Four Decades of Metal Carbonyl Chemistry in Liquid Ammonia: Aspects and Prospects

ADVANCES IN ORGANOMETALLIC CHEMISTRY, VOL. 18

Four Decades of Metal Carbonyl Chemistry in Liquid Ammonia: Aspects and Prospects H. BEHRENS lnsfifute of Inorganic Chemistry University of Erlangen-NUrnberg Federal Republic of Ciermany

.

.

I. Introduction . . . . . . . . . . . . . . . . 11. Reduction of Metal Carbonyls with Alkali Metals and Sodium Tetra. . . . . . . hydridoborate in Liquid Ammonia . . A. Experimental Results . . . . . . . . . . . . . B. Structures of the Carbonyl Metalates [M(CO)JJe-,[M2(CO),,Je-, and . . . . . . . . . [HM&O),,J- (M = Cr, Mo, W) 111. Reactions of the Pentacarbonyl Metalates(-11) and the Decacarbonyl Dimetalates(-I) of Chromium, Molybdenum, and Tungsten . . . A. Monosubstituted Derivatives of the Hexacarbonyls . . . . B. Mono- and Dinuclear Hexacoordinated Pentacarbonyl Metalates(0) . . . . . . . . . . . . . . . . . C. Heterometallic and Inserted Anionic Complexes . . . . . D. Paramagnetic Mono- and Dinuclear Pentacarbonyl Complexes of Chromium . . . . . . . . . . . . . . . . . E. Complexes with M-M‘ Bonds ( M = Cr, Mo, W: M’ = Ge, Sn, In) . . . . . . . . . . , , . . . . . IV. Reactions of Metal Carbonyls and Metal Carbonyl Derivatives with Liquid Ammonia . . . . . . . . . . . . . A. Substitution of CO by NH, and Addition of NH, . . . . . . B. “Base Reactions” and Disproportionations , . . . . . . C. Carbamoyl Complexes of Transition Metals with the Ligand CONH,. . . . . . . . . . . . . . . . . D. Liquid Ammonia as Solvent for Organometallic Reactions . . . E. Cyanocarbonyl Metalates . . . . . . . . . F. Mixed Cyanocarbonyl Metalates Prepared from Metal Carbonyl Derivatives and NaN(SiMeJ, . . . . . . . . . . . V. Reactions of Metal Carbonyls and Metal Carbonyl Derivatives with the Multidentate Nitrogen Ligands bipy, phen, and terpy . . . . . A. Titanium and Vanadium . . . . . . . . . . . . . B. Chromium, Molybdenum, and Tungsten . . . . . . . . C. Iron and Cobalt . . . . . . . . . . . . . . VI. High-Pressure Syntheses . . . . . . . . . . . . . . VII. Conclusions . . . . . . . . . . . . . . . . . References . . . . . . . . . . , . . . . . . .

. .

.

. .

.

.

.

.

. .

.

.

.

.

. . . .

.

2 4 4 7 10 10

12 14 16

17 19 19

20 22 33 36 38 42 43 43 45 46 47 48

1 Copyright @ lW by Academic Press. Inc. All rights of reproduction in any form resewed. ISBN 0-12-0311166

2

H. BEHRENS

I INTRODUCTION

While studying chemistry in the years 1933-1937 at the University of Freiburg and at the Technische Hochschule in Munich, I had the good fortune early in my career to come into contact with many famous and important chemists such as Hans Fischer, Georg von Hevesy, Walter Hieber, Otto Honigschmidt, Hermann Staudinger, Heinrich Wieland, and Eduard Zintl. The lectures on complex chemistry given by Walter Hieber, who, at the age of 39, was appointed Professor in Munich in 1935, made a lasting impression, both from their content and from the didactical excellence of their presentation. I was unusually fascinated by his first review on metal carbonyls ( I ) , which appeared in 1937, as it treated a class of compounds which, at that time, could not be reconciled at all with the normal concepts of inorganic complexes. In the same year I, as a graduate student, was involved with the reactivity of the recently discovered HCo(CO),, which is gaseous at room temperature. In the course of this work, I made my decision to make metal carbonyl chemistry the subject of my dissertation. This decision led to an almost 25year-long collaboration with Walter Hieber. At this time, the chemistry of metal carbonyls was very much in its first stages of development. It is significant that the systematic study of this field, which had hardly begun to be covered in the textbooks, was almost exclusively carried out by Hieber. Other than the mononuclear metal carbonyls of the VIB Group, iron, ruthenium, and nickel, only the polynuclear species FedCO),, RudCO),, CodCO),, FedCO),,, and C O ~ ( C Owere ) ~ ~known (I). Of these, only the structures of the hexacarbonyls of the chromium group (2) and of FedCO), (3) had been determined by X-ray structure analyses. With respect to the derivatives of metal carbonyls, the substituted metal carbonyls of the VIB Group (e.g., Mo(CO)#yJ, the halogenocarbonyls of iron, ruthenium, iridium, and platinum, the hydridocarbonyls H2Fe(CO),and HCo(CO), discovered in 1931 and 1934, and the nitrosyl carbonyls Fe(CO)dNO), and Co(CO),NO were the most important (I). The known anionic CO complexes were limited to [HFe(CO)J- and ICo(C0)J-. For studies of substitution reactions of metal carbonyls at this time, work was almost totally limited to reactions involving the classical N ligands such as NH,, en, py, bipy, and phen. Thus, it is understandable that at the end of the 1930s research in the area of metal carbonyls was, for the most part, concentrated on new preparative and reaction methods in liquid media and via high-pressure syntheses starting from transition-metal compounds. The objectives were

Metal Carbonyl Chemistry in Liquid Ammonia

3

the discovery of new mono- and polynuclear metal carbonyls as well as further halogeno- and hydridocarbonyls. Furthermore, researchers were also concerned with the search for other ligands that could replace the CO of metal carbonyls and their derivatives. A final area of interest involved the reactions of metal carbonyls with various Lewis bases (4). One must bear in mind that at this time none of the spectroscopic methods, which are taken for granted nowadays, were available, so that resort had to be made to extremely careful analytical work. However, the so-called “noble gas rule” was an aid to the research at that time, a concept which should not be underestimated since it played a determining role in planning new experiments and contributed to many successes. I still admire the ability of Hieber, who, with the most modest of means, obtained the wonderful scientific results t h a t 4 0 years later-remain fundamental concepts of organometallic chemistry. Under these influences, my dissertation was concerned, on the one hand, with the reactions of NiS and CoS with CO in aqueous alkaline suspension (4) and, on the other, with high-pressure syntheses of Fe(CO),, CO~(CO)~, and Ni(C0)4 from anhydrous halides and CO in the presence of halogen-absorbing “metal additives” such as Cu and Hg according to (5) Nixl

+ 2Cu + 4CO

2W

Ni(CO),

+ 2CuX (X = C1, Br, I)

It was observed that the reaction of NiS or [Ni(NHdJ[MoSJ with CO in aqueous alkaline suspension could, in one case, be carried out with an absorption of 4 moles of CO per mole of Ni and, in the other, with only 50% of the total nickel being converted to Ni(CO), ( 4 ) . The reaction path, which could not at that time be clarified, was later studied by using CO absolutely free of oxygen, and we were able to determine that the quantitative formation of Ni(CO)4resulted from a reductive cahonylation

-

(6, 7): NiS

+ 5CO + 4 0 H - + Ni(CO), + COSg- + Sa- + 2H20

The corresponding reaction of CoS results in the formation of [Co(CO)J(6):

2C0S

+ I ICO + 120H- + 2[Co(CO)J- + 3COSg- + 2s’- + 6HsO

The reaction of NiSe with CO in aqueous alkaline suspension is analogous to that of NiS (8). The structures of the isoelectronic hydridocarbonyls H,Fe( C0)4 and HCo(CO), were the central theme of numerous studies over a period of many years after their discovery by Hieber. Because of their very similar physical characteristics, these complexes were considered to be “pseudo nickel tetracarbonyls” (H,Fe = HCo = Ni) in which, even then, the

4

H. BEHRENS

bonding of the hydrogen to the transition metals was intuitively postulated ( 1 ) . On looking back, it is surprising that it was relatively late that the definite acid character of HCo(CO), in aqueous solution was recognized (9, 10). The reason for this was undoubtedly the fact that, at that time, neither an isolation of the alkali-metal salts from the "base reaction" solution of CodCO), nor an "esterification" of HCo(CO),, i.e., the formation of CH,Co(CO),, had been achieved. However, in 1941, I was able to show that HCO(CO)~ is extremely soluble in liquid NH, and that NHJCo(C0)J can be isolated at room temperature. This complex can be sublimed at 80°C in vacuum to give long colorless needles (11, 12). With the deep-blue solutions of Li, Na, Ca, or Ba in liquid NHS, colorless alkali or alkaline-earth salts of HCo(CO), are formed with evolution of H2. Also, by the reaction of Cd[Co(CO)J2 with sodium in liquid NH3, the quantitative formation of Na[Co(CO)J takes place. Finally, at that time, we were able to prepare K[Co(CO)J via a reaction analogous to "neutralization" of NHdCo(C0)J with KNH2 in liquid NH3. From these studies, which began in 194 I , we showed that liquid NH3 is an excellent solvent for cahonyl chemistry, and that solutions of the alkali metals in liquid NH3function as extremely strong reducing agents (12). These studies represent the basis of "the chemistry of metal carbonyls in liquid NH3" which has concerned us continuously now for over 35 years and which has made possible the preparation of numerous new, particularly anionic, carbonyl complexes. Furthermore, these studies in liquid NH3 have also provided numerous ideas for further experimentation, by both our own and other research groups, which will be presented here. The following facts concerning the use of liquid NH3 are of particular importance: ( I ) its behavior as a solvent similar to water, but with a much lower proton activity; (2) the access to the reaction temperature range of -78 to +12O"C which made possible the preparation and isolation of complexes at low temperatures; and (3) the strong reducing character of the solutions of the alkali metals in liquid NH3, which especially allowed for access to the carbonyl metalates having very low oxidation states. II

REDUCTION OF METAL CARBONYLS WITH ALKALI METALS AND SODIUM TETRAHYDRIDOBORATE IN LIQUID AMMONIA A. Experimental Results

Although the reduction of such mono- and polynuclear metal carbonyls as Fe(CO),, MndCO),,, Re2(CO)lo,CodCO),, and FeACO),, with finely

Metal Carbonyl Chemistry in Liquid Ammonia

5

divided alkali metals or alkali-metal amalgams in indifferent solvents was first investigated by Hieber's group surprisingly late in the 1950s (13), beginning in 1951 we had success with the quantitative study of such reactions with sodium in liquid NH3 (14). Short reaction times at -78°C led to the formation of Na,[Fe(CO)J or Na[Co(CO)J from Fe(CO),, Fe,(CO),, and Fe3(C0)*,,or CO,(CO)~ and C O ~ ( C O )Halogenocarbonyls ~~. [e.g., Fe(CO)Jz] were also reduced to the respective monomeric carbonyl metalates. The "base reactions" of the Group VIB hexacarbonyls with KOH-methanol led to complicated polynuclear anions whose structures are still not clear. These contain OH- and CH30H ligands as well as CO (13), which means that the preparation of mononuclear pentacarbonyl metalates, such as [M(CO)5]2-or [HM(CO),]- (M = Cr, Mo, W), is not possible by this route because of their extremely strong reducing character. Although it took several years before the first preparation of their isoelectronic counterparts [Mn(CO),]- and HMII(CO)~(13), we had already formed the [Cr(C0),l2- and [HCr(C0)5]- anions by treating Cr(CO),, with sodium in liquid NH3 at -78"C, giving quantitative reduction to the soluble yellow penracarbonylchromate(-11) (15), Cr(CO)(I+ 3Na

4

Na,[Cr-"(CO)d

+ iNaOC&ONa

whereby most of the evolved CO forms the sparingly soluble disodium acetylenediolate (16). A little later, we were able to obtain the pentacarbonyl metalates(-11) N~,[Mo-~'(CO),]and Naz[W-ll(CO),] in low yields (17). The yields were low due to the predominance of further reduction to the sparingly soluble, mononuclear tetracarbonyl metalates(-1V) and formation of NazC2O2: Nh[ M-"(CO)d

+ 3Na 4 Nr4[M-IV(CO)+1+ i N a & a (M = Mo, W)

These complexes have been prepared and spectroscopically characterized by Ellis et al. (18) who used the reduction of (TMEDA)M(CO), (TMEDA = N,N,N',N'-tetramethylethylenediamine)in liquid NH3. The extremely reactive [M-11(CO),]2-anions (M = Cr, Mo, W) allowed the synthesis of numerous new metal carbonyl complexes of Group VIB, so that what was, until then, rather monotonous chemistry became much more versatile. In aqueous solution, the intermediate anion [HM(CO),]- initially formed releases Hz with complementary oxidation to the dinuclear hydrogen-bridged [(OC),M-H-M(C0)J anions (19), the preparation and structure of which will be discussed in detail. Z[HM(CO)J

+ HZ0 + [HMACO)J + OH- + HZ

Within the framework of our investigations into the reducing behavior

H. BEHRENS

6

of the hexacarbonyls, we became especially interested in preparative methods leading to the decacarbonyl dimetalates [Mz(C0),o]2-of Group VIB, which are isoelectronic with the dinuclear metal carbonyls of Group VIIB Mz(CO)lo(M = Mn, Tc, Re). These Group VIB metalates are obtained by closed-system reactions of the hexacarbonyls with NaBH, in liquid NH3 (20, 2 1 ) . 2M(CO)R+ 2NaBH,

+ 6NHs

80 'C

NadMACO),J + 2B(NHJs + 2CO

+ 7H2

It is interesting that, on the one hand, these yellow decacarbonyl dimetalates disproportionate with CO at 150°C, Nal[MI-'(CO)lo] + CO 4 Nal[M-''(CO)6]

+ Mo(CO)6

and, on the other, come to a pH-dependent equilibrium in aqueous solution: [Mz(CO)J-

+ HZ 0 * [HM,(CO)l,]- + OH'

The anions [M2(CO)10]2and [HM2(CO),,]- can easily be separated via their [Et,N]+ salts, as a consequence of the different solubilities of these salts in Et20 (20). In contrast to the reactions of the hexacarbonyls with NaBH, in liquid NH3, several days of reaction in boiling THF results in reduction to the deeply colored trinuclear carbonyl metalates Na2[M3(CO),J (M = Cr, Mo, W),provided the CO evolved is completely removed from the reaction system (22). Quantitative experiments showed that these anionic cluster-complexes react with bipy and H20 to give M(CO),bipy with evolution of exactly 2 moles of CO and 1 mole of H2 per mole of Na,[M,(CO),.J. In this respect, we plan to reinvestigate these complexes, as Churchill and Ni Chang (23) reported that the reaction of W(CO)8 with [Et,N]BH, in boiling THF gives red [Et4N][Wz(CO)8H2], the X-ray structure of which leads to the deduction of an anion with two bridging hydrogen ligands ( C , symmetry). In 1968, we were able to show that at 25°C and in the presence of extremely small (catalytic) quantities of bipy , the hexacarbonyls can be reduced by alkali metals to absolutely pure decacarbonyl dimetalates(- I) (241. 2M(CO)a+ 2Nabipy 4 N@M,(CO),,]

Na

+ bipy

+ 2CO + 2bipy

Nabipy

The 12 anions that we first discovered, e.g., [M(CO)J2-, [HM(CO),]-, [M2(CO)10]2-,and [HM2(CO),o]-(M = Cr, Mo, W), have since been the research object of numerous other groups, and further preparative routes have now been developed (25-30). Besides the alkali metals and alkali

Metal Carbonyl Chemistry in Liquid Ammonia

7

amalgams, sodium-potassium alloys, CsK, and [Et4N]BH4 as well as various solvents, e.g., hexamethylphosphoramide (HMPA) and dimethoxyethane, have been used.

