ANALYST, NOVEMBER 1988, VOL. 113 1613 Analysis of Chrysoidine Dyes using High-performance Thin-layer Chromatography and a Proposed Method for the Measurement of Chrysoidine Y and R in Air Samples Robert D. Foster and John A. Groves Health and Safety Executive, Occupational Medicine and Hygiene Laboratories, 403 Edgware Road, London N W2 61 N, UK A method for the analysis of chrysoidine dyes using high-performance thin-layer chromatography is described. The composition of chrysoidine dyes manufactured in the UK is examined and proposals are made for a sampling and analysis scheme to measure the concentrations of the dyes in air. A modification of the technique for screening of the dyes to determine whether o-toluidine was used in their manufacture is also described. Keywords: Chrysoidine; dyes; high-performance thin-layer chromatography; air sampling; o-toluidine Chrysoidine dyes are used predominantly in the production of brown paper but find several other minor uses, for instance, in the manufacture of wood stains, shoe polish, printing inks, carbon paper and stains for use in microscopy.The dyes came to the attention of the public in the UK as a result of their use as a colorant for maggots intended as bait for fishing. It was suggested that it was this that was responsible for the increased occurrence among anglers of tumours of the urinary bladder.1 An initial limited study of the incidence of bladder cancer in West Yorkshire was able to detect “no major risk to fishermen,” but the confidence limits of the study were wide.2 However, a more recent study in the West Midlands, which included both bladder cancer and upper tract urothelial tumours, reported a correlation with angling and particularly with the use of bronze-dyed maggots for more than 5 years.3 Even so, considering this finding, the DHSS Committee on Carcinogenicity of Chemicals in Food, Consumer Products and the Environment observed that the increased risk for urothelial cancer in anglers only just reached the level of statistical significance and that this might be a result of the greater tobacco usage among anglers or certain sub-groups of anglers .4 The reason for regarding chrysoidines as suspect human carcinogens rests on their chemical similarity to known animal carcinogens such as o-aminoazotoluene and 2,4-diaminotolu- ene and on bacterial experiments in which they demonstrate activity as mutagens .5-7 The evidence for carcinogenicity in animals rests on one published study on mice8 in which it is unclear whether or not Chrysoidine Y was used.9 Subsequent attempts to reproduce the results with another batch of chrysoidine and with synthesised chrysoidine have apparently failed.10 Some azo dyes are potentially carcinogenic because metabolic reduction in the body, either of water-soluble forms by bacteria in the gut or of oil-soluble forms by enzymes secreted by the liver, can produce aromatic amines.l* Some of these amines show marked mutagenicity. Ames tests with Salmonella bacteria indicated little activity for Chrysoidine Y alone, but a marked mutagenicity in the presence of liver enzyme extract.5 The amine believed to be produced by enzyme reduction and to be involved is 1,2,4-triaminoben- zene.This compound alone shows no special activity in the Ames test, but in the presence of liver enzymes it exhibits a mutagenicity similar to that observed for Chrysoidine Y .5 It has been assumed that N-hydroxy compounds generated by further enzyme-induced reactions are the eventual mutagenic agents. Perhaps significantly, mutagenic metabolic activation Crown Copyright of aromatic amines has been demonstrated for enzymes extracted from the bladder urothelium cells of cows.12 Most concern has been expressed about Chrysoidine Y as this used to be the dye recommended for the “bronzing,’ of maggots. In practice, however, Chrysuidine R and mixtures containing similar methylated chrysoidines such as Methic Orange have also been used.The use of these dyes by anglers is now discouraged; ICI, one of the major manufacturers, has advised their distributors to stop the sale of all dyes to the maggot trade.13 One Gurr Chrysoidine supply, which was demonstrated by NMR to consist of methylated species, has been shown to be a more potent mutagen (by 10-fold) of Salmonella typhimurium than Chrysoidine Y in the reductive presence of liver enzymes.7 Analysis of a sample from the same source, discussed in this paper, indicates that the dye is a mixture of chrysoidine species derived from diaminotoluene reacted with a diazotised aniline and toluidine mixture. That the known animal carcinogen o-toluidine can be released from the dye by reduction with tin(I1) chloride, as reported in this paper, might be thought significant in explaining the mutagen- icity indicated by the Ames test results.However, if so, the mechanism requires some explanation because o-toluidine has not been found to be mutagenic in Salmonella typhimurium in the presence of the enzymes of a rat liver alone (as is the Gurr Chrysoidine Y), but only in the presence of both norharman and a rat liver activation system prepared from rats induced with polychlorinated biphenyls. 14 Also, the possible contribu- tion of other methylated chrysoidine species should not be ignored. The amines potentially produced by metabolisation of those chrysoidines synthesised from diaminotoluene, e.g. , Chrysoidine R, are triaminobenzenes with a methyl group ortho to at least one amine group.Ortho-substituted aromatic amines have been noted as being more likely to cause bladder cancer in rats.15 Many, but not all, mutagens are carcinogens and vice versa; the exact level of correlation is still under investigation.14.16 It is not possible to dquate the mutagenicity of a material, measured by experiments with bacteria, with its carcinogenic- ity in man, This is especially true when animal experiments are inconclusive, as in the example of Chrysoidine Y. Nor can it be assumed that dyes are metabolised by man in the same way as by bacteria or animals. Nevertheless, mutagenic aromatic amines, metabolic products of other dyes, have been found in the urine of workers; for instance benzidine in those using benzidine-based dyes.17 Benzidine- , o-tolidine- and o-dianis- idine-based dyes are the subject of a Health Hazard Alert issued by the US Occupational Safety and Health Administra- tion.18 One study of paper workers using the dye Direct Black 38 revealed Chrysoidine Y in urine samples.17 However, it was1614 ANALYST, NOVEMBER 1988, VOL. 113 unclear whether this was a metabolite or a by-product of the synthesis of the dye and unfortunately no special mention is made of those aromatic amines that might have been produced by further metabolisation. A study of occupational bladder cancer in West Yorkshirelg failed to discover any special prevalence of bladder cancer among dye users. However, workers in the dye-manufacturing industry were shown to be significantly more at risk, particu- larly if they were smokers.An epidemiological survey carried out by ICI (Organics Division) covering 33 years of manufacture reported that “no cases of liver or bladder cancer can be attributed to exposure to chrysoidines. ”20 Although this is reassuring, and evidence linking chrysoidines with cancer is disputed, there is still sufficient reason for concern to ensure that chrysoidine dyes should be handled with care in the workplace. Adherence