Extraction, Detection, and Quantification Methods for Analyzing Glyphosate and AMPA in Foods: Challenges and Opportunities

Abstract

Glyphosate (GLYP) is an effective and low-cost broad-spectrum herbicide. However, this herbicide and its primary degradation product, aminomethylphosphonic acid (AMPA), have been linked with adverse human health effects. The global use of glyphosate has significantly increased in recent years, resulting in more direct and indirect human exposure. In this context, GLYP and AMPA are often detected in fresh and processed foods for adults and infants, as well as in drinking water worldwide. Diverse extraction and quantification methods for GLYP and AMPA from foods have been developed. Solid-phase extraction and QuEChERS (Quick, Easy, Cheap, Effective, Rugged, and Safe) are the most frequently employed cleanup strategies, while LC-MS/MS is one of the most sensitive and selective techniques for detecting GLYP and AMPA in food products. Although most studies show that GLYP and AMPA concentrations in foods remain within established maximum residue limits, occasionally they exceed legal limits. Thus, the widespread presence of GLYP and AMPA in foodstuffs is a public concern that may lead to consumers exceeding the acceptable daily intake due to prolonged dietary exposure, even if levels remain within thresholds. Therefore, this review explores the different approaches and techniques used in the extraction, detection, and quantification of GLYP and AMPA in foods.

1. Introduction

Herbicides are chemical compounds widely used in modern agriculture to boost productivity by preventing weed growth, thus ensuring crops receive sufficient nutrients, water, and sunlight for optimal growth [1]. Despite the development of many herbicides, several were later banned due to their harmful effects on the environment and human health [2,3], contributing to global pollution (air, soil, groundwater, and food) [4]. Glyphosate [N-phosphonomethyl-glycine, GLYP] is a broad-spectrum, post-emergent, and systemic herbicide employed for weed control, characterized by high efficacy, ease of application, and cost-effectiveness [2,3], which can be applied to a variety of crops, including soybeans, wheat, cotton, and maize, among others [5,6]. Furthermore, its utilization has increased up to 100-fold since its introduction in the 1970s, and it is projected that between 740 and 920 thousand tons per annum will be applied by 2025 [7,8]. Nonetheless, the use of genetically modified or transgenic crops resistant to GLYP has further promoted its use, where GLYP is commercially available in both liquid and solid forms, enabling direct application without damage to these crops [9,10]. However, its extensive use has raised concerns about its potential adverse effects (direct or indirect) on human health [11]. The carcinogenic potential of GLYP has been a subject of significant scientific debate. In this context, the International Agency for Research on Cancer (IARC) of the World Health Organization (WHO) classified GLYP as a probable carcinogen (Group 2A) in 2015 [12]. Conversely, in 2016, the Joint FAO/WHO concluded that GLYP did not represent a genotoxic or carcinogenic risk in rats [13,14]. Similarly, the European Food Safety Authority (EFSA) and the European Chemicals Agency (ECHA) consider that GLYP does not pose a significant risk to human health in terms of cancer [15], and established an acceptable daily intake (ADI) of GLYP-based herbicides of 0.5 mg/kg body weight per day, and acceptable occupational exposure level of 0.1 mg/kg bd per day [16]. On the other hand, it was determined that no risk exists if the maximum daily intake remains within 1 mg/kg of body weight [17], or used according to the specific label use [18]. However, scientific evidence supports the IARC classification and has highlighted the role of GLYP in carcinogenesis [19,20] and other health conditions such as kidney damage, mental and neurological diseases, and reductions in sperm motility [16].

Additionally, due to the widespread application of GLYP in agriculture, the presence of its residues and their by-products such as aminomethylphosphonic acid (AMPA) in soil, water, and foods is prevalent [18]. AMPA is a chemical compound derived from the degradation of glyphosate; it is more toxic and persistent than glyphosate [4]. These residues can remain in treated crops and transfer to processed foods, increasing consumer exposure. The occurrence of GLYP and AMPA residues in foods raises significant food safety and public health concerns, particularly regarding chronic exposure to low levels of this herbicide [17]. The evaluation of GLYP levels in foods has become an essential task to ensure safety and protect public health due to the cumulative risk of GLYP residues [21]. The detection and quantification of GLYP and AMPA in foods present significant challenges due to the complexity of food matrices and the low concentrations in which they may be present [18]. Food matrices exhibit considerable variation in composition and physicochemical characteristics, which can influence the efficiency of extraction and analysis [2,22]. The complexity of food matrices requires efficient extraction methods that can separate GLYP and AMPA from other food components. In this context, the extraction of GLYP and AMPA from food matrices is the first critical step in their analysis and quantification, followed by cleanup strategies. The most used extraction methods include liquid–liquid extraction, solid-phase extraction, and Quick, Easy, Cheap, Effective, Rugged, and Safe (QuEChERS) [23,24,25]; most of them are based on the EU reference Laboratories for Residues of Pesticides [26]. After the extraction of GLYP and AMPA from the food sample, their detection and quantification are essential. Analytical methods for the determination of GLYP and AMPA in foods encompass chromatographic, spectroscopic and immunological techniques. High-performance liquid chromatography (HPLC) and gas chromatography (GC) coupled to mass spectrometry (MS) are commonly used approaches due to their high selectivity and sensitivity [27,28,29,30]. Consequently, it is imperative to implement extraction and detection methods that are selective and sensitive to the different forms and concentrations of glyphosate and AMPA present in different food matrices.

Therefore, this review explores the different approaches and techniques used in the extraction, detection, and quantification of GLYP and AMPA in foods, emphasizing their advantages, limitations, and applications.

2. Glyphosate and AMPA

Glyphosate is a non-selective, post-emergent, and systemic herbicide developed in the 1970s by the agrochemical corporation Monsanto and introduced into the market under the brand name Roundup^®. Subsequently, this herbicide has been extensively utilized in agriculture, horticulture, and other sectors for weed control. Nonetheless, GLYP is an active component in several organophosphorus herbicides [31]. The mechanism of action resides in its ability to inhibit a key enzyme in the plant metabolic pathway, specifically 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS), which is an essential enzyme in the shikimate pathway [32]. The shikimate pathway produces chorismate, a precursor molecule of essential aromatic amino acids in plants. Inhibition of EPSPS results in plants being unable to synthesize the proteins necessary for growth, development, and reproduction [33]. In the environment, GLYP is degraded to AMPA (its primary metabolite) and CO₂ by microorganisms, which utilize GLYP as a carbon and phosphorus source, resulting in the cleavage of C–N and C–P bonds. Microbial degradation of GLYP may occur under aerobic and anaerobic conditions; however, the degradation rate significantly increases in the presence of oxygen. Furthermore, it has been observed that GLYP degradation can occur within the plants themselves [34]. Additional studies have indicated that AMPA can be produced in aqueous solutions through the interaction of GLYP with metallic ions, often copper ions [35], potentially resulting in the presence of both GLYP and AMPA residues in biological and non-biological samples [2]. It is noteworthy that GLYP is not the sole precursor of AMPA in the environment; AMPA can also be generated from the degradation of amino phosphonates. However, due to the extensive use of herbicides containing GLYP, it remains a significant source of AMPA [35].

Table 1. Physical and chemical properties of glyphosate.

Parameter	Property Value
	Glyphosate	Aminomethylphosphonic Acid
Chemical structure
Family	Organophosphorus compounds	Organophosphorus compounds
Function	Herbicide	Metabolite from glyphosate
IUPAC name	N-phosphonomethyl-glycine	Aminomethylphosphonic acid
CAS number	1071-83-6	1066-51-9
Molecular formula	C3H8NO5P	CH6NO3P
Molecular weight	169.07 g/mol	111.04 g/mol
Solubility	>353 g L−1 at 20 °CNot soluble in acetone, ethanol or xylene	272.3 to 710.7 g L−1 at 20 °CNot soluble in acetone, ethanol or xylene
Melting point (°C)	189	220
Boiling point (°C)	Not defined	Not defined
Temperature of decomposition (°C)	215	290
Dissociation constant	pKa1 = 2.0; pKa2 = 2.6; pKa3 = 5.6; pKa4 = 10.6	pKa1 = 1.6; pKa2 = 5.8
pH	2.5	3.0
Density	1.7 g/cm3	1.51 g/cm3
Octanol–water coeff. (Kow)	−4.6 to −1.6	−2.6 to −2.14
Vapor pressure	1.94 × 10−7 mmHg at 45 °C	2.1 × 10−6 mmHg at 45 °C
Freundlich adsorption coefficient (Kads) (L Kg−1)	0.6 to 303	5 to 500
Degradation half-life in soil (T1/2) (days)	7–60	26–44

Source: [36,37,38,39].

shows the structural, physical, and chemical properties of GLYP and AMPA.

