observed that when the flow rate exceeded 6 μL min−1, there was a decrease in signal intensity, indicating that higher flow rates resulted in reduced ionization efficiency. Therefore, 6 μ L min−1 was selected as the flow rate of solvent for electric spray.
3.1.5 Optimization of the angle between two spray channels. The impact of the angle between the two spray channels, within the range of 30–70°, on the intensity of the MS signal was investigated and is illustrated in Fig. S1e.† As the angle increased from 30° to 60°, there was an increase in signal intensity; however, from 60° to 70°, a decreasing trend was observed, possibly due to reduced ionization efficiency at excessive angles. Therefore, an angle of 60° was selected.
3.1.6 Optimization of the neutral desorption agent. When testing different agents including methanol, methanol/water (1[thin space (1/6-em)]:[thin space (1/6-em)]1, v/v), ethanol, ethanol/water (1[thin space (1/6-em)]:[thin space (1/6-em)]1, v/v) and water for condition optimization, ethanol showed the highest signal intensity. Considering the sample information as much as possible, ethanol was chosen as the neutral desorption agent (Fig. S1f†).

3.2. Optimization of experimental conditions in negative ion mode

When performing heavy metal detection in negative ion mode, we used the parameters optimized for positive ion mode (the temperature of the ion-transport pipe, ionization voltage, etc.), but found that the signal intensity decreased, considering that the root cause was that the acid-containing extractants commonly used in positive ion mode inhibited the ionization efficiency of negative ions. Therefore, while all other parameters remain unchanged, only the extractant composition was optimised for specific purposes.

The optimization results for negative ion mode reveal that the addition of acetic acid significantly suppresses ionization efficiency, leading to reduced signal intensity (e.g., a 32% signal decrease occurs at 5% acetic acid). Consequently, pure methanol (without added acetic acid) is identified as the best extractant for negative ion mode, yielding the highest sensitivity.

3.3. ND-EESI-MS fingerprints of rice soup

Mass spectrometric analysis was conducted on rice soup using both positive and negative ion detection modes under optimized experimental conditions. In positive ion mode, a total of 727 ions were detected, while in negative ion mode, 711 ions were observed (Fig. 2). The identification of a total of 19 active components was accomplished by comparing them with the NIST database, HMDB database, MassBank database, and relevant literature and confirmation using tandem mass spectrometry data. These components were then ranked in descending order based on their signal intensity (Table S2†). Among them, fatty acids including linolenic acid ([M + H]+, m/z 279) and stearic acid ([M + H]+, m/z 285) and amino acids including alanine ([M + H]+, m/z 90), serine ([M + H]+, m/z 106), asparagine ([M + H]+, m/z 133), aspartic acid ([M + H]+, m/z 134), and proline ([M + H]+, m/z 116) were detected. Additionally, sugars, including glucose ([M − H]−, m/z 179) were identified. The present method was employed to establish a comprehensive fingerprint map of rice soup, enabling the simultaneous detection and evaluation of diverse nutrients for quality assessment.
image file: d5ay00420a-f2.tif
Fig. 2 Mass spectra of rice soup acquired by ND-EESI-MS. (a) Mass spectra of rice soup acquired in the positive ion mode; (b) mass spectra of rice soup acquired in the negative ion mode.

