ChemInform Abstract: Isotope Effects in NMR Spectra as a Structural Tool for Organic Molecules

Isotope Effects in NMR Spectra as a Structural Tool for Organic Molecules

Predrag Novak*, Dražen Vikić-Topić, Vilko Smrečki and Zlatko Meić*

Department of Chemistry, Ruđer Bošković Institute, PO Box 1016, 10001 Zagreb, Croatia

* corresponding authors

II

Table of contents

Abstract

III

Abbreviations and symbols

IV

1. Introduction

1

2. Classification and sign convention

3

2.1. Nuclear shielding

3

2.2. Spin-spin coupling

4

3. Theory of isotope effects

5

4. Isotope effects on nuclear shielding

6

4.1. Deuterium isotope effects

6 13

4.1.1. Deuterium isotope effects on C chemical shifts

6

4.1.2. Deuterium isotope effects on other chemical shifts

11

4.2. Isotope effects of other nuclei

13

4.3. Hydrogen-bonding and tautomeric equilibria

16

5. Isotope effects on spin-spin coupling

18

6. Isotope effects in Ph-R and Ph-Z-Ph systems

21

6.1. Deuterium isotope effects on nuclear shielding

22

6.1.1. Short-range deuterium isotope effects on 13C chemical shifts

22

6.1.2. Long-range deuterium isotope effects on 13C chemical shifts

24

6.1.3. Vibrationally induced -electron polarisation

26

1

6.1.4. Deuterium isotope effects on H chemical shifts

28

6.2. 13C isotope effects on nuclear shielding

29

6.3. 15N isotope effects on nuclear shielding

30

6.4. Additivity of isotope effects on nuclear shielding

31

6.5. Deuterium isotope effects on spin-spin coupling

33

6.5.1. Secondary isotope effects

34

6.5.2. Primary isotope effects

35

6.5.3. Geometrical considerations

35

6.6. Deuterium isotope effects on spin-lattice relaxation

37

7. Conclusions

38

8. References

41

III

Abstract

A survey on recent results and developments in the field of isotope effects in NMR spectroscopy is given in the present article. Isotope effects on nuclear shielding, spin spin coupling and relaxation time are discussed. The established empirical trends and correlation of isotope effects with molecular parameters help in better understanding of this phenomenon and also provide a wealth of information which can be very useful in organic and analytical chemistry. It is shown that isotope effects in NMR spectra can serve as an efficient tool for studying structure and conformation of organic molecules. Special attention is devoted to unsaturated systems containing one or two

long-range isotope effects are operative, providing data not only on structure, but also on mechanism of isotope effect transmission. Thus, the alternation in sign and magnitude of long-range isotope effects is interpreted in terms of subtle charge shifts throughout the molecule as a consequence of the bond shortening upon isotopic substitution, while the linear correlation between long-range isotope effects and molecular torsional angle can serve as a conformational probe for molecules of the related type.

IV

Abbreviations and symbols

D

deuterium (2H)

IE

isotope effect

LRIE

long range isotope effect

SRIE

short range isotope effect

DIE

deuterium isotope effect

p

primary isotope effect

s

secondary isotope effect



magnetogyric ratio

J

nuclear spin-spin coupling constant



chemical shift

T1

spin-lattice relaxation time

1

1. Introduction

Isotopic substitution produces changes in the reactivity of a molecule and causes a redistribution of molecular internal vibrational and rotational energy. A study of isotope effects (IE) is of great practical and theoretical importance in chemistry, physics, biology and medicine, since it gives valuable information on the overall molecular structure and dynamics of molecular processes. Generally, two different types of isotope effects in nuclear magnetic resonance (NMR) spectroscopy have been investigated, i.e. intrinsic and equilibrium isotope effects. The former are an intrinsic property of a given molecule, while the latter have to do with a change in chemical equilibrium upon isotopic substitution. Isotopic labelling in NMR has also a practical use. Thus, selectively labelled sites in organic and biomolecules are employed for a simplification of NMR spectra and for assignment purposes. When an isotope is introduced into a molecule, it causes changes in NMR spectral parameters: chemical shifts (nuclear shieldings), coupling constants and relaxation times. By far most studied are the isotope effects on nuclear shielding because they are easily determined from high resolution NMR spectra. Among the isotope effects in NMR spectra, the most frequently reported in the literature are still those of deuterium (2H isotope, commonly used symbol - D), owing to the relative ease of deuterium incorporation into the molecule and the large fractional change in mass on isotopic substitution (1H to 2H mass change is 100%). Hence, it is not surprising that deuterium

2

isotope effects (DIE) on NMR parameters have a long history of experimental and theoretical investigations. The first review and theoretical account of isotope effects in NMR goes back to 1967 1. The substituent effect approach proposed there gained much interest in explaining the phenomenon of DIE 2, particularly in -electron systems 3. This approach is based on Halevi’s theory of secondary isotope effects 4, which treats the phenomenon in terms of inductive and hyperconjugative contributions of deuterium 5. It is not the aim of the present review to report on isotope effects used only for assignment purposes, nor will isotope effects already reviewed by Hansen 6,7, Forsyth 8, Berger 2, Jameson 9, Siehl 10 and Sergeyev 11 be discussed here in detail. We shall restrict our attention only to those previously reviewed observations relevant to the foregoing discussion. The papers reporting on equilibrium IE will be treated only in the cases where they are observed together with intrinsic IE. Kinetic isotope effects will not be discussed in this article at all. The purpose of the review is to cover recent discoveries and developments in isotope effects in NMR spectra as a structural tool for organic molecules, published in the period from 1991 to the present. Special attention will be paid to -electron systems containing phenyl groups. In such molecules isotopic perturbation can be transmitted many bonds away from the isotopic site, giving rise to long-range isotope effects (LRIE). They are potentially important because they could provide information not only on the structure and conformation, but also on the way of isotope effect

3

transmission throughout the molecule, i.e. the intrinsic nature of isotopic perturbation itself.

2. Classification and Sign Convention

2.1. Nuclear shielding Isotope effects on nuclear shielding (chemical shift) can be primary, p or secondary, s. Primary isotope effects refer to the difference in nuclear shielding of two different isotopes, for instance H (1H) and D (2H). However, they are scarcely reported in the literature 6,7. On the other hand, secondary isotope effects refer to the change in nuclear shielding of a particular nucleus caused by a different isotope nucleus. This kind of isotope effects will be reported in the present article. Secondary isotope effects on nuclear shielding are expressed as follows:

sX = X(l) - X(h)

n

(1)

where X(l) and X(h) are chemical shifts of nucleus X with a light and a heavy isotopomer, respectively, while n denotes the number of bonds separating nucleus X and the isotopic site. Thus, the positive sign refers to a shielding and the negative sign to a deshielding isotope effect. For brevity, instead of ns only n will be used for secondary effects on chemical shifts.

4

In the case of equilibrium isotope effects, the observed effect is usually a superposition of intrinsic and equilibrium contributions to the total value.

2.2. Spin-spin coupling In contrast to IE on nuclear shielding, IE on scalar or spin-spin coupling have been much less investigated. This is due to two main reasons. The first is the requirement for high precision of the measurements, and the second is the complexity of NMR spectral analysis and higher-order effects. In the majority of cases, higher-order perturbations take place, which prevents the direct assignment of coupling constants from NMR spectra. Hence, calculations are required to obtain the real values of coupling constants. Drawbacks and difficulties in determining IE on the spin-spin coupling are given in the review article by Sergeyev 11. Isotope effects on coupling constants can also be primary and secondary. Both have been reported in the literature 11. Primary (2) and secondary (3) IE on nJ(X,Y) can be expressed as follows:

pnJ(X,Y) = nJ(X,Y)h - nJ(X,Y)l = (lY/hY)  nJ(XhY) - nJ(XlY)

(2)

snJ(X,Y) = nJ(X,Y)h - nJ(X,Y)l

(3)

where l and h are magnetogyric ratios for a light and a heavy isotope, respectively, while n denotes the number of intervening bonds.

5

3. Theory of Isotope Effects

A thorough theoretical explanation of isotope effects on nuclear shielding and coupling constants has been given by C. Jameson 9,12,13. It has now been established that changes in NMR parameters caused by isotopic substitution are of a rovibrational origin. Isotope effects can thus be interpreted in terms of dynamic and electronic factors, which govern the magnitude and the sign of effects 9. Both factors depend on changes in the averaged nuclear geometry upon isotopic substitution. Therefore, isotope shifts are the difference between rovibrationally averaged nuclear shieldings in heavy and light isotopomers. High-levelled calculations of isotope effects on both chemical shifts and coupling constants have been successfully applied to small symmetric molecules. In larger and less symmetric molecules, where long-range isotope effects exist, calculations are not yet completely feasible, in spite of the advances in computational facilities. For such effects, the dominant contributing terms are not easily estimated 13. Still, the author claims that the theory applies to larger molecules as well. In the case of larger organic molecules and biomolecules some empirical features of isotope effects, such as their extent, sign, additivity, geometric and conformational dependence, as well as already established correlations between isotope effects and geometrical parameters, such as carbon-hybridisation 14,15, carbon-chemical shifts 16,17, spin-spin coupling constants 18, provide further information and enable a better understanding of this phenomenon.

