Simultaneous Separation and Detection of Two Genotoxic Impurities in Urapidil Hydrochloride by HS-GC

Introduction

Urapidil Hydrochloride, is a type of uracil compound as shown in Scheme 1 ^1,2,3, it was developed by German firm BYKGULDEN⁴ and is widely utilized for managing essential hypertension, as well as administered intravenously in cases of hypertensive emergencies^5,6, pheochromocytoma⁷_, renal ischemia reperfusion-related renal injury8. Until now the research of urapidil were focused on stress degradation products², synthetic and hydrolytic impurities^9,10, metabolites¹¹, assay¹²and residual solvent¹³ etc, but there are no reports of genotoxic impurities research.

Benzene is a solvent that may be carried in ethanol¹⁴, dichloromethane, and hydrochloric acid¹⁵. Hydrochloric acid may react with ethanol to form chloroethane¹⁶. Due to the usage of ethanol, dichloromethane, and hydrochloric acid in the production process, urapidil hydrochloride may produce genotoxic impurities such as benzene and chloroethane. A number of studies indicate that neither benzene nor its metabolites are effective mutagens, but they are highly clastogenic, causing sister chromatid exchanges, and the formation of micronuclei in animals, as well as cause cancer in humans in various forms^17,18. Humans are strongly suspected of developing cancer from chloroethane because it is a known animal carcinogen. When high amounts of chloroethane and its metabolites are circulating in most organs, Phase II conjugation metabolic elimination reduces tissue essential glutathione pools, which are then reduced by glutathione transferase^19,20,21. Consequently, the presence of benzene and chloroethane in urapidil hydrochloride should be investigated and controlled.

As outlined in the ICH guideline  Q3C (R9) 2024, benzene is classified as a Category 1 solvent, which has a defined concentration limit( 2 ppm)²². The IARC has classified chloroethane as a Class 3 carcinogen, indicating its potential to pose a carcinogenic risk to humans, and ICH M7(R2) also provided the acceptable intakes based on TD50 linear extrapolation method and requiring a limit of no more than 1810μg/day. There are several studies that use GC¹⁷ or GC/MS²¹ to identify benzene and chloroethane in medicines. In particular, GC/MS are not widely accessible in pharmaceutical manufacturing facilities and need highly skilled personnel. The authors employed the DB-624 column in conjunction with headspace sampling(HS) and a flame ionization detector (FID) for this investigation.

Scheme 1: Chemical structure of urapidil hydrochloride  Click here to View Scheme

Material and Methods

Instruments

A FID detector fitted to the Shimadzu 2010 Pro GC was used to quantify two possible genotoxic impurities in sample. A Sartorius SQP and CPA225D electronic balance were used for weighing references and test samples. The gas chromatography is equipped with a fused-silica capillary column coated with 6% cyanopropylphenyl-94% dimethylpolysiloxane(Agilent B-624, 30m×0.32mm, 1.8μm). HPLC grade distilled water was prepared by Arium® Pro Ultrapure Water Purification System(Sartorius Corporate Administration GmbH).

Materials and Chemical

The investigated sample(Batch No.221001~221003) were provided by Chengdu Aupone Pharmaceutical Co.,Ltd. Chloroethane in Methanol(5000μg/ml, Batch No.22040943) were sourced from Tan-Mo Technology Co.,Ltd. Benzene(Batch No.2019102301), N,N-dimethylformamide(Batch No.2019121128), Methanol(Batch No.2021042926), Ethanol(Batch No.2022092101), Acetonitrile(Batch No. 2021052519) and Dichloromethane(Batch No.2022101401) were obtained through Chengdu Chron Chemical Technology Co., Ltd.

Conditions of chromatographic method

The column temperature profile is designed as follows: Initially, the column is set to 40°C and maintained at this temperature for 1 minute to ensure sample equilibration. Subsequently, the temperature is raised to 80°C at a rate of 20°C per minute, followed by a 10-minute hold phase to achieve optimal separation resolution. After completing this gradient segment, the temperature is further elevated to 220°C at an increased rate of 50°C per minute, with a final 5-minute hold to complete the elution process. A detector and an injection port, both set to 220°C. The headspace vial is also maintained at 85°C for 20 minutes to achieve equilibration. This experiment uses nitrogen as a carrier gas, flowing at 0.5ml per minute with a split ratio of 20:1. The test solution and reference solution should be injected separately onto the column, and the chromatogram should be recorded separately. Accurately dispense 5 ml of each solution into a 20 ml headspace vial and seal it securely.

