Detection of banned amines in leather21 March 2005
SUMMARY There currently exists legislation within Europe banning the use of dyestuffs based on certain carcinogenic amine compounds. Currently a test method exists that is used by the industry (CEN ISO TS 17234 based on DIN 53316). This is able to detect twenty of the compounds listed by the legislation but is not validated for the recently added 2-methoxyaniline and is unsuitable for analysis of -Aminoazobenzene. The method is also known to have problems related to the reproducibility of the analysis both within and between laboratories. Work has been carried out as part of a European Commission funded project to look into the issues surrounding the test method. Within the project there are three key objectives. These are: * To carry out a feasibility study for the development of a certified reference material for banned azo dyes in leather * To develop a method suitable for certifying any proposed reference material * To develop a method suitable for routine testing in industry. Research to develop a method for certifying the reference material has resulted in a procedure based on supercritical fluid extraction for degreasing the leather, followed by sodium dithionite microwave assisted extraction in order to reduce the dyestuffs present. The samples obtained are then cleaned up using solid phase extraction before analysis using High Performance Liquid Chromatography with Diode Array Detection (HPLC-DAD). The technique of standard addition is also used to improve recovery rates. In order to certify a reference material it is necessary to have a method which is very accurate and precise. The method is not planned for general use and, therefore, has incorporated techniques that may not be suitable for all laboratories. A routine test method has also been developed. This method was designed considering that it should be suitable for use in laboratories within the leather industry on a routine basis. It should be relatively straight forward to carry out, providing results with good reproducibility. It is not critical, however, to obtain 100% recovery rates, provided the results are consistent. Also of importance is the cost effectiveness of the method along with the speed in which results can be obtained. A method has been developed incorporating ultrasound based degreasing that can be carried out on the day of analysis. This is followed by an optimised sodium dithionite reduction and solid phase extraction for clean-up. Both HPLC-DAD and Gas Chromatography with Mass Spectroscopy (GCMS) have been evaluated for analysis. During development the methods have been assessed in inter-laboratory trials. Legislation Update Azo dyes, extensively used for colouring a variety of consumer goods such as leather, clothes, and toys, can under certain conditions be reduced to form proven or suspected carcinogenic aromatic amines. As a consequence, the European Parliament recently accepted the 19th amendment of the Council Directive 76/769/EEC relating to restrictions on the marketing and use of certain dangerous substances and preparations (azo colorants) . This amendment, largely based on former German legislation, represents an attempt to control potentially harmful azo dyes that are used in specific textile and leather products that come into direct and prolonged contact with human skin or oral cavity. The azo dyes affected by this legislation are those, which after reduction may form one or more of the 22 aromatic amines listed in the Directive, in detectable concentrations (ie > 30 ppm). These are listed in Table 1 below. Introduction to the project Harmonised testing methods are necessary for the application of the recently introduced EU Directive. Concerning leather, the appropriate analytical procedure to be used is currently the European Standard CEN ISO/TS 17234, recently approved by the European Committee for Standardisation (CEN). This method is very similar to the widely used German method DIN 53316, which is known to suffer from major drawbacks such as low accuracy and precision. Furthermore, sample preparation employs conventional liquid-liquid extraction, which is time-consuming and requires large quantities of hazardous organic solvents. Consequently, two EU funded projects have been initiated since 1998: SMT4-CT97-2194 (1998-2000) and G6RD-CT-2001-00600620 (2002-2004, AALARM). The primary aim of these projects has been the development of test methods while the AALARM project (Aromatic Amines in Leather due to Azo-dyes Reference Material) is also investigating the feasibility of developing a certified reference material that can be used by the leather industry to validate testing methods. As a result of this, test methods are required which should improve on the performance of the current industry standard (CEN ISO TS 17234:2003 which is based on DIN 53316). Within the project, methods have been developed as follows: * A routine test method that allows quantification and identification of the amines restricted as a result of the EC legislation (referred to in this document as restricted amines) in a simple and cost effective way. Reproducibility is important, however recovery rates may be lower than 100% provided they are consistent. * A certification method that allows accurate and precise quantification of restricted amines in any certified reference material produced. This method will not be carried out by test laboratories on a regular basis and so may use advanced techniques such as Supercritical Fluid Extraction (SFE) or Microwave Assisted Extraction (MAE). Sample preparation In order to provide a suitable material for a potential Certified Reference Material (CRM) it is necessary to determine the optimum substrates for use, along with the selection of dyestuffs to be added. It is not possible to include all 22 restricted amines within the samples (as many are found in dyestuffs unsuitable for dyeing leather) and so the dyestuffs were selected to represent amines that are both relevant to the industry and cover the widest range of physical and chemical properties possible. The leather types selected were: * Bovine chrome tanned crust * Bovine chrome tanned crust treated with acrylate type waterproofing agent * Ovine chrome tanned crust treated with silicone type waterproofing agent * Caprine chrome tanned crust - typical of glove leather * Bovine mineral free tanned The restricted amines chosen to include in the leathers are: * o-toluidine * 3,3-dimethylbenzidine * 4-chloroaniline * benzidine * 3,3-Dimethoxybenzidine * 3,3-Dichlorobenzidine * 2-Methoxyaniline * 4 aminodiphenyl The leathers were prepared to contain in the region of 40-60ppm of the restricted amines, with each leather containing different combinations of the amines to represent as many possible interferences and problems as possible. 