(1-13 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY AS A PROBE TO STUDY SUBSTITUTED ARYL CARBOCATIONS A DIsserIaflo-n , for {In Denna of DI», D. MICHIGAN STATE UNIVERSITY William T. Chambers I975 ' L IE R A R Y Michigan State University This is to certify that the I thesis entitled C-l3 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY AS A PROBE TO STUDY SUBSTITUTED ARYL CARBOCATIONS presented by William T . Chambers has been accepted towards fulfillment of the requirements for I Ph . D, degree in Chemis try 97/ [CY/[4; 4/94Cttm 0-7639 Major professor i . E l i I ABSTRACT C—l3 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY AS A PROBE TO STUDY SUBSTITUTED ARYL CARBOCATIONS By William T. Chambers We have determined the carbon—l3 NMR spectra of solutions of a number of aryl substituted carbocations in highly acidic media. We have observed a linear correlation of the carbon-l3 chemical shift of the carbocationic centers in the arylcyclopentyl cations with those in the arylcyclohexyl and symmetrical 1,1—diaryl—l—ethyl cations. However, a plot of the carbon—l3 chemical shifts of the carbocationic centers of the unsymmetrical l—aryl—l—phenyl—l—ethyl cations vs. those of the arylcyclopentyl cations shows two reasonably linear portions of differ- ent slope. The linear plots are as expected if the chemical shifts of the carbocationic centers of the three cation types are responding similarly to the electron donating ability of the aryl groups. The results are consistent with the commonly accepted symmetrically twisted structure gll for symmetrically diaryl substituted carbocations. X \ x - 2 X = e donating N Y = Hydrogen + + X = Hydrogen R x CH ,2/ Y = e_ withdrawing William T. Chambers The change in slope for l-aryl-l—phenyl-l-ethyl cations indicates a change in the response. This change could be accommodated by a change in structure from an unsymmetrically twisted ion in which the more coplanar electron donating aryl group was better conjugated with the carbocationic center than the more twisted phenyl (eg.ng) to one in which the more coplanar phenyl was better conjugated than the more twisted electron withdrawing aryl group (eg.av). The different slopes would then reflect the different responses of the carbon chemical shift of the cationic center to the changing electron donating ability of a more planar aryl group as compared with that of a more twisted aryl group. C—l3 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY AS A PROBE TO STUDY SUBSTITUTED ARYL CARBOCATIONS By ,{sI ‘ \«0‘ ' William . Chambers A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1975 TABLE OF CONTENTS Page INTRODUCTION 0 O O O O O O O 0 O O O O O O 0 O 0 O 0 O O O O O O O 0 0 O O O 0 O O O O O O O O O O O O O O O O O O O O l RESULTSOOOOOO 0.00... 000.00.00.00 0.... 0.0.0.0 000000 .0... 00.00.0000 16 DISCUSSIONOOOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO. 23 EXPERmENTAIJoocoooooooooooooooooooooooooooooooo0.0000000000000000 36 Carbocation Precursors..................................... 44 Carbocation Formation...................................... 44 IR and NMR Spectra......................................... 45 Arylcyclopentanols................................... 45 Arylcyclohexanols.................................... 46 Diarylmethyl carbinols............. ..... ............. 47 l-Aryl—l-phenylethanols...... ....... ................. 47 MISCELLANEOUSOOOOOOOO0.00.00.00.000.00.0000...OOOOOOOOOOOOOOOOOOO 49 Determination of the Spin Lattice Relaxation Times of the Heptamethylbenzenonium Ion Using CNMR................ 49 Synthesis of Some Potential Precursors to the Heptamethyl- cyclohexadienyl Anion................................ 52 Experimental............................................... 56 Preparation of fl-Hexamethylbenzene—fl—Heptamethyl- cyclohexadienyl Iron (II) Hexafluorophosphate.. 56 Preparation of Bis-(hexamethylbenzene) iron (II) hexafluorophosphate O O O O O C O O O O C O O O C O O O C O O O O O O O O O 56 Attempted preparation of bis(fl-heptamethylcyclo— hexadieny1)—iron (II) 0 C C O C O O O O O C O O O O O O O O O O O O O O I 5 7 Attempted exchanges of fl-hexamethylbenzene-W-hepta- methylcyclohexadienyl iron II hexafluorophos- phate with sodium cyclopentadienide............ 59 Attempted reduction of fl-heptamethylbenzene-fl-hepta- methylcyclohexadienyl iron (II) hexafluorophos- phateOOOOOOOOOOOOO0.0.0.000...OOOOOOOOOOOOOOOOO 60 Reduction of 1,2,4,5,6,6-hexamethy1-3-methy1ene-l,4- cyclohexadiene with sodium............... ...... 61 ii TABLE OF CONTENTS—-continued Attempted generation of the heptamethylcyclohexa- dienYl anionOOOOOOO...OOOOOOOOOOOOOOOOOOOOOOO REFERENCES ..... OOOOOOOOOOOOOO ..... 00...... ..... 0.0.00.0... ..... iii Page TABLE 0" o \l . \0 LIST OF TABLES l3C chemical shifts for the carbocationic center is substi- tuted l—aryl-l—phenyl-l-ethyl cations...................... 130 chemical shifts of the carbocationic center for the substituted l,l-diaryl—l-ethyl cations..................... 3C chemical shifts of the carbocationic cent—r in substi— tuted arylcyclopentyl cations.............................. l3C chemical shifts of the carbocationic center in substi— tuted arylcyclohexyl cations............................... Typical chemical shifts for the aryl carbons in substituted cyclic aryl cations........................................ Preparation of substituted arylcyclopentanols.............. Preparation of substituted arylcyclohexanols............... Preparation of substituted l—aryl—l—phenylethyl alcohols... Preparation of substituted diarylmethyl carbinols.......... iv Page l7 18 20 21 22 37 39 41 43 LIST OF FIGURES FIGURE Page 1. Graph of H(l) vs. H(3) chemical shifts in 2—aryl-2—nor- bornyl cations..... ......... .............................. 3 2. Graph of H(l) vs. H(3) chemical shifts in 2—aryl—2- bicyclo (2.2.2) octyl cations............................. 5 3. Correlation of PNMR chemical shift of the methyl protons alpha to the carbocationic center in Z—aryl—Z-propyl cations with O +.......................................... 8 .L\ 0 Correlation of PNMR chemical shift of the methyl protons alpha to the carbocationic center in 2—aryl-2-butyl cations with o .......................................... 9 U1 Correlation of PNMR chemical shift of the methyl protons alpha to the carbocationic center in l phenyl—l—ethyl cations with O +.......................................... 10 6. Correlation of 2-aryl-2—propyl cationic center chemical shift with o +............................................ 12 \1 Correlation of 2—aryl—2—butyl cationic center chemical shift with o + .......... .................................. 13 8. Correlation of l—aryl—l—ethyl cationic center chemical shift with O ..... ................ ..... .................. l4 9. Graph of Arylcyclopentyl vs. Arylcyclohexyl cationic center chemical shifts ..... ............................... 24 10. Graph of substituted arylcyclopentyl vs. substituted 2-aryl—2-propyl cationic center chemical shifts........... 25 ll. Graph of substituted aryl cyclopentyl vs. substituted 1,1—diaryl—l—ethyl cationic center chemical shifts........ 26 12. Graph of substituted arylcyclopentyl vs. substituted 6—aryl—6—bicyclo(3.2.l)octyl carbocationic center chemical shifts—. ........................... ................. ...... 