6

4. Isotope effects on nuclear shielding

4.1. Deuterium isotope effects 4.1.1. Deuterium isotope effects on 13C chemical shifts Among IE on nuclear shielding, the effects of deuterium continue to be most frequently studied. Several papers have recently treated deviations from the additivity rule 19-21. Small but measurable negative deviations of 1-4 ppb (larger calculated values) have been reported in halomethanes 19,20. The negative sign of nonadditivity was in agreement with the idea of the secondary geometrical effect proposed by Jameson and Osten 22. However, an approach has been proposed to explain nonadditivity in halomethanes in terms of quadratic corrections to the secondary geometrical effects 19. Thermodynamic and intrinsic parts of the nonadditivity were distinguished in cyclohexane isotopomers 23. The authors studied small perturbations of conformational equilibrium and concluded that only some IE on chemical shifts could be used for the same purpose in other cycloalkanes. The findings reported in ref. 21 will be discussed later on. A new incremental scheme (4) for the empirical calculations of IE in derivatives of methane was given by Sergeyev et al. 24.

 C(D) = -200 + (Xi)

1 13

(4)

7

where Xi are substituents with increments (Xi), i = 1-3. Relation (4) gave values that were in good agreement with the experimental results. In a series of para-substituted anilines, acetanilides and indoles 25,26 two- and three-bond DIE (D labelled nitrogen), 2 and 3, were correlated with 13C, 15N and 17O chemical shifts. The magnitudes of IE were found to be affected by the nitrogen lone-pair delocalisation and conjugative competitive interaction between the phenyl ring or acetyl group on the one side, and the lone-pair on the other. The 2 effect in para-substituted benzanilides was larger when the electron-attracting character of the substituent was increased 27, which contradicted the results observed in acetanilides 25. The authors tried to rationalise such findings by different contributions of mesomeric structures induced by substituents. Two-bond IE and the proposed correlations in these systems seemed to provide a wealth of information on electron structure and conjugational properties of molecules possessing a nitrogen, which could be perturbed by substituents. One-bond DIE on

13

C shifts were claimed to be correlated with C-H stretching

frequencies, C-H, in IR spectra for a series of monosubstituted methane derivatives for the first time 28. A fair linear correlation was found (5):

1

 (ppb) = 26500 - 8.69 C-H/cm-1

(5)

8

Relation (5) predicts smaller 1 for shorter C-H bonds, which is in agreement with the vibrational model 9. Dependence of 1 on C-H bond length, rC-H (Å), calculated by the semiempirical MNDO method has also been reported in aromatic heterocycles by Nakashima et al. 29. The following relationship was proposed (6):

1

 (ppb) = 8378 rC-H (Å) - 8853

(6)

Large differences in magnitudes of 1 on going from imidazole (191 ppb) to tropylium cation (373 ppb) were claimed to arise from factors other than hybridisation itself, as previously established by Günther and co-workers 14,15. However, the results could not be considered as completely consistent because of the use of a number of different solvents for different compounds and failure to discuss the possible contributions of equilibrium or solvent-induced IE. DIE up to three-bonds were also reported 29. It was concluded that 2 values depend on the nature of the observed sites, while 3 were supposed to be controlled by the heteroatom in the pathway. In their previous paper Nakashima et al. 30 treated the same problem for 3-substituted pyridines. They reported linear correlations between IE over one-, two- and three-bonds and chemical shifts of the observed carbon atom. Mlinarić-Majerski and co-workers studied DIE in rigid adamantane derivatives 31-33. They claimed to have observed, for the first time, IE over five- 31 and even six-bonds 32 in a saturated molecule. Through-space interactions were supposed to

9

cause such a long-range transmission of IE in a rigid molecule. Large 4 effects, up to -23 ppb 32, were claimed to originate from a decreased strength of hyperconjugative interaction of the C-D bond when compared to the C-H bond. The authors found good linear correlations between 1 and the fractional s-character of the corresponding C-D bond 31, as well as between 4 and the chemical shift of the observed carbon atom 33. However, they believed that the observed changes in the chemical shifts of adamantane derivatives were due to different potential surfaces of isotopomers, i.e. that deuterium was behaving like a real substituent, which is nowadays not encouraged in the literature 7,9,13. Nevertheless, IE and proposed correlations can serve as a good basis for studying the electronic structure and conformation of similar rigid systems. Deuterium LRIE in adinazolam mesylate have recently been used to study structural changes under various pH conditions 34. No clear distinction between the intrinsic and equilibrium contributions to the total observed IE was made. On the basis of intrinsic and equilibrium IE, Forsyth et al. 35 discussed possible conformations of dimethyl-9,10-dihydroanthracenes and 1,4,4-trimethylcyclohexyl cations in solution. DIE provided evidence of rapidly equilibrating non-planar conformations for dimethyl-9,10-dihydroanthracenes and trans-9,10-dimethyl-9,10-dihydroanthracenes, while only one conformation was present for cis-9,10-dimethyl-9,10-dihydroanthracenes. Cumulative DIE in the 2-tert-cumyl-2-adamantanyl cation have proved the existence of an unsymmetrically phenyl-bridged carbocation 36. A deshielding effect

10

of 1.0 ppm at the carbocation centre and a shielding effect of 0.9 ppm at the quarternary carbon have been observed in the d6-dimethyl isotopomer of 2-tert-cumyl-2--adamantanyl cation. These effects were temperature invariant, thus pointing to the bridged structure. This was in agreement with ab initio calculations 36 performed by the same authors. Servis et al. 37 studied DIE in

13

C NMR spectra of 2-substituted-2-norbornyl

cations. The effects were found to have appreciably different magnitudes depending on the position of the deuterium in the molecule. The most interesting IE was a shielding three-bond effect, 3, observed in the 2-methyl-2-norbornyl cation, amounting to as much as 1.414 ppm in the monodeuteriated isotopomer. The authors excluded hyperconjugation or perturbation of equilibrium as a cause of this large value. Instead, "isotopic perturbation of resonance" was suggested to be the major contributing factor 37. However, such large IE were not observed in 2-aryl-2-norbornyl cations at all. It was concluded that for the former cation a bridged structure was present in solution, while for the latter, the observed findings pointed to the onset of bridging. Structural properties of 1-(trimethylsilyl) bicyclobutonium ion have been studied using intrinsic and equilibrium DIE 38. Both the sign and the magnitude of IE depend on whether the deuterium is placed at an exo- or endo-position. Using IE, Siehl et al. 38 were able to confirm experimentally their theoretical findings that the bicyclobutonium cation had a bridged structure and that other possible structures might be excluded. The differences in IE between exo- and endo-CHD labelled cations

11

were rationalised by different endo- and exo-C-H bond force constants at the pentacoordinated carbon. Insufficient accuracy led Yonemitsu et al. 39 to the wrong conclusion that IE in polydeuteriated benzoic acids could be estimated on the basis of additivity. The authors failed to observe any LRIE (4 or 5) in monodeuteriated benzoic acid, and their SRIE values (1, 2 and 3) differ appreciably from those reported in the recent paper by Novak et al. 21, who observed significant nonadditivity for some carbons in benzoic acid, which was rationalised by changes in hydrogen-bond dynamics in solution. This will be discussed later in detail. Very large DIE on

13

C chemical shifts, with magnitudes up to 3.06 ppm per

deuterium were also observed in a paramagnetic molecule 40. It was taken that the isotopic substitution may affect the susceptibility of a paramagnetic compound.

4.1.2. Deuterium isotope effects on other chemical shifts Only a few papers reported DIE on proton chemical shifts, 1 1H(D) 40-42. Anet and Park observed 2 effect in 2-D-citrate with the value of 40 ppb (we used here the sign convention opposite to that in ref. 41). They concluded that the reported value was a superposition of approximately two equal contributions, i.e. the intrinsic and equilibrium IE. Very large DIE on 1H, but also on

13

C chemical shifts, were observed in some

paramagnetic compounds 40,42. In deuteriated haemins, Medforth et al. 42 determined LRIE on proton shifts. The authors claimed that those LRIE originated

12

from a change in the contact contribution to the hyperfine shift. They supposed that the deuteriated methylene group had a slightly weaker electron-donating ability than the undeuteriated one. Similar arguments to explain DIE on

15

N shieldings in

alkylcobalamines were used by Brown et al. 43. This reasoning is, however, not supported by the recent literature on IE (see e.g. 7 and 9), where rovibrational averaging was established as the origin of IE. LRIE up to 10 or even 12 bonds from the isotopic site on both 1H and 13C chemical shifts have previously been observed only for some equilibrating 44 or conjugated systems 2. DIE over one bond on

17

O shieldings, 117O(D), in the gaseous oxonium ion was

calculated by Sauer et al. 45 and compared with other similar measured data. A reduction in 1 when going from the gaseous (-1.54 ppb 46) to the liquid phase (-0.3 ppm 47) could partly be accounted for by the much lower shift in protonated H2O relative to the unprotonated one. Similar behaviour was found for NH4+ and NH3 molecules 48.

13

4.2. Isotope effects of other nuclei IE on nuclear shieldings have also been applied for the analysis of isotopically labelled drugs 49. The effects of tritium (3H) on 3H chemical shifts, 3H(3H), (2 = 20 ppb, 3 = 10 ppb) and the effect of

14

C on the 1H chemical shift (-3.5 ppb)

have been determined. For the entire range of functional groups (66) the magnitude of one-bond upfield 18

O isotope effects on

13

C chemical shifts, 113C(18O), ranging from 16 to 52 ppb,

showed a linear relationship with the 17O chemical shifts for the corresponding oxygen atom in the molecule 50. This relationship (7) enables an estimation of the general magnitude of an

18

O isotope effect on the

13

C chemical shift, from the known

17

O

chemical shift value of the oxygen atom under consideration.

