Analyst, August, 1967, Vol. 92, pp. 513-519 513 A Field Method for Determining 4,4'-Di-isocyanatodiphenylme thane in Air BY D. A. REILLY (Imperial Chemical Industries Limited, Dyestuffs Division, Hexagon House, Blackley, Manchester 9) A method has been developed for determining up to 0.04 p.p.m. v/v of di-isocyanatodiphenylmethane (MDI) in a 5-litre sample of air. The test atmosphere is drawn through 0.4 M hydrochloric acid, which hydrolyses the isocyanate to the corresponding amine. This is then diazotised and coupled with 3-hydroxy-2-naphthanilide. After acidification of the solution the resulting pink compound is extracted into chloroform and compared visually with inorganic colour standard solutions. The preparation of three standards is described, corresponding to 0.01, 0.02, 0.04 p.p.m.of MDI, the recommended threshold limit value being 0-02 p.p.m. v/v. There is no interference from the tertiary-amine catalysts that are used with MDI in the production of polyurethanes. ~,~'-DI-ISOCYANATODIPHENYLMETHANE (MDI) is used for the production of rigid and semi- rigid polyurethane foams that have wide applications for thermal insulation, and as packing and building materials. Because of its low vapour pressure it is unusual under normal conditions of use for atmospheric concen- trations of MDI vapour to exceed this value. However, if equipment for dispensing rigid foam in bulk is incorrectly operated, or if the ventilation on spraying equipment is inadequate, then droplets of material containing free MDI can escape into the general atmosphere, where they constitute a hazard.In an earlier method2 the test atmosphere was drawn through a solution of sodium acetate and hydrochloric acid. The hydrolysis product was coupled with 9-nitrodiazobenzene solution to give a yellow azo compound that was extracted into chloroform and matched visually against inorganic colour standards. This suffered from two disadvantages. One, that the standard colours were rather pale and difficult to match, particularly in artificial light, and the other, that interfering orange - red colours were given by some of the tertiary amines that were used as catalysts in certain applications. The method described in this paper avoids both of these disadvantages; the test atmos- phere is drawn through dilute aqueous hydrochloric acid, and MDI is hydrolysed to 4,4'-di- aminodiphenylmethane.This is then diazotised by treatment with sodium nitrite, and the excess of nitrous acid is destroyed with sulphamic acid. The resulting solution is coupled with an alkaline solution of 3-hydroxy-2-naphthanilide to give a pinkish-orange azo com- pound. After acidification, this is extracted into chloroform and the solhtion compared visually with inorganic colour standard solutions. These solutions are deeper in colour and much more easily seen than those described.2 There is no interference from any of the common tertiary-amine catalysts. It is also a constituent of surface-coating lacquers. The recommended threshold limit value for MDIl is 0.02 p.p.m. v/v. METHOD APPARATUS- Sampling pump-Capable of drawing air at 1 litre per minute through the abscrber.AZZ-glass absorber-Of the type shown in Fig. 1. Flat-bottomed specimen tubes-2 x 8 inches (see Notes). Glass-stoppered separating funnel-25-ml capacity, with its stem cut off 2 inch belo v the tap.514 REILLY: A FIELD METHOD FOR DETERMINING [Analyst, Vol. 92 REAGENTS- Dilute hydrochloric acid-Dilute 40 ml of concentrated hydrochloric acid (sp.gr. at 20" C, 1-18> to 1 litre with water. Sodium nitrite solution, 0.7 per cent. w l v . Sulphamic acid solution, 10 per cent. w/v. Sodium hydroxide solution, N. Brenthol A S suspension-Weigh 0.5 g of Brenthol AS (3-hydroxy-2-naphthanilide) Shake vigorously Dilute s ~ d ~ h u r i c acid-Dilute 16.5 ml of concentrated sulphuric acid (sp.gr. at 20" C, 1.84) Chloroform-Analytical-reagent or B .P.grade. Cobalt(l1) chloride colorimetric solution3-Prepare a solution in 0.25 M hydrochloric acid containing 59.5 mg per ml of analytical-reagent grade cobalt (11) chloride, CoCI2.6H,O (purity not less than 97.5 per cent.). Iron (Ill) chloride colorimetvzc solutio~z-Prepare a solution in 0-25 IVI hydrochloric acid containing 45.05 mg of iron(II1) chloride per ml. Colour standavd solutions-Mix the volumes of the cobalt (11) chloride and iron(II1) chloride colorimetric solutions specified in Table I, and dilute to 500 ml with 0.25 M hydro- chloric acid. These solutions have the same quality and depth of shade as the test solutions (containing the equivalent amounts of 4,4'-di-isocyanatodiphenylmethane) prepared as described below.