This dissertation has been 65—3852 microfilmed exactly as received

FARONA, Michael Franklin, 1935— THIOCYANATO DERIVATrVES OF SOME METAL CARBONYLS.

The Ohio State University, Ph.D., 1964 Chemistry, inorganic

University Microfilms, Inc., Ann Arbor, Michigan THIOCYANATO DERIVATIVES OF SOME METAL CARBONYLS

DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By Michael Franklin Farona, B.S., M ‘.Sc.

******

The Ohio State University 1964

Approved by

Department of Chemistry ACKNOWLEDGMENTS

The author is extremely grateful to his adviser^ Dr. Andrew Wojcicki^ not only for the suggestion of such- an interesting and productive project hut also for his invaluable guidance throughout the course of this research and in the preparation of this manuscript. The author is also appreciative of the funds received from Dr. Wojcicki's National Science Foundation grant. The Department of Chemistry is gratefully acknowledged by the author for salaries received through various teaching assistantships. Finally3 the author is appreciative of the patience and understanding of his wife, Jean. He is also indebted to her for her help in preparing the first draft of this dissertation. VITA January 30, 1935 . . B o m - Cleveland, Ohio

1 9 5 6 ...... B.S., Western Reserve University, Cleveland, Ohio

1 9 5 8...... Military Service, U.S. Army Medical Corps

1 9 5 9 -1 9 6 2 ...... Teaching Assistant, Department of Chemistry, The Ohio State University, Columbus, Ohio

1 9 6 3-1 9 6^ - ...... Research Fellow, Department of Chemistry, The Ohio State University, Columbus, Ohio PUBLICATIONS Farona, M. F., Sweet, T. R., and MacNevin, W. M., "A Chemical Investigation of the Lamprey Eel Venom." Arch. Biochem. Biophys. 9 8, 245 (1962). Gray, H. B., Billig, E., Wojcicki, A., and Farona, M. F., "The Electronic Structures and Reactivities of Mn(C0)j-X Complexes." Can. J. Chem. 4l, 1281 (19&3)* Wojcicki, A., and Farona, M. F., "Thiocyanato-S-penta- carbonylmanganese (I) and Some Derivatives." Inorg. Chem. 3, 151 (1964). Wojcicki, A., and Farona, M. F., "Thiocyanato Derivatives of Chromium, Molybdenum and Tungsten Hexacarbonyls." J. Inorg. Nucl. Chem., In press. FIELDS OF STUDY Major Field: Chemistry Studies in Inorganic Chemistry.' Professor Andrew Wojcicki Studies in Analytical Chemistry. Professor Thomas R. Sweet CONTENTS Page

INTRODUCTION...... '--- 1 HISTORICAL ...... 3 EXPERIMENTAL...... 27 Starting Materials ...... '27 Elemental Analyses and Physical Measurements ..... 34 Preparation of Mn(C0)5CNS ...... 36 Attempted Preparations of Mn(CO)cSCN ...... 38 Preparation of the Derivatives or Mn(CO)cCNS ..... 42 Special Preparations of Some Derivatives ...... 59 Attempted Preparation of [Mn(CO)^CNS]o ...... 6l Preparation of the Thiocyanato Derivatives of the Group VIb Hexacarbonyls ...... 62 Attempted Preparations of Fe(C0)i|(CNS)2 ...... 64 RESULTS AND DISCUSSION ...... 68 Investigation of the Mn(CO)cCNS System...... 68 Investigation of the Derivatives of the Mh(C0)5CNS System ...... 107 The Significance of the Metathesis Reactions ..... 131 Linkage Isomerism in the Derivatives ...... 131 Thiocyanato Derivatives of the Group VIb Hexacarbonyls ...... 135 Investigation of the Thiocyanato Iron Carbonyl System...... •...... l40 BIBLIOGRAPHY ...... 142 TABLES Table Page 1. X-Ray Structures of Thlocyanate Complexes ..... 4 2. Infrared Frequencies (cm-1) of Thiocyanate Group ...... 8 3. Analysis of the Derivatives of Mn(COLONS .... 60 4. Infrared Carbonyl Stretching Frequency Region of Mn(CO)^SCN in Various Media ...... 75 5. Comparison of the CO Stretching Frequencies of Mn(CO)^X Complexes ..... 83 6. The CS Stretching Frequency of Mh(CO)cCNS in Various Solvents ...... 87 7. Change in Band Intensities with Rising Temperature ...... 88

8. Approximate Secular Equations for the CO Stretching Modes in the Point Group Cij.v ..... 102

9 . Force Constants (mdynes/A) for Mn(C0)i-X Complexes ...... 103

10. The A 1 (1) Stretching Modes and Values for MoL(CO)^ Molecules ...... 106

11. Infrared Spectra of cis-[Mn(C0 )3L2 CNS] Complexes (cm-1) ...... 113

12. Infrared Spectra of trans-[Mh(C0 )oL2 CNS] Complexes (cm-1) ...... 118

1 3 . Infrared Spectra of Mh(C0)iiLCNS Complexes (cm-1) ...... 121

14. Infrared Spectra of (CH3 )aj.N[M(C0 )^CNS] (cm-1) .. 138

v FIGURES Figure Page I 1. The Carbonyl and CS Stretching Frequency Regions of the Infrared Spectrum of Mn(CO)5CNS ...... 74 2. The Carbonyl and CS Stretching Frequency Regions of the Infrared Spectrum of Mn^CO^CNS in Acetonitrile ...... 7 8 3. Infrared Spectra of the Carbonyl Stretching Frequency Region Showing the Initial, Intermediate, and Final Bands of the Rearranging Mn(C0 )t5CNS In Acetonitrile ...... 85 4. The Carbonyl Stretching Frequency Region of the Infrared Spectra of Mh(CO)trCNS at Different Temperatures in Ethyl Acetate ..... 89 5. The Carbonyl Stretching Frequency Region of the Infrared Spectra of Mh(CO)cCNS in CH3CN at Different Temperatures ...... 91 6 . Assignment of the CO and CN Bands of Mn(C0 )5CNS in 1 ,2 -dichloroethane ...... 95 7. The Carbonyl and CS Stretching Frequency Regions of the Infrared Spectrum of Mn(C0)3py2CNS ...... 114 8. The Carbonyl and CS Stretching Frequency Regions of the Infrared Spectrum of Mn(CO)3 (pfa)2CNS ...... 115 9. The Carbonyl Stretching Frequency Region of the Infrared Spectrum of trans- .[Mn(CO)3 (AsPh3 )2 CNS] in CHCI3 ...... 1 2 -9 10. The Carbonyl Stretching Frequency Region of the Infrared Spectrum of Mn(CO)wAsPhoCNS in CHC1 3 ...... 7 ... 7 ...... 122 11. The Carbonyl Stretching Frequency Regions of the Infrared Spectra of Mh(COKAsPhqCNS in CHC13 and C H g C N ...... 7 .... 133 12. The Carbonyl and CS Stretching Frequency Regions of the Infrared Spectrum of (CH3 )4N[W(CO)5CNS] ...... 1 39 vi INTRODUCTION

Thiocyanato complexes of intermediate oxidation state transition metals have been studied extensively. The fact that so much work has been done on these systems is not surprising because as a ligand, thiocyanate ion is a very interesting species. For example* it is one of the few monodentate ligands which can coordinate to metal ions in one of three different ways. Thiocyanate ion may bond through either the sulfur or nitrogen ends, or bond simultaneously through both ends when it acts as a bridge between two metal ions. Although much time has been spent on synthesis, the truly interesting features of these complexes have been elucidated by means of various physical methods. For example, the purpose of the majority of the work has been to determine the mode of attachment of the thiocyanato group, and to apply modern theories of bonding to examine the factors which influence this attachment.. Until the work of this thesis was undertaken, no attempt had been reported to prepare and examine thiocyanate derivatives of low oxidation state transition metals. In particular, it has been the purpose of this work to

1 synthesize and characterize thiocyanato derivatives of some metal carbonyls. These systems were expected to be analogous in part to the known metal carbonyl halides. It was* then* of great interest to have compared and contrasted the results of this investigation with those of the corresponding carbonyl halides and the known thiocyanato complexes. HISTORICAL

The thiocyanate ion is known to coordinate to transition metal ions in one of three different ways; through either the sulfur or nitrogen ends or as a bridge between two metal ions, when the sulfur and nitrogen ends coordinate simultaneously. These data have been supported by X-ray studies for the three modes of attachment. (l)3

(2), (3 )5 (^) Table 1 (1) shows some terminal thiocyanato complexes whose structures have been determined by X-ray analysis. Certain facts are noticed immediately from the table. For example, the only listed thiocyanato complexes which are N-bonded are those of the first row transition metals. The table also shows that the S-bonded complexes are those of the second and third row transition metals. This table is not complete and the above assignments do not represent strict conclusions. For example, although essentially all first row transition metals bond only 1 through nitrogen of terminal SCN groups, the second and third row metals may in some instances bond through

Two exceptions have been found for Co(III) which bonds a terminal thiocyanato group through sulfur. The first example is in the vitamin Bto complex (5) and the second example is in the complex K^Co^CNj^SCN. (6) TABLE 1

X-RAY STRUCTURES OF THIOCYANATE COMPLEXES

Stereochemistry Arrangement of (CNS) Compound of Central Atom Groups in Complex Ion Isothiopyanates

[Cr(NH3 )2 (NCS)4 ]- Octahedral Co-planar, M-NC co-linear

[Ni(NH3 )2 (NCS)4]2" Octahedral Co-planar, M-NC co-linear [Ni(NH3 )3(NCS)3r Octahedral M-NC co-linear [Ni(NH3 )4 (NCS)2] Octahedral trans-, M-NC co-linear [Nipy4 (NCS)2 J. Octahedral trans-, M-NC = 165° Thiocyanates

[Rh(SCN)6]3“ Octahedral -SCN group parallel (in two sets), M-SC = 120° [Pt(SC.N)4]2" Square Planar -SCN groups approximately parallel (in two sets) [Hg(SCN)4]2- Tetrahedral M-SC approximately 120° [Ag(SCN)2]- Tetrahedral Ag-SC = 110° nitrogen. The following observations have been reported by Mitchell and Williams (5) for those elements which bond through sulfur:

1 . The elements which can bind terminal SCN through sulfur are Co(III) , 1 Rh(I), Pd(II), Pd(IV), Pt(II), Pt(IV) 9 Au(III), Ag(l), and Hg(ll). 2. Thiocyanate ion cannot be bound through sulfur by elements earlier than those of Group VII. 3. Weak bonding to sulfur is also possible for the cations Co(II), Ni(II), and Cd(II), but the bonding is restricted to bridged structures. Several resonance forms may be envisaged when the thiocyanato group bonds to a metal. The bonding extremes for N- and S- attachment may be written as M-N=C=S and M-S-C=N, or as M-N=C-S and M-S=C=N. Since divalent sulfur is preferred over uni- or trivalent sulfur on the grounds of more uniform charge distribution, it seems likely that the first set, namely M-N=C=S and M-S-C=N, represents the two most important extremes. It will be noticed from the above resonance forms that a significant difference exists between the bond orders in the two cases. For example, more double bond character is expected for the CN linkage in N-bonded thiocyanates, whereas more triple bond character should exist for CN in S-bonded species. On the other hand, the 6 CS bond order is observed to increase from that of one for the S-bonded thiocyanates toward that of two for the N- bonded ones. These resonance forms lead one to expect, therefore, that for N-bonded thiocyanates the CN stretching frequency should be lower and the CS stretching frequency higher than in the corresponding S-bonded complexes. Evidence in support of resonance forms in free- thiocyanate ion was brought forward by Jones (7) 5 who carried out infrared studies on KCNS. His results were the following:

^(CN) = 2 0 6 6 cm- 1 £ = 400 t>2 " ^70 cm- 1 = 5 ^)3 (CS) = 743 cm"1 = 11 2 V 2 = 9 41 cm- 1 = 6

He was able to show from infrared data than the CN bond distance in KCNS is 1.17 A. Using these data and

Pauling's (8) formula relating the amount of single, double and triple bond character, he came up with the following contributing resonance forms for free thiocyanate ion:

(a) N=C-S” 71# ‘ (b> ~N=C=S 12io -2 t (c) N-C-S 17#

However, when thiocyanate ion is coordinated to a metal the following resonance forms (1 ) are possible:

(d) M-NsC-S~ (g) M^S-C=N (e) "M=N=C=S (h) M/S=C=N"

(f) m /N=C=S (i) ~M=S=C=N~ 7 Bond angles and bond distances obtained from X-ray studies show that in the case of N-bonded thiocyanates form (d) or (f) is the most significant, whereas in the S-bonded species form (g) is important. These results are consistent with predictions made on the basis of the resonance structures (a), (b), and (c). Because the metal to which the thiocyanato group is attached has a formal positive charge, electron drift will occur toward the coordinating atom of the ligand. Thus for N-bonded complexes, structure (f) will be stabilized, and for S-bonded species, (g) will be favored. Therefore, for an S-bonded ligand, it would be expected that a bent MSC configuration will occur, whereas for an N-bonded thiocyanate, either a linear (d) or a bent (f) structure is possible for the MNC moiety. Several investigators have reported infrared work on the thiocyanato complexes. (1 ), (5), (9 ), (1 0 ), (1 1 ) Table 2 shows some of their results. It is possible to compare at least three different infrared regions from the results in Table 2: the CN stretching frequency region, the CS stretching frequency region and the NCS bending frequency region. Theoretically, it should also be possible to measure the MS or MN stretching frequency, but as of the time of this writing, no work has been reported on this aspect with reference to thiocyanato complexes. TABLE 2

* INFRARED FREQUENCIES (cm-1) OF THIOCYANATE GROUP

NCS (Lending) Compound CN (stretch) CS (stretch) (fundamentalj

Isothiocyanates

K3 [Cr(NCS)6]*4h20 2 1 0 5s 820vw 476w Tl3 [Cr(NCS)6] 2 10 6s 82 Ow > NH^C Cr(NH3) 2 (NCS )i|] 2120s,2 0 5 5sh 823w 466w pyH[Cr(NH3 )2 (NCS)4 ] 2 0 8 5s 8361*7 choline[Cr(NH3 )2 (NCS)^] 2 0 8ls (879w ) K2 [Cr(NCS)i|]*^H20 2 0 9 5s,h,2 0 65sh 820vw 475m K^NifNCSjgMHgO 2 1 2 3sh,2 1 0 1 s 766w 469m K2 [Zn(NCS)j|] *4H20 2 10 1sh,2 0 7 6s 815w 470m

Thiocyanates K3 [Rh(SCN)6] 2 1 10sh,2 0 9 8s,2084s 705w 471,452,438w Tl3 [Rh(SCN)6] 2 0 8 8s,2 0 6 6s 706w 465,428w K2 [Pd(SCN)4 ] 2 1 2 2 s,2 1 l6sh,2098sh, 708sh,703w,694sh ^73^65,439, 2 09 2 sh,2 0 8 8s 43 Ow K2 [Cd(SCN)2i_] *2H20 2 1 1 2sh,2 0 9 3 s 754w,726w 466,458,453,^17w K2 [Pt(SCN)6] 2 1 2 8sh,2 1 2 2 s 6 9 9 sh,694w, 686sh 464,426,4l5w 2115s,2075w K2 [Pt(SCN)4 ] 2 1 2 7s,2 0 9 8s 705sh,700w,692sh

S = strong, m = medium, sh = shoulder, w = weak, v = very, h = broad.

00 9 After examining the resonance forms presented above (p. 6 ), one might expect a clear-cut difference in the CN stretching frequency between N- and S-bonded thiocyanate complexes. However, the difference is not at all straightforward. Although S-bonded complexes in general have a higher CN frequency than N-bonded ones, there is a significant overlap between the two. A general range of the CN stretching frequency for thiocyanates (S-bonded) is from 2060 to 2l4o cm-1. For isothiocyanates (N-bonded), it is from 2070 to 2120 cm-1. Normally, however, most thiocyanates exhibit a CN stretching frequency above

2100 cm-'1', whereas most iso thiocyanates absorb below

2100 cm-'*". Examples of overlap may be seen in the table. There is little difference in the CN frequencies between

Tl3 [Cr(NCS)6] and K3 [Rh(SCN)g] and between

Cj-H[_NH[Cr(NH^)g(NCS)^] and Tl3 [Rh(SCN)g] . Indeed, according to the CN stretching frequency, NHj|[Cr(NH3 )2 (NCS)^3 appears to be S-bonded, which is contrary to all other existing evidence, including X-ray data. Another interesting observation on differences between N- and S-bonded thiocyanates is apparent from Table 2. All of the M-SCN complexes containing at least two thiocyanate groups exhibit two main CN stretching frequency bands in the solid state. This is to be expected because for S-bonding, the bent MSC attachment leads to 10 lower symmetry of the molecule. This lowered symmetry causes additional modes to become active in the infrared since the thiocyanato groups are no longer equivalent within the complex. However, for MNCS bonding, there is only one main band, which broadens in solution (no splitting). Obviously* if only one SCN group is coordinated to a metal, then only one CN band can occur regardless of the mode of bonding. There has been a great deal of interest in the problem of the CN stretching frequency and the factors which cause it to be non-diagnostic in determining the mode of bonding of the thiocyanate group. Mitchell and Williams

(5) have contributed much of the information on this subject. They mention that organic thiocyanates (RSCN) may be differentiated from isothiocyanates (RNCS) by comparison of their respective CN frequencies. The thiocyanates absorb around 2140 cm""1 whereas the corresponding isothiocyanates exhibit CN frequencies from 2060 to 2105 cm""1. However, there are additional factors to be considered for inorganic thiocyanates. For example, one must take into account the coordination number and stereochemistry of the complex in question. It may be seen that octahedral and square planar thiocyanato complexes of the same metal do not necessarily absorb at the same frequency. Comparing the CN frequencies of octahedral 1 1

K2 [Pt(SCN)g] with square planar ^[PtCSCN)^], one finds _n that the former absorbs at 2 1 1 5 cm * whereas the latter absorbs at 2 0 9 8 cm-1. One must also consider the nature of the cation B + which neutralizes the charge in complexes of general formula B+nM(CNS)n“ . For example* the strong bands

for K.^Rh(SCN)g are at 2 0 9 8 and 2084 cm"'1’, whereas for Tl^Rh(SCN)g they are at 2088 and 2066 cm-1. The oxidation state of the cation Mn+ is also a contributing factor which causes the CN stretching frequency to be different over a series of complexes. Here* one will notice the difference

between K2Pt(SCN)^ (2 0 9 8 cm-1) and K2Pt(SCN)g (2 1 1 5 cm"1 ). The following trends were reported for the CN stretching frequency in thiocyanato complexes. 1. The change of the cation* B+ * from a group IA to another group XA* or to a univalent organic base such as pyridinium ion* affects the CN stretching frequency. Comparing NH^ + and pyH + as cations* the N H ^ -‘r4- containing species has a higher CN frequency. 2. An increase in the number of CN's in a complex causes a decrease in the CN stretching frequency. 3. If the oxidation state of the central metal ion is increased* then this seems to increase the CN stretching frequency. However* this relation is not so clear because a change in oxidation state is usually accompanied by other changes such as stereochemistry. 12

4. A change in the cation Mn+, along a row in the periodic table from left to right, brings about an increase in the CN stretching frequency. For example, KCNS and Ba(CNS)2 have CN stretching frequencies about 2020-2050 cm"'*', whereas thiocyanates of Ca(II), Ni(ll), Pd(II),

Pt(II), absorb from 2080 to 2 1 3 0 cm-1. 5. A change down a group usually brings about an increase in the CN vibrational frequency. Thus, there is a small increase from Ni(II) (2080-2100 c m ”'*') to Pt(II) (2110-2120 cm"1). With reference to a change of the cation B+ , it has also been noted by Fujita et al. (12) that the CN stretching frequency of the complex is sensitive to this counter-ion. They point out that Hg[Cr(NH^^NCS)^^ exhibits two bands,

at 2160 and 2 08 0 cm-1, whereas (CH3)3N(CH2)20H[Cr(NH3)2(NCS)4 ]

shows only one band at 2083 cm-1. This difference in the values of the CN stretching frequency led them to suggest

that the band at 2 1 6 0 cm"1 perhaps was due to an interaction which may be pictured as Cr-N+ =C-S •••*Hg2 + . In the other compound, such an interaction is negligibly small, for the positive charge of the cation is screened by the bulky . aliphatic groups. The interesting consequence of Hie above interpretation is an intuitive speculation that the bridged thiocyanates should exhibit a higher CN frequency than either of the terminal ones. Evidence in support of this 13 a priori statement has "been reported by Chatt et al. (2) The results of their infrared studies on bridged thiocyanate compounds show that the bridged CN frequencies are from 30 to 80 cm”^ higher than terminal CN frequencies. Two main conclusions may be stated concerning the

CN stretching frequency: 1. The CN stretching frequency becomes higher in the order MNCS, MSCN, BSCNM when all other factors are constant. 2. For a given mode of bonding (S or N) of the thiocyanate ion, the more polarizing the cation (i.e., the greater its charge or its electronegativity, or the smaller its size), the greater the frequency of the CN vibration. The NCS bending frequencies also offer a possibility of distinguishing S- from N-bonded thiocyanates. For N-bonded compounds, there is a medium-intensity band around 470 cm this band is shifted to lower wave numbers and intensities for S-bonded species. There is also a wider separation and a greater number of bands for S-bonded species. Again, however, there is a certain amount of overlap between the two; even more serious, there is a danger of observing metal-ligand stretching frequencies in this region. Furthermore, many infrared instruments are not equipped to give a resolution in this low frequency range. fortunatelyj the 0.1 stretching frequency la sat; Is factory for differentiating Isnthlocyanatas from thiocyanates, Most ,1-bonded species exhlhlt a weak hawl at

6^0-720 ciri-^ j v/hereas the Isothloeyanateo absorb at yhO- _ •) 860 cm , (11) 'thus the 0,1 stretching frequency 1s clearly diagnostic In distinguishing between N-bonded and bonded thiocyanates, A higher value for N-bonded forms Is obvious, when one considers the previously discussed resonance forms

(see p , 6),

hieetronlc spectra have also been useful In determining the mode of bonding. There 1s a significant difference 1n ligand field spedra for N- and .1 -bonded species, with the N-bonded form absorbing, at a lower wavelength,

Yama.oa and Tsuchlba (J1) observes that the absorption maxima of coualt (31 and 13 J ) aud ebroml um (XT. and III) tb 1 ocyarai.te complexes occur at a higher energy

than these of palladium (XT. and IV) arc platinum (11 and

TV), whose spectra were comparan I e to loose of the

correspond i.ng chloride compl exes . They therefore «• <>n.c3.u.ded

that the investigated compounds of tn.o 3"? rst- row Iransi tioh

elements were N-bonded, a/id those of the second and in i ro

row transi tlon elements were .1-bonded.