B. Structures of the Carbonyl Metalates [M(CO),I2-, [M,(CO) 1012-, and [HM,(CO),,]- (M = Cr, Mo, W)

As a result of the unusually strong reductive character of the pentacarbonyl metalates(-11) [M(CO)5]2- (M = Cr, Mo, W), and the consequent ready oxidizability of the dinuclear anions [Mz(CO)lo12- and [HM,(CO)lo]-, only relatively lately have we been able to publish an interpreted IR spectrum of Naz[Cr(CO),] (31). There is no doubt, however, that the [M(CO)5]2-anions have trigonal bipyramidal structures as the observed intensities and frequency differences of the CO bands are inconsistent with a C4"symmetry of a tetragonal pyramid. Our suggested assignment of the v(C0) valence vibrations is based not only on their positions but more so on the frequency separation of the bands within the isoelectronic series Fe0(CO), (32), [Mn-'(CO),]- (32), and [Cr-11(CO)5]2-.Thus, this series may be compared to the similarly isoelectronic series of tetrahedral complexes Nio(C0)4, [Co-'(CO).J, and [Fe-I1(CO)J2-. Ellis et al. (29) were later able to confirm that Na,[Cr(CO),] prepared by us and by a different route by Kaska (27) is the same complex. This is also, of course, true for this salt which we prepared in 1959 by the reduction of Naz[Cr,(CO)lo]with sodium in liquid NH3 (f 7). The structure of the dinuclear complex anions [Mz(CO)lo]2-(M = Cr, Mo, W) was the subject of much argument for over 10 years. At the end of the 1950s we, in collaboration with Beck et al. (f3),on the basis of the IR spectra, had postulated a nonbridged metal-metal bond, such as was known to occur in the dinuclear metal carbonyls of Group VIIB. A contrasting CO-bridged structure was put forward by Hayter (33) in 1966 for the complexes formed photochemically from the hexacarbonyls and sodium amalgam and described as isomeric with those obtained by us using liquid NH3. The contradictions in the IR spectra were rationalized by newer studies together with Lindner et al. (34), Kaska (27), and Edgell and Paauwe (35).Solvent effects and the nature of the cations both have a marked influence on the frequencies and forms of the CO bands, as less polar solvents (THF) can cause considerable lowering of symmetry due to incomplete dissociation of the dissolved cations and anions (ion pairs). The three research groups just mentioned were able to prove inde-

8

H. BEHRENS

pendently that in more polar solvents, in which complete dissociation occurs, the 3 v(C0) absorptions expected forDId symmetry are observed, in accordance with our findings of 10 years earlier (13). In 1970, Dahl and co-workers (36) carried out X-ray structure analyses on [PPN][CrdCO),,J and [PPN][Mo,(CO),,] (PPN = N(PPh3)3 that conclusively proved that both anions have the same D M structure as Mnz(CO)lo, thus bringing to a successful conclusion an intensive scientific discussion which had lasted for 10 years (see Fig. 1). The same authors (36)carried out X-ray studies on [Et4N][HCrZ(CO),,I for which a linear Cr-H-Cr bond in the sense of 2-electron-3-center bonding was found, with the clearly elongated C r - C r distance leading to Du,symmetry (see Fig. 2). In contrast, the [HW2(CO)lo]-anion exists in linear and bent forms, depending upon the cation. Whereas a] has Du,symmetry with eclipsed equatorial carbonyl [Et4N][HW2(CO), groups, [PPN][HW,(CO),,] has a bent backbone and staggered equatorial ligands (37). With respect to the structures of these H-bridged anions, it is interesting to quote two comments ( 3 8 4 0 ) . Low temperature with single-crystal neutron dithction techniques has shown that [EtSJ][HWACO),J has a nonhydrogen framework which is eclipsed and almost linear (or very slightly bent), but the W-H-W bond is markedly bent and asymmetric with a W-H-W angle of 137.1'. In contrast the structure of [Ph,P][HWdCO),J is nondisordered; the W-H-W linkage is bent and symmetric. (38)The structure is almost identical to that of the isoelectronic HWACO)JVO, which can be prepared from [HWdCO),J- and NaNO,. (39) A combined room temperature X-ray and neutron diffraction study of [PPN][HCrdCO),d has been performed to investigate the effects of crystal packing on the anion and the Cr-H-Cr geometry. A comparison with the earlier work on the [Et,N] salt has shown that the pseudo-D,,, geometry of the anion's nonhydrogen framework is maintained. This result indicates that the [HCr,(CO),J anion is considerably less susceptible than the tungsten analogue to a twisting deformation from a linear-eclipsed to a bent-staggered configuration. (40)

Metal Carbonyl Chemistry in Liquid Ammonia

9

It is fascinating to the author of this historical review that the [HMdCO)lo]- anions prepared by us in liquid NH3 over 25 years ago are still the subject of interesting and exciting structural chemical investigations. Tangential to this theme, I should mention that we have also prepared and characterized by IR spectroscopy nonbridged decacarbonyl dimetalhaving two different Group VIB metals, namely, ates(-I) Na2[CrMo(CO),o] and Na,[CrW(CO),o], from Na,[Cr(CO),] and M(CO)j*THF(M = Mo, W) (31). Anders, a former student of mine, together with Graham (41) was earlier able to obtain decacarbonyl dimetalates with one metal of each of Groups VIB and VIIB, namely, [MnCr(CO)lo]-, [M~MO(CO),~I-, [MnW(CO),oI-, [ReCr(CO)lol-, [R~MO(CO)~OI-, and [ReW(CO)l~l-,by reaction of Na[M(C0)5] (M = Mn, Re) with M(C0)8 (M = Cr, Mo, W). These mixed metalates also have nonbridged structures and may be classified as intermediate between Mn2(CO)loand Re2(CO)loon the one hand and the isoelectronic anions [Mz(CO)lo]2-(M = Cr, Mo, W) and [MM'(CO)lo]2- (M = Cr; M' = Mo, W), prepared by us, on the other. The protonation of [Et4N][ReCr(CO)lo]leads to the neutral complex HReCr(CO),o, which has been shown by the X-ray studies of Graham et al. to have the same structure as [HCr,(CO)lol- (42). The reduction of metal carbonyls or their derivatives with sodium in liquid NH3, begun by us 35 years ago, has become a method often used successfully by numerous other research groups throughout the world. A few examples are the preparation of Na2[$-CsH5V(CO)3] from q5C,H,V(CO), (43) and [V-r11(CO)5]3and [M-1'1(CO)3]3-(M = Co, Rh, Ir). According to Ellis et al. (44), these carbonyl metalates are obtained in the extremely low oxidation state of -111 from the anions [V-'(CO)J and [CO-~(CO)~]and also from the neutral complexes M3(C0)12(M = Rh, Ir). During the 1950s, at the Institute of Inorganic Chemistry at the Tech-

H. BEHRENS

10

nische Hochschule in Munich, there took place an extremely interesting and stimulating scientific competition between the research groups of my most respected teacher Walter Hieber and myself. Whereas Hieber concentrated his efforts, which were extremely successful, on the polynuclear carbonylferrates (13, 45) having 2 , 3 , and 4 iron atoms, we concerned ourselves with the various carbonyl metalates of Group VIB. 111

REACTIONS OF THE PENTACARBONYL METALATES(-II) AND THE DECACARBONYL DIMETALATES(-I) OF CHROMIUM, MOLYBDENUM, AND TUNGSTEN A. Monosubstituted Derivatives of the Hexacarbonyls

Their strong reducing nature is a special characteristic of the pentacarbonyl metalates( -11) and decacarbonyl dimetalates( -I) of chromium, molybdenum, and tungsten that distinguishes them from the other carbony1 metalate anions of the 3d metals, e.g., [Co(CO)J-. The oxidation of the mononuclear species with water 2[M-"(CO)d2-

+ 3HeO -, [HMS-'(CO)lJ- + 3 0 H - + H2

has already been mentioned. In the presence of a monodentate ligand (L), however, Na2[M(CO),], Na2[MZ(CO),,],or Na[HMZ(CO),,] form with HzO monosubstituted hexacarbonyls with evolution of H2 (17, 19-21, 46,

47).

[M-ll(CO)Jp- + 2H20 + L 4 MO(CO)6L+ 2 0 H [M,-'(CO),dZ- + 2He0

[HM$CO),J-

+ %I

4

+ HI

+ 20H- + He + OH- + Ht

2M0(CO)&

+ H,O + 2L + 2M(C0)&

These reactions led to the first preparation of pentacarbonyl complexes of Group VIB, namely, M(CO),NH,, M(CO),py, and M(CO),(NH,Ph) (M = Cr, Mo, W). At higher temperatures, the formation of tetracarbonyl derivatives is also possible (e.g., with py) (48). The analogous reactions using such bidentate ligands as en or phen lead (also via evolution of H2 and CO) to quantitative formation of the tetracarbonyl derivatives M(CO),en and M(C0)phen (M = Cr, Mo, W) (48). Whereas Na,[M(CO),] and Na2[MdCO),,] react with o-phenylenediamine and water to give mononuclear (H2NC6H4NH2)Cr(CO),,with mand p-phen ylenediamine, the dinuclear complexes [(OC),M]H2NC6H4NH2[M(CO)5] are formed, in which two M(CO),

Metal Carbonyl Chemistry in Liquid Ammonia

11

groups are linked through aromatic diamine bridges (47). Thus, o-phenylenediamine acts exclusively as a monodentate ligand. In contrast, the strongly basic I ,2-cyclohexadienediamineforms the chelate complex (47)

In this series of reactions we were also able to isolate a trinuclear complex for the first time, namely

(OC),Cr-N

from Na,=[Cr(CO)J, 1,3,5-triaminobenzene, and water (47). Considerable attention has been given to the reactions of the monoand dinuclear pentacarbonylchromates, -molybdates, and -tungstates with various N ligands, as they were of principal significance to the chemistry of the substituted hexacahonyls of Group VIB metals. Until then, the mononuclear pentacarbonyl derivatives of the type Mo(CO)5L, and the di- or trinuclear, ligand-bridged cahonyl complexes of Group VIB were unknown. The complex Cr(CO)4(PPh3)zwhich we synthesized at that time from Cr(CO),NH, and PPh3 was the first phosphine-substituted derivative of a Group VIB hexacarbonyl(19). This is understandable when it is considered that the direct substitution of CO in the relatively stable hexacahonyls by the stronger donor ligands such as py is not possible at low temperatures and that thermal treatment between 160 and 200°C normally causes the mono compound step to be overshot. It is most impressive how the situation has changed dramatically as a result of the pioneering photochemical studies of Strohmeier et al. (49) and how, since 1960, a really hectic development in the area of substituted hexacarbonyls of Group VIB has occurred. In this respect, the derivatives M(CO),THF have proved to be especially reactive and to constitute a preferred starting material for numerous other pentacahonyl complexes of chromium, molybdenum, and tungsten. Studies by Guttenberger (50) and Bertrand el al. (5f)have shown that the ligand-bridged bis(pentacarbony1 metal) complexes, e.g., (OC),M(NC-CN)M(CO), or (OC),M-P(OCHz),P-M(CO), (M = Cr,

H. BEHRENS

12

Mo,W),may also be easily obtained photochemically from the hexacarbonyls and (CN), or P(OCH&P in suitable solvents. This type of compound was later shown by other authors (52, 53) to be accessible by heating M(C0)6 with suitable ligands in high-boiling solvents (190-210°C). Examples are (OC)5M-PMe,-PMe2-M(CO)5 and (OC),M-AsMez-AsMe2-M(CO)5 (52, 53). The synthesis of the series CNW5NHa,[(OC)5Cr12NzH4, [(OC)SCr12N2H2. and [(OC)5CrlzNzby Sellmann et all (54) in the course of their work on N2 fixation is especially impressive. B. Mono- and Dinuclear Hexacoordinated Pentacarbonyl Metalates(0)

Further to the foregoing reactions, my group and others were able to synthesize numerous new anionic and neutral complexes by various reactions of [M(CO)$- and [M2(CO)lo]a-.Thus, we were able to prepare the cyanocarbonylchromates(0)[Cr(CO),CN]- and cis-[Cr(CO),( CN),I2by oxidation of Na,[Cr(CO),] with aqueous NaCN solution or with (CN), or ICN in THF (55, 56):

+ CN- + 2H,O [Cr-rr(CO)#- + 2CN- + 2H20 [Cr-r'(CO)#- + (CN), [Cr-rr(CO)5]P-+ ICN [Cr+'(CO)#-

[CrO(CO),CN]- + 2OH-

-

+ HP

cis-[CrO(CO)&CN)d2-+ 2OH-

THF

[CrO(CO),CN]-

THF

[CrO(CO),CN]-

+ CO + HZ

+ CN+ I-

For the analogous reactions of Na[M(CO)J ( M = Mn, Re) with ICN, we observed a different reaction, namely, that of oxidative elimination, leading to the cis configurated [M(CO),(CN)I]- anions (152): [M+(CO)d- + ICN

THF

cis-[M(CO),(CN)I]-

+ CO

(M = Mn,Re)

With the preparation of Na[Cr(CO),CN], we not only obtained the first anionic hexacoordinated, pentacarbonyl metalate(0) of Group VIB, but also initiated the extremely extensive chemistry of the [M(CO),X]- anions. Of the many typical redox reactions that we used for the formation of [M(CO),X]- anions, the following are given as examples (57-60): (1961)

(P

~C~P'(CO)IOI*+ 1%~1+2[CrO(C0)dl-

-

Metal Carbonyl Chemistry in Liquid Ammonia (1964)

[M-ll(CO),]*- + Fe+*1(CO)4XZ

THF

[MO(CO)dc]- + [FeOcCO),X]-

(M = Cr, Mo, W; X

+ 2NCS-

13

=

C1, Br, I)

+ 2Fe"

(1966)

[Crg-"(CO)lo]*- + 2Fest

(1969)

[HCr2(CO)lo]- + 2HSC,,H4CHrp --* 2[Cr(CO)JSCeH4CHrpJ- + Ht

--*

2[C?(CO)JNCS]-

+ H2

The anions [Cr(CO),I]- and [M(CO),NCS]- (M = Cr, Mo, W) were prepared by Fischer and Ofele (61) and Wojcicki and Farona (62). respectively, shortly before our own studies. Considerably later, other research groups described further redox reactions of a similar nature, leading to the anions [M(CO),C,F,]- (63), [M(CO),(CsPh)]- (64, [M(CO),SR]- (65), or [M(CO),R]- (R = Me, Et, NCCHZ, PhNH2) (66). For the sake of completeness, reference should also be made to the work of W. Beck, J. K. Ruff, E. W. Abel, and R. B. King, by whose research groups, between 1963 and 1975, a large number of other [M(CO)&]- complexes were synthesized, especially with halide and pseudo-halide anions, by thermal or photochemical reactions of M(CO), or M(CO),THF. Again starting from the decacarbonyl dimetalate( -I) anions, besides the mononuclear [M(CO)&]- anions, we were the first to prepare the dinuclear [M,(CO),&]- species (M = Cr, Mo, W) with the respective metal having the oxidation state 0. These anions can be formally derived from the unknown dinuclear carbonyls M,(CO),, by substituting a CO ligand by X- (X = CN, I, NCS, SH, SMe, SEt, SPh). We obtained the first complex of this type, namely, the [Cr2(CO)loCN]anion, as early as 1959, as a by-product of the oxidation of [Cr(C0),l2with (CN), (55, 67): 2[Cr-11(CO)Jx-+ 2(CN)* + [Crso(CO),CN]-

+ 3CN-

The monoanion [CrO(CO),CN]- also gives this dinuclear anion, however, in acid pH regions (55): 2[Cr(CO),CN]-

+ H+

-P

[Cr&O),,CN]-

+ HCN

The dinuclear species [Cr,(CO),J]- (671, [Cr,(CO),dCS]- (671, and [Cr2(CO)loSH]- (60), however, were prepared by reduction of the paramagnetic neutral complexes CrACO),J (57, 58, 68) and Cr,(CO),,,NCS (59), which will be discussed later, as well as from the mononuclear and also paramagnetic Cr(C0)SH (60) with sodium amalgam in THF: 2Cr(CO).SH

+ 2e + [CrkCO)loSH]- + SH-

The anions [Cr&CO),,SH]- and [Cr&CO),$3R]- (R = Me, Et, Ph) are

H. BEHRENS

14

also formed by oxidation and H2elimination from [HCrz(CO),o]- with water in the presence of SH-or mercaptans (RSH) (60). The IR spectra of these anions show that in all cases the complexes are X-bridged with bent or linear Cr-X-Cr bonds (X = CN, I, NCS, SH, SMe, SEt, SPh) (60, 6 9 ) . Dahl and co-workers (70)found by Xray structure analysis that the Cr-I-Cr angle in [PPN][Cr,(CO),J] is - I 18” (see Fig. 3). It was somewhat later that Ruff et al. (71) turned their attention to the synthesis and structure determination of these dinuclear anions, which they prepared photochemically from M( CO)Band [ M(CO),X]- according ’ to the following process : M(CO), + [M(COhXI-””- [M,(CO),&]-

