Registration Dossier

Administrative data

Endpoint:
monitoring data
Type of information:
migrated information: read-across from supporting substance (structural analogue or surrogate)
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
other: well-documented publication, which meets basic scientific principles

Data source

Reference
Reference Type:
publication
Title:
Fate and Potential Environmental Effects of Methylenediphenyl Diisocyanate and Toluene Diisocyanate Released into the Atmosphere
Author:
Tury, B., Pemberton, D. and Bailey, R.E.
Year:
2004
Bibliographic source:
J. Air & Waste Manage. Assoc. 53:61–66

Materials and methods

Test guideline
Qualifier:
no guideline followed
Principles of method if other than guideline:
Information from a variety of sources has been collected and summarized to facilitate an overview of the atmospheric fate and potential environmental effects of emissions of methylenediphenyl diisocyanate (MDI) or toluene diisocyanate (TDI) to the atmosphere. Laboratory studies show that TDI
(and by analogy MDI) does not react with water in the gas phase at a significant rate. Laboratory studies also show that this oxidation by OH radicals ( the primary degradation reaction of these aromatic diisocyanates in the atmosphere), estimated half-life of one day, is not expected to result in increased ground-level ozone accumulation.
GLP compliance:
not specified
Type of measurement:
other: predictions
Media:
air

Test material

Reference
Name:
Unnamed
Type:
Constituent
Type:
Constituent
Details on test material:
- Name of test material (as cited in study report): toluene diisocyanate (toluene 2,4- diisocyanate (2.4 TDI, 80 %), toluene 2,6- diisocyanate (2.6 TDI, 20 %)
- Molecular formula (if other than submission substance): C9H6N2O2
- Molecular weight (if other than submission substance): 174.2 g/mol
- Smiles notation (if other than submission substance): Cc1ccc(cc1\N=C=O)\N=C=O
- InChl (if other than submission substance): InChI=1S/C9H6N2O2/c1-7-2-3-8(10-5-12)4-9(7)11-6-13/h2-4H,1H3 Y, Key: DVKJHBMWWAPEIU-UHFFFAOYSA-N
- Structural formula attached as image file (if other than submission substance): see Fig.
- Substance type: aromatic diisocyanate
- Physical state: liquid

Study design

Details on sampling:
No details available.

