The Challenge of d and f Electrons - American Chemical Society

Chapter 7

Ab Initio Studies of Transition Metal Dihydrogen Chemistry Edward M . Kober and P. Jeffrey Hay

Downloaded by FUDAN UNIV on January 30, 2017 | http://pubs.acs.org Publication Date: June 8, 1989 | doi: 10.1021/bk-1989-0394.ch007

Los Alamos National Laboratory, Los Alamos, NM 87545

Examples of transition metal complexes containing dihydrogen ligands are investigated using ab initio electronic structure calculations employing effective core potentials. Calculated geometrical structures and relative energies of various forms of WL (H ) complexes (L = CO, PR ) are reported, and the influence of the ligand on the relative stabilities of the dihydrogen and dihydride forms is studied. The possible intramolecular mechanisms for H - D scrambling involving Cr(CO) (H ) are investigated by examining the structures and energies of various polyhydride species. 5

2

3

4

2 2

The recent discoveries of a new class of metal complexes involving molecular hydrogen has spawned numerous experimental(l,2) and theoretical (2—5) investigations to understand the bonding and reactivity of these systems. In this paper we discuss recent theoretical calculations using ab initio electronic structure techniques. In the first part of this article we review calculations on W dihydrogen species of d W complexes with emphasis on predictions of structures and energies of various chemical forms and on comparisons with available experimental information. In the second part a mechanistic problem involving scrambling of H / D mixtures to HD by Cr dihydrogen complexes is addressed. In this section we hope to illustrate the role of for theory in helping to distinguish between various reaction mechanisms in transition metal chemistry. Before proceeding to the specific examples of dihydroge ?hfmistry, it is worthwhile to summarize the particular challenges transition metal and actinide compounds present to this type of approach. There is a striking contrast to most compounds of first— and second—row main—group elements where reasonably accurate bond lengths and bond ancles of stable species can be predicted at the SCF Hartree-Fock level with small basis sets and thermochemical quantities can be computed with reasonable accuracy by perturbative techniques to electron correlation (fi). Accurate studies of molecular properties or transition states of chemical reactions are feasible using multi-configuration SCF ( M C - S C F ) and configuration interaction (CI) techniques. In contrast, for transition metal compounds one encounters a e

2

2

0097-6156/89/0394-O092$06.00/0 c 1989 American Chemical Society

Salahub and Zerner; The Challenge of d and f Electrons ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

7. KOBER & HAY

93

Transition Metal Dihydrogen Chemistry


situation where metal ligand distances are predicted to be 0.05-0.25 Â too long at the SCF level of calculation even with very accurate basis sets (Ζ-1Ω) and such approaches as Moller—Plesset perturbation theory can be unreliable for treating electron correlation effects (11,12). The relatively poor description of the ground state by a single configuration and the presence of numerous low-lying electronic states with very different electron correlation energies often requires sophisticated M C - S C F and CI treatments to obtain reliable molecular geometries and bond energies (13). In addition there are additional challenges arising from the sheer number of electrons in transition metal compounds and the relativistic effects which become increasingly important in second— and third—transition series compounds. These latter difficulties have been largely circumvented by the development of relativistic effective core potentials(li) to replace the chemically inert core electrons and to incorporate the relativistic effects on the valence electrons into the effective potential.

