EVALUATION OF POLLUTANT EMISSIONS RELATED TO THE USEOF MMT IN GASOLINE

JOSEPH ZAYED1,∗, JANIKE PITRE1, MICHÈLE RIVARD2 and SYLVAINLORANGER1

1 Département de Médecine du Travail et d’Hygiène du Milieu;2 Département de MédecineSociale et Préventive, Faculty of Medicine, University of Montreal, P. O. Box 6128, Station

Downtown, Montréal, Canada, H3C 3J7E-mail: zayedj@ere.umontreal.ca

(Received 29 April 1997; accepted in final form 26 November 1997)

Abstract. Since 1976, methylcyclopentadienyl manganese tricarbonyl (MMT), an organic derivativeof Mn, has been used in Canada as an anti-knock agent in unleaded gasoline. In a preliminary eval-uation of the contribution of MMT to urban air pollution, the CO and NOx emissions of 10 vehicleseach having traveled less than 17 000 miles were sampled. Five of these vehicles were filled withMMT-free gasoline and the remaining five with MMT-added gasoline. The emissions were evaluatedaccording to the Federal Test Procedure which simulates urban and highway driving cycles. As for theNOx and CO analysis, it was carried out by means of infrared spectrometry, while the concentrationof Mn in the gasoline was determined through neutron activation. Overall, the amount of CO emittedby the vehicles using MMT-added gasoline was 1.48 g mile−1 ± 1.77. The amount emitted by thevehicles using clear (MMT-free) gasoline was 0.74 g mile−1 ± 0.88. As for the NOx emission level,it was 0.20 g mile−1 ± 0.15 for the vehicles using MMT-added gasoline vs 0.17 g mile−1 ± 0.18for the vehicles using clear gasoline. However, probably due to the small sample size and the veryhigh level of emission of one of the vehicles using MMT-added gasoline, these differences are notstatistically significant. Further studies should take into account variables such as the driving historyand the representativity of the fleet of vehicles tested.

Key words: atmospheric pollutants, CO-NOx, gasoline, MMT

1. Introduction

Methylcyclopentadienyl manganese tricarbonyl (MMT: C9H7MnO3), an organicderivative of manganese (Mn), is used as an anti-knock agent in unleaded gasoline(Abbott, 1987; Cooper, 1984). Used in Canada since 1976 (Environment Canada,1987), MMT has seen a substantial increase in its utilization over the last few years,in particular since it replaced lead in gasoline (Royal Society of Canada, 1986).However, the Canadian government now wants to ban both the interprovincialsale and the importation for commercial purposes of certain manganese-based sub-stances, such as MMT, that are added to unleaded gasoline (Bill C-29). The maincauses for the withdrawal of MMT are the potential public health risk associated

∗ Author for correspondence.

Water, Air, and Soil Pollution109: 137–145, 1999.© 1999Kluwer Academic Publishers. Printed in the Netherlands.

138 J. ZAYED ET AL.

with its combustion products, especially manganese phosphate, manganese sulfateand manganese oxides (Ter Harret al., 1975; Ardeleanuet al., 1997), and claimsthat MMT reduces the effectiveness of antipollution devices.

Recent research has attempted to assess human exposure to Mn (Drolet andZayed, 1994; Loranger and Zayed, 1995; Sierraet al., 1995; Zayedet al., 1994)as well as the environmental Mn contamination resulting from the combustionof MMT in abiotic (Lorangeret al., 1994 and 1995; Loranger and Zayed, 1994)and biotic (Braultet al., 1994) systems. These studies are important to understandthe environmental fate of Mn and to quantify the multimedia human exposure. Ingeneral, the results indicate that the contribution of the Mn emanating from MMTis low compared to contribution of the Mn from industrial and natural sources.Furthermore, the assessment of the multimedia exposure of workers potentiallymore exposed to the Mn emanating from MMT (e.g. mechanics and taxi drivers)shows that food is still the main source of multimedia exposure (> 95%) and thatthe atmospheric concentrations of Mn in the workplace or in the environment arewell below governmental standards or criteria (Loranger and Zayed, 1995; Zayedet al., 1996).

