Pirika logo
JAVA, HTML5 & Chemistry Site

Top page of Pirika

Chemistry@Pirika
 Properties Estimation
 Polymer Science
 Chemical Engineering
 Molecular Orbital
 Chemo-Informatics
 Other Chemistry
 Academia
 DIY:Do It Yourself
 Programing

Hansen Solubility Parameter (HSP)
  Basic HSP
  Applications
  Polymer
  Bio, Medical, Cosmetic
  Environment
  Properties Estimation
  Analytical Chemistry
  Formulating for Cosmetics
  Other
  DIY:Do It Yourself

Other Writing

Ad Space for you

 

Ad Space for you

 

 

 

 

Last Update
02-Jan-2013

Molecular Orbital & Properties: TSL: MOPAC calculation

A problem when building this transition state database by the difference between MOPAC7 and MOPAC97.

MOPAC will optimize structure by a XYZ coordinate system, if XYZ is specified by a keyword. It seems to be easy to arrive at the transition state with righter searching MOPAC97 for a transition state which specified XYZ. When it calculates by specifying this keyword XYZ, although XYZ is contained into the keyword for the output of the summary file of a calculation result, since the structure itself is an internal coordinate system, cautions are required of MOPAC7. That is, if it tries to read test.arc even if XYZ is specified, it will break out that the structures differ by the XYZ coordinate system, an internal coordinate system, or MOPAC7 and MOPAC97. In this database, the keyword XYZ of the summary file after calculating by XYZ specification by MOPAC7 deleted.


SUMMARY OF PM3 CALCULATION
VERSION 7.00

-

IONIZATION POTENTIAL = 9.665508 EV
MOLECULAR WEIGHT = 173.188
SCF CALCULATIONS = 87
COMPUTATION TIME = 10.000 SECONDS

FINAL GEOMETRY OBTAINED CHARGE
PM3 PRECISE TS UHF XYZ
H13C8O4
25
C 0.0000000 0 0.000000 0 0.000000 0 0 0 0 -0.1756
C 1.3714283 1 0.000000 0 0.000000 0 1 0 0 -0.0600
H 1.0917509 1 121.615777 1 0.000000 0 2 1 0 0.1039
H 1.0900503 1 119.868247 1 156.702697 1 2 1 3 0.0858
C 2.1515575 1 101.913576 1 -101.490761 1 2 1 3 -0.1349
H 1.1001726 1 96.826374 1 -11.530850 1 5 2 1 0.1067

-


SUMMARY OF PM3 CALCULATION
MOPAC 97.00

MOLECULAR WEIGHT = 140.161
SCF CALCULATIONS = 118
COMPUTATION TIME = 12.000 SECONDS

FINAL GEOMETRY OBTAINED CHARGE
PM3 PRECISE TS UHF XYZ
H10C7N1O2
20
C -2.6548714 1 1.3222197 1 -1.6414018 1 -0.0497
C -1.8802227 1 0.1968004 1 -1.7677870 1 -0.0937
H -0.8676876 1 0.2678154 1 -2.1664349 1 0.1033
C -1.1655638 1 -0.1910237 1 0.2315340 1 -0.0073
C -0.6492645 1 -1.5833443 1 0.1473097 1 -0.1316

Only one imaginary frequency appears in a transition state. The first lump was recorded on the database in as a result of oscillating calculation (keyword FREQ) of MOPAC. How to attach the title of oscillating calculation by MOPAC7 and MOPAC97 differs.

MOPAC7

ROOT NO. 1 2 3 4 5 6
-614.35131 7.60204 19.78062 27.58241 59.11591 84.04859

1 0.01816 0.03957 0.00669 -0.03887 -0.00665 -0.01537
2 0.00428 0.01417 -0.01356 -0.03534 0.01309 0.03931
3 0.02795 0.00717 0.00872 -0.05752 -0.05663 0.00512
4 -0.09664 0.03957 0.00674 -0.03857 -0.00656 -0.01486
5 0.05081 0.05254 -0.02198 -0.05100 0.02212 0.04283
6 0.25645 -0.00117 0.00662 -0.01928 -0.05886 0.00907
7 -0.01026 0.01299 0.01277 -0.02857 -0.01416 -0.01949

MOPAC97

Root No. 1 2 3 4 5 6 7 8
1 A 2 A 3 A 4 A 5 A 6 A 7 A 8 A
-557.8 18.1 43.0 54.8 69.2 91.6 133.6 173.3

1 -0.0131 -0.0347 -0.0082 0.0128 0.0186 0.0015 0.0016 -0.0069
2 -0.0067 -0.0112 0.0045 -0.0058 0.0253 0.0302 0.0159 -0.0003
3 0.0277 0.0045 0.0117 -0.0161 -0.0501 0.0250 0.0063 0.0641
4 0.0901 -0.0346 -0.0083 0.0128 0.0182 0.0007 0.0015 -0.0090
5 -0.0412 0.0341 0.0123 -0.0179 0.0280 0.0466 0.0153 0.0200
6 0.2456 0.0071 0.0001 -0.0054 -0.0464 0.0109 0.0086 -0.0118

Moreover, MOPAC97 are eight vibration although lumps are six vibration in MOPAC7.

In MOPAC7, "1 (new-line)" appears in some places. although seemingly this is the vestiges when changing a page by a printer -- this -- an obstacle -- since it was stinking, it eliminated.

MOPAC7

36 0.16423 0.06577 -0.02412 -0.06345 0.03391 -0.00448
37 0.01665 -0.01430 -0.00747 0.02330 -0.01450 -0.00203
38 -0.02151 -0.03602 -0.00098 0.06105 -0.01274 0.14048
39 -0.03517 -0.06313 0.04833 0.30010 -0.25089 0.09363
40 -0.01594 0.00896 0.01551 -0.00730 0.00279 0.02738
1
41 0.00999 -0.01623 -0.00336 0.00422 0.01104 0.00968
42 0.02656 -0.01657 0.00272 0.01691 0.01420 -0.00075
43 -0.01842 0.01420 0.00600 -0.01610 -0.01269 0.01400

 

Other versions have not used.

Please inform me if you know other versions.

