J. Phys. Chem. 1986, 90, 583-588 situation. Although it is our plan to further explore the utility of eq 2 within a polyatomic molecule framework, we feel that the observations relating eq 2 to the Landau-Zener picture as well as the physically clear picture or photon absorption afforded by eq 2 or 19 constitute the essential results of this paper.
583
Acknowledgment. We acknowledge the support of the National Science Foundation (Grant No. 8206845) and the US.Army Research Office (Grant No. DAAG2984KOO86). We also acknowledge the Harris Corp. for their generous computer system grant.
ESR of V(CO), ( n = 1 to 3) Molecules in Rare-Gas Matrices R. J. Van Zee, S. B. H. Bach, and W. Weltner, Jr.* Chemical Physics Center and Department of Chemistry, University of Florida, Gainesuille, Florida 3261 1 (Received: July 22, 1985)
Three vanadium carbonyls, two with high spin, were observed and partially identified from hyperfine interactions with 51V ( I = 7/2) and I3C ( I = 1/2). They were formed by the reaction of vanadium metal vapor with low concentrations of I2CO or I3CO in rare gases and condensed at 4 K. VCO is trapped in two conformations of almost equal stability, one linear and the other presumed to be slightly bent; both have S = 5 / 2 . The bent form (6A') has an approximate configuration [
[email protected](pa + d ~ ) ~and. ~a ]zero-field splitting parameter ID1 = 0.45 cm-I. The electronic parameters changed considerably in the linear form. V(C0)2 was observed only in a neon matrix and has a S = 3 / 2 ground state with ID1 = 0.30 cm-I. It is probably slightly bent, but the ESR spectrum provided no definite evidence of nonlinearity. V(CO), was also observed only in neon as an axial molecule with a 2A1'or 2Al ground state depending upon whether it has planar Dgh(considered more probable) or pyramidal C3, symmetry.
Introduction Transition-metal carbonyl molecules continue to be of great interest, partially because of their relevance to catalysis. The simplest molecules, those containing only one metal atom, have been studied spectroscopically, and electron spin resonance (ESR) has been applied successfully in some cases, specifically to V(CO)4, V(CO),,' V(CO)6,2-5Mn(C0)s,6 Co(CO),, C O ( C O ) ~ ,CuCO, ~*~ C U ( C O ) ~ ,and ~ J ~AgCO, Ag(C0)3.11,12(Ionic carbonyls have also been observed via ESRI3,l4but will not be explicitly discussed here.) Theoretical discussions of the geometries, ground states, and bonding in these types of molecules have been given by several authors beginning perhaps with KettleIs and then by DeKock,I6 B ~ r d e t t , ~ ~Elian * ' * and Hoffmann,19 and Hanlan, Huber, and O z h z o Although a number of ab initio calculations have been made on such carbonyls, the vanadium molecules considered here J. R. Morton and K. F. Preston, Organometallics, 3, 1386 (1984). D. W. Pratt and R. J. Myers, J . Am. Chem. Soc., 89, 6470 (1967). K. A. Rubinson, J. Am. Chem. Soc., 98, 5188 (1976). M. P. Boyer, Y.LePage, J. R. Morton, K. F. Preston, and M. J. Vuolle, Can. J . Spectrosc., 26, 181 (1981). ( 5 ) S. W. Bratt, A. Kassyk, R. N. Perutz, and M. C. R. Symons, J. Am. Chem. SOC.,104,490 (1982). (6) J. A. Howard, J. R. Morton, and K. F. Preston, Chem. Phys. Lett., 83, (1) (2) (3) (4)
226 (1981). (7) L. A. Hanlan, H. Huber, E. P. Kiindig, B. R. McGarvey, and G. A. Ozin, J . Am. Chem. Soc., 97, 7054 (1975). (8) S. A. Fairhurst, J. R. Morton, and K. F. Preston, J. Magn. Reson., 55, 453 (1983). (9) G. A. Ozin, Appl. Spectrosc., 30, 573 (1976). (10) P. H. Kasai and P. M. Jones, J . Am. Chem. Soc., 107,813 (1985). (11) D. McIntosh and G. A. Ozin,J . Am. Chem. Soc., 98, 3167 (1976). (12) P. H. Kasai and P. M . Jones, J. Phys. Chem., 89, 1147 (1985). (13) T. Lionel, J. R. Morton, and K. F. Preston, J. Chem. Phys., 76,234 (1982). (14) S. A. Fairhurst, J. R. Morton, and K. F. Preston, Chem. Phys. Lett., 104, 112 (1984). (15) S . F. A. Kettle, J. Chem. SOC.A , 420 (1966); Inorg. Chem., 4, 1661 (1965). (16) R. L. DeKock, Inorg. Chem., 10, 1205 (1971). (17) J. K. Burdett, J . Chem. Soc., Faraday Trans. 2, 70, 1599 (1974). (18) J. K. Burdett, Inorg. Chem., 14, 375 (1975). (19) M.Elian and R. Hoffmann, Inorg. Chem., 14, 1058 (1975). (20) L.Hanlan, H. Huber, and G. A. Ozin, Inorg. Chem., 15,2592 (1976).
