Rate and Equilibrium Constants for the Addition of N-Heterocyclic Carbenes into Benzaldehydes: A Remarkable 2-Substituent Effect**

Rate and equilibrium constants for the reaction between N-aryl triazolium N-heterocyclic carbene (NHC) precatalysts and substituted benzaldehyde derivatives to form 3-(hydroxybenzyl)azolium adducts under both catalytic and stoichiometric conditions have been measured. Kinetic analysis and reaction profile fitting of both the forward and reverse reactions, plus onwards reaction to the Breslow intermediate, demonstrate the remarkable effect of the benzaldehyde 2-substituent in these reactions and provide insight into the chemoselectivity of cross-benzoin reactions.

Acyl anion equivalents generated from the reaction of Nheterocyclic carbenes (NHCs) with aldehydes are important catalytic intermediates that can undergo ar ange of carboncarbon bond forming processes. [1] In this regard, NHCcatalyzed benzoin and Stetter reactions have been widely studied, with an umber of efficient catalytic asymmetric methods available for both intra-and intermolecular reactions. [1,2] However,t he development of cross-benzoin reactions has proven difficult in terms of the chemoselective formation of as ingle reaction product. [3] While efficient chemoselective NHC-catalyzed protocols for both intra-and intermolecular cross-benzoin reactions between aldehydes and ketones have been reported, [4] the reaction between two distinct aldehydes remains as ignificant synthetic challenge. As 2-substituted benzaldehydes are generally poor substrates for homo-benzoin reactions they have been widely utilized in cross-benzoin processes. [5] Fore xample,M iller and Mennen reported the intramolecular cross-benzoin reaction between an arylaldehyde and at ethered aliphatic aldehyde to effect macrocyclization. [5b] Connon and co-workers found that N-C 6 F 5 triazolium NHC precatalyst 3 catalyzes intermolecular cross-benzoin reactions between 2-substituted benzaldehydes and aliphatic aldehydes with high levels of chemoselectivity (Scheme 1a). [5c] As elective cross-benzoin reaction between two benzaldehydes catalyzed by thiamine diphosphate dependent benzaldehyde lyase (BAL) was reported by Müller et al.,with one 2-substituted benzaldehyde aprerequisite for good chemoselectivity. [6] Glorius and co-workers subsequently utilized this phenomenon in arylaldehyde crossbenzoin reactions using thiazolium NHC precatalyst 7 (Scheme 1b). [5e, 7] Gravel et al.h ave reported at riazolium NHC-catalyzed cross-benzoin process between benzaldehydes and alkyl aldehydes,w ith preliminary kinetic studies showing the reaction is at least first-order with respect to both aldehydes and that the chemoselectivity was determined at or after the C À Cb ond forming step. [5h] Current explanations of the observed chemoselectivity in cross-benzoin reactions of arylaldehydes are usually simplistically based upon steric arguments.P revious to this investigation, it was commonly assumed that the presence of a2substituent decreases the rate of NHC addition into an arylaldehyde (Scheme 2). [8,9] TheNHC I therefore preferably adds into aldehyde II to form least-hindered 3-(hydroxybenzyl)azolium adduct IV,w hich undergoes deprotonation to form Breslow intermediate V. [7,10] However,toaccount for the observed selectivity,i ntermediate V must now add into the more "hindered" 2-substituted benzaldehyde VI. [5c,d, 6] This steric argument is therefore inherently contradictory.T here are currently no detailed mechanistic studies that offer insight into the rate of NHC additions into 2-substituted benzaldehydes,the effect of the N-aryl NHC substituent upon the rates of these processes,o rt he role of the 2-substituent in chemoselective cross-benzoin reactions of arylaldehydes. Building upon our previous mechanistic studies of NHCcatalyzed processes, [11] herein the remarkable effect of 2arylaldehyde substitution upon equilibrium constants for 3-(hydroxybenzyl)azolium adduct formation is demonstrated. Fort he first time,i ndividual rate constants for adduct formation have been determined under stoichiometric conditions and the effects of both aldehyde and N-aryl NHC substitution have been probed, with the results offering potential insight into the chemoselectivity of cross-benzoin processes.
