Shift Reagents in NMR Spectroscopyweb.missouri.edu/~glaserr/4210f20/SHIFT-Reagents_anie.19720675… · magnetic shift AF of a specified proton of the sub- strate s.If S = cyclohex~lisonitrile,

Shift Reagents in NMR Spectroscopy

By Reinhard von Ammon and R. Dieter Fischer‘*’

The number of possible applications of N M R spectroscopy has rapidly increased during the past few years. New fields of applications have been opened by the development of supraconducting solenoids and various spin-decoupling techniques and by the method of ‘‘pulsed Fourier transform NMR-spectroscopy”. These methods originate mainly from progress in instrumentation. Recently, another “technique” has been introduced into N M R spectroscopy, which-principally on the basis of chemical and spectroscopic experience- is much less expensive but nevertheless useful. The basic principles, background, and most important applications of this method, known as the “NMR-shift-reagent technique”, form the subject of this paper.

1. Introduction

An NMR-spectrum can be analyzed the more easily the more pronounced the chemical shifts of the nuclei of a sample are in comparison to the spin-spin coupling effects. The ideal case is a first order spectrum where the shift difference of the multiplet centers are large compared to the components of the multiplets. Possible means of simplifying spectra are spin decoupling“] or- in certain cases-the introduction of suitable isotopes into the molecule under investigation. Another way is the variation of the magnetic field acting on the various nuclear magnets. If we discern between increasing the homogeneous, external magnetic field of

the spectrometer and superimposing the “primary field” by additional internal “secondary fields”, two alternative possibilities are evident.

A consequence of choosing the first possibility is the aim of commercial suppliers to produce N M R spec- trometers with higher field strengths and improved resolution. The present status has-by means of supraconducting solenoids-reached field strengths corresponding t o proton resonance frequencies of approximately 300 MHz.

The development of the other alternative is based upon experience of the chemist with materials whose magnetic properties are strongly anisotropic“]. An example is the ring current effect causing chemically not very different

5 L 3 2 1 4 t- 6Cppml position increasing.fieid

of strength - I TMS

, ‘ 4 1 2 3 L 5 6 7 8 9

posltlon 6 [ppml- increasing field of

TMS strength- -

n-C8H1’7 I

Fig. I . (a) ‘H-NMR spectrum of n-octanol (90 MHz); the &scale (ppm) referred to TMS, solvent CCl,; (b) ‘H-NMR spectrum (100 MHz) of the n-octyl part of dioctvlgermanium-porphin I21: &scale referred to TMS, solvent CDCI,.

[*I Dr. R. v. Ammon protons t o exhibit a large separation of their resonance lines. Thus, in the diamagnetic porphin ring of the compound shown in Fig. 1, the primary field induces

1*1 A secondary field is anisotropic if its magnitude and direction change with the direction of the primary field.

Institut fur Heisse Chemie des Kemforschungszentrums 7 5 Karlsruhe, Postfach 3640 (Germany) Prof. Dr. R. D. Fischer Institut fur Anorganische Chemie der Universitat Erlangen- Niirnberg 852 Erlangen. Egerlandstrasse (Germany)

_ _ ~

Angew. Chem. internat. Edit. Vol. I 1 (1972) 1 No. 8 675

an additionalanisotropicfield which acts very differently upon the methylene proton pairs of the two axially coordinated n-octyl chains‘”. Similar signal spreadings are caused by the magnetic anisotropy of the phthalocyanine ( = Pc) system in adducts of the type Fe(II)Pc. 2s (S = amines etc.)[1261 or by the weakly paramagnetic uranyl ion[)’.

Any perceptible change of the structure of the substrate during the adduct formation is, however, to be avoided.

The molar ratio [SI : [R] does not necessarily have to be 1 : 1; it is important, however, that the mean lifetime of the adduct S-R is sufficiently small on the N M R time scale that only one averaged spectrum is observed for S and S-R.

6

1.3 h i “ I

I I

4 3 2 1 0 200 120 11.0 6 0 50 L.0

Fig. 2. ‘H-NMR spectra (60 MHz) of 2-adamantanol (a) before and (b) after addition of Eu(dpm), as shift reagent: molar ratio shift reagent: substrate =0.67, solvent CDCI, [4].

- 6 t ppml t- S [pprnl rzzczl Much stronger secondary fields and therefore a still more pronounced expansion of the spectra can be induced, however, in paramagnetic systems with at least one unpaired electron. Fig. 2 illustrates the changes in the ‘H-NMR spectrum of adamantanol caused by the interaction of this substrate with the paramagnetic “shift reagent” tris(dipivalomethanato)europium(m), Eu(dpm), (cf. Fig. 7), a system with six unpaired 4f e Iec t ron~[~’ .

A further illustration is the H-NMR spectrum of ( - 1- nicotine recorded at 6OC5] and 220 MHzI6] in the absence (Fig. 3 a and 3b) and at 90 MHz (Fig. 3c) in the presence of the organometallic shift reagent tricyclo- pentadienylpraseodymium(II1) I7].

Whereas the enhancement of the external primary field causes a linear expansion of the spectra leaving the relative chemical shifts (in ppm) unchanged, the internal secondary field may also change the sequence of the signals. Since the strength of this field decreases with increasing distance from the origin of the magnetic field lines, “geometric shifts” (see below) become the ordering principle of the spectrum instead of the familiar chemical shifts in all cases where the influence of the secondary field is strong.

One may speak of an N M R shift reagent (R) if the N M R spectrum of a substance S is simplified (with respect to spectral analysis) through interaction with the reagent. The interaction between S and R should be prompted by addition of the reagent to the sample in the sense of a t least one of the three equilibria (1 a)-(l c). S(11 + Rlf) + SIIIRII) (1 a)[‘) Sli)Rll + Sl2) + Sl2)RllJ + Sll) (1b)l’l S(i1RltJ + R(21 + S(tIRl21 + Rll) (lc)l*1

1’1 For simplicity only 1 : 1 adducts are considered at the monent (cf. Section 3.2.1.2).

90 80 70 / I

I ,

8 0 70 40 30 20

li --- 6 [ p p m l

9 8 7 J j 3 2 - 6 [ ppml C I

3.4

I < , i

5 10 1s ZO”50 55 6 [pprn] -+ incre?sing 1ie.d

cxEiiz s1:eyth 4 Fig. 3. ‘H-NMR spectra of (-)-nicotine, recorded under various conditions: (a) 60 MHz in CCI, [S]; (b) 220 MHz in CDC1, referred to TMS [61; (c) 90 MHz in D,-toluene after adduct formation with an esuimolar amount of Pr(C,H,),: F-scale referred to C,H, 171.

676 Angew. Chem. internat. Edit. 1 Vol. I 1 (1972) / N o . 8

2. Theory

Although the principle of the shift reagent technique has been known for some time and, in several cases, has been applied before[8-l’.’551, it was not until 1969 that Hinckley“” initiated the systematic development of this branch of N M R spectroscopy [’ 3. 14.127, 128,1 58. 1601. The following survey of the theoretical background, which has been treated elsewhere in several detailed article^"^ - 2 1 1 , concentrates on the most important as- pects necessary for an understanding of the quantitative correlations between the properties of a shift reagent, the substrate, and the observed spectra.

2.1. The “Isotropic” Shift

The contribution Atso of an internal secondary field to the resonance frequencies 6, of a substance S within the adduct S-R in solution is experimentally obtained as the difference of the resonance frequencies under the influence of and without the secondary field:

Although the internal secondary field is practically always anisotropic, A:,, is generally called the “isotropic” shift, because it is the averaged result over all orienta- tions of a molecule moving freely in solution. In the case of a paramagnetic shift reagent the term “paramagnetic shift” is also used.

If the shift reagent does not contain unpaired electrons, A? is identical with the dipolar or “pseudocontact” term .PIR, which takes into account all magnetic dipolar interactions of the magnetic moment of the ith nucleus with theanisotropicmagneticfield ofthevalenceelectron shell of the shift reagent. If unpaired electrons are present, A? is composed of the dipolar contribution and a (Fermi-)contact term A? taking into account the influence of a possible spin delocalization within the adduct S-R:

2.1.1. Magnetic Dipolar Interactions

It often seems justified to relate the internal secondary field of a molecule in solution for any orientation rela-

Fig. 4. Coordinates r , , 9, and o, of a nucleus i in the coordinate system x, Y, z with the three principal components x , , x y , and x, of the magnetic susceptibility.

tive to the direction of the external primary field to a magnetic “point dipole” of well defined position within the molecule. (The center of this point dipole is generally at the nucleus of a paramagnetic transition metal ion, but not necessarily a t the center of a cyclic n-electron system.) A!’. can then be expressed as a function of the internal molecular coordinates r , , 9, and w, (see Fig. 4) and the three principal components xx, x y and xr of the measurable magnetic suscepti-

( N , = Loschmidt’s constant).

The equations (Sa)-(6b) are valid for certain types of relaxation conditions which are particularly frequently encountered L221.

Particularly important for the shift reagent technique is the special case of an axially symmetric secondary field. defined by: xZ-x ; X ~ = X , , - X ~ . This case applies, for instance, to metal complexes with axially symmetric ligand fields, but also to the example in Fig. 1 b. Since D”=O, equation (4) reduces to

(7) Ap‘“(ax) = - D G ,

Contrary to the non-axially symmetric case the A:’Yax)- values of a system are directly proportional to the corresponding “geometric factors” G,:

A?? A F : A F etc. = G , : G,: G, etc. (8)

Unfortunately, the rigorously axially symmetric case is rarely met in practice[’61.’621. However, in a number of S-R systems with lower symmetry the second term in equation (4) should average to zero because of the rotations of S relative to R which are often very fast on the NMR time scale. It is, nevertheless, most essential that-in accordance with equations (4) and (7)---A.PtR is proportional to r;3 in all cases (cf. Section 4.1).

For the case of a non-axially symmetric complex where the substrate rotates freely around an axis running through the metal ion, the following esuation has been proposed11621 :

A“’” = ---I- 1 D‘(3cos‘a- 1 ) - D“sinLacos2~1(3cos2y- I ) (9) 2 r’

The angle y is included by the radius vector r, (Fig. 4) and the rotation axis of the substrate; a and refer to the position of the rotation axis in analogy to 9 and a.

