Home

GAMESS-UK USER'S GUIDE and REFERENCE MANUAL Version

1. CONF 01234 1516 21 22 27 01235 1516 21 22 27 END The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE CUCL 3 21G ZMAT ANGSTROM CU CL 1 CUCL VARIABLES CUCL 2 093 END BASIS 3 21G RUNTYPE CI CORE 1 TO 14 END MRDCI DIRECT TABLE SELECT CNTRL 18 SPIN 1 SYMM 1 SINGLES 1 CONF 01234 1516 21 22 27 01235 1516 21 22 27 END CI NATORB ENTER 19 DATA FOR SEMEDIRECT TABLE CI SELECTION Example 3 66 Consider performing a valence Cl calculation on the SiH molecule using a 6 31G basis While the molecular symmetry is Tg the symmetry adaptation and subsequent Cl will be conducted in the Ca point group An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 1 1 68 77130710 2 1 6 12943325 3 2 4 23503117 4 3 4 23503117 5 4 4 23503117 6 1 0 73046864 7 4 0 48480821 8 3 0 48480821 9 2 0 48480821 10 2 0 16291387 11 3 0 16291387 12 4 0 16291387 13 1 0 25681257 14 0 33606346 15 3 0 37087856 16 2 0 37087856 17 4 0 37087856 18 al 0 79946861 19 1 0 79946861 20 4 0 86232544 21 3 0 86232544 22
2. 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 1 1 20 58952765 2 1 11 35779935 3 1 1 43525479 4 1 0 87463564 5 3 0 70990765 6 1 0 64751394 7 2 0 53989416 8 3 0 44423257 9 2 0 10853108 10 1 0 25726604 11 1 0 28106873 12 3 0 38903939 13 3 0 40966861 14 2 0 46216570 15 1 0 65466944 16 1 0 82879998 17 2 0 98111608 18 1 0 98701051 19 3 1 07064863 20 1 1 16621340 21 3 1 29856111 22 1 1 82320845 23 1 23 76352004 24 1 43 36689896 55 Thus the orbitals of interest are of common IRrep a1 with sequence numbers 1 2 core and 23 24 complement MOs within the SCF orbital set The following ACTIVE and CORE data would freeze and discard these MOs ACTIVE 3 TO 22 END CORE 12 END The following sequence would be used to simply freeze the orbitals while retaining the complete virtual manifold ACTIVE 3 TO 24 END CORE 12 END Note that if no orbitals are to be discarded as in the example above the user may omit the ACTIVE directive with the CORE specification acting to define this set Thus the above data sequence is equivalent to merely presenting the sequence 17 DATA FOR SEMEDIRECT TABLE CI INTEGRAL TRANSFORMATION 56 CORE 12 END Example 2 In this example we wish to perform a valence Cl calculation on
3. 36641681 41844853 45487336 49648833 49708917 60222488 64242537 76465797 82560838 10140194 20630804 30189632 35219192 50761510 76609415 01493168 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 42 14 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the Cls inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos ag 1 8 1 7 1 7 bau 2 2 0 2 8 9 bou 3 4 0 4 10 13 biu 5 8 1 7 14 20 bay 6 2 0 2 21 22 b3y 7 4 0 4 23 26 13 ITERATIVE NATURAL ORBITAL CALCULATIONS 43 Note that within the DZ basis employed there are no basis functions of bj IRrep 4 or a IRrep 8 symmetry The symmetry re ordered sequence numbers of the ground state orbitals allowing for the effective removal of the la and 1bj orbitals are 1 14 10 2 23 and 8 respectively To perform a single reference 12 electron valence Cl calculation based on the SCF configuration would require the following CONF data CONF 0128 10 14 23 The following data will perform this Cl where e the SCF computation is BYPASS ed e CORE is specified on the TRA
4. THRESH 5 0 5 0 THRESH 5 5 are equivalent causing Tmin and Tinc to be set to 5 microhartree 20 Data for Semi direct Table CI Eigen Solution Data input controlling the semi direct construction and diagonalisation of the Cl Hamiltonian is introduced by the Cl directive The process of extrapolation to zero selection threshold involves the Cl module solving two secular problems the first corresponding corresponding to the selection threshold TMIN the second to the threshold TMIN TINC In default the module will generate NROOT eigenvectors of the Cl matrix on both passes where NROOT is the number of roots specified by the ROOTS directive at selection time Thus the solutions of the zero order Hamiltonian will be used through a maximum overlap criterion in deriving the final Cl eigenvectors Additional data may be specified to provide various convergence and printing controls 20 1 CI The Cl directive comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string Cl e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output FPRINT or DEBUG to produce output suitable for debugging purposes e TEXTB is an optional parameter that shoul
5. e SELECT configuration generation and subsequent selection based on a user specified set of reference configurations and appropriate thresholds e Cl provides pre processing prior to the semi direct evaluation of the Cl eigenfunctions followed by the calculation in semi direct fashion of one or more Cl eigenfunctions of the secular problem In contrast to the conventional module just two secular problems are solved as part of the extrapolation process one at the lowest threshold Tmin and one at the threshold Tmin Tinc e NATORB generates the spin free natural orbitals for one or more of the calculated Cl eigenvectors Note that this module is now executed in default AS with the original code the remaining modules are optional and may be used to further analyse one or more of the Cl eigenvectors 16 DIRECTIVES CONTROLLING SEMI DIRECT TABLE CI CALCULATIONS 53 e PROP to compute various 1 electron properties of the Cl wavefunctions Note that the natural orbitals generated above may be routed to the Dumpfile and examined by the other analysis modules of GAMESS UK in a subsequent job e TM to compute the transition moments between nominated Cl eigenvectors In addition to the Mainfile Dumpfile Scratchfile amd Transformed Integral File ED6 the following data sets will be used by the program e The Tablefile A dataset normally assigned using the local file name table ci will be used as a source of pattern symbolic m
6. ORBITAL 28 TABLE CI CALCULATIONS USING MCSCF ORBITALS 109 COR1 COR1 COR1 DOC1 DOC3 DOC1 DOC2 DOC3 VOC2 VOC1 UDC3 UDCA END PRINT ORBITALS VIRTUALS NATORB CANONICAL 10 FOCK DENSITY FOCK MRDCI DIRECT TABLE SELECT SYMMETRY 1 SPIN 1 CNTRL 16 SINGLES 1 CONF 012345 29 42 43 012345 30 42 43 012345 29 42 44 END ROOTS 1 THRESH 2 2 CI NATORB ENTER Finally we consider the data for performing exactly the same calculation as above but now freezing the oxygen and carbon 1s core orbitals in the Table Cl calculation The following points should be noted e Unlike the conventional Table Cl data specification of the frozen orbitals now requires use of the ACTIVE and CORE directives the CORE directive below specifying MOs 1 and 2 the ACTIVE directive specifying orbitals 3 to 62 e The orbital indices specified on the CONF data lines reflect the removal of these two orbitals with the CNTRL directive now pointing to a 12 electron Cl calculation as distinct from the 16 electron calculation above RESTART TITLE H2C0 DIRECT MRDCI FROM MCSCF NOS SEC 10 FREEZE 1S SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 29 ITERATIVE MRDCI CALCULATIONS 110 END RUNTYPE CI ACTIVE 3 TO 62 END CORE 1 TO 2 END SCFTYPE MCSCF MCSCF ORBITAL COR1 COR1 COR1 DOC1 DOC3 DOC1 DOC2 DOC3 VOC2 VOC1 UDC3
7. to 2e doubly excited configurations 50284201274 18 and 5064301274 19 would require the following CONF data 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 63 CONF 0123 15 013 4 15 0123 16 END The complete data file for performing the SCF and subsequent semi direct Cl would then be as follows TITLE PH3 6 31G VALENCE CI 3M 1R SUPER OFF NOSYM ZMAT P H 1 RPH H 1 RPH 2 THETA H 1 RPH 2 THETA 3 THETA 1 VARIABLES RPH 2 685 THETA 93 83 END BASIS 6 31G RUNTYPE CI CORE 1 TO 5 END MRDCI DIRECT TABLE SELECT CNTRL 8 SPIN 1 SYMMETRY 1 SINGLES ALL CONF 0123 15 013 4 15 0123 16 END CI NATORB ENTER Example 2 In this example we wish to perform a valence Cl calculation on the CuCl molecule using a 3 21G basis While the molecular symmetry is Cooy the symmetry adaptation and subsequent Cl will be conducted in the Ca point group The resolution of the Coy into the Ca orbital species is given in Table 2 An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 2 9 3 9 4 2 1 1 326 84723972 2 0000000 2 1 104 02836336 2 0000000 3 1 40 71695637 2 0000000 4 1 35 46377378 2 0000000 5 3 35 45608069 2 0000000 6 2 35 45608068 2 0000000 7 1 10 42193940 2 0000000 8 1 7 88512031 2 0000000 9 2 7 88222844 2 0000000 10 3 7 88222844 2 0000000 11 1 5 07729175 2 0000000
8. 01257 9 NATORB IPRIN ENTER Example 5 In this example we wish to perform a valence Cl calculation on the CaHa molecule using a 3 21G basis While the molecular symmetry is Don the symmetry adaptation and subsequent Cl will be conducted in the D point group An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 148 37173884 2 0000000 2 1 16 76521275 2 0000000 3 3 13 55586861 2 0000000 4 2 13 55586861 2 0000000 5 5 13 55460610 2 0000000 6 1 2 26357685 2 0000000 7 3 1 36160958 2 0000000 8 2 1 36160958 2 0000000 9 5 1 35089927 2 0000000 10 1 0 34923025 2 0000000 11 5 0 31649941 2 0000000 12 2 0 02334207 0 0000000 13 3 0 02334207 0 0000000 14 1 0 04980631 0 0000000 15 5 0 09478404 0 0000000 16 1 0 12395484 0 0000000 17 3 0 13549605 0 0000000 18 2 0 13549605 0 0000000 19 5 0 28345574 0 0000000 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 23 20 1 1 32404002 0 0000000 21 5 1 45900204 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the nine Ca inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos og 1 7 3 4 1 4 Tue 2 4 2 2 5 6 Tuy 3 4 2 2 7 8 Tu 5 6 2 4 9 12 To perform an 4 electron valence Cl calculation based on the SCF configuration would require the following CONF data CONF 019 The complete
9. 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 TrFrFENODONDRFPWNWNHHWAUAUHAKFAPNDTKEANWBNHNWBREFEBNDKPWNTPNDOTOKFPHAWNHAKrFPKP ITN DE W pb RO0XX x3OoOo00000101010101ds KAA KR RAR EA BRAK PS WWW WWWWUNNNNRR RRA RR RR oo0000o 51942385 53416038 71767715 71767715 74113094 04543157 04543157 15104310 37114036 37114036 53065438 53065438 58044223 89387038 93835376 93835376 02899399 38397418 58126398 58126398 03473543 21032541 21032541 84680067 84680067 93355912 93355912 03730269 03730269 48707133 48707133 52955492 56413770 76466065 76466065 78192405 78192405 91469244 91469244 11932527 42107737 42107737 42155085 42155085 53491607 53491607 70188467 24354532 24354532 31686796 07828789 07828789 42407537 86035408 17202050 o00D0DO0O0O0O0O00000000000000000000000000000000000000000000000000o 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 000
10. Conventional and new Semi direct modules are available and are described in detail below To maintain compatibility with previous documentation we first describe below the data input and file requirements of the older version This is then followed by a description of the new code 1 1 Sub Module Structure of Conventional Table CI An outline of the sub module structure and philosophy behind Table Cl have already been given in Part 2 material that should be taken in conjunction with the present chapter As pointed out previously the module comprises a set of 9 sub modules which must be user driven either implicitly or explicitly see below through data input These sub modules are as follows e ADAPT generation of a symmetry adapted list of integrals derived by a pseudo transformation from the list of raw integrals e TRAN integral transformation using the list of adapted integrals generated above to gether with a molecular orbital coefficient array nominated by the user Note that in contrast to the Direct Cl module transformation is an integral part of the Table Cl mod ule e TABLE generation of the data base of pattern matrix elements required by both the SELECT and Cl sub modules see below this data base will have usually been made available on a given machine but may be generated by the user using this sub module e SELECT configuration generation and subsequent selection based on a user specified set o
11. The latter is 4 MWords in default a value which will not be sufficient to accommodate the above settings Thus an examination of the semi direct Table Cl output will typically reveal the following A A A E A A Insufficient memory for default allocations reduce parameters as follows PARA HEHEHHE HH H H H A A E A A A nteint from 3500001 to 1750001 mdi from 1000001 to 500001 iotm from 2000000 to 1000000 nedim from 2000000 to 1000000 FEET EFEFE EPTFE EFTEPEFT EPTFE EPTET EPTFE EEE EP EET FEET ELE FALE FEFF TELE TEP EEF A A A A Insufficient memory for default allocations reduce parameters as follows FEAF EEEAFEEEFFFEEEFEEEEEEEEEFEFEEEEFEFEEEF EEE EEE nteint from 3500001 to 875001 mdi from 1000001 to 250001 iotm from 2000000 to 500000 nedim from 2000000 to 500000 FEET EPTFE EFTFET EFT EPEFT EPTFE EPTFE EPEEE EEE EP EET This is usually not a problem as the default values are such to accommodate quite demanding applications Should the decreased settings prove inadequate the code should inform the user accordingly then the first action is to increase the memory for the job using the MEMORY pre directive Typically a pre directive setting of 14 MWords will enable the default settings to apply If these defaults prove inadequate the user must resort to specifying the parameter to be increased through data input while increasing the MEMORY setting to compensate for this increase This da
