The Vibrational and Electronic Spectrum, and the Potential Field of Various Manganese Carbonyl Complexes Citation Ottesen, David Keith (1971) The Vibrational and Electronic Spectrum, and the Potential Field of Various Manganese Carbonyl Complexes. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/xbxd-q624. https://resolver.caltech.edu/CaltechTHESIS:11202017-154553422 Abstract I. The infra-red and Raman spectra of Mn(CO) 5 Br as both solid and CH 2 Cl 2 solution from 50 to 2150 cm -1 are presented for the normal isotopic compound, and the 13 CO and C 18 O substituted molecules. The observed bands are assigned, and the force constants and compliance constants of a general quadratic potential field are calculated using the FG method of Wilson, Decius, and Cross. II. The infra-red and Raman spectra of both Mn 2 ( 12 CO) 10 and Mn 2 ( 13 CO) 10 as solids and in CH 2 Cl 2 solutions from 50 to 2150 cm -1 are presented and are assigned. Potential constants are calculated using values transferred from the Mn(CO) 5 Br calculations. Special attention is paid to splitting interactions across the Mn--Mn bond and to the Mn--Mn stretching potential constants. III. Raman intensities and the electronic absorption spectrum of M(CO) 5 X n- compounds are presented and discussed. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Gray, Harry B. Thesis Committee: Unknown, Unknown Defense Date: 16 September 1970 Funders: Funding Agency Grant Number Caltech UNSPECIFIED Atomic Energy Commission UNSPECIFIED NSF UNSPECIFIED Record Number: CaltechTHESIS:11202017-154553422 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:11202017-154553422 DOI: 10.7907/xbxd-q624 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 10564 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 21 Nov 2017 15:08 Last Modified: 25 Jun 2024 00:15 Thesis Files Preview PDF - Final Version See Usage Policy. 59MB Repository Staff Only: item control page

THE VIBRATIONAL AND ELECTRONIC SPECTRUM, AND THE POTENTIAL FIELD OF VARIOUS MANGANESE CARBONYL COMPLEXES Thesis by David Keith Ottesen In Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1971 (Submitted September 16g 1970) Copyright © by DAVID KEITH OTTESEN 1971 ii To Jo Ann iii ACKNOWLEDGMENTS I thank Dr. J. Michael Smith who first suggested this topic of research to me. I am especially grateful to Professor Harry B. Gray at Caltech and Dr. Llewellyn Jones at the Los Alamos Scientific Laboratory for their continued interest and aid in the course of my studiese I am grateful to my colleaguesg especially Mr. De Titus, for their assistance in my research and many enlightening discussions. My thanks are given to my parents for their devotion and support, and to Professors Latimer Evans and Robert Rowan at New Mexico State University for their encouragement of my early scientific endeavors. I am most grateful to my wife, Jo Ann, for her aid in typing the majority of this thesis. Finally I wish to acknowledge the financial support of this Instituteg the Atomic Energy Commission through three summers spent at the Los Alamos Scientific Laboratory, and the National Science Foundation. iv ABSTRACT I. The infra-red and Raman spectra of Mn(Co) 5Br as both solid and CH2Cl2 solution from 50 to 2150cm -l are presented for the normal isotopic compound, and the 1 3co and c 18o substituted molecules. The observed bands are assigned, and the force constants and compliance constants of a general quadratic potential field are calculated using the FG method of Wilson, Decius, and Cross. II. The infra-red and Raman spectra of both Mn 2 ( 12co) 10 and Mn 2 ( 1 3co) 10 as solids and in CH2Cl2 solutions from 50 to 2150cm-1 are presented and are assigned. Poten- tial constants are calculated using values transferred from the Mn(Co) 5Br calculations. Special attention is paid to splitting interactions across the Mn--Mn bond and to the Mn--Mn stretching potential constantso III~ Raman intensities and the electronic absorption spectrum of M(Co) 5xn- compounds are presented and discussed. v TABLE OF CONTENTS Page ACKNOWLEDGMENTS • ABSTRACT • • • • • 8 0 • • • 0 • • • 8 0 iii iv PART I. Chapter 1. Introduction--The Interpretation of Chemical Bonding in Metal Carbonyl Complexes via their Observed Vibrational Spectra • 2. The Vibra~ional Spectrum of Mn(Co) 5Br 2 • 6 • • )2 1 •. The Vibrational Spectrum of Mn2(C0) 1 o • 104 2. Potential Constant Calculation for Mn2(C0)10 • • • • • • • • • 131 3• . Isotopic Substitution and the Potential Field of Mn(CO) 5Br • • • • • PART II. • PART III. 1. Raman Spectra of M(Co) 5xn- Compounds 2. Electronic Absorption Spectra of M(Co) 5Brn- • o • • o ~ • • o o Final Summary and Evaluation of Results. J. • 169 181 Bibliography 0 • • • • • 0 • 0 • • 186 Propositions • • • • • • • • • 8 • 191 1 I. VIBRATIONAL ANALYSIS OF MANGANESE PENTACARBONYL BROMIDE 2 CHAPTER 1 INTRODUCTION -- THE INTERPRETATION OF CHEMICAL BONDING IN METAL CARBONYL COMPLEXES VIA THEIR OBSERVED VIBRATIONAL SPECTRA In the past few years a large number of vibrational analyses dealing with metal-carbonyl complexes have appeared in the literature; some of the more thorough of these investigations are (I- 14). For the main part, however, these studies dealt with assignments of infra-red frequencies which are primarily C--0 stretching vibrations and the calculation of force fields of greater or lesser complexity for these vibrational coordinates. These simple calcula- tions have proved quite useful in the interpretation of bonding changes when one or more ligands is substituted for a CO group or groups. (8, 15 ) We may infer the nature of the substituted ligandto-metal bond by observing changes in the C--0 stretching 'frequencies on the basis of the synergic dTr....rr*COnbackbonding99 model o This is illustrated below in figure 1. A brief explanation is that in the case of octahedrally substituted complexes (with which this work deals exclusively) a given CO ligand will be in competition with the substituted ligand trans to itself for the d-7f electrons in the metal orbital which may combine with the 'TT"* antibonding orbitals of the ligands. CO •M M ,. CO 0-bonding 7rbackbonding Figure 1. 0M<:)+ ~c=o: - @c=o OM~+ <21c,,c/) - C:JM~c=oP ~~o g QQ~ Synergic Model of Metal-CO Bonding (16) As a first approximation it is considered ( 1 ) that the trans interaction will be twice that of a cis interaction simply because the metal may back-bond to two trans-ligands through two d-orbitals, while only one ct-orbital of the appropriate symmetry for back-bonding is common to two cis-ligands. From this postulate we reason that a ligand of less 'tr-accepting character than a CO group will have the effect of increasing the d7T--.Tf * bonding in the CO group trans to itself while exerting a · much smaller effect in those ligands cis to itself. As the bond order of the trans M--C bond rises howeverp the trans C--0 bond order will decrease due to the increasing population of electrons 4 in the 1f* anti-bonding orbital. This should have the net effect of lowering the trans CO stretching frequency and raising the trans MC stretching frequency relative to the unsubstituted compound. A second effect will be changes in the electron density of the sigma bonding system due to differences in the electronegative character of the substituted ligand. The two effects are by no means separable; however, the predominance of one effect over the other is believed responsible for systematic variations in frequency and intensity of the Raman active M--C stretching modes of the M(Co) 5xn- (M =Cr or Mn) systems. This is discussed in conjunction with the electronic spectra of these molecules in Part III. Part I of this thesis is devoted to a complete vibrational analysis and normal coordinate calculation of the molecule Mn(C0)5Bre The motivation for this work was to accurately and fully assign the observed vibrational spectrum of this molecule and from this to calculate a potential field which could be compared with simpler force constant calculationso We may gain by this a knowledge of the accuracy and limitations inherent in previous work concerned only with v•c--o stretching" vibrationso A further extension of this work is the transfer of 5 the calculated force field to the more complicated Mn 2 Cco) 10 system; the remaining interaction constants in this binuclear complex will define the splitting of vibrations whose symmetry coordinates differ from each other only in regard to the rotation axes perpendicular to the metal-metal bondJ ioe•g The magnitude and sign of these interaction constants may give information concerning the electronic structure of the metal-metal bond system. These approximate transfer calculations will be discussed in Part II. 6 CHAPTER 2 THE VIBRATIONAL SPECTRUM OF Mn(C0) 5Br This compound was prepared in accordance with the literature method (17) by the reaction of Br 2 in CCl4 solution with commercially available Mn2(C0) 1o at room temperature for about 15 minutes~ The yellow-orange pro- duct is considerably less soluble in CCl4 than the parent carbonyl and it precipitates out of the reaction mixture. The initial product usually showed only traces of Mn 2 (co) 10 or Mn2(CO)sBr2 present. The former could be extracted by a suitable hydrocarbon solvent such as cyclohexane while the latter, being even less soluble than Mn(C0)5Br in halocarbons, could be removed by recrystallization from a solvent such as CHC1 3 or CH2Cl2• This procedure yielded the pure monomeric compound whose CCl4 solution I. R. spectrum from 2150 to 1950cm- 1 agreed with the literature values (6 ). This was taken as the criterion for purity when accurate spectral measurements were taken& The intense bands in both the I. R. and Raman spectra were found to fall into three energy groups; they are 2150 to 1950cm-1, 650 to 200cm-1, and 150 to 50cm-1. The highest energy vibrations may be assigned to modes involving principally C--o stretching motion and these 7 are discussed below. C--0 Stretching Vibrations The symmetry of this molecule is C4v and the geometry along with the internal vibration coordinates are shown in figure ~. The observed I. R. and Raman spectra are illu- strated in figures 3 and 4 and consist of one moderate band, one very strong band, and one strong band in the infrared in order of decreasing energy; in the Raman spectrum there are two strong bands followed by a moderate band in order of decreasing energy. Group theory predicts the symmetry of the CO stretching modes to be 2A 1 + Bi + E. Of these the B1 mode will I. Ro inactive, and all four modes should be Raman active. These Hsymmetry" coordinates are shown in figure 11. It should be emphasized firmly at this point that these coordinates do not describe exactly the vibrations involved but merely serve as useful approximations. The calculations in the following chapter will yield the amount of other symmetry coordinates which is mixed with each of these in determining the moleculews normal coordinates of vibrationo The neglecting of mixing is precisely the approximation in simpler force field calculations which were mentioned in Chapter lo The assignment of these CO bands has been well established ( oJ 1e ) and is as follows. The mode at 2085 in 8 12 e1' rl 10--! ¢ 10, 3 ~=Br Q =Mn • =c 0 4 0 0 =0 5 Figure 2. Internal Coordinates for Mn(Co) 5Br 211)0 Figure Z/ (cm-1) 2050 2000 3. Raman Spectrum of Mn(Co) 5Br in CH 2c1 2 Solution 2100 Sensitivity = 1500 (Dynode 4) Slit = 10 cm-l;Double Slit;lO cm high '° 21t50 0 ~ ~ Figure 4. 2100 2050 2000 I.Re Spectrum of Mn(C0)5Br in C6Hi2 Solution V (cm- 1 ) 0 1--& 11 the Raman is assigned as (B 1 >co-ri as it appears to be I. Ro inactive. The weak band at 2138 in the I. R. is assigned as (A 1 >co-ri as this band is expected to have the lowest intensity of the three I. R. modes. If the radial-CO ligands were actually planar and if this mode were unmixed with the A:t_axial-CO stretch, this band should have zero intensity. The very intense I. R. band at 2052 is assigned as (E)co-ri as it does not appear in the Raman spectrum. This is reasonable as the Raman intensity of the E mode arises from the azx and Clyz terms of the polarizability tensor which are expected to be considerably smaller than the O.zz diagonal term responsible for the inten- sity of the A1 axial mode. This assignment leaves the band at 2007 as the {A 1 ) 00 _r vibration. The synergic back-bonding model 0 in th~ introduction agrees entirely with these assignments requiring that the (A 1 ) axial-CO mode occur at lowest energy. As Br- is a considerably poorer 1T-elec- tron acceptor than a CO ligand, it is expected that the d7T-P'7T* electron density will increase for the axial Mn--C bond, and the order of the C--0 axial bond will thus decrease relative· to the radial c--o bonds. 12 Mn--C--0 bending, Mn--C and Mn--Br stretching Vibrations The next highest energy-group o:f . bands is :found between 650 and 200cm- 1 and these are shown in figures 5 , 6, 7 , and 8 ~· From studies on the Cr subgroup hexacarbonyls (Fl ) and by comparison with spectra of other pentacarbonyl halide derivatives of Mn (19 - 21) these bands may be assigned as Mn--C--0 bending modes :from 650 to 400cm- 1 and Mn--C stretching modes from 500 to JSOom- 1 1 the Mn--Br stretching mode should be found between 200 and 250cm· 1 • The symmetry of these modes isa MCO Bend ~ A1 + A2 + B1 + B2 + E MC Stretch = 2A 1 + B1 + EJ MBr Stretch = A1 The internal symmetry coordinates for these vibrations are in figure ii~ A very helpful tool in the assignment of these modes to observed frequencies is the phenomenon of Raman depolarization ratios for solution spectra: Experimentally this involves the use of a pola- roid film and a linearly polarized exciting source (which is satisfied by the available He-Ne laser instrumentation)~· Two readings are taken for a given band intensity, first with the analyzer film's electric vector perpendicular to the laser light's electric vector, and second with these vectors parallel. Raman intensity theory (11) predicts a ratio of Figure 5. 650 600 550 Raman Spectrum of Solid Mn( CO) 5Br 1/ (cm-1) Sensitivity = 200 (Dynode 4) Slit = 5 cm-1;Double Slit;5 cm high ~ VJ Figure b • 400 Raman Spectrum of Solid Mn(CO) Br 5 l/ (cm-1) Sensitivity = 200 (Dynode 4) Slit = 5 cm- 1 ;Double Slit;2.5 cm high ....+:" 15 • • ~~ 0 ~ s:: - ·r-1 ...... I a 0 ~ H ~ ~ 0 0 .......... @0 tll Ci-! ·r-1 00 S'D 0 0 '° ::s <Ll H tll +> tll () <Ll <Ll H p.p.. Cf.l H -~ I):; ~ ... • • •• • • • • • ....... • ...... . 0 H • «! ';/!. ,..: Figure 450 e. I.R. Spectrum of Mn(C0)5Br in 400 a KBr Pressed Disc l/ (cm- 1 ) ~ . ...... .. .. . . .. .. ..... 350 • • • • · 120 - - JOO K. K• °' .... 17 . • 75 for all peaks except those modes which are totally symmetric (in this case, A1 vibrations) which may have a ratio somewhere between O and .75. Three ,strongly polarized bands were observed in CH2Cl2 solution at 222, 381, and 470cm- 1 • The lowest of these A1 modes is assigned to (A 1 ~, the Mn--Br stretching vibration. In accordance with the bonding model and the studies on M(C0)6, the axial Mn--C stretching vibration should occur at higher energy than the radial Mn--C stretch, and both should fall between 350 and 500cm-1. Thus v s1 3 is assigned to (A 1 )d. and V4 7o to (A1)a • 1 0 There remains one polarized Raman band in this energy range which was not observed; this is the A1 Mn--C--0 out-of-plane bending mode. This has previously been assigned to the 470cm- 1 frequency (19) but this is deemed highly unreasonable in view of the recent studies on M( CO) 6 compounds (14 ) • For these octahedral complexes the vibration corresponding to (A1)y is the I. R. active Fiu mode observed at 665cm- 1 for Cr(C0)6 in CClL~ solution. Thus the assignment of the two observed Raman polarized bands as ~1n--C stretches is retained as the most reasonable. The moderate I. R. peak at 417cm- 1 is assigned as the (E)ai mode, and the moderately weak Raman band in the solid at 414cm- 1 is assigned to {B1)d.a 1 When the symmetry of octahedral M(Co) 6 is lowered 18 from Oh to C4v in M(C0)5X, the Flu MCO bend is split into components of A1 and E symmetry. This is dia- grammed along with the changes in symmetry for the other MOO bending modes in figure 11. We therefore take the bands at 638 and 645cm-1 in the I. R. CH 2c1 2 solution spectrum as (A 1 )y and (E),e • The band at 6J8cm- 1 is the more intense in the I. R. and the less intense in the Ramani further, the highest energy MCO bend for Mn2(Co) 10 in CH 2c1 2 solution has been observed as a polarized Raman band at 67ocm-l and is the A1 mode analagous to (A 1)1 for Mn(C0)5Bre these facts we take the ordering l! 645 - From (A1 )y and V638 (E)µ • The remaining two I. R. peaks at 604cm- 1 and S45cm-1 must be (E)y and (E), while the Raman active peaks at 627cm-1 and 5J5cm- 1 in the solid are assigned as the ( B1 ~ and ( B2 p modes e To obtain the proper ordering of these modes we make use of the synergic back-bonding model to predict changes in these frequencies between Cr(C0)6 and Mn(C0)5Bri From the symmetry correlation diagram in figure i1 it is seen that both B1 and B2 m~des should lie around 5JOcm-1~ from the observed spectra it is obvious that one of the modes has been shifted to considerably higher energyi The bonding model predicts a higher 19 electron density in the dzx and dyz orbitals between the Mn atom and axial-CO group than in the dxy orbital shared by the radial-CO groups. The B1 bending mode involves the "(, out-of-plane, coordinate, and the main effect of this motion (as illustrated by the isotopic shifts in Chapter 3) is the displacement of a radial C atom in an axial direction. The B2 bending mode will move the same C atom in the in-plane direction perpendicular to the M--C--0 axise Thus the B1 mode will involve primarily distortion of the Tr bonding in the dzx and dyz to CO-Tr* orbitals, while the B2 mode will cause distortions mainly in the dzx .Q! dyz and the less occupied dxy to CO-Tr* systems. This implies the restoring force will be greater for the out-of-plane B1 mode than the in-plane B2 mode and we assigri 116 27 (B 1 )1 and v535 (B2~ o The two remaining E modes have apparently both shifted to higher energy from the analogous modes in Cr(C0)6 indicating that considerable mixing of these two internal coordinates is occurring~ Calculations on both possible assignments for l.16o4 and v 4 with 5 5 (E)y and (E)~ have been carried outo The inactive (A 2 )p mode was not observed and its position is calculated from the potential field described in Chapter 3o 20 C--Mn--C and C--Mn--Br Bending Vibrations These vibrations are expected to occur between 200 and SOcm- 1 and are shown in figures 9 and 10. A complicating factor might be the presence of lattice modes in the spectra of the solid compound. Due to the rather low solubility of Mn(Co) Br in even CH 2c1 2 5 it was necessary to obtain data in the solid phase as only one band was observed for Raman ~olution spectra and no bands were observable in I. R. solution spectra. Fortunately the Mn(Co) Br molecule is neutral and it 5 is reasonable that the intermolecular forces are quite small. Thus the optically active lattice modes may quite reasonably be assumed to lie below SOcm- 1 with the result that all observed bands may be assigned to internal vibrations. Not only did solubility problems limit any information which could be obtained from Raman depolarization experiments, but also the solid I. R. and Raman spectrum proved to be quite complex; further, a lack of trends between the spectra of the Cl, Br, and I pentacarbonyl compounds c~-21) indicates that a large amount of mixing is occurring among the symmetry coordinates chosen to describe the vibrational modes. This simply means that changing the halogen ligand will affect not Figure 9 • 200 1 1SO V(cm- ) 100 Raman Spectrum of Solid Mn(CO) 5Br Sensitivity = 100 (Dynode 4) Slit = 5 cm-1;single Slit;2e5 cm high N ~ 22 0 0,.... s:: •rl H f'.Q M M -:::s 0 :2:: 0 ......,M @.~ ft; 0 s z:::s '"d (!) ;:s.µ H ru -PH 0 ,.... o.µ - (J) en o s:: T"I I s0 ~ • 0 i o::o • H • 0 -4 cu H •). O/o s:: p.. (!) ~ 0 0 •rl (\J 1%.4 cd 23 only the frequency having to do with C--Mn--X bending, but will perturb, to a certain extent, all the deformation modes of the same symmetry. The symmetry of the modes in this energy range ist C--Mn--C bend = A1 + B1 + B2 + 2E C--Mn--Br bend = E. These are shown in figure 11. Experiments on Mn2(Co) 1o in CH 2Cl2 indicate the totally symmetric C--Mn--C bend lies around 111cm- 1 ; this is analogous to (A1)l'_ and 5 thus this band is assigned to the very intense Raman mode at 116cm- 1 o The I. R.