Preparation of solutions

Preparation of diluent

A 1:1 solution of DMF and water was used as diluent.

Benzene stock solution

A 50 mg portion of benzene was accurately quantified and introduced into a 50 mL volumetric flask. Dimethylformamide (DMF) was added incrementally until the solution reached the calibration line, and the mixture was homogenized via vigorous shaking. Subsequently, 1 mL of this solution was aspirated and transferred to a 100 mL volumetric flask. The flask was brought to volume with diluent, and the diluted solution underwent 30 seconds of vortex agitation to ensure complete homogeneity.

Benzene positioning solution

In this phase of the experimental procedure, a precisely measured 1 mL aliquot from the benzene stock solution was carefully transferred into a 25 mL calibrated volumetric flask. Following this transfer, the dilution process commenced by gradually introducing the diluent solution through the flask’s neck. The liquid level was meticulously adjusted until the meniscus aligned perfectly with the flask’s calibration mark at room temperature, and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Chloroethane stock solution

1ml of chloroethane (diluted in methanol solution, 5000μg/ml) was accurately quantified and introduced into a 20 mL volumetric flask The diluent was added until the solution reached the mark, and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Chloroethane positioning solution

A 1 mL aliquot of the chloroethane stock solution was accurately pipetted into a 25 mL volumetric flask. The diluent was added until the solution reached the mark, and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Standard solution

1 mL of the benzene stock solution and 2 mL of the chloroethane stock solution were accurately dispensed using a calibrated pipette, followed by transfer into a 50 mL volumetric flask. The solution was then brought to volume by adding the diluent until the calibration mark was reached, ensuring precise volumetric determination, and the mixture was vortex-mixed for 30 seconds to achieve a uniform and homogeneous solution.

Sample solution

The sample (approximately 500 mg) was carefully measured and introduced into a 20 mL headspace vial. Following this, the diluent was dispensed in an exact volume of 5 mL into the container, which was subsequently hermetically sealed and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Mix stock solution

Based on pre-calculated stoichiometric ratios derived from the formulation protocol, methanol, ethanol, acetonitrile, and dichloromethane were individually metered into designated 100 mL volumetric flasks according to their respective mass percentages. Each solvent was quantitatively delivered, after which the flasks were brought to volume with diluent and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Mix positioning solution

A homogenized three-component solution was prepared by pipetting 1 mL of the stock solution, benzene calibration standard, and chloroethane reference material into a 50 mL volumetric flask. The solution was then brought to volume by adding the diluent until the calibration mark was reached, ensuring precise volumetric determination, and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Mix sample solution

Precisely 500 mg (±0.5 mg) of test material was portioned into a sterile 20 mL headspace vial using analytical-grade spatulas. A 5 mL volume of diluent was dispensed via calibrated pipette into the containment vessel, which was immediately crimp-sealed and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Spiked standard solution

100% Spiked Standard Solution: The standard solution itself was used as the 100% spiked standard solution.

50% Spiked Standard Solution: To obtain the target mixture, a calibrated pipette was used to deliver 0.5 mL of benzene stock solution into a 50 mL volumetric flask. This was followed by the addition of 1 mL chloroethane stock solution, ensuring both components were proportioned as required. The diluent was added until the solution reached the mark, and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

150% Spiked Standard Solution: To obtain the target mixture, a calibrated pipette was used to deliver 1.5 mL of benzene stock solution into a 50 mL volumetric flask. This was followed by the addition of 3 mL chloroethane stock solution, ensuring both components were proportioned as required. The diluent was added until the solution reached the mark, and the mixture was vortex-mixed for 30 seconds to ensure homogeneity.