27 V LIST OF FIGURES--continued FIGURE 13. Graph of substituted arylcyclopentyl vs. substituted l-aryl—l-ethyl carbocationic center chemical shifts....... l4. Graph of substituted arylcyclopentyl vs. substituted l—aryl-l—phenyl-l-ethyl carbocationic center chemical ShiftSOOOO0.0.CC00......0.0.0....0....OOOOIOOOOOOOOOOOOOO. vi Page 29 31 INTRODUCTION Not long after Gomberg discovered the first free radical, two groupsl’2 independently proposed the existence of carbocationic species when they observed that colorless derivatives of triphenylmethane gave deep yellow solutions in concentrated sulfuric acid and formed orange complexes with metal halides. Although much evidence accumulated to support the existence of carbocations, confirmation was not obtained until Hofmann3 and Gomberg4 independently prepared colored anhydrous perchlorate salts, Ar CC104, from triarylcarbinol and crystalline 3 perchlorate acid. Although interest in carbocations has not abated from their discovery to the present time, it has only been since about 1950 that attempts to obtain crystalline salts of carbocations, other than those of triarylmethylcations, have been undertaken. These efforts were undoubtedly inspired by the theories of reaction mechanism which stemmed from investigations concerning solvolysis reactions. Since the presence of unstable carbocations as intermediates was postulated, it has only been natural to try to confirm their existence by isolation as stable salts or preparation of stable solutions. With the development of more highly acid media, the range of carbocationic species available for study has been extended. The belief that alkyl carbocations were very unstable had to be discarded when Olahs’6 was able to isolate alkyl carbocationic hexafluorantimonates by using the very simple reaction: RF + SBF5 = R+ 3th where R = i—Pr, t-Bu. Although classical methods, such as conductometry, cryoscopy, and UV and IR spectroscopy, are still used to detect and establish the structure of stable carbocations, nuclear magnetic resonance spectros— copy, NMR, has more recently proved itself to be of incalculable value. Indeed, all that is necessary even for the layman to comprehend the impact that NMR has had on detection and establishment of structures of carbocation species, not to mention its impact on the rest of the chemical field, is to scan the chemical literature over the past two decades. A detailed review of the applications of NMR spectroscopy to detect and establish the structure of carbocations is beyond the scope of this text (one more than adequate review was published by Olah and Schleyer7). We will focus on those investigations pertinent to and directly relating to our results. Farnum and Wolf8 used proton nuclear magnetic resonance spectros— copy, PNMR, to study a series of substituted 2-aryl-2-norbornyl cations. A plot of the PNMR chemical shifts of H(l) versus H(3) throughout the series of cations showed marked deviations from linearity for substitu- ents on the aryl group more electron withdrawing than hydrogen (Figure l). The results were consistent with the onset of nonclassical partici— pation taking place in norbornyl cations more electron demanding than the 2—phenyl-l-norbornyl cation. A more nearly linear plot was o 3 Lil .IIL I“ w a i 4. . . -..:pt .c:.. is . TIT: INI’I“ . nit" WT 53$“ til; it]; H3('o) H(3) chemical shifts in Graph of H(l) vs. Figure 1. 2—aryl-2-norbornyl cations, similarly obtained for a series of substituted 2—aryl-2-bicyclo (2.2.1) octyl cations (Figure 2). There was also some unusual effect which was present in the case of the parahalogen substituents which caused these points to deviate from the line in bptthhe arylnorbornyl and the arylbicyclooctyl series. Close scrutiny of the PNMR data led these authors to postulate the presence of some other species, (D), in equi- librium with the parahalogen substituted cation. ———-— Where X = Cl, Br, I A possible structure of (D) is the equilibrating dimer is + <9—————— /” X x —————=,- . + : + \ X X + 1 H3 (1:) H(3) chemical shifts in VS. Graph of H(l) Figure 2. 2-aryl-2-bicyclo (2.2.2) octyl cations. However, attempted detection of this equilibrium in the pa£§_iodophenyl derivative by a dilution study over a 50-fold change in concentration showed that if l’is the species present, it is not in equilibrium with significant amounts of the monomer at the concentrations studied. Deno9 has calculated 0+ values for a number of substituents from four independent reactions and has shown that 0+ values for para— halogens do not show good agreement. Thus the anomolous behavior of para—halogen derivatives for the 2-ary1-2-norbornyl and the 2—arylbi— cyclo[2.2.1]octyl cations is not unique. Halonium ions represent a significant class of onium compounds and their role in electrophilic halogenation reactions is well-estab— lished. Evidence for halonium ion formation includes the result of trans addition (so-called anti addition) of the halogens to a double bond.lo’ll With the development of highly acidic solvents with low nucleo- philicity, such as SbFS/SOZ(SOZCIF), Olah12 has been able to prepare, directly observe and in some instances, even isolate stable chloronium, bromonium and iodonium ions of the type: + R-X-R1 Me, Et, iPr, Norbornyl, Adamantyl where R=Rl x Cl, Br, I and where R = Alkyl, R1 = Aryl Olah's7 data lend support to the idea of Farnum and Wolf8 that halonium ion formation was possibly occurring in their work. Proton NMR data are also available for several series of substi- tuted Arylmethyl alkyl carbocations.l3 A plot of the proton chemical shift of the methyl group alph§_to the carbocationic center versus 0+ (Figures 3 through 5) shows a general correlation. However, several features become apparent on closer scrutiny. First, an appreciable amount of scatter is also observed. Second, as the aryl group becomes more electron withdrawing, the correlation becomes worse. In fact, if one looks at the plot of the secondary substituted l-phenyl-l-ethyl cations, one sees a very large deviation from the linearity of the plot for these electron withdrawing substituents. This deviation from linearity may reflect either differences in the anisotropic effect of the various phenyl substituents and/or a leveling effect on the influ- ence of the aryl substituent on the stability and in turn the chemical shift. Olah7 has obtained PNMR spectra on the isopropyl cation and di- isopropyl chloronium ion. The isopropyl cation has two absorptions, one at o 5.00 (6H,d) and another at 6 13.8 (1H, septet), while the di— isopropyl chloronium ion also has two absorptions, one at 6 2.20 (6H,d) and the other at 6 7.1 (H, septet), where the ratio of alkyl chloride to the acid is 2/1. Thus, the proton attached directly to the carbo- cationic center, which feels the effect of the positive charge the most, is moved upfield by 6.7 ppm. Although the carbon-13 (CNMR) spectrum of the diisopropyl halonium ion is not reported, the CNMR spectrum of the dimethyl chloronium ion is reported and has an absorption at 5 48.6 ppm. This value is much .+.b nufis mcoaumu Haaoumlmlahumlm ca umuamo oacowumoonumo mnu on madam mucuoum ahnuoa msu mo uMHnm Hmuaauno mzzm mo mafiumamuuoo .m wusmam .+b nugmcoUmo .TUHEINITCmNIN sun .3”.ch ofiaoUmoonuwo mnu cu «Seam maououm akauma mnu mo umwsm Hmofiamno mzzm mo coaumawuuoo .q muswwm .0 .1 10 .+ b Lug 95.3.60 Htguolalahcmzmla a.“ umuamo UHCOHumoonuwo 9.3 cu madam mucuoua T2308 .23 mo umdim Hmoflamso ~5an mo cowumamuuoo .m musmflm ‘0 a? 4.6... o. o «6- v.0- Q? «6.. Duo- .26- LEI; . . _ . i l i _ o... i i I; “.0 i IIIIA O.” :3.» 