![Page 1: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/1.jpg)
COMPOSITE METHODS
![Page 2: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/2.jpg)
COMPOSITE METHODS• Based on additivity (independence) of
– Basis set improvements– Level of theory improvements
• High level = H [e.g., CCSD(T)]• Low level = L [e.g., MP2]• Big basis = B [e.g., aug-cc-pVTZ]• Small basis = S [e.g., cc-pVDZ]• D1=L(B) - L(S): Basis set improvement
• D2=H(S) - L(S): Level of theory improvement
• H(B)=H(S) + D1 + D2
– Approximate additivity avoids direct calc of H(B)– Mostly applicable to thermodynamics
![Page 3: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/3.jpg)
![Page 4: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/4.jpg)
Predicts heats of formation to ~ 1kcal/molNot good for barrier heights, rxn paths
![Page 5: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/5.jpg)
In GAMESS (Sean Nedd)
G2 (MP2)G3 (MP2)
QCI replaced by CR-CC(2,3)
![Page 6: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/6.jpg)
GAMESS G3 input file
RUNTYP=G3MP2
See EXAM43.inp
![Page 7: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/7.jpg)
OTHER COMPOSITE METHODS
• Weizman Wn methods (n=1,2,3,4) – Jan Martin at Weizman Institute, now UNT
• A. Karton, E. Rabinovich, and J. M. L. Martin, J. Chem. Phys. 125 (144108), 144108 (2006).
– Uses cc basis sets up to aug-cc-pV5Z– No empirical parameters, very time-consuming– Extrapolation to complete basis set limit
• Not yet in GAMESS
Predicts heats of formation to ~ 1kcal/molNot good for barrier heights, rxn paths
![Page 8: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/8.jpg)
OTHER COMPOSITE METHODS• ccCA Methods (Angela Wilson, UNT)
– N. J. De Yonker, B. R. Wilson, A. W. Pierpont, T. R. Cundari, and A. K. Wilson, Molecular Physics 107 (12), 1107 (2009): ccCA-S4• Uses cc basis sets through aug-cc-pVQZ• Extrapolation to complete basis set limit• No empirical parameters
– Modified version in GAMESS• Replaces CCSD(T) with CR-CC(2,3)• Thermodynamics as reliable as ccCA-S4• Accurate for diradical reaction paths due to CR-CC(2,3)• Better than Gn; less demanding than Wn
![Page 9: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/9.jpg)
• SR=scalar relativistic correction• CC=coupled cluster correction
– CCSD(T) in original ccCA-S4– [CR-CC(2,3)] in ccCA-CC(2,3)
• CV=core-valence correlation correction• MAD for ccCA-S4 (G2 test set) ~ 1.1 kcal/mol• MAD for ccCA-CC(2,3) ~0.9 kcal/mol
![Page 10: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/10.jpg)
PATHWAYS FOR PERICYCLIC REARRANGEMENT
![Page 11: COMPOSITE METHODS. Based on additivity (independence) of –Basis set improvements –Level of theory improvements High level = H [e.g., CCSD(T)] Low level](https://reader036.vdocuments.mx/reader036/viewer/2022062500/5697c0281a28abf838cd7195/html5/thumbnails/11.jpg)
- CCCA-S4- CCCA-CC(2,3)