0022-3654/86/2090-0583$01.50/0
amarentlv have not been treated in detail. The baikground for the present investigation was provided by the matrix work of Hanlan, Huber, and OzinZ0who observed the infrared spectra of V(CO), where n = 1 to 5 , in the solid rare gases. Most notably, those authours concluded, from experiment and theory, that (1) VCO is nonlinear, (2) V(C0)2 exists in linear, cis, and trans forms in all three matrices, argon, krypton, and xenon, (3) V(CO)3 is probably of D3,,trigonal planar geometry. It should be emphasized that the supporting theory usually assumed low-spin ground states. Morton and Preston have prepared V(CO)4 and V(CO)s in krypton matrices by y irradiation of trapped V(CO)6.1 From ESR they assign V(CO)4 as high-spin 6Al in tetrahedral ( Td)symmetry and V(CO)5 as 2B2with distorted trigonal bipyramid (C,) symmetry. V(CO)6 is a well-known stable free radical which has been rather thoroughly researched by infrared,21MCD,22ultravi~let?~ electron, and X-ray d i f f r a ~ t i o nand , ~ ~ ESR. It is presumably a Jahn-Teller distorted octahedral (2T2g)molecule at low temperatures leading to a 2B2, ground state. Our ESR findings are only for V(CO),, where n = 1 to 3, and are not always in agreement with conclusions from optical work and semiempirical theory. The most explicit departure is in finding that VCO and V(C0)2 are high-spin molecules. a 1
Experimental Section The vanadium carbonyls synthesized in this work were made in situ by co-condensing neon (Airco, 99.996% pure), argon (Airco, 99.999% pure), or krypton (Airco, 99.995% pure) doped to 0.1-5 mol % with l2C0 (Airco, 99.3% pure) or I3CO (Merck, 99.8% pure) with vanadium metal [99% pure, 99.8% W ( Z = 7/2)] onto a flat sapphire rod maintained at 4-6 K but capable of being annealed to higher temperatures. (21) H. J. Keller, P. Laubereau, and D. Nothe, Z . Naturforsch. B, 24,257 (1969); H. Haas and R. K. Sheline, J . Am. Chem. SOC.,88, 3219 (1966). (22) T. J. Barton, R. Grinter, and A. J. Thomson, J . Chem. SOC., Dalton Trans.. 698 (1978). (23j G. F: Holland, M. C. Manning, D. E. Ellis, and W. C. Trogler, J. Am. Chem. SOC.,105, 2308 (1983). (24) D. G. Schmiddling, J. Mol. Struct., 24, 1 (1975); S. Bellard, K.A. Rubinson, and G. M. Sheldrick, Acta Crysrallogr., Sect. B, 35, 271 (1979).
0 1986 American Chemical Society
584
Van Zee et al.
The Journal of Physical Chemistry, Vol. 90, No. 4, 1986
'1
vco (A) /ARGON
TABLE I: Observed and Calculated Line Positions (in G ) for VCO (X62) in Conformation (A) in Areon at 4 K ( Y = 9.5596 GHz)
VCOIA)
XY I
M,(51V)0
1
'I2 I I
calcd
extra line (8 = l o o ) obsd calcd
5082 5170 5262 5360 5463 5573 5689 5814
7906 (20) 7897 8015 8012 8119 8124 8252 8234 8346 8344 8467 8453 8562 867 1
e=ioo
l 77
(
( 79
(
I 83
1 81
l
l 85
'I2 'I2 -'/2 -3/2 -5/2
obsd
XY 3
calcd
obsd
797 (5) 789 5065 (20) 877 5154 882 97 1 5254 974 1072 5364 1072 1176 5473 1174 1285 5584 1282 1400 5692 1396 1519 5819 1511
AMf = f l transitions 7944 7976 8054 8083 8161 8192 8252 8304 8380 8423 8514 I 08
l
l
1
IO
1 12
1
I
I
I4
H(KG)
-
Figure 1. ESR spectrum of an unannealed argon matrix at 4 K containing 51VCO(A),with hfs of 100 G, and slVCO(a), with hfs of -60 G. For conformation (A) two perpendicular lines and an off-principalaxis line are shown. Y = 9.5585 GHz.