According to the International Union of Pure and Applied Chemistry (IUPAC), GLYP is an organic compound known chemically as N-phosphonomethyl-glycine that belongs to the organophosphate family and whose chemical formula is C₃H₈NO₅P [40]. Its chemical structure consists of a phosphonomethyl group linked to carboxylic acid and an amine group. The phosphonomethyl group (PO(OH)₂) is essential for the herbicidal activity of GLYP as it is involved in the inhibition of the EPSPS enzyme [41]. The carboxyl group (COOH) confers upon GLYP the ability to form hydrogen bonds and interact with other molecules in the environment, which can influence the mobility and bioavailability of this herbicide in soil and water [42]. Furthermore, the presence of a covalent bond between the carbon and phosphorus atom imparts to GLYP chemical and physical properties such as water solubility, high adsorption, and compatibility with other chemical compounds [2]. In relation to its physical and chemical properties, GLYP is a crystalline solid that is odorless and white in color, with a molecular weight of 169.07 g/mol, highly soluble in water (>353 g/L at 20 °C) and practically insoluble in organic solvents including acetone, ethanol, and xylene [36,40]. It is relatively stable under normal environmental conditions, is non-volatile, does not degrade photochemically, and is stable to hydrolysis at pH 3, 6 and 9 (5–35 °C). It is an amphoteric molecule with a pH of 2.5 (1% solution in water), melting point of approximately 189 °C, density of 1.7 g/cm³, and vapor pressure of 1.94 × 10⁻⁷ mmHg at 45 °C, and its decomposition point is 215 °C [2,36].

On the other hand, AMPA (CH₆NO₃P) is an amphoteric and non-volatile organophosphorus compound derived from GLYP degradation. It exhibits a molecular weight of 111.04 g/mol and a density of 1.51 g/cm³. Due to its polar nature, it is highly soluble in water (272.3 to 710.7 g/L), remains stable under acidic conditions (pH 3), and demonstrates a melting point of 220 °C. Research has indicated that AMPA is more persistent in the environment than GLYP and tends to accumulate in soil [34].

Due to the physicochemical properties of GLYP and AMPA, their persistence in the air, water, soil, and foods, and their potential toxigenicity to humans and animals, it is imperative to implement analytical methodologies that achieve the sensitivity and precision for the detection and quantification of this herbicide and its residues in biological and non-biological samples [2].

3. Glyphosate and AMPA in Biological and Non-Biological Samples

The widespread utilization of glyphosate in agricultural and other sectors has elicited considerable debate within health, environmental, academic, and industrial spheres, particularly in countries with extensive maize and soybean production. GLYP has been associated with various adverse human health outcomes, including neurotoxicity, genotoxicity, teratogenicity, and toxicity to gut microbiota [43]. Moreover, research suggests a link between GLYP exposure and neurodevelopmental disorders such as autism spectrum disorder [44]. Nonetheless, GLYP has impacts (directly or indirectly) on ecosystems, leading to a reduction in biodiversity at low or high scales [35]. GLYP and AMPA have been detected in biological and non-biological samples such as soil, blood cells and plasma, air, water, and food matrices [45]. Consequently, certain international agencies have established an acceptable daily intake of glyphosate-based herbicides (0.5 mg/kg body weight per day) and an acceptable occupational exposure level (0.1 mg/kg bd per day) [16]. However, in recent years, the presence of glyphosate and AMPA in different biological and non-biological samples has been frequently detected (

Table 2. Glyphosate and AMPA reports on biological and non-biological samples.

Sample	Glyphosate (µg/L)	AMPA (µg/L)	Ref.
Freshwater	17.5–125	17.5–125	[46]
Drinking water	<0.03–0.225	<0.03–0.184	[47]
Tap water	170–2900	10–80	[48]
Seawater	0.00012–0.0022	NR	[49]
Soil	50–825	238–1182	[50]
Animal feed	* 7.8−2140	ND	[45]
Human milk	0.11–3.32	NR	[51]
Human urine	0.2–5	0.2–5	[52]
Human saliva	0.00038–0.0028	6 × 10−5–0.00031	[53]
Human blood cells	600–7,480,000	100–60,000	[54]
Human urine	1600–400,000	500–1,000,000

ND: Not detected; NR: not reported; * µg/kg.

).

The monitoring of GLYP and AMPA has become a critical concern due to its association with various health issues. GLYP and AMPA have been reported in freshwater (17.5–125 µg/L for GLYP and AMPA) [46], drinking water (<0.03–0.225 µg/L for GLYP and <0.03–0.184 for AMPA) [47], tap water (170–2900 µg/L for GLYP and 10–80 µg/L for AMPA) [48], seawater (0.00038–0.0028 µg/L for GLYP) [49], and soil (50–825 µg/L for GLYP and 238–1182 µg/L for AMPA) [50]. Additionally, GLYP concentrations up to 2140 µg/kg have been reported in animal feed [45].

Regarding human samples, Camiccia et al. [51] reported the presence of GLYP in human milk (0.11–3.32 µg/L). The presence of glyphosate (0.00038–0.0028 µg/L) and AMPA (6 × 10⁻⁵–0.00031 µg/L) in saliva of people non-occupationally and occupationally exposed to pesticides has been reported [53]. Conrad et al. [52] reported the presence of GLYP and AMPA in human urine in values ranging from 0.2 to 5 µg/L. Additionally, Ruiz et al. [55] reported a frequency of 54% for glyphosate (12 µg/L) and 60% for AMPA (14 µg/L) in the urine of lactating mothers, wherein statistical analysis demonstrated a significant correlation between the consumption of eggs and fruits, with an estimated daily intake of 0.31 and 0.37 µg/kg body weight per day for glyphosate and AMPA, respectively. These findings suggest that glyphosate and AMPA can enter the human body with or without direct exposure to Gly-based herbicides, with food serving as a significant vector for GLYP and AMPA, as shown in Figure 1. Furthermore, acute intoxication by ingestion (accidental or non-accidental) of GLYP has been reported [54].

4. Glyphosate and AMPA in Foods

As previously discussed, the extensive use of GLYP for weed control has resulted in the presence of residues of this pesticide and its by-products in water and food products; therefore, consuming contaminated foods is the main pathway through which humans are exposed in a chronic and ongoing manner (directly or indirectly) to herbicides [16], which is considered a significant global health issue today [35]. The overuse and misuse of GLYP, particularly in less-developed nations, can result in environmental contamination and negatively impact human health over time. To ensure compliance with regulations and guarantee food safety, it is crucial to systematically track pesticide residues in the environment and measure human exposure across the entire food production process, including vegetal and animal-based foods. In this context, the regulation of GLYP varies considerably among countries and depends on the foodstuff. Moreover, each country defines acceptable concentrations of GLYP in foods and restricts or prohibits its use due to adverse effects on health and environment [35].

In the United States of America, the Environmental Protection Agency (EPA) is responsible for regulating pesticides, including GLYP. The EPA has established the maximum residue limit (MRL) of 30 mg/kg for most grains and 40 mg/kg for most oilseeds [56]. The European Commission (EC) has set an MRL of 20 mg/kg for barley, oats, sorghum, soybeans, and sunflower seeds; 10 mg/kg for wheat, rye, mustard seeds, lentils, and peas; 2.0 mg/kg for beans; 1.0 mg/kg for corn; and 0.1 mg/kg for cereals and grains if not specified [56]. Similarly, the Food and Agriculture Organization of the United Nations (FAO) and WHO Codex established the MRL at 30 mg/kg for most cereals, 5 mg/kg for beans, and 2 mg/kg for dry peas [57]. The establishment of the MRL aims to maintain a substantial safety margin in food production through the implementation of effective agricultural methodologies; however, prolonged exposure to GLYP and AMPA, particularly when MRLs are exceeded, poses a significant risk to human health [35]. In 2020, a decree was established in Mexico mandating the gradual elimination of GLYP across the country [58]. Although various regulatory agencies have established MRL values for GLYP and AMPA in diverse food matrices, the presence of GLYP and AMPA has been reported in foods globally (

Table 3. Glyphosate and AMPA reports on foods.

