Mo. I dnece mbes Environ Engg and Policy 1 (1999) 235-248 © Springer-Verlag 1999
Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
T.C. Keener, R.K. Jain, R.A. Michaels, J.W. Weber
Abstract ‘his study models and assesses the significance of risks to health potentially posed to individuals exposed to volatile organic compounds (VOCs) due to being occupationally engaged in cleaning petroleum storage tanks. Exposure processes include hydrocarbon vapor inhalation, ingestion via hand-to-mouth contact, and dermal exposure. Under confined space entry regulations, tank cleaners must wear protective clothing and a breathing apparatus. However, such regulations are recent and sometimes violated. Several factors were found to influence exposure of unprotected individuals cleaning tanks. One factor is climate. This paper reports on a study of tanks in south Texas, which is sunny and hot during most of the year. Solar irradiation of metal surfaces may elevate internal temperatures above ambient air temperatures. Vaporization then elevates airborne hydrocarbon concentrations, and ventilation reduces them, producing a dynamic equilibrium. Human entry, however, disturbs this equilibrium by agitating petroleum residuals. Based upon the above considerations, we have modeled petroleum fuel storage tank dynamics in a case study. We considered three cases: manways closed, manways open, and manways actively ventilated. For CASE I, the concentration of each pollutant in the vapor phase was calculated using Raoult’s Law. CASE II and CASE III applied Raoult’s Law and also evaluated a mass balance via a mass transfer coefficient derived using the Reynolds Number and the Schmidt Number. Based upon empirical data, the velocity of the air within the tank was assumed to be 1.6 mph. Outputs of the case study included the steady-state concentration of each constituent and time required to reach it. Health risk assessment was conducted to quantify non-cancer risks posed by individual substances (the hazard quotient, HQ) and by simultaneous and/or sequential exposure to multiple substances (the hazard index, HI). During hot months, gross exceedances of acceptability criteria for acute and
Received: 27 October 1999
T.C. Keener, R.K. Jain (4), J.W. Weber University of Cincinnati, College of Engineering, Cincinnati, OH 45221-0018, USA
R.A. Michaels
RAMT henectady, NY 12309 _ - BAe Seeneee ° Robert A. Mic! AM TRAC Gorporation
S100 ‘ogendae Road Echenectat
dy, NY 12909.
235 chronic exposures to mixed solvent vapors occurred. The HI exceeded the acceptability criterion by more than three orders of magnitude. The HI for chronic occupational neurological risks exceeded the acceptability criterion by more than two orders of magnitude. This result is consistent with observed neurological deficits among tank cleaners, such as depression of performance on memory tests among crew members who had worked during hot months. Finally, in such instances, eight-hour time-weighted average mixed-solvent vapor concentrations can exceed half of lethal levels, suggesting that tank cleaners can potentially approach lethal exposure routinely during hot seasons.
Introduction
Purpose Industry powered by fossil fuels has characterized much of the current millennium. A legacy of this epoch in human history has been global deployment of heating systems, engines powered by petroleum products, and industrial plants using petroleum reactants to synthesize petrochemical products, such as plastics. The global landscape, as a result, has become strewn with the relatively primitive vessels needed to contain petroleum fuels, both above ground and underground, namely, steel tanks. Much attention has been focused in recent years upon leakage of corroded steel tanks as a source of human health risk arising from soil and groundwater contamination. In contrast, relatively little attention has been focused upon health risks posed by direct human exposure to petroleum products contained in steel tanks. To elucidate this issue, we assess the significance of risks to health potentially posed by volatile organic compounds (VOCs) to individuals occupationally engaged in cleaning and refurbishing petroleum storage tanks.
Scope
The size of tanks containing petroleum products is a critical parameter determining human exposure. The most common tanks have low volume, exemplified by automotive fuel tanks. A smaller fraction of tanks have high volume, exemplified by fuel tanks on ships, including tanks powering diesel ship engines, and tanks transporting fuels from refineries to ports where off-loading occurs. The most (in)famous example of the latter is the Exxon Valdez, which spilled its petroleum cargo in Alaska, damaging sensitive coastal ecosystems. Routinely,
Shasls, FhD., ceF|however, off-loading is accomplished via pipelines trans-
Environ Engg and Policy 1 (1999)
porting petroleum products to coastal tank farms for storage prior to distribution, such as by truck to inland consumers. Here we focus upon tank farms, and large tanks, defined as tanks which are large relative to people.
Within the wide range of tank sizes, we differentiate simply between small and large tanks because, historically, small tanks more often have been discarded than cleaned, whereas large tanks typically are cleaned occasionally, and their cleaning has required human entry into the tanks. Human entry into large petroleum storage tanks has been necessary when tanks storing a particular product, such as crude oil, have been rededicated to storage of a different product, such as heating oil, gasoline, or jet fuel. Here we focus upon such large storage tanks, which typically are located within tank farms consisting of multiple tanks, which together are capable of storing diverse petroleum products simultaneously.
This study addresses the case of crude oil storage tanks normally varying in 70-200 ft. in diameter; an enclosure may be considered any area where air flow is constricted. Eight of the most volatile crude oil components are addressed. The methodology developed, however, may be applied very similarly to other situations (such as a worker’s exposure to paint fumes in a paint room). More specifically, this study provides a paradigm for industries (such as crude oil storage and refining) which often must manage risk potentially posed by worker exposure to dangerous volatile organics. The tank bottom constituents were analyzed and worker exposure for various cases was studied.
Findings
The tank cleaning process
The process of cleaning large petroleum storage tanks was found to include a variety of alternative procedures, depending upon tank size and product stored (Lillienberg et al. 1992; Midzenski 1992; Powers 1990). In general, however, cleaning begins with draining and thereby reducing the depth of the residual petroleum product, and exposing deposits known as tank bottoms. Tank bottoms consist of preferentially deposited, relatively heavy constituents of the parent petroleum product(s). Petroleum product depth initially may be reduced via an external pump connected to a floor drain within the storage tank, In recent years, remotely controlled mobile skimmers also have been used to remove lighter, less viscous constituents, until a relatively solid sludge is exposed.
Eventually, human entry into tanks is necessary to complete cleaning. Teams of people equipped with squeegees push remaining liquids and viscous sludges toward the floor drain/pump intake. During this process they may be exposed to petroleum products. Exposure pathways include hydrocarbon vapor inhalation, ingestion via hand-to-mouth contact, and dermal exposure. To protect against such exposures, confined space entry regulations of the U.S. Occupational Safety and Health Administration (OSHA), and similar regulations in other nations, require tank cleaners to wear protective clothing and
breathing apparatus (Lillienberg et al. 1992). However, such regulations are of relatively recent vintage, and sometimes have been violated (U.S. OSHA 1993). For most if not all of the petroleum fuel era, therefore, tank cleaners have been subject to acute and chronic risks potentially posed by exposure to VOCs in petroleum products stored in tanks.
Factors influencing exposure
Several factors have been identified which influence exposure of unprotected or inadequately protected individuals cleaning tanks from within. One factor is ambient conditions. We studied tanks located in south Texas, a region which is sunny and hot during most of the year. Ambient air temperature imposes a baseline condition on internal tank temperature, which in turn drives vaporization of hydrocarbon components of petroleum products inside the tanks. Superimposed on this baseline is the effect of solar irradiation on the metal surfaces of the tank, specifically, the roof and cylindrical walls. These structures may absorb solar energy and radiate it as heat within the tank. Consequently, internal temperatures may exceed ambient air temperatures.