On the other hand, it is known that combustion products accumulate in engines(Lenane, 1990), deposit on exhaust catalysts (Hurleyet al., 1989), and are emit-ted in the atmosphere (Loranger and Zayed, 1994; Lorangeret al., 1995). TheMn particles in an engine can increase in particular the regulated emissions ofhydrocarbons (Hurleyet al., 1992), while on catalysts, they may cause prematuredeactivation by clogging the inlet or by coating it with a layer of impenetrableparticles (Otto and Sulak, 1978; Holiday and Parkinson, 1978).

Nevertheless, the results of a study requested by the Canadian AutomobileAssociation (Energy and Environmental Analysis, 1994 and 1995) indicate that,between 1980 and 2000, the carbon monoxide (CO) and nitrogen oxide (NOx)emissions will have diminished respectively by 14% in the province of Quebecand by 45% in the Montreal region and this, in spite of an increase of approxi-mately 50% in the total distance traveled by vehicles. Moreover, observations haveshown a general decrease in the mean concentrations of these two pollutants in theatmosphere in Canada between 1974 and 1989 (Government of Canada, 1991).

Since transportation is responsible for 62.6% (1 180 000 tons) of the total annualemissions of NOx and 67.1% (7 163 617 tons) of the annual emissions of CO(Canadian Petroleum Products Institute, 1993), any increase in vehicle emissionswould go against the international agreements for the reduction of environmentalpollution.

The general objective of this research is to make a preliminary assessment of thecontribution of MMT from a mobile source to the level of CO and NOx emissionsin the atmosphere.

USE OF MMT IN GASOLINE 139

2. Method

This research was carried out in collaboration with the PMG Test and ResearchCentre (Quebec, Canada). Five Canadian automobiles containing MMT-added gaso-line were compared to five similar American automobile models using MMT-freegasoline. The emissions were evaluated according to the Federal Test Procedurewhich simulates the behavior of an automobile in an urban environment (EPATest Cycle or UDDS-Urban Dynamometer Driving Schedule) and on the highway(HWFET-Highway Full Economy Test). The UDDS comprises 3 different stagesand a 10-min waiting period. The first stage, Highway Cold Motor, lasts 505 sec,while the second stage, Cite Warm Motor, which is followed by a 10 min waitingperiod, lasts 867 sec. Finally, the third stage, Highway Warm Motor, also lasts 505sec. The HWFET is carried out in one single stage which lasts 765 sec.

Before being tested, each vehicle was connected to the following: a dynamome-ter (CTE-50, Clayton, CA) which reproduced actual road conditions by means ofa system of rollers which made the wheels of the automobile turn, a screen linkedto a computer system which ran a program indicating the steering maneuvers tobe carried out for each test, a dilution chamber (to dilute the crude dischargesin order to obtain measurable concentrations), and an infrared Fourrier transformspectrometer.

The tests, carried out in a laboratory with a controlled temperature of 25◦C,including the following 4 stages: (1) preparation of the automobile, (2) adaptationof the automobile to the gasoline to be tested, (3) test per se, (4) collection andanalysis of the data. All the automobiles were drained beforehand to minimisecontamination by gasoline previously used, and the same type of either MMT-added gasoline or MMT-free gasoline was added. The volume of the gasoline usedduring the tests did not exceed 40% of the tank’s capacity. The drained gasoline wasplaced in appropriate containers and disposed of in accordance with governmentregulations.

Preparing the automobiles for the tests consisted in checking the level of thedifferent liquids present in the automobile as well as the condition of the brakesand tires. The exhaust system was also inspected for leaks.

To adapt the vehicles to the gasoline to be tested, they were filled with thenew gasoline and left running for half an hour at the desired ambient temperature(25 ◦C). The motor was then turned off and, while maintaining the temperatureat the same degree, the vehicles were left to stand for at least 12 hr to allow thenew gasoline to soak in. The vehicles were then driven outside for approximately10 km. Finally, the two tests (UDDS and HWFET) were carried out without anyemission measurements being made.

