Impurity Profiling and Quantification of Sudan III Dyes by HPLC-selective UV Detection

Abstract

An analytical methodology was developed for qualitative and quantitative impurity profiling of the coloring agent Sudan III by high-performance liquid chromatography (HPLC)-diode array detection (DAD). The impurities in commercial Sudan III were characterized by comparison of their retention times and UV spectra with those of authentic standards. Four impurities regulated by International Committees in Sudan III were quantified by HPLC-selective UV detection. The impurities in Sudan dye were successfully separated on a reversed phase C18-column within 25 min and sensitively detected by UV spectrometry at two selective wavelengths. Method validation was conducted to determine linearity, precision, accuracy, and limit of quantification (LOQ). The linear dynamic range extended from 0.002 to 4.0%, with a correlation coefficient (R2) greater than 0.995. The LOQs of the impurities ranged from 8.04 to $54.29{\mu}g/mg$. Based on the established method, the levels of regulated impurities in five commercial Sudan III dyes were determined.

Introduction

Sudan III is one of lipophilic phenylazo dyes, employed extensively as an additive in drugs, cosmetics, and medical devices.1,2 The dye is prepared by the coupling of aniline with diazotized phenyl compound. However, impurities can be produced from unreacted starting materials, reaction intermediates, and/or side products during the manufacture of Sudan III.3,4 Some of the impurities in Sudan III, such as 1-phenylazo-2-naphthol (Sudan I), 2-naphthol, aniline, and 4-aminoazobenzene, can cause serious negative effects to human health and can even be carcinogenic.5-7 Because of their toxicity, the levels of these impurities in manufactured Sudan III dye are regulated by the European Union (EU) commission8 and the U.S Code of Federal Regulations (CFR).9 Lipophilic phenylazo analogues with coplanar structures can also manifest serious toxicities similar to benzo[a]pyrene.10 Thus, the characterization of impurities in coloring agents is of crucial importance for ensuring the safety of commercial Sudan III-containing products. Moreover, the impurity profile of Sudan III provides a comprehensive indicator of the manufacturing process and is diagnostic of the overall quality of the final products.

A few reports have appeared on the analysis of impurities in color additives, including 4-methyl-Sudan I in D&C red no. 6 and 7 by HPLC,11 2,4,6-tribromoaniline in D&C red no. 21 and 22 by solid-phase microextraction gas chromatography/ mass spectrometry (GC/MS)12 and 2-naphthol, azobenzene, Sudan I, and isomers of Sudan III (C.I. red 23) by GC and HPLC.13 These reports focused on analyzing a few specific impurities in coloring agents using GC, GC/MS, and HPLC. Recently, overall impurity profiling of Sudan III dye was extensively investigated by GC/MS combined with trimethylsilylation.14

In terms of a comprehensive risk assessment, detailed knowledge of all impurities that might emerge during their production is a prerequisite. However, commercial coloring agents often constitute very complex mixtures, containing several unknown minor impurities together with major compounds that are present at concentrations two or three orders of magnitude higher. HPLC methods15-18 using spectrophotometric detectors have been popularly used for the determination of phenylazo dyes due to their high absorption of visible radiation.19 For profiling of Sudan III by HPLCDAD, two specific wavelengths (230 and 470 nm) were used for overall impurity profiling and the selective detection of phenylazo compounds, respectively.

The aim of this study was to propose a methodology for comprehensive impurity profiling analysis of Sudan III using HPLC-DAD. The levels of certain impurities were further determined by HPLC with selective UV detection. A method validation was performed to determine the intra- and inter-day precision, accuracy, linearity, LOD, and LOQ. Finally, the levels of impurities in Sudan III dyes were successfully profiled by the developed HPLC-DAD method.

 

Experimental

Reagents and Chemicals. HPLC-grade methanol, acetonitrile, acetone and ethyl acetate were purchased from J.T. Baker (Phillipsburg, NJ, USA) and were used for the preparation of standard stock solutions and mobile phase. Acetonitrile and methanol were filtered through a 0.45 μm membrane filter and degassed for 5 min. Deionized water purified using a Milli-Q system (Millipore, Co., Bedford, MA, USA) was filtered through a 0.2 μm membrane filter and degassed for 5 min before use. Ammonium acetate used to make buffer solution was purchased from Sigma-Aldrich (Steinheim, Germany).