The spectrochemlca]. series lists ilgan.bs in the

following order of i/m/'easing; energy: Nr”, <;i , - ~.

Jh;0, , WJh,, <;ctahebral rhodium (1X1) '•uiayl.ez.es illustrate the similarity "between the visible absorption bands of an S-bonded thiocyanate and the corresponding chloride. The ligand field bands for K^[Rh(SCN)^j and

K^[RhClg] occur at 2 0 ,0 0 0 cm ^ and 1 9 * 3 0 0 cm \ respectively. However, for N-bonded thiocyanates, the frequency is often high enough (near 24,000 cm sc tha the ligand field band may be covered up by a charge tran band. (1) At any rate, it is possible to determine the mode of bonding by comparing the electronic spectrum f .? thiocyanate or isothiocyanate with that of a corresp-ndir chloride complex. Many theories have been advanced to explain a particular metal’s preference for the sulfur or nitrsrer. ends and the factors which influence the mode cf attachm? Williams (l4) has developed a general empirical equatim (vide infra) which distinguishes between ionic versus covalent bonding in metal complexes. The equation is ba on the ionization potential of the central metal ion. Applying this to thiocyanato complexes, Lewis (1) made t* assumption that:

1. In the thiocyanate ion the lone pair of electrons on the sulfur atom is more easily polarized ch that on the nitrogen atom. 2. The permanent lone pair dipole on the nitr^r atom is larger than that on the sulfur atom. 16

3. The way in which the thiocyanato group is bound (through S or N) will be decided by the relative bond energies of a covalent M-S and of the more ionic M-N bond. To compare the relative strengths of ionic and covalent bonds, the following equation was used:

R = I.P. r

Here R is the ratio of the strength of a bond if covalent to its strength if ionic; I.P. is the ionization potential of the metal in volts, r is the radius of the metal in A, and ne is the formal charge on the metal ion. Using the above equation for thiocyanato complexes, a number between 5 and 20 is obtained, the higher values of R indicating covalent bonding and hence an MS link. For lower R, a more ionic bond is preferred, and therefore an MN link is expected. This formula has been very useful for uni- and bivalent ions, but not for trivalent species. Other theories have stressed the importance of over­ lap between particular sets of atomic orbitals. Chatt et al. (1 5) have considered the effect of backbonding from the cation to the ligand through the TT-orbitals. The highly polarizing cations would then prefer the S-linkage whereas those which are not highly polarizing prefer a more ionic bond in the metal-nitrogen linkage. The authors classify those metals whose relative bond strength with the halide ions is decreasing in the order F“, Cl“, Br", I", as class 17 (a) metals^ and those whose relative bond strength is reversed as class (b) metals. They note that the change from M-NCS to M-SCN bonding coincides approximately with the change from class (a) to (b) metals in the periodic table. Mitchell and Williams (5) point out that Chatt’s treatment leads one to expect a large stabilization of t^ as opposed to e electrons, and the highly polarizable g ligands should produce a large spectroscopic splitting. However, this is not observed; the heavier halides and the S-bonded thiocyanate are low in the spectrochemical series. These authors offer an alternative explanation in terms of the nephelauxetic series which places ligands in the order of their ability to cause the core electrons to spread. Therefore, the highly polarizing metal ions will prefer the more polarizable atom, namely, sulfur. The most recent explanation has been offered by Pearson (16) who suggests that sulfur in thiocyanate is soft and will prefer to coordinate with soft acids, whereas nitrogen in thiocyanate is hard and coordinates with hard acids. Soft acids correspond to the class (b) metals while hard acids correspond to class (a) metals. The terms soft and hard used by Pearson essentially mean polarizable and nonpolarizable, respectively.

An extremely interesting fact was pointed out by Turco and Pecile. (11) For Pd(II) and Pt(II) complexes, coordinated thiocyanate is either S- or N-bonded depending on the nature of the other ligands present. Thus,

[M(SCN^]2~ and M(NH3 )2 (SCN) 2 are S-bonded, but

M(PR3 )2 (NCS) 2 (R=C2H5, CH3CHCH3) is N-bonded. The authors point out that there must be a critical point where, with a particular ligand, change in coordination can occur. It should then be possible, in the case of complexes with general formula M(CNS) 1^ to have two linkage isomers of the same compound, depending on the correct choice of the ligand L and the ratio m:n. This observation provided Basolo and Burmeister

(1 7) with a novel approach to the synthesis of linkage isomers of the thiocyanate group, theretofore unknown. They succeeded with triphenylarsine as the ligand; the route of synthesis follows:

[Pd(SCN)J + 2AsPh. 2SCN EtOH 'As Ph.

Ph SCN 150 C.^ Ph„As JNCS 5 J?d Nos ^AsPh. 30 min. SCN vAsPh. 3 3

Similar isomers were prepared when 2,2'-bipyridine was the other neutral ligand. 19 In attempting to prepare thiocyanato derivatives of various metal carbonyls, one must first examine the analogous metal carbonyl halides. Since thiocyanogen, (SCN)2, is a pseudohalogen, one may expect the preparation of thiocyanato derivatives to be similar to the preparation of the carbonyl halides. The chloride and bromide derivatives of manganese carbonyl are prepared (1 8) by direct reaction between the halogen and dimanganese decacarbonyl according to the equation:

Mn2 (c°)l0 + X2 (X=ClaBr) --- > 2Mn(C0)5X

The chloride is synthesized at O^C. whereas the bromide is prepared at 4o0C. The iodide, Mn(CO)^I, cannot be prepared in this manner, however, because iodine is not a strong enough oxidizing agent at these conditions. Two methods are known for its preparation. In the first method (19)> Mn2 (C0) and I2 are heated in a sealed tube at 120°C., and the resulting crystals are separated by fractional sublimation. This method produces Mn(C0)^I in low yields. A much better method (2 0 ) is via the preparation of the sodium salt of manganese carbonyl, which may be obtained (2 1 ) by the reduction of Mn2 (C0 )^ 0 with sodium amalgam.

^ 2 ^ C° ^10 + 2Na ~ a(Hg)> 2NaMn(C0) 5 20

Addition of iodine then gives the product according to the equation:

NaMn(C0)5 + Ig > Mn(C0)5I + Nal

Halide derivatives of the iron carbonyl system are also well known. The tetracarbonyls are prepared (22),

(23) simply by the addition of the halogen to iron pentacarbonyl at low temperatures.

Fe(C0)j_ + X2 ----> Fe(C0)^X2 + CO

When the halogen is iodine, this reaction is known to proceed by way of the pentacarbonyl, Fe(C0 )^l2 , intermediate:

0°C. n°r Fe(C0)5 + I2 ----- > Fe(C0)j_I2 — — > Fe(C0)i|I2 + CO

Halide derivatives of the Group VIb carbonyls are also known and have been synthesized recently by Abel et al. (24) The authors produced the anionic species M(C0)p.X (M = Cr, Mo, W; X = Cl, Br, I) by heating the hexacarbonyls with the appropriate trialklammonium halides at temperatures in excess of 100°C. in diglyme. Thus:

RjjNX + M(C0)g 12P. C») R2+N[M(CO)5X] + CO

Derivatives of the above halide systems where various numbers of CO groups have been replaced by neutral ligands are known for manganese and iron. Since the 21 manganese system turned out to be by far the most important one examined in this thesis, and also because the manganese system is the best characterized with respect to its derivatives, it is the intention here to present a brief review of only manganese carbonyl derivatives. Normally, in a Mn(C0)^X system, one or two CO groups, but not more, 1 may be replaced by various neutral ligands.

Thus compounds of the type Mn(CO)^LX and Mn(CO)3L2X have been prepared. For example, Hieber and Schropp (26) have prepared amine derivatives of the type Mn(CO)3L2X and Mh^CO^L'X

(where L = aniline, pyridine; L' = 2,21-bipyridine, o-phenanthroline; and X = Cl, Br, I), by simple addition of an excess of the neutral ligand to the halide at room temperature. Similar derivatives, where the neutral ligands of the disubstituted product are triphenylphosphine and -arsine were prepared by Abel and Wilkinson. (17) These were synthesized in a manner similar to that of the amine- substituted complexes, the reaction proceeding at higher temperatures. Other disubstituted products where the ligands are P(OPh)^a PPhClg, ^(C^Hg)^ and P^C^H^)^ were prepared by Basolo £t al. (27)

1M. Kilner was able to replace more than two CO's when he synthesized ionic complexes such as [Mn(CO)2 (PPh3)2(bipy)] NO^ (25) 22

Monosubstituted derivatives where the ligands are triphenylphosphine, -arsine and -stibine have been prepared by two independent groups. (28), (29) These reactions may be effected at room temperature using stoichiometric amounts of the carbonyl halide and ligand according to the following equation:

Mn(CO)5X + L --- » Mh(C0)jjLX + CO

In addition to substitution of CO by neutral groups, replacement of CO has also been effected by negative ligands. (30) Thus, tetraalkylammonium halides were found to react directly with the manganese carbonyl halides in diglyme at 120°C. according to the following equations (X and Y represent different halide ions):

R^NX + Mn(C0)5X ----- > R^N Mn(C0)^X2 + CO

R^NX + Mn(C0)5Y R^N Mn(C0)i+XY + CO R^NY + Mn(C0)5X

The pentacarbonyl halides may also be caused to

react with themselves (18) at elevated temperatures to form dimers of structure: X (CO^Mn,; >ln(C0)4 X where the bridges are exclusively halides, and never CO. These halides can now react with ligands mentioned above 23 - to give products of substitution which are identical to those obtained from the monomers. One particular derivative of the iron carbonyl system should be mentioned because it represents the first reported octahedral thiocyanato derivative of a metal carbonyl. The compound was prepared by Booth and Chatt (31) and is formulated as F e ^ C S ^ C O ^ P E t ^ ^ * The substituted carbonyl was designated as an isothiocyanate on the basis of a strong CS stretching frequency band at 824 cm"'1'. In summary, it is pertinent to mention the points of extreme interest in attempting to synthesize thiocyanato derivatives which are similar to the carbonyl halides. The most obvious point of interest is the determination of the mode of attachment of the thiocyanate group in these derivatives. Because of several opposing factors influencing the nature of bonding in such complexes, it is extremely difficult to predict a priori the mode of attachment in a thiocyanato metal carbonyl derivative. For example, a general rule is that in halide complexes whose stabilities increase in the order I“, Br , Cl", nitrogen bonding is favored for the corresponding thiocyanato system. In the Mh(CO)^X system, qualitative observations by the author indicate that the stability against the reverse reaction increases in the above order. On this basis,

2Mn(C0),_X----- » ^ ( C 0 );^ + Xg 24 nitrogen■bonding would be expected for the corresponding Mn(COLONS system. However, it is also known that in complexes which contain strongly polarizable ligands, the sulfur end is normally preferred because it is more easily polarized than nitrogen. On this basis one would predict S-bonding for thiocyanato metal carbonyls because CO is a strongly polarizing ligand due to its ability to pi-bond effectively with low valent transition metals. Turco and Pecile (11) have shown that the replacement of two thiocyanato groups in the S-bonded 2_ [M(SCN)^] (M = Pt,Pd) compounds by pi-bonding ligands such as triethyl- or tri-isopropylphosphine brings about a change from the thiocyanato-S to the thiocyanato-N configuration. Replacement of the same thiocyanate groups by ammonia does not result in rearrangement to the N-bonded form. The authors explain the change from sulfur to nitrogen bonding on the following basis. The sulfur end is assumed to participate in pi-bonding with the metal through use of its antibonding orbitals and empty 3d orbitals. The nitrogen end is not believed to participate in pi-bonding with the metal. Strong pi-bonders such as the trialkylphosphines make the t2g electrons on the metal less available for bonding with the pi-orbitals on the sulfur, and hence the metal-nitrogen linkage is preferred. 25 Basolo et al. have pointed out that the observation above may not be strictly correct because one must take into account steric factors as well when considering the changeover between metal-sulfur and metal-nitrogen linkages. The authors point out an example where the change from sulfur- to nitrogen-bonded thiocyanate can be effected by increasing the steric hindrance of the ligands. Sterically less hindered Pd(dien)SCN+ (dien = is S-bonded while the complex

Pd(Eti|dien)NCS+ (Eti^dien = (C2H5)2NC2% m C 2H4N(C2H5)2) is N-bonded, the pi-bonding ability of the two ligands being essentially the same. One also wonders whether substitution of CO by other neutral ligands will bring about a change in the mode of attachment of the thiocyanato group. If the mode of bonding in the derivatives depends on the nature and/or position of the other ligands present, it is then of importance to assess the factors responsible for these variations. Another point of definite interest is the novelty of preparation of linkage isomers in a metal carbonyl system. The hope that linkage isomerism is possible in these systems is strongly supported by the closing state­ ments of the Turco and Pecile (11) paper. . . . our observations suggest that Tr-bonding may be of importance in balancing the equil­ ibrium M-NCS ,=* M-SCN far to the right for cations such as Pd(II) and Pt(ll). If our interpretation is correct, the same change in coordination of the thiocyanate as observed 26

for the complexes with the phosphines may be predicted to occur by different, strong iT-electron acceptor ligands. Finally, the route of synthesis is of interest. The question arises whether thiocyanogen is a strong enough oxidizing agent to react directly with Mn2 (C0)^0 or whether it will be necessary to employ NaMn(CO)^ as an intermediate. The purpose of this work was to synthesize and characterize thiocyanato derivatives of the metal carbonyls. This thesis describes the first reported attempt to investigate thiocyanato derivatives of low oxidation state transition metals and to elucidate the factors which affect the bonding of the thiocyanato group in these systems. EXPERIMENTAL

Starting materials Metal carbonyls. The metal carbonyls used in this work were obtained from several sources. Dimanganese decacarbonyl was obtained as a gift and purchased from the Ethyl Corporation. Additional small amounts were obtained as gifts from Prof. H. B. Gray of Columbia University and Dr. J. Ibers of Brookhaven National Laboratory. Manganese carbonyl was purified by vacuum sublimation (0.01 mm.j 50cC.) before use. Chromium hexacarbonyl was obtained as a gift from Prof. Gray, whereas molybdenum and tungsten hexacarbonyls were donated by the Clinax Molybdenum Co. The hexa­ carbonyls were used as obtained without further purification. Iron pentacarbonyl was purchased from Antara Chemicals, and used as obtained without further purifi­ cation. Mercury tetracarbonylferrate(-II), HgFe(CO)^, was prepared by the method of Hoch and Stuhlman (32),'in which 4,0 g. (2 mmoles) of Fe(CO) were shaken at room 5 temperature with 7.2 g. (2.4 mmoles) of HgSO^ in 25 ml. of 10$ HgSO^. The resulting dark yellow precipitate was collected on a filter and washed with dilute EDjSO^,

27 28

2N HC1, HgOj and acetone and dried' in a vacuum desiccator over CaCl^. Triiron dodecacarbonyl, was prepared according to the method of King and Stone (33) whereby 42 ml. (60 g., 0.3 mole) of Pe(CO)^ in 170 ml. of methanol was reacted with 45 g. of NaOH in 90 ml. of water. The resulting NaHPe(CO)^ solution was buffered with 125 ml. of a saturated solution of NH^Cl, and freshly prepared suspension of MhO^ (from 67 g. (0.4 mole) of KMnOj| in 300 ml. of water and 100 ml. of 95$ ethanol) was added. The excess MnO£ was decomposed by addition of

40 g. of FeSO^^B^O in 250 ml. of 2N sulfuric acid. The mixture was then treated with 300 ml. of lrl'H^SO^ and stirred until the black precipitate appeared. The solid was collected and washed with hot dilute HgSOi}., 95$ ethanol and pentane. It was stored wet under nitrogen. Ligands, solvents, other reagents, and chromato­ graphy supports. Triphenylarsine, p-chloroaniline, p- toluidine, 4-picoline and tetramethylammonium hydroxide pentahydrate were obtained from Eastman Organic Chemicals as reagent grade compounds and were not purified before use. The reagent grade chemicals triphenylphosphine, triphenylstibine, 2 ,2 '-bipyridine and p-fluoroaniline were obtained from Matheson, Coleman and Bell, and were used without further purification. All solvents were of 29 analytical reagent grade, except those mentioned below along with their respective methods of purification. Tetrahydrofuran (THF) and diethyleneglycoldi- methylether (diglyme) were purified by distillation from

LiAlH^ under a nitrogen atmosphere immediately before use. Nitromethane was purified by first shaking the solvent with 1M NaHCO^ until there was no more evolution of CC>2 . It was then dried for 15 hours over CaCl2 and fractionally distilled. Only the fraction which boiled at 101°C. was collected. Nitromethane obtained in this manner was suitable as a solvent for conductivity measurements. Acetonitrile was purified for spectroscopic and conductivity use in the following manner. The solvent

(500 ml.) was stirred over CaP^ (2.5 g.) for 2k hours. After refluxing for 2 hours over CaK?, it was distilled into a flask containing (2.5 g.) and this mixture was refluxed for 2 hours and distilled back into CaHg

(2.5 g»), where it was refluxed for 2 more hours. The distillate from the second CaH2 treatment was collected.

In each distillation, only that fraction which boiled at 84* C. was collected.

Nitrobenzene for cryoscopic measurements was obtained from Mr. James Walther of The Ohio State

University Chemistry Department. It was always stored and

transferred in a dry atmosphere. 30

The alumina which was used as a support for column chromatography was purchased in the acidic , neutral and basic forms of the "Chromatographic Grade" from The Arthur H. Thomas Co. and used without further deactivation. Florisil, also a chromatography support* was donated by Prof. P. Gassman of The Ohio State University. Thiocyanogen. The materials used in the prepar­ ation of a solution of thiocyanogen were lead thiocyanate and bromine. Lead thiocyanate was prepared (34) as follows. To 49 g. (0.5 mole) of potassium thiocyanate in

100 ml. of water at 0°C, were added 82 g. (0.3 mole) of lead nitrate in 300 ml. of water. The reaction proceeds according to the following equation:

Pb(N03 )2 + 2KSCN ---- > Pb(SCN)2 + 2KN03

The white precipitate was collected on a Buchner funnel

and washed with 100 ml. of water in 20 ml. portions, and

then by 100 ml. of 95# ethyl alcohol, also in 20 ml. portions. The resulting lead thiocyanate was dried in a

vacuum desiccator over CaCl2 in the dark. This compound was always stored in the dark and only used so long as it stayed perfectly white. Fresh samples were prepared when

the stored Pb(SCN)2 showed signs of turning yellow. A solution of thiocyanogen, (SCN)2, was prepared immediately before use. In a typical preparation, a

solution of 0.46 ml. (1.36 g., O.OO85 mole) of bromine in 31 4 ml. of hexane and a suspension of 5.3 g. (0.017 mole) of Pb(SON)g in 35 ml. of n-hexane were made up. The above operations were carried out under nitrogen and in carefully dried glassware. The reactants were then transferred to a dry, nitrogen-atmosphere box, and the bromine solution was added in 1 ml. portions to the suspension of Pb(SCN)2 . Only after the previous portion of bromine was decolorized was the next portion added. The reduction of bromine is shown in the following equation:

Pb(SCN)2 + Br2 --- > PbBr2 + (SCN)2

The colorless thiocyanogen solution was filtered from the excess Pb(SCN)^ and resulting PbBr2 into a 250 ml. round bottom flask. It was used immediately upon preparation. A solution of thiocyanogen, prepared in this manner, was usually colorless. However, occasionally a light pink or yellow solution was obtained. The pink color is due to a trace of moisture, and the yellow color comes about through a slight polymerization of (SCN)2 . (34) These weakly colored solutions exhibited normal chemical properties; more intensely colored solutions were discarded.

Thiocyanogen chloride. Thiocyanogen chloride, C1SCN, was prepared (35) j (36), immediately before use by quickly adding a solution of chlorine (0.54 g., 0.0 077 32 moles) to the (SCN)2 solution prepared as above from 0.46 ml. (0.0085 moles) of bromine and Pb(SCN)2 . The chlorine solution was prepared by bubbling Cl2 through concen­ trated HgSO^ and collecting it in 4 ml. of n-hexane at dry ice-acetone temperatures. The n-hexane was cooled to -78°C. in a 10 ml. Erlenmeyer flask. The flask was then wiped dry and weighed rapidly, then returned to the cold 1 bath. Chlorine was bubbled in and the flask was removed from the bath, wiped dry and weighed rapidly. This process was repeated until the desired weight gain was attained. The resulting solution of C1SCN was always bright yellow with a pungent odor. The following equation shows the reaction to produce C1SCN:

Cl2 + (SCN)2 --- » 2C1SCN

Tetramethylammonium thiocyanate. Several materials were necessary for the preparation of (CH^)jjNCNS,"*■ which was produced by neutralizing a solution of tetramethylammonium hydroxide with hydrothiocyanic acid. The acid was prepared by the reaction between barium thiocyanate and sulfuric acid.