(obtainable from I.C.I. Ltd.) into a bottle, and add 50ml of water. immediately before use. to 100ml with water. Store in clean glass bottles with tightly fitting glass stoppers. TABLE I COMPOSITION OF COLOUR STANDARD SOLUTIONS Cobalt( 11) chloride solution, Iron( 111) chloride solution, MDI, ml ml p.p.m. 13.0 15.0 0.0 1 23.0 10.0 0.02 42.0 9.0 0.04 PROCEDURE- Draw 5.0 litres of the test atmosphere, at a rate of about 1 litre per minute, through the absorber containing 3-0 ml of dilute hydrochloric acid. Detach the air-inlet tube, allowing any liquid in it to drain into the body of the absorber and remove the last drops by blowing gently. Add 0-10 ml (3 drops) of sodium nitrite solution, stopper with a B19 glass stopper, mix the contents by inverting the tube and allow it to stand for 1 minute.Add 0.2 ml (6 drops) of sulphamic acid solution, making sure that the joint is wetted by the reagent. Stopper, shake well, and allow the tube to stand for 1 minute. Pour the contents into the 25-ml separating funnel containing 2 ml of N sodium hydroxide and 0-2 ml (6 drops) of Brenthol AS suspension, shaking the bottle containing the latter immediately before use. Stopper, mix well, and allow the mixture to stand for 1 minute. Add 1 ml of dilute sulphuric acid and 3.0ml of chloroform. Stopper the funnel, shake it vigorously for 1 minute, allow the chloroform layer to separate and run it into a 2 x +-inch flat-bottomed specimen-tube. Fill similar tubes to the same depth with the three colour standard solutions. Compare the chloroform extract visually with the three standards in turn by viewing vertically through the depth of the liquids when the tubes are held side by side above a sheet of white paper in a good light, preferably daylight.Assign the appropriate MDI concentration to the extract, interpolating a value between two of the standards, if necessary. NOTES- Flat-bottomed tubes have been recommended for comparison of the colour, because with round-bottomed test-tubes the appearance of the bottoms of similar tubes containing chloroform and water is different and can cause difficulty in colour matching. An alternative means of comparing the colours obtained is to use the Lovibond Special Purposes Comparator. A disc containing glass standards, equivalent to the above colour standard solutions, for use with this instrumcnt is available from Tintometer Scales Limited, Waterloo Road, Salisbury, Wiltshire.A special narrow 20-mm cell is also needed and is obtainable from the same 1. 2 .August, 19671 4,4’-DI-ISOCYANA7rODIPHENYLMETHANE IN AIR 515 CHOICE OF ABSORPTION SOLUTION AND OF STANDARD MATEKIAL- An established absorption solution for isocyanates is 0.4 M aqueous hydrochloric acid. Mixtures of dilute hydrochloric acid and acetic acid have also been used,4 but it has been found that the addition of acetic acid to hydrochloric acid does not improve the degree of trapping of MDI (see “Absorption of MDI from the Atmosphere” below) or give an increased colour yield from a given weight of MDI, and 0-4 M hydrochloric acid has therefore been adopted.Under the conditions used, MDI and 4,4’-diaminodiphenylmethane (DADPM) give equivalent colour yields (see “Comparison of Calibration Graphs obtained from MDI and DADPM” below) and, in establishing the preferred conditions of diazotisation, coupling and colour matching, DADPM has been used as standard because it is much more easily dissolved than MDI in dilute hydrochloric acid. SCALE OF EXPERIMENTS- To