Safety Assessment and Contaminants Detection in Different types of Tea and Tea products
2024 年 12 巻 p. 347-376
詳細
2024 年 12 巻 p. 347-376
Tea crop is highly valued for its aroma, taste, and health benefits. Although it exhibited low water content, it’s essential to consider its safety against contaminants and adulterants such as pesticides, heavy metals, and unlawful chemical additives. Quality, shape, and color are exterior elements, but aroma and flavor are internal core quality factors of tea. Tea leaves are susceptible to contamination during the pre-harvest stage and exposed to many post-harvest factors that enhance the chances of tea quality deterioration. Numerous traditional methods of assessing tea quality exist, but emerging techniques such as spectroscopy, imaging, and electrochemical methods are more accurate. This paper reviews tea types, their consumption, contaminants, and safety issues and evaluates traditional and advanced techniques for detecting tea contaminants and pollutants. This review has the potential to provide comprehensive data about the safety detection of tea utilizing conventional and advanced methodologies, perhaps assisting the tea sector’s expansion.
Tea (Camellia sinensis L.) leaves are well known as one of the prominent ingredients of three global drinks grown mostly in Asia, America, Africa, and Oceania [1]. Its origin is commonly associated with Southwest China, Northeast India, and Myanmar [2]. The six main teas produced globally are green, white, oolong, yellow, black, and Pu-erh tea [3]. Tea plants are botanically classified as shrubs and minor arbors, with leaf diameters ranging from 3.8 cm to 25 cm in length [2]. After harvesting, tea leaves are processed in a variety of ways, yielding four distinct products based on fermentation, including white and green (unoxidized), oolong (partially oxidized), and black (completely oxidized) [4]. According to FAO STAT, global tea production has surpassed 7 million tons annually since 2020, with China (42%) and India (20%) being the major contributors [5]. Tea consumption has increased by 4.5% each year over the past decade. The production of black tea is set to rise by 2.2% annually, while green tea production is expected to increase by 7.5% by 2027 [6]. Black tea accounts for 75% of worldwide consumption in Europe and North America, whereas green tea is more prevalent in Asia, accounting for 15% of global consumption [7]. Oolong tea accounts for 10% of Chinese tea consumption, according to the China Tea Marketing Association (2022), and is popular in China and Southeast Asia [8].
Green tea, produced by rapid boiling or pan-frying newly collected leaves to deactivate enzymes, including polyphenol oxidase, is popular in Pakistan [9]. Henceforth, green tea’s polyphenol composition closely resembles fresh leaves, with the foremost components being epicatechin, galloyl, and their epigallocatechin esters. In black tea production, withered tea leaves (65–70% moisture) are crushed or rolled, and tea catechins are oxidized using polyphenol oxidase (PPO) and peroxidase following inactivation by drying. The reduced catechin levels caused by enzymatic oxidation reduce astringency and bitterness in black tea leaves. This behavior has been found in black tea and green tea aerobic and anaerobic microbial fermentation [10]. Caffeine levels in black tea are identical to those in green tea leaves, indicating that processing procedures may be linked to reduced astringency via lower catechin levels. Although microorganisms are not involved in black tea fermentation, the enzymatic oxidation in the production process is commonly called “tea fermentation”[11]. The widespread appeal of tea consumption arises from the unique color, flavor, and health properties attributed to thousands of secondary metabolites [12]. Secondary metabolites in tea leaves are classified as phenolic compounds (iso-flavonoids, flavonoids, lignin, anthocyanins, and tannins), terpenes (terpenoids), and nitrogen-containing chemicals (alkaloids, glucosinolates, and cyanogenic glycosides) [13]. Flavonoids account for 25 to 35% of tea composition, whereas green and white teas include catechins (30–42%) on a dry basis, and green and white teas contain proanthocyanidins [14]. Oolong and black teas contain theasinensins, theaflavins, and arubigins, while dark teas contain theabrownins. Anthocyanins and new flavoalkaloids are also gaining popularity due to their importance in teas [15]. Because of its distinctive chemical profile, tea is now used not just as a beverage but also for its medicinal and functional properties, such as antioxidant and anti-inflammatory properties [16].
Tea is considered a low-risk food since the water content in the end products is kept low during fermentation. However, pesticide residues, heavy metals, unlawful chemical additions, and various chemical pollutants in tea must not be overlooked. They offer a multitude of food safety risks in tea due to intentional adulteration during cultivation, processing, and storage [17]. Many studies have also found that among the biggest safety problems in tea are microbiological contamination, environmental contaminants, pesticide residues, and dangerous trace elements and heavy metals. Approximately 80% of the tea quality issues are attributable to pesticide residues [18, 19]. To date, tea plant diseases and insect pests wreak havoc on tea production, quality, and crop productivity. Globally, around 1000 different species of tea pests result in 43% productivity losses and 380 different types of tea illnesses reported [20, 21]. Even though tea safety is a critical issue for producers, customers, consumers, and the general public worldwide, there is minimal advanced information on the quality and safety profile of tea and tea products. This lack of combined information is concerning and highlights the need for more research. As a result, a detailed study was required to understand better the modern methods and techniques used to detect tea quality. The current study provides comprehensive information and analyses of advanced literature depicting chemical, conventional, and non-destructive methods being applied worldwide to evaluate the safety of tea and tea products.
The keywords “tea safety”, “tea contaminants”, “adulteration”, “detection techniques”, and “non-destructive detection methods” were used to search and collect material for this critical evaluation, either independently or in combination. Google Scholar, PubMed, ScienceDirect, Scopus, Research Gate, and Web of Science were the databases searched for published studies and literature. For this study, the publication period beginning in 2018 was chosen, with most of the publications concentrating on the most recent research from the last three years. After carefully reviewing over 200 journal papers, around 132 were chosen to be included in this research.
Tea safety begins with crop growth and continues through harvesting and processing. During these intervals, numerous composites, contaminants, pollutants, and environmental factors affect the specific color, flavor, aroma, and chemical profile, i.e., caffeine and phenophenyles, and the safety of tea leaves [22], which are extensively discussed in advanced literature. Most tea production in the world exposes tea to fraud and intentional contamination, either during processing or transportation for import and export [23].
3.1 Effect of pre-harvest condition on tea safetyNumerous biotic and abiotic factors are major limitations to the agricultural production of tea [24]. Notably, fungal infections are a significant biotic factor influencing tea safety [25]. In modern times, using pesticides and chemical fertilizers in tea production has become normal practice and has proved to be a possible cause of heavy metal contamination in tea leaves [26]. Nitrogen (N) is an important component of chemical fertilizers since it enhances caffeine content in tea leaves; hence, it affects tea quality positively for green tea or adversely for black tea. Intense nitrogenous fertilizers resulted in a decrease in polyphenols, specifically thearubigins and theaflavin. Furthermore, high unsaturated fatty acids developed due to the heavy use of nitrogenous fertilizers are said to alter the grassy odor [27]. Soil conditions are crucial in determining tea safety as various nutrient stresses during plantation affect constituent concentration in tea. Heavy metal accumulation, negative effects on secondary metabolites, and organic acid secretion in tea plants have been observed as a result of soil exposure to aluminum (Al) and fluorine (F) stressors [28, 29].
Tiny tea leaves accumulated Al faster when 0.4 mM or 2.5 mM of the metal was applied to the plant for ten days. This resulted in a higher concentration of Al, about 20 and 40 mg/g FW, respectively. The 10 mM F treatment decreased the contents of epicatechin (EC) and epicatechin gallate (ECG), although most treatments did not affect the catechin molecules. Only the 2.5 mM Al treatment marginally raised the EGC and EC contents at 10 days [29]. Moreover, in reaction to insects, intrinsic defense mechanisms in tea plants release secondary metabolites, primarily pyrrolizidine alkaloids (PAs), leading to chronic or acute impairment of human health [30]. PAs toxicity can lead to liver failure, pulmonary hypertension, cardiac and renal damage, and cancer. Severe cases occur when dried teas exceed 150 μg/kg, per European Commission Regulation (EuCR) [31]. PAs accumulation from weeds can also contaminate tea during the growing and harvesting stages [32].
3.2 Effect of post-harvest conditions on tea safetyPost-harvest conditions for tea leaves, such as processing and storage, substantially influence tea’s sensory and nutritional attributes and safety [33]. Around 70% of tea is squandered globally due to inadequate post-harvest storage conditions. The root cause of wastage lies in the absence of cutting-edge technologies for real-time monitoring of tea during storage [34]. Many other factors contribute to the wastage of this appalling percentage of tea. Tea leaf’s vulnerability to fungal contamination results in mycotoxin development, particularly at temperatures ranging from 25 °C to 30 °C or with a long production period, less than 8 °C, and moisture levels exceeding 16% [35].
