inNorthern MSpectrochtmics Acts.1964. Vol.20,pp.721to 727.PO~IIUOII PremLtd. printed
Raman spectra of the aluminum bromide addition compounds in methyl bromide and in ethyl bromide BERNARD RIFE and KENNETH C. BALD* Department of Chemistry, Saint Louis University saint Louis, Mi&3ouri (ReceiuecE 0 Sept. 1963) Abstract-The Raman spectrum of aluminum bromide in cyclohexane was obtained to eetabliah the dimer, Al,Br,, frequenciesin an inert solvent. The spectra of solutionsof aluminumbromide in methyl bromide and ethyl bromide were determined and the frequencies assigned to the addition compounds MeBr : AlBr, and EtBr : Al&,. The carbon-brominefrequency ie lower in a complex than in the parent alkyl halide, the extent of lowering being greatest for the ethyl bromide complex. The resulti support the postuleted mechanismof the FriedelXrafta reactiona in which these compounds participate aa intermediatea.
spectral investigations of aluminum bromide dissolved in methyl bromide and ethyl bromide, respectively, were undertaken to gain information on the addition compounds purported to be formed in these solutions. Studies by JUNQK et al. [l] on the kinetics of the alkylation of aromatic compounds by methyl or ethyl bromides in the presence of aluminum bromide (Friedel-Crafts reaction) led them to conclude that CH,Br: AlBr, and C,H,Br: AlBr, are formed as reaction intermediates. Evidence that these complexes are stable in the absence of aromatio compounds has been obtained by phase equilibrium studies  and dielectrio constant measurements . Ihese complexes are of interest because of their importance in the meohanism of some Friedel-Crafts reactions, and because of the rarity of molecules containing halogen atoms bonded to two other atoms. A Raman study was undertaken to give more detailed knowledge of the structure of these molecules than could be gained through the previously applied experimental methods. It is well known that aluminum bromide exists as a dimer, AlBBr6, in the pure molten state and in inert solvents. The Raman spectrum of molten Al,Br,  has been measured by many investigations but no spectrum has been reported for the salt in an inert solvent. It was considered worthwhile to obtain a spectrum of aluminum bromide in an inert solvent to ascertain that a given frequenoy RAMAN
* Present address: M. W. .Kellogg Company, Researoh and Development Divieion, Jersey City, New Jersey. PI H. JUNGK, C. R. SMOOT,and H. C. Boom, J. Am. Chem. Sot., 78,2186 (1966). PI H. C. BXOWN and W. J. WALIAOE, J. Am. Chem. Sot., 75, 6279 (1963). D. a. WALXEB, J. Phya. Chem., 64,939 (1960). J. J. BIJEBAQEand A. B. GARZUTT,J. P&s. Chem., 68,730 (1952). 131 1. A. SHEKAand Z. A. SHEKA, Dokl. Akud. Nauk., SSSR, 78,739 (1960), (Chcm.AM., 44, 1056Oi). 141 See Table 1. 721
B~IWAXDRICE and KENNETH C. B~CD
ooourring in an a&y1 halide solution is due to compound formation and not a simple solvent effect shift. Therefore the Raman spectrum of aluminum bromide oyolohexane solutions was also measured.
EXPERIMENTAL Ahnninum bromide is very reactive with atmospheric moisture and the impurities commonly present in the alkyl bromides. However, with cam it ie possible to prepare colorlegs solutiona for Raman measurement& A few of the pertinent details will be given below. The 8hIminum bromide (Finher Certified Reagent) w&e given E preliminary purification by dietihation under a nitrogen atmosphere before 3nal sublimation under vacuum conditions. Methyl bromide (The Biatheeon Company) wee condeneed on and distilled from three separate small portione of sublimed aluminum bromide before tinal transfer to the Raman tube containing a known weight of aluminum bromide. Ethyl bromide (F&her Certified Reagent) w&e dried over phosphorus pentoxide and then given the three aluminum bromide treatmenti described above before &al condensation in the Raman tube. This eolvent oontained more color-producing impurities than the methyl bromide. The solution left in the firet flack wae deep red, that in the second wae yellow, and that in the third wae oolorle4le. Cyolohexane (Fisher Spectrograde) wae dried over sodium hydride and given a eingle ahnninum bromide treatment. No color formation wae observed in this solvent. In the preeent work it wae found that the ethyl bromide-aluminum bromide system is photoeensitive. A eolution of aluminum bromide in ethyl bromide wae allowed to stand in a m&d Raman tube for eeveral daye with no apparent color formation. However, a thirty minute expoeure of thie solution to the Raman lamp8 CBu88da