3.4. Evaluation of quality control of the rice soup production process

3.4.1 Endogenous aging leads to reduced quality of rice soup. To gain further insights into the potential correlation between rice aging and characteristic ions, we analyzed both normal rice soup samples and manufacturing defect samples. Untargeted metabolomics analysis of normal rice soup samples and four artificially aged rice samples was performed using ND-EESI-MS. The characteristic spectra of the two samples are shown in the Fig. S3.† The spray reagent consisted of methanol/water/acetic acid (47.5[thin space (1/6-em)]:[thin space (1/6-em)]47.5[thin space (1/6-em)]:[thin space (1/6-em)]5), with ethanol used as the neutral desorption agent, while the aging test was conducted in positive ion mode. Principal Component Analysis (PCA) was employed as an unsupervised method for exploratory analysis, utilizing multivariate data analysis techniques to reduce the dimensionality of raw variables.43 A total of 25 samples (subjects) from five groups with varying degrees of rice aging were subjected to PCA using different combined univariate and multivariate data matrices. In the PCA model, the first two principal components (PC1 and PC2) explained 31% of the total variability in the positive ion mode, and found a clear separation between samples from the aging group and those from the normal group (Fig. 3a). The data changes in the PCA plot describe a clear sample analysis, with only the 28-day aging group showing significant differences from the other three groups, and the remaining three groups were not clearly differentiated (Fig. 3b). OPLS-DA could create a correlation model between various indicators and samples, allowing for the identification of indicators that effectively differentiate the samples based on their variable importance in projection values (VIP value). To assess the accuracy and reliability of the OPLS-DA model, cross-validation and permutation tests were employed. In this analysis, we utilized the OPLS-DA model to further verify whether it can effectively distinguish aging rice soup samples from normal samples. The results show that aging rice soup samples and normal samples could be clearly divided into two groups (Fig. 3c). The explanatory power of the model was represented by R2X and R2Y, while its predictive capability was indicated by Q2. Values of R2 and Q2 that were close to 1.0 suggest a strong fit for the model. As illustrated in Fig. S4a,† R2Y = 0.994 (p < 0.001) and Q2 = 0.866 (p < 0.001), both of which were near 1, indicating that the model demonstrates strong explanatory and predictive capabilities. Significant differences were observed between the aged samples and the normal rice soup samples, with 206 ions exhibiting VIP values greater than 1. Moreover, the volcano plot screened out 20 difference ions (Fig. 3d). The intersection of the two was taken to screen the 20 most representative differential ions, including 101, 111, 79, 97, 251, etc.(Fig. 3d), which represented a unique set of marker ions showing substantial intensity differences and highlighting their impact on rice soup quality due to aging effects. Upon CID, the precursor ions of 3-methyl-2-butenoic acid ([M + H]+) lost H2O to yield a predominant peak at m/z 83; the loss of CH2 and H2O resulted in the formation of fragment ions at m/z 69 and the loss of C3H6 produced fragment ions at m/z 59 (Fig. S6a†).
image file: d5ay00420a-f3.tif
Fig. 3 Metabolomics analysis of rice soup treated under normal and various aging conditions (40 °C RH 80%) by ND-EESI-MS. (a) PCA score plot of the normal group and all aging degree groups; (b) PCA score plot of different aging degree groups (7, 14, 21, and 28 days); (c) OPLS-DA of the normal group and all aging degree groups; (d) volcano plot of the normal group and all aging degree groups.

Furthermore, the metabolite 3-methyl-2-butenoic acid was quantitatively analyzed, revealing a robust linear relationship within the concentration range of 1.10–550.00 ng mL−1. The regression equation Y = 3.9841X + 10.526 demonstrated a strong linear relationship, supported by a high correlation coefficient of R2 = 0.9976. The LOD and LOQ were 0.56 ng mL−1 and 1.71 ng mL−1, respectively. At the same time, we investigated the quantitative results of 3-methyl-2-butenoic acid in normal and aged rice soup samples. The results showed that the content of 3-methyl-2-butenoic acid in normal rice soup samples was 19.95 ng mL−1, while in aged rice soup samples, it was 7.40 ng mL−1. This proves that the biomarker had undergone significant quantitative changes during the aging process.

