JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HYGIENE

, VOL. , NO. , D–D

http://dx.doi.org/./..

OSHA Compliance Issues

Methemoglobinemia resulting from exposure in a confined space: Exothermic

self-polymerization of , -methylene diphenyl diisocyanate (MDI) material
Reported By
Philip A. Smitha , Jeffrey Lodwicka , Joe Darttb , Jenny R. Amanic , and Kathleen M. Faganc
a
U.S. Department of Labor – OSHA, Health Response Team, Sandy, Utah; b U.S. Department of Labor – OSHA, St. Louis Area Office, St. Louis,
Missouri; c U.S. Department of Labor – OSHA, Office of Occupational Medicine and Nursing, Washington, D.C.

ABSTRACT KEYWORDS
A worker attempting to remove solidified material inside a confined space (storage tank) suffered Aniline; MDI;
methemoglobinemia;
severe methemoglobinemia and almost died. The tank contained liquid 4,4 -methylene diphenyl methylene diphenyl
diisocyanate monomer that had solidified after an equipment power failure caused excessive heating. diisocyanate; p-toluidine;
Wearing a full-face elastomeric air-purifying respirator and TyvekR
coveralls, the worker used pneu- thermal degradation
matic air hammers to break up the solid material. After two tank entries totaling slightly less than one
hour, the worker complained of headache and dizziness and within two hours of exiting the tank, he
was admitted to the hospital in severe respiratory distress. During his eight-week hospital course, he
suffered a cardiac arrest among other complications. An investigation into the cause of the worker’s
illness used onsite gas chromatography-mass spectrometry which identified aniline and p-toluidine
vapor within the tank, attributable to overheating that led to formation of the solid material. Both are
well-known causes of methemoglobinemia, and had the initial characterization of the confined space
atmosphere adequately identified the hazards present appropriate engineering controls and personal
protective equipment could have allowed the tank entrant to work safely in the space.

Introduction blanket to eliminate the potential for reaction of isocyano

groups with water which produces an amine in place of
Work within a large tank (Figure 1) which had previously
an isocyano group while evolving a molecule of carbon
held a mixture of 4,4 -methylene diphenyl diisocyanate
dioxide gas. Two tank access points of 18” diameter are
monomer (MDI, Figure 2) and MDI homopolymer
normally sealed shut but may be opened to allow entry or
led to hospitalization of a worker with severe methe-
visual inspection. The floor level access point is shown in
moglobinemia. An inspection was opened by the Occu-
Figure 1. The other access point, not visible in the figure,
pational Safety and Health Administration (OSHA)
is located at the top of the tank. An external pressure relief
to investigate the cause(s) of the worker’s illness. This
valve for this tank limited pressure within the tank to 10
case study describes the methods used to investigate
psig.
the worker’s exposures, provides information regard-
The desired storage temperature for the liquid MDI
ing occupational methemoglobinemia, and illustrates
material ranged from 85–90°F (29.4–32.2°C). Three flex-
little-known hazards related to the storage and use of
ible 480 V resistively heated silicone rubber pads con-
MDI.
nected in series were used to heat the tank contents. The
18 × 72 heaters were placed around the circumference
MDI storage tank background
of the tank immediately above the floor level opening
The storage tank used to store the liquid MDI material seen in Figure 1. This area of the tank surface visible in
measured 22 high with a 7 diameter, providing a use- the figure had been cleaned after the overheating inci-
able storage capacity of 6,000 gallons. The liquid material dent by grinding to remove residual adhesive, in prepa-
in the tank was normally kept under a dry nitrogen (N2 ) ration to replace the heaters. Under normal operation,

