Quantitative determination of PCDD/Fs during combustion of chlorobutyl-lined tires.

by P. Hous, Exxon Chemical Europe and H. Bartelds and E Smit, TNO Environment and Energy Research

In a tubeless tire, the function of an innerliner is to provide an effective air barrier which is capable to minimize the migration of air from the tire cavity into the cord area. The reduction in air (hence oxygen) in the carcass serves to lessen oxidative degradation of the tire fabric and carcass rubbers, thus largely improving tire durability. Moreover, impermeable innerliners keep tires at optimum inflation pressure for a long time. The innerliners with the lowest permeability currently on the market are based on halogenated butyl rubbers, chloro and bromobutyl.

The combination of low permeability, high heat resistance and ability to covulcanize with highly unsaturated rubbers makes halobutyls the rubbers of choice for tire innerliners.

Chlorine and, to a lesser extent, bromine have entered public discussion due to the formation of polychlorinated and/or polybrominated dibenzo-para-dioxins and dibenzofurans (PCDD/Fs, PBDD/Fs) during combustion of halogenated (waste) materials. A number of sources of (chlorinated) PCDD/F emissions have been identified (ref 1). Chlorine is widely available in the environment, e.g. as kitchen salt. Bromine may be used in car fuels (in scavengers) and in polymers (as fire retardant).

Before the present work it was no known whether combustion of compounds with chlorobutyl produces PCDD/Fs. In case of controlled combustion in incinerators, concentrations of PCDDs/Fs above legal limits could be reduced by flue gas cleaning, which is not possible in case of uncontrolled combustion in fires

As fires of car tires have occurred occasionally and as these fires attract public attention, quantitative data on the emission of PCDD/Fs and PBDD/Fs from car tire fires and on the role and contribution of chlorinated/brominated butyl in such an emission was needed. This article presents such data on PCDD/Fs.

While analytical methods for the various polychlorinated dioxins/furans are well established, this is not the case for the brominated derivatives. This is the main reason why this study was limited to chlorobutyl-lined tires.

Dioxins and furans

PCDDs and PCDFs may occur in a large variety of individual dioxins (75 congeners) and furans (135 congeners). Only 17 of these congeners are considered to be toxicologically significant. The most toxic one is 2, 3, 7, 8-tetra-CDD. Table 1 presents the relative toxicity of the relevant congeners. Commonly in literature the term dioxins is used for dioxins and furans together.

Table 1 - international toxic equivalent factors (I-TEF)
for the congeners of the chlorinated dioxins and
furans
Dioxin congener            I-TEF    Furan congener            I-TEF
2,3,7,8-tetra CDD          1       2,3,7,8-tetra CDF          0.1
1,2,3,7,8-penta CDD        0.5     2,3,4,7,8-penta CDF        0.5
                                   1,2,3,7,8-penta CDF        0.05
1,2,3,4,7,8-hexa CDD       0.1     1,2,3,4,7,8-hexa CDF       0.1
1,2,3,6,7,8-hexa CDD       0.1     1,2,3,7,8,9-hexa CDF       0.1
1,2,3,7,8,9-hexa CDD       0.1     1,2,3,6,7,8-hexa CDF       0.1
                                   2,3,4,6,7,8-hexa CDF       0.1
1,2,3,4,6,7,8-hepta CDD    0.01    1,2,3,4,6,7,8-hepta CDF    0.01
                                   1,2,3,4,7,8,9-hepta CDF    0.01
1,2,3,4,6,7,8,9-octa CDD   0.001   1,2,3,4,6,7,8,9-octa CDF   0.001

Multiplication of the concentration of an individual CDD or CDF by its I-TEF gives a 2,3,7,8-TCDD (tetra CDD) toxic equivalent quantity (I-TEQ). The toxicity of any mixture, relative to 2,3,7,8-TCDD, is equal to the sum of the individual I-TEQs.

Facility for fire experiments

The TNO experimental fire facility with flue gas sampling and cleaning systems has been developed and built for experiments on the behavior of fires. In this facility liquid, gaseous and solid materials can be burnt on a 1-2 kg/h scale. The burner plate (diameter 15-30 cm) can be heated electrically or by natural gas combustion to ignite materials or to keep materials burning. In the furnace room of 3 x 3 x 2.5 [m.sup.3] (length x width x height) a system to reduce the combustion air supply can be installed to simulate pyrolytic combustion. A water nozzle system to simulate fire fighting with water can be used.

During the experiments the tire material, milled to pieces of a size smaller than 20 mm, is continuously dropped into a pipe (diameter 25 mm) ending close to the burner plate. No pyrolytic combustion or fire extinguishing has been pursued.

Sampling systems for PCDD/Fs

Sampling of the flue gases occurs after the flue gases have been thoroughly intermixed and the gas temperature has dropped to below 100[degrees]C by mixing with air. The PCDD/Fs pass a heated probe, diluted by clean air and collected in Strohlein- and PUF-filters.