5 CH3“<$: ll closer to that of methylchloride (25.1 ppm) than that expected for the methyl cation (> 300 ppm). Hence, much of the positive charge on the dialkyl halonium ion resides on the halide and not on the carbon to which it is attached. The chemical shift of the dimethyl halonium ion indicates that its carbon is sp3 hybridized. These results clearly demonstrate that carbon-13 chemical shifts are much more sensitive to change in bonding or geometry at the carbocationic center than PNMR chemical shifts. Since more routine carbon-13 instrumentation is now available, we thought that CNMR.was an ideal probe to detect changes in bonding or geometry at the carbocationic center, such as dimeric halonium ion formation. Olah13 has done some preliminary carbon-13 magnetic resonance studies on substituted aryl carbocations. In these studies he plotted the 13C chemical shift of the cationic center, 5130 +, versus Brown's 0+ substituent constants (see Figures 6 through 9). If one looks at these plots, one sees a general overall correlation of 6130 + with 0+. However, there is a considerable amount of deviation from the arbitrary straight line drawn in these plots. It was from this work of Olah's that many of our ideas stemmed. We thought that in order to answer questions about the more intimate structural details of substituted aryl carbocations and in particular the details of the bonding at the carbocationic center, that we needed a series of substituted aryl carbocations as model systems; e.g., the substituted aryl cyclopentyl, which we could use as our standard to plot against all other substi- tuted aryl acyclic, cyclic, and bicyclic carbocations. Our hope was 12 ic O 10!! Correlation of Z-aryl-Z-propyl cat Figure 6. + center chemical shift with 0' l3 +0 ma «Ml -o.1 o 40.3. 40.04 401. -ob -OO' Correlation of Z-aryl-Z-butyl cationic center chemical shift with O'+. Figure 7. l4 4a” -N36 ‘03 eOJL '00‘ -0.‘ ter ionic cen Correlation of l-aryl-l—ethyl cat Figure 8. chemical shift with d'+. 15 that for similar systems a straight line would be obtained by plotting the 613C + of any series of substituted aryl carbocations versus the 6130 + of our model carbocationic system if the electron demands in the systems were comparable. Any deviations from linearity could be inter— preted as a qualitative change in structure of the carbocation. Some factors which.might lead to such a qualitative change in structure in— clude halonium ion formation and changes in geometry around the cationic center. CNMR is an ideal probe to determine whether dimeric halonium ion formation does occur in substituted carbocations as postulated by Farnum and W01f8, since carbon chemical shifts show a large dependence on orbital hybridization as well as electron density. We chose to com- pare substituted aryl cyclohexyl and cyclopentyl carbocationic systems to detect halonium ion formation. Thus, if dimeric halonium ions were formed in these systems, then the different steric constraints in these carbocations should lead to differing amounts of halonium ion formation, and to a deviation from linearity in the plot of the 6130 + of the two systems. Moreover, for the same reasons that it was good for detecting halonium ion formation, CNMR was also thought to be a good probe to detect any geometry change at the carbocationic center by the resultant deviation from linearity of a plot of the 613C + of the system in ques- tion versus the 513C + of an appropriate model system. RESULTS A detailed description of the preparation of the carbocations is given in the experimental section. Fourier transform nuclear magnetic resonance spectroscopy was used to obtain all carbon-13 spectra. The substituted l-aryl-l-phenyl-l-ethyl cation 3(§:5) and the substituted 1,1-diaryl-l-ethyl cations 4(a:d), were generated in FSOBH at —78° from their corresponding carbinols. The spectra for these cations were recorded at ~400. The CNMR chemical shifts of the carbo— cationic centers for these carbocations are listed in Tables 1 and 2 respectively. In addition to the absorptions for the aryl carbons, carbon (or appropriate) aryl substituents, and carbocationic center the carbon-l3 spectra showed a single high field absorption for each of these ions. In the case of the 1,1—diphenyl-1-ethyl cation, a PNMR spectrum was obtained and was found to be nearly identical with that recorded in the literature.14 CH3 l6 17 Table 1. 130 chemical shifts for the carbocationic center is substi— tuted l-aryl-l—phenyl—lvethyl cations. Aryl Group 6C + a) P‘Cflao'céua 209. b) 3,4(CH3)2C6H3 223. c) p—CH3—C6H4 224. d) p—F—C6H4 226. e) p—Cl—C6H4 227. f) p—Br—C6H4 228. g) 06H5 230. h) m—F-C6H4 231. i) m—Cl—CGH4 231. j) p-CF3—C6H4 233. k) 3,5(CF3)2—C6H3 232. * PPM relative to external TMS. 18 Table 2. 130 chemical shifts of the carbocationic center for the substituted 1,1-diary1-l-ethyl cations. Aryl Group 6C + a) p-CH30—C6H4 206. b) p-CH3-C6H4 221. c) 06H4 230. d) p-CF3-C6H4 240 * PPM relative to external TMS. 19 The substituted arylcyclopentyl 5(a-j) and arylcyclohexyl 6(a-h) N N cations were prepared in FSOBH at -78o. The spectra were recorded at -700. The C-13 parameters pertinent for our discussion for the carbo- cations are given in Tables 3 and 4 respectively. Preparation of carbocations 5k and 61 and 6j, required much lower NNA/ temperatures. Those ions were prepared in FSOBH/SOZClF with a trace of SbF5 at -1100. Their spectra were recorded at -90°. Attempts to obtain their spectra at higher temperature only resulted in decomposi- tion of these ions. In addition to the aryl carbons, appropriate aryl substituents, and the carbocationic center absorptions, two high field peaks were observed for cations 5(a-k) and three high field peaks for ions 6(a-j). m N The PNMR spectra obtained for ions 5g and 6g were nearly identical with NN those reported in the literature.11 X X M site 6..-.) The very similar chemical shifts for the aryl carbons in the aryl substituted cyclic carbocations investigated indicate that electron distribution in the aromatic systems are similar. Typical chemical shifts for the aryl cations are given in Table 5. 20 Table 3. 13C chemical shifts of the carbocationic center in substi— tuted arylcyclopentyl cations. Aryl Group 6C + a) p—CH30—C6H4 235. b) 3,4(CH3)2C6H3 257. c) p—CH3—C6H4 259. d) p—F-C6H4 264. e) p-Cl—C6H4 267. f) p-Br-C6H4 267. g) C6H5 270. h) m—F—C6H4 277. i) m—Cl—C6H4 277. j) p-CF3-C6H4 283. k) 3,5(CF3)2C6H4 286. * PPM relative to external TMS. 21 Table 4. 13C chemical shifts of the carbocationic center in substi- tuted arylcyclohexyl cations. Aryl Group 6C + a) p—CH3o—C6H4 227. b) 3,4(CH3)2—C6H3 249. c) p—CH3-C6H4 251. d) p—F—C6H4 256. e) p—Cl—C6H4 259. f) p—Br—C6H4 260. g) C6115 263. h) m—F—C6H4 267. i) p-CF3—C6H4 276. j) 3,5(CF3)2C6H3 279. * PPM relative to external TMS. * Table 5. Typical chemical shifts 22 cyclic aryl cations. for the aryl carbons in substituted Substituent Scpara 6Cortho 6Cmeta 601 p-CH3O 182. 146. 120. 131. p-CH3 173. 143. 134. 135. p-F 180. 148. 121. 134. (JCF=291. Hz) (JCF=15.) (JCF=21.) p-Cl 164. 143. 133. 135. p-Br 156. 142. 136. 136. * PPM relative to external TMS. DISCUSSION Figures 9 through 12 are plots of the carbon-13 chemical shifts of the carbocationic centers of the substituted arylcyclopentyl cations versus the carbon—13 chemical shifts of cationic centers of several series of substituted aryl cations. There are a few features of the graphs that should be noted. First, a good linear correlation is found between the carbocationic center chemical shift of our model system (arylcyclopentyl) and several other series of carbocationic systems. This linear correlation is con- sistent with the interpretation that electron demand is qualitatively fthe same throughout each series. Thus, the 013 chemical shifts of these carbocations reflect the donating ability of the aryl substituent. Second, none of the p-halogen substituents, pF, p—Cl, p-Br (see Figures 9 and 12) show any large deviation from the line. These results imply that either halonium ion formation is not taking place in these systems or that the extent of halonium ion formation is the same in these sys- tems at the temperature and concentration studied. Certainly, the steric constraints in substituted aryl cyclopentyl, cyclohexyl, and bicyclo-[3.2.1110ctyl cations are different. It is our belief that if halonium ion formation occurred in these tertiary cationic systems to any appreciable extent, then much larger deviations from linearity would be observed. 