Annals of Occupational Hygiene Advance Access originally published online on August 26, 2005
Annals of Occupational Hygiene 2006 50(1):15-27; doi:10.1093/annhyg/mei036
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2005 British Occupational Hygiene Society Published by Oxford University Press
Original Article |
MDHS 25 Revisited; Development of MDHS 25/3, the Determination of Organic Isocyanates in Air
Health and Safety Laboratory, Harpur Hill, Buxton SK17 9JN, UK
Author to whom correspondence should be addressed. Tel: +44-1298-218-521; fax: +44-1298-218-570; e-mail: john.hsl.white{at}hsl.gov.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Isocyanates (NCO) are potent sensitizers and the largest cause of occupational asthma in the UK. Probably, MDHS 25/3 is the method most commonly used worldwide for the determination of organic NCO in air. This method uses an electrochemical (EC) detector to quantify oligo-NCO using derivatized NCO monomers as calibrants. This paper gives the results of a validation exercise using this method to quantify industrially used monomeric and oligo-NCO formulations. An expanded uncertainty of
56% was found for seven formulations spiked onto two sampler types at four spiking levels. No differences were found between the analytical results obtained from spiked filters or spiked impinger/filter samplers. Also presented is work on the validation of the EC/UV ratio approach used in MDHS 25/3 to identify NCO-derived peaks. A total of 58 industrially used non-monomeric NCO formulations were studied. These formulations gave 138 peaks that were identified using MDHS 25/3 as oligo-NCO MP derivatives. One of these peaks had an EC/UV ratio outside the range given in MDHS 25/3. Work describing the use of an internal standard to improve the variability of the EC detector is reported. Finally solutions to practical problems encountered during long-term sampling using an impinger are given.
Keywords: EC/UV ratio • electrochemical detector • isocyanates • MDHS 25/3 • validation
|
INTRODUCTION TO ISOCYANATE SAMPLING AND ANALYSIS |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Isocyanates (NCO) are highly reactive molecules widely used in industry, for example, in paints, polyurethane foams, plastics and adhesives. They are known respiratory sensitizers and are the major cause of occupational asthma in the UK (Jarvis et al., 1996
; NIOSH, 1996; HSE, 1997, 2001, 2005a; McAlinden, 2002; Paige, 2002; British Thoracic Society, 2003). Exposure to NCO can occur by inhalation and by contact (Reilly et al., 2002). The UK Health and Safety Executive (HSE) has set a long-term workplace exposure limit (WEL, 8 h time weighted average reference period) of 20 µg m–3 [total isocyanate (NCO) group] and a short-term limit (15 min) of 70 µg m–3 for workplace air (HSE, 2005b). This limit is for total isocyanate, i.e. monomeric and all oligo-NCO (also called poly-NCO, polymeric, oligomeric or pre-polymeric NCO). Sampling and analysis of airborne NCO is not easy. NCO occur in a variety of chemical forms, such as, monomers, oligomers, larger and more structurally complex polymers and mixtures of all these forms. Isocyanate oligomers and polymers are commonly used in industry as they are less volatile than the monomers and so pose less of a vapour hazard. NCO occur in a variety of physical forms, for example, vapours, aerosols and liquids. A sampling method that is suitable for one physical form of NCO is not automatically suitable for another. In the workplace other substances are also present in the air, such as water vapour, dust, amines and alcohols, depending on the product and process that is being used and these can interfere with analysis by liquid chromatography (LC). Standards of the oligo-NCO are not available, but under UK law, these species must be quantified to give a total NCO result.
Because of the reactive nature of NCO, analysis in the workplace is commonly carried out by trapping the NCO with a derivitization reagent to produce a stable derivative. Numerous samplers, derivitization methods and detectors have been proposed for NCO. Factors affecting sampling have been reviewed and comparisons of sampling methods have been carried out (Streicher et al., 1994, 1998; Levine et al., 1995; England et al., 2000; Rudzinski et al., 2001; Ekman et al., 2002; Hext et al., 2003). Guglya (2000) has put forth a useful review of methods for the determination of NCO in air.
|
INTRODUCTION TO MDHS 25/3 |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The Health and Safety Laboratory (HSL) has been involved in developing analytical methods for airborne NCO for over 20 years. The current UK method for isocyanate determination is the HSL method, Methods for the Determination of Hazardous Substances, Organic Isocyanates in Air, #25/3, MDHS 25/3 (HSE, 1999
; ISO, 2001). This method traps the isocyanate with 1-(2-methoxyphenyl)piperazine (MP) reagent to form the stable urea derivative by drawing workplace air through an impinger containing MP solution and/or through an MP impregnated filter. This method has been found to be suitable for the commonly occurring monoisocyanates and diisocyanates, i.e. 4,4'-methylenebis (phenyl isocyanate) (MDI), phenyl isocyanate (PI), toluene 2,6-diisocyanate and toluene 2,4-diisocyanate (TDI), 1,6-(diisocyanato)hexane (HDI), isophorone diisocyanate (IPDI), naphthyl diisocyanate (NDI), methylenebis (cyclohexyl isocyanate) (hydrogenated MDI) and butyl isocyanate, and oligo-NCO based on these monomers.
