Comprehensive coordination chemistry II vol 2

Volume 1: Fundamentals: Ligands, Complexes, Synthesis, Purification, and StructureVolume 2: Fundamentals: Physical Methods, Theoretical Analysis, and Case StudiesVolume 3: Coordination C

Introduction to Volumes 1 and 2

In this first two volumes of Comprehensive Coordination Chemistry II we have endeavored to laydown the fundamentals of coordination chemistry as it is understood in the early part of thetwenty-first century We hope to have provided all the necessary fundamental backgroundinformation needed to prosecute coordination chemistry in the physical and theoretical laboratoryand to appreciate fully the information provided in the remaining volumes of this treatise.These volumes contain 112 contributions from some 130 outstanding,internationally known,contributors They are subdivided into nine major sections whose content is described brieflybelow The contributors were asked to emphasize developments in the field achieved since 1980and since the publication of CCC (1987)

1 LIGANDS – a survey of the syntheses,characterization,and properties of many of the morecommonly employed ligands

2 SYNTHESIS,PURIFICATION AND CHARACTERIZATION OF COORDINATIONCOMPOUNDS – including a detailed survey of aqua metal ions,the use of solvents,chromatographic methods,and crystal growth techniques

3 REACTIONS OF COORDINATED LIGANDS – dealing with the chemistry of moleculessuch as oxygen,nitric and nitrous oxide,carbon dioxide,oximes,and nitriles

4 STEREOCHEMISTRY,STRUCTURE,AND CRYSTAL ENGINEERING – structureand stereochemistry involving lone pair effects,outer sphere interactions,and hydrogenbonding

5 NEW SYNTHETIC METHODS – nine contributions dealing with a wide range of newermethodologies from biphasic synthesis to sol–gel to genetic engineering

6 PHYSICAL METHODS – a very extensive chapter incorporating 34 contributions detailingthe enormous breadth of modern physical methods

7 THEORETICAL MODELS,COMPUTATIONAL METHODS,AND SIMULATION –

17 contributions illustrating the wealth of information that can be extracted from a range ofcomputational methods from semi-empirical to ab initio,and from ligand field theory to metal–metal exchange coupling to topology,etc

8 SOFTWARE – a brief glimpse of some of the packages which are currently available

9 CASE STUDIES – putting it all together – eight studies which reveal how the many physicaland theoretical techniques presented earlier in the volume can be used to solve specificproblems

The creation of these volumes has been an exciting,challenging,time-consuming,and absorbing experience The Editor hopes that it will also be a rewarding experience to the reader-ship Finally,the Editor is greatly indebted to Paola Panaro for her untiring assistance in theconsiderable secretarial work associated with these volumes – without her it would have beenimpossible He is also much indebted to his wife Elaine Dodsworth for her emotional support!

all-A B P LeverToronto, CanadaMarch 2003

xvii

From Biology to Nanotechnology

Second Edition

Edited by

J.A McCleverty, University of Bristol, UK

T.J Meyer, Los Alamos National Laboratory, Los Alamos, USA

Description

This is the sequel of what has become a classic in the field, Comprehensive Coordination Chemistry The first edition, CCC-I, appeared in 1987 under the editorship of Sir Geoffrey Wilkinson (Editor-in-Chief), Robert D Gillard and Jon A McCleverty (Executive Editors) It was intended to give a contemporary overview of the field, providing both a convenient first source of information and a vehicle to stimulate further advances in the field The second edition, CCC-II, builds on the first and will survey developments since 1980 authoritatively and critically with a greater emphasis on current trends in biology, materials science and other areas of contemporary scientific interest Since the 1980s, an astonishing growth and specialisation of knowledge within coordination chemistry, including the rapid development of interdisciplinary fields has made it

impossible to provide a totally comprehensive review CCC-II provides its readers with reliable and informative background information in particular areas based on key primary and secondary references It gives a clear overview of the state-of-the-art research findings in those areas that the International Advisory Board, the Volume Editors, and the Editors-in-Chief believed to be especially important to the field CCC-II will provide researchers at all levels of sophistication, from academia, industry and national labs, with an unparalleled depth of coverage.

Bibliographic Information

10-Volume Set - Comprehensive Coordination Chemistry II

Hardbound, ISBN: 0-08-043748-6, 9500 pages

cws_home/622954/conditionsofsale, and links to our regional sales officeshttp://www.elsevier.com/wps/find/ contact.cws_home/regional

GBP 4,182.50

030/301

Last update: 10 Sep 2005

Volume 1: Fundamentals: Ligands, Complexes, Synthesis, Purification, and StructureVolume 2: Fundamentals: Physical Methods, Theoretical Analysis, and Case StudiesVolume 3: Coordination Chemistry of the s, p, and f Metals

Volume 4: Transition Metal Groups 3 - 6

Volume 5: Transition Metal Groups 7 and 8

Volume 6: Transition Metal Groups 9 - 12

Volume 7: From the Molecular to the Nanoscale: Synthesis, Structure, and PropertiesVolume 8: Bio-coordination Chemistry

Volume 9: Applications of Coordination Chemistry

Volume 10: Cumulative Subject Index

10-Volume Set: Comprehensive Coordination Chemistry II

Volume 2:

Fundamentals: Physical Methods,

Theoretical Analysis, and Case Studies

Edited by

A.B.P Lever

Contents

Section I - Physical Methods

Nuclear Magnetic Resonance Spectroscopy (P Pregosin, H Rueegger).

Electron Paramagnetic Resonance Spectroscopy (S.S Eaton, G.R Eaton).

Electron-Nuclear Double Resonance Spectroscopy and Electron Spin Echo Envelope Modulation Spectroscopy (S.S Eaton, G.R Eaton).

X-ray Diffraction (W Clegg).

Chiral Molecules Spectroscopy (R.D Peacock, B Stewart).

Neutron Diffraction (G.J Long).

Time Resolved Infrared Spectroscopy (J.J Turner et al.).

Raman and FT Raman Spectroscopy (I.S Butler, S Warner).

High Pressure Raman Techniques (I.S Butler, S Warner).

Resonance Raman: Coordination Compounds (J Kincaid, K Czarnecki).

Resonance Raman: Bioinorganic Applications (J Kincaid, K Czarnecki).

Gas Phase Coordination Chemistry (P.B Armentrout, M Rodgers).

X-Ray Absorption Spectroscopy (J Penner-Hahn).

Photoelectron Spectroscopy (Dong-Sheng Yang).

Electrochemistry: General Introduction (A.M Bond).

Electrochemistry: Proton Coupled Systems (K.A Goldsby).

Electrochemistry: Mixed Valence Systems (R.J Crutchley).

Electrochemistry: High Pressure (T.W Swaddie).

Ligand Electrochemical Parameters and Electrochemical-Optical Relationships (B Lever) Mossbauer: Introduction (G.J Long, F Grandjean).

Mossbauer: Bioinorganic (E Muenck et al.).

Optical (Electronic) Spectroscopy (C Reber, R Beaulac).

Stark Spectroscopy (K.A Walters).

Electronic Emission Spectroscopy (J Simon, R.H Schmehl).

Magnetic Circular Dichroism (W.R Mason).

Magnetic Circular Dichroism of Paramagnetic Species E.I Soloman et al.).

Solvation and Solvatochromism (W Linert et al.).

Mass Spectrometry

Neutralization-Reionization Mass Spectrometry

Electrospray Mass Spectroscopy

Magnetism: General Introduction

Electronic Spin Crossover

Excited Spin State Trapping (LIESST, NIESST)

Notes on Time Frames

Section II - Theoretical Models, Computational Methods and Simulation

Ligand Field Theory

Angular Overlap Model (AOM)

Molecular Mechanics

Semiempirical SCF MO Methods, Electronic Spectra and Configurational Interaction (INDO) Density Functional Theory (DFT)

Time Dependent Density Functional Resonance Theory (DFRT)

Molecular Orbital Theory (SCF Methods and Active Space SCF)

Valence Bond Configuration Interaction Model (VBCI)

Time-dependent Theory of Electronic Spectroscopy

Electronic Coupling Elements and Electron Transfer Theory

Metal-metal Exchange Coupling

Solvation

Topology: General Theory

Topology: Assemblies

Electrode Potential Calculations

Comparison of DFT, AOM and Ligand Field Approaches

MO description of Transition Metal Complexes by DFT and INDO/S

Section III - Software

AOMX - Angular Overlap Model Computation

GAMESS and MACMOLPLT

CAMMAG

LIGFIELD

ADF

DeMON

Survey of Commercial Software Websites

Section IV - Case Studies

Mixed Valence Dinuclear Species (J.T Hupp).

