Substituent effects on the formation of benzyl ions from ortho - methoxy substituted 1,1-diarylalkanes under electron ionization: correlations between the abundance of the process and the 13 C-NMR chemical shifts of the neutral precursors

Good linear relationships have been obtained between the abundance of the formation process of benzyl ions from ortho -methoxy- substituted 1,1-diarylalkanes, through consecutive rearrangement reactions induced by electron ionization, and 13 C-NMR chemical shift values of C-1 of the ortho -methoxy- substituted aromatic ring of the neutral precursors for a large number of compounds: 1,1-diphenylethanes 1–19 , 2-methyl-1,1-diphenylpropanes 20–37 , and 1,1,1-trichloro-2,2-diphenylethanes 38–44 . This result is satisfying as good linear relationships between the ion abundances in the mass spectrum and the substituent effects are not easily obtained. Further, the abundance of the process depends on the effects of substituents linked to the ortho -methoxy- substituted benzene ring (ring A) while measurable effects of substituents linked to the other benzene ring (ring B) have not been observed.


Introduction
It has been reported 1 that the electron ionization (EI) mass spectra (MS) of 1,1-diphenylalkanes bearing an ortho-methoxy group are characterized by the presence of abundant benzyl (or tropylium) ions, which are completely absent in derivatives lacking the orthomethoxy substituent.In fact, the EI/MS of such ortho-substituted compounds are dominated by the molecular ion (M + ˙), the diphenylmethyl cation a (arising from loss of the alkyl group linked to the benzyl carbon, as a radical), and the benzyl ion b, as can be seen in the 70 eV EI/MS of 2methoxy-1,1-diphenylethane (1) (Figure 1).
A B The presence of abundant metastable ions generated in the first field-free region, as well as the presence of the appropriate peaks in mass analyzed ion kinetic energy (MIKE)-or linked scan (B/E = constant, B 2 /E = constant) spectra, indicates the ion a as the precursor of the ion b. 2,3 It has been unequivocally demonstrated by experiments with 2 H-and 13 C-labeled compounds that the whole process of benzyl ion formation involves migration of the methylene residue of the methoxy group linked to the 2-position of an aromatic ring (ring A) to the other aromatic ring (ring B). [1][2][3][4] Further investigation revealed that such a reaction occurs also for ortho-alkoxy derivatives (OEt, OiPr) by migration of an alkylidene residue other than methylene, as well as for ortho-alkyl-hetero (NHMe, NMe 2 , SMe) 1,1-diarylalkanes. 4 Finally, it has been determined that the formation of the benzyl-(or tropylium) ions b constitutes the main unimolecular decomposition reaction of ortho-alkylheterosubstituted diaryl-or alkyldiaryl-or triaryl-methyl cations, a, generated either under electron ionization 5 or chemical ionization conditions. 6his previously unreported formation of benzyl ions from diphenylmethane derivatives 7 through skeletal rearrangement induced by proximity effects is therefore to be considered an important general fragmentation reaction of the cations a occurring over a wide range of internal energy.From a mechanistic point of view it has been suggested 1,4 that the formation of the ion b involves consecutive rearrangements (steps 1 and 2) of ortho-methoxy diphenylmethyl cations a, followed by a simple cleavage reaction (step 3) as shown in Scheme 1 for compound 1.In particular, a hydrogen migration as hydride from the α-position of the ortho-methoxy group to the carbenium center affords the rearrangement ion a' (Step 1).This latter undergoes a further rearrangement to a′′′ by electrophilic attack of the charged alkylidene (Step 2), probably through a six-membered transition state (a′′), followed by a carbon-carbon displacement reaction.Finally, the benzylic cleavage of ion a′′′ gives the rearrangement benzyl ion b.
The small kinetic energy release values, calculated from the width of the peak at half-height (T 1/2 ) in the MIKE spectra for the a b process, agree with a last step involving a simple cleavage reaction. 2The isotopic effect indicates that the hydrogen migration affording the rearrangement ion a' is the rate-determining process, and the close values in 70 eV-(k H /k D = 1.4), 1,3 and in MIKE-(k H /k D = 1.6), 2 spectra agree with a loose transition state for the hydrogen migration. 8n appropriate geometry of the transition state that should account for the hydride migration involves the ring A being perpendicular to the sp 2 system of the charged benzyl carbon, with the methyl of the ortho-alkylhetero group oriented toward the carbenium center. 2 In order to achieve detailed mechanistic insight and information on the arrangement of the transition state related to the process involving the hydrogen migration, the determination of substituent effects on the abundance of the benzyl ion formation process could be used as a valuable tool.Hence, in this work we correlate the substituent effects with the values of the ratio Z between the abundance of the benzyl ion b and the sum of the abundances of the benzyl ion b with that of its precursor ion a [Z = I b %/( I a % + I b %)].The ratio Z is, in good approximation, a measure of the fraction of ions a reacting to form ions b.It has been proved previously that Z is a valuable quantitative parameter for correlating the abundance of the process with the degreeof-freedom effect, as well as with the approximate activation energy. 2