Tungsten Dihydrogen Complexes In the molecular dihydrogen species W(CO)3(PR3)2(H2) first characterized by Kubas et al. in both X - r a y and neutron diffraction studies (1,2), the dihydrogen is bonded in τ/ sideways fashion to the d metal center. Over 50 compounds involving many of the transition metals have since been synthesized by various workers. In addition, reanalysis(15) of existing hydrides such as FeH4(PR ) have now been found to be formulated as molecular hydrogen complexes, i.e., as Fe(H )(H)2(PR3)3. In this section we review briefly our previous theoretical calculations(5) on two prototypical dihydrogen complexes W(CO)3(PH ) (H ) and W f P H s M ^ ) . The calculations employed a relativistic effective core potential ( E C P - 1 ) to replace the inner [Xe] (4f ) core on W and a nonrelativistic E C P on Ρ with a flexible gaussian basis to describe the valence electrons of the system. Details of the calculation are given in Ref. 5. Structures 1—3 in Fig. 1 exemplify the modes of H bonding to a W ( C O ) ( P H ) fragment: two sideways bonded (rçS-coordinated) forms (1 and 2) and the end-on bonded (^Coordinated) form (3). Using a rigid W ( C O ) ( P H ^ fragment, the geometries of these three forms of H coordination nave been optimized using Hartree—Fock wave functions. The sideways bonded species (Table I) are found to be stable with respect to the fragments 4 by 17 kcal/mol and more stable than the end-on form, which is bound by only 10 kcal/mol. Little difference in energy is observed between the two-sideways bonded form with the H axis parallel either to the P - W - P axis or to the C—W-C axis. The former orientation is slightly favored, leading to a rotational barrier about the midpoint of the W - H bond of 0.3 kcal/mol. The calculated structure of the lower energy η form (1) shows a slight lengthening (from 0.74 to 0.796 À) of the H - H bond from uncomplexed H with a W—Η distance of 2.15 A . Recent low—temperature neutron diffraction studies show two equal W—Η bonds (1.89 ± 0.01 Â) and with the H lying exactly parallel to the P—W-P axis as predicted by the present calculations and having a Η—Η separation of 0.82 ± 0.01 Â. Although the calculations have correctly described the preferred mode of H binding, there remain some quantitative differences (Table I) between the theoretical and observed bond lengths. These differences are reduced considerably when one employs an effective core potential ( E C P - 2 ) which also treats the outermost 5s and 5p core orbitals of W as valence orbitals. The W - H and Η—Η distances are now calculated to be 1.93 and 0.81 Â, respectively, in much better agreement with the neutron diffraction results. 2

6

3

3

2

3

2

2

14

2

3

3

2

3

3

2

2

2

2

2

2

2

2


94


THE CHALLENGE OF d AND f ELECTRONS

Fig. 1.

Structural forms of W(CO) (PR3)2(H ) species. 3

2


7. KOBER & HAY

95


Table I. Calculated and Experimental W - H and H - H Bond Lengths and Rotational Barriers for W(CO)3(PR3)2(H ) Species


2

Method

R

R(W-H),À

R(H-H),A

Rotational barrier, kcal/mole

SCFcalc. ( E C P - 1 on W)

H

2.153

0.796

0.3

SCFcalc ( E C P - 2 on W)

H

1.932

0.806

1.3

Low-temp X - r a y diffraction

i-Pr

1.95*0.23

0.75*0.16

Low—temp i-Pr neutron diffraction

1.89*0.01

0.82*0.01

2.4

Examination of the Mulliken population analyses for the fragment and the η complex reveals an overall increase of 0.12 e on the W atom upon complexation and the total charge on each hydrogen has decreased slightly from 1.00 to 0.98 e. The σ-bonding orbitals of the W atom (6s, 6p , and 5 d | show a net increase of 0.13 e, while the τ—bonding orbit sus (6p and 5d ) undergo a net loss of 0.03 e. Although the other five ligands also influence the amount of charge on the metal, the above trends are consistent with a mechanism involving some σ-donation from the H2 ligand and a lesser degree of π—back—donation from the metal. Of the possible seven-coordinate dihydrides let us consider the least motion reaction in which the two W—L bonds originally parallel to the H—H axis bend back as two W - H bonds are formed. The energies of these species are compared with the r/S-dihydrogen forms and the fragments in Fig. 2. Both seven-coordinate dihydride species lie higher in energy than the dihydrogen from (17 and 11 kcal/mol, respectively) and are only slightly bound compared to W L + H . For the case of having all P H ligands in the W ( P H ) ( H ) complexes, a much different situation prevails concerning the oxidative addition reaction. In contrast to W ( C O W P H ) ( H ) , the seven-coordinate dihydride W ( P H ) ( H ) lies 3 kcal/mol below the η complex! Replacing the CO ligands by PR3 groups favors the oxidative addition reaction proceeding to completion rather than being arrested in the ^dihydrogen stage. This preference for ^Coordination in W(CO)3(PH ) (H) correlates with the overall stabilization of the 5d orbitals, and the 5 d orbital in particular, by the back—bonding CO ligands. When these ligands are replaced by the less stabilizing PH3 groups, the dihydride is the most favored form. This is consistent with experimental observations that the dihydrogen-dihydride equilibrium can be shifted depending on the basicity of the ligands. For example, in a series of complexes, M o ( C O ) ( R P C H P R ) H , the coordination mode changes from dihydrogen from R = F h to dihydride for the more basic R=Et(2). 2