+ CO

(M = Cr, Mo, W; X = I. CN, NCS, SCH,)

or from M(COla and SCH3-: 2M(CO),

+ SCH,-

*

[M2(CO),,,SCH31- + 2CO

Ruff (71a) was also able to obtain halogen-bridged anionic complexes with different Group VIB metals, namely, [CrW(CO)loBr]- and [CrM0(C0)~dl-.A little later, we used the same method to gain access to dinuclear NCS-bridged anions with heterometals and bent M-NCS -M‘ bonds (72): [M(COXNCSl- + M’(CO), -% [(OC),M-NCS-M’(COhI-

+ CO

(M = Cr, W; M‘ = Mo, W)

C. Heterometallic and Inserted Anionic Complexes

Ruff (73) and our group in Erlangen (31) were successful in preparing dinuclear anionic complexes with metals of Groups VIB, VIIB, and

I C r 2 fCOI,oI1‘

FIG.3

Metal Carbonyl Chemistry in Liquid Ammonia

15

VIIIB, which are characterized by metal-metal bonding (73 and 31, respectively): [MP(C0),J2- + COP(CO)~ -D 2[(OC)@4o(CO)J [M(CO)J2- + FeACO),

+

[(OC),M-Fe(CO),12-

+ Fe(CO)$

(M = Cr, Mo, W)

Our suggested structure for the [MFe(CO),lZ- anions, based on IR data, is given in Fig. 4. With this work, we were able to complete the isoelectronic series based on chromium: [Crz(CO)lo]z-, [CrMn(CO)lo]-, [CrFe(CO),lZ-, and [crco(co)B]-. However, the series cannot be extended to include nickel. Of particular interest are the insertion reactions carried out by Ruff (74) in which SnI,, GeI,, or SOz is inserted into the metal-metal bonds of [M2(C0)10]2- anions to give bridged species of the type [(OC),M(EIz)M(CO)5]2- (E = Ge, Sn) or [(OC),M(SOz)M(CO)5]2-. The insertion of Hg was, however, only successful into [Crz(CO)lo]2-,a result that our own experiments have confirmed. The product is the [(OC)5Cr-Hg-Cr(CO)s]2anion, which is isoelectronic with (OC),Mn-Hg-Mn(CO), and has linear Cr-Hg-Cr bonds (75). In the case of Kz[Fez(CO),], the insertion of all the Group IIB metals into the Fe-Fe bond of the nonbridged [FeZ(CO),]'- anion is also possible (75). The anionic complexes [(OC),Fe-M-Fe(C0) Jz- ( M = Zn, Cd, Hg) so formed are isoelectronic with the neutral complexes (OC),Co-M-Co(CO),, described by Hieber and Teller (76) over 35 years ago. A contrasting reaction is that of Ni(CO)* with [PPN],[W2(CO)lo], which gives the anionic cluster compound [PPN],[WZNi3(CO),,], whose X-ray structure has been determined by Ruff and co-workers (77): [PPN]dWdCO),J + 3Ni(CO), [PPN = (Ph3)ZNl

d

*

[PPN]dW,NidCO),J

b

0

FIG.4

+ 6CO

H. BEHRENS

16

D. Paramagnetic Mono- and Dinuclear Pentacarbonyl Complexes of Chromium

The mono- and dinuclear cahonyl metalates of Cr, Mo, and W, first prepared by us using liquid NHS, have also been of great importance for the synthesis of numerous neutral complexes not accessible by other routes. Special attention is drawn to the deeply colored and very unstable mono- and dinuclear paramagnetic complexes Cr(C0)J. Cr&CO),J, Cr(CO),CN, Cr(CO),NCS, Cr&CO),&'CS, and Cr(CO),SH. We had already seen that oxidation of [Crz-l(CO)lo]z- with Is leads to yellow [Cro(CO)J]- (57). With excess iodine, however, the oxidation goes further, to the deep-blue, hexacoordinated, thermally very unstable Cr+I(CO)J. by which reaction the dinuclear, deep red, paramagnetic, neutral complex of CrZ(CO),J may be isolated as the intermediate. The actual compound formed is solely dependent upon the mole ratio of [Crz(CO)lo]z-:I2 (57, 58, 68): Mole ratio 1 : 1 Mole ratio I : 1.5 Mole ratio I :2

+ I* --* 2 [Cr(CO)Jl[Cr&0),,J2- + 1.512 + Crl(CO),oI + 21[Crl(CO),,j2- + 212+ 2Cr(CO)& + 21-

[CrdCO),#-

Deep-green Cr(CO),CN and deep-red Cr(CO),NCS, which are stable up to -25°C and, like Cr(CO)J, show the paramagnetism of a single unpaired electron, are formed from the reaction of Cr(C0)J with ICN or ISCN (78): Cr(CO)& + ICN

Cr(C0)J

+ ISCN

+ 1, Cr(CO)&4CS+ 1%

+ Cr(C0)fN -P

The dinuclear NCS-bridged neutral complex Cr&CO),,NCS (59) is obtained by oxidation of [Cr2(CO)lo]2-with excess Fe3+ cations in the presence of NCS-:

1

[Cr2(COho]*-+ 2Fea++ 2SCN- + 2[Cr(COXNCS]-

+ 2Fe*+

+Ft+

Cr2(CO),oNCS+ Fe*++ NCS-

All the mono- and dinuclear anionic and neutral complexes of the types [Cr(CO)&I-, [Crz(CO)ld
Metal Carbonyl Chemistry in Liquid Ammonia

17

vents SOz is inserted into the Cr-Cr bond of [CrZ(CO)J- (74), in H,O oxidation of Cr-' to Cr+' occurs with simultaneous reduction of SO, to SH- anion. The latter functions as a complex ligand and thus allows formation of Cr(CO),SH with its bent Cr-S-H grouping. This reductive behavior of Na,[CrZ(CO),,] may be compared with its reaction with 2bromo-2-nitrosopropane Me,C(NO)Br which King and Douglas (79) have shown to lead to (dimethylketimino) pentacahonylchromium Me,C=N(H)Cr(CO),. The mono- and dinuclear paramagnetic complexes are of course extremely reactive. We have shown that the reactions of Cr(C0)J and Cr,(CO),J with various mono- and bidentate ligands lead to the formation of the respective substitution products of CI-(CO)~ with simultaneous oxidation to elemental Iz (80): Cr+'(CO).J

+L

CP(CO)&

+ 41,

(L = MeNH,, EtNHz, PhNHI, py. PPh3 Cr+'(CO)J

+ OCN- + [Cr"(O),NCO]- + 41,

The mechanism of the reaction of Cr(C0)J with NO in EtzO at low temperature is not yet known. All five CO ligands are eliminated, to give a red solution from which we have precipitated violet Cr(N0)2(PPh3)z12 on addition of PPh3. The NO groups are cis-oriented (81). E. Complexes with M-M' In)

Bonds (M = Cr, Mo, W; M' = Gel Sn,

We have also reacted Naz[Mz(CO)lo]with the tetrahalides of germanium and tin (EX3 in THF Monomeric complexes of the type (OC),MEX,.THF are formed, according to (82): NaplMs(CO),J + EX,

THF

(OC)&IEX,*THF + Na[M(CO)&]

+ NaX

(M = Cr, Mo. W; E = Ge, Sn: X = Cl, Br, I)

The Group VIB metals are pseudo-octahedrally coordinated, whereas germanium and tin have a pseudo-tetrahedral environment (82). We obtained the compounds (OC)5MGeCl,.THF and (OC),MSnX,-THF nium and tin (EX3 in THF. Monomeric complexes of the type reaction of M(CO)Bwith CsGeCl, or SnX, (83). With [Et,N]X, these compounds give the anions [M(CO),GeC13]- and [M(C0)5SnX3]- (83). We were also able to obtain such anions by the insertion of SnCl, into the M - C l bonds of [Mo(CO),Cl]- anions (83).These anions with GeCl, and SnCl, ligands had been prepared earlier by Ruff (84) by the photochemical

H. BEHRENS

18

reaction of M(CO), with [Ph.&][GeCl,] or [Ph&][SnCl,]. In 1971, Marks (85) obtained the methyl and butyl derivatives (OC)&-ER2.THF (E = Ge, Sn) from Na2[Cr,(CO),01and RzEC12. New studies by our group have shown that the indium trihalides InX, (X = C1, Br, I) also react with Naz[M2(CO)10] according to (86): Na,[MI(CO),d

+ InX,

(M = Cr, Mo, W; X

THF

(0C)JlMInX.THF

+ Na[M(CO)&]

t NaX

= CI, Br, I).

The structure of Cr(CO)JnBr.THF established by X-ray diffraction shows that in this molecule the Cr, In, and Br atoms and the THF ring form an almost planar structural unit in which the In atom is tricoordinated (86). In the crystals, two additional coordinative bonds to two bromine atoms of neighboring molecules are present, so that the indium is pentacoordinated (see Fig. 5 ) . The molecules are stacked in the c axis, whereby the InBr groups so lie that a double chain having alternating In and Br atoms is formed. In this context, it is relevant that nonbridged K2[Fe2(CO)J with Ph3SnC1 or PhzSnClzforms neutral complexes in which, in contrast to the corresponding reaction of Na,[M2(CO),o] ( M = Cr, Mo, W),the Fe-Fe bond remains intact (87): K.JFedCO)d + 2Ph$nCI+

+

K2[Fel(CO)PJ 2PhSnC1,

Ph3Sn(OC)Je-Fe(CO),SnPh,

+ 2KCl

+ 2KCI

+ CIPhSn(OC)+Fe-Fe(CO)$nPh~l

Both complexes unequivocally occur as trans-trans isomers (87): \

-Sn

/

\ / \ /

/

/ \ / \

\

-Fe-Fe-Sn-

To complete this section, I should like to mention the following remarkable reactions of Na,[M-"(CO),] and Na2[M;'(CO),o]: ( 1) the formation of cyclobutadiene-metal cahonyl complexes (88) Nh[Mo(CO),]

+

-.

L'I

c1

-

IMo(CO).,

+

2NaC1

+

CO

(2) the formation of thiocarbonyl complexes (89) Na,[M&CO),,]

+ CI,CS + M(CO)&S + 2NaCI + {M(CO)$}

(3) the formation of isonitrile complexes (90) N+[Cr(COM

+ C&CNR+ Cr(C0XCNR + 2NaCI

(R = C&i,, C a b ; S O z X J i d

(4) the formation of Cr(CO)4P213 from Na,[Cr(CO),] and Pz14(91).

Metal Carbonyl Chemistry in Liquid Ammonia

19

FIG.5

IV REACTIONS OF METAL CARBONYLS AND METAL CARBONYL DERIVATIVES WITH LIQUID AMMONIA A. Substitution of CO by NH3and Addition of NH3

In numerous experiments we have found that liquid NH3 can replace CO or other ligands in metal cahonyls or their derivatives and also that NH, can become an additional ligand bonded to a transition metal. Thus, the hexacarbonyls of Group VIB can form pentacahonylammine or tricahonylammine derivatives, depending upon the temperature conditions (92): Cr(CO)B+ NH8

5

M(C0)e + 3NHj

lam

Cr(CO)$lH,

+ CO

M(CO)dNH$,

+ 3CO ( M

=

Mo. W)

In liquid NH3, one CO group can also be replaced by NH3 in the case of the tetracahonyl complexes M(CO),bipy and M(CO),phen. "Mixed" species M(CO)&ipy(NH,) or M(CO),phen(NH,) (M = Cr, Mo, W) are formed with the remaining CO ligands being in cis positions (92). Mixed complexes of the type M(CO),(rN)L (N-N = bipy, phen; L = monodentate N, P, or S ligand) have been known since 1935 from the work of Hieber et al. (93); and a large number, described by other authors, have been known for a long time. Our own work in liquid NH, allowed the first preparation of the respective NH3 derivatives. Although Ni(CO), does not enter into reaction with liquid NH, up to -33"C, between -25 and llO"C, CO is substituted by NH, to give an

H. BEHRENS

20

inseparable mixture of pale-yellow Ni(CO)3NH3 and Ni(CO)ANHd, which are only stable as solids up to -60°C (94). Fischer and Jira (95) treated [Ni(NH,)&SCN), with C&,K in liquid NH3 to obtain [Ni(NHdJ(C5H&. We obtained the same ionic complex directly from Ni(+C&,), in liquid NHS (96). We were also able to show similar behavior for Re(CO)& (X = CI, Br, I) which, on treatment with liquid NH3 at 60 or 120°C for at least 40 h, gives the ionic complexes cis[Re(CO)4(NH,)2]Xor [Re(C0)9(NH3)JX (97). We similarly obtained an ionic pentacarbonylammine complex from Re(CO),CI and KSCN in liquid NH3 (97): Re(CO).&I

+ NH8 + KSCN

awc

[Re(CO)$IHJSCN

+ KCI

B. “Base Reactions” and Disproportionations

Completely different behavior toward liquid NHs is shown by the three iron carbonyls Fe(CO)5, Fe2(CO)e, and Fes(CO)12(98, 99) and the two and CO,(CO),~(100). Between -21 and O”C, cobalt carbonyls COACO)~ Fe(CO)band liquid NHS give a homogeneous, pale-yellow solution from which Fe(CO)&may be recovered on evaporating off the NHS. The solution contains the carbamoyl complex NH,[(OC)4Fe40NH2] which cannot be isolated and which is formed by nucleophilic attack of an NH3 molecule on a CO ligand, followed by proton release (101). At 20°C after 14 days of reaction, (NH,),[Fe(CO)J and CO(NHd2 are obtained (99): Fe(CO),

+ 4NH3

POT

(NHJdFe(C0) J

+ CO(NHI),

In contrast to the reactions of M(CO)# (M = Cr, Mo, W), a redox system operates here which is comparable, also in terms of the reaction mechanism, with the so-called “base reaction” of Fe(CO), (13). The reaction of FB,(CO)~in liquid NH3 is similar: FedCO)p + 4NHs

POT

(NHJdFe,(CO)J

+ CO(NH3,

This corresponds to the base reaction of Fe2(CO), with alcoholic alkali according to the following equation: Fe,(CO)@+ 40H-

-B

[FedCO)J*- + C 0 2 -

+ 2Hs0

This also leads to the formation of the octacarbonyldiferrate anion [Fe2(CO)J2-having a nonbridged Fe-Fe bond (13). In contrast, the reaction of Fe3(CO)lain liquid NH3 is comparatively complicated, especially as the time allowed for reaction is of decided

Metal Carbonyl Chemistry in Liquid Ammonia

21

importance (99). The primary reaction is disproportionation: 4FeJ(CO),, + 18NHs + 3[Fe(NH3J[FeJ(CO)lll

+ 15CO

After about 14 days, the trinuclear carbonylferrate undergoes further reaction with the free CO to form the dinuclear carbonylferrate and Fe(CO),. 3[Fe(NH3)J[Pe3(CO),,]

+ 6CO + 3[Fe(NH3)J[Fea(CO)J + 3Fe(CO),

Since, as already described, Fe(CO)Sreacts to give (NHJ,[Fe(CO) J and CO(NH&, the complete reaction equation for Fe3(C0)12and liquid NH3 after 14 days of reaction can be described by the following: 4Fe&CO)u

+ 30NHS

1wc

3[Fe(NH$J[FedCO)d

+ 3(NH&,[Fe(CO)J + 3CO(NH& + 9CO

The polynuclear carbonylferrates and their structure determinations by L. F. Dahl have been dealt with in some detail in my lecture The Chemistry of Metal Carbonyls: The Life Work of Walter Hieber given at the Symposium on Metal Carbonyls in 1974 in Ettal (102). The reactions of C O ~ ( C Oand ) ~ Co4(CO)12in liquid NH3 have also been the subject of our investigations, and in both cases at 20°C disproportionation of the cobalt occurs with CO elimination. The CO then undergoes a secondary reaction with still unreacted carbonyl to give sublimable NH4[Co(C0),I and CO(NH2)2(100): (48%)