Country	Food Product	Number of Samples	Detection Frequency (%) GLYP/AMPA	Glyphosate Range (µg/kg)	AMPA Range (µg/kg)	Ref.
Canada	Honey	200	98.5	1–49.8	NR	[59]
Switzerland	Honey	16	93.8	˂1–15.9	NR	[60]
Estonia	Honey	33	12.1	9–62	NR	[61]
Estonia	Honey	140	18	70	NR	[62]
USA	Honey	85	28.2	15–342	NR	[63]
Brazil	Honey	40	100	40	40	[64]
Mexico	Pollen	120	100	3.71–7.29	NR	[65]
Switzerland	Fruit juice	11	100/18	1.6–1.9	0.2–0.6	[60]
Canada	Juice concentrates	42	NI	4.2–38	NR	[6]
USA	Frozen vegetables, fruit juice, baby fruit puree	83	15–18	3–10	NR	[66]
Spain	Vegetables	221	NI	NI	NR	[67]
Switzerland	Breakfast cereal	10	80/30	˂1–291	2.5–10	[60]
Switzerland	Wheat snacks	11	36.4	˂1–421	NR	[60]
Switzerland	Bread	10	70	˂1–45.8	NR	[60]
Switzerland	Wheat flower	28	28.6	˂1–133	NR	[60]
Switzerland	Pseudo cereals	3	ND	˂1	NR	[60]
Italy	Wheat seed	1	100	243,000	NR	[68]
France	Breakfast cereal	2	100	6–34	NR	[69]
Lebanon	Bread and wheat flour	164	80 and 100	14–52	NR	[70]
Australia	Wheat flour	26	100	˂10	NR	[71]
USA	Oat products	310	100	0.04–1.1	NR	[72]
South Africa	Soy milk	8	100	32–142	NR	[73]
South Africa	Texturized soy protein	7	100	195–2257	NR	[73]
Canada	Pasta	221		5–1400	NR	[59]
China	Fresh maize and soybean	234	11–18	40–290	NR	[74]
South Africa	Maize pasta	3	100	47–62	NR	[73]
South Africa	Maize rice	3	100	28–65	NR	[73]
Switzerland	Potato and vegetables	10	30/0	˂1–7.7	NR	[60]
Europe and UK	Grains, rice, flour, bread, cereal based for infants	2136	19.9	10–267	NR	[75]
Switzerland	Pulses	41	51.2/24	˂1–2948	3.1–25	[60]
Switzerland	Meat and Fish	13	23.1	˂1–4.9	NR	[60]
Ghana	Yam	68	20.5	˂120	NR	[76]
Several European counties	Processed foods	110,829	3.7	Above MRL values	NR	[77]
Brazil	Soy-based infant formula	105	NI	30–1080	20–170	[78]
Nigeria	Fish	75	100	Below ADI and MRL values	NR	[79]

NI: No information; ND: not detected; NR: not reported; MRL: maximum residue level. ADI: acceptable daily intake.

).

GLYP and AMPA have been detected in a wide range of plant- and animal-based food products. Studies conducted in Canada, Switzerland, Estonia, the USA, and Brazil have identified the presence of GLYP and AMPA in honey samples, with detection frequencies ranging from 12 to 100% and concentrations ranging from ˂1 to 342 µg/kg [59,60,61,62,64]. A honey sample from the USA exhibited the highest GLYP content [63]; however, some honey samples from European countries contained GLYP levels exceeding the legally permitted maximum residue limit (MRL) [77]. Furthermore, AMPA was detected in honey samples from Brazil (40 µg/kg) [64]. In Mexico, the presence of GLYP residues has been reported in 120 samples of pollen (3.71–7.29 µg/kg), which could potentially serve as a contamination source for honey [65].

GLYP has been reported in fruit and vegetables (˂1–7.7 µg/kg), pulses (˂1–2948 µg/kg), and fruit juices (1.6 to 1.9 µg/kg) from Switzerland; moreover, AMPA (0.2 to 0.6 µg/kg) was detected in 18% of fruit juice samples and 24% of pulses (3.1–25 µg/kg) [60], while in Spain, GLYP was detected in vegetables [67]. Additionally, GLYP has been detected in processed food products worldwide, for instance, in fruit juice concentrate (4.2–38 µg/kg) from Canada [6], as well as in yam (68% detection frequency, ˂120 µg/kg) from Ghana [76] and frozen vegetables, fruit juice, and baby fruit puree (15–18% detection frequency, 3–10 µg/kg) from the USA [66].

Regarding cereal-based products, GLYP has been detected in breakfast cereal (˂1–291 µg/kg), wheat snacks (˂1–421 µg/kg), bread (˂1–45.8 µg/kg), wheat flour (˂1–133 µg/kg), and pseudo cereals (˂1 µg/kg) from Switzerland. Additionally, AMPA was identified in breakfast cereals (2.5–10 µg/kg). The detection frequency in these product categories ranged from 28 to 80% [60]. Furthermore, GLYP has been identified in one sample of wheat seeds (243,000 µg/kg) from Italy [68] and pasta (5–1400 µg/kg) from Canada [6]. In France, GLYP has been detected in two breakfast cereal samples (6–34 µg/kg) [69]. It has been reported that some grains, rice, flour, bread, and cereal-based products for infants contained GLYP (19% detection frequency, 10–267 µg/kg) [69]. In Australia, GLYP was detected in 26 samples of wheat flour [71]. In Lebanon, 164 samples of bread and wheat flour were analyzed, in which GLYP (14–52 µg/kg) was detected in 80 and 100% of the samples, respectively [70]. In the USA, GLYP (0.04–1.1 µg/kg) was reported in 310 oat products [72]. For its part, a study conducted in China reported the presence of GLYP (40–290 µg/kg) in fresh maize and soybean (240 samples) with a detection frequency of 11 to 18% [74]. Also, a study conducted in South Africa detected GLYP in maize pasta (47–62 µg/kg), maize rice (28–65 µg/kg), soymilk (32–142 µg/kg), and texturized soy protein (195–2257 µg/kg) [73]. Moreover, in Brazil, GLYP (30–1080 µg/kg) and AMPA (20–170 µg/kg) were detected in soy-based infant formula [78].

Additionally, GLYP has been detected in meat and fish (˂1–4.9 µg/kg) from Switzerland [60] and fish (↓ADI and ↓MRL) from Nigeria [79]. Furthermore, according to the most recent EFSA report in 2022 [77], 144,563 food samples were analyzed for glyphosate and its derivatives across European countries, wherein 98% of samples showed no detectable GLYP, 1.7% exhibited values above the limit of quantification (LOQ) but below MRL values and 0.3% of samples exceeded MRL values. Additionally, AMPA was quantified in 0.097% of the samples, predominantly in soybeans. Furthermore, the presence of GLYP in alcoholic beverages such as wine (21 samples, 100% detection frequency, 0.6–18.9 µg/L) and beer (15 samples, 13.3% detection frequency, ˂0.5–6.8 µg/L) has been reported in Switzerland [60], while in the Latvian market, it has been reported in beer (100 samples, 92% detection frequency, ˂0.5–150 µg/L) [80].

Although most studies suggest that GLYP and AMPA remain within the maximum residue limits, the global detection frequency of these compounds in samples is notably high, ranging from 15 to 100%. This prevalence may result in exceeding the acceptable daily intake or “safe GLYP and AMPA doses” for consumers due to prolonged dietary exposure, even within ADI thresholds. Therefore, various methodologies have been developed to identify and quantify glyphosate and AMPA in foodstuffs using chromatographic, spectroscopic, and immunological techniques [25,26,27,28]. However, given the molecular complexity of these substances, sample pretreatment and cleanup strategies must be tailored to the specific characteristics of the food matrix. A proposed diagram for GLYP and AMPA analysis in food matrices is shown in Figure 2, based on the information discussed throughout this work.