Spiked sample solution

Each 20 mL headspace vial received a precisely measured aliquot of approximately 500 mg sample, which had been carefully portioned into individual containers prior to subsequent processing. Different concentrations (50%, 100%, and 150%) of the spiked standard solution were added precisely at 5 mL each to the respective vials. The vials were sealed and shaken thoroughly to ensure proper mixing. Each concentration was prepared in triplicate to ensure reproducibility.

Results and Discussions

Optimization of Chromatographic Conditions

The most important criterion to take into account was the resolution (any two peaks’ resolution should be more than 1.5)^20,21. The peak height (sensitivity), symmetry factor (between 0.8 and 1.5), theoretical plates (more than 5000), and method run time (less than 60min) should be considered^23,24.

Based on the residual solvent method of urapidil in the Chinese Pharmacopoeia II, a preliminary analysis method for benzene and chloroethane was established. It is important to consider whether methanol and ethanol affect the detection of benzene and chloroethane, as ethanol was found in the urapidil hydrochloride sample and chloroethane reference substance was a methanol solution.

Appropriate separation conditions were finally chosen by screening the chromatographic column, optimizing the programmed heating program, and raising the split ratio.

Different polar chromatographic columns were chosen for screening because three residual solvents (ethanol, acetonitrile, and dichloromethane), one reference substance diluent (methanol), and two genotoxic impurities (chloroethane and benzene) must be separated. HP-5, DB-FFAP, DB-624, and DB-WAX were the columns that were analyzed; according to the impurities discovered, each had a different polarity, with methanol being extremely polar and benzene being less polar. Separation would not be possible with a column that is either too polar or too low. The widely used DB-624 chromatographic column for residual solvents was finally selected following experimental comparison, and all six compounds were fully separated on it. The final chromatographic conditions meet the set principles: the resolution of each impurity is greater than 1.5, almost all peaks have a greater S/N than 50, the peak shape is symmetrical, the theoretical plate number is greater than 20000, and the analysis time is less than 30 minutes.

Optimization of diluent

Dimethyl sulfoxide (DMSO), DMF, water, and various DMF-water solution ratios were investigated. Although the sample could completely dissolve in DMSO, DMSO tends to degrade into methyl thioether, dimethyl disulfide, and methyl sulfide under acidic conditions due to the sample being a hydrochloride salt^25,26, which may cause interference. DMF was also considered as a solvent, but it showed very low response for benzene, almost no peaks appeared. Pure water as a solvent could easily damage the gas chromatography column, so various DMF-water solutions were tried instead^27,28. Higher sensitivity for most solvents, especially Class 1 solvents was obtained with DMF-water solutions.Benzene and chloroethane reaction comparisons for various DMF-water solution ratios are shown in Figure 1. The ultimate choice of 10% DMF as the diluent was made because the findings show that while the response of benzene as the ratio of DMF increases, there is almost no difference in the response of chloroethane to various ratios of DMF and water. Results indicate that chloroethane reacts roughly the same way in different combinations of DMF and water, while the response of benzene decreases with an increase in the ratio of DMF, leading to the final selection of 10% DMF as the diluent.

Figure 1: The response of benzene and chloroethane in different DMF aqueous solutions Click here to View Figure

Optimization of chromatographic method

An analytical method for benzene and chloroethane in urapidil hydrochloride was established according to the Chinese Pharmacopoeia (Part II). A polyethylene glycol (PEG)-based capillary column (DB-WAX, 30 m × 0.32 mm, 0.5 μm) was employed. The initial oven temperature was set at 40 °C for 2 min, followed by a temperature ramp at 5 °C/min to 80 °C and held for 10 min, then rapidly increased at 50 °C/min to 220 °C and maintained for 5 min. Preliminary results revealed partial co-elution of ethanol and benzene peaks, with a resolution of only 0.23, indicating potential interference from ethanol in the detection of benzene, see Figure 2 for details. To address this issue, the temperature program was systematically optimized to achieve baseline separation of all target solvents and genotoxic impurities (chloroethane and benzene). Post-optimization chromatographic conditions successfully resolved critical peak overlaps, ensuring accurate identification and quantification of benzene in the presence of ethanol.

