Analysis of 13 Banned Colorants in Cosmetics via Liquid Chromatographic and Mass Spectrometric Techniques

: Even though synthetic colorants can cause side effects such as allergies and pigmentation, they have not been sufﬁciently researched. Herein, high-performance liquid chromatography, liquid chromatography-tandem mass spectrometry, and liquid chromatography-quadrupole time-of-ﬂight mass spectrometry (LC-Q-TOF-MS) were used to detect 13 banned synthetic colorants in cosmetics and characterize their fragmentation. The developed HPLC method was validated following the International Conference on Harmonisation guidelines (speciﬁcity, limit of detection, limit of quantiﬁ-cation, recovery, linearity, accuracy, and precision) and applied to 120 distributed cosmetic products, one of which was found to contain three illegal synthetic colorants, namely Basic Blue 26 (0.33 mg/g), Basic Red 2 (0.53 mg/g), and Basic Yellow 28 (31.50 mg/g). Additionally, based on their fragment ions obtained using LC-Q-TOF-MS, the fragmentation pattern of synthetic colorants was predicted. Thus, our work paves the way for the reliable detection of illegal synthetic colorants and may help to prevent the distribution of cosmetics containing the same.


Introduction
The reliance of image and personality shaping on physical attractiveness creates a high demand for makeup cosmetics, thus driving the diversification of cosmetics consumers and the growth of the domestic cosmetics market [1][2][3][4][5][6].In particular, the increased demand for high-persistence cosmetics (i.e., those that are not smeared upon contact with masks) due to the COVID-19 pandemic has inspired the development of new color cosmetics [3,7].As their name suggests, color cosmetics (e.g., lipsticks, eye shadows, eyeliners, and blushers) contain natural or synthetic colorants and are used to conceal skin defects and create a more attractive appearance [8][9][10].Compared to natural colorants, synthetic ones exhibit the advantages of long-lasting effect, lower price, and higher stability to light, oxygen, and pH, and are therefore typically the preferred choice [10][11][12][13][14][15][16].However, most synthetic colorants contain azo functional groups or aromatic rings and can therefore cause allergies, asthma, DNA damage, and even cancer or mutagenesis [11,12,[14][15][16][17][18][19][20][21][22].Therefore, the types and standards of colorants allowed in cosmetics are closely regulated.
The requirements to be met by cosmetics distributed in the European Union are set forth in the Cosmetics Regulation (EC) No. 1223/2009.Annex IV of this regulation allows 153 colorants, prohibits four from use in products applied to eyes and 19 from use in products applied to mucous membranes, and allows 36 to be used only in rinse-off products.Annex II lists the ingredients prohibited from use in cosmetics, while Annex III lists the restricted ingredients.In the United States (US), color additives are regulated by the Food and Drug Administration (FDA; e.g., Federal Food, Drug, and Cosmetic Act, Sec.721; 21 U.S.C.379e) and must be approved by the same before use in cosmetics.Most color additives must obtain batch certification from the FDA when used in US-marketed cosmetics.The FDA has classified 65 types of color additives as "permissible".Their availability depends on the eye area and general (including lipstick) and external use.The FDA has also set the usage limit for some color additives.
Although color cosmetics directly contact the skin and must therefore meet all safety requirements, many cosmetics with unusable and unsafe pigments capable of causing side effects such as pigmentation have previously been marketed, which highlights the need for appropriate regulation.In the Republic of Korea, cosmetics manufacturers are required to list all the ingredients of a given product.The Ministry of Food and Drug Safety has listed the colorants allowed for use in color cosmetics as well as the related standards and analytical methods.In particular, the Cosmetics Act lists 129 pigments that can be used in cosmetics.However, the use of these pigments is subject to restrictions depending on the type of cosmetics and/or the targeted skin area.In addition, standards for ingredients that cannot be used in cosmetics and those that can be used only under certain restrictions are also designated and managed through timely notices.The above act also prohibits approximately 90 carcinogenic or harmful colorants from use in color cosmetics.
The growing interest in long-lasting cosmetics such as tattoo lipsticks, tattoo eyebrows, and hair tints has increased the incidence of the related side effects.Currently, regulatory authorities manage products by issuing various guidelines; however, no effective analysis methods for illegal colorants in cosmetics have been established.This study contributes to the safety of color cosmetics by developing a method for detecting prohibited colorants therein (Figure 1).Unlike previous investigations, the present work uses HPLC coupled to a diode array detector (HPLC-DAD) and UPLC-MS/MS to develop an analytical method for analyzing illegal colorants and fragmentation of pigments using liquid chromatographyquadrupole time-of-flight mass spectrometry (LC-Q-TOF-MS).