The furnace, Heli-Tran, and IBM/Bruker X-band ESR spectrometer have been previously described.25 Vanadium was vaporized from a tungsten cell at 1975 OC, as measured with an optical pyrometer (uncorrected for emissivity).
gll
g,
IAIl("V)] lAL(5'V)l
Derived Parameters 2.002 (37) ID1 Aiso(51V)a 1.989 (5) Adip("V)' 247 (28.) M H z IAL("C)I 288 (6) MHz
(25) R. J. Van Zee, C. A. Baumann, and W.Weltner, Jr., J . Chem. Phys., 82, 3912 (1985).
0.603 (2) cm-I 274 (13) MHz -14 (11) M H z 17 (3) MHz
"Assuming All and A , are positive. TABLE 11: Observed and Calculated Line Positions (in G ) for VCO (X62) in Conformation (a) in Argon at 4 K (v = 9.5596 GHz) XY I
ESR Spectra VCO. Two ESR spectra of the VCO molecule were observed in matrices prepared by condensing V into CO/argon mixtures at 4 K. We designate these two forms of VCO below as (A) and (a). This symbolism is derived from one of their distinguishing features: one has a considerably larger 51Vhyperfine splitting (hfs) than the other. Only the (a) form survived after annealing the argon matrices and only it appeared in a Kr matrix. Only (A) was observed in solid neon. 51VCO(A)and 51VCO(a)in Argon. Upon d e p i t i n g vanadium metal into an argon matrix doped with 1.0 mol % I2C0, we obtained the 4 K ESR spectrum shown in Figure 1. The two sets of eight strong, sharp lines centered near 1200 G could be attributed to separate species since upon annealing one set [designated by (A)] disappeared. The hyperfine splitting (hfs) in the perpendicular x y , and xy3 lines of the (A) species due to 51V(Z= 7 / 2 ) is approximately 100 G, whereas that in the (a) species is -60 G. The line centered at about 8100 G has been observed with t h a t intensity only once, b u t its appearance, a n d disappearance upon annealing, correlates best with the (A) molecule. Its complex hfs is indicative of an off-principal-axis line where forbidden Am, # 0 transitions can also occur. The observed lines of both (a) and (A) are listed in Tables I and 11. Annealing to 16 K and quenching to 4 K converted the VCO(A) species to (a) which has the spectrum in argon in Figure 2. Again the xyl and xy, lines have the same hfs, now -60 G, and an "extra" line appears but centered at about 6700 G. 5'v'3CO(A)and 5'V"3C0(u)in Argon. These same spectra can be observed when I3CO replaces I2CO and the effect upon the xy, line, which is the same effect for (A) and (a), is shown in
7938 7970 8052 8083 8164 8194 8274 8304 8384 8413 8493 8521 8603 8630
M,('IV)'
3/2
'12
-'/2 -5/2
-'I2
XY 3
extra line (8 = 12')
obsd calcd obsd calcd obsd calcd 940 ( 5 ) 940 4460 (20) 4461 6449 (20) 6448 1000 999 4520 4520 6530 6526 1061 1060 4581 4580 6603 6603 1124 1124 4645 4644 6684 6680 1191 1190 4710 4709 6756 6755 1258 1257 4777 4777 6828 6831 1326 1327 4847 4847 6906 6907 1396 1398 4918 4921 6983
AMf = f l transitions 6476 6502 6563 6590 6637 6665 671 1 6782 6804 6857 6887 6934
gll
g,
IAll("V)l lAL(slV)l
Derived Parameters 2.002 (10) ID1 1.998 (3) Ai,(51V)" 165 (14) M H z Adip("V)' 183 (1) MHz IA,(13C)I
6479 6496 6556 6573 6633 6650 6709 6726 6785 6802 6861 6877 6937 6953
0.452 (2) cm-l 177 ( 5 ) MHz -6 ( 5 ) M H z 17 (3) MHz
"Assuming All and A , are positive.