CONTACT Philip A. Smith smith.philip.a@dol.gov Industrial Hygiene Chemistry Division, OSHA Salt Lake Technical Center,  S. Sandy Parkway, Sandy,
UT, .
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/uoeh.
Joe Dartt’s current affiliation is U.S. Department of Labor — OSHA, Rgion , St. Louis, MO. Jenny R. Amani’s current affiliation is the University of Texas – Houston,
School of Public Health, Houston, TX.
This article not subject to US copyright law.
D14 P. A. SMITH ET AL.
Figure . Tank used to store heated MDI material. Floor level access point has been opened and solidified material is visible at the top of
and at the bottom right of the opening.
internal circulation of liquid MDI material maintained
a fairly uniform temperature with thermostatic control
of current to the heaters. Overheating of liquid material
in the immediate vicinity of the heaters occurred when
power to the internal circulation system was lost at some
point during a weekend storm, but power to the heaters
continued. The thermostat signal to the heaters called for
continuous heating for an unknown period of time (pos-
sibly as long as 30 hr) due to lack of circulation of liq-
uid MDI material. The tank contained about 1,700 gal-
lons of liquid when the power disruption occurred. When
the problem was discovered, about 1,200 gallons of the
remaining liquid material was drained from the tank,
leaving a 60–70 cm thick solidified mass across the entire
tank cross-section. The dry N2 atmosphere inside the tank
was lost when access points were opened for visual inspec-
tion of the tank interior.
Tank entry and hospitalization of tank entrant
Personal protective equipment (PPE) used by the entrant
included hooded coveralls made of Tyvek brand flash-
spun high-density polyethylene fiber, a Survivair Opti Fit
7620 full-face elastomeric air purifying respirator with
Survivair 1058 multi-contaminant/p-100 cartridges, and
MicroFlex Safegrip latex gloves worn over cotton gloves.
The PPE was worn without taping the gaps in the cov-
eralls at the respirator face piece, wrists, or ankles. The
entrant donned new coveralls, latex gloves, and respirator
cartridges between entries.

Figure . Relevant chemical structures. A MDI, B phenyl iso-
cyanate, C p-tolyl isocyanate, D aniline, E p-toluidine; F MDI
uretidinedione dimer (formation of this compound from MDI is
reversible); G MDI isocyanurate trimer (formation of this com-
pound from MDI is irreversible).
Selection of PPE was not based on quantitative sam-
pling information. No evaluation of airborne hazards was
conducted beyond that provided by a multi-gas meter
which showed 20.9 volume percent oxygen, no flammable
gases present (by catalytic combustion sensor), and no
hydrogen sulfide or carbon monoxide present. Coveralls
were chosen to keep what was assumed to be nuisance
dust off of the worker. The multi-contaminant respirator
cartridge was chosen in order to protect the worker from
airborne dust generated during mechanical removal of the
solid material and to provide the worker with relief from
a residual aromatic odor in the tank.
A confined space permit was opened, and testing was
completed with the multi-gas meter prior to each of two
entries. With a coworker stationed on top of the tank, a
32-year old male worker entered the tank at 10:10 a.m.
through the top opening. He descended a chain lad-
der approximately 6 m to the top of the solidified MDI
material and used a jackhammer on the solidified mate-
rial for 30 min. He then left the tank, returning via the
same route at 11:04 am with a bigger jackhammer. The
JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HYGIENE D15
worker remained in the tank for about 20 min and then
left the tank, informing coworkers that he was experi-
encing a headache and dizziness. Believing this to be
caused by heat stress, the tank entrant and coworkers
broke for lunch and drove to a fast food establishment in
an air-conditioned vehicle. The tank entrant rested in the
vehicle while his coworkers went inside. Soon thereafter,
he entered the restaurant complaining of a worsening
headache and difficulty keeping his head up. His cowork-
ers started to drive him to a hospital but pulled over and
called the local emergency response number. An ambu-
lance arrived on-scene at 12:42 pm. Responders noted the
tank entrant’s ashen, gray, sweaty skin, and a “chemical
odor about the patient.” The initial blood oxygen satura-
tion was 88% (normal is 95–100%) but this value dropped
to 85% even after oxygen was administered. Upon arrival
in the emergency department of a local hospital, the tank
entrant’s blood pressure, heart rate, and respiratory rate
were elevated. His oxygen saturation was 86%, unrespon-
sive to supplemental oxygen.
Medical course and clinical outcome
To treat respiratory failure, the worker was sedated and
intubated. A chemical odor was again noted by hospital
staff, as well as yellow flecks of hardened material around
the worker’s head and neck, prompting initiation of an
exposure protocol. Clothing was removed and was later
incinerated, and the worker’s skin and hair were washed
several times. Due to the dusky gray color of the worker’s
skin, lips, palms and soles, and the chocolate color of
his blood, methemoglobinemia was suspected. The initial
methemoglobin level was extremely high at 72.6% (nor-
mal is < 3%). Several treatments with methylene blue
lowered methemoglobin levels to the normal range after
36 hr.
The tank entrant suffered numerous complications,
including hemolytic anemia, serotonin syndrome, and
multi-organ system failure, including kidney failure. Dur-
ing transfer to a tertiary care facility for dialysis (five days
after his entry into the confined space), the tank entrant
suffered a cardiac arrest. He was resuscitated and sta-
bilized, and the transfer was completed. His five-week
course in the tertiary care hospital included treatment
for acute respiratory distress syndrome, kidney failure,
and thrombocytopenia (low platelet count). He was on
a ventilator for several weeks, lost a significant amount
of weight and muscle mass, and suffered brain damage.
He was transferred to a rehabilitation hospital for two
weeks of intensive rehabilitation for multiple health con-
sequences, including cognitive impairment, balance and
coordination problems, trouble swallowing (dysphagia),
trouble speaking (dysarthria), anxiety, and malnutrition.
D16 P. A. SMITH ET AL.