12. gt lt bil Y 2 b gt lt Vil Y y gt lt wil Y ly gt 27 ITERATIVE NATURAL ORBITAL CALCULATIONS 100 The f r and f 7 values are printed out in x y z components and the expectation values for lt Y Y Ti Yi zil gt are also printed 27 Iterative Natural Orbital Calculations We now work through an example of using the natural orbitals generated by the module in a subsequent Cl calculation We consider a DZ calculation on the ground state of C2Ha4 with the computation split into four separate jobs in which we 1 perform the initial SCF 2 carry out an initial Cl where the reference set employed comprises just the SCF configu ration using the SCF MOs of interest 3 based on the output from the initial Cl we augment the reference set to include the leading secondary configuration generating the resulting natural orbitals 4 carry out the 2 reference Cl based on the natural orbitals generated in the previous step We now consider various aspects of each job in turn Job 1 The SCF TITLE ETHYLENE DZ GROUND STATE SCF SUPER OFF NOSYM ZMATRIX ANGSTROM C C 11 4 H 11 1 2 120 0 H 11 1 2 120 0 3 180 0 H 2 1 1 1 120 0 3 0 0 H 2 1 1 1 120 0 3 180 0 END BASIS DZ ENTER The only points to note here is the use of the SUPER directive in suppressing skeletonisation Job 2 The 2M 1R CI An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED
13. working of the module we consider below a number of example calculations 24 1 Calculations on the Formaldehyde Ground State A Semi direct Table Cl calculation is to performed on the formaldehyde molecule in a TZVP basis Given the following data sequence TITLE H2CO TZVP X1A1 DEFAULT TABLE CI OPTIONS ZMAT ANGSTROM C 0 1 1 203 H 1 1 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS TZVP RUNTYPE CI MRDCI DIRECT ENTER then the calculation undertaken will be based on the following 1 Integral transformation will use the set of orbitals from section 1 the integer specified on the ENTER directive i e the closed shell SCF orbitals 2 The table ci data base will be generated rather than restored from a pre existing table ci data set 3 The symmetry spin and number of active electrons will be taken from the corresponding SCF wavefunction In the present case this involves e A Cl wavefunction of Ay symmetry i e SYMMETRY 1 e A singlet Cl wavefunction i e SPIN 1 e The number of active electrons in the Cl will be set to be those involved in the SCF calculation i e CNTRL 16 4 Singly excited configurations with respect to each of the default reference configurations SINGLES ALL will be included regardless of their computed energy lowerings 5 The set of reference configurations to be employed will include the SCF configuration plus those generated from this configuration by including i for each symmetry
14. 1 The SCF TITLE x H20 TZVP DIFFUSE S P MRDCI SUPER OFF NOSYM ZMAT ANGSTROM 0 H 1 0 951 H 1 0 951 2 104 5 END BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END ENTER The only points to note here are i the use of the SUPER directive in suppressing skeletonisa tion and ii use of the default section for eigenvector output section 1 for the closed shell SCF Job 2 The Initial 3M 3R CI An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS and the following orbital assignments characterising the closed shell SCF configuration 1a22a21023a 10 13 12 CALCULATING THE A STATES OF H20 35 1 1 20 56084959 2 0000000 2 1 1 35696939 2 0000000 3 3 0 72200122 2 0000000 4 1 0 58247942 2 0000000 5 2 0 50858566 2 0000000 6 1 0 02724259 0 0000000 7 3 0 04894440 0 0000000 8 2 0 05589681 0 0000000 9 1 0 06133571 0 0000000 10 al 0 20403420 0 0000000 11 3 0 22824210 0 0000000 12 3 0 53700802 0 0000000 13 1 0 56235022 0 0000000 14 2 0 58645643 0 0000000 15 1 0 66887228 0 0000000 16 3 0 74805617 0 0000000 17 1 1 07690608 0 0000000 18 1 1 88545053 0 0000000 19 4 1 92243836 0 0000000 20 2 2 12944874 0 0000000 21 3 2 20541910 0 0000000 22 1 2 34202871 0 0000000 23 3 2 39946430 0 0000000 24 3 2 69788310 0 0000000 25 1 2 72651832 0 0000000 26 2 2 73832720 0 0000000 27 al 3 07664215 0 0000000 28 3 3 26840142 0 0000000 29 2 3 5461657
15. 12 1 3 38247056 2 0000000 13 3 3 35978308 2 0000000 14 2 3 35978307 2 0000000 15 1 01099628 2 0000000 16 3 0 53702948 2 0000000 17 2 0 53702947 2 0000000 18 4 0 49640067 2 0000000 19 al 0 49640067 2 0000000 20 1 0 44715317 2 0000000 21 3 0 39988537 2 0000000 22 2 0 39988537 2 0000000 23 1 0 35127248 2 0000000 24 1 0 00023285 0 0000000 25 3 0 06300102 0 0000000 26 2 0 06300102 0 0000000 27 1 0 12855448 0 0000000 28 1 0 19287013 0 0000000 29 3 0 25729975 0 0000000 30 2 0 25729975 0 0000000 31 1 0 39720201 0 0000000 32 1 0 86197727 0 0000000 33 2 0 88942618 0 0000000 34 3 0 88942618 0 0000000 35 1 1 01877167 0 0000000 36 al 2 16694989 0 0000000 37 3 3 96181512 0 0000000 38 2 3 96181512 0 0000000 39 4 3 98212497 0 0000000 40 1 3 98212497 0 0000000 41 1 4 08851360 0 0000000 42 1 24 51368240 0 0000000 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 65 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the first 14 inner shell orbitals 10220 30 40 11 50 60 21 70 80 3n 20 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos ay 1 22 8 14 1 14 b 2 9 3 6 15 20 ba 3 9 3 6 21 26 as 4 2 0 2 27 28 To perform an 18 electron valence Cl calculation based on the SCF configuration 90740416410075 741107 21 and the doubly excited configuration 90 41 15 100 51 120 22 would require the following CONF data
16. 2 0 86232544 23 1 1 23833149 24 4 1 44033091 25 3 1 44033091 26 2 1 44033091 27 1 3 13181655 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the first 5 silicon inner shell orbitals 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 67 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos a1 1 9 2 7 1 7 b 2 6 1 5 8 12 ba 3 6 1 5 13 17 a2 4 6 1 5 18 22 To perform a 4 reference 8 electron valence Cl calculation based on the SCF configuration and configurations arising from the 2t to 3t3 would require the following CONF data CONF 018 13 18 018 13 19 018 14 18 019 13 18 END The complete data file for performing the SCF and subsequent Cl would then be as follows note that we are retaining all single excitations with respect to each reference function in the final Cl TITLE SIH4 6 31G DIRECT TABLE CI VALENCE CI 4M 1R ZMAT SI H 1 SIH H 1 SIH 2 109 471 H 1 SIH 2 109 471 3 120 0 H 1 SIH 2 109 471 4 120 0 VARIABLES SIH 2 80 END BASIS 6 31G RUNTYPE CI CORE 1 TO 5 END MRDCI DIRECT TABLE SELECT CNTRL 8 SYMM 1 SPIN 1 CONF 018 13 18 018 13 19 018 14 18 019 13 18 END SINGLES ALL CI NATORB ENTER 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 68 Example 4 In this example we wish to perform a valence Cl calculation on the No molecule using a 4 31G basis While the molecular symme
17. 2 0 91131383 20 1 0 91131383 21 1 0 93118300 22 1 1 17900613 23 2 1 45058658 24 1 1 45058658 25 1 3 78674557 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 o00D0DO0O0O0O00O0O00000000NNNNDoNy Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the five inner shell orbitals 1a22a1a 2304a 3 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos a 1 4 14 1 14 a 2 1 6 15 20 To perform an 8 electron valence Cl calculation involving the SCF configuration and two de generate 1e to 2e doubly excited configurations 50280224 Ta and 5076030 Ta would require the following CONF data CONF 0123 15 013 4 15 0123 16 4 The complete data file for performing the SCF and subsequent Cl would then be as follows 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 14 TITLE PH3 6 31G VALENCE CI 3M 1R SUPER OFF NOSYM ZMAT P H 1 RPH H 1 RPH 2 THETA H 1 RPH 2 THETA 3 THETA 1 VARIABLES RPH 2 685 THETA 93 83 END BASIS 6 31G RUNTYPE CI MRDCI TRAN CORE 41 1 TO 4 1 SELECT SINGLES 1 CONF 0123 15 013 4 15 0123 16 NATORB ENTER Example 2 In this example we wish to perform a valence Cl calculation on the CuCl molecul
18. 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 END ENTER The only points to note here again is the use of the SUPER directive in suppressing skeleton isation and use of the default eigenvector section section 1 for storage of the closed shell eigenvectors An examination of the SCF output reveals the following orbital analysis 28 TABLE CI CALCULATIONS USING MCSCF ORBITALS IRREP NO OF SYMMETRY ADAPTED 57768533 34457777 40746540 87003449 69591811 65109519 53687971 44174805 11697212 26220763 27217357 38931080 41757152 46526352 60968525 75001014 86980119 89167074 93051881 07098621 18616042 35370640 52221224 68895841 88480823 97392259 13452238 13982211 19187925 34994146 36355601 67698155 82812279 84696649 97321688 14466153 33275225 51306001 51697350 52974692 oono AUNE FPWWWWWWWWWWNHNNNNNNNNNKP KBR BP RB RP ppp SCOANADAONBRPWNHROWHMOANDAABWNHROWDANDAAKFWNHEHRO BASIS FUNCTIONS NPWRrRRWNHRABNFPWHRERPBRWHENYPBPRPWRPWENFPKPNWBWHRENWNHRPWRRPRP BB WUYW0Y0WYwWUNNNNNNNNNRRRR 24 4 400000000000 o0o00O0O0O0O0O000O00000000000000000000000O0ONNNNNNNN 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000
19. 83958499 44 4 3 85444369 45 2 3 87171104 46 2 4 14123645 47 3 4 16304613 48 1 4 16535425 49 2 4 16620625 50 2 4 32796704 51 3 4 47390388 52 1 4 49325977 53 1 4 68584478 54 4 4 89513471 55 3 5 13208791 56 1 5 14756255 57 2 5 85901984 58 1 5 92954017 59 3 6 06228738 60 1 8 35480844 61 1 27 71561806 62 1 45 72664766 Job 2 The MCSCF o0o00D0DO0O0O0O0O0O000O0000000000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 48 The following data performs a 10 electron in 9 orbital CASSCF calculation using the MCSCF module with the natural orbitals routed to section 10 of the Dumpfile under control of the CANONICAL directive In the absence of the VECTORS directive the SCF MOs will be used as the starting orbitals RESTART TITLE H2C0 MCSCF 10E IN 9 M 0 SUPER OFF NOSYM NOPRINT BYPASS ZMATRIX ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 END SCFTYPE MCSCF 14 TABLE CI CALCULATIONS USING MCSCF ORBITALS 49 MCSCF ORBITAL COR1 COR1 COR1 DOC1 DOC3 DOC1 DOC2 DOC3 VOC2 UOC1 UDC3 UDCA END PRINT ORBITALS VIRTUALS NATORB CANONICAL 10 FOCK DENSITY FOCK ENTER Job 3 The Table CI Job Performing a Table Cl calculation using the natural orbita
20. BASIS FUNCTIONS 27 ITERATIVE NATURAL ORBITAL CALCULATIONS 101 and the following orbital assignments characterising the closed shell SCF configuration 1a71b7 2072b 105 307 103 1b3 35 1 1 11 25533463 2 0000000 2 5 11 25413119 2 0000000 3 1 1 02052567 2 0000000 4 5 0 79195744 2 0000000 5 3 0 64152856 2 0000000 6 1 0 57038928 2 0000000 7 7 0 51438205 2 0000000 8 2 0 36388693 2 0000000 9 6 0 13190687 0 0000000 10 5 0 25441028 0 0000000 11 1 0 25558269 0 0000000 12 3 0 33918097 0 0000000 13 5 0 36641681 0 0000000 14 1 0 41844853 0 0000000 15 3 0 45487336 0 0000000 16 7 0 49648833 0 0000000 17 2 0 49708917 0 0000000 18 6 0 60222488 0 0000000 19 7 0 64242537 0 0000000 20 1 0 76465797 0 0000000 21 5 0 82560838 0 0000000 22 5 1 10140194 0 0000000 23 1 1 20630804 0 0000000 24 3 1 30189632 0 0000000 25 5 1 35219192 0 0000000 26 7 1 50761510 0 0000000 27 1 23 76609415 0 0000000 28 5 24 01493168 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the Cis inner shell orbitals 27 ITERATIVE NATURAL ORBITAL CALCULATIONS 102 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos ag 1 8 1 7 1 7 bgu 2 2 0 2 8 9 boy 3 4 0 4 10 13 biu 5 8 1 7 14 20 bay 6 2 0 2 21 22 b3y 7 4 0 4 23 26 Note that within the DZ basis employed there are no basis functions of bj IRrep 4 or a IRrep 8 symmetry The symmetry r
21. BASIS FUNCTIONS and the following orbital assignments from the converged closed shell SCF 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 21 1 1 15 59983859 2 0000000 2 5 15 59796932 2 0000000 3 1 1 54485796 2 0000000 4 5 0 74550130 2 0000000 5 2 0 63373069 2 0000000 6 3 0 63373069 2 0000000 7 1 0 62012170 2 0000000 8 6 0 20546760 0 0000000 9 7 0 20546760 0 0000000 10 5 0 79186362 0 0000000 11 1 1 16445455 0 0000000 12 2 1 26826720 0 0000000 13 3 1 26826720 0 0000000 14 7 1 43237859 0 0000000 15 6 1 43237859 0 0000000 16 5 1 55279124 0 0000000 17 1 1 83635478 0 0000000 18 5 2 63677794 0 0000000 Note that there are now no MOs of IRREP 4 or 8 Based on the above output the CONF data lines may be deduced from the following table where we again assume that we wish to freeze the two Nis inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos og 1 5 1 4 1 4 Tie 2 2 0 2 5 6 Tia 3 2 0 2 7 8 Tu 5 5 1 4 9 12 Toa 6 2 0 2 13 14 Teg Y 2 0 2 15 16 To perform an 10 electron valence Cl calculation based on the SCF configuration would require the following CONF data CONF 01257 9 The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE N2 3 21G SUPER OFF NOSYM ZMAT ANGS N N 1 NN VARIABLES NN 1 05 END BASIS 3 21G RUNTYPE CI 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 22 MRDCI TRAN CORE 100100 1 1 SELECT SINGLES 1 CONF
22. PR RRE END BASIS DZ CORE 1 2 END RUNTYPE CI MRDCI DIRECT SELECT SYMMETRY 1 CNTRL 12 SPIN 1 CONF 0128 10 14 23 012 10 14 21 23 END THRESH 3 3 ITERATE MAXI 20 C x 2 0 95 WEIGHT 0 002 ENTER The initial Cl calculation concludes with the following analysis Current Energies from MRDCI Iterations State c 2 Energy Energy Davidson T 3 a u T 0 a u a u 1 0 922 78 199076 78 199862 78 213624 Five iterations of the MRDCI module are subsequently required to increase the initial value of c 2 from 0 922 to the requested level of 0 950 involving the addition of 35 secondary configurations as reference functions and a final c 2 value of 0 957 The following analysis appears on completion of these iterations State c 2 Energy Energy Davidson T 3 a u T 0 a u a u gt 1 0 957 78 200915 78 207284 78 211994 Note that merely presenting the data line ITERATE 29 ITERATIVE MRDCI CALCULATIONS 114 would lead to refinement of the ground state wavefunction to the point where all secondary coefficients with c 2 greater than 0 005 are included in the reference state at which point iterations would cease 29 2 The Algorithm for Controlling Multi root Calculations The ITERATE sub directives described above provide for control over the iterative treatment of single Cl wavefunctions This however is not the main purpose of ITERATE controlling the treatment