 C(18O) = 0.05509 ( 17O) + 16.12

1 13

(7)

By extrapolation of the straight line, the occurrence of unusual downfield  C(18O) for organic molecules whose

1 13

17

O chemical shifts are less than -292 ppm

(i.e. upfield with respect to H217O) was anticipated. The first observation of a downfield

18

O isotope effect in

13

C NMR spectrum was reported for the three-bond

effect (amounting to -9 ppb) in 2,6-dimethyl-4-pyrone 51. Oxygen exchange reactions in sugars are usually assayed by the 18O isotope effect on 13C chemical shifts, since the presence of an

18

O directly attached to the anomeric

carbon causes small upfield shifts of its 13C signal. Thus, the following one-bond 18O

14

effect on anomeric

13

C were measured: -ribofuranose, 17 ppb; -ribofuranose,

16 ppb; -ribopyranose, 15 ppb; -ribopyranose, 19 ppb; -2-deoxyribopyranose, 20 ppb; etc. 52. The three-bond downfield 18O isotope effects on 13C chemical shifts (ca. 5 ppb) were observed in acyclic sulfinate esters, while only a line broadening was recognised in cyclic analogues 53. The 113C(18O) produced upon hydrate formation of an aldehyde or ketone were measured for the first time in nucleosides. Namely, in 1-(-D-Glycofuranosyl)uracil-6-carboxaldehydes shielding isotope effects upon the hydrate and aldehyde

13

C nuclei

of 16 and 44 ppb, respectively, were observed 54. One-bond 37Cl isotope effects on 13C chemical shifts, 113C(37Cl), were determined under the conditions of low-power composite pulse 1H decoupling in a series of aliphatic and aromatic chlorohydrocarbons 55. Carbon atoms directly bonded to a chlorine atom showed a characteristic 3:1 doublet, which corresponds to the natural abundance ratio of 35Cl (75.53%) to 37Cl (24.47%). All 1 were upfield effects, ranging from 4.2 to 14 ppb. A number of features of 113C(37/35Cl) were observed: (i) the inverse relationship between the magnitude of 1 and the amount of s character existed, (ii) values of 1 depended strongly on the degree of substitution of the carbon atom, (iii) the magnitude of 1 for the ortho isomer was larger than that for the meta and para isomers, which was correlated with the shorter C-Cl bond in ortho than in meta and para isomers, (iv) 1 was larger for the axial than for the equatorial orientation of the chlorine substituent in cyclohexane ring, suggesting a shorter C-Cl

15

bond for the axial orientation, (v) there was a linear increase in the magnitude of 1 with a decrease in temperature. Raynes et al. 56 experimentally determined and calculated 113C(37/35Cl) in CH3Cl, which amounts to 6 ppb. They found that the changes in the C-Cl bond length and the mean square of the C-Cl bond length are sufficient to account for this isotope shift It was reported that one-bond

37

Cl isotope effects on

13

C shifts in a series of

para-substituted chlorobenzenes are in the range from 4.0 to 5.2 ppb 57. These 1 tend to decrease as the electron-attracting character of the substituent increases, giving rise to a linear correlation between the isotope effect and the chemical shift of the carbon atom bonded to chlorine. The

37

Cl/35Cl-induced isotope effects on

13

C chemical shifts in the spectra of

chlorinated methanes CH3Cl, CH2Cl2, CHCl3, and CCl4 were found ranging from 3.0 to 6.0 ppb per one 37Cl/35Cl replacement and decreasing in magnitude with the number of chlorines 58. The

37

Cl/35Cl isotope effects measured for CH3Cl, CH2Cl2, and

CHCl3 at different temperatures, revealed in all cases a temperature dependence of about 0.015 ppb / C. Authors attributed the

37

Cl/35Cl isotope shifts to the very slight

shortening of C-Cl bonds by about 4 x 10-5 Å at room temperature, on going from C-35Cl to C-37Cl.

16

4.3. Hydrogen bonding and tautomeric equilibria Recently, Hansen has published a doctoral thesis covering his work on IE in hydrogen bonded systems and other equilibrium IE including mostly tautomeric systems. (ref. 59 and references cited therein). He studied IE in ortho-hydroxyacyl benzenes, enamines, -diketones and ketoesters, ortho-hydroxy- and ortho-amino azocompounds and -thioxoketones. He concluded that DIE over two bonds on carbon chemical shifts are very sensitive to hydrogen bonding. Thus, one can distinguish between isomers by observing 2C(ND) in Z- and E-enamines. Similarly, 2C(SD) in -thioxoesters could be a gauge for the percentage of a hydrogen bonded rotamer. Another example of 2C sensitivity to hydrogen bonding is the correlation of isotope effects over two bonds in ortho-hydroxy aromatic esters, ketones and aldehydes with OH. On the basis of the investigations in 2,6-dihydroxyacyl aromatics, it was concluded that the OH group formed a stronger hydrogen bond than the OD group. A study of the isotopic perturbation indicated that the degree of perturbation is proportional to the strength of the hydrogen bond. The main conclusion was that the isotope effects can be used for detection of hydrogen bonding, as well as for indication of the direction of equilibrium perturbation. Additionally, Kozersky ([60] and references cited therein) has reported a comprehensive review of equilibrium isotope effects on chemical shifts and their applications. Only a few papers on equilibrium IE remained, and they will be reported here briefly. Sobczyk and co-workers [61,62] studied DIE on chemical shifts of 13C atoms

17

in the phenyl ring (especially carbon atom directly linked to an OH group) in ortho-Manich bases (systems with a strong intramolecular O-H··· N hydrogen bond). They reported that IE could be very well correlated with the calculated atom charges if the weaker hydrogen bond (when hydrogen is replaced by deuterium) is taken into account. By observing the temperature dependence of the IE for various carbon atoms, the authors found a critical temperature (245 K), where sign inversion of the effect takes place, i.e. contributions of the proton-transfer (PT) and non-proton-transfer (HB) forms compensate each other. On the basis of these results, they estimated that the magnitude of DIE in the PT form is three times bigger than in the HB form at 245 K, which was in agreement with the results observed by UV, IR and dipole moment measurements. DIE

was

used

for

indication

of

the

hydrogen

bonded

structure

in

1H-pyrrolo[3,2-h]quinoline [63]. Penman et al. [64] reported that some four-bond downfield DIE on certain chemical shifts could be attributed to conformational equilibrium perturbations. IE were also used to study the hydrogen-bond structure in biomolecules, such as bovine ferricytochrome [65], but the present article will not treat such molecules in detail.

18

5. Isotope effects on spin-spin coupling

During the last few years, IE on scalar spin-spin couplings in organic molecules have been reported only in a few papers. Detailed surveys of these effects have previously been given by Jameson [9, 12], Sergeyev [11] and Contreras and Facelli [66]. As already pointed out, high precision of the coupling constant measurement is the main prerequisite for determining IE on coupling. Primary IE could be determined less precisely than secondary IE. This is due to the following reasons. First of all, according to equation (2), the experimental error in measuring the coupling constant is always increased by the multiplying factor H/D (6.5144). Furthermore, if the isotope in question possesses a quadrupole moment, such as deuterium, its relaxation usually gives rise to an appreciable broadening of NMR lines. These quadrupolar effects were discussed in detail by Sergeyev [11]. Temperature gradients in the case of decoupled spectra additionally broaden the lines. All this makes the determination of primary IE on coupling rather difficult. On the other hand, determination of secondary IE is not necessarily affected by all the above mentioned facts. However, they are usually much smaller, so that high precision of the measurement is also required. Hence, it is not surprising that many data on IE published so far have suffered from insufficient accuracy. Previous reports claim that primary and secondary deuterium effects on spin-spin coupling could be both positive and negative 67-69, depending on hybridisation and the presence of lone-pair electrons.

