Liquid chromatography–mass spectrometry analysis of dyes formed by in situ oxidative methods then purified by absorption and extraction from hair wefts

The dye mixtures formed from three commercial hair colour formers were purified by absorption onto human hair wefts, washed and dried, extracted with dichloromethane:trifluoroacetic acid (75:25) and then analysed by liquid chromatography–mass spectrometry. Only 1–2 dyes were identified from each complex mixture of commercial aromatic amines along with a broad UV absorption mainly consisting of mixtures of quaternary ammonium salts from shampoos and some surfactants. Mecetronium ethyl sulfate and didecyldimethylammonium chloride were the main ammonium salts.


Introduction
Hair colouration is a popular fashion 1 with different colours sought after. 2 Modern products can be divided into two types, oxidative [3][4][5] and nonoxidative, 6 depending upon the type of chemistry that is used. 1 Permanent or oxidative hair dyeing requires three main components (Figures 1 and 2). The first is an easily oxidised electron rich aromatic amine such as p-phenylenediamine 1 or p-aminophenol 7 which is called a primary intermediate or developer. [7][8][9] The second is called a coupler such as m-aminophenol 3, resorcinol 5 or 4-amino-2-hydroxy toluene 10. The electron rich groups are meta to each other and are not conjugated so these compounds are much less easily oxidised compared to the primary intermediate in which the electron rich groups are para to each other and are conjugated. The third component is an oxidant which is typically aqH 2 O 2 . The oxidant serves to oxidise the primary intermediates and in combination with aqNH 3 , to lighten the natural hair colour. 1 Permanent hair colour formers come in two components. The first of these contains the primary intermediates with couplers in aqNH 3 , and the second contains stabilised aqH 2 O 2 . A mixture of precursors is used, typically from four to ten. The solutions are mixed to give a suspension of pH 9.5 which is then pasted onto the hair and left for 30 min, followed by thorough washing which removes non-absorbed dye. The nature of the oxidised species formed from the primary intermediates has been considered, 6,[9][10][11][12] followed by their electrophilic aromatic substitution with a coupler leading to the dyes. The mixtures of primary intermediates and couplers used suggest that a mixture of dyes will be formed, but this does not give information about the composition of the dyes absorbed onto hair. Here, we report our studies on the liquid chromatography-mass spectrometry (LC-MS) analysis of the dyes which have been purified by absorption onto human hair wefts and then extracted with dichloromethane:trifluoroacetic acid (75:25).

Discussion
The dyes formed from three commercial hair colour formers were analysed and a control was performed using pure p-phenylenediamine 1. Commercial stabilised aqH 2 O 2 in a hair packet was used for all the oxidations. These colour formers are called Honey Blonde, Light Brown Warm and Dark Brown named after the hair colours they give. Table 1 lists the different primary intermediates and the couplers present in these colour formers. Each colour former contains from five to seven precursors. Honey Blonde, Light Brown Warm and Dark Brown each contain two primary intermediates. Clearly, a multiple number of dyes might be produced from these precursors and might be more than necessary or give a range of possible shades. These amines are toxic aromatic amines which some people can have an allergic reaction too and this is an interesting pathway for the industry to have developed. This project provides an understanding of which dyes are selectively absorbing onto the hair wefts so might help influence or simplify the design of the colour former packs.