Seven weeks and five days after he entered the tank, he was variable ionization time (0.12–60 ms), followed by 5 ms
discharged home, where he continued long-term outpa- ion cooling time. Ion ejection over a range of 40–510 m/z
tient physical and speech therapy. His ability to return to was obtained with 60 ms voltage scans.
work remains uncertain.

Laboratory analyses of workplace samples

Materials and methods
Following the qualitative identification of several gas
Field analyses phase aromatic amine analytes within the tank atmo-
sphere using portable GC-MS, further work could be
A person-portable GUARDION GC-MS instrument
completed to provide an exposure intensity estimate.
(Torion Technologies, American Fork, UT) was used to
Samples were collected through the same threaded port
complete on-site analyses to identify possible causative
used for needle trap sampling with phosphoric acid-
methemoglobinemia agents. This instrument employed
treated sorbent tubes (XAD-7, 226–98, SKC Inc., Eighty
a 5 m analytical column (Restek, Bellefonte, PA) with
Four, PA) and typical air sampling pumps. Analysis was
0.1 mm inside diameter and liquid polymeric stationary
completed using OSHA Method PV7029 for aniline,[1]
phase film (MXT-5) thickness of 0.4 µm. The column
and results were also requested for p-toluidine.
temperature was programed (50°C for 10 s, 120°C/min to
A small sample of the solidified MDI material from
300°C, held for 180 s) while the injector and transfer lines
the tank was subjected to solid phase Fourier trans-
were kept at 270 and 250°C, respectively. A high pres-
form infrared (FTIR) absorbance analysis to provide
sure helium cartridge (99.5% purity, Torion Technolo-
functional group information. The HazMatID instru-
gies) provided carrier gas, and both 10:1 and 50:1 injector
ment (Smiths Detection, Edgewood MD) used for this
split settings were on (maximum split) to start an analysis
employed an attenuated total reflectance (ATR) interface
event. Samples were introduced from a needle trap sam-
designed for analysis of solids and liquids.[2] A small
pler or from a solid phase microextraction (SPME) fiber
portion of the material was crushed and held in place
by inserting a loaded 19-gauge needle trap or SPME fiber
on the ATR interface by a pressure arm, and 256 scans
needle directly into the heated injector of the instrument.
(650–40000 cm−1 , resolution of 4 cm−1 ) were averaged to
The SPME fiber used (65 µm polydimethylsilox-
provide an absorbance spectrum.
ane/divinylbenzene coating, Supelco, Bellefonte PA) was
mounted in a Custodion SPME syringe (Torion Technolo-
gies). Passive SPME sampling was used to obtain a 15-
Results and discussion
min sample from the headspace above a small portion of
the solidified MDI material removed from the tank which Field analyses
was heated in a glass sampling jar kept at 110°C. The nee-
dle trap sampler (Torion Technologies) contained 2 mg Field analyses immediately identified aniline and p-
carboxen 1000 behind similar masses of carboxen 1016 toluidine (Figure 3) as vapors within the tank. Trace
and TenaxTA. By applying vacuum to a side-hole in the
needle just above the sorbents, air may be drawn through
the needle tip, passing through the sorbent beds from the
weakest sorbent in the front (Tenax) to the strongest at
the rear (Carboxen 1000). Active needle trap sampling (30
mL/min) was completed from an approximately 2.5 cm
diameter threaded port in the tank about 2.5 m above
floor height (above the solidified material), with sam-
pling duration determined empirically based on chro-
matographic peak size. Backflush desorption of the needle
trap sampler occurred by quickly inserting the needle tip
into the heated injector of the GC-MS instrument where
it bottomed into a receptacle at the base of the injector.
Heated carrier gas entered the side hole above the sorbent
beds and desorbed analytes were flushed out through the
needle tip to the column. Figure . GC-MS chromatogram obtained in the field. This chro-
Electron ionization of analytes (70 eV) occurred within matogram resulted from needle trap sampling ( s) of the tank
the toroidal ion trap mass spectrometric detector, with interior; A Aniline, and B p-toluidine.