First, the catalytic reactions between ar ange of substituted benzaldehydes (0.01m)a nd NHC precatalyst 9-11 (0.002 m,2 0mol %) using Et 3 N( 0.002 m,2 0mol %) in CD 2 Cl 2 were monitored through in situ 1 HNMR spectroscopy.A nalysis of the resulting reaction profiles allowed equilibrium constants for adduct formation (K exp )t ob e determined (Table 1). [12] Theresults demonstrate the remarkable effect of having ah eteroatom substituent in the 2position of the benzaldehyde on K exp .F or example,t he reaction between NHC precatalyst 9 and 2-methoxybenzaldehyde 2 gave K exp = 56 m À1 compared with K exp = 3 m À1 for reaction with benzaldehyde 5 (Table 1, entries 1a nd 2). As observed previously, [11] the 2,6-substituted NHC precatalysts 10 and 11 gave significantly higher K exp values,a lthough 2methoxy aldehyde substitution again led to further prominent increases (Table 1, entries 3-6). Thei mportance of the 2heteroatom for this effect is demonstrated by reaction of NHC precatalyst 10 with 2-tolualdehyde 12,w hich gives K exp = 16 m À1 (Table 1, entry 7). Theeffect is not limited to 2alkoxy substituents,a st he reaction with 2-bromobenzalde-hyde 14 gave K exp = 332 m À1 whereas reaction with 4-bromobenzaldehyde 15 gave K exp = 15 m À1 (Table 1, entries 9a nd 10). Theintroduction of an additional heteroatom substituent in the 6-position further shifted the equilibrium in favor of adduct formation. Fore xample,r eaction of 10 with 2,6difluorobenzaldehyde 17 gave K exp = 785 m À1 whereas with 2fluorobenzaldehyde 16 K exp = 150 m À1 (Table 1, entries 11 and 12). Theu se of 2-pyridinecarboxaldehyde 18 also gave an equilibrium strongly in favor of the corresponding adduct, while reaction with 6-bromo-2-pyridinecarboxyaldehyde 19 exclusively gave 3-(hydroxybenzyl)azolium adduct 33 such that K exp could not be measured (Table 1, entries 13 and 14). In most cases,t he 3-(hydroxybenzyl)azolium salts could also be isolated from astoichiometric reaction between the NHC precatalyst and the corresponding aldehyde in the presence of excess Et 3 N.
To gain further insight into the dramatic effect of 2heteroatom substitution, rate constants for 3-(hydroxybenzyl)azolium adduct formation were measured. First, the effect of the N-aryl NHC substituent was assessed, as no kinetic measurements have previously been made for triazolium-catalyzed benzoin or Stetter processes. [13,14] Reactions of aldehyde 13,which is often employed as amodel substrate for intramolecular Stetter reactions,w ere performed under presteady-state conditions using stoichiometric concentrations of NHC precatalysts in CD 3 OD with aE t 3 N:Et 3 N·HCl (2:1) buffer at 15 8 8C, [15] analogous to the conditions used by Leeper and White in their study of the thiazolium-catalyzedbenzoin reaction. [13a] Kinetic analysis of the reaction profiles obtained before significant product formation (< 5%)allowed pseudo Scheme 2. General mechanism for ac ross-benzoin reaction. . .
second-order rate constants for 3-(hydroxybenzyl)azolium adduct formation (k 1 , m À1 s À1 )and equilibrium constants (K exp , m À1 )t ob em easured ( Table 2). [16] Formation of the 3-(hydroxybenzyl)azolium adduct involves two distinct steps: the initial deprotonation of precatalyst by base and the subsequent reaction of the NHC with aldehyde.A fter the formation of adduct oxyanion, the base can be regenerated upon protonation at oxygen resulting in an overall pseudo second-order process under these experimental conditions. This is confirmed by the excellent fitting of reaction data to ak inetic expression describing as econd-order reaction proceeding to ap osition of equilibrium. [12] Thep seudo firstorder rate constants for adduct dissociation (k À1 ,s À 1 )c ould also be calculated as K exp = k 1 /k À1 .Additional estimates for k 1 and k À1 were obtained from reaction profile fitting,w ith the values used to calculate the corresponding equilibrium constants (K fit ). Pleasingly,t he fitted values obtained are in good agreement with those obtained from kinetic analysis, with the largest discrepancyo ccurring for the reaction using NHC precatalyst 36 where adduct dissociation is negligible (Table 2, entry 4). [17] Next, the reverse decay towards equilibrium was studied. Analysis of the 1 HNMR reaction profiles for dissociation of the adducts of aldehyde 13 allowed rate and equilibrium constants of dissociation to be measured (k d ,s À 1 and K diss , m À1 )and rate constants for association (k a , m À1 s À1 )t ob ec alculated (Table 3). [18] Although k a = k 1 and k d = k À1 adistinction has been made to differentiate between the two methods of measurement. Thed issociation analysis was not possible for the N-2,6-(MeO) 2 C 6 H 3 adduct as the equilibrium lies so far towards the adduct that insufficient data could be obtained. Notably, the values for the equilibrium and rate constants measured from both the forward and reverse reactions at the same temperature are in good agreement with each other, showing that these methods can be used to give reliable measurements.