2

z

y

5

2

3

3

2

3

5

2

2

2

3

5

2

3

2

2

xz

2

2

4

2

2

2


z

yz


V

I/'

o OC

XJ

I

°0. oc υ

H

oc

w

o

20

W + H 1

y

0°

L

ù

(3


(ô) FRAGMENTS L=PR

(0)

LU

(-7)

O υ LU CE

3

(-17)

7] -DIHYDROGEN COMPLEX 2

OCo o DIHYDRIDE COMPLEXES

1

I/

-W + H, υ

20

H

ce

o

(0)

LU Ζ LU

S-

20

(-19)

(-21)

(-33) L

FRAGMENTS L-PRo

η DIHYDROGEN COMPLEX 2

DIHYDRIDE COMPLEXES Fig. 2

Relative energies in kcal/mole of W ( C O ) ( P H ) 2 ( H 2 ) species (above) compared to W ( P H ) ( H ) species (below). 3

3

5

3

2


7.

KOBER & HAY

97


H - D Exchange Involving CrfCO^Dihydrogen Species Background. The gas phase reaction H

+ D -ι 2HD

2

2

occurs only under severe conditions such as in shock tubes with an activation energy (~100 kcal/mole) comparable to the H—H bond energy. In fact, the kinetics have been interpreted in terms of a free radical mechanism involving H atoms rather than the bimolecular process indicated in the above equation. By contrast, several cases of facile H2—D exchange have been observed under thermal or low temperature conditions involving dihydrogen complexes 2

L M H + D -» L MHD + HD η ζ ζ η Downloaded by FUDAN UNIV on January 30, 2017 | http://pubs.acs.org Publication Date: June 8, 1989 | doi: 10.1021/bk-1989-0394.ch007

0

0

Upmacis, Poliakoff and Turner(l£) observed HD exchange for mixtures of Cr(CO)4(H )2 and D but interestingly nol for mixtures ot Cr(CO)5(H ) and D . Kubas et al observe a similar phenomenon where HD is produced from reacting W(CO)3(PR )2(H ) with D either in solution or in the solid state. In addition other cases of H - D scrambling occur readily with metal hydride complexes(U) as in the case of Cp*ScH or Cp2ZrH where Cp* = 2

2

2

2

3

2

2

CsMe . Since the work of Upmacis et al. on Cr(CC>4)(H2)2 complexes is suggestive of an intramolecular mechanism, Burdett et α/.(1£) have examined various polyhydride structures as possible intermediates in this process using extended Huckel theory. These studies have led us to pursue ab initio electronic structure calculations of Cr(CO)4(H )2 species and possible mechanisms leading to H2—D exchange. Implicit in these studies is the assumption that there is rapid equilibrium between 5

2

2

Cr(CO) (H ) + D - Cr(CO) (H )(D ) + H 4

2

2

2

4

2

2

2

which subsequently undergoes intramolecular exchange C r ( C O ) ( H ) ( D ) -, C r ( C O ) ( H D ) 4

2

2

4

2

although this is only inferred from the experimental studies. What is actually observed is Cr(CO) (HD) formation in a mixture of Cr(CO) (D ) and Cr(CO) (D )2 when reacted with H . 5

4

5

2

2

2

Results of ab initio calculations. The calculations have been carried out on stable structures of Cr(CO)4(H )2 or its fragments at the Hartree—Fock level, where the structures have been optimized using gradient techniques with the modified GAUSSIAN82(I£) or the MESA(2Q) electronic structure codes. An effective core potential was used(H) to replace the [Ne] core of Cr with a [3s 3p 2d] contracted Gaussian basis to describe the 3s, 4s, 3p, 4p and 3d orbitals. A flexible [3s lp] bais was used for hydrogen and an S T O - 3 G basis was employed for C and 0 . (Some results are presented using an unpolarized (3s) hydrogen basis.) Of the possible forms for the parent molecule cis—Cr(CO)4(H2)2 the lowest geometrical structure is found to have the H2 molecules oriented in an 2