The test per se was a repetition of the above mentioned tests although this timethe exhaust gas was collected in bags designed for this purpose. The contents of thebags were then analyzed with an infrared spectrometer to determine the concentra-tions of NOx (scale of 10 ppm to 3000 ppm; Horiba chemi-luminescent analyzer

140 J. ZAYED ET AL.

(CIA-22A)) and CO (low scale of 100 ppm to 3000 ppm and high scale of 0.1 to0.2%; Horiba nondispersive infrared detector (AIA-23)).

The Mn present in the gasoline was analyzed through neutron activation (Kennedy,1990). Seven samples of gasoline (5 of which contained MMT) were inserted intopolyethylene vials and, using a pneumatic transfer system, were sent to the coreof the Slowpoke nuclear reactor where they were irradiated for 60 sec at a neutronflux of 5 × 1011 neutrons cm−2 s−1. The samples were then transferred from theirradiation vials to new vials and, after a decay period of at least 3 min, they wereplaced on a large germanium detector for 5 min to detect the gamma-rays emittedby the 56 Mm. Finally, the number of gamma-rays in the 56 Mn peak at 847 keVwas used to determine the total amount of Mn in each sample.

To calculate the concentration of MMT, the Mn concentration was multiplied by4.09, a factor established by Chauet al. (1997) for commercial unleaded gasoline.As for the octane rating, it corresponds to the average of the results obtained fromtwo tests (ASTM D2699 and D2700) carried out according to the standard methodof the American Society for Testing Material.

Because of the small size, a Mann-Whitney nonparametric test was used to com-pare the CO and the NOx concentrations between the two groups of vehicles (withor without MMT). The Mann-Whitney test is an analogue to the two-sample t-testbut it analyzes ranks of the measurements rather than the actual measurements.Its utilization does not require any populational assumptions about normality orequality of variances as does the t-test.

3. Results

For this preliminary study, two of each of the following five automobile modelswere selected (Table I): Buick Century, Ford Taurus, Chevrolet Lumina, Ford Con-tour, and Dodge Stratus. These ten rented vehicles (5 with MMT and 5 withoutMMT) has either 4 or 6 cylinders and each had traveled less than 17 000 miles.The analyses of the MMT-added and the MMT-free gasoline revealed that bothtypes of gasoline had similar octane ratings (Table II).

In general, the amount of CO emitted (Table III) by the Canadian vehicles usingMMT-added gasoline (1.48 g mile−1 ± 1.77) was twice the amount emitted bythe American vehicles using MMT-free gasoline (0.74 g mile−1 ± 0.88), althoughthe difference between these emissions was not statistically significant (p > 0.05).More specifically, the CO emissions of the Canadian vehicles were higher duringeach cycle: 2.12 g mile−1 vs 1.23 g mile−1 during cycle I, 1.43 g mile−1 vs 0.53g mile−1 during cycle II, 1.41 vs 0.75 g mile−1 during cycle III and finally 0.95vs 0.43 g mile−1 during the highway cycle. In each group of vehicles, the DodgeStratus obtained the highest value, its emissions being approximately three timeshigher than the average emission of the other vehicles. This automobile was also the

USE OF MMT IN GASOLINE 141

TABLE I

Fleet of vehicles

Vehicle Test Year Engine Odometer

No Type (miles)

Canadian Vehicles (with MMT)

Buick Century T-01 1996 6 14476

Ford Contour T-02 1996 4 16724

Chevrolet Lumina T-03 1996 6 11024

Dodge Stratus T-04 1996 4 6056

Ford Taurus T-05 1996 6 15560

American Vehicles (without MMT)

Buick Century T-06 1996 6 9756

Ford Contour T-07 1996 4 7776

Chevrolet Lumina T-08 1996 6 15414

Dodge Stratus T-09 1996 4 12565

Ford Taurus T-10 1997 6 924

TABLE II

Mn and MMT in Gasoline

Gasoline Mn concentration MMT concentration Octane rating

(mg L−1) (mg L−1)

With MMT (n = 5) 6.5± 0.9a 26.5 92.0

Without MMT 0.01 – 0.02b – 92.8

(n = 2)

a Mean and SD.b Detection limit = 0.01.

only one which showed a significant difference (p < 0.05) between the emissionsof the American and the Canadian model.