3.4.2 Pollution from exogenous pesticide residues. Similar to the aging experiment, the pesticide detection experiment was conducted in positive ion mode using methanol/water/acetic acid (47.5[thin space (1/6-em)]:[thin space (1/6-em)]47.5[thin space (1/6-em)]:[thin space (1/6-em)]5) as the spray reagent and ethanol as the neutralizer. We analyzed the differential ions associated with pesticide pollution affecting rice soup quality. Upon addition of pesticides, both groups could be effectively differentiated from the normal rice soup sample in the PCA plot (Fig. 4a, d, g and j). The OPLS-DA model shows that the pesticide group was equally well differentiated from the normal group (Fig. 4b, e, h and k). And the results of the permutation test are R2Y = 0.991–1.000 (p < 0.05) and Q2 = 0.920–0.989 (p < 0.01) (Fig. S4b, c, d and e†), which show that the model had good predictive ability and there was no overfitting. The characteristic ions related to pesticides were selected based on VIP > 1 and volcano plots (FC > 2 or FC < 0.5, p < 0.05) (Fig. 4). The blue dots represented the characteristic ions with downregulated content and the red dots represented the characteristic ions with elevated content. Compared with the pesticide group, some characteristic ions show a significant increase and decrease, indicating that the metabolism of rice soup was abnormal after the addition of pesticide. Carbendazim residues resulted in differential ions at m/z 135, 134, 83, 287, 82, 157, 72, 89 and 277 (Fig. 4c). Similarly, molinate led to significant changes in ion intensity at m/z 129, 112, 111, 252, 265, 113, 139, 108, 338, 107, 140, 89, 109, 351, 110 and 324 (Fig. 4f). There were 137 differential ions produced by pirimiphos-methyl (Fig. 4i). After ion identification, the characteristic ions included trans-3-hydroxycinnamic acid ([M + H]+ m/z 165), α-hydroxyisocaproic acid ([M + Na]+ m/z 155), jasmine lactone ([M + H–H2O]+ m/z 151), 3-ethoxy-1-propanol ([M + Na]+ m/z 127), 2′-hydroxy-4′-methoxyacetophenone ([M + H]+ m/z 167), 2-methylglutaric acid ([M + Na]+ m/z 169), and 3,4-dimethyl-o-phenylenediamine (M+˙ m/z 136) (Table S2†). However, cartap produced 52 differential ions (Fig. 4l). After ion identification, the characteristic components included salicylic acid([M + H]+ m/z 139), α-amino-γ-butyrolactone ([M + H]+ m/z 102), 2-methyl-1,4-benzenediol, ([M + H]+ m/z 125), 5-methylcytosine ([M + H]+ m/z 126), and 2-ethoxy-phenol (M+˙ m/z 138), which could be used to identify cartap (Table S3†). These results highlight the effect of pesticides on the quality of rice soup.
image file: d5ay00420a-f4.tif
Fig. 4 Metabolomics analysis of ND-EESI-MS data of rice soup before and after adding pesticides. (a) PCA score plot of the normal group and carbendazim-treated group; (b) OPLS-DA analysis of the normal group and carbendazim-treated group; (c) volcano plot of the normal group and carbendazim-treated group; (d) PCA score plot of the normal group and molinate-treated group; (e) OPLS-DA analysis of the normal group and molinate-treated group; (f) volcano plot of the normal group and molinate-treated group; (g) PCA score plot of the normal group and pirimiphos-methyl-treated group; (h) OPLS-DA analysis of the normal group and pirimiphos-methyl-treated group; (i) volcano plot of the normal group and pirimiphos-methyl-treated group; (j) PCA score plot of the normal group and cartap-treated group; (k) OPLS-DA analysis of the normal group and cartap-treated group; (l) volcano plot of the normal group and cartap-treated group.

Quantitative analysis was performed using characteristic ions (Fig. S7a–d†), establishing a methodology for pesticide residue detection in the rice soup matrix. We observed that salicylic acid exhibited good linearity (R2 = 0.9983) within the range of 1.30–130.00 ng mL−1 when cartap was added as a characteristic component (Table 2), with LOD and LOQ values of 1.30 ng mL−1 and 3.95 ng mL−1, respectively. For carbendazim, molinate, pirimiphos-methyl and cartap in this matrix, linear calibration curves (R2 = 0.9905–0.9991) were observed within the range of 5.00–2000.00 ng mL−1, showing LOD and LOQ values within the range of 0.55–1.07 ng mL−1 and 1.66–3.24 ng mL−1, respectively (Table S4†).

