3.5 ONIOM Method
139
For the 1st layer, high level of the MO calculation is applied, whereas the 2nd
and the 3rd layers are less important in order so that intermediate- and low-level
calculations are employed for these. These situations are schematically shown also
in Figs. 3.21 and 3.22. Note that there is some difference in the calculation level
for the 2nd layer depending on ONIOM2 or ONIOM3 method. For the 2nd and the
3rd layers, it is usual to employ the MO calculations with a simpler basis set or the
semiempirical MO scheme, which is totally called QM/QM framework. Even the MM
scheme might be applied for the low-level calculation called QM/MM framework.
For the two-layer ONIOM (ONIOM2) method, the energy of the real system is
approximated by
E(Real, High) = E(Real, Low) + {E(Model, High) − E(Model, Low)} (3.59)
and for the three-layer ONIOM (ONIOM3) method by
E(Real, High) = E(Real, Low) + {E(Intermediate, Medium) − E(Intermediate, Low)}
+ {E(Model, High) − E(Model, Medium)}
(3.60)
These approximations will be simply extended to multilayer multilevel structures
in the ONIOM framework.
It is most crucial to appropriately deal with the interlayer connection between both
the layers in a seamless manner. Careful manipulations concerning these have been
described in the earlier references (Maseras and Morokuma 1995; Svensson et al.
1996). The total energy, the energy gradient, most of the general electronic properties,
and the vibrational properties of the Real system can thus be well retrieved so that
these properties are acquired in a similar fashion to the ordinary calculation of the
whole molecule. In other words, molecular structural optimization and search of the
transition state during the chemical reaction can be performed are also attainable in
the ONIOM calculation.
3.5.2 Simple Example of the ONIOM Method
Let us examine here a simple ONIOM2 calculation taking 1-propanol molecule as an
example for the Real system and setting methanol part as the 1st layer and ethyl part
the 2nd layer as shown in Fig. 3.23. Several combinations of the calculation method
for each layer have been employed, the results of which are listed in Table 3.3. As
the benchmark test CCSD/6-311G** is employed for both the 1st and the 2nd layers.
Other calculation levels common to both layers are also attempted. It is natural that
the CCSD/6-311G** calculation should take the longest computation time among
others but gives the most stable energy implying the relatively precise result among
the present calculations as expected. As the ONIOM2 calculation combination of
the CCSD/6-311G**: MP2/6-31G* or the CCSD/6-311G**: HF/STO-3G gives a
139
For the 1st layer, high level of the MO calculation is applied, whereas the 2nd
and the 3rd layers are less important in order so that intermediate- and low-level
calculations are employed for these. These situations are schematically shown also
in Figs. 3.21 and 3.22. Note that there is some difference in the calculation level
for the 2nd layer depending on ONIOM2 or ONIOM3 method. For the 2nd and the
3rd layers, it is usual to employ the MO calculations with a simpler basis set or the
semiempirical MO scheme, which is totally called QM/QM framework. Even the MM
scheme might be applied for the low-level calculation called QM/MM framework.
For the two-layer ONIOM (ONIOM2) method, the energy of the real system is
approximated by
E(Real, High) = E(Real, Low) + {E(Model, High) − E(Model, Low)} (3.59)
and for the three-layer ONIOM (ONIOM3) method by
E(Real, High) = E(Real, Low) + {E(Intermediate, Medium) − E(Intermediate, Low)}
+ {E(Model, High) − E(Model, Medium)}
(3.60)
These approximations will be simply extended to multilayer multilevel structures
in the ONIOM framework.
It is most crucial to appropriately deal with the interlayer connection between both
the layers in a seamless manner. Careful manipulations concerning these have been
described in the earlier references (Maseras and Morokuma 1995; Svensson et al.
1996). The total energy, the energy gradient, most of the general electronic properties,
and the vibrational properties of the Real system can thus be well retrieved so that
these properties are acquired in a similar fashion to the ordinary calculation of the
whole molecule. In other words, molecular structural optimization and search of the
transition state during the chemical reaction can be performed are also attainable in
the ONIOM calculation.
3.5.2 Simple Example of the ONIOM Method
Let us examine here a simple ONIOM2 calculation taking 1-propanol molecule as an
example for the Real system and setting methanol part as the 1st layer and ethyl part
the 2nd layer as shown in Fig. 3.23. Several combinations of the calculation method
for each layer have been employed, the results of which are listed in Table 3.3. As
the benchmark test CCSD/6-311G** is employed for both the 1st and the 2nd layers.
Other calculation levels common to both layers are also attempted. It is natural that
the CCSD/6-311G** calculation should take the longest computation time among
others but gives the most stable energy implying the relatively precise result among
the present calculations as expected. As the ONIOM2 calculation combination of
the CCSD/6-311G**: MP2/6-31G* or the CCSD/6-311G**: HF/STO-3G gives a