- band at lJJcm- 1 seems to be very weak in the Raman and is thus assigned as a C--Mn--C bend of E symmetry; due to the weaker bonding and the heavier atom involved 9 it is expected that the (E)8 mode involving mostly C--Mn--Br bending will be at considerably lower energy than 1.3Jcm- 1, thus this band is either (E)~ or (E)d • The intense Raman bands at 129cm· 1 and 99cm- 1 are assigned to (B2)oc and (B1)l'-c5 on the basis of a symmetry correlation diagram with The weak I. R. band at lOOcm- 1 coincident with the Raman frequency at 99cm- 1 is judged Cr(C0)6 (figure 12)~ to be the I. R. forbidden (B1~-s mode; this mode is slightly allowed under the different selection rules prevailing for the lower factor-group symmetry of Mn(C0)5Br in the solid state. 24 The I. R• active--only band at 87cm- 1 is the other member of the (E)~ - (E)~ pair with the peak at lJJcm-1. This leaves the very weak Raman and I. R• band at 1 as the (E) mode. N 53cm· 5 While these assignments are certainly only tentative, it will be seen in Chapter 3 that a quite good agreement of the calculated potential field compared with values transferred from Cr( CO) 6 for this assignment. (J."t) is obtained 25 ?12138 - (Al) ri s 1 = 1/2(r1 + r 2 + r 3 + r4) V2085 - - - I > (Bl)ri v2052 - s 9 • 1/2(r1 - r 2 + S15 r3 - r4) Figure 11. (E) ri = 1/2(r1 + r2 r 3 - r 4) Internal Symmetry Coordinates and Frequency Assignments for C--0 Stretching Modes 26 l/645 __. (A 1 )y s J • 1/2 <Y 1 .+ Y 2 + r J + Y4) not obs.d ...... (A2),d S :8~ a: 1/2(f31 + (3 2 + (3 3 + ~4) ?1'627--=- (Bl )y l.1535 ~(B2)p S10 = 1/2 (Y1 - 'Y2 + 'Y3 'Y4) s 1 J = 1/2 (,B 1 - (3 2 + Figure 110 13 3 -(3 4) Internal Symmetry Coordinates and Frequency Assignments for M--C--0 Bending Modes 27 1-'638- (E),e . s 16 D 1/2 <P 1 + /3 2 - 1-1604 - (E)y sl 7 = 1/2('l 1 + r 2 - (3 3 - f3 4) V545-(E)¢ s1a = 1/ 2 (-¢ 1-¢2> Figure llo Internal Symmetry Coordinates and Frequency Assignments for M--C--0 Bending Modes and M--Br Stretching Mode r 3 -14> 28 Z/470- (A1) d S4 = do 0 Z/381 - (Al)di s 5 = 1/2(d1 + d2 + d3 + d4) Z/414 __..(Bl) di 11417 ~ (E) di s 11 a 1/2(d1 - d 2 + s 19 • 1/2(d1 + d2 - d3 - d4) Figure 11. Internal Symmetry Coordinates and Frequency Assignments for M--c Stretching Modes d3 - d4) 29 ;v116 - - (Al )il/'-8 V 99 __. (Bi )'tf-o S7 • 1/2 2 ('1J!1+¥2+1"3+1P4-S.s-~ S12 a: 1/2 2 (~1-1t~'¥'3- 06-&'7-Sa> Jb4-c55+o6-o7+d'3) V129 ~ (B2)ct. V133 --- (E)"f S14 • 1/2(oc 1 ~-~ 3 + S20 m 1/2(~1 +1?2 - cx:.3'1- Figure llo -cxiJ 'V-'3 - 'Y'4) Internal Symmetry Coordinates and Frequency Assignments for C--M--C Bending Modes JO VB7 - - (E)Ol s21 = 1/../2 COl12 - Ol34) V53-(E)o S22 = 1/2(61 + S2 - 83 - 64> Figure 11. Internal Symmetry Coordinates and Frequency Assignments for C--M--C and c--M-~Br Bending Modes (A 1 + E) I 1g Ff I Figure 12. (cm- 1 ) I 50I 400 I I (A2 + E) Fr (cm-1) Correlation Diagram for Cr(C0)6 and Mn(Co) 5Br (B2 + E) (B1 + E) .600500 Fr I .I Cr(C0)6 C4v Symmetry-- Mn(Co) 5Br 700 I Oh Symmetry-- (B 1 + E) ~ I 100 ~ ~ (B 2 + E) (A 1 + E) e4v Symmetry-- Mn(Co) 5 Br I 150 F2u Cr(C0)6 F2g Flu Oh Symmetry-- .... \..J .32 CHAPTER J ISOTOPIC SUBSTITUTION AND THE POTENTIAL FIELD OF Mn(C0)5Br Instrumentation The vibrational spectra of Mn(Co) Br and the other 5 systems discussed in Part II were obtained with four instruments• I~ R~ spectra in the range JOOO to 200cm-1 were produced on Perkin-Elmer grating spectrometers #225 and #521. Most of these data were reproducible within the limit ± ~Scm-1~ Similar accuracy (except for very weak or poorly resolved bands) was obtainable for the Raman spectra with a Cary 81 monochromator with a 100 mw. He-Ne laser (6328 A.) as an exciting source. Far I. R~ spectra were obtained with an accuracy of 1 on a Beckmann IR-11 grating spectrometero N ± lcmIsotopic Substitution It has been found that the potential field of a given ·m olecule will be invariant to isotopic substitution. This has been a tremendous ai~ in determining the potential fields of the simpler moleculesJ a simple change in the kinetic energy expressions to account for the presence of heavier or lighter nuclei together with 33 observed frequency shifts will furnish additional equations to solve for the unknown potential constants in the vibrational secular equation. When as large a molecule as Mn(C0)5Br is being dealt with, the number of independent potential constants has risen to 78. There are, however, only 22 equations (from the 22 fundamental frequencies) to solve for these 78 constants, and without additional information the situation would be desperate to say the least. Fortunately in the three summers I have spent at. the Los Alamos Scientific Laboratory (LASL) I have had 95% Mn( 12c180)5Br (prepared by Mr. Maxwell Goldblatt at LASL) and 93% Mn( 1 3c 160)5Br (prepared by myself at Caltech from Mn2( 1 3c 16 0>10 on loan from LASL) made available to me during my research program. It was attempted to obtain both I. R. and Raman spectra for these molecules in the common solvent, CH2c1 2 • This proved impossible for the weaker Raman bands and in the far Iu R. where CH2Cl 2 absorbed too greatly (and where the compound also proved to be insufficienty soluble in non polar CCl4 to show a spectrum). In these cases spectral data from the polycrystalline solid and nujol mullg respectively, were usedo These data with their approximate error limits are presented in Tables lg 2 9 and J~ 1 not obsvd. 374.8 ± .2 not obsvd. not obsvd. not obsvd. 462. 3 + • 2 not obsvd. not obsvd. 408e5 ± 1 417.2 ± 1 422eJ ± 2 not obsvd. 545.5 ± .2 604.2 + .2 380 .. 5 ± .3 not obsvd. not obsvd. not obsvd. 470e4 ± .4 not obsvd. not obsvd. not obsvd. not obsvd. not obsvd. not obsvd. not obsvd. not obsvd. obsvd. wkly. 527.1 ± .2 591.3 ± 1 not obsvd. 622.0 ± .7 632 .. 0 + 1.5 1962.6 ± .2 not obsvd. 605 ± 2 not obsvd. 631 ± not obsvd. 638.4 ± .7 645.0 + 1.5 not obsvde not obsvd. 2007.0 + i.O 2007.J ± .2 11962.0 ± .4 2 459.8 ± .6 not obsvd. 1961.8 ± .2 6410 3 ± 1.5 636.2 ± • 7 not obsvd. 600.8 ± .2 542.8 ± .2 no_t obsvd. 416.1 ± 2 411.2 ± 1 402.7 ± 1 I not obsvd. not obsvde not obsvd. not obsvd. Infra-red 366.5 + • 2 220.6 ± .5 Raman Mn(12c180)5Br not obsvd. not obsvd. 411.7 ± 1 396.8 ± 1 not obsvd. 221.6 ± e5 Infra-red 223e8 ± 1 Raman 5 Mn( 1 '.3c 16o) Br 222.4 ± 1 Mn( 12c 16o) Br 5 Raman Infra-red Observed Intense Raman and Infra-red Bands for Mn(Co) 5Br in CH 2c1 2 Solutions TABLE I ""-t:" not obsvd. 2088.2 + .3 2107.6 ± .2 2088.9 ± .6 not obsvd. 2137.7 ± .2 2137.9 ± .2 not obsvd. not obsvd. 2090e6 + e J not obsvd. 2038.3 + .2 not obsvd. 2085.4 ± .2 2038.1 + .4 not obsvd. 2053.1 + • .5 not obsvd .. not obsvd. not obsvd. not obsvd. Raman 2007 .. 0 ± .2 Infra-red not obsvd. 2090.6 ± .2 not obsvd. not obsvd. 2004.o ± .2 Infra-red 5 Mn( 12c 18o) Br not obsvd. Raman 5 2052 .. 2 ± .2 Infra-red . not obsvd. Raman 5 Mn(12ct6o) . Br 1 continued Mn( 13c 16o) Br TABLE \.a) \J\ TABLE 2 Observed Intense Raman Bands for Solid Mn(Co) Bra 5 Mn(12c160)5Br Mn( 1 3c 16 o) Br 5 Mn(12c180)5Br 52.9 ± .1 not obsvd. 51.0 ± .2 99.0 ± .1 98.5 ± 95.6 + .1 115.8 + .6* 109.9 ± .6* 124.3 + 1.5* { not resolved .1 J 116.9 ± 1.5* 122.8 ± 2.0* 129.6 ± 2.0* by hand 218.5 + .1 217.4 ± .1 216.6 ± .1 383.9 ± .1 378.1 ± .1 370.4 + .1 407.0 ± .4 395.5 ± .4 389.5 ± 1.5 414.3 + . .2 407. 7 ± .3 397.2 ± 427.1 ± .3 418.5 ± .5 410.1 ± 1.0 473.2 ± .1 464.5 ± .1 462.1 ± .1 537.4 ± .2 517.4 ± .2 534.7 ± .8 - o3 obsvd., wkly11 527.8 ± 06 obsvd., wkly. 626.8 ± o2 613.,4 ± .1 623 .. 2 ± .3 635.4 ± .2 62402 ± .,2 632.2 ± ~3 647.4 ± .3 634.2 ± .3 644.0 ± .. 6 a * denotes bands lying close together which were resolved by hand. 37 TABLE 3 Observed Far I. R. Bands for Mn(12c16o) Bra in Nujol Mull 5 1 'I/ (cm- ) Intensity 133.0 medium (broad) 121.1 medium 114.J weak 101.7 weak (broad) 87 weak 53 very weak a Bands were obtained only for the plentiful normal species due to the exorbitant amount of sample necessary for each mull. J8 Two of the more interesting isotopic effects are illustrated in figures 13 through 18J the first three drawings indicate the different behavior of the two bending modes around 640cm-1. These were assigned as the out-of-plane (A 1 )r bend and the in-plane (E) 19 bend in order of decreasing energy. Both these bands shift more with 1Jco substitution than with c18o substitution; this is a general trait of MCO bending modes for all metal carbonyls (10, 11 ) and is a consequence of the large motion of the carbon atoms in the bending movement relative to the oxygen atoms. However, it is also noticed that the splitting of these bands is greatest for 1Jco substitution and least for c18o substitution. As both shifts occur to lower energy, this indicates that the (E)~ band has more MCO bending character than does (A 1 )y• It is also general behavior that M--C stretching modes shift further upon c18o substitution than 13co substitution (10)14)w Thus one immediately gains the information that the (A 1 )~ vibration is more highly mixed with M--C stretching modes (mainly (A 1 )d 0 at 470cm-1) than (E)~ is, say with (E)di at 417cm-1e The next three spectra show that the splitting 650 Figure lJ . 645 635 630 625 1 V f (cm- ) 620 615 610 I.R. Spectrum of Mn( 12c 10 o) Br in CH 2c1 2 Solution 5 640 '° \,.) Cl ~ 1.: 650 Figure 14. 645 635 ~.JO VI (cm- 1 6?5 6?0 6i5 610 I. R. Spectrum of Mn( 1 .3c 16 o) Br in CH 2Cl2 Solution 5 640 -'0=" 0 ~ I- 650 Figure 15. 645 615 620 615 610 12 c18 o) Br in CH c1 Solution 2 2 5 625 l/ I (cm- 1 ) 630 I.R. Spectrum of Mn( 640 ....~ Figure 16. 430 ll (cm-1) 410 400 I.R. Spectrum of Mn(Co) 5Br in CH 2c1 2 Solution 420 N ~ 0 ~ ~ Figure 17. 430 V (cm-1) 410 400 I.R. Spectrum of Mn(c 1 3o) 5Br in CH 2c1 2 Solution 420 ~ \....) 0 ~ I- Figure 18 ~ 430 v(cm-1) 5 410 400 I.R. Spectrum of Mn(co 18 ) Br in CH 2c1 2 Solution 420 ~ ~ 45 of the bands around 420cm- 1 is also greatest for 13co substitution. As the band at higher energy, (E)ai exhibits the greatest shift on c18o substitution, it is correspondingly assigned as an M--C stretch. The band at lower energy shows the opposite behavior and it is tempting to assign it as an MCO bending f'undamental11 However, unlike the pair of bands in figures 13, 14, and 15, the band at lower energy is seen to lose intensity drastically as the splitting increases. This is exactly the behavior expected of a combination band near a fundamental of the same symmetry. Ordinarily this band would be forbidden in the I. R. spectrum, but an increasing amount of mixing (Fermi resonance) may occur as it approaches the energy a fundamental of the same symmetry. Thus we may explain this band as the combination .J/13 - V2o ( 537cm- 1 -lJJcm- 1 = 407cm-1)" As the combination must be of E symmetry to interact with (E)di' and as v 13 is of B2 symmetry, we have that Z/20 is necessarily of E symmetry (as only X = E holds for B2 x X = E) thus confirming the symmetry of the 13Jcm· 1 band 9 s assignment in Chapter 2. 46 Potential Field Calculations Excluding the inactive A2 mode we now have 21 x J = 63 observed frequencies. This will be adequate to determine at least fairly well the important constants in the General Quadratic Potential Field {GQPF). This potential field itself utilizes an approximation. The solution of the vibrational equation is in terms of a harmonic oscillator. Thus serious errors may be intro- duced into a solution if any anharmonicity effects are not taken into account. In reference {l3) it is shown by a comparison of the fundamental and combination bands that the C--0 stretching vibrations of Mn{C0)5Br are fairly anharmonic. From the observed binary and tertiary combinations the true harmonic fundamentals may be approximated {2l), and the results for Mn{Co) 5Br in CCl4 solution area l/ Observed Harmonic {A1)ri 2137.9 2158.7 (B1)ri 208504 208Jo8 (E)ri 2052.,2 20?8e5 2001.3 2030.8 (A1)r 0 The I. R. combination spectra for the solid in the regions 2600 to 2400cm-1 and 1200 to 65ocm-l are presented in Table 4 along with their assigned combinations" 47 These assignments result from fundamentals whose combinations possess either the I. R. active A1 or E symmetrye These calculated combinations are shown in Table 5. These combination bands may be fitted, with one exception (the very weak peak at 2445cm- 1 ) to within 15cm- 1 by combinations of observed fundamentals. The average fit is better than 5cm-1, and only 6 bands are fit more poorly than this. This indicates that as long as at least one component of a combination band is not a C--0 stretch, that band will exhibit good harmonic behavior. Therefore anharmonic corrections were made only for the c--o stretching frequencies. The corrections for the isotopically shifted frequencies were obtained from (j) Xkl x(i) (V(j)J/ (j) /H(i)II (i)) kl where z;(i) a = k 1 Vk 1 frequency k for isotopic molecule i k where x~t> = k-lth anharmonicity constant for isotopic molecule i These observed frequencies are then fitted by a potential field of the General Quadratic forme Starting values for this field were obtained by the transfer of the most nearly equivalent element from the M(C0)6 system (L.H. Jones» et alo (11) and L.H. Jones; private m vvw 652 764 w m m w(br) w w 879 I 899 I 940 I 1036 I 1055 I 1077 I I 11414 + V627 = 1041; (E)-I 114 14 + V6JS = 1052; (E) 2 I (V537) = 1074; (A 1 ) I 'V414 + 'V473 = 887; (E) I (V473) 2 = 946; (A1) V384 + ll64S = 1029; (E) vw 813 I (A1) I 11219 + v645 = 864; m 798 I m 788 I I V133 + 11645 = 778; (E) I V384 + V414 = 798; (E) I (V414) 2 = 828; (Ai)-- V219 + V6o4 = 823; (E) 761; (A 1 ) V133 + V627 = V116 + 11645 = V219 + V546 = 765; (E)-- V129 + V638 = 767; (E)-- = 768; (A1)-- 760; (E)-- 2 'V53 + 11604 = 657; (A 1 ) Combination( s); (!') V11 6 + J.S37 = 653; (E)-- I (V334) I Intensity I 11 4 Observed Sum Combination Infra-red Spectrum: Solid Mn(Co) 5Br in KBr Pellet TABLE -'=" co 4 continued V546 + V645 = 1191; (E) 1/2000 + V414 = 24 14 ; (E) V2050 + V384 = 2LtJ4; (E) V2088 + V414 = 2502; (A1,E) V2000 + V546 = 2546; (E) 'V2137 + V414 = 2551; (E) w vw vvw . w vvw w vw w w w w vvw 1168 1191 2419 2430 2445 2468 2488 2497 2506 2545 2555 2584 V2050 + V537 = 2587; (E) V2074 + V4 14 = 2488; (A 1,E) 'V2086 + V4.14 = 2500; (Al'E) V2050 + V414 = 2469; (A1,E) -- ? . V546 + '1627 = 1173; (E) Intensity v V199l + Vlt 7 J = 2464 ; (A 1 ) Combination(s); (f) TABLE '° ~ 887 543 633 798 828 517 603 768 129 133 219 384 B2 E Ai At 1083 1092 1074 ~ 5461 E 1019 960 930 765 679 537 951 667 675 667 1150 1173 1184 1175 1141 1191 1118 1059 1029 1022 1052 864 778 761 744 698 645 Al 857 771 767 759 754 737 691 638 E 1111 1041 760 756 74J 680 627 B2 1077 1018 988 823 737 733 725 720 662 B1 946 692 606 658 654 ?OJ 65? 604 E 645 599 546 E 473 547 589 590 537 B1 Al E,B 1 414 857 535 121 B1 5JO 500 116 E 572 51J 483 99 Al 526 467 437 473 414 384 53 Al E,B 1 A1 E 1/ 5 Calculated Sum Combination Infra-red Spectrum: Solid Mn(Co) Br in KBr Pellet 5 TABLE I \I\ 0 B1 Ai 2587 2521 255i 2610 2502 20881 21371 2623 2500 2086 j 2625 261i 2488 (074 2523 2464 2434 2050 E 2473 24.14 2384 2'683 2634 2632 2620 2596 2546 2537 2535 2462 2403 2464 2532 2459 2400 2405 20001 At 2375 2373 1989 ·{~1991 2370 2741 2692 2690 2678 2654 2604 2593 2593 2590 2677 2775 2726 2724 27i2 2688 2638 2629 2627 2624 2782 2695 2645 2636 2634 2631 1290 645 Ai 1986 1283 638 1249 645 E 1265 1242 638 Ai 1254 627 E 627 604 B2 B2 546 E ·~1231 537 E 604 473 414 384 B1 5 continued E l/ At E,B1 Ai TABLE ~ \./\ 52 communication). Calculations were done by a least- squares fitting computer program originally written by Schachtschneider (24) and since modified by Jones and Ryan at LASL. The problem was then solved on the IBM 360/375 at Caltech and the CDC66oo machine at LASL. The problem was formulated according to the FG method of Wilson, et al. (~5). This method presents the vibrational problem in matrix form, and these series of equations are than solved numerically via high speed computer. Two matrices, .£: and Q, are responsible for the name of the method. Eriefly, the F matrix contains the force constant for each symmetry coordinate (which were presented on pps. 25-30) along the diagonal and the interaction force constant between two symmetry coordinates in the appropriate off-diagonal element. The G matrix is often referred to as the "inverse kinetic-energy" matrix. This is indeed the form of this array in a Cartesian coordinate systemJ however, with internal molecular vibrations a solution in terms of internal coordinates is much more desirable than one in terms of an externally imposed Cartesian .s ystem. We have this situation for the F matrix as the symmetry coordinates are simple combinations of the internal coordinates shown in figure 2. Thus the next step is 53 the unenviable transformation of the G matrix from its Cartesian formulation to symmetry coordinates. This is also accomplished by a computer program by Schachtschneider (~i) and the result is symmetry blocks along the diagonal which correspond to the symmetrized F matrices. The least-squares program forms the product EQ. which in turn is diagonalized and solved. The result- ing eigenvalues are directly related to the frequencies of vibration; the observed frequencies are compared to the calculated frequencies and from the error encountered, a perturbation is induced in each unconstrained force constant and a new calculation is begun. These cyclical calculations continue until a convergence limit selected by the user is reached (specifying either the size of errors in the calculated frequencies or in the size of the perturbations of the force constants for each cycle) or until the problem becomes divergent. The calculations done for Mn(C0)5Br, however, did not involve the above mentioned force constantsa The elements of the F matrix were replaced with compliance constants which are described in (i6). The reason for this departure from the usual procedure in vibrational analyses is that compliance constants 54 are uniqueiy defined for a given molecule while the values of the force constants depend on the internal coordinate system chosen. This is clearly shown in comparison calculations for the H20 and n2o molecules in ('J-7). Furthermore, if redundant symmetry coordinates exist for the molecule in question, additional relationships may be obtained between the compliance constants involved, For example the values of Co.-:, Ctk, C~, c::x_, and Cd,d are uniquely determined for M(C0)6 (L.H. Jones; private communication} while only the force constant F oc - FJ;_', and F~ - Fclct. may be combinations Fd.. - F~, calculated (1~}. As starting values for these calcu- lations utilized either the Cij or Fij from the hexacarbonyl calculations, it was decided to solve the problem in terms of the more uniquely determined compliance constants. The relation between the compliance and force constants is quite simply Q = f- 1 o Thus after a set of compliance constants was obtained as a solution for a given symmetry block, the C matrix was simply inverted to obtain the corresponding force constant solution@ The "kinetic-energy" matrix, ~, which corresponds to the G matrix is simply li = Q- 1 • Throughout the remainder of this thesis the units of all Cij will be 55 angstroms/millidyne, and similarly the units of all Fij are millidynes/a:ngstrom. A1 Symmetry Block Calculations This block contains 7 frequencies and 28 independent compliance constants. Isotopic data raised the number of observed frequencies to 21; however, the problem remained convergent only if the 15 constants (or less} in Table 6 shown with standard deviations were allowed to vary. Although the degree of con- vergence was rather low, the frequencies (also in Table 6) are seen to be fit quite well. The potential energy distribution for each vibration among the various Cij is shown in Table 7. These figures show the nature of the normal vibration coordinates.· For instance, as was predicted on page 38, there is considerable mixing of the (A1)r and (A 1 )d 0 modes. The potential energy for Vr is seen to be 60% c3, 3 and 15% C4 1 4 while the distribution for Vdo is 22% c3, 3 and 70% C4 1 4. The calculations also indicate that the only mixing of any consequence with the 0 Mn--Br" stretching mode is with the C--M--C deformation mode. It is therefore expected that this will also be the case for the "Mn--Mn" stretching mode in Mn2(C0)10 examined in Part 6 c 6, 6 = .. 