Analytical procedures for PCDD/Fs

Collecting of samples for analysis on PCDD/Fs has been performed in accordance with instructions for the sampling and analysis of 2,3,7,8-substituted polychlorodibenzo-p-dioxins (PCDD) and polychlorodibenzofurans (PCDF) emitted into the air from stationary sources (ref. 2). The samples are:

* two samples of tire material;

* six combinations of Strohlein filter and PUF cartridge;

* two samples of residue.

The analysis of the samples includes a determination of seventeen 2,3,7,8-substituted PCDD/F congeners in absolute quantities, in individual I-TEQs and a summation of the individual I-TEQs per sample. The determination has been performed using capillary gas chromatography in combination with high resolution mass spectrometry.

Extraction, clean up and analysis occur in accordance with a standard operating procedure (ref. 2). According to this procedure an internal standard (a mixture of twelve [sup.13C] labeled PCDD/F congeners) has been added to the samples for identification and quantification of the relevant PCDD/Fs. After dissolution and drying the samples have been extracted with toluene. Of the extracts, 80% has been used for analysis. The remaining 20% has been stored for a second analysis (if necessary).

The fractions to be analyzed have been cleaned by extraction with sulfuric acid. The resulting extracts have been led consecutively over an active carbon, multi layer and alumina column. After adding an injection standard the analyses have been performed using gas chromatography (capillary polar GC column) and high resolution mass spectrometry (SIM mode) applying a resolution of approximately 8,000.

Based on internal and external standards, recovery percentages have been determined of the internal PCDD/F congeners, as well as the quantities of PCDD/F congeners in the samples.

Experiments

Three experiments were carried out:

* blank experiment with only air flow through the facility;

* experiment burning tire with no chlorobutyl;

* experiment burning tire with chlorobutyl. The blank experiment with only air flow through the installation (no combustion) is meant to establish background levels of certain pollutants in the experimental systems (signals from environment, installation, sampling system and analytical systems).

Experimental results

Data on tires and residues (table 2)

Table 2 - composition of tires and residues(*)
Experiment                No chlorobutyl          Chlorobutyl
sample                 Tire        Residue      Tire    Residue
composition (mass %)        As received           As received
[H.sub.2]O                 2.71        0.45      2.39      0.31
Ash                       19.30       43.04     13.18     37.95
C                         67.87       59.58     69.67     64.26
H                          6.15        1.55      6.33      1.55
O                          3.30        5.40      2.91      6.29
N                          0.43        0.28      0.47      0.29
S                          1.33        1.12      1.32      1.29
Cl                       <0.004       0.036     0.074     0.059
PCDD/Fs, ng I-TEQ/kg   1.72-6.0   1.05-10.5   3.1-4.9   0.20-10
(*) The number of decimal places for the concentrations do not indicate the
accuracies of the concentrations: see also following tables.

The compositions of the tires are about the same with respect to the organic compounds. Only the ash content of the tire with chlorobutyl is relatively small and the chlorine content relatively high. The compositions of the residues are not very different; only the chlorine content of the residue from the tire with chlorobutyl is relatively high. Comparing the tires with the residues we noticed that the contents of ash and oxygen have increased during combustion, while the contents of hydrogen and nitrogen have decreased. For the other components no clear effects of the combustion process can be observed.

Temperatures offires

Typical temperature distributions of the fires as a function of the axial distance from the burner plate are shown in figure 1. As bare thermocouples have been used, radiation interaction with the fire and surroundings may occur leading to too low readings in high temperature zones and too high readings in low temperature zones. Additionally the thermocouples will be covered with a bit of soot during the fires.

From figure 1 it is clear that the fire temperatures are very low and strongly fluctuating. The maximum fire temperatures are recorded at the burner plate (below 440[degrees]C). Along the fire axis the temperatures drop very quickly. At 120 cm above the burner plate the temperatures are below 70[degrees]C. In general the fire temperatures are very low and drop very quickly through convective and radiative heat transfer but especially through mixing with cold air. The temperatures are about the same for both fires.

Blank and flue gas concentrations

The measured dry (flue) gas concentrations in the flue gas line during the blank experiment and during both experiments with tires (tire without chlorobutyl and tire with chlorobutyl) are presented in table 3 (the concentrations for both tires have been corrected for the blank values).

[TABULAR DATA 3 OMITTED]

From the measured flue gas concentrations it is clear that the mixing of the fires and the flue gases with air has lowered the concentrations considerably. The oxygen concentrations are higher than 20.5% and approach the concentration in air, while the [C0.sub.2] concentrations are below 0.2%.