Trouble when calculate transition state

When looking for the transition state of vinyl acetate and vinylidene chloride, a trouble just for a moment ... When searching for a transition state by MOPAC, it meets well "A trouble and a solution"

Although MOPAC7 does not know, in MOPAC97, the number of times of geometry optimization is 500 times. When it is going to search for a transition state,

MOPAC 7

CYCLE: 739 TIME: 0.00 TIME LEFT: 7096.0 GRAD.: 9.576 HEAT:-63.95119
CYCLE: 740 TIME: 0.00 TIME LEFT: 7096.0 GRAD.: 8.015 HEAT:-63.97669
CYCLE: 740 TIME: 0.00 TIME LEFT: 7096.0 GRAD.: 8.347 HEAT:-63.96488
CYCLE: 741 TIME: 0.00 TIME LEFT: 7096.0 GRAD.: 8.730 HEAT:-63.96032
TRUST RADIUS NOW LESS THAN 0.00100 OPTIMIZATION TERMINATING
GEOMETRY MAY NOT BE COMPLETELY OPTIMIZED
-------------------------------------------------------------------------------
PM3 PRECISE TS UHF XYZ
H9C6O2Cl2
19

MOPAC 97

CYCLE: 497 TIME: 0.000 TIME LEFT: 55.95M GRAD.: 0.103 HEAT:-64.24002
CYCLE: 498 TIME: 1.000 TIME LEFT: 55.93M GRAD.: 0.100 HEAT:-64.24003
CYCLE: 499 TIME: 0.000 TIME LEFT: 55.93M GRAD.: 0.141 HEAT:-64.24004
CYCLE: 500 TIME: 0.000 TIME LEFT: 55.93M GRAD.: 0.104 HEAT:-64.24005
EXCESS NUMBER OF OPTIMIZATION CYCLES CURRENT VALUE OF GRADIENT NORM = 0.103911
CURRENT VALUE OF GEOMETRY
PM3 PRECISE TS UHF XYZ

It may become and calculation may stop at MOPAC97 (calculation is completed in MOPA7). When such, it is in a keyword.

CYCLES=2000

It puts in. They are vestiges when calculation of MOPAC has taken many days.

But as for this VAc-VDC, a transition state can be found. If oscillating calculation is performed using the structure of the transition state,

A FAILURE HAS OCCURRED, TREAT RESULTS WITH CAUTION!!
++++----**** FAILED TO ACHIEVE SCF. ****----++++ PM3 CALCULATION
MOPAC 97.00
2004/ 8/11


FOR SOME REASON THE SCF CALCULATION FAILED.

THE RESULTS WOULD BE MEANINGLESS, SO WILL NOT BE PRINTED.
TRY TO FIND THE REASON FOR THE FAILURE BY USING 'PL'.

CHECK YOUR GEOMETRY AND ALSO TRY USING SHIFT OR PULAY.

A SCF CALCULATION FAILED error will come out and stop. Although this may have truly bad structure and SCF may not be found,

ITRY=2000

When the number of times of SCF was increased, calculation was completed with no problems in this case, and it carried out perfectly. A transition state can be found.

In addition, when the calculation which searches for a transition state goes wrong.

0. Change the structure of a start. (In this case, it makes, for example from VAc-AN.)
1. Replace an internal coordinate system and a XYZ coordinate system.
2. Replace and calculate MOPAC7 and MOPAC97.
3. Move and calculate a part of molecule.

When TS keyword of MOPAC departs from a transition state library like this time by a fairly powerful keyword, the transition state right quite comfortably can be found.

Although the transition state was able to be found, when oscillating calculation goes wrong.

NORMAL COORDINATE ANALYSIS

Root No. 1 2 3 4 5 6 7 8

1 A 2 A 3 A 4 A 5 A 6 A 7 A 8 A

-20.3 17.7 33.1 39.7 85.8 135.9 162.4 194.3

1 0.0017 -0.0013 0.0088 -0.0078 -0.0019 0.0155 0.0023 0.0010
2 -0.0003 0.0127 -0.0091 -0.0013 -0.0127 0.0192 0.0018 0.0167
3 -0.0010 0.0061 0.0041 -0.0126 -0.0111 -0.0509 -0.0120 0.0080
4 0.0017 -0.0015 0.0088 -0.0079 -0.0021 0.0155 0.0019 -0.0001
5 -0.0027 0.0181 -0.0294 -0.0243 -0.0273 0.0237 0.0009 0.0109

For example, such only one imaginary vibration, a solution may be found. Shortly after looking at vibration of this imaginary, it understands, but this is not going in the reaction direction. (Vibration of a twist) Since vibration of imaginary becomes smaller than -200 in case of the transition state right on experience of it, the larger thing than it suspected and has started.

In addition, when two or more vibration of imaginary exists, it is calculation failure clearly.

0. Change the structure of a start.
1. Move and calculate a part of molecule.

These techniques are effective. It seems that TS keyword of MOPAC97 is powerful compared with the thing of MOPAC7, and two or more vibration of imaginary does not exist probably.

If it is the usual calculation, when searching for the first transition state
PM3 PRECISE TS UHF (XYZ)
then
PM3 PRECISE FORCE UHF (XYZ)
Only vibration is calculated by doing. When carrying out this FORCE calculation very occasionally, structure may also move. Let's confirm whether filename.arc (FOR012) etc. is made to the calculation result. In MOPAC7, since like, once [ when such a tendency is especially strong ] calculating TS according to the case, after replacing XYZ and internal coordinates and calculating TS several times, FORCE calculation is performed.

By calculating by the keyword of PM3 PRECISE TS UHF (XYZ) first, in MOPAC7, if FOR012 is made, the portion on it is eliminated, and if TS is calculated by changing it into FORCE, oscillating calculation can be performed. That is, even if the charge is contained in input data, it is calculable satisfactory. It seems that however, the portion of this charge cannot be calculated in MOPAC97 unless it eliminates. It is troublesome to eliminate the portions of a charge one by one. If the Search button is pushed without choosing a radical and a monomer by extraction of structure , empty text area will appear. Only the portion of the structure of a molecule is copied there (every value of a charge), and the Read button is pushed. Probably, it will be good in calculating MOPAC by copying it to a text editor etc., and rewriting and saving TS to FORCE, since the structure where the charge was eliminated will be acquired if the GetStructure button is pushed immediately after that. The meaning which checks the structure of TS is also effective.

The cautions when changing an atomic kind. When changing an atomic kind, the remaining portion is moved using the standard joint length of the new atom. Under the present circumstances, in a transition state, although many are so, since there is no combination between a radical and a monomer, the phenomenon in which it does not move will come out of one of molecules. Even when it is troublesome, in case a molecule is Edit, please make combination between a radical and a monomer. (When changing the kind of some atoms of a ring, joint length does not change.)

