Toxicology Letters, 26 (1985) 187-192

Elsevier

187

TOXLett. 1443

ACUTE TOXICITY OF GASOLINE AND ETHANOL AUTOMOBILE

ENGINE EXHAUST GASES

(Comparative toxicity of engine fuels; ethanol exhaust fumes)

EDUARDO MASSAD, CARMEN DIVA SALDIVA, LUIZA MARIA NUNES CARDOSO,

RUBERVAL DA SILVA, PAUL0 HILARIO NASCIMENTO SALDIVA and GYGRGY M. BiiHM*

Laboratory of Experimental Air Pollution, School of Medicine, University of So Paula, Av. Dr. Arnaldo, 455 - CEP 01246, S-0 Paul0 (Brazil)

(Received September 28th, 1984)

(Accepted May 24th, 1985)

SUMMARY

A comparative inhalation exposure study was performed to investigate the potential health effect of

gasoline and ethanol engine exhaust fumes.

Wistar rats housed in inhalation chambers were exposed to test atmospheres of various concentrations

of carbon monoxide (CO) and gasoline and ethanol exhaust fumes diluted with air. CO level,

temperature, relative humidity and flow rate were monitored continually to control the gas concentration

and the environment. The dilution method gave a concentration within 1.0% of the target.

The LCsas for 3-h exposures were determined for the 3 test atmospheres. The results demonstrated that

the acute toxicity, in terms of LC 50r of the gasoline-fuelled engine was significantly higher than that of

the ethanol-fuelled engine.

INTRODUCTION

The substitution of gasoline by ethanol as an automobile fuel in Brazil aroused

considerable discussion about the impact of exhaust emissions on ambient air quali-

ty and the effects on public health. We started experimental work to assess the

biological consequences of this policy in 1980 [ 1, 21.

Considering that the stationary situation is the most acutely toxic, the lethal con-

centration for 50% of the sample (LCSO) for the exhaust fumes of unloaded and sta-

tionary engines was determined. The expectation was that the acute exposure of

* To whom correspondence should be addressed

0378-4274/85/$ 03.30 0 Elsevier Science Publishers B.V.

188

animals to a wide range of concentrations would provide some insight into the

potential human response to acute hazards due to the new fuel.

MATERIALS AND METHODS

4 Groups of 20 Wistar rats, weighing 19Ok22 g, were used in each of 3 test at-

mospheres: ethanol exhaust gas, gasoline exhaust gas and a CO control, all diluted

with clean air. During the exposure period, the animals were housed in aluminum

115-l inhalation chambers.

Single lots of ethanol and gasoline were used to ensure homogeneity of exhaust

composition. The ethanol used as fuel was 96” proof, with 99% purity. The gasoline

was unleaded, with an octane rating of 73 (Motor Method). The CO (White and

Martins) was bottled, in a nominal concentration of l.O+ 0.3% by vol.

The ethanol and gasoline exhaust fumes were generated by 2 new, previously

unused, Fiat engines of 1300 cm3 displacement, operating unloaded with a cylinder-

head temperature of 97 f 10°C and an outlet exhaust gas temperature of 100 + lO”C,

as measured by thermocouples, in the stationary mode at a constant speed of 1000

rev./min. This speed was selected as it gave experimental results of greater con-

sistency than the recognized value of 750 rev./min for the stationary mode.

Steady-state operation was selected, since it permits better definition and control

of the physical and chemical properties of the gases, and simulates the most hazar-

dous situation in practice.

The emission mode was fixed at 2.0% CO for both engines, although the levels

of this gas in the gasoline exhaust would be 1.5 times higher under normal operating

conditions, according to the manufacturer. On the other hand, it has been shown

TABLE 1

DOSE-RESPONSE RESULTS _~

Atmosphere CO dose (ppm)

CO 2025

2050

2100

2162

Dilution rate Log dose % Response

3.30643 0.20

3.31275 0.45

3.32222 0.55

3.38486 0.70

Gasoline exhaust 1838 10.88 3.26435 0.10

1995 10.03 3.27761 0.60

2000 10.00 3.30103 0.55

2097 9.54 3.32160 0.90

Ethanol exhaust 2007 9.97 3.30255 0.40

2087 9.58 3.31952 0.50

2192 9.12 3.34084 0.90

2275 8.97 3.35698 0.75

189

that the level adopted for the gasoline engine in our studies is the most appropriate

in terms of fuel consumption [3].

