8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

1/7

256 Inorganic Chemistry, Vole14, No. 2, 1975 Wagner, Bates, and PoindexterP d ~ @ l 4 ( N c y ) z , 5 3 1 1 1 - 3 3 - 4 ; P d C h ( P P h 3 ) ( W c y ) , 5 3 1 1 1 - 3 4 - 5 ;P d C l z ( P P h z M e ) ( N c y ) , 5 3 1 1 1 - 3 5 - 6 ; P d C l z ( P P h M e a ) ( N c y ) ,531 11-36-7; Scy , 5 362-78-7; N cy, 3189-56-8; bis(dirnethylpheny1-phosphine)dichloro-k,k -dichloropalladium(II), 15699-80-6; bis-(triphenylphosphine)dichloro-~,~'-dichloropalladiurn(II),5 134-30-2;b i s ( me t hy l d i pheny l phosph i ne ) d i cA l or o~ ~ ,~ '~ d i ch l o r opa l ~ ad i u r n~ I I ) ~29884-90-9; bis(benzonitrile)dichloropalladium(II), 14220-64-5 .References and Notes

1 ) D. K. Mitchell, W . D. Korte, and W. C. Kaska, Chem. Commun., 1384(1970).(2) J. C . Kotz and D. G. Pedrotty, J . Organometal. Chem., 22,425 (1970);D. Cashman and F. J Lalor, ibid., 32, 351 (1971).(3) S. 2. Goldberg, E. N. Duersler, and K. N . Raymond, Chem. Commun. ,826 (1971).(4) K. A . 0 tarzewski, H. T. Dieck, K. D. Franz, and F. Hohmann, J .Organometal. Chem., 42, C35 (1972).(5) A. Greco, J . Organometal. Ch em., 43, 351 (1972).(6) H. Alper and R. A . Partis, J . Organometal. Chem., 44, 371 (1972).(7) D. K. Mitchell and W. C. Kaska, J . Organometal. Che m., 49, C73(1973).(8) H. Berke and E. Lindner, Angew. Chem., 85 668 (1973).(9) R. H. rnup and M. C. Baird, Inorg. N u c l . Chem. Lett., 5, 65 (1969).(10) R. J. Sundberg, R. . Shepherd, and H. Taube, J . Amer. Chem. Soc.,94, 6558 (1972).(1 1) H. Schmidbauer, J. Adlkofer, and W . Buchner, Angew. Chem., 85,448(1973); H. Schmidbauer and R. Franke, ibid., 89, 49 (1973); H. H.Farsch and H. Schmidbauer, ibid., 85, 910 (1973).(12) W. C. Kaska, D. K. Mitchell, and R . F. Reichelderfer, J . Organometal.Chem. , 47, 391 (1973).(13) H. Koezuka, G. Matsubayashi, and T. Tanaka, Inorg. Chem. , 13, 443(1974).

(14) W. J Middleton, E. L. Buhle, J G . McNally, Jr. , and M . %anger,J.Org. Chem., 30, 2384 (1965).(15) A . Rieche and P. Dietrich, Chem. Ber., 96, 044 (1963).(16) W . Gerrard, M . F. Lappert, H . Pyszora, and J. $41. Wallis, J . Chem.Soc., 2182 (1960).(17) M . W. Duckworth, G .W A . Fowles, and R. A Hodless, J. Chem. Soc.,5665 (1963).(18) S . C. Jain and K. ivest, Can. J . Chem. , 41, 2130 (1963).I 9 ) Not all increases in v(CN) epresent coordination in nitriles. For example,in the nitrile-coordinatedpentaammineruthenium(l1) omplexes of severalsubstituted benzonitriles, acetonitrile, and cyanopyridines, v C N ) of acoordinated nitrile shows a decrease from the free-ligand value: R. E.Clarke and P. C. Ford, Inorg. Cihem., 9: 727 (1970); R. E. Clarke andP. C. Ford, ibid., 9, 495 (1970).(20) A . F. Cook and J. G. Moffatt, J .Amer. Chem. Soc., 90, 740 (1968).(21) A. T. Christensen and W . G. Witmore, Acta Crystallogr.,Sect. B , 25,73 (1969).(22) C. Bugg, R. Desiderato, and R. L. Sass, J . Amer. Chem. Soc., 86, 3157(1964).(23) C. Biondi, M . Bonamico, L. Torelli, and A. Vaciago, Chem. Commun.,191 (1965).2 0 ) H. Goetz, B. Klabuhn, F. Marschner, 1%Hohberg, and W. Skuballa,Telrahedron, 27, 999 (1971).(25) J . H. Enemark and R . H. Holm, Inorg. Chem., 3, 1516 (1964).(26) W. El. Baddley, Inorg. Chim. Acta , R e v . ,2, 7 (1968).(27) A = 9.1 X 10-1 ohm-* cm2mol-l for 3.5 X 1 0 4 Msolution in acetonitrileat 250.(28) M . J. Middleton and 1. A. Engelhardt, J. Amer. Chem. Sac., 2788(1958).(29) H. Kohler, W. Eichler, and R . Salewski, Z . Anorg. Allg, Chem. , 379,(30) R, El King and M . S. Saran, J . Amer. Chem. Soc., 95. 1811 (19'73).(31) B. F. Johnson and R . A Ealton, Spectrachim . Acra, 22, 1853 (1966).(32) C. Bugg and K. L. Sass, Acta Crystallogr., 8, 591 (1965).

183 (1970).

Cont r ibution f rom th e LJ S. Army Electronics Technology and Devices Laboratory (EC OM ) ,For t Monmouth , New Jersey 07703

ynarnic Nuclearesidinag ow the

BURKWARD E. W A G N E R . * R I C H A R D D. BATES, J r . , a n d E D W A R D H. POINDEXTERReceived Febmary 13, 1974 A I C 40099X

Th e solvation of inert, low-spin Cr(1) com plexes by nmr-ac tive species in t.he second coordination sphere has been investigatedby dynamic nuclear polar i za tion (dnp) , an nm-e s r double- resonance t echnique. A t 75-G magnet ic f ield, dipolar( through-space) and scalar (Fermi-contac t) interaction between a nuclear spin and a n electron spin on dif ferent moleculescan be dis t inguished. Only dipolar coupl ing was observed for [Cr( CN )sNQ ]3- and the nuclear tes t probes octaf luoro-naphthalene (OFN),r imethyl phosphi te, L if , and BF4-. The re is little unpa ired spin density at the periphery of the nitrosylcomplex. Th e order of observed solvat ion interact ion is BF4- < E l 2 0 < Li+, demon strat ing e lectros tat ic effects . Wi thCr(bipy)3+, moderate scalar coupl ing is observed for OFN and the phosphi te, indicat ing the presence of unpaired spindensi ty in the plane of the bipyr idyl r ing, at the r im of the complex. W ith Cr(C6H6)2+ serving as radical probe, s t rongscalar coupl ing by t ransient bond ing interact ion is observed for t r imethyl phosphi te and for BFn--, Th e absence of scalarcoupling with OF N suggests a lack of unpaire d spin density above the planes of the n-benz ene rings. Th e data a re consistentwith penetrat ion of smal l l igands into the space between the ben zene r ings and t ransient bond format ion directly with themetal. Th e general applicability of dnp to the study of second coordination sph ere molecular interactions having correlationt imes shor ter than 10-8 sec is discussed.

