We limit our discussion here to the HF method, but the general principles outlined also applies to higher level methods.
The correct description of the electronic state of temporary anions usually requires the use of extra diffuse functions in the basis set. You may thus consider having enough extra diffuse functions in your basis set;
It is best you use the core Hamiltonian as guess for the CAP-HF SCF procedure (i.e. set SCF_GUESS to ‘CORE’);
For CAP calculations, Q-Chem first solves the SCF problem using the old SCF drivers in Q-Chem; the solution is then used as the starting point for the CAP part of the calculation. This preliminary SCF solution usually leads to wrong CAP-HF solutions. You may avoid this issue by setting the rem variable SKIP_OLD_SCFMAN to ‘TRUE’. See Example 9.12.3;
Make sure the CAP contribution to the real and imaginary parts of the SCF solution found is negligible. You can get these CAP contributions by asking for an energy decomposition of the complex energy assigning the value ‘1’ to the rem variable CS_SCF_FINAL_PRINT. An alternative is to look at how close the imaginary (and real) parts of the CAP-HF energy and the corrected CAP-HF energy are; the ratio of the imaginary parts should be close to unity. See Example 9.12.3.
For CAP-HF calculations, set the rem variable CS_STRICT to ‘TRUE’ to print out the correct properties (e.g. Mulliken charges and multipole moments) of the solution. See Example 9.12.3.