Figure 2: The chromatogram obtained under unoptimized analytical conditions Click here to View Figure

Optimization of headspace equilibrium temperature

To avoid gas condensation during transmission, which might result in residue, the heating furnace’s temperature is gradually raised from the furnace temperature, transfer line temperature, and quantitative loop temperature to the transfer line temperature²⁹_. Consequently, the ideal range for the furnace temperature is one that is greater than the target substance’s boiling point but lower than the diluent’s.The boiling point of benzene is 80.1°C, chloromethane is -24.2°C, water is 100.0°C, and DMF is 153°C. To evaluate system performance under thermal variation, testing was performed across a controlled temperature gradient spanning 75°C to 90°C, with experimental parameters systematically adjusted in 5°C increments, as detailed in Figure 3. Higher temperatures often resulted in faster dynamic equilibration and greater extraction yields; The procedure reached its optimal condition when mass transfer to the headspace (HS) was nearly instantaneous and complete³⁰. When the temperature exceeds 100°C, decomposition peaks will appear in the sample, affecting its determination; when the temperature is below 80°C, the response of benzene is low. In order to guarantee full volatilization of benzene and chloromethane in the sample while reducing the volatilization of background water, the equilibrium temperature is finally fixed at 80°C after screening.

Figure 3: The result of optimization of headspace equilibrium temperature Click here to View Figure

Optimization of headspace equilibrium time

Another important consideration for HS was equilibration time, which was directly related to the establishment of equilibration between the two phases³¹. The gas and liquid phases are unable to attain equilibrium when the equilibrium time is short. The headspace vial’s airtightness deteriorates once the equilibrium period surpasses 60 minutes, which lowers the chromatogram’s peak area³¹. A comparison was made between the 15-minute, 20-minute, 25-minute, and 30-minute responses, as detailed in Figure 4. Considering both peak response and cost, an equilibrium time of 20 minutes was ultimately selected to meet the detection requirements.

Figure 4: The result of optimization of headspace equilibrium time. Click here to View Figure

Analytical methodology and validation

System suitability

To establish method precision, a six-injection replicate analysis of the reference solution was conducted under chromatographic conditions. System suitability parameters were quantified through statistical determination of relative standard deviation (RSD%) for both retention time precision (t_R) and integrated peak area reproducibility across the benzene-chloroethane analyte pair, together with theoretical plate numbers for both compounds. The theoretical plate numbers for chloroethane were all greater than 50,000, and the RSD values of retention times for both chloroethane and benzene were 0.01%. The RSD values of peak areas were 2.63% for chloroethane and 2.27% for benzene. In Table 1, you can see the theoretical plate counts and %RSD in compliance with ICH guidelines.

Table 1: System suitability of the optimization method

TP*:Theoretical plate; RT^#: Retention time(min); PA^&:Peak Area; TP^@:Tailing Factor; RS^※:Resolution

Specificity

Several chromatograms were recorded with separate injections of Blank solution, benzene positioning solution, chloroethane positioning solution, standard solution, sample solution, and mix sample solution. The results are detailed in Figures 5 to 10. The results show that the blank, test sample, methanol, ethanol, acetonitrile, and dichloromethane do not interfere with the detection of benzene and chloroethane.

Figure 5: The HS-GC chromatogram of blank solution Click here to View Figure

Figure 6: The HS-GC chromatogram of benzene positioning solution Click here to View Figure

Figure 7: The HS-GC chromatogram of chloroethane positioning solution Click here to View Figure

Figure 8: The HS-GC chromatogram of standard solution Click here to View Figure

Figure 9: The HS-GC chromatogram of sample solution Click here to View Figure

Figure 10: The HS-GC chromatogram of mix sample solution Click here to View Figure

(MeOH: methanol; EtCl: choroethane; EeOH: ethanol; ACN: acetonitrile; DCM: dichloromethane; Ben: benzene)

Limits of Detection (LOD) and Limits of Quantification (LOQ)In accordance with ICH Q2(R1) validation guidelines, the limit of detectability(LOD) and quantitation (LOQ) were established to assess the method’s sensitivity³². The LOQ and LOD levels were predicted by the estimation of the S/N NLT10 and NLT3, respectively. The experimental findings have been systematically organized in tabular format (Table 2) for comparative analysis. Six injections of a solution containing two genotoxic impurities from a level of 0.027 μg/ml (benzene, 0.000027%) to 0.100 μg/ml (chloroethane, 0.0001%) were used to show the LOQ; and three injections of a solution containing two genotoxic impurities from a level of 0.013μg/ml (benzene, 0.000014%) to 0.050μg/ml (chloroethane, 0.00005%) were used to LOD.