obtain the maximum intensity of colour in the final chloroform extract, the volumes of reagents were kept as small as possible, i.e., 3 ml of 0.4 M hydrochloric acid and 3 ml of chloroform. This volume of chloroform was sufficient for visual colour matching but in- sufficient for the spectrophotometric measurements that were used to establish the quantitative basis of the method.Therefore, the scale was increased &fold to 15 ml of 0-4 M hydrochloric acid and 15 ml of chloroform for many of the tests. CHOICE OF COUPLING AGENT- Four coupling agents were tried, vix., N-l-naphthylethylenediamine dihydrochloride, 2-naphtho1-3,6-disulphonic acid, sodium salt (R-salt), 3-hydroxy-2-naphthoic acid and 3-hydroxy-2-naphthanilide (Brenthol AS). The last two have the advantage that they produce dyestuffs that can be extracted into chloroform. This is important because field tests showed that the coupled test solution can be turbid owing to spray-borne polyurethane foam or lacquer, and this makes colour matching of the aqueous solution difficult. When the dyestuff was extracted into chloroform the turbidity was left in the aqueous layer.3-Hydroxy-2-naphthoic acid was about 25 per cent. more sensitive than Brenthol AS but produced a fairly strong yellow colour in the chloroform extract, which, although it did not interfere with spectrophotometric measurement of the pink colour due to MDI, made visual matching rather difficult. 3-Hydroxy-2-naphthanilide (Brenthol AS) gave only a pale yellow blank and was therefore preferred as a coupling agent. For most of the subsequent investigation the following conditions were used, with results as shown in Table 11. TABLE I1 EFFECT OF CHANGING METHOD OF ADDING 3-HYDROXY-2-NAPHTHANILIDE Method of adding 3-hydroxy-2-naphthanilide- Solution Suspension A I -l 7 DADPM, pg. . . . . . . . . 3-8 5.5 7.6 10.3 15.2 7.5 12-5 17.5 Optical density a t 540 mp, with 4-cm cells 0.197 0.305 0.394 0.570 0.800 0-342 0.677 0.905 Fifteen millilitres of a solution containing 5 to 30 pg of DADPM in 0.4 M hydrochloric acid were treated with 0-5 ml of 0.5 N sodium nitrite solution, After 1 minute, 1.0 ml of sulphamic acid solution was added, and after a further minute the solution was poured into a separating funnel containing 10 ml of a 0.1 per cent.solution of Brenthol AS in N sodium hydroxide. After 1 minute, 5 ml of dilute sulphuric acid and 15 ml of chloroform were added and the mixture was shaken for 1 minute. The chloroform extract was filtered directly into a dry 4-cm spectrophotometer cell and the optical density measured at 540mp, with a reagent blank as the comparison solution. 0.310 0.560 TIMES OF DI,4ZOTISATION AND COUPLING WITH 3-HYUROXY-2-NAPHTHANILIDE- Tests showed that l-minute periods were sufficient for the diazotisation and coupling stages.EFFICIENCY OF SINGLE EXTRACTION- This was checked by four tests on 5.5-pg amounts of DADPM. In two of the tests, single extractions with 15.0 ml of chloroform were carried out; in the other tests, three516 REILLY: A FIELD METHOD FOR DETERMINING [Aaalyst, Vol. 92 separate extractions with 5.0 ml of chloroform were carried out. The combined extracts were diluted to 15.0 ml, filtered and their optical densities measured with the following results- Number of extractions per 15-0 ml of total extract . . . . 1 1 3 3 These results show that about 95 per cent. of the colour is removed by a single extraction. COMPARISON OF CALIBRATION GRAPHS OBTAINED FROM MDI AND DADPM- As MDI is not easily soluble in dilute aqueous hydrochloric acid, a solution was prepared in dry dioxan, the solvent having been dried by refluxing with sodium, followed by distilla- tion.The solution used contained 52.5 pg of distilled MDI per ml. Aliquots of this solution were diluted to 15.0 ml with 0.4 M hydrochloric acid, diazotised without delay and then coupled with 3-hydroxy-2-naphthanilide, as described under “Choice of Coupling Agent” above ; similar tests were also carried out on a solution of DADPM in dry dioxan containing 35-5 pg per ml, the results of which are shown in Tables I11 and IV. They show that DADPM and equivalent weights of MDI give closely similar optical density levels, indicating that MDI is hydrolysed rapidly and completely to DADPM by 0.4 M hydrochloric acid at room temperature.The level of optical densities obtained from DADPM in the presence of dioxan was lower than that obtained in its absence (see Table 11). Optical density a t 540 mp, with 4-cm cells . . . . . . 