19

The high precision in measuring the coupling constants using an iterative method to analyse quadrupolar effects made it possible to determine IE on J(13C,1H), J(15N,1H) and J(14N,1H) in nitromethane [70]. IE on J(13C,1H) was found to be substantially different (-0.97 Hz) from the value reported by Everett (-0.26 Hz) [68]. Leshcheva et al. [71,72] further demonstrated the usefulness of the iterative lineshape analysis in determining the carbon-deuterium coupling constant and stressed the importance of this procedure for obtaining very accurate coupling constant measurements in toluene and benzaldehyde. They were able to determine the primary effects, pJ, and secondary effects, sJ, very precisely and showed that secondary effects on 1J(C,H) in toluene (-0.319 Hz) were larger than primary (-0.002 Hz), which contradicted the existing theoretical estimates 73. It was later proved by ab initio calculations 74 that primary effects were not necessarily larger than secondary and this is now established as a common phenomenon. The authors used DIE on coupling constants in methane to prove experimentally their theoretical findings of the anomalous behaviour of coupling constants with bond stretching. For instance, stretching of one C-H1 bond in methane produced larger changes in those coupling constants which did not involve proton 1 74. The anomalous effect was ascribed to electron correlation. Primary and secondary DIE on 1J(C,H) in dichlormethane were reported by Sergeyev et al. 20. Their results indicated a substituent dependence of primary IE on 1

J(C,H), but not for secondary IE. pJ increased on going from CH4 (-0.08 Hz 75) to

CH2Cl2 (-0.60 Hz 20) and then to CHCl3 (-0.96 Hz 68), which gave a linear

20

dependence of pJ on the number of chlorine atoms. Both pJ and sJ were found to be negative (-0.60 Hz and -0.25 Hz, respectively) and fairly additive. High accuracy enabled observation of the negative and very small IE on carbon-proton coupling constants induced by

37

Cl/35Cl substitution in chlorinated

methanes 58. It was concluded that these effects seemed to be in the same direction as those induced by deuterium 20. 37Cl/35Cl isotope effect was interpreted in terms of 0.00004 Å shorter C-Cl bond length, as reported in ref. 56. Secondary DIE on the carbon-deuterium coupling constant have been recently used for studies of conformational equilibria in cyclohexanes 23. A clear distinction between the intrinsic and thermodynamic parts of the observed IE was made. For the one-bond and geminal couplings, the intrinsic contribution predominated. In contrast, for the vicinal couplings, IE were found to be completely thermodynamic and thus could be used to estimate shifts of the conformational equilibria in deuteriated cycloalkanes. In

13

C-labelled isotopomers of some diphosphines, constituting an ABX spin

system, Heckman and Fluck 76 speculated on possible secondary vicinal P,P coupling constant. Unfortunately, none were determined.

13

C IE on the

21

6. Isotope effects in Ph-R and Ph-Z-Ph systems

Ph-R and Ph-Z-Ph molecules are benzene derivatives which consist of a phenyl ring Ph, a side-chain group R and a bridging group Z (Fig. (1)).

R 1 6

5

2

1

4

5

R = CH3 (toluene)

Z

3'

1'

4'

 '

3 4

2'

6

3

2

6'

5'

Z = CC (cis- and trans-stilbene)

NH2 (aniline)

CC (tolane)

CHO (benzaldehyde)

C=N (trans-N-benzylideneaniline)

COOH (benzoic acid)

C=O (benzophenone)

COCH3 (acetophenone)

N=N (trans-azobenzene)

Fig. (1) Benzene derivatives of the Ph-R and Ph-Z-Ph types and atom numbering.

Apart from the short range isotope effects, (SRIE) i.e. those transmitted over up to three bonds, 3, long range isotope effects (LRIE) could also be present in these -electron molecules. It was previously reported by Berger and co-workers (see ref. 2 and references cited therein) that, in conjugated molecules containing phenyl rings, deuterium LRIE over up to 10 or even 12 bonds were operative in para-deuteriated stilbene, tolane and 1,4-diphenylbutadiene derivatives. The authors claimed a close parallelism between substituent effects and deuterium effects. They developed the idea

22

of -polarisation, the basic concepts of which were first introduced by Young and Yannoni 77. When deuterium is directly attached to a -system, it polarises the molecule, thus behaving like a real substituent. According to Berger 2, "isotope effects mirror substituent effects" but one has to replace the ppb with the ppm scale. Although interesting, this approach, if strictly taken, is against the Born-Oppenheimer approximation, assuming different potentials for isotopomers. In order to obtain a further insight into the phenomenon of IE and the way of its transmission in -electron molecules, we have systematically incorporated isotopes at different sites in the Ph-R and Ph-Z-Ph types of molecules. We have labelled the ortho-, meta- and para-positions in the benzene ring and the -positions in the side chain and/or bridging group (Fig. (1)). A series of isotopically labelled toluenes, anilines, benzaldehydes, benzoic acids, acetophenones, cis- and trans-stilbenes, tolanes, trans-N-benzylideneanilines, trans-azobenzenes and benzophenones were prepared

and

analysed.

Besides

mono-isotopically

labelled

molecules,

poly-isotopically labelled species have also been synthesised in order to check whether the established additivity rule is preserved or not.

6.1. Deuterium isotope effects on nuclear shielding 6.1.1. Short-range deuterium isotope effects on 13C chemical shifts The effects of deuterium on 13C chemical shifts over one and two bonds, 113C(D) and 213C(D), in both Ph-R and Ph-Z-Ph isotopomers were all found to be positive (shielding) 78-82, which was consistent with the already reported data in molecules

23

of a related type. 1 depends on the C-atom hybridisation and bond order. For example, in -D-trans-N-benzylideneaniline, the effect was found to be 260.0 ppb 79, while in -D-cis-stilbene it amounted to 338.6 ppb 80. 2 in the benzene ring are on average 110.0 ppb, except for those at quaternary carbon C-1 (Fig. (1)). A significant reduction (on average 50 %) of the 2 at C-1 in Ph-R and Ph-Z-Ph isotopomers was rationalised by steric interactions involving ortho-protons, on the one side, and -protons or lone-pair electrons on the other. This was also supported by findings in benzonitrile [83]. Since the molecule has a collinear atom arrangement (Ph-CN), no interaction of this kind is possible. The observed effect of 103 ppb was similar to others found in the benzene ring. Deuterium effects over three-bonds in o-, m- and p-D-isotopomers can be both positive and negative and are spread in the range -17.8 to 62.0 ppb (Fig. (2)). The magnitude of these effects depends on the position of the deuterium in the molecule and on the electronic structure of the substituent in the benzene ring, as well as on the presence of lone-pairs, i.e. they reflect structural relationships. When a heteroatom (nitrogen) is directly attached to the benzene ring, i.e. at the -position, 3 at C-6 in o-D-isotopomers is negative, like in aniline, trans-azobenzene and trans-N-benzylideneaniline (Fig. (2)), while it is positive in all the others.

24

a) H

D 0

14.0

D

24.5

Ph

Ph

N

C

C

Ph N H

N D

D 20.9

0

3.9

H

C

D 14.8

30.8

7.4

O

N D

D -17.8

8.5

-8.2

4.0

0

O

0C

D

-6.9

10.1

Ph

OH

10.9 C

62.0 NH2

62.0 CH3

O

4.4

b) O

H CH3 10.6

7.6

C

NH2 0

D

C

0

O

D

H

C

7.0

D

0

N

13.0

D

Ph

C N

N H

N 8.0

0

9.0

D

6.9

Ph

C 7.4

7.5

3.0

Ph Ph

O

OH

D

0

D

4.7

Fig. (2) Deuterium isotope effects over three-bonds (ppb) in a) o-D- and b) m-D-isotopomers.

D

25

3

 at C-2,6 in p-D isotopomers having a heteroatom with lone-pairs (O or N) at the

-position from the benzene ring have not been observed, while positive values have been found in isotopomers without a heteroatom (Table 1). The steric and inductive influence of lone-pairs is the main reason for the cancelling of the effect at that position. In -D isotopomers, 3 can be positive or negative 78-80,82. Its magnitude depends on the torsional C2=C1-C=C' or C2=C1-N=C angles. If the angle is smaller, the effect is larger. A similar conformational dependence of 3 has previously been observed for deuteriated cyclohexanes, toluenes, protoadamantanes and norbornanes [5, 84-86].

6.1.2. Long-range deuterium isotope effects on 13C chemical shifts For effects over four-bonds, 4, similar structural relationships have been observed. In p-D isotopomers of the Ph-R and Ph-Z-Ph type, 4 at C-1 is always negative when lone-pairs are present in the molecule. Otherwise, a positive value is observed (Table 1). Contrary to this, 5 is negative no matter if lone-pairs are present or not. The only exception is p-D-benzoic acid, where a small positive effect was detected (Table 1). It is ascribed to the additional deuterium effect from the OD group. Namely, in the acetone-d6 solution of benzoic acid a rapid exchange of hydroxyl proton with deuterium is brought about, so that OD instead of OH exists, giving rise to an

26 Table 1. Deuterium isotope effects, n (ppb)a, in para-deuteriated isotopomers of the Ph-R and Ph-Z-Ph types.b  1

2

3

4

5

276.0 289.3 285.0 283.3 280.0 278.0 281.0 279.0 296.0 286.0 281.3 275.0

110.0 110.3 109.8 109.4 109.0 110.5 110.3 111.5 110.0 110.0 109.8 109.0

11.0 0.0 0.0 0.0

7.0 -6.5 -1.6 -4.0

11.5 7.9 7.0 0.0 7.0 0.0 4.0

5.4 1.4 1.4 -3.5 0.0 -6.5 0.0

-3.0 -1.9 3.2 -4.4 -8.0 -4.6 -9.5 -8.6 -6.5

6

7

8

9

0.0

0.0

2.2

-2.4 -2.4 -2.8 -3.0 0.0 0.0 -3.0

0.0 0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0 0.0

10

molecule toluene benzaldehyde benzoic acid benzophenone phenylacetylene diphenylethane cis-stilbene trans-stilbene trans-N-benzylideneaniline (p) trans-N-benzylideneaniline (p') trans-azobenzene tolane

0.0 15.0 0.0 8.7 10.2 7.0

-7.2

15.6

a

n denotes the number of intervening bonds between D and C atoms

b

Standard deviations did not exceed 2 ppb for 1 and were in the range 0.5-1.0 ppb for the other effects

0.6 2.0 2.8 0.0 0.0 3.3 2.7

27

additional deuterium effect of OD. The positive value of 3.2 ppb (Table 1) is therefore a cumulative effect arising from both contributions [87]. The most interesting long-range deuterium effects are those transmitted over six bonds, 6, in p-D isotopomers of the Ph-Z-Ph type. A dependence of 6 and a torsional angles of the intervening C2=C1-C=C', C2=C1-N=C or C2=C1-CC' bond fragments was established (8):

6

 = Acos2 + B

(8)

where  is the torsional angle. We have used calculated (semiempirical and ab initio methods) and experimentally observed values of . Good correlations were found in all cases.† If the values of  calculated by the ab initio 6-31 g* method are taken (Table 2), then the following relation is obtained :

6

 = 14.17 cos2 + 0.28

(9)

This correlation is depicted in Fig. (3). It can serve as a conformational probe in benzene derivatives of a similar type. Once 6 is known,  can be predicted following the expression (9). Similar dependence of the total IE, t, (observed at C-4) on the same angles in pentadeuterio isotopomers of the Ph-Z-Ph type has also been found.† †

unpublished results

28

As already mentioned, the conformational dependence of 3 [84-86], as well as of 2 [89], on the dihedral angle has been reported in the literature.