23 24 .mumwnw Hmofiamao umudmo oadoaumo memnoao%oamu< .m> ahunmaoaohoaxu< mo nacho .¢ muawwm usszmmoaosogsm< +omamw Ixxanoqoioixuv +0615? 25 ahmonmtmlamumlw mounufiumnsm .m> Hmuammoaohoamum umuSuflumndm mo gamma .mumwnw Hmofiamno nousmo oficowumo ...... ............................ -666 .III . ll! .Tnl afizmmoéwufifi + 02 .w . Iii", .OH muswfim TAdOHd-Z-TLHV-Z + OCT 5? 26 .mumwsm Hecaaonu Houses uwsowumo HznuMIHIH%HmfiuIH.H emu:UHumnsm .m> amusomoaoao H%uw wouaufiumnsm mo sumac .HH muvwwm AWHZMMOAUMUAHM< +omH (W 'IIHlH-I—TIHVICI-I ‘ I +3“? 27 .mumanm Hmoflamno umuaoo ofisowumuonumo HmuooAH.N.mv oaomown olakumlo wmuduwumnsm .m> ahucmmoaozoamum vmusufiuwnsm mo nmmuw >— i _. .-.. . . U 9 ET. awazmmoaosuqsm< + oms_kv .Namusmsm (T'Z'E) 038 + OCT .7 28 Olahl7 has shown that formation of tertiary dialkyl halonium ions is at least difficult, when he tried to prepare the unsymmetrical dialkyl halonium ions by the procedure: Rlx + mg 36%— R1 x +R 3th He was able to prepare unsymmetrical dialkyl halonium ions by alkylation of primary alkyl fluoroantimonates with primary and secondary alkyl halides. Attempts to alkylate alkyl fluorantimonates with tertiary alkyl halides led only to tertiary alkyl carbocation formation and symmetrical dialkyl halonium ion formation. Perhaps tertiary carbo— cations are too hindered or too stable to allow halonium ion formation to take place or be detected. Perhaps halonium ion formation could be more readily detected in secondary carbocations such as the substituted l—aryl—l—ethyl cations, which are expected to be "hotter" species. A plot of the carbon~l3 carbocationic center chemical shifts vs. those of our own model systems should be linear with the exception of the p-halogen if halonium ion formation is taking place. Olahl3 has done some preliminary work in forming 1-aryl+ethyl cations. However, he has only obtained C-13 spectra for one p-halogen substituent, para-fluoro, in this series. Figure 13 contains the graph of these cations versus our model system. A good linear correlation is observed for all substituents, both elec— tron donating and withdrawing, except for the p-fluoro substituent which deviates from the line by approximately 5 ppm. This preliminary result indicates that halonium ion formation may even be taking place 29 assumlalamumla vmusufiumnsm .muwfinm Hmofiaozo umusmu oadosumoonumo .m> Hauamnoao%oamum umu=Ufiumnnm mo nmmuu 6:. 5. . on. afizmmodwofima + 03 K . ma enema haw can. .2“ imma-t—imvét + 381 f 30 in the p-fluoro substituted l-aryl-l-ethyl cation. These results sug- gest that further experiments in these cation systems would provide theoretically interesting and fruitful results. It is possible that other para-halogen substituents would show larger deviations as a result of dimeric halonium ion formation. Having already established that the substituted arylcyclopentyl cation is a good model for a number of acyclic, cyclic, and bicyclic systems where no changes in bonding are occurring at the carbocationic center, we thought that it would be theoretically interesting to plot the carbon—13 chemical shifts of our model systems against several series of substituted aryl cations where changes in bonding at the carbocationic center were expected to take place. Any change in bonds ing at the cation center should lead to deviation from linearity in the plots. We first chose to look at substituted l-aryl-l-phenyl-l-ethyl cations (see Figure 14). This graph shows a reasonable linear correla- tion for all substituents showing electron donating ability greater than or equal to para-hydrogen. For more electron withdrawing substitu- ents, large deviations from linearity are observed. We believe that these results indicate that the substituted l-aryl-l-phenyl-l-ethyl cations exist in an unsymmetrically twisted propeller conformation, most easily pictured by the plane propeller conformations/Z and 8, where the twisting of the more planar aryl ring is not considered. The top portion of the graph reflects the conformation/Z’in which the more electron donating group is more nearly coplanar with the cationic center than the phenyl. 31 .muMflsm HmUfiamno umusmo usdoaumoonumo ahnumldra%ao£nlalamumla wousuflumnsm .m> ahusmmoaoaoamum wmuDUHquSm mo guano .H- L. .. . II.” . . .. . .. .. .. . L , UNIT. “IL H - n. L. . All. q ... .IU In . s . I- . In-.. I. II..- -. : ._-. . _ . L , U -. n. 2. .... . .. H. m. u. m . a r . ... q . IJI .. . M Quin. .. .w 3H .. , . m , _ --w.-- .I.. m I INT-AI v p-13. - I I III II ... I”- .7-.-” I --. I ” ...-...... .p I . m . . I III .w w .;.h. H.. .H J... M . I...I I ...... . Wm” .. Lump“ t w ”#1 ......u. .uI MI! . . .H i . “r..- Htw ..u. “HmuInW-. .u w. .“ ...I,u.-.t . . ... . . II.‘ .. . hwy .wws.flnm figsfl: exam” Hr . 1.. i _ 1 . 3 HI-.- .. . .. L .. T. ...._ . .. ... I W ..-... ...: _. . . 1. . In... EH ...mMI- . u . N. .3“ . _ . _ AI . .- . .. 21....- . . u .v .. . 3H . m-I . . . .u... . ......h .I. ..... . Pu . . . . q. . . It..I.-ww+ . A ..I.zw:t It- - A, m a... u .. .. n . . .... can. quzmmOAUWDAHM< + . can ..H .tgwst num—t—mmia-t-nsv-t + DST? 32 7 x = e- donating ,/v y = hydrogen /§J x = hydrogen y = e withdrawing The bottom portion of the graph reflects the conformation}! in which the phenyl is more nearly coplanar with the carbocationic center, and the less coplanar ring has an electron withdrawing substituent. Thus we see two reasonably linear portions to our graph in which the carbocationic center chemical shift is dominated by the aryl ring more nearly coplanar with it. The other aryl ring; i.e., the ring less coplanar with the carbocationic center, acts similar to another alkyl substituent. We have shown above that the carbon-l3 chemical shifts of carbocationic centers of monoaryl carbocations correlate with our model systems where electron demand is thought to be qualitatively the same. The top portion of our graph effectively represents a series of substi- tuted monoaryl carbocations with two substituents, a nonplanar phenyl ring and a methyl group. The bottom portion of the graph effectively represents a series of mon0pheny1 carbocations in which the electron donating ability of one of the other substituents, i.e., the less planar aryl ring, is varied. 33 The entire graph can be interpreted to represent a very subtle change in bonding at the carbocationic center; i.e., it represents a simple conformational change in which one ring becomes more nearly coplanar with the carbocationic center. To support our hypothesis we have prepared a few of the substi— tuted diaryl ethyl cations using a wide range of substituents and plotted their C—13 carbocationic center chemical shifts against our model systems (Figure 11). Since one ring is now restricted to be more or less coplanar with the trigonal carbocationic center. Thus, a plot should be linear with our model systems; e.g., the substituted aryl— cyclopentyl. This plot has already been shown in Figure 11 and a good linear correlation is observed. We believe these results in conjunc— tion with those given in Figure 14, are consistent with the interpreta- tion that the substituted l—aryl-l-phenyl-l-ethyl cations do exist in the unsymmetrically twisted "propeller" conformation and exhibit a simple geometry change; i.e., the conformational change discussed above. Botto18 has used our results to help interpret the result he obtained in several series of bicyclic cations which are known to have a propensity for rearrangement. One such system is the substituted 2-aryl-2-bicyclo(2.2.l)heptyl carbocation. He