The production of 4-methylimidazole (4-MEI), a carcinogen of the 2B category in foods, is associated with their thermal processing, encompassing tea [36]. A study focused on the summer crop of Camellia sinensis (green tea), harvested from New South Wales, Australia, and exposed to varying temperatures (0, 5 and 25 °C) during delays in processing post-harvest from 6 h to 24 h. Despite recording reduced levels of catechins, theanine, and caffeine over these time delays and temperatures, a commercially viable semi-fermented food loss and waste tea was attained after 24 h at 25 °C [37].
3.3 Effect of processing and extraction techniques on tea safetyProcessing has a direct impact on the composition, as well as the sensory attributes of tea [23]. Green and black tea are treated differently, with black tea undergoing a slightly longer processing phase, exposing it to significantly higher contamination [35]. Equipment used in tea processing causes contamination by foreign substances such as copper (Cu), lead (Pb), and iron (Fe). The critical step in oolong tea processing is turning the leaves over after withering [33]. Unlike the simple injury sustained during picking, continuous mechanical damage is caused by leaf turnover. However, it also affects the production of jasmine lactone, one of the potent odors contributing to oolong tea’s fruity, sweet, floral aroma [38].
Controlling the tea production and safety detection systems can ensure that consumers receive safe and high-quality end products. There are several ways to assess tea safety and identify adulteration or contaminants in tea. Thin layer chromatography (TLC), liquid chromatography-tandem mass spectrometry (LC-MS/MS), high-performance liquid chromatography (HPLC), and ultra-high performance liquid chromatography (UPLC) are some of the most regularly utilized analytical technologies [39]. One of the constraints of traditional extraction processes is the efficient release of pesticide residues with tea extracts, which becomes a major safety risk when exporting tea [40].
Elevated temperatures and prolonged extraction can significantly impact tea safety due to the oxidation of bioactive components, resulting in off-flavor tea producing toxic oxidized chemicals [40]. A study used HPLC to assess L-theanine and caffeine levels in green, black, and white tea exposed to varying times and temperatures. In white tea, a theanine to caffeine ratio above 200 was obtained at a low temperature of around 11 °C for 5 min. An elevated temperature of about 90 °C to 100 °C helped extract high L-theanine levels, and at low temperatures, insignificant caffeine content was retained, keeping a constant time of 5 mins for extraction [41]. X-ray fluorescence spectroscopy revealed the nutritional profile of green tea, superfine green tea powder, and tea extract revealed nearly two times more tea polyphenols in an extract with robust antioxidant activity and efficient preservation of two amino acids, L-theanine and glutamic acid [29]. Diverse green technologies are currently applied for efficient extraction, resulting in premium tea extracts with greater bioavailability and antioxidant activity [19].
3.4 Role of contaminants and adulterants in affecting tea safetySafety parameters play a pivotal role in determining the overall quality of tea. The inclusion of lead chrome green (color indicator) to improve tea color and sibutramine, an appetite suppressant, affected consumers’ health negatively, jeopardizing tea safety and ultimately affecting the overall quality. Although talcum powder is a naturally occurring mineral, adding talcum powder due to the inclusion of asbestos in unfiltered talc can be hazardous to human health [42].
Microplastics (MPs) have become a serious environmental concern as they are detected everywhere, including the air, food, and beverages [43]. MPs are plastic particles with a diameter of 5 mm or less found in tea. They are obtained from agricultural plastic films, plastic packaging of tea products and instruments, organic fertilizers, and air deposition. Polyethylene, polypropylene, and polyethylene terephthalate are the most common materials used to make MPs [43]. Despite its nonbiodegradability, polypropylene is widely used to produce tea bags. Many minerals have been found in tea bags, including Al, F, Cu, Pb, arsenic (Ar), radium (Rd) salt, mercury (Hg), cadmium (Cd), barium (Ba), and nitrates (NO3-1) [44]. Plastic tea bags can break down into tiny particles ranging from 1 mm to 5 mm, which can take hundreds of years to decompose in nature. Water heated over 40 °C can release hazardous substances from plastics. As a result, billions of plastic particles can enter human cells through drinking tea [45]. When brewed at the right temperature, plastic tea bags release 11.6 billion MPs and 3.1 billion nanoplastics [46].
The use of non-invasive, non-destructive, rapid, chemical-free, and environment-friendly technologies in tea analysis has piqued the interest of researchers. Near-infrared, mid-infrared, Raman, terahertz, hyperspectral imaging, and vibrational spectroscopy have been widely explored for quality and safety evaluation. In addition, the uses of computer vision technology, as well as electronic tongue (e-tongue) and electronic nose (e-nose), in the evaluation of tea quality and nutrition, are also presented [40, 47]. Chromatography electrochemical fingerprinting has been employed in the taste assessment of tea beverages [48]. Figure 1 compares various degrees of fermented teas, accompanying dangers, and standard and innovative safety detection methods.
Neonicotinoids (NEOs) are nicotine-based water-soluble pesticides and are widely employed on tea plants to combat sucking and chewing pests [49]. Toxicology studies have shown that NEOs residues enhance oxidative stress in rats and animals at sublethal concentrations [50]. A phenolic-based non-ionic solvent mixture of DL-menthol and thymol detected NEOs in tea infusions. The validation results demonstrated good linearity (corresponding correlation R2 ≥ 0.99), high precision (RSD < 11%), and satisfactory recoveries (57.7–98%) at concentrations ranging from 0.05 μg L–1 to 100 μg L–1. The limits of quantification (LOQs) were found to be 0.05 μg L–1. The ingestion risk of NEOs in tea infusion samples was found to be tolerable, with a residual range of 0.1 μg L–1 to 3.5 μg/L [51]. Five black tea samples were tested for contaminants, caffeine, and antioxidants in another study. Findings reported that 80% of the samples contained azo colors, dyes, and sand. All tea samples displayed a wide range of antioxidant activity, ranging from 3.29 ± 1.03 to 15.96 ± 1.2, as measured by the 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radical scavenging method [52].
Tea cultivation that involves organophosphorus pesticides poses health risks to consumers. A study improved the efficacy of carbonized bacterial cellulose (CBC) in removing organophosphorus pesticides. Hydrazine hydrate (N-CBC) treatment showed the highest adsorption effect, removing dicrotophos 13 times better than CBC. Hydrophobic interaction dominated organophosphate pesticides (OPP) adsorption onto N-CBC. The pseudo-second-order kinetic model and Langmuir isotherm model were found to describe the process accurately [53].
4.2 Chromatography techniques for detection of tea contaminants 4.2.1 High-performance liquid chromatographyIn analytical chemistry, HPLC separates, classifies, and calculates each mixture component [54]. One of the most well-known dispersive solid-phase extraction (d-SPE) technologies in food safety is the Quick Easy Cheap Effective Rugged and Safe (QuEChERS) method [55, 56], which was used for ultra-trace detection of pesticides, mycotoxins, and other organic compounds in food matrices [57, 58]. In a study, 100 tea samples were extracted using a QuEChERS and analyzed using UHPLC coupled with tandem mass spectrometry (UHPLC-MS/MS). The analytical technique was found to be suitable for mycotoxins with a recovery rate from 100% to 117%, linearity (> 0.99), and precision (6–29%). The results showed that aflatoxins, namely aflatoxin B1 (AFB1), aflatoxin B2 (AFB2), aflatoxin G2 (AFG2), ochratoxin A (OTA), and zearalenone were not identified in all of the samples tested; nevertheless, the lower quantity (1.72–5.19 g/kg) of AFG1 was detected, which were below the threshold authorized by EuCR (1881/2006) [59]. To assess various tea matrices, a combination method of UPLC-MS/MS preceded by the modified QuEChERS method was used to find insecticide residues, tolfenpyrad, and its metabolites [60]. A combination of HPLC-UV, LC-MS, and headspace solid-phase microextraction coupled with gas chromatography-mass spectrometry (HSSPME-GC-MS) was used to detect polyphenols and volatile compounds in green and black tea [61].
A 5 g tea sample was mixed with a solvent mixture containing acetonitrile, water, and acetic acid (79:20:1). After vortexing, the mixture was centrifuged at 4 °C followed by the addition of 10 mL of this supernatant to a tube with 900 mg magnesium sulfate (MgSO4), 150 mg octadecylsiyl silica gel (C18), and 150 mg primary secondary amine (PSA). After centrifugation, the organic phase was dried under nitrogen and mixed with methanol and water mixture (65:35 v/v) to analyze using UHPLC-MS/MS for mycotoxins detection. The equilibration time between each chromatographic was set at 3.50 min, and analytes were separated on a biphenyl analytical column, maintaining a flow rate of 0.35 mL/min [59]. When afidopyropen is applied to crops, it produces the metabolite M440I007, a novel insecticide for controlling piercing pests in tea gardens. A thin porous teflon (TPT) cartridge-based technique was devised for solid phase extraction of afidopyropen and M440I007 from tea matrices. Both target compounds were extracted with water and acetonitrile in a 4:10 ratio for fresh leaves and an 8:10 ratio for dried tea leaves and analyzed using UHPLC/MS. This technique yielded LOQs of 0.005, 0.005, and 0.002 mg/kg in fresh tea shoots, dried tea, and tea infusion for both target compounds. The mean afidopyropen and M440I007 recovery rates varied from 79% to 101.5%, with 14.7% relative standard deviations [62].
