The low K Enol values of β -sulfonyl-substituted amides

Amides substituted by one β -sulfonyl group and another β -sulfonyl, β -ester or β -CN group, form very low percentages of the corresponding enols, lower than for the β , β -diester and β -cyano, β -ester substituted systems, despite the equal or weaker electron delocalizing ability of the latter groups which help to stabilize the enols more than that of the sulfonyl group. This cannot be attributed to the non-planarity of the enols, since the calculated structures are planar. It is suggested that the sulfonyl-substituted amides are more stabilized than the β -ester- or β -cyano- substituted amides. An amide substituted by β − nonafluorosulfonyl, β − acetyl groups enolizes on the acetyl group, forming a strong, nearly symmetrical intramolecular hydrogen bond. The use of 600 MHz NMR spectroscopy can extend the range of observable enols.

We assume that the lack of a general pK Enol vs. pK a (CH 2 YY′) correlation is significantly affected by the left-hand side of Eq. 1 which includes stabilization of the amide by structure 1b superimposed on destabilization by electrostatic repulsion between the C=O and the C-Y and C-Y′ dipoles, effects which are not correlated with the pK a 's.
In the present paper we want to find out if the observed low K Enol compared with the intuitively expected ones for β-sulfonyl-substituted amides reflects a low extent of promotion of the enolization.We prepared several amides substituted by RSO 2 , R'SO 2 ; RSO 2 , CO 2 R' and RSO 2 , CN EWGs pairs, and tried to observe the derived enols, calculate their K Enol values, and explain the observations by computation.

Structures in solution
In addition to the amide 5, the enols 6 which may have an E-or Z-configuration, the sulfonyl group itself can be a potential enolization site, giving species 8.The three species are shown in Eq. 3, with Z-6 and 8 as hydrogen bonded species.The relative stability of the enol 8 was probed by B3LYP/6-31+G* (B3LYP/6-31G**) calculations.The cyanosulfonyl enols 8 derived from 5l and 5n showed no stable structure.A barrierless proton transfer gave the amides 5l and 5n.The ester-or the acetyl-substituted species (5a, 5c, 5o and the PhSO 2 analogue of 5o) also did not give stable structures 8. Proton transfer from the enols 8a and 8c first converged to the intramolecularly hydrogen bonded enol of ester 9 which was at a minimum, by transferring the proton to the adjacent ester group.It then further transferred the proton to the amido carbonyl to give the enols 6a and 6c on the amide carbonyl.The latter were 10.9 (12.3) and 12.5 (13.7) kcal/mol less stable than 6a and 6c, respectively.In contrast, the enols formed from proton transfer to the acetyl group were stable, and indeed the enol 7 was actually the product isolated.
Comparisons between enolizations at the three sites of acetyl-substituted sulfonylamides 7, R' = i-Pr, Y = COMe showed that enolization on the acetyl carbonyl to give 7 is comparable to enolization on the amido oxygen.For the known analogue with R = C 4 F 9 the enolization is 0.6 (0.2) kcal/mol more favored to give 7 than on the amide group to give 6, which agrees with the observed product 7.The % enol in these cases is at most 2-4%, i.e., K Enol = 0.04 (CCl 4 ) and 0.03 (CDCl 3 ) for 5b/6b and 0.03 (CCl 4 ) and 0.02 (CDCl 3 ) for 5c/6c, but the integration of the small OH signal is not very reliable.In an attempt to obtain a more reliable integration and to detect lower percentages of the enol, if any, the1 H NMR spectra of most of amides 5 were measured in CCl 4 or in CDCl 3 with a 600 MHz instrument.The increased sensitivity enabled us to see the OH signals of 6a, 6b, 6d and 6i more clearly and with a better integration than that in the 400 MHz instrument, and to determine enol percentages of 0.08-0.84%.Moreover, in cases where signals for the enols were not observed, we assume that 0.05% of the enol could have been observed.A drawback is that a weak broad signal at ca. 14.50 ppm, which is ascribed to an unknown impurity, was observed in all the spectra of 5/6 in CDCl 3 and in CDCl  13 C NMR data are given in Figure S1 of the Supporting Information.