IntroductionA metal complex in solution interacts with surroundingsolvent molecules, counterions, or other dissolved species. Th eimportance of this second coordination sphere (SCS) inligand-exchange reactions, and to catalysis in general, isbeginning to be realized.* N m r is one of the few techniquesthat can yield information at the molecular level about in-teractions occurring at the outer fringes of a complex.2+3Dipolar broadening of nmr lines sometim es allows the detectionof preferential orientation of rnoiecules near th e surfa ce of aparamagnetic complex.4~5Distances of closest approach andsolvation geometries can sometimes be deduced from dipolar

(pseudocontact) chemical shifts of solvent molecules positionedin the SCS of transition metal complexes exhibiting strongg-tensor anisotropy.2.6Th e present work examines the applicability of dynamicnuclear polarization (dnp), a double-resonance nmr-esrtechnique, as a tool for studying S@S nteractions of par-amagnetic ions in solution. D np measure s the weak couplingbetween nuclear spins on One molecule and an electron spinon ano ther molecule during encoanters of the two species insolution. As the technique is only responsive to fast molecularinteractions (correlation times shisrler than 10-8 sec), it isespecia ly suited for iravcsfigations o f diffusion-controlled

8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

2/7

Second Coordination Sphere in Inorganic ComplexesS I

\\\ / / I

Figure 1. Com bined energy states and relaxation probabilities for asystem of unpaired electrons coupled weakly t o spin I = 12 nuclei.processes in the SCS. It is an ancillary benefit of low-fielddnp th at dipolar and scalar contributions to the intermolecu larspin-spin coupling can be separated in a clear and unam-biguous fashion for all nuclei studied so far.7-13We wish to report dnp results at 75-G magnetic field fora series of low-spin Cr(1) complexes of differing geometriesand with differing degrees of electron spin delocalization fromthe me tal into the ligand framewo rk. These properties werefound to be important parameters for previous studies of fastmolecular collision processes in solutions of organic freeradicals.7aTheoryDnp is a double-reson ance echnique that measures the weakcoupling of a receptor nuclear spin on one molecule with anunpaired electron spin on another molecule during encountersof the two species in solution.14 Because of diffusion andtumbling of the two molecules, the nucleus and electron ar ein motion with respect to each other. This motion, essentiallyrandom in nature, modulates the electromagnetic couplingbetween the electron and the nucleus. Resultant m odulationwill, in fluids, usually contain frequency components whichma tch the energy differences between levels of the coupled spinsystem (Figure 1) and thereby will allow transitions to beinduc ed between th e four levels of the coupled system. Bysaturation of the purely electronic transitions p (esr transitions)by m eans of radiofrequency power, the nuclear spin populationdistribution across the purely nuclear spin relaxation pathwayq (nmr transitions) will be altered, and the intensity of theobserved nm r signal will change accordingly.Total signal enhancement of nucleus I at a given radio-frequency saturation power P is given byGdP>= [A ?)-A(O)I/A(O) = uc-qfSe p) 1)where A P ) and A 0)are the observed nmr signal intensitieswith and without saturating the esr lines, respectively. Se P)is the saturation function for the esr line,fis the spin leakagefactor, and Um s the ultimate enhancem ent. Th e last is thenmr signal enhancement which would be attained at completeesr saturation, S @ ) = 1, and c omplete domination of nuclearspin relaxation by its interaction with the elec tron,f= 1. Theseterms a re furth er defined in eq 2-5; a m ore com plete discussionf = - T J T l b ) (2)Se Q =S /@ + 1) (3)

(4)U , I = ye /Yr ) r- + c)(2q -t. r + s +c - 5)can be found elsewhere.7 Ti and Tib are the nuclea r relaxationtimes in the presence and in the absence of a radical species,respectively. It should be noted that high leakage factorsfor a nucleus I indicate strong interaction with and nuclearspin relaxation dominated by the radical species. Th e spinrelaxation rates in (5) are depicted in Figure 1, with r - s)being the net dipolar and c the scalar components of the totalcoupling. All other term s have their usual meanings.

Inorganic Chemistry, Vol. 14, No. 2, 1975 257At a magnetic field of 75 G, all relaxation pathways inFigure 1 are fully driven by diffusion-controlled molecularmotions in the solution, with 2q:r:sstill very close to the z erofield limit of 6:2:12. For pure dipolar coupling (c = O), eq5 requires negative signal enh ance ment with Um = -1/2 ye/y1),while for the scalar limit (c>> r , s) enhancem ents are positive,with Um= + re/rI). The difference in sign for the scalar an ddipolar coupling components and the knowledge of theirlimiting values allow a clear sep aration of th e two effects. The

experimentally obtained dnp effect is proportional to thedifference between the dipolar and scalar relaxation com-ponents. For a scalar component smaller than the dipolar term,negative enhancement of the nmr line is observed, while alarger scalar component gives rise to net positive enhancement.When scalar and dipolar terms are of approximately equalmagnitude, the obtained U-I values for a series of closelyrelated solutions may be either sm all and positive or small andnegative, with no intrinsic significance being attache d to thesign of the enhance men t itself. For proton s in solutions oforganic free radicals, U ~ Hs very close to -330, the dipolarlimit;7 for 7Li, on the other hand (for which the exact for-mu lation of Urns similar to the = 1 / 2 case), almost the totalrange in U-Li from the dipolar limit of -846.7 to the scalarlimit of +1693.4 is known.llJ5The respective dynamic strengths of dipolar and scalarcompo nents depend on their static energies and on interactionor correlation times. Th e dipolar term is almost always asimple functio n of molecular collision radii an d diffusion timesand varies in a relatively simple fashion. It depe nds merelyon throug h-spac e spin-spin in teractio n with a d-6 distancedependence. The furthest distance at which radical-nuclearspin relaxation can compete with nuclear spin-nuclear spinrelaxation lies near 10-12 A or at the inner limit of the thirdcoordination sphere. Dipolar dnp enhancemen ts fall off withincreasing fields as well-defined functions of the dipolarcorrelation time, which can then be deduced.7>16The scalar component, however, is transmitted by electronorbital interactions whose strength depends very much on the

