Simultaneous Determination of Ultraviolet Absorbers and Antibacterial Agents in Textiles by Ultra-High Performance Liquid Chromatography / Orbitrap High Resolution Mass Spectrometry

This paper reported a new analytical method for the simultaneous determination of seven benzotriazole ultraviolet absorbers and seven antibacterial agents in textiles. After ultrasonic extraction for the textile samples in methanol, the solutions were analyzed by ultra-high performance liquid chromotagraphy/orbitrap high resolution mass spectrometry (UPLC/Orbitrap HRMS). It showed that a good chromatographic separation for these target compounds was achieved by a Hypersil GOLD column (100 mm × 2.1 mm × 1.9 μm) with a gradient elution of methanol and 0.1% aqueous formic acid solution (containing 0.5 mmol/L ammonium acetate). Triclosan and 4-chloro-3,5-dimethyl phenol (PCMX) were detected by the orbitrap HRMS in an electrospray ionization (ESI) negative mode while the other twelve target compounds were detected by orbitrap HRMS in ESI positive mode. Full scan experiment was performed over the range from m/z 100 to m/z 500. These target compounds were routinely detected with mass accuracy below 2 × 10 (2 ppm) at the optimized conditions. The results showed that the limits of detection (LODs) were in the range from 0.1 to 0.3 μg/kg. The blank samples were spiked at three levels and their average recoveries varied from 80.5% to 96.3% while the relative standard deviation (RSD) changed from 3.2% to 9.9%. The present method was also applied for the determination of those ultraviolet absorbers and antibacterial agents in the commercial textiles.


Introduction
Functional finishing such as ultraviolet-resistance and antibiosisis often applied to textiles in order to improve their performance in use [1]- [8].Benzotriazoles have an excellent absorption capacity to the ultraviolet light due to a phenolic group attached to benzotriazole structure, thus they are widely used as absorbers for enhancing the ultraviolet-resistance property of textiles [9].Isothiazolinones, triclosan and 4-chloro-3,5-dimethyl phenol (PCMX) are the chemicals that have a broad spectrum of activity against fungi and bacteria, they are widely used as the antibacterial agents for the control of microorganisms in textile [10] [11].
However, the quantitative analysis can be carried out only when the components have been separated completely from each other in the conventional HPLC/MS or HPLC/MS-MS method.Recently, an untargeted approach, based on high resolution mass spectrometry(HRMS) using orbitrap analyser, has been used in the determination of organic contaminants [46].In this approach, all ions obtained are monitored without any pre-selection and the identification is achieved according to the exact mass of the analyte(s).A similar method based on anultrahigh performance liquid chromatography/orbitrap high resolution mass spectrometry (UPLC/Orbitrap HRMS) has been used for determining benzotriazoles in textiles [47].We believe that it is possible to determine simultaneously fourteen target compounds (mentioned above) commonly presence in textiles by an improved method based on UPLC/Orbitrap HRMS.
The objective of this study was to develop a method for the simultaneous determination of seven benzotriazoles and seven antibiosis in textiles based on ultrasonic extraction and UPLC/Orbitrap HRMS.The main focus was to explore the capabilities of the orbitrap in the full-scan acquisition mode with high resolution for the quantification of these compounds.The sensitivity, accuracy and related performance characteristics of the present method were also evaluated.and BIT (CAS No. 2634-33-5, purity 99.2%) were purchased from Sigma-Aldrich (St. Louis, MO, USA).Formic acid (purity 98.0%) ammonium acetate (purity 98.0%) were purchased from CNW Technologies GmbH (Dusseldorf, Germany).Methanol (HPLC purity) was purchased from Merck (Darmstadt, Germany).All organic solvents were HPLC-grade.
Eightpositive samples were all available commercially.Sample 1 was white cotton cloth (containing UV-327).Sample 2, 3 and 4 were army green woven polyester, beige woven polyester and white woven cotton, respectively (containing triclosan).Sample 5 was printed silkworm silk (containing PCMX).Sample 6, 7 and 8 were plain woven men's shirt, plain woven dyeing women cotton top and woven silk dress, respectively (containing OI).

Sample Preparation
Textile samples were cut into pieces of about 5 mm × 5 mm by automatic system prototype.Approximately 1.0 g of textile samples were weighed into 250 ml grinding mouth Erlenmeyer flask.20 ml methanol was added into this flask.This flask was caped and ultrasonically extracted at 40˚C for 30 min.The extraction solution was filtered into a heart-shaped bottle and evaporated to near dryness in a vacuum rotary evaporator.The heart-shaped bottle was then transferred to a nitrogen blowing instrument and blown slowly to dry with dry nitrogen.The residue was dissolved in 1ml methanol and the resulting solution was filtered with 0.22 µm membrane and analyzed by UPLC/Orbitrap HRMS technique.Appropriate dilution was carried out before analysis if necessary.