Figure 3. Each line is split into a doublet separated by about 6 G, indicating most importantly that there is only one C O in each species.
ESR of V(CO), in Rare-Gas Matrices
The Journal of Physical Chemistry, Vol. 90, No. 4, 1986 585 6
1 VCO(A)/NEON
I
I
I
53
I
I
L 64
I
I
I
I 46
44
I
I 59
I
57
1
I
I
I 68
66
I
55
70
I
J
I
48
H(KG)
50
Figure 4. ESR lines in a neon matrix at 4 K attributed to 51VC0in conformation (A). Y = 9.5560 G H z .
09
II
13
I 6.I
H(KG)
I
1
I
6.3
Figure 2. ESR spectrum of an annealed argon matrix at 4 K containing Y
I
I
6.7
H (KG)
only 51VC0in conformation (a). Two perpendicular lines and an off-
principal-axis line are shown.
I
6.5
= 9.5585 G H z .
LINE
I
I
1.5
I
I
I .7
I
I
I.9
I 2.1
H (KG) Figure 5. ESR lines in a neon matrix at 4 K attributed to 51V(CO)2.Y
= 9.5560 G H z . TABLE III: Observed Line Positions (in C) for VCO (X62)in Conformation (A) in Neon at 4 K (Y = 9.5560 GHz)
MA5'V)
XY I
MpV)
XY 1
'I2
796 (5) 885 976 1074
-'I2 -312
1177 1286 1398 1514
2
I 90
I
I
I
I1
I 13
I
H(KG) Figure 3. ESR spectrum of the perpendicular x y l line of 5 1 V 1 3 C 0in conformation (a) in an argon matrix at 4 K. Y = 9.5531 G H z .
5'VCO(A)in Neon. In neon only one VCO molecule appears to be trapped, the one designated as (A) in argon with the hfs of 100 G. The ESR spectrum when vanadium was trapped in neon doped with 0.1% ' T O is shown in Figure 4 and the observed lines are listed in Table 111. Again the xy, line is centered at about 1200 G, but the center of the xy3 fine-structure line is difficult to determine. This is because the intensity of the spectrum is lower than in argon, and the lines appear to be split by site structure which is apparently more exaggerated in the high-field line. The off-principal-axis line was not detected in neon. Annealing is difficult in neon matrices, and it only led to loss of the matrix here. At CO concentrations higher than 0.1%this spectrum was not observed, presumably because the higher carbonyls were
-
3 ~ 2 I12
-51~
-'I2
readily formed in this lower-melting solid. 51VCO(a)in Kr. Before annealing there are two series of eight lines of about equal intensity, each with -60 G hfs, one centered at 1 138 G and the other at 1170 G at 4 K. When the sample was annealed to about 30 K, only the series centered at about 1170 G (extending from 967 to 1392 G) remained sharp at 4 K. 51V(12CO)2 and 51V(13C0)2 in Neon. With C O / N e concentrations of 0.1%,in addition to VCO(A), two additional finestructure lines appeared centered at about 1800 and 6400 G with 51Vhyperfine splittings of -60 G. This spectrum, assigned to V(CO), molecules, is shown in Figure 5 and the observed lines are listed in Table IV. With incorporation of I3CO, the line widths appear to be almost the same, perhaps broadened by no more than 20%. As with the monocarbonyl this spectrum is not observed in neon at higher C O concentrations. It has also not been observed in argon with any of the C O concentrations used, which varied from 0.1 to 5%, nor has it been formed by extensive annealing of any argon matrix.
-
-
586 The Journal of Physical Chemistry, Vol. 90, No. 4, 1986 TABLE I V Observed and Calculated Line Positions (in C) for V(CO)* (X4Z,)Isolated in Neon at 4 K (v = 9.5560 CHz)
obsd
'I2
1586 (5) 1646 1706 1766 1830 1896 1960 2030
5/2 3/2
'I2 -'I2 -312 -s12
-'I2
A ' VCOIJNEON 91 I
XY3
XYI
MI(s'V)a
Van Zee et al.
calcd 1586 1646 1707 1769 1833 1898 1965 2032
obsd
calcd
6136 (10) 6199 6259 6324 6387 6452 6522 6597
6140 6198 6259 6321 6385 645 1 6519 6589
Derived Parameters 1.9908 (9) 0.2995 (5) cm-I 132 (56) MHz 178 (3) MHz
Aiso(slV)' Adip("V)'
IA,(13C)I
163 (21) MHz -15 (20) MHZ C42 MHz
"Assuming All and A, are positive. TABLE V Calculated and Observed Line Positions and Magnetic Parameters of V(CO)3 Molecule in a Neon Matrix at 4 K ( u = 9.55498 GHz)
perpendicular lines, G obsd calcd
Mr"
&'I2
3067 (3) 3117 3167 3218 3269 3321 3373 3425
fS/,
&'I2
r1I2 +I2
3068 3118 3168 3218 3269 3321 3373 3425
I
3395 3411 3433 3451 3472 3495
(?)