23. SABF 200 ENTER Job 3 The Natural Orbital CI We show below the data for using the NOs from the 2 reference Cl where the orbitals routed to section 200 are now restored by specification on the ENTER directive The following points should be noted e We assume that the Table Cl data base table ci been saved from the initial Cl job allowing the BYPASS specification on the TABLE directive e In contrast to the conventional Table Cl module restoring the NOs must now be controlled through VECTORS and ENTER specification in the conventional module this was input through TRAN specification within the Table Cl transformation module e The resulting NOs from the natural orbital Cl are now routed to section 210 and could be used in a subsequent Cl in obvious fashion RESTART NEW TITLE ETHYLENE CI GROUND STATE 2M NOS BYPASS SCF ZMATRIX ANGSTROM C C 11 4 27 ITERATIVE NATURAL ORBITAL CALCULATIONS 104 H 1 1 1 2 120 0 H 1 1 1 2 120 0 3 180 0 H 2 1 1 1 120 0 3 0 0 H 2 1 1 1 120 0 3 180 0 END BASIS DZ 3 TO 28 END CORE 1 2 END RUNTYPE CI MRDCI DIRECT TABLE BYPASS SELECT CNTRL 12 SYMMETRY 1 SPIN 1 SINGLES 1 CONF 0128 10 14 23 012 10 14 21 23 END THRESH 30 10 NATORB IPRINT PUTQ SABF 210 VECTORS 200 ENTER 200 We show below the final Cl vector from the natural orbital Cl Description of the XA state EXTRAPOLATED ENERGY 78 20206451 FEO E k k k kkk k k CONFIGURATION WEIGHTS EEEE E E k kk kkk
24. SELECT CNTRL 10 SYMM 1 SPIN 1 SINGLES 1 CONF 01216 24 35 01 2 24 35 50 O 1 2 16 35 58 4 16 24 50 58 1 2 35 END CI NATORB IPRIN ENTER Example 5 73 In this example we wish to perform a valence Cl calculation on the CaHa molecule using a 6 31G basis While the molecular symmetry is Doon the symmetry adaptation and subsequent Cl will be conducted in the D point group An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTE BASIS FUNCTIONS D 1 1 149 34279776 2 0000000 2 1 16 81835576 2 0000000 3 2 13 62701326 2 0000000 4 3 13 62701326 2 0000000 5 5 13 62532936 2 0000000 6 1 2 24495780 2 0000000 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 1 34937153 1 34937153 1 33498314 0 35141213 0 31217469 0 02646555 0 02646555 0 04279408 0 11107022 0 11631297 0 20535510 0 20535510 0 30735852 0 30735852 0 32858094 0 32858094 0 36615374 0 47836712 0 49656774 1 18797119 1 30595420 2 42553536 2 42553536 2 45164750 2 45164750 2 50399980 2 74252653 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 o00D0DO0O0000000000000000000NNNNDN 74 Based on the above output the CONF data lines may be deduced from the fo
25. TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass SELECT processing Such usage is typically associated with restarting Table Cl calculations 6 2 CNTRL This directive consists of one line read to variables TEXT NELEC using format A e TEXT should be set to the character string CNTRL e NELEC is used to specify the total number of active electrons in the Cl calculation Note that any inner shell electrons frozen out under control of the TRAN directive should not be included The CNTRL directive may be omitted when the program will set NELEC to the value char acterising the SCF process implicit within the RUNTYPE Cl processing subtracting out those electrons nominated through the CORE parameter of the TRAN data 6 3 SPIN This directive consists of one line read to variables TEXT NSPIN using format A e TEXT should be set to the character string SPIN e NSPIN is used to specify the spin degeneracy of the Cl wavefunction of the electronic eigenstate s of interest using the values 1 2 3 etc for singlet doublet triplet states etc respectively It is also possible to use one of the character strings SINGLET DOUBLET TRIPLET QUARTET and QUINTET to specify NSPIN 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 10 The SPIN directive may be omitted when the program will set NSPIN to the value specified on the MULTIPLICITY directive see section 4 6 2 Example SPIN 3 S
26. TMIN should be set to the minimum threshold factor in units of micro hartree uH to be used in selection Any CSF with a computed energy lowering greater than TMIN will be retained in the final list of selected configurations e TINC should be set to the threshold increment to be used in the process of extrapolation This process involves solution of the final secular problem at a range of increasing thresh olds defined by TMIN TMIN TINC TMIN 2 x TINC TMIN NEXTRP 1 x TINC TMIN NEXTRP x TINC where NEXTRP is the number of extrapolation passes requested under control of the EXTRAP directive see the DIAG directives 7 DATA FOR CONVENTIONAL TABLE CI H MATRIX CONSTRUCTION 25 The THRESH directive may be omitted when TMIN will be set to 30 0 and TINC to 10 0 With the default EXTRAP setting this would lead to the solution of the T 50 40 and 30 wH secular problem Example THRESH 5 0 5 0 THRESH 5 5 are equivalent causing Tmin and Tine to be set to 5 microhartree 7 Data for Conventional Table CT H Matrix Construction 7 1 CI The Cl directive comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string Cl e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module IPRI
27. TRAN directive are triggered by the presence of the CORE and or DISCARD keywords on the first line above 1 If the CORE keyword has been presented two additional data lines are now required to define the number Line 1 and the sequence numbers Line 2 of the orbitals to be frozen Line 1 is read in l format to the variables NOCORE I 1 NIRREP where NOCORE I specifies the number of orbitals of irreducible representation IRrep that are to be frozen NIRREP is the number of irreducible representations characterising the associated Abelian point group in use containing more than zero orbitals Note again that each IRrep has an associated sequence number see Table 1 and that the input orbital set will be reordered such that e Rreps having zero orbitals are discarded and e orbitals of common Rrep are grouped together these groups being arranged in order of increasing IRrep number and e orbitals of common Rrep are ordered according to their relative disposition in the input orbital set e g by eigenvalue ordering if SCF MOs Line 2 line is also read in l format and specifies the sequence numbers of the frozen orbitals in the spirit of the re ordered sequence above Thus the first NOCORE 1 integers specify the frozen orbitals of IRrep the next NOCORE 2 integers the frozen orbitals of Rrepa and so on until all Rreps have been specified Note that the integer specification within each Rrep refers to the relative ordering w
28. UDCA END PRINT ORBITALS VIRTUALS NATORB CANONICAL 10 FOCK DENSITY FOCK MRDCI DIRECT TABLE SELECT SYMMETRY 1 SPIN 1 CNTRL 12 SINGLES 1 CONF 0123 27 40 41 0123 28 40 41 0123 27 40 42 END ROOTS 1 THRESH 2 2 CI NATORB ENTER EOF 29 Iterative MRDCI Calculations All of the MRDCI examples presented to date involve a single run of the module in which a number of excited states of specified symmetry are typically generated based on a user specified list of main configurations Generating the entire sequence of states required for example in simulating the vertical electronic spectra of a given species is often a labour intensive exercise requiring the repeated refinement of reference configurations The ITERATE directive is de signed to shorten this process by allowing for an iterative sequence of MRDCI calculations in which the initial reference set and associated eigenstates are iteratively refined with the minimum level of user intervention This modus operandi is designed such that the user 1 need present no explicit configuration CONF data 2 may iteratively improve the quality of the Cl wavefunction of a single eigen state by specifying the desired value of c 2 coefficients of the main reference configurations 3 may generate up to 30 eigen states of a specified spin and spatial symmetry in a single 29 ITERATIVE MRDCI CALCULATIONS 111 run of the module with no prior knowledge of the composition of these
29. algorithm While the default memory allocations will prove adequate for small medium cases the user should use the MEMORY pre directive to request at least 8 MWords in calculations with say more than 20 active electrons The overall filespace requirements are significantly reduced compared to the Conventional module note that the FORTRAN unit numbers for some the key data sets have been modified compared to the original settings e An automatic procedure for obtaining a number of excited states within a single given run of the program is now possible under control of the ITERATE directive This is designed to remove much of the labour involved in generating for example vertical excitation spectra 15 1 Sub Module Structure of Semi direct Table CI An outline of the sub module structure and philosophy behind the semi direct Table Cl module has already been given in Part 2 material that should be taken in conjunction with the present chapter As pointed out previously this module comprises a reduced set of 6 sub modules which must be user driven either implicitly or explicitly see below through data input These sub modules are as follows e TABLE generates an input a data base of pattern symbolic matrix elements for use in both the selection process and in solving the secular problem This data base is written to a file with LFN table ci note the name change compared to the LFN TABLE employed in the Conventional module
30. data field Lines between the initiator and terminator define the reference configuration set each line defines a reference CSF by specifying the sequence numbers of the component active orbitals in l format A given reference CSF is defined by 1 the number of open shell orbitals NOPEN NOPEN includes any unpaired orbitals to gether with those non identical spin coupled pairs open to substitution 2 NOPEN integers specifying the sequence numbers of these orbitals 3 the NELEC NOPEN 2 sequence numbers of the doubly occupied orbitals i e the iden tically spin coupled orbitals where the sequence numbers refers to the symmetry ordered orbitals performed at the outset of processing Within the set of open and doubly occupied orbitals the MOs are presented in groups of common Rrep with the groups presented in order of increasing IRrep sequence number A few examples below will help clarify this order of presentation Note again that e all reference function nominated by CONF must be of the same symmetry as that nomi nated on the SYMMETRY directive e at least two reference functions must be specified it is not possible to conduct simple CISD calculations with the semi direct module e note again the requirement for a directive terminator Example 1 Consider performing a valence Cl calculation on the PH3 molecule using a 6 31G basis While the molecular symmetry is Cz the symmetry adaptation and subsequent Cl will be c
31. data file for performing the SCF and subsequent Cl would then be as follows TITLE CAH2 3 21G SUPER OFF NOSYM ZMAT ANGS CA X 1 1 0 H 1 CAH 2 90 0 H 1 CAH 2 90 0 3 THETA VARIABLES CAH 2 148 THETA 180 0 END BASIS 3 21G RUNTYPE CI MRDCI TRAN CORE 322 2 123 12 12 12 SELECT SINGLES 1 CONF 019 NATORB IPRIN ENTER 6 7 ROOTS The ROOTS directive is used to specify those eigenvectors of the root secular problem to be used in the process of selection with the energy contributions of the configurations computed with respect to the nominated vectors The directive consists of a single data line with the character string ROOTS in the first data field Subsequent data comprises integer variables used to specify the number of root eigenstates NROOT and the sequence numbers of these vectors within the matrix of zero order eigenvectors IROOT I I 1 NROOT Two formats may be used in this specification 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 24 1 If the lowest NROOT vectors are to be used then the data line is read to the variables TEXT NROOT using format A e TEXT is set to the character string ROOTS e NROOT is an integer specifying the number of roots to be used where the sequence numbers of the roots will be 1 NROOT 2 If the NROOT vectors to be used are not the lowest in the root eigenvector matrix then the sequence numbers within this matrix must be specified The data line is then read to the variabl
32. k k aoe ROOT 1 RR x C C CONFIGURATION M 0 89794 1 2 8 10 14 23 0 00125 1 8 10 14 15 23 M 0 03203 1 2 10 14 21 23 0 00100 1 2 8 11 14 23 0 00102 1 2 8 10 11 14 0 00279 1 8 15 21 2 10 14 23 0 00106 1 3 14 16 2 8 10 23 0 00499 2 8 15 21 1 10 14 23 0 00143 2 8 16 21 1 10 14 23 0 00186 2 3 14 16 1 8 10 23 0 00119 2 11 16 23 1 8 10 14 0 00101 8 9 10 11 1 2 14 23 0 00144 8 10 21 24 1 2 14 23 28 TABLE CI CALCULATIONS USING MCSCF ORBITALS 105 0 00260 3 8 14 21 1 2 10 23 0 00110 8 9 23 24 1 2 10 14 0 00104 8 12 21 23 1 2 10 14 0 00117 3 10 16 23 1 2 8 14 0 00297 10 11 23 24 1 2 8 14 0 00125 3 11 14 23 1 2 8 10 SUM OF MAIN REFERENCE C C 0 929973257309359 28 Table CI Calculations Using MCSCF Orbitals To conclude our discussion of the semi direct Table Cl module we work through an example of using the natural orbitals generated from the MCSCF module in a subsequent Cl calculation We consider a calculation on the ground state of H2CO with the computation split into three separate jobs in which we 1 perform an initial SCF 2 carry out the MCSCF calculation 3 perform the MRDCI calculation using the MCSCF natural orbitals We now consider various aspects of each job in turn and note that several changes will be required to the corresponding data sets shown above for the Conventional Table Cl study Job 1 The SCF TITLE H2CO DZP F SUPER OFF NOSYM ZMATRIX ANGSTROM N Cc 0 1 1 203 H 11 099 2 121 8 H 1 1 099