LC-MS analysis of the dyes extracted from hair wefts dyed from Honey Blonde, Light Brown Warm and Dark Brown commercial colour formers and a control
The experimental method was much the same as that used with the commercial hair colour former packet and is described in the experimental section. p-Phenylenediamine 1 or mixtures of commercial amines (Table 1) were mixed with the aqH 2 O 2 in a beaker and the hair was mixed in. After washing the hair and drying, it was extracted with dichloromethane:trifluoroacetic acid (75:25) which was allowed to evaporate. The extract was then dissolved in MeOH and prepared in a sample vial. Figure 3 shows the LC-MS chart for the control, and Figures 4-6 show the LC-MS charts for the Honey Blonde, Light Brown Warm and Dark Brown commercial colour former packets, respectively. The LC-MS charts show the UV absorption at 230 nm in black and the horizontal axis is the retention time. The vertical axis has two scales on it. The UV axis is in mAU, and the mass spectral scale is in electro spray mass spectrometer (ES-MS) counts. All the absorption peaks are presented as an accurate mass in the mass spectral data. The neutral dyes all appear as (M + H). Figure 3 shows the LC-MS analysis of the hair extract formed from the in situ oxidation of p-phenylenediamine 1 with aqH 2 O 2 . A strong absorption occurs at m/z 319.1666 (M + H) (calculated m/z 319.1666). This is assigned as compound 2 5,8 (Figures 1  and 7). This control experiment is useful as it verifies that the method of oxidation, absorption and extraction is working. In the retention time range 2-12 min, a broad UV absorption is present with no defined or separated absorption peaks. Luckily, the desired dyes elute faster than this complex material and are free from it. This material is analysed shortly. Figure 4 shows the LC-MS analysis of the hair extract formed from the in situ oxidation of the commercial colour former Honey Blonde with aqH 2 O 2 . A strong absorption occurs at m/z 321.1347 (M + H) (calculated m/z 321.1346). This is assigned as the known compound 6 5,13,14 (Figure 7).  Figure 2. The primary intermediates and couplers present in some different hair colour formers 1, 3, 5 and 7-13.   These are assigned as the known compounds 6 and 4, respectively, (Figure 7).
The dyes observed were weaker in intensity than expected and were swamped out by much stronger chromophores, hence the need to analyse the data in stages. The number of dyes observed, three in total, was much less than expected and many more must surely form. These are the known compounds 2, 5       efficiently and bind strongly to the hair but they can also be extracted. The extracted hair partially lightens in colour but does not lose all its colour so some compounds will still be bound in the hair. The experiment is easy to perform because the absorption onto hair is a method of purifying the dye mixture leaving behind impurities. It would not be possible to analyse the crude mixture of dyes by LC-MS without a preliminary purification so absorption of the dyes onto hair is meaningful and it provides information about which dyes are binding to the hair. The long hair wefts are easy to work with, to wash and to remove excess dye. The method of extraction of the dyes from the hair using dichloromethane:trifluoroacetic acid (75:25) is new and also convenient as it is non-aqueous or non-nucleophilic so it will not degrade the hair or the dyes and it is also volatile so evaporates easily after the extraction has been performed. This keeps the sample treatment as simple as possible. After extraction of the hair, the solvent was evaporated at room temperature in the fume hood air flow until dry and the residue was taken up in MeOH ready for analysis in a sample vial.   Table 1S, compounds 1S-6S). Table 3, compounds 16-19, shows decomposition products believed to come from the quaternary ammonium salts. Table 4, compounds 20-22, shows phthalates and Table 5, compounds 23-26, shows miscellaneous compounds. In summary, the broad UV absorption of the LC-MS charts is dominated by quaternary ammonium salts 14 and 15 and many other compounds. These have not come from the hair colour formers according to the packet ingredients so must be already absorbed on the long human hair weft from washing with shampoos where they are included. The absorption peaks are not scaled to ensure fine detail is seen. This work shows that quaternary ammonium salts have a high affinity for hair wefts and the number of them suggests that they absorb permanently over many years from different shampoos. Called 'quats' for short, quaternary ammonium compounds are used commonly in hair conditioners, shampoos and lotions to impart a slippery feel to the hair and skin. 16 They allow a little dollop of conditioner to easily glide and be distributed throughout the hair. Different quaternary ammonium salts and surfactants are added to shampoos and it is possible they accumulated over a 10-year period or longer as the hair weft is 15 inches long. Some phthalates are also used in shampoos as are some of the miscellaneous compounds. The dominant quaternary ammonium salt is mecetronium ethyl sulfate 14 which has intense peaks in                         The strong affinity of quaternary ammonium salts for hair can be exploited for making new colour extenders. 18,19 The study here verifies that the method has precedent and that a wide variety of quaternary ammonium salts and dyes might be exploited by covalent attachment together. The oxidative in situ formation of dyes, a commercial method which dominates supermarket shelves, remains a method  with some toxicity owing to the toxicity of aromatic amines and the many by-products that form. Some people experience an unpleasant inflammatory reaction, so alternative cleaner and less toxic methods of fixing dyes to hair are desirable.
The common decomposition products 16-19 in Table 3, of the quaternary ammonium salts, were also unexpected. Decomposition of the quaternary ammonium salts requires a dealkylation and loss of a methyl or ethyl group or possibly an aerial oxidation. Either mechanism appears to be slow or the products still bind to the hair in lesser amounts.
Phthalates were also identified as some of the components which can also be used in shampoos. Isononylphthalate 20 (Figures 10, 13