1. Development of a routine test method A test method has been developed based on a standard sodium dithionite reduction and clean-up using relatively simple laboratory apparatus. Each of the stages of analysis have been considered in order to optimise and improve them. The procedure developed is as summarised as follows: * Within the current CEN method the sample is degreased to remove excess fats that may interfere with the analysis and affect the wetting of the leather fibres. This process is carried out using hexane in an ultrasound bath. Hexane is not miscible with aqueous solutions and, as a result, it is necessary to ensure that the samples are totally dry before commencing the reduction stage of the method. This drying time is typically overnight. Within this project it has been determined that suitable levels of degreasing can be achieved using an initial hexane step followed by acetone. Dye removal from the sample is not affected and it can be immediately subject to reduction after decanting the excess acetone. * The various reduction parameters were evaluated (temperature, pH, concentration) and as a result the process was optimised to improve recovery rates and minimise the potential for false positive results (4-aminobiphenyl can in some cases be detected in dyes where it was not used during manufacture). The reductive cleavage of the leather is, therefore, carried out using 800mg/g sodium dithionite in pH 7 buffer at 70oC. * Within the CEN method, sample clean up takes place using a liquid/liquid extraction and large quantities of the solvent, t-butyl methyl ether. Within this work a technique was developed using solid phase extraction (SPE) to clean up the extract which requires only small volumes of the solvent methanol. * One key issue related to the analysis of banned amines in leather is the issue of recovery rates. Matrix effects occur resulting in the amines liberated during reduction, becoming fixed to the leather. To try and maximise the recovery of banned amines, the leather fibres were washed twice using dilute hydrochloric acid with combination of the extracts to allow clean-up on a single SPE column. * Once the samples have been prepared it is necessary to analyse for the restricted amines. Analysis has been carried out predominantly using High Performance Liquid Chromatography with Diode Array Detection (HPLC - DAD). Some partners within the project have also used Gas Chromatography with Mass Selective Detection (GC-MS). This will be discussed in more detail along with the inter-laboratory trials. To illustrate the reproducibility of the test method developed, one of the trial leather samples was analysed and the results are displayed in Table 2. The leather was analysed in triplicate with each sample being injected twice on the HPLC. It is clear from these results that the method is reproducible especially considering the low levels of amines present. (The legislation requires that the detection limit of the method is 30mg/kg in the leather). Only one amine (o-toluidine) showed high standard deviations and this can be attributed to an interfering compound in the sample. This method was, therefore, considered suitable to be evaluated in the inter-laboratory trials. Development of a Certification Method When developing a CRM it is also necessary to have analytical procedures that are suitable to certify the materials, ie identify and quantify all restricted dyes present to high accuracy (recovery rates as close to 100% as possible) and reproducibility. In order to meet this requirement the procedures developed may be quite complex and expensive to carry out. This is in contrast to a method that is required for routine use, whereby precision is important. However, recovery rates may be less than 100% provided they are known and are consistent. A former project resulted in an analytical procedure, based on modern extraction techniques (ie supercritical fluid extraction and microwave-assisted extraction) and external calibration, which gives much higher recoveries in comparison to the German DIN 53316 methodology (10-28%). Accordingly, initial work in the current AALARM project has been focused towards further optimisation of certain parts of this procedure. Even though minor improvements have been accomplished, it does not seem possible to achieve satisfactory accuracy, especially for those azo dyes, which form amines with two phenyl rings (i.e. benzidine, 3,3´-dimethylbenzidine, 3,3´-dimethoxybenzidine, and 3,3´-dichlorobenzidine) after reductive cleavage, using this approach. Since the reason for the low recoveries probably is due to so called matrix effects, the application of standard addition methodology has been evaluated to improve the accuracy. 1. Method of standard addition The term matrix denotes sample components that are different from the analyte(s). Matrices can affect the analyte response essentially by two different mechanisms. Either matrix component(s), present at the time of measuring, or some indispensable sample treatment (eg any reduction or extraction to which the sample is subjected prior to the measurement) can result in matrix effects. To compensate for such effects, the method of standard addition can be applied. This method comprises the addition of defined, varying quantities of the analyte to equal amounts of the sample followed by treatment according to the relevant analytical procedure. The obtained results are then plotted against the amount of added analyte. Extrapolation of the regression line to zero gives the quantity of analyte in the sample, as illustrated in Figure 1. 