Trans-anethole, estragole, coumarin, lawsone, methyl eugenol, pulegone, and trans-cinnamaldehyde were reported in tea samples. Tea samples were extracted using an acetonitrile-water solvent mixture subjected to whirlpool oscillations and dilution. The solvent was then treated with anhydrous sodium sulfate, primary secondary amine (PSA), and C18 to be quantified using UHPLC-MS/MS and a matrix-matched external standard approach. The results exhibited a good linearity from 0.8 to 1600.0 μg/L with R2 > 0.999. The limit of detection (LOD) was 0.020 mg/kg and 0.050 mg/kg, with recoveries ranging from 82.5% to 102.5%. Trans-anethole, coumarin, and estragole were detected at maximum concentrations of 65 mg/kg, 2 mg/kg, and 72.2 mg/kg in green and black tea, respectively [63]. It is possible to identify and quantify various contaminants with great sensitivity and accuracy when using the separation first technique in HPLC, which offers excellent resolution and specificity in separating complicated mixtures. However, this process needs specialized equipment, takes longer since sample preparation and separation are required, and has higher operating expenses because solvents and consumables are needed [64]. In contrast, the direct detection technique in HPLC can be used for quick screening and allows faster analysis by omitting the separation stage. This strategy may also result in cheaper operational expenses. Despite this, there is a greater chance that matrix effects may affect the accuracy of the detection process, and the method’s weaker specificity and resolution will make it difficult to distinguish and measure individual contaminants in complex combinations [65].
4.2.2 Ion exchange chromatographyAccurate selenium (Se) measurement in tea samples is crucial for quality control and safety of Se supplementation. A method for measuring Se in tea samples that combines ion exchange chromatographic separation, mixed acid digestion, and high-resolution inductively coupled plasma mass spectrometry (HR-ICP-MS) was devised. The radioisotope 75 Se was used as a tracer to quantify Se loss during separation. Adopting an automated separation system expedited the operation and cut the analysis time to three hours. Following acid digestion, HPLC measurements on the sample solution eluted 95% of Se. Certified reference materials validated this method; the measured values accord well with the certified values, demonstrating the analytical method’s accuracy [3]. The inorganic anions, including fluoride (F-), chloride (Cl-), nitrate (NO3-1), bromide (Br-) ions, and others, were successfully detected in tea bag samples using an analytical column (Dionex IonPac AS9-HC 4×250 mm). The eluent employed was 12 mM potassium hydroxide (KOH), and LOQ for inorganic anions ranged from 0.02 to 0.20 mgL-1 [66].
In particular, it is useful for complicated matrices where accurate separation is required. Ion exchange chromatography (IEC), the first separation technique, excels in attaining high selectivity for charged species, enabling exact separation and quantification of ions. This technique enables ion exchange resins, which interact with analytes selectively according to their charge, to effectively discriminate between ions in a mixture [67]. Nevertheless, it has some significant disadvantages, such as longer analysis durations since the sample must be carefully prepared, which includes preparing the ion exchange resins, equilibrating the columns, and maximizing the separation conditions. Specialized ion exchange resins and controlled conditions are required, which adds to the complexity and expense and calls for highly trained personnel and advanced machinery [68]. Meanwhile, by removing the separation phase and shortening the analysis time and operational complexity, the direct detection technique in IEC provides a quicker and easier substitute, which is helpful for quickly screening certain ions in simpler matrices.
The direct detection methodology is less accurate and has lower specificity than other methods, which makes it harder to discriminate between comparable ions and more susceptible to interference from other species in the sample matrix. However, it is more economical and requires fewer equipment and refills. This strategy is less effective in complex matrices where overlapping signals can cause unclear data interpretation. Given the particular analytical needs, such as the sample matrix’s complexity, the urgency of the findings, and the available resources and experience, the decision between various methodologies should be made based on methodology [69].
4.2.3 Liquid chromatographyPolycarboxylic acids (PCAs) are major metabolites in all teas, yet due to the complex matrix and physicochemical characteristics of PCAs, contemporary techniques for detecting PCAs in tea beverages are limited. Seven PCAs in tea, including tartaric, ketoglutaric, malic, malonic, cis-aconitic, succinic, and fumaric acid, were detected using LC-MS. The LOQ of PACs was 1–50 ng/mL, with a recovery range of 72.2–122.5%. According to these findings, tea beverages included more malic acid, succinic acid, and malonic acid than non-tea beverages [70]. Tea includes significant polyphenols and amino acids, which interfere with detecting 4-MEI using LC-MS. The research demonstrated a unique approach for sensitive 4-MEI analysis using cold-induced phase separation of an acetonitrile-water system coupled with liquid-liquid extraction (LLE) [36].
Pymetrozine and nitenpyram, two insecticides, were extracted using boiling water, followed by detection with UPLC-MS/MS [47]. Pesticide residual concentrations in 105 Hangzhou-grown green tea samples revealed that only 18 pesticides were present, out of 14 evaluated using GS-MS and 27 tested using LC-MS. Pesticide residues found were imidacloprid (35.2%), acetamiprid (26.7%), carbendazim (21%), bifenthrin (21%), and cyhalothrin (21%) (19.1%) with the mean value less than the LOD (2.64 mg/kg), except for imidacloprid, acetamiprid, carbendazim, bifenthrin, and cyhalothrin [71]. Liquid chromatography (LC) offers the advantage of isolating individual components before detection, providing increased sensitivity and resolution for identifying a wide range of contaminants and handling complex samples. Chromatographic separation allows this method to achieve exceptional accuracy, but it is expensive, time-consuming, and requires specialized equipment and significant sample preparation. Moreover, the direct detection method expedites and streamlines the method by omitting the separation stage, resulting in a quicker analysis and lower operating expenses. Unfortunately, the approach reduces both specificity and sensitivity and raises the possibility of matrix effects. This can make precise determination and quantification difficult. This highlights a trade-off between speed and analytical precision, depending on the specific needs of the investigation. As a result, it may be less suitable for complicated mixtures where exact separation is crucial [72].
4.2.4 Thin layer chromatographyThin layer chromatography (TLC) was used to identify flavonoids in tea, and the multi-imaging and effect-directed reporting method was employed to identify dietary components, bioactive chemicals, adulterants, residues, and pollutants in tea samples [73]. High-performance thin-layer chromatography (HPTLC) is an improved version of TLC that increases chemical separation. HPTLC was used to quantify caffeine, chlorogenic acid, gallic acid, trans-cinnamic acid, and epigallocatechin gallate in the extracts produced by ultrasound-assisted extraction using different extraction parameters. The extraction value was 16.8%, as follows the parameter crude drug-to-solvent ratio was 0.199 g/mL, temperature 39.9°C for 29.9 mins [74]. Caffeine content was also evaluated in a similar study utilizing liquid-liquid extraction and TLC with a 9:1 solvent ratio of chloroform and methanol. Caffeine concentrations in 25 g of each of the five tea samples were reported to be 1.1%, 0.8%, 0.8%, 2%, and 0.9%, respectively [75].
TLC offers a simple, cost-effective method for viewing and separating chemicals, allowing for the simultaneous analysis of many samples. The simplicity and affordability of this approach make it very beneficial. However, it has limitations in terms of sensitivity and quantification accuracy compared to more sophisticated methods like High-Performance Liquid Chromatography (HPLC). Because TLC is manual, it can lead to human errors, which might compromise the accuracy of the results [76]. The direct detection method in TLC requires less equipment and speeds up the initial screening process by reducing analysis time and operating expenses. However, because of its low sensitivity and specificity, this method is unsuitable for thoroughly examining complex sample matrices that have not been separated beforehand. While TLC is efficient and accessible, its limited analytical accuracy restricts its use [77].
4.3 Utilization of nanoparticles for detection of tea contaminantsThe gold nanoparticles-based immunochromatographic assay (AuNPs-ICA) was employed in research to identify tea pollutants and contaminants [39]. A simple hydrothermal technique has been used to synthesize novel fluorescent organic nanoparticles (FONs)-based sensor P-M(w) from 1-Pyrenecarboxaldehyde and L-methionine for the detection of mercuric ion (Hg2+) in dark, green, yellow, oolong, white, and black tea. The FL emission peak of P-M(w) at 380 nm would drop in the presence of Hg2+, whereas the peak at around 425 nm would grow because of PET and the metal ion coordination-induced nano-structure conformational rigidification [78].
Gold nanoparticles (AuNPs) and surface-enhanced Raman spectroscopy (SERS) technologies were coupled to identify and measure pesticides in oolong tea. AuNPs with an average diameter of 15 nm were easily synthesized spherically and monodispersed, resulting in strong electromagnetic fields during SERS studies. Carbendazim in Tieguanyin oolong tea was quickly detected and measured using AuNP substrates. The detected limit was 100 μg/kg, and the R-value for the PLS results was 0.964. These outcomes showed that SERS, in conjunction with an AuNP substrate, is an easy, quick, and accurate analytical method for determining the concentration of carbendazim residues in oolong tea [79].