^It should be mentioned that the thiocyanate group will hereafter be written in one of three ways in con­ junction with a metal or cation. It may be written as MSCN, MNCS and/or MCNS. The first, MSCN, indicates s'ulfur bonding, while MNCS denotes nitrogen bonding to a metal. The third, MCNS, indicates that the mode of attachment of the thiocyanate group is either not known or that it is irrelevant to the discussion at that point. 33 Barium thiocyanate was prepared according to the method of Herstein (37) whereby 76 g. (1 mole) of NH^SCN and 158 g. (0.5 mole) of Ba (OH) 2‘SligO were allowed to react together and then boiled to expel the ammonia. The resulting solution was filtered through Celite on a Buchner funnel. The solution was neutralized with dilute

H^SO^ (6n) and filtered as above. The filtrate was heated until the boiling point reached 125°C. Cooling in an ice

bath brought about crystallization, and the Ba(SCN)2 was collected on a Buchner funnel and dried in the air. A stable solution of hydrothiocyanic acid, HSCN,

was prepared (3 8) by treating 60 g. (0.24 mole) of barium

thiocyanate in 375 ml. of 1 2 .5$ isopropyl alcohol solution in water with 53 ml. of 4.5 M HgSO^. The resulting fine precipitate of BaSO^ was filtered from the solution of the acid using Celite on a Buchner funnel. This solution was prepared immediately before use. Tetramethylammonium thiocyanate, (CH^^NCNS, was prepared from (CH^ijNOH'SI^O by neutralizing the hydroxide with HSCN. Thus, 25 g. (0.l4 mole) of (CH^NOH'SHgO was dissolved in 50 ml. of water and titrated with the HSCN solution to the end point of blue litmus paper. An

excess of 1 ml. of the acid solution was added, and the solvent was removed by means of a rotary evaporator under vacuum (0.01 mm., 80°C.). The crude (CH^)^NCNS was purified by recrystallization from hot absolute ethanol and dried in a vacuum desiccator over CaCl2 . Anal. Calcd.

for (CH^NCNS: C, 45.45; H, 9.09$. Pound: C, 45.365

H, 9.16.

Elemental analyses and physical' measurements~~ Elemental analyses were carried out by the Alfred Bernhardt Microanalytical Laboratories of the Max Planck Institute, Mulheim, Germany, The Galbraith Laboratories Inc., or the Schwarzkopf Microanalytical Laboratory. Manganese was determined spectrophotometrically as

permanganate, using periodate as the oxidizing agent. (3 9 ) In this method, the compound to be analyzed was decomposed

in boiling 3N HNO^, and the final concentration was

adjusted to about 1 mg. of manganese per ml. of solution. The Mn2 + was oxidized to MnO^” by boiling for 451 minutes with KI0j|. After dilution, the extinction coefficient was determined at 525 millimicrons on a Beckman DU spectro­ photometer. Molecular weight measurements were done by the author by cryoscopy in nitrobenzene. Infrared spectral studies were carried out on a Perkin-Elmer Model 337 Grating Infrared spectrophotometer. Solution spectra were taken.in a NaCl cell of 0.1 mm. thickness, a variable thickness reference cell being used when necessary. Spectra of solid samples were taken as either Nujol mulls or thin films. The Nujol was dried , 3 5 over silica gel. In thin film technique, the compound to he studied was dissolved in a solvent such as dichloromethane or chloroform, a drop of solution was placed on one NaCl plate, and the solvent was evaporated, leaving the pure compound behind as a thin film. This process was repeated until the desired amount of product was left on the plate. Visible and ultraviolet spectra were taken on a Cary Model 14 recording spectrophotometer using matched quartz cells which were exactly 1 cm. on a side. Conductivity measurements were carried out by the author on a Model RC 16B2 Industrial Instruments Company conductivity bridge, using a cell with a constant of 1 . 566. Magnetic moment studies were done by the author by the Guoy method in acetone solutions. The correction for the diamagnetism of the tube and acetone was determined as follows. The empty tube and the tube filled with acetone were each weighed three times' in and out of the field of the magnet. The average value obtained from the three weighings was used. The magnetic properties of the carbonyl complexes were determined in a similar way. For infrared temperature dependence studies the solution cell was placed in a small oven which was

constructed from a piece of copper 2 -3/^ inches wide and 36 bent to fit exactly around the cell and holder. The oven was heated by means of a nichrome wire and -the temperature of the solution within the cell was approximated by placing the thermometer directly on the window of the cell. Gas evolution studies were carried out using a calibrated gas buret. A reaction vessel for these

measurements was constructed from a 250 ml. Erlenmeyer flask fitted with a 24/40 tapered ground glass joint. The bottom of the flask had a partition in the middle to provide two compartments of about 50 ml. each. The reactants were placed in the two separate bulbs and the flask connected to the.^gas. buret. After equilibrium had been reached, the contents of the two bulbs were mixed by vigorous shaking. The volume of the gas evolved from the ensuing reaction was measured by displacing mercury in the gas buret.

Preparation of Mn(COLONS

The successful preparation (40) involved the reaction between NaMn(CO),- (21) and C1SCN. The reaction

was carried out in a 100 ml. three heck round bottom flask fitted with a nitrogen inlet, a condenser and a mechanical stirrer. The flask had a 1 ml. bulge in the bottom to facilitate removal of the excess amalgam. The complete operation was carried out under a nitrogen 37 atmosphere. After the flask had been flushed out with nitrogen* 4 ml. (5^ g . 3 0.37 mole) of mercury was admitted and then 0.45 g. (0.026 mole) of sodium was added in about 0;1 g. portions. Considerable heat and fumes result from the amalgamation of the sodium. Local concentrations of solid amalgam were broken up with a glass rod. After the contents in the flask had cooled to room temperature* a solution of 3*0 g. (0.0077 mole) of

Mri2 (C0 )10 in 50 ml. of freshly distilled THF was added and the mixture was stirred for one hour. During the course of this reduction* the initial yellow-orange solution became essentially colorless to faintly yellow-green.

The reaction between sodium and Mn2 (C0)^Q proceeds according to the following equation:

Na(Hg) Mn2 (C0 ) 10 + 2Na ------> 2NaMn(C0)5

The excess amalgam was removed from the flask by means of a hypodermic needle and syringe* the needle being inserted into the bulge at the bottom of .the'flask. The

solution was "washed" with 2 ml. of mercury which was then removed as above. All reactants were then transferred to the dry box. The solution of C1SCN was prepared as described previously (p. 31) by reacting 0.54 g. of chlorine in hexane with a solution of (SCN)2* also in

hexane (prepared from 0.46 ml. (O.OO85 mole) of bromine and excess Pb(SCN)2 . The solution of NaMn(CO)^ was then 38' filtered into the C1SCN solution. Formation of a white precipitate, shown later to he NaCI, was observed, and the solution gradually changed from bright yellow to yellow orange. After all the THF solution of NaMn(CO)^ had been added, the reaction flask and contents were removed from the dry box and the solvent was evaporated (0.01 mm., 25°C.). The residue was taken up into 25 ml. of chloroform and the solution was filtered. The volume of the orange chloroform solution was reduced to about 5 ml. by means of a stream of nitrogen. Slow addition, with stirring, of 200 ml. of low boiling (30-6o°C.) petroleum ether afforded a golden yellow precipitate which was washed several times with petroleum ether (30-60 C.) and dried (0.01 mm., 25°C.) for 30 minutes.

Anal. Calcd. for Mn(C0)5CNS: C, 28.47; N, 5-53; S, 12.67;

Mn, 21.70; 0, 3 1.60; Cl 0.0$; Mol. wt., 253. Found: C, 28.04; N, 5.50, 5.53; S, 12.55; Mn, 21.60; 0, 31.55, 31-53, 31-30; Cl 0.0$; Mol. wt. (cryoscopic in nitro­ benzene) 247, 259; Yield, 2.73 g. (70$); decomp. pt.

1 5 6°C.

Attempted preparations 6 f TfoTCoy5£ c ' r ------

The reaction between Mh2(CO)io and (SCN)2 » The direct reaction between M n ^ C O ) ^ and (SCN)2 was attempted using a solution of (SCN)2 prepared from 0.46 ml. (I.36 g., O.OO85 mole) of bromine and 5.3 g., (0.017 mole) of 39

Pb(SCN)2, in 30 ml. of freshly distilled THF. This solution was added dropwise to 15 ml. of THF containing

1 g. (0.0026 mole) of M r ^ C O ) ^ . The preparation was carried out in a three neck flask equipped with a nitrogen inlet, a condenser and a mechanical stirrer. No apparent reaction occurred during the first hour at 25°C., and therefore the temperature of the reaction mixture was raised to 45°C. At that temperature an orange-red solid precipitated slowly from solution. The solid was collected on a filter and dried (0.01 mm., 45°C.). An infrared spectrum of this solid (KBr pellet) showed no carbonyl bands. Upon evaporation of the filtrate (0.01 mm., 45°C.), a yellow solid remained. The infrared spectrum of this solid was shown to be identical with that of Mn^CO)-^. The reaction between NaMn(CO)^ and (SCN)2 . A second method which was tried employed the use of NaMn(CO)^ and (SCN)2 . The preparation of the THF solution of NaMn(CO)^ was identical to that described previously

(p• 37)* 3 .0 g. (0 .0077 moles) of Mn2 (C0)^0 being used.

A solution of (SCN)2 in 50 ml. of THF, prepared.from 0 .9 2

ml. (2.72 g., 0.017 mole) of bromine in 8 ml. of THF and

10.6 g. (0.034 mole) of Pb(SCN)2 in 50 ml. of THF, was added dropwise to the NaMn(CO)^ solution under nitrogen. 4o

The solution changed from essentially colorless to yellow-orange. The solvent was removed (0.01 mm., 25°C.) and the residue was taken up into 25 ml. of chloroform and filtered. Addition of 200 ml. of low boiling (30- 60°C.) petroleum ether did not give a precipitate. This solution was cooled to -7 8^C. and still no precipitate was obtained at this temperature. The solvent was then removed (0.01 mm., 25°C.) and 2 .8 g. of a yellow solid was recovered. An infrared spectrum (KBr pellet) indicated that the yellow compound was Mr^CO)-]^. A third preparation which was attempted was exactly identical in all respects to the method described above for NaMn(CO)^ and (SCN^a except that the THF solution of NaMn(CO)^ was filtered into the (SCN^ solution. The solvent was removed (0.01 mm., 25°C.) and

the residue taken up into 25 ml. chloroform and filtered. After the volume of the chloroform was reduced to 10 ml., 200 ml. of low boiling (30-60°C.) petroleum ether afforded a golden yellow precipitate in some experiments and in some did not. The golden yellow product was filtered

from the solution and dried (0 .0 1 mm., 25°C.) for 30 minutes. The product was identified as Mn(C0)^CNS by 41 means of its infrared spectrum, which was identical to the spectrum of the product obtained in the reaction between C1SCN and NaMn(CO)^. In the cases where no precipitate was afforded, the solvent was removed (0 .0 1 mm., 25°C.) and remaining solid was identified as

Mn2 (CO)io by infrared spectroscopy. The reaction between Mn(00)^01 and KCNS. An additional attempt at the synthesis of Mn(C0)^CNS was the proposed metathetical reaction between Mn(C0)^Cl and KSCN. Chloropentacarbonylmanganese(I) was prepared by the method of Abel and Wilkinson (18) in which chlorine

was bubbled into a solution of 3*0 g. (0 .0077 mole) of

Mn2 (C0 )1Q in 75 ml. of CCl^ at 0°C. for 1 hour. The solvent was removed (0.01 mm., 30°C.) and the remaining residue was washed with low boiling (30-6o°C.) petroleum ether (5 x 10 ml.) and dried (0.01 mm., 25°C.). The

carbonyl chloride was purified by vacuum sublimation (0 .0 1

mm., 5 0°C.). To the carbonyl chloride (2.3 g., 0.01 mole) in 35 ml. of methanol was added 9.7 g. (0.1 mole) of KSCN in 5 ml. of HgO. The mixture was stirred with a mechanical stirrer at 50°C. under nitrogen for 2 hours. The solvent

0 was drawn off (0.01 mm., 40 C.) and the yellow residue was taken up into 25 ml. of K^O. Addition of 4.4 g.

(0 .0 105 mole) of tetraphenylarsonium chloride afforded a yellow precipitate which was washed well with water and 42 dried in a vacuum desiccator over CaCl2 . An infrared spectrum indicated that this compound was not Mn(CO),_CNS and therefore no further studies were undertaken.

Preparation of the derivatives**- of Mn(C0 )5CM cis-Thiocyanatodipyridinetricarbonylmanganese(I),

Mn(CO)3py2 CNS. To 0.05 g. (0.2 mmole) of Mn(C0 )5CNS in a

32 mm. test tube, which was wrapped with masking tape to protect the reaction mixture from light, was added 0 .0 1 ml. (1.2 mmole) of pyridine in 5 ml. of chloroform. The reaction started immediately as was evidenced by CO evolution, and was allowed to proceed at room temperature for 12 hours. The reaction mixture was filtered and the chloroform and excess pyridine were evaporated under a stream of nitrogen. The residual solid was dissolved in

2 ml. of chloroform and chromatographed on a neutral alumina column, 30 cm. in length and 2 .0 cm. in diameter, using chloroform as the eluent. The solvent was evaporated under a stream of nitrogen, the final traces being removed by high vacuum (0.01 mm., 25°C.), leaving

0 .0 7 g. (100# yield) of the pure yellow powder.

*] -'-The terms cis and trans in the naming of the compounds below refer to the relative positions of the two ligands other than CO or CNS. When the ligands are cis, they are also cis to CNS, whereas when the ligands are trans, they are most likely (27) cis to CNS. 4 3

The reaction of Mn(CO)^CNS with pyridine was also carried out with the purpose of measuring the volume of the liberated CO. The carbonyl thiocyanate (0.0628 g.,

0.248 mmole) in 8 ml. of 2 ,2 -dibromoethane was placed in one of the two mixing compartments of the flask described earlier (p. 36). Into the other bulb was added a solution of 0.2 ml. (2.4 mmole) of pyridine in 4 ml. of 1 ,2 - dibromoethane. This flask was then connected to the calibrated gas buret and 45 minutes were allowed for the system to come to equilibrium at 28°C. and 738 mm. Hg. The contents of the flask were agitated to afford good mixing between the metal carbonyl and pyridine solutions.

After 6 hours, there was no further noticeable displace­ ment of mercury in the gas buret. The mixture was then allowed to stand for an additional hour to ensure completion of the reaction and also to ascertain that there was no decomposition of the product. This derivative was also prepared from the meta- thetical reaction between M^CO^pygCl and KCNS. From Mn(C0)cjCl, prepared as described earlier (p. 4l), the

derivative Mn(C0 )2Py2cl was synthesized as follows. To 0.22 g. (1 mmole) of Mn(C0)^Cl in a 32 mm. test tube was

added a solution of 0..5 ml. (6 mmoles) of pyridine in 15 ml. of chloroform and the reaction was allowed to proceed

for 12 hours at room temperature, after which time the solution was filtered and the solvent evaporated to 44 dryness by means of a stream of nitrogen. The resulting yellow solid was not purified further and was identified as the desired derivative by infrared spectroscopy. For the preparation of Mh^CO^pygONSj 0.33 g.

(1 mmole) of Mr^CO^pygOl were dissolved in 30 ml. of methanol and treated with 1 g. (10 mmoles) of KSCN in 4 ml. of water. The reaction was carried out in a three neck flask fitted with a nitrogen inletj a condenser and a mechanical stirrer. The mixture was stirred at 50 °C. for 8 hours. After that time* the solvent was removed

(0.01 mm. j 50°C.), the residue taken up into 8 ml. of chloroform., and the solution filtered from the resulting KOI and excess KSCN. The volume of the chloroform solution was reduced to 3 ml. Purification was effected by chromatography as described above for this compound (p. 42)a the corresponding chloride remaining on the column. The solvent was then evaporated (0.01 mm. 3 25°C.)

to give 0 .2 5 g. (70$ yield) of the pure yellow compound.

cis-Thiocyanato-2^21-bipyridinetricarbonyl- manganese(I ) 3 Mn(C0 )3bipyCNS. To 0.05 g. (0.2 mmole) of Mn(C0)cjCNS in a 32 mm. test tube was added 0.08 g. (0.5 mmole) of 2^2T-bipyridine in 5 ml. of chloroform. The reaction was allowed to proceed for 12 hours at room temperature in the dark. The solvent was evaporated under a stream of nitrogen and the excess 2 ^2 '-bipyridine was removed by washing (5 x 20 cc.) with low boiling (30-6o°C.) 45 petroleum ether. The compound was dried under a stream of nitrogen. The residual solid was taken up into 2 ml. of dichloromethane and chromatographed on a neutral

alumina column (30 x 2 .0 cm. ), eluting with dichloro­ methane. The solvent was evaporated (0.01 mm., 25°C.) to

give 0 .0 7 g. (100$ yield) of the pure yellow powder. This derivative was also prepared from the metathetical reaction between Mh^O^bipyCl and KCNS. From Mn(C0)^Cl, prepared as described earlier (p. 4l), the derivative Mn(C0)^bipyCl was synthesized as follows. To 0.22 g. (1 mmole) of Mn(C0),_Cl in a 32 mm. test tube

was added a solution of 0.4 g. (2 .5 mmoles) of 2 ,2 f-

bipyridine in 15 ml. of chloroform and the reaction was

allowed to proceed for 12 hours at room temperature, after which time the solution was filtered and the solvent evaporated to dryness by means of a stream of nitrogen. The resulting yellow solid was not purified further and was identified as the desired derivative by infrared spectroscopy. For the preparation of Mr^CO^bipyCNS, 0.33 g. (1

mmole) of Mn^CO^bipyCl were dissolved in 30 ml.of methanol and treated with 1 g. (10 mmoles) of KSCN in 4 ml. of water. The reaction was carried out in a three neck flask fitted with a nitrogen inlet, a condenser and a mechanical stirrer. The mixture was stirred at 50°C. for

8 hours. After that time, the solvent was removed 46

(0.01 iran. 3 50°C.),.the residue taken up Into 8 ml. of chloroform, and the solution filtered from the resulting KC1 and excess KSCN. The volume of the chloroform solution was reduced to 3 ml* Purification was effected by chromatography as described above for this compound (p. 45), the corresponding chloride remaining on the column. After elution of the band from the column, the solvent was evaporated (0.01 mm., 25°C.) to give 0.25 g.

(70$ yield) of the pure yellow compound. cis-Thiocyanatobis-p-toluidinetricarbonyl- manganese(I), Mn(C0)-g(p-tol^CNS. To 0.05 g* (0.2 mmole) of Mn(C0)^CNS in a 32 mm. test tube was added 0.1 g. (1 mmole) of p-toluidine, followed by 5 ml. of chloroform. The reaction was allowed to proceed for 12 hours at room temperature whereupon the solution was filtered and the solvent evaporated under a stream of nitrogen. The excess p-toluidine was removed by several (5 x 20 ml.) washings with low boiling petroleum ether. The solid vias then dried (0.01 mm., 25°C.), to give 0.82 g. (100$ yield) of the pure yellow derivative. cis-Thiocyanatobis-4-picolinetricarbonyl-

manganese(I), Mn(C0 )3 (4-pic)2CNS. For this reaction bromopentacarbonylmanganese(l) was used as the starting material, and was prepared according to the method of Abel and Wilkinson (18) in which Mn^CO)-^ was allowed to react with excess bromine at 4o°C. in ,CC1^ under nitrogen. 47 After the solvent was removed (0.01 mm., 25 C.) the product was purified "by vacuum sublimation (0 .0 1 mm., 4o°C.) yielding pure yellow-orange Mn(C0)^Br. To 0.3 g. (1.1 mmoles) of Mn(C0)^Br was added 0.53 ml* (5*5 mmoles) of 4-picoline. The reaction was allowed to proceed in a 32 mm. test tube at room temperature for 12 hours. The solvent and excess 4-picoline were removed (0.01 mm., 50°C.), and the resulting yellow solid was identified as the desired product by infrared spectroscopy. For the synthesis of Mn(C0)^(4-picJ^ONS, 0.46 g. (1.1 mmoles) of Mn(C0)3(4-pic)2Br was dissolved in 30 ml. of methanol and transferred to a 100 ml. three neck flask fitted with a nitrogen inlet, a condenser and a mechanical stirrer. To the solution was added 1.0 g. (10 mmoles) of KCNS in 4 ml. of water. The resulting mixture was stirred

for 8 hours at 6o°C. under nitrogen. The solvent was removed (0.01 mm., 6 o cC.), and the residue was taken up into 10 ml. of chloroform and filtered from the resulting KBr and excess KSCN. After its volume was reduced to 3 ini* by means of a stream of nitrogen, the solution was chromatographed on a 30 cm. column of acidic alumina using CHCl^ as the eluent. The metathesized thiocyanato derivative band moves more rapidly down the column than does the corresponding bromide band. The solvent was

removed (0.01 mm., 4o°C.) and the resulting solid was washed with low boiling (30-60^.) petroleum ether and 48 redried (0.01 mm., 25°C.). The yield of the pure yellow powder was 0.26 g. (60$). This derivative was also prepared from the reaction of Mn(C0)^CNS with 4-picoline. Thus 0.05 g. (0.2 mmole) of Mn(C0)^CNS in a 32 mm. test tube was

treated with 0 .1 ml. (1 mmole) of 4-picoline in 5 ml. of chloroform. The reaction was allowed to proceed for 12 hours at room temperature, after which time the solution was filtered. The solvent and excess liquid ligand were

evaporated (0.01 mm., 6o°C.). The remaining yellow solid was not purified further, but was identified by means of infrared spectroscopy. cis-Thiocyanatobis-p-chloroanilinetricarbonyl-

manganese(l), Mn(C0 )3 (pca)2CNS. The corresponding bromide,

Mn(C0 )2 (pca)2Br, was prepared by allowing 0 .3 g. (1 .1

mmole) of ^(COj^Br to react with 0 .7 7 g. (5 .5 mmoles) of

p-chloroaniline in 5 ml. of chloroform in a 32 mm. test tube for 12 hours at room temperature. After the solution

was filtered and the solvent evaporated off (0 .0 1 mm., 25°C.), the excess p-chloroaniline was removed by

successive washings (5 x 20 ml.) with pentane, and the resulting yellow solid was dried (0.01 mm., 25°C.). The structure of the compound was elucidated by infrared spectroscopy. For the preparation of M^COj^pca^CNS, 0.52 g.

(1.1 mmole) of Mn(C0 )2 (pca)2Br was dissolved in 30 ml. of 49 methanol and transferred to a three neck flask fitted with a nitrogen inlet and condenser. To this solution was added 1.0 g. (10 mmoles) of KSCN in 4 ml. of 1^0.