to a code of practice such as that givenin the general recommenda- tions of the HSE publication “Safe Handling of Dyestuffs in Colour Stores”21 is advisable. It is desirable to monitor airborne dust levels in order to minimise them. It would also be instructive to be able to ascertain the compositional type of the chrysoidine in use.The analytical method described here was developed in order to measure both Chrysoidine Y and Chrysoidine R levels in chrysoidine dye mixtures used in industry and to be able to monitor the concentrations of these compounds in the air of the working environment. A modification of the method can be used to determine if chrysoidines based on o-toluidine are present in the dyestuffs. Nomenclature The nomenclature used for the chrysoidine dyes in this paper is given in Table 1. Their corresponding chemical structures are shown in Table 1 and in Fig. 1. Chrysoidine R2,6 is the authors’ designation of dye III, i.e., Chrysoidine R derived from 2,6-diaminotoluene. The nomenclature used for these dyes in the dye industry is particularly confusing and requires some elucidation.Chrys- oidine is a generic name applied to a number of chemically related orange azo dyes derivatised from aromatic diamines. I HzN CH3 e N = N - & NH2.HCI 111 CH3 CH3 IV HPN CH3 Fig. 1. Structures of the ch soidine d e components. (I) Chrysoi- dine Y, CJ Basic Orange 2, ‘EI 11270; &I) Chrysoidine R, CI Basic Orange 1, CI 11320; (111) Chrysoidine R2,6, CI 11320 variant derived from 2,6-TDA; (IV) CI 11325 derived from 2,4-TDA; and (V) CI 11325 variant derived from 2,6-TDA Table 1. Nomenclature of chrysoidine dyes. Methic Orange YN and Methylene Orange YN are believed to consist of a mixture of compounds 11, III, IV and V. Chrysoidine YBH, CI Basic Orange 3 is believed to consist of a mixture of compounds I and 11.The table should only be regarded as a guide to composition as the generic names and CI number have been used unsystematically by several manufacturers Commercial/ Structure generic name CI number Systematic name I Chrysoidine Y Chrysoidine Crystal* CI Basic Orange 2 Chrysoidine G (USA) It Chrysoidine Y: base Chrysoidine G (UK) CI Solvent Orange 3 Paper Yellow CI Basic Orange 1 Cerotin Orange nt Chrysoidine R: base CI Solvent Orange 4 m 11 Chrysoidine R - - IV V CI 11270 4-(Phenylazo)-l,3-diaminobenzene monohydrochloride CI 11270 : 1 4-(Phenylazo)-l,3-diaminobenzene CI 11320 4-Methyl-6-(phenylazo)-l,3-diamino- benzene monohydrochloride CI 11320 : 1 4-Methyl-6-(phenylazo)-1,3-diamino- benzene CI 11320 2-Methyl-6-(phenyiazo)-l,3-diamino- benzene CI 11325 4-Methyl-6-( l-methylphenylazo)-l,3- diamino benzene CI 11325 2-Methyl-6-( l-methylphenylazo)-l,3- diaminobenzene * Chrysoidine Crystal is used in the USA to refer to a mixture of Chrysoidine Y and homologues from 0- and p-toluidine. t Free base.ANALYST, NOVEMBER 1988, VOL.113 1615 The most common of these are Chrysoidine Y , also called CI Basic Orange 2 (CI 11270), and Chrysoidine R, also called CI Basic Orange 1 (CI 11320). The Society of Dyers and Colourists allocate CI generic names in the Colour Index.22 Both of these dyes are in the form of the amine hydrochloride or, occasionally, the acetate. When marketed in the form of the free base, they are called CI Solvent Orange 3 (CI 11270 : 1) (in the UK sometimes called Chrysoidine G) and CI Solvent Orange 4 (CI 11320 : l), respectively.Liquid chryso- idine is typically a solution of Chrysoidine Y base in acetic acid. Where this is sold for use in the manufacture of brown paper it is often called Paper Yellow or, in admixture with a blue dye, Paper Brown. Some of the commercial chrysoidine dyes are, however, blended to modify their colouring properties. Commonly, mixtures of Chrysoidine Y and R are produced. The dye is usually designated after the predominant chrysoidine or according to whether the orange dye produced is more yellow (Y) or red (R). This can result in a product designated Chrysoidine R containing substantial amounts of Chrysoidine Y and vice versa. To complicate the matter further the colour of Chrysoidine R may be modified to be more yellow by the partial replacement of aniline by o-toluidine in the manufac- turing process.Such dyes are regarded as equivalent or even superior substitutes for Chrysoidine Y and typically are given, erroneously, the same CI number. Formerly they were marketed as Chrysoidine YN, but in recent years, to reduce confusion, are marketed under the names Methic Orange YN or Methylene Orange YN. Chrysoidine YBH, designated CI Basic Orange 3, appears from analysis to be a mixture of Chrysoidine Y and R similarly modified by partial replace- ment of aniline by o-toluidine in the manufacturing process. Not surprisingly, this situation gives rise to confusion and leads to occasional mislabelling by secondary distributors and even laboratory chemical suppliers. An analysis of the chrysoidine-based products of the major UK manufacturers might therefore be useful to the analyst confronted with a chrysoidine dye of unknown composition.Similarly, such an analysis might be instructive to those considering the carci- nogenic effects of these dyes and who believe them always to be the simple pure products their labelling implies. Method of Manufacture A brief description of the general method of manufacture of the chrysoidine dyes will help to explain the origin of some of the components found in the commercial products. For Chrysoidine Y and Chrysoidine R, aniline is diazotised and the resulting diazonium salt is coupled with an aromatic diamine to produce a dye. The diamines used for coupling are rn-diaminobenzene for Chrysoidine Y and diaminotoluene (TDA) for Chrysoidine R.For example, the manufacture of Chrysoidine Y can be represented as Aniline Diazonium salt (1) Diazonium salt Diaminobenzene Chrysoidine Y Coupling takes place para to the amino groups, which, in the preparation of Chrysoidine Y, results in only one product. Typically, when TDA is used as the coupling agent, a dye that is a mixture of two compounds is produced. This is because commercial grade TDA is usually a mixture of 2,4- and 2,6-&aminotoluene, approximately 80 + 20. Coupling with commercial TDA consequently leads to a mixture of dye products based on 2,4- and 2,6-TDA, respectively. Replacing aniline by o-toluidine as the species to be diazotised in the synthesis of Chrysoidine R results in a chemically similar dye, CI 11325, which is usually a mixture of two compounds, IV and V, which are derived from 2,4- and 2,6-diaminotoluene, respectively. Using a mixture of aniline and o-toluidine in the diazotisation produces a mixture of the R chrysoidines and the dye CI 11325 that is marketed under the names Methic Orange or Methylene Orange.Experimental Principle High-performance thin-layer chromatography (HPTLC) can be used to analyse the dyes as the free bases, hydrochlorides or acetates. The method is based on the fact that the dyes are best chromatographed as the free bases, but are more resistant to degradation on a dry HPTLC plate when in the form of the hydrochlorides. The bulk dye (or the dust on a filter if collected from air) is dissolved in an ethanol - hydrochloric acid mixture to stabilise any free base as the hydrochloride. After application to the HPTLC plate, exposure to ammonia converts the hydrochloride to the free base for chromato- graphic separation The separated dyes on