Authentic impurities of Sudan III were purchased from several chemical companies: 1‒phenylazo‒2‒naphthol (Sudan I), 4‒aminoazobenzene, 1,1‒bi‒2-naphthol, and 4-benzeneazodiphenylamine were from TCI (Tokyo, Japan), and 2‒ naphthol and aniline were from Junsei (Tokyo, Japan). Other standards including azobenzene, 4-hydroxyazobenzene, 1‒ methoxynaphthalene, 1-methoxy‒4‒methylnaphthalene, Sudan II and 2-naphthyldisulfide were purchased from Sigma-Aldrich (Steinheim, Germany). Five Sudan III samples were purchased from Dae-Jung (Siheung, Korea), TCI (Tokyo, Japan), Sigma- Aldrich (Steinheim, Germany), Kishi (Tokyo, Japan) and Wako (Osaka, Japan). Quinoline used as internal standard was obtained from Junsei (Tokyo, Japan).

Stock solutions of 4 regulatory impurities (aniline, 2-naphthol, 4-aminoazobenzene, and 1-phenylazo-2-naphthol), and quinoline were prepared in methanol at 1000 μg/mL. Standard solutions were diluted with methanol as necessary. All standards were kept dry in a refrigerator until used.

Sample Preparation. Approximately 0.5 g of Sudan III dye was dissolved in 10 mL diethyl ether; then, the impurities in the resulting solution were extracted by ultrasonication for 20 min. The ether extract was treated with 1 g of anhydrous Na2SO4 to remove trace water and then filtered through a 0.45 μm membrane filter. The ether extract (50 μL) was added to 50 μL (10 μg of quinoline) of internal standard solution to make a 0.1 mL analytical solution for HPLC analysis. Next, 10 μL of the analytical solution was injected into the HPLC system. The overall analytical process is shown in Figure 1.

Liquid Chromatography Conditions. Impurities in Sudan III dye were analyzed by an Agilent 1200 HPLC instrument (Agilent Technologies, Palo Alto, CA, USA) coupled directly to a diode array detector (DAD). The UV wavelengths were set at 230 nm to monitor for most impurities or at 470 nm to selectively monitor phenylazo-naphthol analogues. The chromatographic separation was performed at room temperature on a Luna C18 column (Phenomenex, 150 mm length × 1.0 mm i.d.) with a particle size of 5 μm. The mobile phase consisted of water (solvent A) and acetonitrile (solvent B) delivered at a flow rate of 1.0 mL/min. The gradient elution program was as follows: 40‒50% B for 4 min and 50‒55% B for 2 min followed by an isocratic hold at 55% B for 6 min, then 55‒90% B for 1 min followed by an isocratic hold at 3 min, and finally 90‒100% for 12 min. B was then returned to 40% over 1 min. The column was reequilibrated with acetonitrile for 10 min. The injection volume was 10 μL for each sample, and the total run time was 30 min.

Figure 1.Overall analytical procedure for the determination of impurities in Sudan III.

Method Validation. The HPLC method was validated by the determination of linearity, limit of detection (LOD), limit of quantification (LOQ), precision, and accuracy. Linearity of the calibration curve was tested by diluting the mixed stock solution to six working concentrations. The LODs and LOQs were evaluated at signal-to-noise (S/N) ratios of 3 and 10, respectively.

The precision of the developed methods was evaluated using intra-day and inter-day variations. The relative standard deviation (RSD) was taken as a measure of precision. Intraday and inter-day repeatability were determined on triplicates performed within one day and over three consecutive days, respectively.

To evaluate the accuracy of the analytical methods, a recovery study was performed by spiking known amounts of standards in known Sudan III (A-company) samples in triplicate at three different concentration levels. The mixture was analyzed according to above described method.