specific atomic species involved in the collision, unlike thedipolar case which is purely a function of the cen ter-to-centerdistance between the electron and the nucleus.17 Sca larcoupling may occur either by exchange polarization or by direc tspin delocalization and orbital overlap during the formationof a transient bond between the radical and receptor molecules.Any exact formulation for the scalar term c depends on thechoice of model for the interaction,lg but for all models at lowmagnetic fields, c is proportional to Aiso2Ts, where Aiso =8a/3)g?P~e~el91 0)12,ith Aiso being the isotropic hyperfinecoupling constant, TS he scalar correlation time, and I9(0)12the induced unpaired electron density at nucleus I.Calculation of the peak scalar coupling energy duringmolecular collisions is complicated and a t best semiquantitative.As an example of calculated values, the induced electrondensity of aromatic fluorines in (r collisions with p-benzo-semiquinone at 2.5-A separation is 0.001 electron; for acollision at 2.5 A, 0.0001 e l e c t r ~ n . ~ ~ ahese sp in densities onthe fluo roca rbon would yield scalar hy perfin e energies of ab out50 and 5 M Hz , respectively. The observed average value forhexafluorobenzene and chlorinated semiquinone is about 5M H z for the scalar component and about 1.5 M Hz for thedipolar hyp erfine energy.1gb It is thus seen tha t even a verysmall induced spin density can rival or overwhelm the dipolarrelaxation mechanism, depending on comparative correlationtimes.Scalar coupling arising from formation of a transientchemica l bond requires distances of closest approac h betweenthe radical source and the receptor nucleus or rece tor molecule

within the respective van der W aals radii, 2-4 Chemical

8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

3/7

58 Inorganic Chemistry, Vol. 14, No. 2, 1975bond formation can have two effects: overiap of orbitals willtransmit a stronger scalar colapiing pulse than is the case insimple nonbonding collisions (increase in Aiso), or alonger-lived attachment will occur which enables the scalarcoupiiiig term to be inore effective through its augmentedinteraction time (increase in 7s). In practice, the strongerelectronic interactions which reflect bonding tendenciesnormally produce a longer-lived collision, so that dynamicscalar coupling is doubly augmented. This dual effect alsoseduces the need for multifkld dnp studies in many situationsof high chemical interest.19

Sca lar coupling by exc hange polarization occurs throughslight unpairing of dectrons around the nm r-active nucleusby the unpaired electron on the radical during an intermo-lecular collision or errcounter. While the exact range overwhich intermolecular exchange coupling i s felt has not beenconclusively established, rcsults from studies with trivalentphospl:orus in solutions of TTBP v ide infra suggest a limitsomewhat above 7 A 20 It s c m s high1.y unlikely that exchangepolarization between a radical and a nuclear probe should bemediated through an intervening molecule, so that this re-laxation process should also require smaller collision radii andbe restricted to SCS interactions.7achemi::al bonding. Po:. in lions of a gi~ien uclear testprobe, an increase in Aisoarp n change of radical would indicategreater unpaired electron density induced by the sw ond radicala t the I nucleus, while an increase in ;rs would indicate atendcncy for the two species to stick together m ore tightly.Conversely, a decrea se in Ais0 or n, as evidenced by diminutiono f scalar coupling compared to a suitable model radical, in-dicates either diminution of in~errnolecu lar nteraction, un-availabi1it.yo f unpaired electron spin at the periphery of theradical, or both.

Dnp cffeccively nonit ors fast, tra nsient bonding betweena radical an d other molecular species in the solution, a processwhich s not easily dctect.ed by other techniques. In order toobserve any dnp in solution, correlation times for nuclearspin-electron spin interactions must be faste r than 10-8 see,while line broadening of esr spectra by the nuclear spin requirescorrelation times slower than about 10-8 sec. D np studies onparam agnetic complexes are limited to the compounds whoseesr line can be at least partiaily saturated with feasible ra-iofreqtnency pow er eveRs ( 7 1 ~ 7 3 ~bove 10-15 secz). As anadditional lirnitatior,, chemical shifts at 7 5 G are small, so thatonly one unresolved nmr line is obtained for each nuclearspecies. H,ikewlse, scalar shifts of the 9F rimr line for radicalconcentrations of 0.01M would be about 0.1 ppm and thusunobs~~d.ble;19bshift of 100 ppm is the minimum detectablea t 75 G owever, the sensitivity of low-field dnp eo scalarcoupiings vrhich elude detection by nmr shift is magnified bythe much more favorable dependenceon conelation or stickingtime, w hich amplifies the scalar component c of the nuclearrelaxation. In contrast, the variation in collision time whichgreatly modifies the dnp signal affects scalar shifts hardly atail. The scalar shift is governed by the overall numericalpartitioning between bound and free molecules, and the exactduration of the collision or complexation has only a very smalleffect in the rapid-mixing regime.It should be noted that th e dnp enhancement process occursfor a given moleeule only while it is in the imm ediate proximityof the radical; but the polarized molecub retains its polarizationfor a substantial time after leaving the radical, because ofrelativeby ineffeciive reiaxation processes in the bulk liquid.The observed signal is therefore originating overwhelminglyfrom molecules in the bulk phase, This memory of collisionaffects gives dnp an unusllial adv anta ge by permitting a post

facto study of molecular interactions. The remainder of the

SCdar CQuphlg S i.hz;S sensitive to many aspects of

ates, and Poindexter

0

TAN0__

Figure 2. Investigated radicals and paramagnetic complexes.species in the bulk of the solution cannot interfere in thedetection of the interacting species, in contrast to other res-onanc e techniques. Dnp thus provides valuable informationon fast intermolecular collisions and interaction processes insolution which are not amenable to study by other resonancetechniques.~ x ~ ~ ~ ~~~~~~~