Prior to human entry in the cleaning process, manways may be opened and air-driven (non-sparking) fans operated to reduce the concentrations of accumulated hydrocarbon vapors. After a period of such ventilation, perhaps one week, human entry is undertaken. Vaporization will continue to elevate airborne hydrocarbon concentrations, and ventilation will continue to reduce them, producing a dynamic equilibrium characteristic of the temperature and ventilation conditions. Once human entry occurs, however, this equilibrium is disturbed by two processes which agitate the surface of the petroleum residuals: the action of booted feet walking through the tank, and the action of squeegees further agitating the liquid hydrocarbon surface. These sources of agitation may increase the rate of hydrocarbon vaporization, whereas the ventilation rate remains constant, producing a net increase in airborne hydrocarbon concentrations.
Based upon the above considerations, we have modeled petroleum fuel storage tank dynamics in a case study. We consider three cases: manways closed, manways open, and manways actively ventilated. The case study is set forth below.
Case study: fuel storage tanks
In the crude oil storage industry, fuel tanks often vary in size. Some of the larger tanks are, 250,000 barrel tanks (56 ft. in height x 180 ft. in diameter), while smaller tanks may be only 26,000 barrel tanks (40 ft. in height x 70 ft. in diameter). Most storage tanks have floating roofs that float on the liquid stored in the tank, but descend only partially towards the tank floor when the tank is emptied, allowing human entry for cleaning. For the purpose of this study, a tank diameter of 180 ft. and roof height of 7 ft. will be used. For smaller diameter tanks, pollutant concentrations would be similar to the larger diameter tanks or possibly somewhat higher due to. reduced indoor air volume. The typical tank has five
T.C. Keener et al.: Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
entrances, called manways which may either be closed or open. The center of a typical manway is located about 2.5 ft. off the ground and the diameter of the manway is about three feet.
Basic assumptions
In the first case considered, CASE I, the manways of the tank are closed. In addition, it is assumed that the tank is at steady state conditions. This is expected because the tank will be saturated with each of the volatile pollutants continuously as the tank is drained. Thus, by the time the tank is empty, the air above the sludge will be saturated. In the two other cases discussed, CASE II and CASE III, the manways are open providing ventilation. In addition, in CASE III the fan is on.
For all three cases, it will be assumed that the storage tank is drained of most of its oil product leaving behind only a bottom sludge about 11 inches deep. For a typical storage tank, there is often a cleaning period when unused sludge is removed and disposed of before refilling. The nature of the cleaning process may also require direct worker involvement. Thus, this situation is a good example for a human exposure study.
Typical oil tank sludge is made up of layers of hydrocarbons. In this case, the bottom three to four inches of the tank consisted of a sticky, slimy, tar-like sludge containing dirt, water and the heaviest crude components. This bottom few inches does not contribute significantly to the vapor phase concentration of the pollutants. The top six to eight inches of the sludge, however, is referred to as the “compromised product” and it does contribute significantly to volatilization into the vapor phase. This is because, although it is the heaviest of the crude that had been stored in the tank, it is still a liquid. In fact, the compromised product is refinable.
Typically, the tank has a floating roof about seven feet high when empty, so the height of the air column in the tank is simply seven feet minus the 11 inches of sludge. In addition, the air and the sludge are assumed to be completely mixed. In the compromised product, the pollutant concentrations are taken from lab test data of typical “sweet” crude oil.
Because evaporation occurs from the surface layer of the liquid, evaporation of the hydrocarbons will be controlled predominantly by the temperature of the atmosphere above the hydrocarbons, which is the . temperature that will equilibrate with the atmosphere
above it. The air temperature in the tank is expected to be at or above the atmospheric temperature. Thus, to be conservative, temperature ranges were chosen which would be typical for the southern United States where higher temperatures would mean more pollutant volatility. Extreme average temperatures were used in order to determine the range. The low temperature was assumed to be about 46°F (7.8°C) which is a typical average night time temperature in January in the southern U.S. The high temperature was assumed to be about 94 F (34.6 °C) which is the average daytime temperature in July in the southern U.S. Finally, the air pressure of the tank is assumed to be at 1 atmosphere.
One of the most important assumptions is the molecular weight of the compromised product. The pollutants make up about 1-2% of the compromised product. To estimate the molecular weight, it is important to know what comprises the other 98-99%. For the purposes of this case study, it is assumed that the rest of the compromised product is straight chain hydrocarbons containing on average 40 carbons in a chain. As shown in Table 1, a conservative assumption is that the compromised product falls in the heavy gas oil category which on average contains 40 carbons per molecule.
Assuming that the compromised product is made of straight chain hydrocarbons with an average of 40 carbons (and thus 82 hydrogens), the molecular weight of the compromised product is determined to be 562 g/mol.
Basic assumptions for CASE Il and Ill
The air is assumed to be completely mixed and initially saturated. It is expected that the air in the tank will be initially saturated because it will have been in use for a while before it had been ready to be cleaned. Thus, there was ample time for the pollutants to enter the vapor phase and saturate the tank air. Therefore, CASE II and CASE III begin with the CASE I situation (a closed tank.) At the start time (t=0 s) the manways are opened and in CASE III the fan is turned on. The models will show how pollutant concentrations decrease from saturation to the final steady state value.
For purposes of calculating a mass transfer coefficient, values for the density and viscosity of clean air are used. As shown in CASE I, the air in the tank contains some amount of each pollutant. However, it is expected that these small pollutant amounts in the air will have little effect on the density and viscosity of the air.
Table 1. Typical crude oil fractions (data compiled from McCain 1973)
Crude fractions Chemical composition Use
Hydrocarbon gases C\-Cy
Petroleum ether C5~Cg Gasoline C,-Cg Kerosene Cro-Ci6 Light gas oil Cis-C30 Heavy gas oil C3o—Cso Residuum (residue) Ceo+
Natural gas, bottled fuel gas
Solvent, paint thinner, cleaner
Motor fuel, solvent
Illuminating oil, diesel fuel, jet fuel, cracking stock
Lubricating oil, medicinal oil, transformer oil, mineral oil, cracking stock Lubricating oil, bunker fuel, road oil
Tars, asphalts, bitumens, waxes, resins, pitch, mineral oil, wood preservatives, roofing compounds, road oil, paving asphalts, coke
Environ Engg and Policy 1 (1999)
Values of diffusivity in air for benzene are used. From this compound a relationship is developed relating the actual concentration in CASE’s II and III to the saturation concentration developed for CASE I. This model shows the concentration of pollutants during air flow as a percentage of saturation.
Methodology for determining the saturation vapor concentration
For CASE I, the concentration of each pollutant in the vapor phase is calculated using Raoult’s Law.
P=P(v)X (1)
Raoult’s Law states that the partial pressure of a substance (P) in the vapor phase is equal to the vapor pressure of that substance P(v) times the mole fraction of the substance in the liquid phase (X).
In this case study, pollutant liquid concentration values are actual values provided from a laboratory analysis of sweet crude oil. The liquid concentration values are expressed in mg of pollutant per kg of compromised product. An example of the calculations used are given in the Appendix.
Methodology for determining the vapor concentration (for
CASE II and CASE III) In CASE II and III the manways are left open allowing air
flow. Performing a mass balance on a storage tank yields the following:
Change in constituent mass with time = Flow in (of pollutant) - Flow out + Input (from sludge)
or
d(VC,
FO 20 QC, RA (CiseeGi) (2)
where:
0 = flow in, because it is assumed that the flow into the tank is clean air,
d(VC;) . sap gs
a the change in the mass of pollutant with time,
QC, = flow out or the amount of mass leaving
through the manways, kA (C;,at~C:) =input or the amount of mass added into the system from the sludge, Vv = Total air volume,
CQ = Actual pollutant concentration (From this equation C, is directly related to C; sat),
C,.a¢ = Saturation concentration of the pollutant (determined for each pollutant in CASE I.),
A = Surface area of the Sludge,
k = mass transfer coefficient.