This solution was stirred magnetically for 8 hours at

6o°C. under nitrogen, after which time the solvent was evaporated (0.01 mm., 50*C.). The residue was taken up

into 10 ml. of chloroform and the solution was filtered

from the resulting KBr and excess KSCN. After its volume

was reduced to 3 ml- hy means of a stream of nitrogen,

the chloroform solution was chromatographed on a 30 x 2 .0

cm. column of Plorisil using a 50-50 (v/v ) mixture of chloroform-acetonitrile as the eluent. The thiocyanato derivative band moves first, followed by the band of the unreacted bromo derivative. The compound was dried (0.01 mm., 50eC.) to yield 0.3 g. (65$) of the desired yellow powder. This derivative was also prepared from the reaction of Mn(C0)^CNS with p-chloroaniline. Thus 0.05 g. (0.2 mmole) of Mn(C0)^CNS in a 32 mm. test tube was

treated with 0 .1 5 g. (1 mmole) of p-chloroaniline in 5 ml. of chloroform. The reaction was allowed to proceed

for 12 hours at room temperature, after which time the solution was filtered. The solvent was evaporated (0.01 mm., 25°C.), and the product was washed with low boiling petroleum ether, and dried in a stream of nitrogen. The 50 resulting yellow solid was not purified further, but was identified by means of infrared spectroscopy. cis-Thiocyanatobis-p-fluoroanilinetricarbonyl- manganese(l), Mn(C0 )3(pfa)2 CNS. The corresponding bromide, Mn(C0 )3 (pfa)2CNS was first prepared by allowing 0.3 g. (1.1 mmole) of Mn(C0)^Br to react in a 32 mm. test tube with 0 .5 2 ml. (0 .6l g., 5*5 mmoles) of p-fluoroaniline

in 5 ml. of chloroform for 12 hours at room temperature. The solution was filtered and the solvent and excess p-

fluoroaniline were removed (0.01 mm., 60°C.). The resulting yellow powder was identified as the desired product by infrared spectroscopy.

In the preparation of Mn(C0 )3 (pfa)2CNS, 0.48 g.

(1 .1 mmole) of the corresponding bromide, MnfCOj^pfa^Br,

was dissolved in 30 ml. of methanol and transferred to a

100 ml. three neck flask fitted with a nitrogen inlet, a condenser and a mechanical stirrer. To the methanol solution was added 1.0 g. (10 mmoles) of KSCN in 4 ml. of

H2O. The resulting mixture was stirred for 8 hours at

6 0°C. under nitrogen, after which time the solvent was drawn off (0.01 mm., 50°C.), and the residue taken up

into 10 ml. of chloroform and filtered from the resulting KBr and excess KSCN. After its volume had been reduced

to about 3 ml. by means of a stream of nitrogen, the

chloroform solution was chromatographed on a 30 x 2 .0 cm. Florisil column using a 50-50 mixture (v/v) of chloroform-acetonitrile as the eluting solvent. The thiocyanate complex moves faster than the corresponding hromo derivative during chromatography. The solvent was removed (0.01 mm., 50°C.) and 0.27 g* of a yellow-green powder was obtained, representing a 60$ conversion. This derivative was also prepared from the reaction of Mn(C0)^CNS with p-fluoroaniline. Thus 0.05 g. (0.2 mmole) of Mn(C0)^CNS in a 32 mm. test tube was.

treated with 0 .1 ml. (1 mmole) of p-fluoroaniline in 5 nil. of chloroform. The reaction was allowed to proceed for

12 hours at room temperature, after which time the solution was filtered. The solvent and excess liquid ligand were removed (0.01 mm., 60°C.). The resulting yellow solid was not purified further, but was identified by means of infrared spectroscopy. trans-Thiocyanatobistriphenylphosphinetricarbonyl-

manganese(I), Mn(C0 )g(PPh3 )2CNS. To 0.05 g. (0.2 mmole) of Mh(C0)^CNS in a 32 mm. test tube was added a solution

of 0.26 g. (1 mmole) of triphenylphosphine in 5 ml. of chloroform. The reaction was allowed to proceed for 12 hours at room temperature whereupon the solution was filtered and the chloroform was removed by means of a stream of nitrogen. The resulting residue was taken up in

2 ml. of benzene and chromatographed on a basic alumina column, with benzene first to remove the small amount of the mono-substituted derivative. The bis-substituted 52 derivative was then eluted with dichloromethane and dried by means of a stream of nitrogen. The yellow powder weighed 0.137 g» (95$ yield). This derivative was also prepared from the metathetical reaction between trans-Mr^COj^PPh^gCl and KCNS. Prom Mn(C0)^Cl, prepared as described earlier

(p. 4l), the derivative trans-Mn(C0 )^(PPh2 )2cl was synthesized as follows. To 0.22 g. (1 mmole) of Mn(C0)ejCl in a 32 mm. test tube was added a solution of 1 .3 g. (5 mmoles) of triphenylphosphine in 15 ml. of chloroform and the reaction was allowed to proceed for 12 hours at room temperature5 after which time the solution was filtered and the solvent evaporated to dryness by means of a stream of nitrogen. The resulting yellow solid was not purified further and was identified as the desired derivative by infrared spectroscopy. For the preparation of trans-Mn(C0 )^ (PPh^ ^CNS *

0.7 g. (1 mmole) of trans-MnfCC^^PPh^^Cl was dissolved in 30 ml. of methanol and treated with 1 g. (10 mmoles) of KSCN in 4 ml. of water. The reaction was carried out in

a three neck flask fitted with a nitrogen inlet5 a condenser and a mechanical stirrer. The mixture was

stirred at 50 °C. for 8 hours. After that time* the solvent was removed (0.01 mm., 50°C.), the residue taken

up into 8 ml. of chloroform^ and the solution filtered from the resulting KC1 and excess KSCN. The volume of 53 the chloroform solution was reduced to 3 ml. Purification was effected by chromatography as described above for this compound (p. 51) j the corresponding chloride remaining on the column. After elution of the band from the column* the solvent was evaporated (0.01 mm.* 25cC.) to give 0.54 g. (75$ yield) of the pure yellow compound. Thiocyanatotriphenylphosphinetetracarbonyl- manganese(I)* Mn(C0 )ij.(PPh3 )CNS. To 0.051 g. (0.2 mmoles) of Mn(C0)^CNS in a 32 mm. test tube was added a solution of 0.52 g. (0 .2 mmoles) of triphenylphosphine in 5 ml. of chloroform* and the mixture was allowed to react for 24 hours at room temperature* after which the solution was filtered and the chloroform removed in a stream of nitrogen. The residue was taken up in 2 ml. of benzene and eluted with benzene on acidic alumina* the bis- substituted product remaining on the column. The benzene was removed (0.01 mm.* 4o°C.) to give 0.009 g. (8$ yield) of a yellow powder. trans-Thiocyanatobistriphenylarsinetricarbonyl- manganese(I)* Mh^COj-^AsPh^^CNS. To 0.05 g. (0.2 mmoles) of Mn(C0)^CNS and 0.31 g. (1 mmole) of triphenylarsine in

a 32 mm. test tube were added 10 ml. of benzene and the

reaction mixture was heated at 8o°C. for 12 hours* after which the solution was filtered from a brown decomposition product* and the volume was reduced to 2 ml. under a stream of nitrogen. This solution was chromatographed on a neutral alumina column (30 x 2 .0 can.), first with benzene to elute the mono-substituted derivative, and then with dichloromethane to elute the desired bis- substituted derivative. The volume of the solution was reduced to about 0 .5 ml. of an oil by means of a stream of nitrogen, and 50 ml. of low boiling (30-60'>C.) petroleum ether were added. Scratching at the insoluble residue with a glass stirring rod brought about crystal­

lization of a yellow powder, which after drying (0.01 mm., 25 C.) weighed 0.08l g. (50$ yield). cis-Thiocyanatobistriphenylarsinetricarbonyl-

manganese(I), Mn(C0 )^(AsPh3 )CNS. To 0.05 g. (0.2 mmoles) of Mn(C0)p_CNS in a 32 mm. test tube was added a solution

of 0.31 g. (1 mmole) of triphenylarsine and the reaction

was allowed to proceed at room temperature for 12 hours, after which the solution was filtered and the volume of

the chloroform solution reduced to about 0 .5 ml. of an oil by means of a stream of nitrogen. After addition of

50 ml. of low boiling (30-6o°C.) petroleum ether, crystallization was brought about by scratching with a glass rod and cooling to -78°C. Chromatography was found to be unsatisfactory as a means of purification because this derivative could not be eluted from an alumina column. Although the product was mostly the cis-disubstituted isomer, there was also a substantial amount of the mono- substituted derivative, as revealed by infrared 55 spectroscopy. The total amount of yellow powder collected was 0 .0 8 g., representing a combined 100$ yield of the mono and cis-disubstltuted derivatives. Thiocyanatotriphenylarsinetetracarbonylmanganese(I),

Mn(CO)i±(AsPh^)CNS. To 0.051 g. (0.2 mmoles) of Mn( COLONS and 0 .06l g. (0 .2 mmoles) of triphenylarsine in a 32 mm. test tube were added 5 ml. of chloroform and the reaction allowed to stand for 24 hours at room temperature. At the end of that time, the solution was filtered and the solvent removed by a stream of nitrogen. The residue was taken up in 2 ml. of benzene and purified by chromato­ graphy on an acidic alumina column (30 x 2 .0 cm.), using benzene as the eluting solvent. The trans and cis disubstituted products remained on the column under this treatment. The volume of the solution was reduced to about 0 .5 ml. of an oil by means of a stream of nitrogen. After the addition of 50 ml. of low boiling (30-60°C.) petroleum ether, the oily residue crystallized as a yellow powder upon repeated cooling to -7 8°C. and scratching with'.' a glass rod. The solid was dried (0.01 mm,, 25^0.) to produce 0.05 g. (48$ yield) of the pure derivative. This derivative was also prepared from the metathet- ical reaction between Mn^O^AsPh^Cl and KCNS. From Mn(C0)^Cl, prepared as described earlier (p. 4l), the derivative Mn(CP)2|AsPhgCl was synthesized as follows. To 0.22 g. (1 mmole) of Mn(C0)^Cl and 0.3 g. (1 mmole) of triphenylarsine was added 5 ml. of chloroform. The reaction was allowed to proceed for 24 hours at room temperature3 after which the solution was filtered and the solvent evaporated to dryness by means of a stream of nitrogen. The resulting yellow solid was not purified further and was identified as the desired derivative by infrared spectroscopy.

For the preparation of Mn(C0)^(AsPh^)CNS5 0.5 g.

(1 mmole) of Mn(C0 )^(AsPh3 )2Cl was dissolved in 30 ml. of methanol and treated with 1 g. (10 mmoles) of KSCN in 4 ml. of water. The reaction was carried out in a three neck flask fitted with a nitrogen inlet^ a condenser and a mechanical stirrer. The mixture was stirred at 50^0.

for 8 hours. After that timea the solvent was removed

(0.01 mm., 50°C.), the residue taken up into 8 ml. of chloroform* and the solution filtered from the resulting KC1 and excess KSCN. The volume of the chloroform solution was reduced to 3 ml. Purification was effected by chromatography as described above for this compound

(p• 55)j the corresponding chloride remaining on the column. After elution of the band from the column* the solvent was evaporated (0.01 mm.., 25°C.) to give 0.21 g. (40$ yield) of the pure yellow compound. trans-Thiocyanatobistriphenylstibinetricarbonyl- manganese(I)* M ^ C O ^ S b P h ^ ^ C N S . . To 0.05 g. (0.2 mmole) of Mn(C0)cjCNS and 0.353 (0.2 mmole) of triphenylstibine 57 in a 32 mm. test tube, benzene (10 ml.) was added and the reaction allowed to proceed at 80°C. for 12 hours. The solution was filtered from a brown decomposition product and the volume of the benzene solution was reduced to 2 ml. by means of a stream of nitrogen. This was chromato­

graphed on a neutral alumina column (30 x 2 .0 cm.), first with benzene to elute the trace of the mono-substituted derivative, and then with dichloromethane to elute the desired product. The volume of the solution was reduced

to about 0 .5 ml. of an oil by means of a stream of

nitrogen, and after addition of 50 ml. of low boiling (30-60°C.) petroleum ether, crystallization to the yellow powder was brought about by repeated scratching with a

glass rod and cooling to -7 8°C. After drying (0 .0 1 mm., 25 C.), 0.14 g. (79$ yield) of the product was obtained. cis-Thiocyanatobistriphenylstibinetricarbonyl-

manganese(I), Mn(C0 )3 (SbPh3 )2CNS. To 0.05 g. (0.2 mmole)

of Mn(C0)^CNS in a 32 mm. test tube was added a solution

of 0 .3 5 g. (1 mmole) of triphenylstibine in 5 ml. of chloroform. The reaction was allowed to proceed at room

temperature for 12 hours, after which the solution was filtered. The volume of the solution was reduced to about

0 .5 ml. of an oil by means of a stream of nitrogen, and after the addition of 50 ml. of low boiling (30-60°C.) petroleum ether, crystallization of a yellow powder occurred on repeated scratching with a glass rod and 58 cooling to -78°C. Chromatography was found to be unsatisfactory for purification because the compound could not be eluted from the alumina column. Infrared spectro­ scopy revealed that the corresponding cis-isomer was also present in the product, but only in very small amounts.

After drying (0.01 mm., 2 5°C.), 0.171 g. (95$ yield) of the yellow powder was obtained. Thiocyanatotriphenylstibinetetracarbonyl- manganese(I), Mn(CO)2j.(SbPh3 )2 CNS. To 0.05 g. (0.2 mmoles) of Mn(C0)^CNS in a 32 mm. test tube was added a solution of 0.l4 g. (0 .2 mmole) of triphenylstibine in 5 ml. of chloroform, and the reaction was allowed to proceed for 24 hours at room temperature. The solution was filtered and the volume of the solvent reduced to 2 ml. under a stream of nitrogen. This solution was then chromatographed on a neutral alumina column (30 x 2 .0 cm.), using benzene as the eluent, the disubstituted products remaining on the column. The solvent was removed (0.01 mm., 4 o uC.), and about 0 .0 0 9 g. (5$ yield) of the yellow product was obtained. Thiocyanat o-l,2-bis(dip henylp ho sp hino)ethanetri-

carbonylmanganese(I), Mn(C0 )3 (diphos)CNS. A solution of

0.°5 g. (0.2 mmole) of Mn(C0)^CNS and 0.15 g. (0.3 mmole) of the diphosphine compound in 5 ml. of chloroform in a 32 mm. test tube was allowed to stand at room temperature for 12 hours. The solution was filtered and evaporated to 59 about 0 .5 ml. of an oil under a stream of nitrogen. Addition of 10 ml. of low boiling (30-60°C.) petroleum ether brought about precipitation of the excess di- phosphine ligand in the upper solvent layer while the desired product remained as an oil in the lower layer of the mixture. The suspension of the excess ligand was poured off leaving the product in the flask. After several similar treatments, the product was cooled to -78°C. and scratched with a glass rod under 50 ml. of low boiling petroleum ether until it crystallized as a yellow powder. After drying (0.01 mm., 25*0.), 0.12 g. (100^ yield) of the product was obtained. Table 3 shows the analytical data for the deriv­ atives of the Mn(C0)^CNS system prepared above.

Special preparations of some derivatives

Reaction between cis-MnfCO)^(SbPh.3 JgCNS and

2 a2 1-bipyridine. To 0.09 g. (0.01 mmole) of cis-Mn(C0)^(SbPh^JgCNS in a 32 mm. test tube was added a

solution of 0 .0 3 g. (0 .0 2 mmole) of 2 ,2 *-bipyridine in 15 ml. of chloroform, and the mixture was heated at 50 C.

for 6 hours. The solution was filtered and evaporated to dryness in a stream of nitrogen. The residue was washed with low boiling (30-60°C.) petroleum ether (5 x 20 ml.) and dried by a stream of nitrogen. The yellow residue was purified by chromatography on a neutral alumina column TABLE 3 ANALYSIS OF THE DERIVATIVES OF Mn(COLONS

Compound Calculated Found C H S to CHS to cis-Mn(CO)3py2CNS 47.35 2.82 9.02 46.99 2.79 8.81 cis-Mn(CO)3bipyCNS 47.66 2.27 9.06 47.82 2.42 8.89 cis-Mn(CO)3(p-tol)2CNS 52.6 4.38 7.79 52.9 4.63 7.50 cis-Mn(CO)3(pca)2CNS 42.4 2.65 42.18 2.44 cis-Mn(CO)3(pfa)2CNS 45.8 2.86 45.55 2.79 cis-Mn(CO)3(4-pic)2CNS 50.02 3.66 10.95 49.98 3.74 10.85 cis-Mn(CO)3(AsPh3)2CNS 59.^ 3.75 1.73 55.4 3.70 2.54 c is-Mn(CO)3(SbPh3)2CNS 53.2 3.3 1.55 52.7 3.7 1.59 Mn(CO)3(diphos)CNS 60.4 4.04 2.34 59.95 4.54 2.30 Mn(CO)4PPh3CNS 2.87 2.79 Mn(CO)4AsPh3CNS 51.9 2.82 2.64 52.2 3.03 2.71 trans-Mn(CO)3(PPh3)2CNS 66.5 4.16 1.94 66.7 4.36 1.79 trans-Mn(CO)3(AsPh3)2CNS 59.4 3.75 1.73 60.8 4.24 1.75 trans-Mn(CO)3(SbPh3)2CNS 53.2 3.3 1.55 53.09 3.41 1.48 61

(30 x 2.0 cm. )> using chloroform as the eluent. The solvent was removed by a stream of nitrogen leaving 0 .3 g.

(90 $ yield) of MnCCO^bipyCNS5 which was identified by an infrared spectrum. Reaction between Mn(C0)^CNS and SbPh^ in acetonitrile. To 0.05 g. (0.2 mmole) of Mn(C0)^CNS in a

32 mm. test tube was added 5 ml. of acetonitrile and the mixture was allowed to stand for 5 minutes * after which 0.35 g» (1 mmole) of triphenylstibine was added. The reaction was allowed to proceed at room temperature for 12 hours and then the solution was filtered. The volume of the solution was reduced to about 0 .5 ml. of an oil (0 .0 1 mm.j 4o°C.)j and crystallization of a yellow powder occurred after addition of 50 ml. of low boiling (30-6ocC.) petroleum ether and scratching with a glass rod. The solid was collected on a filter and dried in a stream of nitrogen. The product was not purified furtherj its infrared spectrum revealed that trans-Mn(C0 )3 (SbPh^)2CNS was the major product3 but that a substantial amount of the cis-disubstituted derivative was also present.

Attempted preparation o ? l f e { C 0')M 2 ------

The preparation of this compound was attempted in - a manner similar to that reported by Abel and Wilkinson

(l8) for the corresponding halide dimers. Thus 0.2 g. (0.8 mmole) of solid Mn(C0)^CNS was heated under high 62 boiling (90-120°C.) petroleum ether at 8o°C. for two hours. The color of the solid gradually changed from golden yellow to orange during the heating. The appearance of a "yellow powdersimilar to that reported by the authors above* was also observed. The solvent was removed (0.01 mm., 5 0 ° C . ) and the resulting solid was washed with water (3 x 20 ml.) and dried in a vacuum desiccator over

CaCl2. Treatment of the residue with 8 ml. of chloroform at 4o°C. resulted in the dissolution of all but the "yellow powder." The solution was filtered and its volume reduced

to 5 ml. by a stream of nitrogen. Addition of 100 ml. of

low boiling (30-6o°C.) petroleum ether afforded a dark yellow powder. Infrared spectroscopy showed that this product was different from its parent* Mn(C0)^CNSa but

because of a critical shortage of Mn2 (C0 ) 10 this compound was not examined further.

Preparation of the thiocyanato derivatives of the Group Vlb hexacarbonyls The anionic thiocyanatopentacarbonyl complexes of tungsten* molybdenum* and chromium were prepared (4l) utilizing a reaction similar to that described by Abel et al. (24) for the halogenopentacarbonyls of these metals. This reaction is:

heat m (c o )6 + (c h 3 )4 n c n s ------► (c h 3 )4 n [m (c o )5 c n s ] + CO 63 where M = W, Mo, and Cr. The preparations were carried out in a 100 ml. three neck flask fitted with a nitrogen, inlet and condenser, the middle neck being stoppered. Tetramethylammonium thiocyanatopentacarbonyl- tungstate(O). Tetramethylammonium thiocyanate (0.5 g.,

3 .8 mmoles) in 20 ml. of diglyme were heated at 100-10513C. under nitrogen until the evolution of CO ceased (3 hours). The mixture was filtered hot in an inert atmosphere.

Addition of 150 ml. of low boiling (30-6o'5C.) petroleum

ether to the cooled (06C.) filtrate gave a deep yellow powder. This was washed (5 x 20 ml.) with low boiling petroleum ether, dried under a stream of nitrogen, and then vacuum treated at 60“C./0.01 mm. for 1 hour to remove

the unreacted hexacarbonyl and yield 1.4 g. (82$) of the product. Anal. Calcd. for (CH3 )2jN[W(CO)5CNS] : C, 26.32; H, 2.63; S, 7.02; N, 6.14$. Pound: C, 26.25, 26.43; H,

2.71, 2.795 s, 6.9 9 , 6.95; n. 6.3 1, 6.32$. Tetramethylammonium thiocyanatopentacarbonyl-

molybdate(0). Tetramethylammonium thiocyanate (0.5 g., 3.8 mmoles) and 1.2 g. (4.5 mmoles) of Mo(C0)g in 20 ml. of diglyme were allowed to react at 80-85°C. until the evolution of CO ceased (2 hours). Addition of low boiling

(30-6o pC.) petroleum ether to the cooled (0VC.) filtrate gave a deep yellow powder. This was washed (5 x 20 ml.) with low boiling (30-60°C.) petroleum ether and dried in a stream of nitrogen. Vacuum treatment at 60°C./0.01 mm. 6 4 for 1 hour to remove the unreacted hexacarbonyl yielded 1.3 g. (85$) of the product. Anal. Calcd. for

(CH3 )^N[Mo(CO)5CNS]: N* 7.6$. Pound: N, 8.6$. Tetramethylammonium thiocyanatopentacarbonyl- chromate(O). Tetramethylammonium thiocyanate (0.5 g.* 3.8 mmoles) and 1.0 g. (4.5 mmoles) of Cr(C0)g in 20 ml. of diglyme were heated at 120°C. until the evolution of CO ceased (3 hours). The mixture was filtered hot in an inert atmosphere. Addition of low boiling (30-60°C.) petroleum ether to the cooled (0°C.) filtrate gave a deep yellow powder. After washing with low boiling petroleum * ether (5 x 20 ml.) and drying in a stream of nitrogen* the residue was vacuum treated at 60°C./0.01 mm. to remove the unreacted hexacarbonyl. Throughout the course of the reaction CrfCO)^ sublimed into the condenser and was '6 returned to the reaction flask intermittently by means of

a glass rod. Amounts of (CH3 )2jN[Cr(C0 )cjCNS] ranged from

0 .6-0 .9 g. (60-80$) depending upon the efficiency of returning the sublimed Cr(C0)g to the reaction vessel.