33. of course to the first orbital in the appropriate IRrep The following sequence would be used to simply freeze the orbitals while retaining the complete virtual manifold TRAN CORE 10001000 11 5 DATA FOR CONVENTIONAL TABLE CI DATA BASE GENERATION 8 Turning to the inner shell complement orbitals we again find the corresponding IRreps 1 and 5 with the orbitals the highest lying member in each case with relative sequence number of 11 Thus the following TRAN data would act to both freeze and discard the inner shells and their complement MOs TRAN CORE DISCARD 10001000 11 10001000 11 11 Further examples of TRAN data will be discussed below in the section discussing Reference Function specification within the SELECT data 5 Data for Conventional Table CI Data base Generation The data base of pattern symbolic matrix elements required by both the Selection and Cl modules may be generated by the user in the course of any Table Cl calculation It is not envisaged that this step will be necessary since the data base will in general be already installed on those machines on which GAMESS UK is available the data set being allocated to the program with LFN TABLE 5 1 TABLE The TABLE directive is used to request and control the data base generator and comprises a single data line read to the variables TEXT TEXTF using format 2A e TEXT should be set to the character string TABLE e TEXTF is an optional parameter that may be
34. of multi root calculations with a view to the automatic handling of e g excitation spectrum requires a far greater level of control that is provided by a number of sub directives MAXROOT SROOT DROOT and RETAIN that are described below Before considering these directives in detail we first outline the algorithm that is used in control ling what is a relatively complex procedure This has involved a number of prototyping exercises that have culminated in a final design criteria dominated by the desire to make the usage and specification as simple as possible while providing the necessary level of robustness to deliver the required solution in a reasonable number of iterations Users who have driven the MRDCI module in the search for excited states will be aware of the typical sequence of calculations that are performed manually This involves starting with a number of reference configurations and obtaining one or more Cl wavefunctions under control of the ROOTS directive whereby selection is carried out with respect to NROOT roots of the root secular problem to be used in process of configuration selection Typically the same number of roots are then generated in the final secular problem The user will then augment the reference set and increase the value of NROOT based on an examination of earlier calculations until the desired number of eigenstates are obtained The current implementation is an attempt to automate the above process Ini
35. on the property with the nominated CSF being typically the corresponding SCF configuration Thus the final data for the properties module comprises a sequence of NVEC data lines each line a sequence of integers defining the single configuration for the Cl vector under consideration The format of these lines is identical to that of the CONF data used in nominating the reference functions and in most instances will be a repeat of that data Example Consider the valence Cl calculation on PH3 described in example 1 of the CONF directive Considering just the Cl data MRDCI TRAN 1 CORE 4 1 1 TO 4 1 SELECT SINGLES 1 CONF 0123 15 013 4 15 0123 16 NATORB then the first data line of the CONF directive specifies the SCF configuration and it is this configuration that should be nominated in the PROP data Thus the Cl data including the property analysis would appear as follows MRDCI TRAN 1 CORE 41 11 DATA FOR CONVENTIONAL TABLE CI TRANSITION MOMENTS 32 1 TO 4 1 SELECT SINGLES 1 CONF 0123 15 013 4 15 0123 16 NATORB PROP CIVEC 1 0123 15 where one such data line is required given the specification of CIVEC 11 Data for Conventional Table CT Transition Moments This Table Cl module will calculate both electrical and magnetic dipole moments as well as oscillator strengths and lifetimes of excited states The module will look to calculate the moment between a specific state typically the ground state and a set of addition
36. out an initial Cl where the reference set employed acts to provide at least a quali tative description of the states of interest 3 based on the output from the initial Cl we augment the reference set to provide a quantitative description of the first three states 4 finally having generated the Cl vectors for the three states we carry out in the final job an analysis of each vector in terms of natural orbitals and one electron properties and generate the transition moments between the ground and two excited states We now consider various aspects of each job in turn Job 1 The SCF TITLE x H20 TZVP DIFFUSE S P MRDCI SUPER OFF NOSYM ZMAT ANGSTROM 0 H 1 0 951 26 CALCULATING THE A STATES OF H30 H 1 0 951 2 104 5 END BASIS TZVP O TZVP H So 1 0 0 02 PO 1 0 0 02 END ENTER 93 The only points to note here is the use of the SUPER directive in suppressing skeletonisation and in the absence of section specification on the ENTER directive the use of the default section for output of the SCF eigenvectors section 1 Job 2 The Initial 3M 3R CI An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS and the following orbital assignments characterising the closed shell SCF configuration 1a72a71b33a 1b 20 56084959 1 35696939 0 72200122 0 58247942 0 50858566 02724259 04894440 05589681 06133571 20403420 2
37. that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output FPRINT or DEBUG to produce output suitable for debugging purposes TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass TM processing Such usage is typically associated with restarting Table Cl calculations The second data line of the TM directive is used to specify the location of the Cl vectors and the number of excited state vectors involved in the subsequent analysis The line is read to the variables IFTNX ISECX IFTNE ISECE NSTATE using format 51 IFTNX defines the FORTRAN data set reference number of the interface holding the Cl vector of the first state Normally this vector will reside on FTNO36 with IFTNX 36 ISECX defines the position of this first vector on the interface defined by IFTNX Typically for the first state of a given symmetry the vector will be located first on the data set i e ISECX 1 IFTNE defines the FORTRAN data set reference number of the interface holding the Cl vector s of the set of additional states Assuming these states are of the same symmetry as the first then we would expect all the states involved to lie on the same interface i e IF TNE will also be set to 36 If however the set of states is
38. the SUPER directive in suppressing skeletonisation and use of the default eigenvector section section 1 for storage of the closed shell eigenvectors An examination of the SCF output reveals the following orbital analysis TABLE CI CALCULATIONS USING MCSCF ORBITALS IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 20 57768533 2 0000000 2 1 11 34457777 2 0000000 3 1 1 40746540 2 0000000 4 1 0 87003449 2 0000000 5 3 0 69591811 2 0000000 6 1 0 65109519 2 0000000 7 2 0 53687971 2 0000000 8 3 0 44174805 2 0000000 9 2 0 11697212 0 0000000 10 1 0 26220763 0 0000000 11 1 0 27217357 0 0000000 12 3 0 38931080 0 0000000 13 3 0 41757152 0 0000000 14 2 0 46526352 0 0000000 15 1 0 60968525 0 0000000 16 1 0 75001014 0 0000000 17 2 0 86980119 0 0000000 18 1 0 89167074 0 0000000 19 3 0 93051881 0 0000000 20 1 1 07098621 0 0000000 21 3 1 18616042 0 0000000 22 1 1 35370640 0 0000000 23 4 1 52221224 0 0000000 24 2 1 68895841 0 0000000 25 1 1 88480823 0 0000000 26 3 1 97392259 0 0000000 27 4 2 13452238 0 0000000 28 1 2 13982211 0 0000000 29 3 2 19187925 0 0000000 30 1 2 34994146 0 0000000 31 2 2 36355601 0 0000000 32 4 2 67698155 0 0000000 33 1 2 82812279 0 0000000 34 2 2 84696649 0 0000000 35 3 2 97321688 0 0000000 36 1 3 14466153 0 0000000 37 1 3 33275225 0 0000000 38 3 3 51306001 0 0000000 39 1 3 51697350 0 0000000 40 2 3 52974692 0 0000000 14 TABLE CI CALCULATIONS USING MCSCF ORBITALS 41 3 3 68500148 42 1 3 79236483 43 3 3
39. then be as follows 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 18 TITLE SIH4 6 31G MRDCI VALENCE CI 1M 1R ZMAT SI H 1 SIH H 1 SIH 2 109 471 H 1 SIH 2 109 471 3 120 0 H 1 SIH 2 109 471 4 120 0 VARIABLES SIH 2 80 END BASIS 6 31G RUNTYPE CI MRDCI TRAN CORE 2111 12111 SELECT CONF 018 13 18 SINGLES 1 NATORB ENTER Example 4 In this example we wish to perform a valence Cl calculation on the N2 molecule using a 4 31G basis While the molecular symmetry is Don the symmetry adaptation and subsequent Cl will be conducted in the Dg point group The resolution of the Daon into the D orbital species is given in Table 2 An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 15 65951533 2 0000000 5 15 65474750 2 0000000 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 19 3 1 1 50615941 2 0000000 4 5 0 75782277 2 0000000 5 1 0 63244925 2 0000000 6 3 0 63135826 2 0000000 7 2 0 63135826 2 0000000 8 6 0 20154861 0 0000000 9 7 0 20154861 0 0000000 10 5 0 63883097 0 0000000 11 1 0 82491489 0 0000000 12 3 0 89634343 0 0000000 13 2 0 89634343 0 0000000 14 1 0 91812387 0 0000000 15 7 1 10036132 0 0000000 16 6 1 10036132 0 0000000 17 5 1 17625689 0 0000000 18 5 1 66995008 0 0000000 19 4 1 70518236 0 0000000 20 1 1 70518236 0 0000000 21 3 1 91001614 0 0000000 22 2 1 91001614 0 0000000 23 8 2 29436539 0 0000000 24 5 2 294365
40. us to bypass the transformation Note also the specific appearance now of ADAPT in the data to enable bypassing RESTART CI TITLE kkk H20 TZVP DIFFUSE S P TABLE CI 12M 3R SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM 0 H 1 0 951 H 1 0 951 2 104 5 END BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI MRDCI ADAPT BYPASS TRAN CORE DISC BYPASS 1000 1 1000 18 SELECT 12 CALCULATING THE A STATES OF H20 39 CONF 0 12 17 23 2 2 3 1 17 23 2 2 4 1 17 23 2 2 5 1 17 23 21718 12 23 21819 12 23 4 17 182325 1 2 4 17 18 2327 1 2 4 2 71718 1 23 4 2 31719 1 23 4 2 32325 1 17 4 2 32327 1 17 THRE 10 10 ROOT 3 ENTER Job 4 The Analysis Assuming that the diagonalisation interface FTNO36 had been saved above then the final analysis job is straightforward again bypassing of the various sub modules involves explicit mention of the ADAPT Cl and DIAG modules in addition to flagging the previous TRAN and SELECT data lines with the BYPASS keyword We have routed the natural orbitals from the 3 A states to the Dumpfile using the PUTQ directive RESTART CI TITLE kk H20 TZVP DIFFUSE S P TABLE CI ANALYSIS SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM 0 H 1 0 951 H 10 951 2 104 5 END BASIS TZVP O TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI MRDCI ADAPT BYPASS TRAN CORE DISC BYPASS 1000 1 1000 18 SELECT BYPASS CONF 12 CALCULATING THE A STATES O
41. used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress output from the module IPRINT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes 6 Data for Conventional Table CI Selection Data for the configuration selection module is initiated with the SELECT directive followed by those directives characterising the symmetry of the state s of interest and reference configura tions CNTRL SPIN SYMMETRY CONF etc and terminated by data ROOTS THRESH controlling the process of selection 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 9 6 1 SELECT The SELECT directive is used to control the configuration selection module and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string SELECT e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module in particular all details of the perturbative energy lowerings associated with the initial set of configurations IPRINT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes This includes the energy lowerings associated with the complete configuration list e
42. 0 0 0000000 30 1 3 58631019 0 0000000 31 4 3 59701772 0 0000000 32 1 3 84174131 0 0000000 33 1 4 84610143 0 0000000 34 3 5 14220270 0 0000000 35 1 7 73115986 0 0000000 36 1 47 56758932 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the Ols inner shell orbitals and discard the inner shell complement orbital IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos az 1 18 1 16 1 16 b 2 6 0 6 17 22 ba 3 10 0 10 23 32 a2 4 2 0 2 33 34 Note that the virtual SCF MOs dominated by the diffuse oxygen basis functions are the 4a the 2b the 2b and the 5a with SCF sequence numbers 6 7 8 and 9 respectively The symmetry 12 CALCULATING THE A STATES OF HO 36 re ordered sequence numbers allowing for the effective removal of the two a orbitals are 3 24 18 and 5 respectively To perform a three root 8 electron valence Cl calculation based on the SCF configurations of the ground and excited Rydberg states involving the single excitations 1b to 2b and 3a to 4a would require the following CONF data CONF 0 12 17 23 223 1 17 23 21718 12 23 The following data will perform this three root Cl where e the SCF computation is BYPASS ed e both CORE and DISCard are specified on the TRAN data line flagging the freezing and discarding of the two a MOs e the default sub module specifications are in eff
