AcademicPressisanimprintofElsevier TheBoulevard,LangfordLane,Kidlington,Oxford,OX51GB,UK 32JamestownRoad,LondonNW17BY,UK Radarweg29,POBox211,1000AEAmsterdam,TheNetherlands 225WymanStreet,Waltham,MA02451,USA 525BStreet,Suite1900,SanDiego,CA92101-4495,USA Firstedition2012 Copyright©2012ElsevierLtd.Allrightsreserved Nopartofthispublicationmaybereproduced,storedinaretrievalsystemortransmittedin anyformorbyanymeanselectronic,mechanical,photocopying,recordingorotherwise withoutthepriorwrittenpermissionofthepublisher PermissionsmaybesoughtdirectlyfromElsevier’sScience&TechnologyRights DepartmentinOxford,UK:phone(+44)(0)1865843830;fax(+44)(0)1865853333; email:permissions@elsevier.com.Alternativelyyoucansubmityourrequestonlineby visitingtheElsevierwebsiteathttp://elsevier.com/locate/permissions,andselecting ObtainingpermissiontouseElseviermaterial Notice Noresponsibilityisassumedbythepublisherforanyinjuryand/ordamagetopersonsor propertyasamatterofproductsliability,negligenceorotherwise,orfromanyuseor operationofanymethods,products,instructionsorideascontainedinthematerialherein. Becauseofrapidadvancesinthemedicalsciences,inparticular,independentverificationof diagnosesanddrugdosagesshouldbemade BritishLibraryCataloguinginPublicationData AcataloguerecordforthisbookisavailablefromtheBritishLibrary LibraryofCongressCataloging-in-PublicationData AcatalogrecordforthisbookisavailablefromtheLibraryofCongress ISBN:978-0-12-397020-6 ISSN:0066-4103 ForinformationonallAcademicPresspublications visitourwebsiteatstore.elsevier.com PrintedandboundinGreatBritain 12 13 14 11 10 9 8 7 6 5 4 3 2 1 CONTRIBUTORS RuudL.E.G.Aspers RadboudUniversity,InstituteforMoleculesandMaterials,BiophysicalChemistry, Heyendaalseweg,Nijmegen,TheNetherlands BruceJ.Balcom MRIResearchCentre,DepartmentofPhysics,andDepartmentofChemistry,Universityof NewBrunswick,P.O.Box4400,Fredericton,NewBrunswick,Canada TeresaBlasco InstitutodeTecnolog´ıaQu´ımica(UPV-CSIC),UniversidadPolite´cnicade Valencia-ConsejoSuperiordeInvestigacionesCient´ıficas,Avda.delosNaranjos,s/n, Valencia,Spain AngelC.deDios DepartmentofChemistry,GeorgetownUniversity,Washington,DC,USA MartinJaeger DSMNutritionalProductsLtd.,AnalyticalResearchCenter,Wurmisweg,Kaiseraugst, Switzerland CynthiaJ.Jameson DepartmentofChemistry,UniversityofIllinoisatChicago,Chicago,Illinois,USA Lu´ısMafra DepartmentofChemistry,CICECO,UniversityofAveiro,Aveiro,Portugal,and DepartamentosdeQu´ımicaF´ısicayAnal´ıticayQu´ımicaOrga´nicaeInorga´nica,Universidad deOviedo,Oviedo,Spain ColleenE.Muir MRIResearchCentre,DepartmentofPhysics,andDepartmentofChemistry,Universityof NewBrunswick,P.O.Box4400,Fredericton,NewBrunswick,Canada Jose´ AlejandroVidal-Moya InstitutodeTecnolog´ıaQu´ımica(UPV-CSIC),UniversidadPolite´cnicade Valencia-ConsejoSuperiordeInvestigacionesCient´ıficas,Avda.delosNaranjos,s/n, Valencia,Spain vii PREFACE As is usual in this series of reports, Volume 77 of Annual Reports on NMR Spectroscopyconsistsofaccountsofrecentprogressinseveralareasofmolec- ularscience.Thevolumecommenceswithadiscussionof‘RecentAdvances inNuclearShieldingCalculations’byA.C.deDiosandC.J.Jameson;thisis followedbyareviewof‘PurePhaseEncodeMagneticResonanceImaging ofFluidsin PorousMedia’by C.E. MuirandB. J.Balcom;M.Jaegerand R. L. E. G. Aspers report on ‘Steroids and NMR’; the final chapter, by L. Mafra, J. A. Vidal-Moya and T.Blasco, covers ‘StructuralCharacteriza- tionofZeolitesbyAdvancedMethodsofSolidStateNMRSpectroscopy’. It is a great pleasure for me to thank all of the contributors for their timely and significant contributions. G. A. WEBB Royal Society of Chemistry Burlington House Piccadilly London UK ix CHAPTER ONE Recent Advances in Nuclear Shielding Calculations † Angel C. de Dios*, Cynthia J. Jameson *DepartmentofChemistry,GeorgetownUniversity,Washington,DC,USA †DepartmentofChemistry,UniversityofIllinoisatChicago,Chicago,Illinois,USA Contents 1. OverviewofThisReview 4 2. AdvancesinMethodsofCalculation 5 2.1 Relativisticcalculations 5 2.2 Densityfunctionalcalculations 11 3. LocalEffectsonShielding:SingleMolecules,Clusters,andFragments 15 3.1 Conformationaleffects 16 3.2 Neighbouringnon-bondedatomeffects 16 3.3 Hydrogen-bondingeffects 18 3.4 Electrostaticfieldeffects 19 3.5 Intermolecularrelativisticeffects 22 3.6 Shieldingandchirality 23 4. ShieldinginExtendedNetworks 25 4.1 Approachestoextendednetworks 25 4.2 Crystallinematerials 32 4.3 Non-crystallinematerials,glasses 34 4.4 NICSinperiodicsystems 36 4.5 Relativisticcalculationsinsolids 36 4.6 Thecaseforretainingclusterapproachesinourtoolbox 37 5. DynamicAveragingofShielding 42 5.1 Whyisaveragingsoimportantfornuclearshieldingcalculations? 42 5.2 Rovibrationalaveraging 43 5.3 Dynamicaveragingincondensedphases 45 6. ExtractingInformationfromNMRChemicalShiftswiththeHelpofTheoretical Calculations 54 6.1 ShieldingtensorsastoolsforNMRcrystallography 55 6.2 Detailsoflocalstructure 57 6.3 Shieldingasaprobeforintermolecularinteractions 59 6.4 Characterizationofsolids 60 References 61 AnnualReportsonNMRSpectroscopy,Volume77 #2012ElsevierLtd. 1 ISSN0066-4103 Allrightsreserved. http://dx.doi.org/10.1016/B978-0-12-397020-6.00001-5 2 AngelC.deDiosandCynthiaJ.Jameson Abstract Nuclearmagneticshieldingcalculationshavereachedagreatdealofsophistication,as thesenowincorporatebothrelativisticandcorrelationeffects.Approachesnowinclude moleculardynamicsaswellaseffectsofthemediumincondensedphases.Withthese computationaltools,calculatedshieldingvaluesarenowobtainedunderconditionsas closeaspossibletothoseofasampleinsideanuclearmagneticresonancespectrom- eter. Indeed, computations are approaching the limits of experimental uncertainty. A brief description of new methodologies of shielding calculations is presented followedbyareviewofthevariousfactorsthatmayinfluenceshielding.Theusefulness ofbeingabletoreproduceexperimentaldataishighlightedbycitinghowshielding calculations in many instances have enabled avenues for extracting information on varioussystems. Key Words: Absolute shielding, Chemical shifts, Chirality, Cluster models, Conforma- tionaldependence,Densityfunctionaltheory,Disorderincrystals,Electroncorrelation, Four-component calculations, Gas phase NMR, Hydrogen bonds, Intra- and inter- moleculareffects,Isotopeshifts,Mediumeffects,Moleculardynamics,NMRcrystallog- raphy,Non-bondedinteractions,Periodicboundarycalculations,Polarizablecontinuum model,Polymorphs,Pseudopotentials,Relativisticcalculations,Rovibrationalaveraging, Shieldingsurfaces,Structuredetermination,Temperaturedependence,Tensor,Torsion angles,Two-componentcalculations ABBREVIATIONS AE all-electron BO Born–Oppenheimer BPPT BreitPauliperturbationtheory CCSD coupledclustersinglesanddoubles CFP chargefieldperturbation CI configurationinteraction COSMO conductor-likescreeningmodel CPMD Car-Parinellomoleculardynamics CSGT continuoussetofgaugetransformation DFT densityfunctionaltheory DHF Dirac–Hartree–Fock ECP effectivecorepotential EIM embeddedionmethod EPR electronparamagneticresonance GAPW Gaussianandaugmentedplane-wavemethod GGA generalizedgradientapproximation GIAO gauge-includingatomicorbitals GIPAW gauge-includingprojectoraugmentedwave HAHA heavyatomheavyatom HALA heavyatomlightatom HF Hartree–Fock HOMO highestoccupiedmolecularorbital IGAIM individualgaugeforatomsinmolecules RecentAdvancesinNuclearShieldingCalculations 3 KS Kohn–Sham LAPW linear-augmentedplanewaves LDA localdensityapproximation LHF-CEDA localized Hartree-Fock-common energy denominator Green’s function approximation LLH localizedlocalhybrid LMF localmixingfunction LR-ESC linearresponseeliminationofsmallcomponent LUMO lowestoccupiedmolecularorbital