Results and discussion

Concentrationopen allclose all
Country:
other: Prediction model: OPS model contained within the European exposure multimedia model, the European Union System for the Evaluation of Substances
Location:
predicted annual average air concentration at 100 m from the source
Substance or metabolite:
substance
Remarks:
TDI
Conc.:
6.7 µg/m³ air
Remarks on result:
other: This is a generic (i.e., not site-specific) model & assumes release directly from a low building with no stack. Predictions are linearly proportional to the emission rate. An emission rate of 1 g/hr continuously over the year (i.e., 8.8 kg/yr) is presumed
Country:
other: Prediction model: OPS model contained within the European exposure multimedia model, the European Union System for the Evaluation of Substances
Location:
prediction: deposition rate 1 km radius from the emitter
Substance or metabolite:
substance
Remarks:
TDI
Conc.:
3.5 other: g/m²/yr
Remarks on result:
other: The model also predicts a deposition rate over a 1-km radius from the emitter of 9 g/m²/yr for MDI and 3.5 g/m²/yr for TDI.
Country:
other: Prediction model: site-specific, EPA, COMPDEP
Location:
Calculation covered MDI aerosol emission of 0.1, 1, 10, & 50 µm diameter, at 1 mg/sec (270 µg/m³), from buildings 10–25 m high, with & without an additional stack. Assumed operation of 12 hr daily, 365 days/yr, corresponding to an annual release of 16 kg.
Substance or metabolite:
substance
Remarks:
MDI (100 m of the buidling)
Conc.:
>= 0.001 - <= 0.008 other: %
Remarks on result:
other: 3 urban locations in USA with PU industry considered: Boston, MA, Jackson, MS, and Phoenix, AZ. Representation of a varied range of meteorology, use of hourly climate data for 1989–1991. Focus: annual average conc. & deposition within 1000 m of the source
Country:
other: Prediction model: site-specific, EPA, COMPDEP
Location:
for a 0.1 -µm aerosol released from a 5-m stack above a 10-m building in Jackson, 1989
Substance or metabolite:
substance
Remarks:
MDI
Conc.:
2.1 other: mg/m²
Remarks on result:
other: maximum combined annual average deposition at 10 m from the source
Country:
other: Prediction model: site-specific, EPA, COMPDEP
Location:
for a 0.1 -µm aerosol released from a 5-m stack above a 10-m building in Jackson, 1989
Substance or metabolite:
substance
Remarks:
MDI
Conc.:
<= 0.3 other: mg/m²
Remarks on result:
other: maximum combined annual average deposition at 100 or more m from the source
Details on results:
Experiments were carried out with a variety of TDI concentrations at 27 °C and relative humidity 7–70 % in a 17-m³, Teflon-lined environmental chamber. Atmospheres were monitored for TDI (by several methods), TDA, TDA-urea, organic carbon, and total bound nitrogen, and for presence of aerosols. A significant TDI loss rate was observed, which was matched by an equal loss of organic carbon and nitrogen, indicative of physical removal from the gas phase by deposition onto the chamber’s walls rather than chemical conversion. No evidence for a gas-phase reaction with water vapour was found, and no TDA or ureas were detected. Review of some earlier literature had suggested rapid reaction of TDI with water vapor and they concluded that the observations were the result of surface reactions or faulty analysis. In more recent studies, no amine was detected in atmospheres generated for toxicological tests and containing TDI vapor or MDI aerosol.
This finding of a slow gas-phase reaction of TDI (and by analogy MDI) with water vapor is consistent with the slow gas-phase reaction of other rapidly hydrolysed compounds. Their results indicate half-lives for atmospheric hydrolysis in air of 40 % relative humidity, at 25 °C, of 13,100, 113, and 455 years for chloroacetyl chloride, phosgene, and trichloroacetyl chloride, respectively.

Any other information on results incl. tables

A simple prediction of the average atmospheric concentrations of MDI and TDI close to a point-source emitter can be obtained using the OPS model contained within the European exposure multimedia model, the European Union System for the Evaluation of Substances. This is a generic (i.e., not site-specific) model and assumes release directly from a low building with no stack. The predictions are linearly proportional to the emission rate. At an emission rate of 1 g/hr continuously over the year (i.e., 8.8 kg/yr), the predicted annual average air concentration at 100 m from the source is 6.7 x10³ g/m³ for both materials. The model also predicts a deposition rate over a 1-km radius from the emitter of 9 µg/m²/yr for MDI and 3.5 µg/m²/yr for TDI. This prediction for MDI is higher than for TDI because the model takes into account the greater tendency of a substance of very low vapor pressure to condense onto atmospheric aerosols, which deposit more readily than vapor. These are reassuringly low predictions, and correlate well with those of more detailed, site-specific studies.

One such site-specific model (EPA, COMPDEP) was used to predict the dispersion of MDI aerosols. Although this study was occasioned by questions about the behavior of these aerosols, the predictions for very small particles (0.1 µm) would be expected to apply equally well to vapor, because the very low settling rate of such particles produces no significant depletion of the atmospheric concentration. The model applies equally well to TDI, because the behavior of the air mass is not affected by the low concentration of pollutant. The predictions of this model are proportional to the emission rate.

The calculation covered the emission of MDI aerosols of 0.1, 1, 10, and 50 µm diameter, at 1 mg/sec (270 µg/m³), from buildings 10–25 m high, with and without an additional stack. Over the assumed operation of 12 hr daily, 365 days/yr, this corresponds to an annual release of 16 kg diisocyanate. Three urban locations in the United States with some PU industry were considered: Boston, MA, Jackson, MS, and Phoenix, AZ. These were chosen to represent a varied range of meteorology, and hourly climate data for 1989–1991 were used.

