Aug. 5, 1960
COMMUNICATIONS TO THE EDITOR
4109
a t virtually any experimental conditions. In or free radical chain process. However, it is turn, since dosimetry gives the rate of initiation and very instructive to speculate on the meaning of conversion data yield the rate of propagation, the the work reported here. If the reaction is indeed radiation technique can give valuable insight into an ionic process, then i t is to our knowledge the first reported ionic chain reaction of hydrocarbons the nature of long chain reactions. Our recent work has emphasized radiation-in- a t these conditions. If the process is free radical duced reactions between low molecular weight in nature, which is in our opinion more likely, paraffins and olefins a t rather high temperatures and then the above initiation step cannot be correct. pressures. Some observations made on the meth- This initiation rate should proceed a t a rate given (CHJ molec. cc.-I ane-ethylene reaction seem to warrant publica- approximately by 1013e-101,000/RT tion a t this time although the work aimed a t de- set.-', where 1013is the normal unimolecular preciding whether this reaction is free radical or ionic exponential factor, (CHI) is the methane concenin nature is still incomplete. The results are of tration, and 101,000 kcal./mole is the CH3-H bond definite interest regardless of which one of these strength. At about 50 atm. of methane this rate is negligibly small a t 800°K., being in the order of alternative mechanisms is ultimately accepted. Experimental.-Static experiments were made in 105-106 molec. cc.-I. sec.-l. Now the high G electrically heated stainless steel systems having a value obtained a t 343 establishes the chain nature surface/volume ratio of 1 cm.-l. The radiation of the reaction and, furthermore, provides an apsource consisted of a 3200 curie Co60 ~ o u r c e , ~proximate value for the chain length a t the condiand the dosimetry was done by Fe++-Fe+++ con- tions of this work. However, the product of the version5 with an uncertainty of 15yG.Radia- above initiation rate and of the known chain length, tion and thermal experiments were made in identi- which should give the value of the reaction rate, cal reactors and a t the same conditions using is smaller by many orders of magnitude than the Phillips pure grade (>99.57,) methane and ethyl- observed thermal rate a t 426". Therefore, equaene. Products were analyzed by gas-liquid parti- tion (1) cannot represent the initiation step. However, a plausible initiation step can be writtion chromatography to *4yo of the amount present. The experimental conditions were : 55 ten which removes the difficulty. This involves atm. total pressure, 10 mole per cent. initial a bimolecular initiation between CHI and CzH4 CzH4 concentration, 343-426 temperature, and CH, CzH4 --+ CH,, + CzH6. (1) 0.12 X lo6 rad./hr. (absorbed). The radiation Such an initiation step is rarely discussed,6 but yields are integral G values obtained directly from the radiation-induced reaction since a t identi- is in fact highly likely since the reverse radical disproportionation reaction is well established and cal conditions no thermal reaction was observed. proceeds with zero or very low activation energy. Results and Discussion.