Using nanoparticles in the separation first approach enhances its specificity and sensitivity when combined with separation methods. This allows for the precise identification and removal of specific pollutants. However, the involvement of multiple stages in the separation process makes it more complex, time-consuming, and costly. Additionally, it requires technical expertise [80]. On the other hand, direct detection using nanoparticles is simpler for quick screening because it can be done quickly, easily, and affordably. However, matrix components can interfere with it and may not give comprehensive details on complex mixtures without separation [81].
4.4 Utilization of nanozymes for detection of tea contaminantsNanozymes (enzyme-mimetic nanomaterials) are new quality and safety detection instruments in agri-food businesses. They meet the inspection requirements of low cost, high sensitivity, specificity, and repeatability [82]. Three different types of nanozymes, iron (II, III) oxide (Fe3O4) nanoparticles, manganese (II, III) oxide (Mn3O4) octahedrons, and protein-linked gold nanoclusters (protein-Au NCs), with LOD at nano-molar (nM), were employed to measure the amounts of ethoprophos, tannic acid, and tea polyphenols [83]. According to the study, polyphenols in green tea can be detected using nanozymes and colorimetric sensors, and an iron carbide/iron-nitrogen-carbon (Fe3C/Fe-N-C) catalyst with excellent oxidase activity was developed. Tetramethylbenzidine (TMB) can be catalytically oxidized to blue oxidized TMB (oxTMB) and used to identify tea polyphenols, including ECG and EC. The chromogenic sensing platform based on TMB, tea polyphenol monomers, and Fe3C/Fe-N-C was used to further detect silver ions (Ag+). These continuous measurement channels serve as sensor arrays at concentrations of 50 𝜇M. The pattern recognition approach can distinguish between polyphenols [3].
The sensor design enabled the precise detection of polyphenols combined in diverse species, concentrations, or ratios. Color changes can be produced by a nanozyme with polyphenol oxidase activity catalyzing the interaction between tea polyphenols and 4-amino antipyrine (4-AAP). A dual output model based on machine learning (ML) was used to predict the classes and concentrations of unidentified samples, resulting in the detection of tea polyphenols. Chinese tea identification was accomplished by merging the sensor array with the dual output ML model to detect specialized tea types [84].
Nanozymes offer high specificity and sensitivity with nanomolar detection limits for substances like ethoprophos, tannic acid, and tea polyphenols. This approach increases detection precision by using catalysts like iron carbide/iron-nitrogen-carbon (Fe₃C/Fe-N-C) and enhancing the efficiency of colorimetric sensors like tetramethylbenzidine (TMB) oxidation. It does, however, need more complicated reagents and equipment, which increases operating costs and adds to the time and complexity of sample preparation [85]. This detection method adds nanozymes directly to the tea samples without separating them first, allowing for quick analysis and economical screening of impurities such as silver ions (Ag⁺) with little equipment needed. In comparison to the separation-first strategy, this method yields data more quickly and at a reduced operation cost. However, it may have poorer sensitivity and specificity, resulting in fewer precise measurements in complex samples because of matrix effects [86].
4.5 Enzyme-linked immunosorbent assay for detection of tea contaminantsThe enzyme-linked immunosorbent assay (ELISA) detects mycotoxins, antibodies, antigens, amino acids, and glycoproteins in large samples quickly [87]. Nanomaterial-based ELISA (nano-ELISA) significantly outperformed the standard ELISA [88]. ELISA was used in a study to analyze ochratoxin A (OTA) and total aflatoxins (AFs) in green and black tea samples obtained from Lebanon’s markets [89]. After screening many coating antigens, a coating antigen/antibody combination with excellent specificity and sensitivity based on an aminopyrine polyclonal antibody was found. The LOD of ELISA was 0.18 ng/mL, much lower than the standard value of 100 ng/mL [90].
The early isolation of target contaminants from the tea matrix minimizes matrix interference and cross-reactivity, resulting in optimum sensitivity and specificity. Therefore, it works well with complicated tea samples where matrix components could normally hamper detection. However, because specific antibodies, reagents, and extra processes are needed, it consumes time and is considered costly [91]. However, due to their quicker and more affordable nature, direct detection techniques such as sandwich ELISA are more suitable for large-scale screening. These techniques offer the advantages of speed and affordability. However, their specificity may be reduced in complex tea matrices, leading to increased potential for cross-reactivity and false positives [92].
4.6 Vibrational spectroscopic techniques for detection of tea contaminantsSpectroscopy is the measurement of radiation strength as a function of wavelength. It is employed in analytical chemistry because atoms and molecules have unique spectra used to detect, identify, and quantify information about samples [83]. Non-destructive spectroscopy can offer information about a food item by scanning without causing it to be damaged. As a result, spectral detection technology has been extensively researched recently and has advanced continuously [83, 93]. The vibrational spectrum can be used as a fingerprint for tea sample quantification, qualification, characterization, and structural elucidation [94].
In the analysis of contaminants in tea, vibrational spectroscopic techniques offer enhanced sensitivity and specificity by integrating detailed structural information of contaminants. This method involves the initial separation of target compounds from the complex tea matrix, thereby improving detection accuracy. However, it is characterized by longer analysis times and higher costs due to the need for sophisticated equipment and specialized expertise [95]. Conversely, the direct detection approach is rapid, cost-effective, and non-destructive, requiring minimal sample preparation. Despite these advantages, it has limited specificity, which can be compromised by matrix interference, making it less effective for complex tea samples where separation is not employed. The choice between these methods depends on the balance between analytical precision and practical considerations such as time, cost, and complexity of the tea matrix [96].
4.6.1 Infrared spectroscopyInfrared (IR) spectroscopy is commonly recognized as one of the most effective techniques for assessing chemical components based on functional group frequency absorption [97]. When the sample is exposed to IR radiations, its molecules absorb a certain frequency, transferring vibrational and rotational energy levels. The IR absorption spectrum of exposed samples can be obtained by measuring the absorption of IR light [97]. The method provides increased penetration into a material due to a spectral window in the IR band ranging from 650 - 900 nm [98]. However, the data produced is frequently large and collinear, so comprehending IR findings may be challenging. Several chemometric methods based on latent variables (LVs) have been developed to address this issue, including principal component analysis (PCA), principal component regression (PCR), partial least squares (PLS), support vector machines (SVM), random forest (RF), and variable selection methods such as competitive adaptive reweighted sampling and successful projections algorithm [97].
Sunset yellow (SY), a prevalent adulterant in tea powders, was investigated in tea samples utilizing Fourier Transform Infrared (FT-IR) spectrum data and ML algorithms. The recursive combination genetic algorithm (RCGA) chose SY distinctive 20, 30, 40, 50, and 60 wavenumbers in the FT-IR spectra. Support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGB) classifiers were used for classification, and among the three, an SVM classifier with 50 variables provided an accuracy of 0.90. PLS, least squares-SVM (LS-SVM), RF, and XGB models were used to quantify SY, with both RF and LS-SVM models outperforming PLS when combined with RCGA to generate 20 fingerprint variables. When using 50 fingerprint variables, the model (RCGA-LS-SVM) exhibited the lowest RMSECV (0.1956) and regression coefficients of RC2 = 0.9989 and RP2 = 0.9979. The results demonstrated that FT-IR and the RCGA-LS-SVM were reliable and rapid methods for recognizing SY in tea powder [99].
Visible and NIR (Vis-NIR) spectroscopy is used to determine the quality of food samples to identify the adulterants present and their sources [100]. This method was also used to determine the amount of Pb in tea leaves through stomata. The treated samples’ spectral variation was measured [101]. An apparatus for detecting polyphenols in fresh tea leaves was designed using a combination of visible short-wave and long-wave NIR and data gathering within the overlapping spectrum (1000–1050 nm). With the development of this advanced instrument, prediction models and sensitive wavelengths for detecting tea polyphenol concentration are also presented [102]. The advantages of this method are its low cost, convenience of use, and ability to establish a direct relationship between absorbance and concentration of organic and inorganic chemicals. This approach has one limitation, as it can only test liquid samples. Furthermore, while employing this approach, factors like pH, pollution, temperature, and contaminants can all significantly influence the findings.
The separation-first method induces rotational and vibrational energy transitions by subjecting materials to infrared light. Because of the IR band (650–900 nm), this approach enables higher penetration; nevertheless, it generates massive, collinear data that need to be analyzed using chemometric techniques [103]. For example, pollutants like sunset yellow can be accurately detected in tea powders using FT-IR spectroscopy in conjunction with machine learning methods like the recursive combination genetic algorithm (RCGA) and SVM classifiers [104]. However, this approach requires complicated sample preparation and is time-consuming, which raises operating costs. This approach establishes a clear correlation between chemical concentration and absorbance and is both easy and economical. Its use is restricted to liquid samples, and its accuracy may be jeopardized by variables such as temperature, pH, pollution, and pollutants [105].
