Study on the Stability of Antibiotics, Pesticides and Drugs in Water by Using a Straightforward Procedure Applying HPLC-Mass Spectrometric Determination for Analytical Purposes

: The stability of analytes is a critical point in chemical analysis, especially in the ﬁeld of trace levels residue analysis. Nowadays, due to advances in analytical technology and in separation sciences, the analyses of water have been improved. Unfortunately, in this context, one of the most critical issues in water analysis include compound stability from sampling station to laboratory procedures. This study was carried out to explore the stability of several compounds in water from sampling to analysis concerning analytes reported in implementing decision 2018/840—Watch List. During method development and validation, the stability of compounds was investigated to detect the best operating conditions concerning sampling, extraction and analysis. In this paper, we report a study on the stability of antibiotics, pesticides and drugs in water determined using a straight-forward procedure applying mass spectrometric detection for analytical purposes. The laboratory tests were performed in Milli-Q water and surface water by analyzing samples through direct injection, solvent mixture (Water/ACN) and solid phase extraction system from time 0 to 168 h. All the analytes of the WL are stable in aqueous solutions with the addition of at least 25% ACN even after 168 h, and the analytes have shown a matrix effect on recovery of some analytes such as Famoxadone from sampling results (recovery in surface water 72%). For all the analytes investigated, recoveries were between 70 and 130% by using SPE procedures before UHPLC-MS/MS analysis, which is in good agreement with method validation procedures.


Introduction
The growth of industrialization and excessive utilization of available resources by humans have resulted in the emergence of water pollution as one of the global environmental issues. Indeed, in the last years, traces of several pollutants, typically at levels of nanograms per litre, were found in environmental matrices such as water, soil and sediments [1][2][3][4][5][6].
Since the adoption of the Water Framework Directive, the EU has taken measures to tackle pollution of freshwater ecosystems (Directive 2000/60/EC) [6] through a system of structured prioritization and, in this context, one of the most important action to preserve water from pollutant was the Watch List (WL) directive. The first WL, published by Decision (EU) 2015/495, included several substances, such as sunscreens, drugs, hormones, neonicotinoids, pesticides and antibiotics [7,8], and subsequent revisions have updated the list of substances to be monitored. In June 2018, the second updated version of the EU Watch List made its appearance (Decision 2018/840) [9], and the third updated version of the EU Watch List was published in August 2020 (EU) 2020/1161 [10].
Outlined in the Watch List directive, member states must analyze surface water in order to identify and quantify several pollutants. Nowadays, due to advances in analytical technology and in separation sciences, water analyses have been improved [11,12]. Unfortunately, in this context, one of the most critical issues in water analysis includes compound stability from sampling station to laboratory procedures. For example, some of compounds reported in Watch List show instability in water depending on several physical chemical properties such as solubility, soil/liquid partition coefficient and pH [13,14].
In detail, quantification of Famoxadone and Metaflumizone in water can be considered a challenge in analytical chemistry due to their instability/precipitation, degradation byproducts and chemical reaction properties [15]. For these reasons, in order to identify and quantify the analytes reported in the Watch List, several organizations have been making efforts aimed at validating analytical methods able to determine all compounds in entire samples [16]. Some researchers have made laboratory tests to study the stability of these compounds in water with regard to sampling and analysis conditions such as stability in water, transformations and time from sampling to analysis.
In this paper, we report a study on the stability of antibiotics, pesticides and drugs in water determined using a straightforward procedure by applying mass spectrometric determination for analytical purposes. The laboratory tests were performed in Milli-Q water and surface water by analyzing samples by direct injection, solvent mixture and solid phase extraction from time 0 to 168 h.
Isotopically labeled compounds (Imidiacloprid D-4, Velanfaxine D-6, Tetraconazole D-9 and Fluconazole D-4 Ciprofloxacin 13 C) were used as internal standards (IS), and were purchased from LabService Analytica. The purity grade of all standards was always above 94%.
For Calibration procedures, a mixed stock solution was prepared by serial dilution in acetonitrile (ACN) and stored at −20 • C in the dark to avoid possible photodegradation. In detail, to obtain a mix solution at 250 µg/L in ACN, 25 µL of each compound at 100 ug mL −1 was diluted with 10 mL of ACN. After that, 200 µL of mix solution at 250 µg/L was once again diluted with 10 mL of ACN to obtain a mix solution at 5 µg L −1 . Calibration standard solutions ranging from 5 to 500 ng L − in water + ACN mixture (75:25) were prepared by serial dilution from mix solution at 5 µg/L.
Intermediate mixed solutions containing all analytes and all labeled compounds were prepared weekly. Aqueous acetonitrile (75:25) working standard solutions were renewed before every analytical run to prevent precipitations.
High purity water was prepared using a Millipore Milli-Q purification system. Stock solutions were prepared in ACN and were stored at −18 • C in amber glassware. To avoid standard degradation, calibration working solutions were prepared by serial dilutions of stock solutions in Milli-Q water before each calibration. After reviewing literature data [9] concerning SPE extraction procedures, 47 mm diameter Empore™ SPE Disks (active group polystyrene-divinylbenzene (SDB-XC)) were used with a SPE-DEX 5000 Horizon Technology for extraction procedures.
LC/MS Acetonitrile grade solvents and formic acid 98% were acquired from (Merck).