4.1. Sample Pretreatment

Sample pretreatment prior to chromatographic separation/spectroscopic and/or antibody-based detection is a critical step to ensure the efficacy of the analysis of GLYP and AMPA, especially when addressing complex matrices such as foods. The sample pretreatment constitutes the most time-consuming phase of a methodology. Typical sample pretreatment processes encompass dissolution, homogenization, extraction, precipitation, and concentration of analytes [81]. The selection of the most appropriate pretreatment technique is essential for achieving optimal accuracy and minimizing potential interference during analysis. The physicochemical properties of glyphosate and AMPA, in conjunction with the complexity of food matrices, are challenging in developing analytical methods for GLYP detection and quantification. GLYP and AMPA are highly polar compounds, rendering them highly soluble in water but virtually insoluble in organic solvents such as methanol or acetonitrile, which complicates the use of traditional extraction solvents [82].

The QuPPe-PO (Quick Polar Pesticides) method edited by the EU Reference Laboratories for Residues of Pesticides serves as the standard guideline for the extraction and measurement of glyphosate and other polar pesticides. Various modifications of this technique have been implemented globally to enhance sample pretreatment in each matrix [26]. For most commodities, including dried and fresh fruits and vegetables, homogenization is essential to evaluate the efficiency of the extraction. When analyzing soybean, oat, maize, rice, and other grains, reducing the particle size constitutes a critical step in the extraction process. Larger particle size may hinder the extraction of the analyte, as the solvent may not effectively penetrate the particle and access the analyte trapped within the grain, while smaller particle size increases the surface area available for solvent interactions, thereby enhancing the efficiency of the extraction process. In this context, reducing the particle size of the food sample is preferable to improve the overall extraction yield of glyphosate [72,82,83].

In the pretreatment of samples with a moisture content below 80%, it is necessary to adjust the water content before the extraction, aiming to minimize analytical errors [72,82,83]. In analysis of honey or foods with higher sugar contents, the sugar concentration should be considered in the adjustment, as it can contribute to the volume when fully dissolved in the extraction solvent; in those cases, the water amount of the sample may be less than usual [26]. Furthermore, the presence of interfering proteins can be treated with the addition of acidified methanol. The addition of methanol to the solution reduces the solubility of proteins by lowering the dielectric constant of the sample proteins, which increases the attraction between charged amino acid residues, facilitating the intermolecular interactions and producing protein aggregates [84]. The addition of weak acids decreases the pH of the medium and provides protons that can be captured by protein amino groups, enhancing the intramolecular interactions. For instance, for LC-MS/MS analysis, volatile organic acids such as formic or acetic acid are recommended to prevent contamination into the ionization source; a low concentration of acid is enough to achieve good recovery percentages; for example, 50 mM of acetic acid [82] or 1.0% formic acid [26] could be used. The process efficiency can be enhanced by resting the sample under freezing conditions. Starch present in corn and other grains can absorb glyphosate, impeding its extraction. Since starch is also soluble in the aqueous medium, it could be dissolved in the water used to extract glyphosate; however, methanol, used to precipitate proteins, may be a helping hand, since starch can be precipitated too. Nevertheless, the polymerization of starch is enhanced at acidic pH and causes starch suspensions that can sequestrate glyphosate [85].

One of the biggest challenges in glyphosate analysis arises from its behavior as a chelating agent. Glyphosate can form chelates with polyvalent metals frequently found in foods (i.e., milk, cheese, cereals, and grains) like calcium, magnesium, and sodium [86], especially in samples with pH around 4.5 to 8.0, where glyphosate is negatively charged [87]. The formation of these metallic complexes can reduce the precision of the method and significantly decrease the recovery yield [23]. Additionally, glyphosate’s chelating capability enhances its interaction with sodium ions in borosilicate glass or iron ions in chromatographic system tubing. To mitigate these interactions, the use of polyethylene or polypropylene materials, along with stainless steel tubing, is suggested [88]. The addition of 10 mM disodium EDTA solution is highly recommended to prevent complexation with metallic compounds in samples, thereby enhancing recovery rates [23]. However, the use of EDTA should be carefully monitored, as concentrations exceeding 5% have been shown to decrease peak signal due to ion suppression [89].

4.2. Sample Cleanup

Foods are complex matrices that contain metallic ions, lipids, proteins, sugars, fibers, starch, and other components, which can result in significant signal suppression during chromatographic analysis of GLYP and AMPA. Consequently, the utilization of an effective sample cleanup technique following GLYP and AMPA extraction is essential for enhancing method accuracy and precision [26]. At present, solid-phase extraction (SPE) and QuEChERS are the most frequently employed cleanup strategies to mitigate the matrix effect when analyzing pesticides such as glyphosate and their derivatives (AMPA) in food samples [82,85,90,91], as listed in

Table 4. Sample cleanup strategies for recovery of glyphosate and AMPA from foods.

Food Sample	GLYP/AMPA	Glyphosate Recovery (%)	AMPA Recovery (%)	Extraction Method	Sorbent	Retention Mechanism	Ref.
Apple	GLYP/AMPA	103.1–115.6	84.2–105.6	SPE	PRiME HLB	Reversed phase	[85]
Cucumber		97.4–112.5	90.5–101.4
Potato		99.3–102.7	95.9–103.4
Celery		90.2–110.5	89.1–99.0
Grape		91.3–107.5	98.1–114.4
Soybean		106.1–111.4	91.4–94.0
Tea		97.0–102.3	92.0–108.0
Kiwi		87.3–105.6	96.0–97.5
Pumpkin		87.0–101.3	99.7–104.5
Orange		104.4–109.2	88.6–100.6
Lettuce		88.6–104.7	91.7–103.1
Rice		96.0–97.2	91.5–103.5
Wheat		101.5–105.5	91.1–98.9
Soybean Corn	GLYP/AMPA	96–98	96–113	SPE	Oasis HLB	Reversed phase	[82]
Oat	GLYP/AMPA	102	NR	SPE	Oasis HLB	Reversed phase	[72]
Beer	GLYP	87.0–123.0		SPE	Strata-XA	Anionic exchange	[80]
Egg	GLYP/AMPA	94.12–139.3	96.83–106.8	SPE	Plexa PCX	Cationic exchange	[91]
Sesame	GLYP/AMPA	87	87	dSPE	C18	Reversed phase	[90]
Lentils		57	60
Wheat		60	74
Cocoa Bean		50	47
Infant Food		71	73
Milk		93	93

NR: Not reported; SPE: solid-phase extraction; dSPE: Dispersive solid-phase extraction.

.

SPE adheres to principles such as liquid–solid extraction, wherein analytes are separated based on their distribution between two phases. The sample, diluted in a liquid system, is passed through a solid-phase sorbent, where the analytes are retained by the sorbent with varying degrees of affinity. Following this retention step, a washing solvent is employed to remove any non-retained impurities. Subsequently, the addition of an elution solvent disrupts the interaction between the analytes and sorbent, facilitating the analyte purification from the matrix. Various approaches, including normal phase, reversed phase, and ion exchange, can be utilized to enhance analyte retention within the SPE system. SPE has been used for GLYP and AMPA extraction from yam and grape [92,93], guava [94], diverse fruit and vegetables [95], soy and corn [82], and beer [80].

Hydrophilic–lipophilic balanced (HLB) SPE cartridges are polymer-based universal sorbents with a sophisticated reversed-phase retention system. Their characteristics enable them to interact with various compounds, including acidic, basic, or neutral analytes. In recent years, HLB cartridges have become the standard cleanup technique for glyphosate and AMPA purification prior to chromatographic analysis. Research has supported HLB as an effective extraction strategy with satisfactory recovery rates for analyzing polar pesticides from food matrices. A study conducted on 13 different food samples utilized a PRiME HLB cartridge in a non-retentive SPE protocol for glyphosate and AMPA determination [85]. Oasis HLB cartridges are also widely employed in sample preparation, with studies reporting glyphosate recovery percentages in soybean (96–98%), corn (96–98%), and oat (102%). AMPA recovery percentages have been reported in soybean and corn (96–113%) [96].