The basic apparatus for low-f ield dnp exper iments has been de-scribed previously.7 All studies were performed at a magnetic fields t rength of 75 G, for which magnet ic resonance f requencies in k H za r e U I H= 319.7, V I T = 300.8, U ~ P- 129.41, and ~ 7 ~ i- 124.01. Atthis low magnetic field, chemical shifts are too small to be detected,so t ha t only one unresolved nmr line is observed for each nucleus. Toachieve a S I N (signal to noise ratio) of 8, unpuinped 3 P and 7Lisignals ( 1 M solutions) typically required signal averaging over3000-600 0 t race s with a computer of averaged t ransients (C AT )following lock-in detection of the s ignal . Qnly 16 and 64 t r acesrespectively were required to ach ieve compa rable SIN for 1H (20-50M and for 1 (4--6 hf . Signals with radiofrequency power appliedwere sufficiently resolved to be usable after four to eight sweeps.Samples of 6-mi total volume were prepared by dissolving thedesired weight of radical or inorganic complex under an iner t at -mosphere. For air-stable radicals samples could also be deoxy genatedby several freeze-pump-thaw cycles. Wh erev er possible, radicalconcentrations of 0.02 M were used to avoid spin leakage problems.13To faci l i tate detection, the nuclear tes t p robe concentrat ion in thesolvent was at leas t 1 M . Th e bulk solvent for K 3C r ( C N ) 5N B w asa 1:1 water-acetone mixtu re to al low dissolution of the organ iccosoivent s; for a l l o ther r adical s and complexes, C H C N was usedto el iminate var iat ions in solvent effects on dnp enhancements .Deu terated solvents or reage nts were used to al low the detect ion ofchemical ly dif ferent protons in the same solut ion. The complexesK3Cr(CN )5NQ,*1 Cr(bipy)3ClQ4,22 a nd Cr(C6H6)2123were preparedby l i terature methods. Esr spectra were taken on a Var ian V-450 2X-ban d spectrometer with dua l sample caki ty.Evdaaatisn of Dwp

Both magnitude a nd sign of th e observed dnp effect dependon the na ture of the radical, on the nature of t he nmr-activenucleus, and on th e specific interaction between th e two speciesduring intermolecular collisions or encounters. Nm r-activenuclei whose dnp enhancements ar e quite sensitiveto the natureof the radical species are used as test probes, and the en-hancements obtained with a given radical are compared toresults with model radicals, whose dnp behavior is char-acteristic for a given type of interaction and whose behavioris already well understood. In thi.s way, the interactioncharacteristics of the unknown radical c an be defined.Qbserved enha nceme nts for selGcted nucleqr test probes insolutions of the mo del radicals BDPA, TTBP, and T A W0 areshown in Table 1, along with corresponding results for theinorganic complexes. T he radicals and complexes studied aredepicted in Figure 2.

8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

4/7

Second Coordination Sphere in Inorganic ComplexesTable I. 30-W Enhancem ents and Extrapolated U, Values for Systems Investigated

Inorganic Chemistry, Vol. 14, No. 2, 1975 259

Receptor moleculeOF N P(OCH,), LiClO, LiBF,SolventaRadical G(30) 'H G(30) I F U , F G(30) ' lP U , 31P G(30) 'Li U , 7Li G(30) I9F U, 19F

B D PA - 246 +219 + 270 + 745 + l o0 0 - 155 -230 - 111 - 175-267TBP -240 -123 -180d +550e -305 'TANO -20 -32.5 -75 + loo f +6OC +525' - 20 - - 2 6 5K,Cr(CN),NO -65 -55 -275 b b -88 -700 -34 -260-670' -130

Cr(bipy),ClO, -39 1-2.3 +4.3 -10.6 -78 -18 -290 b bCr(C,H,) ,I -10.3 -8 .3 -280 +29 +goo -11.2 -450 +1.65 +65a For 0.0 2M rad ica l concentrat ion. Strongly dipolar . Reference 11. Reference 17. e Reference 10. Reference 28b.Nuclear Tes t Probes. For trivalent 31P such as in trimethylphosphite, scalar coupling occurs predominantly via the lonepair, eith er in direct transien t bonding10 or through excha ngepolarization.20 Dn p with phosphine nuclea r probes gauges theextent of lone-pair interaction with a radical or param agneticcomplex.24 For F 9 in the planar octafluoronaphthalene(OFN), strong scalar coupling is correlated with transientbonding interac tion in strong plane-plane collisions withsuitable radicals.25 Res ults with OFN hus m easure interactionwith aromatic T orbitals. For 7Li in the solvated ion, scalarcoupling may arise either through ex change polarization or

through chemical bonding but fully m anifests itself only whentranslational diffusion of the ion is not impeded.llJ5 Dnp with7Li measures, among other effects, interaction of radicals withpositively charged species. For 19F in B F c , scalar couplingdoes not arise to an y appreciable extent from exchange po-larization, as attested by the low scalar component for in-teraction with TTBP, the model radical for exchange po-larization. By comparison of dnp enhancements of Li+ andBF4- in solutions of charged radicals or paramagnetic com-plexes, ion pairing and other electrostatic effects can be de-tected. In almost all cases studied so far, intermolecular spinrelax ation for 1H occurs exclusively by the dipolar mechanism;scalar coupling contributions have only been observed insystems with strong hydrogen bonding, where the proton ispolarized by spin transmission during molecular associationwith su bseq uent diffusion of the polarized species into the bulkof the solution.7J6Model Radicals. Th e efficiency with which a radical m ayinduce scalar coupling depends on steric and electronic factors.Th e availability of unpaired electron density at the peripheryof the radical, the ease with which the radical can be ap-proached, an d th e distance of closest approach ar e principalparameters.20 The m odel radicals below span a wide rangeof electronic spin delocalization and steric accessibility.BDPA . Th e sterically exposed planar n-allyl radical bis-(dipheny1ene)phenylallyl is capable of strong bonding inter-actions. The allyl carbons share 30-50 36 of the unpaired spindensity, with the remainder being distributed throughout thebenzene rings. Because of the wide exposure of electron densityabove and below the plane of the radical, nonstereospecific,rando m-b ounce collisions should be effective in transmissionof scalar coupling.24 Especially high scalar coupling is obtainedfor interactions where random ly occurring plane-plane col-lisions provide good overlap between n orbitals of suitableenergies on the colliding molecules.18919TTBP. About 30% of unpaired spin density in the planartri(ter2-buty1)phenoxy radical is located on the oxygen atom,with the remainder distributed throughout the phenyl ring.7aIn contrast to BDPA, approach of other species to withinbonding distance of the radical is either precluded or hinderedby the shielding tert-butyl groups. Consequently, scalarcontributions to total coupling are usually much smaller thanwith BDPA and arise mostly from exchange polarization.Stereospecific interactions, as exemplified by the interactionof th e oxy1 group with phosphines,20 may raise th e level of

Table 11 Per Cent Contr ib ut ion of Scalar Coupl ing to To talCoupling: lOOcl(2q + P + s + c ) .Receptor molecule