43. 000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 106 28 TABLE CI CALCULATIONS USING MCSCF ORBITALS 41 3 3 68500148 42 1 3 79236483 43 3 3 83958499 44 4 3 85444369 45 2 3 87171104 46 2 4 14123645 47 3 4 16304613 48 1 4 16535425 49 2 4 16620625 50 2 4 32796704 51 3 4 47390388 52 1 4 49325977 53 1 4 68584478 54 4 4 89513471 55 3 5 13208791 56 1 5 14756255 57 2 5 85901984 58 1 5 92954017 59 3 6 06228738 60 1 8 35480844 61 1 27 71561806 62 1 45 72664766 Job 2 The MCSCF o0o00D0DO0O0O0O0O0O000O0000000000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 107 The following data performs a 10 electron in 9 orbital CASSCF calculation using the MCSCF module with the natural orbitals routed to section 10 of the Dumpfile under control of the CANONICAL directive In the absence of the VECTORS directive the SCF MOs will be used as the starting orbitals This data set is just that provided in the conventional Table Cl case shown above RESTART TITLE H2C0 MCSCF 10E IN 9 M O SUPER OFF NOSYM NOPRINT BYPASS ZMATRIX ANGSTROM N C 0 1 1 203 H 11 099 2 121 8 H 11 099 2 121 8 3 180 0 END BASIS DZP 0 D
44. 0000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 71 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 68 1 69 5 70 5 11 65153449 12 12852685 25 31135299 0 0000000 0 0000000 0 0000000 72 Based on the above output the CONF data lines may be deduced from the following table where we again assume that we wish to freeze the two Nis inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos Og 1 16 1 15 1 15 nae 2 8 0 8 16 23 Tuy 3 8 0 8 24 31 dgay 4 3 0 3 32 34 Ou 5 16 1 15 35 49 Tgr 6 8 0 8 50 57 Toy T 8 0 8 58 65 dury 8 3 0 3 66 68 To perform a 10 electron valence Cl calculation based on the SCF configuration 2923 2174 207207905 lTu and associated 7 to 7 excitations 29 29 27172 172 20420490 lna y 1Tg 29 293 2172 172 204204903 lna q 1Tg 207207302 Itu eltg x laity ag T Y based on the SCF configuration would require the following CONF data CONF 01216 24 35 01 2 24 35 50 O 1 2 16 35 38 4 16 24 50 58 1 2 35 END 27 28 29 30 The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE N2 SUPER OFF NOSYM ZMAT ANGS N N 1 VARIABLES NN 1 05 END CC PVTZ NN 19 DATA FOR SEMEDIRECT TABLE CI SELECTION BASIS CC PVTZ RUNTYPE CI CORE 1 2 END MRDCI DIRECT TABLE
45. 000000 4 1 0 87463564 2 0000000 5 3 0 70990765 2 0000000 6 1 0 64751394 2 0000000 7 2 0 53989416 2 0000000 8 3 0 44423257 2 0000000 9 2 0 10853108 0 0000000 10 1 0 25726604 0 0000000 11 1 0 28106873 0 0000000 12 3 0 38903939 0 0000000 13 3 0 40966861 0 0000000 14 2 0 46216570 0 0000000 15 1 0 65466944 0 0000000 16 1 0 82879998 0 0000000 17 2 0 98111608 0 0000000 18 1 0 98701051 0 0000000 19 3 1 07064863 0 0000000 20 1 1 16621340 0 0000000 21 3 1 29856111 0 0000000 22 1 1 82320845 0 0000000 23 1 23 76352004 0 0000000 24 1 43 36689896 0 0000000 4 DATA FOR CONVENTIONAL TABLE CI TRANSFORMATION 6 Thus the orbitals of interest are of common Rrep a1 with sequence numbers 1 2 core and 13 14 complement MOs within the re ordered a set The following TRAN data would freeze and discard these MOs TRAN 1 CORE DISCARD 2000 12 2000 13 14 or assuming the default eigenvector specification is in effect simply TRAN CORE DISCARD 2000 12 2000 13 14 The following sequence would be used to simply freeze the orbitals while retaining the complete virtual manifold TRAN CORE 2000 12 Example 2 In this example we wish to perform a valence Cl calculation on the No molecule using a TZVP basis While the molecular symmetry is Doon the symmetry adaptation and subsequent Cl will be conducted in the D point group The resolution of the Doon into the D orbital species is given below in Table 2 An examination o
46. 18 118 M 0 120 120 0 129 129 M 0 130 130 0 363 363 0 886 887 0 1226 1227 0 1236 1237 0 1246 1247 0 1640 1641 0 1644 1645 0 SUM OF MAIN REFERENCE C C 76 2724087 0 0000790 75 9009069 0 0002121 C C CONFIGURATION 00020304 1 2 17 00284448 1 2 18 06411957 2 3 1 00757395 2 5 1 384004197 17 18 1 00261567 17 19 1 01150383 18 19 1 00221466 1 5 17 00227863 2 5 17 00441966 2 6 17 00550738 2 7 17 00666055 17 18 23 00664455 17 18 23 0 90436458 23 23 17 17 N 18 18 18 18 25 27 23 23 23 23 23 ererrreN 23 23 23 23 12 CALCULATING THE A STATES OF HO Description of the 2 4 state EXTRAPOLATED ENERGY 00017134 78352745 00620129 05602432 07076103 00236797 00343646 00493434 00870751 00638562 00600107 SUM CSF NO 1 1 M 0 118 118 M 0 119 119 0 120 120 0 129 129 M 0 211 211 0 212 212 0 213 213 0 1212 1213 0 1372 1373 0 1376 1377 0 OF MAIN REFERENCE C C 0 85445982 75 8825335 0 0002079 CONFIGURATION 1 2 17 2 3 1 2 4 1 2 5 1 17 18 1 3 5 1 3 6 1 3 7 1 2 3 17 2 3 23 2 3 23 23 17 17 17 17 17 17 19 25 27 23 23 23 23 23 23 23 38 23 17 17 Taking as the criterion for inclusion a weight of 0 005 the final 12 reference set Cl is shown below We have assumed that the FORTRAN interface FTNO31 has been saved from the second job enabling
47. 21 Tg 6 3 0 3 22 24 Togo Y 3 0 3 25 27 vay 8 1 0 1 28 To perform a 10 electron valence Cl calculation based on the SCF configuration 207207302 17 23 and associated 7 to T excitations 202202302177 172 ve 20520 304 174 q 175 y 22 207207307 1ltu eltg c 1tuylrgy 26 would require the following CONF data CONF 0128 11 15 012 11 15 22 012815 25 4 8 11 22 25 12 15 END The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE N2 4 31G SUPER OFF NOSYM ZMAT ANGS N N 1 NN VARIABLES NN 1 05 END BASIS 4 31G 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 70 RUNTYPE CI CORE 1 2 END MRDCI DIRECT TABLE SELECT CNTRL 10 SYMM 1 SPIN 1 SINGLES 1 CONF 012811 15 012 11 15 22 012815 25 48112225 1215 END CI NATORB IPRIN ENTER Now consider the corresponding calculation performed in a larger CC PVTZ basis An exami nation of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 15 66553669 2 0000000 2 5 15 66076870 2 0000000 3 1 1 50580210 2 0000000 4 5 0 76246663 2 0000000 5 1 0 63737662 2 0000000 6 2 0 63425262 2 0000000 7 3 0 63425262 2 0000000 8 T 0 18426391 0 0000000 9 6 0 18426391 0 0000000 10 1 0 40582614 0 0000000 11 5 0 42020784 0 0000000 12 2 0 51942385 0 0000000 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
48. 2824210 53700802 56235022 58645643 ere PRO0O0O0JOQO0dmroyN m N m w B H U WU e H H N WRK oooooooo o a BS DOOD ODO OOO ONNNNN 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 26 CALCULATING THE A STATES OF H30 N a PRRERORRPpRNORNRARAOQOROOQON gt R RR 66887228 74805617 07690608 88545053 92243836 12944874 20541910 34202871 39946430 69788310 72651832 13832720 07664215 26840142 54616570 58631019 59701772 84174131 84610143 14220270 13115986 47 56758932 No PWWWWWWNHNNNNNNRKPRFPRP OO o00D0DO0O0O0O0O0O000O0000000000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 94 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the Ols inner shell orbitals and discard the inner shell complement orbital IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos az 1 18 1 16 1 16 b 2 6 0 6 17 22 ba 3 10 0 10 23 32 a2 4 2 0 2 33 34 Note that the virtual SCF MOs dominated by the diffuse oxygen basis functions are the 4a the 2b2 the 2b and the 5a with SCF seq
49. 3 012345 30 42 43 012345 29 42 44 ROOTS 1 THRESH 2 2 CI DIAG ENTER Finally we consider the data for performing exactly the same calculation as above but now freezing the oxygen and carbon 1s core orbitals in the Table Cl calculation The following points should be noted e The TRAN directive now appends the CORE descriptor after the section specification with the following 2 data lines requesting that the first two orbitals of symmetry 1 be removed from the Table Cl calculation e The orbital indices specified on the CONF data lines reflect the removal of these two orbitals with the CNTRL directive now pointing to a 12 electron Cl calculation as distinct from the 16 electron calculation above RESTART TITLE H2CO MRDCI FROM MCSCF NOS FREEZE 1S MOS SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 15 THE SEMEDIRECT TABLE CI MODULE 51 END RUNTYPE CI SCFTYPE MCSCF MCSCF ORBITAL COR1 COR1 COR1 DOC1 DOC3 DOC1 DOC2 DOC3 VOC2 VOC1 UDC3 UDCA END PRINT ORBITALS VIRTUALS NATORB CANONICAL 10 FOCK DENSITY FOCK MRDCI ADAPT TRAN 10 CORE 2000 12 SELECT SYMMETRY 1 SPIN 1 CNTRL 12 SINGLES 1 CONF 0123 27 40 41 0123 28 40 41 0123 27 40 42 ROOTS 1 THRESH 2 2 CI DIAG ENTER 15 The Semi direct Table CI Module We now consider the data requirements and file structure of the ne
50. 3 98212497 0 0000000 41 1 4 08851360 0 0000000 42 1 24 51368240 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the first 14 inner shell orbitals 107207307407110150760720 707807314 6 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos al 1 22 8 14 1 14 b 2 9 3 6 15 20 ba 3 9 3 6 21 26 a2 4 2 0 2 27 28 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 16 To perform an 18 electron valence Cl calculation based on the SCF configuration 9074016410075 741107 7 would require the following CONF data CONF 01234 1516 21 22 27 The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE CUCL 3 21G ZMAT ANGSTROM CU CL 1 CUCL VARIABLES CUCL 2 093 END BASIS 3 21G RUNTYPE CI MRDCI TRAN CORE 8330 1 T0 8 1 T03 1 TO 3 SELECT SINGLES 1 CONF 01234 1516 21 22 27 NATORB ENTER The inclusion of a second reference configuration corresponding to the doubly excited configu ration 90 47 19 100 571 120 8 would require the following CONF data CONF 01234 1516 21 22 27 01235 1516 21 22 27 Example 3 Consider performing a valence Cl calculation on the SiH molecule using a 6 31G basis While the molecular symmetry is Tg the symmetry adaptation and subsequent Cl will be conducted in the Ca point group An examination of the SCF output reveals the following o
51. 39 0 0000000 25 1 2 84356916 0 0000000 26 7 3 00847817 0 0000000 27 6 3 00847817 0 0000000 28 5 3 37447679 0 0000000 29 sl 3 71753400 0 0000000 30 5 4 09917273 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the two Nls inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos Og 1 8 1 7 1 7 Tag 2 3 0 3 8 10 ma 3 3 0 3 11 13 dejo 4 1 0 1 14 Tu 5 8 1 7 15 21 Tga 6 3 0 3 22 24 Tea 7 3 0 3 25 27 uzy 8 1 0 1 28 To perform a 10 electron valence Cl calculation based on the SCF configuration 29 29 21 4 20320 307 17 9 and associated 7 to 7 excitations 202030 lrg 2m y 10 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 20 2072073021 on y ay 2020 30 ria e rig 2m 12 would require the following CONF data CONF 0128 11 15 012 11 15 22 0128 15 25 4 8 11 22 25 12 15 The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE N2 4 31G SUPER OFF NOSYM ZMAT ANGS N N 1 NN VARIABLES NN 1 05 END BASIS 4 31G RUNTYPE CI MRDCI TRAN CORE 10001000 1 1 SELECT SINGLES 1 CONF 0128 11 15 012 11 15 22 012815 25 4 8 11 22 25 12 15 NATORB IPRIN ENTER Now consider the corresponding calculation performed in a smaller 3 21G basis An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED
52. 8 0 0000000 Thus the inner shell Nis orbitals the lo and lo transform as ay and bj respectively in Da symmetry with Rrep numbers 1 and 5 Both correspond of course to the first orbital in the appropriate IRrep The following sequence would be used to simply freeze the orbitals while retaining the complete virtual manifold CORE 1 2 END Turning to the inner shell complement orbitals we again find the corresponding Rreps 1 and 5 with the orbitals the highest lying member in each case with relative sequence number of 11 Thus the following ACTIVE and CORE data would act to both freeze and discard the inner shells and their complement MOs ACTIVE 3 TO 38 END CORE 1 2 END Further examples of transformation data will be discussed below in the section discussing Ref erence Function specification within the SELECT data 18 Data for Semi direct Table CI Data base Generation The data base of pattern symbolic matrix elements required by both the Selection and Cl modules may be generated by the user in the course of any Table Cl calculation In contrast to the Conventional module we envisage that this step will be executed in each run of the module rather than the user allocating a previously generated version 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 58 18 1 TABLE The TABLE directive is used to request and control the data base generator and comprises a single data line read to the variables TEXT TEXTF TEXTB using f
53. ALCULATING THE A STATES OF H20 33 IFTNX defines the FORTRAN data set reference number of the interface holding the Cl vector of the first state Normally this vector will reside on FTNO36 with IFTNX 36 ISECX defines the position of this first vector on the interface defined by IFTNX Typically for the first state of a given symmetry the vector will be located first on the data set i e ISECX 1 IFTNE defines the FORTRAN data set reference number of the interface holding the Cl vector s of the set of additional states Assuming these states are of the same symmetry as the first then we would expect all the states involved to lie on the same interface i e IFTNE will also be set to 36 If however the set of states is of different symmetry to the first then their vectors will almost certainly reside on a different interface which we will assume reside on FTNO37 i e IF TNE should be set to 37 ISECE defines the position of the first of the excited state vectors on the interface defined by IFTNE Typically if all the states involved are of the same symmetry residing on the same data set IFTNX IFTNE then the first excited state vector will be located second on the data set i e ISECE 2 When the first and excited states are of different symmetry then different data sets will be involved and the first of the excited state vectors will be the first on IFTNE NSTATE defines the number of excited state vectors involved and is usu