Mixed Valence Clusters (Tasuku Ito et al.).

Non-biological Photochemistry Multiemission (A Lees).

Nitrosyl and Oxo Complexes of Molybdenum (M Ward, J McCleverty).

Structure of Oxo Metallic Clusters (R.J Errington).

Iron Centred Clusters (T Hughbanks).

The Dicyanamide System (R.J Crutchley).

Spectroscopy and Electronic Structure of [FeX ] (X=CI,SR)(E.I Soloman, P Kennepohl).

4 n

Nuclear Magnetic Resonance

Spectroscopy

P S PREGOSIN and H RU ¨ EGGER

ETH Ho¨nggerberg, Zu¨rich, Switzerland

2.1.4.2 Principles and Methodologies 20

2.1.1 INTRODUCTION

For more than 50 years, NMR spectroscopy has provided a major aid in solution structure analysis.Starting from modest, 40 MHz machines, one can now measure on instruments approaching thegigahertz range Coordination chemists have been somewhat slow in profiting from this method, asmany of the metal complexes of the first transition series are paramagnetic, and thus only some-times suitable for this methodology Further, sensitivity was initially a problem, i.e., many metalcomplexes are only sparingly soluble; however, the advent of polarization-transfer methods, high-field magnets, and improved probe-head technology have more or less eliminated this difficulty.Measurements of1H,13C,19F, and31P spins on ca 1–2 mg of sample, with molecular weights in therange 500–1,000 Da, are now a fairly routine matter

The spin I¼ 1=2 nuclei with the largest magnetic moments and natural abundance are stillfavored in the inorganic community, e.g., 1H, 13C,19F, 31P,111,113Cd, 195Pt, and 199Hg; however,

15

N, 29Si, 77Se, 103Rh, 107,109Ag, and 183W are now all fairly routine candidates.1–4 The 103Rhliterature is expanding rapidly;5–11however, for other nuclei, e.g.,107,109Ag, the results continue todevelop slowly.12–1457Fe15,16and187Os17–19both represent examples of spins with considerable butnot insurmountable difficulties, primarily due to their small magnetic moments (seeTable 1) Thereare ongoing efforts on quadrupole nuclei,20,21e.g.,67Zn,22,2355Mn,24,2599Ru,26,27and95Mo.28Slowly, multidimensional methods are increasing in popularity within the inorganiccommunity; however, while several of these may be necessary to properly characterize a specificcomplex, they are not all equally useful COSY measurements connect coupled proton spins andare thus useful for assignments However, NOESY data can provide three-dimensional structurefeatures and also reveal exchange phenomena, thereby making these much more valuable for thecoordination chemist

The number of solid-state measurements has increased exponentially, due both to interests inheterogeneous catalysis and to the number of interesting complexes with very limited solubility.Further, relatively new NMR methods are finding application, e.g., PHIP and PGSE diffusion studies,

so that the sections which follow cannot do justice to the individual topics, because of space restrictions

We have tried to emphasize results since about 1990 This will undoubtedly have resulted insome unfortunate omissions

2.1.2.1 Detecting Less Sensitive X-nuclei

The most sensitive and now routinely used method for obtaining spin I¼ 1/2 NMR signals forless sensitive nuclei involves double-polarization transfer (I!S!I ), and uses one of the two-dimensional NMR sequences shown in Figures 1and2.33–35

Table 1 Relative sensitivities for selected nuclei of common interest

3.37 1050.28

IndirectDirectc95

Mo

103

Rh

15.7100

3.23 1033.11 105

DirectcIndirect109

7.20 1041.22 105

Indirect195

199

a

The I-spins are assumed to be a high receptivity nucleus, most often1H, i.e., one needs anJ(X,

1

H) interaction, n¼ 1–4 The data are detected using the proton signals and the spectra are usuallypresented as contour plots, as shown in Figures 3 and 4 Occasionally, 31P o r 19F are suitablealternatives to protons Specifically, for metal complexes containing phosphorus ligands in whichthe31P is directly bound to the metal center, one occasionally has a relatively large1J(M,P) value

of the order of 102103Hz.36–38 Consequently, one need not be restricted to molecules revealingsuitably large proton–metal coupling constants The time  is set to1/(2 J(S, I )), and the time t1

represents the time variable for the second dimension These sequences provide a theoreticalenhancement of (I/S)5/2 For nuclei such as57Fe,103Rh, and183W this means factors of 5,328,5,689, and 2,831, respectively

(b)

I S

(a)

I

Figure 2 Heteronuclear single quantum correlation (HSQC) pulse sequences with optional decoupling of

the S-spin: (a) standard sequence; (b) modified for the I-spin-multiplicity determination

(c)

I S

(d)

I S

decoupling

The diamagnetic screening constant, d, involves the rotation of electrons around the nucleusand is important for proton NMR These electrons may be immediately associated with the atom

in question, or with circulating electrons associated with proximate functionalities, i.e., anisotropiceffects For the paramagnetic screening constant, p(which makes the major contribution to thenuclei13C,15N, 31P,57Fe,103Rh,119Sn,195Pt, etc.), the average energy approximation, for anatom A is often made, i.e.,

Ap/  < r3> BQA;B=E ð4Þ

The term QA,B represents the bond order charge-density terms, r is an average distance fromnucleus A tothe next atoms, and E an averaged energy difference (between suitable filled andempty orbitals).Equation (4)indicates that energies, bond orders, and distances all contribute to

Ap As E can be relatively small (perhaps due toa small n–* o r –* separation), theobserved range of chemical shifts is often hundreds of ppm for donor atoms, and thousands ofppm for transition metals It is not unusual to find several terms inEquation (4)which change as

a function of ligand complexation, so that a thorough understanding of heavy-atom shifts

scale shows the15N chemical shift

requires a more detailed consideration of their source than for proton chemical shifts It isinsufficient to interpret metal chemical shifts using concepts such as ‘‘local electron density’’ atthe atom in question, as this approach can be misleading; e.g., the 13C chemical shift of theanionic carbon in Li(CPh3) is at a higher frequency than that for CHPh3.40 It is clear from theliterature41–49that it is now possible to calculate screening constants (and thus chemical shifts, ,

of heavier atoms) fairly accurately

Often, heavy-atom chemical shifts are considered empirically The range of metal chemicalshifts is usually of the order of thousands of ppm and is very sensitive to changes in, and close to,the local coordination sphere Simple ligand-field-type considerations result in significant changes

in energy levels at a metal center when the donor atoms are changed This will clearly affect the

Eterm inEquation (4), e.g., the Co(H2O)6 þ 59Co resonance is found at ca 15,000, whereas theCo(CN)63– 59Coresonance is at ‘‘0’’ ppm Further, the Rh(H2O)6 þ 103Rh resonance is at 9,924,whereas the Rh(CN)63– 103Rh resonance is at 340 Crude correlations relating the metal chemicalshift with oxidation state or stability50have been found, e.g., for Pt(CN)4 the195Pt resonance is

at 4,746, whereas for Pt(CN)6 

the195Pt resonance is at 3,866 (both vs PtCl6 

); however,ambiguities exist, so that each case should be viewed on its own merits

Solvent effects on metal resonances are routinely tens of ppm, and changes in temperatureduring a measurement result in large enough shifts (often in the range 0.1–0.5 ppm C1) that finestructure on the resonance is readily lost Isotope effects (e.g.,35Cl vs.37Cl,16O vs.18O, or1H vs

2

H) on metal resonance positions51–55are sufficiently large that the different chemical shifts fromthe individual isotopomers are often well resolved These effects are not so marked in donor-atomNMR spectra, i.e., for13C,15N, or31P complexed to a metal center, solvent effects are normally afew ppm or less

