Applications of molecular modeling in
catalysis
From homogeneous to heterogeneous
by
Dr. R. Mahalakshmy
Young Scientist
NCCR
1
Contents
What is molecular modeling?
How do we apply molecular modeling to homogeneous
system?
e.g: Mn-salen (Jacobsen’s catalyst) catalysed epoxidation
of olefin
How do we apply molecular modeling to heterogeneous
system?
e.g: Epoxidation of olefin by immobilized Mnsalen catalyst
in MCM-41
2
Introduction
What is molecular modeling?
It is a broadly generic term that defines the use of computers to
study chemical systems, with an emphasis on the structure,
properties, and activities of molecules.
Who is molecular modeler or computational chemist?
Those scientist who are specially trained in the technologies,
techniques and tools of molecular modeling.
Computational tools:
• Both software and computer hardware
• Computing platform: Ranging from simple desktop computers to
the use of very high performance super computers
• Support tools: Interface to the codes, data visualizers (programs
3
that create images from the computed data)
Methods
(i)
Molecular mechanics (MM)
Properties of the molecule can be calculated by measuring the motion
of the atoms and the changing energies of the spring.
(ii) ab initio quantum chemical methods
Based on the results of calculating the wavefunction, other chemical
properties and activities can be determined.
(iii) Semi-empirical quantum chemical methods
A portion of the calculation comes from experimental data and the rest
comes from mathematics.
(iv) Density Functional Theory (DFT)
Determines molecular properties from calculating the electron density
rather than from the wavefunction.
*Depending on how broadly one defines molecular modeling or
computational chemistry, there are a number of other methods that
can be considered.
4
Types of calculations that can be performed on
the molecule
Single point energies (molecular energies)
Molecular orbital calculations, including determination of
frontier orbitals
Vibrational frequency calculations
Reaction mechanisms and reaction path following studies
Determination of IR and UV-Vis spectra
Transition structures and activation energy diagrams
Electron and charge distributions
Potential energy surfaces (PES)
Thermodynamic calculations
5
Fundamental Uses
Chemical structure, or geometry of the molecule
Number and type of atoms, bonds, bond lengths, angles,
and dihedral angles.
Properties of system of molecules
Basic characteristics of the molecule, such as its
molecular energy, enthalpy, and vibrational frequencies.
The activity or reactivity of a molecule
Those characteristics that describe how the molecule
behaves in the presence of other molecules, such as its
nucleophilicity, electrophilicity, and electrostatic potentials.
6
Other uses
To model a molecular system prior to synthesizing that
molecule in the laboratory.
Understanding a problem more completely.
Some properties of molecule can be obtained computationally
more easily than by experimental means.
e.g. Molecular bonding
Anyone can do calculations nowadays.
Anyone can also operate a scalpel.
That doesn’t mean all our medical problems are solved.
-Karl Irikura
7
Jacobsen epoxidation
Jacobsen Epoxidation: Enantioselective synthesis of epoxides from
isolated alkene.
Jacobsen's catalyst
• The commercially available '1994 reagent of the year.
• Converts achiral olefins to chiral epoxides with enantiomeric
excesses regularly better than 90% and sometimes exceeding
98%.
• To date, the origin of this dramatic selectivity has not
been explained.
8
Catalytic cycle involved in the epoxidation
Step1:An oxidant transfers atomic oxygen to the MnIII catalyst
(the oxygen presumably coordinates to the metal in a site
normal to the salen plane)
Step2:The activated oxygen is then delivered to the alkene.
9
Three possible modes of oxygen delivery
Mechanisms of oxygen transfer from the catalyst to the olefin
Step A: Attack of an oxygen radical intermediate on the
double bond
Step B: Concerted oxygen delivery &
Step C: Formation of a metallaoxetane intermediate
10
Investigation of the stereoselectivity of the Mn(III)salen
catalysed epoxidation reaction mechanism by DFT
Highly enantioselective!!!
What is its origin???
(1) How can the relatively flat catalyst give such impressive
selectivity?
(2) What is the trajectory of approach of alkenes (top, side, or
bottom on 1) to the MnO intermediate?
(3) What is the timing of bond formation between the alkene
and transferred oxygen?
*The above issues can be explored in more details by computing the
geometries of the manganese-Oxo intermediate as a prelude to
11
transition state searches.