Solid State Structures
The solid state structure of "formal" 15 5c, crystallized from EtOAc/petroleum ether and 5o, crystallized from CDCl 3 were determined by single crystal X-ray crystallography.The ORTEPs and the full data for both compounds are given as CIF's in the Supporting Information.This raises the possibility that the static unsymmetrical structures of 7o and 6o at low temperature have a low barrier for reversible hydrogen transfer between the acetyl and the amide oxygens, leading to the observed close to symmetric hydrogen bond.A similar structure with a less symmetric hydrogen bond was obtained for the enol derived from the formal amide PhNHCOCH(COMe)CO 2 Et. 1 A similar difference between calculated and observed hydrogen bond parameters was reported for enols of cyanomalonamides. 12 3) 139.8 o ).We note that a competitive enolization on the COR and CONRR' carbonyls was earlier demonstrated by isolating both solid enols in the 2-carbanilido-1,3-indanedione system. 18The calculated structures of 5c and 7o are given in Figure 1.The calculated B3LYP/6-31+G* thermodynamic data for the barriers of eq. 4 in kcal/mol, kcal/mol and e.u. are ∆H 91.43, ∆G 0.97 and ∆S 1.50 for 7o Transition state and 0.52, 0.29 and 0.80 for the corresponding 6o Transition state, respectively.