Table 2 Regression equations, LODs and LOQs of characteristic ions in rice soupa
No. Analyte Precursor ion (m/z) Quantitative ion (m/z) Linear range (ng mL−1) Regression equation Y = aX + b R2 LOD (ng mL−1) LOQ (ng mL−1)
a X represents the concentration of the solution containing the analyte and Y denotes the signal intensity of the quantitative ion.
1 3-Methyl-2-butenoic acid 101 83 1.10–550.00 Y = 3.9841X + 10.526 0.9976 0.56 1.71
2 Salicylic acid 139 121 1.30–130.00 Y = 0.7096X + 14.802 0.9983 1.30 3.95
3 Rhamnose 163 119 7.50–1500.00 Y = 1.3835X + 5.4402 0.9987 0.95 2.89

The spike recovery experiment is an indicator used to evaluate the ability of a method to maintain satisfactory accuracy in the presence of interference from the biological matrix and other potential interference sources introduced during the sample preparation process. After adding different concentrations of standard to rice soup samples, recovery rates for target pesticides were within the range of 91.86–106.80% (Table S5†), meeting the acceptable range of 80–120% stipulated by the international standard ISO/IEC 17025 for food testing methods. Simultaneously, the RSD values were within the range of 0.79–3.67%, confirming that this method can maintain reliable quantitative accuracy despite interference from the complex biological matrix.

The stability experiment is key to ensuring reliability of analysis results. This study determined intra-day precision through six consecutive intra-day replicate experiments (n = 6). The results showed the RSD value of total ion signal intensity in rice soup was 1.68%, significantly lower than the 5% threshold specified by the International Council for Harmonisation (ICH) for chromatographic methods. These data indicate that the established ND-EESI-MS technology possesses excellent operational stability between different test batches, fully meeting precision requirements for trace contaminant analysis in complex food matrices.

As seen in Table S6,† carbendazim, molinate and pirimiphos-methyl exhibited significant matrix suppression effects in the rice soup matrix, while the effect on cartap was well mitigated. To avoid impact of matrix effects on quantification of targets, a matrix-matched standard calibration method should be employed when plotting standard working curves. Specifically, a blank rice soup matrix was used as the solvent to prepare a series of standard working solutions containing target analytes. This method compensates for or counteracts interference from matrix effects on the ionization efficiency of targets by adding matrix components identical to those in samples into the standards, thereby simulating the matrix environment in actual samples.

Collectively, these validation results demonstrate that the established ND-iEESI-MS methodology provides a sensitive approach for the simultaneous quantification of diverse pesticide residues in complex rice soup matrices.