the HPTLC plate are then exposed to hydrochloric acid fumes in order to stabilise the dyes as the hydrochlorides for quantification.Apparatus Membranefilters. Type DM 800,25 mm (Gelman Sciences). Thin-layerplates. Merck 20 X 10 cm silica gel HPTLC plates (not conditioned) without fluorescent indicator (catalogue No. 5641). Chromatography paper sheets. Whatman No. 3. Horizontal linear developing chamber for HPTLC. Suitable for 20 x 10 cm plates (Camag) fitted with thin glass base-plates. Horizontal glass-conditioning tray for TLC plates Sample applicator.Nanomat (Camag). Micropipettes. Disposable Microcap, 0.5 pl (Drummond Micro-reaction vessels. Tall-form 2-ml Wheaton Reactivials Vortex evaporator. Buchler Instruments. Attached to an Scanning densitometer. Camag Model 1 collimated TLC Integrator. Spectraphysics Model SP 4100. Gas chromatograph. Varian Model 3700. Fourier transform infrared analyser. Nicolet Model 7199. (Cam%). Scientific). fitted with PTFE-faced septum seals (Supelco). oil-filled vacuum pump. scanner. Reagents Analytical-reagent grade reagents were used throughout unless indicated otherwise. Ammonia solution, sp. gr. 0.880 (Fisons). Hydrochloric acid, sp. gr. 1.18 (BDH). Chloroform (stabilised with 2% ethanol) (Fisons). Dichloromethane (stabilised with 0.005 YO amylene) Chrysoidine Y [Chrysoidin (Basic Orange 2, CI 11270)l (Fisons).(Aldrich).1616 ANALYST, NOVEMBER 1988, VOL. 113 Table 2. RF values of chrysoidines in mixed solvents RF ~~ ~ ~ Dye 1* 2* 3* 4* 5* 6* 7* ChrysoidineY . . 0.39 0.46 0.40 0.43 0.43 0.31 0.24 ChrysoidineR , . 0.48 0.54 0.45 0.50 0.44 0.33 0.29 CompoundIV . . 0.54 0.60 0.52 0.57 0.49 0.38 0.33 ChrysoidineRz,a . . 0.66 0.71 0.62 0.67 0.58 0.45 0.40 CompoundV , . 0.72 0.75 0.66 0.72 0.61 0.48 0.44 * Eluents: (1) Chloroform (containing 2% ethanol as stabiliser); (2) cyclohexane - dichloromethane (20 + 80); (3) toluene - dichloromethane (30 + 70); (4) benzene - dichloromethane (30 + 70); (5) toluene - ethyl acetate (60 + 40); (6) toluene - pyridine (80 + 20); and (7) cyclohexane - dichloromethane - acetone (40 + 40 + 20). Table 3.Analysis of chrysoidine samples. All percentages are calculated as free base equivalents. Compounds IV (estimated as Chrysoidine R) and V (estimated as Chrysoidine R2,J are identifications tentatively attributed to spots C and E, respectively (Fig. 6) Chrysoidine, 70 Dye No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Manufacturer?/ Dye supplier$. Y Paper Basic Yellow C (130%) . . C.A. 79 PaperBasicYellowliquid . . . . C.A. 15 Chrysoidine Y * . . . . . . . . C. C . 71 Chrysoidine base * . . . . . . C. C. 96 Chrysoidine base liquid* . . . . C.C. 22 Methylene Orange YN liquid . . C.C. <1 Chrysoidine Y powder . . . . . . DYK. 44 ChrysoidineYliquid . . . . . . DYK. 20 MethicOrangeYN . . . . . . ICI 1 Chrysoidine RN . . . , . . . . ICI 21 Chrysoidine Base A .. , . . . WIL. 97 Chrysoidine Y base . . . . . . ALD. 100 Chrysoidine base . . . . . . . . DYK. 96 Chrysoidine YBH* . . . . . . WIL. 36 Chrysoidine Y . . . . . . . . ALD. 82 Chrysoidine Y , . . . . . . . BDH 8 Chrysoidine R . . . , . . . . BDH 28 Chrysoidine Y . . . . . . . . P&B 66 Chrysoidine R . , . . . . . . P&B 5 R <1 <1 <l <1 <1 11 <1 <1 <l 40 42 45 <I <l <1 32 25 <1 44 IV R2,6 V <1 <1 <1 <1 <1 <1 c 1 <1 <1 <1 el <1 <1 <1 <l <1 <1 <1 <1 <1 <1 <1 <1 <1 14 16 -5 <1 10 <1 5 <1 <1 <1 <1 <1 c1 <1 <1 <1 c1 <1 12 8 3 c 1 11 <1 tl <1 <l c1 7 <1 1.4 4.1 0.9 * Discontinued product. t UK manufacturers: C.A., Clayton Aniline; C.C., Croda Colours; DYK., Dykem (North); ICI, ICI Organics Division; and WIL., Williams $ Laboratory chemical suppliers: ALD., Aldrich; BDH, Gurr supply from BDH; and P&B, Pfaltz and Bauer.§ Poorly separated: compound V peak integrated into the R2.6 result. (Hounslow). 2,4-Diaminotoluene (Aldrich) . 2,6- Diaminotoluene (Aldrich) . 0- Toluidine (Aldrich) . Aniline (Aldrich). Sodium hydroxide (BDH). Sodium nitrite (BDH). Tin(1l) chloride dihydrate (BDH). N-1-Naphthyleihylenediarnine dihydrochloride (BDH). The chrysoidine dyes included in this work are listed in Table 3. Standard Dyes Chrysoidine Y that is sufficiently pure for use as a standard material is available from Aldrich, and Pfaltz and Bauer. No commercially available Chrysoidine R was discovered that was suitable for use as a standard and Chrysoidine R2,6 is not marketed. These were synthesised by the following proce- dure.The method is based on the preparation of diazo- aminobenzene23 and on patented methods for the manufac- ture of pure Chrysoidine Y.24 The reaction of aniline (or an analogue) with sodium nitrite produces a diazonium salt, which is then coupled with a diamine to produce the azo dye. Chrysoidine Y (I), Chrysoidine R (II) and the 2,6-TDA variant of Chrysoidine R (III) were prepared in this way. The preparation of Chrysoidine R is now described. Dissolve 5 g of aniline in a mixture of 18 ml of concentrated hydrochloric acid and 50 ml of water. Cool to -4 "C in an ice - salt bath. Slowly add a solution of 3.7 g of sodium nitrite in 15 ml of water. Add dropwise, with stirring, a solution of 6.54 g of 2,4-TDA (the stoicheiometric amount) in a mixture of 18 ml of concentrated hydrochloric acid and 50 ml of water.Ensure that the temperature of the reaction mixture remains below 0°C. The addition of a solution of 25 g of sodium acetate dissolved in 40 ml of water to the colourless solution produces a red precipitate. After leaving the solution to stand for 30 min, filter the precipitate, wash it with distilled water and either dry in a vacuum desiccator for storage or prepare the free base directly as described below. HPTLC analysis of the product should reveal only one component spot. Standard Solutions It is not advisable to use the chrysoidine hydrochlorides to prepare standard solutions directly, as the stoicheiometric formula cannot be assumed.Solutions prepared from the free bases are the preferred primary standards. Because of the tendency of the free bases to degrade on the HPTLC plate, they are effectively converted to the more stable hydro- chloride by the addition of 2% of concentrated hydrochloric acid to the absolute ethanol used as the make-up solvent for the calibration solutions.ANALYST, NOVEMBER 1988, VOL. 113 1617 Preparation of the chrysoidine free base Dissolve the hydrochloride in the minimum volume of water and filter. Add a 30 g 1-1 solution of sodium hydroxide slowly with stirring until the dark red solution changes to a pale yellow suspension. Leave the suspension to stand for 15 min and filter the precipitate. Wash with distilled water, dry in a vacuum desiccator and store in tightly sealed dark bottles kept under refrigeration. Procedures Application of calibration standards Prepare stock solutions of the base dyes in a solution of 2% hydrochloric acid in ethanol at concentrations of 200 pg ml-1.These are best stored in a refrigerator where they remain stable for several months. To prepare calibration solutions, serially dilute each of the stock solutions five times with an equal volume of the acidified ethanol. These serially diluted solutions (0.5 pl) applied to the plate give standard spots containing 100,50, 25, 12.5,6.2 and 3.1 ng of chrysoidine base. For accurate work, standards and samples are best spotted in duplicate. In this instance it is preferable to spot a separate plate for the analysis of each chrysoidine species of interest, otherwise standards encroach