84J ± .,23 c 7, 7 = 1"784 ± .97 c 1, 2 = -.0016 c 1, 3 = 0 c 1, 4 = .,0044 c 1, 1 = .. 05532 + .. 00029 c 2, 2 = .,06057 ± e00024 c 3, 3 = 1 .. 394 + .. 60 c 4, 4 :::: .. 437 + .. 43 c 5, 5 = .. 413 ± .. 017 Calculated C i , j ± Standard Deviation 0 + e 00 30 c 2, J = .0106 c 2, 4 = -.0186 + .. 0063 c 2, 5 = .0044 c 2, 6 = 0 c 2, 7 = .0107 c 3, 4 = .0775 ± .42 c 1, 7 = -.0107 c 1, 6 = c 1, 5 = -. 0066 Calculated C i , i + Standard Deviation Aa c 6, 7 = -.241 c 5, 6 = 0 c 5, 7 = .14 c 4, 6 = -.088 c 4, 7 = -.233 c 3, 5 = 0 c 3, 6 = - .157 c 3, 7 = -.045 c 4, 5 = 0 ± .23 ± .88 ± .27 + .53 ± .68 Calculated C i , ~ ± Standard Deviation Compliance Constant and Frequency Calculation for ~in(CO)~Br: Symmetry Block (with Z/7 ssigne to 115.8 cm-1) TABLE °' \J\ 221.6 114.o -.1 1.0 222.5 114.8 222.4 115.8 -·9 114.4 221.5 374.4 374.8 0 380.5 380.5 .1 462.2 462.3 -.1 470.5 470.4 .3 631.7 632.0 645 .. 9 645.0 -.4 el .4 .2 1984.o -·3 1984.2 2031.1 2030.8 .1 ~ 2108 .. 1 Vealed. Mn(13c160)5Br 6 continued 2108.2 -.1 2158.,8 2158.,7 V'obsd., /),, Vealed. Vobsd. Mn(12c160)5Br TABLE 109.9 220.6 J66.5 459.8 641.3 1985.0 2110.,6 Llobsd. 110.6 220.6 366.9 459.8 640.8 1984.9 2110 .. 6 -,Vealed. 0 .5 .1 0 /),, -·7 0 -.4 Mn(12c180)5Br \J\ ......., 58 TABLE 7 Potential Energy Distribution for A1 Symmetry Block Calculation (with v7 assigned to 115.8 cm-1) <'requency 645 470 2159 2031 (Approximate (CO- (CO- (MCO) (MCax. Jescription) rad.) ax.) 222 380 (MCrad.) (MX) 116 (CMG) c i, j c 1, 1 .90 .07 0 0 .03 0 0 c 2, 2 c 3, 3 c 4, 4 c 5, 5 c 6, 6 c 7, 7 c 1, 2 c 1, 3 .06 .94 .01 .01 0 0 0 0 0 .60 .22 .01 0 .20 0 .05 .15 • 70 .02 .03 .19 .o4 0 0 0 .96 0 .02 0 0 .03 .04 0 .99 .06 0 0 .21 0 0 .15 .83 .01 -.01 0 0 0 0 0 0 0 0 0 0 0 0 c 11 4 0 0 0 0 0 0 0 c 1, 5 c 1, 6 0 0 0 0 .01 0 0 0 0 0 0 0 0 0 c 1, 7 0 0 0 0 0 0 0 c 2, 3 c 2, 4 c 2, 5 c 2r; 6 0 0 -.01 0 0 0 0 0 ~.05 1101 .. 01 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c 2, 7 0 0 0 0 0 0 0 59 TABLE ? Frequency (Approximate Description) c i, j c 3, 4 c 3, 5 c 3, 6 c 3, 7 c 4, 5 c 4, 6 c 4, 7 c 5, 6 c 5, 7 c 6, 7 continued 222 645 470 J81 (MX) (CO- (CO- (MCO) (MC- (MCax.) rad.) rad.) ax.) 2159 2031 116 (CMC) 0 0 -.04 .06 0 0 -.03 0 0 0 0 0 0 0 0 0 -003 • 0.3 0 0 -.03 0 0 .03 0 0 0 -.03 0 0 0 .01 -.02 0 .01 0 0 -.02 -.06 0 .07 -.04 0 0 .11 -.02 0 -.04 - .. 24 0 0 0 0 0 0 0 0 0 .01 0 -.02 0 -.05 0 0 -.04 0 0 -.19 .11 60 II as this frequency lies even closer to the C--M--C bending mode than the analagous Mn--Br stretch. Finally, equivalent force constants obtained from Q-1 are shown in Table 8, A 2 Symmetry Block Calculation As this frequency was not observed, only an estimate of ca,8 in terms of interal coordinate compliance constants• t c c 8, 8 = cp + 2c~p + op~ From the values in Table ~o this gives Ca,8 = 1.85 (Fa,8 a 1/1.85 = .54) yielding a calculated Va = 449cm· 1 for Mn( 12c16 0)5Bre Bands of A2 symmetry may combine with other A 2 bands or with E bands to give combinations of At or E symmetry which are expected to have the greatest I. R. intensity. These combinations in the energy region covered in Table 5 are predicted to bes 863, 986, 995, 1053, 1076, 1087, 2499, 2523, 2535, and 25J7cm- 1 • These frequencies either are not observed or do not fit an observed band uniquely. Thus the observed combination sp'e ctra neither confirms or denies this frequency prediction~ Further experiments are neces- sary at low temperature in order to reveal some of the weaker combination bands. This assignment for VA 2 61 TABLE 8 Equivalent Force Constants for A1 Block Calculation (V7 = 115.8 cm -1) = 18.144 F 1, 5 = .260 = 16.792 F 3ll 3 = .740 F 4, 4 = 2.622 F 5, 5 = 2.504 F 1, 6 = 0 F 3, 6 F 1, 7 = .07 F 2, 3 = -.16 F 4, 5 F 1, 1 F 2jl 2 F 2, 4 :: .758 F 2, 5 = -.18 F 6, 6 = 1.318 F 7, 7 = .663 F 2, 6 = .05 F 1, 2 = .408 F 2, 7 = .02 F 1, 3 = 0 F F 1, 4 = -.13 J, 4 = -.0970 F 3, 5 = 0 = .135 F 3, 7 = . 0260 = -.14 F 4, 6 = .369 F 4, 7 = .396 F 5, 6 = -.08 F 5, 7 = -.22 F 6, 7 = .236 62 would be greatly substantiated by the presence of (VA 2 >2 around 998cm- 1 as the next nearest predicted combination is 10cm-1 away. B1 Symmetry Block Calculation Good convergence was obtained for this calculation and the calculated ci,j and frequencies are shown in Table 9. Only five of the ten constants could be varied for the calculation to remain stable. c 10 , 12 was adjusted through a range of +.2 to -.8 in steps of .1, and the best fit was obtained for C10,12 = -.5. The fit of v10 is rather poor, the standard deviation being± 2.?cm-1 for the three isotopic species. A calculation assigning this frequency to (B2~ was fit equally poorly and the reason for this is not known. This frequency is a moderately intense Raman band and its position was determined with good experimental precision ( ± o2cm- 1 from Table 2)o The interaction constant between (B1)r and (B1)di also had little effect on the quality of the frequency fittingo Better agreement may possibly be obtained by further very careful refinement of the potential field. + 00055 + .093 c 11, 11 = • 361 C12,12 =2.153 95.1 95. 6 -.1 98.6 98.5 .5 99 .. 5 99.0 ~ 399.1 .397.2 1.0 406.7 407.7 o7 413.6 414.3 - .5 -1.9 1.2 622.0 623.2 2.1 611.3 613.4 -3.4 630.2 626.8 .1 2036.4 20 36 . 5 .4 2036.3 2036.7 -·5 2084.3 2083.8 calcd. calcd. obsd. Mn(12c180)5Br obsd. Mn( 1 3c 16 o) 5Br 0 C11,12 = .07 C10,12 = -·5 = -.0198 + .0043 = -.0053 C10,11 = = -.0053 Calculated C i, j + Standard Deviation cal ed. C9,12 C9,11 C9,10 Calculat ed C i, j ± Standard Dev i ation obsd. Mn(12c16o) 5Br ± ,,0031 C10,10 = e945 C 9, 9 = 005758 ± e000 36 Calculated C i, j ± Standard Devi ation Compliance Cons tant and Frequency Calculation for Mn(C0)5Br: B1 Symmetry Block TABLE 9 °' \...) 64 The other observed frequencies are fit quite well and the Potential Energy Distribution and Equivalent Force Constants are presented in Tables 10 and 12. B~ Symmetry Block Calculation Amazingly enough a stable solution for this small block with 2 frequencies and only J compliance constants could be obtained only by constraining 013,14• As in the case C10,12 above, this constant was varied through the range +.5 to -·5 in steps of .1, and the best value was thus determined to be -.1. A good fit of the frequencies was obtained, and this along with the calculated Cij and the potential ) energy distribution are shown in Table 11. Equivalent Fij are shown in Table 12, E Symmetry Block Calculation This was by far the most difficult and least satisfactory of the calculations. To start with, the three isotopic species give a maximum of 3 x 8 = 24 frequencies to determine the 36 independent constants. Another major problem was that the three M--C--0 bends lay within lOOcm-1 of each other; the mixing of the internal symmetry coordinates shown in figure 2 is thus likely to be quite large and the perturbation TABLE 10 Potential Energy Distribution for B1 Symmetry Block Calculation Frequency 2084 627 (Approximate MCO CODescription) rad. 414 MCrad. CMG 100 c i, j c 9, 9 1.01 0 .01 0 C10,10 0 1.13 0 .01 C11,11 .05 0 .96 .01 C12,12 0 .09 0 1.06 c 9,10 c 9,11 c 9,12 0 0 0 0 -.06 0 .03 0 0 0 0 0 C10,11 0 0 0 0 C10,12 0 -.22 0 -.06 C11,12 0 0 0 -.02 11 537.4 129.6 obsd. 537.2 130~4 ~.8 .2 1.0 -1.0 534.7 122.8 533.7 123.8 -1.2 -1.2 517.4 129 . 6 518.6 (12 9 .8) calcd. cal ed. obsd. Mn(12c180)5Br C13, 14 = -. _1 Calculated Ci,j ± Standard Devia tion obsd. Mn(13c160)5 Br Mn( 12c16o) 5Br cal ed . 014,14 = 1.696 ± .09 Calculated Ci,j ± Standar d Dev i ation C13,13 = 1.692 + .032 Calculated Ci,j ± Standard Deviation Compliance Constant and Frequency Calculation for Mn(Co) 5Brs B2 Symmetry Block TABLE °' °' 67 TABLE Frequency (Approximate Description) c l, . 11 contd. MCO 537 1.30 CMC .75 • 70 • 31 .25 -.05 .05 j C13, 13 014, 14 C13, 14 r 68 TABLE 12 Equivalent Force Constants for B1 Block Calculation F 9, 9 Fto,10 = 17.713 = 1.209 Fll,11 = 2.846 F12,12 = .533 9,10 = .12 Fl0,11 = -.05 F 9,11 = .964 Fl0,12 F 9,12 = .04 F11,12 = .283 = -.1 F Equivalent Force Constants for B2 Block Calculation .593 69 of one compliance constant therefore may be quite complex and poorly determined in its effects. The geometry of symmetry coordinate s 16 (corresponding to (E)fi) very closely resembles the form of the in-plane F1u M--C--0 bend for M(C0)6 (1~). For this reason and the fact that the observed frequencies are quite close together, 1.1~ is assigned with a high degree of certainty to the intense I.R. band at 638cm- 1 • As was discussed in Chapter 2, the assignment of (E)r and (E)~ to the I.R. bands at 545 and 604cm-1 is considerably more uncertain and there should be considerable mixing between these modes. Two series of calculations were done to ascertain the assignment of these modes. The first assigned ~ --.6J8cm- 1 , l/9' ~604cm-1, and Lly --.545cm- 1 • This ordering gave almost no mixing between (E)r and (E)¢ , however, and a large amount of mixing of (E),8 with (E)95 both of which results were unexpected~ Furthermore, c. the internal coordinate compliance constants Cyr and c; were very different from the analogous constants for 11 the M(C0)6 system (1~)o The second assignment was 'Vf3-+638cm- 1,1.lr -.604cm- 1 , and 'll? --.S45cm-1a it produced values for c;r and C~r which agree much more closely with their counterparts in the hexacarbonyls (.112 compared to ~24 and -.02 C16,17 C16,18 + .035 0584 016,20 C16,21 C16,22 017,18 C21,21 = 5.2 022,22 = 3.2 C15,16 = -.01 015,17 = C17,19 C17,20 C15,18 = .007 c 15 , 19 = -.0284 ± .013 0 C16,19 C20,20 = 1.5 C19,19 = C15,22 C1 7, 1 7 = 1. JO 3 C15,21 C16,16 = 2.234 ± .15 ± .081 ± .40 C15,20 019,21 .06 .75 = = .46 C21,22 C20,22 .2 = = C20,21 -.1 = C19,22 C19,20 018,22 .06 .16 = = C18, 21 C18,20 = 0 = -1.04 + .23 017,22 C17,21 C18,19 0 0 0 = -1.05 = -.13 = -.06 = 0 = 0 = .14 = -.9 = -.62 = -.18 = -.11 = = -.46 Calculated C i, j ± Standard Deviation = -.011 = = Calculated C i, j + Standard Deviation C15,15 = .05741 ± .00030 Calculated C i, j + Standard Deviation C18,18 = 2 . 299 13 Compliance Constant and Frequency Calculation for E Symmetry Block TABLE -..J 0 126.7 411.2 ---------- 2.3 ----- --- 409.4 132.1 86.7 52.4 411.7 ------------- -1.6 .2 -.J .1 418.8 132.8 87oJ 52.8 417.2 133.0 87 .. 0 52.9 51.0 539.9 411.8 542.8 529.8 -2.7 527.1 -·5 546.o 545.5 51.1 84.o -.6 600.2 600.8 1.5 589.8 591.3 -1.5 605.7 604.2 -.1 --- --- - .. 6 2.9 1.9 634.3 636.2 623.7 -1.7 622.0 -.8 639 .. 2 638.4 .1 2029.6 2029.7 .5 2032 • .3 2032.8 -.6 2079e1 2078.5 6. Vealed. L'obsd. 6. Vealed. Br Vobsd. Mn( 12c 18o) ti Br Vealed. Mn( 1.3c 16o) l-'obsd. Mn(12c16o) Br TABLE 13 continued ..... -'1 0 0 .24 0 0 0 .02 .11 0 0 .11 0 0 0 0 .o4 0 0 0 0 0 0 018,18 C19,19 C20,20 C21,21 022,22 C15,16 C15,17 C15,18 0 0 .30 .16 0 .01 .02 .02 .01 .30 0 0 0 0 .02 .08 .69 .11 .28 .59 .32 .01 0 .14 0 417 (MCrad. .03 C17,17 .01 .51 0 (/:>-) 546 (MCO 0 0 }'-) 604 (MCO C16,16 0 8-> 638 (MCO 1.02 2079 (COrad. C15,15 c 1, j Frequency (Approximate Description) Potential Energy Distribution for E Symmetry Block Calculation TABLE 14 0 0 0 0 0 0 0 0 .57 .40 .46 .06 .05 .66 .01 0 .26 0 a.-) 87 (CMC .62 0 .42 0 ~MC -) 133 0 0 0 .77 .92 0 0 0 0 .08 0 8-> 53 (CMX N ....., (3-> <P-> y-> -.09 0 -.03 .07 0 .01 004 .01 -.02 0 0 002 .11 -.19 0 0 -.01 008 -.04 .01 0 .16 0 002 0 0 0 0 0 0 0 0 0 0 0 0 C15,22 C16,17 C16,18 C16,19 C16,20 0 16,21 C16,22 c 17, 18 c 17, 19 c 17, 20 0 -.03 -.09 -.21 .01 .01 -.01 .01 .01 .03 .01 .06 -.11 -.03 .02 - 006 -.01 -.01 -.03 .01 -.15 -.03 -.46 .01 0 0 0 0 0 \fl-> 133 (CMC 0 0 . 01 417 {MCrad•) 0 0 0 C15,21 0 0 0 C15,20 0 0 0 546 (MCO 604 (MCO ..,..06 rado) 638 (CO- (MCO 2079 C15,19 c i, j Frequency (Approximate Description) TABLE 14 continued -.01 0 0 -.26 -.08 .10 -.02 -.04 0 0 0 0 -.16 .07 0 0 .01 0 0 0 0 0 0 0 8-> 53 (CMX 0 0 a-> 87 (CMC --.J \,.o.) 0 0 .02 e03 0 0 0 0 -.02 0 0 0 e10l 0 0 .01 0 eOl .01 0 0 0 0 0 0 0 0 0 0 C18,20 C18,21 0 1s,22 C19,20 C19,21 C19,22 0 20,21 0 20,22 0 21,22 .01 eOl 0 0 -.01 0 .01 0 .02 -.02 -.01 .01 C18,19 0 0 0 -.01 C17,22 0 0 4 J -) y-> {3-> $co 546 604 (MCO 638 (MCO 0 I (COrad.) 2079 C17,21 ci Frequency (Approximate Description) -.01 .01 -.03 0 -.68 .38 0 -.OJ -.44 .05 .03 0 -.34 -.02 0 0 -.03 0 0 .01 -.01 -.02 -.03 -.03 0 0 .... 01 0 0 0 0 8-) (CMX 53 0 0 0 0 a-> 87 (CMC .01 0 -.06 -.01 0 0 tj;-) 133 (CMC . 01 0 0 .01 .02 0 -.01 417 (MCrad.) TABLE 14 cont inued .{:"' ~ 75 TABLE 15 Equivalent Force Constants for E Block Calculation F19,19 = 17.879 = .649 = .930 = .565 = 1.913 F20,20 = F21,21 F15,20 = .07 F17,21 = .02 F15,21 = 0 F17,22 = -.05 F15,22 = .037 F18,19 = .031 F16,17 = .06 F18,20 = F16,18 = .275 F18,21 = .011 1.011 F16,19 = -.10 F18,22 = -.04 = .260 F16,20 = -.12 F19,20 F22,22 = .459 F16,21 = -.05 F19,21 F15,16 = -.006 F16,22 = -.18 F19,22 = .07 = -.02 = -.05 F15,17 = -.16 F17,18 F20,21 = F15,18 = -.014 = .914 F17,19 = .. 030 = -0325 F20,22 = .290 F17,20 = -.291 F21,22 = .140 F15,15 F16,16 F17,17 F18,18 F15,19 .035 .210 compared to -.18). The potential energy distribution shown in Table 14 is much more physically reasonable. The in-plane (E)i8 mode is now relatively unmixed except with the (E)~ in-plane C--M--C deformation. (E)'Y is now 32% C17,17 and 24% o1a,18 while (E)ip is 59% C17,17 and 30% 018,18• The modes are so highly mixed that a more accurate set of symmetry coordinates would be {s 17 + s 1a)/-/2 and (s 17 - s 1a)/...J2 • The next major difficulty encountered was with the fitting of the two C--M--C bending modes and the C--M--Br bending mode all of which lie within 80cm-1 of each other. These modes are even more highly mixed than the M--0--0 bending modes and it was necessary to constrain all compliance constants involving them for the calculation to remain stable. Constants 020,20• 021,211 022,221 020,211 020,221 C21,221 and 016,22 were adjusted first individually to determine their effect on the frequency fitting and potential energy distribution and finally "en masse" to give the set which best fit the observed frequencies. Naturally this procedure was extremely tedious and time consuming (not to mention being frustrating and expensive), and the "best values" of the compliance constants finally obtained represent only one possible solution. 77 If time had permitted, a more reasonable approach in the long run would have been to form combinations of the original E block symmetry coordinates which would more closely approximate the normal coordinates of vibration. After the calculation of a new G matrix, the compliance constant calculation might be considerably more stable as the different Cij would be less correlated with each other. The unscrambling of these values into internal coordinate constants then would be only slightly more difficult than it was in the present system. The calculated values for Cij and the frequencies are shown in Table 13. The fit obtained for the seven constants allowed to vary is only moderately good and it is felt that the adoption of the above procedure would significantly improve this calculation. Never- theless the present values are felt to be accurate enough to proceed with the calculation of the internal coordinate compliance constants. The equivalent sym- metry force constants are shown in Table 15~ Internal Coordinate Compliance and Force Constants The calculated values of the symmetry potential constants may now be transformed to internal coordinate potential constants by means of the formulas presented 78 in Tables 16, 17, and 18. These formulas are arrived at by standard group theoretical methods (i~ and 1-S) for a GQPF. The internal coordinate interaction con- stants are in the form Cij where i and j now refer to the respective internal (rather than symmetry) coordinates involved. Thus Cij is the change in the com- pliance of coordinate j when coordinate i is forced through one positive unit displacement. The calculated internal coordinate compliance and force constants are presented in Tables 19, 20, and 21J along with these values are the nearest equivalent elements of the GQPF for M(C0)6 with their calculated values (1~). Where the value of a given element for Cr(C0)6 was substantially different from the mean value of Cr(C0)6, Mo(C0)6, and W{C0)6, this individual value was used in the starting calculations and this is indicated in the tables by a double asterisk. Stretch-Stretch Constants The a-- and 'TT- synergic bonding model is again used to interpret changes in potential constants between M(O)(C0)6 and Mn(+l)(co) 5Br{-l), Lower case cij and f ij will be used when the hexacarbonyls are under discussion to distinguish between the Cij and Fij for the Mn compound. 79 TABLE 16 SYMMETRY AND INTERNAL COMPLIANCE CONSTANTSs STRETCH--STRETCH TERMS c 2 , 2 = cro c 4, 4 = cdo c 5, 5 = cdi + 20°di,di + ctdi,di 0 + ctd·,d1 C11,11 = Cdi - 2c d·,d· t l . l. l. 0 19,19 cdi - cdi,di = c 6, 6 = CA c 1, 2 = 2CCr1,ro 0 c 1, 4 = 2C ri,do c 1, 5 = cri,di + 2CCri,di + ctri,di c 1, 6 = 2C ri, A c 2, 4 = cro,do c 2, 5 = c lg 4 c 2, 6 = cr , A 0 t = 2CC c 4, 5 c 9, 11 = Ori' di - 2C~i,di + Cri,di d 0 ,di c 4, 6 = 0 d , A ct 0 C15,19 = cr·l. ,d·l. - ri,di c 6 a 2Cdi,A s" 80 TABLE 17 SYMMETRY AND INTERNAL COMPLIANCE CONSTANTSs BEND--BEND TERMS c 3, 3 s: c1 + 2c~r + cir C10, 10 = Cr - 2C~y + C~r t C17,17 = Ciy - Cr,Y c t .c 8 , 8 = cf3 + 2c~.~ + cf3,f3 C13, 13 = C,s - 2C~f3 + C~,13 0 16, 16 = 0ra - 0 l~ C18, 18 t = 0¢ - 01>,</J = C¢ c 7, 7 = c1 + 2c~~ + ci~ C12,12 = c~ C20,20 = c~ 0 21,21 - c~s ... 2c~8 - c~& 2C~1 + ~~ - C~s + 2C~$ - ci~ • t 0 ~~ = 06 - 0 ~-o C14 14 = eel - 2CA?cX.. + cY d.. ' C22,22 L-\,J V\.J = c~ - c!,ot.. c 3, 7 = ( 1/f2') (C"f;'lf + 2C~1> + c~?/1 - Cr,& - 2Ctb - c~~) c10 , 12 = <1/(2' ><c'Y,1 - 2c~11 + c~JI> - Cr;s + 20~ 6 - ci,s ) c13, 14 • c~f3 .. 2c;,<3 + c~>(3 C16, 17 t = cp,y - Cfo/Y C16, 18 = /2' Cp,p 81 TABLE 17 continued C16, 20 = 2 C(2>>'1' 0 16,21 = 20 (i3,& 0 16,22 = ~ (Cq(:i "' ci.~) 0 17,1a = -fi' 0 r,<P c17,20 = co/iv - c~'lf . c17,21 =. c'Y,8 - ctr,S c1 7 22 : ../2' (c'J'.;O! - 0J~ I c'1a, 20 = ~ c <t»'IJl C18:r 21 = -J2' Ocp,r;, 0 18fl 22 = 2 0 </>oL ) 82 TABLE 18 SYMMETRY AND INTERNAL COMPLIANCE CONSTANTSs STRETCH--BEND TERMS = Cri,y + 2Ccr ,'Y + ctri,r 1 c 1, 7 = : 16:f'2> <~ri,"1' + 2C~i,"t' + c~i,·"1' - Cri,S - 2C~i,S C 1, 3 ri,6 = c1/fl) <c~i,v + 2c~iJ1f + c~i•lf > c 2, 3 = 2Cro, 'Y c 2, 7 = ./2( 0 ro,'Y' - 0 ro,'f>) = ./2cro,'lfr-8 4 c 3, 5 c 3, 6 c 4, 7 c 5, 7 c 6, 7 c 9,10 c 91112 c 10, 11 0 11, 12 C15,16 C15,17 C15,18 = 2Cdo,'Y 83 TABLE 18 continued C15920 C15, 21 c 15, 22 C16,19 C17,19 C18,19 C19, 20 0 19,21 C19,22 84 TABLE 19 INTERNAL COORDINATE STRETCH--STRETCH COMPLIANCE AND FORCE CONSTANTSa c.J., J. Calcd. for ~in.(C0)5Br Approx. Equiv. Ave. Value for M(C0)6 Calcd. for M(C0)6b .05693(17.904) c00 .05905 **(17.04)** .06057(16.792) c00 .05905 **(17.04)** .499(2.294) CMC .509 **(2.10)** .437(2.622) CMC .509 **(2.10)** e843{ 1. J18) No -.00057(.108) C8o,c•o• -.00077(.175) -.00048(.025) cco,c•o• t -.00025(.039) -.0008*(.204) Cco,c•o• c -.00077(.175) .013(-.086) CMC,MC' c .002J**(-.02)** -.112(.381) CMC,MC' c Cooc,MC' -.114**(.47)** -.0193(.763) Cco,MC ·-.0238(.758) .0033(-.176) Cco,MC' t 0co,MC' c Cco,oocu t O*(O) e0091(-o141) .. 0022*(-0065) is set c Equivalent .002J**(-.02)** .0022(-.088) .0097(-.112) 00022(-.088) = -.0186(.758) -.0238(.758) a * denotes values held constant in calculations; values in parenthesees are equivalent force constants. b ** denotes values for Cr(C0)6; other values are averages for Cr(C0)6~ Mo(C0)6, and W(C0)6• Refo (14 )~ 85 TABLE 19 continued cr•, A 1 Cr o' A cdi'A cd ,A 0 O*(O) No Equivalent o-i~c 0) No Equivalent o·n"( o) No Equivalent -.088(. 363) No Equivalent 86 TABLE 20 INTERNAL COORDINATE BEND--BEND COMPLIANCE AND FORCE CONSTANTSa c.1, J. Calcd. for Mn(C0)5Br cf3 1.970( .. 598) 1. 919-i~*( • 48 )-!~* CY" 1.287(.953) 1.919**( .48)** c'¢ 2.299( .565) 1.919**(.48)** C-y--C"fb 2.149(.61) C"' -CJ.J'' 2. 62 ** ( • 56) .;~* Ci;-C~y. 1.5(1.011) ~ -c;.;_,, 2.62 **(.56)** c<f-c&; 5 .2 ( .26) eel 2 •67 (. 5 3) Approx. Equiv. Ave. Value for M(C0)6 Calcd. for M(C0)6b No Equivalent 1.92 ( .518) • 07* ( - • 023) -C'' ~fd .07(-.023) -.264(-.051) -.25(.068) .112(-.117) c;f3 c'µ(3'. -.017(.022) c• -.25(.068) -.18*(.01) C'ri.rJ..' ' '-C' ' 0(,0(. ~.18(.01) -.093(.04) C' «a -C'oc:O/.' ' - .44( 013) .24(-.029) ~(3 .22* <1!22 - ,,53 -.52 c{J't' -- cc(3"f' = ct(3'r = o No Equivalent a * denotes values held constant in calculations; values in parenthesees are equivalent force constantsu b ** denotes values for Cr(C0)6; other values are average values for Cr(C0)6, Mo(C0)6, and W(C0)6• Rei'. (1't)• 87 TABLE 20 continued c (3¢ - • 37( .134) -C'f3/J' ' -.25(.029) Cr~ -Clp -.0?