+ 12NH3 + ~[CO+'~(NH~)B][CO-'(CO)& + 8CO CoZ0(CO)~+ CO + 4NH3 + 2NHq[C0-'(C0)41 + CO(NHph

(94%)

3c04°(co)lp

(9%) 3C0p'(C0),

(6%)

+ 24NHs ~[CO+"(NH~)BI[CO-'(CO)~]~ + 4CO dc~~"(CO)iz + 3CO + 4NH3 2NHJCo-'(COh] + CO(NHz)z --*

+

Paramagnetic V(CO)ubehaves analogously in liquid NH3, i.e., disproportionation of vanadium occurs, to give [V+"(NH3)J[V-'(CO)& with formation of NH,[V(CO) J and urea (103). We have also observed the disproportionation of vanadium in the r e a d o n of V(CO), with the ammono acid NH4Cl (10.3): 3V(TCO),

+ 2NH4CI + 6NH3

LIP. NHI

[V+"(NHs)&Iz

+ ZNHq[V-'(CO)B] + 6CO

The reaction of Mng(CO)lowith liquid NH3 at 20°C differs from those

of the two cobalt carbonyls in that the disproportionation involves the formation of the cationfa~-[Mn(Co)~(NHd~l+ (104): Mn,o(CO),o

+ 3NH3

+

[Mn+1(CO)3(NH~3][Mn-'(CO)~ + 2CO

--

H. BEHRENS

22

Here, as in isoelectronic Cr(CO),(NH,),, the three CO and the three NH, ligands are infuc positions. Some of the free CO reduces unchanged Mnz(CO)lo, Mnz(CO),,

+ CO + 2NHS + 2HMn(COh + CO(NH&

and some of it carbonylates NH,: CO

+ 2NH3 + CO(NH32 + HZ

The reaction of Rez(CO)lo with NHICl in liquid NH, at 120°C occurs without side reactions, giving quantitative oxidation to f a ~ - [ R e ( c O ) ~ (NH3),]CI with elimination of Hz(97): RedCO),,

+ 2NH,CI + 4NH3

lPW

Z[Re(CO)dNH$JCI

+ 4CO + H2

There is no doubt that most of the mono- and polynuclear metal carbonyls react with liquid NH, via nonisolable carbonyl carbamoyl intermediate complexes in which the acid amide group - C O N H z is directly bound to the transition metal. For the reaction of Mnz(CO)lowith liquid NH,, we have been able to show that, at -78"C, quantitative formation of C ~ S - M ~ ( C O ) ~ ( N H , ) - C Oand N HHMn(CO)5 ~ occurs (105) according to the following reaction: -la t HMn(CO), Mn,(CO),, + 2NH3 Mn(CO),(NH,)C,

-

NH2

At room temperature, the secondary reactions already described play a part and cause the parallel formation of [Mn(CO)3(NH3)3][Mn(CO)5], HMn(CO),, CO(NH,),, Hz,and CO (104).

C. Carbamoyl Complexes of Transition Metals with the Ligand CONH,

Two observations that we made in 1966 have been important in molding the history of the development of this interesting class of compounds. The first involves the reaction of [Mn(CO)8]Cl*HClwith liquid NH, at 20°C (104): [Mn(CO)J+ + 3NH8 + HMn(CO),

+ CO(NH,)* + NH,+

For this reaction, we proposed a mechanism involving a nonisolable intermediate of pentacarbonylcarbamoylmanganese(+I) Mn(CO),CONH2 (104). The second involves the reactions of CO with solutions of KNHz in liquid NH, (106). In the presence of a large excess of CO, formamide

Metal Carbonyl Chemistry in Liquid Ammonia

23

and potassium formamide are formed together, viz: KNH,

+

nCO

+

20-60"

(n-l)NH,

~

HCzo

NHK

+

(n-l)HC, Go

NH,

In contrast an excess of KNH,, depending on the temperature, causes formation of hydrogen together with potassium cyanate or dipotassium urea:

+ CO 2KNH2 + CO KNH,

IO-BWC 12vc

KOCN

+ H2

CO(NHK),

+ H2

If this reaction of KNH, and CO in liquid NH3 is extended to coordinatively bonded CO in cationic carbonyl complexes, instead of the formation of formamide, carbonylcarbamoyl compounds of the type M(CO),-CONH, result, and instead of H2 and KOCN (or CO(NHK),), hydridocarbonyls and NHIOCN [or CO(NH,),] are formed. The first carbonylcarbamoyl complex that we described, cisRe(C0)4(NH3)CONH2,was obtained by the reaction of Re(CO),CI with liquid NH3 at 60°C, after a reaction time of only 1 hour, and without CO evolution (107): Re(CO),Cl

+

3NH,

1 h, 60"

liq. NH,

The salt [Re(CO),NH,]CI certainly occurs as an intermediate in this reaction. As the ammono acid NH4Cl is formed, depending on the temperature, a prolonged reaction time causes cis-Re(CO),(NH,)CONH, to give the ionic complexes [Re(CO)4(NH3)2]CIor [Re(C0)3(NH3)3]CIwith CO elimination: 40 h, 60"

NH,

liq. NH,

c

[Re(CO),(NH3),]C1

+ CO

+

NH,

As early as 1965, Angelici (108) was concerned with the reaction of Mn(CO),CI with primary and secondary aliphatic amines. He initially considered the product with MeNH,, namely, Mn(CO),(NH,Me)(NHMe), to be a complex of heptacoordinated manganese. A subsequent X-ray structure analysis showed, however, that it is in fact a hexacoordinated carbamoyl complex, Mn(CO)4(NH2Me)CONHMe(109). In continuation of these studies, the reaction of cationic CO complexes of various tran-

24

H. BEHRENS

sition metals with primary and secondary amines was investigated by Angelici (110); and, in parallel, independent work, our own Erlanger group looked systematically at the reactions in liquid NH3 between -78 and 120°C. The stabilities and reactivities of the resulting carbamoyl complexes with 4 O N H R and - C O N R z groups are remarkably different from those with the unsubstituted - C O N H 2 ligand. Following our experiments with Re(CO),CI, we investigated the behavior of Mn(CO)SCl, Mn(CO),(PPh3)CI, and [Mn(CO)4(PPh3)z]CIwith liquid NH,; and, after very short reaction times (max. 10 min), we were able to isolate the three carbamoyl complexes ~ ~ S - M ~ ( C O ) ~ ( N H , ) C O N H ~ , ~u~-M~(CO)~(PP~~)(NH,)CONH~, and ~ U ~ - M ~ ( C O ) ~ ( P P ~ , ) ~ The CONH~. stabilities of these complexes increase markedly in this order (111). All of them react at 60-120°C with NH,CI in liquid NH, via CO, or CO and PPh3, substitution to give ionic [Mn(CO),(NH3),]Cl. The ionic species [Mn(CO),N H3]Cl, [Mn(CO),( PPh3)NH3]CI , and [Mn(CO),(PPh,),]CI are formed from these carbamoyl complex series by reaction with HCI in CBHB,according to the following type of reaction. Mn(CO),(NH,)C:

10

NH*

+

2HC1

-

[Mn(CO),NH,]Cl

+

NH,Cl

As a general rule, the carbonylcarbamoyl complexes of rhenium are considerably more stable than those of manganese. In an attempt to isolate the carbamoyl compound Mn(CO),CONH2, which had earlier been postulated by u s as an intermediate in the reaction of [Mn(CO)J+ with liquid NH3, we carried out the reaction at -78°C. However, even under these conditions, the only products were HMn(CO), and NH,NCO (112). Thus, it may be assumed that, in analogy with the Hofmann acid amide degradation, the - C O N H 2 group of the carbamoyl complex is oxidized to NCO-. Although, between 20 and 60"C, [Re(CO),]+ behaves similarly to the manganese compound, giving the hydridocarbonyl complex HRe(CO),, at -78"C, we were able to isolate and characterize, the extremely unstable Re(C0)5CONHz(112): [Re(CO),]+ + 2NH,

-78"

liq. NH,

+

Re(CO),C:o

NH:

m 2

The behavior of cationic penta- and tetracarbonyl complexes of manganese and rhenium toward liquid NH, at various temperatures is summarized in Table 1 (112).

Metal Carbonyl Chemistry in Liquid Ammonia

25

TABLE I Cationic starting compounds

Temperature

[Mn(CO)5PPhsl+ [Mn(CO)d’Phd+ [Mn(COVEt31+ [Re(COWPh31+ [Re(COPW+ tr~ns-[Mn(CO)~(PEt~)l+ trans-[Re(CO)4(PPh3)J+ ~is-[Re(C0)~(PEt~)]+ ~is-[Mn(CO)~bipy]+ ci~-[Re(CO)~bipy]+ cis-[Mn(CO)lphen]+

-70 30 -70 -33

(“c1

Reaction products C~~-M~(CO)~(PP~~)CONH~ HMn(CO)J’Ph3 + NH,”CO HMn(CO)J‘Et3 + N H a C O C~~-R~(CO)~(PP~~)CONH~ C~~-R~(CO)~(PE~~)CONH~

-78 -33 60 60 -33

rner-cis-Mn(C0)3(PEt3kCONHz

mer-Re(CO)s(PPhskCONHz fac-Re(CO)3(PEt3)&ONHz fac-Mn(CO)3(bipy)CONHZ fac-Re(CO)&ipy)CONHZ fac-Mn(CO)&hen)CONH2

60 -33

From these results, it may be concluded that the stabilities increase markedly on going from the penta-, through the tetra- to the tricarbonylcarbamoyl derivatives. As the examples of Re(CO),CONH, and Re(CO),CONHMe ( I 13) show, the complexes with --CONHR and - C O N R 2 ligands are considerably more stable than the comparable - C O N H 2 compounds. With gaseous HCl, Re(CO),CONH2 and the carbonylcarbamoyl complexes listed in Table I revert to the cationic starting compounds. In the case of Re(CO)4(NH3)CONH2in liquid NH3, one can observe the elimination of H20 from the --CONH2 group, Re(CO)XNH$CONH,

+ NH3

Re(CO),(NHJZCN + H@

+ CO

leading to cyanodiamminetricarbonylrhenium(+I) ( I 12). With CH30H, the carbamoyl complexes give the respective carboalkoxo derivatives ( I 14): M(CO)3(PPh3)rCONHZ + MeOH ( M = Mn, Re)

NaOMe

M(CO)3(PPh3)2COOMe+ NHs

When these results are considered, it is understandable that the pentacoordinated [Co(CO)$Ph3]+ cation with NHMez forms the very stable N,N’-dimethylcarbamoyltriphenylphosphinetricarbonylcobalt( +I) species C O ( C O ) ~ ( P P ~ ~ ) C O Nbut, M ~with , , liquid NH3 at -50°C the products (I15). are H C O ( C O ) ~ ( P Pand ~ ~ NH4NC0 ) The initial step in all of these reactions is nucleophilic attack of an NH3 molecule on a CO ligand of the cation, with formation of a carbamoyl

H. BEHRENS

26

-

complex and simultaneous release of a proton: 0

[o=c=co(co),(PPh$]

+

0

[O=C-

Co(CO),(PPh&] +

In analogy with the Hofmann acid amide degradation, the - C O N H 2 ligand can undergo further oxidation to the NCO group. Here, it is assumed that proton elimination takes place,

with reduction of

occurring via intramolecular transfer of a H- ion to give H C O ( C O ) ~ ( P P ~ ~ ) and simultaneous oxidation of the (C0NH)- group to the resonancestabilized cyanate ion:

In contrast to [Co(CO),(PPh&]CI-HCI,the compound Co(CO)dPPh3)I, having fewer CO groups, reacts with liquid NH3 at -35°C to form carbamoyl triphenylphosphineamminedicarbonylcobalt( + I): Co(CO),(PPh,)I

+

3NH,

-

Co(CO),(PPh&NH,)C,/

/o NH2

+

NH41

Infrared data allow us to conclude that the carbamoyl complexes C O ( C O ) ~ ( P P ~ ~ ) C Oand N MC~O ~ ( C O ) ~ ( P P ~ ~ ) ( N H ~ ) Chave O N Htrigonal ~ bipyramidal structures, with the - C O N M e 2 and PPh3 ligands of the tricarbonyl compound in axial positions (115). Many different examples have shown that the Hofmann acid amide

Metal Carbonyl Chemistry in Liquid Ammonia

27

degradation of carbonylcarbamoyl compounds gives hydridocarbonyl complexes [together with NH4NC0 or CO(NH,),]. Thus, this represents a new and interesting synthetic route to this class of compound. If [CO(CO)~(P~~PCH~)~CM~][CO(CO)J, [(Ph2PCH2),CMe = I, 1,l-tris(diphenylphosphinomethyl)ethane] is allowed to react with liquid NH3 for 12 h at 20"C, NH,[Co(CO)J is obtained, together with the monocar(1 16): bonylcarbamoyl complex Co(CO)[(Ph2PCH2)3CMelCONH2 [CO(CO)~Ph~PCH3~CHd[CO(CO)J + 2NHs

eec

If this is heated in liquid NH3 to 60°C, the hydrido complex HCO(CO)(P~,PCH,)~CM~ is formed, together with NH4NC0. On the basis of IR, 'H-NMR, and 31P-NMR spectra, the hydrido complex has the following structure:

We also investigated the reactions of cationic q5-cyclopentadienylcarbony1 complexes of molybdenum, iron, ruthenium, and osmium with liquid NH3. The reactions of [q5-C&5Mo(C0)3NH3]+or [q5-C&5M~(C0)3PPh3]+ at 20 or 60°C lead to the carbamoyl compounds cis-q5C&,MO(CO)~(NH~)CONH,and truns-q5-C5H5Mo(CO),(PPh3)CONH2, respectively (117). Here is further evidence that the tendency to form carbamoyl complexes from cationic carbonyl compounds decreases with increasing substitution of the CO groups by ligands with lower m-acceptor character. This is in keeping with the increasing electron density at the carbon atoms of the remaining CO molecules, which is characteristically manifest in the different reaction temperatures. Thus, this allows for a rationalization of the formation of q5-C,H5Mo(C0)3H from [q5C&,Mo(CO)J+ in liquid NH3, even at -60°C. whereas q5-

H. BEHRENS

28

CJ15Mo(CO)2(PPh3)Honly is formed from [q5-CJ15Mo(CO),(PPh3)]+at 20°C (117). Table I1 shows a list of neutral carbamoyl complexes that we have synthesized from [q5-C&$e(CO),L]+ (L = CO, PPh3, PEt,), [($C&14CHPh2)Fe(CO)2L]+(CJ14-CHPh2 = diphenylmethylcyclopentadienyl; L = CO, PPhJ, and [(q5-C7H,-CPhJFe(CO)d+(C7H,-CPh3 = triphenylmethylcycloheptadienyl)with liquid NH3 (118). On the basis of solid-state, IR-spectroscopic data for these carbamoyl complexes, we had earlier postulated a dimerization of two complex units via hydrogen bridges. This was later confirmed by X-ray structure analyses on other carbamoyl compounds (119,122).The dicarbonylcarbamoyl complexes always show cis-orientation of the CO ligands. /H

OC.-]Fe-C L

/D------H-N k ; N i -H H’

/

______0‘

[L-CQ, PPh3. PEt3: R = 9’-C5H5.

R

C-F.{..cO L

95-C5H4-CHPh2. 95-C,H8-CPh3

1

The oxidation of these carbamoyl complexes with iodine is interesting, as they are converted to the respective isocyanato compounds in accordance with the Hofmann degradation of primary acid amides (118): q’-C&Fe(CO),C:

/o

m,

+

C,%

qs-CsH,Fe(CO),NCO

+

2HI

TABLE I1 Cationic starting compounds

Carbamoyl compounds q6-C&I$e(CO)aCONHz q5-C5HsFe(CO)(PPhs)ONHz q5-CsH8e(CO)(PEtS)ONHp No reaction No reaction

(q~-CsH1-CHPhz)Fe(CO~CONHz (~s-C5H4-CHPhp)Fe(CO)(PPhs)CONHz

No reaction No reaction products could be isolated (qs-C7H8
29

Metal Carbonyl Chemistry in Liquid Ammonia

The liberated HI reacts with part of the still-unoxidized carbamoyl complex to give the ionic starting material [q5-C5H~e(C0),]I: q5--C,HBFe(CO),C”0

“H,

+

2HI

+ [q5-C,H,Fe(CO),]T

~

+

NH,I

We have also been able to prepare new carbamoyl derivatives of ruthenium and osmium from their cationic cyclopentadienylcarbonyl complexes and liquid NH3 (119). For these experiments, it was first necessary to synthesize the required cationic starting compounds (120), as only [q5-C5HJRu(C0),IPF6, [q5-C5H5Ru(CO)zPPh3]+,and [q5C5H5Ru(C0)zCzHI]PF6had been described in the literature. The carbamoyl derivatives obtained and the respective starting compounds are summarized in Table 111. A contrasting reaction is that of [q5-C5H5R~(CO)z(CZHq)]PFg with liquid NH3, in which addition of an NH3 molecule occurs with a change of the rr-bonding of+CzH, into the o-C-bonded compound [q5-C5H5Ru(CO)2CH 2 4 H 2-N HJ PFS (1 19). In the case of iron and ruthenium, we were successful in obtaining the first dinuclear carbamoyl complexes having fi (P^P = PhzPCHzCH2PPhz) bridges and a - C O N H z ligand at both metal atoms (1 19).