Computational details
Program: Gaussian-94
Theory:DFT
Functional:B3LYP
Basis set:
(C, H, Cl, N,O)
Split valence double  (DZ) basis set: 3-21G
Double  plus polarisation basis set : 6-31G*
DZ and TZ valence basis set for Mn
Org. Lett., Vol. 1, No. 3, 1999
12
Model systems employed in calculation
N
N
N
O
1 - Mn(III)
N
N
O
O
2-Mn(V)
N
Mn
O
N
Mn
O
O
N
Mn
Mn
O
O
Cl
3 -Mn(III)
O
O
Cl
4-Mn(V)
13
13
Spin multiplicity
3d7
Mn(0)
Oxidation state range: II to VII
Mn(III)
eg
eg
eg
t2g
t2g
t2g
Low spin
singlet (S=0)
Intermediate spin
Triplet (S=1)
High spin
Quintet (S=2)
Mn(III)
d2(Oh)
Mn(V)
Triplet (S=1
d4 (Oh)
d4 (Oh)
eg
eg
eg
t2g
t2g
t2g
Singlet (S=1)
Quintet (S=2)
14
Quintet and triplet state geometries of the Manganese(III)
model systems
1
H.S (q)
1
L.S (t)
3
H.S (q)
3
H.S (t)
15
Manganese(V) model compounds calculated by the
manganese triple- basis/Becke3LYP/6-31G*
2 (s)
0 kcal/mol
2(t)
3.5 kcal/mol
2 (q)
11.2 kcal/mol
4 (s)
10.2 kcal/mol
4 (t)
0 kcal/mol
4 (q)
2.0 kcal/mol
16
Comparison of X-ray data to calculated values for
model systems
X-ray
B3LYP/
3-21G
B3LYP/
6-31G*
[Mn(dz)]
B3LYP/
6-31G*
[Mn(tz)]
Mn-N
N-Mn-N
Mn-O
O-Mn-O
Mn-Cl
2.001/
1.986
1.945/
1.929
1.958/
1.952
83.8
1.871/
1.853
1.873/
1.867
1.900/
1.900
90.3
2.391
90.6
2.352
92.1
2.239
1.967/
1.961
82.1
1.905/1.9
04
92.1
2.231
82.7
82.6
17
Becke3LYP/3-21G relative energies (kcal/mol) for model
systems 1-4
1
Mn(III)
s
t
3
Mn(III)Cl
q
s
t
2
Mn(V)O
q
s
t
0.8
0.0
2.1
E
48.4 27.6
0.0
53.1 30.5
0.0
S2
2.7
6.0
2.1
6.1
4
Mn(V)O/Cl
q
s
t
q
16.3 11.7
0.0
4.0
6.0
2.9
6.1
The <S2> values measure whether the spin state is pure singlet, triplet or quintet
(S2 = 0, 2, 6, resp)
18
Becke3LYP/6-31G* (CHClNO)/DZ (Mn) relative energies
(kcal/mol) for model systems 1-4
1
Mn(III)
s
t
3
Mn(III)Cl
q
s
t
2
Mn(V)O
4
Mn(V)O/Cl
q
s
t
q
s
t
q
0.0
4.3
10.0
8.2
0.0
1.2
2.1
6.1
2.9
6.1
E
43.6 22.6
0.0
35.4 26.8
0.0
S2
2.6
6.1
2.0
6.1
19
Becke3LYP/6-31G* (CHClNO)/TZ (Mn) Relative energies
(kcal/mol) for model systems 1-4
1
Mn(III)
E
S2
3
Mn(III)Cl
s
t
q
45.7
27.1
0.0
2.6
6.1
S
t
2
Mn(V)O
q
s
t
40.2 19.5
0.0
0.0
3.5
2.1
6.1
2.1
4
Mn(V)O/Cl
q
s
t
q
11.2 10.2
0.0
2.0
6.1
2.9
6.1
•The Mn(III) catalysts are predicted to be high spin (i.e Quintet)
The ligand field does not split the degeneracy of the d
orbitals significantly - No spin pairing
•Mn(V)Oxo is predicted to have nearly degenerate singlet and
triplet states.
20
Conclusion
•A low – spin ( s or t) complex is favoured with weak coordination or
absence of a 6th ligand
• A high spin (t or q) state occurs upon association of a stronger
ligand.