QFN P(OCH,)3 LiClO, LiBF,Nuclear probeLi F9F 31pRadical

BDPA 60 74 24 1756 7 865 86 8TTBP

28TANOK,Cr(CN) ,NOCr(bipy),ClO, 33 30Cr(C6H6)21 7 70 16 40

observed scalar coupling. Hydrogen bonding has beendemonstrated to occur with secondary phosphites, where theproton can overcome the steric shielding of TTB P, leading topolarization of the phosphorus atom by an intramolecular spintransmission pathwa y via the secondary H atom.10-24TANO. Unpaired spin density on 2,2,6,6-tetramethyl-4-piperidone-1-oxy1 is localized on the polar nitroxide, which isonly moderately shielded by the flank ing methyl groups. Asa consequence of the localized nature of the unpaired electron,scalar interactions w ith species such as OFN nd phosphinesare generally diminished compared to those with BDPA,27except where highly stereospecific point-to-p oint bonding28 orelectrostatic interactions11 can occur.ResultsTable I gives the observed enhancements and extrapolatedUrn values for the system s investigated. The ra tio method wasused to obtain Um values for 31P and 19F, using th e tech nique spreviously described.13Observed proton enhancem ents in solutions of param agneticcomplexes are generally smaller tha n in solutions of organicradicals. As seen from eq 1-5, enhancements obtained whilesatur ating an esr line a t a given power level are a function ofthe electronic TleT2e product, which is usually significantlyshorter for paramagnetic metal complexes than for organicradicals.29 Proton enhan cem ents for 0.02 M solutions ofCr(bipy)3+ and Cr(C6H6)2+ indicate TleT2e values that re-spectively are 16% and 4 of the TleT2e product in 0.02 MBDP A solution. However, 0.02 M solutions of [Cr-(CN)5NO]3- have an electronic f e T 2 e value of about twicethat for BDPA, possibly because of the low degree of spindelocalization from the metal atom and minimal intermolecularspi nsp in interaction in the chromium complex vide infra .Error analysis of the obtained data suggests that scalarcoupling components calculated to be below 10% are ofquestionable significance, due to uncertainties in unpumpedsignal intensities A 0 ) and slightly diminished leakage factorseven at the high radical concentrationsemployed in the presentwork. s can be seen in Table 11,scalar coupling comp onentsfor inorganic complexes span about the sam e range a s thosefor organic radicals, with only [Cr(CN)sNO]3- exhibitingalmo st exclusively dipolar coupling. Solvent proton en-hancements in all systems are completely dipolar withinexperimental error, confirming that th e observed polarization

l6 3726

8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

5/7

260 Inorganic Chemistry, Vol. 14, Na. 2, 1975is not due to penetration of the first coordination sphere andbonding to the meta l. For protons in Mn (H28)6 2+, positiveenhancement due to intramo lecular scalar coupling of the metalwith the bound solvent had been observed.30

[cr(c1)5N0]3-. Ultimate enhancements for ail nuclear testprobes lie near or approach the dipolar limit. This absenceof a scalar com ponent has no parallel amo ng organic radicals.Ster ic factor s cannot be held responsible, as the approach ofothe r species both to the equatorial cyanide ligands and to th eaxial nitrosyl oxygen is completely unhindered. However, itis quite obvious that the nitroxide group in TANQ is not asuitable analog for the nitrosyl group in the chromium complex.The results of Table I1 and experiments with other nucleartest probes indicate tha t the unpaired electron must be muchbetter shielded than in T TB P or in TANB, or it cannot reachthe periphery of the complex. Th e latter conclusion is sub-stanti ated by esr evidence.31 Th e unpaired electron is in anequatorial dxLorbital; intramolecular exchange polarizationi s thou ght to be responsible for the small (-0.076) unpairedelectronic spin in th e axial nitrosyl nitrogen p orbitals. Esrhf 14N I = 1) splitting of 5.25 is fa r below the 15.8 Gobserved for aliphatic nitroxides such as TA NQ . Lack ofsignificant 14N hf splitting for th e cyanide ligands com pletesthe evidence that the unpaired electron must be localized onthe me tal ion. Intermolec ular electronic spin-spin interactionis small, as seen by the low degree of dipolar broadening.Nitrosyl 14N hf co mpone nts are still distinct in esr spectra of0.02 A4 aqueou s solutions, while hf compo nents in TA NO havebroadened into a single peak at the same concentration.Dnp enhancements at 10-W esr saturation power over therange of 65-85 G trace out the three hf components of theesr absorption. Ultim ate enhancem ents for protons at thecenter of the hf components extrapolate to -100, -120, and-95, respectively, indicating the expected32 one-third sa turationfor eac h component of a triplet in which the tails of the linesoverlap only slightly. Alip hati c nitroxides, in con trast , ex-trapolate to higher Um values for each hyperfine componentof the esr spectrum.33Th e strength of dipolar coupling interactions between thecomplex and molecules in the SCS is a straightforward functionof distance of closest approach and of corre lation or interactiontimes. The absence of a scalar component makes thischromium (1) nitrosyl complex highly attrac tive for studies ofsolvation, molec ular-caging , and ion-pairing effects in the SCS.As previously concluded from esr evidence in solvated crystals,[Cr(CN)sNQ]3- does not form strong H bonds.3lb Electronspin induced nuclear relaxation for protons in H28-acetone-d6mixture is not significantly more effective than for protons inD2Q-acetone solution of the nitrosyl complex. In contrast tothese ob servations, solutions of aliphatic nitroxide radicals giverise to greater relaxation for H-bond ing solvent protons,33J4and water is known to be a much better ligand than acetonefor inner-sphere solvation.3b Radical-induced nuclear re-laxation times for 7Li and 9F in 1 M solutions of LiBF4 indilute K3Cr(CN)sNO aqueous solution (0.003 na) are muchfaster for t he cation th an th e anion, giving spin leakage valuesof 0.4 for the anion, 0.6 for the solvent protons, and 0.9 forthe cation. For radical concentrations at which nuclearspin-nuclear spin intermolecular relaxation h as become thedom inant m echanism for the anion, 90% of the nuclear spinrelaxation for the cation is still occurring by interaction withthe electron spin on the chromium com plex. Dnp is thus quitesensitive to electrostatic an d ion-pairing effects.Cr(bipy)3+ . Th e planar radical bipyridyl anion by itself iscap able of stron g plane-plane collisions with O F N and ingeneral exhibits dnp enhancements similar to the valuesobserved with BDPA.35 In distinction, the chromium(1)

ates, and Poindexterbipyridyl complex exhibits only moderate scalar coupling withles 1 and 11. The bound rings in Cr-edgewise, with the arom atic T orbitalsinteractions. The most favorable scalarcoupling mech anism, spin tran sfer in plane-plane bondinginteractions, seemingly can no longer take place. Recent nmrresults indicate that only very small solvent molecules maypenetrate into the crevices between the ligand pad dle wheels,36the regions of higher