98


upright position relative to the equatorial plane (Fig. 3). The structure with both H lieands lying in the equatorial plane is 3.1 kcal/mole higher in energy, corresponding to a rotational barrier of 1.5 kcal/mole for the rotation of one H about the metal-H2 bond. Removal of one H 2 to form Cr(CO)4(H ) requires 15 kcal/mole and removal of the second H requires another 14 kcal/mole. At this level of calculation the dihydrogen-dihydride energies 2

2

2

2

Cr(CO) (H ) - Cr(CO) (H) 4

2

4

2

are comparable with the dihydride lower by 1 kcal/mole. The calculated C r - H and H - H bond lengths are 1.787 and 0.77 A respectively for the upright form. (Fig. 4). Some of the possible polyhydride forms—having either a square H 4 or a H 3 — H " species coordinated to Cr(CO)4—are found to be very high in energy (at least 50 kcal/mole) and hence are unlikely intermediates in the H 2 - D exchange reaction. Another pathway is shown schematically in Fig. 5, where the conversion of the bis—dihydrogen complex to a dihydrogen-dihydride complex. The process is symmetry allowed in the sense that the three relevant orbitals in the equatorial plane, the Cr d . and the two H σ orbitals of the bis dihydrogen species transform into the C r - H σ and H σ bonds of the dihydrogen dihydride species. The diagram is oversimplified in the sense that the orbital characters change qualitatively through the course of the reaction. The d - orbital is empty in the dihydrogen dihydride complex and the H bonding orbital actually descends from higher orbitals in the bis—dihydrogen complex. The dihydride form is calculated to lie 10 kcal/mole higher in energy with a C r - H and C r - H bond lengths of 1.729 and 1.923 Â, respectively. The transition state for the reaction


2

4

2

2

x

2

y

2

2

2

x

2

y

2

2

Cr(CO) (H ) HCr(CO) 4

2

2

4

H (H) 2

2

was located at the SCF level assuming C symmetry and treating the H - C r - H bond angle between the two central Η atoms as the reaction coordinate. A n activation barrier of 24 kcal/mole (see Fig. 6) was found for this process—a relatively low barrier compared to some of the other polyhydride species. A non—C2V—pathway was also investigated where a similar barrier was also found. Because of the electron correlation effects can have a significant effect on calculated energies for chemical reactions, calculations on several of the above C r f C O ) ( H 2 ) species were carried out using configuration interaction (CI) techniques. These particular calculations consisted of single and double excitations (SDCI) with respect to the single Hartree—Fock configuration employed in the above studies. The optimized geometries from the SCF calculations were used for the respective species. No excitations were allowed from the inner 16 orbitals corresponding to the carbon Is, oxygen Is and 2s, and chromium 3s and 3p core orbitals. This resulted in 140,642 spin eigenfunctions in C 2 symmetry for the SDCI calculations. 2 v

4

2

V


KOBER & HAY


Cr (CO) (H ) 4

2

2

STRUCTURES AND FRAGMENTS E(kcal/mole)


....

Λ

I

C rr" '

0.0

-..jA

+3.1

Cr

Cr

+H

+15.1

2

Η Cr^.

+H

+14.4

2

Η

Cr

Fig. 3.

+2H

2

+29.4

Relative energies in kcal/mole of Cr(CO)4(H )2 species. 2




OPTIMIZED GEOMETRICAL

PARAMETERS

Bond Lengths (A) Cr - H

2

H - H

1.787

1.772

1.923

0.772

0.779

0.764

Cr - H,

1.729

Cr - C,

1.997

1.980

2.084

Cr - C

1.971

1.992

2.077

Ci - 0 ,

1.147

1.148

1.143

ci - o

1.150

1.146

1.142

2

2

2

Bond Anales idea) 46.9

44.9

42.1

2

90.8

93.6

94.2

«3

90.3

88.9

67.4

i

a

tt

. 4.

Calculated structural parameters for Cr(CO)4(H )2 species. 2


7. KOBER & HAY


CO

oc

H /

\

101

H

•


Cr

/ I V CO

Fig. 5.