too heavily on the area of the plate that can be devoted to sample spotting.If approximate results for the levels of Chrysoidine R and R2,6 are acceptable, only one set of Chrysoidine Y standards need be applied. The resulting calibration graph can be used to determine from the appropriate sample peak areas not only the concentration of Chrysoidine Y, but also the approximate concentration of the other chrysoidine species by multiplying the result by a factor of 1.25 (see Comparison of Peak Areas for Sets of Chrysoidine Y, R and R2,6 Standards on Several Different Plates). Sample application Apply 0.5 p1 of each sample, blank and standard at 5-mm intervals across the HPTLC plate using disposable Drum- mond micropipettes in the Camag Nanomat.Single applica- tions of more than 0.5 pl are not advisable as they result in too much spreading of the chrysoidine spot. Dry the spots with a jet of nitrogen. The outer 25 mm of the plates should be avoided as evaporation of the eluting solvent causes consider- able outward displacement at the edges. A significant reduction in the intensity of the dye spots occurs with increasing residence time on the dry HPTLC plates. In order to keep losses to an acceptable level (<lo%), sample application must be completed in less than 30 min and analysis begun immediately. This places some restrictions on time-consuming processes such as repeated over-spotting, to improve detection limits, and precludes the spotting of both sides of the plate to increase sample throughput per plate.HP TL C considerations Preliminary investigations indicated that the hydrochlorides were difficult to chromatograph, especially on silica gel HPTLC plates, whereas the free bases eluted readily. Conse- quently, a standard procedure was adopted of converting all samples to the corresponding bases on the HPTLC plate by exposure to ammonia vapour before separation. Chloroform is the recommended eluent. Other eluents were investigated, but only chloroform and dichloromethane gave acceptable separation of all five chrysoidine components in the Methic Orange mixture (IV). Longer chain chlorohydro- carbons gave similar separations, but the development times were long, resulting in diffuse overlapping spots.Some mixed solvent systems gave successful separations, but evaporation of the volatile chlorohydrocarbons at the edge of the plate distorted the elution profile. However, in certain circum- stances, such as the presence of some interfering dye spot, these mixtures might prove to be of value. Table 2 shows the RF values obtained. Chromatography It is recommended that the chromatography be carried out in a fume cupboard. Replace the standard thickness glass base-plate in the Camag 20-cm horizontal developing chamber with the thinner version. Place a sheet of Whatman No. 3 chromatography paper (cut to fit and then laterally, in half, to reduce curling) on to the glass plate and wet thoroughly with ammonia solution (about 15 ml). Position the HPTLC plate in the developing chamber in its normal inverted position so that the spotted dyes are exposed to the ammonia.Wait for 1 min before beginning elution in order to allow conversion of the hydrochlorides to the corresponding free bases. Allow the chloroform to develop along the plate for about 50 mm and shield the developing chamber from direct sunlight with a black plastic sheet. Remove the plate after elution and dry thoroughly with a jet of nitrogen (the dye spots are yellow at this stage). Place the plate face down in a Camag horizontal exposure tank above concentrated hydrochloric acid. For better exposure of the silica gel surface, support the plate on two thin glass rods. The acid fumes re-convert the dyes to the hydrochlorides and the chrysoidine spots turn a dark red colour.Remove the plate after 5 min and dry thoroughly with a jet of nitrogen (the spots are now orange). Allow the colour to stabilise for 10 min in a dark cupboard. Measurement With a scanning densitometer, track the line of development along which the component spots of the dyes are now separated. Using an integrator, monitor the absorbance of the reflected mercury lamp emission at 436 nm. (A slit length of 3.5 mm and a slit width of 0.3 mm are suitable when using a Camag Mk 1 densitometer.) From the resulting chromato- grams, obtain values for the component peak areas corre- sponding to the separated chrysoidine spots. Calculation Results are obtained on the basis of a calibration graph constructed for the standards spotted on the same plate.The calibration graph has been shown to be reasonably linear up to 100 ng (Fig. 2). There is usually a small intercept due to the 0 12.5 25 50 C h ry so id i n e/ng 100 Fig. 2. Typical calibration graphs for the amount of chrysoidine s otted vs. the integrated peak area of the scanned HPTLC spot. A, Zhrysoidine Y; B, Chrysoidine R; and C, Chrysoidine R2,61618 inevitable loss of several nanograms of chrysoidine during sample application. However, a significant error is only introduced if very low levels of chrysoidine are to be measured. Experience has shown that the limit of detection is represented by about 3 ng of chrysoidine as a spot on the plate. This is the amount that could be confidently detected with an accuracy of about f25%. However, it is recommended that for all samples for which less than 10 ng of chrysoidine are detected from spotting 0.5 pl, the analysis should be repeated.This time the sample should be over-spotted twice or even four times together with the low-level standards (allow the solvent to evaporate between successive applications). Only small batches of samples can be processed in this manner if spotting times are to be kept acceptably short. The peak-area ratios are sufficiently reproducible within a given batch of HPTLC plates to suggest that estimates of the Chrysoidine R and Chrysoidine R2,6 content can be made by application of the appropriate factors (1.25 in both instances; see Table 4) to the results obtained using a Chrysoidine Y calibration graph. This has the significant advantage of requiring only one commercially available standard, Chrysoi- dine Y. If results of a higher degree of accuracy are required, standards for the other two dyes must be synthesised as detailed earlier.Proposed air sampling method Samples can be collected by drawing air at 2 1 min-f through 25-mm Gelman DM 800 filters using a suitable pump. The sampling time should be no longer than 1 h. Immediately on completion of the sampling, transfer each filter into a 2-ml tall-form Reactivial containing 1.5 ml of a 2% solution of hydrochloric acid in absolute ethanol. Seal the vials with PTFE-lined septa. The samples can be stored under refrigeration in this state for several weeks without deteriora- tion. Allow a minimum time of 1 h to desorb the dye if the samples are to be analysed immediately.Treat an unused ANALYST, NOVEMBER 1988, VOL. 113 blank filter in the same way as the samples. Prepare two loss-monitoring standards in vials by adding 200 pl of the 100 pg ml-1 Chrysoidine Y standard to 1.3 ml of the acidified ethanol (these standards are to be used later to correct for any losses in the sample preparation stages). Remove the filters from the sample Reactivials with the minimum loss of liquid and place them with the blank and the two loss-monitoring standards into the vortex evaporator. A vacuum is applied to the vortex