Results and Discussion
As our starting point, we compared the Z% value of the 2-methoxy-1,1-diphenylethane (1) with those of derivatives 2-11 characterized by an unsubstituted ring B and bearing substituents at the 4-and/or 5-position in the ring A (Table 1).
The EI/MS (see Experimental Section) of the compounds 1-11 are dominated by the peaks corresponding to the molecular ion M + ˙, the ion a, and the rearrangement benzyl ion, b (m/z 91).
Analysis of the 5-substituted derivatives 2-5 shows a Z% = 57 for the 5-methyl-derivative 2, very close to that of 1 (Z% = 53), which agrees with the poor electronic effects of the methyl group.A significant small increase of Z% with respect to 1 is observed for the 5-methoxy derivative 3 (Z% = 63), the 5-nitro derivative 4 (Z% = 64) and the 5-bromo derivative 5 (Z% = 62).These substituents linked at the 5-position exert a small electron-withdrawing effect with respect to the metapositions (C-1 and C-3).
All these findings revealed a meaningful dependence of Z% on the substituents of the ring A. In particular, the substituents which exert an electron-withdrawing effect on C-1 increase the abundance of the process, while strong electron-donating groups produce a dramatic decrease in Z%.In fact, the lower is the electron density on C-1 of the ring A of the ion a is, the higher is the fraction of ions a affording ions b.
In addition to such a qualitative approach, it seemed of interest to verify whether linear relationships occur with the electronic effects of the substituents.However, the use of electronic constants in our substrates, which are characterized by the presence of several substituents, could be strongly affected by severe limitations owing to the occurrence of orthointeractions, which are not easily evaluated, a priori.
The 13 C NMR chemical shift constitutes a physical parameter related to the electronic density of the carbon atoms. 9However, NMR experiments on the diaryl cations a are complicated by the difficulty of generating them in the solution phase.In fact, this should involve the use of superacids on the appropriate precursors, and low temperature experiments. 10Hence, the reasonable assumption that a substituent could play a role in the same direction either in ions a or in their neutral precursors prompted us to correlate the 13 C NMR chemical shift values (used as electronic parameters) of the neutral precursors, the 2-methoxy-1,1-diphenylethanes, with Z%.This approach has been applied successfully in the study of the reactivity of carbanions in solution. 11hen, taking the chemical shift value of C-1 carbon atom of the ring A for 1 as reference, a correlation between the substituent-induced chemical shift variations (cs) for compounds 1-11 (Table 1) and the Z% values was carried out.
The results shown in Figure 3 show that a good linear correlation (s 3.67 ± 0.21, i 57.59 ± 1.40, n 11, r 0.986) has been obtained.Refer to Table 1 for the values.
Once the close dependence of the process on the electronic density of C-1 has been asserted, it seemed of interest to verify a possible influence of substitutions at the ring B. Comparison of the Z% value of the 4'-methoxy derivative 12 (Table 1) with that of 1 shows that the presence of the 4'-methoxy group in the ring B does not produce significant change (Z% = 57).Similarly, the introduction of the 4'-nitro group into the structure 11, i.e., compound 13, leads to a value of Z% = 87, practically identical to the Z% value of 11.Incidentally, the 70 eV EI/MS of the dimethoxy-2 H 6 isotopomer, 13-d 6 , gives evidence of an isotopic effect, calculated by the ratio of I% of m/z 136 (ion b for 13)/ I% of m/z 138 (ion b for 13-d 6 ), K H /K D = 1.5, very close to that previously determined for other derivatives 1,3 lacking electron-withdrawing substituents on the ring B. This means that the hydride migration step 1 (Scheme 1) constitutes the rate-determining reaction also in the presence of a para-nitro group in the ring B, which should enhance the activation energy of the electrophilic substitution involved in Step 2.
Furthermore, good linear relationships between Z% and cs of C-1 for all 1-11, together with the polysubstituted compounds 12-19, have been obtained (s 3.74 ± 0.25, i 60.84 ± 1.50, n 19, r 0.963).The substantial absence of effects of the substituents linked to the ring B on the abundance of the process was also stressed by the lack of any correlation between Z% and the 13 C-NMR chemical shift values of C-1′ of the ring B.