OC

1

CO

Η

Correlation diagram for bis—dihydrogen to dihydrogen-dihydride forms of Cr(CO) (H2)2 species. 4



102

Fig. 6.


Calculated energies from Hartree-Fock (SCF) and configuration interaction (CI) calculations for one possible path for H - D exchange involving Cr(CO)4(H )2. 2


7. KOBER & HAY

103

Transition Metal Dihydrogen Chemistry Table II. Relative Energies (kcal/mole) for Cr(CO) (H )2 and Related Species 2

4

SCE CrfCOVH )2—upright Cr(CO) (H } —coplanar Cr(CO) (H )—coplanar Cr(CO) H (H ) Cr(CO) H + H 2

4

2

4

4

2

4

2

0.0 3.1b 22.5 29.7 43.3

0.0 3.1* 22.5 11.8 14.4

2

4

2

2


*E = -532.823307 a.u. bE = -533.416149 a.u.

The CI results are contrasted with the SCF calculations in Fig. 6 and in Table II. The bis-dihydrogen complex C r ( C O ) ( H ) is 27 kcal/mole and 40 kcal/mole more stable, respectively, than the dihydrogen-dihydride species C r ( C O ) H ( H ) and the C r ( C O ) H + H fragment, respectively. These energies compare to 9 and 11 kcal/mole for the respective SCF calculations. The potential energy surface has also changed in that the C r ( C O ) ( H ) intermediate appeared to be a transition state at the SCF level, but actually lies below the energy of the C r ( C O ) ( H ) product at the CI level. Finally we compare the results ot the above ab initio calculations with the earlier extended Huckel theory (EHT) calculations of Burdett et al (IS). In this work, which helped to stimulate our own calculations, the authors emphasized that they were probing general trends and that reliable thermodynamic stabilities of M H species could not be obtained using this method. With these points in mind we compare the relative energies of tetrahydrogen species of C r ( C O ) in Table III. 4

4

2

2

4

2

2

2

2

4

4

2

2

n

4

Table III. Comparison of extended Huckel theory (EHT) results (lfi) with SCF ab initio calculations on C r ( C O ) tetranydrogen species 4

Rel. energy (kcal/mole) EHT SCF

C r ( C O ) fragment 4

ds-dihydrogen dihydrogen-ndihydride dihydride + H dihydrogen + H planar H linear H 3 triangular H 3 tetrahydride square H tetrahedral H

0

2

2

4

4

4

17 22 35 36 86 160

0 12 14 15 23 22 76 65 54


4

104


While the earlier study did not investigate the H ( H ) species, the other intermediates involving "open " planar H3 and H4 moieties are actually placed at similar energies, while the intermediates involving closed species do not correspond with our findings. In particular, the (H )(H) species previously described as having two—electron triangular H3* and H ' ligands lies at considerably higher energy. There has, in fact, been considerable experimental activity to identify and isolate complexes containing H and H " ligands, but we find no evidence to support these forms for this particular class of complexes. 2

2

3

+

3

3


Discussion of Mechanisms. In summary, the above calculations have identified relatively low-energy (i.e., less than 30 kcal/mole) pathways for the conversion of H 2 - D 2 into HD as exemplified by the dihydrogen—dihydride species originally formed according to the process C r ( C O ) ( H ) ( D ) 1 Cr(CO) (H)(HD)(D) 4

2

2

4

as shown in Figs. 5 and 6. Once the incipient HD bond has begun to form, the reaction could proceed further by (a) rotation of HD about the C r - H D bond followed by collapse to Cr(CO) (HD) , (b) dissociation of HD to form Cr(CO)4(H)(D) followed by insertion of H or D , or (c) dissociation of HD followed by collapse to Cr(CO)4(HD), to mention some of the possibilities. Despite the above low-energy pathways discussed above, it is not clear whether they are actually involved in the liquid xenon experiments of Upmacis et al since relatively small barriers must be involved for any process at these temperatures. We would point out that the possibility of radical processes involving the presence of H atoms should also be examined thoroughly before these mechanisms are definitively understood. We observe that scrambling involving coordinated H and H atoms is much more facile than processes involving two H ligands. In cases such as 4