evaporator, which should be used in the “anti- bump” mode and without heating. Drape black plastic over the vortex evaporator to exclude light. To minimise losses, removal of the solvent should take no longer than 30 rnin.Experiments carried out at levels equivalent to 40, 10 and 2.5 pg of Chrysoidine Y deposited on DM 800 filters demon- strated that, provided the evaporation time in the vortex evaporator is limited to 30 min, loss of Chrysoidine Y will be less than 10%. Immediately after removal of the solvent, add 200 pl of the ethanol - hydrochloric acid mixture to the residue in the Reactivial, re-seal and swirl the vial vigorously for 5 min to re-dissolve the sample. Analyse the samples, blank and loss-monitoring standards as detailed in earlier sections. Correct for any losses occurring during evaporation by applying the factor by which the peak areas of the loss- monitoring standards are reduced compared with those of a 100 pg ml-1 Chrysoidine Y standard.No doubt o‘ther evaporation techniques such as blowing down with nitrogen could also be used. However, it would be advisable to carry out experiments using standard solutions in order to estimate losses before committing samples to this process. Screening for o-toluidine-based chrysoidine components Tin(I1) chloride has been employed in the reductive cleavage of the azo bond of azo dyes to recover and determine the diazotised base used in their manufacture. A macro-scale Table 4. Peak areas obtained by scanning plates spotted with chrysoidine standards Peak area,* units x 103: c o - Peak area/Peak area Y Masslng 100 50 25 12.5 6.25 3.1 6.25-1 00 12.5-100 Chrysoidine Y R Y R Y R Y R Y R Y R R2.6 R2,6 R2,6 R2,6 R2,6 R2,6 Experiment No. Experiment No.1 858 738 71 1 485 405 387 251 199 198 114 86 89 54 35 39 22 19 11 2 897 802 792 536 463 453 277 239 23 1 140 113 112 60 49 50 29 25 26 3 867 802 812 479 386 384 230 171 181 103 71 78 45 27 27 11 10 15 1 2 3 0.86 0.89 0.96 0.83 0.88 0.94 0.88 0.86 0.81 0.84 0.85 0.80 0.79 0.86 0.74 0.79 0.83 0.79 0.75 0.81 0.69 0.78 0.80 0.76 0.65 0.82 0.69 0.72 0.83 0.60 0.86 0.86 0.91 0.50 0.90 1.30 Average 0.90 0.88 0.85 0.83 0.80 0.80 0.75 0.78 0.72 0.72 0.88 0.90 efficient of varia- tion,+ % 5.2 6.3 4.2 4.3 3.2 4.3 7.5 2.9 6.7 6.7 2.6 5.5 12.3 16 32 44 - 3.3 Average: 0.80 Average: 0.80 Average overall: 4.9 * Peak area results are the average of duplicate spottings. + Coefficients of variation are calculated for R and R2.6 from peak-area ratios against spots of Y of equivalent concentration on the same plate and for Y from peak-area ratios against spots of twice the concentration on the same plate.ANALYST, NOVEMBER 1988, VOL.113 1619 method involving reflwing , neutralisation and analysis by conventional TLC has been described for benzidine-, o-tol- idine- (not to be confused with o-toluidine) and o-dianisidine- based dyes.25 The HPTLC method already described for the analysis of chrysoidines lends itself to adaptation for the micro-scale determination of diazotised species used in their preparation, i.e., toluidine or aniline.26 Chrysoidine dyes can be reductively cleaved by tin(I1) chloride in the cold, avoiding a refluxing stage. The exposure of the plate to ammonia during elution obviates the need for a separate neutralisation stage.Dissolve 4 mg of chrysoidine dye in 1 ml of ethanol contained in a vial. Reserve 100 pl for later analysis. Add 100 pl of a 0.25 g ml-1 solution of tin(I1) chloride in concentrated hydrochloric acid. Seal and allow at least 1 h for complete reaction, indicated by decolorisation of the solution. Using 0.5- or 1-pl disposable Drummond micropipettes spot 1 pl of the solution on to an HPTLC plate alongside 0.5 pl of the unreacted solution reserved earlier. Also spot as references 0.5 pl of 1 mg ml-1 o-toluidine and 0.5 pl of 0.5 mg ml-1 aniline solutions in ethanol. By repeated applications, up to 4 pl of sample solution can be applied to improve detection levels. Dry the spots briefly with a jet of nitrogen or air. Proceed according to the HPTLC method already de- scribed, but use the following method to detect the aromatic amine spots.Expose the HPTLC plates for 10 min to concentrated hydrochloric acid fumes, then transfer them directly into a second exposure tank within which is a small test-tube containing sodium nitrite. Allow a few drops of concentrated hydrochloric acid to run into the test-tube and then seal the tank. Expose the plate to the resulting oxides of nitrogen for 2 min. Briefly play a jet of nitrogen over the surface of the plate before dipping it into a 0.5% rnlVsolution of N-1-naphthylethylenediamine dihydrochloride in water - ethanol (1 + 1). Dry rapidly with a jet of nitrogen. Aromatic amines are revealed as purple spots. o-Toluidine (RF, 0.82) is well separated from aniline (RF, 0.70) and is readily distinguishable from rn-toluidine (RF, 0.75) and p-toluidine (RF, 0.72).The last two compbunds tend to be masked by a very intense aniline spot. o-Toluidine yields the least intense spot, presumably owing to a tendency to auto-diazotise. The amine spots are well clear of the spots arising from any residual unreacted chrysoidine. A represen- tation of a typical plate showing the screening of the tin(I1) chloride reduction products of several commercially available chrysoidine dyes is shown in Fig. 3. The method is suitable for the determination of chrysoidines containing not less than 25% o-toluidine-based dyes. An approximate estimate of the amount of o-toluidine-based dye can be made by visual comparison with standard spots. Determinations down to 5% have been performed successfully by reacting up to 16 mg of dye.For single-component dyes, 1 or 2 mg would be sufficient. I[f desired, further confirmation of the presence of o-tolu- idine may be obtained by sealing in a septum vial the 1 ml of reacted solution together with 2 ml of a 0.05 g ml-1 solution of sodium hydroxide and 0.5 ml of dichloromethane and shaking vigorously. The free amines extracted into the dichloro- methane lower layer can be analysed by capillary gas chromatography. A 25 m X 0.33 mm i.d. silica glass capillary coated with 0.5 pm of bonded-phase BP1 was used in our analysis, but no doubt many other phases would be suitable. Identification of component peaks could be based tentatively on retention time data. However, if the instrumentation is available, confirmation from either Fourier transform infrared spectra or mass spectra is advisable.Fig. 4, for example, shows a gas chromatogram and Fig. 5 the associated confirmatory GC - FTIR spectra of aniline and o-toluidine released from Gum Chrysoidine Y. It should be noted that o-toluidine should be regarded, for practical purposes, as if it presented a carcinogenic risk to humans.14 Appropriate precautions should be taken to minimise exposure when handling solutions of this amine , and those solutions of chrysoidine that have been reduced to release it. 0 0 O**OOO 8 6886000 8 0 0 o o 0 0 0 0 0 0 0 0 0 0 a 8 0 0 ~ 0 8 8 0 ~ ~ 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Fig. 3. Screening method for the determination of the presence of 0 5 10 Timdmin Fig. 4.GC - FTIR confirmation of the presence of o-toluidine- and aniline-based components in chrysoidines. Gas chromato ram (flame ionisation detector) of Chrysoidine Y (Gurr) reducefi by tin(I1) chloride, neutralised and extracted into dichloromethane according to the described method. (1) Dichloromethane