The spectral shifting potential of a shift reagent is determined by its anisotropy parameters D‘ and D”, which are usually temperature d e ~ e n d e n t ” ~ ~ ’ . In the series of axially symmetric, organometallic shift reagents of the type (C5H5)3Ln(~~~)127,1441 with Ln = La through Lu, the parameter D’(ax) is proportional to the para-

Angew. Chem. internal. Edil. / Vol. I 1 (1972) / No. 8 677

magnetic shift A F of a specified proton of the substrate s. If S = cyclohex~lisonitrile, the variation of D’(ax) = - A Y / G , is quite pronounced (Fig. 5 ) .

A similar periodicity has been found for a series of homologous chelate adducts L n ( d ~ m ) , S ~ [ ’ ~ ’ . The same very characteristic dependence of AdiP on the number of f electrons of Ln was confirmed with the magnetic anisotropy parameters very recently determined for the crystalline adducts[1621. It is remarkable that, according to available data for homologous series of lanthanoid complexes, the ions of each of the three “triads” [Ce, Pr, Ndl, ITb, DY, Hol, and [Er, Tm, Yb] induce dipolar shifts of the same signf2”. Bleaney et al.”631 have just recently given a theoretical explanation for this periodicity.

1 2 -

11

10

1 - 5 7 - Y

> 6 - 5- r , -

3 -

2 -

1 -

-20 I601 I l l 1 l l l l l i 1 1 ‘ l

la Ce Pr Nd Pm SmEu Gd Tb Dy Ho Er TmYb l u l7zzEJ

+01 +O 075 -0 05 -

01 -

+o 2

-0 2 -03 -0 4 -0 5

-1 0

Fig. 5. Isotropic shifts (in ppm) of the C,H,,-protons in &position of the adducts between cyclohexylisonitrile and the lanthanoid- (iii)-tricuclopentadienides, Ln(C,H,), 171. Solvent D,-toluene: T = 302’K. 90 MHz. The linewidths at half-height (Hz) are given in parentheses.

The sign of A?’. can often be predicted more easily than its absolute magnitude: thus, in the example of Fig. 1 b we have x,,,xI < 0 (diamagnetism) and in the case of the UO;’ ion xI > 0 (temperature-independent paramagnetism). Further, G, > Oand G, < 0, respectivel~, forthetwoexamples. IfX,i

The calculation of A, is even more difficult than the calculation of D’ and D”*261. Since there is n o means of correlating the contact shift with geometric coordinates, as was shown for the pseudocontact shift, and since a reliable factorization of the experimental shift into the dipolar and contact contributions is not possible in many cases, noticeable contact contributions in the shift reagent technique usually prevent quantitative structural evaluations.

Paramagnetic lanthanoid complexes fulfil the requirement of negligible contact interactions particularly

because the radial extension of the 4 f orbitals is exceedingly small. According t o experience, however, it is safe to neglect contact contributions even in 4f- systems only if a t least three atoms (i. e. four bonds) are situated between the nucleus under investigation and the paramagnetic center. From extensive ”0- and ‘ H - N M R studies of lanthanoid a q u o - c o m ~ l e x e s [ ~ ~ ’ it was concluded that AEn%AgD holds for the 0-a toms in the first coordination sphere, whereas the overall shift At;O of the neighboring H-atoms still contains a contact contribution of a t most 60%.

2.2. Dynamic Effects

The equations given in Section 2.1 allow quantitative predictions as to magnitude and sign of the isotropic shifts; they d o not, however, contain a term including the linewidth allowing predictions referring to the chances of a signal to be observed. Unfortunately, drastic signal broadenings have t o be expected frequently in paramagnetic samples containing unpaired electrons. These may be caused not only by intramolecular spin delocalization, but also by time-depend- ent intermolecular in te rac t i~ns[*~l .

These undesirable line broadenings may be partially compensated for if the mean lifetime of the electronic spin states (electron spin-lattice relaxation time Tie) is drastically reduced with respect to the correlation time for the thermal motion of the molecules. Besides extremely low T,, values, rapid changes of the ligand field at the paramagnetic metal ion can also give rise to narrow linewidths.

The term T,’ clearly corresponds to the rate of trans- forming the excitation energy of a discrete spin level into thermal energy of the surrounding “lattice” (i. e. coordination + solvation sphere). T,’ increases with increasing temperature and with the strength of the external magnetic field, and, a t the same time, with increasing magnitude of the spin-orbit coupling constant, but with decreasing energetic difference between the electronic ground state and the next excited ( = ligand field) states. The last two requirements are again fulfilled particularly well by systems containing unpaired 4 f o r 5 f electrons.

Apart from the spin-only system G d 3 + (f’), lanthanold and actinoid complexes actually exhibit N M R linewidths between a few Hz and several hundred H2“4,30,’651 According to systematic study of the linewidths in the

spectra of the adducts Ln(dpm),S,by H o r r ~ c k s “ ~ ’ and according to general experience complexes of Eu and Pr exhibit the sharpest signals. On the other hand, the signals of most d”-transition metal complexes are sufficiently narrow only in exceptional cases (Coz+ and Ni2+). The detection of the N M R signals of paramagnetic organic radicals is usually prevented by their excessive linewidths.

Chemically, the lifetime of the relevant electronic states can also be reduced by rapid reactions of type (1 a)-(1 c). Signal narrowing is also t o be expected if the adduct S-R is able to alternate between several conformations (S-R)’, (S-R)”. . . etc.

3. Systematic Approach t o Shift Reagents

3.1. General Requirements

An effective and versatile shift reagent should fulfil at least four requirements:

1 . Optimal shifting potential combined with minimal signal broadening effect;

2. ability to interact with a great number of substrates, i. e. marked acidic or basic properties;

3 . sufficient solubility;

4. absence of interfering absorptions in the usual range of substrate spectra.

Ideal, universally applicable shift reagents cannot be realized: Lewis-acid shift reagents are effective only with basic substrates (and vice versa). High reactivity (for instance towards extremely weak bases) is not compatible with the chemical stability required at the same time. A high reactivity even excludes the use of polar solvents, since their functional groups would compete with the substrate.

Requirements 1 to 4 d o not yet take into consideration that quantitative information about the molecular structure o r the conformational behavior of a substrate is sometimes desirable. With regard to such studies the following requirements should also be fulfilled:

5 . Pseudocontact should dominate over Fermi-contact interactions;

6 . the secondary fields should have axial symmetry, and the point-dipole approximation should be valid;

7. theadduct S-R should be conformationally uniform (whereas the substrate itself may have several conformations);

8. a straightforward temperature dependence of the isotropic shifts ( e . g . temperature-independence o r linear variation of Also with l/n.

Whereas postulates 5 to 7 are the prerequisites for the applicability of equations (4) to (8). postulate 8 is particularly advantageous for conformational studies within wide temperature ranges1271. This last point, however, has rarely been found to be truec9’, 123, I z 4 . 1521.

Angew. Chem. internat. Edit. 1 Vol. I 1 (1972) 1 No. 8 679

3.2. Types of Shift Reagents Applied

3.2.1. Coordination Compounds Containing Unpaired d o f f Electrons

Although numerous examples of the application of dn- complexes as shift reagents have become

, coordination compounds of the rare earth elements have more favorable properties.

The special features of the 4f radial function^"^] of these elements result in an optimal fulfilment of requirements I and 5 . In addition, the characteristic flexibility of the rare earth ions with regard to their preferential coordination number"461 ensures the spontaneous formation of sufficiently stable adducts.

With respect to their various applications and their chemical nature we classify shift reagents containing unpaired d or f electrons as typical salts, chelates, and organometallics.

known [8 - 11.34 - 38.129- 131,1551

3.2.1.1. Salts

Most representatives of this class are simple salts of transition element or lanthanoid ions forming solvated ions-mainly in aqueous solution-which are capable of additionally coordinating a polar substrate (cf. Table

Table 1. Some examples for the application of shift reagents of the salt type to studies on substrates in aqueous solution.

Shift reagent Substrate Ref.

COCI, n-oropanol [81 CUCI, carnosine [401 (FeNH,),(SO,), amino acids I411 Eu(NO,);6D2O organophosphorus compounds I421 Ln(CIO,),.6 H,O amino acids, carboxylic acids [431

EuCI,. 6 H ,O carboxylic acids [441 EuCI, ribose 5-phosphate I1151 EuCl, mononucleotides 11 561

(Ln : Pr. Eu)

1 ) . Shift reagents of this type are suited only in rare cases to quantitative spectral analyses because of inad- equate fulfilment of postulations 1, 6, and 7"15.1561. Such reagents had already been employed for more than a decadeL8' before Hinckley's first paper[IZ1.

3.2.1.2. Chelates

This category comprises the systems best known and most widely used, including also Hinckley's original reagent Eu(dpm), . 2 p ~ r " ' ~ . All representatives of this class (Fig. 7) constitute b-diketonato complexes of the general formula L,,MX, (L=chelate h a n d ) . In the case of transition metal complexes L2 M at best two coordination sites are available for the substrate. Lan- thanoid complexes L3MX, with n = O provide several free, acidic coordination sites. Because of the varia- bility of the coordination number of rare earth e l e r n e n t ~ " ~ ~ ~ ' ~ ' ~ , complexes with n = 1 or 2 may still exhibit at least one free coordination site, even in the case of Hinckley's reagent (M = Eu, X = pyridine, n = 2, formal coordination number: 8).

Voluminous substituents of simple structure are prefer- ably chosen as substituents R (Fig. 71, e. g . CH,, C6HS, n-C3F7 or C(CH3),. Thus postulations 3, 4, and 7 are accounted for.

dpm or tmhd: K'= R = -C(CH3), fod: 12'= -C(CH3)3, R = -CF2-CF2-CF3

Fig. 7. General structure of shift reagents of the chelate type.

The accumulation of voluminous substituents R should limit the number of possible conformations of the shift reagent, especially if m = 3[451. It also diminishes the affinity of the reagent toward water, which, if absorbed, decreases its shifting a b i l i t ~ ' ~ ~ - ~ ~ ' . Fiuorinated substi-

enhance the solubility and -presumably through inductive effects-the acceptor strength of the reagent.

For the chelate ligand with R' = R" = C(CH,)3f501 the notation "dpm" (from: "dipivalo-methanato-") or "tmhd" (from: "tetramethylheptanedionato-") has come into current use in the literature. The ligand with R' = C(CH,), and R" = n-C3F7[49.1541 is written "fod" (from: "heptafluoro-dinieth4 Ir,ctanetlionato-"). "Fod" complexes are also in use in their perdeuterated forms (postulate 4). A good account of the practical problems connected with the application of lanthanoid shift reagents is given in Ref.[1271.