Pollutant concentrations for CASE |: closed tank, steady state conditions
Vapor concentrations for eight pollutants are analyzed. The liquid concentration values are taken from a labora-
Table 2. Mole fraction
Constituent Liquid Molecular Calculated concentration weight mole (mg/kg) (g/mol) fraction Benzene 2,229 78.11 0.01604 Toluene 5,114 92.13 0.03120 Ethylbenzene 786 106.16 0.00416 Xylenes 4,757 106.16 0.02518 1,2,4-trimethylbenzene 1,843 120.19 0.00862 1,3,5-trimethylbenzene 1,400 120.19 0.00655 naphthalene 1,114 128.16 0.00488 2-methylnaphthalene 2,157 142.21 0.00852
Table 3. Vapor and partial pressure
Constituent Vapor pressure
(mm Hg) (mm HG)
at 7.8°C at 34.6°C at 7.8°C at 34.6°C Benzene 40.30 117.41 0.6464 1.8832 Toluene 10.87 45.55 0.3391 1.4212 Ethylbenzene 3,32 16.21 0.0138 0.0674 Xylenes 2.70 13.57 0.0680 0.3417 1,2,4-trimethylbenzene 0.70 4.02 0.0060 0.0346 1,3,5-trimethylbenzene 0.70 4.02 0.0046 0.0263 naphthalene 0.07 0.52 0.0003 0.0025 2-methylnaphthalene 0.01 0.13 0.0001 0.0011
tory analysis of sweet crude oil and are expressed in mg of pollutant per kg of compromised product. Below, Table 2 shows the mean concentration, molecular weight and calculated mole fraction for each constituent.
Following the methodology for CASE I, the partial pressure in the vapor phase was calculated for each constituent. Table 3 shows the vapor pressure of each constituent and the corresponding partial pressure in the vapor at the two temperature extremes.
Finally, using the ideal gas law, the concentration in air of each constituent can be found as shown in Table 4.
Required saturation concentrations It is necessary to make sure there is enough of each pollutant in the liquid in order to saturate the vapor. To find the mass of each pollutant in the vapor at saturation, the concentration is multiplied by the volume of the air in the tank where V = 4,366.6 cubic meters.
To find the mass of each constituent in the compromised product (CP) the following equation is used:
kgs of constituent = Mean Conc. x Volume of CP x Density of CP (3)
The mean concentration is taken from Table 2 and the volume of the compromised product is the crosssectional area of the tank times the height of the CP (7 inches). The density of the compromised product is not known although it ranges from 0.66 to 1.08 g/cm®., Since the CP is at the bottom of the tank, the density was
'T.C. Keener et al.: Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
Table 4, State concentration and PPM
Parts per million
Constituent Steady st state concentration (mg/m? ) at 7.8 °C at 34.6°C at 7.8°C at 34.6°C Benzene 2,882.26 7,674.89 850.20 2,478.54 Toluene 1,778.11 6,826.83 444,69 1,870.27 Ethylbenzene 83.87 372.62 18.20 88.59 Xylenes 411.90 1,889.65 89.40 449,27 1,2,4-trimethyl benzene 41.10 216.34 7.88 45.43 1,3,5-trimethyl benzene 31.49 164.66 6.04 34.58 naphthalene 2.19 16.66 0.39 3.28 239 2-methyl naphthalene 0.81 8.16 0.13 1.45 —
assumed to be close to 1. Therefore, Table 6 shows the amount of each constituent available in the liquid assuming a density of 0.85 g/cm? and 1.0 g/cm?
Comparing Tables 5 and 6 clearly shows that there is plenty of each constituent available in the liquid CP. This is important because it verifies that the air will be saturated with each pollutant for CASE I. In addition, since there is anywhere from 30 to 10,000 times the amount of each substance in the CP as in the vapor, smaller tanks will also be completely saturated.
CASE Il and CASE Ill: air flow, steady state conditions d(VC,)
dt Eq. (2) from the methodology for CASE II and CASE III,
As time goes to infinity, — becomes zero. Using
Table 5. Mass of each constituent in the vapor
the steady state concentration of each pollutant can be found as given by the following equation:
kA AC, sat
(Q+kA “
1=
Mass transfer coefficient
For mass transfer from a flat surface where the boundary layer is laminar (which occurs if the local Reynolds number is leass than 300,000; the value is assumed to never be achieved in oil storage tanks due to the tank geometry and other physical limitations), the mass transfer coefficient has been determined to be the following (Cassler 1997):
k= (Did) 0.646 (*); (3 }§ 2 \pD
where: = the diffusivity of benzene in air (0.096 cm/s), = the diameter of the tank (54.86 meters),
3 (5)
ie"
) = the Reynolds Number,
}
= the Schmidt Number,
Ble e l=
= the density of air (1.2 kg/m®),
= the viscosity of air (1.75 x 10°° kg/m-s),
__ the velocity of the air in the tank
~ (1.6 mph=0.72 m/s) (Keener et al 1998). The velocity of the air within the tank is assumed to be 1.6 mph based on typical wind velocity within a storage tank. Comparisons of several wind velocity
;
v
Constituent Kilograms of each constituent in the air at saturation conditions at 7.8°C at 346°C
Benzene 12. 586 33. 513
Toluene 7.764 29.810
Ethylbenzene 0.366 1.627
Xylenes 1.799 8.251
1,2,4-trimethyl benzene 0.179 0.945
1,3,5-trimethyl benzene 0.138 0.719
naphthalene 0.010 0.073
2-methyl naphthalene 0.004 0.036
Table 6. Mass of each consti- Constituent
tuent in liquid
Benzene Toluene Ethylbenzene Xylenes
1,2,4-trimethyl benzene 1,3,5-trimethyl benzene naphthalene
2-methyl naphthalene
Kilograms of each constituent ii the CP
Density = 0.85 giem?
Density = 1.0 glem?
795.94 936.40 1,826.13 2,148.39 280.67 330.20 1,698.65 1,998.42 658.11 774.24 c ,499.92 588.14 397.79 467.99 770.23 906.16
Environ Engg and Policy 1 (1999)
patterns showed that data varied significantly for each part of the tank and for each case (fan on and off) so the wind velocity within the tank is taken as 1.6 mph for all parts of the tank and for both cases.
For the values given, k is computed as 0.0219 cm/s, which is taken as the average value of the mass transfer coefficient over the entire liquid surface. Higher air velocities will result in higher average values of the mass transfer coefficient which would result in higher average air concentrations of VOCs within the tank.
Sludge area and flow rate
The surface area of the sludge is determined to be
2,364 m7. Based on typical wind velocity data, it is assumed that in both CASE II and CASE III, only one manway will serve as the air exit. In CASE II (fan off) the flow through the tank is primarily in one manway and out the other one directly across the tank, with the other two manways having little effect on ventilation. In CASE III (fan on) the flow typically comes in two manways and exits through the fan manway. Thus, the volumetric flow rate, Q, is computed as follows:
Q=VAn
where:
V_ = the air velocity through the exit manway,
Am = the cross-sectional area of the exit manway = 0.657 m? (each manway is 3 feet in diameter).
When the fan is off (CASE II), V is estimated as
5 mph=0.224 m/s
When the fan is on (CASE III), V is estimated as
17 mph=0.760 m/s (both estimations are made from
typical storage tank wind data.)
Thus, CASE II: Q=1.47 m?/s while for CASE III:
Q=4.99 m/s
Steady state concentration
The steady state concentrations can now be evaluated by the use of Eq. (3). These are given in Table 7. From Table 7, it is shown that for steady conditions with the manways open, the concentration of each constituent is about 1/4 the saturation concentration calculated for Table 6. After a long time with the manways open and the fan on, the concentration of each constituent is about 1/10 the saturation concentration calculated in Table 6.
