The significant deshielding effects on the 170 chemical shift data for sterically hindered heteroaromatic N-oxides, 17•21 where torsional effects are limited, are useful for identifying certain types of isomeric compounds. However, the identification of the origin of the steric factors that affect 170 NMR data is unclear at present but should be understood to significantly increase insight into molecular structure by use of no NMR methodology. To begin to sort out nontorsional, steric factors, it was essential to study a system which had a less polar functional group with a well-defined, planar geometry and which showed a large chemical shift range. Certain cyclic anhydrides seemed to satisfy these criteria, 28 and, in addition, in some cases the two carbonyls had been reported to exhibit differential reactivity. 29 The phthalic anhydride system was chosen for study since it appeared that large steric interactions were possible. Indeed, the 170 NMR data for a series of 3-substituted phthalic anhydrides (26) and corresponding phthalides (27 to 28) clearly showed30•31 that the chemical shifts are sensitive to nontorsional bond angle deformations.
78 170 NMR Spectroscopy in Organic Chemistry
~ 0 ~ &! 0
26a-j 27a-i 28a-g
The no NMR data (natural abundance) for phthalic anhydride (26a), 3-methyl-phthalic anhydride (26b), 3-t-butylphthalic anhydride (26c), phthalide (27a), 7-methyl-phthalide (27b), 7-t-butylphthalide (27c), and 4-t-butylphthalide (28a) were obtained30 at 75°C in acetonitrile to check the feasibility of the approach and to assign the chemical shift data. The results are summarized in Table 2.9 •10 For both 26b and 26c, two well-defined no signals for the sterically different carbonyl groups, separated by II and 29 ppm, respectively, were ob
served. The single bond oxygen for all the anhydrides appeared at 263 ± I ppm. A representative spectrum of 3-methylphthalic anhydride is shown in Figure 3. The 170 data for the two isomeric t-butyl phthalides (27c, 28a) allowed the assignment of the downfield (deshielded) carbonyl signal in the substituted anhydrides to the carbonyl adjacent to the substituent. Note that the "carbonyl" signal for 27c is downfield of that of 28a by 27 ppm.
The downfield shifts noted for 26b, 26c, 27b, and 27c cannot be due to electronic effects.
Previous work on the benzoate system had shown that the electronic effects of alkyl groups on the carbonyl resonance were modestly shielding (2 ppm).32a Hence, the chemical shift differences between 26a and its methyl and t-butyl analogs (26b and 26c) would be roughly 2 ppm larger if a correction for electronic effects had been made. The deshielding effect in both series (26b, 27b) for an ortho-methyl group was 9 to I2 ppm while a similarly located t-butyl group produced a 22 to 27 ppm shift. Interestingly, the 13C NMR chemical shifts of the two carbonyl carbons in 26b, c were essentially identical (within 1 ppm). Thus, the no
NMR methodology provided new insights into the effect of steric hindrance on the carbonyl function.
TABLE 2
170 NMR Data ( ± 1 ppm) for 3-Substituted Phthalic Anhydrides 26a-c and Phthalides (27a-c, 28a) in Acetonitrile at 75°30
Anhydrides Lactones
Compound 0 0 Compound 0 0
R no. (C=O) (-Q-) R no. (C=O) (-Q-)
H 26a 374 263 H 27a 320 170
Me 26b 383 264 7-Me 27b 332 170
372
t-Bu 26c 396 262 7-t-Bu 27c 346 168
367
4-t-Bu 28a 319 173
As a check of the 170 chemical shift assignments in series 26, qualitative shift reagent studies were carried oue2b on 26a and 26c in CDC13 • The results for the carbonyl signals for 26c showed that the deshielded signal was relatively insensitive to the shift reagent [Eu(FOD)3). 5cã32c The other carbonyl signal showed a sensitivity similar to that observed with the parent anhydride 26a. Thus, the results of the shift reagent study were consistent with the chemical shift assignments. The results are summarized in Figure 4.