O ... H-O values
The results given above, especially the absence of enol signals even in the 600 MHz spectra indicate that the β-sulfonyl-substituted amides, substituted by another β-sulfonyl, β-ester or β-cyano group undergo an inefficient enolization, in contrast with the other Y,Y′ pairs mentioned above.This can be ascribed to three reasons.(a) The sulfonyl together with the other Y group are weaker resonatively EWG than these Y,Y′ groups, and hence amide stabilization due to structure 1b overcomes enol stabilization due to structure 2b.Precedents for such behavior are known for several Y,Y′ pairs. 2 (b) Steric interaction between the β-substituents twist them out of planarity from the C=C(OH)NRR′ plane, thus reducing the maximum resonative stabilization of the enol, which is achieved at full planarity (cf.structure 2b).Precedents for this behavior in diestersubstituted systems were shown by computations. 10(c) Amide destabilization of sulfonylsubstituted systems is lower than in the corresponding esters.The low enolization ability will be discussed in comparison with other EWGs, especially ester groups, in terms of these points.
The following conclusions, based on the B3LYP/6-31+G* (B3LYP/6-31G**) values arise from Table 2: (a) For i-PrNHCOCH(SO 2 Ph)COMe ∆G for enolization on the acetyl group is 0.8 (-0.1) kcal/mol more negative than on the amide carbonyl.A value of 0.8 was calculated for the N-Ph analogue, but the B3LYP/6-31G** value is not available since the enol on acetyl is not a stable structure.For the more EW SO 2 C 4 F 9 derivative the corresponding differences are 0.7 (-0.2) and 0.7 (0.2) kcal/mol for the N-i-Pr and N-Ph derivatives, respectively.The pK Enol values at B3LYP/6-31+G* are more negative for enolization on the acetyl site, as observed experimentally for 7, although the differences are not large.This is noteworthy since when the competition between the two groups is not intramolecular the calculated ∆G difference for H 2 C=CHCOX prefers enolization when X = Me over that when X = NH 2 by 22 kcal/mol. 20The acyl activated enols are among the most stable enols, judged by the pK Enol values.(b) As expected, the most stable enols are those substituted by two SO 2 C 4 F 9 groups, with a larger preference for the N-i-Pr derivative.The N-Ph, (SO 2 C 4 F 9 ) 2 -derivative is 14.5 kcal/mol more favored than the (PhSO 2 ) 2 analogue.(c) The N-substituent effect on ∆G (in kcal/mol) is appreciable.For the CO 2 Me, PhSO 2 combination the ∆G order followed is t-Bu 4.9 (0.8) > 2,4-(MeO) 2 C 6 H 3 4.6 (-0.3) > i-Pr 1.8 (-2.0) > Ph -0.5 (-4.1).Although it may be fortuitous, the three of the four systems measured at 600 MHz are among the few which display observable enols and they are the bulkier substituents.For the SO 2 Ph, SO 2 Ph combination the effect is large and the order of ∆G values is 2,4-(MeO) 2 C 6 H 3 8.9 (5.2) > Ph 6.1 (2.6) > i-Pr 5.0 (0.8), whereas the effect is smaller for SO 2 Me, SO 2 Ph [∆G order is Ph 3.0 (-0.4) >i-Pr 2.2 (0.3)].For systems with no EWG the ∆G order is i-Pr 32.0 (31.4) > 2,4-(MeO) 2 C 6 H 3 31.7 (30.7) > Ph 29.8 (28.7) > t-Bu 27.3 (30.4).We conclude that there is no constant or observed systematic order of the effect of the N-substituents.(d) For N-Ph, Y = SO 2 Ph the order of ∆G for Y′ is SO 2 Ph 6.1 (2.6) > CN 5.1 (2.0) > SO 2 Me 3.4 (-0.2) > COMe 2.6 (-2.9) > CO 2 Me -0.5 (-4.1), and for N-i-Pr, Y = SO 2 Me the order for Y′ = SO 2 Ph 2.2 (-0.3) > CN 1.9 (-1.2).This order differs from the order of σ R -values of these groups. 21(e) The values at B3LYP/6-31G** are consistently more negative than at B3LYP/6-31+G*.(f) For the N-i-Pr derivatives a SO 2 Ph group gives a 0.8 pK Enol units lower values than an SO 2 Me group, both for two SO 2 R groups or CN, SO 2 R combinations.However, for the N-Ph group the trend for the two SO 2 R groups is inverted by 0.8 units.Interestingly, when both SO 2 Me and SO 2 Ph are in the same compound the pK Enol values for the both N-Ph and N-i-Pr derivatives are significantly lower by 2.0 and 1.3 kcal/mol than when for two identical SO 2 R groups.The σ R -values quoted above suggest that order of electron withdrawal is SO 2 Ph ≥ SO 2 Me. 8 (g) The order of EWGs according to σ R -values, i.e., MeSO 2 > CO 2 Me > CN > PhSO 2 20 is not reflected in the order of the calculated pK Enol values, assuming additivity of substituent effects.By calculating the difference ∆pK Enol value for two groups, based on the pK Enol of pairs of Y,Y′-substituted systems, and making the extreme assumption of additivity of substituent effects, i.e., either that the effect of identical group in the compared two pairs is cancelled if two systems are compared, or that the ∆pK Enol values should be divided by two if the groups Y or Y′ appear twice in each pair different values were calculated from different pairs.The following ∆pK Enol values for CO In order to look computationally at simpler systems with fewer interactions, the enolizations of Y-substituted N-unsubstituted acetamides H 2 NCOCH 2 Y, Y = SO 2 R (R = Ph, Me, C 4 F 9 ), CO 2 R' (R' = CH 3 , CH 2 CF 3 ) and CN were computed.The results (Table 2, bottom) indicate that the ∆G and pK Enol values are, as expected, significantly higher than for the Y,Y′ -disubstituted systems.The important result is that the values for sulfonyl and CN groups are much higher than for the ester groups, although among the SO 2 R groups the pK Enol is lower for the much more EW C 4 F 9 than for Ph and Me.The differences are mainly due to the ∆H term, although the ∆S term for the esters is a few e.u.more negative than for the SO 2 R. Consequently, even in the absence of mutual interactions between Y and Y′ and between Y and the N-substituent, the main experimental conclusion that an SO 2 R group is a less enolization promoter than a CO 2 R group remains valid.Since the enols are planar, and the pK Enol values do not follow the σ R -values we conclude that the effect is connected with the amide, which is apparently more stabilized for the SO 2 R-substituted amides than for the CO 2 R-substituted amides.We believe that this (explanation c) holds also for the Y,Y′ disubstituted systems.