CASE II and CASE III: air flow, non-steady state conditions Although Table 7 provides the final concentration of each constituent for CASE II and CASE III, a non-steady state mass balance must be solved to find out how long it will take the system to reach steady state. Referring again to Eq. (2):
Table 7. State concentration
Case Steady state concentration CASE II C, = 0.260 Cisat Cy, =0.094 Cysat
CASE III
a(vc;)
dt =0-— QC, +KA (Ci sar Ci)
The unsteady state solution to this equation assuming that C;=Cj.a at initial conditions is the following:
oe row Cr KAC sae | Q+kA QCr sat
Multiplying by Q+kA and taking the exponential of both side gives:
(Q4KA) C)—KAC oor _ 5 -cavnarvys QC sat
Finally, solving for C, gives:
a QE) Ci sat FRAC sat ‘ Q+kA
—tV . (6)
(7)
Equation (7) provides the relation of C, to Cys: at any time, t. Also, as t becomes large, the exponential term goes to zero and the equation becomes the steady state Eq. (3). A, k, and Q are the same as before.
The only new value in Eq. (7) is the volume of the air which is:
V=AH=4.383 x 10° m?
where H is the height of the air column and is equal to 73 inches (0.185 m.)
Figures 1 and 2 show the relationship of the actual concentration (C,) to the saturation concentration (C;sat) with respect to time for each of the two cases. Figure 1 indicates that it takes about 200 minutes to reach steady state with the manways open and the fan off. Figure 2 indicates that it takes about 50 minutes to reach steady state with the manways open and the fan on.
CASE II: Manways open, fan off
2 © o 8 6 |
|
~ 3
60 507 \
aot \
30 sew
Concentration (in % of saturation concentration - C1sat)
°
a
0 200 Time (min)
vo 400 600
Fig. 1. a CASE III: manways open, fan on. b CASE II: Manways open, fan off
1.C. Keener et al.: Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
CASE III: Manways open, fan on
$1007 ———
Lo}
2 i
i 80+\
\
5 |\
g TN
i \
§ \
s 20 \
; N
Q b ~ a
© % 100 200 Time (min)
Fig. 2. a CASE III: manways open, fan on. b CASE II: Manways open, fan off
Potential risks to health
Table 8 summarizes the estimated concentration of each constituent along with the corresponding National Institute for Occupational Safety and Health (NIOSH) standard. The standard for each pollutant is taken from the Registry of Toxic Effects of Chemical Substances
(RTECS), compiled and updated by NIOSH. NIOSH is one of several institutions which promulgate or recommend time-weighted average (TWA) concentration limits (Threshold Limit Values, or TLVs) limiting for occupational exposure to selected substances. Others include US OSHA, the American Conference of Governmental Industrial Hygienists (ACGIH), and many similar agencies in other countries.
US OSHA, NIOSH, ACGIH, and other groups that recommend or promulgate eight-hour TLV values do so to set upper limits for occupational exposure to individual substances, not multiple substances. When simultaneous and/or sequential exposure to multiple substances occurs, occupational limits may be inadequately protective. Several studies illustrate this point by reporting toxic effects which occur in multiple solvent exposure situations when airborne concentrations of individual solvents were within the range of applicable TLVs (Cherry et al. 1984; Ekberg et al. 1986; Elofsson et al. 1980; Gregersen et al. 1984; Seppaldinen et al. 1978). Therefore, for exposure to multiple substances, individual TLV values must be reduced to maintain mixed airborne chemical concentrations within safe limits.
Standard procedure in health risk assessment quantifies non-cancer risks posed by individual substances via the ‘hazard quotient’, (HQ) which is the ratio of exposure divided by an applicable toxicological benchmark. HQ values exceeding unity (one) typically are deemed unacceptable relative to the selected benchmark. A variety of acceptability benchmarks may be selected, depending
Score (age/educ. norm = 10)
Fig. 3. Memory assessment scales:
IPR Memory Assessment Scale (MAS) parameter
VSep VSip LA LR DLR DPR __INF
Table 8. Concentration summary and NIOSH standards
Constituent Range of concentration
for CASE I (ppm)*
850-2480 220-640
Benzene
Toluene 450-1870 120-490 Ethylbenzene 18-88 5-23 Xylenes 90-450 23-120 1,2,4-trimethyl benzene 8-45 2-12 1,3,5-trimethyl benzene 6-34 2-9 naphthalene 0.4-3 0.1-0.8 2-methyl naphthalene 0.1-1.5 0.03--0.4
“the lowest value in the range is the concentration at 7.8°C and time weighted average
the highest value in the range is the concentration at 34.6 °C. Concentrations are all at steady state and have been rounded off to the nearest whole number
DNF VR
Range of concentration for CASE II (ppm)*
scores of 11 plaintiffs on 12 global memory subtests
IVR DVR
Range of concentration
NIOSH standard (ppm) for CASE III (ppm)* .
80-230 8-hour: 0.1° 15 min: 1° 40-180 10-hour: 100° 2-8 See Xylene 8-40 10-hour: 100° 0.8-4 NA 0.6-3 NA 0.04-0.3 10-hour: 10° 10-hour: 10°
0.01-0.1
*15 minute maximum NA Not Available
Environ Engg and Policy 1 (1999)
upon the focus of investigation. For example, OSHA’s Threshold Limit Value (TLV) may be used to determine whether a ‘hygienic risk’ exists. In that case the HQ value quantifies the degree to which an exposure approaches, or exceeds, an occupationally acceptable exposure level. Exposure levels acceptable for members of the general population typically are quantified via EPA’s risk reference dose (RfD) for ingested substances, or EPA’s risk reference concentration (RfC) for inhaled substances.
Intrinsic to the above HQ parameters are safety factors, quantifying the distance between a non-toxic vs. a toxic dose. Safety factors, therefore, are the toxicology equivalent of pharmacology’s ‘margin of safety’, quantifying the distance between a therapeutic vs. a toxic dose. Safety factors may be removed, however, to assure that HQ values exceeding unity represent toxic doses, rather than just exceedences of desirable safety factors.
Risks potentially posed by simultaneous and/or sequential exposure to multiple substances are quantified via the hazard index (HI), which is the sum of HQ values for all substances of potential concern. Values of HI exceeding unity typically are deemed unacceptable. Applying this procedure to airborne concentration values typical of the hotter months in south Texas (the higher values) and NIOSH TLVs set forth in Table 8 yields the industrial hygiene (‘hygienic’) HQ and HI values set forth in Table 9. As suggested above, alternatives to NIOSH benchmarks exist. One alternative is exemplified in Table 9, specifically, chronic occupational neurological risk reference concentration (RfC) values. Chronic exposures are exposures which occur over a period of months or longer. .
Another type of toxicological benchmark is relevant in
view of the results set forth in Table 9. Table 9 reveals gross exceedances of acceptability criteria for chronic exposures to mixed solvent vapors. That is, the hygienic HI exceeded the acceptability criterion of unity by a factor of 6,406 and the HI for chronic occupational neurological risks exceeded the same acceptability criterion by a factor of 199. These exceedances raise the issue of whether acute toxicity benchmarks also might have been exceeded. Acute exposure refers to exposure occurring within a period of 24 hours or, in the occupational context, within a single shift (normally, eight hours). Accordingly, Table 10 quantifies acute lethality risks potentially posed to individuals occupationally exposed to hydrocarbon vapors while cleaning petroleum storage tanks during hot weather.