The composition of the adducts formed in solution between a shift reagent and a monofunctional substrate is by no means uniform. Obviously it also depends on the nature of the substrate. Thus, the existence of 1 : 1 a d d u c t ~ [ ~ ' ~ . ' ~ ~ ] and 2 : 1 a d d u ~ t s [ l ~ ~ ~ has definitely been proven. So far little is known about the conformations of the adduct in solution. According to recent observa- tions it cannot be excluded that the preferential conformation of an adduct is influenced by the nature of the coordinated s~bs t ra te l"~ .

In the crystalline state evidence exists for well-defined adducts between tris(P-diketonat0)lanthanoid complexes and 1 to 4 additional ligands, especially ~ a t e r ~ ' ~ ~ . ' ~ ~ ] . X-ray structure analyses of the 1 :2 adducts Ho(dpm),- 2 (4-p i~ol ine) '~~I and Eu(dpm),-2 pyr1871 provided evidence that the two substrate molecules d o not have coaxial positions in the adduct. The symmetry of the complexes is only C2"621. Analogous results are provided by the known structures of several crystalline 1 : 1 ad duct^"^^'. If these results would hold also in solution, axially symmetric ligand fields at the central ion could be expected only in rare cases.

Most chelate reagents are more o r less hygroscopic and chemically unstable towards weak acids such as carbonic acids and phenols, and sometimes even alcohols[1641. They are, nevertheless, the shift reagents most widely used today and are certainly the ones best suited for standard applications.

tuents149. 1571

680 Angew. Chem. internal. Edit. 1 Vol. I I (1972) I No. 8

3.2.1.3. Organometallics diamagnetism ( e . g . in cyclic n-electron systems, Fig. 1 b) or through temperature-independent paramagnetism

The tricyclopentadienyl complexes of several lanthan- ( e . g . UO;'). Another shift reagent of this class is oids ( C , H , ) , L ~ ( I I I ) * ~ ~ ~ belong to this category. These iron(Ir)phthalocyanine, FePc, which is paramagnetic. compounds can coordinate only a single unidentate On axial coordination of two bases like amines, how- ligand S: ever, the resulting quasi-octahedral, low-spin complex

FePcS, is diamagnetic"26'.

(C,H,),Ln + S =$ (C,H,),LnS

The acceptor strength of these compounds generally exceeds the acidity of the chelate Thus it is often possible to isolate well-defined 1 : 1-adducts free of the components. On the other hand, the sensi-- tivity of the organometallic shift reagents towards acids, water, and oxygen exceeds the sensitivity of the chelate complexes considerably. Therefore they have to be manipulated in an inert gas atmosphere and in carefully purified solvents.

Apart from the fact that the shifting potential of the tricyclopentadienyl complexes sometimes exceeds the corresponding chelate complexes, a particular advan- tage of many adducts (C5Hs)3LnS is the C3, point symmetry of the effective ligand field a t the central ion"'. Such systems fulfil postulates 6 and 7 and thus often allow quantitative evaluations of the observed spectra with respect to refined structural and conforma- t ionaI analyses [7.27. 1441.

Although the solubility of the free complexes (C,H,),Ln in inert media like benzene o r toluene is often rather limited, the solubilities of the adducts often turn out to be quite high, depending on the substrate 1e.g. (C,H,),Nd: ca. 1 mg/ml; (C,H,),Nd .nicotine: ca. 100 mgiml].

3.2.2. Shift Reagents without Unpaired d or f Electrons

3.2.2.1. Metal-free Radical Systems

Organic nitroxide radicals often exhibit electron spin- lattice relaxation times which-in contrast to other stable organic radicals-allow relatively sharp N M R signals to be observed. Therefore utilization of systems like (CH,),NO or [(CH3),CI2NO as shift reagents was a t t e m ~ t e d [ ' ~ . ~ ~ ] . Incontrast to theacidic metal complexes these compounds tend to interact with acidic substrates. This fact has been made use of in the investigation of hydrogen bonds.

Moreover, nitroxide radicals can weakly interact with other organic radicals thus reducing drastically the originally very large T,, values of these compounds. By this means the N M R spectra of several organic radicals could be o b s e r ~ e d [ ~ ~ . ~ ~ ~ .

3.2.2.2. Non-radical Shift Reagents

This category comprises those systems which develop a noticeable anisotropy either through a strong induced

[*I The assumption of this relatively high symmetry IS justified for the N M R experiment inview of the proven free rotation of the C,H,-rings.

The advantages of shift reagents of this category are sharp signals and temperature-independent isotropic shifts. Their resolving power, however, is only excep- tionally equivalent to shift reagents with unpaired electrons.

Recently, the use of some anionic n-electron systems as shift reagents ( e . g . [B(C6H,),l- o r IP(Cl2H8),1-) has been reported'571. These compounds are especially suited to reaction with cationic substrates in strongly polar solvents. The results published so far deal mainly with short-lived ion pairs of these anions with various- in some cases chiral-quaternary phosphonium ions.

4. Analysis of Shifted Spectra

4.1. Concentration Dependence of the Isotropic Shifts

The concentration dependence of the isotropic shift of a substrate signal averaged over all possible species present in solution can be expressed most clearly by the equilibrium concentrations of the free and complexed substrate :

In this equation, I I ~ - ~ and nS are the molar fractions of the adduct and the substrate, respectively.

Practically, however, only the initial concentrations [Rlo and [Slo of the shift reagent and substrate are usually known. One is forced, therefore, to express 6,,, as a function of L = [Rlo/[Slo instead of the molar fractions. In diagrams of both 6,,, = f(nR-s) and 6,,, =f(L) curves of the type shown in Fig. 8 a are obtained.

In the first case, the extrapolated ordinate intercept and the slope of the linear part are always equal to 6s ( =chemical shift of the free substrate) and 6R--S - 6s, the "molar isotropic shift" of the 1:l-adduct. The analogous correlation in the second case can be approximated only with the restriction that [Rlo< [Sl0 and the formation constant KL3" of the adduct is sufficiently large. The main reason is that, as was observed exper- i r n e n t a I I ~ [ ~ ~ . ' ~ " . the stone of the lineal- scclions in diagrams IS,,, = f(L) depend noticeably on the absolute substrate concentration.

Several authors have derived equations allowing the evaluation of ~ R - - s and K with satisfactory accu- racy[14' - 1 4 ' . I h 9 ] . According to Armitage et a1.[141.'591 . a diagram ISlo = fCS&i) in the concentration range IS10 9 [Rlo at constant [Rlo yields straight lines:

Angew. Chem. internal. Edit. 1 Vol. I 1 (1972) I No. 8 681

From the slope and the y-intercept of such a straight line the chemical shift hRPs and the formation constant K of the adduct can easily be obtained. These two parameters are occasionally termed ''intrinsic LIS parameters" (lanthanoid induced shift)"431. Fig. 8 b is a diagram of [Slo = f(S,A) showing the straight lines obtained for the three different protons of n-propylamine in the presence of E ~ ( d p m ) , " ~ ~ ~ . It is remarkable that for the three protons the intercepts are identical (thus resulting in an identical K value), although the slopes (and thus the values 6R-s) vary markedly.

0 2 OL 06 08 10 12 [ R I ~ I [ S l ~

-ooL 6& x 102[s] - Fig. 8. (a) Idealized dependence of the observed shift 6,,, on the molar ratio of the initial concentrations of shift reagent [RJo and substrate IS]". 65 =Chemical shift of the free substrate, 6, = shift of the 1 : 1-adduct. (b) Plot of 6.~~; = f([Slo) for the three different protons of n-propylamine at constant Eu(dpm), concentration (en. 0 . 0 0 6 ~ in CDCI,) [1591.

Equation (13) is valid only for 1 : 1 adducts. Suitable modifications can be obtained, however, taking into account different stoichiometries. Unknown adduct compositions may thus be determined"51.1591.

Deviations of the idealized curve shown In Fig. 8 a at higher values of L may be caused by the finite magnitude of K o r by the self-association of P r ( d ~ m ) ~ and Eu(dpm), which has been proved independently also in

OccasionaIIv curvatures at low L values are also observed (Fig. 9)1331. However, a determination of the chemical shifts 6s by extrapolation appears to be justified in most cases110.341. If either substrate or shift reagent are polyfunctional, the plots Sex, = f(L) may be completely curved or may exhibit several linear parts (cf. Section 5.2).

4.2. Graphical Signal Assignment

Frequently the problem of signal assignment can rapidly be solved by making use of the fact that the isotropic shifts decrease with increasing distance of the respective nuclei from the bonding site between shift reagent and substrate. An example is the distinction between the cis- and the trans form[58.59.78.145.1531 of geometric isomers, as in the case of a phosphetane (Fig.

trans cis

A'""(H-3) = 160 H z Aiso(H-3) = 224 Hz

Fig. 10. Isotropic shift of the 3-proton in cis- and rruns-2.2.3.4.4- pentamethyl-1-phenylphosphetan-1-one after addition of 0 48 equivalents of Eu(dpm), in CDC1, 1611.

Because of the direct proportionality between and rL-, Icf. equations (4) and (7)l it is usually practical to plot these quantities on a logarithmic In the case of axial symmetry one gets on taking the logarithms of equation (7):

0 1 0 2 0 3 0 4 0 5 06 07 0 8 0 9 Gzg [Eu ldpmI3]/[Substrate~+ Fig 9 'H-NMR shifts of the naohthyl protons of crs-1,3,5,5-tetra- methyl-S-(l-naphthyl)cyclohexan-l 01 as a function of the Eu(dpm), concentration [331

As long as the term in the brackets is constant, equation (14) is the function of a straight line with slope -3 . The essential requirement for this term to be constant is:

If this condition holds, 9 may take all values between 0 and ca. 50". and between 60 and ca. 90". Since with 9 = 54.7"

log(3cos'9,- 1) = - ffi

large deviations from linearity are to be expected in the critical angular range SO" < 9 < 60".

682 Angew. Chem. infernat. Edit. 1 Val. I 1 (1972) 1 No. 8

Fig. 11 shows the diagram logA, = f(logt-,f for the two conformers present at -70°C of the compound (C5W5),PrCNC6H,, with the cyclohexane ring equato- rially or axially substituted[271. The deviation from the best straight line with slope - 3 of the points with 9-values close to the critical range is obvious.