strong yellow coloration. Thie color darkened to a dark brown over a period of several daye after the irradiation. The decomposition initiated by the radiation ie epparently a 13elf-8uet&ningreaction. The ethyl bromide-&m&mm bromide system from which the Raman epe&rum listed in Table 3 was obtained was inspected frequently during the exposure. The exposure was teuminated when color formation became apparent. A three prism epeotrograph (construe by Lane-Wells) wae used in this work. This kmtmment hae 8 diapemhn of 76 om+un et the exciting wave length, 4368 A. The spectra wme~reoorded photographically. The excitation unit, deeigned primarily for Raman m eaeuremente on gaeee, contained four kge vertioal Toronto Type meroury am lampe. The long narrow Raman tube ueed in thie study had an inside diameter of 0.5 cm and an ill uminated length of approximately 100 cm. Polarization meaauremente were made by the method of dual expoeurea. The sample tube ww &at enoomptwed by a aylinder of polaroid that t ranemitted vertically-polarized light for one erpoeum and subeequently by a Polaroid oylinder that transmitted horizontally-polarized light for a aewnd expoeure. A numberof horizontal vanea were placed between the lamps and the Rtunan tube when these expoeurea were made. DATA The Remap frequencies of 8 dilute eluminum bromide in oyclohexene solution are shown in the Srst oolumn of Table 1. The cyclohexane frequencies in solution m nneh8nged from that of pure cyclohexane and 8re therefore omitted. For purpose of oomparison 8ll th e previously reported Raman spectra of the pure molten salt are also listed in Table 1. There is a close oorrespondenoe between the rlaminnm bromide fiequen~ies in this solution and its melt frequencies; the solvent 08uam no signi5e8nt shifts. Three of the wo&k Raman lines reported for the melt were not observed in the ~utk~ but this is most likely due to the low concentration of the salt. It is imbm&bg$het one of these fiequenaies, that ocourring at 407 om-l, was observed
Reman spectra of the aluminum bromide addition compounds
bromide solution. Aluminum bromide is muoh more soluble in methyl bromide than in cyclohexane. Even though oonsiderable complexing occurs in the alkyl halide solution (as will be discussed later) the concentration of unreacted Al,Br, was still greater in this solution than in the cyclohexane and a more intense spectrum of the dimer was obtained.
in the methyl
Table 1. Raman spectrum of aluminum bromide in cyclohexane and of molten aluminum bromide Solution Frequencies (cm-‘)* 68 (7b)
Frequencies (cm-‘) t 67 (6b) 79 (6)
113 (6) 140(6) 186(O) 208(10) 223(l)
107 (3) 136 (4)
@=9§ 67 (6)
73 (6) 112 (3) 140(6) 176(2) 204(10) 221(O) 291(O) 407 (2) 491(3)
112 (3) 140(6) 176(2) 204(10)
291(t) 407 (24) 491(3)
* This work: mole ratio Al,Br&H,, = 0.066 t E. J. ROSENBAUY,J. Chm. Phyu., 8, 843 (1940). $ H. GERDINOand E. SMIT, 2. phy8. CLm., BSl, 217 (1941). 8 K. W. F. KOELRATJSCH and J. WAONEB, 2. p&u. Chem., B#j, 186 (1942).
The Raman. frequencies below 600 cm-l of the methyl bromide-aluminum bromide and ethyl bromide-aluminum bromide systems are reported in Tables 2 and 3, respectively. This region is of primary interest since it includes all the characteristic frequencies of the C-Br and Al-Br bonds. Above 600 cm-l, where C-H and C-C (in the case of ethyl bromide) fundamental frequencies occur, the spectra were found to be very similar to those of the parent alkyl bromides aud were not considered of sufficient interest to be reported here. Polarization measurements were made only on the methyl bromide-aluminum bromide system. Similar measurements were not made on the ethyl bromide-aluminum bromide system because of the istability of this system under irradiation. In the last columns of Tables 2 and 3 are the allocations of the observed frequencies to the oomplexes formed, or to the unreacted aluminum bromide or alkyl bromides, respectively. In most instances the choice is unambiguous but in a few cases the intensity or width of a spectral band can be beet explained if it is assumed to be composed of the unresolved frequencies of two moleoular spe&a. In the allocations cognizance was also taken of the fact that the extent of oomplex formation is greater in the ethyl bromide solution than in the methyl bromide system. This is very clearly indicated by the intensity of the 206 om-1 frequesoJI of unreacted Al,Br,, which has a relative intensity of ten in pure AlsBr,, &o ten in the methyl bromide solution, but-only three in the ethyl bromide solution.