For the two donor atoms nitrogen and phosphorus, the normal chemical-shift range is of theorder of hundreds of ppm A change in hybridization from sp3to sp2will be associated with neworbitals These represent orbitals, e.g.,  and *, whose energy separation will strongly affect thechemical shift As an example, the15N resonance for trialkyl amines, R3N, is at300 to 390,whereas the 15N resonance for pyridines is found at þ 80 to 175, both classes relative to

CH3NO2.56 Moreover, complexation of a sigma donor, e.g., either an aliphatic nitrogen or atertiary phosphine donor, simultaneously changes both the lone-pair energy and the local geo-metry at the donor atom, so that interpretation can be complicated For pyridine (or relatedheterocyclic ligands with sp2 donors14), complexation to a metal usually affords a shift to lowfrequency, whereas for triphenyl phosphine complexation there is normally a high frequencychange There exist compilations of both 14,15N56 and 31P36–38 chemical shifts Electronegativegroups on these donors, and inclusion in various ring sizes, as well as the size of the substituent on

Ru P

P Ph2 O OH Ph

OTf

CH2 CH3

δ δ

Figure 4 31P,1H COSY for the complex shown Note that the two POCH2methylene protons are topic and that one of these happens to fall exactly under the solvent (THF) signal However, the correlationreadily reveals two types of cross-peaks and thus the chemical shift of the hidden proton There are also

diastereo-correlations to POH, and ortho and meta protons of the P-phenyl

the donor atom, all play important roles in determining the chemical shift As there are literallyhundreds of reported nitrogen chemical shifts and thousands of measurements for the 31P spin,the reader is advised to consult the reviews noted.

The special case of carbon as donor, i.e., alkyl, phenyl, alkynyl, allyl, CO, olefin (or arene or

Cp, etc.), and carbene ligands, continues to attract significant attention and several articles57–63have been written on this subject Nevertheless, Equation (4) is valid CO derivatives areoften found in the region ¼ 150–250 Carbene compounds have 13

C positions at relativelyhigh frequency, usually >200 ppm, and this special position is often diagnostic Aryl complexesreveal the coordinated ipso carbon at high frequency, with representative values between 130 ppm and

180 ppm Complexed olefins show their13C positions over a wide range, with coordination chemicalshifts as small as 10–15 ppm, but often 30–70 ppm or more The oxidation state of the metal(and thus the d–* back bonding) is important in determining13C frequencies for these complexes.Given that both the metal centers and parts of the ligands can contain strongly anisotropicregions, ligand complexation often has a significant effect on proton chemical shifts Individualprotons can be forced into environments which result in marked high- or low-frequency resonancepositions The axial positions in square-planar complexes often afford high-frequency protonshifts,64,65 e.g., as in (1); and, of course, phenyl ligands (or aromatic substituents), as well asdonors such as pyridine or triphenyl phosphine—which contain aromatic fragments—canstrongly affect the local environments of proximate protons, e.g., as in (2)

L

M

L

M D

(1)

O Pd

Cl H Me

8.22 ppm

(3)

H H

Even simple ligands such as chloride can influence the position of proximate protons, e.g.,

in structure (3), the proton indicated might be expected at around 7 ppm, but it appears above

A modified form of the Pople and Santry expression is given inEquation (5) and

1JðM; LÞ / MLj sðMÞð0Þj2j sðLÞð0Þj2 S

occ

unocc k

ðEk EjÞ1CðMÞksCðLÞksCðMÞjsCðLÞjs ð5Þ

reveals that the one-bond interaction depends on the metal and ligand atom magnetogyric ratios,

, the s-expectation values, , the occupied, j, and unoccupied, k, molecular orbital energies, andthe s-coefficients of the atomic orbitals used in making up the molecular orbitals

Given that the  and s-expectation values, , depend markedly on the individual metal andligand atoms under consideration, the values of these spin–spin interactions vary over severalorders of magnitude, e.g.,1J(195Pt,1H) is often >1,000 Hz, but 1J(103Rh,1H) is usually <30 Hz

1

J(195Pt, 31P) is often >2,000 Hz, but 1J(103Rh, 31P) is usually <300 Hz Since both the  ands-expectation values for the31P atom are relatively large, one finds spin–spin interactions of theorder of 102–104Hz, depending upon the metal and the nature of the phosphorus ligand There is

an extensive literature on 1J(M, 31P),36–38 although much work has involved platinum andrhodium complexes Metal–metal one-69,70and two-bond71coupling constants can be surprisinglylarge and, in some cases, are 20,000–40,000 Hz.69–72 Complexes (5) and (6) are clear, albeitsomewhat extreme, examples of this idea With the possible exception of complexes containing

31

P, (e.g., see (3) above), long-range coupling constants such as 3J(M, (spin¼ 1/2)) have notreceived as much attention as1J(M, L) Nevertheless, spin I¼ 1/2 metals and ligands can easilycouple to protons and other nuclei over three, four, and sometimes more bonds (and theseinteractions can be quite useful for determining metal chemical shifts of e.g., 119Sn,73 109Ag,74

183

W,75–77 195Pt,78,79and 199Hg.80

ClHg

IrCl

COL

(7)

L

M L

Spin = 1/2 nucleus

(9) (10) (8)

There is sometimes a dependence of either  13C, 15N, or 31P, etc and/or 1J(M,spin¼ 1=2 nucleus) on the trans influence of L.81,82 For stronger L-donors the chemical shift

moves to lower frequencies, whereas the value of 1J(M, spin¼ 1=2 nucleus) can be markedlyreduced This is the case for NH3 (or amine) complexes of Pt(II) and, specifically, for (11)and related amino-acid complexes It has been suggested that O, N, and S donors, in thetrans position, can be distinguished by the chemical shift of the 15NH3 signal Sadler andco-workers83–89 have been very active in this area.

In general, 15N chemical shifts in platinum complexes have received increasing attention andmost reports use indirect methods of detection, even on enriched materials The development

of cis-PtCl2(NH3)2, and related cancer drugs, has been accompanied by a renewed interest in

1

J(195Pt,15N).90–97This parameter varies by ca one order of magnitude between 80 Hz and about

800 Hz.15N data can also be used to recognise unique three-center interactions98in Pt chemistry,e.g., (12), or to identify an enol form of the amide, e.g., (13), an amino-quinoline Ir(III)derivative

H

PtP

PPh

LH

OMe

H

L = PPh3

CF3SO3

In the latter complex, the expected one-bond NH interaction, found in acetamide complexes,

is absent.99This type of complex shows a rarely observed2J(15N,1H) value Given the interest inorganometallic and catalytic chemistry, there are many reports on 1J(M, 13C)100–103 and

H, 13C, or 31Prepresent the most abundant examples There are relatively few modern, detailed studies on thissubject, and Field and co-workers106,107 have used 2-D methods to obtain some useful data onsigns and values ofnJ(X, Y) in organometallic complexes of RuIIand RhI

Both chemical shifts and coupling constants have been used113 tocharacterize the novel Pt–pyrazolyl borate formyl complex, (16) The observed coupling constant from the Pt atom to theformyl proton, 327 Hz, is relatively large

The 13C NMR parameters for the Pt–methyl group provide an interesting contrast whencompared to those of the formyl group The methyl carbon resonance is found at 0.48 ppm,with 1J(Pt, C) expected tobe in the range 620–710 Hz

2.1.2.5 Dynamics

NMR spectroscopy allows the coordination chemist access to a variety of dynamic phenomenavia spin–lattice, T1, and spin–spin, T2, relaxation times, plus line-shape analyses and phase-sensitive exchange (NOE) spectroscopy The spin–lattice relaxation time, T1, can be correlated

to molecular tumbling and rotations Classical T2 measurements, together with the Swift andConnick equations for paramagnetic metal systems, lead to ligand-exchange-rate information.114–116