Please put in a MMOK keyword, in calculating amide.

If many MOPAC97 (thing of attachment in Chem3D) of the Mac version is calculated continuously
Not Enough Time to Complete Hessian
The message to say will come out and stop. Then, application is ended and it rerises. (I think that it is a mere bug.) It seems that cooperation with Chem3D is also amusing, calculation of MOPAC97 is also slow and offset values, such as the number of times of SCF, are not suitable for it, either, although it is its personal opinion. Merit that's right direction it seldom uses. Since there is the restriction [ atomic number / of MOPAC7 which he is using / maximum ] 40, the thing exceeding it is using MOPAC97 reluctantly. In the keyword at that time
T=100000 CYCLES=20000 ITRY = 2000 is added.

When the coordinates which JAVA generates are close to zero, the account of a table of reference may become like 1.23xxxxE-4. Since it will become an error if MOPAC is run as it is (error which says that the atom E does not exist), please rewrite with 0.000123xxx manually.

 

MOPAC PM3

MOPAC PM3 conclusion

Activation energy = Heat of Formation of a transition state - (HF of a monomer + radical HF)

 

AAR

AllCR

AMR

AMDR

ANR

BDR

cHnR

CHOR

GMAR

MAAR

AA

7.32

6.60

7.23

6.70

7.41

11.08

9.32

7.19

10.63

11.18

AllC

8.65

7.67

8.61

7.86

8.44

12.05

11.03

8.54

11.36

11.25

AM

9.60

8.91

9.52

9.18

9.65

13.40

11.72

9.57

13.39

13.74

AMD

9.07

8.74

9.38

8.79

9.25

13.66

11.68

9.10

13.50

13.71

AN

9.07

8.84

9.03

8.47

9.28

12.72

10.58

8.66

12.54

12.62

BD

5.95

6.26

5.96

6.14

6.23

10.01

8.25

5.83

9.35

9.24

cHn

11.30

11.11

11.39

11.82

12.41

18.01

11.91

11.52

15.92

15.78

CHO

9.18

8.69

9.12

8.94

9.35

13.23

11.62

9.00

13.14

13.34

GMA

6.95

7.20

6.89

6.72

6.25

11.81

8.90

6.90

9.91

10.35

MAA

8.50

7.90

8.44

8.28

7.87

12.29

10.04

8.26

13.46

12.02

Mal

11.23

10.45

12.56

10.25

12.07

13.45

11.44

10.94

15.87

14.97

MAN

7.57

7.48

7.62

7.14

7.09

11.31

9.32

7.24

10.70

10.86

MeO

7.38

8.22

8.45

7.91

7.84

12.77

10.79

7.63

11.19

11.20

MMA

8.75

7.57

8.85

8.91

7.75

12.72

11.12

8.84

12.12

12.15

St

5.97

6.10

5.98

6.05

6.02

9.90

5.38

5.91

9.25

9.22

Vac

6.82

7.27

6.82

6.31

6.82

11.68

10.09

6.83

10.41

10.36

VC

8.32

8.57

8.39

8.46

8.75

12.78

10.72

8.41

12.15

11.93

VDC

7.59

7.47

7.45

7.56

7.72

 

 

7.37

11.02

11.05

VDF

8.22

 

8.13

8.04

8.49

 

 

8.08

11.38

11.52

VF

8.48

8.65

8.44

8.47

8.71

12.78

10.86

8.51

11.86

11.91

 

 

 

 

 

 

 

 

 

 

 

 

MalR

MANR

MeOR

MMAR

StR

VacR

VCR

VDCR

VDFR

VFR

AA

6.23

11.07

7.23

11.62

9.88

6.59

6.60

10.41

4.87

5.45

AllC

7.12

11.82

7.84

11.63

12.24

8.27

7.51

10.93

4.54

6.29

AM

9.37

14.69

9.84

13.92

12.41

9.15

9.00

12.70

7.33

7.95

AMD

7.42

13.80

9.94

14.09

12.54

9.03

8.70

12.31

5.87

7.26

AN

8.29

12.70

7.39

13.03

11.62

7.56

8.47

11.71

5.22

6.51

BD

4.94

9.25

6.57

9.66

9.40

7.20

6.03

8.93

 

4.21

cHn

9.09

15.69

11.51

16.27

17.07

11.67

12.91

14.34

7.51

10.00

CHO

7.76

14.38

8.11

13.69

12.23

7.61

8.78

12.23

5.41

6.50

GMA

5.97

11.57

7.75

10.73

9.57

6.17

8.14

10.82

3.16

4.39

MAA

7.00

13.24

8.35

12.39

10.50

7.77

8.08

11.89

5.88

6.45

Mal

10.84

15.93

7.47

15.13

12.55

7.31

9.84

13.66

7.96

7.91

MAN

6.72

10.72

6.29

11.23

9.44

6.61

7.16

10.54

4.12

5.31

MeO

5.27

11.29

7.35

11.68

11.79

7.71

8.08

10.48

4.15

5.77

MMA

6.93

11.84

8.62

12.70

10.93

8.07

8.36

12.06

6.00

6.61

St

4.56

9.41

5.57

9.79

9.19

5.71

6.05

9.21

2.63

4.07

Vac

5.21

10.50

6.34

10.94

10.52

6.53

6.98

11.20

3.28

5.07

VC

7.29

12.00

8.05

12.32

12.14

8.03

8.21

11.27

4.62

6.29

VDC

7.01

 

6.42

11.41

10.50

5.55

7.23

10.54

4.38

5.44

VDF

7.95

 

7.25

11.84

11.05

 

7.74

11.28

4.46

5.86

VF

7.12

11.99

7.80

12.29

11.94

7.79

8.26

11.16

4.48

6.30

For the moment, nine transition states cannot be found. These all returned the error of not having reached SCF and have stopped. In this case, the result was the same even if it enlarged the value of ITRY. This error of characteristic one takes place easily at the time of the monomer (radical) of the symmetry of vinylidene chloride and vinylidene fluoride. When , such as Allyl chloride, butadiene, and cyclohexene, ( when reactivity is scarce) radical, I feel that this phenomenon has occurred for the radical at that time.