5 6 7 8 9 1 0 20 30 40 50 jb892.111 lliSo IpprnI - Fig. 11 . The isotropic shifts A'- of the C,H,,-signals in the 'H-NMR spectrum of (C,H,),PrCNC,H,, at - 70°C as a function of the geometric parameters r(A) and 9; slope of the best straight line: - 3.2 [271.

1 = 6-H (Ee), 9 = 9' 8 = P-H (Aa), 9 = 14' - 2 = 6-H (Ae), 9 = 2' 9 = P-H (Ea), 9 = 24' 3 = y-H (Ee), 9 = 20' 10 =. P-H (Ee), 9 = 24' 4 = 6-H (Ea), 9 = 20 I 1 =P-H(Ae). 9 = 2 6 " 5 = 6-H (Aa) , 9 = 24 - 12=y-H(Aa) . 9 = g ' - 6 = y-H (Ae), 9 = 3' 1 3 = a - H ( E ) , 9 = 1 3 ' 7 = y-H (Ea), 9 = 12 1 4 = a - H ( A ) . 9 = 1 3 '

A = axially substituted conformation E = equatoriallv substituted conformation a = axial proton e =equatorial proton Underlined numbers = points with greatest deviation from the best straight line.

In the case of protons separated from the metal ion by only three bonds (e.g. the a-protons in alcohols or e p ~ x i d e s [ ~ ~ - ~ ~ ] ) , the large deviations frequently encountered from the best straight line have been tentatively interpreted with the assumption of Fermi-contact interactions (see Section 2.1.2). Farid et a/.'661, however, could prove for two such cases[62.671 that these deviations were due t o the neglect of the angular dependence.

From equation (14) it can be deduced that a better fit of all points to the straight line should be achieved by appreciably increasing the anisotropy factor D (i. e. by replacing the Pr ion buTb(rrr) o r DY(III), cf. Fig. 5).

Occasionally, in the diagram log A, = f(logr,) straight lines with slopes smaller than - 3 have been foundL67-75.1371. In spite of the approximations used which are responsible for this effect[761, reliable assignments were achieved in these cases also.

4.3. Position of the Metal Atom

The distances r, can be determined from molecular models (e . g. Dreiding). The metal-ligand distances should be known for this procedure; if they are not, they must be approximated. In some cases sufficiently linear plots of IogA, = f(1ogt-J were obtained even when r, was put equal to the distance between the atom i of the substrate and the periphery of the lone pair a t the donor atom[67-691. Using this method, E r n ~ t [ ~ ~ ] found, e. g . , that the preferential conformation of 2,6-dimethyl-

cyclohexylamine is that shown in Fig. 12. It appears to be somewhat more realistic to express the distance metal-donor atom as the sum of the two covalent radii[641.

Fig. 12. Preferential conformation of 2.6-dimethrlcrclohexrl- amine [751.

Sanders and found the distance between Eu ion and 0 atom in the adduct of 4-tert-butyl- cyclohexanone and Eu(dpm), shown in Fig. 13a by iterative adjustment of the distances r with the isotropic

1- a\

k

Fig. 13. (a) Probable position of the Eu(id ion in the adduct between Eu(dpm), and 4-rert-butvlcyclohexanone 1771; (b) Most probable position ofthe Eu(i~i) ion in the adduct between Eu(dpm), and endo-norbornenol 11391. The 0 atom is at the origin of the internal coordinate system, the 0-C-2 bond is placed along the negative z axis, atom C-1 is placed io the x-z plane with the Dositive x coordinate.

shifts. Such iterations can conveniently be carried out by ~ ~ m p ~ t e r ' ~ ~ ~ ~ ~ ~ ~ ' ~ ~ ~ ~ ~ ~ ~ , especially if the angular term and internal rotations are also taken into consideration. The adduct of Eu(dpm), with endo-norbornen01"~~~'~~~ is an example (Fig. 13b). The distances Eu-0 or Yb-0 thus obtained in adducts with shift reagents of the chelate type are generally between 3.0 and 3.5 A,L78-801. In the case of fluorenone and 1-indanone, however, a satisfactory correlation was obtained with a distance Yb-0 of 1.5 k 0.2 (Fig. 14). The results of two structuredeterminations by X-ray methods in the solid state are, for comparison: Ho-N in Ho(dpm),(4-pi~oIine)~ 2.53 A[451 and Eu-N in Eu(dpm),(pyr)r 2.65 A[871.

Angew. Chem. internal. Edit. 1 Vol. I1 (1972) No. 8 683

- E, t 15! 10

5 7 I

3 L 5 6 7 8 9 - r CAI- Fig. 14. The isotropic shifts as a function of the distance r in the adducts between Yb(dpm), and fluorenone and I-indanone, respectively [81].

4.4. Influence of the Angular Term

Iftheangular dependenceof GI [equation (6)1 is taken into account, the correlation between the A,-values and the geometric factors is generally improved [ 2 7 . 4 8 . 6 6 . 8 2 . 8 3 . 1 3 8 1 .

Thus, in the case of the axially symmetric adduct (C,H,),PrCNC,H, 1[271 the standard deviation of the best straight line logAl=f(logr,) is 25.6% (Fig. l l ) , whereas it decreases to 2 .8% in the plot A,=f(G,) (Fig. 15). A similar improvement by nearly an order of magnitude could be achieved with an adduct of a shift reagent of the chelate This example shows that it is obviously advantageous not to neglect the

A-Form

aA a E&

2 L 6 8 10 12 li, 16 18 lo3 G [8?1 -

Fig. 15. Correlation between the isotropic shifts measured at - 70°C of the C,H,,-protons and their G-values in the two chair conformers ofthe adduct (C,H,),PrCNC,H,, (solvent: D,toluene) 1271. The nomenclature used to label the protons belonging to the points of the plot is given in the caption to Fig. 11.

term 3cos2s- 1 also in cases of lower than axial symmetry.

The change of sign of this term at 9 = 54.7" is responsible for the occasional observation of shifts to high field using Eu complexes, which generally cause shifts to lower field[33.44.85.86~8s.891. For instance, in the alcohol shown in Fig. 9 the signals of the naphthyl protons 3' and 4' are shifted, contrary t o the signals of all other protons, to high field[331. These results are explained in a straightforward manner by the coordination of the Eu ion at the axial OH-group.

4.5. Internal Rotations

In substrate molecules with internal rotational degrees of freedom which cannot be frozen-in in the temperature range used for the measurements, the Ai-values have to be correlated with mean G, values taking into account all possible rotamers. The mean Gi value is influenced most strongly by internal rotations whose rotational axes d o not coincide with the principal axis of symmetry of the complex. Equation (9) has been proposed for the case that the metal ion lies within the rotation axis"621.

From the good correlation in the diagram A, = f(G,) of (C,H,),PrCNC,H, 12'] (Fig. 15) a linear Pr-CN=N arrangement along the threefold principal rotation axis of the complex can be deduced. For the analogous alkoxy-complex (C,H,),UC,H however, a quantitative correlation between A, and G , was not achieved, since the rotation axis 0-C in the bent U-0-C group does not coincide with the principal molecular axis U-0. In most cases of this kind an exact determination of the mean Gi values is quite tedious and requires the use of a computer. In unfavorable cases, therefore, only qualitative signal assignments are possible, even if a powerful shift reagent is used.

5. Special Applications

5.1. Determination of Spin Coupling Constants

Shift reagents lowering the order of the spectrum of a substrate without broadening the signals excessively, simplify the determination of coupling constant^'^^-^^].

Table 2 . Coupling constants of rmns-5-rert-butyl-2-methyl-1,3,2- dioxa~hosahoran-2-one (Fig. 16). Upper line: Numbers obtained with computer program LAOCN3. Lower line: Numbers directly taken from the expanded spectrum.

[Eu (dam ,)I [substrate] JAB J A X J B , JAP JBP

0 -11.1 10.5 4.47 4.14 20.2 0.5 11 11 4 3 20

Table 2 shows the coupling constants. of a heterocyclic AA' BB'XY-system (Fig. 16)L941. The data obtained from an iterative computer analysis of the spectrum of

684 Angew. Chem. internat. Edit. 1 Vol. 11 (1972) 1 No. 8

the free substrate are listed in the upper line, and in anilines and the shift of the phenyl protons in the the lower line the constants directly read off from the 0- and rn-position obtained at constant Eu(dpm), con- spectrum recorded in the presence of a shift reagent. centration (Fig. 17a).

I

H, H, I I t-Bu

As important as the basic strength of the substrate are the steric conditions in the vicinity of the basic center[791. Thus, in the series aniline, N-methylaniline. and N,N- dimethylaniline the shifts decrease in spite of increasing pK,-values (Fig. 17 b)[991.

I 7

b l

-5OHz----i

Fig. 16. 'H-NMR spectrum of rru~s-5-rer/-butyl-2-methyl-l.3,2- dioxophosphoran-2-one [941. (a) 0.2 M in CCL: (b) With 0.5 equivalents of Eufdpm),. The coupling constants directly taken from the spectrum are indicated.

The influence of adduct formation on the magnitude of the coupling constants of a substrate is comparable to a substituent effect in all cases where a change of the substrate conformation is negligible""'. A linear increase of up to 1 0 % in several geminal coupling constants in camphor and 3,5,5-trimethyl-3-Cp-chloro- pheny1)cyclohexane has been noted upon increasing the concentration of the reagents Eu(dpm)3 o r Eu(fodI3 up to a molar ratio L: 0.7"57'. In the case of noticeable contact interactions additional changes of the coupling constants may possibly O C C U ~ [ ~ ~ . ~ ~ ' . A closer examina- tion of this effect is generally prevented by line broadening.

5.2. Influence of the Basicity of the Substrate; Polyfunctional Substrates

The formation constant of a n adduct between a substrate and a shift reagent (cf. Section 4.1) and thus the isotropic shifts of the substrate depend markedly on the basicity of the substrate. Thus, the "molar shift" induced in substrate RX by Eu(dpm)3 decreases in the order X = NH, > O H > C=O > -0- > COzR > CNr771. This series could be extended by some very weakly basic molecules, such as M-SO,CH,, o r the organometallics (C,H,),Sn o r [(C,H,)Fe(C0),l,[981.