4.6.2 Molecular spectra and chemometricsInteractions between electromagnetic waves and samples are studied in molecular spectroscopy (MS) using radio waves, microwaves, ultraviolet light waves, X-rays, and gamma [106]. Chemometrics is the quantification of chemical matrices that deliver suitable information about chemical data after analysis. It analyses data using mathematical, statistical, and ML methods [107]. In Indonesia, MS was used to verify Java Preanger steaming green tea (JPGT). The UV spectra of tea samples were scanned at 190–400 nm, and chemometrics such as PCA, partial least squares discriminate analysis (PLS-DA), linear discriminate analysis (LDA), and PCA-LDA were calculated. The results demonstrated that chemometric PCA can separate JPGT samples from non-JPGT samples [108]. This approach was used in several nations to monitor tea safety, including the verification of geographical tea suppliers [109].
Tea fraud is common and involves altering the labeling of inferior items or those without geographical origin certification and blending them with graded teas. A Chemometrics-Assisted Color Histogram-based Analytical System (CACHAS) is used to screen the quality of teas as a simple, cost-effective, and environmentally friendly analytical approach. Class Analogy software was used to confirm their geographical origin and category simultaneously, precisely recognizing all Argentinean and Sri Lankan black teas and Argentinean green teas. PLS exhibited good predictive value for the assessment of moisture, total polyphenols, and caffeine, with RMSEP values of 0.50 mg/kg, 0.788 mg/kg, and 0.25 mg/kg, and REP values of 6.38%, 9.031%, and 14.58%, respectively [110]. Molecular spectra and chemometrics offer high specificity, sensitivity, and detailed analysis but are time-consuming, complex, and costly [111]. The choice of this method depends on the specific analysis requirements and available resources [112].
4.6.3 Fluorescence spectroscopyA tried-and-true method that is easy to use, quick, precise, and non-destructive is fluorescence spectroscopy. Because of its great sensitivity and non-destructive nature may be used to continuously measure quality at different stages of the tea manufacturing process [113]. For both qualitative and quantitative research, it can offer a sizable number of physical parameters. Additionally, the sensitivity of the fluorescence analysis is typically 2–3 orders of magnitude higher than that of a spectrophotometer, which is a characteristic benefit. Under different experimental settings, it also permits nondestructive assessments for compounds with low concentrations [114]. With minimal use of hazardous chemical solvents, this approach has demonstrated exceptional sensitivity and selectivity for organic contaminants, indicating its potential as a green analytical methodology [115].
A laser beam is used in fluorescence spectroscopy (FS) to excite electrons in test compounds, causing the electrons to emit light. That light is directed at a filter and a measuring detector, which detects the molecule or any changes in the molecular structure [116]. This technique directly analyzes the quality of solid tea infusions. A luminous europium-based metal-organic framework (Eu-TFPA-MOF) with two ligands, 1,10-phenanthroline (Phen) and tetrafluorophthalic acid (H2TFPA), was developed to detect fipronil in genuine tea samples. Fipronil extinguished the sensor’s red fluorescence, and the detection process was characterized by excellent sensitivity and selectivity, a low detection limit (4.4 nM, 1.9 μg/kg), a wide linear range (0–0.15 mM), and a fast reaction time (2 min). The findings reported FS can be used to detect fipronil in tea samples. The sensor’s superior ethanol resilience allows it to have remarkable recovery properties (98.33–106.17%) for green and oolong tea with repeated ethanol washing [117]. The grade of tea is an essential criterion for determining its commercial value. For this, integrated excitation-emission matrix-fluorescence spectroscopy (EEM-FS) with three different algorithms models, including PLS-DA, multi-way partial least squares discriminant analysis (N-PLS-DA), and unfolded partial least squares discriminant analysis (U-PLS-DA) was used with a recovery rate 85.1% to 97.8% for both training and test sets. Complete EEM-FS paired with multi-way classification algorithms has shown to be an excellent method for assessing the grade of green tea [3].
4.6.4 Laser induced breakdown spectroscopyLaser-induced breakdown spectroscopy (LIBS) is a rapid atomic-based emission spectrometry [118]. LIBS is a versatile elemental analytical approach used for industrial applications and scientific research in recent years. Compared to conventional elemental analytical methods, LIBS significantly benefits material composition identification [119]. With the benefits of simultaneous multi-element analysis, real-time and quick detection, easy 98detection technique. Element analysis and detection for solids, liquids, gasses, and aerosols can be done with LIBS [120]. LIBS’s unique benefits are that it requires very less material for analysis, requires little to no sample preparation, and produces no waste or consumables [3]. Although LIBS can be used to analyze any sample, direct multi-elemental analysis of solid materials has been its most common usage [121].
LIBS was used to analyze the spectra of four distinct tea samples and performed a semi-quantitative analysis of magnesium (Mg), calcium (Ca), aluminum (Al), manganese (Mn), sodium (Na), and potassium (K) [122]. Comparable research also examined the spectra of four different tea samples with a quantitative analysis of Mg, Mn, Ca, Al, Na, and K [123]. LIBS was employed to assess the elemental composition of Tieguanyin tea and lead chrome green (LCG). Fe, Ca, and Mg were revealed to be present in both LCG and tea, with chromium (Cr), Pb, and Cu particular to LCG and Na, K, Al, silicon (Si), Mn, and strontium (Sr) specific to tea. The free radicals, cyanide (CN), calcium oxide (CaO), and carbon dimer radical (C2) were studied in the LIBS spectra of tea and LCG-tea samples. The CN and CaO spectral lines were used in PCA to identify ordinary tea from LCG tea. Compared to PCA results using all atomic spectral lines, the cumulative contribution rate of the two PCs increased from 90.7% to 99.4% [124].
4.6.5 Inductive couple plasma-mass spectrometryInductively coupled plasma mass spectrometry (ICP–MS) quickly leads to determining various elements in plant samples. ICP-MS can measure various elements simultaneously with speed and accuracy, and it has the widest linear range (nine orders of magnitude), maximum sensitivity, and lowest detection limit for metals [125]. However, using ICP-MS directly to identify trace elements in an actual sample can only provide information on the element’s overall concentration. Before analysis, an efficient separation and pre-concentration approach for various species is typically needed to gain information about their chemical form [126].
The increasing consumer demand for tea with authentic geographical origins needs trustworthy geographic authenticity measures. ML algorithms and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) were employed in a study to evaluate the geographical authenticity of Xinyang Maojian tea. When algorithms were used for pattern recognition of the mineral components of collected tea samples in this study, the accuracy in discriminating geographical authenticity was 100%, suggesting that ML algorithms combined with ICP-MS can be utilized to detect the geographical authenticity of Xinyang Maojian tea [127]. In a study, 35 tea (Camellia sinensis) samples were assessed to detect fluoride through fluoride ion selective electrodes and Al via ICP-optical emission spectroscopy (ICP-OES) [28].
4.7 Metabolomics for ensuring tea safetyThe complex chemical constituents of tea can be measured more thoroughly due to the application of metabolomics technology [128]. It is an effective method for objectively screening chemicals and has been applied to distinguish between tea samples from different origins or processing techniques. Metabolomics is commonly used in conjunction with either post-separation chromatographic techniques, such as gas chromatography-mass spectrometry (GC–MS) and liquid chromatography-mass spectrometry (LC–MS), or direct spectroscopic measurement methods, such as nuclear magnetic resonance (NMR) and infrared (IR). It effectively identifies the phytochemical makeup of various tea origins, varietals, or products. When identifying quality signals in various tea grades, metabolomics can be far more helpful than traditional analytical methods [23]. However, there is a shortage of knowledge about assessing food adulteration by metabolome profiles despite the broad and relatively extensive discussion of the applications of metabolomics in evaluating the safety and quality of foods [129].
Metabolomics provides comprehensive knowledge about key variables, such as the chemical makeup of different teas and the processing effect on tea chemical composition. Hence, metabolomics is well recognized as an effective tool for evaluating the quality of tea products [23]. Metabolomics is often used in conjunction with direct spectroscopic measurements such as nuclear magnetic resonance (NMR) and IR and post-separation chromatographic techniques such as GC-MS and LC-MS [130]. Multivariate data analysis, such as PCA and hierarchical cluster analysis (HCA), can assist in understanding the influence of cultivation and processing procedures on green and black tea features when resolving plant metabolomes using either hyphenated or spectroscopic approaches.
The influence of varied sun-withering degrees on black tea sensory quality was investigated using metabolomics analysis. Moreover, 65 non-volatile components were discovered using Ultra Performance Liquid Chromatography-Quadrupole-Time of Flight-Mass Spectrometry (UPLC-Q-TOF/MS). Results showed that the freshness and sweetness of black tea can be enhanced by boosting the amount of amino acids and theaflavins. The aroma of the tea was analyzed using a combination of Solvent Assisted Flavor Evaporation-Gas Chromatography-Mass Spectrometry (SAFE-GC-MS) and Headspace-Solid Phase Micro Extract-Gas Chromatography-Mass Spectrometry (HS-SPME-GC-MS), and 180 volatiles were identified. Statistical analysis revealed 11 volatiles as potential major aroma differential metabolites in black tea, including volatile terpenoids (linalool, (E)-citral, geraniol, and myrcene), amino-acid-derived volatiles (benzene ethanol, benzeneacetaldehyde, and methyl salicylate), carotenoid-derived volatiles (jasmone and damas) [131].