Anal. Calcd. for (CHg)ij.N[Cr(CO^CNS]: S* 9.875 N* 8.64$.

Pound: S* 10.06; N* 8.69$.

Attempted preparations of Pe’l CO )‘ij.XCWs) 2------

The reaction between Fe(C0)^ and (SCN)2 * To 1 ml. of Pe(C0)^ (7.4 mmoles) in 25 ml. of low boiling (30-6o°C.) petroleum ether at 0°C. a solution containing 9 mmoles of 65 thiocyanogen (from 0.5 ml. of Br2 and excess Pb(SCN)2 ) in

35 ml. of low boiling petroleum ether was added dropwise. The reaction mixture was stirred magnetically. An immediate reaction occurred as was evidenced by the formation of a brown precipitate, and also by evolution of CO. After the evolution of gas had ceased., the product was collected on a filter, washed with low boiling petroleum ether, and dried in a stream of nitrogen. Infrared spectroscopy revealed that no CO was present in the solid. When the same reaction was attempted at -20PC., there was also an immediate appearance of a brown precipitate, but without noticeable evolution of CO. The product was collected and dried with a stream of nitrogen. The solid contained no carbonyl bands in the infrared. A quantitative determination of the volume of CO evolved in this reaction was carried out in the apparatus described earlier (p. 36). Iron pentacarbonyl (0.1 g.,

0 .5 mmole) in 10 ml. of chloroform was added into one bulb of the flask and a solution of 1.5 mmoles of (SCN)^ (from

0.22 ml. of Br2 and excess Pb(SCN)2 ) was added to the second bulb. The flask and reactants were cooled to -20°C., and after allowing 30 minutes for equilibrium to be reached, the contents were shaken vigorously to afford good mixing. Formation of the solid occurred immediately;

15 minutes were allowed for the completion of the 66 reaction. The contents of the flask were allowed to gradually warm up to room temperature. There was no evolution of gas from -20 to 2°C. However, steady evolution of CO occurred at 2 aC. and continued as the temperature was being raised to 10°C. The corrected1 volume of mercury displaced was 60 ml., corresponding to a complete loss of CO by the product. The solid was collected and dried by a stream of nitrogen. A subsequent infrared spectrum revealed no CO bands.

Reaction between HgFe(C0 )4 and (SCN)2 * A

suspension of 0.37 &• (1 mmole) of HgFe(CO)^ in low boiling (30-60*C.) petroleum ether, kept uniform by magnetic stirring, was treated by dropwise addition of a

solution of 2 mmoles of (SCN)2 (from 0.3^ ml. of Br2 and

excess Pb(SCN)2 ) in 30 ml. of low boiling petroleum ether. This procedure was carried out several times at various

temperatures ranging from 0 to 25°C. There was an immediate reaction, as was evidenced by the change in color of the solid suspension from the yellow of HgFe(CO)^ to the brown of the product. Gas evolution was also observed in each experiment. In all cases the solid was collected and dried in a stream of nitrogen. No carbonyl

^ h e correction to allow for the natural expansion of the gas in the flask was determined by carrying out an experiment which employed identical conditions, but without the reactants. containing product was isolated, as was shown by infrared spectroscopy. Reaction between Fe3(C0)i2 and (SCN)g. To 0.5 g.

(1 mmole) of Fe^CCO)^ in 20 ml. of low boiling (30-6 0^0 .) petroleum ether was added dropwise a solution of 4 mmoles

of thiocyanogen (from 0.2 ml. of bromine and excess

Pb(SCN)2 ) in 30 ml. of low boiling petroleum ether. Since no reaction occurred at 25°C., the temperature of the reaction mixture was slowly raised to 50°C., whereupon gas evolution was observed. After the evolution of CO

had ceased (6 hours), the solution was filtered and the solid obtained was washed with low boiling (30-60PC.) petroleum ether and dried by a stream of nitrogen. The solid contained no CO, as was shown by infrared spectro­ scopy. RESULTS AND DISCUSSION

Investigation of the M n (00)5CNS system

Attempted preparations of Mn(C0 )5CNS. Several methods of preparation of a monomeric thiocyanato derivative of manganese carbonyl were attempted, but only one method proved reproducibly successful. The first attempt involved the direct reaction between (SCN)2 and Mn^CO)-^, and was based on an analogy to the known reaction between manganese carbonyl and chlorine or bromine. (18) Whereas the reaction between M r ^ C O ) ^ and Clg to produce Mn(CO)^Cl proceeds at 0°C.; heat (ca. 4o°C.) is required to produce

the corresponding bromide from Mn2 (co)]_o and Br2 . As was mentioned previously (p. 1 9 )3 I2 does not react at all ■under similar conditions because of its low oxidizing power, although the direct reaction between I2 and MhgCCO)^ may be effected in a sealed tube at 120cC. Comparing the oxidizing power of thiocyanogen with

that of the halogens, one finds that (SCN)^ is between Br2

and I2 : F2 > CI2 > Br2 > (SCN^^ Ig* This series is based on the oxidation potentials of the halogens (42), which are given below.

I -- > 1/2 I2 + e e 298 = _0*535 volt SCN" — > 1 /2 (SCN)2 + e" = -0 .7 7 Br~ — ) 1 /2 Br2 + e" = - 1 .0 7 68 69

In view of the position of (SCN)2 among the halogens in oxidizing power* it is not surprising that a reaction between (SCN)2 and Mn2 (C0)10 did not occur at room temperature. At higher temperatures, the attempted synthesis yielded an orange-red solid, Mn2 (CO)io being recovered unchanged. The infrared spectrum of the orange- red solid, which was insoluble in common organic solvents, exhibited no CO bands in the 2000 cm"-1- region. It is known that although (SCN)2 is stable in most dry, inert solvents, it polymerizes in solution, especially under the initiating influence of heat, light, moisture, or oxygen. (34) It was therefore concluded that the orange-red solid was parathiocyanogen, (SCN)X , a polymer of thiocyanogen of unknown composition. (43) In the second attempted method of preparation, which involved dropwise addition of (SCN)2 to a solution of NaMn(CO)^, only Mn2 (C0)^0 was recovered as a major carbonyl containing product. This result can readily be explained when one considers the known reaction of metal carbonyl halides and sodium carbonylates to produce mixed metal carbonyls. For example, the equations

NaCo(C0)4 + Mn(C0)5Br — > (CO)4Co-Mn(CO) 5 + NaBr (44) and

NaMn(C0)5 + (C0)5ReCl— » (CO)5Re~Mn(CO)5 + NaCl (45) show the oxidation and reduction of the metals involved from

-1 and +1 states, respectively, to the zero oxidation state. 70

When (SCN)2 is being added to NaMn(CO)^, there is excess NaMn(CO)^ present throughout most of the reaction. The thiocyanato derivative formed can therefore react with the excess NaMn(CO)^ in a manner similar to that shown above for the mixed carbonyls. The following set of equations illustrates the formation of Mn2 (CO)io*

NafHs') Mh2(CO)io + 2Na 2NaMn(CO)5

NaMn(C0)5 + (SCN)2 ----- > NaCNS + Mn(C0)5CNS

NaMn(C0) 5 + Mn(C0)5CNS --- > NaCNS + Mn2 (C0)i0

The successful preparation of Mn(C0)^CNS involves the use of C1CNS and NaMn(CO)^. The physical properties of thiocyanagen chloride in solution and its reactions have been studied by Angus and co-workers. (35)* (36) They have definitely established the monomeric nature of C1CNS and describe it as a very polar molecule; the thio- cyanate end being polarized positively enough to undergo electrophilic substitution reactions on benzene rings.

Since chlorine constitutes the negative end3 C1CNS is expected to react with NaMn(CO)^ to give exclusively NaCl and Mn( COLONS., without any formation of Mn(C0)^Cl and NaSCN as byproducts. Physical properties of Mn(COLONS. Thiocyanato-

pentacarbonylmanganese(I) 5 the first simple thiocyanato

derivative of a metal carbonyl5 is readily soluble in 71 dichloromethane, 1,2-dichloroethane, ethyl acatate, acetone, acetonitrile, methanol, nitrobenzene, tetrahydro- furan and nitromethane, moderately soluble in chloroform, slightly soluble in chlorobenzene and diethyl ether and insoluble in benzene, petroleum ether, cyclohexane, carbon tetrachloride, carbon disulfide and water. This compound, which is obtained as a golden yellow powder, is a non­ electrolyte in nitromethane (molar conductivity = 0.7 ohm"-1-) and is also diamagnetic.

Infrared spectroscopic studies of Mn(C0 )5CNS. Infrared spectra of Mn(C0)^CNS were taken in Nujol,

chloroform, dichloromethane, 1 ,2 -dichloroethane, ethyl acetate and acetonitrile. In the Nujol spectrum both CO and CN bands are clearly detected in the 1900-2200 cm'-*- region, but the two strongest CO bands are fairly broad. . The infrared study in Nujol is completely reproducible from spectrum to spectrum, indicating that the broadness of the intense CO bands is not due to faulty technique in making the mulls. The significant feature in the Nujol

mull spectra is the presence of a weak band at 676 cm“^, which has been assigned to the CS stretching mode. Re­

calling that N-bonded thiocyanates absorb from 780 to

860 cm“^, and S-bonded thiocyanates absorb between 680 and

720 cm"-*-, (1 1 ) this characteristic band indicates that the thiocyanato derivative of manganese carbonyl, Mn(C0)^CNS, has an S-bonded structure in the solid state. 72 Infrared spectra of Mn(CO)^CNS in dichloromethane, chloroform, 1 ,2 -dichloroethane and ethyl acetate are very similar and fairly complex. These spectra by no means represent molecular Cij. symmetry such as that of Mn(CO)^Cl,

Mn(C0 )5Br, etc. Indeed, in view of the S-bonded structure of Mn(C0 )5SCN, the point group C^ is not expected. Because of a bent M-S-C moiety, the highest symmetry that the compound Mn(C0 )5SCN can attain is that of the point group Cs (only a plane of symmetry). The group Cs symmetry arises if the thiocyanate ligand is either directed at one equatorial CO group, or is pointing exactly half way between two equatorial carbonyls. If the SCN group is pointing at random within the molecule, then group symmetry arises. The number of CO fundamentals expected to be active in the infrared may easily be determined from the character tables below for the symmetry groups C]_ and Cs, respectively.

C1 E cs E A 1 A' 1 1

A" 1 - 1

For either symmetry, five infrared active modes are ex­ pected. If the molecule Mn(CO)^SCN belongs to the point group C^, then five A fundamentals are expected. Similarly, group theoretical calculations show that five infrared active modes, namely 3A' + 2A", are predicted if the 73 molecule Mn(CO)n>SCN belongs to the point group Cg. Figure 1 shows the spectrum of Mn(CO)^SCN in dichloromethane. It should be mentioned that an additional very weak band, which is always observable in a spectrum of Mn(CO)^CNS in 1,2-dichloroethane, does not appear in the other solvents at concentrations less than 3 * 5 mg./ml. and at temperatures below 22 °C. However, this weak peak can be seen in the latter solvents at higher concentrations and/or temperatures. Table 4 gives the absorption bands in the CO region of Mn(CO)^SCN in the aforementioned solvents. The numbers in parentheses indicate frequencies of the extremely weak bands mentioned above. No assignment of the CO and CN bands is made in the table. The large number of bands in the above spectra by no means supports a monomeric structure of the thiocyanato derivative. In fact, solution spectra are strikingly similar to those expected for the thiocyanatotetracarbonyl- manganese dimer, [Mn(C0 )ij.CNS]2 . Nevertheless, the elemental analyses and molecular weight, determined by cryoscopy in nitrobenzene, give values very close to those calculated for the monomeric thiocyanato pentacarbonyl. The latter results were interpreted with caution, however, since the elementary compositions of the monomer and dimer are not sufficiently different and examples are known where cryoscopic measurements have given unreliable results for the molecular weights of simple metal carbonyls. For 74

2200 20,00

Figure 1. The carbonyl and CS stretching frequency regions of the infrared spectrum of Mn(C0)cCNS. The 1900-2200 cm- 1 and 6 5 0 -7 0 0 cm- 1 regions were taken in CHgCl-g and Nujol* respectively. TABLE 4

INFRARED CARBONYL STRETCHING FREQUENCY REGION OF Mn(CO)5SCN IN VARIOUS MEDIA

Solvent CO and CN Bands (cm-1)

Chloroform 2162m 2144m (2110vw) 2099w 2 0 6 0s 2037s 1957s

Dichloromethane 2156m 2l40m (2110vw) 2096w 2 0 5 8s 2 0 3 6s 195 0s lj2-Dichloroethane 2162m 2139m (2110vw) 2094w 2 0 6 0s 2 0 3 1s 1960s

Ethyl acetate 2159m 2145m (2110vw) 2096w 2 0 5 8s 2034s 1 9 50s

Nu j ol 2 l6 0sh 2138m 2084w 2043sb 1 9 5 8sb

m = medium; w = weak; s = strong; sh = shoulder; b = broad; v = very.

ui 7 6 example3 Calderazzo £t al. (46) found that the molecular weights of V(CO)g and Cr(C0)g were anomalously high when determined by cryoscopy in cyclohexane and benzene. In order to unambiguously determine the monomeric nature of Mn(C0 )5CNS, the volume of CO evolved in the quantitative reaction between pyridine and Mn(CO),-CNS was measured. This reaction proceeds at room temperature according to the equation:

Mn(CO)5CNS + 2C5H5N --- > Mn(CO) 3 (C5H5IQ2 CNS + 2C0

From the reaction of 0.0628 g. (0.248 mmole based on the

monomer5 0.140 mmoles based on the dimer) of the thiocyan­ ato metal carbonyl with excess pyridine at 28°C. and 738 mm. Hg, 12.8 ml. (0.504 mmole) of CO were evolved. For

the monomeric pentacarbonyl complex, evolution of 1 2 .6 ml. (0.496 mmole) of CO is expected, whereas for the tetra- carbonyl dimer, slightly more than half of that amount,

i.e., displacement of 7 .1 ml. (0 .2 7 9 mmole) is predicted from the following equation:

[Mn(C0)^CN^2 + 2C5H5N ---- > 2Mn(C0)3 (C5H5N)2CNS + 2C0

Thus the result clearly indicates the monomeric nature of

Mn(C0)5SCN. The first clue in attempting an explanation of the

unexpectedly large number of bands in the 1 9 0 0 -2 2 0 0 cm' 1 region was obtained from the spectral studies of Mn(C0)^SCN 77 in acetonitrile. This spectrum (Figure 2) is essentially representative of a carbonyl compound belonging to the point group C4v . The only feature of the spectrum which does not occur with complexes belonging to the point group

C4 V is the presence of a slight shoulder on the most intense band. There are two possible explanations for the

C4 symmetry of the derivative. Either the thiocyanato group had been replaced by a solvent molecule, resulting in the ionic compound [Mn(C0 )5CH3CN]+ SCN", or a rearrangement from the sulfur-bonded form to the nitrogen-bonded form had occurred. The replacement of the thiocyanato group by the solvent was immediately ruled out on the grounds of the following evidence. First of all, acetonitrile solutions of Mn(CO)pjCNS show a peak at 813 cm"*L, which is most probably due to the CS stretching frequency of an N-bonded thiocyanate. Free thiocyanate absorbs at much lower frequencies, around 750-770 cm-1. Since there is solvent absorption below 77 0 cm"-1-, possible presence of a band in the 750-770 cm"-1- region could not be investigated. Although there is no plausible explanation for the 813 cm- 1 absorption in terms of free thiocyanate ion, this band could possibly be due to a mode other than a C-S stretch. To guard against the latter possibility, the infrared spectrum of Mn(C0)^Br in acetonitrile was run in the 8 0 0- 820 cm-"1- region. Since no absorption bands due to the 78

2 2 0 0 2000 1800

900 800 I__ _1_ cm

Figure 2. The carbonyl and. CS stretching frequency regions of the infrared spectrum of Mn(C0)5CNS in acetonitrile. 79 bromo carbonyl compound were detected, it is concluded that the assignment of this band to the CS stretching frequency of a thiocyanato-N compound is correct. Additional evidence in support of the rearrangement rather than the solvent replacement viewpoint is the

presence of a medium intensity band at 2 1 1 3 cm"in the spectrum of Mn(CO)^CNS in acetonitrile. The CN stretching mode of free thiocyanate ion gives rise to an absorption at 2075 cm-1, which is not present here. Thus, the most plausible assignment of the band at 2113 cm--5- is to the CN stretching mode of an N-bonded thiocyanate. The most convincing evidence in support of a rearrangement from the S-bonded to the N-bonded form was obtained through conductivity measurements in acetonitrile. If the thiocyanate ion had been replaced by the solvent,

then the ionic compound [Mn(C0 )^(CH2 CN)]+ SCN“ will be present in solution. Conductivity measurements of such a species should indicate uni-univalent electrolytic behavior. Conductivity studies were carried out on Mn(C0)^CNS and Mn(C0)^Cl in acetonitrile at 25°C. and at concentrations

of- about 1 0 - 3 M. The measurements on the thiocyanato compound were taken after allowing three minutes for the system to come to equilibrium. The conductivities were taken over a period of one hour and were observed to Increase slowly for both the thiocyanato and chloro derivatives. The increase in conductivities amounted to 8o about 6 ohm”'L for each compound, and is probably due to either a slow decomposition of the samples or to a small amount of ionization of the derivatives. Molar conduct­ ivity data were plotted vs. time, and extrapolation to A t = 3 min. for the thiocyanato complex, and to zero time for chloro carbonyl derivative, gave the initial con­ ductivities of the samples. These initial values for

Mh(C0)^CNS and Mn(C0)^Cl were found to be 8 .6 and 6.0 ohm-'1', respectively. The molar conductivity of Ph^AsCl, a uni-"univalent electrolyte, was found to be 138 ohm"-*- in acetonitrile. These results show that while there is a small amount of dissociation (ca. 6.2$ for Mn(CO)^CNS) in acetonitrile, the compounds Mn(C0)^CNS and Mn(00)^01 are essentially nonelectrolytes. The results of the conductivity study indicate that a rearrangement of the thiocyanato group had occurred. On the basis of the presented evidence that the compound is N-bonded in acetonitrile, one immediately concludes that for Mn(C0)^CNS, the energy difference between the S- and N-bonded linkage isomers is small indeed, since even a change of solvent causes a dramatic change in spectrum. It is interesting to compare the spectrum of Mn(C0)^NCS with those of the manganese carbonyl halides, Mn(C0)^X, all of which belong to the point group C^y . Using the following character table for group , the 8l number of infrared active modes expected for these com­ pounds may be readily predicted.