Table 2. Deuterium isotope effects over six bonds, 6 (ppb), and torsional angles,  (), for para-deuteriated benzene derivatives 6



1

molecule 12.02

18.9

CH3-styrene

9.0

2

43.0

(CH3)3-styrene

0.02

91.3

phenylacethylene

15.0

0.03

cis-stilbene

8.7

43.6

trans-stilbene

10.2

23.0

tolane

15.6

0.03

trans-N-benzylideneaniline

7.0

44.64

styrene

0.05 r6 1

values calculated using the ab initio 6-32 g* method

2

values taken from ref. 88

3

planar molecule

4

refers to C2=C1-N=C angle

5

refers to C2=C1-C=N angle

6

correlation factor

0.97

29

1.2 1.0

cos2

0.8 0.6 0.4 0.2 0.0 0

2

4

6

8 6

10

12

14

16

18

 (ppb)

Fig. (3) Dependence of 6 in p-D isotopomers on cos2 ( calculated by the ab initio method).

It is of interest to note that, in contrast to 6 in p-deuteriated isotopomers, 6 in

-deuteriated isotopomers does not show a linear dependence on cos2. Moreover, their magnitudes and also signs differ appreciably. For example, 6 measured at C-4 in -D-trans- and -D-cis-stilbenes amounted to 12.9 ppb and -9.0 ppb, respectively, while 6 measured at C- in p-D-trans and p-D-cis-stilbenes were 10.2 ppb and 8.7 ppb, respectively 80,82. It demonstrated that, besides the importance of the isotopic position within the molecule, the way and direction of isotopic transmission play an important role with respect to the magnitude of the effect. Thus, different values are obtained when the effect is transmitted from the para to the alpha position

30

and vice versa. This directionality of DIE has also been observed for other LRIE, as well as for 3 ( Fig. (2)).

6.1.3. Vibrationally induced -electron polarisation LRIE over seven, eight, nine and even ten bonds were also determined in p-D- (Table 1) m-D- and o-D-isotopomers of the Ph-Z-Ph type [21,79-81]. Such LRIE are usually found in fully or cross conjugated -electron molecules. In p-D isotopomers (Table 1), the longest possible deuterium effect is the one over ten bonds, 10

 (or 9 in the case of benzophenone). They were determined in all p-deuteriated

isotopomers at C-4 or C-4' positions (Fig. (1)). Furthermore, in o-D (Fig. (4)) and m-D benzophenones, the corresponding effects at the same site, i.e. 7 and 8, respectively, were also detected 21. The same was found in o'-D- and m'-D-trans-N-benzylideneaniline (see Fig. (4)).† This indicates that whenever D is attached to the benzene ring at o- m- or p-positions, the isotopic perturbation is transmitted to the neighbouring aromatic moiety, in spite of the large torsional C2=C1-C=C' or C2=C1-N=C angles 79,80 and even if there is no direct conjugation between the two aromatic moieties 21. Deuterium LRIE in Ph-Z-Ph molecules usually display an alternation in sign and magnitude, thus resembling the previously mentioned -polarisation mechanism of transmission [2]. However, -polarisation could also be understood within the †

unpublished results

31

2.7 1.7 -3.1

N

C

1.0

H D

O 14.0

C -4.2 3.9

D

-1.3

Fig. (4) The sign alternation of LRIE in o'-D- trans-N-benzylideneaniline and o-D-benzophenone.

Born-Oppenheimer approximation if one assumed that subtle charge shifts take rise only from a shortening of the C-D bond, as compared to the C-H bond (ca. 0.07 Å). This "vibrationally induced" polarisation of a -system was supported by calculations of the atomic charge differences modelled by the C-H bond shortening in p-deuteriated isotopomers [90]. Since polarisation of the -system occurs both through the bond and through space, LRIE were observed also in distorted molecules where delocalisation is perturbed by a large torsional angle (cis-stilbene, trans-N-benzylideneaniline), (benzophenone) and also in molecules where conjugation between the two phenyl rings is broken by two (benzophenone) or more consecutive single bonds (diphenylethane). The redistribution of charge within a molecule is considered to be of

32

vibrational origin, i.e. it arises from changes in bond lengths and angles upon the isotopic substitution.

6.1.4. Deuterium isotope effects on 1H chemical shifts Deuterium isotope effects on 1H chemical shifts, n1H(D), were determined in isotopomers of cis-stilbene, trans-stilbene and trans-N-benzylideneaniline [82,91].  H(D) were measured for olefinic protons in cis-stilbene and trans-stilbene, and for

n 1

azomethine proton in trans-N-benzylideneaniline (Fig. (1)). In -D-cis-stilbene, the effect over three bonds is much smaller (2.3 ppb) than in the corresponding trans-stilbene (6.3 ppb), which is consistent with the observed trend, i.e. 3-cis is smaller than 3-trans [92]. This again demonstrates a conformational dependence of isotope effects. In -D-styrene, the respective effects are significantly larger (3-cis = 6.0 ppb and 3-trans = 13.0 ppb) [92]. Moreover, in D-acetylene, the value of as much as 16.0 ppb was determined [93]. The reduction of 3 in cis-stilbene and trans-stilbene was attributed to the inductive influence of the second phenyl ring. However, deuterium effects on C- in the mentioned molecules exhibit an opposite trend, i.e. the effect increases in the opposite direction. It is proposed that ring current effects which give rise to an increase in shielding anisotropy should be responsible for such behaviour [91]. In contrast to cis-stilbene and trans-stilbene, all deuterium effects in trans-N-benzylideneaniline are negative, which was ascribed to lone-pair interactions. In p-D-trans-N-benzylideneaniline, the effect over as many as six-bonds

33

was determined by the value of -0.88 ppb [82]. Such LRIE on 1H chemical shifts were very seldom observed, e.g. in paramagnetic molecules [41,42] or some hydrogen-bonded systems such as macrolide antibiotics [44].

6.2. 13C isotope effects on nuclear shielding We have determined the effects of 13C isotope on 13C shifts, 113C(13C). 13C isotope effects are much smaller than deuterium effects, owing to the low 13C/12C mass ratio. It is therefore not surprising that reports on 13C effects are scarce in literature [6,7]. We have studied

13

C effects in

13

C labelled carbonyl compounds, such as benzaldehyde,

benzoic acid, benzophenone and acetophenone (see Table 3) [94].

Table 3. 13C isotope effects, n (ppb), on 13C chemical shifts in 13C-carbonyl labelled compounds 1



2



3



molecule benzaldehyde

16.36 (0.07)

benzoic acid

16.00 (0.30)

acetophenone

[C-1] 14.24 (0.50)

-1.32 (0.26)

-1.37 (0.04)

0.57 (0.22)

-0.72 (0.11)

0.69 (0.08)

[CH3] 10.10 (0.70) benzophenone

13.40 (0.65)

Standard deviations are given in parentheses. Solvent = acetone-d6. T = 294 K.

34

The effects over one bond are positive and in the range 10.10 to 16.36 ppb (Table 3), depending on the electronic structure of the substituent group in the benzene ring and the site of observation. Contrary to this, the effects over two-bonds at C-2,6 are all negative and dependent on molecular conformation. Namely,

 in

13

C-benzaldehyde

13

C-benzophenone, owing to steric interactions in the latter, involving ortho-protons

and

13

2

C-acetophenone

are

twice

as

large

as

that

in

from the adjacent phenyl rings. No effect over two or three bonds was observed in benzoic acid, which was due to hydrogen-bond dynamics in benzoic acid solution [87]. We have for the first time determined 13

 C(13C) in

3 13

13

C-acetophenone and

C-benzophenone, both being positive and with similar magnitudes (Table 2). Sign

alternations of 13C isotope effects in 13C-acetophenone and 13C-benzophenone support the previously discussed charge shifts upon isotopic substitution. The apparent solvent dependence was revealed for

13

C-benzoic acid owing to

perturbations in hydrogen-bond dynamics in different solvents, acting as hydrogen-bond donors or acceptors [87].