Figure . GC-MS chromatogram obtained in the field. This chro-
matogram resulted from a  min SPME sample collected in the
headspace above solidified material removed from the tank at
°C: 1 phenyl isocyanate, 2 aniline 3 p-tolyl isocyanate, 4 p-
toluidine, 5 p-toluonitrile, 6 butylated hydroxytoluene.
amounts of phenyl isocyanate were also observed in sev-
eral of the needle trap samples, but the amine compounds
dominated all of the chromatograms that resulted from
sampling of the gas phase tank contents. Field analyses
also showed that phenyl isocyanate, aniline, p-tolyl iso-
cyanate, p-toluidine, (Figure 2), benzonitrile, and p-tolyl
nitrile were present in the solidified material retrieved
from the tank, although to avoid the use of solvents in the
field, heating was needed to volatilize these analytes for
SPME sampling (Figure 4).
Laboratory analyses of workplace samples
Results for quantitative laboratory analyses of the atmo-
sphere within the tank are provided in Table 1. Break-
through occurred for an initial sampling event (results not
shown) due to the high concentration of both analytes
within the tank atmosphere and sampling time of about
175 min. The results shown in Table 1 are for the sec-
ond event with very short sample duration. The concen-
trations of aniline detected in the tank atmosphere were
about twice the immediately dangerous to life or health
(IDLH) value determined by the National Institute for
Occupational Safety and Health (NIOSH).[3]
An ATR/FTIR absorbance spectrum across the range
700–2400 cm−1 is shown in Figure 5 for a sample of
Table . Quantitative laboratory analysis results for  min sam-
ples collected simultaneously from the tank atmosphere through
a threaded port (about . cm diameter) in the MDI tank.
Aniline, -- (ppm) p-Toluidine, -- (ppm)
. .
. .
JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HYGIENE D17
Figure . ATR/FTIR spectrum obtained for a sample of the solidified
polymer material removed from the tank.
the solidified material removed from the tank. Abundant
absorbance is noted from 2000–2400 cm−1 due to the
continued presence of unreacted isocyanate functional
groups in the solid matrix, which would mask charac-
teristic absorbance peaks expected for a carbodiimide
polymer, if present. The peaks seen at 1705, 1506, and
1407 cm−1 are consistent with the presence of the MDI
isocyanurate trimer,[4] and the peak at 1774 cm−1 is con-
sistent with the presence of the uretidinedione dimer[5]
(Figure 2). Urethane linkages within the solid material are
also suggested by the peak seen at 1223 cm−1 .[5]
Methemoglobinemia and its occupational causes
Methemoglobin is produced through oxidation of heme
iron atoms in hemoglobin from the normal ferrous (Fe2+ )
state to the ferric (Fe3+ ) state. Methemoglobin will not
bind oxygen, and when present as one of the hemoglobin
tetrameric subunits, also causes oxygen to be bound
to the remaining ferrous heme groups more strongly.
The combination of decreased oxygen binding to ferric
heme groups and stronger retention of oxygen in fer-
rous heme groups results in reduced delivery of oxy-
gen to body tissues.[6] Some methemoglobin arises natu-
rally in the body, and several reductive enzyme pathways
exist to maintain high levels of ferrous hemoglobin iron.
However, exposure to strong oxidizers may overwhelm
these protective mechanisms. Signs and symptoms of
methemoglobinemia are related to hypoxia (low oxygen)
and include a blue-gray coloring of the skin (cyanosis),
headache, shortness of breath, dizziness, fatigue, and
the classic chocolate-colored blood. Organ systems, such
as the central nervous system (CNS), with high oxy-
gen requirements are particularly vulnerable.[7] Severe
cases of methemoglobinemia may lead to permanent CNS
damage or death.
D18 P. A. SMITH ET AL.
Medications are the most common cause of methe-
moglobinemia, but some occupational exposures, par-
ticularly to aromatic compounds containing amino
or nitro groups, are also known causes. Occupational
methemoglobinemia cases have been reported in the
chemical, dye, and rubber industries with exposure to
aniline, paranitroaniline, chloroaniline, p-toluidine, and
o-toluidine, among others.[8] There are many published
reports of aniline-induced methemoglobinemia, both
work-related and otherwise.[8-13] Most exposure routes
described in these reports were dermal or dermal plus
inhalation, with resulting methemoglobin levels ranging
from 19–73%. All cases in these reports survived. Sekimpi
and Jones[12] described 325 cases of industrial methe-
moglobinemia that occurred in the United Kingdom from