4.8 Utilization of biosensors for ensuring tea safetyAgriculture, food processing, animal breeding, and trade have all benefited from biosensor technology. Using biosensors enables the rapid, convenient, and cost-effective capture of precise quality data detection of contamination in tea at various stages of manufacturing and processing [132]. The electronic nose (E-nose), or artificial olfaction, is a gas-sensitive sensor that can detect a wide range of chemical components while mimicking the olfactory function of the human nose. When an olfactory stimulus is provided, an e-nose system generates fingerprints from a sensor array of heterogeneous gas-specific sensors. A pattern recognition computer is trained using fingerprints acquired from recognized odors, followed by identifying unfamiliar odors to assess the degree of degradation [133]. The gases released are absorbed by the sensor in the e-nose, which produces a signal. This signal may be recognized using several statistical approaches; however, with the advent of artificial intelligence (AI), ML algorithms are being utilized to assess food quality [134]. Seven cross-sensitive electrodes detect umami, richness, sourness, saltiness, sweetness, bitterness, and astringency [135].
E-nose, responsible for the tea’s distinct flavor, color, and aroma, was used to detect the tea’s fermentation stage. A fast gas phase electronic nose (GC-E-Nose) combined with multivariate statistical analysis was used to examine the fragrance quality of 44 Dianhong black tea (DBT) infusions. Aldehydes were the most common of the 61 volatile compounds found [2]. To identify polyphenols in tea, e-nose sensors were designed to detect changing polyphenol concentrations [3]. Given the inconsistency of findings obtained while evaluating carbendazim, a fungicide, with modern electrochemical sensors, a novel electrochemical sensor helped by ML was developed. In the presence of additional pesticides, copper (II) oxide, reduced graphene oxide, and poly N-phenylglycine electrodes demonstrated high carbendazim sensitivity [136].
From March to May, skilled workers picked Longjing tea samples as their quality is linked to plucking time. Tea leaves gathered early have a better tea grade and a higher price. To prepare tea leaf samples, 5 g of tea leaves were inserted in a 500 mL beaker, and for infusion, samples were combined in 250 mL boiling water with 5 g of tea leaves and brewed for 5 mins. E-nose data was then classified using SVM and logistic regression (LR). LR showed 83% accuracy for tea leaf but only 72% for tea infusion. When the data dimension was reduced using LDA and classified using LR, tea leaf and tea infusion accuracies increased to 86% [137]. E-nose uses various sensor arrays to gather information about aroma compounds. This information is useful in identifying changes in tea odor throughout processing and differentiating the tea’s aroma quality [138]. E-nose does not require difficult sample preparation and is not time-consuming, expensive, labor-intensive, or constrained. Many materials/sensing principles, including conducting polymers, quartz crystal microbalance, amperometry, electrochemical, surface acoustic wave, and metal oxide semiconductor sensors, are typically used in conjunction with this non-invasive, intelligent online instrument. Therefore, each e-nose’s sensitivity, selectivity, efficiency, and response speed can be significantly impacted by the materials and sensors that are used [135].
4.9 Computer vision and imaging techniques for detection of tea contaminantComputer vision develops autonomous systems capable of mimicking human visual abilities. It emerged as a core notion to check categorized tea products based on characteristics [83]. It involves picture collecting, pre-processing, feature extraction, detection, high-level processing, and decision-making. A novel model, You Only Look Once version 5 (YOLOv5), combined with self-attention and convolution (ACmix) and convolutional block attention module (CBAM), was introduced to address the difficulty of diagnosing tea tree leaf ailments and insect pests. YOLO-Tea outperformed YOLOv5s by 0.3% to 15% across all the data tested, with better results than Faster region-based convolutional neural network (R-CNN) and single shot multibox detector (SSD). Its superior performance suggests that YOLO-Tea is ideal for real-world tree disease monitoring systems [139].
Hyperspectral imaging (HSI) methods can be used as a quality control tool for subsequent differentiation in tea bag blends or loose tea. Scanning the tea bag and inserting the picture into the PLS-DA calibration model makes it feasible to quantitatively regulate the quantity of each raw material integral component using an appropriate chemometric method. The precision with which duplicates were calculated determined the strength of these models. It is feasible to collect 3D data, which is uncommon to occur using typical analytic procedures. HIS combines traditional imaging with vibration spectrum technology to collect spatial and spectral information about a subject, its surface quality, and biological makeup using tea leaves.
Changes in plant structure, material composition, and physiological and biochemical status can be detected by a hyperspectral imaging system (HSIS). HSI collects information in the form of pictures using clear spectrum bands and merges photos into three-dimensional hyperspectral data cubes that are processed and analyzed. Coordinates can be any value in the electromagnetic spectrum, depending on the availability of light sources, detectors, and specialized applications [140]. The tea quality components (polyphenols, amino acids, and volatile organic compounds) and the appearance and texture of black tea products can be examined in real-time during black tea fermentation, drying, and fragrance extraction. Tibetan tea is a black tea native to Ya’an, and its flavor quality is strongly related to the concentration of tea polyphenols (TPs) and free amino acids present (FAAs). Chemometrics was employed to identify TPs and FAAs and to minimize feature dimensionality. HIS data were collected and pre-processed using PCA. Savitzky Golay (SG)-Standard Normal Variable (SNV)-PCA-Extratree has the best prediction ability for TPs (Rp2=0.9248, RMSEP=0.4842, and RPD=3.646). The greatest predictive capacity for distinguishing FAAs was demonstrated by SG-Multiplicative Scatter Correction (MSC)-PCA-Extratree (Rp2=0.8736, RMSEP=0.159, and RPD =2.813). Furthermore, the SG-MSC-PCA-Support Vector Machine can evaluate tea grade with 100% accuracy [141]. One notable feature of hyperspectral imaging (HSI) technology is its capacity to deliver rich and real-time information [142]. It can obtain hundreds of spectra with great spectral and spatial resolution; it can provide spectral and spatial information on each pixel across a certain wavelength range [143].
4.10 Terahertz algorithms for detection of tea contaminantThe non-destructive, non-contact transmission-type terahertz spectrometer is capable of quickly identifying samples. Terahertz waves are characterized by two properties: electromagnetic and light electronics. Because of these properties, the terahertz time-domain spectrometer (THz-TDS) is an extremely helpful tool and method for examining material behavior. It may also be used in biomedicine, non-destructive materials testing, and security [144]. However, detecting trace chemicals with high toxicity has shown to be challenging with conventional terahertz spectroscopy. The divergence between the THz wave’s wavelength and the target molecule’s size may constrain the sensitivity of terahertz radiation used to detect trace analytes in free space [145].
Terahertz (THz) is an electromagnetic wave between microwave and IR light with a wavelength range of 0.03 – 30 mm. The energy level in this region is about 1–10 meV, equivalent to one-millionth of an x-ray, so THz will not lead to photoionization radiation damage to the biological tissues. THz wave can pierce food packaging and capture spectral and image information from the food matrix, which is an effective food safety detection approach [146]. THz time-domain spectroscopy (THz-TDS) simultaneously measures target objects’ amplitude and time delay signals. Fast Fourier transform THz can analyze parameters such as refractive index, absorption coefficient, and dielectric constant for qualitative and quantitative analysis [147, 148]. Before calculating the absorption coefficient, it is crucial to determine the thickness of each sample, as the THz absorption coefficient inversely correlates with it [149]. Current X-ray technology detects hard metal objects effectively, but soft foreign bodies remain a safety challenge in tea. THz-TDS plus baseline correction and variable selection identified insect and small-density foreign bodies in tea. However, the THz-TDS system can be installed on the tea packaging line to detect foreign bodies [150].
4.11 Stochastic multi-level bayesian machine learning methods for tea quality detectionIn food safety investigations, Bayesian approaches employ prior distributions and efficient algorithms to estimate post-distributions of parameters in a model, allowing for uncertainty induced by numerous factors such as sampling and processing procedures. These methods are reliable for inferring phenomena of interest based on sample data [151]. Bayesian multi-level models estimate random effects and other parameters [152]. In contrast to conventional frequentist statistics, which examine a parameter of interest using a point estimate, Bayesian approaches integrate previous distributions and provide effective procedures to estimate post-distributions of model parameters. Such techniques are thought to be trustworthy for deducing the phenomenon of interest across the entire population based on the observed data for samples because they account for the uncertainty of results caused by various factors (such as sampling, processing, and detecting methods and procedures in the field of food safety investigation). Given these benefits, Bayesian techniques are frequently used in the fields of quality control and research on food safety [153].