6.3. 15N isotope effects on nuclear shielding The effects of 15N [6,7] have been even less investigated than that of

13

C owing to

the same reason. We have determined 15N isotope effects over one- and two-bonds in 15

N-labelled trans-N-benzylideneaniline and trans-azobenzene [82]. Effects over one

bond are positive, while those over two-bonds can be both positive and negative. The

35

previously established dependence of 1 15N(13C) on the C=N bond order [95-97] was confirmed here. With increasing the bond order, effects also increase. 15

15

N effects

in

15

N-trans-azobenzene

are, on average,

smaller

than in

N-trans-N-benzylideneaniline. This reduction is caused by the influence of a

lone-pair of the second nitrogen in 15N-trans-azobenzene. It was also demonstrated that lone-pair electrons reduce the one-bond carbon-nitrogen couplings 1JC,N, also observed by Jameson [98], which was not true of the two-bond couplings, 2JC,N, in trans-azobenzene.

15

N-isotopomers of trans-N-benzylideneaniline and

36

6.4. Additivity of isotope effects on nuclear shielding Isotope effects were shown to be additive in most cases. When two or more isotopes are present in a molecule, the total observed effect is additive. The linear dependence of IE on the number of isotopes in equivalent positions has been thoroughly discussed by Jameson and Osten 22 for some small symmetric and linear molecules. The authors related the question of additivity to the mass effects. Nonadditivity in larger molecules was usually found in equilibrating systems (7 and ref. cited therein). Ph-R and Ph-Z-Ph types of molecules generally obey the additivity rule. However, in some cases, significant deviations from the additivity rule were observed at the specific sites in the molecules. Accordingly, these molecules can be divided into two groups. The first group consists of isotopomers where nonadditivity was accounted for by unequal rotamer distribution of isotopomers and the second group includes hydrogen-bonded systems. Deviations from the additivity in benzophenone 21, trans-N-benzylideneaniline, 99, cis- and trans-stilbene 91 were attributed to different rotamer population distribution for isotopomers. Namely, in perdeuteriated phenyl rings, the strain between the ortho-ortho or ortho-alpha positions (Fig. (5)), being the main reason for the nonplanar structure of Ph-Z-Ph molecules, is partly released when compared to undeuteriated rings. It is due to a shorter C-D than C-H bond, which affects rotations about Ph-C or Ph-N axes. This, in turn, brings about differences in the distribution of rotamer populations between deuteriated and undeuteriated molecules, slightly changing the chemical shifts.

37

O C

'



Fig. (5) Conformation of benzophenone.

The nonadditivity was established for atoms (C or H) most sensitive to such perturbations 21,91,99. In all isotopomers where deviations from the additivity rule were found, dependence of isotope effects on temperature was observed, supporting the aforesaid conclusion. If changes in the rotamer population upon the isotopic substitution in molecules that may have many rotational isomers are the reason for nonadditivity, then additivity should be obeyed in rigid systems possessing only one conformer. To check this, we have recently studied IE in rigid benzocyclobutane isotopomers (Fig. (6)).† The magnitude and sign of IE in 2, 3 and 5 are in good agreement with those reported by Berger 2. We found effects perfectly additive for all carbon atoms in all isotopomers of benzocyclobutane. Furthermore, measurements of IE showed no temperature dependence, corroborating the rotamer redistribution as the cause of nonadditivity in nonrigid Ph-Z-Ph molecules.

†

unpublished results

38

D

D

D

D

D

4

Fig. (6)

3

2

1

D

D

D

D

D

5

Deuteriated isotopomers of benzocyclobutane.

Benzoic acid belongs to the second group where nonadditivity was observed. The apparent deviation from the additivity rule was found for carbonyl carbon (C=O) and a slight deviation was detected for C-4 as well. Perturbations in hydrogen-bond dynamics were suggested as a probable cause of nonadditivity 21. We have demonstrated that in acetone solution, which is a relatively good H-bond acceptor, the dimer structure of benzoic acid was suppressed in favour of the monomer bonded to acetone 87, thus giving a benzoic acid-acetone complex, as depicted in Fig. (7). Thereby, two equilibria exist, one in the dimer and the other in the benzoic acid-acetone complex, depending on the concentration 87. Changes in H-bond dynamics give rise to a redistribution of electron density, primarily at the carbonyl site which affects the magnitude of IE.

39

Fig. (7) Hydrogen-bond structure of the benzoic acid dimer and the benzoic acid acetone complex.

6.5. Deuterium isotope effects on spin-spin coupling Primary and secondary deuterium effects, nJ(Hz), were determined on the olefinic coupling constants, 1J(C,H) and 3J(H,D), according to expressions (10), (11) and (12) in a series of deuteriated cis- and trans-stilbenes [100].

p1J (C,H) = (H/D)  J(C,D) - J(C,H)

(10)

s1J (C,H) = 1J(C,H) - 1J(C,H)

(11)

s3J (H,D') = 3J(H,D') - 3J(H, D')

(12)

where H and D are the magnetogyric ratios of hydrogen and deuterium, respectively, while the asterisk denotes the deuteriated molecule.

40

The olefinic parts of both cis- and trans-stilbene constitute a typical ABX spin system (1a). The exact values of J(C,H) could only be derived by a complete spectral analysis. We have analysed both the AB and X parts of the spectra using the PANIC program. All -deuteriated isotopomers constitute an AMX spin system (1b) and the coupling constant could be measured directly from spectra.

HA

HB Cx

HA

C

Ph

DM Cx

Ph

Ph

1a

C Ph

1b

6.5.1. Secondary isotope effects Secondary effects on 1J(C,H) are negative, i.e. reduced values of coupling constants were observed in deuteriated molecules (from -0.06 to -0.20 Hz). In contrast, secondary effects on 3J(H, D') are all positive and in the range 0.04 - 0.13 Hz, thus being relatively larger than those on 1J(C,H). In D11-trans-stilbene, we have observed the biggest change in the coupling constant [11], originating from a secondary isotope effect (5.9%). Secondary deuterium effects on both coupling constants are cumulative, i.e. the additivity rule holds within the experimental error.

41

6.5.2. Primary isotope effects We have determined only two primary effects in -D-trans-stilbene and

-D-cis-stilbene in order to check whether the trends found for secondary effects are preserved also for primary effects. As already pointed out, primary effects could be less precisely determined due to the multiplication by factor H/D and the quadrupolar deuterium relaxation which broadens the lines in proton-decoupled 13C spectra. Like in the case of secondary effects, the primary effect is larger in trans-stilbene (-0.60 Hz) than in cis-stilbene (-0.44 Hz).

6.5.3. Geometrical considerations Isotope effects on coupling constants could be understood in terms of traditional physico-chemical concepts. The C-D stretching vibrational mode makes the greatest contribution to the change in rovibrational averaging [9,12,13]. Possible angle deformations could sometimes be significant too [101]. Both 1J(C,H) and 3J(H,D') depend on C-atom hybridisation and bond or dihedral angles [102] and, therefore, any change in these parameters will affect the coupling. Conformations of cis-stilbene and trans-stilbene are a consequence of steric repulsions between ortho- and -protons and competitive -conjugation, as previously discussed, leading to a nonplanar atom arrangement in both isomers (Fig. (8)).

42

H H

 '

C H

H

H

C 1'

1

H

H 4

H

2 3

H

H

6'

6 5

H

H

5' 2' 3'

H

H H

H

3'

'

1'

C

H

2'

H

H 2

5 4

H

6'

1

6

4'



C

H

H

Fig. (8)

4'

5'

H

3

H

Molecular models (AM1) of cis-stilbene and trans-stilbene, and atom numbering.

By the semiempirical AM1 and ab initio methods, we calculated that the shortest of all the nonbonding proton-proton distances is the one in trans-stilbene involving ortho-' protons (H-6', H- or H-2, H-'). One should, therefore, expect larger secondary deuterium effects in -deuteriated trans isotopomers, which was confirmed in our experiments. The ortho- strain is released by using D instead of H, because a C-D bond is shorter than a C-H bond, leading also to a better transmission of isotope effect in the molecule. A greater relative change in 3J(H,D') than in 1J(C,H) indicates a higher sensitivity of the former to a subtle conformational change upon isotopic substitution. Therefore, changes in relative rotamer populations associated with different chemical bond lengths and dihedral angles, as well as rovibrational averaging over these, should account for the measured isotope effects on couplings.

43

6.6. Deuterium isotope effects on spin-lattice relaxation Isotopes with a spin greater than 1/2 possess, besides the dipole moment, the electric quadrupole moment. This influences the spin-lattice relaxation time (T1) of the neighbouring carbon nuclei. The effects of isotopes on relaxation have been scarcely reported in the literature owing to the complexity of relaxation mechanism [103]. To the best of our knowledge, there has not been any recent observation of the isotope effects on the relaxation times, except for that reported in ref. [104]. For a 13C nucleus, four possible relaxation mechanisms may occur (13):

1/T1obs = 1/T1DD + 1/T1CSA + 1/T1SR + 1/T1SC

(13)

where T1obs is the observed spin-lattice relaxation time, whereas T1DD, T1CSA, T1SR and T1SC refer to dipole-dipole (DD), chemical shift anisotropy (CSA), spin-rotation (SR) and scalar coupling (SC) contributions, respectively. Those carbons which bear a hydrogen atom are predominantly relaxed by the dipole-dipole mechanism. However, for carbons that have no directly attached hydrogen, and for those bearing a deuterium, other mechanisms may operate [105-107]. Deuteriation could drastically affect the spin-lattice relaxation time because of a much smaller magnetic moment of deuterium than that of hydrogen [104,108]. We have focused our attention on deuteriated isotopomers of benzophenone and have measured their T1 at the magnetic field of 7.0 T (Table 4)[104].