Preparation of Spiked Samples
Approximately 1.0 g of three different kinds of blank textiles swatches, such as white cotton substrate, white polyester substrate and cotton/polyester blended fiber, which were cut into pieces of about 5 mm × 5 mm by automatic system prototype, were weighed into 250 ml grinding mouth Erlenmeyer flask.A certain amount of standard stock solutions of fourteen target compounds was added to submerge the blank textile samples.After an equilibration time of 24 h, the solvent was eliminated using a gentle stream of dry nitrogen.Then these samples were used for the spiked recovery experiments.

Orbitrap HRMS Conditions
The Dionex Ultimate 3000-Q Exactive ultra-high performance liquid chromatography/orbitrap high resolution mass spectrometry system (Thermo Scientific, MA, USA), equipped with an electrospray ionization (ESI) source in positive and negative mode, a Dionex ultimate 3000 RS pump (Thermo Scientific, MA, USA) and a Dionex ultimate 3000 RS autosampler (Thermo Scientific, MA, USA) was chosen for the qualitative and quantitative analysis of fourteen target compounds.The spray voltage, capillary temperature, and auxiliary heating gas temperature were set at 3500 V, 320˚C and 350˚C, respectively.The rate of flow of sheath gas and auxiliary gas were set at 30 arbitrary units and 10 arbitrary units, respectively.Chromatograms were recorded under the full scan mode with the resolution of 70000, over a range of m/z 100 -m/z 500.The width of ion extraction window of Xcalibur 2.2 software was 5 × 10 −6 (5 ppm).The whole analytical parameters of UPLC/Orbitrap HRMS for fourteen target compounds were shown in Table 1.

Results and Discussion
In this work, fourteen target compounds were studied, including seven kinds of benzotriazole ultraviolet absorbers, five kinds of isothiazolinone antibacterial agents, PCMX and triclosan.To obtain the best analytical conditions, the optimization of the extraction procedure, the chromatographic separation and mass spectrometry parameter was necessary, especially when the method developed includes a mixture of compounds with a wide range of physicochemical properties.

Procedure for the Sample Preparation
The subjects of this study could be divided into four categories: benzotriazole ultraviolet absorbers, isothiazolinone antibacterial agents, PCMX and triclosan and their physicochemical properties varied greatly.Various methods aiming to extract these target compounds from textiles, such as microwave-assisted extraction, ultrasonic extraction and accelerated solvent extraction, were reported previously, among which ultrasonic extraction was a common method for the extraction of target compounds from textile samples.Target compounds in eight positive samples were first ultrasonically extracted using methanol as the extraction solvent and the extraction conditions were optimized.As for triclosan, the results demonstrated that the optimal extraction conditions were ultrasonic extraction at 40˚C for 30 min, using 25 ml methanol as the extractions solvent.As for PCMX, the results demonstrated that the optimal extraction conditions were ultrasonic extraction for 15 min, using 20 ml methanol as the extraction solvent.
On the other hand, the extraction temperature had no evident influence on the extraction.For isothiazolinone antibacterial agents, the results demonstrated that the optimal extraction conditions were ultrasonically extracted at 45˚C for 20 min, using 20 ml methanol as the extraction solvent.As for benzotriazole ultraviolet absorbers, the results showed that the optimal conditions were ultrasonically extracted at 45˚C for 30 min, using 20 ml methanol as the extraction

Optimization of Chromatographic Separation and Mass Spectrometric Detection
The qualitative analysis was carried out by the exact mass of quasi-molecular ion and the retention time.For example, the formula of OI is    Two mobile phases such as methanol/water and acetonitrile/water were usually used in LC/MS analysis, and appropriate formic aid was added into water to promote ionization, to enhance the peak intensity and improve the peak shape [48].It had been found by the comparative experiments that there existed better separation among the target compounds when using acetonitrile/water system as the mobile phase.On the other hand, the intensity of each target compounds increased significantly when using methanol/water system as the mobile phase.Similar phenomena were observed in literature [49].Baseline separation is not necessary when UPLC/Orbitrap HRMS technique was applied in the quantitative analysis.Therefore, methanol/water system was chosen as the mobile phase.The peak area and peak shape of the extracted ions were observed when changing the content of formic acid in the range of 0.05% -1.00%.The results demonstrated that the peak area of these extracted ions all reached the maximum when 0.1% formic acid added into water.At the same time, the peak of the extracted ions was sharp and symmetric.A certain concentration of ammonium acetate could significantly increase the peak responses of quasi-molecular ion [M + H] + .Ammonium acetate of various concentrations in the range of 2 to 10 mmol/l were added into the mobile phase to observe the change of the peak area in extracted chromatogram.The results showed that the peak area in the chromatographic measurement increased significantly when 5 mmol/l ammonium acetate was used.Therefore, the mobile phase was at last determined as methanol/0.1% aqueous formic acid solution containing 5 mmol/l ammonium acetate.
To achieve the best separation, a series of trials were performed on the original composition of the mobile phase and the gradient of elution and the determined separation conditions were shown in segment 2.  an unknown sample, qualitative analysis could be performed according to the retention time and the detected exact mass of extracted ion.Obviously, the extracted ion chromatograms of DCOI distinguished completely from that of UV-P.Similar phenomena occurred between UV-327 and UV-328, UV-320 and UV-326, respectively.Quantitative analysis was carried out by the peak area in extracted chromatogram, therefore, overlapping between DCOI and UV-P, UV-327 and UV-328, UV-320 and UV-326, respectively, did not affect the accuracy of quantification.

