IBERS, AND DAVISON 1928 LA P L A C A , HAMILTON, these bonds are single in character. In Rh12(CH3)(P(CsHb)&,22which has a tetragonal pyramidal configuration about Rh with CH3 a t the apex, the Rh-CH3 distance is 2.081 (9) A. If we take the normal covalent then a covalent radius of Rh of radius of C as 0.77 1.31 A may be derived. If one assumes a normal covalent radius of S of 1.04 then one predicts the length of a Rh-S single bond to be 2.35 A,some 0.1 A shorter than observed here. However, if one takes the radius of S in the SOz- ion to be half the S-S distance in the dithionite ion, namely, 1.20 A, then one predicts a Rh-S02- distance of 2.51 8,closer to that observed here. It would thus appear that the rationalization presented in the Introduction is a useful one in describing the varying geometries exhibited by NO and , 5 0 2 in transition metal complexes. However, it is far from clear why NO and SO2 may act either as Lewis acids or Lewis bases and why in their attachment to the MX(CO) (P(CeH5)& systems (M = I r or Rh, X = halogen) they should exhibit a Lewis acid character. This be(30) L. Pauling, "The Nature of the Chemical Bond," 31d ed, Coinell University Press, Ithaca, N. Y., 1960.
CONTRIBUTIOS FROM
THE
Inorganic Chemistry havior must certainly be contrasted with the bonding of trans-1,2molecular ~ x y g e n , tetracyan~ethylene,~~ ~~'~~ dicyanoethylene,31 and tetrafl~oroethylene~~ in these systems, where the a-bonding scheme of Chatt and Dewar seems most applicable. Of even more interest, the contrast between the structures of IrCl(C0)(NO)(P(C,&)&+ and the formally isoelectronic IrCI(C0)2( P ( C ~ H S )molecule33 ~)~ is striking. As already indicated, in the NO complex the iridium coordination polyhedron is a tetragonal pyramid, with the apical NO group bonded to iridium through an Ir-N u bond; in the bis-CO complex a trigonal-bipyramidal configuration is found, with trans-phosphine ligands and normal Ir-C-0 geometries. Acknowledgment.-We are indebted to Dr. N. C. Payne for the preparation of the crystals used in this study. This work was supported by the National Science Foundation. (31) Lj. Manojlovi&Muir, K. 1%'. Aluir, and J. A. Ibers, Discussions F a v a day Soc., in press.
(32) N. E. Kime, Doctoral Dissertation, Izorthwestern University, 1068. (33) N. C. Payne and J. A. Ibers, t o be submitted for publication.
CHEMISTRY DEPARTIIESTS, BROOKHAVEN SATIONAL LABORATORY, UPTOS, XEW YORK 11973, XORTHWESTERN UNIVERSITY, EVANSTON, ILLINOIS 60201, A S D MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASSACHUSETTS 02138
Nature of the Metal-Hydrogen Bond in Transition Metal-Hydrogen Complexes: Neutron and X-Ray Diffraction Studied of p-Pentacarbonylmanganese Hydride BY SAM J. LA P L ! L C A ,WALTER ~ C. HAMILTON,Z JAMES A. IBERS,3 ASD ALAS DAVISON4 Received February 17, 1969 An X-ray and neutron diffraction study of a new crystalline form of HMn(CO)s, hereafter designated p-HMn(CO)5,is described. The material crystallizes with etght molecules in space group Csh6-C2/cof the monoclinic system in a cell (measured a t -75') of dimensions a = 12.16 (3) A, b = 6.28 (2) A, c = 19.34 (4) A, and p = 91.2 (5)'. The structural details involving the nonhydrogen atoms derived from 596 visually estimated independent X-ray intensities collected a t - 75" and from 398 independent neutron diffraction counter data collected a t - 105' agree within the assigned limits of error. I n addition, a precise position for the hydrogen atom has been derived from the neFtron data. The p-HMn(CO)smolecule deviates only slightly from C4"symmetry. The Mn-C,, bond length is 1.882 (12) A, while the average Mn-C,, bond length is 1.853 (12) A (ax = axial, trans to H; eq = equatorial, cis to H). The equatorial carbonyl groups are bent toward the bond angles range from 94.9 ( 5 ) to 98.2 (5)'. The Mn-C--0 linkages do not deviate hydrogen atom; the C,,-Mn-C,, significantly from linearity. Thus the molecular structures of 01- and p-HMn(CO)aare essentially identical, but the crystal structures differ markedly and these differences are dicussed. The Mn-H distance in p-HMn(CO)j is 1.601 (16) b. This value, the first to be determined with such precision in a parent transition metal carbonyl hydride, is consistent with the usual covalent radii sums and with earlier, less accurate X-ray determinations of metal-hydrogen distances, but it is totally inconsistent with some theoretical studies and the various interpretations of the broad-line nmr spectra obtained from this and related compounds. From the present study it is concluded that the hydrogen atom is a stereochemically active ligand and that the metal-hydrogen distance is that of a normal covalent bond.
