170 NMR data on a series of hindered 3-substituted phthalic anhydrides and correspond
ing phthalides have been shown30•31 to correlate with in-plane bond angle distortions. The nontorsional 170 chemical shift effects observed for the anhydrides, thought to be indicative of repulsive van der Waals interactions, could be used to rationalize the regiospecificity of reduction reactions. Sterically hindered imide systems also show regiospecificity35 in re
ductions and, thus, should be investigated by 170 NMR techniques. 170 NMR data for a series of N-substituted phthalimides 29a-d and a series of N-substituted succinimides 30a
d and maleimides 31a-d presented below show36 that the 170 NMR chemical shift data can provide additional insights into structure and reactivity in relation to steric phenomena.
170 NMR data were obtained36 (natural abundance) for a series of N-substituted phthal
imides (29a-g), in acetonitrile at 75°C.
29a-g
The data for the phthalimides are summarized in Table 6. With the exception of the parent compound 29a, the signal for the carbonyl oxygens was deshielded as the size of the N
substituent increased, despite similar electronic effects (shielding) of the alkyl groups. For example, the substitution of anN-isopropyl group for anN-methyl group yielded a 9 ppm downfield shift while the similar effect of the N-t-butyl group was 20 ppm (compare 29b to d). In the unsymmetrical compound 29f separate signals for both carbonyl oxygens were detected. Since electronic effects are again negligible, the large difference (deshielding) observed was indicative of significant repulsive van der Waals interactions. The magnitude of this shift (28 ppm) suggested in-plane distortions caused by the partial relief of the steric interactions of the substituent on the aromatic ring with the carbonyl oxygen similar to that observed30 in the analogous phthalic anhydride system ( -25 ppm). The results for the doubly hindered compound 29g showed that the ring substituent deshielding effect and that due to theN-substituent were additive (vide infra).
The large deshielding effects noted in 29b to d were surprising. The 170 chemical shift data for N-substituted succinimides (30a-d) and maleimides (31a-d) were examined36 to test the generality of this finding. The 170 NMR chemical shift data for the N-substituted succinimides and maleimides showed deshielding effects of similar magnitude to those for the phthalimides. In addition, 170 NMR data for the analogous N-substituted phthalamides (32a-c) showed deshielding effects with large N-substituents in agreement with the above three imide systems. The data are summarized in Table 7. The signals for compounds with
TABLE 6
170 Chemical Shift Data36 ( ± 1 ppm) for Substituted Phthalimides 29 in
Acetonitrile at 75°C
Compound l) l)
no. R. R, (C=O). (C=O)z
29a H H 379.0 379.0
29b H Me 374.0 374.0
29c H i-Pr 383.0 383.0
29d H t-Bu 394.0 394.0
29e H Ph 378.3 378.3
29f t-Bu H 407.3 370.6
29g t-Bu t-Bu 423.3 385.3
TABLE 7
170 NMR Chemical Shift Data ( ± 1 ppm) for N-Substituted Malimides, Succinimides, and Phthalamides in Acetonitrile at 75°C36
Succinimides Malimides Ph thalami des
Compound l) Compound l) Compound
no. N-R (C=O) no. N-R (C=O) no. N-R
30a H 373.5 31a H 411 32a H
30b Me 371 31b Me 407 32b Me
30c t-Bu 392 31c t-Bu 426 32c t-Bu
30d Ph 376 3ld Ph 412
l)
(C=O) 282 281 300
N-t-butyl groups were deshielded 20 ± 1 ppm relative to those for theN-methyl compounds for all four cases. The signals for theN-phenyl compounds were deshielded by 5 ± 1 ppm relative to those for theN-methyl compounds. The chemical shift data for the parent com
pounds (R =H) of each group (29a, 30a, 31a, 32a) are complicated by the presence of a hydrogen-bonding component. Simple hydrogen bonding to a carbonyl group should clearly result in an upfield shift of the 170 signal. 4•ã20•37•38 The effect of N-H donation to another system on the carbonyl of the donor imide is not as clear. 38 •39 The overall effect of differential hydrogen bonding is complex.38 Thus, the chemical shift differences between these com
pounds and those for the N-substituted compounds were difficult to interpret.