has plotted the carbon— 13 chemical shifts of this carbocationic system vs. that of the sub- stituted 6-aryl-6-bicyclo(3.2.1)octy1 carbocationic system as shown in Figure 15. These results show a good linear correlation for electron donating substituents. A breaking in the plot at para-hydrogen indi- cates that a change in bonding at the carbocationic center is taking 34 .mumasm Hmoflamno GOHumoonumo H%uoo AH.N.mv oao%owmlolaxu thuonHOZINlakuHum>fiHmQ mnuoomm mam: amazooaoz .mHocmuamaoaokoH%um condufiumnsm mo coaumummmum .o magma 38 .Amnav scams: umnum Sufi: .AHH mufl>wuowum>wumo muuommm mom: smasomaoz vosafiucoollo manna 39 woncwunoo AwHIav gas «as «as oquNm III oamaamao sssoIaIa AwHIsV mad was was oasIos III coammao mass AwHIav om~.wmm sm~.om~ smu.sm~ osaIwa III oamnHmNHoIa smooIaaIa AwaIav Naa.saa oHN.NHN OHN.NHN omnlnm III oaonammau «woolaolm AmHIav was «as «as omaIma III oamasmao smooIaIa AwHIav «as omH omH ommlmm III omammao «*qmoolmmolm AwHIav ems AmHIaV ass as H.o as «om sow III ooaa ooumsao masomfimmovs.m sow sou ommIsm III Nomasmao smoo-oases Awaaomv m\a A.oamov o\a as an maaauom m>wum>wuoa muuoomw mmmz Hmaaomaoz .maocmxmnoao%oamum vousuaumnsm mo soaumumaoum .n manna i ; 40 .Hmnum\mamxon fiufiB wmusao aHH kufl>fluoflum>flumn IIIIIIIIIImHMMMNWlmmmeIIIIIIIIIIIII Hmaaomaoz U033 UEOUIN MHn—MH 41 coacwuaoo awsIMV was AmsI v www ass oaaINa III oaasm«so «aQUIaIa AmsImv oms sms v MM. was osmIaa III o«sa«so ammo Naa.«a~ Na~.«m~ III as Mamsam osomsm«so «asoIsoIa AmsIMV was I . m Ams v WWW osm III as Mamss ommsm«so «sooImIa AmsImV «as amsI v WNW ass III as Mamssm casmmso «:aoIamoIa AmsImv was 37 v WWW was I as Hows: omsmoso amoumammovim osN osN omaIi III o«smmso ..«mooIommoIa chsomv m\a A.uamov m\a QB an masahom o>flum>wuom muuommm mmmz “masooaoz .maofiooam Husumahconmlalamumla wmusuflumnsm mo coaumummoum .w manna 42 .Gflwmao * AasIav msa smsIaV ssm AmsIav asm «mm «mm o«oINs III assassoso assessmaoom.a smsIav m«m AmsIav saw as s.o as com com II ooss ommmsmmso «mooImaoIs as s .0 us NmN.«ma Nm~.«ma III cams osoasa«so «msUIsOIa Awasomv m\a A.0Hmov m\a me an masahom o>Hum>HHmm msuommm mmmz. HmHSUmHoz wmscfluaoollw manna 43 .Aaumv umauo\mcwxM£ aufl3.vmusao was .AHH huH>HuoHum>Humn .mHoaHnumo Hanumaahumflw wouSuflumnsm mo Gowumnmmmum .a mHQMH 44 ~Carbocation Precursors The alcohols were prepared by reacting 1.1 moles of the appro— priate Gringnard reagent with 1 mole of the corresponding ketone; cyclopentanone, cyclohexanone, and substituted aryl methylketones or by reacting 2.2 equivalent of the Gringnard reagent with 1 mole of ethyl acetate. In all cases a 20% molar excess, with respect to the carbonyl compound, of magnesium was used. Yields of the appropriate alcohol range from 45% to 90% based on the ketone. The important physical constants for the compounds are summarized in Tables 6 through 9. Carbocation Formation In order to facilitate complete ionization, one of the following methods was chosen to form the carbocations: 1) Approximately 8x10-4 moles of precursor was placed in a jacketed dr0pping funnel at 0 to -200 in the apparatus described by Hart.19 Enough Freon II was added to dissolve the precursor. The resulting solution was added over approximately a twenty minute period to 1.57 ml of rapidly stirred FSOBH at -780 under a N2 blanket. The resulting colored, nonhomogenous solution was allowed to stir for fifteen minutes at -780 and was then blown with N2 into a CNMR tube cooled at —78°. The Freon II was allowed to separate and was removed with a micropipette. 2) The precursor was added in small amounts to 1.57 of the acidic media, FSO H or FSO3H/SOZClF (1/3) with a trace of SbF at the 3 5 45 appropriate temperature (-780 to -1100). A homogenous solution was obtained by intermitant vibra stirring and cooling. IR and NMR Spectra Arylcyclopentanols CH3O-C6H4 (olefin): nmr (CC14) 56.85 (4H, AA'BB'Zh)= 30 Hz, J = 9 Hz) 5.83 (1H, t, J = 2 Hz), 3.67 (3H,s) 2.80-1.77 (6H, m); ir (nujol) p 6.15, 9.67. 3,4(CH c nmr (0014) 67.17—6.63 (3H, m), 2.3 (1H, s) 2.13 3’2 6H3: (6H, br. s.), 1.77 (8H, br. s.); ir (neat)u 2.93, 6.05, 9.91. p-CH nmr (CC14) 66.90 (4H, AA'BB' Av = 17 Hz, J = 7 Hz), 3-C6H4: 2.12 (3H,s), 1.77 (9H, br. 3.); ir (neat)1I2.90, 6.20, 9,35. p-F-C6H4: nmr (C014) 67.20 (2H, distorted quartet, J = 5 Hz, J = 8 Hz), 6.75 (2H, distorted triplet, J1 = J = 8 Hz), 2.73 (1H, br. s.), 1.80 (8H, br. s.); ir (neat)112.95, 6.18, 9.95. p-Cl—C6H4: nmr (0014) 57.20 (4H, AA'BB' Av 2’0, J = 10 Hz), 1.90 (9H, br. s.), ir (nujol) u 3.01, 6.21, 9.88. “J nmr (CCl4) 57.17 (4H, AA'BB' Av'= O, J p-Br-C 10 Hz), 1.83 6H4: (8H, br. s.), 1.43 (1H, s); ir (neat)u 2.95, 6.08, 9.90. C6H4: nmr (CC14) 57.4-6.90 (5H, m), 2.15 (1H, s), 1.33 (8H br. 8.); ir (neat)u 2.92, 6.20, 9.90. m-Cl-C6H4: nmr (CC14) 67.73-6.90 (4H, m), 2.40 (1H, br. s.), 1.80 (8H, br. s.); ir (neat)¢12.95, 6.23, 9.92. m-F-C6H4: nmr (CC14) 67.20-6.49 (4H, m), 1.80 (8H, br. 3.), 1.53 (1H, br. 5.); Lr(neat)u 2.99, 6.18, 6.28, 9.89. 46 p-CF3-C6H4: nmr (CC14) 67.48 (4H, s), 1.93 (9H, br. 3.); ir (nujol) H2.95, 6.17, 9.82. 3,5(CF3)206H3: nmr (0014) 67.80 (2H, br. 3.), 7.63 (1H, br. 3.), 1.97 (1H, s), 1.73 (1H, 8); ir (nujol)112.98, 6.10, 8.80. Arylcyclohexanols p-CH30—C6H4: nmr (0014) 66.83 (4H, AA'BB' Av = 34 Hz, J = 8) 3.70 (3H, s), 1.66 (10H, br. 3.), 1.20 (1H, s); ir (nujol)112.90, 6.19, 9.60. 3,4(CH3)2C6H3: nmr (CC14) 66.87 (3H, m), 3.33 (1H, s), 2.23 (6H, br. 3.), 1.67 (10H, br. 3.); ir (neat)q12.87, 6.60, 9.60. p-F-C6H4: nmr (CC14) 67.23 (2H, distorted quartet, J = 8 Hz, J = 5 Hz), 6.77 (2H, distorted triplet, J = 8 Hz), 1.63 (11H, br. 3.); ir (nujol)Lr2.99, 6.19, 9.60. C6H5: nmr (CC;4) 67.37-6.93 (5H, m), 1.70 (10H, br. 3.), 1.37 (1H, 3.); ir (nujol)}l3.05, 6.22, 9.68. p-CH ' rmu:(CCl4) 67.0 (4H, AA'BB', Av = 16 Hz, J = 8 Hz), 2.25 3-C6H4. (3H, 3.), 1.63 (10H, br. 3.), 1.45 (1H, s); ir (nujol)L12.95, 6.60, 9.60. m-F—C6H5: nmr (CC14) 67.2—6.53 (4H, m), 1.67 (11H, br. 3.); ir (nujol) p 3.00, 6.19, 6.29, 9.59. p-Cl—C6H5: nmr (0014) 67.13 (4H, AA'BB;, 00'2’0, J = 8 Hz), 1.67 (10H, br. 3.), 1.43 (1H, br. 3.); ir (neat)113.00, 6.25, 9.89. CF3—C6H4: nmr (CC14) 67.43 (4H, 3.), 2.67 (1H, 3.), 1.63 (10H, br. 3.); ir (neat)112.90, 6.13, 8.85. 47 3,5(CF3)2-C6H3: nmr (CC14) 67.73 (2H, br. 3.), 7.57 (1H, br. 3.), 1.67 (10H, br. 3.), 1.60 (1H, 3.); ir (nujol)¢13.00, 6.15, 8.75. p-Br-C nmr (0014) 67.20 (4H, AA'BB', Av”; 0, J = 8 Hz), 1.67 6H4: (11H, br. 3.); ir (nujol)142.92, 6.23, 9.91. Diarylmethyl carbinols p-CH30-C6H4: (6H, 3.), 2.67 (1H, br. 3.), 1.70 (3H, 3.); ir (neat)112.95, 6.19, nmr (C014) 66.77 (8H, AA'BB;, Av - 32 Hz, J = 9 Hz), 2.53 9.70. p-CH3-C6H4: nmr (C014) 66.90 (8H, AA'BB',ZM)= 14 Hz, J = 8 Hz), 2.23 CIH, br. 3.), 1.70 (3H, 3.); ir (neat)L12.90, 6.19, 9.15. p—CF3-06H4: nmr (CC14) 67.40 (8H, 3.), 2.73 (1H, 3.), 1.87 (3H, 3.); ir (neat)112.88, 6.10, 9.80. l-Aryl—lvphenylethanols 3,4(CH3)2C6H4: (1H, br. 3.), 1.73 (3H, 3.); ir (neat)112.85, 6.18, 9.68. nmr (CC14) 67.27-6.73 (8H, m), 2.10 (6H, br. 3.), 2.0 p-CH nmr (CC14) 66.67-7.27 (9H, m), 2.23 (3H, 3.), 2.07 3—C6H4: (1H, br. 3.), 1.73 (3H, 3.); ir (neat)L12.90, 6.20, 9.35. p-F—C6H4: nmr (CC14) 67.30-6.50 (9H, m), 2.67 (1H, 3.), 1.73 (3H, 3.); ir (neat)112.91, 6.21, 9.35. p-Cl-C6H4: nmr (CCl4) 67.00 (9H, distorted br. 3.), 2.67 (1H, 3.), 1.70 (3H, 3.); ir (neat)112.89, 6.19, 9.12. C6H5: nmr (CClA) 67.33-6.87 (10H, m), 2.53 (1H, br. 3.), 1.73 (3H, 3.); ir (nujol)112.95, 6.25, 9.40. 48 m—F—C6H4: nmr (C014) 67.26-6.50 (9H, m), 2.00 (1H, 3.), 1.80 (3H, 3.); 1r (neat)112 99, 6.12, 9.45. m-Cl-C6H4: nmr (CClA) 67.30—6.80 (9H, basically 3 broad peaks with maxima 7.22, 7.07 and 6.97), 2.67 (1H, br. 3.), 1.70 (3H, 3.); ir (nujol)1£2.92, 6.22, 6.35, 9.35. p-CF3C6H4: nmr (CC14) 67.33 (5H, 3.); 7.12 (4H, AA'BB' Av = 0, J = 10 Hz), 2.80 (1H, 3.), 1.90 (3H, 3.); ir (neat)112.90, 6.15, 8.90 (br.). 