The focus was on annual average concentrations and deposition at locations within 1000 m of the source. The dominant effect noted was that of a stack in reducing building downwash. Without a stack, maximum groundlevel air concentrations ranging from 0.001 to 0.008% of the release concentration were predicted (for 0.1 m particles), at positions all within 100 m of the building.

Addition of a 15-m stack to a 10-m building reduced these maximum ground-level air concentrations by a factor of 10 or more and moved their position to 300–400 m distance from the source. Predicted air concentrations were little affected by aerosol particle size.

Deposition comprises both wet and dry processes and is therefore a more complex function of site and climate. For example, Phoenix, with the highest maximum air concentrations, had the lowest maximum deposition rates because of the low rainfall. In general, although the maximum deposition rate was found to increase with particle size, as would be expected, it was low in all cases and very localized close to the source. The deposition process was not sufficient to significantly deplete the air concentrations. As an example, for a 0.1 µm aerosol released from a 5-m stack above a 10-m building in Jackson, 1989, the maximum combined annual average deposition was 2.1 mg/m² at 10 m from the source. At 100 m or more from the source, the annual average deposition was = 0.3 mg/m².

A similar model (ISC Short-Term Dry Deposition) has also been used to calculate annual average MDI concentrations close to two manufacturing plants in the United States, a boat manufacturing facility and an oriented strand board facility. In this study, the highest measured emission levels were assumed to prevail continuously throughout the year (thereby presenting a worst-case scenario) and actual plant configurations were used.

For the boat manufacturing facility, with the lower emission rate (73 µg/sec; 11 µg/m³) and lower release height (11 m), highest annual average concentrations of about 0.0006 µg/m³ were predicted at about 100 m from the site, and less than 0.0001 µg/m³ at 800 m, for 50 µm aerosols. With smaller particles, or gas, maxima of 0.0005–0.0006 µg/m³ were predicted within 300 m of the site.

For the oriented strand board facility, a high emission rate of 12 mg/sec (280 µg/m³) at a height of 30 m led to highest annual average concentrations in the 0.003– 0.004 µg/m³ range at 1000 m and less than 0.0008 µg/m³ by the 3000 m limit for 50 -µm particles. Here again rather

more disperse concentrations and somewhat lower maxima (0.0024–0.0032 µg/m³) were predicted for smaller particles or gas.

Reaction with Atmospheric Water

The knowledge that MDI and TDI are easily hydrolysed and degraded in aqueous solution, forming polyureas and small amounts of diamine raises the possibility that a similar reaction might occur in the vapor phase. The first phase of a study of the atmospheric fate of TDI, at Battelle Laboratories, was devoted to an investigation of the extent of any gas-phase reaction with water and of any production of toluenediamine (TDA). Experiments were carried out with a variety of TDI concentrations at 27 °C and relative humidity 7–70 % in a 17-m³, Teflon-lined environmental chamber. Atmospheres were monitored for TDI (by several methods), TDA, TDA-urea, organic carbon, and total bound nitrogen, and for presence of aerosols. A significant TDI loss rate was observed, which was matched by an equal loss of organic carbon and nitrogen, indicative of physical removal from the gas phase by deposition onto the chamber’s walls rather than chemical conversion. No evidence for a gas-phase reaction with water vapor was found, and no TDA or ureas were detected. The authors reviewed some earlier literature that had suggested rapid reaction of TDI with water vapor and concluded that the observations were the result of surface reactions or faulty analysis. In more recent studies, no amine was detected in atmospheres generated for toxicological tests and containing TDI vapor or MDI aerosol.