Comparing the N-aryl NHC precatalysts,t he rate of adduct formation (k 1 or k a )i ncreases with more electron-withdrawing N-aryl substituents (4-F > 4-H > 4-MeO). This reflects the trend in pK a for the NHC precatalysts (pK a 4-F < 4-H < 4-MeO), [19] suggesting that the rate of 3-(hydroxybenzyl)azolium adduct formation is more influenced by the equilibrium for precatalyst deprotonation. However, N-Mes precatalyst 10 is an exception as its pK a is similar to N-Ph precatalyst 9 (pK a 17.7 and 17.8, respectively) but it reacts 2.5 times faster. This is postulated to be due to the orthogonal orientation of the mesityl substituent to the triazolium ring providing amore favorable approach of the aldehyde. [20] In all cases 3-(hydroxybenzyl)azolium adduct formation shows adegree of reversibility,however the kinetic data shows the rate of dissociation for the adduct derived from 13 and 10 is particularly slow,m eaning that adduct formation is effectively irreversible in this case. [21] Having established reliable methods for measuring equilibrium and rate constants for adduct formation this analysis was extended to look at substituted benzaldehydes (Table 4). [22] Ther eactions were performed using NHC precatalyst 9,w ith comparable data obtained from both kinetic analysis and reaction profile fitting in all cases.T he presence of aheteroatom in the aldehyde 2-position again has am arked effect, leading to significantly higher equilibrium constants for adduct formation. [23] Thek inetic data gives an insight into the origin of this trend. Fore xample,t he rate of NHC addition into 2-methoxybenzaldehyde 2 is over 2.5

Angewandte
Chemie times faster than addition into benzaldehyde 5,a nd over ten times as fast as addition into 4-methoxybenzaldehyde ( Table 4, entries 1-3). As imilar trend is seen comparing intramolecular Stetter substrate 13 with its 4-substituted analogue,d emonstrating that the 2-substituent effect is not purely electronic in nature (Table 4, entries 4and 5). In both cases the rate of the reverse process is also up to five times slower for 2-substituted benzaldehydes,r eflecting the increased stability of these adducts.T he importance of the heteroatom substituent is highlighted by the use of an analogue of 13 without the oxygen atom linker and 2-and 4-tolualdehyde,which all give equilibrium and rate constants comparable with benzaldehyde 5 (Table 4, entries 6-8).
However,e ven in this case the rate of NHC addition into 2tolualdehyde is nearly twice as fast as addition into 4tolualdehyde (although the effect is smaller compared with heteroatom substituents). Further kinetic analysis of the reaction profiles following the decreasing concentrations of the 3-(hydroxybenzyl)azolium adducts over time allow estimation of the pseudo firstorder rate constants for deprotonation (k 2 ,s À 1 )i nto the transiently formed Breslow intermediates (Table 4). Therate constants for deprotonation are of the same order of magnitude for all the aldehydes,i ncluding those containing a2-substituent. Unlike the observed substituent effect on the first step (k 1 and K), the observed order of reactivity on k 2 reflects normal through-bond electronic effects on carbon acidity where electron-donating groups on the aldehyde decrease the rate of deprotonation. This is in agreement with our previous observations of normal electronic effects of the NHC N-aryl substituent on this deprotonation step. [11] Rate constants for deuterium exchange at the benzylic position of O-methylated 3-(hydroxybenzyl)azolium adducts were observed to decrease in the presence of electrondonating substituents (for example,2 -MeO) on the N-aryl ring.