3,5(CF3)2C6H3: nmr (CC14) 67.67 (2H, br. 3.), 7.60 (1H, br. 3.), 7.20 (5H, br. 3.), 2.03 (1H, 3.), 1.93 (3H, 3.); ir (nujol) u 2.78, 6.10, 8.80. MISCELLANEOUS Determination of the Spin Lattice Relaxation Times of the Heptamethylbenzenonium ion using CNMR The heptamethylbenzenonium ion has long been a system of interest. Proton NMR has shown it to undergo a very rapid 1,6—methyl migration at +700.20 / ~l,6 Me ‘— x \ \ The availability of more routine instrumentation has made carbon“ 13 nuclear magnetic resonance spectroscopy, CNMR, another important 49 50 tool for structure elucidation. Although the literature involving CNMR is rapidly expanding, very few spin lattice relaxation, (T 1), studies have been undertaken. We have obtained the CNMR spectrum and determined the spin lattice relaxation time for the carbons of the heptamethylbenzenonium ion at probe temperature. We have undertaken this preliminary study for two reasons; first, to familiarize ourselves with the technique for determining spin lattice relaxation time; and second, to see if the spin lattice relaxation time can give an indica- tion as to whether a species might be either rapidly equilibrating or a bridged species. The solution of the heptamethylbenzenonium ion was prepared by gradual addition, with intermitant vibra-stirring and cooling, of 200 mg of 4 methylene-l,l,2,3,5,6-hexamethyl-2,S—cyclohexadiene to approxi- mately 1 ml of a mixture of 4.8 parts CF3COZH and 5.2 parts H2804 at -30°. The solution was then diluted to 1.57 ml with the acid mixture. The resulting solution gave proton spectra for the heptamethylbenzenon- ium ion nearly identical with those reported in the literature.24 The CNMR spectra was run on a Varian CFT-20 with lock being obtained on an external capillary of d6-DMSO. The resulting probe temperature spectrum showed absorption at -37.50, -31.04, +20.94, +113.52, +136.22, +138.59, +115.06 relative to the carbonyl carbon in CFBCOZH' These chemical shifts are assigned C1, 03’ C2’ C4, C5, and C7, C6’ and 08 respectively as in 2} It should be noted that there should be four high field absorptions for the methyl carbons; however, there are only three. We believe that if our assignments for C1 and C3 51 are correct, then the methyl carbons most likely to be coincidentally equivalent are those attached to C1 and C3; i.e., 05 and C7. The spin lattice relaxation times were obtained using the inver— sion recovery method with a pulse delay of 70 seconds and are recorded below: Carbon Tl Cl 11.8 sec. C2 9.60 sec. C3 9.88 sec. C4 14.0 sec. C5,7 1.7 sec. C6 2.0 sec. C8 2.7 sec. Finding appropriate models for relaxation times for these carbons is difficult since few 130 relaxation studies have been done on uncharged species, and, to our knowledge, none have been done on carbo- cations. Levy3 has reported the T for the quarternary ring carbon of l toluene (Cl) to be 51 sec. for an undegassed sample. The T1 values obtained for the ring carbons of the heptamethylbenzenonium ion, “’10 sec., indicate something unusual about these carbons. One factor which could possibly influence the rate of relaxation is the fluctuating positive charge in the molecule. Further experi— ments are needed. 52 Synthesis of Some Potential Precursors to the Heptamethylcyclohexadienyl Anion The heptamethylbenzenonium ionlg’has been shown to undergo a rapid methyl migration at 70°. The heptamethylcyclohexadienyl anion l9) unlike the heptamethylbenzenonium ion, is not expected to undergo the rapid rearrangement thermally; however, it can be predicted that it would undergo the rearrangement photochemically. 10 N For reasons of comparison with the heptamethylbenzenonium ion, we have synthesized some potential precursors to the anion and have attempted a few experiments in order to generate the anion itself. Our first approach was patterned after that of two groups of 23,24 workers: 1) Braitsch and Helling and 2) Nesmayanovzs. Braitsch and Helling23’24 have reported the synthesis of N-hexa- methylbenzene—fl—exo-t—butylhexamethyl cyclohexadienyl iron (II) hexafluorophosphate,‘ll. We have prepared and isolated the methyl analogue N—hexamethyl— benzene-fi-heptamethylcyclohexadienyl iron (II) hexafluorophosphate, l2; by reacting bis (hexamethylbenzene) iron (II) hexafluorophosphate dihydrate, £3, with three equivalents of methyl lithium in ether. 53 The PNMR spectrum showed absorption at't7.48 (3H, s.), 7.70 (18H, 8.), 8.17 (I6H, 8.), 8.73 (3H, s.), 8.80 (6H, s.), and 10.16 (3H, s.). This spectrum compares very favorably with that of the exo— t-butyl compound 11 made by Helling and Braitschzz’23 where resonances at T 7.63 (3H, s.), 7.79 (18H, 3.), 8.20 (6H, s.), 8.32 (3H, s.), 8.73 (6H, s.), 9.73 (9H, s.) are reported. The only resonance which has any appreciable chemical shift difference is the exo-t-butyl and the exo- methyl, in compound 11 and 12 respectively. , R+ / a ’l\l/ R = t-Bu Fe X 13 R.= Me $39 ... .1 If one compares these resonances with those of fl-mesitylene-w-exo— alkyl-1,3,5-trimethyl cyclohexadienyl iron II hexafluorophosphates 14’ and lé_where the alkyl substituents are t-butyl and methyl one finds resonances of T 9.53 and 9.84 respectively for the exo substituents. If this 9.31 ppm chemical shift difference were to hold in the hexa- fluorophosphate salts 11 and 12, then the chemical shift of the exo methyl in 12 should come at 10.04. The value of 10.16 is in reasonable agreement. We had then hoped to remove the heptamethylcyclohexadienyl anion from the metal by one of two methods: 1) Exchanging it with sodium cyclopentadienidezs. This method of exchange was known to work for complexes of the type 13, to give the free arene and ferrocene. -—p + (CL); d. 14 N L Our hope was not only that the arene would be exchanged off, but also, the cyclohexadienyl anion. We have found that reaction of lg’with sodium cyclopentadienide gave no ferrocene type products; i.e., neither the arene ring nor the heptamethyl cyclohexadienyl ring is exchanged off the metal. 2) Reducing;the anion off the metal. It was known that complexes of the type lflrwere converted into ferrocene and the free arene by the action of reducing agents as Na/Hg amalgam. We hoped that the complex _13 could be converted either to the bis-(fi—heptamethylcyclohexadienyl)— iron (11) from which we could attempt to remove the metal by some similar process or the free heptamethylcyclohexadienyl anion and/or arene would be liberated from the metal. Attempted reduction of Tailed only to recovery of starting material an isolation of small amounts of hexamethylbenzene. 55 Our second approach involved the synthesis of l,2,3,3,4,5,6- heptamethyl-1,4-cyclohexadiene by reduction of 1,1,2,3,5,6-hexamethy1—4- methlene-Z.5-cyclohexadiene with Na/NH3(1). The PNMR of the product shows resonance at 6 2.17 (1H, q, J = 6H2), 1.58 (12H, 3.), 1.0 (3H, s.), and 9.95 (3H, d, J = 6H2). Irradiation of the signal at 2.17 collapsed the signal at 0.95 to a singlet. This signal at 2.17 is unusually high for a doubly allylic methine hydrogen. However, the doubly allylic protons in 1,4-cyclohexadiene and l,4—dimethy1-l,4-cyclohexadiene came at 2.63 and 2.45 respectively. If this shift to higher field by about 0.2 ppm per set of methyl groups located 1,4 on the diene system were to continue, then the value of 2.17 is reasonable. The mass spectrum with a parent peak of 178 and fragments at 163, 148, and 133, is consistent with the assigned structure. This does not, however, exclude the conjugated isomer, 1,2,3,4,5,5,6-heptamethy1-1,3-cyclo- hexadiene, which might have a similar PNMR spectrum if some of the signals were accidently coincidental. The UV spectrum and the CNMR spectra ruled out the conjugated diene. No UV maximum is observed above 200 nm. The CNMR shows only nine absorptions as expected for the nonconjugated diene while the conjugated diene should show thirteen. Bates26 has described a procedure for generating anions from 1,4—dienes. Attempts at generating the anion by this method are described in the experimental. 