3.4.3 Exogenous heavy metal pollution. Heavy metal ions typically exist as inorganic salts that are not readily volatilized under typical conditions of organic mass spectrometry analysis. Consequently, they cannot be ionized and subsequently analyzed in the mass spectrometer, posing a significant challenge for the detection of heavy metal ions using organic mass spectrometry techniques. Heavy metals (copper, lead, chromium and cadmium) were detected in negative ion mode by employing methanol plus EDTA-2Na as the extraction reagent due to the ability of EDTA to chelate with over 90% of heavy metals and form complexes.44 These complex ions facilitated rapid detection of heavy metal ions using organic mass spectrometry. We analyzed differential ions associated with heavy metal pollution in rice soup. Principal component analysis demonstrated clear differentiation between samples containing added heavy metals and normal rice soup samples (Fig. 5a, d, g and j). The OPLS-DA model was also applied to heavy metals analysis, and the results show that the heavy metal group can be well distinguished from the rice soup group (Fig. 5b, e, h and k). And the results of the permutation test are R2Y = 0.995–1.000 (p < 0.01), Q2 = 0.841–0.971 (p < 0.01), which show that the model had good predictive ability and there was no overfitting (Fig. S5a–d†). OPLS-DA and volcano plot analyses (Fig. 5c, f, i and l) (VIP > 1, FC > 2 or FC < 0.5 and p < 0.05) revealed characteristic ions indicative of heavy metals pollution alongside excimer ions. Residues of Cu+ led to the appearance of differential ions such as m/z 185, 269, 217, 227, 241, 281, 121, 136, 228, 199, 443, 79, 242, 166, 122, 253, 73, 91, 211, 286, etc. Rhamnose ([M − H]− m/z 163) and mannose-6-phosphate ([M − H]−; m/z 259) were characteristic ions when the concentration of Cu2+ exceeded standard limits (Fig. S6b and c†). Upon CID, the precursor ion of rhamnose ([M − H]−) lost CO2 to yield the predominant peaks at m/z 119 (Fig. S6b†). Moreover, precursor ions of mannose-6-phosphate ([M − H]−) lost H2O to yield the predominant peaks at m/z 241 (Fig. S6c†). Similarly, the presence of Pb2+ resulted in marked changes in the ion intensity at m/z 298, 299, 203, 275, 247, 315, 152, 245, 568, 179, 157, 259, 235, 181, 205, 219, 217, 213, 189, 135, 280, 178, 553, 215, 233, 519, 158, 333, 351, 261, 569, 464, 164, 145, 239, 525, 249, 567, 267, 286, 349, 248, 166, 216, 265, 365, 393, 332, etc. The differential ions associated with Cr3+ were detected at m/z 242, 255, 119, 122, 194, 106, 151, 92, 241, 132, 254, 253, 141, 469, 478 and 229. The differential ions associated with heavy metal Cd2+ were detected at m/z 167, 201, 219, 382, 381, 187, 207, 338, 335, 277, 336, 185, 122, 318, 164, 188, 173, 203, 231, 395, 232, 238, 176, 383, 239, 538, 174, 190, 150, 199, 432, 136, 781 and 200. CID identification revealed that m/z 187 represented azelaic acid ([M − H]−) (Fig. S6d†). The precursor ion of azelaic acid ([M − H]−) yielded predominant peaks at m/z 12, due to the cleavage of COOH and OH (Fig. S6d†). The data confirmed the successful detection of azelaic acid.
image file: d5ay00420a-f5.tif
Fig. 5 Metabolomics analysis of ND-EESI-MS data of rice soup before and after adding heavy metals. (a) PCA score plot of the normal group and Cu2+; (b) OPLS-DA of normal group and Cu2+; (c) volcano plot of the normal group and Cu2+ group; (d) PCA score plot of the normal group and Pb2+ group; (e) OPLS-DA of the normal group and Pb2+ group; (f) volcano plot of the normal group and Pb2+ group; (g) PCA score plot of the normal group and Cr3+ group; (h) OPLS-DA of the normal group and Cr3+ group; (i) volcano plot of the normal group and Cr3+ group; (j) PCA score plot of the normal group and Cd2+ group; (k) OPLS-DA of the normal group and Cd2+ group; (l) volcano plot of the normal group and Cd2+ group.

Furthermore, we conducted quantitative analysis of characteristic ions to establish a detection method for daily rice soup quality inspection. It was observed that the characteristic component rhamnose, produced by adding copper solution exhibited a strong linear relationship within the range of 7.50–1500.00 ng mL−1, with a high linear correlation coefficient (R2 = 0.9987). The LOD value was 0.95 ng mL−1, while the LOQ value was 2.89 ng mL−1 (Table 2). Additionally, in anion mode, we performed quantitative analysis experiments on copper, lead, chromium and cadmium in rice soup. The results clearly demonstrate that these elements show excellent linearity within the concentration range of 0.05–1500.00 ng mL−1, with linear correlation coefficients ranging from R2 = 0.9918 to R2 = 0.9996 (Table S7†). The LOD and LOQ values were 0.27–2.18 ng mL−1 and 0.82–6.61 ng mL−1, respectively.

