PB' Done Carefully'

1
PB. Done Carefully.

• (How Else?)

2
(No Transcript)
3
With PB
4
Trying with PB
5
Parameterized ZAP Radii
SAS6 cals/A2, Single atom fix H20, CH4, SH2,
NH3
6
Remainders 20 of Variance
7
Why Pyrazine?
8
AM1BCC charges on Pyridine and Pryazine
ZAP -5.5 kcals, Expt -4.7
ZAP -8.4 kcals, Expt -5.6
9
(No Transcript)
10
DFT/cc_pVTZ charges on Pyridine and Pyrazine
ZAP -4.0 kcals, Expt. -5.6
ZAP -4.2 kcals, Expt.-4.7
11
Why Acetamide?
12
Why (N-methyl) Acetamide?
13
The N-methylamine Problem
NH3 NH2CH3 NHC2H6 NC3H9
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
Published 1963 7,000 Values 6,000 Compounds 2,178
Citations
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
Cyclopentanal
Benzaldehyde
22
DFT-cc_pVTZ
AM1BCC
23
Back of the Envelope

• N3 correction failed 1.8kcals
• Unnecessary carbonyl correction added 1.2kcals
• Expected error 3 kcals, Actual error 4 kcals

24
Nitros
25
(No Transcript)
26
Adjustments were also made to the N3hdelocdeloc
O1,2 BCC used for the nitro containing
functionalities which appear to be
problematicC. Bayly (italics added)
27
N3hdelocdeloc O1,2 BCC-0.1500
N3hdelocdeloc O1,2 BCC 0.1203
28
(No Transcript)
29
N3hdelocdeloc O1,2 BCC-0.06
30
(No Transcript)
31
The Halogens, aka Stalingrad
32
All Halide RMS Error1.30
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
My 2nd Favorite Graph
All Halide RMS Error1.43
40
Conclusions

• What did I learn?
• Peter knows how to chose wicked compounds
• Big outliers are always
• Associated with parameter failures
• Inevitable with parameters
• AM1BCC is parameterized for simulations not
• PB
• Reality (dipole moments)
• Make progress? Ditch the parameters
• Cant wait until next year!

41
The story so far..

• AM1BCC is geared to help simulations, not PB or
  to agree with vacuum dipoles
• Better dipoles (Aldehydes, Nitros) Better ZAP
• Better dipoles (Iodine) Thats funny

42
My 2nd Favorite Graph
All Halide RMS Error1.43