I (M = Fe, Ru)

At higher temperatures, all of the cyclopentadienylcarbamoyl compounds of these Group VIIIB metals in liquid ammonia give the respective hydrido complexes q5-C5H5M(C0),H (M = Fe, Ru, Os), q5C,H&u(CO)(L)H (L = PPh,, PEt,), and (q5-C5H5)Ru(CO)(H)*-(H)(OC)Ru(q5-C5H5) (1 19). It should be mentioned that the already known cationic complexes [M(CO)3(PPh3)zCl]PF6 (M = Fe, Ru, 0s) also react with liquid NHS to give the “noble gas” configurated carbamoyl compounds M(CO)dPPh& (Cl)CONHz(M = Fe, Ru, 0s). With these compounds, we had made the

H. BEHRENS

30

TABLE 111 Cationic starting compounds

Reaction products ~lS-CsH,Ru(CO).$0NHz q5-C$I5Ru(CO)(CNCH&C0NHz ([q'--C,H,-Ru(CO),NHdPF~ No reaction q5-C$IsR~(CO)(NCCH&CONHz T)~-C$I~RU(CO)( PPhJCONHZ *)lS-C$I~Ru(C0)(PEtdCONH~ No reaction

first carbamoyl complexes having a halogen ligand and the metals in the oxidation state +I1 (If9). Understandably, these compounds having cis CO groups do not undergo a Hofmann acid amide degradation to give the respective hydrido complexes. The stability of these Group VIIIB carbamoyl complexes increases in the order Fe, Ru, 0 s . The last class of carbamoyl complexes with which we have been concerned has been that with NO+ ligands. For preparative and structural studies in this area, besides the cationic cyclopentadienylnitrosylcarbonyl complexes of Group VIIB (121),the nitrosylcarbonyl metal cations of Mo and W offer interesting starting materials (121). The results of our work with Mn and Re are summarized in Table IV. TABLE IV Cationic starting compounds

Carbamoyl compounds q5-C$15Mn(CO)(N0)CON H , qs-C$Ipe(CO)( N0)CON H, q 4 MeC&,Mn(CO)( NOICON Hz q5-C&Mn(CNMe)(NO)CONHZ q5-C$I&4n(CNEt)(NO)CONH, qW$IH,Mn(PPh&(NO)CONHz qS-C&Mn(PEt,Ph)(NO)CONH, qW$IHyn(AsPhJ(NO)CONH2 [q5-C&IsMn(NO)CONH&fi

a

Fjrst time prepared. PP = MepPCHzCHzPMez,PhoPCH2CHzPPhz.

Metal Carbonyl Chemistry in Liquid Ammonia

31

FIG.6

The crystal structure of q5-C5H5Mn(CO)(NO)CONH2, given in Fig. 6, shows that the Mn atom is octahedrally coordinated by the C5H5,CO, NO, and CONH, ligands and that the planar carbamoyl group lies in the same plane as the Mn atom. Hydrogen-bridge bonds between the carbamoyl ligands cause the coupling of two molecules to give dimers (122). In the course of this work, we were also able to convert q5C,H,Mn(L)(NO)CONH2 (L = CNEt, PEt,Ph) with MeOH into the corresponding carboalkoxonitrosyl complexes T)~-C,H,M~(L)(NO)COOM~. The nitrosyl ligand containing carbamoyl complexes of Mo and W are given in Table V (121). These compounds could also be converted into their carboalkoxo derivatives with MeOH. Finally, a comparison of the behavior of the three alkyl carbonyl complexes $-C5H,Mo(CO),Me (123, q5-C5H5Fe(CO),Me (123), and (124) toward liquid NH3 shows completely different C3F7Fe(C0)3(PPh3)1 reaction character. Whereas $-C5H,Mo(CO),Me with liquid NH3leads to $-C,H,Mo(CO), TABLE V Cationic starting compounds

Carbamoyl compounds

H. BEHRENS

32

(NH3)H (as the trans isomer) and acetamide, r)5-C5H5Fe(CO)2Me gives the acetyl complex r)5-C5H5Fe(CO)(NHs)COMevia an insertion type reaction (123).

+ 2NHS

+C&Mo(CO),Me vf-C&$e(CO),Me

+ NH,

llq. N&

Ilq. NHI

$-C,,H&b(CO)dNHJH

+ MeCONH,

qs-CJ&Fe(CO)(NHJCOMe

Our preparation of v5-C5H5Mo(CO),(NHs)Hconstituted the first isolation of a mixed hydridoammine complex, which cannot be obtained from r)5-CJH5Mo(CO)sHand NHS. The NH3 can be replaced by py to give r)5-C5H5Mo(CO)2(py)H. The complex r)5-CsH5Fe(CO)(NH3)COMe is also interesting in being the first solvent-substituted acetyl complex of iron ever prepared. In the case of C3F,Fe(CO),(PPhdI, even after only a very short reaction time, no carbamoyl derivative is formed; but, instead, depending upon the reaction time, either CO and PPhs substitution to CsF7Fe(CO),(NH3).J or formation of the ionic complex [C3F7Fe(C0)2(NH3)s]I by addition of a further NHS molecule occurs. Both compounds have the CO groups in cis positions (124). To complete this treatment of the manifold types of reactions of the paramagnetic chromium complexes Cr(C0)J and Cr2(CO)l,,I(cf. Section III,D), it remains to discuss their behavior toward liquid NHS (80).With Cr(CO)J, substitution of three CO ligands by NHsand addition of another NHS molecule gives truns-[Cr(CO)2(NHs)JI, which constitutes the first preparation of a cationic CO complex of chromium: Cr+'(CO)& + 4NH3 + [Cr+'(CO)ANH,)JI

+ 3CO

Based on the mesomerism of dinuclear Cr,(CO),J, (OC)5Cr+'-I-Cro(CO)5], its reaction with [(OC)5Cro-I
Cr&O), J

+ 6NH3 + [Cr+'(CO),(NH~)JIC~(CO)s(NH~)~Il + 5CO

Based on the foregoing experimental results, the versatility of metal carbonyls and their derivatives in their reactions with liquid NH3 may be summarized as follows: (I) substitution of CO or other ligands by NH3 without change in the oxidation number of the transition metal in question; (2)conversion of covalent carbonyl complexes into ionic compounds by addition of NH3 molecules; ( 3 ) "base reactions" in which the transition metal is reduced to a carbonyl metalate with complementary oxidation of a CO ligand to CO(NH2),: (4) valence disproportionations with

Metal Carbonyl Chemistry in Liquid Ammonia

33

formation of carbonyl metalate anions and hexammine metal cations or carbonyl ammine cations; (5) simultaneous valence disproportionation and "base reaction"; ( 6 ) formation of carbamoyl complexes from cationic carbonyl complexes, whereby the former may undergo Hofmann acid amide degradation; and (7) insertion of NH3in metal carbonyl complexes. D. Liquid Ammonia as Solvent for Organometallic Reactions

Liquid NH3 is an excellent solvent for a variety of reactions of metal carbonyls. Of particular importance are the reactions with the N-heterocyclic ligands bipy, phen, and terpy, and with CN-. Thus, in liquid NH3 at 6O"C, we were able to react cisCr(C0)2[(PhzP)2CH2]2(125) with bipy and phen to obtain cisCr(CO)z(bipy)zand cis-Cr(CO)2(phen)2(126). These complexes cannot be obtained from Cr(CO), and the bidentate N ligands directly, as complete substitution of all six CO groups occurs to give Cr(bipy), and C r ( ~ h e n ) ~ (127), respectively. These reactions are, however, remarkable in that the and main products are the mixed species cis-Cr(CO)2[(Ph2P)2CHz]bipy cis-Cr(CO)2[(Ph2P),CHz]phen (126). We further found that, in liquid NH3 at 6O"C, trans-Cr(C0)2[(Ph2P)2CHz]2 isomerizes to the cis compound (126).

In the case of cis-K4[Cr(CO)dCN)J, with phen a ligand exchange also occurs (126): [Cr(CO),(CN)J'-

+ 3phen

120°C

Cr(phen), + 4CN-

+ 2CO

K[V(CO)J reacts with 2,2',2"-terpyridyl (terpy) in liquid NH3 to give complete ligand exchange with simultaneous oxidation of the metal (128), whereby V(terpy), could be prepared for the first time. K[VlCO),l

I2WC

+ 3 t e v y ~ N H J V o ( t e r p y+)K(terpy) ~ + 6CO

We had earlier observed a similar exchange of bipy with terpy in the case of chromium (129): Cr(bipy),

+ Zterpy

I2WC

Cr(terpYh + 3bipy

With bipy or phen, the complex ( Q ~ - C & ~ )in~ liquid N ~ NH3 at 120°C gives, quantitatively, Ni(bipy), or Ni(phen), (96). We were also able to show that K,[Cr(CN),] or K4[Ni(CN)J in liquid NH3 more easily undergoes substitution reactions than the corresponding isoelectronic Cr(CO), and Ni(CO)4(130, 131). Thus, even above -60°C all six CN ligands of K,[Cr(CN),] can be replaced by bipy, phen, terpy,

34

H. BEHRENS

(Ph2P),CH2, (Ph2P),C2H4, or (Ph2As)2C2H4to give the complexes Cr(bipy), and Cr(phen),, first prepared by reductive methods by Herzog and Taube (132), as well as C r ( t e r p ~ ) (129), ~ Cr[(PhzP)2CH213, Cr[(Ph2P)2C2HJ3,and Cr[(Ph,As),C,HJ,. In these reactions, neither the formation of such mixed complexes as K4[Cr(CN)dKN)], Kz[Cr(CN)dcN)J (KN = bipy, phen) or KdCr(CN),terpy], nor the occurrence of the analogous salts with @ [FP = (Ph,P),CHJ, can be observed. It is significant that, even under the most extreme conditions, it is impossible to replace more than four of the CO ligands in Cr(CO), by the aforementioned ditertiary phosphines or arsines (125). The isoelectronic counterpart of Ni(CO)4, the complex K,[Ni(CN)J, undergoes reactions similar to those of KB[Cr(CN),J in liquid NH, (131). At room temperature, the complexes Ni(PPh3)4, Ni(A~ph,)~,and Ni(SbPh,),, first prepared by Wilke er al. (133)by other methods, can be obtained. The tendency to complex formation falls markedly along the series as the dipole character of the incoming ligands decreases (131). Although Ni(C0)4 first reacts with (Ph2P)2C2H4at 200°C to give not quite pure Ni[(Ph2P),C2HJ2 (1341, the same compound can be obtained analytically pure from K,[Ni(CN)J in liquid NH, even at > -33°C. With (Ph2As),C2H4,the corresponding complex Ni[(Ph2As),CZHJ2 is formed (131).

Whereas in Ni(C0)4 only two CO groups can be replaced by bipy or phen, in [Ni(CN)J4-, all of the CN- ligands can be eliminated to give Ni(bipy), and Ni(phen),, which were prepared for the first time by this route (131). An area that we have examined in some detail is that of the behavior of three isoelectronic nitrosylcarbonyls Mn(NO),CO, Fe(NO),(CO),, and Co(NO)(CO), in solutions of KCN in liquid NH,. With Mn(NO),CO, the diamagnetic, dimeric anion [Mn(N0),(CN),le4- with cis NO ligands and a Mn-Mn bond is obtained; heating in ethanol causes isomerization to the trans form (135). Reduction with potassium in liquid NH3 causes rupture of the Mn-Mn bond to give the extremely unstable monomeric [Mn(N0),(CN),I3- which is tetrahedral and isoelectronic with the [Fe(NO),(CN)J2- anion (135). We isolated the latter from Fe(N0)2(C0)2 with KCN in liquid NH3 at 60°C, although the singly negatively charged anion [Fe(NO)2(CO)CN]- could not be obtained (136). We have also used KCN in liquid NH3 to prepare the compounds [CO(NO)(CO),_,(CN),]~- (x = 1, 2, 3) from Co(NO)(CO), by control of the respective molar ratios (136). An excess of KCN leads to the redbrown K,[Co(NO)(CN)J, whereas a molar ratio of nitrosylcarbonyl :KCN of 1 : 1 gives K[Co(NO)(CO),(CN)]. This monocyano complex is very

Metal Carbonyl Chemistry in Liquid Ammonia

35

unstable and readily disproportionates into the dicyano complex (136). K,[Co(NO)(CO)(CN),] and CO(NO)(CO)~ The v(N0) frequencies for these three anions decrease with increasing CN- substitution so that the [CO(NO)(CN),]~-anion shows the very low value of 1485 cm-'. This anion is isoelectronic with [Ni(NO)(CN)3]2prepared by Hieber and Fuhrling (137). The nitrosylcyano metalates K2[Fe(N0)2(CN)2]and K,[Co(NO)(CO) (CN),] react with tetrakis(diphenylphosphinomethy1)methane in liquid NH3 at 120" or 60"C, to give spiroheterocyclic compounds (138):

Fundamentally, optical isomers of the spiroheterocyclic cobalt complex must exist. Since these compounds could not be separated, we tried to synthesize the dicyano derivative under CO substitution by reaction with KCN in liquid NH3 in the hope that a separation into the optical antipodes might be possible. At 120°C. however, the reaction gave, besides K,[Co(NO)(CO)(CN),], a complex in which the C(CH2PPh2)4ligand is tridentate (139):

The complex q5-CsHsMo(NO)(CO),reacts with KCN in liquid NH3 at 120°C to give the pentacyanonitrosyl complex K,[Mo(CN),NO] by elimination of the CJ3, ligand and the two CO groups (140).

H. BEHRENS

36

T~-C&I,MO(NO)(CO)I + 5KCN

+ KdMo(CN)$JO]

+ 2CO + KCSH,

The structure of the [Mo(CN),NO]'- anion has been the subject of attention of many authors (141). E. Cyanocarbonyl Metalates

The cyanocarbonylchromates(0) [Cr(CO),CN]and cis[Cr(CO)4(CN)#-, which are formed by the oxidation of [Cr-"(C0);l2with aqueous solutions of KCN or with (CN), or ICN have been discussed earlier (55, 56). Further studies have shown that the reaction of metal carbonyls or their derivatives with solutions of KCN in liquid NH, is a most advantageous method for the synthesis of new cyanocarbonyl metalates of transition metals, since, in liquid NH3, alkali cyanides are not subject to solvolysis, and also since the broad temperature range of -78 to +12O"C may be employed. Thus, we have been able to obtain quantitative yields of the tricyano-tricarbonyl metalates (O)~UC-[M(CO)~(CN),]~(M = Cr, Mo, W) by direct reaction of the hexacarbonyls (21): M(CO),

+ 3KCN

lWC

KJM(CO)$CN)d

+ 3CO

The bipy-substituted tetracarbonyl complexes M(CO)4bipy (M = Cr, Mo, W) also give the cyanocarbonyl metalates(0) K,[Cr(CO),(CN),] and K,[M(CO)3(CN)3](M = Mo, W) when treated similarly with KCN in liquid NH, at 60°C (92). It is, however, impossible to use this route to obtain cyanocarbonyl metalates having more than three CN- ligands. The rerrucyanodicarbonyl metalates(0) of chromium, molybdenum, and tungsten (126) are readily accessible from the dicarbonyl complexes M(CO)&ipyz (M = Cr, Mo, W),Cr(CO)e[(PhpP)zCH,]2,or Cr(C0)2[(PhzP)2C2H J a (125) with KCN in liquid NH3 at 120°C. cis-M(CO)dbipy), + 4CN-

4

cis-[M(CO)dCN)J4- + 2bipy

Jr + 4CNcis- or trans-Cr(CO)d(PhArC~

+

cis-[Cr(CO)&N) J*- + 2(Ph8P)gC,H,

The starting materials M(CO),bipyz (M = Mo, W) are obtained in quantitative yield when M(CO)8 (M = Mo, W)reacts with bipy in such high-boiling solvents as decalin or tetralin at 190°C (127). The synthesis of Cr(CO),bipy, in liquid NH, was mentioned earlier (Section IV,D). With these studies, and with the exception of the [M(CO)(CN)s]sanions, we have achieved access to all of the cyanocarbonyl metalates(0) and can thus present the following isoelectronic series: M(CO)8, [M(CO)sCNl-, [M(COMCN)aIz-, [M(CO)~(CN)sIs-, [M(cO)z(CN)J'-. [M(CO)(CN),]'-, and [M(CN)J8- (M = Cr, Mo, W).