For heavy metal detection, the spike recovery experiment showed that after adding different concentrations of standards to rice soup samples, recovery rates for target heavy metals were stable within the range of 90.72–106.40%, and the RSD values were within the range of 0.42–3.75% (Table S8†). Matrix effect analysis indicated that certain heavy metals including copper, chromium and cadmium exhibited matrix suppression, while the effect on lead was relatively minor (Table S9†). The matrix-matched standard calibration method should be employed for correction.

3.5. Matrix effects for characteristic ions in rice soup

3-Methyl-2-butenoic acid, salicylic acid and rhamnose exhibited matrix suppression effects (85.22%, 88.60% and 92.17%, respectively) in rice soup (Table S10†). These values confirm consistent ionization suppression across endogenous biomarkers, necessitating matrix-matched calibration for accurate quantification.

4. Conclusions

Before testing, we need to perform complex sample processing, which carries a certain risk of introducing contamination. In this study, we employed the ND-EESI-MS method combined with untargeted metabolomics for the rapid detection of endogenous and exogenous metabolites in rice soup without any sample pretreatment. This approach offers the advantages of simplicity and fast detection speed, enabling swift identification of rice soup quality. It is anticipated that this method will serve as a reference for high-throughput detection of differential endogenous and exogenous metabolites in other foods, presenting promising potential for direct and rapid analysis in food safety and the pharmaceutical industry.

Open access

This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper. Should any data files be needed in any other format, they are available from the corresponding author upon reasonable request.

Author contribution

Manman Qin, Leting Wang and Kaixin Xu: performed data acquisition by ND-EESI-MS. Manman Qin, Zhehao Liang, and Dan Shan: conducted data preprocessing and statistical analysis. Manman Qin: drafted the manuscript. Chao Zhong, Yanfei Xie, Hao Fan, and Jun Yu: supervised data interpretation. Jiang Wang: conducted partial mass spectrometry data analysis and facilitated communication among authors. All authors provided critical input, revised, and approved the final version.

Conflicts of interest

The authors declare no competing interests.

Acknowledgements

This work was supported by the Jiangxi University of Chinese Medicine School-level Science and Technology Innovation Team Development Program (No. CXTD22005), High-level Talents of Chinese Medicine (No. 13030599), the Science and Technology Program of the Department of Jiangxi Administration of Traditional Chinese Medicine (2021A394), and the University-level Open Experimental Program of Jiangxi University of Chinese Medicine.

