Thirty-six pigments have been pyrolyzed at 800 °C to create a library of specific pyrolysis decomposition patterns (Tables S1–11). The pigments investigated were chosen according to their popularity in use for tattoo inks taken from ink declarations and published surveys (BAG 2009; CVUA 2011).
To determine a suitable pyrolysis temperature, six pigments which cover the most abundant organic structures used in tattoo inks were pyrolyzed at 200, 400, 600, 800 and 1000 °C. Areas of the extracted molecular mass ions, normalized to the total chromatogram area, were found increasing and thus confirm the expected temperature-dependent formation of decomposition products (Fig. 1). Some pigments, such as pigment orange (P.O.) 13 and pigment red (P.R.) 170, decompose at rather low temperatures (<400 °C) which becomes apparent by rising peak ratios of cleavage products and color changes in sample holders (Fig. 1). In these cases, the thermal instability is caused by incorporated azo bonds which are prone to cleavage already at temperatures starting at 200 °C (Az et al. 1991). On the other hand, extremely stable pigments such as pigment violet (P.V.) 19 and pigment blue (P.B.) 15 remain more or less unaffected below 800 °C. Based on these observations, a pyrolysis temperature of 800 °C was chosen for the generation of a pyrogram library and the following tattoo ink analyses to ensure cleavage of all targeted pigments.
Occasionally it was impossible to identify all pyrolysis products using either an MS library, as provided by the US National Institute of Standards and Technology (NIST), or via judgement of the mass spectrum taking into account the pigment’s molecular structure. Since the occurrence of such as yet unknown molecule descendants is unique to certain pigments, we added these fragments as unknowns to the lists of decomposition products. Basically all pyrolysis products representing >1 % of the total peak area were included in the pyrogram library, with a few exceptions only (Tables S1–11). In the following, all different classes of organic pigments used in tattoo inks are discussed in terms of their main pyrolysis products and accompanying toxicological hazards.
1,2-Benzenedicarbonitrile and its halogenated derivatives are the most abundant products emerging upon pyrolysis of the phthalocyanines P.B.15, pigment green (P.G.) 7 and P.G.36, respectively (Table S1). Because these molecules do not occur in the pyrograms of any other pigment analyzed, they can be used for the identification of phthalocyanines. In addition, benzonitrile and phthalimide result from pyrolysis of P.B.15. Since the latter compound is used in the synthesis of the pigment, its presence likely indicates incomplete purification. Highly toxic cyanide compounds including hydrogen cyanide, cyanogen chloride and cyanogen bromide are of special concern among all pyrolysis products of phthalocyanines. Recently we have demonstrated that hydrogen cyanide is also released upon ruby laser irradiation of P.B.15 (Schreiver et al. 2015).
Copper-containing phthalocyanines are the only blue- and green-colored organic pigments present in tattoo inks so far (BAG 2009; CVUA 2011). However, P.B.15 and P.G.7 are listed in annex 1 of the cosmetics regulation in Germany whose substances are forbidden for usage in tattoo inks according to the German law (TätoV 2008). Since the comment “when used as a substance in hair dye products” was added to both pigments, their actual legal status concerning an application in tattoo inks is not always interpreted in the same way. Nonetheless, P.G.7 was listed as forbidden pigment in a governmental market survey (CVUA 2011).
For identification of azo pigments, cleavage of azo and amide bonds of specific residues attached to the core structures such as naphthols or biphenyls result in the appearance of characteristic fragments (Tables S2–6). Since similar coupling groups are used in many different pigments, identification depends on the occurrence of specific fragmentation patterns. For small pigments such as P.R.4 and P.O.5, even the unfragmented molecule ion can enter the gas phase and thus can be easily detected via py-GC/MS (Table S2). To our knowledge, P.O.34, which was found in tattoo inks lately, has not been described in pyrolysis studies before (Table S3) (BAG 2009; CVUA 2011). Similar to other diazo pigments, P.O.34 is cleaved at the azo bonds, thereby releasing the carcinogenic primary aromatic amine (pAA) 3,3′-dichlorobenzidine. Other carcinogenic pAAs such as aniline, o-anisidine or o-toluidine originate from the cleavage of coupling groups of azo and diazo pigments, too. In the past, the release of carcinogenic pAAs upon sunlight exposure and laser irradiation, along with the occurrence of allergic reactions mainly reported with red and yellow tattoos that presumably contained azo pigments, shed bad light on their usage (Cui et al. 2004; Engel et al. 2007; Gaudron et al. 2015; Hauri and Hohl 2015; Vasold et al. 2004; Wezel 2013).
Most abundant fragments of both P.O.73 and P.R.254 pigments are benzonitriles with either tert-butyl or chlorine residues depending on the respective parent compound (Table S7). In these cases, cleavage occurs inside the ketopyrrol rings as indicated. In recent years, diketopyrrolopyrroles were preferentially used by German tattoo ink manufacturers to replace the disputed azo pigments (Hauri 2011). Just as by any other organic pigment, carcinogens like benzene or, in the case of P.O.73, naphthalene are formed at temperatures of 800 °C and higher.