This finding of a slow gas-phase reaction of TDI (and by analogy MDI) with water vapor is consistent with the slow gas-phase reaction of other rapidly hydrolysed compounds. Their results indicate half-lives for atmospheric hydrolysis in air of 40% relative humidity, at 25 °C, of 13,100, 113, and 455 years for chloroacetyl chloride, phosgene, and trichloroacetyl chloride, respectively.

Oxidative Degradation

In the second phase of their study, the Battelle workers used their environmental chamber to investigate the effects of irradiation, photochemically produced agents, and other pollutants on the atmospheric stability of TDI. A substantial loss rate of TDI caused by adsorption onto the walls was again observed. Irradiation substantially increased the loss rate. The incremental loss caused by irradiation was not affected by the presence of common urban pollutants: hydrocarbon mixtures, NH3, and ammonium sulfate aerosol. It was enhanced by the presence of relatively high concentrations of diazabicyclooctane (DABCO, a PU catalyst) but was suppressed by nitric oxide, a free radical scavenger. No TDA was detected in any experiment, corresponding to less than 0.05% conversion of TDI. Overall, the study clearly indicated degradation by photolytically generated radicals, rather than by direct photolysis, and the absence of any hydrolysis in the vapor phase.

The degradation of most trace organic gases in the atmosphere is initiated by reaction with OH radicals, themselves generated photochemically. This was confirmed as the dominant process for TDI by a study of the degradation rate of 80/20 TDI in a large photoreactor, at ambient temperature and pressure, in the presence of photolytically generated OH radicals. The reaction rate of TDI with OH radicals, measured relative to that of toluene, was estimated as=7.4 x 10E12 cm³/molecule/ sec. This was in satisfactory agreement with a rate calculated from the Battelle results.

Carter et al. found a reaction rate of 5.9 x 10E12 cm³/molecule/sec for p-toluene isocyanate (PTI), a close structural analogue of MDI, in a smog chamber study. A reaction rate for MDI about double this, 1.2 x 10E11 cm³/molecule/sec, would be expected because the predominant reaction of aromatics with OH radicals is addition to the ring and MDI has two rings available for reaction, each with alkylene para- to the NCO, as in PTI. When these reaction rates are combined with an assumed typical global seasonal average concentration of OH radicals in the troposphere, 1.1 x 10E6/cm³ atmospheric half-lives can be estimated. These are 24 hr for TDI (based on the experimental reaction rate) and 15 hr for MDI.

Prediction of the actual lifetime of a chemical that is not evenly mixed throughout the atmosphere is considerably more complicated because the OH radical concentration can vary by several orders of magnitude depending on the intensity of sunlight, temperature, and other factors. However, these estimates provide a useful way to compare the persistence of different chemicals in the atmosphere.

Smog Formation

Although a chemical that degrades readily in air will have low persistence, the consequent reacting mixture of organics and nitrogen oxides (NOx) can enhance the accumulation of ozone and peroxyacetyl nitrate (PAN) in the troposphere. This unhealthy mixture is commonly referred to as photochemical smog. The reactions taking place in a polluted atmosphere are so complex that the complete reaction sequences have been elucidated for only a few chemicals. Simplified summaries of the reactions have been adopted for modeling of photochemical smog formation, for which the primary indicator is ozone concentration.

Carter et al. assessed the potential of both 2,4- and 2,6-TDI for the formation of atmospheric ozone in a study comprising environmental chamber experiments and computer modeling. The incremental effects of added 2,4- and 2,6-TDI on ozone formation, NO oxidation, and OH radical concentration in three different simulated photochemical smog systems were determined. These systems covered a wide range of ratios of reactive organic gas to NOx. Both isomers of TDI were found to inhibit ozone formation and radical levels in all experiments. Although the precise mechanism of TDI reaction with OH is unknown, parameterised models showing the overall processes in simple but chemically reasonable terms were made and used in simulations of the environmental chamber reactions. Overall, the comparison of simulation with experiment suggested the formation of radical and NOx sinks in the atmospheric reactions of TDI. The best-fitting model assumed 70% radical inhibition, no conversions of NO to NOx, and formation of significant yields of products (such as cresols and nitrophenols) whose subsequent reactions cause removal of NOx. An 18-hr irradiation showed that the ozone inhibition effects extended over at least two days. Application of the smog chamber reaction parameters to model scenarios representing ozone pollution episodes throughout the United States led to predictions that TDI is unlikely to have a positive effect on ozone formation under any atmospheric conditions in the United States, and it should not be considered an ozone precursor.