MAS magicanglespinning MB-GIAO magneticallybalancedgauge-includingatomicorbitals MC MonteCarlo MD moleculardynamics MM molecularmechanics NAO numericatomicorbital NBO naturalbondorbital NICS nucleusindependentchemicalshift NMR nuclearmagneticresonance OEP optimizedeffectivepotential OOD occupied-orbitaldependent PAW projectoraugmentedwave PBC periodicboundaryconditions PCM polarizablecontinuummodel PES potentialenergysurface PW planewave QM quantummechanics RMB restrictedmagneticbalance RPA randomphaseapproximation RSC relativisticsmallcore SCF self-consistentfield SCREEP surfacechargerepresentationoftheelectrostaticembeddingpotential SO spin–orbit SPARTA shiftpredictionfromanalogyinresiduetypeandtorsionangle SPC simplepointcharge SSNMR solid-statenuclearmagneticresonance STO Slater-typeorbitals SWNT single-wallednanotube USPP ultrasoftpseudopotential XC exchangecorrelation XRD X-raydiffraction ZORA zeroth-orderregularapproximation 4 AngelC.deDiosandCynthiaJ.Jameson 1. OVERVIEW OF THIS REVIEW Monographsonshieldingcalculationshavetracedtheadvancesinthis field over the years: Nuclear Magnetic Shieldings and Molecular Structure (1993),1 Modeling NMR Chemical Shifts: Gaining Insights into Structure andEnvironment(1999),2CalculationofNMRandEPR(electronparamag- netic resonance) Parameters: Theory and Applications (2004).3 Reviews on shielding calculations have been published in various venues.4–8 In addition, a review of ab initio shielding calculations as an emerging tool for protein structure determination9 and of the PAW/GIPAW approach for the study of solids10,11 have been published. The physical and theoretical aspects of nuclear shielding are reported on an annual basis in Specialist Periodical Reports on NMR.12 Earlyreviewsofshieldingcalculationsincludethetworeviewsinthispar- ticularseriesofvolumes(vol.29,1994).13,14Excellentrecentreviewsinthis series have detailed the theoretical basis for the relativistic and density functional theory (DFT) methods for nuclear shielding computations.15,16 We do not repeat a development of these theoretical foundations here. Rather we update with recent developments in exchange-correlation functionals (since the Wilson review in vol. 49, 2003) and in relativistic computations (since the Autschbach review in vol.67, 2009). To minimize duplication, we have limited this review primarily to papers published in 2004 and later. Acrucialroleoftheoreticalcalculationsisforunderstandingphysicalsys- tems,thuswestayconnectedtotheconceptsandmotivationsofexperimen- talistsbutusealimitednumberofexampleswhichillustratethequalityand reliabilityoftheinformationthatcalculationsprovide.Forexample,inpart 2ofthisreview,wediscusshowimportantarerelativisticeffectsonshielding andwhichpartsofthecorrectionsarelargestforwhatsystems.Weconsider herethecurrentthinkingaboutwhichoftheexchangeandcorrelationfunc- tionalsworksbestforwhichtypesofsystems.Inreviewingtheoreticalcalcu- lationsinparts3–6,wehavegroupedexamples,suchastoserveasaguideto approachesthathaveworkedwellforvarioustypesofsystems.Thus,inpart3, weconsidercaseswheresinglemoleculecalculationsmaybeused,andcases where model fragments are the best systems for understanding the depen- dence on conformation separately from other factors, and cases where using supramolecular clusters including interacting neighbours is the most logicalapproach,especiallywhenhydrogenbondingisinvolved.Wereview RecentAdvancesinNuclearShieldingCalculations 5 various approachesinincluding subtlequantumeffectssuchas electrostatic fieldeffectsfromdistantneighboursthatinfluencetheelectrondistribution ofthelocalneighbouratomsandthereforeaffecttheshieldingofthenucleus inquestion.Localeffectsareparamountinunderstandingshieldinginama- jorityofcases;thus,approachesthatfocusonlocaleffects,usingclustersand