Table 2: The LOQ and LOQ for the optimized method

Linearity and range

In order to validate linearity, the stock solution of benzene and chloroethane was quantitatively diluted to 200%,150%, 100%, 50%, 20% and LOQ. For the two impurities, a correlation coefficient greater than 0.990 was observed, demonstrating that the proposed GC method exhibits linearity within the tested concentration range. The measured peak area are plotted versus concentration to create calibration curves. Figures 11 to 12 show the calibration curves for both impurities, and Tables 3 to 4 provide the linearity parameters associated with them.

Table 3: The linearity and range for chloroethane(n=3)

Figure 11: Calibration curves for chloroethane Click here to View Figure

Table 4: The linearity and range for benzene

Figure 12: Calibration curves for benzene Click here to View Figure

Repeatability

By injecting six duplicates of the sample solution with preset acceptance criteria for a percent RSD of < 5.0% for repeatability. As shown in table 5, the %RSD of outcomes was within the limits.

Table 5: Repeatability results

ND: Not Detected; NA: Not Applicable; the same below.

Intermediate precision

By injecting six duplicates of the sample solution by different person, instrument, and date with preset acceptance criteria for a percent RSD of < 10% for intermediate precision. As can be seen from Table 6 of the results, the %RSD of the results was within the acceptable boundaries of what was considered acceptable.

Table 6: Intermediate precision results

Accuracy

Method accuracy was verified through spiked recovery studies for two genotoxic impurities at 50%, 100%, and 150% spiked levels. As shown in Table 7, recoveries ranged from 90% to 110%, meeting the validation criteria of ICH Q2(R1).

Table 7: Accuracy results

Solution stability

The stability of sample and standard solutions was evaluated by storing both solutions at ambient temperature (25 ± 2°C) over a 24-hour period. Aliquots were analyzed at predetermined intervals (0, 3, 9, 12, 16, and 24 h) under the validated method conditions. Solution stability was considered acceptable if the relative standard deviation (RSD) across all time points remained below 10%, aligning with the predefined analytical acceptance criteria of the International Council for Harmonisation (ICH) Q2(R1) guidelines. In order to determine the peak area of each impurity, the chromatograms of each impurity were recorded and the peak area was determined over time. Table 8 summarizes the complete results.

Table 8: The standard solution and sample solution stability

Robustness

The method’s robustness was systematically verified in accordance with ICH Q2(R1) regulatory framework by implementing controlled variations of core analytical variables. Key system parameters underwent adjustments encompassing: (1) initial column temperature (±5°C), (2) carrier gas delivery rate (±0.1 mL/min), (3) injection port temperature (±5°C), and (4) FID thermal settings (±5°C). Additional assessments involved one distinct chromatographic condition (different chromatographic column, instrument, and analysts). All experimental deviations met the predefined acceptance criteria (RSD <10.0%), with comprehensive data presented in Table 9. This protocol aligns with the ICH Q2(R1) recommendations for robustness testing in analytical method validation.

Table 9: Robustness results

Real sample detection

The suggested method was used to analyze many lots of urapidil hydrochloride(221001~221003). Peak areas were used to assess each chromatogram, and the outcomes are shown in Table 10. The chromatograms revealed that all samples no included benzene, and the detected amount of chloroethane is far below the 5% control threshold. It was thus shown that the suggested method might be used to the regular quality control of genotoxic impurities in urapidil hydrochloride.