It appears that the concentrations for [0.sub.2], [C0.sub.2], HCl, [Cl.sub.2] and PCDD/Fs are of the same order of magnitude during the three experiments. For [C0.sub.2], HCl and [Cl.sub.2] the measured concentrations are in the range of the detection limit of the measuring methods. The concentrations of products of incomplete combustion (CO, [C.sub.x][H.sub.y], particles/soot) and of [SO.sub.2] are as expected higher for the tire fires compared to the blank experiment. The concentrations of PCDD/Fs may be compared with measured concentrations in the range of 0.000001-0.000113 ng I-TEQ/[m.sup.3] in ambient air in The Netherlands (ref 3). It is clear that the blank concentration is higher than the ambient air concentration. So in the blank concentration contributions from the sampling systems, the sample holder and/or the analytical systems will be present. The flue gas concentrations of PCDD/Fs of the tire fires are of the same level as the result of the blank experiment.

The concentrations of CO, [C.sub.x][H.sub.y] [NO.sub.X], PCDD/Fs, particles/soot are approximately the same for both tire experiments. The amounts of solid residue are about the same as well. The concentrations Of [SO.sub.2], HCl and [Cl.sub.2] are somewhat higher for the tire with chlorobutyl. For PCDD/Fs the difference in flue gas concentration for the tire without and the tire with chlorobutyl are within the inaccuracy of the measurements.

In figure 2 an example of the stoichiometric dry flue gas concentration of CO during tire fires is given. From the figure it is clear that the concentrations of CO fluctuate quite strongly. The same is true for [C.sub.x][H.sub.y], [NO.sub.x] and [SO.sub.2]. A steady fire does not exist, which is common to many fires with continuously fluctuation position. The calculated emissions (g/kg input tire) for the tires are given in table 4. The emissions from both fires are very similar. Only the emissions of [SO.sub.2] and [Cl.sub.2] are relatively high for the tire with chlorobutyl.

[TABULAR DATA 4 OMITTED]

Discussion

Comparing the experiments

In table 5 conversion rates for different components of the tire producing air pollutants and residues are given.

[TABULAR DATA 5 OMITTED]

Although the calorific values of the tires are as high as those of fuel oils (>40 MJ/kg) combustion is far from complete during the experiments. This is due to the bad mixing between tire (gases) and air. Combustion occurs in the stoichiometric zone at the boundary between air and tire (gases). Inside the combustion zone air is short and the tire (gases) decompose without oxidation. From carbon and hydrocarbons soot is formed. Outside the combustion zone temperatures are very low (<<500[degrees]C) due to the infinite excess of air. Enough time at a high enough temperature is not available to oxidize the soot, CO and other products of incomplete combustion. Therefore compared to controlled combustion processes large amounts of carbon remain unburnt (in CO, [C.sub.x][H.sub.y], emitted particles/soot and solid residue).

The temperatures in the fire are so low that most probably no thermal [NO.sub.x] can be formed. The nitrogen compounds from the tire produce fuel-[NO.sub.x], but the conditions are unfavorable for a high conversion of N into [NO.sub.x]. A lack of oxygen exists in the flame root so N may be converted into molecular nitrogen. The nitrogen content of the tire is quite small. Therefore the [NO.sub.x] concentration in the flue gas is limited. The emissions for [NO.sub.x] vary around 18% of the maximum values for complete conversion of nitrogen into [NO.sub.x].

The conversion of S into [SO.sub.2] varies between 50 and 90%. Possibly other flue gas components with S have been formed, while the solid residue contains S as well.

The conversion of Cl into HCI varies between 60 and 150%. As [Cl.sub.2] has been emitted and Cl is found in the solid residue as well, the chlorine balance will be in error. The concentrations of HCl and [Cl.sub.2] in the flue gases are that small, that the results are quite close to zero. Inaccuracies in the results of the measurements are probably the main cause of the discrepancies.

Mass balances of PCDD/Fs

Mass balances on PCDD/Fs are given in table 6 for the fires of the tire without chlorobutyl and the tire with chlorobutyl. No conclusions can be drawn concerning production or destruction of PCDD/Fs in the fire. For PCDD/F levels minimum and maximum values are given. The rest terms vary around zero. The data in table 7 show that the concentrations in the tires and in the residues are quite similar.