An example of the latter type of application concerns proton and phosphorus T2 relaxationenhancement in several phosphite ester anions by manganese paramagnetic complexes Thesecompounds contain the fragment (17) shown, and analysis of the various relaxation data allowsthe determination of metal/ligand association rate constants without temperature studies.117Theimportant subject of NMR studies on paramagnetic complexes in biological systems, i.e., therather special consequences of porphyrin, phosphate, and amino-acid derived ligands, has beenreviewed several times.118,119

MnOOH

PHO

line-as -allyl anions,124–126 all stem from these types of measurement Their contributions to metalcarbonyl dynamics and rearrangements in cluster compounds is even more pronounced, and wecite selected studies in this very large area of organometallic chemistry.127–148 For slow exchange

–4,750.0–4,700.0 –4,800.0 –4,850.0 –4,900.0 –4,950.0 –5,000.0

117

195Pt NMR

119117

Figure 5 195Pt NMR spectrum of the [Pt(SnCl3)3(2-methylallyl)]2dianion The intensities of the117,119Sn

satellites reflect the number of complexed tin ligands

NPt

Me

OH

between two sites and negligible overlap of the signals, expressions such as Equation (6)120 (ormore complicated versions149) have served well:

(W¼ bandwidth at half height, k ¼ first-order rate constant)

In the early reports, line-shape studies predominated; however, many of the more recent reportsuse 2-D exchange spectroscopy

2.1.2.6 NOE and Exchange Spectroscopy

Nuclear Overhauser effects, NOEs, involve dipole–dipole relaxation phenomena which result insignal enhancements.150

max is:

max¼ ð5 þ !2 2c  4!4 4cÞ=ð10 þ 23!2 2c þ 4!4 4cÞ ð7Þ

(!¼ frequency, c¼ correlation time)

For small molecules with short c values (extreme narrowing limit), this equation reduces to

max¼ þ50% This is rarely achieved for a single proton in coordination compounds, as there areoften a number of spins contributing to the relaxation of an individual proton and the cvaluesare not always so short Clearly, if the quantity (5þ !2 c2– 4!4 c4)¼ 0, then there is noNOE It

is well known150

maxcan pass through zero and the limiting value is100% This can be thecase for biological or other macromolecules Further, a negative NOE is also possible for highermolecular weight metal complexes, e.g., MW > 1,000, and/or in viscous media (perhaps due tolow-temperature studies) In these cases ROESY spectra150,151can be useful

Although selective1H NOE studies and magnetization-transfer experiments are still frequently

in use,150 the simple three-pulse (phase-sensitive) 2-D NOESY sequence, given in Figure 6, isfinding increasing popularity.152–159 The mixing time should be chosen such that exchange cantake place without losing too much signal intensity Practically, this often means values in therange 0.4–1.0 seconds, although individual T1 values and temperature will require that thisparameter be constantly adjusted to suit the coordination compound in question

Wherever coordination chemistry problems overlap with those of organic chemistry, e.g., formational analysis,1H NOE studies will have their classical value Chiral inorganic complexeshave been studied with emphasis on inter- and not intra-ligand NOEs.155,156,160–166These resultsallow the determination of the 3-D structure of the complex and thus, for enantioselective catalysts,the shape of the chiral pocket offered by a chiral auxiliary to an incoming organic substrate Sincemany such auxiliaries possess phenyl phosphine donors, the interactions between the ortho protons

con-of the P-phenyl group and those from a second ligand make a decisive contribution to the structuredetermination Structure (18) shows a hypothetical Pd(chiraphos)-(allyl) cation, and it is easy to seehow NOEs, from the three allyl protons to the P-phenyl ortho protons, can provide useful structuraldata The four phenyl groups, two pseudo-axial and two pseudo-equatorial, are all nonequivalent

For complexes of modest size which tumble relatively rapidly, 2-D NOESY methodsdistinguish between NOE and exchange phenomena via the phases of the signals The diagonaland exchange peaks have the same phase in contrast tothose due toNOE Since the 2-Dmethodology is not selective, i.e., all of the spins are excited simultaneously, the exchange mapcan reveal several species in exchange with each other as well as two or more different processes.

A nice example is provided by the tetranuclear Ir-cluster anion (19).167 The CO ligands areinvolved in several temperature-dependent exchange processes

+

(18)

One of these, the so-called ‘‘merry-go-round,’’ selectively exchanges the bridging and terminal

CO ligands, a, d, b, and f, in the pseudo-equatorial direction Since the various13CO signals can

be assigned, 2-D 13C NOESY spectroscopy reveals exchange cross-peaks connecting all four ofthese signals, thus identifying this selective process, as indicated in the drawing

A unique aspect of this form of exchange spectroscopy concerns the ability to detect specieswhose concentration is so low that they escape detection in a conventional one-dimensionalexperiment Figure 7 shows a section of the 1H NOESY spectrum for a mixture of isomericpalladium phosphino–oxazoline, 1,3-diphenylallyl complexes.168 One observes a major com-ponent in exchange with a visible minor component (ca 10% of the more abundant isomer).However, there are additional, very broad, exchange cross-peaks from the main isomer to an

‘‘invisible’’ species, which would easily have gone undetected

Interest in19F,1H NOEs in coordination chemistry is developing,169,170and several interestingexamples of31P,31P exchange spectroscopy have been reported.171–173

2.1.2.7 Special Topics

2.1.2.7.1 High-pressure studies

NMR studies under high pressure have increased markedly in the last twodecades Technically, thesemeasurements are most frequently carried out using sapphire NMR tubes, and this methodology hasbeen modified over the years.174–179These pressure experiments are usually carried out with the jointaims both of determining activation volumes and of shifting chemical equilibria Occasionally, detailswith respect tothe pressure dependence of NMR parameters are published.180

Measuring rate constants vs pressure allows the determination of activation volumes, and thusgives a hint as to whether the reaction mechanism is associative or dissociative

Much work has been done on solvated metal complexes by Merbach and co-workers.178,181–189These pressure studies have been extended toorganometallic CO190–192 and SO2193 complexesplus, interestingly, the first dihydrogen aqua-complex, Ru(H2)(H2O)5 þ

, (20),194 produced asshown inEquation (9):

The 1J(H, D) value of 31.2 Hz in the H2ligand allows an estimation of the H–D separation,

ca 0.90 A˚, using the relationship:

suggested by Maltby et al.195 It is probably useful to remember196 that correlations with vation volumes may not be straightforward Elsevier and co-workers27,170,197–199 have used highpressures in connection with supercritical fluids, and have studied effects on line widths and otherNMR parameters

acti-Homogeneously catalyzed hydrogenation chemistry, often under an overpressure of gas, hasbeen followed by proton NMR for decades, and frequently important intermediates go unde-tected due to their relatively low concentration Since the para hydrogen induced polarization,(PHIP) signal magnification can be several orders of magnitude, Bargon,200–207 and the Duckettand Eisenberg groups208–222plus others have studied in situ reactions using parahydrogen undermild hydrogen pressure The major limitation arises from the necessity for the two parahydrogenatoms to be transferred pairwise The PHIP effect has also been recently shown to be usefulfor 13C, as well.200

The PHIP approach has been used to help identify the cationic Rh(I) dihydrido-bis-solventocomplex shown, (21).222

This type of dihydrido-phosphine chelate complex is often mentioned in mechanisticdiscussions on enantioselective hydrogenation, but was previously thought to be not very stable

8a8c8c

8b8b8a

4.5

5.0

5.5

ppm4.5

5.05.5

OMeOMe

H

PP

2.1.2.7.2 Molecular hydrogen and agostic complexes

Much effort has been invested in the use of NMR methods to study molecular hydrogencomplexes.223–241 The identification of an Lm 2–H2) often requires variable temperature,deuterium enrichment, and T1 studies The deuterium incorporation is useful in that the value

of1J(H, D) can be diagnostic, as noted above

In polyhydride complexes, exchange between hydride and complexed molecular hydrogen oftenleads toobservable dynamics in their 1H NMR spectra In many cases these processes areassociated with relatively low activation-energy barriers.242 In the complex IrH2X(H2)(PR3)2,(22), X¼ Cl, Br, or I, the exchange can proceed via either hydride/hydrogen exchange leadingto(23), or oxidative addition leading to (24)