Figure 1.
Experimental process for analysis of 13 illegal synthetic colorants.

Standard Stock Solutions
Stock solutions of Basic Blue 26, Basic Red 2, Disperse Brown 1, Disperse Orange 3, Disperse Yellow 3, HC Blue No. 2, HC Yellow No. 5, Solvent Orange 4, Solvent Yellow 1, and Solvent Yellow 3 were prepared in 100% MeOH.Stock solutions of Solvent Orange 7 and Solvent Red 24 were prepared in 100% CHCl3.Basic Yellow 28 was dissolved in 100% water.All stock solutions were stored in a refrigerator (2-8 °C).Working standard solutions for HPLC-DAD, LC-MS/MS, and LC-Q-TOF-MS analyses were prepared by diluting stock solutions with 100% MeOH.

Sample Preparation
The samples were purchased through an online store and categorized into tattoo lipstick, tattoo eyebrow, and hair tint.Each sample (100 mg) was placed in a 50 mL corning tube, and the volume was adjusted to 50 mL using 100% MeOH.Ultrasonication-assisted

Standard Stock Solutions
Stock solutions of Basic Blue 26, Basic Red 2, Disperse Brown 1, Disperse Orange 3, Disperse Yellow 3, HC Blue No. 2, HC Yellow No. 5, Solvent Orange 4, Solvent Yellow 1, and Solvent Yellow 3 were prepared in 100% MeOH.Stock solutions of Solvent Orange 7 and Solvent Red 24 were prepared in 100% CHCl 3 .Basic Yellow 28 was dissolved in 100% water.All stock solutions were stored in a refrigerator (2-8 • C).Working standard solutions for HPLC-DAD, LC-MS/MS, and LC-Q-TOF-MS analyses were prepared by diluting stock solutions with 100% MeOH.

Sample Preparation
The samples were purchased through an online store and categorized into tattoo lipstick, tattoo eyebrow, and hair tint.Each sample (100 mg) was placed in a 50 mL corning tube, and the volume was adjusted to 50 mL using 100% MeOH.Ultrasonicationassisted extraction was performed at room temperature for 30 min and was followed by centrifugation at 3100× g for 5 min.Then, the supernatant was filtered with a 0.2 µm PTFE syringe filter and used for HPLC-DAD, LC-MS/MS, and LC-Q-TOF-MS analyses.

Method Validation
The HPLC method was validated for specificity, limit of detection (LOD), limit of quantification (LOQ), recovery, linearity, precision, and accuracy according to the guidelines of the International Conference on Harmonisation (ICH) of Technical Requirements for Pharmaceuticals for Human Use [31].Specificity was confirmed by spiking the blank matrix sample (tattoo lipstick, tattoo eyebrow, and hair tint) with the 13 illegal colorants.The LOD and LOQ were evaluated for the spiked samples using signal-to-noise ratios of 3:1 and 9:1, respectively.The concentration range of linearity was set to six points, considering the analyte LOQ.Calibration curves were obtained in triplicate at each concentration.Recovery was determined by comparing the responses of the colorants in the spiked samples and standard solutions.For precision and accuracy evaluation, the experiment was conducted thrice on the same day (intra-day) and thrice over three days (inter-day) at the three concentrations.The recovery, precision, and accuracy were measured at low, medium, and high concentrations.

Results and Discussion
The detection wavelengths for the simultaneous analysis of the 13 prohibited synthetic colorants by HPLC-DAD were set in consideration of the maximum absorption wavelength (λ max ) for each compound.As shown in Figure 2d V, gas temperature = 300 °C, nebulizer = 45 psi, drying gas (N2) = 10 L/min, fragmentor = 175 V, and skimmer = 65 V.