Method Validation
A validation was performed and several parameters such as linearity, matrix effect, limit of detection (LOD), spiked recovery and precision were studied.The obtained results were shown in Table 5 and Table 6, respectively.
In order to evaluate the linearity, mixed standard solutions of fourteen target compounds were investigated.The calibration curve was based on the peak area in extracted chromatogram versus concentration.As a result, all standards exhibited a good linearity in its own linear range and the correlation coefficients all larger than 0.998.   5.
Recoveries were evaluated at three different spiked concentration levels as shown in Table 6.The spiked blank samples were analyzed using the established method and nine parallel assays were also carried out.Recoveries at three spiked levels ranged from 80.5% to 96.3% while the relative standard deviation (RSD) varied from 3.2% to 9.9%.The method precision was assessed by repeatability and reproducibility studies, expressed as the percent relative standard deviation (RSD%).Repeatability was assessed by the recovery study and values were shown in Table 6.Reproducibility was carried out by extracting and analyzing a positive textile sample available commercially containing 295.4 mg/kgOI in nine different laboratories.
Replicate (n = 2) samples were run in each laboratory and the RSD% value (n = 18) was calculated for nine laboratories, which the RSDs were within 1.5%.

Analyses of These Compounds in the Commercial Samples
The high mass accuracy and full-scan data of the Orbitrap HRMS allow the developed method to assess virtually all of the compounds present in a sample.387 textile samples available commercially were analyzed using the established method.The analysis results showed that different target compounds at various content levels were detected in sixteen samples as shown in Table 7.The

Conclusion
The present study established a new approach for the simultaneous determination of the contents of seven benzotriazole ultraviolet absorbers and seven antibacterial agents in textile samples based on UPLC/Orbitrap HRMS system.The results showed that the present method provided good limits of detection, precision and accuracy in the quantification of these ultraviolet absorber and antibacterial agent compounds.The applicability of the present method was also demonstrated by the analysis of 387 textile samples from commercial sources.