44

Table 4.

13

C spin-lattice relaxation times T1 (s) of benzophenone and its deuteriated

isotopomers. T1 C-atom

C-

C-1

C-1'

C-2

C-2'

C-3

C-3'

C-4

C-4'

BPN

31.00

25.87

25.87

3.45

3.45

3.41

3.41

2.57

2.57

o-D-BPN

31.62

28.12

25.80

10.88

3.10

3.01

3.03

2.32

2.32

D5-BPN

39.46

35.59

39.81

23.15

3.60

24.69

3.56

15.19

2.65

D10-BPN

41.99

38.37

38.37

22.62

22.62

21.93

21.93

16.80

16.80

isotopomer

Solvent was chloroform-d1. T = 293 K.

An increase of T1 by up to 550 % has been observed for carbon atoms in perdeuteriated benzophenone. Surprisingly, in o-D-benzophenone, besides the three-fold increase of T1 at C-2, a decrease in T1 for some neighbouring carbons was observed. It was not quite clear what had caused that decrease, but a change in the relaxation mechanism should have had an impact. We have also determined the Nuclear Overhauser effects (NOE) in benzophenone to obtain the dipole-dipole contributions to the total T1, and confirmed that carbons bearing hydrogen predominantly relax by the dipole-dipole mechanism. Measurements of T1 in a higher magnetic field (9.4 T) have shown that CSA gives a significant contribution to the relaxation rate of C- and C-1,1' as well as of carbons having a directly attached deuterium.

45

7. Conclusions

In the present article, we have reported recent observations and findings in the field of isotope effects in NMR spectra of organic molecules. The effects have been discussed in terms of the vibrational theory. However, in the absence of rigorous high-level calculations for moderate and larger size organic molecules, the established empirical trends and correlations provide a further insight into this phenomenon and could safely be used in structural investigations. Isotope effects on both nuclear shielding and coupling constants were related to structural parameters, such as hybridisation, bond-order, bond-length, bond-, torsional- or dihedral angles, conjugation, resonance, etc. In conjugated -electron molecules, isotopic perturbations were observed many bonds away from the isotopic site. Long-range isotope effects reflect the overall molecular geometry and also provide information on the transmission pathway. The observed sign and magnitude alternations of long-range isotope effects were accounted for by subtle charge shifts throughout the molecule as a consequence of bond shortening upon isotopic substitution. This could polarise the aromatic moiety and thus affect the magnitude and sign of the effect. Linear correlations between long-range isotope effects and molecular torsional angles were found. They could serve as a possible conformational probe in benzene derivatives of a related type. Isotope effects in molecules containing a heteroatom were modulated by lone-pairs, which usually decrease the magnitude and affect the sign of the isotope effect.

46

The additivity of isotope effects on nuclear shielding and coupling constant generally holds in the majority of cases. Deviations found in small molecules were attributed to secondary geometrical effects, or to quadratic corrections to secondary geometrical effects, while those observed in larger molecules, such as those containing phenyl rotors, were due to a rotamer redistribution or perturbation in hydrogen-bond dynamics upon isotopic substitution.

Acknowledgement We are indebted to Prof. P. E. Hansen, Prof. N. Müller, Dr. E. Gacs-Baitz, Dr. J. Plavec and Prof. H. Sterk for helpful discussions and/or providing some experimental measurements. Thanks are also due to Dr. G. Baranović for carefully reading the manuscript and for helpful comments. We thank Prof. A. P. Marchand for English improvements. This research was supported by the Ministry of Science and Technology of the Republic of Croatia (Project No. 00980802).

47

References

1. Batiz-Hernandez, H.; Bernheim, R.A. Prog. Nucl. Magn. Reson., 1967, 3, 63. 2. Berger, S. In NMR Basic Principles and Progress; Diehl, P.; Fluck, E.; Günther, H.; Kosfeld, R.; Seelig, J., Ed.; Springer-Verlag: Berlin, 1990; Vol. 22, pp. 1-29. 3. Berger, S.; Künzer, H. Angew. Chem., Int. Ed. Engl., 1983, 22,321. 4. Halevi, E.A. Prog. Phys. Org. Chem., 1963, 1, 109. 5. Wesener, J.R.; Günther, H. Tetrahedron Lett., 1982, 23, 2845. 6. Hansen, P.E.; Annu. Rep. NMR Spectrosc., 1983, 15, 105. 7. Hansen, P.E. Prog. Nucl. Magn. Reson. Spectrosc., 1988, 20, 207. 8. Forsyth, D.A. In Isotopes in Organic Chemistry, Buncel, E.; Lee, C.C., Eds.; Elsevier: Amsterdam, 1984, Vol. 6, 1. 9. Jameson, C.J. In Isotopes in the Physical and Biomedical Sciences; Buncel, E.; Jones, J.R., Ed.; Elsevier: Amsterdam, 1991; Vol. 2, pp. 1-54. 10. Siehl, H.-U. Adv. Phys. Org. Chem., 1987, 23, 63. 11. Sergeyev, N.M.; In NMR Basic Principles and Progress; Diehl, P.; Fluck, E.; Kosfield, R.; Günther, H., Ed.; Springer-Verlag: Berlin, 1990; Vol. 22, pp. 31.-80. 12. Jameson, C.J.; Osten, H.J. Annu. Rep. NMR Spectrosc., 1986, 17, 1. 13. Jameson, C.J. In Specialist Periodical Reports. Nuclear Magnetic Resonance; Webb, G.A., Ed.; Royal Society of Chemistry: Cambridge, 1994; Vol. 23, pp. 47-78. 14. Wesener, J.R.; Moskau, D.; Günther, H. J. Am. Chem. Soc., 1985, 107, 7307. 15. Wesener, J.R.; Moskau, D.; Günther, H. Tetrahedron Lett., 1985, 26, 1419. 16. Berger, S.; Diehl, B.W.K. Tetrahedron Lett., 1987, 28, 1243. 17. Balzer, H.H.; Berger, S. Magn. Reson. Chem., 1990, 28, 437. 18. Everett, J.R. J. Chem. Soc., Perkin Trans. 2, 1984, 1151. 19. Sergeyev, N.M.; Sergeyeva, N.D.; Raynes, W.T. Chem. Phys. Lett., 1994, 221, 385.

48

20. Sergeyev, N.M.; Sergeyeva, N.D.; Raynes, W.T. Magn. Reson. Chem., 1994, 32, 381. 21. Novak, P.; Vikić-Topić, D.; Meić, Z.; Gacs-Baitz, E. Magn. Reson. Chem., 1996, 34, 610. 22. Jameson, C.J.; Osten, H.-J. J. Phys. Chem., 1984, 81, 4293. 23. Uvarov, V.A.; Chertkhov, V.A.; Sergeyev, N.M. J. Chem. Soc. Perkin Trans. 2, 1994, 2375. 24. Sergeyev, N.M.; Sergeyeva, N.D.; Raynes, W.T. J. Chem. Soc., Chem. Commun., 1994, 485. 25. Morales-Rios, M.S.; Joseph-Nathan, P. Magn. Reson. Chem., 1991, 29, 49. 26. Morales-Rios, M.S.; Joseph-Nathan, P.; Wrackmeyer, B.; Kupče, E. Magn. Reson. Chem., 1993, 31, 238. 27. Morales-Rios, M.S.; Perez-Alvarez, V.; Joseph-Nathan, P.; Zepeda, L.G. Magn. Reson. Chem., 1994, 32, 288. 28. Nakashima, Y.; Takahashi, K. Bull. Chem. Soc. Jpn., 1991, 64, 3166. 29. Nakashima, Y.; Kanada, H.; Fukunaga, M.; Suzuki, K.; Takahashi, K. Bull. Chem. Soc. Jpn., 1992, 65, 2894. 30. Nakashima, Y.; Sandori, B.; Itoh, S.; Takahashi, K. Bull. Chem. Soc. Jpn., 1991, 64, 3694. 31. Mlinarić-Majerski, K.; Vinković, V.; Meić, Z.; Gassman, P.G.; Chyall, L.J. J. Mol. Struct., 1992, 267, 389. 32. Mlinarić-Majerski, K.; Vinković, V.; Chyall, L.J.; Gassman, P.G. Magn. Reson. Chem., 1993, 31, 903. 33. Vinković, V.; Mlinarić-Majerski, K.; Marinić, Ž. Tetrahedron Lett., 1992, 33, 7441. 34. Ogata, M.; Tahara, T.; Nishimura, T. J. Pharm. Sci., 1995, 84, 186. 35. Forsyth, D.A.; Moussa, A.M.; Panyachotipun, C. Croat. Chem. Acta, 1992, 65, 703. 36. Stoelting, D.T.; Forsyth, D.A., Fry, J.L. J. Org. Chem., 1995, 60, 2841. 37. Servis, K.L., Koh, E.V., Baine, P. Croat. Chem. Acta, 1992, 65, 679.