Table 10: Real sample detection results

Conclusion

The establishment and validation of a headspace gas chromatography (HS-GC) methodology facilitated concurrent resolution and quantitative monitoring of two genotoxic impurities – benzene and chloroethane – in urapidil hydrochloride drug substance. Method validation studies revealed exceptional linear correlation (R² >0.990) across the validated concentration range (LOQ-200%), and precision (RSD<5.0%) conforming to stringent validation thresholds prescribed in ICH Q2(R1) for chromatographic impurity methods. The accuracy results, which range from 96.17% to 101.33% for benzene and 94.28% to 99.92% for chloroethane. Method validation, including specificity, sensitivity and robustness, also complied with ICH Q2(R2) recommendations for genotoxic impurity control. This cost-effective analytical approach has been successfully implemented for routine quality monitoring of urapidil hydrochloride in commercial batches.

Acknowledgement

The authors acknowledge Chengdu Aupone Pharmaceutical Co., Ltd. management.

Funding Sources

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Conflict of Interest

The author(s) do not have any conflict of interest.

Data Availability Statement

This statement does not apply to this article.

Ethics Statement-

This research did not involve human participants, animal subjects, or any material that requires ethical approval.

Informed Consent Statement

This study did not involve human participants, and therefore, informed consent was not required.

References 

1. Badert, A.; Zechs, K.; Cromet, O.; Christianst, U.; Ringes, B.; Pichlmayrs, R.; Sewing, K-FR.; , 1994, 24(7), 623-633. https://doi.org/10.3109/00498259409043265
   CrossRef
2. Laximan, V.; Vivek, D.; Bhoopendra, S. K.; Vijaya, M.G.; Gananadhamu, S.; J Pharmacrut Biomed., 2022, 22, 114612. https://doi.org/10.1016/j.jpba.2022.114612.
   CrossRef
3. Ma, J.; Lin, G.; Wang, X.; Li, J.; Wang, X.; Hu, L.; J Liq Chromatogr Relat Technol.,. 2011, 34, 307-316. https://doi.org/10.1080/10826076.2011.547413.
   CrossRef
4. Wang, C.J.; Lu, S.H.; Zhou, Z.G.; Qin, Y.; Drug Eval Res. 2022, 45(8), 1613-1621.
5. Zhang, L.Z.; Liu, F.L; Xiang, H.Y.; Zhan, Z.Y.; Shen, J.; Chem Pap., 2024, 78, 4395-4400. https://doi.org/10.1007/s11696-024-03398-2.
   CrossRef
6. Xie, Y.Z.; Ni, J.M.; Zhang, S.J.; Ding, G.-R.; Feng, J.F.; Medicine, 2019, 98(41), e17352. https://doi.org/10.1097/MD.0000000000017352.
   CrossRef
7. Kumar, A.; Gupta, N.; Gupta, A.; J Anaesth Clin Pharm., 2020. 36(1), 55-56. https://doi.org/10.4103/joacp.JOACP_328_18.
   CrossRef
8. Derya, G.E.; Ayhan, T.; Mustafa, C.G.; Ersen, E.; Selim, Ç.; Songül, D.; South Clin Ist Euras., 2022, 33(2),198-202. https://doi.org/10.14744/scie.2022.89421.
   CrossRef
9. Baker, R.; J Chromatogr.,1987, 393(3), 447.
   CrossRef
10. Jia, X.D.; Zhang, W.; Sun, Z.; Li, Z.L.; Chin J Pharm Anal., 2021, 41(7), 1226-1236.
11. Sturm, E.; Zech, K.; Biomed Mass Spectrom., 1984, 11, 211-216.