Derivation of risk reference concentrations
a. Benzene
Chronic occupational exposure, neurological effects. ‘'wo epidemiological studies are relevant to deriving an inhalation neurological effects occupational RfC for benzene. In one study, an airborne benzene concentration of
210 ppm-v (671 mg/m*) in the workplace was associated with neurological abnormalities (Baslo and Aksoy 1982, cited in ATSDR 1997a). However, toluene was also present in the workplace. In the other study, benzene at 20-50 mg/m? was associated with neurological disorders, and competing toluene levels were relatively low (Kahn and Muzyka 1973, cited in ATSDR 1997a). Inasmuch as this value accords with the ACGIH TWA of 32 mg/m? (10 ppm-v) reported above, that value also may be used as an‘ occupational neurological effects RfC.
Table 9. Potential exposure of unprotected tank cleaners in hot climates to hydrocarbon vapors compared with industrial
hygienic and neurological risk benchmarks
analyte CAS concentration no. case 2 manways open (ppm-v) benzene and alkyl benzenes benzene 71-43-2 640 toluene 108-88-3 490 ethylbenzene 100--41-4 23 xylenes 1330-20-7 120 1,2,4-trimethyl-benzene 95-63-6 12 1,3,5-trimethyl-benzene 108--67-6 9 naphthalene and alkyl naphthalenes . naphthalene 91-20-3 0.8 2-methyl-naphthalene 91-57-6 0.4
hazard index (HI)*
* NIOSH TWA TLV unavailable for ethylbenzene. Value for toluene and xylenes was used based upon similarity of chemical structure
> See text for derivations
“ The hygienic HI value quantifies the degree of exceedance of acceptable occupational exposures to mixed solvent vapors according to industrial hygiene criteria recommended by the
hygienic neurological NIOSH hazard chronic, hazard TWA TLV* quotient occupational quotient (HQ) RFC (HQ) (ppm-v) on (ppm-v) os 0.1 6400 10 64.0 100 4.90 5 98.0 100 0.23 4 5:75 100 1.20 4 30.0 32 0.38 "32 0.28 10 0.08 3 0.27 10 0.04 3 0.12 6406 199
U.S. National Institute of Occupational Safety and Health (NIOSH). The HI value for the chronic occupational neurological risk reference concentration (RfC) quantifies the degree to which occupational exposure of tank cleaners to mixed solvent vapors via inhalation over a period of months or longer exceeds exposure limits which would be expected to protect against adverse neurological effects
TC. Keener et al.: Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
Table 10. Potential exposure of unprotected tank cleaners in hot climates to hydrocarbon vapors compared with lethality risk benchmarks
analyte CAS concentration 8-h lethality 8-h lethality no. case 2 ——_ — = - manways occupational, hazard RfC, with hazard open no safety quotient safety quotient margin* (HQ) margin* (HQ) (ppm-v) (ppm-v) te (ppm-v) . benzene and alkyl benzenes benzene 71-43-2 640 1949 0.33 20 32.0 toluene 108-88-3 490 3295 0.15 33 14.9 243 ethylbenzene 100-41~4 23 1725 0.01 24 0.96 — xylenes 1330-20-7 120 4331 0.03 137 0.88 1,2,4-trimethyl-benzene 95--63-6 12 1831 0.01 18 0.67 1,3,5-trimethyl-benzene 108-67-6 9 1831 0.00 18 0.50 naphthalene and alkyl naphthalenes naphthalene 91-20-3 0.8 734 0.00 15 0.05 2-methyl-naphthalene 91-57-6 0.4 301 0.00 6 0.07 hazard index (HI) we 0.53 we 50.0 * See text for derivations shorter breaksduring morning and afternoon portions of. the > The 8-h occupational HI value without safety factors work shift. The HI for the 8-h lethality risk reference concentra-
compares the 8-h time-weighted average mixed solvent concen- _ tion (RfC) quantifies the degree to which exposures of tank tration to which unprotected tank cleaners working during hot cleaners would exceed concentrations which would protect
weather might be exposed via inhalation with concentrations against lethality with a margin of safety. Such exposure, even if which would be expected to be lethal following eight hours of non-lethal, would not be acceptable unless it also reliably continuous in-tank exposure. Actual patterns of occupational protected against adverse toxic effects other than lethality
exposure would be expected to include a lunch break and
Table 11. Memory assessment scales (MAS): plaintiff global memory subtest scores
indiv. VS ep VS i ip LA IR DLR IPR DPR INE DNF VR IVR DVR acronyms and abbreviations
1 5 6 7 10 8 9 9 ll 9 ll 12 ul VS ep verbal span
2 5 5 10 9 10 9 15 8 7 6 9 10 VS ip visual span
3 6 9 6 5 9 9 8 120 «11 7 ll 9 LA list acquisition
4 6 6 7 10 7 7 8 ll ll 9 10 9 IR list recall
5 1 6 6 7 9 10 9 7 9 7 9 4 DLR delayed list recall
6 8 6 13 8 #8 7 8 7 7 9 5 6 IPR _ immediate prose recall
7 11 6 8 12 9 6 6 8 9 5.5 3 DPR delayed prose recall
8 4 9 9 7 9 10 10 6 5 6 5 2 INF __ immediate names-faces
9 9 6 7 4 6 7 8 8 7 30 9 7 DNF delayed names-faces 10 2 6 4 5 5 3 4 2 3 13, 10 )«11 VR visual reproduction ll 9 6 5 4 4 3 4 3 6 5 10 7 IVR immediate visual recognition age/ 10 10 10 10 10 10 10 10 10 10 10 10 DVR __ delayed visual recognition education norm
Lethality, acute occupational exposure. Examination of good models for use in extrapolating to humans, revealed multiple acute toxicity reports produces an eight-hour a lethality value (LCs) of 31,947 mg/m®. Examining the inhalation lethality benchmark of 3,981 mg/m?* (6,226 mg/ data holistically suggests a four-hour lethality benchmark m? if a ‘mammal’ and ‘human’ report are identital) based of 12,451 mg/m’, with uncertainties around that value to
upon the geometric mean of the product of dose and be overcome by application of safety/uncertainty factors. exposure duration (RTECS 1997). A five-minute exposure Thus, an inhalation lethality RfC for acute occupational to an estimated airborne benzene concentration of exposure to benzene may be derived by applying two 63,894 mg/m? was reported to be lethal to humans. This _safety/uncertainty factors of 10 to extrapolate from an would suggest a four-hour lethality concentration of effect level to a lethality NOAEL (LC), and from animals 1,331 mg/m*, assuming validity of Haber’s rule (dose x to humans (because only one of the reports considered exposure duration = constant) for this substance over above involved humans). This yields RfC = 125 mg/m°, this exposure duration range. However, a four-hour which may be equated with an eight-hour shift value of
bioassay involving rats, which are widely recognized as 63 mg/m’.
Environ Engg and Policy 1 (1999)
b. Toluene
Chronic occupational exposure, neurological effects. The neurological effects inhalation RfC may be derived based upon the U. S. EPA’s inhalation RC value of 0.4 mg/m? (0.1 ppm-v; IRIS 1997). This value was derived based upon bioassay and occupational studies in which neurological effects were reported. The occupational RfC may be upwardly adjusted by multiplying it by 24/8 h/h and 7/ 5 d/d to reflect shorter-term exposure in the workplace, and by a factor of 10 to reflect relatively healthy individuals in the workplace. This yields R(C=17 mg/m?
(5 ppm-v).