Results and Discussion
The detection wavelengths for the simultaneous analysis of the 13 prohibited synthetic colorants by HPLC-DAD were set in consideration of the maximum absorption wavelength (λmax) for each compound.As shown in Figure 2d  The HPLC-DAD method was validated in terms of specificity, LOD, LOQ, linearity, recovery, accuracy, and precision.The specificity was confirmed as peaks of interfering compounds were not observed at the retention times of the colorants in all blank samples.The LODs and LOQs of colorants in tattoo eyebrow samples were 0.10-1.20 µg/mL and 0.30-3.60µg/mL, respectively, while those for tattoo lipstick and hair tint samples were 0.01-0.30µg/mL and 0.02-0.90µg/mL, respectively (Table 1).The calibration curves for all prohibited colorants showed excellent correlation coefficients (R 2 > 0.999, Table 1).The recovery was determined at three concentrations (low, medium, and high) based on the LOQ concentration of each compound.Low, medium, and high concentrations were set to the LOQ, 10 times the LOQ, and 50 times the LOQ, respectively.The recoveries of the tattoo lipstick, tattoo eyebrow, and hair tint samples were 86.11-108.39%,83.32-105.33%,and 93.49-106.48%,respectively, and the corresponding relative standard deviations (RSDs) were <4.66%, 3.58%, and 9.63%, respectively (Tables 2-4).The recoveries and RSDs were within the appropriate ranges.Accuracy was evaluated as the average of recovery, and precision was evaluated as RSD% (Tables 2-4), with satisfactory results obtained in all cases.For tattoo lipstick samples, the intra-and inter-day accuracies were 85.97-106.50% and 89.22-108.28%,respectively, while the intra-and inter-day precisions were 0.53-8.96%and 0.04-6.96%,respectively.For tattoo eyebrow samples, the intra-and inter-day accuracies were 80.23-103.90% and 80.70-103.74%,respectively, while the intra-and inter-day precisions were 0.14-9.68%and 0.02-7.64%,respectively.For hair tint samples, the intra-and inter-day accuracies were 81.20-107.44%and 86.34-104.02%,respectively, while the intraand inter-day precisions were 0.08-9.99%and 0.45-9.80%,respectively.In addition, to check the illegal colorant ions, multiple reaction monitoring (MRM) conditions for each compound were confirmed using LC-MS/MS.Each standard solution was directly infused to find the precursor ion.Then, the product ion was obtained by adjusting the collision energy.Two or more transitions were selected for each analyte.The optimized MRM transitions are listed in Table 5. Disperse Yellow 3 and HC Yellow No. 5 were detected in the negative-ion mode, while the other 11 compounds were detected in the positive-ion mode.The 120 purchased samples were classified into tattoo lipsticks (40), tattoo eyebrows (41), and hair tints (39).Figure 3 shows the results of illegal pigment detection by HPLC-DAD, revealing that three peaks with retention times and UV spectra matching those of Basic Red 2, Basic Yellow 28, and Basic Blue 26 were detected in a tattoo eyebrow product.The contents of these three components (HPLC-DAD) were determined as 0.53 mg/g (Basic Red 2), 31.50 mg/g (Basic Yellow 28), and 0.33 mg/g (Basic Blue 26).The identities of these three compounds were confirmed by positive mode using LC-MS/MS.The total ion chromatogram (Figure 3) of the tattoo eyebrow sample showed peaks at 5.13, 5.97, and 7.62 min corresponding to the peaks for Basic Red 2, Basic Yellow 28, and Basic Blue 26 standards, respectively.The precursor and product ions of these three colorants were the same as those of the respective standards.Basic Red 2 showed a precursor ion at m/z 315.32 and product ions at m/z 299.30, 210.24, and 237.26.Basic Yellow 28 showed a precursor ion at m/z 322.34 and product ions at m/z 160.24 and 136.18.Basic Blue 26 showed a precursor ion at m/z 470.42 and product ions at m/z 349.30, 454.38, and 333.28.Based on the above, unknown peaks 1, 2, 3 were tentatively identified as Basic Red 2, Basic Yellow 28, and Basic Blue 26, respectively.Hence, the developed method was found to be suitable for the detection and quantitation of illegal synthetic colorants added to cosmetics.In the case of Basic Red 2, several peaks ascribed to impurities were observed and marked as such in the chromatograms of Figure 3 [32].
The MS/MS fragmentation patterns of the three detected compounds (Basic Blue 26, Basic Red 2, and Basic Yellow 28) were examined via LC-Q-TOF-MS at different collision energies.The optimized MS/MS spectra and the related ion structures are presented in Figures 4-6.Basic Blue 26 exhibited a precursor ion at m/z 470.26 (C 33 H 32 N 3 + ), and the difference from the theoretical value was estimated to be within 0.64 ppm.Fragment ions were observed at m/z 454.23 ) and fragment ions at m/z 166.06 and 137.04.To the best of our knowledge, the fragmentation patterns of the 11 out of the 13 examined dyes have not been reported before, i.e., only the fragmentations of   To the best of our knowledge, the fragmentation patterns of the 11 out of the 13 examined dyes have not been reported before, i.e., only the fragmentations of Solvent Orange 7 and Solvent Red 24 have been described [33].The precursor ions of Solvent Orange 7 and Solvent Red 24 were observed at m/z 277.13 (C18H17N2O + ) and 381.17 (C18H17N2O + ), respectively, as expected for these analytes.The fragment ions of Solvent Orange 7 were observed at m/z 121.09, 153.04, and 260.13, while those of Solvent Red 24 were observed at m/z 91.05, 106.07, 156.04, 224.12, and 276.11.