Dispersive solid-phase extraction (dSPE), also known as QuEChERS, is an alternative methodology initially developed for pesticide residue analysis in fruits and vegetables [90]. The primary distinction from classical SPE cartridges is that in dSPE, the sorbent is presented as a loose powder in predetermined quantities. In this context, polar pesticides such as GLYP are mainly analyzed using single-residue methods instead of multi-residue procedures [97]. Traditional QuEChERS methodologies use C18 or PSA (primary or secondary amines) as retention sorbents, acetonitrile as the sample diluent, and MgSO₄ for sample cleanup. Due to the chelating nature of GLYP and AMPA, the addition of MgSO₄ is not recommended [96]. It is mainly recommended as a cleanup strategy for food samples containing a single type of pesticide. Official guidelines, such as QuEPPe-PO, establish the use of dSPE with C18 sorbent as the standard for GLYP and AMPA cleanup in food samples [26]. Notwithstanding regulatory guidelines, several researchers have reported lower recovery rates with the C18-based dSPE extraction technique compared to SPE cartridges. Schafer et al. [90] evaluated the QuEPPe-PO method for the analysis of glyphosate and AMPA, reporting recovery percentages of 60–93% for glyphosate and 47–93% for AMPA. These recoveries were lower than those reported by other researchers who utilized SPE cartridges to analyze similar samples.

Other extraction methods for recovering GLYP and AMPA from food matrices have been explored with or without applying cleanup strategies in combination with the QuPPe method [26]. For instance, Chiesa et al. [98] extracted GLYP from honey, fish, and beer samples using ultrasound-assisted extraction (15 min) and acidified water (1%) and methanol (7:3) before cold centrifugation (4 °C) at 2500× g for 10 min, prior to IC-HRMS analysis, with an LOQ of 43, 51, and 65 µg/kg, respectively. Similarly, ultrasound-assisted extraction (10 min) using water as solvent was applied for extracting GLYP from cereals [99].

Additionally, Melton et al. [100] homogenized food samples (melon, pea, pepper, or pineapple) with a mixture of water and methanol (1:1) before centrifuging at 2500× g for 10 min, prior to IC-MS/MS analysis, with an LOD of 25 µg/kg. Similarly, centrifugation and acidified methanol at 1% have been used for extracting GLYP from fruits and vegetables (4000× g for 5 min) [66], oil (3700× g for 10 min) [101], and rice and corn (4000× g for 5 min) [83].

The selection of the most appropriate cleanup strategy should prioritize the reduction of potential interferences within the sample matrix. An inadequate extraction process can lead to matrix effects that either enhance or suppress the analytical response, leading to inaccurate results. Furthermore, a non-optimized extraction protocol may compromise analyte recovery, thereby reducing the sensitivity and detection capability of the analytical technique. Based on this, parameters such as recovery rates are critical when comparing different cleanup strategies.

SPE is a highly selective, specific, and robust cleanup strategy that can be tailored to a wide range of food matrices. It is particularly useful for analyzing complex samples, especially those with high levels of fat, sugar, or pigments. The use of reversed-phase cartridges effectively reduces non-polar interferences, making them suitable for lipid-rich samples. Conversely, ion-exchange cartridges are highly recommended to minimize ionic interferences, such as metal ions. Given the chelating properties of GLYP and AMPA, reducing ion concentrations in the sample matrix can significantly mitigate the instrumental signal suppression [102]. Nevertheless, the application of SPE presents certain limitations. The highly polar nature of GLYP and AMPA is a challenge for retention on reversed-phase cartridges. In such cases, non-retentive SPE protocols must be employed, which can compromise analyte recovery and result in lower efficiency compared to conventional SPE methods.

QuEChERS is often the preferred approach among analytical chemists, as it is the standardized methodology recommended by QuPPe-PO. Additionally, it is cost-effective, straightforward, and highly adaptable to a wide range of sample matrices by modifying its composition according to specific analytical requirements, if the method is validated. However, despite these advantages, QuEChERS has been associated with lower recovery rates in the analysis of GLYP and AMPA in food matrices and lacks complete validation traceability or formal protocol references [103].

One of the biggest challenges in pretreatment and sample cleanup is the loss of the analytes. These processes are rarely perfect and often involve multiple steps to reduce matrix interferences. Each of these steps is susceptible to analyte loss, decreasing the recovery of the method. This, combined with matrix effects, can significantly reduce the analytical response and compromise method performance. Procedural calibration is a suitable alternative to offset biases caused by pretreatment losses. Guidelines such as SANTE [104] foster the use of spiked samples that undergo the full extraction procedure in order to minimize analytical errors.

According to the above, the selection of the most appropriate retention method and the development of a suitable extraction protocol are crucial for achieving optimal recoveries and minimizing the matrix effect [102].

4.3. Detection and Quantification

The physicochemical properties of GLYP (high polarity, low molecular weight, absence of ultraviolet absorption, low ionization, high solubility in water, lack of chromophores, and low volatility) render its detection challenging when employing conventional analytical methods [2,26]. Reversed-phase chromatography has become the standard methodology for small-molecule analysis; nevertheless, the high polarity and zwitterionic nature of glyphosate and AMPA reduce the retention capacity of non-polar C18 columns, leading to reduced resolution and repeatability [105]. Therefore, a wide array of alternative analytical methods have been developed to identify and measure glyphosate levels in food products, including chromatographic techniques, as listed in

Table 5. Chromatographic and spectroscopic strategies for the detection and quantification of glyphosate and AMPA in foods.

Food Sample	GLYP/AMPA	Analytical Methodology	LOD	LOQ	Recovery (%)	RSD%	R2	Ref.
Tea leaves	GLYP/AMPA	LC-MS/MS	0.0028 and 0.046 µg/mL	0.0093 and 0.046 μg/mL	98.69–106.26/79.95–83.84	0.91–1.18/6.4–7.45	0.999/0.991	[106]
Commercial breakfast cereals	GLYP/AMPA	LC-MS/MS	1–5 ng/g	5 and 40 g/g	92–111	˂8	0.9989/0.9987	[72]
Baby formula, bovine liver and kidney	GLYP/AMPA	LC-ESI- MS/MS	NI	10–25 µg/Kg	104	5–25/11–38		[28]
Commercial corn flour	GLYP	UPLC-MS/MS	0.1 and 0.2 µM	0.2 µM 1.0 µM	58.48–109	NI	0.9976/0.9980	[107]
Beer	GLYP/AMPA	CE-TOF-MS	<5 μg/L/3.3 and 30.6 μg/L	NI	94.3–110.7/80.2–100.4.	~8.1%.		[30]
Grapes, orange, lettuce, oat and soya beans.	GLYP/AMPA	LC-ESI-QTRAP-MS	NI	0.02–0.05 mg/kg	83–118/93–120	5–30/ 3–19		[108]
Pome fruits, berries, pulses and cereals	GLYP	FI-MS/MS and LC-MS/MS	NI	0.5–2 mg/kg	78–111	˂20%		[92]
Cow milk, liver (bovine), kidney (bovine) and meat/egg chicken.	GLYP/AMPA	LC-ESI-TQ-MS	NI	0.01–0.02 mg/kg	70–120	≤20%		[109]
Maize and rice	GLYP	LC-MS/MS	0.002 and 0.004 mg/kg	0.01 mg/kg	70–105	<20%	0.9982	[83]
Edible oils	GLYP/AMPA	LC-MS/MS	NI	10 μg/kg 5 μg/kg	81.4–119.4	<20%	0.996	[101]
Honey, fish (bass) and bovine muscle	GLYP/AMPA	IC-HRMS	NI	4.30–9.26 ng/g	75–100/75–96	7–13/2–12		[98]
Soybean	GLYP/AMPA	HPLC	0.002 mg/kg 0.001 mg/kg	NI	85.4–94.1/87.3–95.2	3.1–4.7/3–4.4	0.999	[110]

NI: No information.

.

Currently, LC-MS/MS is recognized as the most sensitive and selective technique for detecting glyphosate in food products due to its precise (sensitivity, specificity, and accuracy) identification and quantification of GLYP in complex samples. This method is also endorsed by the European Union Reference Laboratory for Pesticide Residues as the preferred approach for such analyses [26]. When analyzing glyphosate in complex samples such as food, matrix effects often arise. The influences of different substances have an impact on the accuracy and sensitivity of determination. In those cases, matrix-matched strategies can be employed, and the matrix effect can be calculated to minimize errors in quantification [98,111]. Matrix matching is a frequently used technique in analytical chemistry to consider the influence of matrix effects, which can either enhance or suppress the analytical response. To create matrix-matched calibration curves, major components of the sample are added to a series of known-concentration standard solutions. The use of isotopically labeled internal standards (IL-ISs) is commonly employed to offset error and matrix interferences; in those cases, any instrument variation or matrix signal suppression affects both analytes and IL-IS equally, making the quantification process more accurate. Glyphosate 2-13C,15N and AMPA 13C, 15N D2 are the preferred IL-ISs [72,85,112].