R3P

Ir

HH

has commented on how these interactions are related The name ‘‘agostic’’

is often used244–246 for the case of X¼ a suitably substituted carbon atom There are also anumber of examples of X¼ a suitably substituted silicon atom.247–250

The agostic interaction of a

CH bond, (25), results in a low-frequency shift of the proton resonance (due to the development

of ‘‘hydride-like’’ character) and substantial reduction in the one-bond coupling constant,1J(13C,

1

H) This reduction can be 50% or more Similarly, for X¼ SiR3, the one-bond,1J(29Si,1H) valuedecreases In the solid state one finds the CH bond as a donor to the metal There are manyexamples of this type of interaction.251–261

2.1.2.8 Relaxation

Relaxation times can be useful for coordination chemists For our discussion it is sufficient to expressthe longitudinal relaxation rate of a nucleus, R1, (¼ 1/T1), as the sum shown inEquation (11):262

R1¼ RDD þ RCSA þ RSR þ RSC þ RQ þ REN þ Rother ð11Þ

with the various contributions defined as follows: DD¼ dipole–dipole; CSA ¼ chemical-shiftanisotropy; SR¼ spin rotation; SC ¼ scalar coupling; Q ¼ quadrupole; and EN ¼ electron–nuclear It should be noted, however, that for coupled-spin systems, this simple sum is no longervalid.263–265

The measurement of the relaxation rate gives the coordination chemist access to parametersrelated to the anisotropic interactions described by the spin Hamiltonian which, in solution, areaveraged to their isotropic values or even zero In principle structural information can thus beretrieved, since R1DDand R1Erender information with respect to the separation to other nuclei orunpaired electrons, respectively, e.g., via the 1/r6 distance dependence shown in Equation (12).Frequently, a number of dipoles contribute to relaxation, so that a sum is necessary:

In practice, however, it is often difficult to separate or exclude some of the contributingpathways unless one of the interactions is clearly dominating

The R1DD term can usually be evaluated Consequently, data from one- and two-dimensionalnuclear Overhauser spectroscopy studies contribute to the coordination chemists understanding

of three-dimensional solution structures152–166and molecular association phenomena such as ionpairs.169,170,266–269Distance constraints are usually qualitatively established, based on cross-peakintensities or volumes Occasionally monitoring the build-up rates is preferred, in order toquantify internuclear distances.266,268

The determination of R1DD, and in particular the maximum rate, i.e., the T1 minimum, ispopular for the determination of the HH distance in molecular hydrogen complexes, as theintraligand HH separation is much shorter than other interproton distances.270 The HHdistances are calculated from the T1 minima according to two models involving static271 orfast-rotating hydrogen ligands,272 respectively Distances thus derived should be considered assemiquantitative, as additional spins (e.g., other hydride ligands in polyhydrides) or dipolarcoupling to NMR active metal centers may shorten T1.273 Other relaxation contributions, such

as the spin-rotation mechanism, may not be ruled out Moreover, the exact nature of liganddynamics (classical vs quantum-mechanical rotation and tunneling of hydrogen) is not settled.270

R1CSA is an important contributor to the relaxation of heavy nuclei, particularly for thetransition metals, and can be separated from the other contributions due

toits unique B02dependence (see Equation (13)) Structural conclusions have been derived fromthis parameter, e.g., linear, trigonal, and tetrahedral Pt(PR3)n complexes can easily bedistinguished from their 195Pt T1CSAvalues.29

R1Q is normally the dominating relaxation pathway for quadrupolar nuclei For a series ofmetal deuterides, quadrupole coupling constants have been determined using this method, thusshedding light on the size of the electric-field gradient at the D nucleus These results reflect thecharacteristics of the MD bond, in particular ionic vs covalent contributions.274–276

R1

EN

is responsible for relaxation of the nuclei in a paramagnetic complex and dependsstrongly on the relaxation rate of the unpaired electrons, correlation times for molecular reorien-tation, ligand-exchange rates, the bonding situation, and the electron–nucleus distance The study

of various enzymes containing paramagnetic metal centers,118,119,277–283and the use of complexes

of rare-earth metal ions as contrast agents in magnetic resonance imaging,284–287 represent twoimportant applications of this methodology

The term R1othersummarizes other possible contributions to spin–lattice relaxation, e.g., a spin–photon Raman scattering mechanism has been proposed for relaxation of the 205Pb nucleus inlead nitrate and other heavy spin-1=2 nuclei in solids.288

2.1.3 NMR DIFFUSION MEASUREMENTS

2.1.3.1 Introduction

The determination of relative molecular size in solution remains a subject of considerable interest

to the coordination chemistry community, in particular with respect to the formation of nuclear complexes, ion pairs, and otherwise aggregrated species Apart from classical methodssuch as mass spectrometry289(see Chapter 2.28) and those based on colligative properties290—boiling-point elevation, freezing-point depression, vapor and osmotic pressure—the Pulsed FieldGradient Spin-Echo (PGSE) methodology291,292has recently resurged as a promising technique.PGSE measurements make use of the translational properties, i.e., diffusion, of molecules andions in solution, and thus are directly responsive to molecular size and shape Since one canmeasure several components of a mixture simultaneously,293,294PGSE methods are particularlyvaluable where the material in question is not readily isolable and/or the mixture is of especialinterest

poly-PGSE methods were introduced in 1965 by Stejskal and Tanner295and, since then, have beenwidely used In the 1970s this approach was used to determine diffusion coefficients of organicmolecules.296 In the following decade, variants of the technique were applied to problems inpolymer chemistry.297–300 Since then, diffusion data on dendrimers301–306 and peptides,307–310 aswell as on molecules in various environments, e.g., in porous silica311 and zeolites,312have beenobtained Recent applications of PGSE methods in coordination and/or organometallic chemistryhave emerged.169,313–326

2.1.3.2 Methodology

The basic element of an NMR diffusion measurement consists of a spin-echo sequence,327 incombination with the application of static or pulsed field gradients.295,328 Several commonsequences are shown inFigure 8, and we discuss these only briefly as the subject is covered inseveral reviews.291,292,329–334

2.1.3.2.1 Spin-echo method

In the Stejskal–Tanner experiment,295 Figure 8a, transverse magnetization is generated by theinitial /2 pulse which, in the absence of the static or pulsed field gradients, dephases due tochemical shift, hetero- and homonuclear coupling evolution, and spin–spin (T2) relaxation Afterapplication of an intermediate  pulse, the magnetization refocuses, generating an echo At thispoint the sampling (signal intensity measurement) of the echo decay starts Fourier transform-ation of these data results in a conventional NMR spectrum, in which the signal amplitudes areweighted by their individual T2values and the signal phases of the multiplets due to homonuclearcoupling are distorted Both effects are present in the diffusion experiment; however, due to thefixed timing, these are kept constant within the experiment

, respectively, and strength G

The application of the first pulsed linear field gradient results in an additional (strong) dephasing

of the magnetization, with a phase angle proportional to the length () and the amplitude (G) o f thegradient Because the strength of the gradient varies linearly along, e.g., the z-axis, only spinscontained within a narrow slice of the sample acquire the same phase angle In other words, thespins (and therefore the molecules in which they reside) are phase encoded in one-dimensionalspace The second gradient pulse, which must be exactly equal to the first, reverses the respectivephases and the echoforms in the usual way If, however, spins move out of their slice intoneighbouring slices via Brownian motion, the phase they acquire in the refocusing gradient willnot be the one they experienced in the preparation step This leads to incomplete refocusing, as inthe T2dephasing, and thus to an attenuation of the echo amplitude As smaller molecules movefaster, they translate during the time interval  intoslices further apart from their origin, thusgiving rise tosmaller echointensities for a given product of length and strength of the gradient

2.1.3.2.2 Stimulated echo method

The second experiment, shown in Figure 8b,328works in quite the same way, with the differencethat the phase angles which encode the position of the spins are stored along the z-axis in therotating frame of reference by the action of the second /2 pulse Transverse magnetization andthe respective phases are restored by the third /2 pulse This method is advantageous in thatduring time , T1as opposed to T2is the effective relaxation path Since T1is often longer than

T2, a better signal/noise ratio is obtained Furthermore, phase distortion in multiplets due tohomonuclear coupling is attenuated

2.1.3.2.3 Derived sequences

The accurate determination of diffusion coefficients for large, slow-moving species requires stronggradient amplitudes The resulting eddy-current fields can cause severe errors in the spatial phaseencoding The sequence shown in Figures 8c335,336and8d337,338contains an additional, so-calledlongitudinal eddy-current delay (LED) element, i.e., magnetization is again stored along the z-axisduring the decay time of the eddy currents Disturbance of the field-frequency lock system can beminimized by the use of bipolar field-gradient pulses, Figure 8d

Technically, all of the above experiments are performed by repeating the sequence whilesystematically changing the time allowed for diffusion (), or the length () or the strength (G)

of the gradient Mathematically, the diffusion part of the echo amplitude can be expressed by

The diffusion constant D can be related to the hydrodynamic radii of the molecules via theStokes–Einstein Equation (15):

volumes,322,324,325analogous complexes, or from calculated structures.324The agreement betweenthe twoparameters—seeFigure 10andTable 2—is generally acceptable: perhaps even too good,given the assumption that all of the complexes have spherical shapes.