Introduction The nature of the metal-hydrogen bond in transition metal hydride complexes has been of considerable in(1) Research carried Out a t Brookhaven National Laboratory under contract with the LJ. S. Atomic Energy Cominission. (2) Brookhaven Xational Laboratory. (3) Noi thvestern University. (4) Massachusetts Institute of Technology.
terest for a number of years. A brief history of the structural aspects Of the subject U p to the end of 1964 is a ~ a i l a b l e . ~As of that time it was clear from crystallographic studies on Various stable hydride Complexes' ( 6 ) J. A. Ibers, A?%%. Rev. P h y s . Chem., 16,3T.5 (1965).
( 6 ) S. J. La Placa, W. C. Hamilton, and J. A. Ibers, Inoig. Chenz., 3, I491 (19641, and references contained therein.
VoL. 8, No. 9, September 1969 that the hydrogen is a stereochemically active ligand. However, reliable data on the metal-hydrogen bond length were sparse. As of that time only the Rh-H bond length in RhH(CO)(P(CsH&' (1.60 (12) A by X-ray diffraction) and the Re-H bond length in the ReHg2- ions (1.61-1.72 A by neutron diffraction) were known. No diffraction data on the metal-hydrogen bond lengths in the parent carbonyl hydrides were available, although the results of theoretical calculations@ and a broad-line nmr studyJ0on these systems were interpreted as support for an extremely short metal-hydrogen bond, perhaps 1.1-1.5 k , in keeping with the original notionll that the hydrogen atom was "buried" in the metal orbitals and was not a stereochemically active ligand. It is perhaps worthwhile recalling that it was not until after an X-ray study on a-HMn(CO)& showed unequivocallys that the molecule possesses essentially Clv symmetry with H as a stereochemically active ligand that the infrared data on this compound could be interpreted12 in a manner consistent with this symmetry. The evidence from the X-ray diffraction study of a-HMn(CO)&and from other studiesI3 was strongly in favor of a metal-hydrogen bond of normal length. Yet, as noted then, it still seemed worthwhile to settle the question beyond any doubt through a neutron diffraction study of HMn(CO)j. That study, described in this present paper, was delayed by a variety of experimental difficulties. I n the interim the need for such a study became even more apparent as broad-line nmr studies on polycrystalline samples of parent carbonyl hydrides14*16gave values which were exceptionally short (1.28 f 0.02 A for MnH in HMn(CO)eand 1.42 f 0.05 A for Co-H in HCo(C0)r). These studies used the Van Vleckls secondmoment expression which assumes that the nuclei are quantized about the direction of the applied magnetic field. This treatment neglects any quadrupole coupling effects a t the 55Mn and 59C0nuclei. However, this approximation has been criticized in two recent c ~ m m u n i c a t i o n s lin ~ ~which ~ ~ quadrupolar coupling a t the metal nucleus was considered. However, both interpretations place relatively short values of 1.42 and 1.44 A on the Mn-H distance in HMn(C0)5. These distances are in remarkable agreement with that of 1.425 f 0.046 A derived from electron diffraction (7) S. J. L a Placa and J. A. Ibers, J . Am. Chem. Soc., 86, 3501 (1963); Acta Crysl., 18, 511 (1965). (8) S. C. Abrahams, A. P. Ginsberg, and K. Knox, Inorg. Chem., 5 , 558 (1964). (9) F. A. Cotton, J . Am. Chem. Soc., 80, 4425 (1858). (10) E. 0. Bishop, J. L. Down, P. R. Emtage, R. E. Richards, and G. Wilkinson, J . Chem. Soc., 2484 (1959). If only intramolecular interactions are assumed, the value of 1.88 i 0.05 A found for the H-H distance calculated from the second-moment data gives a value of 1.33 It 0.05 b for the Fe-H distance in Fe(C0)aHz (assuming cis-octahedral coordination). (11) W. Hieber, Die Ckemie, 66, 24 (1942). (12) L). K Huggins and H. D. Kaesz, J . Am. Chem. Soc., 86, 2734 (1964). (13) L. L. Lohr, Jr., and W. N. Lipscomb, Inoug. Chem., 8, 22 (1964). (14) T. C. Farrar, W. Ryan, A. Davison, and J. W. Faller, J . Am. Chem. Soc., 88, 184 (1966). (15) T. C . Farrar, F. E. Brinckman, T. n. Coyle, A. Davison, and J. W. Faller, Inorg. Chem., 6, 161 (1967). (16) J. H. Van Vleck, Pkys. Res., 74, 1168 (1948). (17) D. L. Van der Hart, H. S. Gutowsky, and T. C. Farrar, J . Am. Chem. SOL,89, 5056 (1967). (18) G. M. Sheldrick, Chem. Contmun., 751 (1967).