1961–1980. Of these cases, 30% were due to aniline expo-
sure and 12% to p-toluidine exposure. As over half the
cases occurred during the hot summer months (June–
September), the authors hypothesized that peripheral
vasodilation and decreased use of protective equipment
may have contributed to increased skin absorption of
the chemicals. A recent report by Lee et al.[14] described
two cases of occupational methemoglobinemia from an
MDI manufacturing company in Korea. One case had
inhalational exposure to nitrobenzene, and the other case
had direct skin exposure to liquid aniline. Their methe-
moglobin levels were 19.8% and 46.8%, respectively.[14]
Wellner et al. described the potential for significant der-
mal absorption of aromatic amines (including aniline)
in rubber workers.[15] Bradberry reviewed two published
cases of methemoglobinemia due to p-toluidine exposure,
with a methemoglobin level of 71% in one of the cases.[8]
Exposure assessment, exposure route, and personal
protective equipment
As no chemical known to cause methemoglobinemia
was stored in the tank, an important question to be
addressed with the initial exposure assessment (prior
to entry) related to identification of potential unknown
stressors. The presence of unknown chemical stressors,
such as reaction products in a poorly-understood or
uncontrolled process complicates completion of a cor-
rect exposure assessment;[16] “each potentially hazardous
chemical, physical, and biological agent in the workplace
should be identified.”[17] This includes “the presence of
process off-gases, byproducts, waste products and prod-
ucts of pyrolysis, combustion, or thermal degradation”
which “should be identified to make the inventory of
chemical agents complete.”[17] While not a universal solu-
tion, in this case a proper characterization could have
been completed by using a simple flame ionization detec-
tor (FID) or photoionization detector (PID) in addition
to the multi-gas meter actually used prior to tank entry.
At aniline concentrations near several hundred parts per
million (ppm), either of these detectors would have indi-
cated the presence of an unknown component in the
tank atmosphere. This should have then prompted a
whole-air analysis approach using laboratory-based GC-
MS instrumentation to identify the unknown atmosphere
components.
Through additional work completed after the exposure
it is apparent that the tank entrant was likely exposed to
aniline and p-toluidine vapors by either inhalation or skin
absorption, or through both routes. However, at the time
of the entry the tank was assumed to contain only the
solidified material that was to be mechanically removed.
Subsequent analyses also showed MDI (and other chem-
icals) to be present in the solidified material itself. Expo-
sure routes for MDI include inhalation, ingestion, skin,
and/or eye contact. For work with MDI the U.S. NIOSH
recommends prevention of skin contact and the use of
a supplied-air respirator or and self-contained breathing
apparatus (SCBA), depending on the measured airborne
exposure level. Air purifying respirators are not recom-
mended for MDI exposures at any level.[18]
For both aniline and p-toluidine NIOSH recommends
the use of an SCBA at concentration > 5 ppm, and any
detectable concentration respectively, and prevention
of skin contact.[19,20] The importance of the skin expo-
sure pathway for exposure to liquid aniline and solid
p-toluidine is noted in documentation of the exposure
limits recommended by the American Conference of
Governmental Industrial Hygienists (ACGIH),[21,22] and
skin absorption has been demonstrated as a significant
pathway for aniline vapor.[23] The ACGIH Threshold
Limit Value for exposure to both aniline and p-toluidine
as an 8-hr time-weighted average is 2 ppm, and both
substances carry the “skin” notation.[24] The Threshold
Limit Value for vapor exposure is based on avoidance of
methemoglobinemia.[21,22]
Exothermic MDI polymer formation
Extreme heating of MDI polymer has been shown to pro-
duce aromatic amine methemoglobinemia agents. Ren-
man, Sangö, and Skarping showed that both isocyanates
and amines (including aniline) were evolved from MDI-
based polyurethane foundry mold binders upon heating.
Phenyl isocyanate and aniline were the dominant air-
borne isocyanate/aromatic amine compounds seen in air
samples collected from a laboratory test system used with
MDI-based polyurethane binder for casting molten alu-
minum (740°C) or grey iron (1350°C).[25]
Sato et al. showed that MDI can exothermically self-
polymerize if overheated, and described an incident