For example, if a monomer called cyclohexene regards as horizontally and goes, it needs a high value for activation energy with everybody 11kcal /mol or more. A Mal radical, a VDF radical, etc. may be able to react barely. It seems that it reacts and the made cyclohexene radical (cHnR) can react well with a styrene monomer, butadiene, GMA, etc. Although a styrene monomer reacts very easily, a styrene radical cannot react easily. This is explained to be because for the styrene radical to be stabilized by resonance. It says since what vinyl acetate (VAc) generally cannot receive a radical attack in easily has the unstable generated radical. But the calculation result of MOPAC has brought a result that VAc monomer react with some type of radicals . On the other hand, the reactivity of a Vinyl Acetate radical (VAcR) is high. The radical which will be generated once it opens, although it is hard to open double bond of VAc is very unstable, it is rich in reactivity, and it is explained that it has very high homopolymerization velocity.

I think that this talk will shift greatly when making an electron donor acceptor complex from the talk applicable to Free Radical Polymerization and reacting.

The reaction of such a radical and a monomer A textbook explains as follows.

Radical polymerization theory

"radical-polymerization theory" (Imoto P61- Japanese Polymer Text Book)

Radical SOMO, HOMO of a monomer, and orbital mixture start. Next, the anti-bonding orbital then made causes LUMO of a monomer, and orbital mixture.
Then, I think that a natural thing thinks that dE can be estimated from radical SOMO, HOMO and LUMO of a monomer.

Then, make such a table.

name

dE

RaSOMO

MonHOMO

MonLUMO

MAA-Mal

14.97

-9.682

-11.708

-1.548

AN-VDC

7.72

-9.855

-9.744

0.327

AA-AMD

12.31

-8.803

-9.794

0.143

AN-AM

9.65

-9.855

-11.067

-0.078

MAA-AA

11.18

-9.682

-11.147

-0.17

AN-AN

9.28

-9.855

-10.886

-0.187

CHO-VDF

8.08

-9.93

-10.759

0.227

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Vac-cHn

11.67

-8.785

-9.593

1.168

MeO-MeO

7.35

-8.352

-9.515

1.302

MMA-MeO

11.68

-9.566

-9.515

1.302

VC-AllC

7.51

-8.928

-10.269

0.621

MAA-Vac

10.36

-9.682

-10.083

0.687

BD-AA

11.08

-8.848

-11.147

-0.17

First of all, calculate this by multiple regression. Since a correlation coefficient is 0.29, there is almost no correlation.

Formula

dE=0.963*RaSOMO-1.099*MonHOMO+0.621*MonLUMO+6.852

It comes to be alike. This is the plot.

In a correlation coefficient 0.29, it is useless truly. If considering why being useless radical activity is high (RaSOMO being large), stable and low reactant monomer (MonHOMO -- small) can react well.
Probably, it will be natural that accuracy does not come out, since all of such an interaction are not taken into consideration in multiple regression. Then, analyze by the neural network in consideration of the interaction among the inputs.

The following results were brought, when the number of the middle layer neuron was set to 3 and it learned 715500 times with the bias neuron (reconstruction learning method).

It can't be said very much that precision is still high. Do the various tuning such as to increase the number of middle layer neurons, probably a great improvement will not be carried out.

For this reaction said to decide by radical SOMO and monomer HOMO,LUMO in schoolbook, This results means some factors of other are involved.

Other factor

For this reaction said to decide by radical SOMO and monomer HOMO,LUMO in schoolbook, this results means some factors of other are involved.

For example, it is the charge of Head carbon and the radical energy level of LUMO, and the charge of Tail carbon. It is also one means to put in all the factors to invent and to narrow down by principal component analysis. Since I consider the energy level of radical LUMO should participate in this reaction, I reconstruct a neural network once again by radical SOMO, LUMO, and HOMO,LUMO of a monomer.

The following results were brought, when the number of the middle layer neuron was set to 5 and it learned 490000 times with the bias neuron (reconstruction learning method).

If compared with multiple regression, it turns out that it was improved fairly. Next, what should you do? A lot of plan to improve result exist, such as changing the number of middle layer neurons or putting in a charge. However, if you wants to learn the neural network method, please hold your ground just for a moment.

At first, confirm what kind of data are shifted from the straight line.

name

dE

NN Answer It obtains.

Error Absolute value

RaSOMO

RaLUMO

MonHOMO

MonLUMO

cHn-AMD

11.68

7.69

3.99

-8.895

1.047

-9.794

0.143

AA-MeO

10.48

14.33

3.85

-8.803

0.160

-9.515

1.302

AMD-AA

6.70

10.53

3.83

-9.647

-0.480

-11.147

-0.170

AMD-MeO

7.91

11.67

3.76

-9.647

-0.480

-9.515

1.302

BD-cHn

18.01

14.46

3.55

-8.848

0.211

-9.593

1.168

Mal-MeO

5.27

8.68

3.41

-10.963

-1.509

-9.515

1.302

MAN-MeO

11.29

14.69

3.40

-9.486

-0.617

-9.515

1.302

VFR-AMD

7.26

3.96

3.30

-9.343

0.704

-9.794

0.143

Vac-Mal

7.31

10.61

3.30

-8.785

0.553

-11.708

-1.548

VDF-MeO

4.15

7.24

3.09

-9.874

0.113

-9.515

1.302

MeO-MeO

7.35

10.41

3.06

-8.352

1.433

-9.515

1.302

MMA-MeO

11.68

14.74

3.06

-9.566

-0.598

-9.515

1.302

AN-MeO

7.84

10.89

3.05

-9.855

-0.719

-9.515

1.302

AMD-GMA

6.72

9.71

2.99

-9.647

-0.480

-10.587

0.143

VFR-cHn

10.00

7.06

2.94

-9.343

0.704

-9.593

1.168

Vac-MeO

7.71

10.63

2.92

-8.785

0.553

-9.515

1.302

AMD-Vac

6.31

9.21

2.90

-9.647

-0.480

-10.083

0.687

cHn-AllC

11.03

8.14

2.89

-8.895

1.047

-10.269

0.621

cHn-CHO

11.62

8.73

2.89

-8.895

1.047

-10.681

-0.098

MAA-MeO

11.20

14.07

2.87

-9.682

-0.717

-9.515

1.302

Arrange the error (|neural network Calc. - MOPAC PM3 Calc.|) in the descending order. It turns out that ten of worst 20 pieces are the system of a methyl-vinyl-ether (MeO) monomer (marked yellow). And the neural network always predicts the value high about 3kcal/mol. It means that the MeO monomer structure whose heat of formation is about 3 kcal/mol low, will exist. If the structure where transition state is high 3kcal/mol exists, it is the same. But the reaction will pass the lowest level of TS, so you need not consider higher route of TS.

