Theoretical Studies of HeavyAtom NMR Spinspin Coupling Constants

1
Theoretical Studies of Heavy-Atom NMR Spin-spin
Coupling Constants

• With Applications to Solvent Effects in Heavy
  Atom NMR

Jochen Autschbach Tom Ziegler, University of
Calgary, Dept. of Chemistry University Drive
2500, Calgary, Canada, T2N-1N4 Email
jochen_at_cobalt78.chem.ucalgary.ca
1
2
What is interesting about Heavy Metal Compounds ?

• Spin-orbit coupling, scalar relativistic effects
• Relativistic theoretical treatment sizeable
  effects on bonding for 6th row elements (bond
  contractions, De,ne,IP, ) are already textbook
  knowledge (e.g. Au maximum)
• Simple estimates propose absolute (!) scalar
  relativistic effects of gt100 for 6th row
  elements for NMR spin-spin coupling constants
• Coordination by solvent molecules possible

2
3
Methodology

• Spin-spin coupling constants

Indirect coupling K(A,B)
Electrons with orbital- and spin- magnetic moments
Direct coupling (vanishes for rapidly rotating
molecules)
Nucleus B Spin magnetic moment creates magnetic
field
Nucleus A Spin magnetic moment creates magnetic
field
3
4
Reduced spin-spin coupling tensor
Reduced coupling constant
Coupling constants in Hz from the NMR spectrum
4
5

• The ZORA one-electron Hamiltonian

Variationally stable two-com- ponent relativistic
Hamiltonian
Tnrel relativistic corrections of T and V ,
spin-orbit coupling
Molecular effective Kohn-Sham potential if used
in DFT
Magnetic field due tonuclear magnetic moments
Replacement to account for magnetic fields
5
6
The ZORA Hyperfine Terms
Requires solution of 1st-order pertur- bation
equations
Nuclei A and B, directions j and k, point-like
magnetic dipoles
6
7
Description of the code

• Auxiliary program CPL for the program ADF
  (Amsterdam Density Functional, see www.scm.com)
• Based on nonrelativistic, ZORA scalar or ZORA
  spinorbit 0th order Kohn-Sham orbitals
• Analytic solution of the coupled 1st order Kohn-
  Sham equations due to FC-, SD-, and PSO terms
  (instead of finite perturbation)
• Accelerated convergence for scalar relativistic
  calculations (lt 10 iterations)
• Spin-dipole term implemented
• Currently no current-density dependencein V, Xa
  or VWN approximation for 1st order exchange
  potential

7
8
Results I scalar ZORA
One-bond metal ligand couplings Hg-C Pt-P W-C
, W-H, W-P, W-F Pb-H ,Pb-C, Pb-Cl FC PSO
DSO terms included
JCP 113 (2000), 936.
8
9
Tungsten compounds
Lead compounds


W(CO)6 W(CO)5PF3W(CO)5PCl3W(CO)5WI3cp-W(CO)3HW
F6
PbH4 Pb(CH3)2H2 Pb(CH3)3H Pb(CH3)4 PbCl4
exp. extrapolated from Pb(CH3)xHy not
directly measured
9
10
Platinum compounds
Mercury compounds
Pt(PF3)4
Hg(CN)2
Hg(CN)42-
(CH3)Hg-X
PtX2(P(CH3)2)
Hg(CH3)2
cis-PtCl2(P(CH3)3)2trans-PtCl2(P(CH3)3)2cis-PtH2
(P(CH3)3)2 trans-PtH2(P(CH3)3)2 Pt(P(CH3)3)4 Pt(PF
3)4
Hg(CH3)2 CH3HgCl CH3HgBr CH3HgI Hg(CN)2 Hg(CN)42
-
10
11
Results II spinorbit coupling
Spin-orbit (SO) coupling causes cross terms
between the spin-dependent ope- rators (FC,SD)
and the orbital dependent ones (here PSO). The
differences between Scalar and SO in the table
above is mainly caused by these cross terms, and
by the SO effects on the PSO contribution itself.
Tl-I is the first example where SO coupling was
demonstrated to cause the major contributions to
heavy atom spin-spin couplings.
JCP 113 (2000), 9410.
11
) VWN Becke86 Perdew 88 functional, Tl-X
coupling constants
12
Results III solvent effects
Experimental results on pages 9 and 10 obtained
from solution. The cases where results are
unsatisfactory are marked red (linear Hg and
square planar Pt complexes)
SO coupling yields only minor corrections in all
these cases! Is coordination of the heavy atoms
by solvent molecules important?
Some structures that were optimized, explicitly
including a number of solvent molecules
12
13
) Hg-C coupling, VWN functional, scalar ZORA
(numbers in parentheses ZORA spin-orbit)
JACS 123 (2001), 3341.
13
14
cis-PtH2(PMe3)2
trans-PtH2(PMe3)2
Pt complexes
) K / 1020 kg/m2C2Pt-P coupling, VWN
functional. scalar ZORA (in parentheses ZORA
spin-orbit)
14
15
Results III more solvent effects
Two heavy nuclei A Pt-Tl cyano complex
Complex I
Two-bond coupling much larger than one-bond
coupling
Experiment ) 1J(Tl-Pt) 57 kHz 1J(Tl-CB)
2.4 kHz 2J(Tl-CA) 9.7 kHz 2J(Tl-CC) 0.5 kHz
Four water molecules can coordinate to Tl in
aqueous solution (exp. confirmed)
) Optimized bond distances, experimental bond
lengths in parentheses (in Å) ) J. Glaser et
al., JACS 117 (1995), 7550.
15
16
Results III more solvent effects
The unintuitive experimental result 2J(Tl-CA) gtgt
1J(Tl-CB) questions the proposed structure with
a direct Tl-Pt bond (page 15). However, our
computations confirm the structure and the
unusual coupling pattern. The solvent
coordination effect on J(Pt-Tl) and the Tl-C
cpouplings is remarkably large.
16
17
Results III more solvent effects

• free complex both couplingsare comparably large
  in magni-tude but of opposite sign
• inclusion of solvent moleculesshifts both
  couplings. The one-bond coupling is as
  expected influenced much stronger.
• As a result, the two-bond coup-ling is much
  larger than the one-bond coupling
• Delocalized bonds along theC-Pt-Tl-C axis are
  responsiblefor the large magnitude of
  thetwo-bond Tl-C coupling in thefree complex

17
JACS 123 (2001), in press.
18
Summary

• NMR shieldings and spin-spin couplings with ADF
  now available for light and heavy atom systems
• Based on the variationally stable two-component
  ZORA method
• Relativistic effects on spin-spin couplings are
  substantial and recovered by the ZORA method
• Spin-orbit effects are rather small for many
  cases, but dominant for Tl-X
• Coordination by solvent molecules has to be
  explicitly taken into account for coordinatively
  unsaturated systems. Saturating the first
  coordination shell yields satisfactory results in
  these cases.
• Further solvent contributions within the DFT
  error bars

18