The pyrogram of pigment yellow (P.Y.) 138 (3,4,5,6-tetrachloro-N-[2-(4,5,6,7-tetrachloro-2,3-dihydro-1,3-dioxo-1H-inden-2-yl)-8-quinolyl]phthalimide) contained a characteristic molecule with an m/z ratio of 426 and some smaller peaks of different chlorinated benzenes. Among them, the human category 1B carcinogens hexachlorobenzene and 4,5,6,7-tetrachloro-1,3-isobenzofurandione (TIF) can be found (Table S8). The relative peak areas of TIF, the unknown 426 m/z molecule and xylene were not increasing at decay temperatures of 600 °C and higher (Fig. 1). In the case of TIF, the molecule abundance even strongly decreased at higher temperatures. It is thus likely that all three molecule species represent impurities degradable only at higher temperatures. The quinophthalone P.Y.138 is especially used in tattoo inks manufactured in Germany. To our knowledge, data on the pyrolysis of this pigment have never been presented before.
Due to their high stability (i.e., lack of weak bonds), temperatures above 600 °C are needed to induce pyrolysis of quinacridones (Fig. 1, Table S9). Cleavage occurs at the indicated sites within the pyridone rings, thereby forming benzenes with pigment-specific residues attached (Table S9). Besides that, detected fragments were due to the loss of carbons and hydrogens or resulted from the rearrangement of the parent compound leading to molecule species that could not be further characterized. Absolute peak areas of quinacridone pyrolysis products were relatively low, thus making high amounts of pigments necessary for py-GC/MS analysis. Since pigment concentrations in tattoo ink mixtures were too low quinacridone identification via py-GC/MS remained unfeasible.
The only known toxic compound evolving from pyrolysis of quinacridones is the human carcinogen benzene. On the other side, as yet unknown fragments emerging during pyrolysis leave an uncertainty in the toxicological assessment of these pigments.
Unlike chlorinated P.R.202, the pigments P.R.122 and P.V.19 are listed in annex I of the cosmetics regulation with the addition “when used as a substance in hair dye products”; thus these pigments are accounted as being prohibited for usage in tattoo inks in Germany as well (CVUA 2011). Nonetheless, all three pigments are still frequently used to create magenta to violet color shades due to their high color brilliance.
P.V.23 is mainly cleaved into the class 2 carcinogen 9H-carbazole and a highly abundant unknown product with an m/z ratio of 211 (Table S10). P.V.37 is cleaved into a variety of molecules due to the multitude of weak bonds in its structure. Among pyrolysis-induced degradation products the carcinogen benzene emerged at high levels with >8 % of the total peak area (Table S10).
In contrast to P.V.37, P.V.23 was banned from use in tattoo inks by German legislation. Nonetheless, both pigments can be found in tattoo inks sold elsewhere in Europe and thus need to be monitored and analytically separated from each other (Hauri 2011). Pyrolysis of both triphendioxazine pigments has previously been shown by Ghelardi et al. (2014).
Other polycyclic organic pigments
The replacement of azo pigments led to the introduction of a variety of chemical classes as coloring agents, namely perinone (P.O.43), anthraquinones (e.g., P.R.177), perylenes (e.g., P.R.179) and rhodamines (e.g., rhodamine B) (Table S11). Due to their compact polycyclic structure, P.O.43 and P.R.179 give rise to rather unspecific decomposition products, whereas P.R.177 can be easily identified through the occurrence of 1-amino-9,10-anthracenedione. Rhodamine B is cleaved at both its amine moiety and at the carboxyl group to yield a variety of different pyrolysis products. Benzene is the only known carcinogen formed from these polycyclic organic pigments at high pyrolysis temperatures.
P.O.43 is prohibited for use in tattoo inks, whereas P.R.177 and P.R.179 are not regulated. Rhodamine B (C.I. 45170) and its hydroxide form (45170:1) are forbidden in accordance to the German tattoo regulation (TätoV 2008). It still can be found in tattoo inks though (Hauri 2011).
Polymers and additives
Besides pigments, common polymers and thickeners used in tattoo inks have also been analyzed. Polymers decompose into their primary building blocks and other secondary pyrolysis products (Table 1). Some primary structures such as the carcinogen N-vinylpyrrolidone are of major concern. Residues of this monomer remaining in the polymer polyvinylpyrrolidone (PVP) upon synthesis or emerging during metabolism or degradation of PVP are thus to be excluded (Klimisch et al. 1997). The polymer PVP as such is non-toxic; however, when administered in large amounts and at molecular weights above 20,000, it might lead to localized cutaneous PVP storage disease (Chi et al. 2006).