 

Results and Discussion

Analysis of Impurities in Sudan III dyes by HPLCDAD. To optimize the separation conditions for regulated impurities such as aniline, 2-naphthol, 4-aminoazobenezene, and 1-phenylazo-2-naphthol (Sudan I), as well as other authentic impurities by HPLC-DAD, the composition of mobile phase and the concentration of the ion-pair agent were examined. An acetonitrile-water mixture as mobile phase was found to provide better separation and shorter LC run-time than a methanol-water mixture on a C18 column. Several lipophilic components, such as Sudan I and other impurities in Sudan III extract, could be rapidly eluted from the C18 column using this mobile phase. As the portion of acetonitrile increased, the LC run-time decreased, but 4- aminoazobenzene in Sudan III extract could not be successfully separated from other co-extracted impurities (Figure 2(a)). Gradient elution with increasing percentage of acetonitrile was applied to successfully separate 4-aminoazobenzene from the other co-extracted components (Figure 2(b)). Thus, an acetonitrile-water mixture using gradient elution provided reasonable separation efficiency and appropriate LC run-time for four regulated impurities of interest.

The effect of buffer solution and pH variation of mobile phase on the separation of impurities in Sudan III dye was also investigated. The concentration of buffer solution and the pH of the mobile phase (pH from 4 to 9) did not significantly affect the peak shape and retention times of authentic impurities. In addition, the UV detection wavelength for the simultaneous determination of impurities was a compromise between the UV absorption maxima of these compounds and those of interfering components. The overall maximum sensitivity of impurities was achieved at 230 nm. Although the UV response of 4-aminoazobenzene was lower at 230 nm than those of other impurities, this wavelength was still sufficient for measuring the regulated concentration by the EU Commission and US-CFR (Table 1).

Figure 2.HPLC chromatograms of impurities of Sudan III extract using gradient mode (a) 15‒17 min: 80‒95% ACN and (b) 12-13 min: 55‒90% ACN. Peak identities as follows: 1. Aniline, 2. 2- napthol, 3. 4-hydroxyazobenzene, 4. 4-aminoazobenzene, 5. 1,1'- bi-2-naphthol, 6. methoxy-naphthalene, 7. methoxy-methylnaphthalene, 8. azobenzene, 9. 4-benzeneazodiphenylamine, 10. Sudan I, 11. Sudan II, 12. 2-naphthyldisulfide, 13. Sudan III, I.S. quinoline.

Based on the optimized LC conditions, 12 authentic impurities and Sudan III were successfully separated within 25 min and detected with reasonable sensitivity (Figure 3(a)) and typical Sudan III dye (A-company) was measured at UV 230 nm (Figure 3(b)) and 470 nm (Figure 3(c)), respectively.

As seen in Figure 3(b), peaks 1, 2, 4, and 10 (regulated impurities) in Sudan III dye could be successfully separated within 27 min and sample chromatograms showed a stable baseline without significant drift under the used LC conditions. Although aniline and 4-aminoazobenzene were observed as minor components, their identification and quantitation could be performed with UV spectra and enlarged LC chromatograms. In addition to authentic impurities, two naphthol analogues, two aminoazophenyl analogues, and six phenylazonaphthol analogues were observed at 230 nm. However, several peaks that were mainly phenylazo analogues and related compounds based on their UV spectra were closely eluted in the retention time range of 16 to 26 min. To overcome this problem, the profiling of phenylazo impurities in Sudan III dyes was performed at 470 nm because of their high absorption of visible radiation.18 Seven phenylazo analogues including Sudan I and Sudan III were selectively detected at 470 nm, enabling precise quantification of Sudan I without interference (Figure 3(c)). Thus, certain impurities and other phenylazo analogues in Sudan III dye could be simultaneously quantified and identified under the optimized HPLC conditions with selective UV detection.

Figure 3.HPLC chromatograms of (a) authentic standards at 230 nm, Sudan III extract (A company) at (b) 230 nm, and (c) 470 nm. Peak identities as follows: 1. Aniline, 2. 2-napthol, 3. 4-hydroxyazobenzene, 4. 4-aminoazobenzene, 5. 1,1'-bi-2-naphthol, 6. methoxy- naphthalene, 7. methoxy-methylnaphthalene, 8. azobenzene, 9. 4-benzenzeneazodiphenylamine, 10. 1-phenylazo-2-naphthol, 11. Sudan II, 12. 2-naphthyldisulfide, 13. Sudan III, I.S. quinoline, 2': naphthol analogues, 5': bi-naphthol analogues, and 13': phenylazonaphthol analogues.

Figure 4.HPLC chromatograms (at 230 nm and 470 nm) of B-E company Sudan III samples manufactured from four different chemical companies. Peak identities are as Figure 3.