*(.021) -.07(.023) C1f~ -C~i; -.9(.210) c •• ~/3 c,,, , -C, , c \ta( I -.22*(.11) c~~ -.22(.14) I -.25(.07) c;_e1. -.22(.14) Cra'l/- .03*(-.06) .03(-.03) C136 .03*(- .03) c. t o<.p ct •• cf~ .21(-.05) c:~ .55(-.11) c - • 26( .06) c~tt cf3o1. cl,1f-&+ctr,v--& -.20(.11) c Cry .08( .05) t .46*(-.291) cr"f'-crl/I cr& -c;s - • 46 "~ ( • 02) I O*(O) c,oc rld- clcx. t .03(-.03) cl~ - -.18**(.01)** -.09(.04) C'' ol~ c°'(3• +cd.(J •• ct .64(-.152) COlp• -C'' d.(3 .46(-.068) -C'ol.~ +C'' ri.(J -.46( .068) t • ol.(3 No .03(-.015) Equivalent C~-y, - • 13"~ ( • 03) -c•ti.(!> -.55(.11) c~~ -.09~~( .008) -Cd..(J •' ~.09(.042) c ~ol -.OJ*(-.02) -C''' o((J -.03(.015) 88 TABLE 21 INTERNAL COORDINATE STRETCH--BEND COMPLIANCE AND FORCE CONSTANTSa Calcd. for Mn(Co) 5Br Ci 1J. Approx. Equiv. Ave. Value for M(C0)6 Calcd. for M(Co) 6b +2Cc C Ij_,,8 +ct~iti8 O* ( 0) No Equivalent C =Cc =Ct = O*(O) risr r1.r IJ...'Y No Equivalent ri•{J cIJ_,cp C' e005*(0) +Ct ' Ii>1' ccc r'j_,11' O* ( 0) -cco,p cco,a. -c00,oc O*(O) Il,,,., cr.•].H -C ~1H O* ( 0 ) '\U 'Ill No -.0076(0) .0038(0) -.004*(.037) cro, r .0053*(-.oa> Cro,1f-S .0038*( .02) - 0co,oc ].H> co,~ Cd· a+2cg. A ~r+c~,pr .16*(-.05) Crl:t,'Y =C ~;y =C ~.r cdi,~ Cdi.t1" -C <i_,S e 1 O* cdj,1J1-cdi>s-·10* -.0038 .0053(0) .16(-.102) = O* ( 0) -.08*(o011) Equivalent -.0076(0) cri,~ l.' 0(0) • c •• CCO,cx. - CO,ct. c.CO,cx. -c CO,Ol. •• c•co, ol c •• Cr.· 8 -C~ c- O*(O) .0053(0) Equivalent - 0 r~c, /3 -.08(.05) 2C' .10 -20'MC,ct -.10 MC ,OI. a * denotes values held constant in calculationsJ values in parenthesees are equivalent force constants. b ** denotes values for Cr(C0)6; other values are averages for Cr(C0)6, Mo(C0)6, and W(C0)6• · Ref. (14). 89 TABLE 21 continued c~.l.~:'f -c~.l.' s .10* 2CMC,ct. .10 Cd]_,oc -C ~1'oc .10*(-.05) CMC,a. • -C MC,ix. •' .10(-.20) c do,'Y .0388(-.049) .08(-.051) cdo,'f-s c•MC, /3 -.12(.20) -cMC,d.. -.05(.10) CA,,, -.078(.068) No Equivalent cA ,"1'~8 -.09(.09) No Equivalent 90 It is immediately obvious that Cri < Cr 0 (i.e. Fri > Fr 0 ) and Cdi > Cd 0 (Fdi < Fd 0 ) • to the effect of greater d Tr - This is due 1T* bonding from Mn to the axial-CO ligand relative to the radial-CO ligands .. This is also obvious in the difference of Cd 0 for Mn(I) · and cMC for M(O)(Cd 0 = 11437 and CMC = .509). The radial M--C constants are about the same (Cdi = .. 449 and c000 = ~509) indicating little net change has occurred in these bonds. This is somewhat deceptive, however, as the radialCO stretching constant is a good deal lower for Mn(+l) than c 00 (Cri = .05693 and c00 = .05905) while the axial-CO stretching constant is approximately the same as cco (Cr 0 = .06057 and c00 = .05905). Thus sigma- bonding changes and their consequent effect on the pi-bonding system must not be ignored in these systems as is tacitly done in the majority of "approximate" force fields for C--0 stretching only ( 1, 5, 15' ) • The presence of a larger positive charge on the Mn atom relative to the neutral hexacarbonyl system shrinks the occupied d Tf- orbitals sufficiently so that the population of the axial-CO Tr* level is roughly the same as in M(C0)6 (as Cr ~ c00 )g 0 This increase in c--o bond order is clear in the case of 91 the radial-CO groups where Cri < c 00 where there is less compensating effect due to the substituted ligand's much lower 1T-electron accepting capability. The greater positive charge on the Mn atom is also partly responsible for the greater Mn--C bonding as both Cdi and Cd 0 are<~· The importance of sigma-bonding (30) in the interpretation of bonding changes will be discussed further in connection with the Mn(C0)5X and Cr(C0)5x- systems in Part III. Another approximation inherent in many simplified potential field calculations (1,5,8,15,18) is that the magnitude of the various CO, c•o• interaction constants is determined primarily by the amount of 1T-bonding d-orbitals common to CO and c•o•. As force constants were used in that formulation, this discussion will also be in terms of force constants. t c The approximation that Fr.,r. = 2Fr ,r. (1) l. l. 1 l. is shown to be in fortuitous agreement with the approximate calculations of reference (16) by the fact that F~ifi ~ .25F~ifi in the present calculation. As was illustrated by comparison calculations in (1~), this error is introduced primarily by the neglect of the anharmonic nature of the c--o stretching modes and the neglect of mixing primarily with the appropriate M--C stretching modes. 92 It might also be expected that the interaction of radial-MC with radial-CO would be smaller than the interaction of axial-MC with axial-CO due to the larger relative amount of d1T-+axial-C01T* bonding. This is definitely not the case as Frifi = .763 and Fr 0 ~ 0 = .758. It is even more interesting that fco,MC = .758 for M(C0)6 supporting the suggestion in (1~) that these interactions may be due to induction effects of the changing CO dipole moments rather than changes in the pi-bonding system. Little can be said of CA. The observed M--x stretching frequency for Cr(Co) Br- is 188cm-1, and by 5 a crude approximations 2 CA( Cr) = C..o. (Mn) ( ll Mn--Br) 2 ( ll Cr--Br) = .843 (222/188) 2 = .995 with a corresponding decrease in FA from 1.32 to 1.12; the higher bond order of Mn--Br compared to Cr--Br correlates well with the higher positive charge on Mn than Cr" The greatest interaction of /:j. with another stretching coordinate is C4 p 0 = -.088; the standard error is ± e27, however6 and thus this calculated value is judged to be not very significantly different from zero~ 93 The implications of a slightly negative interaction compliance constant with do (and thus a slightly positive interaction force constant) may be interpreted in terms of a dominant sigma-bonding effect. When d0 is increased (stretching the axial-MC bond), the population of electrons in that sigma bond decreases creating an increased positive charge on the Mn atom. This in turn strengthens the Mn--Br bondJ hence the sign of ~,do and FA,do• Bend-Bend Constants The most striking difference between the M--c--o bending constants for Mn(Co) 5Br and M(C0)6 is for c7 and cp (Cr = 1.287 and cp = 1.919). This was more or less assumed to be the case earlier when (E)'Y was assigned to the higher of the observed I. R. bands at 604 and 545cm- 1 and it should be emphasized again that the choice of this assignment is supported by the reasonable potential energy distribution obtained and the far better agreement of C!Y,'Y and c~1 with values transferred from the M(C0)6 . system (see pages 75 and 76). It is also apparent that the magnitudes of c~r and C~~ are considerably less than their hexacarbonyl equivalent. Although the assignment of this mode has 94 been successful through the argument that the vibration would involve distortion of the strongest aw-. Tr* bonds, the interaction constants are less than those for a system with presumably weaker equivalent d7T-+1T* bonds. Although it is a more tenuous argument than that advanced above for the more precisely determined stretchstretch interactions, it seems that the M--c--o .bend- · bend interactions may also arise primarily through electrostatic effects rather than changes in the pibonding system. 1,r >> C~ C~r , This is further supported in that as would be expected for a through- space electrostatic interaction, The compliance constant C~ for the in-plane MCO bending coordinate is found to be quite near the value determined for Cr(C0)6• The order Cp > Cr was predicted above due to changes in 1T-bonding between the M(C0)6 and Mn{Co) 5Br molecules and the calculated value of C~ agrees with this prediction. This similarity in the in-plane compliance constants for the systems is also reflected in the interaction constant .ct p,~ • = -.25. while c~~ = -.264 Due to the fact that the A2 frequency 2 ~ was not observed, insufficient information was available for the calculation of both CPif and c~ , hence 13 (A C~~ was held fixed at .07. 95 According to the "TT-bonding model we should expect that the bending of the axial M--C--0 angle would show the least compliance (or greatest restoring force) of all the MCO bending coordinates. However, the calcu- lated value of C~ is slightly more than the in-plane constant, Cp • One possible solution for this dilemma is the reversal of the assignments for (E)i and (E); • As was mentioned above, this leads to an extremely undesirable potential energy distribution and solutions for Cr, ci,1 , and c;,r~ Furthermore, the value of C~ only drops from 2.30 to 1.96 for this reversed assignment. We may be more certain of the values calculated for or and c~ because of the additional equations for these coordinates obtained in the accurately assigned and determined A1 , B1 , and B2 blocks. The ¢ MCO bending coordinate, however, is involved only in the E symmetry block, and thus significant errors in C~ may have resulted from the imprecision in the .present calculation for this symmetry blocke Hopefully further calculations for this block will refine the value presented here t for c4'. Another mechanism which might account for the differences among the MCO compliance constants is the electrostatic repulsion between the moving carbon atom and the bromine atom; this is larger for the out-of-plane r coordinate than the p and +coordinates and possibly • c'~ = Cp > Cr• coul d account f or t he ordering IV corro- borative evidence for this mechanism is found in that for Mn2(C0)10• which is analogous to (E)r for 1 Mn(Co) 5Br, has shifted to a position within 10cm- 1 (E1) of (E1)¢ (see figure 2~ in Part II). This implies a smaller electrostatic interaction for the out-of-plane bending coordinate which seems consistent with the much more covalent nature of the binuclear species. The I.R. spectrum of Mn(Co) 5H and its deuterium analogue have recently been reported (13) with a brief vibrational analysis; unfortunately the locations of lip , Vy , and V<f1 for this compound are quite obscured by the large amount of mixing of the C--Mn--H bending mode (also of E symmetry) with the Mn--C--0 bending modes. A normal coordinate calculation for this molecule would be most interesting for the purpose of determining the importance of the electrostatic interaction mechanism for Mn--C--0 bending modes. Recent neutron diffraction studies (18) have shown the Mn--H bond distance is 1.601 A.which is in good agreement with the covalent radii of Mn and H. However, the axial Mn--C is not foreshortened relative to the equatorial Mn--c,, distancesa This is not expected on the basis of 7r-bonding considerations alone, and thus a fair amount of positive 9? charge on the central Mn atom must be hypothesized to account for this situation. Considerable electrostatic interactions of the MCO bending and the polarized H ligand are therefore expected, and if this interaction is the dominant one in the determination of Cfo, c1 , and c;, it is predicted that these values for Mn(Co) 5H will more closely resemble those of Mn(C0)5Br than Mn2(C0)10• Ve-ry little may be said about the calculated compliance constants for the C--Mn--C bending coordinates ~ and ~. and the C--Mn--Br bending coordinate 8 • The values for C~, C "/' - C"f,8, and c1 - c,;,y, lie fairly close to their counterparts for the M(C0)6 calculations. That the constants involving the out-of-plane coordinate 'lJ' - 8 generally exhibit ~ compliance than their hexacarbonyl analogues may be related to the behavior noted above for the MCO bending coordinates. This self-consistency is a good indication that at least the frequency assignments for this low energy region are correot. . Further refined calculations will be necessary, however, before a more definite relationship between MCO and CMC bending constants can be establishede 98 The values calculated for c~,'V' - c~,s and ci_,d. are seen to be in fair and good agreement, respectively, with their M(C0)6 counterparts. There remain to be discussed only those interactions between different symmetry coordinates. c13,r, cp,r, and C~~ had no equivalents among the Cij for M(C0)6 and were all constrained equal to zero. Cp,~ was cal- culated = -.37 in fairly good agreement with the transferred value of -.25. The CMC and MCO coordinates ci and ~ are highly mechanically coupled and proved to be very sensitive to the values of the off-diagonal constants coupling them. The values for C~,rJ.. and c,B:oL are in fair agreement with transferred hexacarbonyl estimates. The out-of-plane coordinates1-8 and~ were also very sensitive to their off-diagonal coupling constants and i ~ is seen that the value for Cr,'l}l.t ci'Y'-s is quite different from the hexacarbonyl equivalent. This is yet another indication that the major distinction among the bending coordinates of the two systems is between the out-of-plane modes. The remainder of the bend-bend constants were judged to be of less physical significance than those above (indeed most of these were held constant during 99 the calculations) and no attempt will be made to discuss them at this point. Stretch-Bend Interactions All constants involving ri and r 0 were held at constant values in the calculations. These values proved to be quite small and poorly determined in the calculations for M(C0)6 (1~) so that the above approach was deemed most reasonable for the more complex Mn(Co) 5Br system~ A similar procedure was adopted with regard to constants involving di• Values for Cdo'Y and Cdo'f-8 are in fair agreement with values transferred from M(C0)6• The remaining constants, . cA,r and c61,f-B are both negative and their interpretation is essentially the same as that for Cd o,~A (which was found= -.088); that is, the sign of these interactions is best explained by changes in the electron density of the appropriate metal-ligand bond due to the change in the charge of the central Mn atom; this in turn is caused by the stretching pf the Mn--Br bond through a unit displacementQ The negative sign of these compliance constants can not be explained on the basis of changes in '7T-bondingQ An increasing amount of positive charge on the Mn atom due to stretching the Mn--Br bond would be expected 100 to shrink the metal d-orbitals causing weaker Mn--C 'Tr-bonds; this mechanism would thus predict a more compliant C--Mn--C bending coordinate and positive values for Cfl ,'Y and CA ,1f'-o. This is contradictory with calculated values for these constantse Conclusions The following will summarize briefly the conclusions on the chemical bonding within the Mn(Co) Br 5 · system on the basis of the vibration analysis presented in this chapter. le Changes in C--0 bond order from M(C0)6 to Mn(Co) 5Br are seen primarily in the radial ligands which are approximately 5% less compliant than the C--0 bonds in M(C0)6• The compliance of the axial-CO bond is very nearly the same as that of a CO group in M(C0)6. 2. These changes in C--0 bond order are produced by two effects& a selectively increasing population of the CO 7T* orbital in the axial-CO ligand due to the decreased pi-acceptor capability of the Br- ligand; and by a concurrent decreasing population of both radial- and axial-CO 7r* orbitals due to the shrinkage of the filled metal d-orbitals which is caused by the 101 relatively higher positive charge on the Mn atom compared to the M(O) atom in M(C0)6• 3. The pi-effect described above is responsible for making the axial-MnC bond less compliant than the radial-MnC bonds; however, both axial- and radial-MnC bonds are less compliant than the M--C bonds in M(Co) 6 due to the greater amount of Mn--CO sigma bonding presents 4e Stretch-stretch interaction constants were found not to agree with values predicted from only simple IT-bonding effects (1, 18). At least some contribution to these interactions must arise from electrostatic effects (or w• oscillating dipoles") described in (8, 14, 29). 5. Changes in 'Tr-bonding did not fully explain the ordering of the primary MCO bending constants although it is correctly predicted that the compliance of the out-of-plane bend is less than the compliance of the in-plane bend. An alternative mechanism was proposed to explain the anomalously high compliance of the axial-MCO bending coordinate. This mechanism is based on elec- trostatic interactions and successfully predicts the order c r <. c,s -::: c</>. 102 6. CMC bending constants are rather poorly determined but seem to reflect the dependence of the MOO modes on electrostatic interactions. 7. Bend-bend interactions also are found not to change noticeably between the different 7r-bonding situations in Mn(Co) 5Br and M(C0)6 except for those interactions involving out-of-plane bending. This further supports the electrostatic interaction mechanism above. 8. Changes in the amount of charge on the cen- tral metal atom due to the difference in O--bonding between Mn(Co) Br and M(C0)6 are necessary to explain 5 the signs of the interaction constants between Mn--Br stretching and1-MnCO bending, d 0 -MnC stretching, and (f-8)-CMnC bending, respectively, all of which constants are negative. 103 II. VIBRATIONAL ANALYSIS OF DIMANGANESE DECACARBONYL 104 CHAPTER 1 VIBRATIONAL SPECTRUM OF DIMANGANESE DECACARBONYL Preparation and Instrumentation Normal Mn 2 (co) 10 was furnished by the Pressure Chemical Company and the Mn2( 1 3co) 10 used in this study was prepared by Mr. Maxwell Goldblatt at LASL by the reaction of Mn(C 2H;02)2 with 1 3co under pressure. Both compounds were quite impure on arrival and were sublimed in vacuo at 60° c. to obtain the pure, bright yellow Mn2 (co) 10 • The purity of the resultant product was checked by comparison with the reported I. R. spectrum in the 1900 to 2100 cm- 1 region (3 ) and this criterion was used in all subsequent investigations. Although the solid is stable in air over a period of weeks, solutions in CH2Cl2 were found to decompose rapidly within the course of a day. Therefore fresh solutions were prepared for each spectroscopic determination. The instruments used were the same as those mentioned in Part I with the exception that far I. R. measurements with the Beckmann IR-11 were not made. Instead this portion of the spectrum is taken from Reference (20). 105 C--0 Stretching Vibrations The internal coordinates for Mn2(C0) 10 are shown in figure 19 and the symmetry coordinates for the C--0 stretching modes are shown in figure 20. For the Ai and B2, A2 and B1, and Ei and EJ modes the internal coordinates will be similar to A1 , A2 and E modes, respectively, of Mn(C0)5Bro Each of the modes for the monomeric compound will have split into a pair of modes in the dimeric system. These modes will differ only in their symmetry with respect to those operations which interchange the two halves of the molecule and they correlate as follows1 Mn2(C0)10 In Phase Mn(C0)5Br Out of Phase Ai ~ A1 + B2 A2 ~ A2 + B1 E1 + EJ E ...... The components of the degenerate E2 mode are restricted to the separate halves of the molecule and resemble exactly their B1 and B2 analogues in the Mn(C0)5Br system$ Of these modes the A1 , E2, and E3 types are Raman active only, the B2 and E1 types are IG R. active only, while the A2 and B1 types are totally inactive. 8 I 3 0 I 130 Figure 19. 90_-2-- 19 r 7 I 8 18 Internal Coordinates :for Mn (co) 2 10 G 11 I 0 1 0 8 01 0 ~ °' 0 107 Sit= 1/2{2(r1+r2+r 3+r4, -r5-r6-r7-ra. 2/1992 - (Al) r 0 - (B2) ro S2 = 1//2(r9+r10> Figure 20e Internal Symmetry Coordinates and Frequency Assignments for C--0 Stretching Modes 108 111978 - (E3)ri s 31 = 1/2;12Cr1+rz;;.r 3-r4 -{2r_s+;'2r ) 7 Figure 20~ · Internal Symmetry Coordinates and Frequency Assignments for C--0 Stretching Modes and M--c:-o Bending Mode 109 Figure 200 Internal Symmetry Coordinates and Frequency Assignments for M--C--0 Bending Modes 110 l/649 Figure 20. Internal Symmetry Coordinates and Frequency Assignments for M--~-~~ Bending Modes 111 Figure 20@ Internal Symmetry Coordinates and Frequency Assignments for M--c--o· Bending Modes and M--C Stretching Mode 112 Figure 20 • . Internal Symmetry Coordinates and Frequency Assignments for M--C Stretching Modes 113 Figure 20" Internal Symmetry Coordinates and Frequency Assignments for M--C Stretching Modes, C--M--M--C Torsional Mode 114 V14o (E2)cl V95 (E2>1-s S29 = 1/2(ci1f-d2,J+l:X3f!