Lethality, acute occupational exposure. Examination of multiple acute toxicity reports produces an eight-hour inhalation lethality benchmark of 12,415 mg/m? based upon the geometric mean of several bioassays (RTECS 1997). Although this value will be only approximate with respect to toluene lethality to humans, it is likely to be sufficiently accurate to justify deriving a lethality RfC, which should overcome uncertainties by applying two safety/uncertainty factors of 10 to extrapolate from an effect level to a lethality NOAEL (LC,), and from animals to humans. This yields 4-h Rf{C=248 mg/m’, equivalent to 8-h RfC= 124 mg/m’, assuming validity of Haber’s rule (concentration x exposure duration = constant) for this substance over this exposure duration range. This value may be compared with the Short-Term Exposure Limit (STEL) of 150 ppm-v (565 mg/m?) promulgated by NIOSH, the Ceiling Limit of 300 ppm-v (1,131 mg/m*) promulgated by OSHA, and the peak limit of 500 ppm-v (1,884 mg/m*) for 10-minute exposure also promulgated by U. S. OSHA.
c. Ethylbenzene
Chronic occupational exposure, neurological
effects. Neurological effects constitute an effect category shared by many groups of organic solvents. According to Andrews and Snyder (1991):
“Many organic solvents, including hydrocarbons, chlorinated hydrocarbons, alcohols, ethers, esters, and ketones, have the potential on acute high-level vapor exposure to cause narcosis and death. A typical exposure scenario is the worker who enters a reaction vessel or holding tank without appropriate respiratory equipment. In such a confined space, solvent vapor concentrations may reach many hundreds or thousands of parts per million, and the worker may be quickly overcome. Of course, the experimental animal analogy to this scenario is the acute inhalation toxicity study (LCs).
Workers exposed to solvents under these conditions will typically show signs of central nervous system disturbance. While there is some variation in signs and symptoms with solvent structure, results of high-level exposure are quite similar. The scenario of disorientation, euphoria, giddiness, confusion, progressing to unconsciousness, paralysis, convulsion, and death from respiratory or cardiovascular arrest is typically observed... In the majority of subjects, recovery from central nervous
system effects is rapid and complete following removal from exposure...
The similarity of the narcosis produced by solvents of diverse structure suggests that these effects result from a physical interaction of a solvent with cells of the central nervous system. If a purely physical interaction is assumed, then the narcotic effect of the solvent will depend only on the molar concentration of the solvent in the central nervous system cell. Equimolar concentrations of different solvents will result in narcotic effects of equal intensity...” (source document, pages 682-3).
Based upon considerations including the above, and the structural similarity of ethylbenzene and the xylenes and toluene, the neurological effects RC will be assumed equal to that of the xylenes and toluene. Thus, the neurological RfC for the occupational setting is 17 mg/m?
(4 ppm-v).
Lethality, acute occupational exposure. Examination of multiple acute toxicity reports produces an eight-hour inhalation lethality benchmark of 7,492 mg/m* (ATSDR 1997b; RTECS 1997), based upon the geometric mean of several reports of bioassays involving mice and rats, and assuming applicability of Haber’s rule for this substance, over the reported range of exposure durations. This value may form a basis for estimating a risk reference concentration (RfC) for lethality. The acute toxicity inhalation RfC may be derived by applying two safety/uncertainty factors of 10 to extrapolate from an effect level to a NOAEL, and from animals to humans. This yields an acute (four-hour) occupational lethality inhalation RfC value of 150 mg/m’. This may be compared with a ShortTerm Exposure Limit (STEL) value of 543 mg/m? set forth by ACGIH (31.5).
A lethality NOAEL of 400 ppm (1,737 mg/m?) has been reported for both mice and rats exposed 6 h/d, for four days, equivalent to 2,605 mg/m? with daily exposure of 4h/d (ATSDR 1997b). The RfC may be derived from this value by applying a single safety/uncertainty factor of 10 to extrapolate from animals to humans, yielding RfC = 261 mg/m. To derive the RfC from the lethality benchmark, in contrast, requires applying two safety/ uncertainty factors, to extrapolate from an effect level to a NOAEL, and from animals to humans, yielding RfC = 150 mg/m?. A reasonable approach to resolving this difference is to select the midpoint: RfC = 205 mg/m? for four hours. This may be equated approximately with 103 mg/m? (24 ppm-v) for an eight-hour shift, or 34 mg/ m? over 24 hours, assuming validity of Haber’s rule (concentration times exposure duration = constant) for this substance over this exposure duration range.
d. Xylenes
Chronic occupational exposure, neurological
effects. Neurological effects constitute an effect category shared by many groups of organic solvents (Andrews and Snyder 1991). Based upon considerations including the above, and the structural similarity of xylenes and toluene, the lowest reasonable value for the neurological effects RFC would be 10 times the U.S. EPA RfC for
T.C. Keener et al.: Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
toluene, which was based upon neurological effects, though the value may be upwardly adjusted to reflect occupational exposure 5 d/wk and 8 h/d, yielding 17 mg/ m? {4 ppm-v; calculation: 4 mg/m? x 1/[(5d/7d) x (8 h/ 24h)]}.
Lethality, acute occupational exposure. Examination of multiple acute toxicity reports produces an eight-hour inhalation lethality benchmark of 24,749 mg/m?
(18,805 mg/m? omitting multiple reports of a single study), based upon the geometric mean of the two remaining studies cited (RTECS 1997). One study, pertaining to humans specifically, gives a six-hour lethality value of 43,427 mg/m? (10,000 ppm-v). This is equivalent to a 4-hour value of 65,141 mg/m, assuming validity of Haber’s rule (dose x exposure duration = constant) for xylenes over this exposure duration range. The acute lethality RfC may be derived either from the 4hour lethality benchmark or the single report pertaining to humans. In the latter case, the RfC results from applying a safety/uncertainty factor of 10 to extrapolate from an effect level to a NOAEL (LC ). This yields 4-h RfC=6,514 mg/m’, equivalent to an eight-hour shift value of 3,257 mg/m®, assuming validity of Haber’s rule. The RfC may be derived from the 4-h lethality benchmark by further adjusting the 4-h lethality benchmark by applying a safety/uncertainty factor of 10 to only the rat study to extrapolate to humans, yielding a 4-h lethality benchmark of 11,893 mg/m’, based upon the geometric mean of the rat and human study. The RfC may finally be derived by dividing this by a factor of 10 to extrapolate from an effect level to a lethality NOAEL (LC,), yielding 4-h RFC = 1,189 mg/m? or 8-h RfC=595 mg/m? (137 ppm-v). This RC will be selected, and may be compared with an ACGIH Short-Term Exposure Limit (STEL) of 651 mg/m? (150 ppm), corresponding to exposure durations within 15 minutes.
e. 1,2,4-Trimethylbenzene and 1,3,5-trimethylbenzene Chronic occupational exposure, neurological effects. A neurological effects RfC may be derived based upon either of two criteria. One is a study of rats administered mixed trimethylbenzenes at 1,700 ppm-v (8,357 mg/m*), resulting in marked central nervous system. The chronic R£C for occupational exposure may be derived by dividing the concentration to which rats were exposed by two safety/uncertainty factors of 10 to extrapolate from a LOAEL to a NOAEL and from animals to humans. This yields RfC = 17 ppm-v (84 mg/m’). This value is close to the second available criterion. Specifically, NIOSH has set forth a 10-hour TWA of 25 ppm (125 mg/m*) (HSDB 1997; NIOSH 1994), equivalent to 156 mg/m? (32 ppm-v) for eight hours. This value constitutes an appropriate neurological effects RfC for chronic occupational exposure. This value may be regarded as being ‘preferable, though somewhat higher than that derived from the rat study, because the rat study involved an isomeric mixture of trimethyl benzene, possibly with some benzene as well.