-The homogeneous gas Therefore, step (1) should have an activation phase reaction between CH4 and C:H4 is a notori- energy not appreciably greater than its endoously slow process, presumably as a result of the thermicity. From available data on heats of forslow unimolecular formation of CH3. and H. from m a t i ~ n , the ~ ' ~endothermicity of (1) is estimated CH4. The results of thermal experiments showed to be 60 kcal./mole. Therefore, initiation by step no conversion in six hours a t 343". In the pres- (1) should proceed a t an approximate rate of ence of radiation, the reaction between CH4 and e-60s000/RT(CH4) is the normal (C!H4), where CzH4 gave 3.1% conversion of CH4 and 27.0% con- bimolecular collision number. At 50 atm. methversion of CzH4. The total reaction product con- ane, 5 atm. ethylene, the initiation rate calculated tained, in mole per cent., 3.0 C2H6, 30.4 C3H8, from this expression is approximately that required 1.4 C3H6, 15.7 iso-C4H1o1 21.2 C4H8, 12.0 iso-CjHlO, to explain the measured thermal rate corresponding and 5.8 C6H12. The radiation yield a t these to the chain lengths estimated from this work. conditions is 1200 molecules of C3 and higher (6) N. N. Semenov, "Some Problems in Chemical Kinetics and Remolecular weight products per 100 ev. absorbed. Princeton University Press, Princeton, New Jersey, 1958. At 426 both the thermal and radiation-induced activity," (7) M. Szwarc, Chem. Rea., 47, 75 (1950). reactions are appreciably rapid. For example, P. J. LUCCIIESI RESEARCH AND ENGINEERING the thermal reaction gives in six hours CH4 and Esso COMPANY, LINDEN,NEWJERSEY W. BARTOK CzH4 conversions of 5.0% and SO%, respectively. RECEIVED JUNE 20, 1960 At this temperature a CHI conversion of 4.1% and a C2H4 conversion of il.0Y0 were obtained after two hours of irradiation. There is, therefore, a AN INVESTIGATION OF T H E REACTION PRODUCTS thermal reaction a t 426" that is accelerated by OF BIS-CYCLOPENTADIENYLTITANIUM radiation. Above 426" the thermal reaction is so DICHLORIDE AND VARIOUS ALUMINUM ALKYLS BY ELECTRON SPIN RESONANCE rapid that acceleration by radiation is difficult to notice. Sir: A'ork now in progress is designed to obtain Bis-cyclopentadienyl titanium dichloride was detailed product composition versus conversion allowed to react with about a thousandfold excess data for both the thermal and the radiation- of several aluminum alkyls in benzene solution and induced reaction. X strict comparison a t the the electron spin resonance (es.r.) spectra of these same conditions and conversions is, of course, solutions were observed a t 3.2 cm. wave length and the best direct evidence one can use in deciding a t 20". The concentration of the titanium whether the radiation-induced reaction is an ionic compound was between and molar in (5) N. J. Miller, J. Chem. Phys., 18, 70 (1950). each case. The reactions with several aluminum
*
+
4110
COMMUNICATIONS TO THE EDITOR
alkyls were investigated : Ale&, Et%%l, isoBu3.41, EtZAlC1, and EtAlC12. The tri-isobutyl system has been investigated previously by e.s.r. The nature of the e s r . spectrum was found to depend upon the aluminum alkyl which was allowed to react with the bis-cyclopentadienyltitanium dichloride. Furthermore, the e.s.r. spectrum of the EtAlClz system was found to change with time ( t ) . The intensity of the signal of this system was observed to increase slowly during the period t = 25 rnin. to t = 240 rnin. and had the appearance of Fig. IC with an average g value of 1.9753 a t t = 90 min. The line width, AH, is 20.3 gauss. After a period of 51 days the appearance of the spectrum had changed to that shown in Fig. Id. The spectrum shows six equally spaced and approximately equally intense components suggesting an isotropic hyperfine interaction with the AIz7 nuclear moment ( I = 5 / 2 ) . The slight irregularity in the apparent intensities of the six hyperfine components is consistent with the presence of a small amount of the species producing the spectrum of Fig. IC. At t = 8 days the e.s.r. spectrum had an appearance intermediate between Fig. I Cand Id.
Vol. 82
The g values and resonance line widths, AH (as defined in Fig. l), are given in Table I, a t t = a. TABLE I Aluminum compound
Me3.41 Et& EtZAlC1 Eta%lClz 1~0-B~s.11 Iso-BuaAl'
Fig. , .
la lb Id ..
..