Figure 1 .
Figure 1.Experimental process for analysis of 13 illegal synthetic colorants.
, three wavelengths were selected to effectively separate the compounds: 254 nm for Basic Red 2, Disperse Yellow 3, HC Blue No. 2, Solvent Orange 7, and Solvent Red 24; 430 nm for Basic Yellow 28, Disperse Brown 1, Disperse Orange 3, HC Yellow No. 5, Solvent Orange 4, Solvent Yellow 1, and Solvent Yellow 3; and 590 nm for Basic Blue 26.The ultraviolet (UV) spectra are shown in Figure 2d.Appl.Sci.2023, 13, x FOR PEER REVIEW 5 of 17 , three wavelengths were selected to effectively separate the compounds: 254 nm for Basic Red 2, Disperse Yellow 3, HC Blue No. 2, Solvent Orange 7, and Solvent Red 24; 430 nm for Basic Yellow 28, Disperse Brown 1, Disperse Orange 3, HC Yellow No. 5, Solvent Orange 4, Solvent Yellow 1, and Solvent Yellow 3; and 590 nm for Basic Blue 26.The ultraviolet (UV) spectra are shown in Figure 2d.

Figure 3 .
Figure 3. Chromatograms for 3 illegal synthetic colorants in standards and tattoo eyebrow samp (a) HPLC chromatograms and UV spectra and (b) total ion chromatograms (TICs) and extracted io chromatograms by LC-MS/MS (the peaks (*) are observed due to impurities).

Figure 3 .
Figure 3. Chromatograms for 3 illegal synthetic colorants in standards and tattoo eyebrow sample: (a) HPLC chromatograms and UV spectra and (b) total ion chromatograms (TICs) and extracted ion chromatograms by LC-MS/MS (the peaks (*) are observed due to impurities).

Figure 4 .
Figure 4. Optimized MS/MS spectrum (a) and proposed fragmentation pattern of Basic Blue 26 (b) by LC-Q-TOF-MS.

Figure 4 .
Figure 4. Optimized MS/MS spectrum (a) and proposed fragmentation pattern of Basic Blue 26 (b) by LC-Q-TOF-MS.

Figure 5 .
Figure 5. Optimized MS/MS spectrum (a) and proposed fragmentation pattern of Basic Red 2 (b) by LC-Q-TOF-MS.

Figure 5 .
Figure 5. Optimized MS/MS spectrum (a) and proposed fragmentation pattern of Basic Red 2 (b) by LC-Q-TOF-MS.

Table 1 .
LOD, LOQ, and linearity of colorants in cosmetic samples using HPLC-DAD.
LOD, limit of detection; LOQ, limit of quantification.

Table 2 .
Recovery, accuracy, and precision of colorants in tattoo lipstick sample with different concentrations using HPLC-DAD.

Table 3 .
Recovery, accuracy, and precision of colorants in tattoo eyebrow sample with different concentrations using HPLC-DAD.

Table 4 .
Recovery, accuracy, and precision of colorants in hair tint sample with different concentrations using HPLC-DAD.
16333.14, and 349.17.Basic Red 2 showed a precursor ion at m/z 315.16(C 20 H 19 N 4 + ), and the difference from the theoretical value was estimated to be within 1.90 ppm.Fragment ions at m/z 237.11, 210.10, and 299.13 were observed.Basic Yellow 28 showed a precursor ion at m/z 322.19 (C 20 H 24 N 3 O + ), and the difference from the theoretical value was estimated to be within 1.24 ppm.Fragment ions were observed at m/z 160.11 and 136.08.Figures S1-S10 present the fragmentation patterns of 10 colorants other than those detected in cosmetics samples.Disperse Brown 1 exhibited a precursor ion at m/z 433.02 (C 16 H 16 Cl 3 N 4 O 4 + ) and fragment ions at m/z 356.97, 197.05, and 153.02.Solvent Yellow 3 exhibited a precursor ion at m/z 226.13 (C 14 H 16 N 3 + ) and fragment ions at m/z 91.05, 106.07, and 121.08.Disperse Yellow 3 presented a precursor ion at m/z 268.11 (C 15 H 14 N 3 O 2 − ) and fragment ions at m/z 134.06 and 92.05.HC Yellow No. 5 exhibited a precursor ion at m/z 196.07 (C 8 H 10 N 3 O 3 −