Lethality, acute occupational exposure. Examination of multiple acute toxicity reports produces an eight-hour
inhalation lethality benchmark of 18,000 ppm-v
(88,483 mg/m?) in rats (RTECS 1997). This may be approximately equated with 9,000 ppm-v (44,242 mg/m’) for an eight-hour shift, assuming the validity of Haber’s rule (concentration times exposure duration = constant) for this substance, over this exposure duration range. The acute lethality inhalation risk reference concentration (RfC) may be derived by dividing the concentration to which rats were exposed by two safety/uncertainty factors of 10 to extrapolate from an effect level to a lethality ‘NOAEL’ (LC,), and from animals to humans. This yields
RfC = 90 ppm-v (442 mg/m’). 245
f. Naphthalene
Chronic occupational exposure, neurological
effects. Neurological effects - including headache, malaise, and confusion ~ were reported among individuals exposed to mothballs in their homes (ATSDR 1997c). Although concentrations and exposure duration data were not provided, one home was reported to exhibit an airborne naphthalene level of 20 ppb-v (0.02 ppm-v). A higher concentration (30 ppm-v) was reported to have produced no histopathological lesions in the brains of mice exposed for two years. One procedure for deriving the RfC is to apply a safety/uncertainty factor to 30 ppmv as an animal No Observed Adverse Effect Level (NOAEL), to extrapolate to humans. This yields an inhalation neurological effects RfC value of 3 ppm-v (15 mg/ m?°). An alternative approach is to apply a safety/uncertainty factor to 20 ppb-v as a Lowest Observed Adverse Effect Level (LOAEL) for a clinically less-serious effect to extrapolate to a NOAEL. If a value of three for the safety/ uncertainty factor is selected, the RfC would be 6.67 ppbv (33.3 g/m’). A factor discouraging use of the latter approach is uncertainty about the 20-ppb-v (0.02 ppm-v) value on which it is based. Thus, the neurological RfC for the occupational setting may be set higher, at 3 ppm-v (15 mg/m’).
Lethality, acute occupational exposure. The acute lethality RFC may be derived based upon a report of a 4h lethality benchmark of >85 mg/m? in rats (RTECS 1997). However, this derivation is unsatisfactory because values preceded by the > symbol may be of indeterminate magnitude. In this case, data pertaining to oral exposure and to inhalation of structurally related compounds suggest a lethality benchmark orders of magnitude higher than that RTECS. The RfC alternatively might be derived from a 4-h LC, of 78 ppm-v (408.95 mg/ m?°) in rats. The RfC is derived by applying a safety/ uncertainty factor of 10 to extrapolate from animals to humans, yielding 4-h R{C=41 mg/m’, equivalent to 8-h RfC=21 mg/m®*. However, this study likewise seems too low, and failed to report a level which was lethal to any animals. The inhalation RfC, therefore, may be derived by extrapolating from an oral lethality benchmark, such as 2,200 mg/kg in rats (which are often used for extrapolation to humans). This is equivalent to 7,700 mg/m? (calculation: 2,200 mg/kg x 70 kg/20 m?/d = 7,700 mg/m’). The RfC may be derived from this lethality benchmark value by applying two safety/uncertainty factors to extra-
Environ Engg and Policy 1 (1999)
polate from an effect level to a NOAEL (LC,) and from animals to humans. This yields Rf{C=77 mg/m? (15 ppmv). The applicable exposure duration is uncertain, but may be assumed applicable to an eight-hour workshift.
g. Methylnaphthalene
Chronic occupational exposure, neurological
effects. Neurological effects constitute an effect category shared by many groups of organic solvents (Andrews and Snyder 1991). Based upon considerations including the above, and the structural similarity of 2-methylnaphthalene and naphthalene, the neurological effects RfC will be assumed equal to that of naphthalene. Thus, the neurological RfC for the occupational setting is 3 ppm-v (15 mg/m’).
Lethality, acute occupational exposure. Although no inhalation lethality studies were available, two lethality values (one oral, one intraperitoneal) were reported (ATSDR 1997c), and may form a basis for estimating a risk reference concentration (RfC) for lethality. A mouse LD,o value of 1,000 mg/kg following intraperitoneal injection was reported, with toxic effects on the respiratory system. Also, an oral rat LDso value of 1,630 mg/kg was reported. The acute toxicity inhalation RfC may be derived based upon the lower value, by assuming that a reference individual weighs 70 kg and breathes 20 m?/day of air. This yields an airborne concentration of 3,500 mg/ m? (calculation: 1,000 mg/kg x 70 kg/20 m? = 3,500 mg/ m?). Two safety/uncertainty factors of 10 may be applied to this value to extrapolate from an effect level to a NOAEL (LC,), and from animals to humans. This yields RfC = 35 mg/m? (6 ppm-v). The applicable exposure duration is uncertain, given the oral dosing used in the study on which the RfC is based, but may be assumed applicable to an eight-hour work shift.
Discussion and conclusions
Characterization of toxic risks to unprotected tank cleaners Applicable regulations require protection of petroleam storage tank cleaners via use of impervious boots and clothing, and self-contained breathing apparatus (SCBA) or a remote source of clean air delivered via mask. However, such regulations are of relatively recent vintage (U.S. OSHA 1993), Even though promulgated, these regulations are sometimes violated. This raises the issue of whether toxic injury to tank cleaners is likely to have occurred, and whether such injuries might be longlasting. This issue, and the issue of liability for damages and long-term care of those claiming injury, has been addressed in litigation. Although available literature has quantified concentrations of selected petroleum hydrocarbons in storage tanks (Lillienberg et al. 1992; Midzenski et al. 1992), the tank locations have been cooler than south Texas, thereby underestimating concentrations to which tank cleaners in hot climates would be expected to encounter for prolonged durations, seasonally.
Table 9 indicates gross exceedances of acceptability criteria for chronic exposures to mixed solvent vapors.
That is, the hygienic HI exceeded the acceptability criterion of unity by more than three orders of magnitude; and the HI for chronic occupational neurological risks exceeded the same acceptability criterion by more than two orders of magnitude. These exceedances raise the issue of whether acute toxicity benchmarks also might be exceeded. Table 10 reveals that eight-hour time-weighted average mixed-solvent vapor exposure was more than half (53%) of a lethal exposure for that duration, and more than 50 times a value which would be expected to protect reliably against lethality.
The lethality results suggest that tank cleaners might approach lethal exposure routinely during hot seasons. In fact, sworn testimony by an expert witness revealed a lethality incident, in which an individual entered a tank which was not being actively ventilated, and was found dead near the tank center, apparently overcome by petroleum hydrocarbon vapors. Tank cleaners testifying in litigation revealed multiple occasions when coworkers had lost consciousness within tanks, and described various strategies for avoiding loss of consciousness. These included taking a mid-shift lunch break and several shorter breaks before and after lunch, as well as leaning out of open tank manways whenever possible. Notwithstanding these strategies, tank cleaners also described visual and auditory hallucinations whose onset would signal the urgent need to exit the tank. Indeed, exposures were described as being sufficiently intense for the odors and tastes of hydrocarbon vapors to be experienced even when the tank cleaners were on break outside the tank. Presumably this residual effect occurred because hydrocarbon levels were so elevated in the blood of tank cleaners emerging from the tank that they excreted the vapors via their lungs as they exhaled.
Exceedance of the neurological effects RfC by a factor of 199 (Table 9) suggests that adverse neurological impacts should be experienced by tank cleaners employed for a sufficiently long time. Indeed, hydrocarbon vapors are known to damage nerve cells in the central and peripheral nervous systems. Redundancy of brain cell functions may mask the damage until the extent of damage overcomes the reserve capacity of the brain to preserve functionality. Thus, functional tests, such as neurological and neuropsychological tests, may underestimate hydrocarbon-mediated damage to the nervous system.