Table 6 - mass balances on PCDD/FS for both
tire experiments
Flow         Input, ng           Flow          Output, ng
        1-TEQ/kg input tire                1-TEQ/kg input tire
Tire without chlorobutyl
Tire    1.72-6.0              Flue gases        0 - 1.43
                               residue        0.5 - 5.01
                                 rest        1.22 - (-) 0.44
Total   1.72-6.0               Total         1.72 - 6.0
Tire with chlorobutyl
Tire    3.1 - 4.9             Flue gases        0 - 1.66
                               Residue       0.09 - 4.43
                                 Rest        3.01 - (-) 1.19
Total   3.1 - 4.9              Total          3.1 - 4.9
Table 7 - comparison of PCDD/FS in tires,
residues and flue gases
                    Unit              No         Chlorobutyl
                                  chlorobutyl
Tire         ng I-TEQ/kg input     1.72 - 6.0     3.1 - 4.9
Flue gases   ng I-TEQ/kg input        0 - 1.43      0 - 1.66
             tire
Residue      ng I-TEQ/kg residue   1.05 - 10.5   0.20 - 9.98

Comparing conversion rates of chemical fires, pesticide fires and tire fires Comparing conversion rates for pesticides (parathion, chlorfenvinphos, dichlobenil, 2,4-D) and chemicals (1,2-ethanediamine, toluene-2.4-diisocyanate, 1,2-dichloroethane, ammonium nitrate) of previous projects (refs. 4 and 5) and tires it appears that large variations, dependent on the composition of the material on fire, occur (table 8). Compared to the maximum figures in the table for the tire fires the conversion rates are relatively small for CO, [C.sub.x][H.sub.y] and particles/soot. The amount of solid residue is quite large for tire fires. The formation of nitrogen oxides is relatively high for tire fires and the [Cl.sub.2] formation as well. Finally the emission of PCDD/Fs (I-TEQ) is extremely small compared to the emissions of fires of pesticides and chemicals. This is possibly caused by the high sulfur availability in tires which may make catalytic materials ineffective in the dioxin production process and reduces the [Cl.sub.2] content, necessary for dioxin formation, through reaction with [SO.sub.2] and [H.sub.2.O] to produce S03 and HCl (refs. 6-8).

[TABULAR DATA 8 OMITTED]

Comparing with emission limits for waste combustion The emission of PCDD/Fs from municipal solid waste combustion installations has been limited to 0.1 ng I-TEQ/[m.sup.3] at 11% [O.sub.2]) in a number of countries (e.g. Germany, The Netherlands). The emissions of the tire fires corrected to 11% [O.sub.2] would be (ng I-TEQ/[m.sup.3]): 0-0.067 (tire without chlorobutyl): 0-0077 (tire with chlorobutyl).

Of course a comparison with waste combustion installations is not fair, as those installations are meant to incinerate waste for several decades during some 8,000 hours per year, while fires normally are extinguished as quickly as possible.

Conclusion

The emissions of PCDD/Fs from fires of tires with and without chlorobutyl are in the same range.

The emissions of dioxins from tire fires for tires with and for tires without chlorobutyl are considered to be negligible compared to the emissions from fires of pesticides and chemicals and to the emissions from municipal solid waste combustion installations.

References

[1.] Bartelds, H., Broker, G., Ham, J. van and Vicard, J.-F., Expertise on the measurement and control of dioxins, Directorate General XI, EC, Society for Clean Air in the Netherlands/Vereniging Lucht, Brussels/Delft, 1991. [2.] Anonymous, Instructions for the sampling and analysis of 2,3,7,8-substituted polychlorodibenzo-p-dioxins (PCDD) and polychlorodibenzofurans (PCDF) emitted into the air from stationary sources, Ministry, of Housing, Physical Planning and Environment, Directorate for Chemicals, External Safety and Radiation Protection/IPC 655, The Hague, 1993. [3.] Schutter, M.A.A. and Jaarsveld, JA.: Dispersion and deposition of dioxins in The Netherlands, Report number 730501036, National Institute of Public Health and Environmental Protection, Bilthoven, The Netherlands, 1993. [4.] Molag, M., Bartelds, H. and Weger, D. de: Toxic combustion products from pesticide fires, Executive summary, Reference number 92-366, TNO Institute of Environmental and Energy, Technology, Apeldoorn, 1992. [5.] Bartelds, H., Smit, E. and Molag, M.: Emissions from medium scale fires of chemicals, Proceedings Industrial Fires, Major Industrial Hazards, Eds. Cole, S.T. and Wicks, P.J., ISBN 92-826-7014-7, EUR 15340 EN, Brussels/Apeldoorn, 1993. [6.] Gullet, B.K., Bruce, K.R., and Beach, L O.: Effect of sulfur dioxide on the formation mechanism of polychlorinated dibenzodioxin and dibenzofuran in municipal waste combustors, Environ. Sci. Technol, Vol 26, No. 10,1938-1943,1992. [7.] Griffin, R.D., A new theory of dioxin formation in municipal solid waste combustion, Chemosphere, Vol. 15, Nos. 9-12, pp. 1987-1990, 1986. [8.] Geiger, T, Hagenmaier, H., Hartmann, E, Romer, R. and Seifert, H.: Einflu[beta] des Schwefels auf die Dioxin- und Furanbildung bei der Klarschlammverbrennung, VGB Kraftwerkstechnik 72, Heft 2, 159-165, 1992.