Calculated geometries of the enols
The calculated geometries of few of the enols are given in Figures 2 and 3.In Figure 2 few calculated and observed bond lengths and angles are compared for the amide 5c and the enol on the acetyl group 7o for which X-ray data are available.The crystallographic parameters are mostly similar, especially for the amide 5c, except for the hydrogen bond parameters of 7o, where we interpret the observed structure as resulting from a dynamic equilibrium between enols 6o and 7o, whereas the calculated structure represents the static most stable structure.In Figure 3 the calculated structures of few enols on the amide carbonyl are shown.Additional structures are given in the Supporting Information.The important conclusion is that the NHR, OH, Y, Y′ and C=C bond of the enolic moiety are all in the same plane, as demonstrated by a side view of each of the enols.This excludes suggestion (b) above that the low % of enolization is due to twisting of the β-Y,Y′ substituents from planarity.Consequently, although the full negative charge delocalizing ability of these substituents is operating to stabilize the enols, this is insufficient to observe a significant percentage of the enols.
Suggestion (c) is therefore the remaining explanation.To investigate it we need to dissect the total effect of the substituents on K Enol to the separate effects on the amide and the enol.This was performed by using the bond separation isodesmic equations, in which the effect of substituents on the total amide/enol equilibria is dissected to the effect of the substituent on the stabilization of the amide (Eq.5) and the enol (Eq.6) in comparison with the parent system.In these hypothetical isodesmic equations the Y,Y′ groups are no longer conjugated with the substituents on Cα(NHR)OH.Eq. 7 (the difference of eqs. 5 and 6) gives the ∆G for the difference between ∆G for the Y,Y′ substituted system and the parent N-substituted acetamide and its enol.Table 3 display the results of equations 5 and 6 using ethylene as the "deconjugating" reagent, at both B3LYP/6-31+G* and B3LYP/6-31G**.Similar calculations when CH 4 is used instead of H 2 C=CH 2 are given in the Supporting Information, while the energies of the parent reactions required for comparisons are given in Table 4.

Experimental Section
General.Melting points are uncorrected. 1H and 13 C NMR spectra were recorded as described previously. 22Precursors for synthesis, solvents and deuterated solvents for NMR measurements were purchased from a commercial supplier and used without further purification.
Calculations.The geometries were fully optimized a the B3LYP/6-31+G* and B3LYP/613G** levels of theory, with normal convergence using the Gaussian 03 program, 23 Vibrational normal mode analyses were performed at the same level to ensure that each optimized structure was a true minimum on the potential energy surface, no imaginary frequency, and to calculate the thermal correction needed to obtain the Gibbs free energies.H, G and S values obtained at 298.25 K are given in the Supplementary information along with Cartesian coordinates of the optimized structures at respective levels of theory.Chemicals.5a-j/6a-j were prepared by the reaction of the active methylene compounds with sodium followed by reaction with the organic isocyanate.The procedure of the preparation of 5c/6c is representative of that for all derivatives.Sodium pieces (0.12 g, 5 mmol) were added to a solution of methyl phenylsulfonylacetate (1.07 g, 5 mmol) in dry THF (20 mL) and the mixture was stirred overnight.The colorless precipitate was dissolved on addition of isopropyl isocyanate (0.5 mL, 5 mmol) and the mixture was heated at reflux for 2 h.The solvent was evaporated giving the yellow solid sodium salt, which was dissolved in DMF (5 mL) and the solution was poured into ice-cooled 2N HCl solution (50 mL).

9 Figure 3 .
Figure 3. Calculated structure of several enols with different Y and Y′ groups.Planarity is shown by the side view on the right hand side of the structures.