in Japan where “an MDI tank burst” in 2007 due to
overheating.[26] The proposed runaway exothermic reac-
tion proceeds via formation of carbodiimide polymer,
with elimination of carbon dioxide which will lead to
an increase in pressure within a sealed tank. Figure 5
ATR/FTIR data obtained for the solidified material in
the present case show evidence for the presence of unre-
acted isocyanate functional groups, MDI uretidinedione
dimers, MDI isocyanurate trimers, and uretonimines,
consistent with overheated MDI material.[26]
Zhang et al. described the thermal degradation of
MDI isocyanurate trimers, which led to the production
of aniline, p-toluidine, benzonitrile, and p-toluonitrile at
a temperature of 580°C.[4] Chambers, Jiricny, and Reese
describe the same products from heating of material con-
taining free MDI and MDI-based polycarbodiimides. Evi-
dence that the MDI-based isocyanurate trimer degrades
at 400°C to produce a solid carbodiimide/isocyanurate
polymer was provided by these authors. At temperatures
above 500°C an MDI-derived polycarbodiimide material
decomposed to give “a complex mix of volatile products”
shown to be, “mainly aniline, benzonitrile, p-toluidine,
and p-toluonitrile.”[27]
While aniline and p-toluidine were easily detected as
vapors within the tank and from headspace sampling
above the heated solid material in the field, evidence in
the field for the presence of the two nitriles was only
seen in the samples collected above the heated material.
Benzonitrile was detected only by carefully examining
expected extracted ion current (m/z 103) in GC-MS chro-
matograms. As the two nitriles are produced from the iso-
cyanurate by heating (although to much higher tempera-
tures than were used to release adsorbed analytes in the
field for SPME sampling, e.g., >500°C[4,27] ), acetonitrile
extraction of the solid material was also performed in the
laboratory. Analysis of the extract using laboratory GC-
MS instrumentation (not described here) confirmed the
presence of these compounds in the solidified material.
The various analyses performed in the field and lab-
oratory indicate the following as a potential explana-
tion for the presence of aniline and p-toluidine vapors
within the tank. The ATR/FTIR data indicate that isocya-
nurate trimers and uretidinedione dimers were present
in the solidified material, and while the dimer forma-
tion is reversible, “formation of the trimer is not eas-
ily reversed.”[28] Sato et al. showed that while not gen-
erally recognized as a hazard, upon excessive heating of
MDI, exothermic self-polymerization is possible.[26] Dow
Chemical Company warns that isocyanurates will form
from MDI at temperatures greater than 140°C, in an
exothermic reaction which will “cause the formation of
carbodiimides and the subsequent formation of carbon
JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HYGIENE D19
dioxide gas that, in a closed or restricted vessel, could lead
to an explosive rupture.” Dow further warns that “when
temperatures reach 230°C or higher, decomposition will
be extremely rapid.”[29]
We postulate that uncontrolled heating near the edge
of the tank initiated exothermic self-polymerization
which further raised the temperature of nearby MDI
material, and a solid polymer plug of complex compo-
sition was formed beginning at the outside edge of the
tank. The thermal mass of MDI liquid above and below
the immediately heated zone moderated the temperature
and confined the continued exothermic polymerization
reaction to the MDI in close contact with the volume
where polymerization was occurring. The detection
of aniline, p-toluidine, benzonitrile, and p-toluonitrile
in the solidified material indicate that extremely high
temperatures may have occurred within the reaction
mass.[6,27] Alternatively, lower temperatures (but suffi-
cient to cause thermolytic cleavage of MDI) could also
have produced phenyl isocyanate and p-tolyl isocyanate,
which were found within the solidified material removed
from the tank. Reaction with atmospheric water vapor
could then have produced aniline, p-toluidine, and car-
bon dioxide gas. The two known methemoglobinemia
agents formed from either of these two possible processes
(or a combination of both processes) were found in the
gas phase within the tank and appear to have caused
the unfortunate medical incident. An explosive pres-
sure release resulting from self-polymerization was not
observed, as the tank was equipped with a pressure relief
valve.
The earlier incident in Japan[26] and this incident
bring attention to the need for careful control of MDI
temperature to avoid self-polymerization. The medical
case discussed in the current setting underscores the
need to fully understand chemical exposures before they
occur in order to protect worker health. The use of a
multi-gas meter limited to detection only of oxygen con-
tent, flammability, and several common toxic gases was
inadequate in this instance to identify the extreme danger
presented by the high-concentrations of aromatic amines
likely present in the tank atmosphere following overheat-
ing and the self-polymerization reaction. In any situation
where an unknown chemical reaction or reactions have
occurred in a confined space prudence requires adequate
testing, and in this case FID or PID sensors would have
indicated an unknown airborne analyte was present
in the tank. This should have then triggered work to
determine the identity of the unknown tank atmosphere
contaminant(s) such as through a whole-air sample