--ct4,1) S30 = l/2:/2(1f 1-1f2+13-1f4 -s1+B2-s3+S-4> Figure 20e . Internal Symmetry Coordinates and Frequency Assignments for C--M--C Bending Modes 115 Figure 20~: Internal Symmetry Coordinates and Frequency Assignments for C--M-~~ Bending Modes 116 · Figure 20. Internal Symmetry Coordinates and Frequency Assignments for C--M--M Bending Modes 117 The three intense I. R. modes in the 1900 to 2100 cm- 1 region have been reported and assigned a number of times (3, 5, 31, 32, 33, 34, 35). These assignments are l.12046 ~ (B2)ri' Zl2012 -... (E1)r111 and V1979 ,..... (B2>r 0 • The Raman spectrum exhibits four bands and is shown in figure 22. The two at higher energy correlate with the (A 1 )ri and (B 1 )ri modes of Mn(C0)5Br and are thus assigned as Zl2114 ~ (A1)ri and Zl2020 _., (E2>ri • The assignment of the two bands at lower energy has been quite controversial among the various Raman studies of Mn2(C0) 1o• The assignment Z1197 8 --.. (A1)r 0 and Zl1992 -+- (E ) ri has been argued for in (5) on the basis of 3 positive CO-C'O' interaction force constants across the metal-metal bond. The authors in (31) and {34) favor the reverse assignment on the basis of the observed binary combination I. R. spectrum. The present investigation has shown that the band at 1992 cm-1 in solution is very slightly polarized indicating that this is indeed the (A 1 )r 0 mode and thus the second assignment above is the preferred one. M--C--0 Bending Modes The Raman and I. R. spectra of Mn2(C0)10 from 700 to 500 cm-1 are shown in figures 23 and 24. The 118 • s::: 0 ·r-1 .p ~ ...-1 0 ti) (\J r-i 0 0 0 0 (\J ::r:: (\J 0 i:: •r-I -- ..... 's 0 ......... 0 ..... 0 0 ......... (\J ~ 0 \/"\ 0 (\J ~ 'H 0 s ~ +> 0 <I> p. ti) • 0::: • H 0 0 .... • ..... (\J C\l Cl> ~ &i or! ~ Figure 22. I 2100 I V ( -1) cm 2010 Raman Spectrum of rv'ln 2 (co) 10 in CH 2c1 2 Solution. 2050 Sensitivity = 750 (Dynode 4) Slit = 10 cm- 1 ;Double Slit;10 cm high '° ..... ..... 120 three Raman bands around 625cm- 1 for Mn(Co) 5Br have apparently been replaced by the single peak at 672cm- 1 for Mn2(C0)10• Depolarization measurements indicate that this is the (A 1 )r' out-of-plane mode. As three modes are expected in this region it may be that either or both (E 2 \, and (E 3 ),8 with this mode. are accidentally degenerate Further experiments with a more power- ful exciting source than the 100 rrNI. He-Ne laser used in this study and low temperature apparatus will be necessary to establish the presence of weaker lines in the vicinity of 625cm-1. I. R. bands in this region consist of two strong peaks at 649 and 642cm- 1 • Preliminary experiments by R. Levenson (J6) reveal that the assignment of these bands should be 'l/64 9 --9' (B 2 )'( and l.164 2 --. (E1),s. This ordering was obtained from polarized I. R. spectra of Mn2(C0) 10 in a nematic liquid crystal as described in (35)11 Two pairs of Raman-I. R. splittings are observed at 555 and 558cm- 1 and at 541 and 547cm- 1 • One of these pairs is analogous to the (E)s6 mode of Mn(CO) 5Br at 545cm- 1 and thus we assign V547 -.. (E 1 )s6 and l/ 541 _,,.. (E3)¢ • Figure 23. 600 550 Raman Spectrum of Solid Mn2(C0)10 (cm-1 Sensitivity = 260 (Dynode 4) Slit = 10 cm-l;Double Slit;2.5 cm high ~ I\) ~ 122 0 0 ..::}' s:: •rl 0 ~ ,...... 0 0 .......... C\l s:: ~ c.rl 0 0 s:: 0 s •rl \.{'\ $-i~ ~..p ..p r-l 00 (l){/) .... 's () A ti) .:}- r-l •O ~o H • • ..::t 0 '° '-1- o/o C\l <D H ~ QO •rl J:'r.i cO 123 As there are no bands around 604cm-1 corresponding to the (E)'Y mode in Mn(Co) 5Br we assign the remaining pair as l-1558 ~ (E1)r' and l.1555 -+- (E3)r • Upon 13co substitution a weak band appears between the two moderately strong Raman peaks at 473 and 45Jcm-1. This is shown in figure 25 and the position of the weak band is indicated by an arrow. In the normal compound, however, only two peaks are observed in this region. It is noted that the peak at higher energy is considerably more intense than its counterpart in Mn2(1Jco) 10 • This is due to the accidental degeneracy of the weakest bands which are clearly resolved in the spectrum of the 1 3co substituted molecule. Both of the stronger peaks may be assigned as Mn--C stretching modes as is discussed below. It was seen in (10), (14), and Part I that MCO bending modes shift further upon 1 3co substitution than MC stretching modes, thus the weakest band is assigned as (E 2 ~ analogous to l.1535 ~ (B2),e for Mn(C0) Br. 5 The (A 2 ~ and (B 1 ?,B modes are I. R. and Raman inactive and were not observedo Their positions were calculated with transferred potential constants from Mn(C0)5Br as is discussed belowo ~ ti Mn 2 25. • • & 7. .../ "'- ... • .:' j • - • 460 • • • e • • • l/ 41 • • • • • • w • . • I • • 440 . . .. (cm- 1 .. Raman Spectrum of Isotopically Substituted Solid Mn 2 (co) 10 • . · ·4ao c13c16o) 10 . . . . . .. 5.00 e Figure '1 Mn2(12c160>10 • .• •. .{:" f\) ~ 125 M--C Stretching Modes The Raman and I. R. spectra for this region are shown in figures 26 and .24,, Only two I. R. bands are observed in this region and they are assigned as 2/468 __.... (E1)di and v399 _.... (B 2 )di on the basis of similarity with the I. R. bands of Mn(C0)5Br regarding both position and intensity. Four Raman bands were observed in solution two of which were quite polarized. These two polarized bands are thus assigned as the A1 modes v482 ~(A 1 )d 0 and If there were no interactions across the Mn--Mn bond, (E 1 ) di and (E 3 ) di woul·d be expected to occur at the same frequency; hence we assign 2/465 ...-. (E3)ai leaving V419 ~ (E2)ai • Using this same argument we might reasonably expect the (B2)a mode to occur in the vicinity of its Raman A1 count erpar t a t 482cm- 1 . Th.is 1 eads t o th e . ac t ive conclusion that either this mode has very little in0 w tensity in the infra-red or that it is accidentally degenerate with (E 1 )a at 465cm- 1 • The latter inter1 pretation is preferred in spite of the fact that this band is observed to be quite symmetrical in a variety of solvents. As B2 modes develop intensity in the Figure 26. 400 -z/ (crn- 1 ) 350 Raman Spectrum of Solid Mn 2 (co) 10 Sensitivity = 260 (Dynode 4) Slit = 10 cm- 1 ;Single Slit; 2.5 cm high °' l\) ~ 127 infra-red along the Z-axis (i.e. the Mn--Mn bond direc• tion), it is expected that the axial-MC mode should be more intense than the radial-MC mode (as is the case for Mn(C0)5Br). Confirmation of this hypothesis may perhaps be obtained by resolution of the two bands at liquid nitrogen temperature or a shift in band position and shape in polarized single crystal experiments. M--M Stretch and C--M--C Bending Modes The Raman spectrum from 200 to 50 cm-1 is presented in figure 27. This region is quite complex and no complete assignment will be attempted here in view of the uncertainties regarding this region in the simpler 5Br molecule. Fortunately Mn2(C0) 10 is a good deal more soluble Mn(Co) in CH2Cl2 than the simpler pentacarbonyl halides and excellent polarization results were obtained. These are shown in figure 28 and clearly indicate that the bands at 160 and 113cm-1 are of A1 symmetry. This is of considerable importance as one of the more fundamental questions to be answered by vibrational analyses of metal-metal bonded systems is how much mixing occurs between the pure metal-metal stretching coordinate and ligand-metal-ligand deformation coordinates of the same Figure 27. 200 7/ (c.m-1) 100 Raman Spectrum of Solid Mn2(Co) 1 o Sensitivity 1 130 (Dynode 4) Slit = 5 cm- ;Single Slit; 2.5 cm high 00 l\) ~ e • • ~ .• (Al )A 150 • • • 0 • v cm-1 (Al )'f'- S 100 i ... ......... • ....... . 5,0 .• .. .• Polarized Raman Spectrum of Mn 2 (co) 10 in CH2c1 2 Solution. . .•. ••. . ... 1... .... .... ••• Polaroid Analyzer Perpendicular to Laser E. Figure 28. 200 D Polaroid Analyzer Parallel to Laser E. \0 [\) ~ 130 symmetry. Obviously nothing too definite can be said in this regard if these frequencies are not accurately assigned. For the purposes of potential constant calculations the remaining frequencies were assigned as similarly as possible to the assignments for Mn{C0)5Br from an energy standpoint. Thus we assign Zl14o --... {E3)~, l/131 -... {E2)ct., V95 ~ {E2>1-a• and 1/75 _,.. {E3)ct_. Zlt60 and v113 are assigned as above to {A 1 )A and CA 1 )~_ • 8 The ordering of these A1 modes is quite cer- tain as the polarizability change with the change in the normal coordinate of vibration {which determines the Raman intensity of a given mode {22,25)) should be considerably greater for the stretching of the bond between two diffuse heavy atoms than for the ligandmetal-ligand deformation mode. The importance of this concept in the interpretation of Raman spectra of metalmetal bonded systems in this region is discussed further in (37). The C--M--M deformation {similar to (E) 8 for Mn(C0)5Br) is expected to occur at an energy lower than 5Jcm- 1 • That no band was observed in this immediate area gives additional support to the assignment for the monomeric system. The torsional mode (B 1 )r is expected to lie well below 50cm·1 and is also both I. R. and Raman inactive. 131 CHAPTER 2 POTENTIAL CONSTANT CALCULATION FOR DIMANGANESE DECACARBONYL Isotopic Substitution The observed Raman and infra-red bands for Mn2(Co) 10 and Mn2( 1 3co) 10 in both CH2Cl2 solution and polycrystalline phases are presented in Tables 22, 23, and 24. Unfortunately time limitations did not permit the recording of spectra for Mn2(c 180)10 (also prepared by Mr. M. Goldblatt of LASL). Due to the approximate nature of the calculations below the lack of these data is not too serious a detriment; however, more detailed. future calculations should definitely include this information for the most accurate fitting of the M--C stretching and C--M--C bending modes which are particularly sensitive to c18o substitution. Anharmonic Corrections The binary combination region of CO stretching bands in the Io R. spectrum of Mn 2 (C0) 1o has been reported and assigned a number of times (5, 34, 38). Unfortunately reasonably correct harmonic frequencies can not be calculated until data for the ternary and 132 TABLE 22 Observed Intense Raman and Infra-red Bands for Mn 2(co) 10 in CH 2Cl 2 Solutions Mn 2 c12 0160 ) Raman 10 Infra-red Mn 2 Raman c130160 ) 10 74.6 ± @5 74.2 ± .4 95.2 ± .2 113.4 ± .6 93.7 ± .7 112.5 ± .6 160.3 ± .1 158.9 ± .1 392.4 ± .1 398.7 ± • 3 407.7 ± .2 401.6 ± .1 419.0 ± .8 413.0 ± .6 464.6 ± .2 456.3 ± .5 460.5 ± .1 467.9 ± .1 476.1 ± .4 481.8 ± .2 546.5 ± 2.0 not msrd. 558.0 ± 2.0 642.8 ± 1.0 not msrd. 649.0 ± 1.0 not msrd. not msrd. 672. 7 ± 01 658.3±.l 1978.2 ± 1.0 1933.1 ± 1.0 1933.6 ± .3 1978e7 ± .1 199108 ± ·3 Infra-red 1947.2 ± .3 133 TABLE 22 continued Mn <12 0 16 0 ) 2 10 Raman Infra-red Mn (130 160 ) 2 10 Raman Infra-red 2011.5 ± .1 1966.9 ± .1 2020.3 ± .3 1975.7 + .3 2045.9 + .1 2113.5 + .2 - - 1999.0 ± .1 2066.6 ± .2 134 TABLE 23 Observed Intense Raman Bands for Solid Mn 2 (co) 10 a Mn c12c16o) 2 10 70. 7 ± . 2 l Mn ( 13c160 ) 2 10 70. 3 ± .1 81.7 ± 1.5* not } resolved 91.6 ± 1.5* by hand 98.5 ± 1.5* 114.8 ± .1 115.5 ± .1 129.8 ± 1.0* {not resolve] by hand 138.5 + 1.0* 160.0 ± .1 158.8 ± .1 411.0 ±. ·3 405.3 + .2 421.1 ± .3 414.5 ± .7 465.1 ± .2 456.8 ± .3 not obsvcl.. 464.5 ± .7 480.1 ± o5 474.1 ± .4 540~7 ± .4 522.3 ± .4 535.7 + .2 55408 ± .6 67206 ± .2 657.9 ± .2 a * denotes bands lying close together which were resolved by hand. 135 TABLE 24 Observed Intense Infra-red Bands for Solid Mn2(C0)1oa Reference (20a) 1:30 s 121 s 115 vs 107 a s Reference (20b) 129 sh 118 sh 109 m Frequencies given are for Mn2(1 2c160)10 only and are taken exclusively from References (20a) and (20b) as noted above. quaternary CO stretching combinations are obtained. As a first approximation the harmonic corrections for the observed C--0 stretching frequencies of Mn(Co) 5Br will be applied to the observed frequencies for Mn2(C0) 1o using the symmetry correlations for C4v-+D4d given in Thus we obtain: Chapter 1. v Observed Harmonic (A1)ri 2113.5 2134.3 (B2)r. 2045.9 2066.7 (E2)ri 2020.3 2018.7 (E1)ri 2011.5 2037.8 (E3)ri 1978.2 2007.2 (A1>ro 1991.8 2012.6 (B2>ro 1978.7 2002.2 l Admittedly this is a dubious procedure but these corrections will probably be correct to a first approximation; as these calculations themselves are a first full approximation, the transfer of anharmonic corrections from Mn(Co) Br to Mn 2 (co) 10 is justified on 5 the basis that a more accurate harmonic potential field would result than if only the observed frequencies of Mn 2 (co) 10 were used. As was done for Mn(Co) Br, only observed frequen5 cies for modes other than CO stretching were used in potential constant calculations. 1.37 Potential Constant Calculations The calculations for Mn 2 (co) 10 followed the same procedures outlined in Part I. A new G matrix was calculated using data from the crystal structure determination of Dahl and Rundle (39). The formulation of the symmetry coordinates was carried out in a manner which would most facilitate the direct transfer of symmetry potential constants from Mn(Co) 5Br to Mn2(C0)10• It was hoped that the values of the transferred off-diagonal constants in Q would be accurate enough that a satisfactory fit of the frequency data could be obtained by perturbing only the diagonal constants. A1 and B2 Symmetry Blocks As in the case of the monomeric pentacarbonyl bromide, these are by far the most accurately assigned frequenciese A splitting diagram in figure 29 shows the steps in the calculation. First, the frequencies are calculated from a direct transfer of compliance constants from QA 1 of Mn(Co) 5Br. Second, the Cij are adjusted for the best fit of the data for Mn2 (co) 10 still with no splitting between pairs of A1 and B2 modes. Finally the split A1 and B2 frequencies are 138 compared with the observed values. Convergence for the six frequencies of the B2 block was obtained rapidly and with excellent fitting of the data by adjusting only the diagonal Cii• A considerable amount of difficulty was encountered with this procedure in the A1 block. As can be seen in the direct transfer calculation (figure 29) (A 1 )A was too high and (A 1 )f\/l-S was too low. Calculations by Spiro (37) have also shown this behavior. There a "reasonable amount" (to be specified in a later more detailed publication) of mixing between these two coordinates is assumed and the (A 1 )A mode is fitted to a frequency of 160cm-1. This yields a calculated value for {A 1 )~-S of 57cm-1 which is a factor of two lower than the observed value of 11Jcm-1. It was found in the present study that a satisfactory fit of these two frequencies could not be obtained with the value of C 6, 7 transferred. The problem was complicated by the fact that {A 1 )r and {A 1 )~-s are very highly mechanically coupled. Thus the large changes required in C 7, 7 to fit (A 1 )~_ 8 had a most undesirable effect on the fit of (A 1 )r· The final adjusted Cij are presented in Table 25. Symmetry coordinate compliance and force constants are presented in Table 26 while the internal coordinate \ \ Figure 29. (..,,; c.rn- 1900 1 I • Gl. 700 I I '-oo 0 bserve.cl .I I I I I I I A, - 13,_ 500 11 . II I I I I I I I I I I ( ( . 300 I Spl•' ttinJ Splitf:,'njS t.{oo J I 1 I I 1 I I • I .100 • A1 and B2 Vibrational Calculations for Mnz(CO)io• ) 1 I .tooC> I • I ' I . 11. ' I 11 I I 2H>o I I I I' I I I I I I I I I II I I I 1 I I ( 1 1 100 II I II I I I I I I I I I I ( \ I \ \ \ I \ I I II II (<.<l)~ 8 r Spt~-t.tinJ) YVln (A)A (q 1 4B~\-~ ~ I \ I \ I II 11 11 - I I (N,, I I I I I I I I I I 1 II 1 I .fro,.,,.. I II A1 Ci l c .. 'J 0 I \ I\ /\ I I I I I I I e.f 0 lAJ.1 (BJ.t (~J.i; lB )cl.i 11 Co.lc~lc.;tecl ~ c 0... dj u.~t Tro.n$ fer I I \ \ i I 11 I \ \ /\ \ \ I\ I I \ \ \ Oirec.t J \ I lA.\(S~r I \ I (A)l"i (BJr; lA1)r.lS~r0 I ..... '° \...iJ 140 TABLE 25 Compliance and Force Constanta Calculation for A1 Symmetry Block c 1, 1 = .0569 c 2, 2 = .0621 c J, .3 = 1.311 c 4, 4 = .435 c 5, 5 = .359 c 6, 6 = 1.594 c 7, 7 = 1.7 c 6, 7 = -1.0 F 1, 1 F 2, 2 = 17.64 = 16.38 F J, 3 = .794 F 4, 4 = 3.16 = 2.99 F 6, 6 = 1.26 F 7, 7 = 1.29 .88 F 6, 7 = F 5, 5 a Force constants were derived from elements of c- 1 • 1.9 -2.9 158.4 116.3 160.3 113.4 157.2 115.5 -3.0 112.5 1.7 .1 1.9 -4.6 .2 1.5 158.9 401.5 401.6 -.2 407.9 407.7 474.2 476.1 -1.8 483.6 481.8 662.9 677.5 672.7 1967.3 658.3 2014.7 2012.6 2085.1 2086.6 -4.8 2136.0 2134.3 Vealed. Vobsd. 1967.5 -1.7 V'calcd. V'obsd. 6. Mn c13016o) 2 10 -2.1 6 Mn c12 016o) 2 10 --- 111.0 156.1 J9J.4 --- --- 471.9 671.1 1970.1 2089.2 --- --- ----------- 10 D. Vealed. ----- a.--- ----- Vobsd. 2 Mn c120 18o) Frequency Calculation for A1 Symmetry Block TABLE 25 continued .... .{::" .... 142 TABLE 25 continued Compliance and Force Constanta Calculation for B2 Symmetry Block 0 11,11 = .0617 Fll,11 .. 0612 F12,12 = 16.26 = 16.60 = 1.138 F13,13 = .891 F14,14 = 2.39 F15,15 = 2.76 F16,16 = .41 0 12,12 = C13,13 = .. 451 = .371 0 16116 = 2.59 c14,14 0 15,15 a Force constants were derived from elements of c-1. .8 -·3 0 1955.9 627.6 460.8 392.4 112.7 1956.0 628.4 460.5 392.4 --- -.2 -.7 .3 0 .2 2002.4 643.5 467.6 398.7 642.8 467.9 398.7 113.4(?) 113.2 --- .1 --------- 109.1 .384.J 456.2 638.2 1957.0 2021.5 ------- .6 2018.4 2019.0 2002.2 l/calcd. l/obsd. 6. Vealed. 2 --------- --- --- 6. 10 Mn c120 180 > Vobsd. -·5 10 2067.2 c130 160 ) 2066.7 2 ~ Mn Vealed. 10 Vobsd. 2 Mn c120 16 0 ) Frequency Calculation for B2 Symmetry Block TABLE 2 5 continued \,,.) ~ ~ 144 TABLE 26 Symmetry Coordinate Splitting Constants A1-B2 Blocks Internal Coord. r·l. .0569 - .0617 = -.0048 17.64 - 16.26 = +1.38 .0621 - .0612 = +.0009 16.38 - 16.60 = -.22 1.311 - 1.138 = +.17 .794 - .891 = -.10 .435 - .451 = -.02 3.16 - 2.39 = +.77 .359 - .371 = -.01 2.99 - 2.76 = +.23 .0596 - .0616 -.0020 17.21 - 16.67 = +.54 E1 -E '3 Blocks r·l. = 145 TABLE 27 Internal Coordinate Splitting Constants .6cri = 4 <cri,45 + 0r1,13;> .6c = 2(0 .) ro ro,J ~c,, = 4 <0r,45 + cr,135> .6Fri = 4(Fri,45 + Fr1,135) 6.F = 2(F .) .6cdo = 2(Cdo, j) .6Fdo = 2(Fdo,j) 6.Fdi = 4 (Fdi,45 + Fdi,135) .6cdi =, 4 <cd 1,45 + 0d1,135> c i, j ro 6.Fry = 4(Fr,45 + Fr,135) This F i, 0r1,45 -.0010 Fri,45 0r ,135 1 cro,J. Ciy 45 + 6r,13s Cd o•J. -.0002 .oa Fri,135 Fro,J. -.11 Work .0004 .o4 -.01 Cdi,45 + cdi,135 -11003 Fy j This (5) (31) .27 .33 .22 () .03 -.12 .20 Work 45 + Fr,135 -.03 Fd o•J. ,,77 Fdi,45 ro•J + Fdi,13.5 .06 not cal ed. " " " " 146 constants are shown in Table 27. The value of c7 , 7 should be compared with that for c16,l6 in the B2 block calculation shown in Table 25. The compliance of the A1 C--M--C bending mode is much less than its B2 counterpart and this suggests that perhaps a higher frequency should be assigned to ( B2 ).y, _8 e The "metal-metal" stretching force constant is determined to be 1,26 and the mixing constant with the C--M--C bending mode is o.88, This very large value for FA,~-o indicates a considerable interaction of the A and ~-S coordinates as predicted in Part I. This is at variance with conclusions drawn ear- lier (37, 40, 41) and the present results are compared with some previous approximate calculationss F .. This Work Ref. (37) Ref. (40) Ref. (41) Fll. 1.26 .59 1.4 1.45 F6,1/J-5 .,88 ? ? ? F1'-8 1.7 ? ? ? 'l/L!i. 160 160 157 159 ~-s 113 cald. 57 ? ? - l. J z./i The calculations from (40) and (41) were done on a diatomic approximation with the mass of each "particle" 147 equal to the mass of a Mn(Co) unit. These two re5 sults are much more accurate than the normal coordinate calculation in .(37), but this agreement is seen to be purely fortuitous as neither the off-diagonal force constant nor the mechanical coupling is explicitly taken into consideration. Values for FA were also calculated in (37) for Tc 2 (co) 1o, Re2(Co) 10 , and Re(C0)5Mn(C0)5, and the trend in the force constants obtained was discussed. However, in (41) an entirely different trend was obtained by use of the "diatomic" approximation. Thus by the incorrect assignment of (A 1 )~_ 8 , not only are the absolute values of FA incorrect for the M2(C0)10 series in (37), but the relative values may also be incorrect for these F6 and this renders somewhat premature any conclusions drawn from these results. From the general agreement of these FA with conclusions drawn from electron impact measurements (42) it is expected that the trends in metal-metal bonding observed by Spiro (37) will be borne out; however, a complete normal coordinate analysis on Re2(C0) 10 based on calculations for isotopically 148 substituted Re(Co) 5x systems should be a minimum requirement to establish the nature of the F~ for the Mz(Co) 10 system. Calculated values for the Cij which were adjusted are presented in Table 25. For both Mn--C and C--0 stretching modes the A1 frequency for a given coordinate is at higher energy than its B2 counterpart. This is reflected in the larger Fii and smaller Cii for the A1 modes compared to the corresponding B2 modes except for (A 1 )r 0 and (B 2 )r 0 • These results give rise to positive interaction force constants across the metal-metal bond. These constants are Cri, (+r1 with +r5 or +rs), Cri, 135 (+r 1 with 45 +r6 or +r 7 ), and Cro., j(+r9 with +r10>. The nomenclature is similar for di and d0 coordinates. In these calculations interactions between axial coordinates on one Mn atom and radial coordinates on the other Mn atom were assumed to be equal to zero., If these interactions are assumed to arise through a conjugated 7r-electronic system involving the filled metal d-orbitals, valence bond theory (32) would predict that the stretching of a given CO bond (or MnC bond) on one Mn atom would produce weaker CO bonds (or MnC bonds) in the other Mn(C0)5 group. 