fragmentsandembeddingtheminachargefieldmayoftenbethemostlogical theoretical approach. On the other hand, some aspects of shielding in ex- tendednetworkscannotbeproperlytreatedbyusinglocalapproachessuch asclusterandfragmentmodels,andwehavetoconsiderthelong-rangeeffects oftheentirenetworkbydoingcalculationsinperiodicsystems;thisisespe- ciallytrueforcovalentsolids.Thus,inpart4,wediscusscalculationswhich use methods which have been developed for these systems, theoretical methods that have become an important tool for experimentalists making measurementsinthesolidstate.Inpart5,weconsiderdynamiceffects.For some systems, dynamic averaging is such a significant part of the observed chemicalshiftthatcalculationsusingastaticgeometry(atomiccoordinates) donotprovidethewholeshieldingstoryandmayevenleadtoincorrectas- signmentsofchemicalstructure.Weconsiderbothrovibrationalaveraging, whichleadstoisotopeshiftsandtemperaturedependenceofshieldingevenin thegasphase,anddynamicaveragingincondensedsystems,notablyinliquids. Therearedifferentapproacheswhichhavebeenused,andweconsiderwhich oftheapproachesismostlogicalforthetypeofsystemathand.Intheearlier parts,wediscusshowthephysicalaspectsofthesystem(bondlength,bond angles, torsion angles, presence of neighbours, dynamics of bonded and non-bonded atoms, relativistic neighbours bonded and otherwise) have to be introduced into the model so as to carry out a proper calculation that can be compared with experimental measurements. In part 6, we discuss how to turn the information arrow the other way around: in favourable circumstances, what can the observed differential shieldings tell us about the nature of the physicalsystem? Here, we provide some examples where theoretical calculations have been used to verify hypothetical information aboutthephysicalsystemunderstudy. 2. ADVANCES IN METHODS OF CALCULATION 2.1. Relativistic calculations In the calculations of nuclear magnetic shielding, the vector potential pro- ducedbythenuclearmagneticmomentisverylocalandthereforeaffordsa weightedsamplingoftheelectronicwavefunctionintheclosevicinityofthe 6 AngelC.deDiosandCynthiaJ.Jameson nucleus. Thus, the nuclear magnetic shielding in a molecular system is in- trinsically an all-electron relativistic property. Methods and application of relativistic approaches to the calculation of nuclear shielding have been reviewedbyAutschbachinthisseriesofvolumesin2009.Therefore,inthis review, we consider developments from 2009 on. Conceptual and computational difficulties have conspired to delay the availability of fully four-component all-electron treatments at correlated levelsusingdistributedgaugeorigins.However,thesituationislargelyclear- ingup.Adifficultyparticulartorelativisticcalculationsofmagneticproper- tiesoriginatesfromthefactthattheinclusionofavectorpotentialaffectsthe balancebetweenthelargeandthesmallcomponentsofthefour-component spinors.Thismagneticbalancemustbetakenintoaccount,explicitlyorim- plicitly, in order to obtain accurate results for magnetic properties. An im- portant differencebetween relativistic and nonrelativistic theory is the lack of an explicit diamagnetic term in the Hamiltonian in relativistic theory. However,thediamagnetictermentersthroughthecontributionofnegative energystates.Thenegativeenergystatesareimportantfornuclearshielding becausemagneticfieldsareintroducedthroughoperatorswhichcouplethe large and small components, so the negative energy states cannot be neglected in relativistic treatment of magnetic properties, but it has been suggestedthattheymaybetreatedinasimplifiedmanner.17Extremelycom- pact negative energy states require additional steep s and p functions in the basissetwhenhighlyaccurateabsoluteNMRshieldingsaresought.Effects of highly compact negative energy states essentially cancel out