30a-d
c$-0 31a-d
32a-c
86 170 NMR Spectroscopy in Organic Chemistry
Molecular mechanics (MM2) calculations33 were carried oue6 on 29a to d; see Table 8 for selected bond angles (entries 1 to 9). The calculations predicted in-plane angle distortions of the planar molecules with the larger N-substituents. The bond angle of the carbonyl group opened toward the ring (entries 1 and 2), and the internal bond angle of the imide (entry 4) diminished as theN-substituent increased in size. In contrast to the 3-substituted anhydride system, the distortion for the N-substituted imides resulted in a symmetrical opening of the carbonyl angles (entries 1 and 2, Table 8). The X-ray structures of phthalimide40 29a and N-t-butylphthalimide36 29d have been reported. Both structures were found to be planar in agreement with molecular mechanics calculations. Unfortunately, the X-ray results showed that the crystal structures for 29a and 29d were not symmetrical around an axis through the nitrogen bisecting the molecule, making quantitative comparisons with the molecular me
chanics calculations difficult. Despite this, the bond angle distortions noted were in qualitative agreement with those predicted by the calculations (Table 8).
TABLE 8
Molecular Mechanics Calculated Bond Angles(± 1°) and X-ray Data for Phthalimides 29a, d, g36
29a 29d 29g
Entry Angle MM2 (X-ray)• MM2 (X-ray)b MM2 (X-ray)•
1 O,C2N3 124° (124.8°) 126° (126.3°) 124° (125.5°) 2 CllCzN, 105° (105.2~ 107° (108.6°) 108° (106.8°) 3 C10CuC2 131° (130.00) 131° (131.9°) 133° (132.8°) 4 C2N,C. ll4° (112.2°) ll1° (ll0.6°) ll1° (110.5") 5 N,c.o, 124° (125.4°) 126° (129.6°) 126° (126.6°) 6 N,c.c6 105° (106.2°) 107° (104.6°) 107° (107.4°)
7 c.c6c7 131° (120.3°) 131° (125.9°) 128° (127 .2°)
8 C2N3Rz 122° 122° (122.1 °) 122° (122.6°)
9 CuC,oR, 126° (125.2°)
a Values taken from Reference 40.
b Sigma values = 1.5.
c Sigma values = 0.5.
The 170 signal for the double-hindered carbonyl of 29g was deshielded by 50 ppm, which was much larger than any other effects seen. Moreover, the magnitudes of the deshielding effects on the carbonyl signals were consistent with expectations based upon the 170 data for 29a, 28b and 29e. The X-ray structure of 29d was obtained36 (Figure 6) for comparison with molecular mechanics calculations (Table 8) and again confirmed the molecule to be planar. The calculations predicted essentially identical values for the angles for (C = 0)2 in both structures (29d and 29g), and the X-ray results were qualitatively in agreement; compare entries 1 and 5 for compounds 29a and 29g. The doubly hindered carbonyl, (C = 0)1, is being influenced in opposing directions by the two t-butyl groups
FIGURE 6. X-ray structure of N-t-butyl-3-t-butylphthalimide (29g). See Table 8 for data on selected bond angles.
such that no unusual distortion is apparent in the structure. However, the 170 results showed that this carbonyl, (C=O),, was subject to severe van der Waals interactions.
Interestingly, the 13C NMR signals of sterically hindered imides were found36 to be extremely insensitive to compression effects. For example, the two carbonyl signals for 29g were within 0.5 ppm of one another, whereas the 170 NMR data for the double-bonded oxygens attached to these carbons differed by 50 ppm.
The 170 NMR results for imide systems gave interesting new insights into ground state structure. In certain systems (cf. 29g) the 170 method provided detailed information not accessible by other methods. The results for planar amides (32a-c) show that theN-substituent deshielding effects are not limited to imides. However, it is not clear that these effects will be predominant in conformationally mobile (acyclic) systems. The 170 NMR results may provide insights into the regiospecificity of imide reductions. 35 For reductions in which electron transfers are rate determining, one would expect the carbonyl which shows the most deshielded 170 chemical shift value to undergo reaction preferentially. For example, the regiospecific zinc reduction of a hindered imide41 is consistent with the above expectation.