Wang et al. [106] developed a method for detecting GLYP and AMPA in tea leaves using a GLYP dummy molecular synthesized polymeric template by SPE-coupled LC-MS/MS, which exhibited high sensitivity and specific detection (R²: 0.999 and 0.991) of GLYP and AMPA (0.05 to 4 µg/mL), with good GLYP (98–106%) and AMPA (79–83%) recoveries, and RSD values of 0.81–1.18% and 6.40–7.45% for GLYP and AMPA. According to the authors, the synthesized template exhibited more binding sites for GLYP absorption (28.6 µg/mg). Cruz and Murray [72] used LC-MS/MS for GLYP and AMPA quantification from 30 breakfast cereal samples. The LOQ was 5 ng/g (R²: 0.9989) for GLYP and 49 ng/g (R²: 0.9987) for AMPA; moreover, they reported LOD values (detected but not quantified) from 1 to 5 ng/g. According to the authors, the developed method could be applied for screening commercially available oat-based products. Santilio et al. [83] used LC-MS/MS for GLYP quantification in corn and rice samples; moreover, they used a GLYP isotope (10 µg/mL as standard) and phosphate buffer as the mobile phase. The recovery was 70 to 105% for both samples with acceptable precision (RSD: ˂20%), good linearity (R²: 0.9982), and adequate LOD (0.01 mg/kg). Previously, GLYP and AMPA were detected in soybean samples by HPLC using 3,6-dimethoxy-9-phenyl-9H-carbazole- 1-sulfonyl chloride as a fluorescent labeling reagent (R²: 0.999), with LOD values of 0.02 ng/mL for GLYP and 0.01 ng/mL for AMPA and recovery values higher than 95%.

Chiarello et al. [101] developed a fast method for GLYP and AMPA detection and quantification in edible oils using LC-MS/MS by electrospray ionization in negative mode. They reported an LOQ of 10 µg/kg for GLYP and 5 µg/kg or AMPA (R²: 0.996), with recoveries from 81.4 to 119.4% and RSD values less than 20%. Méndez-Barredo et al. [107] quantified GLYP and AMPA using UPLC-MS-QqQ without derivatization steps. They reported a low-cost and rapid method with high sensitivity and specificity and high recovery percentages (58.48–109%) for GLYP and AMPA detection and quantification (LOD: 0.1 and 0.2 µM; LOQ: 0.2 and 1 µM) in corn samples. Furthermore, LC-ESI-MS/MS has been used to quantify GLYP and AMPA in baby formula and bovine liver and kidney with an LOQ of 10–25 µg/kg and recovery values of 104% (RSD: 5–25/11–38%, respectively) [28].

During GLYP and AMPA analysis, it is possible to use methodologies based on underivatized and derivatized processes aimed at increasing the sensitivity, selectivity, accuracy, reproducibility, and robustness of the method. The underivatized strategies offer a faster, cheaper, and simpler analytical approach for glyphosate and AMPA determination, while the derivatization approach facilitates chromatographic separations using reversed-phase columns and provides a more specific fragmentation pattern, resulting in a more robust strategy [113]. In the underivatized methodology, negative-mode ESI is preferred since both compounds tend to lose a hydrogen ion, resulting in easier detection and producing a higher abundance. For glyphosate, the quantifier transition is 168→150 with 63 and 79 m/z as qualifier ions. For AMPA, the quantifier transition is 110→81 with 63 m/z as the main qualifier ion [85]. Direct analysis of polar pesticides requires the use of specialized polar columns, such as the Waters Anionic polar pesticides [114] or Luna Polar Pesticides [115]. The use of polar columns offers several advantages in underivatized methods, including higher reproducibility and resolution, and more efficient retention of both anionic and cationic analytes, showing that the same column can be used either in positive or negative modes. Ding et al. [111] developed an underivatized method for detecting GLYP in foods using SPE followed by separation with a hydrophilic interaction/weak anion-exchange column and LC-MS/MS coupled with negative electrospray ionization. The authors reported acceptable LOQ (16–26 µg/kg) and LOD (5–8 µg/kg) values, and recoveries ranged from 83 to 100%, with good accuracy and precision. Similarly, Botero-Coy et al. [116] investigated the feasibility of directly detecting and quantifying GLYP in vegetables using LC-MS/MS. They employed an Obelisc N column and SPE-Oasis HLB cartridges in their methodology.

The high sensitivity of LC-MS/MS makes it particularly valuable for detecting GLYP and AMPA at trace levels, and it should therefore be considered the primary option when low-concentration quantification is required [97]. Despite its analytical power, LC-MS/MS also presents certain limitations. Its high sensitivity makes it more susceptible to matrix effects, which can introduce errors in quantification. Additionally, when analyzing complex samples, the presence of impurities may contaminate the instrument, necessitating the use of more efficient extraction and cleanup procedures. Some studies validated LC-MS/MS methods across multiple food matrices, consistent with AOAC and FDA requirements. These cases highlight the benefits of robust and guideline-driven validation in ensuring the quality of results.

On the other hand, HPLC systems represent one of the most accessible alternatives for many laboratories. However, due to the lack of chromophore and fluorophore functional groups in the chemical structure of GLYP and AMPA, their detection with conventional chromatographic detectors, such as fluorescence detection (FLD) and diode array detection (DAD), is limited. This necessitates the use of derivatization processes to form chromophores/fluorophores able to be detected [117]. The derivatization approach can be carried out using reagents such as dansyl chloride [80]; nevertheless, the most used derivatizing agent is 9-Fluorenylmethyl chloroformate (FMOC-Cl), which provides an easier extraction and separation process. The reaction mechanism involves the nucleophilic attack of the amine group in glyphosate on the acyl chloride in FMOC-Cl, with chloride as the leaving group, in which HCl is produced (Figure 3). Since free HCl could react with the alkaline amine group, adding a base is necessary to improve the reaction yield; a borate buffer solution with a pH of 9.0 is required to achieve the reaction conditions [118].

The analysis of glyphosate under derivatization can be performed on LC-MS/MS systems. In these cases, the lower polarity of the glyphosate–FMOC compound addresses the poor retention encountered with polar pesticides in routine analysis. The aryl groups in FMOC enhance both AMPA and glyphosate’s capability to interact with reversed-phase columns such as C8 or C18. Electrospray ionization (ESI) can be applied in both positive and negative ion modes. A method for quantifying glyphosate and AMPA in agricultural samples using the positive ESI mode has been reported. In multiple reaction monitoring mode (MRM), for glyphosate, a precursor ion of 392 m/z was reported, with product ions at 179, 88, and 214 m/z, and 392→179 as the quantifier transition. For AMPA, the quantifier transition was 334→179, with 156 and 112 as the qualifier’s ions [24]. When using negative ESI mode, the quantifier transition for glyphosate was reported as 390→168, with 150 m/z as the qualifier [50]. The derivatization approach has been used to analyze glyphosate in food matrices (

Table 6. Detection of glyphosate in food using derivatization approach.

Food Sample	GLYP/AMPA	Analytical Methodology	Column	Derivatization	Results	Ref.
Soy-based products	GLYP/AMPA	LC-MS	ACQUITY UPLC BEH C18 1.8 μm, 2.1 × 100 mm, column	FMOC	LOQ was 50 and 5 µg/kg	[112]
Corn	GLYP	UHPLC-MS/MS	HSS T3 (1.8 μm, 2.1 mm × 100 mm) from Waters	FMOC-Cl	LOQ was 0.005 mg/kg	[119]
Baby formula, bovine liver and kidney	GLYP/AMPA	LC-ESI-MS/MS	Luna column (150 × 2 mm) with bonded 3 μm C18	FMOC-Cl	LOQ: 10–25 μg/kg for both	[80]
Liver of animal origin	GLYP	UPLC-MS/MS	ACQUITYTM UPLC BEH C18 (1.7 μm, 2.1 × 100 mm)	FMOC-Cl	LOQ and LOD: 0.025 mg/kg	[120]
Oat and rye wheat	GLYP/AMPA	LC-MS/MS	ACQUITY™ BEH C18 column (1.7 μm; 2.1 × 100 mm)	FMOC-Cl	LOQ: 10 ng/mL	[121]
Honey	GLYP/AMPA	HPLC-FLD	Aminex-A9 potassium exchange column (100 mm × 4.6 mm, 5 μm)	OPA-MERC	LOQ: 10–25 μg/g for both	[64]
Soybean	GLYP/AMPA	HPLC	ODS column (150, 4.6 mm I.D., 5 mm)	DPCS-Cl	LOQ: 0.002 mg/kg for GLYP and 0.001 mg/kg for AMPA	[110]

DPCS-Cl: Fluorescent labeling reagent 3,6-dimethoxy-9-phenyl-9H-carbazole1-sulfonyl chloride.