MDI could not be studied directly because of its low vapor pressure. However, in a similar study, PTI was tested as a surrogate for MDI. The reactivity of PTI was predicted to mimic that of MDI because of their close structural similarity. PTI was found to inhibit ozone yields and OH radical levels in lower NOx experiments but had very small positive effects under higher NOx conditions. The results could best be fitted by a model assuming 10% radical inhibition, some conversion of NO to NOx, and some formation of a photoreactive product such as methylglyoxal. It was concluded that MDI would probably have a negative impact, reducing ozone formation, under atmospheric conditions where NOx is limited, which constitute most “base case” scenarios used to represent high ozone episodes in U.S. urban areas. It might have a positive impact equal to or greater than ethane on ozone formation under high NOx conditions where ozone is more sensitive to the level of volatile organic compounds, though there were reasons to suspect that the model overestimated these effects, in particular, because the early degradation products of MDI would still contain an aromatic ring and, hence, could be less photoreactive than those from PTI.

Conclusions:

The only releases to air of any significance arise during the use of MDI and TDI in manufacturing processes and are generally very low. Typical environmental loadings are less than 1 g/t of MDI used and about 25 g/t of TDI used. Models of the physical dispersion of such releases predict very low concentrations in air at and beyond fenceline limits. In any situation where atypically large releases are possible, site-specific modeling could be used to evaluate a worst-case scenario. Although MDI and TDI do not react to any appreciable extent with water vapor in the air, they are not persistent in the troposphere, being readily degraded by reaction with OH radicals. Half-lives are predicted to be on the order of 15–24 hr. A more quantitative appreciation of this lack of persistence can be obtained by incorporation of the degradation rates into the EUSES generic multimedia model. Predictions are for a region of area 4 104 km² within a continent similar to western Europe. It is assumed that MDI and TDI are emitted at 1 and 25 g/t used, respectively, and that usage of each within the region is 100,000 t/yr. This leads to regional yearly emissions of 100 kg of MDI and 2500 kg of TDI. On these assumptions, the model predicts background air concentrations throughout the region of only 5 106 g/m³ for MDI and 1 104 g/m³ for TDI.

This atmospheric degradation of MDI and TDI by OH radicals raises concerns that the process might lead to ozone or smog formation. Smog chamber studies have shown that this is not the case. TDI, in particular, should not be considered an ozone precursor. MDI might enhance ozone formation under specific (high NOx) conditions but is unlikely ever to be present in significant concentrations.

Physical deposition of diisocyanate from such low atmospheric concentrations should be very low. Such transfer, and the subsequent environmental fate in water or soil, is not fully amenable to multimedia modeling because of the reactivity of these compounds in the aquatic environment, but an appreciation can be obtained from the EUSES dispersion calculation referred to earlier. This predicts that even within 1 km of a source emitting 8.8 kg/yr, the annual deposition rates of MDI and TDI will be only 9 and 3.5 g/m², respectively. If this 9 g/m² MDI were to be mixed into the top 10 cm of soil, density 1.3 g/cm³, the concentration of MDI would be about 0.07-g/kg soil. In addition, any material so deposited would react or hydrolyze rapidly. The major products of hydrolytic breakdown under most environmental conditions are stable, nontoxic polyureas. Any diamines produced would be at correspondingly low concentrations and are known to be photodegradable in water, to be aerobically biodegradable, and to bind strongly and irreversibly to soil. Such deposition of MDI and TDI is therefore most unlikely to lead to adverse effects. Overall, it can be concluded that no significant long-term or wide-ranging environmental effects would be expected from current emissions of MDI or TDI to air.