So I examined the structure of methyl vinyl ether (MeO).

In this case, the above structures of this figure are taken. Since the surroundings of ether oxygen were free rotation, when they were turned round and the following structures were taken, they were stable 1.607kcal/mol. So, I fix the dE table of every MeO monomer. Since the value of HOMO and LUMO also changes delicately, this is all also corrected.

And a neural network is calculated once again by the newly remade data. Thus, if it corrects and calculates and corrects, calculation is repeated for a long time.

If MeO is corrected the result will become such.

name

dE

NN Guess value

Absolutely Error

somo

lumo

MonHOMO

BD-cHn

18.01

13.17649712

4.833502884

-8.848

0.211

-9.593

cHn-AMD

11.68

6.985349722

4.694650278

-8.895

1.047

-9.794

BD-AMD

13.66

9.284698015

4.375301985

-8.848

0.211

-9.794

Vac-Mal

7.31

11.17093357

3.860933572

-8.785

0.553

-11.708

cHn-CHO

11.62

7.787294954

3.832705046

-8.895

1.047

-10.681

MMA-AMD

14.09

10.33804494

3.751955055

-9.566

-0.598

-9.794

AMD-AA

6.7

10.33514547

3.635145471

-9.647

-0.48

-11.147

MAA-AMD

13.71

10.1018309

3.6081691

-9.682

-0.717

-9.794

cHn-AM

11.72

8.193541698

3.526458302

-8.895

1.047

-11.067

cHn-MMA

11.12

7.658936543

3.461063457

-8.895

1.047

-10.553

GMA-AMD

13.5

10.1030179

3.396982101

-9.699

-0.74

-9.794

cHn-AllC

11.03

7.760870639

3.269129361

-8.895

1.047

-10.269

AMD-Vac

6.31

9.552774897

3.242774897

-9.647

-0.48

-10.083

Vac-MeO

9.311734

12.52751168

3.215777684

-8.785

0.553

-9.619

AA-AMD

9.07

5.94931352

3.12068648

-10.118

-0.814

-9.794

BD-CHO

13.23

10.1212422

3.108757801

-8.848

0.211

-10.681

AM-AMD

9.38

6.291948181

3.088051819

-10.017

-0.711

-9.794

Vac-AA

6.59

9.610234461

3.020234461

-8.785

0.553

-11.147

Vac-VDC

5.55

8.46468061

2.91468061

-8.785

0.553

-9.744

Vac-Vac

6.53

9.404180911

2.874180911

-8.785

0.553

-10.083

VDC-AMD

12.31

9.490990594

2.819009406

-8.803

0.16

-9.794

And again I examined which TS was bad. The result is that the TS using the monomer of acrylamide was bad. And all of these cases, the result of NN becomes lower 3kcal/mol respectively. Supposing the result of this NN is right, The transition state should be lower.

Mathematically it is same if the heat of formation of AMD monomer is higher 3kcal/mol but higher energy structure is not stable structure, so I need not take it account.

I use MOPAC 7 and MOPAC97, but version difference will not lead 3kcal/mol .

What I found is that the position of HOMO of amide monomer mistook in this case, and it led to the calculation error.

 

 

Root No.

12

13

14

15

16

 

 

 

12A

13A

14A

15A

16A

 

 

 

-10.925

-10.786

-9.794

0.143

1.62

S

C

1

-0.0217

-0.0922

0.0167

-0.0071

0.2072

Px

C

1

0.0252

0.1019

-0.0167

-0.008

-0.0712

Py

C

1

-0.0574

-0.2712

0.0426

0.0031

0.186

Pz

C

1

0.657

-0.1608

-0.0569

0.5128

0.1986

S

C

2

-0.0022

-0.0013

-0.003

0.019

-0.0901

Px

C

2

-0.0144

-0.0457

0.006

-0.0186

0.0878

Py

C

2

0.0132

0.0701

-0.0134

-0.0036

0.0475

Pz

C

2

0.6367

-0.1571

-0.0652

-0.6589

-0.1594

S

H

3

0.017

0.0793

-0.0135

-0.0067

-0.0308

S

H

4

-0.02

-0.1008

0.0161

-0.0097

0.097

S

C

5

-0.0038

-0.0389

-0.0204

-0.0005

-0.3267

Px

C

5

-0.0086

-0.0792

0.0425

-0.1061

0.2679

Py

C

5

0.0225

0.089

-0.0139

-0.0753

0.1662

Pz

C

5

-0.0868

-0.0111

0.052

0.3687

-0.2106

S

O

6

0.0041

-0.0018

-0.0001

-0.0037

0.0315

Px

O

6

0.259

0.7406

0.015

0.1013

-0.1683

Py

O

6

0.0315

-0.3488

0.1241

0.0823

-0.0908

Pz

O

6

-0.2592

0.1421

-0.4444

-0.3015

0.138

S

H

7

0.0301

0.1215

-0.0138

-0.0194

-0.0008

S

N

8

0.0484

0.1695

0.1169

-0.0214

0.4984

Px

N

8

0.0941

0.1827

-0.3405

0.0955

0.0597

Py

N

8

0.0089

0.0164

-0.2064

0.0427

-0.0007

Pz

N

8

-0.0116

0.1789

0.7713

-0.1682

-0.0181

S

H

9

-0.0273

-0.1036

-0.0452

0.0268

-0.358

S

H

10

-0.0331

-0.0871

-0.0438

0.0165

-0.3695

The electron filled level-- RootNo14 -- therefore the energy level of LUMO ( 0.143) is O.K. Although RootNo.14( -9.794) was adopted HOMO, this was Pz orbital of nitrogen of amide (Marked yellow 0.7713). The orbital of double bond of C=C was located in the place of -10.925 ( RootNo.12). So I fixed this error.

Moreover, the monomer of lower energy found acrylic acid as well as MeO, and it also corrected this.

Systematic error was now corrected.
Next, Activation energy error individually.

individual error

Systematic error was now corrected.
Next, Activation energy error individually.