References

  1. S. W. Qing, Dietary Cookbook of Suixiju, Sanqin Publisher, 2005 Search PubMed .
  2. B. Kaur, V. Ranawana and J. Henry, Crit. Rev. Food Sci. Nutr., 2016, 56, 215–236 CrossRef PubMed .
  3. M. Kianmehr, A. Saber, J. Moshari, R. Ahmadi and M. Basiri-Moghadam, Nurs. Midwifery Stud., 2016, 5, e25852 Search PubMed .
  4. Y. Cao, W. Yan, L. Lu, Y. Tao, W. Lu, Y. Chen, Q. Tang and W. Cai, Asia Pac. J. Clin. Nutr., 2016, 25, 631–635 Search PubMed .
  5. H. Z. Zhang, Chin. Tradit. Pat. Med., 1991, 2, 46 Search PubMed .
  6. X. Dong, X. T. Sui, L. M. Hu, R. G. Wang, Z. G. Li, J. L. Wang, S. M. Wang and L. Liang, Sci. Technol. Food Indust., 2024, 45, 358–367 Search PubMed .
  7. C. L. Luo, H. X. Chen, W. J. Wang and Z. Zhao, Chin. J. Mod. Dev. Tradit. Med., 2016, 18, 501–504+508 Search PubMed .
  8. L. F. Wang, C. Chen, J. H. Yang, J. Y. Chen, H. F. Wang and X. R. Ju, Sci. Agric. Sin., 2015, 48, 1807–1817 Search PubMed .
  9. Z. J. Lu, S. S. Zhao, F. Cai, S. Cai, Y. Sui, J. B. Shi, T. Xiong, J. J. He, X. L. Chen, C. H. Fan, Z. W. Jia and X. Mei, Mod. Food Sci., 2022, 38, 126–135+128 Search PubMed .
  10. Z. Q. Yang, Q. L. Lin, X. X. Cheng and X. Liu, Food Sci. Technol., 2010, 35, 164–167 CAS .
  11. C. Lv, J. Jin, P. Wang, X. Dai, Y. Liu, M. Zheng and F. Xing, Food Chem., 2019, 293, 472–478 CrossRef CAS PubMed .
  12. S. Y. Sun, C. Zhang and R. F. Hu, China Agric. Econ. Rev., 2020, 12, 367–379 CrossRef .
  13. G. Chen, L. Shi, J. Wang, S. Zhu, J. Sheng, X. Yang and H. Xu, Food Addit. Contam. B, 2023, 16, 176–184 CrossRef CAS PubMed .
  14. E. S. Harris, S. Cao, B. A. Littlefield, J. A. Craycroft, R. Scholten, T. Kaptchuk, Y. Fu, W. Wang, Y. Liu, H. Chen, Z. Zhao, J. Clardy, A. D. Woolf and D. M. Eisenberg, Sci. Total Environ., 2011, 409, 4297–4305 CrossRef CAS PubMed .
  15. S. F. Ahmed, P. S. Kumar, M. R. Rozbu, A. T. Chowdhury, S. Nuzhat, N. Rafa, T. M. I. Mahlia, H. C. Ong and M. Mofijur, Environ. Technol. Innovat., 2022, 25, 102114 CrossRef CAS .
  16. Y. Xu, B. Seshadri, B. Sarkar, H. Wang, C. Rumpel, D. Sparks, M. Farrell, T. Hall, X. Yang and N. Bolan, Sci. Total Environ., 2018, 621, 148–159 CrossRef CAS .
  17. Q. Li, W. Huang, C. Xiong and J. Zhao, Chemosphere, 2018, 201, 294–302 CrossRef CAS PubMed .
  18. C. Li, Y. Lin, X. Li, J. J. Cheng and C. Yang, Sci. Total Environ., 2023, 866, 161411 CrossRef CAS .
  19. J. Zhang, Z. Tang, E. Agathokleous, G. Zheng, L. Xu and P. Li, Sci. Total Environ., 2023, 889, 164328 CrossRef CAS PubMed .
  20. W. Dai, B. Inumbra, P. Y. Wong, A. Sarmiento, Y. Yau, J. Han, G. Mao, Y. K. Peng and J. L. Chen, Biosensors, 2023, 13, 523 CrossRef CAS .
  21. X. Li, Y. Wu, H. Wang, J. Wen and M. Zhu, Chemosphere, 2023, 337, 139385 CrossRef CAS .
  22. B. C. Lv, Z. X. Yang, L. B. Fu, J. F. Chen, M. Yin, H. Chen, W. Wang, Y. X. Wang, Z. Y. Wang, W. Yang and Z. X. Su, J. Food Sci. Technol., 2023, 36, 224–233 Search PubMed .
  23. Y. Q. Li, R. L. Li and Z. Q. Shi, J. Jiangxi Agric. Univ., 2008, 24, 977–978 Search PubMed .