2 CO analyzers were used, one monitoring the CO concentration in the engine ex-

haust pipe (Sun, EPA-73) and the other the CO inside the inhalation chambers

(Hartmann and Braun, Uras-2t). This analyzer was on-line with a microcomputer

(DEC, PDP 1 l/03). The ratio of the concentrations read by the 2 CO monitors gave

the dilution rate (Table I).

The gas conduction lines were insulated to ensure a temperature of approx. 9O”C,

to keep the gases at typical exhaust-pipe temperature until being rapidly diluted with

clean air near the top of the chamber, which caused the temperature to drop to ap-

prox. 35°C. The ambient temperature was maintained at 20+2”C.

Relative humidity and gas flow rate were also controlled during the experiment,

to ensure that each group of animals experienced the same environmental condi-

tions, except for the test atmospheres. The ranges of relative humidity and flow rate

were set at 60t 20% and 15.5 to.5 l/min, respectively.

RESULTS

Dose-response data are shown in Table I. We adopted the following statistical

model in these experiments:

Yij = olj + 6 Xij (1)

where:

Yij = probit of the expected Pij;

(Y and 0 = unknown parameters;

Xij = log10 of the ith dose of the jth atmosphere;

i = i, . . . . kj; j = 1,2,3

The parameters of the model (1) were estimated by the probit analysis method [5].

The observed value of the xi statistics with 6 degrees of freedom, appropriated

to the linearity test of the 3 dose-response curves, is:

xi = 7.28

This is not significant at the 5% level, hence the hypothesis of linearity of the dose-

response curves cannot be rejected.

The value of the x$ statistics, with 2 degrees of freedom to the parallelism test is:

& = 4.17

This is not significant at the 5% level.

190

TABLE II

ESTIMATES OF THE PARAMETERS OF THE MODEL (1)

Test atmosphere Estimated equation

Gasoline Y = - 103.5505 + 32.954 X

Ethanol Y = - 104.3183 + 32.954 X

co Y = ~ 104.4346 + 32.954 X

Therefore, the model (1) is appropriate for the analysis of our experiments.

Estimates of the parameters of the model (1) are given in Table II.

Considering the estimated straight lines for ethanol and gasoline (Table II), we

estimated the log LCsO and constructed the confidence interval for this parameter

with a coefficient of 0.95. Calculating the antilogarithms of these estimations, we

obtained estimates and confidence intervals for the L&O of the test atmospheres

(Table III).

DISCUSSION

These experiments demonstrate that the acute toxicity of the exhaust of the

gasoline-fuelled engine is greater than that of the ethanol-fuelled engine; the higher

LCSO value of ethanol exhaust fumes indicates a lower acute toxicity (Table III).

An interesting point is the superposition of the confidence intervals of the CO and

ethanol LC~O values, showing an equality of both gaseous mixtures and suggesting

that the acute effects of ethanol exhaust gases depend mainly on their CO content.

Gasoline and ethanol exhaust fumes are complex gaseous mixtures containing CO

(an important common component) and many other substances. By comparing their

acute toxicity based on CO levels, we have attempted to appraise, indirectly, the

other components which may or may not be common to both types of exhaust gases.

Therefore, one of the main conclusions of this work is that gasoline exhaust fumes

contain noxious substances other than CO, which are responsible for its greater tox-

icity. This does not seem to be true for ethanol exhaust fumes, at least in the case

of acute toxicity.