2. Comparative study between external calibration and standard addition To take into account the influence of matrix effects, a method based on external calibration has been applied to bovine leather samples and the results have been compared with those obtained using a second method of quantification: the standard addition method. The standard addition has been performed using both the relevant azo dyes as well as the corresponding carcinogenic amine as spiking agents. This enables investigations regarding the efficiency of different parts of the analytical procedure. Whereas spiking with the amines only compensates for inefficiency in the extraction step, spiking with azo dyes compensates for inefficiency in both the reduction and the extraction step. These experiments have been carried out using six different leather samples. These samples are dyed with the restricted azo dyes listed in Table 3 (along with the corresponding amine released during reduction). Accordingly, the leather samples have been spiked separately with the spiking agents before the reduction step and subsequently treated according to the analytical procedure illustrated in Figure 2. The analytical procedure based on quantification by means of external calibration is almost identical to the one described in Figure 2 with the exception that the actual spiking step has been omitted. The results obtained using the different approaches can be seen in Table 4. As can be seen in Table 4, the approaches based on standard addition methodology result in higher recovery rates than obtained when using external calibration, especially when the azo dyes themselves are used as spiking agents. The results obtained using standard addition of the amines reveal that some losses occur during the extraction step since these are higher than those obtained using the external calibration procedure. However, since the improvements are accentuated when azo dyes are used for standard addition, it appears that the losses due to side reactions or incomplete reduction are more severe. Nevertheless, there may be possibilities for improvement of the methodology since certain amines (eg 3,3´-dichlorobenzidine, benzidine, and p-chloroaniline) still show unsatisfactory results. For example, the large recovery (167%) obtained for p-chloroaniline could be due to evaporation losses during the analytical procedure because of the volatility of this rather small compound. Acidifying the extraction medium and, thus, making the compound less volatile, in the analytical procedure could probably eliminate these losses and give recoveries of p-chloroaniline closer to 100%. In the case of benzidine and 3,3´-dichlorobenzidine, these amines show recoveries lower than 100% after standard addition of the azo dyes. This might be due to the fact that the spiked fraction does not behave exactly like the native molecules, since aged molecules might be more strongly bound to the leather. Inter-laboratory Trials When developing new test methods, it is necessary to ensure that they are reproducible not only within a laboratory, but between laboratories also. Therefore, a series of inter-laboratory trials have been carried out to assess the performance of the methods. The methods developed have two key stages: sample preparation and chromatographic analysis. It was decided to evaluate the chromatography as a separate parameter due to reports from the previous project suggesting that this may be a major source of variability. A test solution was prepared for each of the 13 laboratories participating in the trials which contained a mixture of the restricted amines of interest to the leather industry. These were analysed by HPLC and the concentration calculated based on calibration curves prepared by the individual laboratories. For each amine the mean and 95% confidence limits obtained for each of the amines were determined for each laboratory. Also evaluated was the mean of the means obtained. During an initial trial it was highlighted that considerable variability was occurring due to the chromatography conditions being used by the partners. For this reason many of the chromatography parameters were defined for use within the trial (unlike the current CEN method, where the chromatography conditions are recommendations only). These included specifying the buffer type and concentration, the column type and the need to carry out injection using an overfilled injection loop. In order to assess the effect of the improvements in the chromatography, the test solution results were evaluated. This showed that a core group of the partners were obtaining good results. In many cases the relative standard deviations were low. This resulted, however, in few laboratories having the target amine concentration within their error range. Typically, a relative standard deviation (rsd) of 5% would be expected from laboratories carrying out this type of analysis. Table 5 illustrates the number of laboratories achieving the target value within the reported relative standard deviation or a 5% relative standard deviation. From these results it is clear that there is considerable variation between the laboratories with regards to the analysis of a test solution and the effect is larger for some of the amines. At the time of preparation of this paper, the final test methods described above are being assessed in a final validation inter-laboratory trial. Conclusion The research described in this paper is a summary of work that has taken place over the past two years resulting in the development of two test methods for the determination of restricted azo-dyes in leather. The research carried out to date indicates that the use of standard addition techniques is the most appropriate mechanism for dealing with the profound matrix effects that influence the analysis of restricted amines. This technique although too complex and time consuming for use on a routine basis, shows potential for development into a certification method for a CRM. A separate method has been developed for routine use in the laboratory. While standard addition is not feasible to account for matrix effects, it is not an issue for a routine method. The method developed shows good reproducibility during the development stage and once the data is obtained from the inter-laboratory trials, it is intended that this be submitted to the IULTCS IUC Commission, ISO and the CEN committee TC 289 - Leather for consideration as an International and European standard. Acknowledgement The authors would like to acknowledge the European Commission for financial support (project No. G6RD-CT-2001-00620, AALARM). Thanks also to all the partners involved in the AALARM project.