(5) (5) (3)
4
loo
3356 3373 3391 3410 3430 3451 3472 3495
gll
IA11(51V)l
2.1024 (9) 1.9923 (9) 55 (3) MHz
IA,(S'V)I
lAi,(slV)(" JAdip(slV)I'
I
I 32
I
I
34
i 36
SIMULATION
-40
150 (3) MHz 82 (3) M H z 68 (2) M H z
'Assuming A,l and A, are of opposite sign.
51V('2C0)jand 51V(13C0)3 in Neon. Using CO concentrations of 0.1 to 1.O% in neon gave a complex signal extending from 3000 to 3500 G containing a pattern of eight strong hyperfine lines centered about g = 2.12 with A(51V)r 50 G (see Figure 6 and Table V). There are also two very weak signals at 2948 and 3008 G (not shown) and two at 3560 and 3643 G which appear clearly in Figure 6. These weaker lines did not appear to have the same intensity variation from matrix to matrix as the strong features. Using I3COessentially doubled the width of the eight strong lines. The region between about 3430 and 3530 G becomes approximately one broad line with phase down, indicating a broadening of each of the narrow I2COlines by a factor of at least three. Even with extensive annealing this spectrum was not observed in argon matrices. Analysis VC0, ( A ) and ( a ) . The low-field fine-structure line of each of these molecules has an effective f = 6 and each then may be assigned a 6L: ground state. This is in accord with the further analysis shown in Tables 1-111. The x y , line at about 1100 G in each case corresponds to the transition within the Kramers doublet ( M , = +l/z The other lines in argon diverge further for the two species, indicating that their zero-field splitting (zfs) parameters are quite different; as the tables show these are 0.60 and 0.45 cm-' for (A) and (a), respectively, in argon. Although the hfs in neon indicates that the trapped molecule there is (A), there was insufficient data for analysis. The same was true of the krypton spectra; however in that matrix the hfs identified the trapped molecules as (a), presumably in two sites before annealing. The two perpendicular-line positions in argon yield good values 0 1 and g, for (A) and for (a), and in principle gll could of lb$ = 1 then be determined reliably from the high-field extra lines in each
-
Y
-20
Derived Parameters g,
I
30
parallel lines, G obsd calcd 3358 (?)
t
911
-60
+ +
+ t
+ +
+ t
+ +
-00
t +
+ +
+ +
'1
+ti++++
-100
3100
3300
3500
FIELD STRENGTH (G) Figure 6. (Top) ESR spectrum near g zz 2 in a neon matrix at 4 K attributed to an axial 51V(CO)3molecule. u = 9.5584 GHz. (Bottom) Simulated spectrum using g, A(slV) parameters and line widths given in
the text. case if the b: parameter is neglected in the spin Hamiltonian for these presumably linear molecules:26 7f = g,lPff,Sz + g , P W J * + ffpJ+ b W , 2 - (1/3)S(S + l ) ] AIlI,S, + A,(ZJ, ZpY) + (1/60)bs[35S,4 30S(S 1)s: 25s: - 6S(S + 1) + 3S2(S 1)*]
+
+
+
+
+
where b! = D and the hyperfine parameters here refer to 5'V(Z = 7/2). b: parameters have been determined for at least two axial high-spin m o l e c ~ l e s and, ~ ~ ~although ~* they were small, they had a pronounced effect upon the positions of high-field lines. This may not be the case here but it makes the establishment of g,, uncertain. However, neglecting 6,: one finds that gll values of 2.039 (A) and 2.012 (a), with the admittance of forbidden AM, = &1 transitions, gave reasonable fits to the lines. For S = 5 / 2 , I = 7/2, the above Hamiltonian generates a 48 X 48 spin matrix which was diagonalized to obtain eigenvalues and thereby magnetic parameters which best fit the observed transitions at the resonance frequency. If interaction with the quadrupole moment of the 51V nucleus (-O.05b) is involved then the coupling constant must be less than 14 (A) and 6 MHz (a). These Ag,, values are large, and (26) See W. Weltner, Jr., "Magnetic Atoms and Molecules", Van Nostrand-Reinhold, New York, 1983, for a discussion of the spin Hamiltonian, zero-field splitting parameters, and basic references. (27) R. J. Van Zee, R. F. Ferrante, K. J. Zeringue, and W. Weltner, Jr., J . Chem. Phys., 75, 5297 (1981). (28) L. B. Knight, Jr., R. J. Van Zee, and W. Weltner, Jr., Chem. Phys. Lett., 94, 296 (1983).