BERNARDRICE and KENNETHC. BALD
Table 2. Reman frequenciesbelow 600 cm-l of an aluminum bromide-methyl bromide solution* Frequencies (cm-l)
69 (7b) 103 (7) 138 (6) 153 (3b) 194 (10) 206 (10) 219 (1) 271 (4b) 407 (1) 452 (3b) 485 (2) 550 (lob) 591 (9)
DP DP DP P P P DP P P
Complex; AlzBra Complex; Al,Br, AlaBr, Complex Complex AI,Br, Al,Br, Complex Al,Br, Complex Al,Br, Complex Methyl bromide
* Mole ratio: AlBr$CH,Br
Table 3. Raman frequenciesbelow 600 cm-l of an aluminum bromide-ethyl bromide solution* Compound
69 (6b) 103 (8)
127(2) 138(2) 171(O)
EthylBromide Complex Complex Complex
340(Ob) 400(0) 449(3b) 473(lb)
Complex Complex Ethyl Bromide
502 (9b) 555 (4)
* Moleratio: AlBr&H,Br = 0.90. ASSIGNMENT OF FREQUENCIES OF CH,Br : AlBr,
AND C,H,Br : AlBr,
A very useful guide for the assignment of some of the frequencies of the vibrational modes of the complexes were the assignments for silicon tetrabromide . The mass and electronegativity of silicon is very close to that of aluminum. If one considers the simple X : AlBr, model for the complexes, the effective mass of X is close to that of bromine. The X: AlBr, model may be considered a distorted tetrahedron. The antisymmetrical stretching frequency of SiBr, occurs at 487 cm-i. Analogous frequencies occur at 452 and 449 cm-l for the methyl and ethyl complex, 
Infrared and Rawn
New York: (1945).
Spectra of Polyatomic
167 Van Nostrand
Raman spectra of the aluminum bromide addition compound8
respectively. The totally symmetric stretching frequenoy of SiBr, has a magnitude are of 249 cm-l. Comparable frequencies for CH,Br: AlBr, and C,H,Br:AlBr, the intense, polarized frequencies at 194 and 196 cm-l, respectively. The two bending modes of vibration of SiBr, have the frequenoies 137 and 90 cm-l which suggest the assignment of the 103 and 69 cm-l frequencies whioh occur in both complexes to terminal Br-Al-Br deformations. At this point the comparison of the complexes to SiBr, has served its purpose. The 650 and 502 cm-l frequencies observed in the respective complexes can be assigned unambiguously to a carbon-bromine stretching mode. The related frequencies in the parent alkyl halides occur at 691 and 666 cm-l. The most reasonable candidates for the frequency characteristio of the bond between the aluminum atom and the bromine atom of the alkyl halide appear to be the 271 (polarized) and 281 cm-r frequencies of the methyl and ethyl complexes, respectively. There are no spectral assignments available in the literature whioh could be used to anticipate the magnitude of the frequency of this bond. Table 4. Raman frequencies of CH,Br : AlBr, C,H.Br : AlBr. below 600 cm-l CH,Br : AIBr, (cm-‘) 69 103 153 194 271 452 650
C,H,Br: AlBr, (cm-‘) 89 103 127 196 221 261 449 602 171 340 460 413
Dwxiption Br-Al-Br Br-Al-Br C-BlcAl Al-Br C-C-B? -BPAl Al-Br C-B+ 69 + 103 69 + 281 2 x 196 196 + 261
Tcwminal deformstion Temdnal defwmation Defomlation Terminal sym. &r&ah Deformetion Internal atre&.h Terminal antisym. rtretmh Stretch
The only experimental frequency that remains to be assigned in the CHsBr: AlBr, spectrum occurs at 153 cm-l and could very probably correspond to C-Br-Al deformation. An analogous mode should occur in the ethyl complex. There is no frequency of the same magnitude in the latter compound, the two closest occurring at 171 and 127 cm-l. A better intensity correspondence between the two complexes is achieved if the 127 cm-’ frequency in the spectrum of C,H,Br: AlBr, is given this assignment, and this is done tentatively. Of the remaining frequencies in the ethyl bromide complex, the line at 221 om-1 is assigned to C-C-Br deformation. The frequency of this motion in unreaoted ethyl bromide is 295 cm-l. All the other frequencies are of very low intensity and can be accounted for as overtone or combination bands. The above assignments of the complexes are summarized in Table 4. It was brought to the authors’ attention by one of the referees that TAWE and MOYER [a] had obtained the Raman spectrum of the methyl bromide-aluminum bromide complex. Their  R. C. TAYLOR and J. R. MOYER, Abetructe 135th Natiotaal Meet&g ACE.,Boston, April 1969; J. R. MOYER, Ph.D. Thesis, University of Michigan, 1958, D&e&a&m abebreots 19, 3149 (1959).