solvent; (2) aniline; and (3) o-toluidine. Column: BP1, silica, 25 m X 0.32 mm i.d., thickness 0.5 pm. (BPI is a general-purpose non-polar bonded phase. Its use gave rise to more peak tailin than would one of the more polar phases specifically recommended for amine analysis.) Temperature pro- gramme: held at 50 "C for 1 min then increased at a rate of 10 "C min-1 to 90 "C. Injection: 5 p l Grob splitless. (A large injection was chosen in order to obtain strong infrared s ectra when the GC peak components were eluted through the lig\t-pipe gas cell of the FTIR, consequently the GC conditions are far from ideal, the column being heavily overloaded.)1620 ANALYST, NOVEMBER 1988, VOL.113 4000 3500 3000 2500 2000 1500 1000 4000 3500 3000 2500 2000 1500 1000 Wavenum ber/cm -1 Fig. 5. Infrared spectra of the eluting GC peaks (lower traces) compared with library spectra selected by computer search as the best match (upper traces). ( a ) Aniline (Fig. 4, peak 2); and (b) o-toluidine (Fig. 4, peak 3) Results and Discussion Elucidation of the Composition of UK Chrysoidine Dyes Samples were obtained from the major producers of chryso- idine dyes in the UK and from several laboratory suppliers to the UK in 1983. The spot distribution on the HPTLC plate obtained for a selection of these dyes using the described procedure is illustrated in Fig.6. Spot Attribution Five principal spots were revealed (Fig. 6, A-E). Spots A, B and D were assigned (on the basis of having identical RF values to synthesised compounds using a number of different eluents) to Chrysoidine Y, R and R2,6, respectively. Spots C and E are believed to be the mefhylated chrysoi- dines IV and V, respectively, on the following evidence. Milligram amounts of the component bases corresponding to spots B, C, D and E were obtained using preparative-scale TLC from the chrysoidine mixture marketed as Gurr Chrysoi- dine Y by BDH (component A, already identified as Chrysoidine Y, was also present in the dye, but was inadequately separated and so was not isolated).Infrared spectra for all the components proved to be very similar to a Fig. 6. Spot distribution on the HPTLC plate obtained using the described procedure for a selection of chrysoidine dyes manufactured in the UK or available as laboratory supplies. The diagram is based on black and white photographs taken with a Polaroid Land Camera mounted on a Camag Reprostar TLC photographic unit. The s ot sizes represent those which would be obtained for 1 pg of dye (solifor liquid). The spots are larger than wou!d be optimum for quantitative work for (he major dye components. Standards are spotted at 100-n levels. (A) Precipitated base from Chrysoidine Y (Aldrich); (Bj synthesised Chrysoidine R base; (C) component separated by preparative TLC from the Gurr supply labelled Chrysoidine Y obtained from BDH [assumed to be com ound IV (Fig.l)]; (D) synthesised Chrysoidine R2 6 base; and (I37 component derived by preparative TLC from BDh, Gurr Chrysoidine Y believed to be compound V (Fig. 1). The identities of the numbered dye spots are given in Table 3 library spectrum of Chrysoidine Y. The spectra for com- ponents C and E, however, significantly lacked an absorption band at 71CL690 cm-1 attributable to a monosubstituted benzene ring, suggesting that a substituted aniline was involved in their manufacture. Using the described reductive screening method, the Gurr chrysoidine mixture was demon- strated to be based on a diazotised mixture of aniline and o-toluidine (see Fig. 3). The reductive release of these components was confirmed using GC - FTIR (see Figs.4 and 5). Application of the method to the single-component samples confirmed that components B and D, corresponding to Chrysoidine R and Chrysoidine Rz,~, respectively, were aniline based. Components C and E were shown to be o-toluidine based. From these results, it is reasonable to deduce (assuming any Chrysoidine Y to be present as a later contaminant) that the bulk of the mixture of chrysoidines in the Gurr dye was produced by reacting a diazotised mixture of aniline and o-toluidine with commercial diaminotoluene, which is usually an 80 + 20 mixture of the 2,4- and 2,6-isomers. On the basis of relative abundances, it is reasonable to attribute component C to structure IV, the reaction product of o-toluidine and 2,4-TDA.This is corroborated by component C having a similar RF value to the chemically similar component D, the reaction product of aniline and 2,4-TDA. By the same arguments, component E is attributed to structure V, the reaction product of o-toluidine and 2,6-TDA. Again this is corroborated by E having a similar RF value to the chemically similar component D, the reaction product of aniline and 2,6-TDA. Composition Revealed by the Spot Profile of the Eluted Samples Many of the products labelled as Chrysoidine Y, or as Basic Orange 3, Paper Yellow or Chrysoidine Liquid, gave the simple single spot of the expected Chrysoidine Y. Laboratory supplies of Chrysoidine Y obtained from Aldrich and Pfaltz and Bauer also yielded this single spot. However, not all the dyes proved to be of such simple composition. Chrysoidine YBH, a discontinued product from Williams, yielded two similar sized spots corresponding to Chrysoidine Y and Chrysoidine R with an additional spot corresponding to the further methylated chrysoidine, IV.The presence of 5-10% of species based on o-toluidine was demonstrated using the described screening method and visual estimation (see Fig. 3). Methic Orange YN was found not to contain any Chrysoi- dine Y. The spot profile indicated that the bulk of the dye was consisted of Chrysoidine R and Chrysoidine R2,6 in the expected proportions together with further methylated chrysoidines, IV and V. Methylene Orange YN, manufac- tured by Croda Colours, was shown by its spot profile to be of a very similar composition to Methic Orange YN.A visually estimated 2630% of the chrysoidine in both dyes was demonstrated to be o-toluidine based using the screening method (see Fig. 3). The Gurr supply of Chrysoidine Y for microscopy staining obtained from BDH also proved to be mostly a mixture of Chrysoidine R isomers and the further methylated chrysoidine, as already stated. Chromatographic- ally this was very similar to Methic Orange YN (ICI) and therefore seems likely to have been derived from this source and later mixed with Chrysoidine Y. This is consistent with the correspondence between Searle and BDH reported in refer- ence 27. Chrysoidine RN (ICI), in addition to the expected isomers of Chrysoidine R, was shown to contain substantial amounts of Chrysoidine Y.A Gurr supply of Chrysoidine R for microscopy staining obtained from BDH was, from its spot profile, evidently Chrysoidine RN (ICI), re-packaged. A laboratory supply of Chrysoidine R from Pfaltz and Bauer contained Chrysoidine Y, in addition to the expected Chrys- oidine R isomers.ANALYST, NOVEMBER 1988, VOL. 113 1621 Quantitative Analysis of UK Manufactured Chrysoidine Dyes Examples of the scanning densitometer chromatograms obtained for the dyes are shown in Fig. 7. A quantitative analysis of the dyes based on the integrated peak areas is given in Table 3. Because dyes vary from batch to batch, the analysis does not purport to be definitive and is presented only as a guide to the composition of UK manufactured dyes. The dyes were analysed for Chrysoidine Y , Chrysoidine R and