).

Ehling et al. [112] developed a derivatized method using FMOC for detecting GLYP and AMPA from soy-based products by LC-MS/MS by multiple transition monitoring (fluorescence detector) and reported accuracy and intermediate precision (91–106%, RSD ˂10%) in soy protein isolate, where LOQs were 50 µg/kg for powders and 5 µg/kg for liquids. They used fragment ions of 170 and 214 m/z for GLYP and 112 and 156 m/z for AMPA. Zhang et al. [119] established a sensitive method for determining GLYP from corn by UHPLC-MS/MS, using FMOC-Cl as the derivatizing agent, and reported LOQ of 5 µg/kg with good recovery (90–95%) and RSD of 1.24 to 3.35%. They used product ions of 167.8 and 150.0 m/z. Nonetheless, after purification and derivatization, samples were diluted to reduce matrix effects. On the other hand, it has been reported that excess FMOC reagent may affect the reaction mixture [119]. Additionally, FMOC-Cl has been used as a derivatizing agent during the analysis of GLYP and AMPA from baby formula, bovine liver and kidney [80,120], and oat and rye wheat [121], among others.

Thompson et al. [59] reported a methodology for the analysis of GLYP and AMPA in honey using FMCO-Cl derivatization prior to LC-MS/MS determination. In this case, two important challenges related to the derivatization process were identified. Firstly, the quite acidic nature of the concentrated honey sample (40% w/v) reduces the buffering capacity of the sodium tetraborate solution, lowering the final solution pH to 9.0. This suboptimal pH decreases the efficiency of the reaction. To address this problem, a 0.1 M sodium carbonate solution was used to raise the optimal pH value of 9.5, ensuring proper reaction conditions. The second major challenge involves the low concentration of FMOC-Cl (1–10 mg/mL) used in the reaction. At these levels, the amount of derivatizing agent is not enough to fully derivatize the analytes. Honey is a complex matrix rich in amino acids; thus, competitive reactions can occur, reducing the availability of FMOC-Cl for GLYP and AMPA. To overcome this problem, the authors recommend using 0.2 mL of a 50 mg/mL FMOC-Cl solution to guarantee the complete derivatization of the analytes.

Chamkasem and Harmon [82] developed a method for determining glyphosate, glufosinate, and AMPA in organic soybean and corn using LC-MS/MS. The analytes were separated on a silica-based column and 50 mM ammonium formate as the mobile phase. Different pH values were tested, with 2.6 found to be ideal for glyphosate retention; in addition, higher pH produces more retained and tailing peaks. Fortification levels of 0.1, 0.5, and 2.0 μg/g were selected for all analytes. Recovery percentages for glyphosate ranged from 89 to 107% in corn, and 96 to 103% in soybean. For AMPA, recovery values were low when no IL-IS was added. However, with the addition of IL-IS, recoveries increased to 96–113% for corn and 101–108% for soybeans. In the analysis of honey, the derivatization of glyphosate faces some challenges: the acidic nature of concentrated honey solutions results in a low efficient buffering capacity of sodium tetraborate solution, keeping the pH lower than necessary for the reaction to have a good yield. In this situation, 0.1 M sodium carbonate (pH = 9.5) works better for glyphosate derivatization [59].

Furthermore, additional derivatizing agents have been employed to enhance the detection and quantification of GLYP and AMPA in food samples. Ferreira de Souza et al. [64] developed an effective protocol that involves the oxidation of GLYP with calcium hypochlorite, followed by a reaction with OPA-MERC to yield a fluorescent derivative with a limit of quantification (LOQ) of 40 µg/kg and a coefficient of variation of less than 20%. Moreover, Zhang et al. [110] used a sensitive pre-column derivatization HPLC method for GLYP and AMPA detection in soybean samples. They used 3,6-dimethoxy-9-phenyl-9H-carbazole- 1-sulfonyl chloride (DPCS-Cl) as a labeling agent under alkaline conditions. The LOD was 0.02 and 0.01 ng/mL for GLYP and AMPA, respectively, with recovery values higher than 95%.

In this context, the use of derivatization methods constitutes a significant drawback. These reactions typically require extended reaction times, are challenging to control, and may leave residual derivatizing agents in the reaction medium. Additionally, the poor stability of the resulting derivatives often hinders the achievement of the desired levels of accuracy and precision in the analysis [103].

In addition to chromatographic methods, antibody-based immunoassay techniques have been developed to identify and quantify pesticides in food samples. These techniques offer high sensitivity, specificity, and rapid analysis capabilities for routine screening and monitoring of pesticides in complex matrices. Notable examples include enzyme-linked immunosorbent assays and surface plasmon resonance (SPR) biosensors. These methodologies rely on specific binding interactions between antibodies and target analytes [35]. Selvi et al. [122] developed a straightforward competitive immunoassay utilizing avian antibodies for the detection of GLYP in foods, achieving a limit of detection (LOD) of 2 µg/kg and recovery values ranging from 9 to 134%. Bettazzi et al. [123] introduced an electrochemical competitive immunoassay for detecting GLYP in liquid foods, employing antibody-modified magnetic nanoparticles with an LOD of 5 ng/L and an LOQ of 30 ng/L. Côco et al. [124] developed a biosensor based on the localized SPR of gold nanoparticles for the detection of GLYP in food samples. This sensor represents a cost-effective alternative with high specificity for detecting GLYP in foodstuffs. Moreover, gold nanorods have been used to detect GLYP from tomato juice [125]. Similarly, quantum dots capped with thioglycolic acid and charged with gold nanoparticles for GLYP detection have been synthesized [126], while graphitic carbon nitride has also been explored for detecting GLYP in orange juice [127]. Moreover, other absorption and emission methods have been used to detect GLYP and AMPA in complex matrices, including surface-enhanced Raman scattering, nuclear magnetic resonance, and chemiluminescence–molecular imprinting sensors, as well as electrochemical sensors (amperometry and voltammetry methods), capillary electrophoresis, and cell biosensors. However, most of them have been used for detecting GLYP and AMPA in soil and water [86]. These immunoassay techniques showed high recovery rates but did not evaluate selectivity, specificity, or matrix interference.

4.4. Analytical Method Validation

Analytical method validation is an essential process to guarantee the quality of the results obtained. In the analysis of pesticide residues in food samples, the use of validated methodologies is crucial as far as satisfactory analysis is concerned [128]. Several quality standards such as ISO/IEC 17025 and regulatory bodies like the FDA require analytical laboratories to validate all newly developed analytical methods to ensure they comply with the accuracy, precision, and applicability criteria for producing reliable results [129].

The SANTE guidelines are the standard document for analytical quality control and method validation for pesticide residue analysis in food samples. The objectives of this document include defining the procedures and criteria for method validation. SANTE requires the estimation of linearity, matrix effect, LOQ (quantitative methods), LOD (qualitative methods), specificity, recovery, precision and robustness [104].

The limit of detection (LOD) refers to the lowest concentration of analyte that can be detected by the analytical method, while the limit of quantification (LOQ) is the lowest analyte concentration that can be reliably measured with acceptable repeatability and trueness [130,131]. It is important to note that the matrix effect can influence LOD and LOQ values. To determine the actual limits, experiments should be conducted using samples analyzed under the method conditions. While the LOD of a method describes its ability to detect low concentrations of analytes, if an analyte is not detected, this does not imply its absence in the sample but rather that the method is unable to detect it [128]. Additionally, it is important to consider that LOQ is usually higher than LOD, meaning that, in some cases, the analytes may be detected but not quantified. According to the SANTE guidelines, the LOQ is preferred over LOD to avoid errors. In the validation of analytical methods for pesticide residue analysis in food, the LOQ must be lower than the established MRL for each pesticide [104].