In an effort to gain insight into the influence of substituents of varying electronic character on the properties and reactivities of the two carbonyl functions of 3-substituted phthalic
2-Butanone (INTERNAL STANDARD)
a b
c
26b
550 500 450 400 350 300 250 200 17o CHEMICAL SHIFT (PPM)
150
FIGURE 3. 170 NMR spectrum of 3-methylphthalic anhydride (26b) in acetonitrile at 75°.
0
anhydrides, the 170 chemical shifts of additional series of phthalic anhydrides (26d to j) and related phthalides (27d to i and 28b to g) were studied. 31 170 chemical shift data for the 3- substituted phthalic anhydrides (26b to k), 7-substituted phthalides (27d to i) and 4-substi
tuted phthalides (28b to g) in acetonitrile at 75°C, measured at natural abundance, are given in Table 3 and Table 4. All the substituted anhydrides showed two carbonyl signals in the 370 to 395 ppm region and one signal near 263 ppm.
The assignment of the two carbonyl signals for the remaining anhydrides of series 26 was made by using the results from the corresponding phthalides 27 and 28 (Tables 2 to 4) as discussed above. 30 It is clear from the data for the phthalide series 27 that the effect of a substituent ortho to the planar carbonyl was deshielding, regardless of the electronic character of the substituent. Consequently, the assignments for the two carbonyl signals of the anhydrides shown in Tables 2 to 3 are made by analogy with the corresponding phthalides.
With the exception of the 3-t-butyl and the 3-nitro compounds in series 26, the effect of substituents was a uniform deshielding value of 9 ± 1 ppm. This, in part, is presumably a result of repulsive van der Waals interactions.12 The single bonded oxygen signal for all 3- substituted anhydrides, 26, was insensitive to substituents with all chemical shift values at 263 ± 1 ppm. The 170 NMR signal for the carbonyl group meta to the substituent was relatively insensitive to the 3-substituent ranging in value from 367 to 377 ppm. The chemical shift of the carbonyl ortho to the substituent for the 3-methoxy compound 26e was deshielded
80 170 NMR Spectroscopy in Organic Chemistry
400 • ã-ã-
390 380 &>ã
"
0 A
("")
~~
:X: ,..., 3 370
• O=C 26a
("") ):>
'
(/) -~. ã~
:X:
~
""T1
360
-1
" ~ .,
"
3:
~
350 ~ .,
340
0.1 0.2 0.3 0.4 0.5
[Eu(FOD) 3]/[Anhydride]
FIGURE 4. Shift reagent studies on phthalic anhydride (26a) and 3-t-bu
tylphthalic anhydride (26c) in CDCI3 at 55°C.
TABLE 3
170 NMR Data ( ± 1 ppm) for 3-Substituted Phthalic Anhydrides 26d-j in Acetonitrile at 75°C31
Compound I) I) I) I)
no. R (C=O), (C=0)2 (-Q-) (R)
26d 3-C2H50 370 384 264 91
26e 3-CH30 371 383 264 65
26f 3-F 376 385 264
26g 3-Cl 375 386 264
26h 3-Br 374 385 265
26i 3-1 373 382 263
26j 3-N02 377 395 262
relative to the parent 26g by 9 ppm, and the shift data for the 3-nitro compound 26j was deshielded by 22 ppm relative to the parent. Thus, the carbonyl groups ortho to the 3- substituents were deshielded relative to the parent compound by all substituents regardless of the electronic character of the groups. The deshielding effect seemed to be the result of steric interactions of the 3-substituent and the lone pairs on the carbonyl group.
The chemical shift data for the carbonyl group of the 7-substituted phthalides (27) were
TABLE 4
170 NMR Data ( ± 1 ppm) for 7- and 4-Substituted Phthalides (27d-i, 28b-g) in Acetonitrile at 75°C31
7 -Substituted
Compound 8
R no. (C=O)
8 (-Q-)
0 (R)
4-Substituted
Compound 8
no. (C=O)
8 (-Q-)
0 (R)
CH,O 27d 333 168 59 28b 323 170 49
F 27e 334 171 28c 325 169
Cl 27f 335 170 28d 327 170
Br 27g 334 170 28e 327 171
I 27h 330 171 28f 327 171
N02 27i 337 172 616 28g 325 174 582
all deshielded by 10 to 26 ppm compared to that of the parent molecule 26a, which exhibited a carbonyl signal at 320 ppm. In series 27, electron-donating or electron-withdrawing sub
stituents resulted in deshielding as observed in series 26. The resonances for the lactone ( dicoordinate) oxygen of series 27 were essentially invariant, all appearing at 170 ± 2 ppm.