China is the world’s largest pesticide user to reduce the severity of insect, sickness, and weed infestations on teas [154]. China employed a Multi-level Bayesian approach to provide reliable information for analyzing the quality and presence of additives and pollutants in different tea varieties available in the market [153]. The study found that the overall rate of nonconforming tea was 2.6%. Pesticides were the main cause of nonconformity in yellow, green, and oolong teas. Illegal additives were more common in black tea. Another study conducted in China used descriptive statistical analysis of heavy metals in tea and tea plantation soils by searching the China National Knowledge Infrastructure/Institute for Scientific Information (CNKI/ISI) website with the keyword tea or heavy metals in tea plantation soils. The results showed that the average contamination of six heavy metals in tea was about 0.21, 0.14, 1.17, 14.6, 0.04, and 1.09 mg/kg for As, Cd, Cr, Cu, Hg, and Pb, respectively, however, the heavy metals risks in tea were all under acceptable ranges [154].
A revolutionary d-SPE-based approach is the multi-plug filtration cleaning (m-PFC) technique. In the m-PFC column, the sorbents are packed tightly between two sieve plates in a short column comparable to an SPE cartridge. While pushing and withdrawing the piston, the extract interacts with the sorbents, allowing the interfering chemicals to be adsorbed and successfully used to identify pesticide residues from complicated matrices. Because it does not need solvent evaporation, vertexing, or centrifugation, the m-PFC method is quicker than the d-SPE method in identifying pesticide residues [155]. Further, details on the tea safety studies using combined and tandem analytical and detection approaches are presented in Table 1, and their advantages, disadvantages, and validity parameters are highlighted in Table 2.
| Samples/Product | Parameters or adulterant | Detection Techniques | Detection Parameters | Results | Ref. |
|---|---|---|---|---|---|
| 45 black tea samples and 15 green tea samples | Mycotoxins and aflatoxins (B1, B2, G1 and G2) and ochratoxin | HPLC | C18 as reversed phase. Mobile phase of acetonitrile methanol-water (2:3:5 v/v) Fluorescence detector excitation wavelength of 365 nm and an emission wavelength of 435 nm |
18 aflatoxins and 41 ochratoxins-positive black tea samples 2 aflatoxins and 11 ochratoxins positive green tea samples |
[156] |
| Tea brew (Black tea leaves) |
Heavy metals (Co, As, Ag, Cd, In, Hg, and Pb) | Perkin Elmer Ne×Ion-300× ICP-MS | RF power/ 1500 W, Plasma gas flow (18 L/min), Auxiliary gas flow (1.2 L/min) | Co, As Cd, In, and Pb (μg/kg) detected high among samples Co (71.9–214.3) As (1.2–53.7) Cd (1.2–53.7) In (1.2–53.7) Pb (61.0–2404.3) |
[157] |
| Oolong tea | Flusilazole residues (pesticide) | Surface-Enhanced Raman Spectroscopy | Nanocomposites consisted of CNF coated with AgNPs CNF-AgNP substrates used for measurement of flusilazole in Oolong tea samples by SERS |
LOD of 0.5 mg/kg for flusilazole was obtained | [158] |
| Green tea | Imidacloprid (Insecticides) | Surface-enhanced Raman Scattering coupled with chemometric algorithm | Imidacloprid yielded a SERS signal after adsorption on Ag-NF under laser excitation GA-PLS (Genetic algorithm-partial least square) used to quantify imidacloprid residue |
The model exhibited an Rp of 0.9702 RPD of 4.95 % in the test set RSD for precision to 4.50 % |
[159] |
| Tea based | Talcum powder | FT-IR spectroscopy coupled with chemometrics | A hybrid of biPLS regression, CARS algorithm, and SPA to select optimal 19 wavenumbers | Among wavenumbers, 1016 (Si-O-Si), 1182 and 1249 (CO-O), 1340 (CO32–), 1631(OH) and 2296 (Si-H) cm−1 were wavenumbers of talcum powder | [160] |
| Tea infusion | Lead Chrome Green (LCG) | Raman spectroscopy combined with chemometrics methods | PLS and least square support vector machine (LS-SVM) were used to build the model Raman wavenumbers, 1341, 1451, 1527, and 1593 cm-1 used for LCG |
LS-SVM presented better results with R2 of 0.964 and the root mean square error of prediction of 0.535 Tea sample detected with illegally added LCG |
[161] |
| Dark tea | Mycotoxins (AFs, OTA, ZEA, DON, FBs) | HPLC | Multi-functional column and immunoaffinity column combined with HPLC | Aflatoxins were detected below the acceptable carcinogenic risk level | [162] |
| Green tea | 15 mycotoxins and 4 aflatoxins | LC-MS | For method validation parameters used recoveries, matrix effect, linearity, LOD and LOQ | 56% of samples were contaminated Mycotoxins in samples were AOH (40%), ZEN (35%), AFG1 (2%), AFB2 (2%), ENB (2%), TENT (1%), and ZEN with a value of 45.8 ng/g |
[163] |
| Flavored teas |
Pyrrolizidine alkaloids (PAs) | LC/MS coupled with chemical extraction | Extraction using sulfuric acid and methanol Cleaning using Oasis MCX SPE cartridges C18 column with gradient elution |
PAs in teas were relatively high in lemon, balm, peppermint, and mixed teas, having senecionine and senecionine N-oxide | [33] |
| Tea leaves | Polyphenols, amino acids | NIR spectrometer coupled with smartphone | PLS modeling coupled with spectral pre-processing | The coefficients of the prediction set for tea polyphenols, amino acids, and the P/A value were 0.90, 0.91, and 0.91 Residual predictive deviations were 2.24%, 2.43%, and 2.42%, respectively |
[164] |
| Tea samples Flavored teas |
Neonicotinoid (insecticide) Pyrrolizidine alkaloids (PAs) |
LC-MSLC/MS coupled with chemical extraction | Analytical Accucore C18 column for separation of NEOs Extraction using sulfuric acid and methanol Cleaning using Oasis MCX SPE cartridges C18 column with gradient elution |
High detection frequencies and concentration residues of NEOs in collected tea samples PAs in teas were relatively high in lemon, balm, peppermint, and mixed teas, having senecionine and senecionine N-oxide |
[32, 49] |
| Tea samples Tea leaves |
Caffeine, Polyphenols, amino acids | TLC-NIR spectrometer coupled with smartphone | Lead (II) acetate used to separate tannins from caffeine liquid-liquid extraction dichloromethane and sodium sulfate as a drying agent Silica gel plates Mobile phases were glacial acetic acid and ethyl acetate (95:5, v/v) Second mobile phase ethyl acetate and ethanol (80:20, v/v) PLS modeling coupled with spectral pre-processing |
Rf value of the first phase was 0.36 Rf value for the second phase was 0.86 pH of boiled sample teas was 4.85 to 5.80 Green tea reported abundant caffeine of 2.04 % The coefficients of the prediction set for tea polyphenols, amino acids, and the P/A value were 0.90, 0.91, and 0.91 The residual predictive deviations were 2.24%, 2.43%, and 2.42%, respectively |
[164, 165] |
| Green tea Tea samples |
Quality grading Neonicotinoid (insecticide) |
E-nose, LC-MS |
PLSR model MBPNN, SVM, and RF Analytical Accucore C18 column for separation of NEOs |
99%, 99%, and 97% accuracy MBPNN and SVM achieved 99%, and RF exhibited 97% accuracy high detection frequencies and concentration residues of NEOs in collected tea samples |
[49, 166] |
| Pu-erh teas Tea samples |
Volatile metabolites and aroma quality Caffeine | GC-E-Nose, GC–MS, GC-IMSTLC | 43 volatile components were identified by GC–MS 91 volatile substances were detected by GC-IMS Model PLS-DA Lead (II) acetate used to separate tannins from caffeine liquid-liquid extraction dichloromethane and sodium sulfate as a drying agent Silica gel plates |
9 flavor compounds, including linalool, (E)-2-hexenal, 2-hexenal, 2-methyl butyl acetate and terpinene-4-ol, cyclohexanone, 2-butoxyethanol, 2-octanol, and 2-isopropyl-3-methoxypyrazine detected Rf value of the first phase was 0.36 Rf value for the second phase was 0.86 pH of boiled sample teas was 4.85 to 5.80 Green tea reported abundant caffeine of 2.04 % |
[164, 167] |
| Green tea | Quality grading | E-nose | PLSR model MBPNN SVM and RF |
99%, 99%, and 97% accuracy MBPNN and SVM achieved 99%, and RF exhibited 97% accuracy |
[166] |
| Pu-erh teas | Volatile metabolites and aroma quality | GC-E-Nose, GC-MS, GC-IMS |
43 volatile components were identified by GC–MS 91 volatile substances were detected by GC-IMS Model PLS-DA |
9 flavor compounds, including linalool, (E)-2-hexenal, 2-hexenal, 2-methyl butyl acetate and terpinene-4-ol, cyclohexanone, 2-butoxyethanol, 2-octanol, and 2-isopropyl-3-methoxypyrazine detected | [167] |
| Sample | Analyte | Analyte Extraction Mode | Technique/ Method |
LOD | LOQ | Advantages | Disadvantages | Ref. |
|---|---|---|---|---|---|---|---|---|