Table 1 . Composition of 5a-n/6a-n and 5o/6o/7o in
12itself and may have prevented the observation of the OH signals of other enols.Small signals which appear in the region of the enol NH signals of other systems were occasionally observed in these cases, but their assignment is only tentative.Consequently, it is difficult to estimate the precision of the K Enol values, except that the values are low.The fact that the % of enol 6a is higher in CCl 4 than in CDCl 3 as was observed with other enols12increases the reliability of the assignment.The data are given in Table1and spectra are shown in the Supporting Information.several solvents at room temperature a

6o/7o system
displays in the 1 H NMR spectrum in CDCl 3 two low field 1:1 signals at 19.08 and 9.80 ppm ascribed to OH and NH signals, respectively, of the same tautomer, as well as Me and Ph signals.The 13 C NMR spectra displayed two low field signals at 198.3 ppm (t, J 6 Hz) and 168 ppm ascribed to C α of the enol on the COMe group and to the amide CO, respectively.The very low δ(OH) value and the similar δ values to those of the enol PhNHCOC(CO 2 Et)=C(OH)Me on the acetyl group 2b argue strongly that the species is the enol 7 (see below).The full 1 H and

DFT calculations of K Enol values
19, CO 2 R, NO 2 and CN groups can be compared by using substituent parameters, especially σ R -values which measure negative charge delocalizing ability.Slightly differing values are available in the literatue, and our values are taken from a recent compilation.19ForCO 2 Me, CO 2 Et, CN, NO 2 , MeSO 2 and PhSO 2 the σ p values are 0.MeSO 2 is a better resonatively negative charge delocalizing than CO 2 R or CN, which is only exceeded by that of a NO 2 group.A PhSO 2 group is less EWG than a MeSO 2 .Based on this argument alone, the MeSO 2 group should give higher K Enol values than corresponding systems with CO 2 R EWGs.The calculated thermodynamic parameters and pK Enol values for all our systems, a few others, as well as several values for diesters, a cyano ester and dicyano substituted systems, and systems activated by only the single groups SO 2 R (R = Me, Ph, C 4 F 9 ), CO 2 R (R = Me, CH 2 CF 3 ) and CN, as well as R'NHCOCH 3 systems [R = i-Pr, t-Bu, Ph and 2,4-(MeO) 2 C 6 H 3 ] at both B3LYP/6-31+G* and B3LYP/6-31G** are given in Table2.Earlier calculated ∆H, ∆G and pK Enol values for (MeO 2 C) 2 CHCONHPh at B3LYP/6-31G** are respectively -5.7 and -2.7 kcal/mol and 1.98.1 observed ratiosThe problem of quantitative comparison is that for most of the sulfonyl-substituted systems the K Enol values are ≤ 0.005.Consequently, for observable sulfonyl-substituted enols in CDCl 3 , a CO 2 Me group exceeds enol-promotion ability than SO 2 Ph or CN as shown by the following ratios K Enol [PhNHCOCH(CO 2 Me) 2 ]/K Enol [PhNHCOCH(CO 2 Me)SO 2 Ph] = 17.4,and for the N-i-Pr analogue > 5.For K Enol [i-PrNHCOCH(CN)CO 2 Me]/K Enol [i-PrNHCOCH(CN)SO 2 Ph] = > 9000 and two methoxycarbonyl groups are better than two RSO 2 groups: K Enol [PhNHCOCH(CO 2 Me) 2 ]/K Enol [PhNHCOCH(SO 2 Me)SO 2 Ph] = 88.It is clear that it is difficult to observe trends with the few available accurate values.A more extensive comparison will be achieved by calculating many more K Enol values by the DFT method.