ESR of V(CO), in Rare-Gas Matrices it seemed more reasonable to see if a fit could be made with gll z g,. This necessitated a slight alteration in the All hyperfine parameter, which is very uncertain in any case, but the approximate fit of the “extra-line” transitions shown in Tables I and I1 could be made. This leaves the gll and b: parameters in considerable doubt. There is no doubt about both of these molecules being VCO since the 51Vand 13Chfs clearly establish that. The spectra appear to be those of linear molecules since the perpendicular lines are narrow and any splitting would have to be very small. However, as can be seen from the lines at 1200 G in Figure 1, the widths of the VCO(a) lines, while still quite narrow (-8 G), are about twice as wide as those of the (A) form. This may be significant in suggesting an unresolved splitting in the (a) molecule lines (see below). It is also possible that the molecules are bent but appear linear because they are rotating about axes of least moment of inertia. However, the similarity of the spectra in three matrix environments may be evidence against such motional averaging. V(CO)2. The same Hamiltonian applies to a linear V(C0)2 Again as was used for VCO except that S = 3/2 instead of neglecting b:, one solves a 32 X 32 eigenvalue problem to arrive at the parameters given in Table IV. With only two perpendicular fine-structure lines observed, only g, and 1 0 1 could be derived. V(CO)3.Because of the complexity of the spectrum in Figure 6 near g = 2.0, a simulation program for S = ‘I2utilizing second-order perturbation theory29 was employed to derive the magnetic parameters. The final fit shown in the bottom of that figure used the following parameters: g, = g,, = 2.1 189, g, = 2.0057; A, = A, = 148.0 MHz, A , = 56.0 MHz; line widths W, = W,, = 25 MHz, W, = 12 M H z for Lorentzian line shapes; v = 9.5584 GHz. To further confirm these derived parameters, exact diagonalization of the matrix was done, yielding almost the same results, as shown in Table V. There was concern about the weak lines at low and high fields mentioned earlier in that perhaps they were part of the V(CO)3 spectrum and implied a nonaxial molecule. However, attempts to simulate the spectrum with all g and all A components distinct were not successful. Also, the observation that the intensities of these weak lines appeared to vary relative to the strong ones when the C O concentration was varied implied that the two sets of lines belonged to separate molecules. V 2 C 0 is a possible source of the weak lines, although, from the 3E ground state of V2,30one might expect the carbonyl to also be triplet (or singlet), but a fit to an S = 1 molecule appears unlikely. It is not the spectrum attributed to V(CO)41which is a “derivative-shaped line at g = 1.9583” in Kr, and it did not grow but diminished in relative intensity with increasing CO/Ne concentration.
The Journal of Physical Chemistry, Vol. 90, No. 4, 1986 587
VATOM
-
Discussion VCO. Both molecules observed at g‘ z 6 in unannealed argon matrices are established as VCO by the 5iV and I3C hyperfine structure observed. The sharpness of the lines and the lack of additional features indicate that their g tensors are nearly axial and the molecules are therefore close to linear. Trapping in neon and krypton provides no evidence to the contrary and makes the possibility of axial spectra due to molecular rotation, as mentioned earlier, unlikely. V atoms have a 3d34s24Fground state but the 3d44s 6D state lies only about 2000 cm-’ higher.31 Thus, it is not difficult to justify S = 5 / 2 for the VCO molecule. The ten valence electrons of C O plus the five of vanadium fill the levels, as indicated in Figure 7, such that the 30 and 1~ provide the shared orbitals. (Figure 7 was derived from a similar figure for NiCO given by (29) R. L. Belford and M. J. Nilges, ”Computer Simulation of Powder Spectra”, EPR Symposium, 21st Rocky Mountain Conference, Denver, CO, Aug, 1979; M. J. Nilges, Ph.D. Thesis, University of Illinois, 1981; T. E. Altman, Ph. D. Thesis, University of Illinois, 1981; A. M. Maurice, Ph.D. Thesis, University of Illinois, 1982; E. P. Duliba, Ph.D. Thesis, University of Illinois, 1983. (30) P. R. R. Langridge-Smith, M. D. Morse, G. P. Hansen, R. E. Smalley, and A. J. Merer, J . Chem. Phys., 80, 593 (1984). (31) C. E.Moore, Natl. Bur. Stand. US.Circ., No. 467 (1949).