LC-MS Instrumentation
Analyses were performed without extraction procedures except for SPE investigation studies. Separations were performed by using ultra high-performance liquid chromatograph (UHPLC) consisting of binary pump EXION LC Sciex pump, a DGU-20A 5R de-gassing unit, a SIL-30AC autosampler equipped with a 50 µL loop, a CTO/20AC thermostat column compartment and a CBM-20A module controller, as well as other components.
The EXION LC SCIEX system was coupled to a 6500 plus Q-Trap mass spectrometer (Sciex), equipped with a Turbo V interface by an ESI probe.
The experimental operating conditions were optimized by standard infusion to detect best ionization conditions and fragmentation.
Mass spectrometry optimal parameters and transitions used are reported in Tables 1 and 2, respectively. Two transitions were used for all analytes: one for quantification and one for qualification. The compounds were separated by using a CORTEX T3 analytical column (150 mm; 4.6 mm; 5 µm) using as mobile phase a mixture of water formic acid 0.05% + ammonium formate 5mM and methanol formic acid 0.05% + ammonium formate 5mM in gradient mode. Elution conditions are reported in Table 3. Analysis was performed by multiple reaction monitoring (MRM) both in positive and negative ionization mode by using m/z, decluttering potential and collision energy, as reported in Tables 2 and 3.
Analytes were identified both by comparing their retention times (RT) with the RT of the standards and by qualifier ions. Two selected reaction monitoring (SRM) transitions were recorded for each compound: one for quantification and the other one for confirmation. Time-specific SRM windows were set for each retention time in order to enhance sensitivity.
The entire system was controlled via the Analyst software (SCIEX,), while quantifications of the analytes were performed with Multi-quant 3.0 (SCIEX,).
Quantification was based on the peak area for each compound, and baselines were adjusted manually when necessary.

Data Validation
The method was validated in terms of detection limits, accuracy and precision usually reported in literature [17]. The limits of quantification (LOQs) were estimated by 10-time standard deviation calculated at the first level of calibration curve prepared in Millipore Milli-Q. Accuracy and precision (RSD%) were calculated on a data set of six analyses conducted on water-spiked samples at LOQ. All data obtained were subject to different statistical tests. Normal distribution was evaluated by using the Shapiro-Wilk test, while outlier data were checked by the Dixon test. Table 4 summarizes the results obtained for these performance parameters. For all the analytes, accuracy at LOQ Levels ranged from 70 to 130%, while precision (in term of RSD) were less than 20%.