The sampling device recommended in MDHS 25/3 depends upon the physical form of the isocyanate. Isocyanate vapours can be sampled by an impregnated filter in a suitable sampling head (Institute of Medicine sampling head or similar). The recommendation in MDHS 25/3 is that isocyanate aerosols are most effectively sampled by a combination of impinger containing MP solution backed up by an MP impregnated filter. This recommendation is based on a study carried out by the International Isocyanate Institute (Hext et al., 2003; ISO, 2005). This work found that impregnated filters alone were unsuitable for collecting large particle sizes (>10 µm) because of the particles sticking to the filter cassette and local depletion of sampling reagent on the filter. However, it is widely acknowledged that impingers are difficult to use for personal monitoring, and several groups are currently researching the limitations and applicability of filter only sampling methods because they are easy to use (Bello et al., 2002; Ekman et al., 2002; Sennbro et al., 2004a,b; HSL, unpublished data). These groups are investigating such issues as local depletion of reagent at high concentration, the need for field desorption, and the effect of extended sampling times and any associated problems with self-reaction of the isocyanate. The Hext study (Hext et al., 2003) found that reagent filled impingers did not effectively sample small particles (<1 µm).
In MDHS 25/3 the impinger/filter sampling train is recommended as a ‘fail-safe’ method. The impinger/filter sampling train is suitable for the wide range of reacting isocyanate aerosols and aerosol particle size distributions that could possibly be encountered in the workplace. The high concentrations of reagent present in the impinger and filter mean that the sampler is unlikely to be overloaded; however, if very high concentrations of NCO are expected then higher MP concentrations can be used for the impinger solution and doped onto the filter. After sampling the urea derivative (NCO-MP) is analysed by LC with electrochemical (EC) and ultraviolet/visible (UV) detection.
It is a requirement under UK law to quantify the oligo-NCO as well as the monomeric ones. For some industrial uses, i.e. oligo-HDI paints, the monomeric fraction is very small (<1% of total NCO) so determination of the monomer only can lead to serious underestimation of the amount of NCO present. In addition, LC chromatograms of workplace isocyanate samples commonly contain many peaks. Not all of these are isocyanate derived; e.g. pigments, solvents, amines, catalysts, etc. will also be sampled. Correct identification of the isocyanate-derived peaks is necessary for accurate quantification. Analysis and characterization of an MP derivatized bulk sample for the formulation under study is useful for the identification of isocyanate-derived peaks and to develop the LC conditions (mobile phase composition) necessary to elute all oligo-NCO from the column. Typically, MDI formulations require a faster (more acetonitrile) mobile phase than HDI formulations and oligo-NCO require a faster mobile phase than monomeric NCO.
MDHS 25/3 identifies isocyanate-derived peaks by dividing the EC response of a given peak by the UV/vis response of the peak (ECPOLYMER/UVPOLYMER). This ratio is then divided by the same ratio calculated for the parent monomer (ECMONOMER/UVMONOYMER) and compared with an empirically derived range of values. This is the EC/UV ratio approach. It should be noted that the EC/UV ratio approach is an aide to identification only. In earlier (1987 and 1994) versions of MDHS 25, NCO peaks were identified by retention time and EC–UV response ratio. The use of this approach, the requirement to use the correct sampler and the need to quantify both oligomeric and monomeric species are not well understood by sections of the analytical community (Piney et al., 2002). Questions have been raised regarding the suitability of the response ratio approach for NCO pre-polymers (Key-Schwartz, 1995; Streicher et al., 1995). MDHS 25/3 answered these criticisms by introducing the use of UV spectral matching, comparison of samples with a derivatized bulk and the use of other confirmatory techniques, such as, FT-IR, mass spectrometry (MS) and titration. Work demonstrating the validity of the EC/UV ratio approach is reported below.
More modern techniques, e.g. MS, give more information and so better identification. However, mass spectrometers are expensive in comparison with the EC/UV combination, and are not yet ubiquitous in all laboratories, and so there is still a role for the EC/UV combination as a ‘poor man's mass spectrometer’. In addition, as discussed below, the EC detector is also able to quantify oligo-NCO using the available monomer MP derivatives; although, MS methods require a sample of the bulk NCO or a specific purified oligo-isocyanate. Modifications to the method, to enable determination of light NCO (methyl isocyanate and isocyanic acid) found during thermal degradation of polyurethanes have been described (Henriks-Eckerman et al., 2000, 2002; Zweigberk et al., 2002; HSL, 2003a). The use of MS with MDHS 25/3, or a similar method, has also been described (White et al., 1997; Östin et al., 2002; Gagne et al., 2003; HSL, 2003b; Vangronsveld and Mandel, 2003; Marand et al., 2004).
NCO for which a standard exists or can be prepared may be quantified using an UV/vis detector. However, for the majority of industrially used oligo-NCO no standards exist. This can pose a problem for quantification by MS-based methods if no bulk material or purified oligo-isocyanate derivative is available. MDHS 25/3 quantifies these compounds using the EC detector, which oxidizes the –OMe group on the MP derivatization reagent to give the EC signal. This group is common to all MP derivatized NCO and has been found to give the same response irrespective of the isocyanate it is bonded to (Warwick et al., 1981; Bagon et al., 1984; Schmidtke and Seifert, 1990; HSL, 1997). Therefore, the oligo-species may be calibrated using the corresponding NCO monomer, i.e. HDI monomer for oligo-HDI, MDI monomer for oligo-MDI. Therefore, standards of the various oligo-NCO are not required. The NCO monomer LC peak also acts as a useful retention time marker for MP derivatized bulks and samples. Work to extend the scope of validation for the use of EC for the quantification of oligo-NCO is reported here.