Given favorable receptivity and T1and/or T2relaxation times, one is not limited in the choice

of the nucleus measured Although most studies in coordination and organometallic chemistryinvolved the observation of1H, the use of alternative or additional nuclei often gives a comple-mentary view Studies based on7Li,325 13

C,324and 19F169,316,322,326 have appeared

2.1.3.3 Study of Complex Nuclearity

The determination of molecular size in solution is a frequent problem for coordination pounds; e.g., in lithium and copper, as well as in transition-metal carbonyl and hydroxo/oxochemistry, one finds numerous examples of polynuclear species Increasingly, use is made ofNMR diffusion measurements to directly assess molecular volumes in solution

com-Venanzi and co-workers320 characterized the equilibrium between the monomeric RuCl2(mesetph)—mesetph¼ (C6Me3H2)P{CH2CH2PPh2}2—and the dinuclear [Ru2(-Cl3)(mesetph)2]Clcomplexes, based on the 1.23 ratio of their diffusion coefficients indicating a doubling of thevolume for the latter The structures of the mixed-ligand dinuclear complexes (MeOBiphep)-

were postulated from identical diffusion rates for both subunits and their larger volumes

com-6

-p-cymene) RuCl]Cl and Ru2(-Cl)2(Cl)2 6-p-cymene)2 Interestingapplications in zirconocene chemistry involved (i) the characterization of the dinuclear intermediate[{Cp2ZrCl}2(-O2CH2)] in the course of the CO2 reduction with Cp2Zr(H)Cl324; and (ii) theobservation of ion quadruples for Cp2Zr(Me)2in the presence of a Lewis acid like B(C6F5)3.313Diffusion measurements also showed that addition of isonitrile to coordinatively unsaturatedtetrameric copper(I) complexes proceeds with the retention of the square CuS core.321

L¼ AsMexPh3x, x¼ 3, 2, 1, 0 (increasing molecular volume from left to right) The absolute value of the

slope decreases with increasing molecular volume

Experimental diffusion coefficients for the dimeric and tetrameric THF solvated n-butyllithiumaggregates, [n-BuLi]2THF4 and [n-BuLi]4THF4, respectively, agree well with those calculatedfrom X-ray or PM3 structures.325 In terms of larger species, Valentini et al.317 investigateddendrimeric ferrocenylphosphine ligands, while in a bioinorganic application, Gorman et al.301estimated the hydrodynamic radii of iron–sulphur-cluster-based dendrimers with the cube Fe4S4

core The largest examined particles containing coordination compounds result from the assembly of 30 {R3P}2{CF3SO3}Pt–(C6H4)n–Pt{CF3SO3}{R3P}2, n¼ 1 and R ¼ Et, or n ¼ 2 and

self-R¼ Ph; and 20 tri(40-pyridyl)methanol into dodecahedra with 55 A˚ and 75 A˚ diameter.314

2.1.3.4 Study of Ion Pairing

Occasionally one can determine the diffusion coefficients for cations and anions separately, andthus determine whether they move together as ion pairs or separately as free ions Frequently,coordination compounds in use in homogeneous catalysis possess anions such as PF6 

minimized gas-phase structures (PM3)

Table 2 Hydrodynamic and crystallographic radii

Macchioni and co-workers,315in pioneering work, measured diffusion coefficients of the structurallyclosely related complexes trans-[Ru(PMe3)2(CO)(COMe)(pz2CH2)]BPh4and trans-[Ru(PMe3)2(CO)-(COMe)(pz2BH2)], and found clear evidence for ions in nitromethane, ion pairs in chloroform at lowconcentrations, and ion quadruples in the latter solvent at high concentration Ion quadruples were alsomentioned by Beck et al.313 for zirconocene compounds in benzene solution An interesting solvent

6

-benzene)(PBu3)2]PF6 and [Pd(diphenylallyl)(Duphos)](CF3SO3).322,326 In CD2Cl2 the diffusion coefficients are quite different, with the anionsmoving much faster, whereas in CDCl3they are the same within experimental error, suggesting thatfree ions and ion pairs are present in the two solvents, respectively Effective doubling of the volume was

6

-p-cymene)]Cl with BArF, a containing derivative of tretraphenylborate.322The conclusion that the BArF analog is present as arelatively tight ion pair is supported by the19F diffusion measurement on the anion giving almost thesame diffusion coefficient derived from the1H study.322Ion pairing with the BArF anion has also beenreported for a series of iridium complexes, whereas analogs with PF6 are separated.169

fluorine-Given that there are several known examples of anions that affect results from homogeneouslycatalyzed reactions,349–351 studies of ion-pairing effects assume new significance

2.1.3.5 Study of Hydrogen Bonding

Yet another promising area concerns hydrogen bonding in metal complexes

The twotriflates in complex (28), were found to move at almost the same rate.322 Althoughtight ion pairing could be an explanation, it was concluded that hydrogen bonding to the P–OHfragment carries the noncoordinated triflate effectively along with the cation.322

in solving problems in coordination chemistry

2.1.4.1 Introduction

Since about 1980, NMR spectroscopy of coordination compounds in solution has been ingly used on a routine basis to address a multitude of new and older chemical problems Theintroduction of two-dimensional correlation methods afforded quick access to parameters forrelatively rare spin-1=2 nuclei Further, three-dimensional solution structures can now routinely

increas-be solved, including not only their constitution but also all aspects of configuration, ation, and intra- and intermolecular dynamics

conform-Although there are now hundreds of publications on the applications of solid-state NMRspectroscopy in coordination chemistry, this technique has not yet made a similar transition Itstill remains mostly in the realm of ‘‘specialists,’’ often more interested in the physical propertiesitself than in their chemical significance This is certainly partly due tothe additional equipmentand knowledge required, but also due to the neglect of chemists who define structural chemistry

as X-ray crystallography At the moment, solid-state NMR exists only as a tool for bridging thegap to solution studies, thereby overlooking the inherent wealth of information available.Naturally, but certainly not exclusively, solid-state NMR spectroscopy is the method of choicefor all those materials that are neither crystalline nor soluble, e.g., coordination complexes

adsorbed or covalently linked to organic or inorganic polymer supports, and compounds inamorphous or glass phases.