56 Experimental Preparation of flfiflexamethylbenzene-fl—Heptamethyl- cyclohexadienyl Iron (II) Hexafluorophosphate In a flame dried three necked round bottom flask equipped with stopper, nitrogen inlet, a serum cap and magnetic stirrer, was placed 1.0 gm of bis-(hexamethylbenzene) iron (II) hexafluorophosphate under a blanket of N2. Ether (3.0 ml), dried over sodium was added using a syringe. The resulting suspension was stirred and cooled to -78°. Methyl lithium, 3.2 ml (1.24 M), was added to the reaction flask all at once. The reaction mixture was allowed to stir for one minute and was allowed to warm slowly to room temperature. As the vessel warmed the reaction mixture turned a dark purple, almost black, color. The reaction flask was allowed to stir for two hours at room temperature. Ether (10 ml), saturated with H 0 was added to react with the excess 2 methyl lithium. The resulting mixture was filtered through a fritted funnel. The solid was placed in a beaker and extracted three times with metnylene chloride (10 ml). The resulting red solution was gravity filtered. Ether was added to precipitate the light brown product (0.59 gm, 51%) NMR (CDC13): T 7.40 (3H, s.), 7.70 (18 H, s.), 8.17 (6H, s.), 8.73 (3H, s.), 8.80 (6H, 3.), 10.16 (9H, s.); ir (KBr): 2950, 1437, 1005, 845 cm-1; mp (sealed evacuated capillary): gradual decomposition as temperature raised above 1000. Preparation of Bis-(hexamethylbenzene) iron (11) hexafluorophosphate A flame dried three necked round bottom flask, equipped with stoppers, magnetic stirrer, and a condenser with a nitrogen inlet on 57 the top, was purged with nitrogen. Hexamethylbenzene (6.85 gm), ferrous chloride (2.15 gm), and cyclohexane (2.5 ml) were placed in the flask and stirred. Aluminum chloride (5.40 gm, freshly sublimed) was added to the stirred suspension. The suspension was refluxed for twenty-four hours under a blanket of nitrogen. The reaction mixture was allowed to cool to room temperature and then cooled to 00 in an ice bath. Hydrolysis was accomplished by adding 30 m1 of iced water. The organic layer was discarded. The aqueous layer was extracted with two 25 m1 portions of petroleum ether and filtered. To the aqueous solution was added an aqueous solution of 13.88 gm of sodium hexafluoro- phosphate to precipitate the product which was collected by filtration. The crude product was dissolved in 95% aqueous acetone and precipitated by adding ether. Yield: 7.68 (64%) light orange solid; mp: gradual decomposition on heating; NMR (d6-acetone):'r7.18 (3.6H, s.), 7.5 (36H, 3.); ir (KBr) 3450, 3000, 1470, 1445, 1300, 1010, 850 cm’l. Attempted preparation of bis(fi-heptamethylcyclo— hexadienyl)-iron (II) The procedure used here was basically the same as that used for the preparation of fi—hexamethylbenzene-fi-heptamethylcyclohexadienyl iron(II)hexaf1uorophosphate. All experiments started with 1.0 gm of bis(hexamethylbenzene) iron (II) hexafluorophosphate. The following modifications were made: 1) Five equivalents of methyl lithium were added at «780 and then allowed to warm to room temperature and kept there for two hours with constant stirring. Work—up followed. 58 2) Five equivalents of methyl lithium were added at -780 and then the reaction mixture was refluxed overnight. Work—up followed. 3) Reaction was run in di-n-butyl ether as a solvent. The reac- tion mixture was cooled to ~00 before adding the methyl lithium (eight equivalents). It was allowed to warm to room temperature. The ether was distilled out. The reaction mixture was refluxed overnight. Work-up followed. The work—up procedure was also a little different from that used in the fl—hexamethylbenzene-fl-heptamethylcyclohexadienyl iron (11) hexa— fluorophosphate preparation. It was as follows: After the reaction with methyl lithium was thought to be complete, the reaction mixture was cooled to 0°. Fifty ml of ether saturated with water followed by 10 ml of water were added to quench any unreacted methyl lithium. The reaction mixture was filtered. The solid was collected and extracted with 2 x 25 ml of Et20 and then saved for recrystallization. The ether was combined with the mother liquor from the reaction. The combined two phase water-ether mixture was put in a separatory funnel and the two layers were separated. The ether solution was again extracted with 10 ml of H 0. The ether layer, which should contain any bis(N-hepta— 2 methylcyclohexadienyl) iron (II) was dried over MgSO filtered, and 4: rotovap distilled. None of the desired product was obtained in any of the cases mentioned above. The solid obtained from these reactions was recrystallized by dissolving it in CH2C12 and reprecipitating it with ether. The solid had an NMR identical to that of fl—hexamethylbenzene—fl—heptamethylcyclo- hexadienyl iron.(11) hexafluorophosphate. The yield of 59 fl—hexamethylbenzene—fl—heptamethylcyclohexadienyl iron (11) hexafluoro— phosphate was 41-68% based on the bis arene iron complex. Attempted exchanges of fl-hexamethylbenzene—fl-hepta— methylpyclohexadiepyl iron II hexafluorophosphate with sodium cyclopentadienide In a one necked flame dried flask with a Teflon stopcock in a side arm, equipped with a magnetic stirrer, and condenser, was placed 0.51 gm fl—hexamethylbenzene—W—heptamethylcyclohexadienyl iron (II) hexafluorophosphate (0.51 gm) and 3 ml of THF (dried by distilling from LAH) under argon. The reaction mixture was also purged of oxygen by bubbling the argon through the solution, and then kept under a blanket of argon. The side arm of the flask was equipped with a serum cap and the reaction was cooled to —78°. Sodium cyclopentadienide (2 ml of a 2.5 M solution in THF) was added all at once. The reaction mixture was stirred at -780 for five minutes and allowed to warm to room tempera- ture where it was then stirred for twenty hours before 2 ml of H20 were added to quench the reaction mixture at 00. Most of the THF was roto- vap distilled, leaving a wet brown solid. Ten ml of H20 were added and then the mixture was extracted with petroleum ether to attempt to remove any ferrocene type products. The brown solid was filtered from the mother liquor and saved for recrystallization. The two phase mother liquor was separated in a separatory funnel and the petroleum ether layer was dried and rotovap distilled. No observable products were found. The solid and aqueous layer was then extracted with methylene chloride giving a red brown solution. Et20 was added to precipitate a 60 a light brown solid (0.38 gm). The NMR of the solid was identical with that of the starting material. Other attempts to exchange were made using 2 to 15 equivalents of sodium cyclopentadienide. Temperatures ranged from room temperature to refluxing THF for times up to twelve hours. In all cases neither ferrocene type products nor any hydrocarbon type products, with the exception of small amounts of hexamethylbenzene were observed. In all cases, recovery of starting material was observed (70-95%). Attempted reduction of fl-heppamethylbenzene-fl-hepta— methylcyclohexadienyl iron (11) hexafluorophosphate fl-hexamethylbenzene-fi-heptamethylcyclohexadienyl iron (11) hexa- fluorOphosphate (0.805 gm) was placed in a dried flask containing approximately 80 ml of dried THF, degassed with argon. In a second flame dried flask was placed 1 ml of mercury. Sodium (0.15 gm) was added to the mercury to make Na/Hg amalgam under