Ernst and M a n n ~ c h r e c k [ ~ ~ ] found a linear correlation between the pKa-values of a series of p-substituted

2.46

pK. = 5.06 pK, ,= 4.85 ib892.171 (b)

Fig. 17. (a) Correlation between the pK,-values of p-substituted anilines and the isotropic shifts (in ppm) of the 0- and m-protons (molar ratio Eu(dpm),/substrate = 0.2; solvent CDCI,). (b) Influ- ence of the substitution at the nitrogen of aniline on the isotropic shifts (in ppm) of the phenyl protons 1991.

If a substrate molecule contains two donor functions whose equilibrium constants of complex formation with the shift reagent differ greatly (like -OH and -CO,R['ooJ o r P=O and C=O"o'l), in the diagram 6 = f(L), where 6 is the additive isotropic shift"371, pronounced changes of the slope at the equimolar ratio ( L = 1.0) for nuclei close t o one of the two basic centers a re observed. Evidently the weaker basic group does not coordinate before most o r all of the stronger function has reacted. Figure 18 shows the results obtained with a substrate carrying a n alcohol and an ester function""'.

According to H i n c k Z e ~ [ ~ ~ ' the contributions of the two donor centers of a bifunctional substrate t o the overall shift can be separated graphically in the following manner (Fig. 19): in the diagram logAA = f(logrA) of testosterone the AA value of proton i very likely originates from Eu(dpm), coordinated at the OH-group. It seems justified to draw a straight line of slope - 3 through point i, since such a line was experimentally found for the protons of the structurally very similar, but monofunctional cholesterol['21. The vertical differences of the experimental shifts of protons j and 1 from this line are equal t o the contribution AB resulting from

Angew. Chem. internal. Edit. 1 Vol. 11 (1972) / No. 8 685

OH H-3 I

H-6 I COOMe

, H-1 x

/ / x

, / x

L I I L L 0 0.4 0.8 1.2 1 6 2.0

Eu Idprn131f[Subs t ra te l~

Fig. 18. The chemical shifts (ppm relative to TMS) of the protons of methyl-trans-9-hydroxybicyclo(3.3.l)nonane-endo-3-carbox~l- ate as a function of the molar ratio Eu(dpm),isubstrate 11001.

Eu(dpm), coordinated at the carbonyl group. These shifts again fall approximately on a straight line log AB = f(logr,) of slope - 3. The shifts- AA and AB

500

Fermi contact 200 . contribution

tog rb- \og rn - Fig. 19. Graphical determination of the relative coordination strength of the OH and -0 functions in testosterone: double- logarithmic plot of the isotropic shifts A (in Hz) of protons i, 1. k, and 1 as a function of the distances rA and r, (in A) (cf. text) 1631.

corresponding to the same distance r are proportional to the complex formation constants K,, and KoH. In

Hart and LoveL'021determined the relative basic strength of functional groups by inter- or intramolecular competition: under the influence of Eu(dprnI3 an esuirnolar mixture of acetone-tetrahydrofuran exhibits the molar shifts indicated in Fig. 20a and b. From the changes of the shifts one can deduce that Eu(dpm), coordinates tetrahydrofuran about 8 times better than acetone.

this case the ratio AA/AB = KcoiKon =0.47.

0 H , C K C H 3 1.2

(11.1)

(a)

24.9 (28.0) 11.1 (12.7)

I H 170

( C ) (d)

Fig. 20. Extrapolated molar isotropic shifts (in ppm) of the protons oftwo monofunctional substrates in an equimolar mixture and of two bifunctional substrates on addition of Eu(dpm), 11021. In parentheses: values of the pure substances. (a) acetone, (b) tetrahydrofuran, (c) thioxane. (d) morpholine.

Similar competition experiments were carried out with bifunctional molecules (Fig. 20c and d). From a large number of data a series of basicity was deduced which was practically identical to the one cited above.

5.3. Polymers

Solutions of polymers such as polyethylene oxide""31 and poly(methy1 m e t h a c r ~ l a t e ) " ~ ~ * ' ~ ~ ~ also display induced shifts upon the addition of lanthanoid chelates leading to a substantial improvement of the resolution of their rather complex NMR spectra. In the case of polypropylene glycol"o51 it was possible to separate the 'H-NMR bands of the end methyl groups from the bands of the methyl groups in the polymeric chain, thus allowing determination of the molecular weight by direct comparison of signal intensities.

5.4. Dynamic Processes

Dynamic processes whose reaction rates are in a favorable range on the N M R time scale (rate constant k : 3 to 300 s-') cause line broadening, and finally line splitting, if the temperature is lowered. At the splitting o r coalescence temperature T'the following approximate correlation holds between the rate constant k, and the maximum peak separation AV"~~':

Apart from the fact that the rate of such a process can be changed by complex formation with a shift reagent,

686 Angew. Chem. internat. Edil. 1 Vol. 11 (1972) 1 No. 8

the temperature rises if Av is increased by the paramagnetic effect. Some processes can therefore be observed better in the presence of a shift reagent, because the broadening and splitting of the signals occurs in a more suitable temperature range.

One has to distinguish between intra- and intermolec- ulardynamic processes. Intramolecular processes within the substrate are sometimes obscured in the N M R spectrum by intermolecular ligand exchange phenom- ena1107. 1081

Ligand exchange in a system lanthanoid chelate-substrate was observed for the first time by Evans and Wyatt11331: in a mixture of Eu(fod), with an excess of dimethyl sulfoxide in CD2C12 the CH3 signal broadens considerably on cooling the solution to - 45 "C (Fig. 21 ).

CHDCI,

(50 to 127"C120.1091). On addition of Yb(dprn), Beaute et a/."09' noticed an appreciable increase in the coalescence temperature, but n o significant change in the

I

0 10 20 30 Av [Hzl-

CHDCI,

free OMS0

L 5 6 7 IA89221/ z -

Fig. 21. 'H-NMR spectra of a 0 . 1 3 1 ~ solution in CHDCI, of dimethyl sulfoxide (DMSO) after addition of 0 . 0 4 4 ~ Eu(fod), at (a), 38°C; (b) - 45°C; and (c) - 80°C 11 331.

At - 80°C this signal splits into signals corresponding t o the free and complexed substrate. A stoichiometric ratio S: R = 2: 1 can be deduced from the relative intensities of these lines.

The determination of solvation numbers of metal ions ( e . g . UO:+) in solution by the separate observation of the N M R signal of the free and the shifted signal of the complexed solvent a t low temperature is a well established

An intramolecular process observable without the inter- ference of ligand exchange is the rotational isomerism of amides whose coalescence temperature is rather high

I II

Fig. 22. Coalescence temperatures ( T J and free enthalpies of activation (AC:) for N,N-dimethrlformamlde (I) and N,N-di- methylacetamide (11) without and with addition of Yb(dpm),. The curves represent the temperature dependence of the frequency difference Av of the methyl signals 11091. Molar ratio IYb(dpm),l/[Il: (1) = 0 , (2) =0.06; IYb(dpm),l/lIIl: ( 3 ) =0, (4) =0.05. The numbers at the formulas represent the extrapolated molar shifts (in ppm) of the respective protons.

a1 30°C

C l -30°C

l l r \

20 30 40 50

Fig. 23. Temperature dependence of the 'H-NMR spectrum (C,H,,-part) of the adduct between (C,H,),Pr and C N C , H , , 1271; 6-scale (abscissa) relative to C,H,: solvent: D,-toluene. Nomenclature of the protons the same as in Fig. 11.

6 [ppml-

Angew. Chem. internat. Edit. / Vol. I1 (1972) / No. 8 687

free enthalpy of activation of this rotation (Fig. 22). Sch warzhans[20', however, found the coalescence temperature in the spectrum of N,N-dirnethylacetamide unchanged upon complex formation with UBr4"loJ. Thence he deduced that the C O group and not the nitrogen atom is the center of coordination.

The chair-chair interconversion of a monosubstituted cyclohexane (cyclohexylisonitrile) upon adduct formation with (C,H,),Pr[271 was studied by observing the temperature dependence of seven signals of the C6Hl , protons. The coalescence temperatures of several signal pairs with particularly large Av are raised so much that large signal broadenings occur already at a little below room temperature (Fig. 23).

5.5. Reaction Mechanisms

The mechanism of a reaction which is catalyzed by lanthanoid ions can be studied with the shift reagent technisue if these ions can act a t the same time as paramagnetic shift reagents on the N M R spectrum of the substrate. Thus, during the decarboxylation of oxaloacetic acid either of the complexes I and I1

I IT

JTZ O=C-C'C\CO,Et 66 11

I11 l\i Fig. 24. Possible intermediates ( I and 11) during the decornposi- tion of oxaloacetic acid catalyzed by Ln(ii3 ions 1441. 111 and IV: Substances of comparable structure.

(Fig. 24) can be formed as an intermediate. For the elucidation of the mechanism, Reyes-Zamora and Z ~ a i ' ~ ~ ] prepared the half esters I11 and IV (Fig. 24). Upon addition of EuCI, the CH, protons of 111 ex- hibited theusuallowfieldshift in the 'H-NMR spectrum, in contrast t o the signals of IV which were shifted to high field. As oxaloacetic acid shows a low field shift on addition of EuCI,, the authors concluded that the intermediate must have structure I which is the only possible one that the half ester 111 can form with the metal.

The kinetics of the catalytic deuteration of 4-tert-butyl- hexanone were studied by measuring the intensity decrease of the various 'H-NMR signals of that substrate after spectral simplification with E u ( d ~ m ) , { ' ~ ~ ' .

5.6. Pulsed Fourier-Transform 13C-NMR Spectroscopy

Shift reagents have hitherto been employed almost ex- clusively in 'H-NMR spectroscopy, not only because the proton is still the nucleus most widely used, but also because the spectral resolution efficiency of the shift reagents is greatest here. However, there are interesting applications in the NMR-spectroscopy of other

nuclei, e. g . in pulsed Fourier-transform ( = P.F.T) ~3C-NMR-spectroscopy~1131. The investigation of cholesterol in the presence of E ~ ( d p m ) , [ ' ~ ~ ' is an example.