Applicant's summary and conclusion

Conclusions:
No significant long-term or wide ranging environmental effectets are expected from the current emissions of MDI and TDI to air.
Executive summary:

Information from a variety of sources has been collected and summarized to facilitate an overview of the atmospheric fate and potential environmental effects of emissions of methylenediphenyl diisocyanate (MDI) or toluene diisocyanate (TDI) to the atmosphere. Atmospheric emissions of both MDI and TDI are low, both in terms of concentration and mass, because of their low volatility and the need for careful control over all aspects of their lifecycle from manufacture through disposal. Typical emission losses for TDI are 25 g/t of TDI used in slabstock foam production. MDI emission losses are lower, often less than 1 g/t of MDI used. Dispersion modeling predicts that concentrations at the fenceline or beyond are very low for typical releases. Laboratory studies show that TDI (and by analogy MDI) does not react with water in the gas phase at a significant rate. The primary degradation reaction of these aromatic diisocyanates in the atmosphere is expected to be oxidation by OH radicals with an estimated half-life of one day. Laboratory studies also show that this reaction is not expected to result in increased ground-level ozone accumulation.

Experiments were carried out with a variety of TDI concentrations at 27 °C and relative humidity 7 – 70 % in a 17-m³, Teflon-lined environmental chamber. Atmospheres were monitored for TDI (by several methods), TDA, TDA-urea, organic carbon, and total bound nitrogen, and for presence of aerosols. A significant TDI loss rate was observed, which was matched by an equal loss of organic carbon and nitrogen, indicative of physical removal from the gas phase by deposition onto the chamber’s walls rather than chemical conversion. No evidence for a gas-phase reaction with water vapor was found, and no TDA or ureas were detected.

Overall, the study clearly indicated degradation by photolytically generated radicals, rather than by direct photolysis, and the absence of any hydrolysis in the vapor phase.

A substantial loss rate of TDI caused by adsorption onto the walls was again observed.

Irradiation substantially increased the loss rate. The incremental loss caused by irradiation was not affected by the presence of common urban pollutants: hydrocarbon mixtures, NH3, and ammonium sulfate aerosol. It was enhanced by the presence of relatively high concentrations of diazabicyclooctane (DABCO, a PU catalyst) but was suppressed by nitric oxide, a free radical scavenger. No TDA was detected in any experiment, corresponding to less than 0.05% conversion of TDI. Overall, the study clearly indicated degradation by photolytically generated radicals, rather than by direct photolysis, and the absence of any hydrolysis in the vapor phase.

The degradation of most trace organic gases in the atmosphere is initiated by reaction with OH radicals, themselves generated photochemically. This was confirmed as the dominant process for TDI by a study of the degradation rate of 80/20 TDI in a large photoreactor, at ambient temperature and pressure, in the presence of photolytically generated OH radicals. The reaction rate of TDI with OH radicals, measured relative to that of toluene, was estimated as = 7.4 x 10E12 cm³/molecule/ sec.

Both isomers of TDI were found to inhibit ozone formation and radical levels in all experiments. An 18-hr irradiation showed that the ozone inhibition effects extended over at least two days. The prediction is made, that TDI is unlikely to have a positive effect on ozone formation under any atmospheric conditions in the United States, and it should not be considered an ozone precursor.

This atmospheric degradation of MDI and TDI by OH radicals raises concerns that the process might lead to ozone or smog formation. Smog chamber studies have shown that this is not the case. TDI, in particular, should not be considered an ozone precursor.

Overall, it can be concluded that no significant long-term or wide-ranging environmental effects would be expected from current emissions of MDI or TDI to air.