Studies on surface properties of asbestos

ENVIRONMENTAL RESEARCH 41, 296--301 (1986)

Studies on Surface Properties of Asbestos IV. Catalytic Role of Asbestos in Fluorene Oxidation R.

ZALMA,

J.

GUIGNARD, E. COPIN, AND H. PEZERAT

Laboratoire de rdactivit~ de surface et structure, U.A. 1106, CNRS, Universitd P. et M. Curie, 4 place Jussieu, 75252 Paris Cedex 05, France Received M a r c h 26, 1984 To determine w h e t h e r a s b e s t o s is a basic catalyst, catalytic oxidation of fluorene to fluorenone in a h e t e r o g e n e o u s s y s t e m was tested. It was s h o w n that oxidation was quantitatively possible on the surface o f all basic minerals, s u c h as asbestos (chrysotile and crocidolite) and magnesia, but was not possible with acidic mineral materials such as silica. T h e effects of different factors are discussed. © 1986AcademicPress, Inc.

INTRODUCTION In several studies carried out in our laboratory by Bonneau and Pezerat (1983) and Bonneau et al. (1984) the active sites of the asbestos surface were identified as reductor and basic sites, i.e., electron-donor sites. In order to study the ability of these sites to play a catalytic role in some reactions at low temperatures, we selected a model reaction in basic catalysis: the oxidation of polycyclic aromatic hydrocarbons (PAH) with acidic character (Pines and Stalick, 1977; Pratt and Trapasso, 1960). More precisely, we chose the oxidation of fluorene to fluorenone, since absorption studies on asbestos of PAH, performed by J. Fournier and Pezerat (1984), revealed the existence of fluorene chemisorption, both for crocidolite and for chrysotile. Quantitative oxidation of fluorene in basic medium (Triton B) in a homogeneous phase was studied by Sprinzack (1958) and an ionic mechanism was proposed. More recently, Russel et al. (1968) have studied fluorene oxidation, again in basic medium, but in the presence of an electrophilic molecule. In this case, they showed a radical process after the initial formation of a carbanion. Lee-Ruff and Timms (1980), and Tezuka et al. (1975) oxidized fluorene directly by a preformed superoxide anion; in this case, they also suggested a radical process. MATERIALS AND METHODS Asbestos fibers used were provided by the UICC. Before use, the fibers were washed twice with hot benzene. Other solids in addition to asbestos were used: magnesia selected as a basic solid reference; silica (Spherosil XOB 15, Rh6ne Poulenc); and silica-alumina (Rh6ne Poulenc, 40% alumina). Solvents and chemical products employed were Merck or Prolabo products of high purity. Analysis by high-pressure liquid chromatography (HPLC) was carried out with 296 0013-9351/86 $3.00 Copyright © 1986by AcademicPress, Inc. All rights of reproduction in any form reserved.

R O L E OF A S B E S T O S IN BASIC CATALYSIS

297

isocratic elution on a C8 column (dimethyl-octyl-silyl) with a methanol-water (80/20) eluant. Fluorenone was identified with a reference product. Quantities of fluorene and fluorenone were obtained after standardization and evaluation of peak surfaces. The precision of the results was 3%.

EXPERIMENTAL Three methods were used; the oxygen gas was always introduced for six hr at room temperature or at 70°C: (1) in a solution of fluorene in either n-hexane or tetrahydrofuran (THF), with the solid in suspension (gas-liquid-solid system); (2) directly on a solid with preimpregnated fluorene (gas-preimpregnated solid system); or (3) as a carrier of fluorene on the solid (gas-solid system). In the gas-liquid-solid system, there was 5 g of solid in 150 ml of 4 x 10 -3 M solution. In the gas-preimpregnated solid system, the quantities of fluorene retained (adsorbed and deposited by evaporation) on the solid support were approximately as follows: 6.77 x 10 -4 mole/g for chrysotile (2.82 x 10 -5 mole/m2); 1.59 x 10 -5 mole/g for crocidolite (1.99 x 10 -6 mole/m2); and 8.6 x 10 -6 mole/g for silica alumina (6.1 x 10 -8 mole/m2). After reaction, the absorbed products were extracted with chloroform.

RESULTS The main results are summarized in Table 1. It was not possible, in this table, to compare the values given by the three methods for the same support. In fact, fluorenone quantities obtained were compared with introduced fluorene, which differs in the three methods and does not correspond to the same physical state. Consequently, only the column-by-column evolution according to different supports could be compared. The results show an oxidation on basic solids, asbestos or MgO, and not on solids with an acidic surface (silica or silica-alumina). Additional results were also obtained: Influence of temperature. With chrysotile in n-hexane solution, the yield of TABLE 1 FLUORENE OXIDATION IN THE PRESENCE OF SOLID SUPPORTS Method

Support

Gas-solidliquid system (n-hexane, dry, 70°C)

Gas-preimpregnated solid system (70°C)

Gas-solid system (70°C)

Chrysotile Crocidolite Magnesia Silica-alumina Silica Without support

99 a 23 19 -0 0

99 81 98 4 4 --

93 84 78 0 0 1

Note. Results are given as percentages fluorenone/(fluorene + fluorenone). a 30o(2.