Sampling of NCO in workplace air can be difficult. In addition to the sampler considerations described above other problems are also commonly encountered. Many jobs using NCO require sampling for a short time only (<10 min); however, if long-term sampling is required then the impinger will require re-filling with toluene MP solution during the sampling period. This is inconvenient and time consuming. A ‘field solution’ is to use 30 ml of the MP reagent in toluene solution instead of the 10 ml stated in MDHS 25/3. Another solution is the use of the involatile solvent, butyl benzoate, with silica gel solid phase extraction (SPE) as described by Bello et al., 2002. These approaches were investigated as a means of simplifying extended sampling times.
During analysis of WASP QA (Workplace Analysis Scheme for Proficiency, isocyanate quality assurance scheme) filters and analysis of research and enforcement samples by HSL it has been noticed that the EC detector is more variable than the UV/vis detector. A method of reducing this variation is the use of an internal standard. The ideal internal standard is very similar to the analyte in order to correctly model the behaviour of the analyte throughout the work-up process. Deuterated internal standards are commonly used in MS but for the NCO under study they are either not available, i.e. NCO bulk formulations, or difficult to procure and expensive. As a cheaper, more readily available, alternative the reagent 1-(2-ethoxyphenyl)piperazine (EP) available from Aldrich as the hydrochloride salt was investigated as an internal standard for MP derivatized MDI (HSL, 2004). For analysis using MDHS 25/3, EP derivatives also have the advantage that the LC retention time of the NCO-EP is increased by 50%, compared with that of the NCO-MP, giving good peak resolution. The use of EP to derivatize NCO formulations (bulks) thereby giving an internal standard for each NCO-MP peak and EP doped filters as sampling media, coupled with MS analysis are also being investigated (HSL, unpublished data). Details of the preparation of the free EP from the EP HCl salt are given elsewhere (HSL, 2004).
In summary, the aim of the current paper is to publicize the work HSL has carried out on MDHS 25/3 since its publication in 1999. A second paper, covering the research being carried by HSL into LC/MS analysis and alternatives to the impinger/filter sampling train is in preparation.
The work reported in the current paper falls into four areas.
- Results obtained from the routine use of the EC/UV ratio for peak identification work are given. Conclusions regarding the validity of this approach for the identification of NCO derived peaks are given.
- The results of a method validation exercise using EC detection for seven monomeric and oligo-isocyanate workplace formulations are presented and an expanded uncertainty based on these results calculated. This level of validation data has not previously been published in the scientific literature for MDHS 25/3.
- Advice on the use of MDHS 25/3 and field strategies to avoid some commonly cited sampling difficulties is given.
- The use of internal standards as a means of improving EC sensitivity is investigated.
|
EXPERIMENTAL |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Reagents
Solvents used (acetonitrile, methanol, isopropanol, cyclohexane, diethyl ether and toluene) were of LC grade or better (from Rathburns, Walkerburn, UK or Fisher Scientific, Loughborough, Leics, UK). Buffers were prepared using water purified in-house using an Elgastat UHQ II system (Millipore Ltd, High Wycombe, Bucks, UK). Acetic acid, sodium acetate, butyl benzoate, bromophenol blue, 1 M hydrochloric acid (Convol), MP, EP, PI, TDI, NDI, HDI, MDI and oligo-HDI bulks [Aldrich product numbers 41,799-5 (Desmodur N 100) and 41,801-3 (Desmodur N 3300) and 41,800-5] were purchased from Aldrich Chemical Co (Gillingham, Dorset, UK). Diisocyanate and monoisocyanate MP and EP derivatives were prepared at the HSL from the relevant monomers and from samples submitted by HSE Inspectors during routine occupational hygiene monitoring and enforcement. The validation work used industrial isocyanate products kindly supplied by Mr Vangronsveld (Huntsman Polyurethanes) and Drs Pilger and Dislich (Bayer AG) and purchased from Aldrich Chemical Co.
Instrumentation
LC work was carried out on Waters Millennium (Waters Ltd, Watford, Herts, UK) and Hewlett-Packard HP1090 (HP, Stockport, Cheshire, UK) LC systems. Electrochemical (Coulochem II, ESA Analytical Ltd, Aylesbury, UK) and diode-array detectors (Waters and Hewlett-Packard) were used.
LC columns used were C18 e.g. 150 x 4.6 mm S3 ODS2 or 250 x 4.6 mm S3 ODS2 with an equivalent 30 x 4.6 mm guard column (Thames Chromatography, Windsor, UK and Jones Chromatography, Hengoed, UK).
A variety of sampling pumps were used (Gilian Instruments Ltd, SKC Ltd and Rotheroe and Mitchell Ltd) Impingers (SKC Ltd, Dorset, UK) and glass fibre filters (Fisher Scientific) were used as sampling media. Preparation of sampling equipment was as described in MDHS 25/3. IOM and Swinnex sampling heads were used for the filter only and impinger/filter sampling trains, respectively (SKC Ltd).
Samples were collected from workplaces as part of HSE's normal enforcement activities. These samples were analysed as described in MDHS 25/3. The results obtained by the identification procedure described in MDHS 25/3 (comparison with bulk derivative, UV spectral library matching, EC/UV ratio and other confirmatory techniques as required) were compared with using the EC/UV ratio only (as described in earlier versions of MDHS 25).
Silica gel SPE cartridges were Waters Sep-Pak plus (Waters Ltd).
Evaluation of EC/UV ratio for identification of NCO peaks
Isocyanate containing samples from a variety of industrial uses, i.e. car body shops, foam factories, fabric works, flooring application, aircraft workshops and manufacturers, component and packaging manufacturers, laminating and printing works and foundries were analysed during routine work carried out at HSL. The EC/UV ratios for these peaks were calculated and compared with the ranges given in MDHS 25/3 to assess the validity of the EC/UV ratio. Qualitative LC/MS/MS analysis of the peaks in the MP derivatized formulations was also carried out to assist identification.