z5 +-----@ IC
+-----p
Figure 7. Molecular orbital scheme for the 6 2 VCO molecule (modeled after Figure 5-43 in DeKock and Grayz6).
DeKock and Gray.32) The five unpaired spins are then essentially 3d orbitals on vanadium with some hybridization of the 3du with the 4su and 4p0, and with some small population of C O orbitals. This orbital picture is corroborated by the observed hfs, as will be discussed below. It is also in accord with the energy level scheme described by Hanlan et a1.,20but here corrected to the high-spin case. The unusual aspect of the VCO molecule is the appearance of two forms of the molecule in argon, designated as (A) and (a) here, with only form (A) appearing in neon and only form (a) 0 1 = 0.60 cm-I and IA1(51V)l in krypton. (A) is characterized by 1 = 288 MHz and (a) by 1 0 1 = 0.45 cm-’ and IA,(51V)l = 183 MHz, as given in Tables I and 11. The derived values of Ah(51V) in the two cases, although less exact, also lie in that order: for (A) it is 274 MHz and for (a) 177 MHz. For a (d?r)2(dS)2(su)1 configuration one would derive from 51Vatomic data33Ah = 1/5 X 4165 = 833 MHz, and comparison with the two above values yields the unpaired u electron in VCO as 33% (A) and 21% (a) s character. The remainder of the u character in each case is then 3d and 4p, and the values of Adip can also be considered in this way. Since (ud, a d ) and bd contributions to Adip have opposite signs, the small value of that parameter in both VCO molecules can then be accounted for. For example, for molecule (A) and assuming su + pu hybridization &,(A)
[ ”( - ”) +I 5 7
’1
1 X 0.67 X - 438 = -2 MHz 5 5
where 438 MHz is the atomic radial factor,27as compared to the observed value of -14 (1 1) MHz. Only A1(I3C) z 6 G was definitely observed but there is no indication of distinct All(13C)splittings (see Figure 3) so that the A(13C) tensor of VCO may be assumed to be approximately isotropic. Then, accounting for the 2 s spins, one finds that the s character at 13C is only 5 X 6 X 2.8/3777 = 0.02, where the atomic Aisofactor for I3C is from Morton and Preston’s table.33 (Corresponding values in CuI3CO and Agi3C0were ps = 0.05,’O and p s = 0.01, pp = 0.02,12 respectively.) Spin density on the C O ligand is then apparently quite small, although of course measurements of the hfs at the oxygen nucleus are needed to confirm that presumption. The effective distortion of these -3d5 electrons from spherical symmetry leads to both the large zfs parameters and the shifts in the g components from g,. If the orbital angular momentum (32) R. L.DeKock and H. B. Gray, “Chemical Structure and Bonding”, Benjamin/Cummings, Menlo Park, CA, 1980, Figure 5-43, p 329. (33) J. R. Morton and K. F. Preston, J . Mag. Reson., 30, 577 (1978).