RICJEand l[(ENNE!l’JiC. BALD
experimental frequencies designed to the complex are 71(w), 102(m), 149(w), 199(e), 277(m), 464(w), 467(w) and 664(s) cn-1. These are in general agreement with the data reported by the euthorcof this paper. However,some of the assignments by TAYLOB and Movuu differ from those presented here. They have no aeeignmentfor the internal -Br-Al stretchingmode. Their 277 cm-1 frequency (which is e possible candidate for this motion) is instead assigned to the terminal Al-Br asymmetric stmt&mg mode of vibration. This frequency appears to the author of this paper to be too low for ass&ment to the latter motion by analogy with the related frequency in SiBr, and from analysis of the data available for AlCls addition complexes. On the other hand, TAYLOB end MOYEBaeoribethe 464 cm-l frequency to a difference band (6643-101~7). The intensity (3b) reported hereinof the comparable frequency at 462 cm-l appears to the authors of this paper to be too great to support such an assignment. Its intensity and magnitude make the 462 cm-l frequency a more likely candidate for the terminal Al-Br stretching mode of vibration. The 71,102and 149cm-l frequencies are assigned by TAYLOR aud MOYER to deformations M was done with the related frequenciesin this paper. However the descriptionsof the respective deformations differ in the two papers. Since more elements of choice are usually present in the a#[email protected]
of low-lying frequencies,and with the limited experimental data now available, it is not too pro&able to evaluete here the pros and cons of the alternative assignments.
DISCUSSION The Rsman spectra of the methyl bromid~aluminum bromide and ethyl bromid~shnuinum bromide systems support the existence of addition compounds with the enticipated RBr: AlBr, structure. If ion pairs, R+AlBr,-, were the principal form of the complexes, no frequencies characteristic of the carbonbromine bond would be observed. The intense bands at 550 and 502 cm-l in the methyl and ethyl bromide solutions, respectively, are most certainly assigned to carbon-bromine stretching in the complexes . The carbon-bromine characteristic frequency is lower in the complexes than in the respective parent alkyl bromides. The frequency is 9.5 per cent lower in the ethyl complex than in ethyl bromide (502 cm-l vs. 555 cm-l) while the decrease for the methyl complex is 0.9 per cent (550 cm-l vs. 591 cm-l). The decrease in the carbon-bromine frequency on complex formation is too greet to be accounted for by a mass effect alone. This can be shown by a simple calculation based on the diatomic model. Changes of the magnitudes observed for the respective complexes would require that the BrAlBr, group acts as a rigid mass in this mode of vibration. It seems reasonable to conclude that part of the lowering of this frequency on complex formation is due to a decrease in the carbon-bromine stretching force constant. JENQK el al. [l] in their study of the kinetics of the Friedel-Crafts reaction concluded that the experimental results are best explsined by a reaction mechanism in which the aromatic compound attacks a polarized compound, R”+Br”- : AlBr, (when R is methyl or ethyl), and that the positive character of the alpha carbon atom is increesed on complex formation. The Rsmtln study supports this conclusion since an increase in the polar nature of the carbon-bromine bond is in sooord with the decrease in the force constant previously postulated. [7’l Infraredmeasurementshave been used to conihm a similar complex for methyl chloride in J. Whys. Chern., 66, antimony pen&chloride solution, i.e., CH,Cl:SbCl,. H. M. NELSON, 1330(1962).
Raman spectra of the aluminum bromide addition compounds
It was further proposed by JUNQKet al. that the greater rate of ethylation as compared to methylation in the Friedel-Crafts reaction is principally due to the more positive character of the alpha carbon atom in the ethyl complex. The support for this contention from the Raman spectra is very tentative. As was shown, the per cent decrease in the carbon-bromine frequency on oomplex formation is greater for the ethyl complex. However, a complete force constant treatment would be necessary before a valid oomparison could be made of the oarbon-bromine bonds in the two complexes. This is not feasible as long as such important parameters as the carbon-bromine-aluminum bond angle and the alkyl brominealuminum bond distance cannot be estimated within reasonable limits of certainty. Aoknowledgement In the process of this reearch it W&Blearned that D. CL WALKEE, Humble Oil and Refining Company, Baytown, Texas, had made Raman meesurementa on the methyl bromidsaluminum bromide system. Appreciation is expressed here for the opportunityto examinehis experimental reaulta and also his unpublished data on the diethyl ethtw-eluminum bromide s-m.