2

2

2

2

2

Cr(H )(H)(H) - C r ( H — H — H ) ( H ) -+ Cr(H)(H )(H) 2

2

and V ( C O ) ( H ) ( H ) -, V ( C O ) ( H — H — H ) -, V ( C O ) ( H ) ( H ) 5

2

5

5

2

the barriers involved in open H intermediates are only 10-13 kcal/mole. A second possibility to be considered is that the Cr(CO)4H dihydride species actually possesses a triplet ground state much lower in energy than the singlet species discussed in Fig. 5 and Table II. Such a species could react with H to form H—atom containing species. In summary, the nature of some likely reaction intermediates involved in HD formation from H and D complexes of Cr(CO)4 have been identified here. Unraveling the further details of the mechanisms involved in these fascinating complexes will require more extensive experimental and theoretical studies. 3

2

2

2

2

Acknowledgment This work was carried out under the auspices of the U.S. Department of Energy.


7. K O B E R & I I A Y


105

Literature Cited


1.

Kubas, G.J; Ryan, R.R.; Swanson, B.I.; Vergamini, P.J.; Wasserman, H. J. Am. Chem. Soc. 1984, 106, 451. (b) Kubas, G.J.; Ryan, R.R.; Wrobleski, D. ibid. 1986, 108, 1339. (c) Kubas, G.J.; Unkefer, G.J.; Swanson, B.I.; Fukishima, E. ibid. 1986, 108, 7000. 2. Kubas, G.J. Acc. Chem. Res. 1988, 21, 120 and references therein. 3. Saillard, J.-Y.; Hoffmann, R. J. A m . Chem. Soc. 1984, 106, 2006. 4. Jean, Y . ; Eisenstein, O., Volatron, F.; Maouche, B.; Sefta, F. J . Am. Chem. Soc. 1986, 108, 6587. 5. Hay, P . J . J . A m . Chem. Soc. 1987, 109, 705. 6. Hehre, W . J . ; Radom, L.; Schleyer, Paul v.R.; Pople, J . Α. Ab Initio Molecular Orbital Theory. Wiley: New York, 1986. 7. Spangler, D.; Wendoloski, J.L.; Dupuis, M . ; Chen M . M . L . ; Schaefer III, H.F. J. A m . Chem. Soc. 1981, 103, 3985. 8. Faegri, K . ; Almlof, J . Chem. Phys. Lett. 1984, 107 121. 9. Dobbs, K . D . ; Hehre, W . J . J . Computational Chem. 1987, 8, 861. 10. Williamson, R . L . ; Hall, M . B . Int. J. Quantum Chem. Symp. 1987, 21, 503. 11. Rohlfing, C . M . ; Martin, R.L. Chem. Phys. Lett. 1985, 115,104. 12. Rohfling, C . M . ; Hay, P . J . J . Chem. Phys. 1985, 83 4641. 13. Bauschlicher, Jr., C.W.; Walch, S.P.; Langhoff, S.R. Quantum Chemistry: The Challenge of Transition Metals and Coordination Chemistry; Veillard, Α., Ed.; Reidel: Dordrecht, Holland, 1985; p. 15. 14. (a) Hay, P.J.; Wadt, W.R. J. Chem. Phys. 1985, 82, 270. (b) Wadt, W. R.; Hay, P . J . ibid. 1985, 82, 274. (c) Hay P. J.; Wadt, W . R . ibid. 1985, 82, 299. 15. Crabtree, R.H.; Hamilton, D . G . J . A m . Chem. Soc. 1986, 108, 3124. 16. Upmacis, R . K . ; Poliakoff, M . ; Turner, J.J. J . A m . Chem. Soc. 1986, 108, 3645. 17. Thompson, M.E.; Baxter, S.M.; Bulls, A.R.; Burger, B . J . ; Nolan, M . C . ; Santarsiero, B.D.; Schaefer, W.P.; Bercaw, J.E. J . A m . Chem. Soc. 1987, 109, 203. 18. Burdett, J . K . ; Phillips, J.R.; Pourian, M . ; Upmacis, R. Inorg. Chem. 1987, 26, 3061. 19. Modified GAUSSIAN82 Program: J.S. Binkley and R.L. Martin. 20. M E S A program: P.W. Saxe and R.L. Martin. RECEIVED December 9, 1988


The Challenge of d and f Electrons - American Chemical Society

Recommend Documents