  24. T. Shan, W. Sun, H. Liu, S. Gao, S. Lu, M. Wang, W. Sun, Z. Chen, S. Wang and L. Zhou, Int. J. Mol. Sci., 2012, 13, 11275–11287 CrossRef CAS .
  25. Q. Zhao, J. Xi, D. Xu, Y. Jin, F. Wu, Q. Tong, Y. Yin and X. Xu, Food Chem, 2022, 385, 132701 CrossRef CAS .
  26. D. K. Lim, C. Mo, D. K. Lee, N. P. Long, J. Lim and S. W. Kwon, J. Food Drug Anal., 2018, 26, 260–267 CrossRef CAS PubMed .
  27. M. Arslan, M. Zareef, H. E. Tahir, Z. Guo, A. Rakha, H. Xuetao, J. Shi, L. Zhihua, Z. Xiaobo and M. R. Khan, Food Chem,, 2022, 368, 130783 CrossRef CAS PubMed .
  28. W. Jin, Z. Zhang, S. Zhao, J. Liu, R. Gao and P. Jiang, Food Res. Int., 2023, 169, 112879 CrossRef CAS .
  29. S. Feng, L. Wang, P. Shao, B. Lu, Y. Chen and P. Sun, Food Chem,, 2022, 387, 132742 CrossRef CAS PubMed .
  30. N. R. Harischandra, M. S. Pallavi, M. Bheemanna, K. PavanKumar, V. Chandra Sekhara Reddy, N. R. Udaykumar, M. Paramasivam and S. Yadav, Food Chem,, 2021, 347, 128986 CrossRef CAS .
  31. W. Jira and S. Münch, Food Chem,, 2019, 275, 214–223 CrossRef CAS PubMed .
  32. I. Romero-Sánchez, L. Ramírez-García, E. Gracia-Lor and Y. Madrid-Albarrán, Food Chem,, 2022, 395, 133611 CrossRef PubMed .
  33. J. Verstraete, S. Strobbe, D. Van Der Straeten and C. Stove, Talanta, 2021, 224, 121905 CrossRef CAS PubMed .
  34. Z. Luo, L. Zhang, Y. Mou, S. Cui, Z. Gu, J. Yu and X. Ma, Anal. Bioanal. Chem., 2019, 411, 2447–2460 CrossRef CAS PubMed .
  35. J. Ding, H. Gu, S. Yang, M. Li, J. Li and H. Chen, Anal. Chem., 2009, 81, 8632–8638 CrossRef CAS .
  36. X. Y. Huang, X. W. Fang, X. Zhang, X. M. Dai, X. L. Guo, H. W. Chen and L. P. Luo, Anal. Bioanal. Chem., 2014, 406, 7705–7714 CrossRef CAS .
  37. Y. Liu, X. Zhang, Y. Ouyang, Z. Hu, L. Ma, J. Zhang, J. Lin and H. Chen, J. Mass Spectrom., 2011, 46, 794–803 CrossRef CAS .
  38. D. Wu, M. Cui, Y. Hao, L. Liu, Y. Zhou, W. Wang, A. Xue, K. Chingin and L. Luo, J. Agric. Food Chem., 2019, 67, 12945–12952 CrossRef CAS PubMed .
  39. Z. Wu, K. Chingin, H. Chen, L. Zhu, B. Jia and R. Zenobi, Anal. Bioanal. Chem., 2010, 397, 1549–1556 CrossRef CAS PubMed .
  40. M. Tamura, H. Tsujii, T. Saito and Y. Sasahara, LWT--Food Sci. Technol., 2021, 150, 111887 CrossRef CAS .
  41. H. Y. Zhang and Y. O. Yang, J. Mass Spectrom., 2008, 43, 1181–1190 CrossRef CAS PubMed .
  42. J. H. Ding, H. W. Gu, S. P. Yang, M. Li, J. Q. Li and H. W. Chen, Anal. Chem., 2009, 81, 8632–8638 CrossRef CAS PubMed .
  43. C. Gertz, A. Gertz, B. Matthäus and I. Willenberg, Eur. J. Lipid Sci. Technol., 2019, 121, 1900281 CrossRef CAS .
  44. S. S. Shen, J. Q. Xu, H. W. Chen and X. W. Fang, Paper Presented at the the Third National Symposium on Mass Spectrometry Analysis, Xiamen, China, 2017 Search PubMed .

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ay00420a
This journal is © The Royal Society of Chemistry 2025
Click here to see how this site uses Cookies. View our privacy policy here.