Firefly and PC GAMESS-related discussion club

Learn how to ask questions correctly

Alex Granovsky
gran@classic.chem.msu.su

this message means that the current approximation to solutions
of TDDFT transition energies (and most likely the exact solutions
as well) are imaginary. Most often, this is caused by the
instability of solution (orbitals) of the reference KS equations.

You can try to use TDA approximation to TDDFT equations.
If you are lucky, you'll get some positive number of
questionable value. If you are not, you'll get negative
excitation energy.

Alternatively, if you are interested in excited states of
some specific symmetry type, use proper ISTSYM and solve TDDFT
problem for specific irrep only. This may or may not help depending
on the exact reasons causing TDDFT excitation energy to become imaginary.

See also the following link:

http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C3d4763b0bbHW-6340-846+00.htm

Regards,
Alex Granovsky




On Tue Jun 29 '10 12:36pm, Andrii Kulinich wrote
------------------------------------------------
>Dear All,
>trying to calculate the theoretical spectrum of D2-symmetry anionic squaraine dye, I've faced with the problem which I cannot solve.
>The checking of input-file (is attached here) ends normally. But full processing interrupts after calculation of basic molecular properties with the lines:

>>                    ------------------------------------
>>                    ADIABATIC TIME DEPENDENT DFT ENERGY
>>                    PROGRAM WRITTEN BY ALEX A. GRANOVSKY
>>                    ------------------------------------
>>
>> # CORE ORBITALS      =   56
>> # OCCUPIED ORBITALS  =  139
>> # MOLECULAR ORBITALS =  880
>> # BASIS FUNCTIONS    =  880
>>
>> NUMBER OF TDDFT SPIN-ADAPTED ANTISYMMETRIZED PRODUCTS (SAPS) IS   95215
>>
>> (RE)GENERATING XC FUNCTIONAL SECOND DERIVATIVES FOR TDDFT CALCULATIONS...
>>
>>                         TIMING STATISTICS ON NODE 0:
>> CPU        TIME:   STEP =     14.39 ,  TOTAL =     1644.3 SECONDS (   27.4 MIN)
>> WALL CLOCK TIME:   STEP =     14.41 ,  TOTAL =     1644.7 SECONDS (   27.4 MIN)
>> CPU UTILIZATION:   STEP =     99.83° TOTAL =      99.97SCBR>>
>> MIN MEMORY REQ. FOR TDDFT ENERGY FOCK-LIKE BUILDS =   1548804 WORDS
>> MEMORY REQ. FOR SINGLE BATCH BUILDS               =  18585604 WORDS
>> MEMORY AVAILABLE                                  = 189213536 WORDS
>>
>> SINGLE BATCH ENERGY CALCULATION WILL BE PERFORMED
>>
>> UNIT VECTOR GUESS AT TDDFT COEFICIENTS ...
>> Error: non positive-defined subspace A-B matrix!

>Please, explain me my mistake.
>

[ This message was edited on Thu Jul 1 '10 at 7:52pm by the author ]