Metal Carbonyl Chemistry in Liquid Ammonia

37

The [M(CN),]'- ions cannot, however, be prepared from carbonyl complexes, and only the deep-green chromium compound K6[Cro(CN)6] can be obtained by reduction of K3[Cr+"1(CN)6] with potassium in liquid NH3 (142). The foregoing series of metal(0) complexes of Group VIB corresponds to that of the isoelectronic metal(+I) compounds of Group VIIB: [M(CO)J+, M(CO)&N, [M(CO)XCN)il-, [M(Co)ACN)d2-[M(CO)2(CN)J3-, [M(CO)(CN),]'-, and [M(CN)J5-(M = Mn, Re). Before we began our own studies, the only known cyanocarbonyl complexes of manganese were Mn(CO),CN (143) and K[Mn(CO),(CN)d (144). We obtained the colorless tricyanotricarbonylmanganate(+I) from Mn(CO),Cl and KCN in EtOH at 120°C (145). Mn(CO)&I

IZOOC + 3KCN y

K,[Mn(CO).&N)d

+ KCI + 2CO

This complex is readily prepared in pure form since, in contrast to K,[Mn(CO),(CN)J and KCI which are also formed in the reaction, it is extremely soluble in liquid NH3. Tetracyanodicarbonylmanganate(+I) is easily isolated if Mnz(CO)loreacts with KCN in liquid NH3 at 120°C (149, MndCO),,, + 4KCN

14vc

K,[Mn+'(CO),(CN)rl

+ K[Mn-'(CO)J + 3CO

whereby the slightly soluble cyano complex is precipitated. This reaction is comparable to the disproportionation of Mn,(CO),, in pure, liquid NH3, the complex [Mn+'(CO)3(NH3)3][Mn-'(CO)s] being formed instead of K,[Mn+'(CO),(CN)J and K[Mn-'(CO)J (104). As with the pentacyanomonocarbonyl metalates(0) of Group VIB, all attempts to prepare [Mn(CO)(CN)J4- were unsuccessful. With the exception of [Re(CO)(CN),I4-, we have also been able to use new reactions in liquid NH3 to complete the series of possible cyanocarbonylrhenates(+I) (146). Despite numerous attempts by many research groups, preparation of Re(CO),CN, isoelectronic with [W(CO)5CN]-,has remained elusive. In contrast, we were able to prepare the NH3-substituted derivativef~c-Re(CO)~(NH~)&N from Re(CO)5CIand KCN (molar ratio 1 : 1) in liquid NH3 at 12O"C, and by removal of H,O and CO from1 the carbamoyl complex cis-Re(CO)4(NH3)CONHzin liquid NH3 at the same temperature (112, 146). Re(CO)XNH3)CONHz+ NH3

llVC

Re(CO)dNH&CN

+ HzO + CO

Although Re(CO),Cl reacts with KCN in methanol at 100°C to give cisK[Re(CO),(CN),] (147), as our experiments have shown, with excess KCN in ethanol the product isf~c-K~[Re(Co),(CN)~] (146). This tricyano

38

H. BEHRENS

complex cannot be prepared in liquid NH3 even with excess KCN and at 120°C. These conditions lead to ~ U C - K [ R ~ ( C O ) ~ ( C N ) ~which N H ~ is ], also formed from [Re(CO)3(NH3)3]CIand KCN at 120°C in liquid NH3. For the preparation of C~S-K,[R~(CO)~(CN) J, our starting material was the salt {Re(C0)2[(Ph2P)2C2H J2}Cl, which we treated with KCN in ethanol at 200°C (146). All attempts to obtain cyanocarbonylvanadates starting from V(CO)e have been unsuccessful, since, with KCN in liquid NH3, disproportionation occurs (103). 3V(CO),, + 6KCN

119. NHI

K/V+'VCN) J

+ 2K[V4(CO) J + 6CO

This different behavior of V(CO)e toward KCN in liquid NH3, compared to Mn2(CO)lo,can be explained in terms of the especially high stability of the [V(CO),]- anion. However, Rehder (148) has been able to spectroscopically characterize [PhJ'12[V(CO),CN], which he precipitated from the yellow solution formed by UV irradiation of Na[V(CO)s] and NaCN in liquid NH3 at -50 to -60°C. Thus, the isoelectronic series [V-'(CO),CNl2-, [Cro(CO),CN]-, and Mn+'(CO),CN can be considered complete.

F. Mixed Cyanocarbonyl Metalates Prepared from Metal Carbonyl Derivatives and NaN(SiMe3)

In the course of our studies on the preparation of cyanocarbonyl metalates, especially those of Groups VIB and VIIB, encouraged by our successful synthesis of the "mixed" cyanocarbonyl complexes Re(C0)3(NH3)2CNand [Re(CO),NH3(CN)J, we extended our work to include the preparation of further complexes containing ligands other than CO and CN-. For such experiments, however, liquid ammonia is not suitable. A much more promising reaction offered by that discovered by Wannagat and Seyffert (149) involves treatment of sodium bis(trimethylsily1)amide with mononuclear metal carbonyls in which nucleophilic attack of the silyl amide on a CO ligand leads to monocyanocarbonyl metalates. wc

+ O[SiMe3I2

Fe(COk

+ NaN(SiMe&

Ni(CO),

+ NaN(SiMe& 3Na[Ni(CO)FN] + O[SiMe31p

Na[Fe(CO)CNI

awc

In the first instance we used this method on many different monomeric metal carbonyl derivatives, and thereby obtained numerous new mono-

Metal Carbonyl Chemistry in Liquid Ammonia

39

nuclear monocyanocarbonyl metalates not accessible by other routes. These are summarized in Table VI. Most of these new monocyanocarbonyl metalates react with acids, MeI, [Me30]BF4,[Et30]BF4,R3EX (R = Me, Et, Ph: E = Si, Ge, Sn; X = C1, I), and thus are protonated, deuterated, silylated, germylated, or TABLE VI Starting compounds Cr(CO)5~~ Cr(CO),CNCeHI I (CNCeHll = cyclohexyl isonitrile) Cr(CO)SNCCMe3 (Me3CCN = pivalonitrile) Mn(CO)&(X = CI, Br) Re(CO)J(X = CI, Br) CSIRFe(C0)3 ( C a e = butadiene) CSH8Fe(CO)s (C5H8= 1,3-pentadiene) C&Fe(CO)a (C5H8= isoprene) CeH,Fe(C0)3 CeH8= 1.3-cyclohexadiene) CPISe(C0)3 (CeHlo= 1,3-hexadiene CeH10Fe(CO)3 CeHlo = 2,3-dimethylbutadiene) C7HlSe(C0)3 (C7Hlo= 1,3-cycloheptadiene) CBH~F~(CO)~ (C8H8= cyclooctatetraene) C8H14Fe(CO)3 (CaH14 = 2,5-dimethyl-1,3-hexadiene) CeHJWCO)s (CeH8 = 1,3-~yclohexadiene) CeHI&u(COh C,H,, = 2,3-dimethylbutadiene) C~HI&~(CO~ (C7H10 = 1,3-cycloheptadiene) C$IPU(CO)3 ( C P , = cyclooctatetraene) C,H,Mn(CO), C3HSRe(CO), C 3H$e( CO),N 0 (C3Hs = +allyl)

Monocy anocarbon y I metalates

Reference

cis-Na[Cr(CO),(CN)py] cis-Na[Cr(CO)4(CN)CNCeHIIl

150 150

cis-Na[Cr(CO),(CN)NCCMe3]

151

C~S-N~[M~(CO)~(CN)X] cis-Na[Re(CO)4(CN)X] Na[(C,HdFe(CO),CN ]

152 152 150

Na[(C5H8)Fe(CO),CN1

153

Na[(C5HB)Fe(CO)zCN]

154

Na[( CeHB)Fe(CO),CN]

150

Na[(CBH3Fe(CO)aCNI

153

Na[(CeHIo)Fe(CO)zCNI

153

Na[(C7HI3Fe(CO),CN1

154

Nal(C8HdFe(CO)&N1

150

Na[(C8H~ F ~ ( C O ) Z C N I

154

Na[(CeH8)R~(CO)zCNl

155

Na[(CeHIo)Ru(CO)~CNl

155

Na[(C7Hlo)Ru(CO)~CNl

155

Na[(C8Hs)Ru(C0)zCNl

155

Na[(C3HdMn(CO)3CN] Na[(C3HdRe(CO)3CN] Nat(C3HdFe(CO)(NO)(CN)1

157 157 157

40

H. BEHRENS

stannylated, giving the corresponding neutral hydrogen isocyanides or isonitrile compounds with CNH(CND), CNR, or CNERSligands, respectively (152, 1 5 5 4 5 7 ) . Spectroscopic data show that the olefin monocyanocarbonylferrates and -ruthenates and their derived isonitrile complexes occur as two isomeric forms, in which the CN- or isonitrile ligand occupies either an apical or basal position in a square-pyramidal structure. The X-ray structure of the monoclinic C$I,Fe(CO)&NEt (Fig. 7) ( C a , = 1,3-cyclohexadiene)shows that the isonitrile ligand CNEt, a CO group, and two C atoms of the diene part of the C&, ring occupy the basal positions of a square pyramid, with the second CO molecule at the apex (158). We have also been interested in the behavior of dinuclear metal carbonyls and their derivatives toward NaN(SiMed, With MnACO),, and ReACO),,, we obtained the monocyano-enneacarbonyldimetalates(0) Na[MndCO),CN] and Na[RedCO),CN] (153). Their structures can be derived from those of the parent carbonyls by replacement of a CO group by the isoelectronic CN- ligand. The corresponding isonitri!e complexes MnACO),CNR (R = Et, SiMe3, GeMe, SnMe,, PPhd also have linear Mn-Mn-CNR units (159). Thus, it is also characteristic of the dinuclear metal carbonyls that only

FIG.7

Metal Carbonyl Chemistry in Liquid Ammonia

41

one CO ligand is replaced by a CN- anion. The nucleophilic attack of NaN(SiMeJ, on the mixed carbonyl MnRe(CO),, occurs only at the equatorial CO of the rhenium, so that the CN- ligand in the [MnRe(CO),CN]- anion is undoubtedly bonded to the rhenium and not to the manganese (160). In the cases of the olefin-bridged, iron complexes (OC)$e(olen)Fe(CO), (olen = cyclooctatetraene, 1, I '-di-2,4-cyclohexadienyl, I , 1 '-di-2,4-cycloheptadienyl), we have again shown that reaction with NaN(SiMe& leads exclusively to the corresponding monocyano complexes [(OC)r Fe(olen)Fe(C0)2CN]- (155). In its reactions with metal carbonyl derivatives having ligands bearing acidic C-H bonds, NaN(SiMe& behaves as a proton acceptor as the silyl amide is a strong base (161). We have thus been able to show that the acetonitrile-pentacarbonyl complexes M(CO),NCMe (M = Cr, Mo, W) and cycloheptatriene-irontricarbonyl C7H,Fe(CO), are deprotonated as follows, without attack at the CO ligands (151). M(CO)JNCCH3+ NaN(SiMeJ2

6lPC

Na[M(CO)JNCCHJ + HN(SiMe3,

( M = Cr, Mo, W) C7Hpe(C0)3+ NaN(SiMes), + Na[C7H7Fe(CO)d+ HN(SiMeJ2

On the basis of the 'H-NMR spectra of the monocyanomethanidopentacarbonyl metalates of Group VIB, we consider that these anions exist as a tautomeric mixture:

The [C,H,Fe(CO)J anion, with its eight T electrons in the C7H7ring was for many years the subject of numerous theoretical treatments and discussions. Molecular orbital calculations by Hofmann (162) indicated that an 77,allyl anion complex with a noncomplexed butadiene group is more stable than an q4-butadiene complex, in which the allyl anion portion of the seven-carbon ring is not bonded to the metal. We have now been able to confirm this result by an X-ray structure analysis (see Fig. 8) on [Ph&l[C7H7Fe(CO)d (163).

H. BEHRENS

42

FIG.8

An ORTEP plot of the [C7H7Fe(C0)3]-anion shows that the C7H7ring indeed consists of an allyl anion and a diene system. Because only the four T electrons of the allyl anion are involved in bonding to the iron, it has the krypton configuration. Of the numerous reactions of the [C7H7Fe(C0)3]- anion, those with chloroformic esters are particularly interesting in the present context. These reactions lead to ester-substituted cycloheptatriene-irontricarbonyls, namely, 7-R-C7H7Fe(C0)3and l-R-C7H7Fe(CO)3 (R = MeOCO, EtOCO) (164). The 7-isomers are saturated esters, and the 1-isomers are a,@-unsaturated esters. With Et3ECI (E = Si, Ge), only the 7-isomers Et3EC,H,Fe(CO), are formed with the Et3E group at the sp3-hybridized C atom of the ring (164). The reactions of the extremely nucleophilic [C7H7Fe(C0)3]- with the electrophilic [C$I,Fe(CO),]+ or [CsH7Fe(C0)z CNEt]+ cations led to the first preparation of a dinuclear iron complex having the metal atoms bridged by a cyclohexadienylcycloheptatriene group (165):

w

LIOC12Fe

Fe(C013

I L = CO.CNEt1

V REACTIONS OF METAL CARBONYLS AND METAL CARBONYL DERIVATIVES WITH THE MULTIDENTATE NITROGEN LIGANDS bipy, phen, AND terpy Our reactions with liquid NH3 have repeatedly prompted us to investigate the behavior of metal carbonyl derivatives toward the N-hetero-

Metal Carbonyl Chemistry in Liquid Ammonia

43

cyclic ligands bipy, phen, and terpy in suitable solvents, which is the subject of the following section.

A. Titanium and Vanadium Quantitative replacement of the cyclopentadienyl and CO ligands occurs in the reactions of (Q~-C,H,),T~(CO),(166) and r15-CJ-15V(CO)4(128) with phen or terpy to give the metal(o) complexes T i ( ~ h e n )Ti(terpy)z, ~, and V(terpy)z. The complex V(terpy), is also formed by the reaction of terpy with V(CO),[(Ph2P),C2H+Ior V(C0)2[(Ph2P)zCzH+Iz (167). The latter phosphine-substituted paramagnetic vanadium compounds were first prepared by u s from v(C0)s. In contrast to bipy and phen, below 80°C, the complex V(CO),[(Ph2P),CzHJ undergoes disproportionation to give [V+11~bipy~,l{V-'~CO~~[~PhzP~2C2~41~2 or [v+''(~hen)~l{v-'(C0)~1(Ph2P)~CzH4l}z (167).