In addition to the higher emission rate of CO in the gasoline exhaust fumes, which

was intentionally equalized in our experiments, there are some important points

TABLE III

L&o AND CONFIDENCE INTERVAL OF THE TEST ATMOSPHERES

Test atmosphere I-C50 (rvm) Confidence interval (y = 0.95)

Gasoline 1967.8863 [1940.4390-1991.59031

Ethanol 2076.3473 [2041.7379-2101.8417]

co 2093.1483 [2067.7594-2120.8014]

191

regarding the qualitative and quantitative differences between ethanol and gasoline exhaust [6-g]. For example, there are high levels of aldehydes in ethanol exhaust which are almost absent in gasoline; relatively high concentrations of sulphur oxides in gasoline which are practically nonexistent in ethanol; important qualitative dif- ferences in the emitted hydrocarbons; and the carcinogenetic properties of gasoline, which are apparently absent from ethanol exhaust gases [lo]. Thus, the toxicity dif- ferences shown in our experiments, in which CO emissions were equalized, may be explained by the substances that are not common quantitatively of qualitatively to both engine exhausts. This matter is discussed in another paper concerning the chronic effects of gasoline and ethanol gases [lo].

Although it is hazardous to extrapolate from data obtained in experimental models to real situations affecting humans, these results plus the known levels of CO in gasoline emissions indicate that, at least considering traffic congestion, tun- nels, garages and other potentially toxically intense conditions, ethanol cars are ap- parently less dangerous than gasoline-fuelled automobiles. This is a very important point for petroleum-dependent countries, like Brazil and others, which are exploring renewable sources of energy.

ACKNOWLEDGEMENTS

This work was realized with grants from FAPESP (75/1200), CNPq (400721/81), STI-MIC (110/82) and HCFMUSP.

REFERENCES

1 G.M. BGhm, E. Massad, P.H. Saldiva, M.A. Gouveia, C.A. Pasqualucci, L.M.N. Cardoso, M.P.R.

Caldeira and D.F. Caiheiros, Comparative toxicity of alcohol and gasoline fueled automobile exhaust

fumes, in A.N. Nayes, R.C. Schnell and T.S. Miya (Eds.), Development in Science and Practice of

Toxicology, Efsevier, Amsterdam, 1983, pp. 479-482.

2 M.A. Gouveia, C.A.G. Pasqualucci, P.H. Saldiva and G.M. B(ihm, Pathological aspects of intoxica-

tion by ethanol car exhaust emission, J. Pathol., 142 (1984) A9.

3 2. Linke, I. Cotaite Neto and W. Weishampt, Efeitos da regulagem da emisdo da concentra@.o de

mon6xido de carbon0 de veiculo motor ciclo Otto, em regime de marcha lenta, le Simpbsio de

Engenharia Automotiva, Brasilia, Brazil, 1983.

4 D.J. Finney, Statistical Method in Riological Assay, Charles Griffin, London and High Wycombe,

1978.

5 D.J. Finney, Probit Analysis, Cambridge University Press, Cambridge, 1971.

6 M. Matsuno, Y. Nakano, H. Kachi, K. Kimura, F. Tsuruga, N. lwai, H. Suto, H. Enokido and T.

Itoh, Alcohol engine emission, Proc. 111 Int. Symp. Alcohol Fuels Technol., California, 1979.

7 K. Oblander and J. Abthoff, The Influence of Methanol and Ethanol in the Pipe Composition of

Engines, Daimler-Benz Press, Stuttgart, 1981.

8 R.L. Furey and M.W. Jackson, Exhaust and evaporative emissions from a Brazilian Chevrolet fueled

with ethanol-gasoline blends, Proc. XII Intersociety Energy Conversion Engineering Conference,

Washington, DC, 1977.

192

9 N. Nefussi, J.V. Assunqlo, M.P. Toledo and AS. Castelli, Comparaq$o entre emissdes de poluentes

de veicufos a &lcool e a gasolina, XI Congress0 Brasileiro de Engenharia Sanitkia e Ambiental, For-

taleza, Brazil, 1981.

10 E. Massad, P.H.N. Saldiva, C.D. Saldiva, M.P.R. Caldeira, L.M.N. Cardoso, A.M.E. Moraes, D.F.

Calheiros, R. Silva and G.M. BGhm. Toxicity of prolonged exposure to ethanol and gasoline exhaust

gases, Environ. Res., in press.