54. CONTENTS i Computing for Science CFS Ltd CCLRC Daresbury Laboratory Generalised Atomic and Molecular Electronic Structure System i 3 _ _ _ _ _ _ _ l __ GAMESS UK USER S GUIDE and REFERENCE MANUAL Version 8 0 June 2008 PART 6 TABLE Cl CALCULATIONS M F Guest J Kendrick J H van Lenthe and P Sherwood Copyright c 1993 2008 Computing for Science Ltd This document may be freely reproduced provided that it is reproduced unaltered and in its entirety Contents 1 Introduction 1 1 1 Sub Module Structure of Conventional Table Cl 1 2 Directives Controlling Conventional Table Cl Calculations 2 2L Os iran de a du BE we hk Oe Rd aa 2 3 Data for Conventional Table Cl Symmetry Adaptation 2 eee Se ie ie oe AN 4 Data for Conventional Table Cl Transformation 4i TRAN gee ae a ch Gi Se ck ee we Hd a a nO oe a CONTENTS 5 Data for Conventional Table Cl Data base Generation S1 AR e g e Athen a wh a de AR Ae hd Bae e ee td 6 Data for Conventional Table Cl Selection OL SELECT sica SPS Oe raras 02 CNTRE gt y ESOS He ee eee eee 4 EES a Oe Wes a a oe ee ee ta ta 64 SYMMETRY AI 6S SINGLES osas a a a BA A AR RAE a ee i GO CONF 2 46 655 6448 2b a aaa e oR ERS OES EK Ewe 7 ROOTS in 2 Oe re oC ew Oe See ew OEE Ses EOE OG WHRESEs 6 460 boone eS daa ho a e 7 Data for Conventional Table Cl H Matrix Construction Tb OP ne Aba eee hehe Bao OME be RRA et eh hh ek 8 Data for Conventional Table Cl Diag
55. F H20 40 17 23 17 23 17 23 2 3 2 4 2 5 17 18 18 19 17 18 23 25 17 18 23 27 2 7 17 18 2 3 17 19 2 3 23 25 2 3 23 27 THRE 10 10 ROOT 3 CI BYPASS DIAG BYPASS NATORB IPRIN 1 1 1 1 1 1 23 23 17 17 SRR ER RNNNNNO Rp CIVE 1 2 3 PUTQ AOS 50 51 52 PROP CIVE 1 2 3 0 12 17 23 21718 12 23 2 2 3 1 17 23 MOMENT 36 1 36 2 2 ENTER Description of the Output for MRDCI Moments The MRDCI module calculates the oscillator strength using both the dipole length formalism f r 2 3 lt Wrjw gt AE and the dipole velocity formalism f y 2 3 lt Y y v gt AE The most significant contributions due to individual molecular orbitals are printed out as a table containing the largest coefficients of the transition density matrix and the following correspond ing integrals lt Vilrl gt lt dilyly gt lt wil2 hj gt lt bil Y 2 113 gt lt bil Y y gt lt wil Y ly gt The f r and f y values are printed out in x y z components and the expectation values for lt Y Yi Zi Yi zily gt are also printed 13 ITERATIVE NATURAL ORBITAL CALCULATIONS 41 13 Iterative Natural Orbital Calculations We work through an example of using the natural orbitals generated by the module in a sub sequent Cl calculation We consider a DZ calculation on the ground state of C2Hx4 with the computation split into four separate jobs in which we 1 perform the initial SCF 2 carry out a
56. IRREP the doubly excited configuration arising from excitation of the highest occupied DOMO 24 SEMI DIRECT TABLE CI USING DEFAULT OPTIONS 85 of that symmetry to the lowest virtual orbital VMO of the same symmetry and ii the lowest singly excited configuration again arising from the highest occupied DOMO to the lowest VMO of the same symmetry In the present example this will correspond to the SCF configuration the double and single excitation arising from the DOMO 5a to VMO 6a the double and single excitation arising from the DOMO 1b to VMO 2b and the double and single excitation arising from the DOMO 2b to VMO 3b gt No reference configurations will be included involving orbitals of ag symmetry given the absence of such orbitals involved in the occupied manifold This results in a total reference set of 7 functions as shown thus in the job output numbers of open shells and corresponding main configurations 0 1 2 3 4 5 28 37 38 T SCF configuration 0 1 2 3 4 6 28 37 38 es 5al gt 6al double 2 5 6 1 2 3 4 28 37 38 5al gt 6al single 0 1 2 3 4 5 29 37 38 i 1b1 gt 2b1 double 2 28 29 1 2 3 4 5 37 38 1b1 gt 2b1 single 0 1 2 3 4 5 28 37 39 ts 2b2 gt 3b2 double 2 38 39 1 2 3 4 5 28 37 2b2 gt 3b2 single 6 The default selection process subsequently undertaken is equivalent to the following ROOTS and THRESH directives THRESH 10 10 ROOTS 1 Thus this default selection process involves construction of an e
57. MMETRY 1 SPIN 2 CNTRL 11 THRESH 2 2 ROOTS 5 ITERATE MAXI 20 MAXROOT 8 SROOT 0 30 DROOT NATORB BYPASS ENTER The following points should be noted 1 The C and O 1s inner shell MOs together with the two highest virtual orbitals are excluded from the calculation under control of the CORE and ACTIVE directives 2 The SELECT directives requesting the 11 electron doublet Cl states of A symmetry are as follows SYMMETRY 1 SPIN 2 CNTRL 11 3 The ITERATE directive ITERATE MAXI 20 MAXROOT 8 SROOT 0 30 DROOT requests derivation of the lowest 8 roots allowing a maximum of 20 MRDCI iterations to derive these roots The initial Cl will be based on the lowest 5 roots of the Cl Hamiltonian ROOTS 5 derived from the default set of generated reference configurations 29 3 4 The 2A States Treating the 2A states is somewhat more complex given the need to perform an additional SCF calculation on the lowest state of that symmetry We now perform an initial SCF calculation on the ground state and initiate the Cl calculation in the second step by restoring these MOs interchanging the appropriate orbitals under control of the SWAP directive and conducting the SCF prior to the Cl TITLE HCO DZP BOND SP 2AP SCF HARMONIC 29 ITERATIVE MRDCI CALCULATIONS 121 MULT 2 ZMAT C BQ 1 RCO2 X 21 01 90 0 D 2 RCO2 3 90 0 1 180 0 X 11 0 2 90 0 3 0 0 H 1 RCH 5 40 0 4 180 0 VARIABLES RCO2 1 125 RCH 2 076 END BASIS DZP H
58. N data line flagging the freezing and discarding of the la and 1b MOs The default section for retrieval of the closed shell eigenvectors is assumed e the default sub module specifications are in effect with no specific need to reference ADAPT Cl or DIAG activity we also assume that the TABLE data set is available to the job RESTART TITLE ETHYLENE CI GROUND STATE SCF MOS BYPASS SCF ZMATRIX ANGSTROM Cc 4 1 2 120 0 1 2 120 0 3 180 0 11 120 0 3 0 0 El C H H H H 120 0 3 180 0 NNRPRP PRO BPRPPPPR END BASIS DZ RUNTYPE CI MRDCI TRAN CORE 100100 11 SELECT SYMMETRY 1 SPIN 1 SINGLES 1 CONF 0128 10 14 23 ENTER Job 3 The 2M 1R CI We show below the final Cl vector for the ground state 13 ITERATIVE NATURAL ORBITAL CALCULATIONS 44 Description of the X A state EXTRAPOLATED ENERGY CSF NO 1 1 M 43 43 142 142 374 375 78 1944869 0 0001635 C C CONFIGURATION 0 90889665 1 2 8 10 14 23 0 01645740 1 2 10 14 21 23 0 00543252 21 22 1 2 10 14 23 0 00328957 2 8 17 21 1 10 14 23 Taking as the criterion for inclusion a weight of 0 01 the 2 reference set Cl job is shown below We have assumed that the FORTRAN interface FTNO31 has been saved from the second job enabling us to bypass the transformation Note also the specific appearance now of ADAPT in the data to enable bypassing The natural orbitals are routed to section 200 of the Dumpfile Note that subsequent usage of the
59. NOs by the Conventional Table Cl module requires the SABF specification on the PUTQ directive RESTART NEW TITLE ETHYLENE CI GROUND STATE 2M SCF MOS BYPASS SCF ZMATRIX ANGSTROM 120 0 Taca NNrRR RO BPRPRPP Pe END BASIS DZ RUNTYPE CI MRDCI ADAPT BYPASS TRAN CORE BYPASS 100100 11 SELECT SYMMETRY 1 SPIN 1 SINGLES 1 CONF 0128 10 14 23 012 10 14 21 23 NATORB IPRINT PUTQ SABF 200 ENTER 120 0 3 0 0 4 12 1 2 120 0 3 180 0 11 1 1 120 0 3 180 0 Job 4 The Natural Orbital CI 13 ITERATIVE NATURAL ORBITAL CALCULATIONS 45 We show below the data for using the NOs from the 2 reference Cl where the orbitals routed to section 200 are now restored by specification on the TRAN directive The following points should be noted e We assume that the adaptation interface FTNO22 has been saved from the initial Cl job allowing the BYPASS specification on the ADAPT directive e Restoring the NOs must be controlled via TRAN specification an attempt to restore such orbitals through VECTORS and ENTER specification will lead to an error condition e The resulting NOs from the natural orbital Cl are now routed to section 210 and could be used in a subsequent Cl in obvious fashion RESTART NEW TITLE ETHYLENE CI GROUND STATE 2M NOS BYPASS SCF ZMATRIX ANGSTROM C C 11 4 H 1 1 1 2 120 0 H 1 1 1 2 120 0 3 180 0 H 2 1 1 1 120 0 3 0 0 H 2 1 1 1 120 0 3 180 0 END BASIS DZ RUNTYPE CI MRDCI ADAP
60. NT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass Cl processing Such usage is typically associated with restarting Table Cl calculations 8 Data for Conventional Table CI Diagonalisation Data input controlling the diagonalisation of the final Cl Hamiltonian is introduced by the DIAG directive The process of extrapolation to zero selection threshold involves the diagonalisation module solving not just one but several secular problems corresponding to a range of selection thresholds The number of so called extrapolation passes is specified by the EXT RAP directive In default the module will generate NROOT eigenvectors of the Cl matrix on each pass where NROOT is the number of roots specified by the ROOTS directive at selection time Thus the solutions of the zero order Hamiltonian will be used through a maximum overlap criterion in deriving the final Cl eigenvectors Additional data may be specified to override this default and provide various convergence and printing controls 8 DATA FOR CONVENTIONAL TABLE CI DIAGONALISATION 26 8 1 DIAG The DIAG directive comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string DIAG e TEXTF is an optional parameter that may be used to control the quantity
61. PIN TRIPLET are equivalent the wavefunction will be three fold spin degenerate 6 4 SYMMETRY This directive consists of one line read to variables TEXT NSYM using format A l e TEXT should be set to the character string SYMMETRY e NSYM is an integer parameter used to specify the spatial symmetry of the Cl wavefunction and is set to the appropriate sequence number of the required irreducible representation see Table 1 The SYMMETRY directive may be omitted when the program will set NSYM to 1 i e the totally symmetric representation see section 4 6 2 Example In a system of Ca symmetry the data line SYMMETRY 3 would be required when performing calculations on states of B2 symmetry Failure to present the directive in such cases will lead to the default Ay symmetry 6 5 SINGLES This directive consists of one line read to variables TEXT NREF using format A l e TEXT should be set to the character string SINGLES e NREF is an integer parameter used to nominate a particular configuration within the set of reference functions The selection module will then retain in the final Cl all single excitations with respect to the nominated function regardless of their computed energy lowerings 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 11 Table 1 Resolution of the Co Species into the C2 Species Orbital IRrep Casi Coy Sequence No o ay 1 bx2 y2 Ta b 2 Ty ba 3 Oxy a2 4 Table 2 Resolutio
62. S BQ 1 0 0 02 P BQ 1 0 0 02 D BQ 1 0 0 02 DZP C DZP 0 END OPEN 1 1 ENTER CORE 20000000 TIME 300 RESTART NEW TITLE HCO DZP BOND SP 2APP SCF CI HARMONIC MULT 2 ZMAT C BQ 1 RCO2 X 2 1 0 1 90 0 D 2 RCO2 3 90 0 1 180 0 X 1 1 0 2 90 0 3 0 0 H 1 RCH 5 40 0 4 180 0 VARIABLES RCO2 1 125 RCH 2 076 END BASIS DZP H S BQ 1 0 0 02 P BQ 1 0 0 02 D BQ 1 0 0 02 DZP C DZP 0 END REFERENCES 122 OPEN 1 1 CORE 1 2 END ACTIVE 3 TO 42 END RUNTYPE CI MRDCI DIRECT SYMMETRY 2 SPIN 2 CNTRL 11 THRESH 2 2 ROOTS 5 ITERATE MAXI 20 MAXROOT 8 SROOT 0 30 DROOT NATORB BYPASS VECTORS 5 SWAP 8 10 END ENTER 4 5 References 1 R J Buenker in Proc of the Workshop on Quantum Chemistry and Molecular Physics Wollongong Australia 1980 R J Buenker in Studies in Physical and Theoretical Chem istry 21 1982 17
63. T BYPASS TRAN 200 CORE 100100 11 SELECT SYMMETRY 1 SPIN 1 SINGLES 1 CONF 0128 10 14 23 012 10 14 21 23 NATORB IPRINT PUTQ SABF 210 ENTER We show below the final Cl vector from the natural orbital Cl Description of the X A state EXTRAPOLATED ENERGY 78 2021125 0 0002347 CSF NO C C CONFIGURATION 14 TABLE CI CALCULATIONS USING MCSCF ORBITALS 46 1 1 M 0 89795112 1 2 8 10 14 23 29 29 M 0 03202257 1 2 10 14 21 23 194 195 0 00279273 1 8 15 21 2 10 14 23 308 309 0 00499259 2 8 15 21 1 10 14 23 430 431 0 00259858 3 8 14 21 1 2 10 23 514 515 0 00297122 10 11 23 24 1 2 8 14 SUM OF MAIN REFERENCE C C 0 92993884 14 Table CI Calculations Using MCSCF Orbitals To conclude our discussion of the Conventional Table Cl module we work through an example of using the natural orbitals generated from the MCSCF module in a subsequent Cl calculation We consider a calculation on the ground state of H2CO with the computation split into three separate jobs in which we 1 perform an initial SCF 2 carry out the MCSCF calculation 3 perform the MRDCI calculation using the MCSCF natural orbitals We now consider various aspects of each job in turn Job 1 The SCF TITLE H2C0 DZP F SUPER OFF NOSYM ZMATRIX ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 END ENTER The only points to note here is the use of
64. TORB CIVEC 1 VECTORS ATOMS SWAP 78 END ENTER 4 5 25 Memory Specification for the Semi direct Table CI Mod ule The memory requirements of the semi direct module are typically greater than those associated with the conventional algorithm While the default memory allocations will prove adequate for small medium cases the user should use the MEMORY pre directive to increase this allocation in more demanding cases e g at least 8 MWords in calculations with say more than 20 active electrons In this section we outline the main demands on memory and the mechanisms for increasing this allocation should the default allocations prove inadequate The overall memory required is determined by four integer quantities namely 1 NTEINT the number of transformed two electron integrals The default value is 3 500 001 2 IOTM the field length for the selected configurations The default value is 2 000 000 25 MEMORY SPECIFICATION FOR THE SEMEDIRECT TABLE CI MODULE 91 3 NEDIM the dimension of internal arrays required by the direct Cl algorithm The default value is 2 000 000 4 MDI the maximum dimension of the hamiltonian The default value is 1 000 001 In practice the code will attempt to use these default values and computes the overall memory requirement based on the settings above If this memory is not available each of the above values will be halved until the calculation can proceed within the memory allocated to the job