The chemical shift values for the carbonyl oxygen for series 28 were only slightly affected by substituent, appearing at 323 ± 4 ppm, as were those of the single bonded(-(}-) oxygen which appeared at 171 ± 3 ppm and could be explained as normal electronic effects.
The molecules in series 26 and 27 were expected to be planar; however, since the magnitude and direction of these shift differences in series 26 and 27 were similar to those attributed to torsional variations,6•9 this possibility was evaluated and could be ruled out.
Results27 for sterically hindered ketones showed that a torsional rotation of the carbonyl of at least 20° would be required to yield the deshielding of 25 ppm observed for the chemical shifts in 26c and 27c. Molecular mechanics calculations33 (MM2) for 26 to 28 predicted30 that all the ring systems were planar. In addition, the calculations indicated that no significant changes in bond lengths were anticipated within each series of compounds. However, the calculations did predict substantial in-plane distortions of the bond angles in both rings in close proximity to the R group at the juncture of the two rings. Representative results for the alkyl-substituted anhydrides and phthalides are shown in Table 5. To obtain independent corroboration of the predicted geometry, the crystal structure (Figure 5) of the most distorted anhydride, 3-t-butylphthalic anhydride, 26c, was obtained. Analysis of a single crystal showed that all the atoms for 26c with the exception of t-buty 1 methyl groups, were completely planar.
The trends discernible in Table 5 showed that for compounds with the larger R groups, the angles represented, particularly by entries 1, 2, 3 and 4, were larger and those represented by entries 5, 7 and 13 were smaller. The carbonyl group peri-like to the R-group showed significant bond angle variation with R-group size, whereas the relatively unhindered car
bonyl group was essentially unaffected. Presumably the in-plane distortions observed reflect (partial) minimization of van der Waals interactions. The trends noted in Table 5 paralleled the carbonyl 170 chemical shift data. The predicted distortions for the phthalides 27a to 27c also paralleled those observed for 26a to 26c, whereas little distortion of the carbonyl angles for 28a was estimated. The carbonyl 170 data for the phthalides reflected this trend (Tables 2 to 4). For all five compounds no distortion of the bond angle involving the single bonded (-(}-)oxygen (entry 6) was predicted by the calculations; the 170 data for 09 is essentially constant in each series (Tables 2 to 4). Comparison of the MM2 data and the X-ray results for 26c showed they were in reasonable agreement. The two methods were in poorest agreement for the angles represented by entries 3, 6 and 9; MM2 calculations underestimated the bond angles although the trends were consistent with 170 data.
It is of interest to note that there is no appreciable difference in the data for the compounds
82 170 NMR Spectroscopy in Organic Chemistry TABLE 5
Calculated (MM2) Bond Angles for Anhydrides 26a-c and Phthalides 27a-c, 28a30
Oa
Angle degrees Angle degrees
Entry Type 26a 26b 26c 27a 27b 27c
1 C2C3R 121 122 125 (124.7)' 121 122 125
2 C,CzCw 129 130 133 (132.9) 128 129 131
3 CzCwOII 126 126 128 (133.3) 125 126 128
4 C2C100 9 109 llO Ill (107.4) Ill 112 ll2
5 OIICI009 123 123 120 (119.3) 123 122 ll9
6 C709CIO 103 103 103 (II 1:3) 107 107 107
7 c,czc10 108 107 105 (105. 7) 108 107 105
8 C2C,C7 108 108 109 (109.6) 109 110 Ill
9 c,c1os 126 126 126 (132.0)
10 C,C709 109 109 109 (107.0) 102 102 101
1l o.c7o9 123 123 123 (121.0)
12 c2o,c. ll6 ll5 113 (113.7) 116 ll5 113
13 c6c,c7 129 128 126 (126.8) 130 129 125
14 C,C6R" 120 120 121
. X-ray data .