for relative shieldings(chemicalshifts).18Kutzelniggproposedafield-dependentunitary transformation of the four-component relativistic Hamiltonian in order to introduceanexplicitdiamagneticterm.19Whenmagneticbalancebetween the small and large components is explicitly built in, the otherwise missing diamagnetism arises naturally20 and the heavy demand on the basis sets of high angular momenta is also greatly alleviated. Several groups considered a unified approach to four-component rela- tivistic treatment of nuclear shielding nearly simultaneously. Kutzelnigg and Liu presented the formulation of a logical and systematic classification of existing methods of calculations of NMR parameters within relativistic quantumchemistry,togetherwithnewvariantsandpresentednewmethods aswell.21Variousmethodshavebeenreportedseparatelyinthisseriesover several years; the Kutzelnigg and Liu analysis puts all systematically in the propercontext.Theyconsidertransformationsatoperatorlevelversusma- trixlevel,thepossibleformulationsoftheDiracequationinamagneticfield, RecentAdvancesinNuclearShieldingCalculations 7 traditional relativistic theory, field-dependent unitary transformation, bispinordecomposition,equivalenceofthemethodsatoperatorlevel.They then consider relativistic theory in a matrix representation, expansion in unperturbed eigenstates, expansion in a kinetically balanced basis, and ex- pansioninanextendedbalancedbasis.Theauthorsexploredecomposition of thelowercomponent,decompositionofthefull bispinor, unitarytrans- formationatthematrixlevel.Theypaycarefulattentiontosingularities.First theydiscussmethodswhichareexactinthesensethattheiraccuracyisonly dependentonthequalityofthechosenbasis.Inthelimitofacompletebasis, all these methods yield the same results, but the rate of convergence to the limitcanbedifferent.21Amongthesemethods,theyconsidertheonesbest suitedforeachofthemagneticproperties.Forthecaseofnuclearmagnetic shieldingwhereonevectorpotentialisduetoanexternalfieldandoneisdue to a nuclear magnetic dipole, they suggest to discard the method based on theuntransformedDiracoperatorbecauseitdoesnotgivethecorrectnon- relativisticlimit.Theyalsosuggesttodiscardthemethodbasedonaunitary transformationofthefullmagneticfieldbecauseitisplaguedbysingularities. Theyofferthegoodchoiceasthemethodbasedonaunitarytransformation oftheexternalfieldonlyandformalismsequivalenttoitbecausetheseleadto the correct nonrelativistic limit and are not plagued by singularities. They alsorecommendasafurtherpossibilitythemethodcalledFFUTm(full-field unitary transformation “at matrix level”) by Xiao et al.,22 where one starts formallyfromaunitarytransformationatmatrixlevelbutevaluatesthedia- magnetic term exactly. They suggest that the restricted magnetic balance (RMB)17shouldbeused.Theperformanceofthevariousmethodswithre- spect to the basis set requirement has recently been investigated by Cheng etal.18Theresultsdifferverylittle,evenforasmallbasis.Finally,Kutzelnigg and Liu21 consider various approximations previously proposed which do not give the exact results in the limit of a complete basis, for example, see Xiao et al.22 All of these approximations are based on methods which give the correct nonrelativistic limit and use a pseudo sum-over-states formula- tion with the restriction of the intermediate eigenstates to those with pos- itive energy. Along the way, various commonly used approximations such as the Douglass-Kroll-Hess approximation23 and the zeroth-order regular approximation (ZORA)24 are discussed in context. Chengetal.showedthatthesevariantsofapproachesusingmagneticbal- ance can all be recast into one unified form.18 The various schemes previ- ouslyproposedforincorporatingthemagneticbalancedependencearethen showntobeequivalenttothisnewapproachandthereforecanbecombined