Gansow et ~ l . 1 " ~ ~ assigned the P.F.T. "C-NMR spectrum of isoborneol, recorded by the common 'H- broadband decoupling technique (Fig. 25). British

Fig. 25. Isotropic T - N M R t1141 and 'H-NMR [64, 1141 shifts (in ppm) of isobomeol ( 0 . 3 4 ~ ) in an eauimolar mixture with Eu(dpm), in CCL. CH, groups=*; all shifts are to low field; the ' H data are written in italics.

expanded the relatively complex l3C-NMR spectrum of ribose-5-phosphate by adding Eu3+ ions in aaueous solution. The assignment of the signals was achieved by means of hetero-spin-decoupling: the selective irradiation of definite 'H-decoupling frequencies was feasible only in the spectrum simplified by the Eu ions Fig. 26).

bl l i l

3a :" 1 bl liiil

30 20 10 0 - S [pprnl Fig. 26 (a) IH-NMR spectrum (100 MHz) of ribose-5-~hos~hate ( 0 . 7 ~ in D,O) in the absence (above) and in the presence of 0.97 M E u ~ + ions at pH = 1.1. DSS =sodium salt of 2,2-dimethyl-2- silapentane-5-sulfonic acid; I a to 5 B=nuclei in position 1 to 5 with respect to the phosphate group of the a- and B-anomer, respectively. (bf 13C-NMR spectra (25.2 MHz. P. F.T. techniaue) of ribose-5-phosphate (0.6 M ) with an addition of 0.72 M Eu' ions 11 151; (i) 'H-broad band decoupliny: Cii) selective irradiation of the 21)-'H frequency; (iii) selective irradiation at the 3a-'H freauency .

688 Angew. Chem. internat. Edit. 1 Vol. I 1 (1972) No. 8

5.7. Chiral Shift Reagents The absence of chiral splitting in the case ofenantiotop- ic[1501 pairs of nuclei in otherwise nonchiral substrates ( e . g . the methyl groups in 2-propanol or dimethyl sulfoxide)[1201 as well as the dependence of Ach,r on the concentration ratio reagent: is consistent with

Pirkle et a1.["6.1321 noticed that the NMR spectra of the enantiomers of asymmetric compounds (particularly alcohols,a-hydroxy-acids, amines, amino acids and sulfoxides) differ appreciably In optically active solvents, e. g . (R)- ( - )-2,2,2-trifluorophenylethanol. These differences are presumably caused by the formation of diastereomeric adducts between the substrate and the optically active solvent. The latter is to be considered as a chiral shift reagent. The splitting of the spectra results from a different position of the substrate nuclei in the diastereomers with respect t o the magnetic field produced by the ring current of the x-system of the solvent.

Prior t o the introduction of more efficient chiral shift reagents of the chelate type numerous N M R studies were conducted on paramagnetic complexes of the d- transition elements with asymmetric chelate ligandsf"7."s1. Primary concern in investigations of this type is directed, however, at the spectra of the chelate ligands. The first chiral shift reagent of the chelate type with a lanthanoid central ion was prepared by Whiteside and Other, similar reagents with praseodymium and europium as central ions were de- scribed later"20.1211. Fig. 27 illustrates the effect of the chiral shift reagent ( 1 ) (Table 3) o n the ' H - N M R

Table 3. Chiral shift reagents of the B-diketonato type

( I ) R = -C(CH,), (2) R = -CF, (3) R = R'

( 5 ) R' = R" = R' ( 6 ) R' = R', R" = R'

(7) R ' = l l % R 2 + 2 3 % R ' ; R " = R3 (8) R' = R' ; R" = 11% R2 + 23% R'. ( 4 ) R = l l % R 2

+ 23%R1

spectrum of pure (S)-a-phenylethylamine and on a mixture of both enantiomers of this substrate. All signals in the spectrum of the pure (S)-enantiomer appear doubled in the spectrum of the mixture with the splitting Achlr being largest for nuclei in the vicinity of the asymmetric center of the substrate. Whereas the utilizability of ( I ) is limited to relatively strong donors like amines, chiral splitting is also induced in the spectra of alcohols, ketones, esters, and sulfoxides by the more acidic chiral shift reagents (2) t o ( 8 ) (Table 3). Table 4 gives a survey of the chiral shifting potential of such reagents. A particularly straightforward synthesis of the h a n d of chiral shift reagent ( 2 ) (Table 3) is given in Ref."49'.

I

180 170 160 150 1LO 130 120 110 100 9 0 8 0 - 6 [pprnl Fig. 27. 'H-NMR spectra of a-phenvlethylamine in CCI, after addition of trisf~3-tert-butylhydrox~methylene~-o-cam~horatol- europiurn(1ri) [I 191. Molar ratio ~Eu-corn~lexl/tsubstratel: 0.5 to 0.6. (a) Pure (S/-form of the amine; (b) mixture of (9- and (Rbform.

Table 4. Signal shifts 4h,r induced by chirdl shift reagents (R).

Substrate Proton

2-octanol

1 -phenulethyl acetate

1 -rnethyl-2-nor- bornanone

1 -phenulethuI- amine

2-phenyl-2- butanol

benzyi methyl sulfoxide

2-butyl formate

u-CH, 121 0.11 I I

-C02CCH3 (21 0.18

(21 0.11

' 31 1.13 I ( 8 ) 1.65

-+, (31 0.32 ( 8 , 0.25

-CHzS- i . 0.6

z 0.6

> 0.6

ca 1.3 ca. 0 . 8

ca. 1.3 ca. 0.8 ca. 1.3 ca. 0.8 ca. 1 3 ra. 0.8

[a1 The numbers refer to the compounds in Table 3

the assumption that the formation of conformationally stable diastereomers is an essential requirement for the spectral splitting. In particular, if the formation constants of the diastereomers differ measurably, the assumption seems justified that different geometric factors correspond t o analogous nuclei of the substrate in the diastereomeric pair. A comparison of the Achir values given in Table 4 with comparable data of diastereomeric systems without unpaired electrons

Angew. Chem. internal. Edil. / Vol. I1 (1972) / N o . 8 689

shows that the anisotropic field of the lanthanoid ion improves the splitting of the spectrum by approximately an order of magnitude.

5.8. Isotope Effects

A further interesting application of shift reagents of the chelate type was found by Hinckley et a1.[‘35.’361 and other author^"^'^'^^^'^^^: the I H-NMR spectra of alcohols partially deuterated at the position geminal to the OH-group are doubled under the influence of Eu complexes. The isotropic shifts of the doubled signals differ by 1 to 3.5 %. The authors discuss as possible reasons for this splitting a change in the basicity of the hydroxyl oxygen by the deuterium substitution or differences in the strength of a secondary hydrogen bonding between the geminal H (or D) and the ring oxygens of the metal chelate ligands.

Received: Mar 9. 1972 [A 892 IEI German version: Angew. Chem. 84, 737 (1972)

[ I 1 W . Y. Philipsborn, Angew. Chem. 83, 470 (1971); Angew. Chem. internat. Edit. 10, 472 (1971). [21 J . E. Maskuskv and M . E. Kennev, J . Amer. Chem. SOC. 93, 2060 (1971). [3] T. H . Siddull. 111 and C. A . Prohusku, Inorg. Chem. 4, 783 (1965). [4] G. H . Wuhl, Jr. and M . R . Peterson, Jr., Varian Instr. APPI. 4, 6 (1970). 151 NMR Spectra Catalog, Varian Associates, Palo Alto (1963). Spectrum No. 269. [61 M . Ohashi, I . Morishima, and T. Yonezawa, Bull. Chem. SOC. Japan 44, 576 (1971). 171 R. D. Fischer and R. v. Ammon, unpublished work. [81 W . D. Phillips. C. E. Looney, and C. K. Ikeda, J. Chem. Phys. 27, 1435 (1957). [91 R. W . Kluiber and W. D. Horrocks, J . Amer. Chem. SOC. 87, 5350 (1965). 1101 a) K. Fricke and H . Suhr, Ber. Bunsenges. phys. Chem. 72, 434 (1968); b) K. Fricke, Tetrahedr. Letters 1971, 1237. [I 11 W . A . Szurek, E. Dent, T . B. Grindlev. and M . C. Bird, Chem. Commun. 1969, 953. (121 C. C. Hinckley, J. Amer. Chem. SOC. 91, 5160 (1969). [131 L. Ernst, Nachr. Chem. Techn. 18, 459 (1970). [141 W . D e w . Horrocks, Jr. and J . P. Sipe, III, J . Amer. Chem. SOC. 93,6800 (1971). 1151 D. R. Eaton and W . D. Phillips, Advan. Magn. Resonance I , 103 (1965). I 1 61 E. de Boer and H. van Wiflixen, Progr. NMR Spectroscopy 2, 1 1 1 (1967). 1171 D. R . Euton in H . A . 0. Hill and P. Day: Physical Methods in Advanced Inorganic Chemistry. Interscience, New York 1968. [181 R. H. Holm, Accounts Chem. Res. 2, 307 (1969). 1191 H . J . Keller and K. E. Schwarzhuns, Angew. Chem. 82, 227 (1970); Angew. Chem. internat. Edit. 9, 196 (1970). I201 K. E. Schwurzhuns, Angew. Chem. 82, 975 (1970); Angew. Chem. internat. Edit. 9, 946 (1970).

I211 H . J . Keller in P. Diehl, E. Fluck, and R. Kosfeld: NMR- Basic Principles and Progress. Springer, Berlin 1970. [221 W . D e w . Horrocks, Jr., Inorg. Chem. 9, 690 (1970). [231 W . D e w . Horrocks, Jr. and D. D e w . Hull, Coord. Chem. Rev. 6, 147 (1971). 1241 The formulation of D’ and D” as a function of spectroscopic splitting factors (g-factors [25, 261). which is still frequently encountered in the literature, neglects the influence of the Van Vleck paramagnetism and is therefore inferior to the formulation used here 1221. 1251 J . B. Jesson, J. Chem. Phys. 47, 579, 582 (1967). [261 A. J . Kurlund and B. R. McCurvev, J. Magn. Resonance 2, 286 (1970). 1271 R. v. Ammon, R . D. Fischer, and B . Kanellakopulos, Chem. Ber. 104. 1072 (1971).

[281 R. D. Fischer in R. H . Holm, W . D. Horrocks, Jr., and G. N. LaMur: Chemical Applications o f NMR-Resonance in Paramagnetic Molecules. Academic Press, in press. [291 J . Reuben and D. Fiat, J. Chem. Phys. 51, 4909 (1969). [301 R. v. Ammon, B. Kanellakopulos, R . D. Fischer, and P. Laubereuu, Inorg. Nucl. Chem. Lett. 5, 315 (19693. 1311 a) H. Huber and J. S e e k , Helv. Chim. Acta 55, 135 (1972); b) K. Roth, M . Grosse. and D. Rewicki, Tetrahedron Lett. 1972, 435.