28a 121 126 125
Ill
123 107 107 109
103 ll6 131 123
26f to 26i even though the substituents (halogen) appear quite different sterically. Molecular mechanics calculations (MM2) carried out on this series (26f to 26i) showed31 similarities in plane distorations to those found for 26b. Apparently, the increase in halogen size is offset by the lengthening halogen-carbon bond distance resulting in the effective steric interactions remaining roughly constant. A possible explanation for the relatively invariant deshielding effects noted for 170 chemical shifts for the anhydrides is that the repulsive van der Waals interactions between substituents and the ortho carbonyl group are roughly con
stant. The two substituents which lead to the large downfield shifts, 3-t-butyl 26c (22 ppm) and 3-nitro 26j (21 ppm), exhibit opposite electronic effects - the former a moderate electron donor, the latter strongly electron withdrawing. In the case of 26c, the large effective size caused larger in-plane distortions and gave rise to increased repulsive van der Waals interactions. 12 The large downfield shift for 26j was a composite of the electronic deshielding effect of the nitro group as well as the influence of repulsive van der Waals forces.
Due to steric congestion, the nitro group of 3-nitrophthalic anhydride 26j is forced to rotate from the plane, and a reciprocal effect from this interaction is noted on the nitro group
170 signal. On comparing the 170 shifts of the nitro signal for 3- and 4-nitrophthalic anhydride, some information regarding torsion angle rotation for the 3-nitro group can be deduced.
Note that the nitro groups of both isomers are electronically equivalent. Thus, the difference in chemical shifts (N02 group) for the two isomers (27 ppm) can be attributed to torsion angle rotation, which minimizes the lone pair repulsion of the carbonyl oxygen. Using the slope (0. 76 &/angle0 ) for the nitro group 170 chemical shift-torsion angle relationship pre
viously reported, 6 •9 the 3-nitro group can be estimated to be rotated 35° from the plane of the aromatic ring. This is in contrast to the steric effect on the carbonyl which appears to be only in-plane distortions in nature.
FIGURE 5. X-ray structure of 3-t-butylphthalic anhydride (26c).
See Table 5 for data on selected bond angles.
The effect of substituents on the 170 chemical shifts for the 3-substituted phthalic anhydrides (26) and ?-substituted phthalides (27) were similar, both experiencing a com
parable magnitude of deshielding in relationship to the respective parent molecules. 30•31 The magnitude of the deshielding effect for series 27 was surprisingly constant for all the groups except the large t-butyl group and to a lesser extent the nitro group. The relative consistency in chemical shift for the substituents of varying size and electronic effects can be explained in terms of competition between simple electronic effects and van der Waals interactions (in-plane). Series 28 has a meta relationship between the carbonyl group and the substituent.
(The small 170 chemical shift differences can be explained by normal electronic effects.) An interesting conclusion to be made from these results for all the 3-substituted compounds listed in Table 2 to 3 is that the 170 chemical shift values suggest that reactivity difference29•34 for the two carbonyl groups of these compounds cannot be explained by simple electronic effects alone.
Differential reactivity of carbonyl functions in similar systems has been reported, and explanations for the differences have included steric blocking of the attacking reagents as well as the influence of electronic effects of the substituents. 29•34 The present study clearly
84 170 NMR Spectroscopy in Organic Chemistry
showed that steric interactions in these systems were not limited to steric blocking but also included molecular distortions, which should affect the reactivity of the carbonyl functions.
An interpretation of the 170 data for the anhydrides suggests greater double bond character for the hindered carbonyl which should lower its reduction potential, whereas the other carbonyl oxygen shows greater charge density and, thus, should be more likely to form a complex with Lewis acids. This combination of effects must be considered in explanations of reactivity data. Furthermore, these results suggested that the 170 NMR data reported17 •21
for sterically hindered N-oxides may involve a similar deshielding mechanism indicative of analogous deformations.