Middle layer 5

 

 

 

 

Middle layer 4

 

 

 

 

Middle layer 3

 

 

 

name

dE

Answer

d

 

name

dE

Answer

d

 

name

dE

Answer

d

Vac-Mal

7.31

10.45

3.14

 

Vac-Mal

7.31

11.45

4.14

 

St-cHn

17.07

11.81

5.26

VDC-MeO

12.09

15.17

3.08

 

Vac-MeO

9.31

13.40

4.09

 

BD-cHn

16.90

12.02

4.88

AMD-Vac

6.31

9.16

2.85

 

cHn-CHO

11.62

7.58

4.04

 

GMA-cHn

15.92

11.52

4.40

VDF-GMA

3.16

5.95

2.79

 

cHn-AllC

11.03

7.23

3.80

 

VFR-cHn

10.00

5.77

4.23

AMD-GMA

6.72

9.40

2.68

 

AN-GMA

6.25

9.85

3.60

 

Mal-Vac

5.21

9.16

3.95

VDF-MeO

5.75

8.44

2.68

 

cHn-MMA

11.12

7.52

3.60

 

MMA-cHn

16.27

12.34

3.93

MeO-Mal

7.47

10.09

2.62

 

Vac-GMA

6.17

9.73

3.56

 

MAA-cHn

15.78

11.85

3.93

AN-GMA

6.25

8.74

2.49

 

AMD-GMA

6.72

10.07

3.35

 

GMA-Mal

15.87

12.02

3.85

VC-cHn

12.91

10.43

2.48

 

cHn-AMD

11.68

8.40

3.28

 

VC-cHn

12.91

9.26

3.65

VDF-Vac

3.28

5.76

2.48

 

AMD-Vac

6.31

9.56

3.25

 

AMD-Vac

6.31

9.91

3.60

Mal-MeO

6.87

9.23

2.36

 

Vac-Vac

6.53

9.76

3.23

 

MAN-Mal

15.93

12.46

3.47

MMA-MeO

13.28

15.63

2.35

 

cHn-VC

10.72

7.52

3.20

 

AN-cHn

12.41

8.96

3.45

BD-cHn

16.90

14.55

2.35

 

AMD-Mal

10.25

13.38

3.13

 

MAN-cHn

15.69

12.27

3.42

AMD-MeO

9.52

11.83

2.32

 

AMD-MeO

9.52

12.58

3.07

 

AM-Mal

12.56

9.17

3.39

VDF-VF

4.48

6.76

2.28

 

cHn-VF

10.86

7.82

3.04

 

Vac-VDC

5.55

8.90

3.35

MMA-GMA

10.73

12.97

2.24

 

VC-MeO

9.68

12.65

2.97

 

AMD-GMA

6.72

10.06

3.34

AMD-Mal

10.25

12.47

2.22

 

cHn-AM

11.72

8.77

2.95

 

Vac-GMA

6.17

9.44

3.27

GMA-cHn

15.92

13.70

2.22

 

Vac-VDC

5.55

8.41

2.86

 

Mal-GMA

5.97

9.23

3.26

St-cHn

17.07

14.86

2.21

 

BD-cHn

16.90

14.06

2.84

 

cHn-MeO

12.40

9.27

3.13

AA-MeO

8.98

11.17

2.18

 

cHn-Vac

10.09

7.26

2.83

 

VDF-GMA

3.16

6.16

3.00

It is the way said that the way which he does well then changes the number of a neural network's middle layer neurons, prepares a result a number of points, is investigated in an order from what has an always big error, and goes in it. In this case, the middle layer neurons was examined by 5, 4, and 3. the middle layer neurons of a with error absolute value decreases -- be alike and hang -- although it becomes large -- for example, AMD-VAc, AMD-GMA, and BD-cHn -- always -- worst one -- it enters in 20. Three, in the large thing, an error is purple and is marked yellow for two large. And regardless of such an error is large and the number of the middle layer neurons, it reexamines in an order from a bad thing.

It is a bad miss, when considering and beginning reexamination. He forgot to correct the level of SOMO of an AMD radical to the level of HOMO of an AMD monomer coming to the nitrogen atom, and having corrected. It seems that therefore, the AMD radical also had the systematically large error. The middle class is rejected once again and it is reinquiring.

Middle class 3

 

 

 

 

Middle class 4

 

 

 

 

Middle class 5

 

 

 

name

dE

Answer

d

 

name

dE

Answer

d

 

name

dE

Answer

d

VC-cHn

12.91

8.05

4.86

 

BD-cHn

16.90

12.68

4.22

 

St-cHn

17.07

13.72

3.35

cHn-CHO

11.62

7.13

4.49

 

AN-GMA

6.25

9.89

3.64

 

Vac-Mal

7.31

10.45

3.14

cHn-MeO

12.40

8.05

4.35

 

AMD-cHn

11.82

8.45

3.37

 

GMA-cHn

15.92

13.07

2.85

BD-cHn

16.90

12.68

4.22

 

BD-CHO

13.23

9.86

3.37

 

BD-cHn

16.90

14.08

2.82

cHn-MMA

11.12

7.06

4.06

 

Vac-Mal

7.31

10.29

2.98

 

MeO-Mal

7.47

10.17

2.70

cHn-cHn

11.91

7.92

3.99

 

BD-VC

12.78

9.91

2.87

 

AM-Mal

12.56

9.95

2.61

AMD-cHn

11.82

7.84

3.98

 

Mal-MeO

6.87

9.70

2.83

 

GMA-MAA

13.46

10.85

2.61

cHn-AllC

11.03

7.08

3.95

 

BD-AMD

13.66

10.90

2.76

 

VDC-MeO

12.09

14.68

2.59

St-cHn

17.07

13.19

3.88

 

St-GMA

9.57

12.31

2.74

 

MAN-CHO

14.38

11.79

2.59

cHn-AMD

11.68

7.89

3.79

 

AN-MeO

9.44

12.17

2.73

 

MAA-cHn

15.78

13.23

2.55

VDF-Vac

3.28

6.97

3.69

 

VFR-GMA

4.39

7.10

2.71

 

cHn-CHO

11.62

9.10

2.52

MMA-cHn

16.27

12.62

3.65

 

BD-MMA

12.72

10.03

2.69

 

AA-VDC

7.59

5.13

2.46

VDF-GMA

3.16

6.80

3.64

 

AN-Vac

6.82

9.46

2.64

 

AMD-cHn

11.82

9.36

2.46

cHn-VC

10.72

7.14

3.58

 

St-Mal

12.55

15.09

2.54

 

Mal-MeO

6.87

9.32

2.45

CHO-cHn

11.52

8.02

3.50

 

BD-MAA

12.29

9.75

2.54

 

AA-CHO

9.18

6.73

2.45

MeO-cHn

11.51

8.04

3.47

 