P-PENTACARBONYLMANGANESE HYDRIDE 1929 studieslga on gaseous HMn(C0)6.19bUnfortunately, both the published electron diffraction determination and the theoretical approximations made in the nmr studies appear to be incorrect, for the present study provides unequivocal evidence that the Mn-H distance in HMn(C0)5is 1.601 (16) 8. It is, however, interesting to note that a recent theoretical calculationz0predicts a distanceof 1.60 (+0.07) A.
X-Ray Structure Manganese pentacarbonyl hydride was prepared and purified as previously described.z1 Samples were sealed in quartz capillaries of nominal 1-mm diameter, sharply tapered to a point a t the end to facilitate nucleation. Crystals were grown by slo~vcooling from the liquid phase in a stream of cold nitrogen gas using a modification of a previously described apparatus.2z X-Ray data were collected a t -75" using the precession method. On the basis of the observed Laue symk # 2n for metry 2/m and the systematic absences h (hkL) and h # 2n, 1 # 2%for (h01),the structure may be assigned to the space group C2h6 -C2/c or Cs4-Ccz3 of the monoclinic system. The former (centrosymmetric) space group was assumed. Cell constants were determined from precession photographs and are a = 12.16 (3) A, b (unique axis) = 6.28 (2) A, c = 19.34 (4) ,8 = 91.2 (5)' (X(Mo KO) = 0.7107 A). The calculated density for eight molecules in the cell is 1.762 g ~n1-3,~4Although the cell constants are very similar to those obtained in our original X-ray study of HMn(C0)5, the symmetry of the cell is different, and i t is clear that we are dealing with a second crystalline phase. Previously we had found a = 12.18 (2) A, b = 6.35 (1) k , c = 19.20 (3) A, /3 = 93.3 (S)', and space group Czh6-I2/a. All efforts in this present study to grow crystals of the original modification, hereafter designated a-HMn(CO)S,were unsuccessful. Consequently, it was highly desirable to determine the structure of the new form, pHMn(CO):, by X-ray diffraction techniques in preparation for a neutron diffraction study. A crystal was aligned with the (170) direction coincident with the goniometer head axis. The reciprocal 2, 1 were lattice nets hk0, h k l , hk2, hk3, hLl, and h, recorded photographically using Zr-filtered Mo K a radiation, and in this way intensities of 741 reflections, of which 596 are independent, were obtained. The Xray data were all collected from one crystal, but many
+
A,
+
(19) (a) A. G. Robiette, G. M. Sheldrick, and R. S . F. Simpson, ibid., 506 (1968). (b) T h e preliminary gas electron diffiaction value of 1.425 (o = 0.046 b) has now been revised to 1.50 8, and the estimated standard deviation has been assigned a more conservative value of 0.07 A ( A G. Kobiette and G. M . Sheldrick, private communication, Jan 1969). Thus, the revised electron diffraction value-although quite imprecise-is not in essential disagreement with the present neutron diffraction value. I n any case, it would seem quite unreasonable t h a t there is a substantial difference in bond length between the gas and solid phases. (20) W. G. McDugle. Jr., A. I?. Schreiner, and T. L. Brown, J . A m . Chem. Soc., 89, 3114 (1967). (21) A. Davison and J. W. Faller, I n o v g . Chem., 6, 845 (1967). (22) B. Post, R. S . Schwartz, and I. Fankuchen, Rev. Sci. Instr., 22, 218 (1951). (23) "International Tables for X-Ray Crystallography," Vol. I , The Kynoch Press, Birmingham, England, 1952. (24) T h e density of the a form is 1.755. Owing to a typographical error, i t was quoted as 1.70 in ref 6. Thus V a / V o = 1.004.