    CrossRef
12. Sun, S.J.; Wei, Y.F.; Long, C.J.; Deng, B.Y.; J Chromatogr B., 2015, 1006, 146-150. https://doi.org/10.1016/j.jchromb.2015.10.011.
    CrossRef
13. Wang, C.H.; Lei, K.; Wang, S.X.; Chin Hosp Pharm J., 2013, 33(3), 218-221.
14. Santosh, K.; Neetu, S.; Ram, P.; Renew Sust Energ Rev., 2010, 14(7), 1830-1844. doi:10.1016/j.rser.2010.03.015.
    CrossRef
15. Xi, D.C.; Bu, L.K.; Liang, S.; Chem Enter Manage., 2021, 12, 144-145.
16. Zhai, Z.R.; Zhang, X.W.; Li, X.R.; Li, J.Q.; Zhou, A.A.; Chin J Pharma., 2020, 51(10), 1324-1328.
17. Shinkichi, N.; Yoshiharu, I.; Shigehiko, T.; Tatsuya, K.; Chem Pap., 2018, 73, 673-681. https://doi.org/10.1007/s11696-018-0608-2.
    CrossRef
18. Snyder, Robert.; Witz, G.; D.Goldstein, B.; Environ Health Perspect., 1993, 100, 293-306. https://doi.org/10.1289/ehp.93100293.
    CrossRef
19. Roll, E.; Norbert, F.; Hans, C.; Hans-Jiirgen, W.; Jean-Francois, R.; Richard, M.; Stephen, W.D.; Mutat Res., 1994, 322, 33-44. doi:10.1016/0165-1218(94)90030-2.
    CrossRef
20. Holder, J.W.; Toxicol Ind Health., 2008, 24, 655-675. doi:10.1177/0748233708100371.
    CrossRef
21. Frédéric, S-J.; Travis, R.; Rémy, A.; Diane E, C.; Danial, B.; Sushant, M.; Andrew, M.; Archana, K.; Gerd, O.; Andreas M, S.; Francis, G.; Org Process Res Dev., 2019, 23(5), 783-793. https://doi.org/10.1021/acs.oprd.9b00050.
    CrossRef
22. Shinkichi ,N.; Yoshiharu, I.; Shigehiko, T.; Tatsuya, K.; Chem Pap., 2019, 73, 673-681. https://doi.org/10.1007/s11696-018-0608-2.
    CrossRef
23. Hela, L.; Federico, G.; Patience, S.; Dalma, S., Can J Chem Engin., 2022, 100(11), 3123-3144. https://doi.org/10.1002/cjce.24395
    CrossRef
24. Blumberg, L.M.; Gas Chromatography (Second Edition). doi:10.1016/b978-0-12-820675-1.00026-5
    CrossRef
25. hao, K.S.; Fang, Z.L.; Peng, S.A.; Zhao, X.X.; Liu, Y.G.; Sun, B.G.; Tian, H.Y.; Liang, S. J Sulfur Chem., 2024, 45(3), 315-329. https://doi.org/10.1080/17415993.2023.2280671.
    CrossRef
26. Yang, Q.; Sheng, M.; Li, X.Y.; Tucker, C.; Suhelen, V. C.; Webb, N.J.; Whiteker, G.T.; Yu, J Org Process Res Dev.,2020, 24(6), 916-939. https://dx.doi.org/10.1021/acs.oprd.0c00159.
    CrossRef
27. Liu, H.; Tang, Q.L.; Markovich, R.J.; Rustum, A.M.; J Pharm Biomed Anal., 2011, 51, 417-421. doi:10.1016/j.jpba.2010.09.006.
    CrossRef
28. Li, M.P.; Li, R.; Wang, J; Zhang, Q.; Bai, H.; Lv,Q.; J Sep Sci., 2019, 2(1):26-37. https://doi.org/10.1002/sscp.201800125
    CrossRef
29. Sriseadka, T.; Wongpornchai, S.; Kitsawatpaiboon, P.; J Agric Food Chem., 2006, 54(21), 8183-8189. doi:10.1021/jf0614490.
    CrossRef
30. Li, M.P.; L,i R.; Wang, Z.J.; Zhang, Q.; Bai, H.; Lv, Q.; Sep Sci Plus., 2019, 2, 26-37. DOI: 10.1002/sscp.201800125.
    CrossRef
31. Wang, L.; Results Chem., 2024, 7, 101269. https://doi.org/10.1016/j.rechem.2023.101269.
    CrossRef
32. Anantha, L.V.;Sathapathy, P.V.; Duvvuri, S.; Orient J Chem., 2024, 40(1), 228-238. http://dx.doi.org/10.13005/ojc/400128.
    CrossRef