Conservativeness of case study
Based upon this study of crude oil storage tanks with known concentrations of a variety of, residual petroleum hydrocarbon chemicals, a model has been developed to estimate the vapor concentration of these substances from crude oil tank bottoms, including sludges, inside the tanks. The model was designed using conservative assumptions. Indeed, hydrocarbon vapor concentrations would be expected to exceed estimated concentrations, especially in CASE II and CASE III. To illustrate, turbulent air flow through the tank, non-uniform air flow near the edge of the tank, and increased petroleum hydrocarbon evaporation rates due to air flow across the hydrocarbon surface each would be expected to increase
‘T.C. Keener et al.: Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds
concentrations above those estimated using the model. Further, during cleaning, tank bottoms are agitated, producing atmospheric supersaturation with hydrocarbon vapors during the human entry phase of tank cleaning. Moreover, the model assumes that the air is completely mixed, whereas in reality a concentration gradient would be expected to result in higher hydrocarbon vapor concentrations near the surface of the sludge and in poorly ventilated areas of the tank. Hydrocarbon concentrations to which unprotected tank cleaners are exposed, therefore, may be higher than those predicted by the model. Finally, Elkins (1963) showed that the vapor pressure of benzene in mixtures deviates from Raoult’s Law (NIOSH, Refined Petroleum Solvents, pg. 211). As a result, according to Elkins, “the benzene vapor concentration resulting from handling such mixtures will frequently be higher than would be expected from the composition of the solvent.” Benzene concentrations in petroleum storage tanks, therefore, could be higher than those estimated here.
Appendix 1. Mole fraction calculation for benzene at 34.6 °C: Mean Conc.(mg/kg) x (mol wt. of Comp. Product)/ (mol wt. of Benzene) So: 2,229 mg/kg x (kg/1,000,000 mg) x (562 g/mol)/ (78.11 g/mol) =X =0.01604 2. Partial pressure of benzene in the air: P(v)=117.41 mm Hg So: P=P (v) X= 1.8832 mm Hg There are 760 mm Hg per atm so: P=2.48 x 107° atm 3. Using PV=nRT or n/V=:P/RT the moles per volume is found: n/V =2.48 x 10°? atm/R(307.6 K) where R= 82.054 cm? atm/K mol = 8.2054 x 10-° m? atm/K mol Thus: n/V = 9.82 X 10-? moles/m?* 4. Maximum Conc. in mg/m? is: (using the molecular weight of benzene) 9.82 x 10 moles/m? (78.11 g/mol) (1000 mg/g) = 7674.89 mg/m? 5. Parts per million: ppm =(n/V) x (RT/P) x 10° R= from above. P=latm. ppm =9.82 x 10°? moles/m? x 8.2054 x 10~° m? atm/K mol x (307.6 K/1 atm.) x 10°= 2,478.54 ppm
References
Andrews LS, Snyder R (1991) Toxic effects of solvents and vapors. Casarett and Doull’s toxicology - the basic science of poisons, fourth edition. New York: Macmillan, pp 681-722
ATSDR (1997a) Toxicological profile for benzene. Toxicological profiles on Cp-ROM. Atlanta, Georgia; U.S. Public Health Service, Agency for Toxic Substances and Disease Registry
ATSDR (1997b) Toxicological profile for naphthalene. Toxicological profiles on Cp-ROM. Atlanta, Georgia; Agency for Toxic Substances and Disease Registry, Updated 1997
Cherry N, Venables H, Waldron HA (1984) British studies on the neuropsychological effects of solvent exposure. Scandinavian Journal of Work, Environment, and Health; Supplement 1:310-2
Cassler EL (1997) Diffusion: mass transfer in fluid systems cambridge. Cambridge University Press, p 344
Ekberg K (1986) Chronic and acute effects of solvents on central nervous system functions in floorlayers. British Journal of Industrial Medicine 43:101-6
Elkins HB, Comproni EM, Pagnotto LD (1963) Industrial benzene exposure from petroleum naphtha-II. Pertinent physical properties of hydrocarbon mixtures. J Am Ind Hyg Assoc 24:99-102
Elofsson SA (1980) Exposure to organic solvents: a crosssectional epidemiologic investigation on occupationally exposed car and industrial spray painters with special reference to the nervous system. Scandinavian Journal of Work, Environment and Health 6: 239-73
Gregersen P (1984) Neurotoxic effects of organic solvents in exposed workers: an occupational, neuropsychological, and neurological investigation. Am J Ind Medic 5:201-25
HSDB (1997) Hazardous substances data bank web page. October 17, 1997
IRIS (1997) Integrated risk information system web page. June 30 1997
Keener TC, Jain RK, Weber JW (1998) Ingleside fuel tanks-air quality modeling and analysis. University of Cincinnati, College of Engineering
Lillienberg L (1992) Health effects of tank cleaners. American Industrial Hygiene Association Journal 53:375-80
McCain WD Jr (1973) The properties of petroleum fluids. Petroleum Publishing Company
Midzenski MA (1992) Acute high dose exposure to benzene in shipyard workers. American Journal of Industrial Medicine 22:553-65
NIOSH (1974) Criteria for a recommended standard occupational exposure to Benzene. NIOSH Publication No. 74~137
NIOSH (1977) Criteria for a recommended standard occupational exposure to refined petroleum solvents. NIOSH Publication No. 77-192
NIOSH (1973) Criteria for a recommended standard occupational exposure to Toluene. NIOSH Publication
NIOSH (1975) Criteria for a recommended standard occupa-, tional exposure to Xylene. NIOSH Publication
NIOSH (1979) Revised recommendation for an occupational exposure standard for Benzene. NIOSH Publication, recommendation on Aug. 25, 1976
NIOSH (1994) Pocket guide to chemical hazards. DHHS (NIOSH) Publication No. 94-116. Washington D.C., U.S. Government Printing Office
Pankow JF (1991) Aquatic chemistry concepts. Lewis Publishers
Patty FA (1958) Industrial hygienic and toxicology. New York, Interscience Publishers, Inc
Powers BS (1990) Evaluation of residual hydrocarbon vapors following the cleaning of above-ground gasoline storage tanks. Masters thesis essay. Baltimore, Maryland, Johns Hopkins University, School of Hygiene and Public Health, Division of Environmental Health Engineering
Environ Engg and Policy 1 (1999)
RTECS (1997) Registry of toxic effects of chemical substances. U.S. EPA (1988) Interim methods for derivation of inhalation
October 17, 1997 reference concentrations and application of inhalation dosiSawyer, McCarty, Parkin (1994) Chemistry for environmental metry. Research Triangle Park. Environmental Criteria and
engineering, Fourth Edition. McGraw Hill Assessment Office, Office of Research and Development Seinfeld JH (1986) Atmospheric chemistry and physics of air U.S. EPA (1990) Methods for derivation of inhalation reference
pollution. John Wiley & Sons concentrations’ and application of inhalation dosimetry. Seppalainen AM, Husman KAJ, Martenson C (1978) Neurophy- Research Triangle Park, Environmental Criteria and Assess-
siological effects of long-term exposure to a mixture of ment Office, Office of Research and Development
organic solvents. Scandinavian Journal of Work, Environment U.S. OSHA (1993) Standard on confined spaces. Federal
and Health 4:304-14 Register, 58:4,462
Yaws CL (1994) Handbook of vapor pressure. Volumes 2 and 3, Gulf Publishing Company
Publication citation
Keener, Tim C.; Ravi K. Jain, RA Michaels, and John W. Weber. Exposure and health risks potentially posed to petroleum storage tank cleaners by volatile organic compounds. Environmental Engineering and Policy, 1(4):235-48, December 1999;