1930
K
t
FOR5 I.
FCALC
3
0
4
66
67
0
6
32 33
35 3c
19
4,
0 8 0 12
0 18
7, 217 69 13
z
n
2
1
2 2
2 3
2
4
(7
2
6
38 bn 54
46
2 7 2 8 2 11 2 13 2 :4
7n
65 64
2 15 2 16 2 17 2 18 4 0 4
9:
8 4 11 4
6 6 6 6
50 36
54 52 70 58 66 811
62
60
33
?e
I.
13 4R
65 40
'1
3
82 220 79
34 35 35 69 57
db 37
4
21 7v
0
bP
3
24
23
1
V9
5
31'
Pn 28
62
I
4
1
5 8
1 I
: 11
3 3 3 3
H'
26 42
32 15 61 108
21 10 31 46 60
105 31
3P
36
(7
5
45 119
2 6 2 10
3 .1 3 .7
36
36
57
58
47 3n
12
2 17 +-:l 4-10
47 25
46
26
20
5 6
113 95 28
4 .9 I .a 4 -2 4 .I
4"
8
109 27 9b 25
46 54 65
21 25 107
13
41
41
4 1
43
4:
4
23
26 5n 25
1
4
29
1 q 3 10 3 11 3 12 3 13 3 I. 5.10 5 .2 5 0 5 I 5 2 5 5
57 24
2: 49
39 i(=
0 .4
n -2 0
0
0
2
:8l 54 155 25 28 43 62
58 59
18
2-:7
19
13 32n
36 J:?
2'
7-12
28 5, 37 32
: .1 1 .3 : .Z
33
.,
2.16
35
.e
2.11 2 .9
4)
I
.(
33 54
23
17
1
0 I 2
31 53
2 .d 2 .I 2 -6
IS
49
1
3
12
47
2.ll
2 .5 2 .1 2 -1
l45b 37
2 3B
0.14 0.12 0 .R 0 -6
9278
98 2,
48 4: 34 97 247
4
:
I
2 3
I1 65 I5 25 3" 95 36 27
4
5 0
116
1
35
4 4
3:
9 12 I3 1 1' b -2 6 -1 6 2 (I 3
48
5" 46
41 24 26
91 35 25
113 31
50 75
61
52
8n
75
ns 24
1-16 1-15
17 87
7 85 19 72 19
28
:.I4
49
1.13
bl 62 58 30
is12 1-11 1-10
24
80
78 19
I .9
bb
24
1 -0 1 -7
32 131
33 61
1 .2
55 45
36 51
21 16
31
:
0
93
1 1 1
I 2
3
22 49
34 23
1 1
4 6
95
1
7
236
1
0
71
38 b5 70 267 I9 5e 5:
129
.Sb 47
36 68 78 275 I' 59 $7
7
2
9
2 10 2 I1 2 13 2 I4 2 :h 4 2 4 J
60
62
a2
97 129
13
15
77
80
17
38
3 -1 3 0 3 2 3 3
29 97 32
25 97
18
3
4
3
5
47 27 2b 61 35 39 51
17 51 24 66 27 34 35 55
1)
..
3 3
7 8 3 I. 3 I4 5 -3 5 s:
5 5
68
0
Ill
3
24 U.
I-ll
7r 6.
@.IO
2
n.:r
43
81
I
46
2
3 5 5
79 5n 67 22
(25) T h e crystal length w a i greatel' than ilir diameter o I t h e X-lay collim a t o r iiud i i x < i I the Iapeie