298

ZALMA ET AL.

oxidation of fluorene rose from 73% at room temperature to 99% at 30°C. For crocidolite, the yield varied from 1.5% at room temperature to 23% at 70°C. Influence of water. In the test in solvent medium, the influence of water on the reaction was important. For example, with chrysotile in n-hexane, the yield fell from 37 to 0.5% when water-saturated oxygen instead of dry oxygen was used. On the other hand, if the solvent n-hexane was dried on zeolite, the yield rose to 73% at room temperature and 99% at 30°C. Influence of solvent. The role of solvent was significant. All other parameters being equal, experiments with chrysotile in solvents dried on zeolite gave results of 0% with CHC13, 43% with THF, and 73% with n-hexane. Without a solid, fluorene oxidation was not obtained. Amphiboles: Problem of fluorenone desorption. During testing under standard conditions in n-hexane medium at room temperature, with crocidolite for 6 hr, the solution was filtered and the fluorenone of the solution and that of the adsorbed phase on fiber were analyzed separately; 36 and 64%, respectively, were obtained, showing that in spite of the weak concentration of fluorenone, it tended to stay strongly adsorbed on crocidolite. For chrysotile, these proportions were inversed and the main part of the fluorenone was in the solution. Catalytic character of the reaction. Instead of carrying out the reaction for 6 hr in a gas-solid system (column 3, Table 1), it was continued for 5 days. Desorbed fluorene and fluorenone were extracted by the carrier gas, recovered in a bubbler at the end of the apparatus, and analyzed every 6 hr (with a variation of 15 rain). Fluorenone continuously appeared in the case of chrysotile and crocidolite. (Cf. Fig. 1.) For chrysotile, the graph showed that the desorbed fluorenone quantity increased with time and then reached a plateau. We can deduce that there was no "poisoning" of active sites. The solid played the role of a catalyst. In the case of crocidolite, the graph of desorbed fluorenone also presented a plateau after 80 hr, which suggests that the solid also played a catalytic role,

30 ~

~mmoles x 10 2

20

10

d 30

60

90

120

Hours

1~44~

FIG. 1. Fluorene and fluorenone recovered beyond the reactor at 6-hr intervals. (a) Fluorene in chrysotile. (b) Fluorenone in chrysotile. (c) Fluorene in crocidolite. (d) Fluorenone in crocidolite.

299

ROLE OF ASBESTOS IN BASIC CATALYSIS

though it differed from the chrysotile graph. The fluorenone quantity obtained reached a maximum and then decreased prior to stabilizing at a low level. This latter feature is probably due to the "poisoning" of some sites: the active sites of the fluorenone formed did not desorb from the strong sites of crocidolite. Study of mechanisms. For this reaction, two mechanisms--an ionic one and a radical o n e - - a r e possible. To discriminate between these hypotheses, the adsorption of fluorene on chrysotile was studied by electron spin resonance (ESR) on a VARIAN X3 spectrometer. A sample of chrysotile, pretreated as previously reported, was placed under a vacuum (10 -4 Tort) at room temperature, and the ESR spectrum was recorded (Fig. 2a). The field set, modulated at 100 kHz, varied in the region of 2500 G. A large signal appeared which was due to iron (Fe 3+) present in the sample, principally in the form of particles of adsorbed magnetite. This signal, badly resolved, had a g value of 3.18 for a resonance field fixed at 2123 G and a peak-to-peak width of 1500 G. If we submitted the sample to a vapor pressure of fluorene, the effect of the organic molecule was revealed by the appearance of a new signal at g = 5.31, due to disturbed Fe 3+ (peak-to-peak width: 100 G; Fig. 2b). But no signal due to an organic radical species appeared (near g = 2). The addition of oxygen (20 Torr) to the sample revealed no modifications in the above spectrum. After pumping (10 -4 Torr), the reversible effect of the organic molecule on Fe 3+ was revealed by the reappearance of the initial spectrum (Fig. 2c). The expansion of the scale around 3400 G confirmed the absence of an Organic radical species (Figs. 2a',b',c'). The same results were obtained at 77°K.

~'~

[DPPH

FIG. 2. ESR spectra: (a) Chrysotile, gain 2.0 x 103; (b) chrysotile + fluorene vapor, gain 1.25 × 103; (c) chrysotile + fluorene vapor + 02, gain 2 x 103 after pumping. (a') Chrysotile, gain 5 x 104; (b') chrysotile + fluorene vapor, gain 5 x 104; (c') chrysotile + fluorene vapor + O2, gain 5 x 10 4 after pumping.

300

ZALMA ET AL.