g-factor (i0.0005)
AH (gauss) ( 1 0 . 5 gauss)
1.9762 1,9760 1.9750 1,9745 1.9872 1.988
13.3 10.7 20.3 38.7 28.0
The product of the reaction between Et& and CpzTiC12 has been shown to be the compound CpzTiCIzA1EtZ(I), which has one unpaired electron.2 A preliminary X-ray structure investigation showed that there is a tetrahedral disposition of groups around both the titanium and aluminum atoms and that these atoms are linked by chlorine bridges3 The final products of the reactions based upon EtzAICl and EtAIClz are most probably the compounds CpzTiClzAIEtCl (11) and CpZTiC12AlC12 (111), respectively, with structures similar to (I)." The rates of formation of these compounds from CpzTiCIZand the appropria b r ate aluminum alkyl have been shown to be in the order Et3A1>Et2,41C1> EtA1C12, both by absorption spectral measurements in the visible r e g i ~ n , ~ . ~ and by electrical conductivity and polarization measurements.6 The first step in the reactions is the rapid formation (within a few minutes) of a complex containing aluminum, which is thought to be a diamagnetic species of the type CpZTiClZ.AlEt2CI5J or Cp2TiC1Et.A1EtC1z,7and which changes slowly to the final, paramagnetic, product. Our C d measurements of the EtA1ClZsystem are consistent with this view since we observe a slow increase in radical concentration after the initial mixing of the reactants. The spectra also indicate that more than one paramagnetic species is present. The spectrum of the Et2AlC1 system contains poorly hyperfine structure of six comI ponents.resolved 1 1 This is also consistent with a smaller I 1 hyperfine interaction with the AIz7 nuclear moment. Fig. 1.-E.s. spectra a t 3.2 cm. wave length and If we assume that AH is determined mainly by the 20' of mixtures of CpeTiCla in benzene with (a) EtsAl, (b) hyperfine interaction with A127,then the increase EtZAlC1, and (c) and (d) EtAICL. of AH in the sequence I < I I < I I I would indicate that the unpaired electron distribution favors the None of the other systems investigated showed region of the AI atom in this same order. This a similar time dependence of the e.s.r. This ob- interpretation is also consistent with the observaservation indicates that the final paramagnetic tion that A H for the Paramagnetic compound product was already formed a t the time of the first CpTiClz, which contains no aluminum, is only e.s.r. measurement: t = 1 day for Et& and t = about 6 gauss in solution and exhibits no hyper30 rnin. for EtzAICl. The e.s.r. spectra of the para- fine structure.8 The unpaired electron distribumagnetic products of these reactions are shown in tion in the complex may be influenced by the Fig. l a and Fig. l b , respectively. The e.s.r. inductive effect of the unbridged C1 atoms; the spectra of the product of the reaction of the titan- presence of this inductive effect is apparent from ium compound with Me& are essentially identical with those observed in the EtA1 system. The (2) G. N a t t a , P. Pino, G. Mazzanti and U. Giannini, THISJ O U R N A L , e.s.r. spectrum given by the isoBuAl system was 79, 2975 (1957). (3) G. N a t t a , P. Corradini and I. W. Bassi, ibid., 8 0 , 209 (1958). similar to that observed by Shilov, et al.,l and ex( 4 ) G. Mazzanti, private communication. hibits a hyperfine structure of eight equally spaced ( 5 ) D. S. Breslow and N. R. Newburg, ibid., 81, 81 (1959). components with the approximate intensity rela( 0 ) E. W. Randall and L. E. Sutton, paper given a t the International Congress of Pure and Applied Chemistry, Munich, September, 1959: tion 1 : 3 : 4 : 4 : 4 : 4 : 3 : 1 . E . W. Randall, D. Phil. Thesis, Oxford, 1959. ( 1 ) A. E. Shilov, A . K . Zefirova and H. H . Tikhomirova, Z h x v Az K h i m . , 39. 2113 (1959).
(7) W. P. Long and D . S. Breslow, THISJ O U R N A L , 8 2 , 1965 (19(jO) (8) P. D. Bartlett and B. Seidel, private communication.
Aug. 5, 1960
COMMUNICATIONS TO THE EDITOR
4111
*
dipole moment measurements ( ~ I I = 6.4 0.5D andp, = 4.6 f.0.50).6 The compositions and structures of the products of the systems based upon MeAl and isoBu&J apparently are not known, but the formulations Cp2TiC12AlR2, where R = alkyl, may well apply. For R = Me, the e.s.r. spectrum is consistent with this assumption, whereas the large width of the spectrum observed in the case R = iso-Bu cannot be explained in this way, if the treatment used to explain the widths of the other spectra is correct. The hyperfine pattern given by the iso-BuAl product is consistent with the interaction of the electron spin with the A12' nuclear moment and with two equivalent protons which have equal coupling constants.' The e.s.r. results suggest that the final paramagnetic product of this reaction is not of the same structural type as given by the other aluminum alkyls which we have studied. We gratefully acknowledge the support of this work by the Office of Naval Research and the National Science Foundation.