Interaction betweenecies with moleculesin the SCS occurs via the para position of the pyridine ring,36and not by d i m t interaction with the metal scalarWtronspin-nuclear spin coupling a t the rim of rings.The unpaired electron is in an ai orb ital of the D3 molecule.373Electronic speclra,37 esr,3* and infrared data39 suggest th atconsiderable electron delocalization intoin the electronic ground state and in r,complex. Esr dat a indicate a 1 ~ 1 % dcharac ter for the un paired electron and strong bonding betweenthe metal 4s orbitals and ligand MO states of like symmetry.37There is also amp le evidence for o- delocalization of unpairedspin density in m etal complexes with pyridine-type ligands.40As intermolwcular transfer or induction of only 0.001 unpairedspin density onto the nucleus is required to give some observablescalar coupling, the edges of the bipyridyl rings should indeedbe able to ind uce a m oderate sc alar coupling pulse.As a consequence of the described steric and electronicfactors, the scalar coupling component is smaller than in BDPAbut s by no means insignificant. Th e observed positive en-lriancements with OFN make scalar coupling by exchangepolarization an unlikely mechanism, although the averagedegree of scalar coupling with the recepto r test probes in Table11resembles the values obtained with TTBP, the m d e l radicalfor exchange polarizat ion. ~ e q u ir e m e n ~ sor scalar couplingby exchange polarization are a delocalized electron spin, theexistence of low-lying excited states, and steric hindrance th at

prevents close approach and bonding interaction betweenradical and nuclear spin s p i e s 2 0 OFN exhibits a larger sca larcoupling c o ~ p o n e n ~ith @r(bipy)x+ han with TTBP,whichis sterically shielded, or with TANO, where the unpaired spinis sterically available but quite localized. The amoun t anddistribution of unpaired spin density at the rim of the sphericalCr(bipy)P is obviously SM induce electron spintransfer in collisions with th ethe presence of unpaired spithe plane of the bound bipyridyl ligand and indicate substantialinteraction between the coniplex and species in the SCS.U p to now it has not been possible to measure directly theeffect of collision attitude upon dnp enhancements for a planarradical such as BDPA. Collision of a molecule with the planeof the radica l, the sterically preferred mechanism, leads tostrong scalar coupling by transient bonding interactions.25From the results with Cr(bipy)3+, it mzy be concluded thatcollisions with the edge of a planar radical lead only tomod erate scalar coupling. As previously assumed, the ex-perimentally obtained strong scalar coupling of planar radicalswith OFW must therefore be due to interactions with thespin-rich planes of the radical.7Cr C61Cg6)2+. Th e complex exhibits strong scalar couplingwith the phosphite and fluorob clear test probes andalmost pure dipolar coupling wi The bis(ar-beniene)complex is sterically accessible 1 sides; the inter-ringdistance of abo ut 3.2 even exposes the chrom ium ion toatta ck by solvent molecules.41 In contrast to CrQbipy)3+,hestacked ben iene rings provide am ple o

plane collisions with species such as

coupling with th e C r(bipy)3+ complex also

8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

6/7

Second Coordination Sphere in Inorganic Complexesbis(benzene) complex is known to form charge-transfercomplexes with nitrobenzenes and tetracyanoethylene to givethe chromium complex cation and planar anions.42 Th eabsence of scalar coupling with O F N must be a consequenceof electronic factors, as the geometry of the complex is wellsuited f or the type of plane-plane collision which gives riseto the strong scalar coupling of O FN with BDP A. The ob-served dipo lar coupling with O F N indicates the absence ofunpaired electron density in the plane of the benzene rings andabove these planes. Accord ing to esr evidence, the unpairedelectron is in an aig (3d9) orbital of the C r atom, with about2.2%metal 4s admixture.43 In agreem ent with our dnp results,the esr studies indicate th at th ere is little char ge on the ringsand little delocalization of unpaired electron density into therings.The extremely high degree of scalar coupling with thephosphine and w ith BF4- m ust be accounted for by a differentinteraction mechanism. A priori, both excha nge polarizationand ch emica l bonding interactions may be expected to con-tribute t o scalar coupling in the case of 3 P and 7Li. However,BF4- does not show significant scalar coupling with TTBP,with which any propensity for exchange polarization shouldbe fully realized. Th e high scalar component in the 19Fcoupling with the Cr(C6H6)2+ complex must be attribu ted tobonding interactions. In order for scalar coupling by bondinginteraction to b e observable , distances of closest approach mu stbe about 2-3 A , and correlation or sticking times for theinteractio n m ust ex ceed 10-10 sec.7315 T he most obvious modeof interaction by which OFN alone among the nuclear testprobes cannot be polarized efficiently is direct intrusion of amolecular species into the space between the benzene ringsand dire ct bonding w ith the metal located in the xy plane ofthe complex, the region of high unpaired spin density. Asimilar bonding interaction with the metal has been postulatedfo r th e liga nd-ex change reactio ns of C T ( C ~ H ~ ) ( C O ) ~ . ~ ~Bond ing with the BF4- anion is most likely due to electrostaticattraction to the chromium cation. The slight scalar componentfor 7Li in turn may be due to ion pairing with the fluoroborateanion, attrac ting Lif to an area of high unpaired spin density.Sca lar coupling of the phosphine 3 P is most likely a resultof bonding interaction of the exposed chromium a tom withthe phosphine lone pair.Molecular interaction between Cr(C6H6)2+ and scs speciesis thus characterized by weak solvation of the benzene ringsand by strong transient bonding of th e sandwiched chromiumatom with negatively charged groups or atoms and with lonepairs of suitable ligands.Conclusions

Intermolecu lar e ncounter or collision between an inorganicor organometallic paramagnetic complex and other species insolution may give rise to both dipolar and scalar coupling ofelectron and nuclear spin. Th e relative contribution of eithercom ponent can be measured by low-field dnp. Protonsgenerally exhibit only dipolar coupling in intermolecularinteractions, while 7Li, 19F, and 3 P may exhibit moderate tostrong scalar coupling. The exten t of the latter relaxationprocess depends on the availability of unpa ired elec tron densityat the periphery of the complex, the possibility of close ap-proach to within the effective limits for scalar coupling, andthe presence of chem ical or physical forces encouraging closeand protracted approach.The presence of scalar coupling either from exchangepolarization effects or from the presence of bonding interactionsbetween the paramagnetic complex and molecules in the SCSallows a qualitative description of these molecular interactions.Nuclear test probes can trace out the availability of unpairedelectron density at th e surface of a param agnetic complex inwhich the unp aired electron resides primarily on the metal.