The developed method was effectively applied for the determination of impurities in other commercial Sudan III dyes (Figure 4). Some of the analyte peaks in the chromatograms of Sudan III extracts were identified from the retention times and UV spectra of authentic standards. For example, 4-benzeneazodiphenylamine and 2-naphtyldisulfide were observed as relatively abundant peaks for all samples and 4 of the samples, respectively. Interestingly, Sudan II was observed for two Sudan III dyes Company B and E, while regulated impurities except for aniline were observed in all of the samples. Additional impurity peaks were also detected in LC chromatograms at 230 nm. However, the peak areas of Sudan I (peak 10) and Sudan II (peak 11) could not be precisely determined at 230 nm because of the presence of several phenylazo- or phenylazonaphthol-analogues. In this case, the determination of phenylazonaphthol analogues was effectively performed by selective detection at 470 nm. Sudan I was typically detected and precisely quantified without any significant interference in LC chromatograms at 470 nm (inset, Figure 4).

Table 1.The calibration curves, linear correlation coefficients, LODs, and LOQs of impurities obtained by HPLC

Method Validation. Calibration curves of four impurities were generated using standard solutions at concentrations between 0.002 and 4.0%. These calibration ranges were found to be adequate for the quantitation of impurities in Sudan III extract according to regulations set by the EU commission and the US CFR. The calibration curves of the analytes showed good linearity with given concentration ranges. Correlation coefficients were over 0.995, indicating excellent linearity. Line equations representing the calibration curves and their correlation coefficients are summarized in Table 1.

The limit of detection (LOD) and limit of quantification (LOQ) were calculated by statistical methods using the ratios of signal-to-noise of 3 and 10, respectively. The LODs and LOQs of four compounds are also represented in Table 1. Considering the regulatory concentrations of International Committees, the HPLC method provided reasonable sensitivities for the evaluation of regulatory impurities in Sudan III dyes.

Table 2.Intra and inter day precision and accuracy for authentic impurities spiked into Sudan III by HPLC (n = 3)

To test the accuracy and precision of both analytical methods, authentic standards were fortified at 0.08, 0.2 and 0.4% for aniline, 2-naphthol, and 4-aminoazobenzene, and 0.2, 0.4, and 1.0% for 1-phenylazo-2-naphthol in Sudan III dye (company A). Intra-day precision and accuracy were determined from the variability of triplicate analyses of spiked authentic standards within the same analytical run. Interday precision and accuracy were evaluated from the variability of triplicate analyses of spiked authentic standards for three consecutive days. The precision of the method for simultaneously determining the four compounds was acceptable because the relative standard deviation (RSD) did not exceed 10% at any given concentration for HPLC method (Table 2). At the same concentrations, the intra-day accuracy ranged from 92.00 to 108.75% for both methods, while the inter-day accuracy ranged from 94.60 to 110.00%, indicating excellent accuracy. These inter-day and intra-day data demonstrated that the developed methods were highly reproducible and precise for the four compounds of interest.

Method Application. The validated HPLC method was applied for the quantification of impurities in 5 commercial Sudan dyes (Table 3). As seen in Table 3, the amounts of 2- naphthol, 4-aminoazobenzene and 1-phenylazo-2-naphthol for the company A sample were over the regulatory limits set by the EU and the US-CFR. Moreover, the relative standard deviations (RSD) of 4-aminoazobenzene and Sudan I obtained by HPLC were relatively higher than other impurities due to the low sensitivity and the poor resolution of 4-aminoazobenzene on HPLC. The UV detection at 470 nm (inset of Figure 4) could successfully screen the phenylazo-impurities in complicate sample matrix. Thus, the established method could apply for the safety evaluation of Sudan III dyes.

Table 3.ND: not determined. aMean and relative standard deviation (%RSD) of three measurements (n = 3)

 

Conclusion

For impurity profiling, sonication using diethyl ether enabled the rapid and efficient extraction of impurities from Sudan III. LC-conditions were optimized to separate possible impurities in Sudan III dye. The HPLC-DAD method was developed and successfully applied to measure the regulated impurity levels in commercial Sudan III samples. Furthermore, the LC-selective UV detection method was useful for the discrimination and determination of phenylazo impurities from other impurities in commercial Sudan III. The established method can provide very useful information for the quality control of Sudan III dye and evaluation of its manufacture.