Accuracy is typically represented as the percentage of analyte recovered compared to the actual amount spiked into the sample. Recovery studies are usually conducted by spiking pesticide-free samples (sample blank) with solutions of known concentration. The samples are then analyzed following the established protocol, and recovery percentages are calculated [132]. According to SANTE, to achieve adequate accuracy, recovery levels between 70% and 120% must be reached as the average recovery for each spiked level tested [104]. On the other hand, precision is defined as the degree of agreement between independent analytical results when the analysis is carried out under defined conditions, on different portions of the same sample by the same analyst (repeatability), by a different analyst (reproducibility) or in a different laboratory (intermediate precision) [130,131]. Precision studies should be conducted by repeating the experiment on different days and calculating the relative standard deviation (%RSD). In pesticide residue analysis, a %RSD value of <20% for both repeatability and reproducibility must be reached for each spike level tested to demonstrate the precision of the methodology [104].

Recent analytical studies on GLYP and AMPA analysis in food matrices reveal substantial variability in methodological rigor, particularly regarding adherence to internationally recognized validation protocols. Although most studies report essential parameters such as LOD, LOQ, accuracy, and repeatability, many fail to discuss matrix effects, uncertainty estimation, or full compliance with guidelines. This undermines reproducibility and complicates comparisons between laboratories.

Table 7. Summary validation practices in glyphosate and AMPA detection studies.

Analytical Method	Matrix Effect Handling	Uncertainty Estimation	Guideline Referenced	Validation Summary	Ref.
LC-MS/MS	Controlled with standards	Estimated measurement uncertainty provided	SANTE/11312/2021	Fully validated for GLYP and AMPA in water; high accuracy and reproducibility.	[133]
LC-MS/MS	Matrix effects assessed; internal standards used	Not specified	Not specified	Fast and reliable for GLYP and AMPA; suitable for biomonitoring.	[134]
LC-MS/MS	Matrix effects evaluated; internal standards used	Not specified	Not specified	Acceptable accuracy and precision good for GLYP/AMPA.	[135]
LC-MS/MS	Matrix effects evaluated; internal standards used	Not specified	Not specified	Validated for GLYP/AMPA in soil; applicable to field samples.	[136]
LC-MS/MS	Matrix effects evaluated; internal standards used	Not specified	Not specified	Robust method for large-scale GLYP/AMPA.	[137]
LC-MS/MS	Matrix effects evaluated; internal standards used	Not specified	Not specified	Suitable for GLYP/AMPA exposure assessment in workers.	[138]
LC-HRMS	Controlled with standards	Not elaborated	Not specified	Sensitive for GLYP in water; limited validation data.	[139]

summarizes the validation practices in GLYP and AMPA detection studies in food samples.

Despite methodological advancements, some of the reviewed studies omitted critical aspects of method validation, particularly the matrix effect assessment and uncertainty estimation. Moreover, references to regulatory standards such as SANTE/12682/2019 (European Commission) [104], AOAC [140], ICH Q2(R1) [141], and ISO 17025 [142] were often absent. These omissions pose challenges for regulatory acceptance and data reproducibility. To advance the field, future validation efforts should explicitly reference recognized guidelines, incorporate a comprehensive evaluation of matrix effects, and quantify uncertainty. This would enhance analytical comparability and increase confidence in reported results for regulatory and public health purposes.

In this context, the ideal method for GLYP and AMPA quantification needs to be suitable for a wide range of matrices and easy to carry out for most laboratories while being robust, selective, accurate, and precise. However, most developed methods for GLYP and AMPA detection are generally suitable for only one matrix or a specific type of matrix in which they were validated. It is a challenge to expand the scope of food samples and complex matrices, and low concentration levels pose a barrier [68].

5. Challenges and Opportunities

The extensive and continuous use of glyphosate (GLYP) in agricultural practices has resulted in its frequent detection, along with its primary metabolite aminomethylphosphonic acid (AMPA), across a wide array of fresh and processed food products and drinking water. Although most detected concentrations remain within the maximum residue limits (MRLs) established by regulatory agencies, the potential for cumulative exposure, particularly in vulnerable populations such as infants and children, continues to raise public health concerns. In this context, the reliable analytical determination of GLYP and AMPA remains a central component of food safety assessment. However, the robustness and applicability of current methods are constrained by numerous physicochemical and methodological limitations.

GLYP and AMPA exhibit high polarity, lack chromophoric groups, and exhibit metal-chelating properties. These characteristics impede their retention in reversed-phase chromatographic systems and make them undetectable using conventional ultraviolet or fluorescence detectors. Their chelation with metal ions in food matrices presents further complications regarding extraction efficiency and detector response. Derivatization methods, including those based on FMOC-Cl, have enabled detection by introducing fluorescent groups, but these protocols require strict pH and temperature control, are time-consuming, and often yield unstable derivatives. Alternatively, liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) has enabled more sensitive and selective detection without the need for derivatization. However, its susceptibility to matrix-induced ion suppression, high operational cost, and reliance on specialized expertise limit its routine use in many laboratories.

The performance of analytical methods is also highly dependent on the nature of the food matrix. In plant-derived samples, derivatization-based HPLC methods can be effective when reaction parameters are rigorously optimized. In contrast, complex food matrices, such as milk, cheese, honey, or processed foods often require LC-MS/MS methods, typically combined with solid-phase extraction (SPE) or QuEChERS to improve selectivity and recovery. These enhanced protocols have demonstrated improved accuracy and sensitivity but also require additional time, resources, and validation efforts for each matrix. The reviewed literature highlights that matrix complexity not only affects extraction and cleanup efficiency but also impacts the overall method precision, particularly in the absence of matrix-matched calibration or isotopically labeled internal standards.

Some major challenges remain central to the determination of GLYP and AMPA in food samples:

• The high polarity of glyphosate and AMPA prevents their retention on reversed-phase chromatographic columns and limits the use of traditional extraction solvents, such as ethanol, methanol, and acetonitrile.
• Their chelating ability facilitates the formation of complexes with metallic elements commonly found in food matrices.
• The lack of chromophore groups on the GLYP structure hinders its detection using fluorescence and diode array detectors.
• The low concentrations of GLYP and AMPA present in foods necessitate the implementation of more sensitive and robust methodologies, with detection and quantification limits tailored to each food group and the established maximum residue limits.
• Compliance with international regulations for glyphosate and AMPA residue limits presents a significant analytical challenge.

Solvent-based extraction methods using water or methanol are favored for their simplicity and cost-effectiveness, particularly in less complex plant matrices. However, their performance is significantly reduced in high-fat or high-protein matrices. More advanced methods like SPE and QuEChERS have been developed to address these shortcomings by improving cleanup and reducing co-extractive interference. Yet the need for matrix-specific optimization can limit reproducibility across food types. Importantly, derivatization-based HPLC methods are less suitable for high-fat and animal-based matrices, where recovery variability tends to be higher, making LC-MS/MS the preferred option despite its cost and operational constraints.

International guidelines such as the QuPPe method have been instrumental in standardizing the analysis of polar pesticides in food products. Nonetheless, their broader applicability to animal-derived or processed foods remains insufficient. As evidenced in multiple studies, protocols developed for products such as infant formula, eggs, and milk require significant modifications and validation to achieve consistent results. Moreover, the lack of harmonized calibration strategies contributes to the limited transferability and comparability of results across laboratories and regulatory contexts. Many researchers agree that the complexity of food matrices and the variety of samples make it difficult to apply these guidelines universally.

To address these limitations, future efforts should focus on developing derivatization-free methods that are less sensitive to matrix variation. The integration of internal standards and matrix-matched calibration protocols would significantly enhance interlaboratory comparability and method robustness. Method validation must be extended to include a broader array of animal-derived and processed foods to more accurately reflect real-world dietary exposure. Streamlining sample preparation workflows is also crucial for facilitating high-throughput applications in routine testing laboratories, particularly in low-resource settings. Additionally, high-resolution mass spectrometry and non-targeted screening approaches should be employed to identify coformulants or transformation products that may contribute to a cumulative concentration of GLYP and AMPA in food matrices. Finally, greater international cooperation will be crucial in developing harmonized protocols that ensure global consistency in GLYP and AMPA monitoring.

Taken together, while substantial progress has been achieved in the analytical determination of GLYP and AMPA, existing methodologies remain limited by challenges related to analyte chemistry, matrix variability, and operational demands. A coordinated interdisciplinary approach is necessary to refine and validate methods that are both sensitive and broadly applicable, thereby supporting effective food safety surveillance and regulatory compliance worldwide.