Chrysoidine R2.6.Analysis for these compounds was based on calibrations obtained from synthesised standards spotted on to the analysis plates. The concentrations of the components corresponding to spots C (assumed to be compound IV) and E (assumed to be compound V) (Fig. 6) were only of incidental interest in this study. No standards were synthesised for these compounds. For completeness, however, these were calculated from calibrations for Chrysoidine R and Chrysoidine R2,6, respec- tively. This approximation was justified on the basis that compounds differing in composition by only one methyl group are likely to give a similar response. The separation of Chrysoidine R2,6 and compound V was not ideal, as is evident from Fig. 7. Consequently, the results for these two dyes are sometimes quoted as their total per- cent ages.Comparison of Peak Areas for Sets of Chrysoidine Y, R and R z , ~ Standards on Several Different Plates This exercise, the results of which are given in Table 4, was useful in that it demonstrated the consistency in the shape of the calibration graphs and gave an approximation to the limit of detection. Of particular interest was the variability of the ratio of the response for the Chrysoidine R and R2,6 standards to that of Chrysoidine Y of the same concentration. Absolute repeatability was not sought, because differences in spot intensity due to the variation in the thickness of the absorbent layer from one HPTLC plate to another are (a) Time - fig. 7. Examples of integrator - output chromatograms for densi- tometer scans of elution tracks of I-pg spottings of chrysoidine dyes eluted with chloroform.(a) Gurr Chrysoidine Y (BDH); and (b) Chrysoidine RN ICI). I, Chrysoidine Y; 2, Chrysoidine R; 3, front compound IV; 4, L hrysoidine R2,6; 5, cornpound V; and 6, solvent inevitable, The utility of the method depends not on absolute calibration, but on a calibration based on standards spotted together with the samples on to the plate. The calibration graphs obtained were reasonably consistent and sufficiently similar in shape to be useful up to 100 ng; one set of graphs is shown in Fig. 2. Typically, there was a small intercept on the concentration axis (corresponding to about 1 ng), presumably due to losses from the plate. The intercept was sufficiently close to zero to permit measurement down to 3 ng.Coefficients of variation of the ratios of average areas of duplicate spottings to those obtained for equivalent spots on the same plate were of the order of 5%, except at the lowest levels of concentration. The average value of the peak area to peak area Y ratios (the chrysoidine mass being in the range 6.25-100 ng for all three plates) was 0.80 for Chrysoidine R and 0.80 for Chrysoidine R2,6. The ratios were, to an extent, dependent on the concentration of the spot. Nevertheless, for each com- pound, only two values of the 15 determined fall outside rtl5%0 of this value. Frequently, the main concern in analysis will be to determine the concentration of Chrysoidine Y, and the concentrations of Chrysoidine R and R2,6 will be of secondary interest.Also, when monitoring dye levels in air, absolute accuracy will not be important. In such instances an error as small as this would be acceptable and it would be necessary merely to spot Chrysoidine Y standards and to use the calibration graph so obtained to determine Chrysoidine R and R2,6 by multiplying the concentration obtained by 1/0.80, i.e., 1.25. It should be borne in mind, however, that the results obtained were from plates all taken from the same batch. It is possible that a different batch of plates might give a different ratio. It is wise, therefore, to spot at least one standard of each of Chrysoidine R and R2,6 as a check, if these are available. Results Affecting the Choice of Procedure Losses due to prolonged exposure of the spots to air on the HPTLC plate Losses are acceptable (<lo%) using the described procedure provided that sample application is completed within 30 min.In comparison, it can be seen from Fig. 8 that losses are serious if hydrochloric acid is not routinely added to solutions of the dyes as the free bases. After 1 h, only 40% of Chrysoidine Y, 47% of Chrysoidine R and 61% of Chrysoidine R2,6 remain. Fig. 9 shows the scanning densitometer chromatograms obtained for 100 ng of the free bases of Chrysoidine Y, Chrysoidine R and Chrysoidine R2,6 spotted from solution in unacidified ethanol and either eluted immediately or eluted after the plate had been left exposed to the air. Note the marked reduction in the chrysoidine peak area and the appearance of other spot peaks due to degradation products.Although, for chromatography, the dyes are converted to the free base they are only exposed to the air on the dry plate for periods immediately before and after elution, which are deliberately kept short. During elution they are protected from attack as the air is excluded from the silica gel. No significant losses seem to occur. Choice of filter material for sample collection As chrysoidine is a dyestuff for fabrics and plastics, irrevers- ible absorption of the dye by the filter material is a potential prob€em. A range of potentially suitable filters of different compositions was investigated for their suitability. Glass-fibre GFA filters were not considered as the fibres tend to block the micropipettes during HPTLC sample application.To test their suitability, 5 pg of Chrysoidine Y base (two 12.5-pl applica- tions of a 200 pg ml-1 solution in ethanol) were applied to each1622 750 m 2 2 500 m L 250 0 12.5 25 50 Ch rysoid i nelng 100 Fig. 8. Plot of the integrated peak areas for scans of eluted HPTLC spots vs. amount of Chrysoidine Y (base) spotted, indicating the improved stability of Chrysoidine Y (base) standards to which hydrochloric acid has been added; exposure on the HPTLC plate is shown in parentheses. A, Base + HCl (5 min); 0, base + HCl (30 min); X , base + HCl (1 h); 0, base (5 min); and 0, base (1 h) Scan time - Fig. 9. Scans showing the marked reduction in peak height when 100 ng of chrysoidine base spots are eluted after prolonged exposure on the HPTLC plate.Chrysoidine Y after a) 5 rnin and ( b ) 1 h. 5 rnin and (f) 1 h Chrysoidine R after (c) 5 rnin and (d) 1 h. L hrysoidine R2,6 after (e) ANALYST, NOVEMBER 1988, VOL. 113 25-mm diameter filter and allowed to dry in air. The filters were then processed according to the proposed air-sampling method and a correction was made for losses during process- ing. The filters examined and the percentage recoveries of the dye are given in Table 5 . Two types of FTFE membrane filters produced by Millipore demonstrated an expectedly high recovery of 90%. This level would be acceptable for a field analybcal method but the filters proved to be difficult to handle and were particularly prone to static effects. Gelman Science DM 800 membrane filters composed of a PVC - acrylonitrile copolymer gave an identical recovery and were easier to handle.Gelman Science DM 800 membrane filters were, therefore, adopted for subsequent experimental investigations and are the recom- mended collection filters for this application. The fact that the measured losses were identical for three types of filter suggests that even this acceptable loss is more likely to be attributable to degradation of the base when exposed to the air than to filter absorption effects. Losses during sampling The