Conclusions
The choice of the 13 C-NMR chemical shift values of C-1 of the ring A of the neutral precursor of the benzyl ions b, formed by consecutive rearrangements of the diarylmethyl cations a, leads to good linear relationships with the fraction of ions a affording the ions b, measured by the ratio Z%, for a large number of compounds.This result is quite satisfying, as good linear relationships between the ion abundances in the mass spectrum and the substituent constants are not easily obtained.In fact, this approach does suffer severe limitations owing to: (i) the fact that the ion abundances in the mass spectrum results from competitive and consecutive reactions of precursors with a relatively large range of internal energy; (ii) there is the possible occurrence of isolated electronic states, not easily predictable a priori and, (iii), there is the degree-of-freedom effect on the fragment-ion abundances. 8urther, our findings give evidence that the abundance of the formation process of benzyl ions b depends on the effects of substituents linked to the ortho-methoxysubstituted benzene ring (ring A).In particular, the lower is the electron density of C-1 of the ring A of the neutral precursor, the higher is the fraction of ion a which give benzylic ion b.On the other hand, measurable effects of substituents linked to the other benzene ring (ring B) have not been observed.
Finally, it seems possible to estimate the ion-abundances in the EI/MS of ortho-alkylheterosubstituted 1,1-diarylalkanes by means of a unique parameter which reflects the overall effects of the substituents, that is, the C-1 13 C NMR value of the neutral precursor.Table 3. 13 C NMR data a and Z% b of 38-44

1-(2-Methoxyphenyl)-1-phenylethane (1)
. 13  A solution of the appropriate alcohol (10 mmol) in glacial acetic acid (25 ml) was added to a stirred solution of appropriate aromatic substrate (50 mmol) in glacial acetic acid (25 ml) / 70% H 2 SO 4 (25 ml) at 20 °C.After standing at room temperature overnight, the mixture was poured onto crushed ice and the oil obtained was extracted with ethyl acetate, neutralized and concentrated under reduced pressure.The unreacted aromatic substrate was removed by steam distillation.Finally the residue was extracted with ethyl acetate, dried and concentrated at reduced pressure.The utilized alcohol and aromatic substrate, together with the purification method, are reported below.For compounds 2, 3, 5, and 7, commercial sec-phenylethanol was utilized and aromatic substrates were 4-methylanisole, 1,4-dimethoxybenzene, 4-bromoanisole, and 1,3dimethoxybenzene, respectively.The crude products were chromatographed over silica gel, employing cyclohexane:diethyl ether (95:5) as eluent.
After stirring overnight at room temperature, the reaction was quenched with solution of ammonium chloride, then extracted with diethyl ether, neutralized and dried.After solvent evaporation in vacuo, the oily residue was chromatographed over silica gel.Elution with cyclohexane gave pure 6 (yield 65 %) as a white oil. 13 To a solution of 4-methylanisole (81 mmol) and isobutyraldehyde (27 mmol) in glacial acetic acid (50 ml), 98% H 2 SO 4 (30 ml) was added dropwise with stirring.External cooling with ice was necessary in order to maintain the reaction temperature below 20 °C.After standing at room temperature overnight, the mixture was poured onto crushed ice and the oil obtained was extracted with ethyl acetate, neutralized, and concentrated under reduced pressure.The unreacted 4-methylanisole was removed by steam distillation.Finally, the residue was extracted with ethyl acetate, dried and concentrated at reduced pressure.Pure 28 was obtained by chromatography over silica gel, employing cyclohexane: diethyl ether (90:10), as eluent.White crystals from EtOH (yield 60%), m.p. 101 °C. 13 After cooling, the solid material was removed by filtration and the solution was diluted with water and extracted with diethyl ether.The ethereal extracts were treated with 10% aqueous sodium hydroxide solution, then neutralized, dried and evaporated in vacuo.The residue obtained was chromatographed on silica gel employing cyclohexane:ethyl acetate (90:10) as eluent.The elution order of the compounds was: 10, 9.

-6. 23 a
Chemical shift values for C-1 are in ppm; ∆cs(C-1) values are calculated as differences between δ(C-1) of each compound and δ(C-1) of compound 1. b Z% are calculated as I b %/( I a % + I b %) per cent.

33a
Chemical shift values for C-1 are in ppm; ∆cs(C-1) values are calculated as differences between δ(C-1) of each compound and δ(C-1) of compound 20.b Z% are calculated as I b %/(I a % + I b %) per cent.

58a
Chemical shift values for C-1 and C-1' are in ppm; ∆cs(C-1) values are calculated as differences between δ(C-1) of each compound and δ(C-1) of compound 38.b Z% are calculated as I b %/( I a % + I b %) per cent.