DISCUSSION The results in Table 1 show that in a heterogeneous system, fluorene oxidation to fluorenone is possible in a quantitative manner only in the presence of solids whose surfaces have a dominant basic character. This oxidation was only possible through the contribution of the oxygen gas. Indeed, in a test on gas-preimpregnated chrysotile in degassed solvent with a dry nitrogen stream of high purity (Air liquide N60, 02 < 0.1 ppm) at 70°C, a very low yield of fluorenone was obtained, falling from 99 to 2%. The small quantity of fluorenone obtained is probably due to traces of oxygen in the reactants. Water plays the role of a reaction inhibitor, and many of our results are proof of this. Yet, such an effect is only sensitive at a high concentration, since the yield in solvent medium is 99% at 30°C, though each fluorene molecule oxidized generates one molecule of water. Solvent influence has been studied very little, but results obtained with the three solvents (CHC13, THF, n-hexane) appear to agree with Gutmann's theory (1968) on the acceptor number (AN) of solvents: 0 for n-hexane, 8 for THF, and 23. ! for CHCI 3. The variation in AN and acidity is of the same order, so the very low yield obtained in CHC13 medium is probably the result of a blocking of the basic sites by solvent molecules. In solvent medium, at room temperature, yields are lower for amphiboles than for chrysotile; the analysis for crocidolite of the distribution of fluorenone between absorbed phase and solution implies that it is the desorption of this compound which limits the reaction. The differences noted between chrysotile and crocidolite at room temperature in their affinity for fluorenone are probably correlated with the strength of electron-donor sites on these minerals. Fluorenone, which has an acceptor character, can be trapped on this type of site to form a complex. It is probable that the crocidolite surface had stronger sites than chrysotile; this would agree with the results of Fournier and Pezerat (1984). Taking into account the ESR results (no radical species was detected) we propose a mechanism of a basic rather than a reductor type. Experiments with other PAH, carcinogenic or not, such as benzo[a]pyrene, 7,12-dimethylbenzanthracene, and phenanthrene, under the same experimental conditions, reveal no oxidation, since these compounds do h o t , a v e a sufficient electron-donor character. It is clear that the mechanism suggested here does not play a role in the oxidation of carcinogenic PAH in biological medium, which agrees with the works of Eastmann et al. (1983) and Chang et al. (1983). However, it is also clear, and it is our conclusion, that asbestos is a material capable of presenting catalytic activity of the basic type.

ACKNOWLEDGMENTS The authors thank the CNRS for financial support (Department of Physics); P. Dansette, G. Bram, J. Normant, and M. E Ruasse for help and advice; and L. Bonneau for the ESR study.

ROLE OF ASBESTOS IN BASIC CATALYSIS

301

REFERENCES Bonneau, L., and Pezerat, H. (1983). Etude des sites donneurs et accepteurs d'un 61ectron en surface des amiantes. J. Chim~ Phys. Phys.-Chim. Biol. 80, 275-280. Bonneau, L., Suquet, H., Malard, C., and Pezerat, H. (1986). Studies on surface properties of asbestos. I. Active sites on surface of chrysotile and amphiboles. Environ. Res. 41, 251-267. Chang, M. J. W., and Hart, R. W. (1983). The effect of asbestos on the uptake and metabolism of benzo(a)pyrene by normal human fibroplast. In "Seventh International Symposium on PAH, in press. Eastmann, I., Mossman, B. T., and Bresnik, E. (1983). Influence of asbestos on the uptake of benzo(a)pyrene and DNA alkylation in hamster tracheal epithelial cells. Cancer Res. 43, 1251-1255. Fournier, J., and Pezerat, H. (1982). Etude du mode d'adsorption des hydrocarbures polycycliques aromatiques sur les amiantes. Cas du ph6nanthr~ne. J. Chim. Phys. Phys.-Chim. Biol. 79, 589-596. Fournier, J., and Pezerat, H. (1986). Studies on surface properties of asbestos. III. Interactions between asbestos and polynuclear aromatic hydrocarbons. Environ. Res. 41, 276-295. Gutmann, V. (1968). "Coordination Chemistry in Nonaqueous Solutions. Springer-Verlag, Vienna. Lee-Ruff, E., and Timms, N. (1980). Superoxide mediated autoxidation of hydrocarbons. Canad. J. Chem., 58, 2138-2141. Pines, H., and Stalick, W. M. (1977). "Base Catalysed Reactions of Hydrocarbons and Related Compounds," Ch. 13, pp. 508-515. Academic Press, New York. Pratt, E. E, and Trapasso, L. E. (1960). The autoxidation of selected organic molecules in the presence of alumina. J. Amer. Chem. Soc. 82, 6405-6408. Russel, G., Benis, A. G., Geels, E. J., Janzene, G., and Morgan, A. J. (1968). Oxidation of carbanions. Oxidation of diarylmethanes and diarylcarbinols in basic solution. Adv. Chem. Ser. 75, 174-202. Sprinzack, Y. (1958). Reactions of active methylene compounds in pyridine solution. I. The ionic autoxidation of fluorene and its derivatives. J. Amer. Chem. Soc. 80, 5449-5455. Tezuka, M. Okrnatsu, Y., and Osa, T. (1975). Reactivity of electrogenerated superoxide ions. I. Autoxidation of 9,10-dihydroanthracene. Bull. Chem. Soc. Japan, 48, 1471-1474.