65. XSTATE is an integer used to specify the maximum number of eigenstates to be gen erated from the MRDCI treatment The directive may be omitted when MXSTATE will be set to the default value of 8 Example ITERATE MAXROOT 12 Thus the iterative procedure will from an initial point of generating NROOT roots of the secular problem attempt to generate roots NROOT 1 up to and including root MAXROOT 29 2 2 The SROOT Directive This directive may be used to establish the criterion whereby the number of roots to be used in selection is increased from that in effect NROOTS during the preceding MRDCI iteration The directive consists of two data fields read to the variables TEXT ROOTDEL using format A F e TEXT should be set to the character string SROOT e Selection will be extended from the NROOTS of the zero order problem to NROOTS 1 if the difference in the zero order eigenvalues ABS EIGV AL N ROOTS 1 EIGV AL NROOTS lt ROOTDEL 38 The directive may be omitted when ROOTDEL will be set to the default value of 0 0 i e no increase in NROOTS will be undertaken during the MRDCI iteration process Example SROOT 0 20 29 ITERATIVE MRDCI CALCULATIONS 116 29 2 3 The DROOT Directive This directive should be used to confirm that the number of eigenstates derived from the diagonalisation process is to be increased during the iterative MRDCI cycles until the number specified by the MAXROOT directive has been derived at
66. ZP C DZP H FC 11 0 FO 1 0 1 0 END 28 TABLE CI CALCULATIONS USING MCSCF ORBITALS 108 SCFTYPE MCSCF MCSCF ORBITAL COR1 COR1 COR1 DOC1 DOC3 DOC1 DOC2 DOC3 VOC2 VOC1 UDC3 UDCA END PRINT ORBITALS VIRTUALS NATORB CANONICAL 10 FOCK DENSITY FOCK ENTER Job 3 The Table CI Job Performing a semi direct Table Cl calculation using the natural orbitals generated in the previous step is fairly straightforward The following points should be noted e The MCSCF data presented in the preceding step must remain as part of the input data set with that computation now BYPASS ed e While specification of the input orbital set in the conventional Table Cl calculation requires use of the TRAN directive it is assumed in the present case that the orbital set required is that nominated on the CANONICAL directive i e no explicit section specification is required e With no frozen or discarded orbital the orbital indices specified on the CONF directive follow in obvious fashion from the list of IRREPs given above We are performing a simple 16 electron 3 reference calculation deriving just the first root and using a 2 micro hartree threshold RESTART TITLE H2C0 MCSCF 10E IN 9 M 0 DIRECT MRDCI FROM MCSCF NOS SEC 10 SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM N C 0 1 1 203 H 11 099 2 121 8 H 11 099 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 END RUNTYPE CI SCFTYPE MCSCF MCSCF
67. age differs from that described in the Conventional module An alternative form of the SINGLES directive is also possible comprising a single data line read to variables TEXT NREF using format A e TEXT should be set to the character string SINGLES e NREF is an integer parameter used to nominate a particular configuration within the set of reference functions The selection module will then retain in the final Cl all single excitations with respect to the nominated function regardless of their computed energy lowerings The SINGLES directive may be omitted when the program will include all single excitations with respect to ALL nominated reference functions i e SINGLES ALL Example Presenting the data line SINGLES 1 in a Table Cl calculation of a closed shell system where the SCF configuration is the first in the CONF list will lead to the inclusion of all single excitations with respect to the SCF function in the final Cl Such inclusion leads of course to a marked improvement in the quality of one electron properties computed from the Cl wavefunction 19 6 CONF The CONF directive is used to specify the reference CSFs for the Cl expansion The first line of the CONF directive is set to the character string CONF In contrast to CONF specification in the conventional module the last line of the directive the directive terminator now consists of the 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 61 text END in the first
68. al states typically the excited states It is assumed that the Cl eigen vectors have been generated and are available on the appropriate FORTRAN interfaces While in most cases the ground and excited state Cl vectors will reside on the same interface FTN036 the module will allow the use of differing data sets for these vectors a situation most likely to occur when the ground and excited states are of different symmetry 11 1 TM The TM directive is used to request Transition Moment analysis and comprises two data lines The first line is read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string TM e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass TM processing Such usage is typically associated with restarting Table Cl calculations The second data line of the TM directive is used to specify the location of the Cl vectors and the number of excited state vectors involved in the subsequent analysis The line is read to the variables IFTNX ISECX IFTNE ISECE NSTATE using format 51 12 12 C
69. ally equal to the number of transition moment calculations to be performed Thus if we wished to calculate the transition moment between the two lowest states of H20 then NSTATE would equal 1 and the TM data would appear as follows TM 36 1 36 2 1 Calculating the A states of HO To clarify our discussion of the Table Cl module we work through a typical example of using the Table Cl method in calculating the energetics and properties of the three low lying tA states of the H20 molecule The basis set employed is the TZVP triple zeta plus polarisation set this is augmented with a diffuse s and p orbital on the oxygen to provide a reasonable description of the known Rydberg character of the states of interest The computation is split into four separate jobs in which we 1 2 perform the initial SCF carry out an initial Cl where the reference set employed acts to provide at least a quali tative description of the states of interest based on the output from the initial Cl we augment the reference set to provide a quantitative description of the first three states 12 CALCULATING THE A STATES OF HO 34 4 finally having generated the Cl vectors for the three states we carry out in the final job an analysis of each vector in terms of natural orbitals and one electron properties and generate the transition moments between the ground and two excited states We now consider various aspects of each job in turn Job
70. arameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings 10 DATA FOR CONVENTIONAL TABLE CI ONE ELECTRON PROPERTIES 30 NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass PROP processing Such usage is typically associated with restarting Table Cl calculations 10 2 CIVEC The CIVEC directive is used to specify those eigenvectors of the Cl matrix to be analysed The directive consists of a single data line with the character string CIVEC in the first data field If the properties associated with NVEC eigenvectors of the secular problem are to be generated subsequent data fields should contain NVEC integers the integers specifying the numbering of the Cl eigenvectors on the FORTRAN interface FTNO36 If the CIVEC directive is omitted under PROP processing an analysis of the first Cl vector will be performed Example The data line CIVEC 1 3 may be used to analyse the first and third Cl eigenvector generated by the DIAG sub module 10 3 AOPR The AOPR directive may be used to request printing of the property integrals in the basis function AO representation If specified the directive consists of a single data line with the chara
71. atrix elements in the SELECT and Cl phases of the Table Cl procedure The space requirements of the Tablefile are now about 6 MBytes e n contrast to the other post Hartree Fock modules of GAMESS UK the Table Cl routines make extensive use of unformatted sequential FORTRAN data sets or interfaces The data set reference numbers and associated LFNs of these files have been given in Table 8 of Part 2 16 Directives Controlling Semi direct Table CI Calculations Data input characterising the Semi direct Table Cl calculation commences with the MRDCI data line and is typically followed by a sequence of directives terminated by presenting a valid Class 2 directive such as VECTORS or ENTER A fairly thorough overview of the data structure has been given in Part 2 we provide additional detail on the directives associated with each sub module below 16 1 MRDCI The Table Cl data initiator for the semi direct module consists of a single line containing the character string MRDCI in the first data field and the string DIRECT in the second It acts to transfer control to those routines responsible for inputing all data relevant to the MRDCI calculation Termination of this data is achieved by presenting a valid Class 2 directive that is not recognised by the Table Cl input routines for example VECTORS or ENTER 17 Data for Semi direct Table CI Integral Transformation In contrast to the Conventional Table Cl code integral transformation is now p
72. avidson procedure requested by DROOT Note that these criteria are executed such that 3 is not invoked before 2 is satisfied and 2 is not invoked before 1 is satisfied e This approach is designed to ensure approximate convergence in the structure of the lower eigenstates before attempting to extract solutions for higher states in the spectrum 29 2 4 The RETAIN Directive This is the most complex of the ITERATE directives in terms of appreciating the role it plays in ensuring that all the states of interest are obtained by the iterative procedure outlined above The key to obtaining these states is that at some point the associated leading configuration is identified as a reference species and is incorporated as a main configuration in the zero order 29 ITERATIVE MRDCI CALCULATIONS 117 space While the process outlined above has been found to be extremely effective when all eigenstates are of similar character e g Rydberg States this is not always true when the derived eigenstates are of quite different character e g valence and Rydberg states In such cases it is quite common for such a valence configuration not be included in the generated set of reference configurations at the outset and while low lying in energy to have an extremely low value of c 2 in for example each of the derived Rydberg states i e it will never be incorporated in the set of main configurations and hence never appear as one of the derived states A ke
73. b 3 The Final 12M 3R CI An examination of the output from the initial Cl reveals that the dominant configurations have as expected been included We show below the final Cl vectors for each of the states not 26 CALCULATING THE A STATES OF H30 96 surprisingly the ground state is more accurate by virtue of its SCF MOs having been em ployed Augmenting the reference set to improve the description of the two excited states follows straightforwardly from the statistics below Description of the X 4 state EXTRAPOLATED ENERGY 76 27252960 FER kk k CONFIGURATION WEIGHTS FEO E k k kkk k k a ROOT 1 Rx C C CONFIGURATION M 0 94942 1 2 17 23 0 00141 1 2 19 23 0 00112 2 7 17 19 1 23 0 00125 2 7 23 27 1 17 0 00147 17 19 23 27 1 2 SUM OF MAIN REFERENCE C C 0 949613377982042 Description of the 1 4 state EXTRAPOLATED ENERGY 75 90090942 FE k kk kk kk k CONFIGURATION WEIGHTS EEEE k k k kk kk k aK ROOT 2 Kx C C CONFIGURATION 0 00283 1 2 18 23 M 0 06414 2 3 1 17 23 0 00758 2 5 1 17 23 M 0 84001 17 18 1 2 23 0 00263 17 19 1 2 23 0 01150 18 19 1 2 23 0 00221 1 5 17 18 2 23 0 00136 1 8 17 18 2 23 0 00226 2 5 17 18 1 23 0 00444 2 6 17 18 1 23 0 00550 2 7 17 18 1 23 0 00667 17 18 23 25 1 2 0 00664 17 18 23 27 1 2 SUM OF MAIN REFERENCE C C 0 904355147296805 Description of the 2 4 state 26 CALCULATING THE A STATES OF H30 EXTRAPOLATED ENERGY FE RAK CONFIGURATION WEIGHTS FEO OA kkk k
74. bitals associated with NVEC eigenvectors of the secular problem are to be generated subsequent data fields should contain NVEC integers the integers specifying the numbering of the Cl eigenvectors on the FORTRAN interface FTNO36 as generated by the Cl sub module If the CIVEC directive is omitted under NATORB processing the natural orbitals of the first Cl vector will be generated Example CIVEC 1 3 The above data line may be used to generate natural orbitals from the first and third Cl eigenvector generated by the Cl sub module 21 3 Natural Orbital Data PUTQ The PUTQ directive may be used to route spin free natural orbitals to the Dumpfile and consists of a single dataline with the first two fields read to variables TEXT TYPE using format 2A e TEXT should be set to the character string PUTQ e TYPE should be set to one of the character strings AOS A O or SABF defining the basis representation required for the output NOs The character string AOS and A O will yield the NOs in the basis function representation suitable for subsequent input to the other analysis modules of GAMESS UK The string SABF will result in the NO expansion in the symmetry adapted basis representation and should be used when performing iterative natural orbital calculations see section 6 13 below The remaining data consists of a sequence of NVEC integers between 0 and 350 inclusive specifying the section number of the Dumpfile where the spin f