Pyrolysis of silicones (polydimethylsiloxanes) resulted in the formation of cyclic dimethylsiloxanes (Table 1). Silicones, but also different dimethylcyclosiloxanes are used for suspension and as anti-foaming agents. However, py-GC/MS analytics cannot distinguish between linear and cyclic siloxanes such as D4 (octamethylcyclotetrasiloxane) and D5 (decamethylcyclopentasiloxane), respectively. While cyclic D4 siloxanes have revealed only low estrogenic activity, D5 siloxanes increased the rate of uterine tumors in animal studies (OEHHA 2008). Also, D5 is presumably interfering with the neurotransmitter dopamine and the hormone prolactin. Based on these data the use of silicones should be further evaluated in terms of tattoo regulation.
Polyethylene glycol (PEG) can be metabolized into low molecular weight oligomers and its hydroxy acid and diacid derivatives, or even toward the monomer ethylene glycol (MAK 1998). This metabolic degradation pattern of PEG could be recapitulated through py-GC/MS (Table 1). The hydrophilic metabolites of PEG will be excreted via urine but can also trigger acidosis at high concentrations due to an increased serum osmolarity and the formation of calcium complexes. Ultimately this may lead to renal and heart failure (MAK 1998; Parry and Wallach 1974).
Since styrene was found after pyrolysis of tattoo inks (data not shown), its generation has been verified via pyrolysis of polystyrene (Table 1). Pyrolysis of polystyrene indeed resulted in the formation of styrene, α-methylstyrene, styrene dimers and higher building blocks. Among these degradation products, styrene has been shown to be metabolized into styrene 7,8-oxide, an intermediate categorized as carcinogen 1B according to GHS classification that can also trigger contact allergy (Ohtsuji and Ikeda 1971; Sjöborg et al. 1894). Polystyrene can be used in pigment synthesis to facilitate particle distribution in aqueous dispersions, an application explaining its occurrence in tattoo inks (Tsubokawa et al. 1999).
Some manufacturers also use natural thickeners such as shellac. Shellac is an organic resin which only fragmented into unspecific products such as benzene, toluene and naphthalene during pyrolysis (Table 1). Interestingly, also styrene was formed upon pyrolysis of shellac, but not α-methylstyrene which only appeared in polystyrene pyrolysis (Table 1).
Pyrolysis of tattoo inks
In total, we looked into the pyrolysis-mediated degradation of 28 tattoo inks which were, according to their labeling, supposed to contain pigments that have been included in our pyrogram library (Table 2). Additionally, 18 self-made mixtures along with some rather “challenging” pigments were pyrolyzed.
Exemplarily, a pyrogram of a blue tattoo ink is displayed in Fig. 2. The product 1,2-benzenedicarbonitrile indicates usage of P.B.15 as main pigment. Also, PVP and PEG have been used, which is verified by the occurrence of pyrrolidinone, N-vinylpyrrolidone and various polyethylene glycol derivatives, respectively. Apart from these rather rare inks made of a short ingredient list, combinations of more than one organic pigment are frequently used on the market. A greater variety and higher amounts of components in the inks make pyrograms more complex and identification of ingredients only achievable for experienced personnel. Therefore, we compared two different data evaluation approaches for an easier and faster pyrogram interpretation.
In the first approach, pyrolysis products were manually compared with the fragments compiled in Tables S1–11 (see “Materials and methods” section). Depending on the pigment, 1–3 fragments were required to emerge in the respective pyrogram to sufficiently clarify the identity of the pigment. In total, 80 % of all declared pigments could be identified using this approach (Table 2). As already discussed, tattoo inks are often labeled incorrectly, thereby leaving the possibility that the analysis would be in better agreement with the true composition than the product declaration. In self-made mixtures, the pyrolysis approach did not result in false identification, yet all quinacridone pigments have been missed.
The second approach for pigment identification was a modified version of the method published by Yang et al. (2014). They used a statistical comparison of average mass spectra (AMS) of vehicle top-coatings to describe hierarchical cluster similarity with reference samples of different manufacturers. Unfortunately, this kind of data processing would be only suitable to identify a “brand” rather than single components of the inks. We therefore modified this data evaluation approach using AMS to create a mass spectral library of the 36 pigments pyrolyzed in our study. The AMS of unknown samples were then compared to the AMS library using the NIST MS 2.0 program. High abundant masses from column bleed or other column noises were excluded and masses in the range of 30–400 Da were included in the search. The highest match between a certain tattoo ink and the AMS library was taken as pigment hit. By that, we were able to identify the most abundant pigments in a few seconds in 92.9 % of all tattoo inks and around 80 % in self-made mixtures (Table 2). Hence, this method would facilitate fast and easy screenings, which then can be manually re-assessed using the evaluation approach explained above.
Identification of some pigments in commercially available inks, namely the diazo pigments P.O.13 (ink no. 9–10) and P.Y.14 (ink no. 22–23), the polycyclic P.O.73 (ink no. 12–14) and the quinacridone P.R.202 (ink no. 6 and 14) failed in either of the two evaluation approaches described above (Table 2). Since P.Y.14 and P.O.73 were successfully identified in self-prepared mixtures, their concentrations in the inks were probably too low or they were not present at all. However, the identification of quinacridones and P.O.43 also remained unsuccessful in self-suspended pigment mixtures.