E 2 C4 C2 CM 2

Al 1 1 1 1 1 A-2 1 1 1 -1 -1 B1 1 - 1 1 1 - 1 B2 1 - 1 1 - 1 1 E2 2 0 -2 0 0

Group theory considerations show that the CO vibrational modes of the molecules Mn(CO)^X belong to the representation 2A^_ + B-^ + E. In order for any of these modes to be infrared active, a change in dipole moment within the molecule must occur during vibrational motion. Examination of the above representation reveals that one of the Aj modes and the mode are infrared inactive since the CO stretches which they represent do not result in a change of dipole within the complexes. However, as Orgel (47) points out, when a molecule has two stretching modes of the same symmetry, one active and one inactive, then these vibrations will couple. The result of this coupling is that the active mode transmits some of its intensity to the inactive mode, thus making it infrared active. There­ fore, one expects to observe three CO bands, namely 2A^ + E, in the infrared. Furthermore, it is expected that the relative intensities of these bands should be weak

(A1(2)), strong (E), and medium (A1 (l)). The assignments have been made (48) for pentacarbonyl halide complexes; 82

Table 5 shows the comparison of the CO stretching frequen­ cies of the halogeno pentacarbonyl compounds with those of Mn(CO)ejNCSj whose band assignments have been made in an analogous manner. It can be seen that the spectrum of Mn(CO)^NCS closely resembles those of halogeno derivatives*

except that the frequency of the A-^C1) band is about 50 cm"-1- lower for the thiocyanato complex. Upon reexamination and comparison of the spectra of Mn(C0)^CNS in the various solvents, one notices small differences in the carbonyl stretching bands among the

solutions in chloroform, dichloromethane, 1 ,2 -dichloroethane and ethyl acetate, and a vast difference between these spectra and the one in acetonitrile. In addition, the

slight deviation from the C4V symmetry in acetonitrile is difficult to explain unless one assumes that in every solvent there is an equilibrium between the N- and S-bonded forms, i.e.,

The position of this equilibrium depends, among other factors, on the solvent in which the sample is dis­ solved. Thus in the four solvents whose Mn(C0)^CNS TABLE 5

COMPARISON OF THE CO STRETCHING FREQUENCIES OF Mn(CO)5X COMPLEXES

CO Bands (cm"^) Compound Solvent AX(2) E A 1 ( D

Mn(CO)5I (48) chci3 2136w 2056vs 2 0 1 7s

Mn(CO)5Br (48) CHCI3 2l40w 2050vs 2 0 1 0 s

Mn(CO)5Cl (48) chci3 2l45w 2055vs 2 0 0 5s

Mn(CO)5NCS ch3cn 2l4lw 2 05 3vs 1958s

w = weak; s = strong; v = very.

co UJ 84 spectra are similar, the equilibrium is toward the S-bonded form, while in acetonitrile, the N-bonded form predominates. These observations immediately suggested several experi­ ments from which direct evidence in support of the equilibrium interpretation could be obtained. Evidence in support of an equilibrium. The equil­ ibrium hypothesis was first tested by carrying out infrared spectral studies on the thiocyanato compound in acetoni­ trile in order to detect the rearrangement from the thiocyanato-S to the thiocyanato-N form. Solutions of several concentrations of Mn(CO)^CNS in acetonitrile,

ranging from 1 mg./ml. to 10 mg./ml., were prepared. Immediately upon dissolution of the sample the first spectrum was recorded. Subsequent spectra were taken as soon as the previous one had been recorded. In a similar

study, the 7 8 0 -8 6 0cm region was scanned in order to

detect the increase in Intensity of the CS band at 8 1 3 cm- 1 as the rearrangement took place. Attainment of the final spectrum was found to be dependent on the concentration of the solution. Thus, when the concentration was 1 mg./ml., the bands other than those due to the N-bonded isomer

essentially disappeared in less than 5 minutes, whereas

when the concentration was 1 0 mg./ml., about 2 0 minutes were required for the spectral change to be no longer noticeable. Figure 3 shows the change in the carbonyl stretching frequencies in the course of isomerization in

acetonitrile at 2 2 °C. 2200 2000 , 1800 2200 , 2000 1800 2200t 2000 1800 J 1------1----- 1------L cm"1 L cm 1 — Lem’1 J - ' I 1

1 f

Figure 3. Infrared spectra of the carbonyl stretching frequency region showing oo the initial, intermediate, and final bands of the rearranging Mh(C0)^CNS in ui acetonitrile. 86

Studies in the 7 8 0 - 8 6 0cm"’1' region were also carried out at the same conditions. Using Mn(CO)^CNS solutions of approximate concentration of 5 mg./ml., the increase in intensity of the CS band at 813 cm"^, although small, was clearly observable. The maximum intensity of the CS stretching frequency was reached in about 10 minutes. In all the above experiments, the compound could be recovered from acetonitrile, even after the solution had been standing for periods longer than one hour. Additional evidence in support of the equilibrium interpretation was obtained from spectra of Mn(C0)^CNS in mixed solvents. An equilibrium between the N- and S-bonded forms should depend on the nature of the solvent. Therefore, relative intensities of the carbonyl bands are expected to change upon altering the nature of the medium, for example, gradual addition of a different solvent. A pronounced shift in equilibrium was indeed t observed when infrared studies were carried out on the compound dissolved in mixtures of chloroform and aceto­ nitrile of variable composition. When chloroform was being added gradually to a solution of Mn(C0)^CNS in acetonitrile the change of spectrum indicated an increasing amount of the S-bonded form. Addition of acetonitrile to a chloroform solution brought about the opposite result. This operation could be carried out indefinitely, the equilibrium shifting from one form to the other with the 87 appropriate solvent. Little, if any, decomposition was observed on recovery of the compound from these solutions. It was also felt that the CS stretching frequency due to the N-bonded isomer would be observed in an infrared spectrum if the concentration of the N-bonded form was sufficiently high. Since an equilibrium was presumed to exist, some of the N-bonded species were present in all solvents. Whereas the strong absorption of CHCl^ and

CH2 CI2 in the 7 8 0 -8 6 0cm- 1 region precluded the use of

these solvents, 1 ,2 -dichloroethane and ethyl acetate were satisfactory as solvents in this spectral region when used with a reference cell. The CS stretching frequencies were indeed observed,

but were very weak. Table 6 shows the frequency of the CS absorption in those solvents where the band was observed. From each of the three solvents the compound could be recovered without significant decomposition.

TABLE 6 THE CS STRETCHING FREQUENCY OF Mn(CO)c-CNS IN VARIOUS SOLVENTS 3

Solvent Frequency (cm-1)

1 ,2 -Dichloroethane 8l0w

Ethyl acetate 80 1w

Acetonitrile 8 1 3m

w = weak; m = medium. . 88 Additional supporting evidence for the equilibrium between the N- and S-bonded forms was obtained from the infrared studies of solutions of Mn(C0)^CNS at various temperatures. These temperature dependence studies were carried out in solvents of relatively high boiling points, namely ethyl acetate* l*2 -dichloroethane and acetonitrile* at temperatures ranging from room temperature (17°C.) to 4o°c. A dramatic difference in the spectra was observed as a function of increasing temperature; in all cases the shift was toward the spectrum of Mn(CO)j-CNS in acetonitrile. Figure 4 shows these changes in ethyl acetate; they are essentially the same as those in l*2-dichloroethane. When the solutions are allowed to cool back to room temperature* the original spectra are observed again* indicating that the system regains its previous equilibrium. The compound was recovered unchanged from these solutions by precipi­ tation with petroleum ether. Table 7 shows the change in the spectra of Mn(CO)^CNS in l*2-dichloroethane and ethyl acetate.

TABLE 7 CHANGE IN BAND INTENSITIES WITH RISING TEMPERATURE

Solvent Increasing (cm-'1-) Decreasing (cm-**')

1* 2-Dichloroethane 2 1 1 0 * 2 0 3 1, 1973 2139, 2 0 9 4 , 2 0 6 0* i960 Ethyl acetate 2107, 2034* 1971 2145, 2 0 9 6 * 2 0 5 8, 1950 2200 20.00 2200 2000 1800 2200 2000 18.00 citT1

Figure 4. The carhonyl stretching 00 VO . 90

The spectrum of Mn(C0)^CNS in acetonitrile was also observed to alter with rise in temperature. The only pronounced change, however, was a decrease in intensity of the slight shoulder on the E band mentioned previously (p.

7 7). As was pointed out earlier, this slight shoulder was the only deviation from the spectrum expected for a carbonyl of group C4V symmetry. However, with a small rise in temperature, the band is virtually eliminated. Figure

5 shows this small change in the spectrum of the aceto­ nitrile solution with increasing temperature. £n an attempt to isolate the N-bonded form, a solution of Mn(C0 )5CNS in acetonitrile was prepared and allowed to stand a sufficient time for rearrangement to occur. The solvent was then removed by means of a vacuum pump. Spectra taken of the remaining solid in Nujol were identical with the spectrum of the solid S-bonded isomer. Spectra taken by the thin film technique from acetonitrile revealed that rearrangement back to the S-bonded form occurred as the solvent was removed. When solvent bands were no longer visible, the thin film spectrum was identical to the Nujol spectrum of the S-bonded linkage isomer. The fact that no N-bonded form could be isolated is not surprising. Since the energy difference between the S- and N-bonded forms is presumed to be very small, a return to the more stable S-bonded isomer in the solid state would be expected. 2200 2000 1800 2200 2000 1800 _l I L. c m 1 L— I— 1

(b)

Figure 5. The carbonyl stretching frequency region of the infrared spectra of Mn(C0)cjCNS in CH^CN at different temperatures: (a) 22°C.; (b) 38°C. 92 Conductivity measurements were carried out in acetonitrile at 3 8°C. in order to ascertain that there was no solvent replacement of thiocyanate ion at that temper­ ature. A 10"3 m solution of Mh(CO),_CNS in acetonitrile was allowed to stand for 3 minutes for equilibrium to be attained. Then conductivity readings were taken over a period of one hour; they were found to have increased by about 6 ohm"-*- during that time. After plotting these data vs. time., extrapolation back to 3 min. = 0 time gave

8 .2 ohm“^- as the initial molar conductivity of the complex. Recalling that the molar conductivity of Mn(CO)^CNS in acetonitrile at 25°C. was 8 .6 ohm"-1-, it is seen that no significant difference exists for the molar conductivities of Mn(CO)cjCNS at the two temperatures. Additional studies involving visible-ultraviolet spectra were carried out with the purpose of obtaining further supporting evidence of the existence of an equil­ ibrium in solution between the N- and S-bonded forms of Mn(CO)^CNS. As Schaffer (^9) points out3isothiocyanates may be distinguished from thiocyanates by comparison of the energy differences of their respective ligand field spectra. Y/hereas the d-d transitions of a thiocyanato-S compound occur at a higher wavelength than those of a metal containing a thiocyanato-N attachment, the ligand field spectra of the former are similar in energy to those obtained for the corresponding chloride complexes. 93 Therefore* a comparison of the ligand field spectrum of a thiocyanato metal complex with that of the corresponding chloro metal complex may he used as a criterion for differ­ entiating the mode of attachment of the thiocyanate group. Visible-ultraviolet spectra were recorded for Mn(CO)^CNS in chloroform and acetonitrile solutions. Similar studies were also carried out on Mn(CO)^Cl and Mn(CO)^Br in the same solvents. Whereas a comparison of the spectra of the thiocyanato and halogeno complexes

revealed them to be similar, Gray et al. (^8) have assigned the bands in the visible region to charge transfer trans­ itions for the carbonyl halide complexes. Because of the similarity in intensity of the visible region absorption bands of Mn(CO)^CNS and those of the halogeno carbonyl complexes, it is assumed that the former are also due to charge transfer transitions. Since no bands resulting from d-d metal transitions are observed for either the thio­ cyanato or halogeno carbonyl complexes, it is concluded that they are weak and hence covered up by the stronger charge transfer absorptions. Therefore, visible-ultraviolet studies were of little value in support of the equilibrium hypothesis. The evidence which has been presented above, while not unequivocal, certainly strongly supports the existence of an equilibrium in solution. Whereas the Mn(CO)^CNS system represents the first known example of reversible linkage isomerism of the thiocyanate group in inorganic chemistry* equilibria involving the thiocyanate group are known among organic compounds.

Iliceto and co-workers (50)* (51)s (52) have described studies on some thiocyanate-isothiocyanate isomerizations in organic systems. Equilibria between S- and N-bonded forms of benzhydryl* allyl* y-methylallyl* and tf* -dimethylallyl thiocyanates were definitely detectable in polar, but not in nonpolar solvents such as cyclohexane (except for V* V-dimethylallylthiocyanate). It is of interest to note that the solvent in which these equilibria were most clearly demonstrated was acetonitrile* which also best promotes Mn(C0)^SCN to Mn(C0)^NCS isomerization. Band assignment of the infrared frequencies of

Mn(C0)5CNS. Although no Raman spectra of Mn(C0)^CNS are available to. complement the infrared studies, it is still possible to make reasonable band assignments on the basis of the spectral variations among solutions of the penta- carbonyl in several solvents and at different temperatures.

In Figure 6 the band assignments are made on the spectrum of Mn(C0)^CNS in 1,2-dichloroethane. Since an equilibrium between the S-bonded and N- bonded forms of Mn(C0)^CNS is presumed to exist in solution* the spectra taken in various solvents consists of absorption bands due to both linkage isomers. 95 2200 2000 1800 i 1 I * i c m 1

2162 CN(S) 2139 Ap\s) 2110 CN(N)

2094 B,.(S)

2060./' E,(S)

Figure 6 . Assignment of the CO and CN hands of Mn(C0)cCNS in 1 ,2 dichloroethane. 96 Using the data on the change in spectra with temperature in 1 ,2 -dichloroethane, some bands may be assigned immediately. Recalling that the shift in spectra of Mn(C0 )5CNS with rising temperature was toward that of the N-bonded form, it may then be assumed that those bands which increase in intensity are due to the thiocyanato-N isomer. Conversely, those bands which decrease in intensity with increasing temperature may be assumed to be due to the thiocyanato-S form. On this basis, the bands

increasing in intensity with increasing temperature at 2 0 3 1 and 1973 cm"-1- are assigned to the E (N) and A^C-1-) (N) modes, respectively. The corresponding bands which decrease in intensity with increasing temperature are those

at 2060 and i9 6 0 cm"-1-, and are assigned to the E (S) and (S) stretching modes, respectively. The weaker absorptions may be assigned with the aid of the data on variations in band intensities with increasing temperature and in conjunction with the spectrum of Mn(CO)^CNS in acetonitrile. The latter is assumed to be representative of the N-bonded isomer, which greatly predominates over the S-bonded form in that solvent. Since the CN stretching frequency of Mn(CO)^CNS in acetonitrile

occurs at 2 1 1 0 cm"*1-, the 2 1 1 3 cm"-1- band, which increases in

intensity with increasing temperature in 1 ,2 -dichloroethane, is therefore assigned to the CN (N) stretching mode. . 97 The highest frequency hand (2162 cm"1 ) in 1*2- dichloroethane does not change appreciably in intensity with rising temperature. However, it disappears completely when the spectrum of Mn(CO),_CNS is taken in acetonitrile. It must therefore be due to an absorption by the S-bonded form; because of its high frequency the band is assigned to the CN- (S) stretching mode. The peak at 2139 cm"3-, which decreases in intensity with increasing temperature, and which also is not present in the acetonitrile spectrum of Mn(CO)ejCNS, has been assigned to the carbonyl (S) stretching mode. It is conceivable that the above assignment for the 2162 and 2139 cm bands are actually reversed. How­ ever, investigation by the author of the medium and weak intensity bands in this and other related systems supports the former conclusions. The CN bands are usually broader than the normally sharp CO bands, and the band at 2139 “1 —1 cm” is considerably sharper than that at 2 1 6 2 cm” . \ The weakly intense band at 2094 cm- 1 was found to considerably decrease in intensity with increasing temperature; moreover, it was absent in the spectrum of Mn(C0)^SCN in acetonitrile. This band is assigned to the B-^ (S) stretching mode. For a carbonyl complex belonging

to the point group C2j.v, this stretching mode is infrared inactive. However, the angular arrangement of the M-S-C moiety in Mn(C0)^SCN actually reduces the symmetry of the 98 molecule from C2j.v to no higher than Cs. Now the stretching

mode (B^ in C4 vj A in Cs) becomes infrared active due to the resulting unequivalence of the equatorial CO groups. An analogous problem was encountered by Cotton and

Kraihanzel (5 3)j who observed a weak band in the spectrum of (PPhg)Mo(CO)cj. They reasoned that this molybdenum

carbonyl did not belong to the point group C4 v j indeed* since the highest symmetry attained by the triphenyl- phosphine group is the symmetry of the whole complex

is reduced from C4 v to Cl3 thus causing the mode to become active in the infrared. This assignment received support from Raman spectroscopy* which definitely estab­ lished the band as due to the B^ stretching mode. The above assignments for Mn(CO)^CNS are made in complete accord with arguments presented by Orgel (47) and El-Sayed (54)* on the basis of relative frequencies' and intensities of the absorption bands. It should be mentioned in all fairness that the shoulders at 2014 and

2 0 1 5 cm"-1- which appear at higher temperatures in the spectrum of Mn(C0)^CNS in l*2-dichloroethane and ethyl acetate* respectively* cannot be explained with the data on hand. One cannot assign them to the CN stretching mode of free thiocyanate ion* since free SCN- was found by the

author to absorb around 2 0 7 0 cm-'*" in ethyl acetate. The solid state spectra of Nujol mulls and thin films of Mn(C0)^CNS lend support not only to the band 99 assignment above, but also to the hypothesis of the equilibrium between the two linkage isomers. On the basis of a weak CS stretching frequency at 6 7 6 cm"1 and a complete absence of a CS band in the 7 8 0 -8 6 0cm- 1 region, the thiocyanato group was shown to be attached exclusively through sulfur in the solid Mn(CO)^SCN. Looking back to Table 4, it is seen that none of the bands assigned to the N-bonded structure are observed in the Nujol or thin film spectra. Furthermore, all the bands observed in the solid state spectra correspond to those assigned to various stretching modes of the S-bonded form in solution. Upon assignment of the bands in the spectrum of Mn(CO)^SCN, it is possible to use the appropriate frequen­ cies to calculate CO force constants. The groundwork for force constant calculations in carbonyl compounds has been laid, inter alios, by Cotton and Kraihanzel (53)a (55)a

(5 6), who have rather extensively developed the procedure for these calculations. The following deductions are made by Cotton (53) in deriving the force constants: 1. All (CO-stretching)-(CO stretching) interactions should, give rise to terms in the potential energy expression with positive coefficients (interaction force constants). That this is true can readily be seen when one considers that as a CO bond is stretched, the TT-bonding between the

C and 0 atoms decreases, and thus the tt* (antibonding) 100 orbital decreases in energy, and more nearly matches the energy level of the metal dTr orbitals (which are involved in back-bonding to the *tt* antibonding orbitals on the C atom). The metal-carbon monoxide interaction thus increases and causes a "drift" of the drr electrons to this M-CO grouping. This means that d electrons are less available to the other M-CO groups in the molecule. There­ fore, the carbon pT T orbitals in these other CO groups will participate more fully in CO TT-bonding. This new IT-bonding with the carbon p orbitals strengthens the other C-0 bonds, thus increasing their resistance to stretching. 2. If kc and k^ are the interaction constants between pairs of cis and trans CO groups, respectively, we should expect that k-fc~2kc . This follows because a pair of cis CO groups directly share only one dir orbital whereas a pair of trans CO groups directly share two d~rr orbitals. 3. The CO stretching force constants should decrease steadily as CO groups are successively replaced by other ligands which make less demand for metal dTr electrons. 4. CO groups cis to substituents of the type considered above should have higher stretching force constants than those trans to such substituents. This is

a direct corollary to 1 and 3 * 101

5. Stretch-stretch interaction constants should probably increase with increasing replacement of CO by ligands of lower TT-bonding ability. This is to be expected

since as the total number of dn electrons per CO increases3 the effect responsible for the interaction constants* described in lj should be magnified. The exact secular equations may be set up using the following constants: k-^ and stretching constants

of CO's trans to a ligand other than CO and C03 respec­ tively; kc and k ^ representing stretch-stretch interaction

of cis pairs of CO groups; and k^ 3 representing interaction of a trans pair of CO groups. These may be simplified to approximate equations by making the substitutions:

ki — kc — kc — k^./2

which are dictated by rule 2 above and by the assumption that in practice it will be impossible to detect any meaningful difference between kQ and k^. It should be mentioned that from rule 4, ls>, must be greater than k^ if the results are to be meaningful. Taking the bands as assigned for Mn(C0)^SCN and

Mn(C0)^NCS5 the force constants may now be calculated with the aid of approximate secular equations and given below

in Table 8. (53)

The value of k2 is obtained from the solution of the secular equation for the B-^ stretching mode. The absorption band was present in the spectrum of the S-bonded 102 ' thiocyanate., but was absent in the acetonitrile spectrum of Mn(C0)^NCS because of the rigid Cij. molecular symmetry of the latter isomer. However* it is possible to calculate the frequency of the mode in Mn(C0 )5NCS from a -simul­ taneous solution of the three secular equations3 using the known values of the E, and A^(^) stretching frequencies.

TABLE 8 APPROXIMATE SECULAR EQUATIONS FOR THE CO STRETCHING MODESa IN THE POINT GROUP Ci|v

Symmetry Species of CO Stretching Modes Approximate Secular Equations < * 1 a J 1) H 2//. k± w CM = 0 < H 2pk± ^(1^+4^)- A K Bi 11 E A =

SL Force constants in dynes/cm.;M represents the reciprocal of the reduced mass of the CO group, viz. 0 (1 6 .0 0 + 1 2 .oi)/(i6 .oo x 1 2 .0 1 ) = 0 .1 4 5 8 3 5A = (0 .0 5 8 8 9 0. where \> is the frequency in cm-1.

The value of k^ may be calculated by substitution of k£ (calculated as described above) and the stretching frequency of the E mode into the secular equation for the doubly degenerate vibration.

Finally5 the value for the force constant k^ is calculated by substitution of k ^ ki5 and the frequencies 103

(cm-1) of the (2 ) and (1 ) modes into the remaining secular equation.

It is then of interest to compare the values of the force constants obtained for Mh(C0)^SCN and Mn(C0)^NCS with those reported for the corresponding halogeno complexes.

(5 6) Table 9 shows this comparison.

TABLE 9 FORCE CONSTANTS (MDYNES/A) FOR Mn(C0)5X COMPLEXES

Compound Symmetry B^(cm“-*-) kl k2 ki

Mn(C0)5Cla C4y 2 0 9 0 d 16.27 17.63 0 .2 2 0

Mn(C0)5Bra > 2083d 16.31 17.53 0.23 Mn(C0)5Ia 2 0 7 6d 17-41 0 .2 2 °4v 16.33 Mn(C0)5NCSb 2 0 8 3d 15.64 17.52 0.25

Mn(C0)5SCNc cs 2094 15.47 17.52 0.32

aInfrared frequencies obtained in chloroform. Infrared frequencies obtained in 1,2-dichloro- ethane. cInfrared frequencies obtained in acetonitrile. ^Calculated from the infrared vibrational frequencies.