Inorganic Chemistry, Vol. 14, No. 2, 1975 261Results from dnp experiments are in agreement with con-clusions reached from esr evidence.Although dnp is not sensitive to angular variations in in-termolecular encounters in the sam e way as pseudocontactchemical shift techniques,za stereospecific intermolecularbonding effects can frequently be identified. Ion-pairing andion-repulsion effects are graphically demonstrated in theintermolecular encounters between electrically charged radicalsor paramagnetic complexes and ionic nuclear test probes.Detailed information on fast molecular interactions betweenradical inorganic complexes and molecular species in thesolvation shell can be ob tained as in previous studies on organicradicals.7a

Acknowledgment. The authors gratefully acknowledge thehelp of Anthony J. Montedoro in sample preparation and inoperation of instrumentation.Registry No. BDPA, 21 52-02-5; TTBP, 2525-39-5; TA NO ,2896-70-0; K3Cr(CN ) sNO, 14100-08-4; Cr(bipy ) 3C104, 14524-27-7;Cr(C6H6)21, 12089-29-1; O F N , 7789-25-5; P(OCH 3)3, 121-45-9;LiC104, 77 91-03-9; LiBF4, 14283-07-9; 3*P, 7723- 14-0; 7Li,13982-05-3.

References and Notes.Presented in part at the 166th National Meeting of the AmericanChemical Society, Chicago, Ill., Aug 1973; see Abstracts, No. INOR107.(a) D. R. Eaton,Advan. Che m. Ser., No. 100, 174 (1967), and referencestherein; (b) V. Gutman, Coord. Chem. Rev., 2, 239 (1967) ; (c) J.Bjerrum, Advan. Chem. Ser. , No. 62, 178 (1967).(a) P. Beck, Coord. Chem. Rev., 3, 91 (1968); (b) A. Fratiello, Progr.Znorg. Chem ., 17, 57 (1972); (c) G. N. La Mar and G . R . Van Hecke,J . Chem . Phys., 52, 5676 (1972).I. Solomon, Phys. Rev. , 99, 559 (1955) .(a) T. R . Stengle and C. H . Langford,J . Phys. Chem., 69, 3299 (1965);(b) S. Behrendt, C . H. Langford, and L. S. Frankel, J . A mer . Chem.SOC. , 1, 2236 (1969); (c) L. S. Frankel, J . Phys. Chem., 73, 3897(1969); (d) L. S. Frankel, ibid., 74, 1645 (1970).H. M. McConnell and R. E.Robertson, J . Chem. Phys., 29, 1361 (1958).(a) J. A. Potenza, Advan. Mol. Relaxati on Processes, 4, 229 (1972);(b) K. H. Hausser and D. Stehlik, Advan. Magn. Resonance, 3, 79(1969).W. Muller-Warmuth, 2 Naturforsch. A , 21, 153 (1966).E. H. Poindexter, J. R. Stewart, and P. J . Caplan, J . Chem. Phys. , 47,2862 (1967).E. H . Poindexter, R . A. Dwek, and J. A . Potenza, J . Chem . Phys., 51,628 (1969).J . A. Potenza and J. W . Linowski, J . Chem . Phys. , 54, 4095 (1971).R. D. Bates, Jr., B. E. Wagner, and E. H. Poindexter, Chem. Phys. Left.,17, 328 (1972).R. D. Bates, Jr., E. H. Poindexter, and B. E. Wagner, J . Chem . Phys. ,59, 3031 (1973).A. W . Overhauser, Phys. Rev. , 92, 411 (1953).B. E. Wagner, R. D. Bates, Jr., and E. H . Poindexter, paper presentedat the 5th Northeast Regional Meeting, of the American ChemicalSociety, Rochester, N . Y ., Oct 1973; see Abstracts, No. 046.(a) J. Leblond, J. Uebersfeld, and J. Korringa, Phys. Rev., 4, 1532 (1971);(b) J. Leblond, P. Papon, and J. Korringa, ibid., 4, 1539 (1971).E. H . Poindexter, J. A. Potenza, D. D. Thompson, N. V . Nghia, andR . H . Webb, Mol. Phys. , 14, 3 8 5 (1968).(a) G. J. Kruger, W. M uller-Warmuth, and R . van Steenwinkel, Z .Naturforsch. A , 22, 2102 (1967); (b) E. H . Poindexter, P. J . Caplan,B. E. Wagner, and R. D. Bates, J . Chem. Phys. , 61, 3821 (1974).(a) J. A. Potenza and E. H . Poindexter, J . Amer. Chem. Soc., 90, 6309(1968); (b) R. A. Dwek, J. G. Kenworthy, D. F. Natusch, R. E.Richards, and D . J. Shields, Proc. Roy . Soc., Ser. A , 291, 487 (1966).R. A. Dwek, R. E. Richards, and D. Taylor, J . Chem. SOC.A, 1173(1970).W. P.Griffith, J. Lewis, and G. Wilkinson, J . Chem. Soc., 872 (1959).Fr. Hein and S. Herzog, Z . Anorg. Allg. Chem., 267, 337 (1952).G . Brauer, Handbook of Preparative Inorganic Chemistry, Vol. 2 ,2nded, Academic Press, New York, N. Y., 1965, p 1397.J. A. Potenza, E.H. Poindexter, P. J. Caplan, and R . A. Dwek,J . Amer.Chem . Soc., 91, 4356 (1969).R. H. Webb, N. V. Nghia, M . R. Perlman, E. H. Poindexter, P. J.Caplan, and J. A. Potenza, J . Chem. Phys. , 50, 4408 (1969).R. A. Dwek, J . G. Kenworthy, and R. E. Richards, Mol. Phys. , 10,529(1966).R. L. Glazer and E. H . Poindexter, J . Chem . Phys. , 55, 4548 (1971).E.H. Poindexter and R. L . Glazer, J . Amer. Chem. Soc., 92,4784 (1970).J. A . Pople, W. G . Schneider, and H . J. Bernstein, High-ResolutionNuclear M agnetic Resonance, McGraw -Hill, New Yo rk,N. Y., 1959,pp 199-217.

8/13/2019 ic50Investigation of the second coordination sphere in inorganic complexes by dynamic nuclear polarization (DNP).