The amount of NOx emitted (Table V) by the Canadian vehicles (0.20 g mile−1

± 0.15) was 18% higher than the amount emitted by the American vehicles (0.17g mile−1 ± 0.18), but the difference was not significant (p > 0.05). More specifi-cally, the NOx mean concentrations of the Canadian and American vehicles wererespectively 0.24 and 0.19 g mile−1 during cycle I, 0.15 g and 0.09 g mile−1 cycleII, 0.29 and 0.27 g mile−1 during cycle III, and 0.15 and 0.13 g mile−1 duringthe highway cycle. Only the emissions of the Canadian and the American Contourdiffer significantly (p < 0.05), the emissions of the latter being higher.

142 J. ZAYED ET AL.

TABLE III

CO emission rates of automobiles running either with MMT-added or MMT-freegasoline

Test Cyclesa Mean SD Min. Max. CV

No (g mile−1)

Canadian T-01 4 0.35 0.22 0.15 0.66 62.4

automobiles T-02 4 0.65 0.33 0.28 1.08 51.1

with MMT- T-03 4 1.39 0.89 0.29 2.41 64.0

added gaso- T-04 4 4.64b 1.10 3.99 6.28 23.6

line T-05 4 0.35 0.37 0.05 0.81 104.1

All 20 1.48 1.77 0.05 6.28 120.1

American T-06 4 0.45 0.37 0.15 0.99 83.8

automobiles T-07 4 0.37 0.30 0.07 0.78 79.2

with MMT- T-08 4 0.60 0.56 0.11 1.32 92.9

free gaso- T-09 4 2.05b 1.17 0.49 3.31 57.0

line T-10 4 0.22 0.17 0.01 0.38 75.7

All 20 0.74 0.88 0.01 3.31 118.8

a Cycles: HWFET (n = 1) and UDDS (n = 3).bP < 0.05, Mann-Whitney test.

4. Discussion

The approach chosen to compare the CO and NOx emissions associated with theMMT-added and the MMT-free gasoline led to the use of both American and Cana-dian automobiles. The specific choice of American automobiles was motivated bythe need to avoid potential bias caused by the presence of MMT residues or ofits combustion products. Nevertheless, these cars were rented in a town close toCanada (Burlington, VT) and they may well have been driven into Canada andfilled with gasoline there. Moreover, the difference of 0.8 in octane rating anddifferences on based fuels could have a significant effect on emissions.

The national mean concentration of Mn in unleaded gasoline is approximately9 mg L−1 whereas the maximum mean concentration allowed is 18 mg L−1 (Cana-dian General Standards Board, 1986). The Mn concentration of 6.5 mg L−1 foundin the MMT-added gasoline is therefore approximately 25% below the Canadianaverage.

The specific results relative to the CO and NOx emissions differ from thoseof other studies (Lenaneet al., 1994); however, they are in accordance with theemission standards of Canada (Government of Canada, 1996) and those proposed

USE OF MMT IN GASOLINE 143

TABLE IV

NOx emission rates of automobiles running either with MMT-added or MMT-freegasoline

Test Cyclesa Mean SD Min. Max. CV

No (g mile−1) %

Canadian T-01 4 0.22 0.09 0.10 0.31 42.2

automobiles T-02 4 0.03b 0.02 0.01 0.05 63.8

with MMT- T-03 4 0.43 0.09 0.34 0.54 21.2

added gaso- T-04 4 0.14 0.08 0.05 0.21 54.9

line T-05 4 0.19 0.12 0.07 0.33 6.6

All 20 0.20 0.15 0.01 0.54 76.7

American T-06 4 0.04 0.07 <0.01 0.18 156.1

automabiles T-07 4 0.09b 0.02 0.07 0.12 24.1

with MMT- T-08 4 0.42 0.17 0.17 0.54 41.0

free gaso- T-09 4 0.06 0.06 <0.01 0.14 97.1

line T-10 4 0.23 0.14 0.07 0.41 61.2

All 20 0.17 0.18 <0.01 0.54 102.9

a Cycles: HWFET (n = 1) and UDDS (n = 3).b P < 0.05, Mann-Whitney test.