The composition of the LC mobile phase was varied to give the best compromise between acceptable run time (up to 35 min) and peak separation. For slow eluting formulations (i.e. samples containing large, oligo-NCO) more acetonitrile was used in the mobile phase. The use of a less retentive column (i.e. C8) or elevated column temperature (60°C) can also speed up and improve oligo-NCO analysis (HSL, 1998). Typical LC and EC conditions for the analysis of an oligo-NCO containing formulations are mobile phase: isocratic, 60% acetonitrile, 40% sodium acetate buffer (5 g l–1), pH 6, flow rate 1 ml min–1; ESA Coulochem II: guard cell +1500 mV, E1(screen) + 450 mV, E2(analytical) + 800 mV, 1 µA scale; UV detection at 242 nm; LC column: suitable C18 analytical column and guard column.
Validation of the use of an EC detector for the quantification of oligo-NCO by MDHS 25/3
As discussed above, an advantage of MDHS 25/3 is the ability to quantify oligo-NCO by EC detection using the appropriate monomer MP derivative as calibrant. This approach was used in a validation exercise carried out as part of an ISO (International Standards Organization) working group project (HSL, 2004). The purpose of the ISO exercise was to validate the analytical part of the methods under study and the technically simple option of spiking underivatized bulks onto the samplers was decided upon. Further work to evaluate the sampling characteristics of the various methods for atmospheres of oligo-NCO aerosols is planned.
Compounds used for the validation exercise were Suprasec 2234, MDI based prepolymer (Huntsman Polyurethanes); Suprasec 5030, MDI based prepolymer (Huntsman Polyurethanes); Desmodur T80, TDI based mix, 2,4 and 2,6 monomers (Bayer AG); Desmodur N 3300, HDI based prepolymer (Bayer AG) and Desmodur H, monomeric HDI (Bayer AG). Also used, but not part of the ISO survey, was poly(toluene-2,4-diisocyanate), product #41, 805-6 (Aldrich).
Isocyanate concentrations of the bulk products were determined by titration (HSL, 1993). Each bulk was spiked at four levels, corresponding to 2x, 1x, 0.5x and 0.1x the UK short-term limit value (UK STEL) of 70 µg NCO m–3 (assuming a 15 l sample volume, 15 min at 1 l min–1). This equates to spiked amounts of 2.1, 1.05, 0.525 and 0.105 µg NCO per filter. Approximately 100 µl of each underivatized bulk, in toluene solution, was spiked onto the sampling devices and six samplers or more were spiked for each level and sampling device type. The sampling device used in MDHS 25/3 is dependent upon the physical form of the NCO as discussed above. Sampling flow rates suggested in MDHS 25/3 are 2 l min–1 for filter only sampling and 1 l min–1 for the impinger/filter combination.
For the filter samples, an impregnated filter, in an IOM sampling head, was spiked with a solution of the underivatized NCO bulk in toluene and air pulled through the filter at 2 l min–1 for 8 min to give an 15 l sample volume (to replicate the UK STEL). For the filter spikes, the spike was observed to spread out to cover most of the filter, suggesting that local depletion of the MP on the filter was minimized. The filter was then put into a few millilitres of MP solution and processed as described in MDHS 25/3.
For the impinger/filter samples, 10 ml of MP solution were placed in the impinger. The impinger was then spiked with a solution of the underivatized NCO bulk, the sampling head (Swinnex) containing the filter connected (after the impinger) and air pulled through the impinger/filter combination at 1 l min–1 for 15 min to give an 15 l sample volume (to replicate the UK STEL). The back-up filter was not spiked. For workplace samples taken by HSL using the impinger/filter combination the vast majority of isocyanate is found in the impinger with the back-up filter samples usually blank and the spiking regime employed in this work will replicate that finding. Sample work-up was as described in MDHS 25/3. The results for the impinger and filter are added together to give a total impinger/filter result.
The sampling pumps were measured before and after the sampling period, with a calibrated digital flow meter, to ensure they were delivering the required flow rate.
MP derivatized bulk NCO samples were used to identify the isocyanate-derived peaks, using UV/Vis library matching and the EC/UV ratio as described in MDHS 25/3. LC/MS/MS was also used to assist in identification.
After work-up the samples were analysed by LC/EC/UV and quantified by EC, using the NCO monomer MP derivative as calibrant, as described in MDHS 25/3. The NCO monomer MP standard solutions were made up in acetonitrile from stock solutions of 100 µg g–1 NCO/solvent. The standards covered a concentration range of 0.01–0.5 µg g–1 NCO/solvent and were stored in a freezer.
The settings of the ESA Coulochem II EC detector were varied according to the NCO bulk to give the best balance of EC sensitivity and EC stability as each NCO bulk/sampler type combination required an LC sequence of 16 h. The composition of the LC mobile phase was varied to give the best compromise between acceptable run time (up to 35 min per injection) and peak separation. For slower eluting mixes more acetonitrile was used in the mobile phase. Typical LC/EC/UV conditions were given above.
The uncertainty of the method was calculated following the ISO GUM approach (ISO, 1993). The calculations were based on the approach used by Karlsson and Skarping (ISO, 2004a,b) as agreed by the ISO committee looking at NCO in workplace air (ISO/TC146/SC2/WG4). This ‘uncertainty budget’ approach sums the contributions for each individual source of uncertainty for an analytical method. The combined uncertainty squared is the sum of the squares of the uncertainty in volume sampled plus the uncertainty in analyte determination plus the uncertainty in the blank determination plus the uncertainty for between laboratories. The full breakdown of these component uncertainties is given in the references quoted above (HSL, 2004, ISO, 2004a,b). This approach was also used to calculate the uncertainty for Desmodur N 3390, an HDI isocyanurate (cyclic HDI trimer) based formulation, using data obtained previously by HSL (HSL, 1997).