• High spin oxo intermediate could lead directly to high spin product
in a concerted fashion with conservation of spin.
• Low spin species would have to undergo a change of spin
multiplicity during reaction which could give stepwise processes.
*Nature of the ligand will influence spin multiplicities and
potentially the relative rates of stereo specific concerted and
stepwise nonconcerted processes.
21
Investigation of the origin of enantioselectivity in the
epoxidation of alkenes catalysed by anchored
oxo Mn(V)-salen into MCM-41 channels
DFT and QM/MM approach
Journal of molecular catalysis A: Chemical 271(2007) 98-104
22
Objective
Based on DFT and QM/MM calculations,
To rationalise the effect of immobilization and show how that
correlates with the linker and substrate choices.
To evaluate the enantioselectivity of the catalyst with respect to
the
energy surfaces along the epoxidation reaction pathway.
23
Model systems employed in calculation
N
N
N
Mn
L
N
Mn
O
O
O
O
O
L
I
II
The full (I) and truncated (II) models of the Mn-salen complex
L(axial linker) = Cl-, Phenoxyl
24
Model systems employed in calculation
(a) Visualized Mn-salen complexes
anchored inside a MCM-41 channel
using phenoxyl group as the
immobilizing linker
(b) The model for the
immobilized Mn-salen
complexes using phenoxyl
axial linkage.
25
Relative energies of spin states for oxo-Mn-salen II complex
vs. axial linkage (Calculated by Mn(tz) basis using
B3LYP/6-31G* functional)
Method
Spin state
Mn-salen-Cl
(truncated)
Mn-salenphenoxyl
(truncated)
B3LYP
B3LYP
B3LYP
Singlet
Triplet
Quintet
10
0
2
50
0
10
•The triplet state is found to be the ground state for both Cl- and
phenoxyl linkers.
• Epoxidation reaction for both homogeneous catalyst (i.e., Cl- is
the axial ligand) and heterogeneous catalysts (i.e., phenoxyl group
is the axial ligand) occurs on a triplet surface.
26
Effect of axial linkage on the geometry of oxo-Mn-salen II for
the triplet spin sate using B3LYP method: distances in A° ,
angles in degrees
Bond/angle
Mn-salen-Cl
truncated)
Mn-salen-phenoxyl
truncated)
Mn= O(oxo)
1.82
2.11
Mn=O(phenoxyl)
2.34
1.85
Ligand
plane…Mn
0.184
0.555
Mn–O(oxo)–L
56. 940
124.930
27
DFT-calculated structures for the truncated Oxo-Mnsalen II with Cl−
Optimised structure
HOMO
LUMO
28
Optimized geometry of TS1 and TS2 corresponding to
the attack of TBMS for homogeneous Mn-salen catalyst
29
General mechanism scheme for asymmetric epoxidation
of olefins using Mn-salen complexes, L = axial linker
30
Reaction profiles for the oxygen transfer reaction on the
triplet surface for homogeneous Mn-salen catalyst
1. It is the mixture of CBMS
and the catalyst in gas phase
without any interaction in
between.
2.Olefin enters the coordination
sphere of complex from the
Mn=O center.
Activation energy for TS1=22
kJmol
3. Formation of radical
intermediate.(90 kJ/mol more
stable than 2)
Activation energy for TS2=35
kJ/mol
Trans--methyl styrene (TBMS)
4.Formation of Epoxide
31
Cis--methyl styrene (CBMS)
complex.
Results of DFT calculation
• The attack of TBMS is overall about 7 kJ/mol less in favour
compared to that in CBMS.
•The epoxide complex formed by TBMS lies 2.5 kJ/mol below
the epoxide complex formed by CBMS.
•These findings are in agreement with a general assumption in
homogeneous Mn-salen catalyst that TBMS is a less suitable
substrate than CBMS.
•The more stable t rans-epoxide complex 4 is in qualitative
agreement with the experimental observation that epoxidation of
CBMS leads to a thermodynamically more stable trans-epoxide
32
DFT-calculated structures for the truncated oxo-Mn(II)salen
with phenoxyl, located trans- to the oxo group
HOMO
LUMO
33
Optimised geometry of the transition states TS1 and TS2
corresponding to the attack of TBMS at the Mn=O center of
heterogeneous Mn-salen catalyst
34
Energy profile for the epoxidation reaction of CBMS and
TBMS catalyzed by immobilized Mn-salen complex.