| Green tea | Pesticides | Direct solvent extraction using acetonitrile and methanol | UHPLC-HRMS with multivariate analysis on C18-PFP column | 10 g/kg | 100 g/kg | Perform rapid screening of contaminants. The blind detection rate of tracers annotated in positive mode was about 44% and 38% in negative mode. Overall, the analyte detection rate is 66%. |
Pretreatment of samples chemicals Can produce noisy peaks |
[168] |
| Green tea | Lead chrome green | Extraction using boiling water | Raman Spectroscopy | 0.651mg/g | - | Non-destructive and time-effective technique. Integrated intensity of full range (2804 cm−1 to 230 cm−1) was proven effective for quantitative detection of lead chrome green color. |
High installation cost Can better perform qualitative analysis of analytes Can detect only one particular ion or analyte at once |
[169] |
| Tea samples | Aflatoxins (μg/kg) (Afs; AFB1, AFG1, AFB2 and AFG2), Ochratoxin A (OTA) |
Solvent extraction using 1 g of NaCl and 80% methanol | HPLC | AFB1 (0.1), AFB2 (0.2), AFG1 (0.3), AFG2 (0.2), Total Afs: (0.6) OTA (0.47) |
AFB1 (0.4), AFB2 (0.7), AFG (0.9), AFG2 (0.6) Total Afs: (1.8), OTA: (1.23) |
High extraction and well separation of all AFs (B1, B2, G1, and G2) in the standards and samples chromatogram Good recovery and accuracy. |
Cost intensive due to procurement of separate analysis chemicals and standards | [156] |
| Green and black tea | Pesticides (μg/kg) |
Solvent extraction with acetonitrile | GC/MS | 1.0 –500 | 2.0–1000 | Exhibited good repeatability and reproducibility. | Pretreatment of the sample with chemicals is required | [170] |
| Green and black tea | Pesticides (μg/kg) |
Solvent extraction with acetonitrile | GC/MS/MS | 1.0 to 900 | 2.0 to 1800 | A laboratory repeatability of 100% shows that the method's repeatability is very good. | A collaborative approach is highly sensitive and can loss of samples during analysis. |
[170] |
| Tea samples | Pesticides (μg/kg) |
Solvent extraction with acetonitrile | HPLC/MS/MS | 0.03–4820 | 0.06–9640 | For 91% of pesticides determined by HPLC/MS/MS, the average recoveries reported between 60 and 120%. | Required optimization of HPLC/MS/MS conditions and selection of pesticides that are suitable for analysis |
[170] |
| 5 Tea products | 15 polycyclic aromatic hydrocarbons (PAHs) contaminant |
Solvent extraction using methanol | In-tube solid-phase microextraction coupled with HPLC-FLD | PAHs (pg/mL): Nap: 2.63 Ace: 1.25 Flu: 1.74 Phe: 1.78 Ant: 1.29 Flt: 2.20 Pyr: 1.68 BaA: 1.35 Chr: 1.29 BbF: 1.27 BkF: 0.32 BaP: 0.53 DahA: 0.46 BghiP: 0.50 IP: 4.63 |
PAHs (ng/g): Nap: 1.7 Ace: 0.8 Flu: 1.16 Phe: 1.18 Ant: 0.86 Flt: 1.47 Pyr: 1.12 BaA: 0.9 Chr: 0.86 BbF: 0.84 BkF: 0.21 BaP: 0.35 DahA: 0.31 BghiP: 0.33 IP: 3.08 |
Simple, rapid, and sensitive method of extraction; No pre-treatment of sample required. | High instrument calibration and installation cost | [171] |
| Tea leaf infusions | PAHs (μg/kg) | Dried tea infusions were made with freshly boiled water | HPLC-FLD | Benzopyrene (0.25), Benzoanthracene: (0.15), Benzo-fluoranthene: (0.15), Chrysene (0.25) |
Benzopyrene (0.75), Benzoanthracene (0.50), Benzo-fluoranthene (0.50), Chrysene (0.75) | Speedy and easy sample preparation method. Environmental safety due to low consumption of chemicals during sample preparation. Lower analysis cost. |
Optimization and validation are required for the method | [172] |
| Puer tea | 4 pesticides (μg/kg) (chlorpyrifos, indoxacarb, triazophos, and tolfenpyrad) |
Solvent extraction with acetonitrile and methanol |
MWCNTs, UHPLC-QTRAP-MS/MS | Chlorpyrifos:0.03 Triazophos: 0.15 Tolfenpyrad: 0.06 Indoxacarb: 0.15 |
Chlorpyrifos: 0.10 Triazophos: 0.50 Tolfenpyrad:0.20 Indoxacarb: 0.50 |
Quick, easy, cheap, effective, rugged, safe method High detection rate. |
Optimization and calibration of the method required | [173] |
| Green tea beverage | Organophosphorus pesticides (μg/L) |
GC-FPD | Hydrophobic deep eutectic solvents used in combination with vortex-assisted dispersive liquid-liquid microextraction | 0.05–0.3 | 0.17–1 | Highly sensitive method Selectivity of analyte can be achieved at lower detection limits. |
Consume multiple chemicals Require special instruments that must be operated by trained personnel Cost incentive and environmentally unfriendly |
[174] |
| Tea samples | Pesticides (μg/kg) |
Solvent extraction with acetonitrile | GC/MS | 1.0–500 | 2.0–1000 | High (94%) average recoveries of pesticides. Results indicated good repeatability, reproducibility, and yield Can simultaneously determine hundreds of pesticides in tea. |
Require strict cleanup Optimization of GC/MS conditions and Require column selection of pesticides suitable for analysis |
[174] |
| Tea samples | 653 pesticides (mg/kg) | Solvent extraction with acetonitrile | HPLC/MS/MS | 0.03–4820 | 0.06–9640 | For 91% of pesticides determined by HPLC/MS/MS, the average recoveries reported between 60 and 120%. | Required optimization of HPLC/MS/MS conditions and selection of pesticides that are suitable for analysis | [174] |
The amounts of free amino acids, catechins, and caffeine in fresh tea leaves using visible to short-wave infrared hyperspectral reflectance data and machine learning algorithms can be used as non-destructive assessment techniques of tea quality-related metabolites in fresh leaves. The data collected from about 200 fresh leaves of varying status showed mean ratio to performance of deviation (RPD) values between 1.1 and 2.7, the majority above the acceptable or accurate criterion (RPD = 1.4 or 2.0, respectively). Using data-based sensitivity analysis, significant hyperspectral zones between 1500 and 2000 nm were found. By merging the Visible–Near Infrared–Short-Wave Infrared (VIS-NIR–SWIR) (400–2500 nm) hyperspectral reflectance data with machine learning methods, successful estimation of the nitrogen and chlorophyll contents in tea leaves was also possible without causing any damage [175].
The applicability of conventional techniques in quick, on-site detection of tea safety and adulteration is limited due to its long run time, high cost, complicated equipment, and expert operators. Tea products should have a maximum MRL of 10 mg/kg of chemicals due to the high toxicity of contaminants. As a result, current detection techniques have insufficient LOD and LOQ, which must be improved to increase sensitivity and accuracy. The safety of tea leaves can be reliably measured via chemical analysis. However, chemical analysis is time intensive and needs many people, in addition to numerous chemical reagents and specific apparatus, making it difficult to determine the quality components of fresh leaves in tea factories quickly and in situ. This method has certain drawbacks, including limited sensitivity and accuracy. Advanced lab techniques, like microbial analysis, microscopy, and chromatography, are costly and require technical expertise.
This review has extensively explored various novel and emerging technologies for detecting tea safety and quality, particularly emphasizing techniques employed in recent years. A focus on the safety of tea is crucial as it is one of the most widely consumed beverages worldwide. Tea leaves are vulnerable to pre- and post-harvest contamination, leading to quality degradation. Considering these vulnerabilities, this review aimed to explore recent data on techniques promising better attributes of tea. Among conventional methods, chromatography is the most popular technique. Various types of spectroscopy, including visible and ultraviolet, fluorescence, and atomic spectroscopy, have gained attention among researchers. Other emerging technologies include imaging techniques, microbial methods, nano-technology, and electrochemical methods. These techniques are efficient and useful in detecting tea’s safety and quality parameters. Some of them are portable and can be used in the field, thus increasing the efficiency of tea production. These techniques enable lab-based results to be obtained in the field. Some methods have been designed by incorporating two or more techniques (LC-MS/MS, UPLC-MS/MS, etc.), resulting in an efficient assessment of tea through a tandem approach. Microplastic contamination in tea is a concern nowadays. To ensure its safety, extensive research is required to detect and manage the risk of microplastic contamination in tea. This includes the development of safety thresholds for microplastic in tea, as well as the implementation of effective control and management strategies.