Thek inetic and equilibrium data of NHC addition into the benzaldehydes potentially offers insight into the observed chemoselectivity of cross-benzoin reactions.Arepresentative cross-benzoin reaction between benzaldehyde 5 and 2methoxybenzaldehyde 2 was performed using NHC precatalyst 3 (20 mol %) in CH 2 Cl 2 at 45 8 8C( Scheme 3a). The observed chemoselectivity is consistent with that previously reported, [5e] with cross-product 37 favored and smaller amounts of homo-benzoin 38 and benzoin 39 also formed (Scheme 3a). Similar product ratios were observed using NHC precatalyst 11,a lthough the conversion was lower (ca. 15 %). Monitoring the cross-reaction at 25 8 8Cu sing NHC precatalyst 11 revealed a1 0:1m ixture of 3-(hydroxybenzyl)azolium adducts 25:24 at equilibrium, again demonstrating ap rominent 2-substituent effect in this system (Scheme 3b). However,d espite formation of adduct 25 being favored, cross-product 37 is derived from reaction of minor adduct 24, indicating the chemoselectivity must be determined later in the reaction pathway. [24] This leads to three main possibilities for the origin of the observed chemoselectivity:1)formation of the Breslow intermediate;2 )onwards reaction of the Breslow intermediate;3 )dissociation of the resulting tetrahedral adducts (Scheme 3c).
Them easured rate constants for Breslow intermediate formation show that 2-MeO substitution decreases k 2 by af actor of about two relative to benzaldehyde 5 (Table 4, entries 1a nd 2), however this does not outweigh the tenfold increase in equilibrium constant for adduct formation with a2 -MeO substituent and cannot account for the observed chemoselectivity.Adifference in rate of the onwards reaction of the two Breslow intermediates 40 and 41 would account for the cross-benzoin selectivity.I nb oth cases reaction with 2methoxybenzaldehyde 2 will be comparatively fast over  . . reaction with benzaldehyde 5 owing to the previously described 2-substituent effect. However, the increased steric hindrance around the nucleophilic carbon of 41 compared with 40 may decrease its relative rate of addition sufficiently to explain the formation of cross-benzoin 37. [25] Alternatively,N HC dissociation from tetrahedral intermediate 43 may be slow compared with 42,again resulting in preferential formation of cross-product 37.T his would be consistent with the measured rate constants for dissociation (k À1 )o ft he related 3-(hydroxybenzyl)azolium adducts in which a4-fold difference was observed (Table 4, entries 1and 2). However,accumulation of intermediates such as 42 and 43 have not been observed in any of our NMR experiments to date,o ri ne arlier NMR studies by Leeper and White of the thiazolium-catalyzed benzoin reaction, [13a] suggesting afaster rate of breakdown relative to the rate of formation from the relevant Breslow intermediate and aldehyde.F urthermore, monitoring reactions of NHC precatalyst 11 with either 37 or 38 gave about 10 %retro-benzoin products but no observable products consistent with formation of the corresponding tetrahedral adducts. [26] Additionally,acontrol experiment reacting NHC precatalyst 11 with acetophenone gave no observable products,s uggesting that any NHC-ketone adducts formed rapidly dissociate.T herefore,i ts eems more likely that the chemoselectivity in cross-benzoin reactions is determined by the onwards reaction of the Breslow intermediate.
Although the increased rate of nucleophilic addition into benzaldehydes bearing a2 -heteroatom substituent is clearly evident, the origin of this phenomenon is unclear. [27] One possibility is that the presence of alone pair on an atom in the 2-position changes the conformation of the aldehyde carbonyl such that it twists out of conjugation with the aryl ring. This ground state destabilization of aldehyde could result in increased reactivity towards nucleophiles.A lternatively increased product stability due to hydrogen bond formation between the 2-heteroatom substituent and the OH group of the 3-(hydroxybenzyl)azolium adducts could also contribute to the observed increase in both rate and equilibrium constants.T hese ground and product state effects could be realized in any nucleophilic addition to 2-substituted aldehydes of this type,i ncluding in the onward reaction of Breslow intermediates in cross-benzoin reactions.
In conclusion, measurements of equilibrium and rate constants for the reaction of triazolium NHC precatalysts with substituted benzaldehydes to give 3-(hydroxybenzyl)azolium adducts under both catalytic and stoichiometric conditions have been made.The results obtained from kinetic analysis and fitting data for both the forward and backwards processes show that nucleophilic addition into benzaldehydes bearing a2 -heteroatom substituent is particularly fast. By contrast, smaller substituent effects are observed on the rate of deprotonation of 3-(hydroxybenzyl)azolium adducts,which fall within the same order of magnitude regardless of aldehyde substitution. Ther esults offer insight into the apparent inconsistency over the second aldehyde addition in cross-benzoin reactions,o verturning the assumption that 2substituted benzaldehydes are less reactive based upon steric arguments.