potentially large loss of Chrysoidine Y as the free base when in the finely divided state on an HPTLC plate prompted an investigation of similar potential losses from the filter during the collection of an air sample. Films of finely divided Chrysoidine Y base (5 ng) were deposited on 25-mm DM 800 filters by evaporation of an ethanolic solution.Similar films of the Chrysoidine Y hydrochloride were deposited by first adding 2% hydrochloric acid to the solution. The amount deposited, although sufficiently low and hence distributed thinly enough to provide a representative test, was neverthe- less sufficiently above the detection limit to produce statisti- cally significant results. Filters, so prepared, had air drawn through them at the recommended sampling rate of 2 1 min-1 for 30 min, before being transferred into Reactivials containing acidified ethanol. Similar batches of filters dosed with Chrysoidine Y base were exposed for various times without pumping air through them. A percentage recovery of chrysoidine was obtained using the described analytical procedure.A correc- tion for filter-absorption and processing losses was made. The correction factor was derived from analyses of dosed filters that had been transferred into Reactivials immediately after preparation. Losses due to exposure on the filter were calculated and are given in Table 6. Pumping air through the filter was found to make little difference to the result. The rate of loss of Chrysoidine Y base, corrected for losses due to absorption on the filter, etc. , was about 20% h-1, correspond- ing to 1 pg h-1. As expected, the corresponding figure for the hydrochloride was much less, about 10% h-1, corresponding to 0.6 pg h-1. The experiment was repeated with smaller amounts of base on the filter (1.25 and 2.5 pg) and lower rates of percentage loss were obtained.These results were less Table 5. Recovery of chrysoidine from membrane filters Filter Recovery, YO Comments DM 800 PVC - acrylonitrile VM1 PVC Mitex PTFE Fluoropore PTFE (Gelman-Hawksley) . . . . . . 90 Recommended filter (Gelman-Hawksley) . . . . . . 75 Slightly soluble in ethanol - HCl (Millipore) . . . . . . . . . . 90 Difficult to handle (Millipore) . . . . . . . . . . 90 Difficult to handle MF cellulose acetate - nitrate (Millipore) . . . . . . . . . . - Disintegrated in ethanol - HC1 - Celotatecelluloseacetate . . . . . . 75ANALYST, NOVEMBER 1988, VOL. 113 1623 Table 6. Recovery of Chrysoidine Y dosed on to a DM 800 filter and exposed to the air (corrected for losses due to filter absorption and analytical processing).For experimental details, see section on Losses During Sampling Recovery, % Free base Free base Hydrochloride exposed on exposed on exposed on filter to filter to filter to Exposure/min static air drawn air drawn air 0 100 100 100 15 94 30 91 87 94 45 60 80 78 88* - - - - - * Extrapolated value. reliable as the analyses were close to their limit of detection. However, they do demonstrate that a concentration-indepen- dent rate of loss of 1 pg h-1 does not continue to occur as this would have been readily detected. These results are significant as regards the sampling strategy. The detection of low concentrations of chrysoidine in air requires a sampling time that is as long as possible but still consistent with losses being at an acceptable level, i.e., 4 0 % averaged over the complete sampling time.The maximum recommended sampling times are 1 and 2 h for the free base and the hydrochloride, respectively. When the dye is of uncertain composition, a sampling time of 1 h should not be exceeded. Limits of detection The minimum amount of chrysoidine that can be detected on the HPTLC plate is 3 ng. Translated into air-sampling terms, 3 ng collected at an air-sampling rate of 2 1 min-1 for 60 min (i.e., 120 1 total volume) in a final sample volume of 200 @ correspond to limits of detection of 10 pg m-3 for single spotting, i.e. , for 0.5-pl spots; 5 pg m-3 for double spotting, i.e., for 1.0-p1 spots; and 2.5 pg m-3 for quadruple spotting, i.e., for 2.0-pl spots. The authors express their appreciation to Mr.G. M. Sole of the Department of Urology, Hereford County Hospital, for the loan of his MSc Thesis,” “The Use of Synthetic Dyes by Coarse Fishermen and the Risk of Urothelial Cancer,” and for his comments on the validity of the statements in this paper relating to his work on the mutagenicity of chrysoidine dyes. They also thank the following companies, both for their generosity in supplying samples of their dye products for investigation and for their helpful comments on the text of the paper: Clayton Aniline (Division of Ciba-Geigy); Croda Colours; Dykem North; ICI Colours and Fine Chemicals (Organics Division); and Williams (Division of Morton Thiokol). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21 22. 23. 24. 25. 26. 27. References Searle, C. E., and Teale, J., Lancet, 1982, 1, 564. Cartwright, R. A., Robinson, M. R. G., Glashan, R. W., Gray, B. K., Hamilton-Stewart, P., Cartwright, S. C., and Barham-Hall, D., Carcinogenesis, 1983,4, 111. Sole, G., and Sorahan, T., Lancet, 1985,1, 1477. Whitney, R., Hansard, Written Answers, 19 November, 1985, column 162. Garner, R. C., and Nutman, C. A., Mutat. Res., 1977, 44,9. Sole, G. M., Fielding, J. W. L., and Chipman, J. K., Eur. J. Surg. Onc., 1985,11,315. Sole, G. M., and Chipman, J. K., Carcinogenesis, 1986, 7 , 1921. WHO, “IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Man,” Volume 8, World Health Organ- ization, Geneva, 1975, p. 91. Albert, Z., and Orlowski, M., J . Natl. Cancer Inst., 1960, 25, 443. Searle, C. E., and Teale, J., Lancet, 1984,1,563. Chung, K., Mutat. Res., 1983, 114,269. Hix, C., Oglesby, L., MacNair, P. , Sieg, M., and Langenbach, R., Carcinogenesis, 1983,4,1401. Williams, M., Angling Times, 24th November, 1982,2. WHO, “IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Man,” Volume 27, World Health Organization, Geneva, 1982, p. 155. Weisenburger, E. K., Basic Life Sci., 1983,24,23. McCann, J., Choi, E., Yamasaki, E., and Ames, B. N., Proc. Natl. Acad. Sci. USA, 1975,72,5135. Lowry, L. K., Tolos, W. P., Boeniger, M. F., Nony, C. R., and Bowman, M. C., Toxicol. Lett., 1980,7,29. “Health Hazard Alert: Benzidine- , o-Tolidine- and 0-Diani- sidine-based Dyes,” DHHS (NIOSH) Publication No. 81-106, US Department of Labor, Occupational Safety and Health Administration and US Department of Health and Human Services, National Institute for Occupational Safety and Health, Cinchatti, 1980. Cartwright, R., Scand. J. Work. Environ. Health, 1982, 8, Supplement 1 , 79. Hinton, A. J., Lancet, 1984, 1, 1179. “Safe Handling of Dyestuffs in Colour Stores,” Health and Safety Executive, Wool and Textiles National Industry Group, North and West Yorkshire Area, Leeds, 1980. “Colour Index,” Society of Dyers and Colourists, Volumes 1-8, 1971-87. Hickinbottom, W. J., “Reactions of Organic Compounds,” Third Edition, Longmans, London, 1957, p. 506. Maximoff, A. T., US Pat. , 2053 095,1936. Pindar, A. G., and Tinsley, H. M., Analyst, 1984,109,1101. Foster, R. D., “An Analytical Technique Using High-perfor- mance Thin-layer Chromatography for the Screening of Chrysoidine Au, Dyes for o-Toluidine-based Components,” Internal Report IR/L/SP/87/05, Health and Safety Executive, London, 1987. Sole, G. M., MSc Thesis, University of London, 1986. Paper 8101 985A Received May 19th, 1988 Accepted June 21st, 1988