TABLE V

CO and NOx emissions and current standards (g mile−1)

CO emissions NOx Emissions

With MMT Without MMT With MMT Without MMT

Present study 1.48 0.74 0.20 0.17

Lenaneet al. (1994) 2.75 2.84 0.42 0.49

Standardsa 3.40 3.40 1.00 1.00

Standardsb NS 2.50 NS 0.30

a Standards established by the Government of Canada, 1996.b Standards proposed by the American Automobile Manufacturers Association (Cadleetal., 1996)NS: Non specified.

144 J. ZAYED ET AL.

by the American Manufacturers Association (Cadleet al., 1996). The data can befound in Table V.

Although the CO and NOx emissions increased respectively by 100 and 18%with the use of MMT-added gasoline, the emission increases were not significantlydifferent. This is probably due to the small sample size and to the very high levelof CO emitted by one of the vehicles using MMT-added gasoline. Our results aretherefore similar to the results obtained by Hurleyet al. (1992) which did not showconclusively the effects of MMT. However, they are different from those of Lenaneet al. (1994) who found that, over a distance of 40 000 miles, the CO emissionswere essentially the same for both the clear and the MMT-added gasoline, but thatthe NOx emissions were significantly different. The amount of NOx emitted by thevehicles using MMT-added gasoline was 20% lower than the amount emitted bythe vehicles using clear gasoline.

On the other hand, the emissions measured during our research are approxi-mately two times lower than those measured by Lenaneet al. (1994). This is prob-ably due to the low mileage of the vehicles used (less than 17 000) and to recenttechnological progress (the vehicles tested were either 1996 or 1997 models).

Furthermore, it is evident that the variations between the different automobilemodels are more important than the variations attributed to the use or nonuse ofMMT in gasoline. A striking example is the Dodge Stratus whose CO emissionswere 15 times higher than those of other models using the same type of gaso-line. When studying the reduction of emissions at the source, this aspect shouldtherefore be given considerable importance.

It should also be noted that the CO and the NOx emissions were more impor-tant during stage II of the UDDS urban driving cycle (data not included). This isprobably due to the frequent accelerations and decelerations of the vehicles duringthis stage. The emission pattern was in fact similar to the speed pattern

Between 1979 and 1993, the average levels of CO and NOx in Canadian citieshave decreased respectively by 56 and 28% despite a 13% increase in the dis-tance traveled by automobiles (Environment Canada, 1996). However, we can notinfer from the results of our research the impact of MMT on the atmosphericconcentrations of both gases.

Further studies should take into account the variability resulting from the me-chanical condition of the vehicles as well as the representativity of the fleet ofvehicles tested.

Acknowledgements

This research was funded by the Natural Sciences and Engineering Research Coun-cil of Canada, the Ministry of Renewable Resources of Quebec and the Cana-dian Petroleum Products Institute. The authors wish to thank Lise Gareau for thetechnical and scientific assistance.