A very interesting observation can be made from comparison of the force constants and infrared spectra of the thiocyanato and halogeno derivatives of manganese carbonyl. As was stated earlier, refers to the CO group trans to the halogeno or thiocyanato group, and the A^C-O 104 band is also due to the vibrational mode,of the same carbonyl. Comparing the k-^ constant and frequencies for the thiocyanato and halogeno derivatives, one finds that both of them are substantially lower in the thio­ cyanato derivative.. The k-^ values are. 15.47 for Mn(CO)^SCN and about 16.3 for the halogeno derivatives. The A^C-1-) stretching frequencies are from 40-50 cm”-1- lower for Mn(CO)^SCN than for the Mn(CO)^X complexes. The comparison of these values leads one to conclude that the metal-sulfur Tf-bond is very weak indeed. This may result from polarization of the sulfur atom, which now becomes a strong electron donor and is not competing effectively for TT-bonding with the CO groups. In other words, the polarizing effect of the CO groups has caused the (T-bonding between manganese and sulfur to be at a maximum, whereas the TT-bonding has been reduced to a minimum. The arguments in support of these statements are given below. The fact that the values of both k^ and the stretching mode are lower for Mn(CO)^SCN than for the corresponding halogeno complexes indicates that the C-0 bond order in the carbonyl group trans to the thiocyanate is less than the C-0 bond orders in the carbonyls trans to the halide ligands. As was stated in the third deduction of Cotton (p. 100), the force constant of a CO group decreases with decreasingTT-bonding (increasing (T-bonding) 105 of the group trans ot it. It follows that as the

TABLE 10 THE A-^1 ) STRETCHING MODES AND kx VALUES FOR MoL(C0) 5 MOLECULES

Compound Ai^ ) (cm--1-) k^ (mdynes/A)

m o (p c i 3 )(c o )5 2 0 0 1 1 6 .3 8

m o (p c i 2 o c 2h 5 )(co)5 1987 16.15

M o [P(OC2H5)3](c o )5 1975 15.97

m o [p (o c h 3 )3 ](c o )5 1967 1 5 .6 2

m o [p (c 2 h 5 )3 ](c o )5 1941 15.40

m o (c 6h 11n h 2 )(c o )5 1895 14.65 Mo(DMF)(C0) 5 1847 13.93

From the above considerations, it is now possible to explain why the thiocyanate group is bound through sulfur in the Mn(C0)^SCN complex. Since the five CO groups exert such a great polarizing effect on the ligand in this system, the more polarizable atom in the thiocyanate group, namely sulfur, is preferred for bonding to the metal. It will also be noticed that the force constant k-^ for Mn(C0)cjNCS, while slightly higher than that for Mn(C0)^SCN, is also substantially lower than the k-^ for the corresponding halogeno complexes. Moreover, the 107 stretching frequency due to the A-^(-0 mode for Mn(C0)^NCS is also significantly lower (about 40 cm"-'-) than that of the chloro* bromo* and iodo carbonyl compounds. These observations indicate that the nitrogen end of the thio­ cyanate ligand* although not as polarizable as the sulfur atom, is still a very good electron donor in this complex. The fact that the force constant and vibrational frequency values for the thiocyanato-S and -N forms are similar lends

support to the author's previous conclusion (p. 8 0) that the energy difference between these two forms is small. With the knowledge of these values* it is not difficult to understand the existence of an equilibrium in solution.

Investigation of the derivatives of the ^(d'OjejCNS system

Properties of the derivatives. The physical properties of the derivatives of Mn( COLONS* including color* electrical conductivity* and solubility are strik­ ingly similar. For example* the complexes are all yellow powders. The molar conductivities of the pyridine* 2*2'- bipyridine* p-fluoroaniline* p-chloroaniline* 4-picoline and trans-bistriphenylphosphine derivatives were found to

be less than 1 ohm--*- in nitromethane.

The complexes Mn(C0 )3py2 CNS* Mn(C0 )3 (4 -pic)2 CNS*

Mn(C0 )3 (p-tol)2 CNS* and Mn(C0 )3 (pca)2 CNS are readily soluble in dichloromethane* nitromethane* methanol* and acetonitrile. The solubilities of the 2*21-bipyridine and - 108 p-fluoroaniline derivatives parallel closely those of the above .tricarbonyls, but are much lower in chloroform. All the above derivatives are slightly soluble in benzene and insoluble in petroleum ether. The cis and trans-disubsti- tuted and monosubstituted derivatives containing triphenyl- phosphine, -arsine and -stibine have a slightly' wider range of solubilities. In addition to dissolving in dichloro- methane, chloroform, nitromethane, methanol and acetonitrile, they are also readily soluble in benzene and slightly soluble in petroleum ether. Structures of the derivatives. The diagrams of octahedral structures of the cis and trans-disubstituted and monosubstituted derivatives are given below. The compounds cis[Mn^O^LgCNS] (L = pyridine, p-chloroaniline, p-fluoroaniline, 4-picoline, p-toluidine, triphenylarsine, and triphenylstibine) have the following structure:

/s (Clr a) x 0C: \ / Ml1V 0C J-Jr \ CO y The chelate derivatives of 2,21-bipyridine and 1,2-bis- (diphenylphosphino)ethane are of the structure:

The structures of the trans-disubstituted complexes, trans­ it Mn( CO) ^I^CNS] (L = triphenylphosphine, -arsine, and -stibine), are presumed to be:

rather than 110 on the basis of the evidence and arguments presented by Basolo et al. (27) The monosubstituted derivatives, Mn(C0)^LCNS (L = triphenylphosphinea -arsine, and -stibine), are of the structure:

The above structures have been included in order to facilitate discussions of the infrared spectra, given later. Infrared spectra of the derivatives. The spectra of the derivatives of the thiocyanato manganese carbonyl system are readily predicted from group theoretical considerations. It was therefore of interest to examine both the carbonyl and CS stretching frequency regions of the derivatives in order to elucidate their structures and determine the mode of attachment of the thiocyanato group. The cis-disubstituted derivatives of the thiocyanato manganese carbonyl system all belong to the point group Cs. Using the character table given earlier (p. 72) for group Cg symmetry, one predicts three CO stretching frequencies,

2A1 + A". Two bands result from stretching modes which Ill are symmetric with respect to reflection in the plane of symmetry, whereas a third hand arises from a CO stretch which is antisymmetric with respect to such reflection. As was mentioned earlier, in order for a vibrational mode to he infrared active, a change in dipole within the molecule must occur with the stretch. Examination of the three stretching modes indicates that all of them are infrared active, and hence expected to he strong in intensity. No unequivocal assignment can he made for the three frequency hands observed in the infrared spectra. However, as was pointed out in the presentation of the assumptions for force constant calculations, and as was mentioned again in the discussion of the band assignments of Mn(CO)^CNS, a decrease in the it-bonding capacity of a ligand causes a shift in the stretching frequency of the trans CO group toward lower wavenumbers. Therefore, in the complexes cis- [^(CO^I^CNS], where L tt-bonds appreciably with the metal, the lowest frequency absorption band is due to the A* mode of the CO group trans to the thiocyanate ligand. Thus, for the molecules whose cis ligands are pyridine, 4-picoline,

triphenylarsine, triphenylstibine and the chelates 2 ,2 '-

bipyridine and l,2 -bis(diphenylphosphino)ethane, the above statement should hold true since these ligands participate to an appreciable extent in IT-bonding with the metal. In

complexes of the type cis-[Mn(C0 )3L2 CNS] where the ligands 112

L do not IT-bond significantly with the metal, at least two CO bands should occur at similar frequencies. This feature is expected to occur in the spectra of the derivatives containing p-toluidine, p-chloroaniline and p-fluoroaniline. No unequivocal assignment can be made for the CO bands in the spectra of the latter molecules. In addition to the CO and CS bands, these carbonyl derivatives also exhibit a CN stretching frequency. The bands due to CN vibrations occur at wavenumbers higher than those for any of the CO vibrational modes because the CN bond order in SCN is much less sensitive than that of CO to the nature of other ligands in the complex. Table 11 shows the values for the CN, CO, and CS stretching frequencies in the infrared.

Figures 7 and 8 show the CO and CS spectral

regions of Mn(C0 )3py2 CNS and Mh(C0 )3 (pfa)2 CNS, respectively. In the cases of the p-toluidine, p-chloroaniline and p- fluoroaniline complexes, ligand absorption occurs in the

7 8 0 -8 6 0cm--*- region. In order to unambiguously assign the correct band to the CS stretching mode, the spectra of the corresponding bromo complexes were taken in that region.

The dashed line in Figure 8 represents the spectrum of

Mn(C0 )^(pfa)2 Br in the CS stretching frequency region. The

derivative Mr^CO)^(4-pic)2 CNS also exhibits ligand ab­

sorption in the 7 8 0 -8 6 0cm"’*' region. However, the spectrum of the corresponding bromo complex was essentially TABLE 11

INFRARED SPECTRA OF cis--[Mn(C0)3L2 CNS] COMPLEXES (cm"-1-)

Compound CNa C0a CSb 2A» + A" Mn(C0 )^py2 CWS 2097m 2039s, 1954s, 1924s 813m

Mn(CO)3bipyCNS 2107m 2 0 3 8s, 1951s, 1941s 813m

Mn(CO)3 (4-pic)2CNS 2097m 2039s, 1953s, 1 9 22 s (808m)

Mn(CO)3 (p-tol)2CNS 2l42m 2 0 3 6s, 1954s, 1942s 800m

Mn(CO)3 (pca)2CNS 2l42m 2039s, 1953s, 1948s 800m

Mn(C0)3 (pfa)2CNS 2142m 2039s, 1952s, 1946s 803m

Mn(CO)3 (dlphos)CNS 2110m 2033s, 1965s, 1930s 8l8m

Mn(CO)3 (AsPh3)2CNS 2148m 20 3 0s, 1957s, 1931s —

Mn(CO)3 (SbPh3)2CNS 2l48m 2 0 3 0s, 1955s, 19 3 0 s

aTaken in CHC1 5 btaken in Nujol.

m = medium, s = strong.

H H UO 114

2200 , 20,00 , 18,00 J 1------1------1------1— cm -1

900 800 i. cm'

Figure J. The' carbonyl and CS stretching frequency regions of the infrared spectrum of Mn(C0 )3py2 CNS. The 1800-2200 cm"1 and 7 5 0 -9 0 0 cm”l regions were taken in CHClo and Nujol, respectively. 115

2200 , 2000 18.00 J L cm

r

900 800 cm

Figure 8. The carbonyl arid CS stretching frequency regions of the infrared spectrum of Mn(C0)ofpfa)oCNS. The dashed line represents the CS region of •Mn(C0)otp£a )2Br• The 1800-2200 cm“l and 750-900 cm-1 regions were taken in CHCl^ and Nujola respectively. 116 identical to that of the thiocyanato derivative. Spectra of Mn(C0 ) 3 (4-pic ^CNS, in the 6 8 0 - 8 6 0cm- 1 region were then taken using higher concentrations of the complex in Nujol. No new bands appeared either in the 680-720 cm-'*' or 780-860 cm"'*’ regions. However, the 1900-2200 cm"'*' region of this complex is practically identical with that for

Mn(C0 )2Py2 CNS; it is therefore assumed that the CS band of the 4-picoline derivative is covered up by a stronger

absorption band of the ligand which occurs between 8 0 2- 820 cm"'*'. The value for the CS stretching frequency of

Mn(CO)^(4-pic)2 CNS is given in parentheses in Table 11. The pattern of the CO bands in the infrared spectra of the trans-disubstituted derivatives can readily be explained from group theoretical considerations. These

molecules belong to the point group C2 V* which is repre­ sentative of a species containing a twofold axis and two vertical planes of symmetry. Calculations from the character table for group Co symmetry, given below, show that 3 CO bands, namely 2A^_ + fundamentals, are expected in the infrared spectra of these complexes.

E C2 v C2

’ A1 1 1 1 1 a 2 1 1 - 1 - 1 B1 1 - 1 1 - 1 B2 1 - 1 - 1 1

The A-^ bands result from those CO stretches which are symmetric with respect to the operation C2- (rotation 117 by 1 8 0° about the twofold axis). The band arises from a CO stretch which is antisymmetric with respect to the above rotation. One A-^ stretching mode represents vibration of the CO group directed along the twofold axis. The other stretching mode arises from a symmetric stretch of the two trans CO groups in xy plane. This symmetric stretch does not result in a change in dipole moment within the molecule, and hence the mode is infrared inactive. As pointed out previously for the molecules belonging to the point group

C4 V (p. 8 1 )5 these two A-j_ modes couple and some of the intensity is transmitted to the infrared inactive mode, causing it to become observable in the spectrum. This coupling therefore results in a decreased intensity of the infrared active Astretching mode. An alternative

explanation of the appearance of a weak A1 stretching mode can be presented as follows. Since the two axial ligands are different, inevitable distortion exists in the equatorial plane of the molecules. Therefore, a slight change in dipole moment within the complex is expected when a symmetry stretch of the trans CO groups in the xy plane takes place. Hence presence of a weak band due to this stretch is expected in the infrared spectra of the trans­ disubstituted derivatives. The band arises from the antisymmetric stretch of the two trans CO groups in the xy 118 plane. A change in dipole within the molecule occurs with this stretch, and hence the mode is infrared active. The infrared bands of these carbonyl derivatives

of group C2 V can be unequivocally assigned to the corres­ ponding vibrational modes. As explained by Orgel (47)*

the weak A ^ band is always higher in frequency than the other absorptions. In addition, all the bands may be assigned on the basis of their relative intensities. Therefore, for these systems one expects to observe a weak, strong, and medium intensity band, corresponding to the A-^C2 ), Bi, and A1 (1) stretching modes, respectively. Table 12 shows the CN, CO and CS stretching frequencies of the trans-disubstituted derivatives. Figure 9 shows the

spectrum of trans-[Mn(C0 )^(AsPh2 )2 cNS], which is represent­ ative of this group of carbonyl complexes.

TABLE 12 INFRARED SPECTRA OF trans-[Mn(CO^I^CNS] COMPLEXES (cm"1 )

Compound CNa C0a csb H Ax (2) Bi H Mn(CO)3 (PPh3 )2CNS 2 0 9 6m 2039w 1957s 1 9 2 6m 8 2 0m Mn(CO)3 (AsPh3 )2CNS 2 1 0 3m 2046w 1963s 1 9 2 7m • 8l4m

Mn(CO)3 (SbPh3 )2CNS 2097m 2033w 1 9 2 6 s 1 9 2 7m 8 2 0m

aTaken in CHCI3 . bTaken in Nujol. w = weak, m = medium, s = strong. 119

2200 | 20,00 , l8P°r.m-i

Figure 9« infrared 120

The monosubstituted derivatives of the type

Mn(C0 )4LCNS belong to the point group Cs . With the aid of the Cs character table* presented earlier (p. 72)* one predicts the appearance of four CO bands* 3A* + A" funda­ mentals* in the infrared spectra of these molecules.

Symmetry considerations dictate that two A 1 modes and the

A" mode are infrared active* whereas a third A 1 mode is infrared inactive. The three A' modes combine and transmit some intensity to the infrared forbidden fundamental. However* because there are two active A r modes available to transmit intensity* the transfer process will not affect the intensities of these bands to an appreciable extent* and hence both active A ’ modes will give rise to strong absorptions. The alternative argument* based on the inevitable distortion of the equatorial plane in these molecules* and invoked previously for the molecules belonging to point groups C4V and C2 V * is also applicable here. Therefore* the infrared carbonyl spectra of the monosubstituted derivatives are expected to consist of one weak and three strong intensity bands. Table 13 gives the spectra of these complexes* whereas Figure 10 shows the

spectrum of Mn(CO)2jAsPh3CNS* which is typical of these derivatives.

Interpretation of infrared data in the CS stretching region. As was pointed out in the introduction (p. 14)* the truly diagnostic evidence for the mode of attachment of 121 the thiocyanato ligand (N or S) is the presence of the CS stretching frequency "between 7 8 0 and 8 6 0 cm"1 for the IT- bonded forms or between 6 8 0 and 720 cm"1 for the S-bonded complexes. Prom Tables 11 and 12, it may be noticed that the compounds which exhibit a CS stretching frequency in the 7 8 0 -8 6 0cm"1 region are the trans-disubstituted deriv­ atives and all the amine containing cis-disubstituted derivatives. In addition, the complex Mn(C0)3 (diphos)CNS also exhibits a band of medium intensity in the region

7 8 0 -8 6 0cm . On the basis of this common and diagnostic spectral feature, the above compounds are all assigned the thiocyanato-N structure.

TABLE 13 INFRARED SPECTRA OP Mn(C0)i|LCNS COMPLEXES (cm"1 )

Compound ON3, C0a CSb 3A* + A"

Mn(C0 )4 (PPh3 )CNS 2 1 5 1m 2097w 2045s 1963 s 1927s Mn(C0)^(AsPh3 )CNS 2 1 ^9 m 2 0 9 3 w 2037s 1 9 6 0s 1931s

Mn(C0 )4 (SbPh3 )CNS 2 1 5 6m 2 0 9 9 w 2046 s 1 9 6 2 s 1 9 3 8s

aTaken in CHC13 . ^Taken in Nujol. m = medium, w = weak, s = strong.

The cis-disubstituted derivatives of triphenylarsine and -stibine, along with the monosubstituted derivatives 2200 2000 t 1800 —* 1---|—— I--- L c m ‘‘

+,JfU?S+-10' ^The cart,onyl stretching frequency reeion the infrared.spectrum of.Mn(CO)4AsFh3CNS in CHCl^ 123 presented in Table 13* show no bands in the 7 8 0 - 8 6 0cm-1 region* and they are therefore assigned thiocyanato-S structures. No CS frequency could be detected in the 680-

7 2 0 cm-1 region due to strong phenyl ring absorption of the ligands. The same problem was encountered by Burmeister

and Basolo (5 8) with the complexes Pd(SbPh3)2(SCN)2 and

Pt(SbPh3 )2 (SCN)2 . °n the basis of a negative evidence* the two compounds were therefore assigned thiocyanato-S structures. The reason why some of the above derivatives have thiocyanato-N and others thiocyanato-S structures can readily be explained* at least in qualitative terms. The arguments for the particular modes of attachment of the thiocyanato group are given below. The thiocyanato-S structure of the parent complex* Mn(CO)cjSCN* was explained on the basis of the large polarizing effect of the CO groups exerted on the thio- cyanate ligand through the metal. By analogous reasoning* the bonding of the thiocyanato group in the monosubstituted derivatives* Mn(CO)ijLSCN* may be explained. Whereas the polarizing effect of five CO groups is certainly expected to be greater than that of four CO groups* there is still a substantial amount of it displayed in the monosubstituted derivatives. In addition* the ligands triphenylphosphine* -arsine* and -stibine can participate in TT-bonding with the metal* thus strengthening the polarizing effect which ' 124 is felt by the thiocyanato group. Therefore, as in the case of the parent MntCO^SCN, the more polarizable end of the thiocyanate group, namely sulfur, is preferred in bonding with the metal. Examination of the bonding in the derivatives reveals that in all the tricarbonyls, with the exceptions

of cis[Mn(C0 )3 (AsPh3 )2 SCN] and cis-CM^CO^SbPh^^SCN], the thiocyanate group is N-bonded. This apparent inconsistency can be readily resolved. Stuart and Briegleb molecular scale models show that a great deal of steric hindrance is inherent in the cis- disubstituted triphenylarsine and -stibine derivatives. Furthermore, no scale model of the complex cis- [Mn^CO^CPPh^^CNS] could be put together. It was observed that two triphenylphosphine groups cannot be placed in cis positions because the phosphorus atoms are too small in size; they are so small in fact that the bulky phenyl groups are drawn toward the metal, getting in the way of one another, the metal and the thiocyanate group. It is therefore significant that all attempts to synthesize this derivative were completely unsuccessful. The same diffi­ culty was encountered by Basolo and co-workers (27), who were unable to prepare the complex cis-CMh^CO^tPPh^^Br], presumably for the same reasons as given above. Increasing the size of the central ligand atom from phosphorus to arsenic moves the phenyl groups away from the 1 2 5 metal; now they are much less sterically hindered than in the corresponding triphenylphosphine complex. In the cis- bistriphenylarsine derivative, the bulky phenyl groups allow room for only slightly more than one large atom in axial positions. The angular -SCN fits in perfectly., the CN group pointing away from the phenyl groups. However, the molecular model of the corresponding nitrogen-bonded isomer cannot be prepared since in the linear MNCS attach­ ment, the carbon and sulfur atoms are severely hindered by the overlapping phenyl groups. The steric hindrance in the complex cis-

[Mn(C0 )3 (SbPh3 )2 SCN] is considerably less than in the corresponding bistriphenylarsine derivative, mentioned above. The larger size of the antimony atom relative to arsenic removes the phenyl groups far enough from the metal and thus relieves steric crovrding in the complex. However, models clearly indicate that there is still some interference between the phenyl groups and a nitrogen- bonded thiocyanate ligand in the complex cis- [Mn(C0)3(SbPh3)2NCS]. The S-bonded derivative, as observed from molecular scale models,is definitely less sterically hindered than its N-bonded counterpart. The photograph on the following page represents the molecular model of cis- [M^CO^CSbPh^^SCN]; the hindrance above the sulfur atom (yellow) is noticeable. The Stuart and Briegleb molecular scale model of cis-[Mn(CO)3 (S-bPh3)2SCN] . 1 2 7 The importance of steric hindrance in these systems is clearly established by the fact that the thiocyanato-N linkage is preferred in the complex Mn(C0 )2 (diphos)NCS. In spite of the relatively small size of the phosphorus atoms, steric hindrance in this molecule is not a signifi­ cant factor since only two phenyl groups are bound to each phosphorus atom. Indeed, molecular models of this chelate derivative indicate that steric hindrance is negligible for both the N-bonded and unknown S-bonded isomers. Further evidence in support of the importance of steric factors was obtained from the results of the synthesis of Mn(C0)3(SbPh3)2CNS in acetonitrile'. A sample of Mn(CO)^SCN was dissolved in acetonitrile and allowed to stand a sufficient length of time for the rearrangement to the N-bonded isomer to occur. Upon addition of triphenyl- stibine, the reaction was allowed to proceed at room temperature for 12 hours. Under identical conditions the reaction between Mn(C0 )5SCN and SbPh^ in chloroform resulted in a 90fo yield of cis-[Mn(CO)3 (SbPh3 )SCN] . However, infrared data showed that the product obtained from the reaction in acetonitrile was largely the trans­ disubstituted isomer (in which the thiocyanate group is N- bonded), with *the cis-disubstituted derivative present in t j — • smaller amounts. Recalling that trans-tMn^O^CSbPh^^UCS] could be synthesized only upon heating the reaction mixture of Mn(CO)^SCN and SbPh^ at 80°C., one obtains a definite 128 appreciation of the importance of steric factors in these systems. Steric factors are of key importance in explaining differences in the relative amounts of various products of the reactions involving Mn(CO),-SCN and AsPh^, and Mn(CO)^SCN and SbPh^. It was mentioned in the experimental section that the major product obtained from the reaction between the parent carbonyl and excess triphenylarsine at room temperature was the cis-disubstituted derivative. However, it was noted that invariably a substantial amount of the monosubstituted derivative was also present. Higher temperatures were required for the formation of the trans- disubstituted derivative, which could be obtained in good yields. Steric hindrance in cis-[Mn(C0)3(AsPh3 )2 SCN] is insufficient to prevent formation of the compound, but large enough to cause difficulty in introducing a second triphenylarsine group in a position cis to the first one. The Stuart and Briegleb model of the cis-disubstituted triphenylarsine derivative can be assembled; however, it clearly demonstrates the crowding of the bulky phenyl groups in the. molecule. On the other hand, a very good yield of the cis- disubstituted triphenylstibine complex is obtained under conditions similar to those above for the syntheses of the triphenylarsine derivative. The complex cis- [Mn^O^CSbPh^^SCN] is much less sterically hindered than 129 the corresponding cis-[Mn(CO)3(AsPh3)2SCN] 3 and hence the former derivatives may be obtained as virtually the sole product under the proper reaction conditions. The evidence presented above in support of the importance of the role of steric hindrance in these systems clearly strengthens the contention made previously (p. 124). In summary, aside from those systems where steric factors appear to be significant, all the disubstituted thiocyanato derivatives are of an N-bonded structure. Investigation of the CN vibrational frequencies. The infrared data presented in Tables 11, 12, and 13 include values for the CN stretching frequencies of the N- and S-bonded derivatives. Examination of these data reveals that the CN absorption in all thiocyanato-S derivatives occurs above 2148 cm--*-, whereas, with the exception of the aniline derivatives, the CN bands are below 2110 cm"*1- for the N-bonded complexes. The CN absorption for the derivatives Mn(C0)3 (p-tol^NCS, Mn(C0)3 (pca)2NCS, and Mn(C0)3 (pfa)2NCS, all at 2142 cm-1, are much closer to the frequencies found for the S-bonded complexes than to those for the other N-bonded thiocyaiiates. This unusual feature of the three derivatives cannot be readily explained. In an attempt to formulate an explanation for the anomalously high frequency of these CN bands, it is of importance to examine the CO and CS stretching frequency 130 regions in order to determine whether other irregularities are also present. As is revealed in Table 11* differences in the spectra between the aniline derivatives and the other N-bonded complexes do indeed exist. First of all, the band due to the lowest frequency A* stretching mode is about