Extended sampling times
Impinger/filter combinations were filled with MP solution in toluene as described in MDHS 25/3 and air pulled through at 1 l min–1 for a variable time period. MP was dissolved in butyl benzoate (instead of toluene as used in MDHS 25/3), this solution spiked with MP derivatized oligo-HDI (Desmodur N 3300) and air pulled through at 1 l min–1 for a variable time period. The butyl benzoate was then removed by passing the sample through a silica gel SPE cartridge. The oligo-HDI MP derivative was then eluted from the SPE cartridge with methanol and analysed as described MDHS 25/3.
Different impinger designs may have different solvent evaporation rates. Therefore, two impinger designs were investigated, a ‘standard’ glass mini-impinger and a ‘spill-proof’ mini-impinger.
Use of an internal standard for EC analysis using MDHS 25/3
The EP derivative of 4,4'-methylenebis (phenyl isocyanate), (MDI-EP) was spiked onto filters that had been prepared for the WASP QA scheme (MDI-MP spiked onto MP impregnated glass fibre filters). These filters were then worked up and analysed as described in MDHS 25/3.
Similar experiments were carried out with filters spiked with oligo-HDI (isocyanurate) using 4,4'-MDI-MP as an internal standard and on workplace air samples (oligo-MDI sampled onto an impinger/filter combination) using NDI-MP as an internal standard. The specific isocyanate-EP or isocyanate-MP derivative used as an internal standard for any particular isocyanate formulation depends on factors such as interfering/co-eluting peaks and ease of identification. Any isocyanate MP/EP derivative with acceptable performance may be used as an internal standard.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Evaluation of EC/UV ratio for the identification of NCO peaks
A total of 58 industrially used formulations containing non-monomeric NCO were studied. The majority of isocyanate samples received were HDI-based and MDI-based with smaller numbers of TDI, IPDI and NDI formulations. The samples investigated were a mixture of bulks (NCO formulations) and air samples. The number of peaks seen in the chromatograms varied for each formulation. Typically there are >10 peaks per chromatogram of which only two or three are eventually identified as isocyanate derived, i.e. the majority of peaks in the chromatogram are not NCO. Commonly observed non-NCO peaks such as toluene and various xylenes have distinctive UV spectra and poor EC responses. Other non-NCO peaks seen in the chromatogram are caused by resin components (e.g. in two pack paints), pigments and other additives. Chromatograms of MP derivatized NCO formulations are given in Figs 1 and 2.
|
= 250 nm.
|
Chromatograms for the oligo-HDI formulations used in motor vehicle repair work gave the most complex chromatograms, both in terms of the number of non-NCO species present and in terms of the greater number of polymeric species found in the oligo-HDI formulations compared with those based on aromatic NCO. Chromatogram for TDI and oligo-MDI formulations were usually much simpler with fewer non-NCO peaks and fewer oligo-NCO peaks. TDI is usually present as a mix of 2,4 and 2,6 isomers.
The isocyanate-derived peaks are identified by UV library matching, LC/MS/MS and the EC/UV ratio approach as described above. The 58 formulations gave 138 peaks that were identified using MDHS 25/3 as oligo-NCO MP derivatives. One peak was identified in a bulk (formulation) derivative (by UV spectral library matching and LC/MS/MS) with an EC/UV ratio outside the range given in MDHS 25/3 (0.6–1.7). This peak was identified by LC/MS/MS as an oligo-HDI diisocyanurate and had an EC/UV ratio of 1.8, possibly because the UV peak used to calculate the ratio was small and hard to integrate accurately. The HDI diisocyanurate peak was a sizeable peak in the chromatogram and in order to provide a worst-case value for NCO concentration in these samples this peak would have been included in the list of calibrated NCO peaks using the EC/UV ratio. This peak illustrates the importance of getting good size UV peaks for the ratio calculation. There is a difference in response between the EC and the diode array detector, which usually means running a concentrated bulk MP derivative to get the UV peaks, diluting 100x to get the EC peaks and then working out the EC/UV ratio using this data.
The conclusion is that the EC/UV ratio approach, when used in conjunction with the other techniques described in MDHS 25/3, is an acceptable method of identifying NCO derived peaks.
Validation of the use of an EC detector for the quantification of oligo-NCO by MDHS 25/3
The results obtained from the spiking experiments are given in Table 1.
