Related Experiment Videos
A natural linear scaling coupled-cluster method
1Quantum Theory Project, Department of Chemistry and Physics, University of Florida, Gainesville, FL 32611, USA.
The Journal of Chemical Physics
|January 7, 2005
Summary
A new natural linear scaling coupled-cluster (NLSCC) method enables accurate calculations for large molecular systems. This approach uses localized molecular orbitals (LMOs) to sum correlation energies from small subunits, achieving linear scaling for electronic correlation energy computations.
Area of Science:
- Computational Chemistry
- Quantum Chemistry
- Materials Science
Background:
- Coupled-cluster methods are accurate for electronic structure but computationally expensive for large systems.
- Scaling limitations hinder the application of high-accuracy methods to extended molecular structures.
Purpose of the Study:
- To develop a linear scaling coupled-cluster (NLSCC) method for efficient calculation of electronic correlation energies in large systems.
- To demonstrate the transferability of local correlation energies using a localized molecular orbital (LMO) basis.
Main Methods:
- Implementation of a natural linear scaling coupled-cluster (NLSCC) approach using a localized molecular orbital (LMO) basis.
- Calculation of coupled-cluster singles and doubles (CCSD) wave functions and energies on small system subunits.
- Summation of local occupied orbital correlation energies from subunits to obtain total correlation energy for extended systems.
Main Results:
- The NLSCC method achieves natural linear scaling for electronic correlation energy calculations.
- Demonstrated accuracy by comparing results for polyglycine molecules with large-scale LCCSD calculations.
- Successfully applied to alkane and polyglycine systems, showing the method's applicability to nonperiodic extended systems.
Conclusions:
- The NLSCC method provides an efficient and accurate way to compute correlation energies for very large molecular systems.
- The use of LMOs and local correlation energies enables linear scaling, overcoming limitations of traditional CCSD methods.
- This approach allows for the treatment of extended systems with effectively infinite basis set sizes.