2.1.4.2 Principles and Methodologies

Comprehensive treatments of solid-state NMR spectroscopy are available elsewhere.352–358 Forthis discussion, it is sufficient to express the principal spin interactions as the following sum:

H¼ HZ þ HCS þ HQ þ H IS

D þ HIS

where the subscripts in the Hamiltonians denote the following relevant interactions: Z, Zeeman;

CS, chemical shift;359–361 Q, quadrupolar;362,363D, direct dipolar coupling; and J, indirect orscalar spin–spin coupling.364,365 As NMR measurements are usually carried out in strongmagnetic fields, the Zeeman interaction is dominant and the other terms represent only modestperturbations Only in cases where a quadrupolar nucleus is involved will the magnitude of thequadrupolar coupling constant, , be comparable to, or greater than, its Larmor frequency, L.All of the above interactions transform as tensors under rotation, and thus their magni-tudes depend on molecular orientation.355 The values of the familiar chemical shifts aredetermined not only by the position of the nucleus within the molecule, but also by theorientation of each molecule or crystallite with respect to the external magnetic field Insolution, where molecules are tumbling fast with respect to the Larmor frequency, thussampling all possible orientations on a short timescale, chemical shifts and scalar couplingconstants average to their isotropic values, and to zero for the traceless quadrupolar anddipolar interactions

Given the angular dependence mentioned, it is obvious that the anisotropic spectra obtainedfrom the condensed phase must be much richer in information, if more complicated They containall the essential geometrical information describing a molecule in terms of angles betweenthe chemical shift, electric-field gradient, direct and indirect dipolar tensors The angulardependence of the resonance frequencies can be separated from the molecular contribution bymonitoring line positions as a function of a systematic rotation of the sample in the threeorthogonal directions Because of resolution restrictions, this can generally be realized only withsingle crystals, and the most accurate results are still obtained using this method

In a powder sample, the orientations of the molecules are fixed within the rigid lattice but thecrystallites are randomly distributed Their resonance frequencies generally sum to form a broad

‘‘powder line,’’ representing the sum of their individual contributions The frequency spanencountered for such lines is often larger by orders of magnitude than the differences based onthe position within the molecule itself, and the limited resolution is of general concern Twodifferent approaches to address this problem have been proposed: (i) line-narrowing techniqueswhich simulate the tumbling of molecules either in the spin space using elaborate multipulsesequences,366–368 or in real space with macroscopic rotation of the sample around specific,so-called ‘‘magic,’’ angles369–372with the aim of observing an isotropic spectrum; (ii) application

of additional frequency dimensions in two- and multidimensional experiments which correlate

or separate the different anisotropic spin interactions.373 A combination is also possible, i.e.,correlating an anisotropic spectrum with one of its isotropic counterparts, thus retaining thegeometric information while benefitting from the high resolution of the latter.373

The first approach is well established and experiments employing magic-angle spinning (MAS)constitute the bulk of all reported studies Fast rotation, relative to their frequency span, around

an angle of 54 440 reduces the chemical shift and scalar coupling interactions to their isotropicaverages and the dipolar interaction to zero Averaging of the quadrupolar interaction, described

by a rank 4 tensor, to zero requires an additional simultaneous rotation around the angle of

30 340which is achieved in the double-rotation (DOR) experiment.370,374,375Alternatively, in thedynamic angle-spinning (DAS) 2D experiment, the rotor hops sequentially between the angles 37 230and 79 110.376–378Anisotropies at one angle cancel those at the other

A potential problem in obtaining solid-state NMR spectra involves the longitudinal relaxationtimes which tend to be long, thus compromising sensitivity To overcome this, cross-polarizationfrom abundant nuclei, such as1H, tothe dilute spin, S, under observation may be employed.379–383Recycle delays are then determined by the T1of1H rather than of S In addition, the magnetiza-tion of S can be increased up toa maximum of  /

The quadrupolar-echo experiment represents the most widely used experiment for theobservation of quadrupolar nuclei.384 For half-integer nuclei, it may be tuned to observe only thecentral transition (1/2–1/2), which is perturbed by the quadrupolar interaction only to secondorder, thus allowing the observation of less dominant anisotropic interactions A significantimprovement in sensitivity can be obtained by ncorporating a spin-echo method such as theCarr–Purcell–Meiboom–Gill sequence into the detection period.385–388The powder line-shape splitsinto a manifold of sidebands, from which information on the homogeneous and inhomogeneousinteractions can be extracted from the line-shape of the sidebands and their envelope, respectively.One- and two-dimensional multiple-quantum techniques have been introduced for the observa-tion of quadrupolar nuclei with half-integer spin These methods have proven powerful inresolving overlapping resonances of multiple sites.389–395

2.1.4.3 Spin-1/2 Metal Nuclei

The importance of NMR for molecular structure determination rests primarily on the phenomenon ofchemical shielding effects, which are particularly large for the heavy atoms Isotropic metal chemicalshifts obtained from MAS studies are often sufficient to solve most chemical problems, as there is now

a large empirical body of data derived from solution NMR However, taking the orientationdependence of the chemical shielding into account provides considerably more insight into thebonding at the metal center Studies concerned with establishing the span of the metal chemical-shift anisotropy and its relation to coordination geometries and oxidation states have beenreported.396–419 The topic was reviewed in the 1990s for the d- and p-block elements,396,397and inparticular alsofor mercury compounds.398Procedures for intrumental set-up have been suggested.399The span,420 ¼ 11–33, of the 199Hg chemical shift tensors in [Hg(S-2,4,6-iPr3C6H2)2],[PPh4][Hg(S-2,4,6-iPr3C6H2)3], and [NMe4]2[Hg(SC6H4Cl)4] are found to be 4479, 1548, and

178 ppm, respectively.398 This decrease, which follows the sequence linear ML2>trigonal

ML3>tetrahedral ML4, was alsoestablished for 195Pt shielding, based on the CSA contribution

to T1relaxation in Pt0phosphine complexes.29 Interestingly, the sequence found for sp, sp2, and

sp3 hybridized carbon is similar.421–423 Octahedral coordination environments generally showsmaller spans, e.g., <120 ppm for M2PtCl6, M¼ Na, K, NH4 complexes,409–412 whereas thoseassociated with square-planar geometries can be very large,409,413–415 e.g., 10,414 ppm in

K2PtCl4.409 Large shielding anisotropies were also observed for 2-coordinate tin(II) pounds,406–408as opposed to moderate ones in tin(IV) complexes.404–406

com-In bioinorganic applications,113Cd has attracted considerable interest,400–402due toits function

as an effective spin-spy in monitoring the active site in metalloproteins where the native metalshave poor spectroscopic properties The elements of the shielding tensors as a function of thedonor atoms, N, O, and S, and the coordination number have been determined, and empiricalcorrelations to structure deduced.401–403

2.1.4.4 Quadrupolar Metal Nuclei

NMR spectra of quadrupolar nuclei are of particular interest, because they offer a uniqueopportunity to investigate both the electric-field gradient and the chemical-shift tensors at acommon nuclear site simultaneously Furthermore, many of the biologically relevant nuclei,e.g., 23Na, 25Mg, 39,41K, 55Mn, 59Co,424–427 61Ni, 63,65Cu,428 67Zn,22,429,430 95,97Mo,431,432 arequadrupolar The measurement and interpretation of solid-state NMR spectra of these nucleiobtained on powders is still a challenge New spectroscopic techniques385–395 and especially theimpressive advances in computational chemistry, which make possible the calculation of bothelectric-field gradient and chemical-shift tensors, provide much-needed assistance Solid-stateNMR on quadrupolar d- and p-metals was reviewed in the early 1990s.396

The Euler angles relating the electric-field gradient and chemical shielding tensors in theacetylacetonato complexes of beryllium,433aluminum,434and cobalt424,427have been determined

In the case of 59Cothe results arise from single-crystal NMR.427 Euler angles relating the twointeractions have also been retrieved from the manifolds of the spinning sidebands of the centraland satellite transitions, observed in 51V MAS spectra of a series of ortho- and meta-vanadates.434,436 Reports on the chemical shielding anisotropies for quadrupolar nucleiare becoming more frequent in the early 2000s.424–428,432–439

2.1.4.5 Ligand Nuclei

2.1.4.5.1 1H NMR

The concentrated presence, high abundance, high gyromagnetic ratio, and low chemical-shiftdispersion of the proton make this nucleus a difficult one: homogeneous broadening due tostrong homonuclear dipole–dipole couplings lead to featureless absorptions much broader thanits chemical-shift range As a consequence of the homogeneous nature of the broadening, slow-spinning MAS only scales the interaction and does not resolve the spectrum into an isotropic partand a spinning side-band manifold Ultrafast spinning is mandatory; otherwise homonucleardecoupling sequences—either alone368 or in combination366,368 with magic-angle spinning—would have to be employed Fortunate cases exist where one particular interaction dominatesthe dipolar Hamiltonian