TBMS
CBMS
35
Result DFT calculation for the attack of CBMS
•The activation barrier for the formation of adsorbed CBMS radical is overall
8 kJ/mol less than that for the homogeneous salen catalyst.
(Homogeneous catalyst =22 kJ/mol; heterogeneous catalyst=14 kJ/mol)
•The activation energy for epoxide complex formation is about 5 kJ/mol
lowered for the immobilized Mn-salen complex than that for the homogeneous
salen catalyst.
(Homogeneous catalyst=35 kJ/mol; heterogeneous catalyst=30 kJ/mol)
•On the triplet energy surface, the calculations suggest that epoxide formation
is clearly more preferred by the immobilized catalyst compared to the
homogeneous catalyst.
•These observations provide an explanation for the effectiveness and
importance of additional ligand as the immobilizing linker.
36
DFT calculation for the attack of TBMS
•In contrast to the homogeneous epoxidation reaction, a trans olefin
can be a suitable substrate for epoxidation by immobilized Mn-salen catalysts.
•Epoxide complex 4 formed of CBMS is slightly (3 kJ/mol) more stable than its
trans-epoxide counterpart, despite of the fact that there is a lower barrier
(5 kJ/mol) towards epoxide formation starting from TBMS radical adsorbed
on Mn=O center.
•The energy profile strongly depends on electron donor/acceptor properties of the
axial linker.
37
Modelling of MCM-41 silica channel
•MCM-41 silica channel was modelled based on a straight, three-dimensional
channel.
Pore length= 3.4 nm
•The structure was represented by the pseudo cell, Si6O12, consists of hexagon
arrangements of Si–O–Si units.
•Oxygen atoms saturate all silicon atoms at the pore surface.
•Oxygen atoms with fewer than two silicon atoms attached to them (at the
inlet, outlet and outer surface) were then saturated by hydrogen atoms.
38
Modelling of MCM-41 silica channel
•This model places the silicon and oxygen atoms in a simple geometrical
arrangement and does not reproduce the real amorphous structure of MCM-41.
•In this MCM-41 model, all hydroxyl groups were located at the outer surface.
•The pore length is 3.4 nm and the pore diameter is 2.3 nm.
• Immobilization was performed by attaching the Si atom of the linker to oxygen
atoms connected to two Si atoms on the wall (SiLinker–O–SiMCM).
39
Model used for ONIUM based QM/MM calculation
• A two-layer ONIOM protocol
was used to couple the QM and
MM parts in the full Mn-salen-L
(L = pheoxyl) calculations as well
as for the immobilized Mn-salen
catalyst into MCM-41 channel.
•The high-level model system
includes Mn, N, O, carbon atoms
at the bridge and the full
substrate and linker atoms.
•The low level calculation
includes the rest of salen ligand
and the atoms of MCM-41
channel.
40
Reaction profiles for the epoxidation of CBMS on the triplet
surface for Mn-salen immobilized into MCM-41 channel
using a phenoxyl linker
•Energy barrier are lower
because of the confinement
effect inside MCM-41 channel.
• The more restricted space
inside the mesopore in
combination with the effect
of
immobilizing
linker
hinders the free-movement
of the attached olefin.
• Besides, a full-salen ligand
with substituents at 5, 5and 3, 3-positions exhibits
stronger steric influences
compared to that for the
model complex II.
41
Conclusions
•
Catalyst
Homogeneous
Heterogeneous
Suitable substrate
cis-olefin
Trans-olefin
Main product
trans-epoxide
cis-epoxide
•Comparison of the optimized structure of the intermediate oxo-Mn-salen with the
chloride and phenoxyl linker at the trans-position leads to the following
suggestions:
Any coordination in the trans-position causes the Mn atom to move into the plane
of the ligand.
The movement of the Mn atom can improve the enantioselectivity in view of the
shielding of the oxygen by the equatorial ligands.
The Mn=O bond becomes lengthened –facilitates oxygen transfer to the olefin.
•A trans-substrate has a higher level of asymmetric induction to the immobilized
Mn-salen complex than that to a homogeneous catalyst, but the reaction path is
more in favor of the cis-substrate.
•The MCM-41 channel reduces the energy barriers and enhances the
enantioselectivity by influencing geometrical distortions of the Mn-salen
complex.
42
43
Questions?
44