(e.g., in an evacuated canister) for GC-MS analysis in a
laboratory.
D20 P. A. SMITH ET AL.
Conclusions
Incomplete characterization of an unknown atmo-
sphere within a tank which contained overheated MDI
material led to a worker exposure that caused severe
methemoglobinemia. Weeks after the exposure person-
portable GC-MS identified aniline and p-toluidine vapors
within the tank, and quantitative sorbent tube sampling
measured extremely high concentrations of these ana-
lytes in the tank atmosphere. The health outcomes that
resulted from exposure to these compounds were seri-
ous and nearly fatal. A comprehensive characterization
approach is needed for any exposure scenario where
an unknown chemical reaction has occurred, including
an approach to screen for dangerous toxic analytes that
would not be detected by a simple multi-gas meter. Such
an approach is especially important for confined space
entry protocols where an unknown chemical reaction has
occurred.
Disclaimer
The views expressed here are those of the authors. This
document is not a standard or regulation. It creates
no new legal obligations and alters no existing obliga-
tions created by OSHA standards or the Occupational
Safety and Health Act. It may contain recommenda-
tions that are advisory in nature, informational in con-
tent, and are intended to support the provision of safe
and healthful workplaces. Information presented was col-
lected under OSHA’s enforcement authority pursuant to
the Occupational Safety and Health Act [1970]. Per 45
CFR 46.102 under the Protection of Human Subjects rule
[DHHS, 2009], information collected under other gov-
ernment authority is excluded from coverage. Therefore,
this report does not require Institutional Review Board
approval.
References
[1] Aniline: “Method PV2079: OSHA Sampling and Ana-
lytical Methods.” Available at http://www.osha.gov/
dts/sltc/methods/partial/t-pv2079-01-9406-ch/t-pv2079-
01-9406-ch.html (accessed March 7, 2016).
[2] Heikkilä, A.: Field-Portable Fourier Transform Infrared
(FTIR) Spectroscopy for Gas and Vapor Analysis. In
Important Instrumentation and Methods for the Detection
of Chemicals in the Field, P.A. Smith and G.W. Cook
(eds.). Falls Church, VA: American Industrial Hygiene
Association, 2013. pp. 101–113, Ch. 8.
[3] U.S. National Institute for Occupational Safety and
Health (NIOSH): “Chemical Listing and Documenta-
tion of Revised IDLH Values (as of 3/1/95).” Available
at http://www.cdc.gov/niosh/idlh/intridl4.html (accessed
March 7, 2016).
[4] Zhang, J., Tang, Y-F., Liu, J., and Y. Chen: Thermal sta-
bility and thermal degradation reaction kinetics of 4,4 -
diphenylmethane Diisocyanate trimer. Asian J. Chem.
26:1527–1529 (2014).
[5] Dillon, J.G.: Infrared Spectroscopic Atlas of Polyurethanes.
Lancaster, PA: Technomic Publishing Co., Inc., 1989.
[6] Bloom, J.C., A.E. Schade, and J.T. Brandt: Toxic
Responses of the Blood. In Casarett and Doull’s Toxicol-
ogy, the Basic Science of Poisons, 8th ed., C.D. Klaassen
(ed.). New York: McGraw-Hill Education, 2013. pp. 527–
558, Ch. 11.
[7] Moser, V., M. Aschner, R.J. Richardson, and M.A.
Philbert: Toxic Responses of the Nervous System. In
Casarett and Doull’s Toxicology, the Basic Science of Poi-
sons, 8th ed., C.D. Klaassen (ed.). New York: McGraw-
Hill Education, 2013. pp. 733–766, Ch. 16.
[8] Bradberry, S.M.: Occupational methemoglobinemia.
Toxicol Rev. 22:13–27 (2003).
[9] Liao, Y.P., D.Z. Hung, and D.Y. Yang: Hemolytic anemia
after methylene blue therapy for aniline-induced methe-
moglobinemia. Vet. Hum. Toxicol. 44:19–21 (2002).
[10] Demirel, H., V.S. Koster, M.J. Koot, H.H. Ponssen, and
A.C.M. van Vliet: Methemoglobinemia as an uncommon
cause of cyanosis. Neth. J. Med. 55:19–22 (1999).
[11] Phillips, D.M., R. Gradisek, and D.E. Heiselman:
Methemoglobinemia secondary to aniline exposure. Ann.
Emerg. Med. 19:425–429 (1990).
[12] Sekimpi, D.K., and R.D. Jones: Notifications of indus-
trial chemical cyanosis poisoning in the United Kingdom
1961-80. Br. J. Ind. Med. 43:272–279 (1986).
[13] Kearney, T.E., A.S. Manoguerra, and J.V. Dunford, Jr.:
Chemically induced methemoglobinemia from aniline
poisoning. Western J. Med. 140:282–286 (1984).
[14] Lee, C.H., S.H. Kim, D.H. Kwon, K.H. Jang, Y.H.
Chung, and J.D. Moon: Two cases of methemoglobine-
mia induced by the exposure to nitrobenzene and aniline.
Ann. Occup. Environ. Med. 25:31 (2013).
[15] Wellner, T., L. Lüersen, K.H. Schaller, J. Angerer, H.
Drexler, and G. Korinth: Percutaneous absorption of
aromatic amines – a contribution for human health risk
assessment. Food Chem. Toxicol. 46:1960–1968 (2008).
[16] Smith, P.A., M.T.A. Roe, C. Sadowski, and E.D. Lee:
Unknown exposures: gaps in basic characterization
addressed with person-portable gas chromatography-
mass spectrometry instrumentation. J. Occup. Environ.
Hyg. 8:129–138 (2011).
[17] Mulhausen, J., J. Damiano, and E.L. Pullen: Basic Char-
acterization and Information Gathering. In A Strategy for
Assessing and Managing Occupational Exposures, 3rd ed.,
J.S. Ignacio, S.D. Jahn, and W.H. Bullock (eds.). Fairfax,
VA: AIHA Press, 2015. pp. 21–34, Ch. 3.
[18] U.S. National Institute for Occupational Safety and
Health (NIOSH): “NIOSH Pocket Guide to Chemical
Hazards, Methylene Bisphenyl Isocyanate.” Available at
http://www.cdc.gov/niosh/npg/npgd0413.html (accessed
April 22, 2016).
[19] U.S. National Institute for Occupational Safety and
Health (NIOSH): “NIOSH Pocket Guide to Chem-
ical Hazards, Aniline (and Homologs).” Available at