588 The Journal of Physical Chemistry, Vol. 90, No. 4, 1986
in the ground state is coupled to the same excited states to produce ID1 and g, then one should find that34
where X is the spin-orbit coupling constant of the vanadium atom with S = 5 / 2 . X is difficult to estimate for a d’ vanadium but might be taken the same as for Cr’ as 190/5 = 38 cm-’ 35 so that
D = f19[2.002 - 1.9891 = f0.25 cm-l as compared to 0.60 cm-’ observed. The idea that VCO(A) and (a) might be isomeric VCO and VOC can be dismissed because A,(13C) is the same for both forms. One would expect quite different spin densities on 13Cfor these isomers. Of course, the ESR spectra, particularly in the absence of spin density information on oxygen, cannot eliminate the possibility that VOC and not VCO is being observed, but bonding considerations indicate that the metal-carbon bond is favored. Earlier it was noted that the line widths differ for the two forms in Figure 1. The increase in line width of (a) could be due to a slight bending of VCO in that form. This implies that a third g tensor component and a second zero-field parameter, E , are strictly required in the analysis of the VCO (a) spectra, but the effect is too small to make any significant change in the analysis. However, if (a) is a bent form of (A), the electronic parameters are very sensitive to angle, since the zfs and A(I5V) are considerably lowered and the g tensor made more isotropic. [This 0 1 and g, - gl,is in accord with eq 1.1 concomitant decrease in 1 A decrease in su character in the bent molecule is not unexpected since it implies decreased su pu hybridization which in the linear molecule helps to lower the energy by placing some unpaired spin density on the side of the metal atom away from the CO. Then one can understand the detection of two forms of VCO if the V-CO potential energy curve is relatively flat with two shallow, but apparently distinct, minima. A rationale for the preference of particular forms (A) or (a) for particular matrices or matrix conditions might be given as follows: If (A) contains the more distorted vanadium, as judged by the larger s hybridization and larger zfs, then it is understandable that it would be the form found in unannealed strained sites in argon and in neon where the sites are smaller and less accommodating. Upon annealing, the (a) form becomes the preferred species in argon, presumably now being surrounded by a more relaxed environment. It is then reasonable that this form would also be found in solid krypton where the sites are larger than in argon. If asked which form would VCO take in the hypothetical gas phase at 4 K, one would then choose (a), the bent 6A molecule. If the V-CO bond is relatively weak, then the C O stretching frequency would be expected to be relatively high, as Hanlan et a1.20have observed. However, in an unannealed argon matrix one might also expect to observe two C-O stretching frequencies from VCO(A) and (a). These apparently were not observed by those authors, but perhaps this could be due to the difference in the experimental conditions during preparation of the matrices:0 since the sapphire rod was always cooled by liquid helium in these ESR studies, or perhaps to the low concentrations of C O used here. V(CO),. The molecule observed here definitely contains only one vanadium atom, but the number of attached CO’s is not definite since only line broadening was observed when 13C0was substituted. However, the addition of one C O molecule to S = s / z VCO can be expected to lower the spin to S = 3 / 2 . The only possible evidence of any nonlinearity in the molecule is the small splittings of about 10 G on each of the xy hyperfine lines in neon (see Figure 5 ) , but this could also be due to multiple sites in the matrix. With the lower spin, the molecule also has lower values of ID1 = 0.30 cm-I and Ais,(5iV) = 160 MHz compared to the monocarbonyl. pEat the vanadium nucleus is calculated to be 160
+
(34) B. R. McGarvey, ‘Transition Metal Chemistry“, Vol. 3, Marcel Dekker, New York, 1966, p 118. (35) T. M. Dum, Trans. Faraday Soc., 57, 1441 (1961).
Van Zee et al. 3/4165 = 0.12, considerably smaller than in VCO. Although the ESR spectra are not definite, it seems very likely from the observed quartet multiplicity that the molecule is at least applies slightly bent. If the usual u bonding and a ba~k-bonding~~ to each attached C O so that a completely double-bonded structure is obtained, then one can only assume the unpaired spins to be essentially (dQ3which would yield only a doublet state. Bending the molecule avoids this problem since it removes the degeneracies in the d orbitals. Then the 10 G splitting in the lines could be due to a small zero-field splitting E term. If the bent symmetry is Cz, then the ground electronic state might be 4Bz. Our observations are not in agreement with the IR assignments in Ar, Kr, and Xe matricesz0where the presence of three forms of the molecule was inferred. V(C0)3.As with V(C0)2, the assignment of the spectrum to a tricarbonyl is somewhat arbitrary since 13C0substitution only led to broadening of each of the 51Vhyperfine lines. However, the molecule does contain only one vanadium atom, and the spectrum does not correspond to any of those observed for the higher carbonyls.’-’ There is no ambiguity in the analysis of the spectrum as that of a doublet axial molecule with Agl large and positive. The parameters are reminiscent of a d9 Cuz+ ion with the unpaired spin in a d,i orbital in a distorted ~ c t a h e d r o n . ~A~ similarly large g tensor anisotropy is found for the linear CuF2 molecule.37 Here, of course, the molecule is considered as planar D3*or pyramidal C3, with its lowest state being ZA1’or zA,. The addition of four more C O electrons to VCO has resulted in lowering the spin to S = No matter what the relative signs of All and A , are the small percent s character at the vanadium atom (