0.305 0.310 0.315 0.325 TABLE I11 CALIBRATION POINTS FOR DADPM IN THE PRESENCE OF DIOXAN DADPM solution, ml . . . . . . . . . . 0.1 0.2 0.3 0.4 0.5 DADPM, pg. . . . . . . . . . . . . 3.55 7.10 10.65 14.2 17.75 Optical density a t 540 mp, with 4-cm cells . . . . 0.140 0.275 0.395 0.500 0.600 TABLE IV CALIBRATION POINTS FOR MDI IX THE PRESENCE OF DIOXAN MDI solution, ml . . . . . . . . . . . . . . 0.1 0.2 0-3 0.4 MDI, pg . . . . . . . . . . . . . . . . 5.25 10.5 15.75 21.0 DADPM, equivalent p g assuming complete hydrolysis .. 4.2 8.4 12.6 16.8 Optical density a t 540 mp, with 4-cm cells . . . . . . 0.165 0.300 0.460 0.535 DADPM, pg equivalent to optical density (from Table 111) . . 4.3 7.9 12.8 15.6 Percentage recovery of original MDI as DADPM . . . . 102 94 102 93 ABSORPTION OF MDI FROM THE ATMOSPHERE- Because of its low volatility and the very low concentrations involved, attempts to prepare accurately known atmospheric concentrations of MDI vapour were unsuccessful. Steady concentrations were, however, prepared by passing 1 litre of pure nitrogen per minute from a cylinder through a flow-meter and a bubbler containing 15 ml of Suprasec DN, which was immersed in a water-bath. By raising or lowering the temperature of the water-bath, the MDI concentration in the nitrogen leaving the bubbler was increased or reduced.Two bubblers of the type shown in Fig. 1, each containing 3 ml of 0.4 M hydrochloric acid, were connected in series to the outlet of the bubbler containing the MDI. Glass-to-glass contact between the three bubblers inside rubber sleeves was ensured. After each test the contents of the bubblers were diluted separately to 15.0 ml, and diazotised and coupled with 3-hydroxy- 2-naphthanilide, as described under “Choice of Coupling Agent’’ above. Results of the tests (Table V) show that an average of 75 per cent. of the total MDI vapour trapped was con- TABLE V RECOVERY Test number . . . . .. Temperature of water-bath, “C Time of bubbling, minutes . . Optical density, 1st bubbler . . Optical density, 2nd bubbler .. MDI, pg, 1st bubbler (see Table I1 for calibration points) MDI, pg, 2nd bubbler . . . . MDI, total pg . . . . . . MDI, p.p.m. . . . . .. Percentage of total MDI in in 1st bubbler.. . . . . OF MDI 1 20 85 0.325 0.135 7.6 3-15 10.75 0.013 71 VAPOUR FROM THE ATMOSPHERE 2 3 4 5 20 20 27 44 90 101 96 67 0.400 0.430 0.345 0.835 0.095 0.145 0.140 0.250 9.35 10.1 8.05 19.7 2.3 3.4 3.3 5.9 11-65 13.5 11.35 25.6 0.013 0.013 0.012 0.038 71 77 75 80 6 35 120 0-975 0.290 23.1 6.8 29.9 0.025 77 7 42 75 0.550 0.190 12-9 4.4 17.3 0.022 75August, 19671 4,4’-DI-ISOCYANATODIPHENYLMETHANE IN AIR 517 tained in the first bubbler. Attempts were made to improve this by incorporating either acetic acid or a surface-active agent (Lissapol NX) but little, if any, improvement resulted.Pairs of bubblers were mounted close to the spraying head of a machine producing rigid foam-laminated building board. Each pair consisted of an absorber of the type used in the method, containing 3.0 ml of 0.4 M hydrochloric acid, followed by a sintered-plate absorber containing 15 ml of 0.4 M hydrochloric acid, the two being joined together by glass-to-glass joints inside rubber sleeves. Recovery tests were also carried out on spray-borne MDI. -9 mni i.d. -6 mm 0 . d . ,jet 1.0 mm i.d. Fig. 1. All-glass absorber For each test air was drawn through the absorbers at 1 litre per minute for times ranging from 2 to 5 minutes. After sampling, the contents of the first absorber were diluted to 15.0ml with 0.4 M hydrochloric acid, and the contents of both absorbers were diazotised and coupled with 3-hydroxy-Z-naphthanilide, etc., as described under “Choice of Coupling Agent’’ above. The results are given in Table VI.The average percentage recovery was 76 per cent., i.e., close to that for vapour, although the spread was greater. TABLE VI RECOVERY OF MDI FROM SPRAY IN THE ATMOSPHERE Test number . . . . . . .. Optical density at 540 mp, with 4-cm cell, Optical density at 540 mp, with 4-cm cell, Time of sampling, minutes . . . . 