49

38. Siehl, H.V., Fuss, M.; Gauss, J. J. Am. Chem. Soc., 1995, 117, 5983. 39. Yonemitsu, T.; Tsuzuki, H.; Mukumoto, M.; Mataka, S.; Tasiro, M. In Synthesis and Applications of Isotopically Labelled Compounds; Allen, J., Ed.; John Wiley & Sons, 1995, pp. 487-488. 40. Heintz, R.A.; Neiss, T.G.; Theopold, K.H. Angew. Chem. Int. Ed. Engl., 1994, 33, 2326. 41. Anet, F.A.L.; Park, J. J. Am. Chem. Soc., 1992, 114, 411. 42. Medforth, C.J.; Shiau, F.Y.; La Mar, G.N.; Smith, K.M.J. Chem. Soc., Chem. Commun., 1991, 590. 43. Brown, K.L.; Zou, X.; Webb, B.M. Inorg. Chem., 1994, 33, 4189. 44. Everett, J.R. J. Chem. Soc., Chem. Commun., 1987, 1878. 45. Sauer, S.P.A.; Špirko, V.;Paidarova, I.; Oddershede, J. J. Chem. Phys., 1994, 184, 1. 46. Lutz, O.; Oehler, H. Z. Naturforsch., 1977, 32a, 131. 47. Mateescu, G.D.; Benedikt, G.M.; Kelly, M.P. In Synthesis and Applications of Isotopically Labelled Compounds; Duncan, W.P.; Susan, A.B., Eds.; Elsevier, Amsterdam, 1983. 48. Jameson, C.J.; Osten, H.-J. J. Am. Chem. Soc., 1985, 107, 4158. 49. Kaspersen, F.M.; Vader, J.F.; Funke, C.W.; Sperling, E.M.G. Recl. Trav. Chim. Pays-Bas, 1993, 112, 191. 50. Risley, J.M. Magn. Reson. Chem., 1991, 29, 143. 51. Arias, W.; Risley, J.M. J. Org. Chem., 1991, 56, 3741. 52. Cortes, S.J.; Mega, T.L.; Van Etten, R.L. J. Org. Chem., 1991, 56, 943 53. Okuyama, T.;

Senda, K.;

Takano, H.;

Ando, N.;

Ohnishi, K.;

Fueno, T.

Heteroatom. Chem., 1993, 4, 223 54. Groziak, M.P.; Koohang, A.; Stevebs, W.C.; Robinson, P.D. J. Org. Chem., 1993, 58, 4054. 55. Aliev, A.E.; Harris, K.D.M. Magn. Reson. Chem., 1993, 31, 54. 56. Raynes, W.T.; Grayson, M.; Sergeyev, N.M.; Sergeyeva, N.D. Chem. Phys. Lett., 1994, 226, 433.

50

57. Morales-Rios, M.S.;

Garcia-Martinez, C.;

Joseph-Nathan, P.;

Zepeda, L.G.

Magn. Reson. Chem., 1995, 33, 149. 58. Sergeyev, N.M.; Sandor, P.; Sergeyeva, N.D.; Raynes, W.T. J. Magn. Reson. Chem., Series A, 1995, 115, 174. 59. Hansen. P.E. Ph.D. Thesis; Isotope Effects on Chemical Shifts as a Tool in Structural Studies, Roskilde University Press, 1996. 60. Kozersky, L. J. Mol. Struct., 1994, 321 89.. 61. Rospenk, M.; Koll, A.; Sobczyk, L. J. Mol. Liq., 1995, 67, 63. 62. Rospenk, M.; Koll, A.; Sobczyk, L. Chem. Phys. Lett., 1996, 261, 283. 63. Krasnokutskii, S.N.;

Kurkovskaya, L.N.;

Shibanova, T.A.;

Shabunova, V.P.

J. Struct. Chem., 1991, 32, 10. 64. Penman, K.G.; Kitching, W.; Wells, A.P. J. Chem. Soc. Perkin Trans. 1, 1991, 721. 65. Lee, K.B.; McLachlan, S.J.; La Mar, G.N. Biochem. Biophys. Acta, 1994, 1208, 22. 66. Contreras, R.H.; Facelli, J.C. Annu. Rep. NMR Spectrosc., 1993, 27, 331. 67. Hansen, P.E.; Led, J.J. Org. Magn. Reson., 1981, 15, 288. 68. Everett, J.R. Org. Magn. Reson., 1982, 19, 169. 69. Luzikov, Y.N.; Sergeyev, N.M. J. Magn. Reson., 1984, 60, 177. 70. Strelenko, Y.A.; Torocheshnikov, V.N.; Sergeyev, N.M. J. Magn. Reson., 1990, 89, 123. 71. Leshcheva, I.F.;

Torochesnikov, V.N.;

Sergeyev, N.M.;

Chertkhov, V.A.;

Khlopkov, V.N. J. Magn. Reson., 1991, 94, 1. 72. Leshcheva, I. F.;

Torochesnikov, V.N.;

Sergeyev, N.M.;

Chertkhov, V.A.;

Khlopkov, V.N. J. Magn. Reson., 1991, 94, 9. 73. Sergeyev, N.M.; Solkan, V.N. J. Chem. Soc., Chem. Commun., 1975, 12. 74. Raynes, W.T.; Geertsen, J.; Oddershede, J. Chem. Phys. Lett., 1992, 197, 516. 75. Benett, B.; Raynes, W.T.; Anderson, C.W. Spectrochim. Acta, Part A, 1989, 45, 821. 76. Heckman, G.; Fluck, E. Magn. Reson. Chem., 1995, 33, 553.

51

77. Young, W.R.; Yannoni, C.S. J. Am. Chem. Soc., 1969, 91, 4581. 78. Vujanić, P.; Gacs-Baitz, E.; Meić, Z.; Šuste, T.; Smrečki, V. Magn. Reson. Chem., 1995, 33, 426. 79. Smrečki, V.; Müller, N.; Vikić-Topić, D.; Vujanić, P.; Meić, Z. J. Mol. Struct., 1995, 348, 69. 80. Meić, Z.; Novak, P.; Vikić-Topić, D.; Smrečki, V. Magn. Reson. Chem., 1996, 34, 36. 81. Vikić-Topić, D.; Novak, P.; Smrečki, V.; Meić, Z. J. Mol. Struct, 1997, in press. 82. Novak, P. Doctoral Thesis, University of Zagreb, 1995. 83. Bell, R.A.; Chan, C.L.; Sayer, B.G.J. Chem. Soc., Chem. Commun., 1972, 67. 84. Aydin, R.;

Wesener, J.R.;

Günther, H.;

Santillan, R.L.;

Garibay, M.E.;

Joseph-Natan, P. J. Org. Chem., 1984, 49, 3845. 85. Majerski, Z.; Žuanić, M.; Metelko, B. J. Am. Chem. Soc., 1985, 107, 1721. 86. Aydin, R.; Frankmölle, W.; Schmalz, D.; Günther, H. Magn. Reson. Chem., 1988, 26, 408. 87. Novak, P.; Vikić-Topić, D.; Meić, Z.; Sekušak, S.; Sabljić, S. J. Mol. Struct., 1995, 356, 131. 88. Künzer, H.; Berger, S. J. Am. Chem. Soc., 1985, 107, 2804. 89. Berger, S.; Diehl, B.W.K. J. Am. Chem. Soc., 1989, 111, 1240. 90. Vikić-Topić, D.; Becker, E.D.; Hodošček, M.; Graovac, A. Croat. Chem. Acta, 1995, 68, 193. 91. Novak, P.; Meić, Z.; Vikić-Topić, D.; Sterk, H. J. Mol. Struct., 1997, in press. 92. Baird, M.C. J. Magn. Reson., 1974, 14, 117. 93. Rossi, M.; Colleta, F.; Rigatti, G. Lincei. rend. Sci. Fis. Mat. Nat., 1972, 12, 80. 94. Vujanić, P., Meić, Z.; Vikić-Topić, D. Spectrosc. Lett., 1995, 28, 395. 95. Berbard, C.; Mouchel, B.; Sueur, S. J. Chem. Soc., Chem. Commun., 1982, 300. 96. Wasylishen, R.E.; Mulchew, D.H.; Friesen, K.J. J. Magn. Reson., 1980, 41, 341. 97. Sano, M.; Yoshikawa, Y.; Yamatara, H. Inorg. Chem., 1982, 21, 2521. 98. Jameson, C.J.; Osten, H.-J. J. Am. Chem. Soc., 1986, 108, 2497.

52

99. Smrečki, V.; Vikić-Topić, D.; Meić, Z.; Novak, P. Croat. Chem. Acta 1996, 69, 1501. 100. Novak, P.; Meić, Z.; Sterk, H. J. Chem. Soc., Perkin Trans. 2, 1996, 2531. 101. Lazzeretti, P.; Zanasi, R.; Raynes, W.T. J. Chem. Soc., Chem. Commun., 1986, 57. 102. Karplus, M. J. Am. Chem. Soc., 1963, 85, 2870. 103. Levy, G.C.; Lichter, R.L.; Nelson, G.L. Carbon-13 Nuclear Magnetic Resonance Spectroscopy, Wiley Interscience: New York, 1980. 104. Vujanić, P., Meić, Z. J. Mol. Struct., 1993, 293, 163. 105. Olivson, A.; Lipmaa, E. Chem. Phys. Lett., 1971, 11, 241. 106. Spiess, H.W.; Schweitzer, D.; Haeberlen, U. J. Magn. Reson., 1973, 9, 444. 107. Shapiro, M.J.; Kahle, A.D. Org. Magn. Reson., 1979, 12, 4. 108. Echols, R.E.; Levy, G.C. J. Org. Chem., 1974, 39, 1321.