|
Generally the recoveries for the impinger/filter and filter samplers are in the acceptable range usually applied by HSL to these sampler types (>80%, <110%), for all the NCO formulations and both sampler types tested (HSL, 2002
). Six or seven replicates were carried out per sampling type. For the results where n < 6, this is because the sample was lost owing to an EC malfunction or spiking error. No results have been excluded as outliers. Most of the recoveries were 100% ± % relative standard deviation (%RSD) suggesting that the MP derivatives are being efficiently extracted. Significance testing (two tailed t-tests) on the data showed that none of the formulations, when all the spiking level and sampler types were grouped together, gave recoveries that were significantly different from 100% recovery (at significance level P = 0.01, n = 8). No significant difference was found when all the impinger/filter and the filter sampler results were compared (at significance level P = 0.01, n = 24). Significance testing (paired two tailed t-tests) of the sampler types (impinger/filter versus filter) for the formulation/spiking level pairs gave 19 results that were not significantly different and five that were (at significance level P = 0.01, 24 pairs tested, six replicates per sampler type). For the 24 sampler pairs tested, the five significantly different sampler pairs were Suprasec 5030 (2x, 1x, 0.5x LV spikes) and the 2x LV spikes (Suprasec 2234 and Desmodur T80). All five of the significantly different results had the impinger/filter less than the filter sampler. However, because of the small numbers of results involved these results were not significant at the 99% confidence level (at significance level P = 0.01, t-tests and sign tests, P = 0.125 for three out of three impinger/filter less than filter (NCO formulation and Suprasec 5030), P = 0.031 for five out of five significantly different results having impinger/filter less than filter). The conclusion is that for these spiking tests the analytical performance of the impinger/filter and filter only samplers are not significantly different.
The high %RSD values seen in Table 1 have been found to be usual for the EC detector as operated at HSL following MDHS 25/3. Typically, repeat injections of the same sample have a %RSD of 2%. For replicate samples, a %RSD of 5% is possible if the LC system and EC are operating well (clean LC column, new mobile phase, new EC cell, etc.). As the system ages the %RSD values increase as system performance declines. At HSL an ‘analytical variance’ of 25% is routinely quoted for results calculated using the EC detector. The above data set gives an ‘average standard deviation’ of 11 ± 6 (n = 48). During this project one of the EC cells ‘fouled’ and was replaced by a new cell. This was probably because of non-NCO components in the concentrated crude bulk MP derivatives used to characterize and identify the NCO formulations.
For the expanded uncertainties, pooling the data for the formulations tested gave a combined uncertainty for MDHS 25/3 of 28% (7 formulations, 4 levels, 2 sampler types). The corresponding expanded uncertainty (two times the combined uncertainty) was 56%. No significant differences were observed between the impinger/filter and filter only results (at a significance level of P = 0.01). A list of the data used to calculate the combined and expanded uncertainties is given in Table 2. The data in Table 2 show that the largest contribution to uncertainty in this method is the ‘uncertainty in analyte mass’ (%RSD). The work reported above did not use an internal standard, as the object of this validation work was to evaluate MDHS 25/3 as currently written. The use of an internal standard reduces the variability of quantification by EC (see Table 5). For the specific formulations, the range of expanded uncertainties is given in Table 3. The full results (by level, sampler type and formulation) are given elsewhere (HSL, 2004).
|
|
|
Extended period sampling
The results of this work are given in Table 4. From these results it can be seen that the effect of the impinger is quite small. The standard and spill-proof impingers would need re-filling after
1 h. The field solution of using more toluene MP solution extends sampling time 3-fold. These results are in agreement with those found by HSL field scientists who usually begin to consider re-filling the impingers after 30 min (from discussions with HSL field scientists). Butyl benzoate is hardly evaporating at all and could easily be used for an 8 h whole shift sample. A disadvantage of using butyl benzoate would be the requirement to remove the solvent by SPE.
|
The recovery of the oligo-HDI (Desmodur N 3300) after silica gel SPE (to remove the butyl benzoate) was found to be 94% (%RSD 8, n = 6, 1 µg NCO spiked).
In conclusion, if extended sampling periods were required, either using more toluene or using butyl benzoate with silica gel SPE would be recommended. The involatile butyl benzoate has also been found to be useful as a solvent in situations when the use of toluene is not recommended, i.e. air sampling in a spray-bake booth during the heated (80°C) re-circulation phase.
Use of an internal standard for EC analysis using MDHS 25/3
The results in Table 5 suggest that an internal standard greatly reduces the variability associated with EC detector. For EC detection with an internal standard the %RSD are comparable to those obtained for the UV detector. This effect was much more pronounced at the lower level samples. The variance of the higher level samples would not be expected to be so sensitive to detector type. Care must be taken when choosing an internal standard to make sure it does not interfere with the sample peaks, this is particularly important for oligo-NCO samples where many isocyanate peaks may be present e.g. NDI-MP was not suitable as an internal standard for HDI because of co-elution. The distinctive UV spectra of MDI-MP and NDI-MP are useful for identification purposes.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This paper describes work carried out to extend the scope of validation and improve the performance of MDHS 25/3. The work described was carried out in four areas:
- Validation of the EC/UV ratio for identification of NCO peaks.
- Validation of the EC detector for the quantification of oligo-NCO.
- Use of an internal standard to reduce EC variability.
- Solutions to problems encountered during long-term sampling using the impinger/filter combination.
A wide variety of NCO formulations and air samples from a variety of industrial uses were characterized using the EC/UV ratio approach and the results were compared with those obtained from LC/MS. Although LC/MS undoubtedly gives superior information for the identification of NCO-derived peaks the use of the EC/UV as a ‘poor man's MS’ was found to be valid. The EC/UV ratio approach does require a skilled analyst to perform it accurately.
Validation data for the analytical part of MDHS 25/3 have been presented. Acceptable recoveries and %RSD were found for the spiking experiments for all formulations and samplers under study. The EC detector has been shown to be able to accurately quantify oligo-NCO and validation data has been presented for workplace formulations. Acceptable recoveries from spiked filters and impinger/filter combinations have been found for six industrially used NCO formulations. No difference between the analytical performance of the impinger/filter sampler and the filter only sampler were found for any of the formulations studied. The sampling effectiveness of these samplers requires further study. The combined uncertainty of the method was calculated as 28%.