1st absorber . . . . . . . . 2nd absorber . . . . . . . . MDI, pg, 1st absorber . . . . . . MDI, pg, 2nd absorber . . . . . . MDI, total p g . . . . . . . . MDI, p.p.m. . . . . . . . . . . Percentage of total MDI in 1st absorber . . 1 3-0 0.355 0.100 8.2 2.3 10-5 0.35 78 2 3 8.0 2.0 1-13 0.445 0.470 0.145 26.4 10.4 11.0 3.4 37.4 13.8 0.47 0.69 71 76 4 5 0.365 0.095 8.2 2.2 10.4 0.21 79 5 4-5 0.600 0.260 14.0 6.1 20.1 0.44 70 6 4-25 0.390 0.070 9.1 1.6 10.7 0.25 85 STABILITY OF 3-HYDROXY-2-NAPHTHANILIDE SOLUTION AND ITS USE INSTEAD OF AN AQUEOIJS SUSPENSION- In all of the work described previously, freshly prepared solutions of 3-hydroxy-2-naph- thanilide were used, which were pale yellow initially, but darkened.Old solutions produced much deeper blank solutions in the test procedure than fresh ones, even when they were kept in darkness. This was of little importance when the final colour was being measured518 REILLY: A FIELD METHOD FOR DETERMINING [A~~aZyst, Vol. 92 on a spectrophotometer, but in the usual matching procedure the use of solutions of 3-hydroxy- 2-naphthanilide more than 4 hours old resulted in colours that were too yellow to be matched with the inorganic colour standard solutions.It would be very inconvenient to have such an unstable solution when testing atmospheres that are remote from laboratory facilities. This problem was overcome by using a 1 per cent. w/v aqueous suspension of 3-hydroxy- 2-naphthanilide. The coupling solution was prepared shortly before use by adding to the separating funnel 10 ml of N sodium hydroxide and 1-0 ml of the 1 per cent. aqueous suspension of 3-hydroxy-2-naphthanilide. Table I1 shows calibration points obtained by using either this suspension or the solution; all the points lie close to the same curve when plotted on a graph. PRODUCTION OF PERMAKENT INORGANIC COLOUR STANDARD SOLUTIOKS- I t was desired to prepare inorganic colour standards that would match the chloroform extracts obtained when the method was applied to MDI concentrations of 0.01, 0.02 and 0.04 p.p.m.v/v. In establishing the standards, colours were developed with DADPRI and the amounts of DADPM corresponding to the desired MDI concentrations were calculated, assuming that only 75 per cent. of the MDI is trapped in a single absorber, but that the trapped MDI gives the same colour as an equivalent weight of DADPM. Thus 0.01,0.02 and 0.04 p.p.m. v/v of MDI on a 6-litre sample (at 20” C and 760 mm of mercury) correspond to 0.31, 0.62 and 1.24 pg of DADPM, respectively. For convenience in handling, the preparation was scaled-up 5 times to that 15 ml of chloroform extract could be obtained.From a solution of DADPM, containing 5.0 pg per ml in 0.4 M hydrochloric acid, 0.31, 0.62 and 1.24-ml portions were taken and were diluted to 15.0 ml with 0.4 M hydrochloric acid and diazotised, coupled and extracted with chloroform, as described under “Choice of Coupling Agent” above. Portions of each extract were run into 2 x $-inch flat-bottomed specimen tubes. Similar tubes were filled with coloured solutions made by diluting mixtures of standard iron( 111) and cobalt(I1) chloride solutions3 with 0.25 M hydrochloric acid. The proportions of the standards were varied by trial and error until an exact match was obtained when the tubes containing chloroform extract and standard were held side by side above a sheet of white paper and viewed vertically through the depth of the liquid in good daylight, but not bright sunlight.Table I shows the volumes of cobalt(I1) and iron(II1) chloride solutions, diluted to 500 ml with 0.25 M hydrochloric acid, which will produce standards equivalent to 