http://www.cdc.gov/niosh/npg/npgd0033.html (accessed
April 22, 2016).
[20] U.S. National Institute for Occupational Safety
and Health (NIOSH): “NIOSH Pocket Guide
to Chemical Hazards, p-Toluidine.” Available at
http://www.cdc.gov/niosh/npg/npgd0624.html (accessed
April 22, 2016).
[21] American Conference of Governmental Industrial
Hygienists (ACGIH): “Aniline” in Documentation of the
TLVs and BEIs. Cincinnati, OH: ACGIH Worldwide,
2001.
[22] American Conference of Governmental Industrial
Hygienists (ACGIH): “p-Toluidine” in Documentation of
the TLVs and BEIs. Cincinnati, OH: ACGIH Worldwide,
2001.
[23] Dutkiewicz, T.: English language summary of “The
absorption and metabolism of aniline in man.” Societas
Scientarium Lodziensis, Lods, Poland (1962).
[24] American Conference of Governmental Industrial
Hygienists (ACGIH): 2016 Threshold Limit Values for
Chemical Substances and Physical Agents and Biological
Exposure Limits. Cincinnati, OH: ACGIH Worldwide,
2016.
JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HYGIENE D21
[25] Renman, L., C. Sangö, and G. Skarping: Deter-
mination of isocyanate and aromatic amine emis-
sions from thermally degraded polyurethanes
in foundries. Am. Ind. Hyg. Assoc. J. 47:621–628
(1986).
[26] Sato, Y., K. Okada, M. Akiyoshi, S. Murayama, and
T. Matsunaga: Diphenylmethane diisocyanate self-
polymerization: thermal hazard evaluation and proof of
runaway reaction in gram scale. J. Loss Prevent. Proc.
24:558–562 (2011).
[27] Chambers, J., J. Jiricny, and C.B. Reese: The thermal
decomposition of polyurethanes and polyisocyanurates.
Fire Mater. 5:133–141 (1981).
[28] Woods, G.: Isocyanates and Polyols for Flexible
Polyurethane foams. In Flexible Polyurethane Foams,
Chemistry and Technology. The Netherlands: Applied
Science Publishers, 1982. p. 4.
[29] The Dow Chemical Company: “Dow Polyurethanes –
MDI Temperature Control Handling Precautions.” Avail-
able at http://dowac.custhelp.com/app/answers/detail/
a_id/13324/∼/dow-polyurethanes—mdi-temperature-
control-handling-precautions (accessed February 18,
2016).