A means of improving the precision of the EC detector using various internal standards has been evaluated. The use of an internal standard greatly reduces the variability of the EC detector especially at the lower spiking levels.
Strategies for extended period impinger/filter sampling were suggested and evaluated. Several solutions to the problem of long-term sampling using an impinger have been described.
In summary, MDHS 25 has proved to be an excellent method for the determination of organic NCO in air and has been found to be capable of expansion to cover a variety of NCO analysis problems.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This work was funded by the Health and Safety Executive.
Received January 11, 2005; in final form June 23, 2005
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTION TO ISOCYANATE...INTRODUCTION TO MDHS 25/3EXPERIMENTALRESULTSCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Bagon DA, Warwick CJ, Brown RH. (1984) Evaluation of total isocyanate in air method using 1-(2-methoxyphenyl)piperazine. Am Ind Hyg Assoc J; 45: 39–43.[Medline]
Bello D, Streicher RP, Liu YC et al. (2002) Field comparison of impingers and treated filters for sampling of total aliphatic isocyanates with the MAP reagent. Am Ind Hyg Assoc J; 63: 790–6.
British Thoracic Society. (2003) British guideline on the management of asthma. Thorax; 58 (Suppl. 1): i1–94.[CrossRef]
Ekman J, Levin J-O, Lindahl R et al. (2002) Comparison of sampling methods for 1,6–hexamethylene diisocyanate (HDI) in a commercial spray box. Analyst; 127: 169–73.[Medline]
England E, Key-Schwartz R, Lesage J et al. (2000) Comparison of sampling methods for monomer and polyisocyanates of 1,6–hexamethylene diisocyanate during spray finishing operations. Appl Occup Environ Hyg; 15: 472–8.[CrossRef][Medline]
Gagne S, Lesage J, Ostiguy C, Tra H-V. (2003) Determination of unreacted 2,4 toluene diisocyanate (2,4 TDI) and toluene 2,6 diisocyanate (2,6 TDI) in foams at ultratrace level using HPLC-CIS-MS-MS. Analyst; 128: 1447–51.[Medline]
Guglya EB. (2000) Determination of isocyanates in air. J Anal Chem; 55: 508–29.
Henriks-Eckerman M-L, Välimaa J, Rosenberg C. (2000) Determination of airborne methyl isocyanate as dibutylamine or 1-(2-methoxyphenyl)piperazine derivatives by liquid and gas chromatography. Analyst; 125: 1949–54.[Medline]
Henriks-Eckerman M-L, Välimaa J, Rosenberg CK et al. (2002) Exposure to airborne isocyanates and other thermal degradation products at polyurethane processing workplaces. J Environ Monit; 4: 717–21.[CrossRef][Web of Science][Medline]
Hext PM, Booth K, Dharmarajan V et al. (2003) A comparison of the sampling efficiencies of a range of atmospheric samplers when collecting polymeric diphenylmethane diisocyanate (MDI) aerosols. Appl Occup Environ Hyg; 18: 346–57.[Medline]
HSE. (1997) Isocyanates, health hazards and precautionary measures, EH16. HSE, Sudbury, UK: Health and Safety Executive.
HSE. (1999) Methods for the determination of hazardous substances #25/3 organic isocyanates in air (MDHS 25/3). Sudbury, UK: Health and Safety Executive.
HSE. (2001) Health and safety statistics 2000/01. Sudbury, UK: HSE Books.
HSE. (2005b) Workplace exposure limits, EH40/2005. Sudbury, UK: HSE Books.
HSE. (2005a) http://www.hse.gov.uk/mvr/main-illhealth.htm, HSE website.
HSL. (1993) Preparation of samples for the HSE diisocyanate AQUA scheme. Report # IR/L/SP/93/01. UK: Health and Safety Executive.
HSL. (1997) Evaluation of MDHS 25/2, the determination of isocyanates in a reacting paint mix. HSL Report # IACS/97/19. UK: Health and Safety Executive.
HSL. (1998) Development of MDHS 25/2. Report # IACS/98/25. UK: Health and Safety Executive.
HSL. (2002) Determination of pesticides in solutions and residues in Tenax, GF/A, pad, glove and sock sampling devices, standard operating procedure, OMS-003. UK: Health and Safety Executive.
HSL. (2003a) The determination of isocyanates in the presence of amines. Report # OMS/2003/10. UK: Health and Safety Executive.
HSL. (2003b) Isocyanate Call-Off contract. Report # OMS/2003/12. UK: Health and Safety Executive.
HSL. (2004) International Standards Organisation—intercomparison of methods for organic isocyanates in air, results for the method MDHS 25/3, Report # OMS/2004/07. UK: Health and Safety Executive.
ISO. (1993) Guide to the expression of uncertainty in measurement. Geneva: International Standards Organisation.
ISO. (2001) Workplace air quality–determination of total isocyanate groups in air using 1-(2-methoxyphenyl)piperazine and liquid chromatography. ISO Method 16702:2001. Geneva: International Standards Organisation.
ISO. (2004a) Determination of organonitrogen compounds in air using liquid chromatography and mass spectrometry, part 1, isocyanates using dibutylamine derivatives. ISO/TC146/SC2/WG4/N334. Geneva: International Standards Organisation.
ISO. (2004b) Determination of organonitrogen compounds in air using liquid chromatography and mass spectrometry, part 2, amines and aminoisocyanates using dibutylamine and ethylchloroformate derivatives. ISO/TC146/SC2/WG4/N335. Geneva: International Standards Organisation.
