The ONIOM (our Own N-layered Integrated molecular Orbital and molecular Mechanics) method [1] is a multi-layered hybrid QM/MM method designed to calculate energies and optimized geometries for large molecules. It is currently available in GAUSSIAN 03 [2].

Our development has centered on excited state reaction path calculations involving "non-vertical" excitation energies, which depend upon molecular geometry changes on excited state potential energy surface(s). Key features of potential energy surfaces to be identified to understand any subsequent reactivity are transition structures (controlling barriers to reaction paths) on a single potential energy surface, and conical intersections (giving rapid radiationless decay) between potential energy surfaces.

To date (2006), few other excited state calculations have appeared in the literature that use the ONIOM method [3-6], and none of these [3] have fully optimized a conical intersection between potential energy surfaces. Our initial results show that such calculations are possible while only a small part of a molecule (a chromophore) is described using an excited state method.

Calculations on ergosterol, for example, were up to one hundred times faster and less demanding of resources, with little loss of accuracy compared to a full non-ONIOM calculation. (Ergosterol has also been studied with MMVB).

The ONIOM method is a hybrid method, which uses an extrapolation to approximate the TARGET energy E(HIGH,REAL):

E(ONIOM) = E(HIGH,MODEL) + E(LOW,REAL) - E(LOW,MODEL) (1)

The MODEL is a small fragment of the full REAL molecule, as shown for example in Figure 1, while the HIGH method is the more accurate level of theory.

Figure 1: Ergosterol, showing the MODEL and REAL partitions used in our ONIOM calculations.

The absolute energy E(ONIOM) is not expected to be equal to the TARGET energy E(HIGH,REAL), but energy differences between two points on a potential energy surface will be reproduced, provided that the effects of changing size, given by:

E(LOW,REAL) - E(LOW,MODEL) (2)

and level of theory, given by

E(HIGH,MODEL) - E(LOW,MODEL) (3)

are separable.
From the equations above and Figure 1, there are two interpretations of how ONIOM works:

Adding substituent effects (REAL) to a HIGH level MODEL calculation;
Improving a LOW level REAL calculation in the region of a (MODEL) active site.

Based on our experience to date, we favour the first interpretation: ONIOM is useful in that it saves computational time and resources, by limiting expensive/slow accurate (HIGH level) calculations to a small molecule fragment, where they are essential. The surroundings can be described by much cheaper/faster (LOW level) computational methods, which may give poor results on their own, yet have a beneficial effect on the fragment calculation. Even though three energy calculations are performed for a single ONIOM point, it is the E(HIGH,MODEL) calculation that dominates, and because of the reduced MODEL size (Figure 1), the calculation is faster overall.

References:
[1] T. Vreven, K. Morokuma, J. Comp. Chem. 21, 1419-1432 (2000).
[2] Gaussian Development Version, Revision D.02, M. J. Frisch et al; www.gaussian.com.
[3] F. Blomgren, S. Larsson, J. Phys. Chem. B. 109, 9104-9110 (2005).
[4] A. Yamada, T. Ishikura, T. Yamato, Proteins-Structure, Function and Bioinformatics 55, 1063-1069 (2004)
[5] T. Vreven, K. Morokuma, J. Chem. Phys. 113, 2969-2975 (2000)
[6] Other published ONIOM excited state calculations are for vertical excitations only.

previous :: next

Home - General Information - Group Members - Research - Publications - CASSCF - Available Positions