an argon blanket. The solution of the iron complex was added to the vessel containing the amalgam. The reaction mixture turned from its initial red color to a dark brown. The reaction was allowed to stir for three hours at room temperature. If an aliquot was removed and exposed to the air, the solution immediately turned back to the red color. The reaction mix- ture was then either allowed to stir for three more hours at room temperature or sodium cyclopentadienide (4 ml of 2.5 M in THF) was added and then allowed to stir for three hours more. The reaction was then quenched with water. The starting material was precipitated by adding ether and isolated by filtration (0.6 gm and 0.49 gm recovered 61 respectively). The mother liquor was evaporated at room temperature under vacuum. The residue was chromatographed on silica gel with hexane as the eluant. The only insoluble materials were mineral oil (250 mg) when the sodium cyclopentadienide was added and hexamethyl- benzene (0’40 mg). Reduction of 1,2,4,5,6,6-hexamethyl-3—methylene- 1,4—cyclohexadiene with sodium The entire system consisting of a 100 ml, three necked round bottom flask with N2 inlet, Dry Ice condenser with a soda lime drying tube, ammonia inlet, and magnetic stirrer was flamed while being flushed with N2. l,2,4,5,6,6 hexamethyl-3-methylene-l,4-cyclohexadiene (0.176 gm) was then placed in the flask and approximately 10 ml of NH3 was condensed into the apparatus. Half of the sodium (0.046 gm) was added and a blue solution resulted. The rest of the sodium was added and 20 ml more of NH3 were condensed into the apparatus. When the solu- tion lost its blue color, 1 gm of ammonium chloride was added slowly. Water ( 10 ml) was added and the NH3 was allowed to evaporate. Et20 (20 ml) was added and the two layers were separated in a separatory funnel. The ether layer was dried over MgSO4, filtered, and rotovap distilled. A light yellow liquid (9.150 gm) was obtained. Chromatog- raphy on a 6 ft., 20% SE—30 column at 1500 showed two peaks (ret. ratio 5/3) of an unknown compound and starting material. T.L.C. on 20% AgNO impregnated silica with hexane ether, (lO/l eluent showed two f1 = 0.55). The second spot had an Rf identical to 3 spots (Rf = 0.87, R that of the starting material. 62 Chromatography on 20% silver nitrate impregnated silica gel (70-135 mesh) lead to the isolation of 102 mg of compound whose spectra were consistent with the structure l,2,3,3,4,5,6 heptamethyl—1,4-cyclo- hexadiene PNMR (CC14): 2.17 (1H, q, J = 6Hz), 1.57 (12H, 5.), 1.0 (3H, s.), 0.95 (1H, d, J = 6H2); irradiation of the signal at 2.17 collapsed the signal at 0.95 to a singlet; no UV max above 200 nm; CNMR (CDC13) 14.34, 18.06, 19.96, 24.77, 39.99, 43.31, 75.45, 128.46, 131.49; mass spectra (m/e) 178, 153, 148, 133 and 0.23 gm of starting material. The l,2,3,3,4,5,6 heptamethyl—l,4-cyclohexadiene must be stored in a freezer (below -30°) under an inert atmosphere. Attempts at obtaining a satisfactory microanalysis were unsuccessful. Attempted_generation of the heptamethylcyclo— hexadienyl anion A) In a flame dried NMR tube with a serum cap was placed 1,2,3,3,4,5,6-heptamethylcyclohexadiene (0.356 gm) and 0.6 ml of THF. The solution was degassed by bubbling through argon. The mixture was cooled to -780 and 1.2 m1 of a 1.67 M solution of n—BuLi in hexane was added. The mixture was allowed to warm to room temperature. A faint yellow color appeared but no layer separation occurred as expected. Quenching with water at 00 followed by extraction with hexane led to the recovery of 0.327 gm of starting material after chromatography on 20% AgNO impregnated silica gel with hexane ether (10/1) eluent. 3 B) The reaction was run identically as in A only with 2.4 m1 of n—BuLi in hexane. Again no layer separation occurred. Quenching with D20, followed by extraction with hexane led to the recovery of 0.313 gm 63 of starting material. Mass spectra of the crude recovered starting material showed no deuterium incorporation. C) The reaction was run as in either A or B with one or two equivalents of a BuLi in hexane with one and two equivalents, relative to the butyl lithium, of either dry tetramethylethylene diamine or dry HMPA. The reaction was quenched with water (either H 0 or D20), at 00. 2 In those experiments where tetramethylethylene diamine was used, the quenched reaction mixture was diluted with water (3 ml) (H20 or D20) and then extracted with three 5 m1 portions of hexane. The hexane solution was then extracted with two 5 m1 portions of 1% HCl, and then with one 5 ml portion of H20. The hexane solution was dried over MgSO filtered and rotovap distilled. The crude recovered starting material 4: ("“100 mg), showed no deuterium incorporation. In those cases where HMPA was used, a fleeting red color was observed on addition of the butyl lithium. After quenching with water at 0°, the reaction mixture was diluted with 3 ml of water and then the mixture was extracted with five 2 ml portions of hexane. The hexane extracts were dried with MgSO filtered and rotovap distilled. 4: Attempted chromatography on silica gel or AgNO impregnated silica gel 3 led to no insoluble product. As material was eluted down the column with hexane/ether (lO/l), the column turned every color of the rainbow. 10. 11. 12. REFERENCES a) J. F. Norris, Am. Chem. J".Z§’ 117 (1901); Chem. Zentr. 1901 (I) 699. b) J. F. Norris and W. W. Saunders, Am. Chem. J., 25, 54 (1901); through Chem. Zentr. 1901 (I) 463. F. Tehrmann and F. Wentzel, Chem. Ber., 33, 3815 (1901). M. Gomberg and L. H. Cone, Ann. Chem., 370, 142, 193 (1909); 376, 183 (1910). K. A. Hofmann, and H. Kirmreuther, Chem. Ber., 42, 4856 (1909). G. A. Olah g£_§1,, J. Am. Chem. Soc., 86, 1360 (1964). G. A. Olah ggfl§1., J. Am. Chem. Soc., 86, 2198 (1964). "Carbonium Ions", Vol. I—IV, G. A. Olah and P. V. R. Schleyer, Eds. John Wiley and Sons, New York, London, and Sidney (Especially Chapters 7 and 28). D. G. Farnum and A. D. Wolf, J. Am. Chem. Soc., 96, 5166 (1974). N. G. Deno and W. L. Evans, J. Am. Chem. Soc., 12, 5804 (1951). McKenzie, J. Chem. Soc., 101, 1196 (1912). a) W. G. Young 25 ii}, J. Am. Chem. Soc., 51, 2528 (1929). b) S. Winstein and H. J. Lucas, J. Am. Chem. SOC",§}’ 1576 (1939). c) H. 0. House and R. S. Ro, J. Am. Chem. Soc., 89, 182 (1958). d) J. H. Rolston and K. Yates, ibid., 21, 1496, 1477, 1483 (1969). a) G. A. Olah and J. M. Bollinger, J. Am. Chem. Soc., 947 (1969). ,9. b) C. A. Olah and Yoke K. Mo, ibid., 26, 3560 (1972). c) G. A. 01ah.g£_§1,, ibid., 96, 3565 (1972) d) G. A. Olah g£_§l,, ibid., 680 (1975) and reference cited in the above. 64 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 65 a) G. A. Olah g£_§l,, ibid., 24, 2044 (1972). b) G. A. Olah §£H§1.,.ipid.,‘2§, 464 (1971). a) D. C. Farnum, ipgg,,‘§9, 2970 (1967). b) D. C. Farnum and C. F. Wilcox, ibid., 5379 (1967). D. C. Farnum and Goverdhan Mehta, 1219': 91, 3256 (1969). G. A. Olah and Gao Lang, 121$}, 26, 195 (1974). G. A. Olah g£_§1,, 32293,,23: 2112 (1970), 94, 156 (1972). R. Botto, Ph.D. Thesis, Michigan State University, 1975. Harold Hart and Masayuki Kuzuya, J. Amer. Chem. Soc., 26, 6436 (1974). M. Saunders, Magnetic Resonance in Biological Systems, A. Ehrenberg, B. G. Malmstrom, and T. Vanngerd (editors), 90 (1967). Harold Hart g£_al,, J. Am. Chem. Soc., 88, 1005 (1966). C. C. Levy and I. R. Prat, J. Magnetic Resonance, 18, 500—521 (1975). ' J. F. Helling and D. M. Braitsch, J. Amer. Chem. Soc., 92, 7209 (1970). N D. M. Braitsch, Ph.D. Thesis, University of Florida, 1971. A. N. Mesmeyanov g£_§1,, Akademi Nauk, SSR Vol. 190 No. 2, pp. 354 (1970). R. B. Bates g£_§1,, Tetrahedron Letters, No. 3, 199 (1967). IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 1113111191111mulwmlnummnuln