AN-MAN

7.09

9.62

2.53

 

MeO-Vac

6.34

8.76

2.42

cHn-VF

10.86

7.41

3.45

 

St-MeO

13.40

15.89

2.49

 

AA-cHn

11.30

8.93

2.37

AM-cHn

11.39

7.96

3.43

 

MMA-cHn

16.27

13.92

2.35

 

AA-AllC

8.65

6.31

2.34

AA-cHn

11.30

7.93

3.37

 

GMA-cHn

15.92

13.61

2.31

 

AM-cHn

11.39

9.10

2.29

GMA-cHn

15.92

12.57

3.35

 

cHn-CHO

11.62

9.33

2.29

 

VC-cHn

12.91

10.64

2.27

Three, in the large thing, an error is purple and is marked yellow for two large. In the error, the cyclohexene (cHn) monomer is involving greatly from door to door this time. And the value is wholly larger than the predicted value of the neural network(NN) method. If activation energy has shifted to the larger one, it is in the place where a transition state is lower, or is the case that the heat of formation of a monomer is higher. The neural network will think that it should have a lower transition state since a thing when the heat of formation of a monomer is higher does not need to think. Though regrettable, the ways of coping of this are not found. Having marked on the place of dE in orange expresses that in which the still lower transition state was found. For example, the mark has been carried out to 13.46 of GMA-MAA by the middle layer neurons 5.

If this structure is seen, as shown in the upper figure, hydrogen of methacrylic acid has turned to the outside. If a transition state is calculated by turning an inner side for this as follows, it will become 1.5kcal /mol stability. Then, 13.46 is set to 12, and since prediction of a neural network is 10.85, what an error checks each structure to the 1.15, and is corrected corrects and goes. Most things will have improved at least mol in 1.5kcal /from 0.5, if the twist angle adjusted it bad.

When a neural network returns a big value, it is the case where there is structure with more stable monomer and radical. When the monomer (radical) is merely used in that case, only the always same quantity becomes large. When it sees roughly and there is no such tendency, it ignores for the time being.

The relative position of double combination of double combination of BD radical and a monomer, C=O of aldehyde, the spatial relationship of double combination, etc. are adjusted, and a new table is obtained.

Next, use this new table. Radical SOMO and Head carbonCharge, HOMO, LUMO, and Tail carbon charge of a monomer and Build a neural network.

neural network

use this new table. Radical SOMO and Head carbon Charge, HOMO, LUMO, and Tail carbon charge of a monomer and Build a neural network.

For example, when a vinyl acetate radical reacts to a maleic anhydride monomer, P orbit of the same phase as the place of double bond of maleic anhydride comes. The orbit interchanged and the knot is made of LUMO. vinyl acetate -- although there is the orbital biggest Orbit P into the portion of oxygen of maleic anhydride in lap when radical SOMO reacts -- here -- a charge ---like -- minus -- large -- Head of vinyl acetate -- it is disadvantageous that radical carbon also reacts here since a charge is minus. Then, a vinyl acetate radical reacts to double combination of maleic anhydride. Therefore, if the charge of radical Head carbon and the charge of Tail carbon of a monomer are put in and a neural network is constructed, thinking that activation energy can be guessed with more sufficient accuracy makes sense.

(As for vinylidene fluoride (VDF), in having calculated this time, only Head radical carbon is actually added.) And the activation energy of VDF-St is the minimum in 2.63 and this time. Activation energy with other monomers also brings the result that 7.96 (at the time of maleic anhydride) and which monomer react well at the maximum. If it merely does so, the radical of VDF cannot explain to itself with what whether oxygen of maleic anhydride is attacked. Although it is the thing which seems to be advantageous also in lap also in charge as for an orbit ..... strange about there is the hardness of those who have not done the molecular orbital method specially. If the direction which someone can explain comes, I will need your help.

name

dE

RaSOMO

RaHCharge

MonHOMO

MonLUMO

MonTCharge

AllC-AA

8.910044

-9.321

-0.2344

-11.147

-0.17

-0.0547

AA-MeO

8.984206

-10.118

-0.1489

-9.619

1.25

-0.2969

MMA-MeO

13.28327

-9.566

-0.1579

-9.619

1.25

-0.2969

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AN-Vac

6.82

-9.855

-0.0478

-10.083

0.687

-0.2102

VDF-AMD

5.87

-9.874

0.0572

-10.925

0.143

-0.0608

AllC-GMA

7.2

-9.321

-0.2344

-10.587

0.143

-0.095

The following results were brought, when 391 data was prepared in such a form, the number of the middle class was set to 5 and it learned 1490000 times with the bias neuron (reconstruction learning method).

Now, the result that it was worse than the case where radical SOMO, LUMO, and HOMO of a monomer and LUMO are put in, fairly was brought. If radical LUMO does not have this for an important numerical value too, activation energy is being unable to guess, even if it uses a neural network.

Then, it re-calculated by extracting what finally puts in radical SOMO, LUMO, HOMO of a monomer and LUMO, Head, and Tail carbon charge, and constructs a neural network, and a calculation result does not suit (repeatedly). 133 of 391 of the activation energy in the first table pieces were corrected after all.

The neural network finally obtained thinks that it can turn out that activation energy can be guessed only from the molecule information on a radical and a monomer very good. What it has separated from most greatly is BD-AMD. A reason is not known well.

Even if the activation energy therefore obtained by calculating a transition state by MOPAC uses a neural network, it can be guessed in the accuracy of there there. Then, how many realities were reproducible with PM3 of MOPAC primarily? I want to verify in relation to [ for photoresist ] a monomer especially.

 

PM3 acuracy

What calculation accuracy is it as molecular orbital calculation of MOPAC PM3? Although it may say that accuracy is not expected from bases, such as a semiempirical molecular orbital method, there are also one cup of merits, such as speed of calculation and a molecular size which can be treated. Moreover, if it is a relative value even if the absolute value of a rate of reaction is not known, it is made to compare with one r1r2 value of a radical reaction to what extent it goes.

r1 and r2 are the elementary processes of radical polymerization.

M1 and +M1-> M1- Reaction rate constant k11

M1, +M2-> M2, and k12

M2, +M1-> M1, and k21

M2, +M2-> M2, and k22

When it carries out,

It defines as r1=k11/k12 r2=k22/k21, and there are some which were collected into the polymer handbook, the textbook, etc. about various combination.