Molecular hydrogen complexes are again special, since the strong dipolar coupling due to theshort HH distance dominates all other interactions Zilm et al.270,440

reported the solid-state1HNMR spectrum of W(H2)(CO)3(PCy3)2 showing a distorted Pake-pattern, from which inform-ation on the HH separation and the chemical-shift anisotropy of the hydrogen atoms could beobtained The temperature dependence indicated an in-plane motion of 16 440 Similar resultswere also obtained for Mo,270Mn441, and Ru270hydrogen complexes

2.1.4.5.2 2H NMR

The deuterium quadrupolar coupling constant is a sensitive measure of the magnitude of theelectric-field gradient at the nucleus, and is therefore affected by the bonding situation Thedeviation from axial symmetry sheds light on the bonding mode, e.g., in distinguishing molecularhydrogen complexes from classical hydrides The facile isotope substitution for hydride andhydrogen ligands renders the observation of deuterium an alternative to1H In general, motionalproperties of ligands are easily assessed with the 2H NMR method

A highly asymmetric quadrupolar tensor has been found for OsCl2(D2)(CO)(PiPr3)2, togetherwith characteristic line-shapes in the MAS side-bands, originating from interference of dipolarand quadrupolar interactions.442 From the strength of the dipolar interaction, DD distancescould be obtained as well as the relative orientations of the tensors.442 Solid-state 2H NMRdata have also been reported for W(D2)(CO)3(PiPr3)2and interpreted as a motionally averagedquadrupolar tensor of axial symmetry due to significant molecular motion.443 Coherent D2

rotational tunneling and incoherent D2dynamics were shown to affect the2H NMR lineshapes ofnonclassical Ru(D2) complexes.444,445

2

-C2D4)(PPh3)2ruled out rotation of the ethyleneligand based on the observed 2H quadrupole coupling constants, which are comparable inmagnitude to those found in rigid olefins.446

2.1.4.5.3 13C NMR

Structural data on coordination compounds can be gathered from 13C CP–MAS spectra of thecarbon nuclei in the backbone of ligands.447Here, however, we restrict the discussion to carbonnuclei directly bound to the metal center, i.e., -bonded CO,448–459 CN,428,429,460–465acetylide,414,415 alkyl466,467 and aryl ligands,447,466 and the -bonded alkenes, alkynes,446,466–469cyclopentadienyls, and arenes.470–474

Metal carbonyl complexes are at the center of many areas of chemistry Interest has focussed ontheir dynamic processes,447,449but also on the shielding tensor itself.454–458With respect tothe latter,the three modes (1to 3) of CO bonding have been investigated;455CO coordination is characterized

by 13C tensors spanning ca 400 ppm and 200 ppm for terminal and bridging COs, respectively

A large deviation from axial symmetry is observed for the 2-mode, and it is interesting to note that inFe(CO)5the shielding of the equatorial ligands deviates slightly from axial symmetry.455

Copper(I) cyanides have been extensively studied by 13C solid-state NMR, and reveal axialchemical-shift tensors plus the novel coupling constants 1J(63,65Cu, 13C) and Cu–C separationsfrom the dipolar coupling.428,460 CuCN has a linear polymeric structure [–Cu–CN–]n, with thecyanides subject to‘‘head–tail’’ disorder.460

The low-frequency isotropic chemical shift normally observed for -bonded alkylorganometallic carbon is due to one particularly shielded component, whose orientation withrespect to the carbon–metal bond could not be determined.466For methyl groups, the span,  ofthe shielding tensor is larger than for organic compounds and depends strongly on the otherligands present in the complex.466

2

-bonded olefins.446,466–469The spans

of the chemical-shift tensors are reduced with respect to the free olefins, which is discussed interms of the Dewar–Chatt–Duncanson model of -donation and -back-bonding.446,468The CCbond lengths and the orientations of the shielding tensor elements are available from dipolar-chemical shift methods and 2-D spin-echo experiments on the doubly13C labelled olefins.445,469

-benzene ligands of transition-metal complexes,466,470–473 but alsosome derivatives of alkali or main-group elements,474exhibit single13C resonances and shieldingtensors of axial symmetry at room temperature Both observations point to relatively fastrotations around the respective 5- and 6-fold local rotor axis.470,471

2.1.4.5.4 15N NMR

The question of nitrogen vs carbon in cyanides as a donor atom has been investigated by 15Nsolid-state NMR.428,460,461This problem could also be addressed from the metal side, given that asuitable spin-1/2 metal is involved.475 Interest in the antitumor reagent cis-[PtCl2(NH3)2]prompted studies on the15N characteristics of this and related types of platinum complexes.476–478Cobaloximes with aniline and pyridine ligands have been investigated, yielding the orientations

of the nitrogen shift and the cobalt electric-field gradient tensors with respect to the molecularframe, plus the signs and magnitudes of1J(59Co, 15N) scalar and the59Co quadrupole couplingconstants.479Nitroso and aryl-nitroso metalloporphyrins have been studied with interest towards

2.1.4.5.7 31P NMR

31

P in phosphines, phosphates, and their metal complexes remains the most studied ligandnucleus Developments up to 1992 have been summarized by Davies and Dutremez.483 MASstudies often show isotropic chemical shifts and isotropic scalar coupling constants very similar tothose derived from solution studies.484–488

Static disorder in crystals represents a problem which can lead to a crystallographic symmetryhigher than its molecular symmetry, thus hiding important structural features In this respect thepresence of a PHPd agostic interaction in the dinuclear Pd(I) complex (31) was established492

insolution and shown to remain intact in the solid phase, as evidenced from the four quite different31Pparameters of the four P ligands, which were in good agreement with their solution values.493Thecentrosymmetrical crystal structure suggested that only half of the molecule would be independent.493

The magnitudes of the one-bond metal–phosphorus coupling constants represent a rich source ofstructural information.483–490,495–498 The anisotropy of the indirect 199Hg–31P spin–spincoupling constant has been determined.497–499 Additional information on metal-ligand one-bondcoupling constants can be gained from MAS spectra of phosphorus ligands coordinated toquadrupolar metal nuclei In solution, T1values for the latter nuclei are extremely short, unlessthey are in a cubic environment, leading to ‘‘self-decoupling’’ effects In the solid phase, these T1

CoN

NPPh3

values are long, even for quadrupolar nuclei, and coupling to phosphorus ligands is observed.One-bond scalar couplings have been reported for the following spin >1/2 nuclei: 55Mn,500–502

59

Co,500,50263,65Cu,504–51193Nb,500,51295,97Mo,501,51399Ru,514105Pd,515115In.517Within the limits

of the high-field approximation, i.e., L> Q, the observed 1J values are readily interpreted.Occasionally, first-order perturbation theory or even a full treatment of the Hamiltonians isnecessary.500,504,517–520It is worth emphasizing that, given the utility of1J(M,L), such information

is otherwise not available

Twonice examples are shown in (32) and (33) The cobaloxime complexes of the type (32) areespecially informative,503as they reveal the expected change in 1J(59Co,31P) as a function of thetransinfluence of the ligand X.81,82

70

80

90

δ(31P)70

8090

δ(31P)

Figure 11 Contour plot showing the isotropic part of the 162.0 MHz 2D31P CP/MAS exchange spectrum( ... Berger, S Angew Chem Int Ed 20 02, 41, 107–109.

325 Keresztes, I.; Williard, P G J Am Chem Soc 20 00, 122 , 1 022 8–1 022 9.

326 Chen, Y.; Valentini, M.;... Brookhart, M Organometallics 20 01, 20 , 526 6– 527 6.

25 2 Shultz, L H.; Tempel, D J.; Brookhart, M J Am Chem Soc 20 01, 123 , 11539–11555.

25 3 Hii, K K.; Claridge,... Chem Soc 20 01, 123 , 11 020 –11 028 .

26 7 Macchioni, A.; Zuccaccia, C.; Clot, E.; Gruet, K.; Crabtree, R H Organometallics 20 01, 20 , 23 67? ?23 73.

26 8 Zuccaccia,