2 5 cm-1 higher for the aniline compounds than for the pyridine derivatives. Furthermore, the CS stretching frequencies for the p-toluidine, p-chloroaniline and p- fluoroaniline derivatives are about 13 cm"’*' lower than those for the pyridine and 4-picoline complexes. However, the other CO stretching frequencies occur at similar wavenumbers for both aniline and pyridine derivatives. The appearance of CN and CS stretching frequency bands at higher and lower wavenumbers, respectively, is consistent with the resonance forms given earlier (p. 6). Furthermore, presence of the carbonyl A ’ band at higher wavenumbers is indicative of less -bonding between the metal and the CO group trans to the thiocyanate ligand. It is extremely difficult to offer an explanation which can take into account all the above observations without introducing inconsistencies in the argument. Perhaps when more examples of this type of behavior are reported in the literature, it will be possible to give a meaningful explanation. 131 The significance of the metathesis reactions ~ The significance of the reactions which produced the derivatives Mn(C0 )gpy2 NCS, Mn^O^bipyNCS, trans-

CMn(C0 )3 (PPh3 )2NCS]3 and Mn(CO)4AsPh3SCN from the corresponding chloro1'complexes, and Mn(CO)3 (4-pic^NCS,

Mn^O^Cpca^NCS, and Mn(C0 )^(pfa)2 NCS from the corres­ ponding bromo complexes and KCNS cannot be overlooked. Since these reactions were carried out over several hours and at elevated temperatures, the free thiocyanate ion had an equal opportunity to bond through either end to the metal. In.all cases, the product obtained from metathesis was identical to that isolated from the reaction of Mn^O^SCN with the appropriate ligand. The methatetical reactions clearly indicate that the thermodynamically stable product is being formed in each case. Therefore, These results eliminate the possibility that the complexes obtained from direct reactions between the parent thio­ cyanate and the ligands are the kinetically favored linkage isomers which could later, under the proper conditions, rearrange to the thermodynamically stable forms.

Linkage isomerism in the derivatives The incidence of the thiocyanato-S linkage in both the parent thiocyanato carbonyl and the monosubstituted derivatives was explained through arguments which were similar for the two cases (p. 123). It then followed that 132 perhaps the energy difference.between the N-bonded and S- bonded forms of the monosubstituted derivatives, like those of the pentacarbonyl, would also be small. The compound Mn^O^AsPhcjSCN was the first to be investigated for possible linkage isomerism. I A sample of the mono-arsine derivative was dissolved in acetonitrile and an infrared spectrum of this solution was taken in the' carbonyl stretching frequency region. A dramatic difference was observed between this spectrum and the one recorded in chloroform. In addition to changes.in the CO stretching frequencies, the CN band observed at 2149 cm"-'- in the spectrum of the derivative in chloroform was shifted to 2 1 0 6 cm- 1 in the spectrum in acetonitrile. This shift on the CN band to 2106 cm"-*-, while not diag­ nostic, strongly suggests linkage isomerization from the thiocyanato-S to thiocyanato-N complex. Figure 11 shows the two extreme spectra, i.e., in CHCl^ and in CH3CN. Examination of the infrared spectrum of the derivative in acetonitrile in the 7 8 0 -8 6 0cm"-*- region revealed the presence of a weak band at 8l4 cm-’1'. This region was clear in the spectrum of the S-bonded complex in NuJ ol. The spectrum of the S-bonded isomer, taken in

1 ,2 -dichloroethane, was found to shift slightly with rising temperature. The shift was toward the spectrum of

Mn(CO)2jAsPh.3CNS in acetonitrile, and hence parallel to that 2 2 0 0 , 20,00 , 1800 J—J- 1 I L cm ' 22,00 . 20.00 cm*"'

CH,CN

CH Cl:

Figure 11. The carhonyl stretching frequency regions of the infrared spectra of Mn(CO)j^AsPh^CNS in CHCl^ and CH^CN, 133 134 observed for Mn(CO)^CNS under similar conditions. The spectrum of the arsine derivative was also found to be sensitive to the nature of mixed solvent systems. Thus, spectra taken in mixtures of CHCl^ and CH^CN of variable composition were found to exhibit features characteristic of either of the structures, depending upon the relative amounts of the two solvents present. The evidence presented above clearly indicates that the thiocyanato-N isomer, Mn(C0)2jAsPhgNCS, is present in acetonitrile. Furthermore, the behavior of this derivative in the various tests of linkage isomerism closely resembles that of Mn(CO)5SCN. Attempts to isolate M^CO^AsPh^NCS from aceto­ nitrile solutions were unsuccessful. Thin film spectra showed that rearrangement from the N-bonded back to the S- bonded. isomer occurred immediately upon evaporation of the solvent. When acetonitrile bands were no longer observable, the spectrum of the derivative was identical to that initially recorded in Nujol, and representative of the S- bonded isomer. In these experiments, the derivative Mn^Oj^AsPh^SCN could be recovered from solutions, showing little, if any, decomposition. A similar investigation was carried out on the

derivative Mn(C0 )i|PPh3SCN with the hopes of perhaps discovering yet another linkage isomer system. 135 The spectrum of this compound in acetonitrile was indeed different from the one obtained in chloroform. However, it indicated that a complete rearrangement to the N-bonded form did not occur; instead, an equilibrium between the thiocyanato-S and thiocyanato-N isomers was attained with appreciable amounts of both forms being present. That the equilibrium was reached rapidly is supported by the lack of change in the spectrum as a function of time. Indeed, infrared spectra of solutions which were allowed to stand for periods up to 75 minutes were identical to those obtained five minutes after dissolution in acetonitrile. The spectrum of this derivative was also found to. be sensitive to the nature of solvent systems such as

CH3CN-CHCI3 . Thus, spectra taken after addition of

chloroform to a solution of Mn(C0 )i|PPh3CNS in acetonitrile showed a shift in equilibrium toward the S-bonded form.

Thiocyanato derivatives of the group Vlb h exa c a rbony1 s It was mentioned in the discussion of the derivatives of the Mn(CO),_CNS system that a negative charge buildup on the metal led to nitrogen bonding of the thiocyanato group. This view was supported by appropriate arguments. However, in order to unambiguously assess the effect of the metal charge on the nature of the thiocyanato-metal attachment, the anionic thiocyanatopentacarbonyl complexes of tungsten,, 136 molybdenum and chromium were prepared and examined with regard to the bonding. Properties of these derivatives. The compound (CH3)4N[W(CO)5CNS] is a deep yellow powder, very stable in the cold and under nitrogen, and decomposing only slowly on prolonged exposure to air at room temperature. The compound is a uni-univalent electrolyte in nitromethane, the value for its molar conductivity being 95 ohm-1. This anionic thiocyanato carbonyl derivative was found to be readily soluble in the more polar solvents such as nitromethane, diglyme and dichloromethane, moderately soluble in chloro­

form, and insoluble in low boiling (3 0-6ocC.) petroleum ether. The molybdenum derivative, also obtained as a deep yellow powder, decomposes readily to a red brown solid, which was shown to contain no carbonyl groups by infrared spectroscopy. This color change becomes noticeable after

a 1 0 -1 5 minute standing in air at room temperature. Although the compound could not be characterized accurately through elemental analyses, the analytical results obtained were in the region expected for the pentacarbonyl decom­ posing with loss of CO. An attempt was made to measure the volume of CO evolved on addition of a solution of iodine in pyridine to the complex. However, the best results

showed that only 80% of the calculated amount of CO had been collected. This compound has the same solubility 137 properties as the tungsten complex mentioned above, except that it is readily soluble in chloroform. The chromium derivative is similar to the tungsten complex in stability; it is indefinitely stable in the cold and under nitrogen, and decomposes slowly on standing in the air at room temperature. This pentacarbonyl, obtained also as a deep yellow powder, was found to be a 1:1

* 1 electrolyte in nitromethane (molar conductivity 96 ohm ). It is readily soluble in the more polar solvents, including chloroform, but insoluble in hydrocarbons. Infrared data and interpretation. The CN, CO and CS infrared stretching frequencies of these compounds are given in Table l4. Figure 12 shows iirqportant regions of the infrared spectrum of (CH2 )2jN[W(CO)^CNS] ; the other two spectra (chromium and molybdenum) are similar. It should be mentioned that the shoulder which appears at about

1 9 9 0 cm-^ on the strongest band is due to a trace of the starting material, W(CO)g. In order to obtain a meaningful spectrum of the unstable (CH3 )Z|N[Mo(CO)5CNS] compound, the crude thiocyanato carbonyl was extracted with chloroform, leaving behind the insoluble red-brown residue. The pure compound was then precipitated from solution by addition of petroleum ether, and the spectrum in the 6 8 0 -8 6 0cm”'1' frequency region was run immediately before any appreciable decomposition had occurred. The only absorption band due to the anion was 138 observed at 794 cm"-1-. However, when a Nujol suspension of the carbonyl was allowed to stand at room temperature, there was a gradual disappearance of the band at 794 cm"1 and a simultaneous appearance of one at 758 cm"-1-. The latter has been identified with the red-brown, non-carbonyl decomposition product. Since no new absorption bands were observed in the 6 8 0 -8 6 0cm- 1 spectral region, one may conclude that the molybdenum complex does not undergo linkage isomerization, but decomposes directly to the red- brown material.

TABLE 14

INFRARED SPECTRA OF (CH3 )4N[M(C0 )5CNS] (cm"1)

Compound CN c o a CSb

* , ( 2 ) E A ^ 1 )

(c h 3 )4n [w (c o )5c n s ] 2105m 2067W 1 9 2 0 s 1 8 7 3m 802m

(CH3 )4N[Mo (CO)5CNS] 2 1 0 5m 2 0 7 3w 1 9 4 1 s 1 8 7 6m 794m

(CH3 )4N[Cr(CO)5CNS] 2 1 1 5m 2 o 6 6 w 1934s 1 8 82m 791m

8* Taken in CHCI3 for the Mo and Cr compounds\ in CHgClg for the W compounds. ^Taken in Nujol. w = weak, m = medium, s = strong.

To guard against a possibility that the observed bands at 7 9 0 -8 1 0 cm- 1 might be due to modes other than the

C-S stretch, the spectrum of (CH3 )4N[W(CO)3Br] was examined 139

2 2 0 0 , 20,00 , 18,00 J 1------1------1------L cm"'

9 0 0 800 J _ J __ 1 cm "'

Figure 12. The carbonyl and CS stretching frequency regions of the infrared spectrum' of (CH3 )4 N[W(C0 )5CNS]. The 1800-2200 cm-1 and 750-900 cm"1 regions were taken in CHgClg and Nujol, respectively. i4o carefully in this region. No absorption due to the brorao complex was found. For all three compounds* the weak., strong* and lower frequency medium intensity bands in the 1 8 0 0 -2 1 0 0 cm“’L region are due to the carbonyl stretching modes

(2A^ + E) and indicate the C4v symmetry of the ions. The frequencies of these bands correspond closely to those for the anionic chromium* molybdenum and tungsten halogeno- pentacarbonyls* reported by Abel et al. (24) Medium intensity absorptions in the 2100-2120 cm"'1' and 7 9 0 -8 1 0 cm-1 regions are assigned to the CN and CS stretching modes* respectively. Since the known N-bonded thiocyanates absorb in the 7 8 0 -8 6 0cm-1 frequency range (11)* the spectral data support the M-NCS bonding assignment for all three complexes. Thus it may be concluded that a decrease in the charge on the metal from Mn(C0)^SCN to the isoelectronic [Cr(C0)^NCS]” changes the bonding from M-SCN to M-NCS. This observation is in complete accord with the results on substituted thiocyanato manganese carbonyl derivatives* and unambiguously shows the influence of the metal charge on the nature of M-CNS attachment.

Investigation of the thiocyanato iron carbonyl system With the knowledge that the thiocyanato-S linkage occurs with manganese(l) in the complex Mn(C0)^SCN* and l4i that the thiocyanato-N attachment is preferred with chromium(O) in the isoelectronic [Cr(CO)^NCS]“, it was then of interest to attempt a synthesis of a simple thio­ cyanato carbonyl of a metal in the +2 oxidation state, namely Fe(C0 )ij.(SCN)2 . The interest in this iron carbonyl derivative concerns the mode of attachment of the thio- cyanato group. Having established the fact that in the d electronic systems the zero oxidation state promotes N- bonding in carbonyl thiocyanato metallates and that manganese(I) in Mn(CO)^(CNS) may bind the thiocyanato

ligand through either end, it was thought that Fe(C0 )^(CNS)2 would perhaps prefer exclusively a sulfur attachment with the thiocyanato group. Unfortunately, all attempts to prepare the iron thiocyanate carbonyl compound were unsuccessful. Several different approaches were tried; all of these reactions yielded products containing no carbon monoxide. It is pertinent to add that a structurally similar complex of the isoelectronic Co(III), [Co(CN)^SCN]^”, was

recently prepared by Burmeister (6 ), who assigned to it a thiocyanato-S structure on the basis of a weak band in the 690-720 cm”’1' spectral region. This synthesis and the fact

that the complexes Fe(C0 )^X2 (X = Cl, Br, I) are known tend to complicate the problem of elucidation of the factors responsible for the lack of success in the attempted

preparations of Fe(C0 )||(CNS)2 . BIBLIOGRAPHY . 143 Lewis, J., Nyholm, R. S., and Smith, P. W., J. Chem. Soc. 4590 (1961). Chatt, J., Duncanson, L. A., Hart, P. A., and Owston, P. G., Nature 1§1, 43 (1 9 5 8). Owston, P. G., and Rowe, J. M., Acta. Cryst. 13, 253 (I9 6 0 ).

Chatt, J., and Hart, F. A., J. Chem. Soc. l4l6 (1 9 6 1). Mitchell, P. C. H., and Williams, R. J. P., J. Chem. Soc. 1912 (I960). Burmeister, J. L., Inorg. Chem. 3* 919 (1964).

Jones, L. H., J. Chem. Phys. 25, 1069 (1 9 5 6). Pauling, L., Nature of the Chemical Bond, Cornell University Press, Ithaca, 1.945, p. 195* Chamberlain, M. M., and Bailar, J. C., J. Am. Chem. Soc. 81, 6412 (1959). Chatt, J., and Duncanson, L. A., Nature, 178* 997 (1956).

Turco, A., and Pecile, C., Nature 1913 66 (1 9 6 1). Fujita, J., Nakamoto, K., and Kobayashi, M., J. Am. Chem. Soc. 7 8, 3295 (1956). Yamada, S., and Tsuchida, R., Bull. Chem. Soc. Japan 2 6, 492 (1953).

Williams, R. J. P., Proc. Chem. Soc. 20 (i9 6 0 ). Ahrland, S., Chatt, J., and Davies, N. R., Quart. Rev. 12, 265 (1 9 5 8).

Pearson, R. G., J. Am. Chem. Soc. 8 5, 3533 (1 9 6 3). Basolo, F., and Burmeister, J. L., J. Am. Chem. Soc. 85, 1700 (1963). Abel, E. W., and Wilkinson, G., J. Chem. Soc. 1501 (1959). Brimm, E. 0., Lynch, M. A., and Sesny, W. S., J. Am. Chem. Soc. 7 6, 3 83 1 (1954).

Hieber, W., and Wollman, K., Ber. 94, 305 (1 9 6 1). 144 (21) King, R. B.* and Stone* F. G. A.* Inorganic Syntheses* Vol. VII* McGraw-Hill Book Co.* Inc.* New York* “ 1 9 6 3* p. 198. (22) Hieber* W.* and Bader* G.* Ber. 6l* 1717 (1928).

(2 3 ) Hieber* W.* and Bader* G.* Z. Anorg. Allgem. Chem.* 190j 193 (1930); 201* 329 (1931). (24) Abel* E. W.* Butler* I. S.* and Reid* J. G.* J. Chem. Soc. 2068 (1 9 6 3). (25) Kilner* M.* Private communication. (26) Hieber* W .* and Schropp* W. * ZNaturforsch. l4B* 460 (1959). (2 7) Basolo* F .* Angelici* R. J.* and Poe* A. J.* Nature 195, 993 (1 9 6 2 ). (28) Angelici* R.' J.* and Basolo* F.* J. Am. Chem. Soc.* 84* 2495 (1 9 6 2 ). (29) Schropp* W.j Doctorate Thesis* Technische Hochscule Munchen* i9 6 0 * referred to in Hieber* W.* Beck* W.* and Tengler* H.* Z. Naturforsch. 15B* 4ll (i9 6 0 ).

(30) Abel, E. ¥.* and Butler* I. S.* J. Chem. Soc. 434 (1964). (31) Booth* G.* and Chatt* J.* J. Chem. Soc. 2099 (1962).

(32) Hoch* H.* and Stuhlman* H.* Ber. 62* 431 (1 9 2 9 ). (33) King* R. B.* and Stone* F. G. A.* Inorganic Syntheses* Vol. VII* The McGraw-Hill Book Co.* Inc.* New York* 1963, p. 193. (34) Wood* J. L.* Organic Reactions * Vol. Ill* John Wiley and Sons* Inc.* New York* 1946, p. 240. (35) Angus. A. B.* and Bacon* R. G. R.* J. Chem. Soc. 774 (1958); ibid* 778.

(3 6) Angus* A. B.* Bacon* R. G. R.* and Guy* R. G.* Chem. and Ind.* 4l* 1281 (1 9 6 3). (37) Herstein* K. M.* Inorganic Syntheses* Vol. Ill* The McGraw-Hill Book Co.* Inc.* New York* 1950* p. 24. 145 Weidner, C., German Patent 964,902, Chemical Abstracts 54, p6o47c (i9 6 0 ). Sandell, E. B., Colorimetric Determination of Traces of Metals, 2d ed., Interscience Publishers, Inc., New York," 1 9 5 0, p. 430. Wojcicki, A., and Farona, M. F., Inorg. Chem. 3j 151 (1964). Wojcicki, A., and Farona, M. F., J. Inorg. Nucl. Chem., In press. Moeller, T., Inorganic Chemistry, John Wiley and Sons, Inc., New York^ 1952, pp. 465-4o6. Cotton, F. A., and Wilkinson, G., Advanced Inorganic Chemistry, John Wiley and Sons, Inc., New York, 1962, p. 457. Joshi, K. K., and Pauson, P. L., Z. Naturforsch. 17B, 5 6 5 (1 9 6 2 ). Nesmeyanov, A. N., Anisimov, K. N., Kolobova, N. E., and Kolomniker, A. K., Invest. Akad. Nauk. S.S.S.R., Otdel. Khin. Nauk., 194 (1 9 6 3). Calderazzo, F. Cini, R., Corradini, P., Ercoli, R., and Natta, G., Chem. & Ind. (London), 500 (i9 6 0 ). Orgel, L. E., Inorg. Chem. 1^, 25 (1962). Gray, H. B., Billig, E., Wojcicki, A., and Farona, M. F., Can. J. Chem. 4l, 1281 (1 9 6 3). Schaffer, C. E., International Conference on Coordin­ ation Chemistry, Chem. Soc. Special Publ. No. 13, 153 (1959). Iliceto, A., Fava, A., and Mazzucato, U., Tetrahedron Letters 11, 27 (i9 6 0 ). Iliceto, A., Fava, A., Mazzucato, U., and Rosetto, 0., J. Am. Chem. Soc. 8 3, 2 7 2 9 (1 9 6 1 ). Iliceto, A., Fava, A., Mazzucato, U., and Radici, P., Gazz. Chim. Ital. 9 0 , 9 1 9 (i9 6 0 ). Cotton, F. A., and Kraihanzel, C. S., J. Am. Chem. Soc. 84, 4432 (1962). 146

(54) El-Sayed, M. A., and Kaesz, H. D., J. Mol. Spect. 9> 310 (1962). (55) Kraihanzel, C. S., and Cotton, F. A., Inorg. Chem. 2, 533 (1963). (5 6) Cotton, F. A., Inorg. Chem. 3, 702 (1964). (57) Nakamoto, K., Infrared Spectra of Inorganic and Coordination Compounds, John Wiley and Sons, Inc., New York, 1 9 6 3> P* 10. (5 8) Burmeister, J. L., and Basolo, F., Inorg. Chem., In press.