It is a reaction rate constant kij here.

kij=Aij*e(-dE/RT)

It is made the formula of Arenius to say and MOPAC and the activation energy for which it asked by PM3 are put into the place of dE. Thinking that Aij called a frequency factor will cancel oscillating calculation of MOPAC by the denominator and a molecule since it will not change a lot by the thing with low accuracy, and two kinds of monomers although originally computed from the entropy acquired from oscillating calculation, Aij ignores here. If it does so

r1=e(-dE1/RT)/e(-dE2/RT)

It will come out and the clause of temperature will also disappear. If the calculation result of PM3 is arranged, it will become like this.

PM3

StR

MMAR

ANR

VDCR

VCR

VacR

malR

AMR

VFR

St

1.00

0.05

0.04

0.26

0.11

0.44

0.00

0.03

0.11

MMA

5.72

1.00

0.22

4.59

1.16

4.66

0.02

0.51

1.36

AN

11.40

1.39

1.00

3.22

1.29

2.78

0.08

0.61

1.23

VDC

3.73

0.28

0.21

1.00

0.38

0.37

0.02

0.13

0.42

VC

19.12

0.68

0.59

2.08

1.00

4.46

0.03

0.32

0.99

Vac

3.80

0.17

0.09

1.95

0.29

1.00

0.00

0.07

0.29

Mal

28.78

11.39

16.34

22.74

5.10

2.18

1.00

20.83

4.96

AM

25.05

3.41

1.45

8.70

2.21

13.77

0.23

1.00

5.19

VF

15.75

0.66

0.57

1.86

1.05

3.52

0.02

0.34

1.00

When a styrene radical (StR) sets styrene (St) and velocity which reacts to 1 when done so for example, 5.72 can make in MMA the table which is expressing and carrying out that St tends to react 5.72 times compared with MMA.

The reference value of a reactant ratio

r1r2

StR

MMAR

ANR

VDCR

VCR

VacR

MalR

AMR

VFR

St

1.00

0.46

0.33

0.09

0.02

0.10

0.10

0.18

 

MMA

0.52

1.00

0.15

0.25

0.07

0.02

0.02

0.34

 

AN

0.40

1.20

1.00

0.60

0.04

0.06

0.00

0.95

0.005-1

VDC

2.00

2.75

0.80

1.00

0.30

0.05

0.00

0.65

0.16

VC

17.00

16.10

2.70

3.20

1.00

0.23

0.01

4.40

0.11

Vac

55.00

20.00

4.05

6.70

1.68

1.00

0.00

9.00

0.25

Mal

0.01

3.50

6.00

9.00

0.30

0.06

1.00

2.80

 

AM

0.75

1.69

1.40

0.85

0.12

0.10

0.02

1.00

0.01

VF

 

 

24-44

6.00

11.60

3.50

 

43.00

1.00

It becomes such. A reference value also changes a lot by the difference in a solvent, the difference in an initiator, and the difference among polymerization conditions. The pink color was given for the thing included in less than (1/3) 3 times. Since the percentage of victories is 20/66 noting that it is disregarded, since a diagonal clause is 1.0, it exceeds 30 percent just for a moment. Especially it returns as a general trend, it did not necessarily say that it was good and is scattered uniformly.

In the case of maleic anhydride prominent especially as alternating copolymerization, to the monomer of maleic anhydride, since a styrene radical is 0.01, by r1 of an experimental value, it reacts to maleic anhydride easily 100 times. It turns out that it turns out that it reacts, and it reacts to a styrene monomer easilier 10 times than a maleic anhydride monomer since the radical of the made maleic anhydride is 0.1, and alternating copolymerization of this system is carried out after this. In the calculation result of MOPAC PM3, since a styrene radical is 28.78, it reacts to a maleic anhydride monomer easily in the direction of a styrene monomer. Therefore, if it is concentration of the same monomer and 28-piece styrene will be connected, it is greatly shifted from reality in the result of PM3 that it returns to a styrene chain immediately since the maleic acid radical which was able to react and do one maleic anhydride is 0.00 to styrene.

Although vinyl acetate maleic anhydride is also the system of alternating copolymerization typical at r1=0.06 and r2=0.00, since it is r1=2.18, by the calculation result of PM3, it does not become alternating copolymerization.

In the transition state database, after searching for a transition state by MOPAC, it is re-calculating by B3LYP/6-31G**. When arguing about fixed-quantity nature too, in a semiempirical molecular orbital method, it is impossible. Then, it recalculated by B3LYP/6-31G**.

B3LYP

StR

MMAR

ANR

VDCR

VCR

VacR

malR

AMR

VFR

St

1.00

0.34

0.20

1.10

0.16

13.13

0.00

0.59

0.03

MMA

0.30

1.00

0.70

0.13

0.07

0.05

0.00

0.61

0.00

AN

0.05

0.09

1.00

0.11

0.03

0.01

0.04

2.13

0.00

VDC

1.22

1.20

2.42

1.00

0.24

0.12

0.02

2.26

0.02

VC

8.98

7.16

9.18

1.66

1.00

1.01

0.06

16.86

0.17

Vac

20.36

5.69

5.86

1.61

2.66

1.00

0.02

11.42

0.29

Mal

0.03

0.38

229.14

4.89

0.08

0.00

1.00

2.87

0.00

AM

0.11

0.30

0.80

0.11

0.10

0.11

0.01

1.00

0.01

VF

121.45

14.60

34.43

2.94

6.33

7.71

0.06

25.80

1.00

It is over 60 percent by the percentages of victories 40/66 this time. A result bad at the time of the monomer of St, AN, and Mal is brought. Since it can still predict that both maleic anhydride and vinyl acetate maleic anhydride also becomes alternating copolymerization, it can be said that it is so so accuracy. Does the thing with bad calculation result in this MOPAC PM3 and coincidence of the reactant ratio r1 depend on the frequency factor having been disregarded? Or is the activation energy dE itself bad? Plot dE of PM3, and dE of B3LYP/6-31G** to a trial.

As long as this is seen, in calculation of PM3, it turns out that the dE itself is not good. It turned out that a reactant ratio is unreproducible by the cause of both that the accuracy of dE is low, and the thing for which the frequency factor was disregarded.

Then, it is in order to guess a reactant ratio by calculation. Which Level Ab Initio calculation?

Top page
Frequency View of TS
Edit TS Structure
 How to Use
Check Frequency

 How to use
MOPAC Calculation
Ab initio calculation
Monte Carlo simulation
Polymer properties estimation