7/7

262 Inorganic Chemistry, Vol. 14, No. 2, 1975 obin W erutz and James J Turrxer(30) R. S. Codrington and N. Blombergen, J . Chem. Phys., 29 , 600 (1958).( 31 ) (a) J. J. Foitm an and. R . G . Hayes, J. Chem. Phys. ,43, 15 (1965); (b)H . A . Kuska and M . T. Rogers, J . Chern. Phys., 42, 3034 (1964); (c)1.Bernal and S . E. Harrison, ibid., 34, 02 (1961).(32) A . Abragam, "The Principles of Nuclear Magnetism," Clarendon Press,Oxford, England, 1961, p 340.(33) R . D. ates, Jr. , B . Wagner. and E . oindexter, in preparation.(34) (a) I. Morishima, K. Endo, and T.Yonezawa, J . Chem. Phys. ,58 3146(1973); (b) PX.S. Davis and R. W . Kreilick, J . Amer. Chem. Soc., 95,5514 (1973).(35) B. E . Wagner and E. P i Poindcx ter, unpublished rcsults.(36) 6 . La Mar and 6 . Van Hccke, Inorg. Chem., 12, 1'767 (1973) .( 3 7 ) E. Konig, Z . Matw~orsorsch.A , 19. 1139 (1964).

(38) E. Konig and S . Herzog, J . Inorg. Wucl. Chem. , 32, 585 (1970) .(39) Y. Saito, J. Takemoto, B utchinson, and K. Nakamoto, Inorg. Chem.,11, 2003 (1972) .(40) (a) J. A. Happeand R. L. Ward. J. Chem. Phys. , 39, 1211 (1963); (b)R .H. Holm, G. W. Everett, Jr.. and 6 .. Horrocks,Jr., J . Amer. Chem.So,., 88, 1071 (1966); (e) W . D. Horrocks, Jr., and U. I,. Johnston,Inorg. Cheni., I D , 1835 (1971).(41) F. A. Cotton, W . A. Dollase, and J. S . Wood. J . Amer. Chem. SOC., 5,1543 (1963).(42) J . W. Fi tch, 111, and 3. J . Lagowski, Inorg. Chem., 4, 864 (1965).(43) R. Prins and F. J. Reinders, Chem. Phys. Lelt. , 3, 45 (1969).(44) J . D. Holmes, D. A. K. Jones, and R . Pettit, J . Orgammetnl . Chem. ,4, 324 (196.5).

Cont r ibut ion f rom the Depar tment of Inorganic Chemistry,The Univers i ty , Newcas t l e upon Tyne, NE1 7KU, England

ROBIN W. PERUTZ and JAMES J. TURNER*Received Ju n e 4, 197-3 ATC40357Q

The infrared spectra of the 13CO-enriched metal hexacarbonyls and pentacarbonyls have been studied in Ar and CH4 matricesa t 20 K. The hexacarbonyl spect ra can be f i t t ed very accurately in f requency and intensi ty us ing a CO-factored forcefield. The spectra of the pentacarbonyls a re inconsistent with a D3h structu re but can be fitted accurately using a C4 tructure.Using intensi ty d ata , axial -radial bond angles C ~ Vt ructure) between 90 and 95 are calculated. The s t ruc ture and forcecons tant s ar e a lmos t i ndependent of t he mat r ix mater i a l .Chromium, molybdenum and tungsten pentacarbonyls havebeen generated by u-f photolysis of the hexacarbonyls inhydrocarbon glasses* a t 77 and in argon m atrices1 at 20 K.In the ir spectra three bands were observed in the C - 0stretching region e.g.,Cr(@ 5 in Ar: 2093 (ww), 1965.6(s), 1936.1 (m) cm-1). The bands were assigned to the Ai ,E, and Ai bands, respectively, of a square-pyramidal (C4,)molecule. The evidence for this structure a s opposed to tha tfor a D3h structure was based on the presence of the very weakhigh-frequency band and on the intensity ratio of the low,-frequency bands. During investigation of the uv spectra ofthe pentacarbonyls in different m atrices,3 it becam e essentialto have stronger evidence that the I A h structure was incorrectand to estimate bond angles for a C4v structure.In the experiments described below, the ir spectra oflU 2- en ric he d metal pentacarbonyls were studied using themeth ods of analysis developed by Ha as a nd Sheline,'?and Dar ling 2nd Ogden.6 Th e results prove tha t the C4vstructure for photociiemically generated M ( C 0 ) s is indeedcorrect.

Analysis o f SpectraWhen a metal carbonyl is enriched to approximately 50%all possible isotopic molecules M 1 2 C O ) x 1 3 Cx il 0, I , ...) E ) are present in a scrambid mixture whosecomposition is determined by the statistical weights o f thediffercnt molecules. Since the vibrational symmetry is loweredby 1 0substitution, the ip. spectrum in the e-o stretchingregion c onsists of the sum of the su perimpose d spectra of all.the individual molecules, each of which has a different vi-

gden6 have calculated the patterns o f thef metal carbonyls for: those cases in whichparent molecule has no permanent dipole moment and allgroups ar e equivalent i t . , M(12CQ)x(13CO)n-x: n 2,. Dmh gt.,Ometrgi; ? Lz 3?B 3 h ge5rtnebry; ?2 -= 4,Td Or D4h; Iz

4 Oh). En such molecules there is only one C-O stretchingforce constant and not more than two Cconstants. Such calculations have beenspectra of a number of metal carbon yls and dinitrogen species.7M(GO)s presents a m ore complex case because it must haveat least two different GO stretching force constants and threeinteraction constants whether the molecular symmetry is D 3 hor C4v. In intensity calcula tions for the C4, geometry we mustinclude bond angle and bond moment data.* The distinctionbetween C4v, B 3 h , or any geometry of lower symmetry restson detailed comparison of experimental spectra and a rangeof possible theoretical spe ctra. We assume that the moleculecture for which the experimental spectrathe structure with the minimum numberSuc h an assignment depends on having air estimate of theerrors expected in the prediction of the isotopic spectra. Therehave been a number of studies of isotopically enriched car-bonyls in solution but most have involved only partial en-richment.9 Ni(C0)4 and eo(eo)3NCF have been examinedwith more extensive ewric ese compounds haverelatively few bands; Nsac studied Fe(C895 butdid not optimize the force constants. Johnson, et ab. l1 haveexamine d the spectrum of in chloroform duringCW l-C1*0exchange and have refined the Cotton--Kraihanzelparame ters and calcarlated intensity da ta; the spectra howeverwere only of moderate resolution. Da.rling12 has generatedCl W-substituted Cr(C0)6 in Kr matrices by cocondensationof Gr atoms and CO-ICr mixtures? ut the sp xtra are not sharpplete analysis. W e therefore decided 10species in low-tem perature maErices asa check on our calculations a.nd to obtain an estimate of thesystem atic errors in the method. (Details of the methods ofcalculation of frequency and intensity a re given in Appendix2.)~ ~ o ~ ~ ~ ~ e ~ p ~ ~ a ~ ~nakices have the advantage that the