Test Performed
In order to investigate the chemical stability of the analyte, three different systems were spiked with an analyte mixture at a concentration of 100 ng L −1 and analyzed.
All samples were stored at 4 • C. The surface water used was collected from a previously analyzed river in the Lombardia region.
The result obtained are reported in Table 5 and as percentage recoveries. The data reported in Table 5 show that analytes Clotrimazole, Miconazole, Ipconazole, Famoxadone and Metaflumizone are unstable in aqueous solution without organic solvents, while they show partial stability in the solutions obtained from aqueous solution + ACN 25%.
This trend confirms the insolubility of several substance in water as reported in literature [18].
In the first approximation, it is possible to hypothesize that low recoveries can be ascribed to the difference in solubility of the analytes in water and water with added organic solvent.
In fact, as reported in literature, several substances such as Metaflumizone, Fomaxadone and Miconazole show hydrophobic behavior.
The best stability results were obtained in the mixture Milli Q water + ACN 25% (percentage recoveries from 104 to 124%), while in surface water + ACN 25% some analytes show a decrease in stability. However, in the system surface water + ACN 25%, the analyte shows recoveries in the range from 70 to 130% in accordance with requirements for validation procedures. This behavior would seem to highlight a possible matrix effect that decreases the recoveries of some analytes in surface water.
Furthermore, the analytes that show lowest recovery values in water with added ACN appear to have ideal behavior in the QC control sample prepared in milli-Q water.
To investigate the effects of particulate matter concerning analyte stabilities, several tests on surface water that was unfiltrated, filtrated and unfiltered were performed. The analytes were extracted by ultrasound in order to migrate analytes from particulate matter to solutions, and solutions were analyzed during time (0-24 h). The tests we performed are reported in detail below, and the results are reported in Table 6.
(j) Surface Water filtrated + ACN 25% + analyte at 100 ng L −1 analysis performed after 24 h; (jj) Surface Water + ACN 25% + analyte a 100 ng L −1 analysis performed after 24 h; (jjj) Surface Water + ACN 25% + analyte at 100 ng L −1 a after 24 h sonicated by ultrasound and analyzed; (jjjj) Surface Water + analyte a 100 ng L −1 analysis performed after 24 h. Data reported in Table 6 highlight that, in the systems investigated, the best recoveries were obtained by system j (Surface Water filtrated + ACN 25% + analyte at 100 ng L −1 analysis performed after 24 h GREEN COLOR).
These data confirm that the stability of compounds is greater by the addition of ACN even after 24 h both in filtrated water + ACN 25% and in unfiltrated water + ACN 25%.
Furthermore, in order to underline possible effects after 24 h from sampling to analysis, different percentages for spiked samples at 100 ng L −1 of "water + ACN 25% at t0" and "water + ACN 25% at t24 h" are reported in Table 7. If uncertainty values are taken into account, the differences detected cannot be considered relevant.  Finally, according to EU decision, which requires samples to be analyzed in the entire water sample, quantification tests were also carried out by solid phase extraction by using SPE Horizon SPEDEX 5000 system equipped with an extraction disk in SDB and which involves the elution of the sample, the washing of the particulate collected on the filter with ACN and the reunification of the organic phase with the aqueous eluate.
In detail, operating conditions for solid phase extraction procedures using SPE-DEX 5000 Horizon Technology are reported in Table 8. By using this analytical procedure, all analytes, without solubility differences, were collected in water + ACN 100 + 25% v/v.
Analysis was carried out after 168 h from spiked samples, and data obtained are summarized in Table 9. For all analytes investigated, the recoveries after 168 h from spiked operations are in the range from 70 to 130%, in good accordance with requirements from guidelines in analytical chemistry.
In detail, by particulate washing with organic solvent, analytes absorbed from particulate matter can be desorbed from solid matter, eluated, collected and reunified with the aqueous solution.
Based on these considerations, it is possible to conclude that by using SPE procedure, it is possible to analyze an entire water sample according to EU guideline [19,20].

Conclusions
This study has highlighted some critical aspects in the analysis of several chemical compounds during Watch List analyses on water samples.
In detail, the results obtained can be summarized as follows: -Several analytes show a matrix effect: In surface water samples, the analyte Clotrimazole, Miconazole, Ipconazole, Famoxadone and Metaflumizone, if not properly treated, are not stable in aqueous solution even within a few hours from sampling or sample preparation; -All the analytes of the WL are stable in aqueous solutions with the addition of at least 25% ACN even after 168 h from sampling; - The analysis of Watch List compounds can be carried out through the use of the SPEDEX Horizon 5000 system; -For all analytes investigated, the percentage recoveries were between 70 and 130% by using SPE procedures followed by a UHPLC-MS/MS analysis, in good agreement with method validation procedures. Funding: The study was supported by Regione Lombardia financial resources.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: All data were collected in laboratory book and stored in a personal computer adoperated during analyses.