0.01, 0.02 and 0.04 p.p.m. v/v of MDI in a 5-litre sample when tested by the procedure described above. The validity of these standards was checked by tests on a second standard DADPM solution. STRENGTH OF NITRITE SOLUTION REQUIRED FOR DIAZOTISATION- In some of the tests described under “Production of Permanent Inorganic Colour Standard Solutions” above, the chloroform extracts were much yellower than in others, making matching very difficult. This was traced to the failure to destroy completely with sulphamic acid all of the excess of nitrous acid left after the diazotisation.The difficulty was overcome by reducing the strength of the nitrite used from 0.5 to 0.1 N and by ensuring that the ground surfaces of the joint and stopper of the bubbler tube were wetted with sul- phamic acid when the latter was added. Reduction in the strength of the nitrite did not affect the colour of the final chloroform extracts, as shown when the calibration points with 0-1 N sodium nitrite given below are plotted on the same graph as those in Table 11. DADPM, pg . . . . . . .. . . 3.15 6.3 9.45 12.6 Optical density a t 540 mp, with 4-cm cells . . 0.185 0.350 0-465 0.620 INTERFERENCE FROM TERTIARY-AMINE CATALYSTS- Several tertiary-amine catalysts are used to promote the formation of polyurethanes and it is possible that atmospheric concentrations of these might arise together with MDI.Six commercial catalysts were tested as follows: 1.0 ml of a 0-1 per cent. w/v solution of distilled MDI in dry dioxan was diluted to 100 ml with 0-4 M hydrochloric acid and 5.0 ml of this solution were further diluted to 50 ml with 0-4 M hydrochloric acid. Five-rnillilitre volumes of this solution were mixed in turn with 10 mg of each catalyst and diluted to 15 mlAugust, 19671 4,4’-DI-ISOCYANATODIPHENYLMETHANE I N AIR 519 with 0.4 M hydrochloric acid. Each solution was then diazotised and coupled with 3-hydroxy- 2-naphthanilide, as described under “Choice of Coupling Agent” and the results are shown in Table VII. In no instance did the optical density in the presence of the catalyst differ significantly from that obtained with MDI alone.TABLE VII INTERFERENCE FROM TERTIARY-AMINE CATALYSTS Name of catalyst Optical density at 547 mp, with 4-cm cells NN-Dimethylcyclohexylamine . . . . 0.210 Di-azabicyclo-octane . . . . .. .. 0.216 Dimethyl- /3-phenylethylamine . . .. 0.203 N-Ethylmorpholine . . .. . . .. 0.209 NN-Dimethylbenzylamine . . .. . . 0.225 None, i.e., 5 p g of MDI only . . .. . . 0.215 NN-Dimethylaminoethanol . . .. . . 0.225 DISCUSSION OF RESULTS The work described above presents the experimental basis of a method for determining MDI in air at concentrations in the region of the recommended threshold limit value of 0.02 p.p.m. v/v. Conditions for sample collection, colour development and colour matching have been established. It has been shown that the colours produced can be reproduced artificially by stable reproducible inorganic coloured solutions, thus facilitating visual com- parisons in the field. The colours are stronger and, being pink, more easily matched than the pale yellow colours given by the previous method.2 The range of inorganic standards chosen is sufficient for field use in checking the level of safety in an industrial atmosphere. It represents from half to twice the threshold limit value. The equipment required is portable and independent of laboratory facilities. With the exception of the sampling pump or aspirator, it can all be packed into a wooden carrying-case 19 x 9 x 8 inches. A determination can be completed in 12 to 15 minutes, including sampling time. REFERENCES 1. 2 . 3. United States Pharmacopoeia, 17th Revision, p. 1071. 4. Marcali, K., Analyt. Chem., 1957, 29, 552. “Ministry of Labour, Safety, Health and Welfare New Series No. 8, Dusts and Fumes in Factory “The Determination of Toxic Substances in Air, A Manual of I.C.I. Practice,” Second Edition, Atmospheres,” Third Edition, H.M. Stationery Office, London, 1966. W. Heffer & Sons Ltd., Cambridge, 1965, p. 118. Received Januav?, 20th, 1967