Its ultraviolet spectrum is very similar to that of 1,1'-dimethylferrocene,2 containing a weak maximum a t 320 mp (Emax 492), a shoulder a t 250 mp ( E 4,920), and high end absorption ( E Z ~13,600). Although the ultraviolet and infrared spectra of I are characteristic of alkyl ferrocenes, the n.m.r. spectrum is The n.m.r. spectrum of the close analog, 1,l'-diisopropylferrocene contains a single peak due to the ferrocene ring protons ( T = 5.97),5 together with the expected doublet (7 = 8.86) and multiplet4 (T = 7.36) due to methyl and methine protons, respectively. Similarly the spectrum of 2,3-diferrocenyl-2,3-dimethylb~tane~ contains three singlet peaks a t 7 = 8.69 (methyl), 5-94 (unsubstituted ring) and 5.86 (substituted ring). The spectrum of I contains the expected methyl proton singlet ( T = 8.74)) but the ring protons are found in two widely separated triplets ( T = 6.11 and 5.37; J = 1.6 c.s.). The considerable chemical shift is best explained as arising from the ring protons' unequal distance from the iron nucleus in a molecule where the rings are DEPARTMENT OF CHEMISTRY AUGUSTH . MAKI tilted; this interpretation is thus in accord with HARVARD UNIVERSITY EDWARD UT.RANDALL the mono-iron molecular weight above. CAMBRIDGE 38, MASSACHUSETTS Since protons close to a transition metal nucleus RECEIVED JUNE2, 1960 have been shown to give n.m.r. peaks a t relatively higher field in similar compounds (cf. cyclopenta1,I '-(TETRAMETHYLETHYLENE)-FERROCENE. dienylcyclopentadienecobalt),7 the ring 2-protons OBLIQUITY AND N.M.R. IN BRIDGED FERROCENES are assigned to the higher field peak ( r = 6.11) Sir: and the ring 3-protons to the lower ( r = 5.37). Similar two-carbon bridged ferrocenes are formed We wish to report the preparation and properties of 1,l'-(tetramethylethy1ene)-ferrocene (I), con- from other fulvenes*; however, an alternative reactaining an unusual two-carbon bridge linking the tion path can predominate. Thus, the major product from treatment of cyclopentylidenecyclotwo cyclopentadienyl rings.
i
2.00
3.00
4.00
5.00
7.00
6.00
8.00
9.00
10.00
7
Fig. 1.
Treatment of 6,6-dimethylfulvene' with sodium pentadiene (6,6-tetramethylenef~lvene)~ with sometal dispersed in tetrahydrofuran, and then dium, then ferrous chloride, is 1,l'-di-(1-cycloferrous chloride, gave a mixture in which the major pentenyl) -ferrocene (11), m.p. 108-109 O [Anal. product (albeit in low yield) was the bridged comJ. Bellamy, "The Infra-red Spectra of Complex Molecules," pound (I); red crystals, sublimes 130' (8 mm.), 2nd(3)e dL.. , John Wiley and Sons, New York, N.Y.,1958, p. 13. m.p. 168-169' (crystal modification 129-130') (4) Additional differences between bridged and non-bridged alkyl [Anal. Found: C, 71.83; H, 7.62; Fe, 20.68; ferrocenes, to be discussed in the complete paper, are lower stability (standing, as crystals or in solution) and decreased reactivity (toward mol. wt. (cryoscopic), 2851. acylation) of the bridged compounds. Detailed disThe infrared spectrum of I is generally similar Friedel-Crafts cussion of n.m.r. spectra, in particular of splitting, is also deferred to to that of 1,1'-diisopropylferrocene, with bands the full report. a t 3100, 917 and 850 cm.-I (alkylferrocene),2 (5) G.V. D.Tiers, J . Phys Chem., 69, 1151 (1958). 1451, 1377 and 1367 crn.-l (gem-dimeth~l).~ (6) K . L. Rinehart, Jr., P. A. Kittle and A. F . Ellis, THISJOURNAL, sa,
(1)
G.Crane, C. E. Boord and A . L. Henne, THISJOURNAL, 61, 1237
(1945). (2) K . L. Rinehart, Jr., K . L. Motz and S. Moon, i b i d . , 1 9 , 2749 (1957).
2082 (1960). (7) Cf., c.g., G.Wilkinson and F. A. Cotton, Progr. in Inorg. Chem., 1, 1 (1959), and Ref. 1.3 therein. ( 8 ) R . L. Pruett and J. E. McMahon, unpublished results. (9) E.P. Kohler and J. Kable, THISJOURNAL, I T , 917 (1935).