149 This predicts negative interaction force constants across the Mn--Mn bond which is in disagreement with this study and three previous works (5, 31, 34). An attractive mechanism proposed by Cotton (5) predicts correctly the signs of these interaction constants by assuming the major effects involved are changes in repulsion between the various metal d-'li orbitals. Stretching of a radial-CO bond would in- crease the associated d-Mn~1T*CO back-bonding and lower the electron density in this metal d-orbital. The interaction across the Mn--~.n bond would then be a stabilization in the appropriate d-orbitals of the other Mn atom causing a decrease in d-Mn---+1T*CO bonding and a consequent strengthening of those CO bonds. Through-space electrostatic interactions as proposed in (14) and Part I of this work form a satisfactory alternative mechanism. Electrostatic effects in this molecule are shown to be significant by the large asymmetry parameter in the electric field gradient about each Mn atom as found by an investigation of the pure quadrupole resonance spectrum of polycrystalline 55Mn2(C0) 1o (43). This a- symmetry presumably arises from distortions of the charged radial-CO ligands from D4d symmetry as revealed 150 by the crystal structure determination of Mn2(C0) 10 ( 39). This X-ray diffraction study also revealed that the radial-CO groups on one Mn atom were bent an average of 4 degrees out of an MnC4 plane toward the other Mn(Co) group. This electronic repulsion be5 tween the axial-CO group and the radial-CO groups may be expected to have its counterpart in interactions across the metal-metal bond. This would pre- dict that the largest A1-B2 splitting due to electrostatic effects would occur for the radial-CO mode. This is seen to be the case from the splitting diagram in figure 29. Thus electrostatic interaction between nearby CO groups must be of considerable importance for the correct interpretation of spectra and potential field calculations. The nature of these interactions may be thought of as follows. The change in dipole moment along r 1 by stretching this coordinate will oppose the oscillating dipoles in ra and rs during in-phase (A1 and E1 ) modes and result in the lowest energy during the out-of-phase B2 and E3 modes. The net repulsion or attraction 151 of these dip:>les satisfactorily explains the signs of the splitting constants Cri, 45 and Fri,45 in Table 27. This mechanism predicts values for Cr o, J. and Fr o, J. of nearly zero (or very slightly negative) due to the large distance and number of atoms separating the axial coordinates r 9 and r 10 • This is indeed the case in this work where Cr 0 ,j = +.0004 and Fr o, j = -.11. The value of Fr 0 ,j can not be compared to the value of -.12 obtained in (5) due to the uncertainties in the input frequency (A 1 )r 0 in that work • . Recently Bor (31) has calculated a value of +.20 for Fro,j• The large disagreement between that value and the value in this work must arise principally used. from differences in the value of F c ro,ri This constant determines the mixing of the axial A1 and B2 modes with the respective radial A1 and B2 modesJ due to the large differences in "symmetry" splittings of the two Ai modes and the two B2 modes, the value used for F ~ 01 r· will be 1 very important in determining the value of Fr o, J•• 152 That the values of Fro, j and Fri, are much 135 smaller than the value of +.35 for Fri, 45 is taken as good evidence that the dominant mechanism of CO-C'O' interaction across the Mn--Mn bond involves primarily the oscillating dipole moments in the CO ligands themselves rather than the more internal mutual repulsion of metal d-orbitals. These considerations also offer an explanation for the very large anharmonici ty corrections and solvent shifts found for the CO stretching modes when compared to the other internal vibrational modes. The signs of the interaction constants for di and d 0 coordinates are predicted by either of the two mechanisms discussed above. However, the larger magnitudes of Cd o, j and Fd o, J· compared to Cdi, 45 and Fdi, 45 would indicate that electrostatic inter- actions play a very small role across the metal-metal bond in these "internal vibrations." It seems that the mutual influence of metal 1T- symrne1ry d-orbi tals (34) is more successful in predicting that the d 0 -d~ interactions should be larger than the di-di interactions. This mechanism would be more effective in the axial mode (d 0 -d~) than the radial mode (di-di), if perturbations in bond order are due primarily to changes in Mn~co -Tr back-bonding 153 rather than changes in Mn+-CO a- bonding. This is because there are more d-orbitals involved which are appropriately aligned for interaction across the Mn--Mn bond for the axial-CO ligands than the radial-CO ligands. In Part III it will be shown in a study of Raman intensity measurements that the role of the metald~1T*CO back-bonding system becomes increasingly important relative to the role of the CO~metal <Y bonding system in determining changes in bonding electron density for the series Mn(Co) c1, Mn(Co) 5Br, 5 Mn(Co) 5I, and Mn 2 (co) 10 • These independent findings would seem to favor changes in the Mn d-orbitals arising primarily through perturbations in the d-Mn~ 1T*CO system. The compliance constant for (A 1 )~ is actually larger than the value of c1 3,13 for (B2)~ • This is probably due to the misassignment of (B 2 )~-S mentioned earlierJ reassigning this mode to a higher frequency in accordance with the discussion on page 146 would have the desirable effect of raising the value of A2 and B1 Symmetry Blocks Calculations for the (A2)~ and (B 1 ~ MCO bending 441.3 = --- l Vealed. 448.8 42.3 Vobsd. --- --- Vealed. 427.8 Vobsd. --- --- ~ --- (130 16 0 ) 10 ~ ----- ----- 42.0 434.8 Vealed . Mn2(13c160>10 . ) Vobsd. " ( " ) ~ " ( 1c85 (held fixed) Mn2(12c160>10 0 10,10 2 Mn --- 2 435.9 --- --- ~ ----- Vobsd. ( " " .. " 4o.4 443.6 Vealed. --- --- 6 Mn2(12c180)10 0 = F 9 1 10 ( ) ) ,54 (held fixed) .1 9 = F 10, 10 = F 91 --- 10 ~ Vealed. Mn c12c18o) l = c54 (held fixed)} Vobsd. F B, 8 Frequency Calculation for B1 Symmetry Block = 10.0 0 c 9, 10 = c 9, 9 Vealed. Vobsd. Mn c12c16o) 2 10 { c 8, 8 = 1.85 (held fixed) Frequency Calculation for A2 Symmetry Block TABLE 28 t-a .{::" \..r\ 155 modes utilized a direct transfer of compliance constants from Mn(C0)5Br. A compliance constant of 10.0 (i.e. a force constant of 0.1) was arbitrarily chosen for the torsional mode (B 1 )~. The calculated frequencies are presented in Table 28. E2 Symmetry Block The components of these degenerate modes are localized on one or the other Mn(Co) group, there5 fore no interaction constants across the metal-metal bond exist for this block. The calculations are quite similar to the B1 and B2 blocks for Mn(C0)5Br. Off-diagonal Cij between E2 modes were transferred directly from these calculations and it was found necessary to alter only C25,27 to obtain a satisfactory fit. Cij were constrained equal to zero for those pairs of elements in which one mode was analogous to a B1 Mn(C0)5Br mode and the other a B2 Mn(Co) 5Br mode. The resulting C matrix had the following forms (__.B2) (E2 )J""d (_.Bl) (E2) ci. (E2)di (--l>Bl) (--r>-B2) (E2)/J - (-a>B1) (E2)'( .. (E2)ri (-t=-B1) I C23, 24 . IC23,23 0 24,24 (E2 )1' (_.Bl) (E2)ri (--+-B1) C25,25 0 0 (E2 ),8 (-.B2) ' 0 0 26,26 C27,27 C25,27 0 0 (E2 )ct (-.....B2) 0 0 24,26 023,26 (~Bl) (E2)di 0 28,28 0 0 26,28 0 0 24,28 023,28 (E2 )'¥'- o (-+Bl) °' .... \J\ 157 TABLE 29 Compliance and Force Constanta Calculation for E2 Symmetry Block C25,25 = 00615 F25,25 C26,26 = .802 = 2.73 F26,26 C27,27 c28,28 C29,29 C30, JO 0 26,28 a F27,27 = 16.58 = 1.44 = .369 F28,28 = 2.94 .83 .50 .05 = .349 = 1.21 = 2.34 FJO, JO = = = -.167 F26,28 = F29,29 Force constants were derived from elements of c- 1 • -1.8 -·3 412.5 131.6 94.o 413.0 129.8 -1.8 -.4 -.6 .3 483.6 419.4 131.9 94.9 481.8 419.0 131.3 95.2 -2.9 465.6 466.2 93.7 655.4 658.3 .,5 .6 2.9 1.1 675.6 1973.0 672.7 1974.1 -1.1 ~ 2019.8 Vealed. 2018.7 Vobsd. 2 c130 160 ) 10 !:J. Mn Vealed. 10 Vobsd. 2 Mn c12 0 16 0 ) 482.1 404.6 124.7 --- --- 90.7 666.8 ----- --- 1973.8 ---- Vobsd. ------------- 10 6. Vealed. 2 Mn ( 120 180 ) Frequency Calculation for E Symmetry Block TABLE 29 continued ~ co \J\ 159 The full value of these calculations will be realized only when less approximate values for the potential constants of the remaining symmetry blocks are obtained making it possible to determine the internal coordinate potential constants as was done for Mn(C0)5Br. As large uncertainties exist in the assignments of (E 2 )ry, (E 2 )ct., and (E 2 >1-s, further discussion of these calculations is thus unwarranted at this time. It is noted in closing that the values expected for this block should be close to the B1 and B2 calculations as the analogous E2 frequencies are not shifted· too far from the position of the Mn(Co) 5Br modes. E1 and E Symmetry Blocks 3 The calculations for these blocks were extremely unsatisfactory as no good fit of the three MCO bending modes and the MC stretching mode could be obtained by adjusting only the diagonal Cii• Apparently (E 1 )o/ and (E 3) 1 have shifted to lower energy while (E 1 )di and (E 3)di have shifted to higher energy. Mixing among the 'Y, ~' and di coordinates is expected to be even more extensive than for Mn(Co) 5Br. 160 It is unreasonable at this point to offer explanations for the large shifts of the'Y and di modes relative to their counterparts in Mn(Co) Br due to 5 the uncertainties regarding the potential field of the simpler molecule (see pps. 93 to 9 7). Potential constants were calculated for CO stretching modes yielding values of c1 7, 17 = .0596, 031,31 = .0616, F17,17 = 17.21, and F3l,Jl = 16.67. From these we deduce ~ Cii = -.0020 and AF ii = +.54. These result in numerical values for the internal coordinate splitting constants Cri, , Cri, 135, 45 Fri,45 , and Fri, 135 when combined with results from the A1 - B2 block results. this work Cri, 45 = -.0010 this work (5) ( 31) Fri,45 = .27 Fri,135 = .08 .33 .22 0 Cri,135 = -.0002 These values are in fair agreement with the .03 results of Bor (31) and indicate that the assumption by Cotton (5) that Fr , = O was fairly good. 1 136 Conclusions 1. From the observed frequency shifts and cal- culated potential constants it is seen that all C--0 bonds are more compliant in Mn2(C0)10 than in Mn(C0)5Br while the reverse is true for all Mn--C bonds. 2. These changes in bond order may best be explained by the increasing importance of Mn-d--+'7T*CO back-bonding from Mn(Co) Br to Mnz(C0)10• 5 The more covalent nature of the Mn--C bonds in the binuclear species due to the negligible induced positive charge on the Mn atoms will lead to a relatively greater expansion of the Mn d-orbitals and consequent greater Mn-d__,.7T"*CO back-bonding. 3. Splittings of A1 and B2 modes by interaction constants across the Mn--Mn bond offer very strong evidence that CO-C'O' interactions arise primarily through electrostatic effects while Mn.C-Mn'C' interactions come about by mutual repulsion of the Mn d-orbitals of '1T symmetry. 4~ Mixing of the metal-metal stretching coor- dinate with ligand-metal-ligand deformation coordinates is quite large and serious errors in the estimation of F~ for various metal-metal bonded systems may arise if this mixing is not taken into account. 162 III. RAMAN AND ELECTRONIC ABSORPTION STUDIES OF M(C0)5xn- COMPOUNDS CHAPTER 1 RAMAN SPECTRA OF M(C0)5xn- COMPOUNDS These compounds were prepared in accordance with the literature method (44); commercially available M(C0)6(M = Cr, Mo, or W) is reacted with N(c~ 5 )4X (X = Cl or Br; no iodides were prepared in this study) in hot diglyme for approximately one hour, then filtered from the remaining solids consisting of unreacted salts and decomposition products. The complex in solution was then precipitated by the addition of petroleum ether. At this point large amounts of impurities were present in most of the complexes prepared (except · for the Cr(C0)5Br- derivative) as evidenced by their infra-red spectra in the 2100 to 1800cm- 1 region. Before any experimental measurements were undertaken, the compounds were purified by recrystallization or by separation in CH2Cl2 or CHCl3• Intensity Studies of Raman Modes As shown in Table JO there is evidence for a reversal of the trend in intensity ratios for the 164 metal--carbon stretching modes (A1)d 0 /(A1)di going from the Cr subgroup pentacarbonyl halides to the Mn pentacarbonyl halides. For the Mn compounds the trend seems to be an increase in this ratio for the halide order I< Br< Cl while the intensity of the radial A1 M--C stretching mode seems to remain constant with respect to the intensities of the C--0 stretching modes. Thus the major change in intensity may be seen to be occurring in the vibration of the Mn--C bond trans to the halide atom. This is accompanied by an increase in the frequency of this mode in the order I< Br<. Cl. As the intensity of this mode increases with the electronegative character of the halide ion, it is suggested that the mechanism responsible for these changes is primarily an increasing stabilization of the axial-MnC o--bond by the induction of an increasing positive charge on the Mn atom. The increased electron density in the o--bond would then give rise to a greater change in polarization on vibration, 'dd.-/ a Qdo' which in turn is responsible for the inten- sity of this Raman mode. This induction of a positive charge also implies a shrinkage effect in the metal 'Tr-symmetry d-orbitals which in turn must reduce the amount of metal-ligand Ai (MC-rad) Ai (MC-ax) Ai (CO-ax) B1 (CO-rad) A1 (CO-rad) not msrd. 2062 387 498 i856 1969 2060 It n Br " " " " Cl " Cr " " " Mn " (1.00) ( .20) ( • 56) .70 .19 (1.00) i.oo .18 " not msrd. .50 aFor CH2Cl2 Solution; Values in parenthesees for Solid " .. Ai (MC-rad) not msrd. ( 063) Ai (CO-rad) i857 J89 not msrd. ( .58) A1 (CO-ax) B1 (CO-rad) 498 " " " " 387 Cl Cr 1.00 A1 (MC-rad) Ai (MC-ax) 1/ x M Assignment Intensity Relative to A1 (MC-rad)a Raman Intensities for CO and MC Stretching Modes: M(Co) 5xn- TABLE 30 ~ \.}'\ ~ A (CO-ax) 1 B1 (CO-rad) 384 1989 " " Br " " " " I " " " Mn tt " et Mn et A1(CO-rad) 2131 .. 2073 " " .. A1(CO-ax) B1(CO-rad) 2005 A1 (MC-rad) A1(MC-ax) A1 (CO-rad) B1 (CO-rad) A1 (CO-rad) . 464 389 2137 2080 473 2080 + 2095 1989 " et A1 (MC-rad) A1 (MC-ax) 2151 ff A1 (MC-ax) A1 (CO-ax) " 479 Cl Mn Assignment -- x M 1/ .,25 .17 ( • 53) ( • 98) ( .68) ( .19) (1.00) ( • 50) ( .90) ( • 54) ( • 34) 024 .10 (1.00) 1.00 ( • 51) (1.26) ( .65) Intensity relative to A1 (MC-rad) ( .44) TABLE 30 continued ..... °' °' d7f ~7f*CO bonding. This effect seems to be secon- dary to the a- effect in the Mn compounds; however, for the case of Cr(Co) c1- and Cr(Co) Br- it seems that 5 5 the importance of the er- and 'Tr-electron effects is reversed. This trend was also observed for the case of Mo(C0)5c1- and Mo(C0)5Br- but the data were deemed not as reliable as those for the Cr compounds due to the large amount of impurities present and the rapid decomposition of the material in solution. Only the W(Co) 5Br- derivative was prepared which obviously eliminates the observation of any trends for the tungsten series. It is suggested here that the increasing dipole induced in these M(O) compounds for Br to Cl is governed by the '1T back-bonding destabilization effect thus causing a net decrease in electron density within the axial Cr--C bond and a corresponding decrease in the intensity ratio of the Raman bands. Although one must necessarily have reservations about this interpretation until the M(Co) I- compounds 5 are prepared and measured, the trend is also reflected in the fact that there is now no increase in the frequency of (A 1 )d 0 mode going from Br to Cl. Further research must establish the relation of these M--C 168 stretching intensities to the intensities of other molecular vibrations (as was done above for the Mn compounds). Intensity ratios for the Cr compounds were determined with respect to only one C--0 stretching vibration, and thus it is deemed that further work with purer starting substances would be necessary to more fully substantiate these conclusions. It would also be helpful to extend the series of compounds to include the hydride derivative, M(Co) Hn-, 5 and possibly the binuclear species M(Co) --[M(C0)~2n-. 5 (, It was found that the intensity ratio of (A 1 )d 0 /(A 1 )di for Mn 2 (CO) 10 fitted into the series ~Cl; Mn( CO) 5 < I <Br however, here the shift in (A 1 )d 0 could not be directly compared with the other compounds due to a change in symmetry and frequency shifts due to interaction potential constants across the metal-metal bond. To summarize briefly, Raman studies indicate that the halide atom's influence in the Cr subgroup compounds is felt primarily through the M--C 'TT-bonding system. Further, the influence of the halide ion in the Mn compounds is most obvious in the M--C O--bonding system. This also implies the relative amount of 'Tr-bonding increases from Mn(C0) 5x to Cr(Co) x-e 5 CHAPTER 2 ELECTRONIC ABSORPTION SPECTRA OF M(Co) 5BrnRecording of the Spectra The observed electronic absorption spectra of )iJ Cr(Co) BrJ!<c 2H5 and Mn(Co) Br in various solvents 5 5 are illustrated in figures 30, 31, and 32. The band positions (with molar extinction coefficients for the Mn(I) compound in parenthesees) are given in Table 31. These spectra were obtained with a Cary 14 spectrometer specially modified for low temperature work. The difficulties with impurities and rapid decomposition of the Cr compound which were experienced in the Raman experiments were met with here also. The compound seemed especially unstable in the pure hydrocarbons used as solvents for Mn(C0)5Br, hence only the CHCl; solution data are presented here. Due to the impurities present (even following recrystallizat~on) only a rough estimate of the con- centration of the Cr compound in solution could be made and this is listed in figure 32. The observed spectra will be assigned on the basis of two possible models discussed belowi the 170 tetragonally distorted ligand field model (TDLF) and the pseudo octahedral ligand field model (POLF). Ligand Field Models for M(Co) Brn5 When the symmetry of a metal hexacarbonyl molecule is lowered from Oh to c4v on going to a M(Co) 5xncomplex the eg and t 2g ligand field levels are split as followsa ~ ~ a 1 (dz2) + b 1 (dx2 _ y2) ~ ~ e(dzx,yz) + b2(dxy> In the complex under discussion the X group is the er-electron acceptor Br. It may be argued that replacement of a CO group by a Br atom would considerably stabilize the a 1 (dz2) level relative to the bi(dx2 _ y2) level. This results in a ligand field which is substantially different from the octahedral case and gives rise to the TDLF model. X-ray diffraction studies have shown, however, that the axial Mn--C bond is shorter than the radial Mn--C bonds for both Mn(Co) 5H (28) and Mn2(C0) 10 (39). Presumably this is also the case for Mn(Co) 5Br and thus it may be surmised that the tighter bonding of the axial-CO group will largely destabilize the a 1 (dz2) level causing the a 1 - b 1 splitting to be 171 considerably smaller than expected in the TDLF model. This is the basis of POLF model. In (45) Slater-Condon energy parameters were calculated for various electronic configurations of d6 M(CN) 5xn- complexes with a strong crystal field. These are listed belows Electron Configuration Designation Slater-Condona energy (+ 15 F 0 ) e 4 b~ 1A1 - JO F2 - 60 F4 e4b2b1 1A 2 - .30 F2 - 95 F4 e4b2a1 1B2 1E( 1) - 14 F2 - 175 F4 - 26 F2 - 115 F4 1E(2) - 18 F2 - 155 F4 e3b~a 1 e .3b 2 2b 1 a Configuration interaction between 1E(1) and 1E(2) is neglected. On the basis of electron repulsion the 1A2 and 1E(1) transitions are expected to be the lowest observed ligand field bands. In the case of the TDLF model we expect the lA2 transition to lie at higher energy than the 1E(1) band due to the large splitting of dz2 and dx2 _ y2• On the basis of the POLF framework we expect the lA2 and 1E(1) transition might occur at quite similar energies. Due to the unfavorable electron repulsion terms for 1B2 and 1E(2), both these bands are expected at higher energies. ~ 0.1 .0 w 0 .0 J.