ISO. (2005) Working Group 4 (Organic Vapours), Technical Committee 146 (Air Quality)/Science Committee 2 (Workplace Atmospheres), Isocyanates, a guide to sampling and analysis, ISO TC146/SC2/WG4, Document N371 (in preparation). Geneva: International Standard Organisation.
Jarvis J, Agius R, Sawyer L. (1996) Odds on for asthma. Chem Br; 32: 51–3.
Key-Schwartz R. (1995) Analytical problems encountered with NIOSH method 5521 for total isocyanates. Am Ind Hyg Assoc J; 56: 474–9.[Medline]
Levine SP, Hillig KJD, Dharmarajan V et al. (1995) Critical review of methods of sampling, analysis and monitoring for TDI and MDI. Am Ind Hyg Assoc J; 56: 581–9.
Marand A, Dahlin J, Karlsson D et al. (2004) Determination of technical grade isocyanates used in the production of polyurethane plastics. J Environ Monit; 6: 606–14.[Medline]
McAlinden JJ. (2002) Isocyanate exposures in the United Kingdom. In Lesage J, Graff ID and Danchik R, editors. Isocyanates: sampling, analysis and health effects. ASTM STP 1408. West Coshohocken, PA: American Society for Testing and Materials.
NIOSH. (1996) Preventing asthma and death from diisocyanate exposure. Publication No. 96-111. Available on http:/www.cdc.gov/niosh/asthma.html.
Östin A, Sundgren M, Ekman J et al. (2002) Analysis of isocyanates with LC/MS/MS. In Lesage J, Graff ID, Danchik R, editors. Isocyanates: sampling, analysis and health effects, ASTM STP 1408. West Coshohocken, PA: American Society for Testing and Materials.
Paige J (2002) Safety rep survey on asthma at work; no substitute for action on asthma. Available on http://www.tuc.org.uk/h and s/tuc-4227-f0.cfm (accessed on 19 January 2002).
Piney M, Helps J, White J, Wolfson H. (2002) Measuring isocyanate exposure—a warning and some guidance. Br Occup Hyg Soc Newslett, November, 16–18.
Reilly MJ, Rosenman KD, Peck KH. (2002) Work related asthma from exposure limits below the Michigan OSHA permissible exposure limit. In Lesage J, DeGraff I and Danchik R, editors. Isocyanates: sampling, analysis and health effects. West Coshohocken, PA: American Society for Testing and Materials.
Rudzinski WE, Yin J, England E et al. (2001) A comparison of solid sampler methods for the determination of hexamethylene-based isocyanates in spray painting operations. Am Ind Hyg Assoc J; 62: 246–50.
Sennbro CJ, Lindh CH, Östin A et al. (2004a) A survey of airborne isocyanate exposure in 13 swedish polyurethane industries. Ann Occup Hyg; 5: 405–14.
Sennbro CJ, Ekman J, Lindh CH et al. (2004b) Determination of isocyanates in air using 1,2-methoxyphenyl piperazine impregnated filters: long term sampling performance and field comparison with impingers with dibutylamine. Ann Occup Hyg; 5: 415–24.
Schmidtke F, Seifert B. (1990) A highly sensitive high performance liquid chromatography procedure for the determination of isocyanates in air. Fresenius J Anal Chem; 336: 647–54.[CrossRef]
Streicher RP, Kennedy ER, Loberau CD. (1994) Strategies for the simultaneous collection of vapours and aerosols with emphasis on isocyanate sampling. Analyst; 119: 89–97.[CrossRef][Medline]
Streicher RP, Arnold JE, Cooper CV et al. (1995) Investigation of the ability of MDHS 25 to determine urethane-bound isocyanate groups. Am Ind Hyg Assoc J; 56: 437–42.[Medline]
Streicher RP, Reh CM, Key-Schwartz R et al. (1998) NIOSH manual of analytical methods, Part K, determination of airborne isocyanate exposure. National Institute of Occupational Safety and Health, Washington DC, pp. 115–40.
Vangronsveld E, Mandel F. (2003) Workplace monitoring of isocyanates using ion trap liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom; 17: 1685–90.[Medline]
Warwick CJ, Bagon DA, Purnell CJ. (1981) Application of electrochemical detection to the measurement of free monomeric aromatic and aliphatic isocyanates in air by high performance liquid chromatography. Analyst; 106: 676–85.[CrossRef]
White J, Brown RH, Clench MR. (1997) Particle beam liquid chromatography/mass spectrometry analysis of hazardous agricultural and industrial chemicals. Rapid Commun Mass Spectrom; 11: 618–23.[CrossRef]
Zweigbergk P, Lindahl R, Östin A et al. (2002) Development of a diffusive sampling method for the determination of methyl isocyanate in air. J Environ Monit; 4: 663–6.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. H. E. Arts, W. H. de Jong, J. J. van Triel, M. A. Schijf, A. de Klerk, H. van Loveren, and C. F. Kuper The Respiratory Local Lymph Node Assay as a Tool to Study Respiratory Sensitizers Toxicol. Sci., December 1, 2008; 106(2): 423 - 434. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. White Reply Ann. Hyg., October 1, 2007; 51(7): 648 - 649. [Full Text] [PDF]
|
|
|
|
|
|
M. J. Walker Isocyanate Sampling and Analysis Ann. Hyg., October 1, 2007; 51(7): 645 - 646. [Full Text] [PDF]
|
|
|
|
|
|
K. S. CREELY, G. W. HUGHSON, J. COCKER, and K. JONES Assessing Isocyanate Exposures in Polyurethane Industry Sectors Using Biological and Air Monitoring Methods Ann. Hyg., August 1, 2006; 50(6): 609 - 621. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||