Table 2 (continued)
a Hydrogen-bonding with an oxygen atom of the SO 2 Me group.b Hydrogen-bonding with an oxygen atom of the SO 2 Ph group.c Untable structure, converged to the enol on the amide.d 2 Me -SO 2 Ph are obtained based on the following two pairs of Y,Y′ groups: -1.83 (2 CO 2 Me -2SO 2 Ph), -1.33 (2 CO 2 Me -CO 2 Me, SO 2 Ph), -0.44 (CO 2 Me, CN -SO 2 Ph, CN), -2.33 (CO 2 Me, SO 2 Ph -2SO 2 Ph), for the N-i-Pr derivative and -4.90 (CO 2 Me, CN -SO 2 Ph, CN) for the N-Ph derivative and -3.1 (CO 2 Me, SO 2 Ph -2SO 2 Ph) for the N-2,4-(MeO) 2 C 6 H 3 derivative.The CO 2 Me -SO 2 Me values are -2.33 (2CO 2 Me -2SO 2 Me), -3.37 (CO 2 Me, CN -SO 2 Me, CN), for the N-i-Pr compounds and -5.09 (CN, CO 2 Me -CN, SO 2 Me) for the N-Ph compound.The CN -SO 2 Ph values are -0.58(2CN -2SO 2 Ph) and -1.68 for (CN, CO 2 Me -SO 2 Ph, CO 2 Me), for the N-Ph derivative, and CN -SO 2 Me value of -0.98 (2CN -2MeSO 2 ) for the N-i-Pr derivative.For CO 2 Me -CN, the pair 2CO 2 Me -2CN gives -1.25 for the N-i-Pr derivative.The crude ability of the groups to promote enolization obtained from these values is therefore CO 2 Me > CN > SO 2 Ph > ?SO 2 Me.(h) Finally, the difference between the calculated gas phase K Enol values and the observed values in CCl 4 is not large: For 5b/6b and 5c/6c the experimental ∆G values are 1.9 and 2.0 kcal/mol, compared with the respective calculated values in Table 2 of 1.8 and 4.6, respectively.

Table 4 .
Energy Difference (kcal/mol) between Enol and Amide Calculated atThe following conclusions arise from Table3: (a) All the ∆G (and ∆H) values for the reaction of the amides (Eq.5) are negative, indicating that the overall interaction between the αand β-substituents is destabilizing and they prefer to be in different molecules.(b) In contrast, all the reactions of the enols (Eq.6) give positive ∆G and ∆H values, whose values are much larger than those of Eq. 5).(c) The values based on the B3LYP/6-31+G* basis set are less positive than those based calculated at B3LYP/6-31G**.(d) The differences are substituent dependent; for Eq. 5, the order of destabilization for N-i-Pr is CO 2 Me, CO 2 Me ,SO 2 Me, COMe, COMe,COMe < CO 2 Me, SO 2 Ph ,SO 2 C ,COMe, CO 2 Me,CN < CO 2 Me, COMe < SO 2 Me,CN < CO 2 Me,CN < SO 2 Me, SO 2 Ph < SO 2 Ph,CN < SO 2 Ph,SO 2 Ph.With Y′ = COMe, Y= SO 2 C 4 F 9 gives a more negative values than SO 2 Ph.For N-Ph the order is: COMe,COMe < CO 2 Me,CO 2 Me < CO 2 Me,CN < SO 2 C 4 F 9 ,COMe < SO 2 Ph, CO 2 Me < SO 2 Ph,SO 2 Ph < SO 2 Ph,CO 2 Me < COMe,CN < CO 2 Me,COMe.(e) The ∆G values for enols (Eq.6) are consistently higher for CO 2 Meactivated systems than by sulfonyl systems.The order of ∆G values when R' = i-Pr is: CO 2 Me, COMe > CN, COMe > COMe, COMe > CO 2 Me, CO 2 Me > CO 2 Me, CN > SO 2 Me, COMe > CO 2 Me, SO 2 Ph > SO 2 Me, CN > SO 2 Ph, CN > SO 2 Me, SO 2 Ph > SO 2 Ph, SO 2 Ph >SO 2 C 4 F 9 , COMe.For the N-Ph derivatives the order is: CO 2 Me, COMe > COMe, CN > COMe,COMe >SO 2 C 4 F 9 , COMe > CO 2 Me, CO 2 Me > CO 2 Me, SO 2 Ph > CO 2 Me, CN > SO 2 Ph, SO 2 Ph.