-i ro ~ 0 (!) t 1.0 77 K. K• Figure 30. 50,000 40,000 •. v ....\ Path= 1.0 cm (cm-1) 30,000 1.0 mm 6/1 Pentane/Isopentane Solution 20,. 000 ... Concentration =1·5 x 10-J molar Electronic Absorption Spectrum of Mn(Co) 5 Br in Path=~ mm o. ... .. --JOO • t ~ I\) -..J Path = 0.1 mm 40,000 ,; (cm-1)30,000 Path = 1.0 mm I 20,000 Path = 10.0 mm Concentration = 1.45 x 10-3 molar Electronic Absorption Spectrum of Mn(Co) 5Br in CHClJ Solution 50,000 0.1 1.0 Figure 31. <i: .0 w 0 ~ .0 m ~ 0 (!) f J--- I-' ..._,, w t 40,000 11 (cm- 1 ) 30,000 1.0 mm 20,000 10.0 mm Concentration = .6 x 10-J molar Figure 32. Electronic AbsorptionSpectrum of (Et4N)Cr(C0)5Br in CHc1 3 Solution \ Path = 0.1 mm 50,000 --0.1 ~ .0 {I) 0 .0 S-i ro ~ Q) () 1.0 .{::" ~ ~ ( 390) " not obsd. 1A1 -> 1A2 " 22,900 1A1 - > 1E( 1) " 37, 500 1Al -> 1E( J) not msrd. 41,800 Cr( CO~ 22,000 25,800 1Al -> 1E(4) ( JO) 21 1 600 V 'Nk 1 br 25 , 600 not obsd. not obsd. 1A1 -> 1B2 1A1 - > 1E( 1) 1 1 Al -> A2 36,700 J6,600sh (600) 1A1 -> 1E(2) 42,600 cyclohexane soln ( J00°K) 43,100 (19,000) CHCl3 soln (300° K) 1A1 -> 1E(J) Assignment f11n (CO) 5Br ft " " not msrd. 23,100 26,300 not obsd. 37,600 not msrd. 6/1 pentane/isopentane soln ( 77°K) Electronic Absor~"t ion Spectra of M(C0)5xn- in CHCl3, cyclohexane, / 1 pentane/isopentane solutions TABLE 31 t-4 \,/\ -...J 176 The spacing of these ligand field levels should be represented by the following diagram& Octahedral LF Levels POLF Levels ___.....:;.__ -- TDLF Levels ....._ . . . . _ _a.-'-,'--- Band Assignments for Mn(Co) Br 5 The first intense band in the Mn(C0)5Br spectrum is found at 25,600cm- 1 and is assigned to the 1E(1) d-d transition. Recent work on the Mn(C0)6+ complex by Gray and Beach(46) has made untenable a very early assignment (47) of this band as part of the Mn~/T*CO charge transfer system. The second intense band is a shoulder at 36,600cm-1 which is assigned to 1E(2). observed value This is close to the of J8,5oocm- 1 for 1A1g_.,,. 1T1g in Mn(C0)6+ (46) • An extremely intense band at 43,1oocm-l is assigned as the first Mn~'7T*CO charge transfer band, probably 177 of 1E symmetry due to its very large E value. A very weak and broad feature at N 2t,600cm- 1 was observed clearly at liquid nitrogen temperature at 2J,100cm- 1 • The two possible assignments for this band are 1A 1 - . 1A 2 or 1A 1 --"> %( 1) as its low intensity indicates that it should be either orbitally or spin forbidden. The third possibility 1A ~ 1B2 is rejected on the basis of the very large 1 electron repulsion as discussed above. The assignment of this weak band is crucial as it allows us to ascertain the relative importance of both TDLF and POLF models in this system. If the TDLF model is predominant this must be a singlettriplet transition as the 1A2 state would be expected at a higher energy than the observed 1E(1) band; if the POLF model is predominant the nearness of this weak absorption to the 1E(1) band would indicate the very small splitting of the a 1 (dz2) and bl(dx2 _ y2) levelso To confirm either assignment the spectrum of the analogous Re complex should be obtained. Considerable enhancement of this band (as was the case (46) for the absorption at ca. 29,000cm-1 in Mo(C0)6 and W(C0)6) would tend to favor its assignment as the spin-forbidden 178 1A --.. ~(1) transition. A comparable intensity 1 (E 30) would support the 1A2 assignment. l'V It is of interest to note that a POLF model is compatible with the ligand field spectra in the Co(CN) x3- system (48). 5 This type of model has also been invoked to account for the magnetic properties of the Fe(CN) x3- system (49). 5 Band assignments for Cr(C0)5BrThe lower positive charge on the metal atom going from Mn(I) to Cr(O) will result in a considerable energy lowering of the charge transfer system. Thus the two intense bands at 41,800 and 37,5oocm-1 are both assigned as d-Cr--:;>o'Tf*CO charge transfer bands. Due to this lowering 1E(2) is not observed in Cr(Co) Br5 as it is expected to fall underneath the lowest charge transfer band. E(1) was observed at 22,900cm- 1 but the weak band expected at slightly lower energy was not seen. This may be due to either very low intensity or accidental degenracy with 1E(1). Discussion The electronic structure of metal carbonyl complexes has just recently become the subject of many 179 thorough t~eoretical and experimental electronic structural investigations including electronic absorption spectral and electron impact experiments (7, 46, 50, 51), and molecular orbital calculations (46, 52, 53). It is thus of considerable interest to determine what insight into these problems the foregoing vibrational analyses contribute. The conclusions reached in Part I favor the POLF model discussed above for the following reasons: 1. It was found that the calculated interaction constants agree very well with equivalent constants from the Cr(C0)6 octahedral system. This tallies quite well with the observation that although considerable changes in bonding from M(C0)6 have taken place, these changes have been complementary in such a way that the ligand field of Mn(Co) Br has remained very nearly 5 octahedral. 2. Those interaction constants which show a large change from their octahedral equivalents are involved with coordinates whose atomic motion is along or parallel to the four-fold axis of Mn(Co) Br. 5 The relative magnitudes of these constants and the orientation of their coordinates indicate that they may be electrostatic in nature; these differences may thus be thougllt of as external to the resultant ligand field. 180 3. The repeatedly emphasized importance of CJ-bonding effects in the interpretation of trends in potential constants from M(C0)6 to Mn(C0)5Br may be taken as evidence that the changes in 7f bonding and charge distribution caused by substitution of a Br for a CO group are being compensated for, and the departure from an octahedral field for this molecule may be quite small indeed. 181 CHAPTER 3 FINAL SUMMARY AND EVALUATION OF RESULTS As were outlined in the Introduction, the main goals of this work were (1) the evaluation of the assumptions underlying the many approximate force fields presented for metal carbonyls of the form M(C0)5X in the literature since about 1962; (2) the relation of M(C0)5X systems to M(C0)6 systems regarding changes in bonding, also the determination of the highly debated nature of vibrational interactions across the metalmetal bond in M2(C0)10 systems; (J) andfllially the evaluation of several calculations of the metal-metal stretching force constant in Mn2(C0) 10 to determine their relative merits in light of a full normal coordinate analysis and to establish a minimum set of requirements for calculations on other M2(C0)10 systems and more complicated polynuclear carbonyls. The main con- clusions on these points are summarized below. 1. The assumptions involved in the force field proposed by Cotton and Kraihanzel (1) to account for the observed infra-red C--0 stretching modes have been found in this work to be unjustifiable. (1) that The approximation in 182 present value F~. r· ~ 4 Frt· r·• 1' 1 1' 1 Although this does not discount the results of many later studies which make use of the ~tive values of these parameters for a series of M(Co) 5x derivatives, any discussion of the absolute values of these parameters or their dependance on changes in the metal back-bonding d-'7T orbitals as formulated in (1) is incorrect. 2. The fact that Frc· r 1' and that Fri,4 5 i t • is much larger than Fr· r· 1' 1 is much larger than Fri, 135 and Fr 0 ,j for Mn2(C0)10 indicates that interactions among the various CO ligands are primarily electrostatic in nature. The opposite behavior is observed for the metal-carbon stretching coordinates, di and d~ where jFtd. d.j » j F~i• di} for Mn(CO) 5Br and likewise \ Fdi,4 5 + Fdi. l 35 j for Mn 2 (CO) 10 • 1• 1 IF do, j j )) These results support an interaction mechanism involving changing electron populations in the metal d-orbitals of 1T symmetry as proposed in (5). Raman intensity studies and electronic absorption investigations show that considerable changes in the <Tbonding system of M(Co) xn- have occured from M(CO)~n+l)-; 5 the substitution of a halide atom for a CO group is apparently responsible for these effects and they appear to play a major role in the transmission of metal-ligand 183 stretching interactions. These Raman intensity studies reveal further that changes in metal d~7T*CO back-bonding are the dominant factors in transmitting metal-ligand stretching interactions across the metal-metal bond in Mn2 (co) 10 • This is analogous to the mechanism of Cotton (5) which was incorrectly applied to interactions among the CO stretching modes. Both the vibrational and electronic spectral studies indicate a striking similarity in the ligand field of the Mn(Co) 5Br molecule to the octahedral Cr(C0)6 and Mn(C0)6 systems. Apparently the changes in O--bonding mentioned above on the substitution of Br for CO compensate to a large degree for changes in the 'TT electronic structure. It is hoped that this satisfactory correla- tion of results will serve as an incentive for further combined spectral studies in this field. 3. A large amount of interest exists today in the nature of the metal-metal bonding in polynuclear carbonyls. The calculation of accurate metal-metal stretching potential constants is thus of major importance in the interpretation of these systems. This work has shown that the metal-metal stretching coordinate,~, is highly mixed with the C--M--C bending i84 coordinate of Ai symmetry. calculat.e d here, F Comparison of the value = i. 26 md/A, with that of Spiro (37), FMn-Mn = .59 md/A, indicates that a mere estimate of this mixing is insufficient to obtain accurate potential constants. Both Ai modes must be definitely assigned and the mixing constant FA,~-8 allowed to vary for the most accurate fit. Values for these parameters in the Tc 2 (CO)io• Re2(C0) 1 o, and MnRe(CO)io systems should definitely be re-evaluated on the basis of polarized Raman solution spectra before any trends in the metal-metal bonding of these systems or more complicated polynuclear systems are discussed. 185 B I BL I 0 GRAP HY 18E> BIBLIOGRAPHY 1. F. A. Cotton and c. s. Kraihanzal, Inorg. Chem., 2, 533 ( 1963). < 2. F. A. Cotton, ibid., J, 702 (1964). N. Flitcroft, D. K. Huggins, and H. D. Kaesz, ibid., J, 1123 (1964). 4. J. C. Hielman, D. K. Huggins, and H. D. Kaesz, ibid., 1, 933 (1962). F. A. Cotton and R. M. Wing, ibid., ~. 1328 (1965). 6. H. D. Kaesz, R. Bau, D. Hendrickson, and J. M. Smith, J. Am. Chem. Soc., .§.2, 2844, (1967). E. w. Abel, et al., J. Molecular Spec., JQ, 29 (1969). . 8. H. Haas and R. K. Sheline, J. Chem. Phys., !t.Z, 2996 (1967). I. J. Hyams, D. Jonesv and E. R. Lippincott, J. Chem. Soc. (A), 1.2.211 1987. 10. L. H. Jones, R. s. MacDowell, and M. Goldblatt, J. Chem. Phys., 48, 2663 (1968). 11" K. L. Watters, J. N. Brittain, and W. M. Risen, Inorg. Chem., 8, 1347 (1969}. 120 I. J. Hyams and E. R. Lippincott, Spectrochimica Acta, 25A, 1845 (1969). 13 .. w. F. Edgell, J. w. Fisher, G. Asato, and w. M. Risen, 14. L. H. Jones, R. s. MacDowell, and M. Goldblatt, ibid., 8, 2349 (1969); and L.H. Jones, private communicatTon. Inorg. Chem., 8, 1103 (1969). (a) w. A. G. Graham, Inorg. Chemo, 1 1 315 (1968)1 (b) R. P. Stewart and M. D. Treichel, ibid.,· 7, 1942 (1968). 187 16. F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 2nd Ed., Interscience Publishers, New York City, N. Y., 1966, P• 728ff. 17. Abel and Wilkinson, J. Chem. Soc. (A), J:.2i2, 1501. 18. F. A. Cotton, A. Musco, and G. Yagupsky, Inorg. Chem., 6, 1357 (1967). I. s. Butler and H. K. Spendjian, Canad. J. Chem. !±11 4117 ( 1969). 20. R. J. H. Clark and B. c. Crosse, J. Chem. Soc. (A), 12.§.2, 224. (b) D. M. Adams and A. Squire, ibid., 1968, 2817. 21. v. Valent, et al., Inorg. Nucl. Chem. Letters, J, 22. G. Herzberg, Infrared and Raman Spectrar D. Van Norstrand Co., Inc., Princeton, N.J., 1945. 23. L. H. Jones, Inorg. Chem.,£, 1269 (1967). 24. J. H. Schachtschneider, SD-9032-I and SD-9032-VII, Shell Development Corp. (a) 237 (1967,. E. B. Wilson, Jr., J. c. Decius, and P. c. Cross, Molecular Vibrations, McGraw-Hill Book Co., New York, N. Y., 1955· 26. J. c. Decius, J. Chem. Phys., 1§, 241 (1963). 27. L. H. Jones and R. R. Ryan, ibid., ~' 2003 (1970). 28. s. J. la Placa, et al., Inorg. Chem.,~' 1928 (1969). M. A. El-Sayed and H. Do Kaesz, J. Molecular Spec., 2• 310 (1962). 30. D. J. Darensbourg and T. L. Brown, Jnorg. Chem.,1, 959, (1968). 31. G. Bor, Chem. Communications, .12.§.2, 641. 32. J. Lewis, A. R. Manning, and J. R. Miller, J. Chem. Soc. (A), 1966, 845. 188 33. D. J. Parker and M. H. B. Stiddard, ibid., 1966, 34. J. M. Smith, Ph. Do Dissertation, University of California, Los Angeles, Cal., 1966. 695. G. P. Ceasar, R. A. Levenson, and H. B. Gray, J. Am. Chem. Soc., 21• 772 ( 1969). 36. R. Levenson, private communication. 37. c. o. Quicksall and T. G. Spiro, Inorg. Chem., 8, 38. J. Lewis, et al., Nature, 207, 142 (1965). 39. L. F. Dahl and R. E. Rundle, ~ta Cryst . , l.§, 40. H. M. Gager, J. Lewis , and M.• J. Ware, Chem. Communications, 1966, 616. 41. G. o. Evans, w. T. Wozniak, and R. K. Sheline, Inorg. Chem., 2• 979 (1970). 42. H. J. Svec and G. A. Junk, J. Am. Chem. Soc., ~. 2363, (1969). 419 (1963). 2836 (1967). s. L. Segel, J. Chem. Phys., 21• 848 (1969). 44. Abel, Butler, and Reid, J. Chem. Soc. (A), 1.2§J, 2068e H. B. Gray and c. J. Ballhausen, J. Chem. Phys., ]£, 1151 (1962). 46., N. A. Beach and H. B. Gray, J. Am. Chem. Soc., .2.Q, 5713 (1968). H. B. Gray, E. Billig, A. Wojcicki, and M. Farona, Canad. J. Chem . ,~. 1281 (1963). 48. D. F. Gutterman and H. B. Gray, to be published. 49. D. F. Gutterman and H. B. Gray, to be published., 189 50. 51. D. R. Lloyd and E. w. Schlag, Inorg. Chem., 8, 2544 (1969). R. E. Sullivan and R. w. Kiser, J. Chem. Phys., !±2· 1978 (1968). .. 52. I. H. Hillier, J. Chem. Phys.,~. 1948 (1970). 53. A. F. Schreiner and T. L. Brown, J. Am. Chem. Soc., .2Q, 3366 (1968). 190 P R 0 P 0 S I T I 0 NS 191 PROPOSITIONS I. It is proposed to study the adsorption, reaction, and desorption mechanisms of several simple diatomic and triatomic gases on various transition metals and their oxides by means of isotopic fractionation analysis. II. Oriented single crystal Raman experiments are proposed to unequivocably assign the observed spectrum of M2 (C0) 10 and M2(C0)3X2 (where M = Mn or Re and X = Cl, Br, or I). III. Metal-matal interactions have been suggested to explain the observed Raman spectrum of Bi6(0H)~~; it is proposed here to further elucidate the nature of any Bi--Bi interactions through the use of nuclear quadrupole resonance experiments. 192 IV. The reaction mechanisms for the pyrolysis and disproportionation of the simple silanes and halosilanes is very incompletely understood. Fundamental research on this topic is proposed using the techniques of isotopic fractionation analysis of 28si JOsi isotopic moleculese v. It is proposed to obtain the polarized Raman and infra-red spectra of single crystals of K4Mo(CN)a•H20• Accurate assignment of these vibrational spectra will be very beneficial in the characterization of numerous poorly understood derivatives of the eight-coordinate Mo(CN)i- ion. 193 PROPOSITION I (1) Isotopic fractionation study of the catalytic hydrogenation of carbon monoxide. There has been very little use made of isotopic fractionation as a tool for studying reactions catalyzed by metal and metal oxide surfaces. It seems reasonable that since isotopic fractionations are characteristic of the detailed structure of the reactants and intermediates during the course of chemical reactions, investigations dealing with isotopic fractionations should give information regarding the structural details of various heterogeneous catalytic processes. The simple reaction of CO + 2 H2 - CH4 + H2o as catalyzed by Co and Ni metal subgroups affords an opportunity to investigate the isotopic fractionations associated with the reaction of all three elements: H, C, and o. These are elements whose isotopic ratio variations may be measured to within ±0.01% for carbon and oxygen and ±0.1% for hydrogen with contemporary high precision mass spectrometers. Research may be initiated by determining fractionation factors for the adsorption of CO and H2 separately on the metal catalysts mentioned above. This data may then be used in conjunction with the available studies on chemisorbed H and CO species in interpreting fractionation experiments as H2-CO mixtures are catalytically reacted to form CH4 and HzO. The values of the fractionation factors, a.., for each element and their temperature dependence should indicate the relation of rate of adsorption to the rate of reaction on the surface; this is due to the fact that the value of Cl. can be expected to vary depending on whether the adsorption of H2 and CO is in equillibrium with the gas phase or whether the adsorption of one or the other gas is the rate determining step. A study of the temperature dependence of the fractionation involved in each step would provide a measure of the activation energy of that step, and it might also indicate the presence of more than one mechanism by changes in the slope of temperature versus a, or an "impossible0 increase of a. observed with increasing temperature as was found in Reference (1). Previous work (2,3,4) indicates the rate determining step of this process is the formation of H-COH complexes on the surface of the catalyst and it would be very interesting to see how the isotopic fractionation work will compare with this. Even more interesting would be the fractionations involved in the products CH4 and H20. This data will be 195 a key to the mechanisms involved in the reactions of the adsorbed complexes themselves. For even a simple system as this the problems in analysis of the resulting data will still be immense; it is therefore desirable to do precise determinations of CL H' a. c, and CL 0' simultaneously, thus utilizing all the available information from these systems. This multiple approach to the problem may in turn lead to the definite acceptance or rejection of certain of the mechanisms proposed in the literature. (2) Isotopic fractionation study of the catalytic decomposition of nitrogen oxides. A parallel study is anticipated on the decomposition of NO and N20 on metal oxides such as Crzo 3, Fe 2o 3, NiO, CuO, and ZnO at somewhat elevated temperatures with the formation of N2 and Oz. Analysis of the mechanism occurring on the catalyst surface will be more complex than in the preceding case due to interaction (and thus possible fractionation effects) between the oxygen atoms of the nitrogen oxides and metal oxideso Very interesting studies are possible here in terms of examining various layers of the catalyst after reaction for evidence of o18/o 16 fractionation. Also, comparison experiments using the analogous metal sulfides would be an intriguing avenue of research; noticeable differences in o18/o 16 fractionation on the catalyst surface would be valuable in deducing the role of the surface oxygen anions in the adsorption and decomposition mechanisms (as in the recent EPR experiment (5) indicating the existence of the species N02 in NO adsorbed on ZnO and the absence of N02 in NO adsorbed on ZnS). 197 BIBLIOGRAPHY lG Brown, Dole, and Lane; J. Chem. Phys. ,g,z, 251 (1957). 2. Vlasenko, et al.; Kinet. Kataliz. 2• 337 (1964). 3a Kozub, et al.; ibid. 6, 244 (1965). 4. Vlasenko, Yuzefovich, and Rusov; ibid.