Conclusion
The dyes formed from three commercial hair colour formers have been purified by selective absorption on long human hair wefts. After washing and drying, the dyes were extracted with a non-nucleophilic and volatile solvent mixture of dichloromethane and trifluoroacetic acid. After ambient evaporation of the extractant, the extract was taken up in MeOH with no further treatment and analysed by LC-MS. Only one or two dyes, compounds 2, 4 or 6, were identified despite the more complex mixture of aromatic amines present in the starting mixtures (Table 1) which were oxidised with the aqH 2 O 2 supplied. This suggests that the complex mixture of toxic aromatic amines 20 in cosmetic hair dye which is oxidised might be simplified with the knowledge of which dyes bind efficiently to the hair wefts. A broad overlapping UV absorption was also present which was analysed in stages because of its complexity. The broad UV absorption consisted of a number of quaternary ammonium salts, some amines formed by their decomposition, and some phthalates and miscellaneous surfactants. Health and safety data and further information on these compounds are easily found by searching the molecular formular online. Long Remy human hair wefts must absorb quaternary ammonium salts from shampoos over a number of years building up a mixture of them. Ammonium salts and other compounds bind tightly to hair wefts from shampoos and might be a basis for the development of novel colour extenders 18,19 like anthraquinones which can disperse and bind tightly on hair wefts and polyacrylonitrile (PAN) fibres. [21][22][23][24][25][26][27][28][29][30][31][32]

Methods
For analytical separation, an Agilent 1290 Infinity high-performance liquid chromatography (HPLC) system consisting of a quaternary HPLC pump, cooled auto sampler compartment, column compartment and diode array UV-Vis detector was used. An F5-Kintex C-18 column (2.1 mm × 150 mm, Thermo Scientific, UK) was used for the separation with an acetonitrile/methanol gradient (both 0.1% v/v formic acid) from 40% acetonitrile to 100% in 10 min with 2 min hold and then reset (gradient 1); from 5% acetonitrile to 100% in 10 min with 2 min hold and then reset (gradient 2). Data are shown in the Supplemental section (Supplemental Figures  S18-S36). The flow rate was 0.5 mL min −1 , the column temperature was 40 °C, and the sample volume was 5 μL. The mass spectrometer (ES-MS) used was a MaXis II UHR-TOF LC-MS System (Bruker UK Ltd, UK) with electrospray ionisation (ESI) source connected to the UV-Vis detector by a short length of PEEK tubing. The ES-MS was operated in positive ion mode with a capillary voltage of 4.5 kV using sodium formate clusters for calibration and methyl stearate as lock mass. Mass spectra were recorded automatically. Long blonde human hair wefts (15 inches) from China were purchased from Remy and were used for dyeing in accordance with the relevant guidelines and regulations of the University of Aberdeen Health and Safety committee.

LC-MS analysis of the dyes extracted from hair wefts dyed from Honey Blonde, Light Brown Warm and Dark Brown commercial colour former packets and a control
The experimental method was much the same as that used with the commercial hair colour former packet. An equal volume of aqH 2 O 2 (5 mL), provided, was mixed with an equal volume of the mixture of primary intermediates and couplers (5 mL) in a beaker. For the control, pure p-phenylenediamine 1 was oxidised. Tied long extensions from a blonde Remy human hair weft (1 cm in width) were pasted in the beaker onto the hair with a spatula and left for 30 min. The hair was then rinsed under a water tap and left to dry. It was then immersed in a solvent mixture of dichloromethane:trifluoroacetic acid (75:25) for 5 min in a fume hood with the fan off. If the mixture of hair and solvent goes dry, then more of the solvent mixture was added. The extractant becomes darker but not excessively. The hair was removed from the extractant and left to dry in the fume hood. The beaker was allowed to slowly evaporate without any treatment and then mixed with a small amount of MeOH. This MeOH was placed in a small vial ready for LC-MS analysis (four vials in total). Two different gradients G1 and G2 of MeOH/CH 3 CN were used for the analyses (see section 'Methods').