B. Chromium, Molybdenum, and Tungsten

Although all six of the CO ligands of Cr(CO)6can be replaced by bipy or phen to give Cr(bipy), or Cr(phen),, for Mo(CO),, and w(co)6, even extreme temperature conditions only allow for a maximum substitution of four CO groups (127). With terpy, on the other hand, total replacement occurs with all three hexacarbonyls, giving M(terpy)z(129). The reactions of [$-C,H,Mo(CO)J, with bipy, phen, or terpy follow a similar pattern, giving cis-Mo(CO)z(bipy),, cis-Mo(CO)Aphen), or Mo(terpy), (166). The (MO=) ~Mo, W; C7H8= cycloheptatriene) with reaction of T ~ - C ~ H ~ M ( C bipy or phen (NN) in nonpolar solvents allowed us to obtain the CObridged diamagnetic complexes

which probably contain metal-metal double bonds (168). Thus, besides the mononuclear types M(CO)4(flN)and M(CO),(N-N), (127), it was also possible to prepare for the first time dinuclear CO-bridged Group VIB carbonyl complexes with bipy and phen (168). In polar solvents (L = MeOH, Et,O, THF, MeCOMe, or MeCN), however, the T ~ - C ~ H ~ M ((CMO=) ~Mo, W) compounds with N-N (bipy, phen) give the mixed mononuclear complexesfuc-M(CO)dN-N)L (168);

44

H. BEHRENS

and, in the presence of X- (X = C1, Br, I, CN, NCS, N3, SH), the same reaction leads to the anions f u c - [ M ( C O ) d ~ ) X ] -(169). These observa tions opened a route for the synthesis of numerous "mixed" neutral complexes of the type M(CO)XKN)L, where X- is replaced by NH3, py, PPh3, PPh2Cl, or SO2, as has already been partially described (170). The coordinatively unsaturated complexes [ M ( C 0 ) d m ) l 2 are extremely reactive. With polydentate N and P ligands, new mono- or dinuclear bridged mixed-ligand complexes are formed by fission of the CO bridges. Examples are given in Table VII (171, 172). The CO ligands are in all cases facial to the molybdenum. The comand [Mo(CO)ArN)] plexes [Mo(CO)drN)l2[NH,-C$I.,-NHJ [NHACH&,NHJ have the following structures:

Remarkably, [M~(CO)~bipy], is reduced by Nabipy to [bipy(OC)sMoM~(CO)~bipy]~with a nonbridged metal-metal bond and cis CO ligands (173). With this reaction, we were able to prepare the first example of a derivative of the [MO~(CO),,]~-anion in which four CO groups are replaced by two bipy molecules. The neutral complex bipy(ON)(OC),Mo-Mo(Co),(NO)bipy, prepared from [M~(CO)~bipy], and NO, is isosteric with the [bipy(OC)sMoMo(C0),bipyl2- anion (174). This also has a nonbridged structure having a Mo-Mo bond with the NO ligands occupying the axial positions in the double octahedron. Further examples of dinuclear CO-bridged compounds of Group VIB are the complexes [MO(CO)~L,],, which we have synthesized from C,H,Mo(CO), (C,Hs = cycloheptatriene) and maleic and fumaric esters (L) (175, 176).

Metal Carbonyl Chemistry in Liquid Ammonia

45

TABLE VII Mononuclear complexes

Dinuclear complexes

When treated with I2 (molar ratio 1: l), the CO-bridged species

[ M(C0)3(m)I2is oxidized to the diamagnetic iodine-bridged molybde-

num( +I) complexes

which we also obtained (177) from M(CO)Xm) and I2in the molar ratio of 2: 1. In contrast, the reactions of M(CO)dI@J),with iodine lead to the ionic dicarbonyl metal(+11) compounds [M( CO)2( RN),I]I or [M( C 0 ) dflN),I]I, (M = Mo, W) with heptacoordinated metals in the cations (178).

C.

Iron and Cobalt

Studies by Hieber and Lipp (179) showed that with bipy Fe(CO), undergoes disproportionation to [Fe(bipy)J[Fe,(CO),,]. Similar reactions occur with other N bases (131, which means that Fe(CO), cannot be used as a starting material to obtain substitution products with N ligands. We have, however, found that Fe(CO),terpy can be formed from [v5CJ-I,Fe(CO)J, or C$I,Fe(CO), ( C a , = cyclooctatetraene) and terpy (180, 181). In the trigonal-bipyramidal complex, both CO molecules are cis-oriented. It is well known that CodCO), disproportionates on treatment with the N bases NH3 and N% = bipy or phen) to give cobalt( +II)bis[tetracarbonylcobaltate( -I)] with the cations [Co(NHJd2+or

(m

H. BEHRENS

46

[C0(flN)J2+(13). Using the reaction of (nor-C,HB),CoACO), (nor-C,Hs = in nonpolar solvents, we were able, however, norbornadiene) with CN),, which are the first real substitution products to prepare CO&CO)~( of Co,(CO), with N ligands (182);they have the same, CO-bridged structures as CodCO), with each of the metal atoms having two terminal CO in cis positions. The same method allowed us to ligands replaced by synthesize the diphosphine-substituted complex CoAC0),[(Ph2P),C2H4],, JI and reduced by sodium which is oxidized by I2 to CO(CO)~[(P~,P),C~H to N~(CO(CO),[(P~,P)~C,HJ} (182). In contrast, (C,HB),CoACO), (CJ-I, = 1,3-~yclohexadiene)with K N does not undergo disproportionation into Co(+II) and Co(-I), but into Co(+I) and Co(-I) (182).

+

( C & & € O ~ C O ) ~ 3(NeN) + [Co+Ym)J[Co-YCO)J

+ 2C&

We have also reacted Co,(CO), with 1-diphenylphosphino-2-diethylaminoethane [Ph2P-C2H4-NEt2 (FN)] and obtained [Co(CO)@N)J[Co(CO)J or CodCO)dP"N),, depending upon the reaction conditions (183). The I?N ligand is always monodentate, bonding to the cobalt exclusively via the phosphorus. Both ( P x ) ligands in the [Co(CO)AFN)J+ cation are in the axial positions of a trigonal bipyramid and are also in trans positions in the nonbridged CodCO)&PN), (183). That the PIN ligands in C O ~ ( C O ) ~ Pwhich ~ ) ~ ,can also be obtained from CodCO),(nor-C,HJ, (nor-C,H, = norbornadiene) and FN via elimination of metallic cobalt, are monodentate is readily shownpy N-methylation with Me1 to give the [Co2(CO)dPh2P-CH2CH2-4V Et2Me).J2+ cation (183).

VI

HIGH-PRESSURE SYNTHESES At the beginning of this review, I mentioned that in my dissertation, over 40 years ago, I was concerned with high-pressure synthesis of metal carbonyls from anhydrous metal halides. As is often the case, the later periods of one's scientific life often see one remembering and returning to the first areas of interest. This has also been true for myself, and in recent times my group has been able to prepare numerous known and unknown derivatives of Co2(C0),, in one step by high-pressure synthesis using cobalt metal, the pertinent ligand, and CO. Several examples are given in Table VIII (184).

Metal Carbonyl Chemistry in Liquid Ammonia

47

TABLE VIII Starting compounds

Co + bipy + CO Co + phen + CO c o + (€9. + co CO + PPh3 CO Co + (FN)b+ CO Co + SnCI, + CO

+

+

Co + InCl CO Co + PPh3 + SnCI, + CO Co + ( f N ) + SnCI, + CO Co + (FN) + Cd + CO C o + (FN) + HgBr, + CO

co

(“c)

pressure (bar)

Time (h)

I80 180 I50 200 I80 200

200 200 200 200 350 200

24 24 12 72 48 72

200 200 I80 I80 I80

200 200 300 300 350

72 72 36 48 48

Temperature

Reaction products [Co(bipy)&Co(CO)& [Co(p~en)dCo(CO)J, [CO~PP)~CO)J[CO(CO)J, Co&CO)&FhJ, Co&CO)&PN)z [Co(CO)J$nCI, [Co(CO)&SnCI + CoCI, {[Co(CO).J,In(p-CI)h [Co(CO),PPh&SnCI, [Co(CO)APi\l)l,SnCl, [Co(CO)dFN)],Cd [Co(CO)dFN)],Hg + CoBr,

FP = Ph2P-C,H4-PPh,.

’P% = Ph,P-C,H,-NEt,.

VI I CONCLUSIONS

I hope to have shown in this review that our work on reactions in liquid NH3 has contributed to the development of the preparative chemistry of metal carbonyls, especially in the field of anionic carbonyl metalates, and that it has initiated new impulses and ideas for further advances in this still very active area of organometallic chemistry. One’s own scientific activities are, as a matter of course, always dependent upon and closely bound to the situations of the times. Thus, it is understandable that the fateful developments in Germany in the 1930s and 1940s had grave effects upon my own generation. The beginning of World War I1 occurred at exactly the time when I began my first independent studies in the area of metal carbonyls. This not only caused numerous plans for the future to be abandoned.but also meant that many valuable research years were lost, particularly as a result of the fact that, by the end of the war, the Munich Institute was almost totally destroyed. In this respect, Hieber is to be sincerely thanked, for it was largely through him that his Institute, of which I was a member until 1962, so rapidly regained its international reputation, which could not have been expected in 1945.

40

H. BEHRENS

Looking back at the tempestuous and fascinating progress in the chemistry of metal carbonyls that has occurred over the past 40 years, and which, at the beginning, were unimaginable, the question automatically arises of how the future might be. In this connection, I can vividly remember the comments of Hieber, particularly at the beginning of the 1950s, as he thought that the chemistry of metal carbonyls would soon be exhausted. This was, of course, during the period prior to the onset of the exciting developments in the organometallic chemistry of the transition metals, beginning in the mid- 1950s. which naturally had an enormous influence on the study of metal carbonyls and which allowed this area to break away from its more classical character. Whether this rate of advancement will increase, or perhaps decrease, is difficult to predict. Based on the experience of the past, however, it seems definite that chemists all over the world will continue to offer new and fundamental ideas for further growth in organometallic chemistry, even though one sometimes thinks that nothing new can possibly develop. If this does occur, it is certain that, as in the past, many other areas of chemistry will be markedly influenced. In this general context, I am sure that studies of reactions in liquid NH3 can also play an interesting role in the future, particularly when it is increasingly recognized that experimentation with liquefied gases presents no great difficulty. Finally, I wish to acknowledge that this research has been primarily the result of the efforts and enthusiasm of very many co-workers and students during both good and bad times.

REFERENCES I . W. Hieber, Z . Elekrrochem. 43, 390 (1937). W. RudorfFand U . Hofmann, Z. Phys. Chem. Abr. B 28,351 (1935). R. Brill, Z. Krista//ogr. 77, 35 (1934). W. Hieber, Angew. Chem. 55, 1 (1942). W. Hieber, H. Behrens. and U . Teller, Z . Anorg. Allg. Chem. 249, 26 (1942). H. Behrens and E. Eisenmann. Z. Anorg. Allg. Chem. 278, 155, 166 (1955). H. Behrens, Tagungsber. Chem. Ges. DDR, Haupfjahresfag. 1954, Akademie-Verlag Berlin, 1955. 8. H. Behrens and G. von Taeuffenbach, Z . Anorg. Allg. Chem. 315,259 (1962). 9. W. Reppe, "Neue Entwicklungen auf dem Gebiet der Chemie des Acetylens und Kohlenoxids." p. 120, Springer-Verlag, Berlin and New York, 1949. 10. W. Hieber and W. Hubel, Z. Elekrrochem. 57, 235 (1953). 1 I . FIAT Review of German Science, Naturforsch. Med. Deutschland 1939-1946, Band 24, "Anorganische Chemie," p. 112. 12. H . Behrens, Angew. Chem. 61, 444 (1949). 2. 3. 4. 5. 6. 7.

Metal Carbonyl Chemistry in Liquid Ammonia

49

13. W. Hieber, W. Beck, and G. Braun, Angew. Chem. 72, 795 (1960), and references

cited therein.

14. H. Behrens and R. Weber, Z. Naturforsch. Teil B 7, 321 (1952); 2. Anorg. Allg. Chem. 281, 190 (1955). 15. H. Behrens and R. Weber,Angew. Chem. 67,521 (1955); 2. Anorg. A&. Chem. 291, 122 (1957). 16. E. Weiss and W. Buchner, Helv.Chim. Acra 46, 1121 (1963). 17. H. Behrens and W. Haag, Z. Naturforsch. Teil B 14,600 (1959). 18. J. E. Ellis, C. P. Parnell. and G. P. Hagen. J. Am. Chem. Soc. 100, 3605 (1978). 19. H. Behrens and W. Klek,Z. Anorg. Allg. Chem. 292, 151 (1957). 20. H. Behrens and W. Haag, Chem. Eer. 94,312 (1961). 21. H. Behrens and J. Vogl, Chem. Eer. 96,2220 (1963). 22. H. Behrens and W. Haag, Chem. Ber. 94, 320 (1961). 23. M. R. Churchill and S. W.-Y. Ni Chang, J. Chem. Soc. Chem. Commun. p. 691 (1973). 24. E. Lindner, H. Behrens, and S. Birkle. J . Organomet. Chem. 15, 165 (1968). 25. H. E. Podall, H. B. Restridge, and H. Shapiro,J. Am. Chem. Soc. 83, 2057 (1961). 26. U. Anders and W. A. G. Graham, J. Chem. Soc. Chem. Commun. p. 499 (1965). 27. W. C. Kaska, J. Am. Chem. SOC. 90, 6340 (1968); 91, 2411 (1969). 28. R. G. Hayter. J. Am. Chem. Soc. 90, 6340 (1966). 29. J. E. Ellis and G. P. Hagen, J. Am. Chem. Soc. 96, 7825 (1974); J. E. Ellis, S . G. Hentges, D. G. Kalina, and G. P. Hagen, J. Organomet. Chem. 97, 79 (1975): J. E. Ellis and G. P. Hagen, Inorg. Chem. 16, 1357 (1977); J. E. Ellis and E. A. Flom, J. Organomet. Chem. 99, 263 (1975). 30. C. Ungurenasu and M. Palie, J. Chem. Soc. Chem. Commun. p. 388 (1975). 31. E. Lindner, H. Behrens, and D. Uhlig, Z. Naturforsch. Teil B 28, 276 (1973). 32. L. H. Jones and R. S. McDowell, Spectrochim. Acta 20, 248 (1964). 33. R. G . Hayter, J. Am. Chem. Soc. 88, 4376 (1966). 34. E. Lindner, H. Behrens, and S. Birkle. J . Organornet. Chem. 15, 165 (1968). 35. W. F. Edgell and N. Pauuwe, J. Chem. Soc. Chem. Commun. p. 284 (1969). 36. L. B. Handy, J. K. Ruff, and L. F. Dahl, J. Am. Chem. Soc. 92, 7312 (1970). 37. R. D. Wilson, S. A. Graham, and R. Bau, J. Organomet. Chem. 91, C49 (1975). 38. D. W. Hart, R. Bau, and T. F. Koetzle. manuscript in preparation. 39. M. Andrews. D. L. Tipton. S. W. Kirtley, and R. Bau,J. Chem. SOC.Chem. Commun. p. 181 (1973); J. P. Olsen, T. F. Koetzle. S. W. Kirtley, M. Andrews, D. L. Tipton, and R. Bau, J. Am. Chem. SOC.96,6621 (1974). 40. J. L. Petersen, P. L. Johnson, J. O’Connor, L. F. Dahl, and J. M. Williams, Inorg. Chem. 17, 3460 (1978); R. Bau, R. G. Teller, S. W. Kirtley, and T. F. Koetzle, Acc. Chein. Res. 12, 176 (1979). 41. U. Anders and W. A . G. Graham, J. Am. Chem. Soc. 89,539 (1967). 42. A. S . Foust. W. A. G. Graham. and R. P. Stewart, Jr., J. Organomet. Chem. 54, C22 (1973). 43. E. 0. Fischer and S. Vigoureux, Chem. Ber. 91, 2205 (1958). 44. J. E. Ellis and M. C. Palazzotto, J. Am. Chem. SOC. 98, 8264 (1976); J. E. Ellis, P. T. Barger, and M. L. Winzenburg, J. Chem. Soc. Chem. Commun. p. 686 (1977). 45. H. Behrens, J . Organomet. Chem. 94, 139 (1975). 46. H. Behrens, W. Klek, and J. Kohler, Angew. Chem. 69, 716 (1957). 47. H. Behrens and J. Kohler, Z. Anorg. Allg. Chem. 300, 51 (1959). 48. W. Hieber and E. Romberg, 2. Anorg. Allg. Chem. 221, 349 (1935): W. Hieber and F. Mdhlbauer, Z. Anorg. Allg. Chem. 221, 337 (1935).