£, 688 (1965). 5. Lunsford; J. Chem. Phys • .zg,, 2141 (1968). 198 PROPOSITION II A most interesting consequence of the vibrational analysis of polynuclear metal carbonyl systems would be the evaluation of interaction parameters which define the amount of coupling between vibrations originating on different metal centers. This has stimulated much interest and a number of papers have appeared dealing with the metal-metal or metal-bridging group(s)-metal vibrations of these systems. (1·5) These results unfortunately yield only an idea of the strength of the bond between metallic centers and no knowledge of the electron distribution about theme The evaluation of vibrational coupling constants acting between the metallic centers will provide important information through their magnitude and symmetry indicating which molecular orbitals may be instrumental in transmitting vibrational effects from one metallic center to the othero These calculations of coupling constants depend on precise assignment of fundamental vibrations if any accurat~ conclusions are to be drawn at all; and this has stymied workers in this area for some time as the very interesting and complex polynuclear carbo- 199 nyl complexes possess correspondingly interesting and complex vibrational spectra. However, with a highly polarizable laser light source available on contemporary Raman spectrometers and the known crystal structures of many of these systems, one may obtain polarized Raman spectra of their various single crystals; (~-1l) The rigid selection rules which obtain for these systems will provide unequivocal assignment as to symmetry type for most observed modes and thus furnish an excellent starting point for vibrational calculations on these complicated systems. A case in point are the M2(C0)10 and M2(C0)3X2 (M = Mn or Re; X = Cl, Br, or I) complexes; the former involves a direct metal-metal bond and the latter a metal-bridging halogens-metal system. Both types form regular shaped monoclinic crystals (sli,1$'), and preliminary experiments by myself and Mr. Jeffery Hare of this laboratory have produced highly polarized Raman spectra for crystals of Mn2(C0)101 Re2(C0)10• and Mn2(C0)3Br2• The main difficulty in definitive experiments will be the determination of the crystals' precise orientation in the spectrometer's laser beam. The individual M2(co) 1o and M2(CO)aX2 molecules possess almost exact 200 D4a and Dzh point symmetry, respectively, and are arranged in the crystal lattice as shown belows 1 ~ 0 0 0 0 ......., N 10 s:: ::E 0 0 ......., 0 N s:: s G-1 0 1.0 Cl) r-l r-l <l.) 0 .µ ·r-1 s:: ::J G-1 0 s::0 •r-l .µ 0 0 <l.) .-I •r-j 0 0 0 ......., N s:: ::E H p.. • 201 Since the M--M axes and the M<~M axes, respectively, lie parallel to one another in the crystal, one may ignore as a first approximation the lower symmetry of the unit cell's factor group and derive selection rules for ·the molecular vibrations directly from the molecular point symmetry involved. These follow from the appro- priate character tabless TABLE I D2h D4d Ai A2 axx + ayy' azz Rz B1 Alg B1g Rz axy Bzg Ry azx Rx ayz B2 Tz B3g El Tx, Ty Au E2 E3 axx - ayy•axy Rx, Ry axx' ayy' azz B1u Tz Bzu Ty B3u TX Thus if the appropriate orientation of the crystals is achieved, simply rotating the plane of polarization of the incident laser beam relative to a polaroid film used to "analyze" the scattered radiation previous to its entering the spectrometer should produce the following polarization effectss 202 y polrzd. laser beam -x -z f.y Raman Light : j (due to a,YY) @=--t----t-0 ¥:I Figure l· -~toayzl tz Raman Light (due "" Polaroi£ Analyzer (&) Single Crystal Raman Set-Up for Mn 2 (C0) 10 In this figure the molecular orientation in the crystal is given by the x, y, and z axes, and the crystallographic axes are indicated by a, b, and c. · The polaroid analyzer will pass Raman scattered light Cy or ez depending on the orientation of E. To indicate the various experimental orientations involved we use the standard notation given in Reference ( 8 ) ; this is X( ij) X where x = direction of propagation of laser beam i = orientation of electric vector of laser Raman j = " light x = direction of propagation of Raman light II fl 203 Considering then only Raman active modes under n4d symmetry we haves A1 acti~e for x(yy)x and x(zz)x. If IntensitYx(yy)x ) IntensitYx(zz)x then a'}'}> a z:z. for the particular mode in- volved which gives information as to the axial or radial character of the vibration. inactive for x(yz)x. E2 E 3 inactive for x(yy)x, x(zz)x, and x(yz)x. active for x(yz)x. inactive for x(yy)x and x(zz)x. In the case of the M 2 (co) 8x2 systems one must modify figure 2 so that the a crystallographic axis now lies parallel to the laser beam's direction of propagation. This again will align the M(i>M axes perpendicular to the x axis, and under n2h symmetry the following selection rules for Raman active modes may be derived& Ag active for x(yy)x and x(zz)xo The same argument above for the intensities of the two orientations will apply here also. inactive for x(yz)x. B1g inactive for x(yy)x, x(zz)x, and x(yz)x. 204 B2g Bg 3 inactive for x(yy)x, x(zz)x, and x(yz)x. active for x(yz)x. inactive for x(yy)x and x(zz)x. 205 BIBLIOGRAPHY 1. F. A. Cotton and R. M. Wing; Inorg. Chem. !J:., 1328 (1965). 2. I. J. Hyams, D. Jones and E. R. Lippincott; J. Chem. Soc. (London) 1967A, 1987. 3. c. o. Quicksall and T. G. Spiro; Inorg. Chem. · ~, 2 36 3 ( 196 9 ) • F. J. Farrell, V. A. Maroni 11 and T. G. Spiro; Inorg. Chem. 8, 2638 (1969). 5. G. o. Evans, W. T. Wozniak, and R. K. Sheline; Inorg. Chem. 2• 979 (1970). 6. L. Couture-Mathieu and J.-P. Mathieu; Acta Cryst. _5., 571 ( 1952). P. J. Hendra and E. R. Lippincott; Nature 212, 1448 (1966). ~ 8. T. c. Darnen, s. P. s. Porto, and B. Tell; Phys. Rev. 142, 570 (1966). Vo Ananthanarayanan; J. Chem. Phys.!±..§, 573 (1968). 10. L. Colombo; ibid • .±21 4688 (1968). 11. I. Nakagawa and J.L. Walter; ibid • .il1 1389 (1969). 12. B. Fanconi, et. al.; ibid • .il• 3993 (1969). 13. R. K. Khanna, c. w. Brown and L. H. Jones; Inorg. Chem.~; 2195 (1969). 14. L. F. Dahl and R. E. Rundle; Acta Cryst • .12., 419 (1963). 15. L. F. Dahl and c.-H. Wei; ibid. 16, 611 (1963). 206 PROPOSITION III Recently a localized molecular orbital scheme was proposed (1) for the Bi6(0H)f~ cluster compound; the basis for this was a vibrational analysis which indicated that a substantial amount of metal-metal bonding was present. structure of this cation has been deduced (2) by X-ray scattering from a 5.1 molar aqueous solution of its c104 salt and this is illustrated in figure 1. 0 = 0 Q = Bi Bond Distances: Bi-Bi = 3. 71 A Bi-0 87° Figure 1. Structure of Bi6(0H)~~ in Aqueous Solution = 2 • .33 A The 207 The M.o. scheme outlined in (1) is similar in terms of bonding orbitals to that proposed for Ta6c1i~ (3) differing in that there are fewer complications in the Bi compound due to the completely filled 5d shell in contrast to the transition metal complex. Thus the bonding in the Bi--OH cluster may be discussed as a first approximation in terms of M.o.•s con- J orbitals. The · structed solely fro-m the Bi [6s] and [6p empty [6Px] and [6Py] orbitals are used to form the slightly bent (164°) bonds with the OH- ions while the remaining [6s] and [6Pz] orbitals are hybridized to form 6 bonding orbitals which interact at the center of the Bi6 cage. This interpretation is at variance with the usual behavior of the Bi(III) ion (4) which may be described in terms of an "inert" pair of electrons in the [6s] orbital. This is due primarily to the rather large energy separation of 8.8 e.v. between the [6s] and [6pJ levels in the free Bi3+ ion (5)a However, sufficient stabilization might arise from the metal-metal interaction to promote some s-pz hybridization. To estimate this quantity it is pro- posed to obtain and interpret the nuclear quadrupole resonance (NQR) spectrum of Bi6(0H)12(Cl04)6• A fair amount of success has been obtained in the interpretation of NQR results for various metal-halide systems (excluding fluorides) using the "p-shell hole" 208 theory of Townes and Dailey (6). In brief, they showed that the major contribution to the electric field gradient (EFG) about a nucleus exhibiting an NQR spectrum was due to an imbalance in the number of p valence electrons for that nucleus. Effects arising from s-electrons and all filled shells were zero due to their spherical symmetry, and contributions from d- and f-wave functions were judged to be neglegible. This is assumed to be the case here also in view of the equally large energy sep- @ct] aration of the ~P] and levels. The 209Bi nucleus has a spin of I ~ 9/2 and thus it will exhibit an NQR spectrum of 4 lines corresponding to the splittings of the spin states by the EFGa 'l I = +9/2 = +7/2 I = ±5/2 I = +J/2 I = +1/2 I ,, •Ir ' (The selection rules for obsd. NQR frequencies are 6 I = ±1• ) , The separation of these frequencies yields the quadrupole coupling const.eQqzz (where eQ =the nuclear quadrupole moment and qzz = EFG in an axially symmetric field)o qxx In the case of axial symmetry we have also = qyy, and the imbalance in p-electrons is 209 Up = unbalanced p-electrons = (Npx,y - Npz) Npx = No. of electrons in Px, etc. and it can be shown (6) that eQqzz = UpeQqzz(p) where eQqzz(p) is the coupling constant arising from 1 unbalanced p-electron. In reference (7) this was cal- culated from atomic hyperfine structure to be eQqzz(p) = 1500 me/sec. However, before estimating the amount of s-p hybridization, one must consider changes in Up due to the fact that NPx,y f O; i.e., the Bi--0 bond is far from being completely ionic. It may be estimated from Pauling (8) that the ionic character of this bond is 47%; thus NPx,y = (1.00 - o.47) = 0.53, and Up = 0.53Npz• The s-p bonding orbital's wave function may be written as where Ais the fraction of s hybridization; as each s-p hybridized bonding orbital will be filled, the final expression for Up is estimated to be Up= 0.53 - 2 (1 + A2)V2. and expressing the ratio of coupling constants as eQqzz/eQqzz(p) = P 210 we may write Up = 0.53 - (l ! >.._2f,_= p This expression is plotted in figure 2 for values of Afrom .1 to 4.o. This encompasses a fractional s character of 9% to 80% (or consequently a fractional p character of 91·% to 20%). If the p character should fall below 20% then the function is no longer analytic due to the ambiguity of the sign in Up. Thus one must be continually more cautious in regard to interpretation of the amount of hybridization as the value of p falls toward o. p I 0 + .2 + .4 + .6 + .8 +1. +1.2 +1.4 +1.6 +1.8 +2.0 r .75 r .80 A 5.0 r .85 ~ 6.o Figure 2. A versus p Fraction of "s" character - - - .10 rrr r 1 1 1 r 0 .25 .40 .50 .60 .65 .70 .5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 .ts 7.0 B.o I-' I-' N 212 BIBLIOGRAPHY v. A. Maroni and T. G. Spiro; Inorg. Chem. 1, 183 (1968). 2. Ho A. Levy, M. D. Danford, and P. A. Agron; J. Chem. Phys. J1, 1458 (1959). 3. F. A. Cotton and T. E. Haas; Inorg. Chem. J, 10 (1964). 4. L. E. Orgel; J. Chem. Soc. (London) .12221 3815. 5o C. E. Moore, "Atomic Energy Levels," Vol. III, Circular No. 467, N.B.s., u.s. Govt. Printing Office, Washington, D.C., 1958. 6. c. H. Townes and B. P. Dailey; J. Chem. Phys. 1.11 782 7. H. G. Robinson, H. G. Dehrnelt, and w. Gordy; Phys. Rev • .§2, 1305 (1953). 8. L. Pauling, "Nature of the Chemical Bond, .. Jrd Ed., Cornell Univ. Press; Ithica, N. Y., 1960. (1949). 213 PROPOSITION IV Although the chemistry of silanes and their derivatives has been in existence for approximately 100 years, surprisingly little is known regarding the mechanisms involving breaking of Si--H or Si--Si bonds. The smaller silanes and their halogen derivatives are either gases at room temperature or extremely volatile liquids. As such they lend themselves well to mass spectrometric studies, and it is proposed to initiate isotopic fractionation studies of reactions involving these molecules to elucidate the nature of the operative reaction mechanisms. This approach utilizes only natural isotopic abundances and thus avoids the stigma of entirely changing the nature of the system under investigation which is the fate of isotopic enrichment studies. One may also take advantage of the relatively large amount of 30si (N5%) and obtain fractionation factors, a, not only for D/H ratios, but also for 28si/30si ratios. One motivation for this proposal is a recent publication ( 1) indicating very large differences in asi for lunar and terrestial silicates as compared with the Si material contained in fossilized organisms (2,3). These differences are on the order of 5% with an 0 experimental error of± .08%0• As of yet, however, 214 the characterization by means of isotopic fractionation of even the simplest reactions in Si chemistry has not been done. A logical starting place would be with the pyrolysis reaction of monosilane (1) 2SiH4 ~ Si2H6 + H2 and the disproportionation of monohalo derivatives of monosilane liq. (2) Current mass spectrometric research (4) on reaction 1 has includedan isotopic enrichment study involving equimolar mixtures of SiH4 + SiD4 in order to distinguish between 2 proposed mechanisms (5,6) involving, first, the silyl radical SiH4 __. SiH3• + H• H• + SiH4 __..., SiH3• + H2 2SiH3• - Si2H6 and, second, the silane diradical (analogous to carbene) By the absence of SiH 3--SiD3 in the final product they eliminate the first mechanism; however, a large amount of HD is present which should not be according to the second mechanism. This difficulty and the fact that isotopic enrichment itself may have significantly altered 215 the reaction mechanism make isotopic fractionation studies extremely attractive. Also, a large amount of disagreement (7,8,9) over the dissociation energy of the Si--Si bond has existed, thus the a30si;28si will be quite interesting in view of the fact that a Si--Si bond is being formed. The dependence of a on temperature will be a function of the different activation energies for Si--H bond breaking or Si--Si bond forming depending on which step is the rate determining process. A similar amount of conjecture exists for reaction 2; the commonly accepted mechanism for this process (10) involves a polynuclear intermediate This does not, however, explain the complication that this reaction (where X = F) is catalyzed by HF (11). Changes ina. factors for the reaction with and without HF present could indicate that a second preferential mechanism is operative when HF is in the system. Hopefully studies of this nature may yield a firmer understanding of Si--H, Si--X, and Si--Si bond breaking and formation processes and can be extended to the higher silanes and their derivatives, and to reaction in more complicated systems. 216 BIBLIOGRAPHY 2. s. Epstein and H.P. Taylor; Science 1.§1, 533 (1970). J, Reynolds; Geochim. et Cosmochim. Acta J, 224 (1953). 3, D. Tilles; Geophys. Research 66, 3003 (1961). 4o M.A. Ring, et al.; Inorg. Chem. 8; 2033 (1969). 5. H. Niki; J. Phys. Chem. Phys. Jl, 207 (1962). 6. B. H. Mahan and R. Mandal; J, Cham. Phys. Jl, 207 (1962). 7. s. R. Gunn and L. G. Green; Ju Phys. Chem • .§2, 779 8. I. M. T. Davidson and I. L. Stephanson; J. Chem. Soc. (London), A, 282 (1968). 9. W. C. Steele and F. G. AG Stone; J. Am. Chem. Soc. 84, 3599 (1962)G 10. MacDiarmid; Advan. Inorg. Chem. Radiochem. J, 207 (1961). 11. J. K. Wolfe and N. C. Cooke; Abstr. 128th Meeting A. c. s. Minneapolis, 1955, 48M. 1. (1961). 217 PROPOSITION V 4- The Mo(CN) 8 ion has been of continued interest to the inorganic chemist as it is one of the few examples of a stable eight-coordinate complex involving unidentate ligands. Further, a large number of re- lated eight-coordinate compounds have been prepared 4- f ro m Mo(CN) 8 by the substitution of one or more ligands (1, 2, 3, 4). This provides a large potential know- ledge to be gained concerning the energetics of the eight-coordinate state. Unfortunately this potential has largely been unrealized due to the difficulties encountered in characterizing these systems. Vibrational spectroscopy which is usually a great aid in structural characterizations has so far been of only limited value in these investigations due to the large uncertainties surrounding the vibrational spectrum of the parent 4- Mo ( CN) s ion. It is therefore proposed to undertake a complete normal coordinate analysis of this system with the aid of 1 3c and 15N substitution and appropriate polarized vibrational spectra in the solid crystalline state. The structure of K4Mo(CN) 8 .2H 2o has been solved and refin~d by X-ray diffraction (5, 6) and the structure of the anion is shown below in figure 1. 218 I ff I lo Figure 1. Structure of the Mo(CN)~- Ion. The molecular symmetry is D2d and it is found that the spectra of the CN stretching region (6 ex• pected Raman bands and 4 expected I. R. bands) may be assigned under this symmetry ( 7, 8, 9, 10, 11). This indicates that interactions between ions in the unit cell may be ignored thus considerably simplifying the effective spectroscopic selection rules which would obtain for the D2h symmetry of the unit cell's space group (12). It was shown in (9) that excellent crystals of this material may be grown from aqueous solution and it is thus a simple matter to obtain the polarized Raman spectrum (the workers in Reference 9 observed by accident that two of the CN stretching modes showed strong polarizations for their particular experimental orientation, but these results were not discussed or 219 pursued further). On the basis of the prevailing n2d symmetry it will be possible to distinguish between A1 , B2, and E modes. A2 modes are totally inactive and the re- maining B1 modes will polarize similar to the A1 modes. It was also noted in (9) that attempts to obtain polarized I. R. data were hampered by crystals which were too thick. A new technique which is applicable in the case of large crystals is Attenuated Total Reflectance (ATR) spectroscopy (13) which analyzes reflected I. R. light from the surface of a crystal thus making its thickness immaterial. On this basis it will be possible to assign the B2 and E I. R. active modes. Very useful structural information may be obtained from the MCN bending and MC stretching regions of the spectrum which lie from 600 to 250 cm- 1 in general (14). Although modes in this region are very susceptible to intermixing with one another they do possess the advantage of occupying a large energy region compared to the approximately 60cm-1 span in which the CN stretching modes are found. 220 The symmetries of these lower energy modes may be determined from the polarization experiments above, however frequency shifts from isotopic substitution will be necessary to determine the MCN-bend or MCstretch character of each individual vibration. Substitution of 1 3cN- will be most helpful regarding frequency shifts for MCN-bends which should be about 15cm- 1 to lower energy; MC-stretching modes will shift only about 6cm-1 to lower energy (14). Substi- tution of c15N- will aid especially in the determination of MC-stretching character of a given mode as the expected shifts are approximately 6cm-1 for these vibrations while MCN bending motions should shift only about 2cm- 1 (14). With this information in hand a reasonable calculation of the normal coordinates of vibration for K4Mo(CN)a~H20 may be obtained. Vibrational spectro- scopy may then possibly realize a fuller role in the characterization of this and other related systems. 221 BIBLIOGRAPHY 1. 2. 3. 4. W. R. Bucknell and Wo Wardlaw, J. Chem. Soc. (A), .12,gl, 2981. A. w. Adamson and J. R. Perumareddi, Inorg. Chem., !±, 247 (1965). J. R. Perumareddi, z. Naturforch., 21b, 22 (1966). G. Rossman, private communication. J. L. Hoard and H. H. Nordsieck, J. Am. Chem. Soc., 61, 2853, ( 1939). 6. J. L. Hoard and J. 235 (1963). v. Silverton, Inorg. Chem., £, S. F. A. Kettle and R.. V. Parish, Spec.trochimica ~, 8. 21A, 1087 (1965). R. v. Parish, ibid., 22A, 1191 (1966). K. o Hartman and F. A, Miller, ibid., 24A, 669 (1968). 10. v. H. Stammreich and o. Sala, z. Elektrochem., .§.!±, 741, (1960). 11. 12. v. H. Stammreich and o. Sala, ibid., ,22, 149 (1960). w. B. White and B. A. de Angelis, Spectrochimica Acta, 23A, 985 (1967). 13. A. M. Deane, E. W. T. Richards, and I. G. Stephen, §.nactrochimica Acta, 22A, 1253 (1966). 14 .. L .. Ho Jones, Inorg. Chem., 4, 1472 (1965).