2-( 2 '
-hydroxy-5'-methylphenyl) benzotriazole (UV-P, CAS No. 2240-22-4, purity 98.0%), UV-326 (CAS No. 3896-11-5, purity 98.0%) and UV-350 (CAS No. 36437-37-3, purity 98.0%) were purchased from AccuStandard Inc. (New Haven, Connectinut, USA).CMI (CAS No. 26172-55-4, purity 99.0%), triclosan (CAS No. 3380-34-5, purity 99.5%), PCMX (CAS No. 88-04-0, purity 99.0%) and UV-320 (CAS No. 3846-71-7, purity 99.0%) were purchased from Dr. Ehrensterfer GmbH (Augsburg, Germany).UV-327 (CAS No. 3864-99-1, purity 98.0%) was purchased from ChemService Inc. (West Chester, PA, USA).UV-328 (CAS No. 25973-55-1, purity 98.0%), 2-(2H-benzotriazol-2-yl)-4-(1,1,3,3-tetramethylbutyl)phenol (UV-329, CAS No. 3147-75-9, purity 98.0%) and DCOI (CAS No. 64359-81-5, purity 99.8%) were purchased from Kasei Co. Ltd. (Tokyo, Japan).MI (CAS No. 2682-20-4, purity 99.9%), OI (CAS No. 26530-20-1, purity 99.9%) solvent.In order to comprehensively assess the impact of three factors such as extraction time (Factor A), extraction temperature (Factor B) and solvent volume (Factor C) on the extraction amounts, orthogonal experiments were performed as shown in Table 2.The extraction amounts of eight positive samples were detected under condition No. 1 to condition No. 9.The optimal extraction conditions varied from each other for these eight positive samples.The orthogonal experimental data in Table 2 were analyzed to obtain the maximum gap and optimal conditions for each positive samples, as shown in Table 3. Obviously, extraction time (Factor A) has the largest effect on the extraction amounts while extraction temperature (Factor B) has the smallest effect on the extraction amounts.Taking three factors into consideration, two programs such as A1B2C1 and A2B2C1 were chosen to be alternate optimal conditions and marked as No. 10 and No. 11, respectively, in Table 2.The extraction amounts of eight positive samples were all larger in condition No. 10 than in condition No. 11.At the same time, the extraction amounts of eight positive samples were larger than or close to the maximal values under conditions No. 1 to No. 9. Therefore, the optimal ultrasonic extraction conditions were as follows: the extraction time was 30 min, the volume of extraction solvent was 20 ml, and the extraction temperature was 40˚C.Eight positive samples were ultrasonically extracted under such optimal conditions, using methanol, ethanol, acetone, acetonitrile, dichloromethane, trichloromethane, t-butyl methyl ether, petroleum ether, acetone/n-hexane (1:1, V/V) and acetone/dichloromethane (1:1, V/V), respectively, as the extraction solvents and the results were shown in Table 4.For Samples 1, 6, 7 and 8, the extraction amounts reached the maximum values when using methanol as the extraction solvent.For Samples 2, 3 and 4, the extraction amounts reached the maximum values when using trichloromethane as the extraction solvent.For Sample 5, the extraction amount reached the maximum value when using ethanol as the extraction solvent.For all these eight positive samples, the extraction amount reached the maximum value or close to the maximum value, or at least reached alarger value when using methanol as the extraction solvent.Therefore, methanol was chosen as the extraction solvent taking all factors into consideration.
Figure 1(a).There appeared a strong peak at m/z 214.12580.The quasi-molecular ion peak of OI was shown in Figure 1(b), with a resolution of 70,000.The

Figure 3 .
Figure 3. Extracted ion chromatograms of fourteen target compounds in the mixed standard solution.

4 A
expressed as the matrix-matched calibration slope to solvent calibration ratio in the whole calibration range.Blank textiles were extracted and diluted with different dilution factors such as 1200, 3000 and 6000, and then a series of matrix-matched standards of fourteen target compounds were prepared with the blank extracts mentioned above.The results showed that the matrix effects of fourteen target compounds in three different dilution factors, ranging from 97% to 109%, were not significant and could be neglected.To obtain a better sensitivity, this work chose 1200 as the dilution factor for the textiles.Previous studies have typically calculated LOD as the concentration level as a signal-to-noise ratio of 3 (S/N = 3), regardless of the low noise level of ion chromatograms extracted by high resolution mass spectrometry (HRMS) with a mass extraction window of ±5 ppm.Therefore, the establishment of an LOD based on S/N value was not realistically feasible.This work used an alternative approach to estimate the LOD, that is to say, matrix-matched standard solutions were diluted successively to obtain the lowest concentration that could be repeatedly determined with a low RSD value during a longer time period.The LOD values of fourteen target compounds varied from 0.1 to 0.3 µg/kg.RSD values calculated from six repeated injections at the LOD level were as low as 3.5% -11.7%.The retention time, linear range, linear equation, correlation coefficient, and LOD of fourteen target compounds are shown in Table UPLC/Orbitrap HRMS chromatogram of a positive sample was shown in Figure4(a), in which a sharp peak appeared at 6.087 min, nearly close to the retention time of UV-327 (t R = 6.091 min).The corresponding full-scan mass spectrometry was shown in Figure4(b), in which only the region from m/z 357 to m/z 359 was demonstrated.The theoretical exact mass of UV-327 was m/z 358.16807 and its extracted ion window was 5 ppm, that is to say, from m/z 358.16628 to m/z 358.16986.Only one peak appeared at m/z 358.16745 in this region and the error of the accuracy between this peak and the theoretical exact mass of UV-327 was −1.73 ppm.Therefore, this peak was the quasi-molecular ion of UV-327.UV-327 was proved to exist in this positive sample according the retention time and the detected exact mass of the quasi-molecular ion.

Figure 4 .
Figure 4. (a) UPLC/MS chromatogram and (b) full-scan mass spectrum of a real sample.

Table 3 .
Analysis of the orthogonal experimental data.

Table 4 .
The extraction effects of different solvents (mg/kg).

Table 5 .
Linear relationship and limits of detection.

Table 6 .
Recoveries of fourteen target compounds in three different kinds of blank textiles (n = 9).

Table 7 .
Analysis results of real samples.