USE OF MMT IN GASOLINE 145

References

Abbott, P. J.: 1987,Sci. Total Environ.67, 247.Ardeleanu, A., Loranger, S., Kennedy, G. and Zayed, J.: 1997,Sci. and Technol.(Submitted).Brault, N., Loranger, S., Courchesne, F., Kennedy, G. and Zayed, J.: 1993,Sci. Total Environ.153,

77.Cadle, S. H., Bailey, B. K., Belian, T. C., Carlock, M., Gorse, Jr., R. A., Haskew, H., Knapp, K. T.

and Lawson, D. R.: 1996,J. Air and Waste Manage. Assoc.46, 355.Chau, Y. K., Yang, F. and Brown, M.: 1997,Appl. Organomet. Chem. (in press).Canadian Petroleum Products Institute: 1993,Composition of Canadian Summer and Winter

Gasolines(Sulphur Manganese, T(90)) 1992.Canadian General Standards Board (CGSB): 1986,An Assessment of the Effect of MMT on Light-

Duty Vehicle Exhaust Emissions in the Canadian Environment, Working Group of the CanadianGeneral Standards Board, Gasoline and Alternative Fuels Committee, Ottawa.

Cooper, W. C.: 1984,J. Toxicol. Environ. Health55, 184.Drolet, C. and Zayed, J.: 1994,Can. J. Nutr.55, 184.Energy and Environmental Analysis: 1994,Clearing the Air: Urban Smog Emission Trends in

Selected Canadian Metropolitan Areas, Final Report, CAA, Ottawa, Ontario.Energy and Environmental Analysis: 1995,Rapport sur les Prévisions des Émissions de Gaz

Provenant des Automobiles et des Camions Légers, CAA, Ottawa, Ontario.Environment Canada: 1987,National Inventory of Sources and Emissions of Manganese: 1984, EPS

5/MM/1, Conservation and Protection, Environmental Analysis Branch, Ottawa, Ontario.Environment Canada: 1996,Urban Air Quality, Bulletin 96-1, National Environmental Indicator

Series.Gouvernement du Canada: 1991,L’État de l’Environnement Canada, Ottawa, Ministère des Appro-

visionnements et Services Canada.Government of Canada: 1996,Motor Vehicle Safety Regulations-Amendment, Motor Vehicle Safety

Act, Canada Gazette Part 1.Holiday, E. P. and Parkinson, M. C.: 1978,Ford Motor Company Advanced Engine Engineering

Office Report, June.Hurley, R. G., Watkins, W. L. H, and Griffs, R. C.: 1989,Society of Automotive Engineers, Paper No.

890582.Hurley, R. G., Hansen, L. A., Guttridge, D. L., Hammerle, R. H., Matzo, A. D. and Gandhi, H. S.:

1992,Society of Automotive Engineers, Paper No. 920730.Kennedy, G. G.: 1990,Metallization of Polymers, ACS Series440, 128.Lenane, D. L.: 1990,Society of Automotive Engineers, Paper No. 902097.Lenane, D. L., Fort, B. F., Ter Haar, G. L., Lynam, D. R. and Pfeifer, G. D.: 1994,Sci. Total Environ.

146/147, 245.Loranger, S. and Zayed, J.: 1995,Int. Arch. Occup. Environ. Health67, 101.Loranger, S. and Zayed, J.: 1994,Atmos. Environ.28, 1645.Loranger, S., Zayed, J. and Kennedy, K.: 1995,Atmos. Environ.29, 591.Loranger, S., Forget, E. and Zayed, J.: 1994,Water, Air, and Soil. Pollut.74, 385.Otto, K. and Sulak, R. L.: 1978,Environ. Sci. and Technol.12, 181.Royal Society of Canada: 1986,Lead in the Canadian Enmvironment, Royal Commission, Ottawa,

Ontario.Sierra, P., Loranger, S., Kennedy, G. and Zayed, J.: 1995,Am. Ind. Hyg. Assoc. J.56, 713.Ter Haar, G. L., Griffing, M. E., Brandt, M., Oberding, D. G. and Kapron, M.: 1975,JAPCA25, 858.Zayed, J., Gérin, M., Loranger, S., Sierra, P., Begin, D. and Kennedy, G.: 1994,Am. Ind. Hyg. Assoc.

J. 55, 53.Zayed, J., Mikhail, M., Loranger, S., Kennedy, G. and L’Espérance, G.: 1996,Am. Ind. Hyg. Assoc.

J. 57, 376.