1321 M . K. Archer, D. S. Fell, and R. W . Jothum, Inorg. Nucl. Chem. Lett. 7, 1135 (1971). 1331 B. L. Shaairo, J . R. Hlubucek, G. R. Sullivan, and L. F. Johnson, J. Amer. Chem. SOC. 93, 3281 (1971).

1341 R. W . Kluiber and W . D. Horrocks, J. Amer. Chem. SOC. 88, 1399 (1 966). 1351 I . Morishima and T. Yonezuwa. J. Chem. Phys. 54, 3238 (1971).

[361 I . Morishima, K. Okudu, M . Ohashi, and T. Yonezuwa, Chem. Commun. 1971, 33. 1371 Y . N. Molin and E. E. Zaev, Zh. Strukt. Khim. 8 , 140 (1967). I381 E. E. Zuev, V . K . Voronov, M . S. Shvartsberg, S. F. Vasilevskv, Y. N. Molin. and 1. L . Kotliarevsky, Tetrahedron Lett. 5, 617 (1968). C391 H. C. Friedman, Jr., G. R . Choppin, and D. G. Feuerbacher, J. Chem. Educ. 41, 354 (1964). 1401 M . lhnat and R. Bersohn, Biochemistry 9, 4555 (1970). 1411 A. M . Bowles, W. A . Szarek, and M . C. Baird, Inorg. Nucl. Chem. Lett. 7, 25 (1971). 1421 J. K. M . Sanders and D. H . Williams, Tetrahedron Lett. 1971. 2813. f431 F. A . Hart. C. P. Moss. and M . L. Stunijorth, Tetrahedron Lett. 1971, 3389. I441 C. Reyes-Zamoraand C. S . Tsai, Chem. Commun. 1971,1047. 1451 W . D e w . Horrocks, Jr., J. P. Sipe, I l l , and J . R. Luber, J . Amer. Chem. SOC. 93, 5258 (1971). 1461 J. K. Przvstal, W . G. Bos, and I . B. Liss, J. Inorg. Nucl. Chem. 33. 679 (1971). 1471 D. R. Crump. J. K. M. Sanders. and D. H. Williums, Tetra- hedron Lett. 1970. 4419 and 4949. [48] J. Armitage and L. D. Hall, Canad. J . Chem. 49, 277 (1971). I491 R. E. Rondeau and R. E. Sievers, J. Amer. Chem. SOC. 93, 641 (1971). [501 K. J. Eisentraut and R. E. Sievers, J. Amer. Chem. SOC. 87. 5254 (1965). I511 M . McCrearv, cited in Ref. 11201. 1521 E. 0. Fischer and H. Fischer, J. Orpanometal. Chem. 6. 141 (1 966). 1531 I. Morishima, K. Endo, and T. Yonezawa, J. Amer. Chem. SOC. 93, 2048 (1971). 1541 I. Morishima. K. Endo, and T. Yonezawa, Chem. Phys. Lett. 9, 143, 203 (1971). [ 5 5 ] H . Brunner, K. M . Hausser, and F. A . Neugebauer, Tetra- hedron 27, 361 1 (1971). [561 F. A . Neugebauer. H . Brunner. and K. M . Hausser. Tetra- hedron 27, 3623 (1971). [571 G. P. Schiemenz. J. Magn. Resonance 6, 291 (1972). I581 H . G. Richey, Jr. and F. W. von Rein, Tetrahedron Lett. 1971, 3781. I591 A. H . Lewin, Tetrahedron Lett. 1971, 3583. 1601 H. Huber and C. Pascual, Helv. Chim. Acta 54, 913 (1971). 1611 J. R. Corfield and S. Trippetf, Chem. Commun. 1971, 721. I621 P. V. Demarco. T. K. Elzey, R . B. Lewis, and E. Wenkert, J. Amer. Chem. SOC. 92, 5734 (1970). [631 C. C. Hinckley, M . R. Klotz. and F. Patil, J. Amer. Chem. SOC. 93, 2417 (1971). [64] C. H. Wahl,jun. and M . R. Petersen, iun., Chem. Commun. 1970, 1167. [651 L. H. Keith, Tetrahedron Lett. 1971. 3. f661 S. Furid. A. Ateya. and M . Maggio, Chem. Commun. 1971. 1285. I671 A. F. Cockerilland D. M. Rackham, Tetrahedron Lett. 1970, 5149.

1681 A . J. Rafulski, J . Barciszewski, and M . Wiewiorowski, Tetra- hedron Lett. 1971, 2829. [691 Z. W . Wolkowski, Tetrahedron Lett. 1971, 825. f701 A. F. Cockerilland D. M . Rackham, Tetrahedron Lett. 1970, 5153.

.

690 Angew. Chem. internat. Edit. Vol. I 1 (1972) / No. 8

1711 J . Barcisrewski, A. J . Rafalski, and M. Wiewiorowski, Bull. Acad. Pol. Sci, Sir. Sci. Chim. 19, 545 (1971). 1721 M . R. Willcott, J . F. M . 0 th . J . Thio, G. Plinke, and G. Schrdder, Tetrahedron Lett. 1971, 1579. [731 J . Paasiwirta, Suomen Kemistilehti B 44, 131, 135 (1971). [741 J . Paasivirta and P. J. Malkonen, Suomen Kemistilehti B 44, 230, 238 (1971). 1751 L. Ernst, Chem.-Ztg. 95, 325 (1971). 1761 J . Goodismanand R. S. Matthews, Chem. Commun. 1972,127. (771 J . K. M . Sanders and D. H. Williams, J . Amer. Chem. SOC. 93, 641 (1971). 1781 R. R. Fraserand Y. Y. Wigfield, Chem. Commun. 1970,1471. [791 C. B e a d . Z . W . Wolkowski, and N . Thoai, Tetrahedron Lett. 1971, 817. [801 C. Beauth, Z. W. Wolkowski, J. P. Merda, and D. Lelandais, Tetrahedron Lett. 1971, 2473. [811 Z. W . Wolkowski. Tetrahedron Lett. 1971, 821. 1821 R. v. Ammon, B. Kanellakopulos, and R. D. Fischer, Chem. Ber. 105, 45 (1972). [831 P. H. Mazrocchi, H. J . Tamburin, and G. R. Miller, Tetra- hedron Lett. 1971, 1819.

1841 R. Caple and S. C. Kuo, Tetrahedron Lett. 1971, 4413.

[85] T. H . Siddall, I I I , Chem. Commun. 1971, 452. 1861 S. B. Tjan and F. R. Visser, Tetrahedron Lett. 1971, 2833. 1871 R. E. Cramer and K. Sef f , Chem. Commun. 1972.400. [881 M . Kishi, K. Tori, and T. Komeno, Chem. Commun. 1972, 3525. 1891 N. S. Bhacca and J . D. Wander, Chem. Commun. 1971.1505. 1901 A. F. Bramwell, G. Rierebos, and R. D. Wells, Tetrahedron Lett. 1971, 2489.

[911 H:D. Scharfand M.-H. Feilen. Tetrahedron Lett. 1971,2745. 1921 S. G. Levine and R. E. Hicks. Tetrahedron Lett. 1971, 311. (931 M . Gielen, N. Coffin, and J. Topart, J. Organometal. Chem. 32, C38 (1971). (941 K. C. Yee and W . G. Bentrude, Tetrahedron Lett. 1971, 2775. 1951 H. Katoand H. Karo. Bull. Chem. SOC. Japan 44, 1734 (1971). 1961 J. A. Peters, J . E. Remijnse, A. van der Wiele, and H. van Bekkum, Tetrahedron Lett. 1971, 3065. 1971 C. Beauti. S. Cornuel, D. Lelandis, N . Thoai, and Z. W . Wolkowski. Tetrahedron Lett. 1972, 1099. [981 T. J . Marks, J . S. Kristoff, A. Alich, and D. F. Shriver. J . Organometal. Chem. 33, C35 (1971). 1991 L. Ernst and A. Mannschreck, Tetrahedron Lett. 1971, 3023. [lo01 I . Fleming, S . W . Hanson, and J. K. M . Sanders, Tetrahedron Lett. 1971, 3733. 11011 Y. Kashmanand 0. Awerbouch.Tetrahedron27.5593 (1971). 11021 H . Hurt and G. M . Love, Tetrahedron Lett. 1971, 625. [lo31 A. R. Katritzky and A. Smith, Tetrahedron Lett. 1971, 1765. 11041 J . E. Guillet, I . R. Peat, and W . F. Reynolds, Tetrahedron Lett. 1971, 3493. [lo51 F. F.-L. Ho, Polymer Letters 9, 491 (1971). [lo61 H . Kessler, Angew. Chem. 82. 237 (1970); Angew. Chem. internat. Edit. 9, 219 (1970). 11071 A. C. Adams and E. H. Larsen, Inorg. Chem. 5, 814 (1966). I1081 T. R. Stengle and C. H. Langford, Coord. Chem. Rev. 2, 349 (1967).

11091 C. Beauth, Z. W . Wolkowski, and N . Thoai, Chem. Commun. 1971, 700. [1101 F. Lux, G. Wirth, K. W . Bagnall, and D. Brown, Z . Natur- forsch. 24b. 214 (1969). 11111 A. Frutiello, V. Kubo, R. E. Lee, S . Peak, and R. E. Schuster. J. Inorg. Nucl. Chem. 32. 3114 (1970). 11 121 A. Fratiello. V. Kubo. S. Peak, B. Sanchez, and R. E. Schuster, Inorg. Chem. 10, 744 (1971) and literature cited therein. 11131 E. Breitmaier, G. Jung, and W . Voelter, Angew. Chem. 83, 659 (1971); Angew. Chem. internat. Edit. 10, 673 11971). 11141 0. A. Gansow, M . R. Wilcott, and R. E. Lenkinski, J. Amer. Chem. SOC. 93, 4295 (1971). 11151 B. Birdsall, J . Feenev. J . A. Glasel, R. J . P. Williams, and A. V. Xavier, Chem. Commun. 1971, 1473. (1161 W . H. Pirkle and S. D. Beare, J . Amer. Chem. SOC. 91, 51 50 (1969). and literature cited therein. 11171 C. R. Powers and G. W. Everett, Jr., J . Amer. Chem. SOC. 91. 3468 (1969). [1181 R. E. Ernst, M. 1. OConnor, and R. H. Holm, J. Amer. Chem. SOC. 89, 6104 (1967); 90. 5305 (1968).