H. BEHRENS 49. W. Strohmeier and K . Gerlach, Z. Naturforsch. Teil B IS,413 (IW);Chem. Ber. 93, 2087 (1960);W.Strohmeier, K. Gerlach and G. Matthias, Z. Nafurforsch. Teil B IS, 621 (1960);W.Strohmeier, K. Gerlach, and D. von Hobe, Chem. Ber. 94, 164 (1961); W. Strohmeier and G. Schonauer, Chem. Ber. 95,1767 (1%2); W.Strohmeier,Angew. Chem. 7 6 , 873 (1964); W. Strohmeier, J. F. Guttenberger, H. Blumenthal, and G . Albert, Chem. Ber. 99,3419 (1966). 50. J. F. Guttenberger, Angew. Chem. 79, 1071 (1%7). 51. R. D. Bertrand, D. A. Allison, and J. G. Verkade, J. Am. Chem. Soc. 92,71 (1970). 52. R. G. Hayter, Inorg. Chem. 3,711 (1964). 5 3 . J. Chatt and D. T. Thompson, J. Chem. Soc. p. 2713 (1964). 54. D.Sellmann, A. Brandl, and R. Endell. J. Orgunomet. Chem. 111, 303 (1976). 55. H. Behrens and J. Kohler, Z. Nururforsch. Teil B 14,463 (1959). 56. H. Behrens and J. Kohler, Z. Anorg. Allg. Chem. 306,W (1960). 57. H.Behrens and H. Zizisperger, Z. Naturforsch. Teil B 16, 349 (l%l). 58. H. Behrens and R. Schwab, 2. Naturforsch. Teil B 19, 768 (1964). 59. H. Behrens, R. Schwab, and D. Herrmann, Z. Naturforsch. Teil B 21, 590 (1%). 60. H. Behrens, E. Lindner. and S. Birkle, Z. Anorg. Allg. Chem. 369, 131 (1969). 61. E.0.Fischer and K. Ofele, Chem. Ber. 93, I156 (1960). 62. A. Wojcicki and M. F. Farona, J . Inorg. Nucl. Chem. 26, 2289 (1%). 63. W. J. Schlientz and J. K . Ruff, J. Orgunomet. Chem. 33, C64 (1971). 64. W.J. Schlientz and J. K . Ruff, Synth. Inorg. M e t . Org. Chem. 1, 215 (1971). 65. W.J. Schlientz and J. K. Ruff, Inorg. Chem. 11, 2264 (1972). 66. J. E. Ellis and G. P. Hagen, Inorg. Chem. 16, 1357 (1977). 67. H.Behrens and D. Henmann, Z. Nuturforsch. Teil B 21, 1234 (1966). 68. H.Behrens, Angew. Chem. 80,244 (1968). 69. E. Lindner and H. Behrens. Spectrochim. Acra Purr A 23, 3025 (1%7). 70. L. 9. Handy, J. K. Ruff, and L. F. Dahl, J . A m . Chem. Soc. 92, 7327 (1970). 71. J. K. Ruff, Inorg. Chem. 7 , 1821 (1968);8, 86 (1%9); J. K. Ruff and R. B. King, Inorg. Chem. 8 , 180 (1969). 72. H.Behrens, D. Uhlig. and E. Lindner, Z. Anorg. Allg. Chem. 394, 8 (1972). 73. J. K. Ruff, Inorg. Chem. 7 , 1818 (1968). 74. J. K. Ruff, Inorg. Chem. 6 , 2080 (1%7). 75. H. Behrens, H.-D. Feilner, E. Lindner, and D. Uhlig. Z. Nafurforsch. lei1 B 26, 990 ( 1971). 76. W.Hieber and U. Teller, 2. Anorg. A//g. Chem. 249, 43 (1942). 77. J. K. Ruff, R. P. White, and L. F. Dahl,J. Am. Chem. Soc. 93,2159 (1971). 78. H. Behrens and D. Herrmann, Z. Naturforsch. 7eil B 21, 1236 (1%6). 79. R. B. King and W. M. Douglas, J . Am. Chem. Soc. 95, 7528 (1973); R. B. King and C. A. Harmen, J. Orgunomel. Chem. 86, 139 (1975). 80. H.Behrens and D. Herrmann, Z. Anorg. A&. Chem. 351, 225 (1%7). 81. H. Behrens and H. Schindler, Z. Nuturforsch. l e i / B 23, 1109 (1968). 82. H. Behrens, M. Moll, and E. Sixtus, Z. Naturforsch. Teil B 32, 1105 (1977). 83. D. Uhlig, H. Behrens. and E. Lindner, 2. Anorg. Allg. Chem. 401, 233 (1973). 84. J. K. Ruff, Inorg. Chem. 6 , 1502 (1967). 85. T. J. Marks, J. Am. Chem. Soc. 93, 7080 (1971). 86. H.Behrens, M. Moll, E. Sixtus, and G. Thiele, Z. Nafurforsch. Teil B 32, 1109 (1977). 87. H. Behrens, M. Moll, E. Sixtus, and E. Sepp, Z. Nuturforsch. Teil B 32, 1114 (1977). 88. R. G. Amiet, P. C. Reeves, and R . Pet1it.J. Chem. SOC. Chem. Commun. p. 1208 (1%7). 89. B. D. Dombeck and R. J. Angelici, Inorg. Chem. 15, 1089 (1976).

Metal Carbonyl Chemistry in Liquid Ammonia

51

90. W. P. Fehlhammer, A. Mayr, and B. Olgemiiller, Angew. Chem. 87, 290 (1975). 91. G. Schmid and H.-P. Kempny, Z. Anorg. Allg. Chem. 418, 243 (1975). 92. H. Behrens and N. Harder, Chem. Ber. 97,433 (1964). 93. W. Hieber and E. Romberg, Z. Anorg. A&. Chem. 221, 349 (1935); W. Hieber and F. Miihlbauer, Z. Anorg. Allg. Chem. 221, 337 (1935). 94. H. Behrens and H. Zizlsperger, J . Prakt. Chem. 14, 249 (1961). 95. E. 0. Fischer and R. Jira, Z. Naturforsch. Ted B 8, 217 (1953). %. H. Behrens and K. Meyer, Z. Nafurforsch. Teil B 21,489 (1%6). 97. H. Behrens and P. Passler. Z. Anorg. Allg. Chem. 365, 128 (1969). 98. H. Behrens, Angew. Chem. 74, 120 (1%2). 99. H. Behrens and H. Wakamatsu, Z. Anorg. Allg. Chem. 320, 30 (1963). 100. H. Behrens and H. Wakamatsu, Chem. Eer. 99, 2753 (1%6). 101. H. Behrens, unpublished results. 102. H. Behrens, J. Organomet. Chem. 94, 139 (1975). 103. H. Behrens and K. Lutz, Z. Anorg. Allg. Chem. 354, 184 (1967). 104. H. Behrens, E. Ruyter , and H. Wakamatsu, Z. Anorg. Allg. Chem. 349, 241 (1967). 105. H. Behrens and H. Schlenker, unpublished results. 106. H. Behrens and E. Ruyter, Z. Anorg. Allg. Chem. 349, 258 (1%7). 107. H. Behrens, E. Lindner, and P. Passler, Z. Anorg. AIIg. Chem. 365, 137 (1%9). 108. R. J. Angelici, J. Chem. Soc. Chem. Commun. p. 486 (1%5). 109. R. J. Angelici and D. L. Denton, Inorg. Chim. Acta, 2, 3 (1968). 110. R. J. Angelici, Ace. Chem. Res. 5, 335 (1972). and references cited therein. 111. H. Behrens, E. Lindner, D. Maertens, P. Wild, and R.-J. Lampe, J. Organornet. Chem. 34, 367 (1972). 112. H. Behrens, R.-J. Lampe, P. Merbach. and M. Moll. J. Organomel. Chem. 159, 201 ( 1978). 113. R. W. Brink and R. J. Angelici, Inorg. Chem. 12, 1062 (1973). 114. T. Kruck and M. Noack, Chem. Ber. 97, 1693 (1964). 115. H. Krohberger, H. Behrens, and J. Ellermann, J. Organomet. Chem. 46, 139 (1972). 116. J. Ellermann, J. F. Schindler, H. Behrens, and H. Schlenker, J. Organornet. Chem. 108, 239 (1976). 117. A. Wister, H. Behrens, and M. Moll, Z. Anorg. Allg. Chem. 428, 53 (1977). 118. J . Ellermann, H. Behrens, and H. Krohberger, J . Organornet. Chem. 46, I19 (1972). 119. H. Behrens and A. Jungbauer, Z. Nafurforsch.. Ted E 34, 1477 (1979); H. Wagner, A. Jungbauer. G. Thiele, and H. Behrens, Z. Naturforsch., Teil B 34, 1487 (1979) [Crystal structure of qs-C,H,Ru(CO)&CONH&]; A. Jungbauer and H. Behrens. J . Organomet. Chem., in press: A. Jungbauer and H. Behrens, Z. Naturfnrsch.,in press. 120. A. Jungbauer and H. Behrens, Z. Naturforsch. l e i / B 33, 1083 (1978). 121. H. Behrens, G. Landgraf, P. Merbach, and M. Moll,J. Organomet. Chem., in press; H. Behrens and G. Landgraf, unpublished results. 122. D. Messer, G. Landgraf, and H. Behrens, J. Organomel. Chem. '172, 349 (1979). 123. H. Behrens, A. Pfister, M. Moll, and E. Sepp, Z. Anorg. Allg. Chem. 428, 61 (1977). 124. H. Krohberger, J. Ellermann, and H. Behrens. Z. Naturforsch. Teil E 27,890 (1972). 125. J. Chatt and H. R. Watson, J . Chem. Soc. p. 4980 (l%l). 126. H. Behrens, E. Lindner, and J. Rosenfelder, Chem. Eer. 99, 2745 (1966). 127. H. Behrens and N. Harder, Chem. Eer. 97, 426 (1964). 128. H. Behrens, H. Brandl, and K. Lutz, Z. Nafurforsch. Teil B 22, 99 (1967). 129. H. Behrens and U. Anders, Z. Narurforsch. Teil E 19, 767 (1964). 130. H. Behrens, K. Meyer, and A . Miiller, Z. Nafurforsch. Teil B 20, 74 (1965). 131. H. Behrens and A. Miiller, Z. Anorg. Allg. Chem. 341, 124 (1965).

52 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151.

H. BEHRENS S. Herzog and R. Taube, Z. Chem. 2, 208 (1962). G. Wilke, E. W. Miiller, and M . Korner, Angew. Chem. 73, 33 (1961). J. Chatt and F. A. Hart, J. Chem. SOC. p. 1378 (1960). H. Behrens, E. Lindner, and H. Schindler, Z. Anorg. Allg. Chem. 365, 119 (1969). H. Behrens, E. Lindner, and H. Schindler, Chem. Eer. 99, 2399 (1966). W. Hieber and H. Fiihrling, Z. Anorg. AlIg. Chem. 373,48 (1970). J. Ellermann. H. Behrens, H. Dobrzanski, and F. Poersch, Z. Anorg. Allg. Chem. 361, 306 (1968). J. Ellermann. H. Behrens, and H. Dobrzanski, Z. Naturforsch. Teil B 23, 560 (1%8). H. Behrens and H. Schindler, Z. Narurforsch. Teil B 23, 1 I10 (1968). N.G. Conelly, Inorg. Chim. Acra Rev. 6.47 (1972). and references cited therein (page 65). E. A. Heintz, J. Inorg. Nucl. Chem. 21, 262 (1961). B. Chiswell and L. M. Venanzi, J . Chem. Soc., A , p. 417 (1966). W. Hieber and W. Schropp, Jr., Z. Nafurforsch. Teil E 14, 460 (1959). H. Behrens, E. Ruyter, and E. Lindner, Z. Anorg. Allg. Chem. 349, 251 (1%7). H. Behrens, E. Lindner, and P. Pilssler. Z. Anorg. Allg. Chem. 361, 125 (1%8). W. Hieber and L. Schuster. Z. Anorg. Allg. Chem. 287,214 (1956). D. Rehder, J. Organomer. Chem. 37, 303 (1972). U. Wannagat and H. Seyf€ert. Angew. Chem. 77,457 (1965). H. Behrens and M. Moll, Z. Anorg. ANg. Chem. 416, 193 (1975). M. Moll, H. Behrens. R. Kellner, H. Kniichel, and P. Wiirstl, Z. Narurforsch. Teil B 31, 1019 (1976). H. Behrens, H.-J. Ranly, and E. Lindner, Z. Anorg. Allg. Chem. 409, 299 (1974). H. Behrens, M. Moll, W. Popp, and P. Wiirstl, Z. Narurforsch. Teil B 32, 1227 (1977). M. Moll, H. Behrens, and W. Popp, 2. Anorg. ANg. Chem.. 458, 202 (1979).

152. 153. 154. 155. H. Behrens and W. Popp, unpublished results. 156. H. Behrens, M. Moll, and P. Wiirstl. Z. Naturforsch. Ted B 31, 1017 (1976). 157. H. Behrens, M. Moll, W. Popp, H.-J. Seibold, E. Sepp, and P. WUrst1.J. Organomet. Chern., in press; M. Moll, H.-J. Seibold, and W. Popp, J. Organomet. Chem., in press; P. Merbach, P. Wiirstl, H.-J. Seibold, and W. Popp, J. Organornet. Chem., in

press; H. Behrens and H.-J. Seibold, unpublished results.

158. H. Behrens, G. Thiele, A. Piirzer, P. Wiirstl, and M. Moll, J. Organomet. Chem. 160, 255 (1978). 159. H. Behrens, P. Wiirstl, P. Merbach, and M. Moll, Z. Anorg. Allg. Cham. 456, 16 (1979). 160. H. Behrens and V. Schneider, unpublished results. 161. C. Kriiger, J . Organomet. Chem. 9, 125 (1967). 162. P. Hofmann. Z. Naturforsch. Teil B 33, 251 (1978), and references cited therein. 163. E. Sepp, A. Piirzer, G. Thiele, and H. Behrens, Z. Narurforsch. Teil B 33,261 (1978). 164. H. Behrens, K. Geibel, R. Kellner, H. Knochel. M. Moll, and E. Sepp, 2. Naturforsch. Teil E 31, 1021 (1976). 165. M. Moll, P. Wiirstl, H. Behrens, and P. Merbach, Z. Naturforsch. Teil B 33, 1304 (1978). 166. H. Behrens and H. Brandl, Z. Naturforsch. Teil B 22, 1216 (1967). 167. H. Behrens and K. Lutz, Z. Anorg. Allg. Chem. 356,225 (1968). 168. H. Behrens, E. Lindner. and G. Lehnert, J . Organornet. Chem. 22,439 (1970). 169. H. Behrens, E. Lindner, and G. Lehnert, J. Organornet. Chem. 22, 665 (1970). 170. H. Behrens, G . Lehnert, and H. Sauerborn, Z. Anorg. Allg. Chem. 374, 310 (1970). 171. H. Behrens, W. Topf, and J. Ellermann,J. Organomet. Chem. 63,349 (1973). 172. H. Behrens. W. Topf, and J. Ellermann, J. Organomet. Chem. 63,369 (1973).

-

Metal Carbonyl Chemistry in Liquid Ammonia 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184.

53

G. Lehnert, H. Behrens, and E. Lindner, Z. Nuiurforsch. Teil E 25, 106 (1970). E. Lindner, H. Behrens. and G. Lehnert, Z. Narurforsch. Teil E 25, 104 (1970). H. Schaper and H. Behrens, J . Organomei. Chem. 113, 361 (1976). H. Schaper and H . Behrens, J. Orgunomer. Chem. 113, 377 (1976). H. Behrens and W. Ziegler, Z. Anorg. Allg. Chem. 365, 269 (1969). H. Behrens and J. Rosenfelder, Z. Anorg. AlIg. Chem. 352, 61 (l%7). W. Hieber and A. Lipp, Chem. Ber. 92, 2075 (1959). H. Behrens and W. Aquila, Z. Nafurforsch. Teil E 22,454 (1967). H. Behrens, H.-D. Feilner, and E. Lindner,Z. Anorg. Allg. Chem. 385, 321 (1971). H. Behrens and W. Aquila, Z. Anorg. Allg. Chem. 356, 8 (1%7). H. Behrens. A. Jungbauer, and M. Moll, Z. Nururforsch. Teil E 32, 1222 (1977). H. Behrens, R. Huller. A. Jungbauer, P. Merbach, and M. Moll, Z. Nafurjorsch. Teil E 32, 1217 (1977).