I1 191 G. M . Whiteside and D. W . Lewis, J . Amer. Chem. SOC. 92, 6979 (1 970). 11201 H. L . Goering, J . N. Eikenberry, and G. S. Koermer, J . Amer. Chem. SOC. 93, 5913 (1971). 11211 G. M . Whitesides and D. W . Lewis, J . Amer. Chem. SOC. 93. 5914 (1971). 11221 E. A. Berends and J . G. Brushmiller, Inorg. Nucl. Chem. Lett. 6, 847 (1970). 11231 N. Ahmad, N . S. Bhacca. J. Selbin, and 1. D. Wander. J . Amer. Chem. SOC. 93, 2564 (1971). I1241 L . TomiC, 2. Majerski, M . TomiE, and D. E. Sunko, Chem. Commun. 1971, 719. [1251 A. E. Crease and P. Legzdins, Chem. Commun. 1972, 268. 11261 J . E. Maskasky, J . R. Mooney, and M . E. Kenney, J . Amer. Chem. SOC. 94, 2132 (1972). [1271 L. Pohl, Kontakte (Merck) I , 17 (1972). 11281 A. Ahmadi, Bericht CEA-R-4235 (Saclay) (1971). I1291 K. Fricke, Spectrochim. Acta 28A. 735 (1972). [1301 I. Morishima, K. Okada, and T. Yonezawa, J. Amer. Chem. SOC. 94, 1425 (1972). [131] R. Claramunt. J . Elguero. and R. Jacquier, Org. Magn. Resonance 3, 595 (1971). 11321 W. H. Pirkle, R. L. Muntz, and 1. C. Paul. J . Amer. Chem. SOC. 93, 2817 (1971). [1331 D. F. Evans and M . Wyatt. Chem. Commun. 1972. 312.

11 341 Y . Takagi, S. Teratani. and J . Uzawa, Chem. Commun. 1972, 280. 11351 G. V. Smith, W . A. Boyd, and C. C. Hinckley, J . Amer. Chem. SOC. 93, 6319 (1971). 11361 C. C. Hinckley, E. A. Boyd. and G. V. Smith. Tetrahedron Lett. 1972. 879. [137] A. Ius, G. Vecchio, and G. Carrea, Tetrahedron Lett. 1972, 1543. 11381 W. L. F. Armarego, T. 1. Batterham. and J . R. Kershaw. Org. Magn. Resonance 3, 575 (1971).

I1391 M . R. Wfllcott, III , R. E. Lenkinski, and R. E. Davis, J . Amer. Chem. SOC. 94, 1742 (1972).

[1401 R. E. Davis and M . R. Willcott, I I I . J. Amer. Chem. SOC. 94, 1744 (1972). 11411 I . Armitage. G. Dunsmore, L. D. Hall, and A. G. Marshall. Chem. Commun. 1971. 1281; Chem. Ind. 1972, 79.

[1421 D. R. Kelseu, J. Amer. Chem. SOC. 94. 1764 (1972).

11431 D. E. Williams, Tetrahedron Lett. 1972, 1345.

[1441 R. v . Ammon and B. Kanellakopulos, Ber. Bunsenges. Phrs. Chem.. in press. 11451 L. R. Isbrandtand M. T. Rogers,Chem.Commun. 1971.1378. [1461 Th. Moeller in K. W. Bagnall: MTP International Review of Science, Inorg. Chem. Ser. 1, Vol. 7, p. 275, Butterworths. London 1972.

[1471 L. A. Aslanov, E. F. Korytnyi, and M. A. Porai-Koshils, J . strukt. Chim. (Russ.) 12, 661 (1971); Engl. transl. 12, 600 (1971).

[1481 B. Kanellakopulos. E. Dornberger, R. v. Ammon, and R. D. Fischer, Angew. Chem. 82, 956 (1970); Angew. Chem. internat. Edit. 9, 957 (1970). 11491 V. Schurig, Inorg. Chem. 11, 736 (1972). I1501 M . van Gorkom and G. E. Hull, Quart. Rev. 22, 14 (1968).

[1511 A. M . Grotens. J . Smid, and E. de Boer, J . Magn. Resonance 6, 612 (1972). 11521 R. D. Bennett and R. E. Schuster, Tetrahedron Lett. 1972, 673. [1531 G. J . Martin. N. Naulet, F. Lefevre, and M . L. Martin. Org. Magn. Resonance 4, 121 (1972). [I541 H. E. Francis and W. F. Wagner. Org. Magn. Resonance 4, 189 (1972). 11551 M. Fre.smunri and R. Frevmunn, C. R . Acad. Sci. 250, 3638 ( I 960). [ 1561 c'. L). Burt-y. A. C. T. North, J. A. Glusel, R. J . P. Willrums, and A. V. Xavier, Nature 232, 236 (1971). 11571 B. L. Shapiro, M. D. Johnston, and R. L . R. Towns, J . Amer. Chem. SOC. 94,4381 (1972). 11581 J . K. M. Sanders, S. W. Hanson, and D. H. Williams, J . Amer. Chem. SOC., in press. i1591 I. M. Armitage, V. Gibb, L. D. Hall, and A. G. Marshall, 2"' Int. Symp. NMR Spectroscopy, Guildford (England), June 1972.

,

Angew. Chem. internat. Edit. I Vol. I 1 (1972) / N o . 8 69 1

(1601 B. E. Mann in Spectroscopic Properties of Inorganic and Organometallic Compounds. Swcialist Periodical Report. The Chem. A d a 43,267 (1971).

11651 L. TomiC, 2. Maierski, M . TomiC, and D. E. Sunko. Croat.

Ch&ical Society, London 19721 [I611 C. L. Honeybourne, Tetrahedron Lett. 1095 (1972)

11661 W. B. Smith and D. L. Deavenport, J . Magn. Resonance 6, 256 (1972). _._ ,~

11621 W . D e w . Horrocks. Jr. and J. P . Sipe, 111% X W h Internat. Conf. Coord. Chem., Toronto, June 1972; Science, in press. I1631 B. Bleaney, C. M . Dobson, B. A. Levine, R. B. Martin, R. J . P. Wil(ioms, and A. V. Xavier, Chem. Commun., in press. 11641 A. S. Beryland and L. I . Martinenko, Vestnik Mosk. Univ. Khim. I2, 731 (1971).

[,671 A. M . Grotens, c. ,+,. E. de Boer, and J . smid, Tetrahedron Lett. 1972, 2067.

11681 J . K . M . Sanders and D. H. Williams, Chem. Cornmun. 1972,436.

I1691 J. Bouquant and J. Chuche, Tetrahedron Lett. 1972. 2337.

Organic Syntheses uia Free-Radical Displacement Reactions of Organoboranes[**l

By Herbert C. Brown and M. Mark M i d l a n d [ * ]

In the few short years since the discovery that organoboranes undergo facile free-radical substitution, application of these reactions has become a major new area of synthetic utility. Organoboranes undergo a wide variety of free-radical reactions such as autoxidation to the alcohol or hydroperoxide, 1.4 addition to a,P-unsaturated derivatives, addition to disulfides, and oxygen-induced radical coupling. It is evident that organoboranes constitute a versatile new source of free radicals and that these reactions can be readily controlled t o give very clean syntheses. The application of these reactions will be discussed.

1. Organoboranes as Excellent Sources of Free Radicals

Recognition that the organoboranes are excellent sources of free radicals and can participate in free-radical chain reactions cameabout through detailed mechanistic studies of the reaction of organoboranes with OXY- gen"- and with u,B-unsaturated carbonyl corn- pounds'41.

Originally, it was suggested that the autoxidation of a trialkylborane proceeds via an intermediate "borine- peroxide" / 1 ' ' ' More recently, it was recognized that

e e R,B-0=0 / I )

these peroxides are true alkyl peroxide derivatives, containing the structure ROOB:. Thus, the reaction of oxygen with tri-sec-butylborane in dilute solution yields the diperoxide (2) '". However, the formation of these

I*] Prof. H . C. Brown and M. Mark Midland R. B . Wetherill Laboratory Purdue University Lafayette. Indiana 47907 (USA)

I**] Based in part o n Chapter XIX of a forthcoming book (Baker Lectures) by H. C . Brown: Boranes tn Organic Chemistry. Cornell University Press, Ithaca. N. Y . (USA)

peroxides was long considered to involve non-radical processes because many of the usual radical inhibitors, such as hydroauinone, had n o apparent effect upon the r e a ~ t i o n ~ ' ~ .

The oxidation of optically active 1 -phenylethylboronic acid gave a racernic product and this was considered to besuggestiveofa process involving radicals[". Indeed, i t was observed that a remarkable induction period occurs in the autoxidation of this compound in the presence of added inhibitors such as copper(r1) N , N - dibutyldithiocarbamate and galvinoxyl 12.6-di-rert- butyl-a-(3,5-di-rerr-butyl-4-0~0 -2.5 -cyclohexadien - 1 - ylidene)-ptolyloxyll['l. It was then discovered that galvinoxyl also effectively inhibits the autoxidation of tri- i s o b u t ~ l b o r a n e [ ~ ~ , the epimeric trinorbornylboranes'*' and other organoborane~*~] . These findings led to the suggestion that the autoxidation of trialkylboranes proceeds via a free-radical chain processr' - 3 1 .

R I B + 0 2 --t R - + R2B0, . R . S O ? 4 RO,. RO?. i BR, - RO,BR: + R .

Perhaps the most unusual feature about this proposed process is the assumption that a free alkylperoxy radical is capable of rapidly attacking boron with displacement of an alkyl radical, the latter then continuing the chain. Evidence for this reaction course has accumulated rapidly. For example, when di-rert-butylperoxide is

692 Angew. Chem. inrernat. Edit. / Vol. I1 (1972) 1 No. 8

Shift Reagents in NMR Spectroscopyweb.missouri.edu/~glaserr/4210f20/SHIFT-Reagents_anie.19720675… · magnetic shift AF of a specified proton of the sub- strate s.If S = cyclohex~lisonitrile,

Documents