75. consists of a single dataline with the first two fields read to variables TEXT TYPE using format 2A e TEXT should be set to the character string PUTQ e TYPE should be set to one of the character strings AOS A O or SABF defining the basis representation required for the output NOs The character string AOS and A O will yield the NOs in the basis function representation suitable for subsequent input to the other analysis modules of GAMESS UK The string SABF will result in the NO expansion in the symmetry adapted basis representation and should be used when performing iterative natural orbital calculations see section 6 13 below The remaining data consists of a sequence of NVEC integers between 0 and 350 inclusive specifying the section number of the Dumpfile where the spin free NOs derived from the NVEC Cl vectors nominated by the CIVEC directive are to be placed Example PUTQ AOS 100 120 The spin free NOs in the basis set representation are output to sections 100 and 120 respectively of the Dumpfile A section setting of 0 on the PUTQ directive will act to suppress natural orbital output to the Dumpfile 10 Data for Conventional Table CI One electron Properties 10 1 PROP The PROP directive is used to request the computation of one electron properties and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string PROP e TEXTF is an optional p
76. cter string AOPR in the first data field Subsequent data fields are used to specify those integrals to be printed Valid character strings include S T X Y Z XX YY ZZ XY XZ and YZ requesting in obvious notation printing of the components of the overlap kinetic energy dipole and quadrupole moments respectively Example AOPR X Y Z would result in printing of integrals of the x y and z components of the dipole moment 10 4 MOPR The MOPR directive may be used to request printing of the property integrals in the molecular orbital MO basis If specified the directive consists of a single data line with the character 10 DATA FOR CONVENTIONAL TABLE CI ONE ELECTRON PROPERTIES 31 string MOPR in the first data field Subsequent data fields are used to specify those integrals to be printed Valid character strings include S T X Y Z XX YY ZZ XY XZ and YZ requesting in obvious notation printing of the components of the overlap kinetic energy dipole and quadrupole moments respectively Example MOPR XX YY ZZ would result in printing of integrals of the diagonal components of the quadrupole moment 10 5 Configuration Data Lines In addition to evaluating the properties of a given Cl vector the module will also look to eval uating the corresponding properties of a nominated single configuration typically the leading term in the Cl vector the idea here of course is to provide a guide to the effect of the Cl treatment
77. ctive specifies the SCF configuration and it is this configuration that should be nominated in the PROP data Thus the Cl data including the property analysis would appear as follows MRDCI DIRECT TABLE SELECT CNTRL 8 SPIN 1 SYMMETRY 1 SINGLES 1 CONF 0123 15 013 4 15 0123 16 END CI NATORB PROP CIVEC 1 0123 15 where one such data line is required given the specification of CIVEC 23 Data for Semi direct Table CT Transition Moments This Table Cl module will calculate both electrical and magnetic dipole moments as well as oscillator strengths and lifetimes of excited states The module will look to calculate the moment between a specific state typically the ground state and a set of additional states typically the excited states It is assumed that the Cl eigen vectors have been generated and are available on the appropriate FORTRAN interfaces While in most cases the ground and excited state Cl vectors will reside on the same interface FTN036 the module will allow the use of differing data sets for these vectors a situation most likely to occur when the ground and excited states are of different symmetry 23 DATA FOR SEMI DIRECT TABLE CI TRANSITION MOMENTS 83 23 1 TM The TM directive is used to request Transition Moment analysis and comprises two data lines The first line is read to the variables TEXT TEXTF and TEXTB using format 3A TEXT should be set to the character string TM TEXTF is an optional parameter
78. d e The OPEN directive is now present specified prior to the Table Cl data e In the absence of section specification on the ENTER directive the set of vectors used in the Table Cl transformation will be restored from the default open shell SCF section containing the energy ordered eigenvectors section 5 The SWAP directive is to generate initial starting MOs for the B state e The symmetry and spin of the Cl wavefunction will be deduced from the preceding SCF calculation i e a Cl wavefunction of By symmetry corresponding to SYMMETRY 2 and a doublet Cl wavefunction corresponding to SPIN 2 e The number of active electrons in the Cl will be set to be those involved in the SCF calculation i e CNTRL 15 e Singly excited configurations with respect to each of the default reference configurations SINGLES ALL will be included regardless of their computed energy lowerings e The set of reference configurations to be employed will be generated using the same algo rithm used in the open shell case above i e the SCF configuration plus those generated from this configuration by including i for each symmetry IRREP the doubly excited configuration arising from excitation of the highest occupied DOMO of that symmetry to the lowest virtual orbital VMO of the same symmetry and ii the lowest singly excited configuration again arising from the highest occupied DOMO to the lowest VMO of the same symmetry In the present example this
79. d be set to the string BYPASS if the user wishes to bypass Cl processing Such usage is typically associated with restarting Table Cl calculations 20 2 ACCURACY This directive may be used to define the diagonalisation thresholds for the two extrapolation passes consists of a single line read to variables TEXT THRESHE using format A F e TEXT should be set to the character string ACCURACY or DTHRESH e THRESHE On both extrapolation passes the diagonalization is converged to an energy threshold THRESHE 21 DATA FOR SEMI DIRECT TABLE CI NATURAL ORBITALS 78 The THRESH directive may be omitted when THRESHE will be set to 0 000001 Example Presenting the data line ACCURACY 0 0000001 will result in a diagonalisation threshold of 0 0000001 for the two extrapolation passes 20 3 PRINT The PRINT directive may be used to control the printing of Cl coefficients and weights through out the extrapolation passes and in the final analysis This directive consists of a single data line read to variables TEXT PTHR PTHRCC IFLAG using format A 2F I e TEXT should be set to the character string PRINT e Cl coefficients less than PTHR in absolute magnitude will not be printed during the extrapolation passes e Cl weights coefficients less than PTHRCC in absolute magnitude will not be printed in the final analysis of the Cl wavefunctions e IFLAG may be used to control the printing of the Cl wavefunctions in the event that the d
80. deally perform selection with respect to at least the corresponding NVEC roots of the root secular problem to ensure a consistent treatment of each of the required states The choice of the reference set will clearly prove crucial and should be such as to ensure a one to one correspondence between each of the final Cl vectors and a certain vector of the root problem Indeed the whole process of extrapolation to zero threshold is meaningless if this condition is not obeyed 19 8 THRESH This directive defines the threshold factors to be used in the process of configuration selection and consists of a single line read to variables TEXT TMIN TINC using format A 2F e TEXT should be set to the character string THRESH e TMIN should be set to the minimum threshold factor in units of micro hartree uH to be used in selection Any CSF with a computed energy lowering greater than TMIN will be retained in the final list of selected configurations e TINC should be set to the threshold increment to be used in the process of extrapolation This process now involves solution of the final secular problem at just two thresholds defined by TMIN and TMIN TINC Note the restricted number of calculations here compared to the conventional module The THRESH directive may be omitted when TMIN will be set to 10 0 and TINC to 10 0 leading to the solution of the T 10 and 20 uH secular problem Example 20 DATA FOR SEMI DIRECT TABLE CI EIGEN SOLUTION 77
81. e noted e The RETAIN directive may be omitted when ERETAIN will assume a default value of 0 10 au and be used in selecting configurations as reference functions from the outset e A maximum of 30 configurations may be retained on a given iteration e Setting IRETAIN to a high value e g 100 provides a potential mechanism for the iterative process to focus on just a sub set of the final Cl eigen states 29 ITERATIVE MRDCI CALCULATIONS 118 29 3 Examples of Excited State Generation To clarify our discussion of iterative processing using the Semi direct Table Cl module we work through a number of example of using the method in calculating the energetics and properties of the lying states of a variety of molecules of increasing complexity 29 3 1 Calculating the A states of Formaldehyde Data for performing an iterative MRDCI calculation on the ten lowest A states of formaldehyde is given below The calculation is initiated with an SCF calculation on the ground state The following points should be noted 1 The C and O 1s inner shell MOs together with the two highest virtual orbitals are excluded from the calculation under control of the CORE and ACTIVE directives 2 The SELECT directives requesting the 12 electron singlet Cl states of A symmetry are as follows SYMMETRY 1 SPIN 1 CNTRL 12 3 The ITERATE directive ITERATE MAXI 20 MAXROOT 10 SROOT 0 10 DROOT WEIGHT 0 005 requests derivation of the lowest 10 roots allowin
82. e one of the character strings SINGLET DOUBLET TRIPLET QUARTET and QUINTET to specify NSPIN Example SPIN 3 SPIN TRIPLET are equivalent the wavefunction will be three fold spin degenerate 194 SYMMETRY This directive consists of one line read to variables TEXT NSYM using format A l e TEXT should be set to the character string SYMMETRY e NSYM is an integer parameter used to specify the spatial symmetry of the Cl wavefunction and is set to the appropriate sequence number of the required irreducible representation see Table 1 Example In a system of Ca symmetry the data line SYMMETRY 3 would be required when performing calculations on states of B symmetry Failure to present the directive in such cases will lead to the default Ay symmetry 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 60 19 5 SINGLES This directive consists of one line read to variables TEXT TEXTS using format 2A e TEXT should be set to the character string SINGLES e TEXTS may be set to one of the following character strings ALL The selection module will retain in the final Cl all single excitations with respect to ALL nominated reference functions specified under the CONF directive regardless of their computed energy lowerings OFF The selection module will use the computed energy lowerings as the sole criteria for including configurations in the final Cl with no automatic inclusion of single excitations Note that this us
83. e ordered sequence numbers of the ground state orbitals allowing for the effective removal of the la and 1bj orbitals are 1 14 10 2 23 and 8 respectively To perform a two reference 12 electron valence Cl calculation based on the SCF configuration 24 203 103 347103 103 36 and the doubly excited configuration 2a52b3 103 3a5103 103 37 would require the following CONF data CONF 0128 10 14 23 012 10 14 21 23 END The following data will perform this Cl where e The SCF computation is BYPASS ed e The freezing and discarding of the la and 1b MOs is accomplished using the CORE and ACTIVE directives e The default sub module specifications are in effect with no specific need to reference TABLE or Cl activity i e the table ci data set is to be constructed in the job e The natural orbitals are routed to section 200 of the Dumpfile Note that subsequent usage of the NOs by the Table Cl module requires the SABF specification on the PUTQ directive RESTART NEW TITLE ETHYLENE CI GROUND STATE 2M SCF MOS BYPASS SCF 27 ITERATIVE NATURAL ORBITAL CALCULATIONS 103 ZMATRIX ANGSTROM pa o 20 0 0 0 3 180 0 0 0 3 0 0 120 0 3 180 0 Ropa 2 2 mommaa NNRPRP RO BPRPrPP PR 4 12 12 11 11 END BASIS DZ CORE 1 2 END ACTIVE 3 TO 28 END RUNTYPE CI MRDCI DIRECT SELECT SYMMETRY 1 CNTRL 12 SPIN 1 SINGLES 1 CONF 0128 10 14 23 012 10 14 21 23 END THRESH 30 10 NATORB IPRINT PUTQ
84. e sequence of data lines defining the Semi direct Table Cl calculation is again terminated by the ENTER directive Note at this stage that the full data specification corresponding to the defaults generated from the above data file is as follows TITLE H2C0 2B2 TZVP EXPLICIT DATA FOR DEFAULTS MULT 2 CHARGE 1 SUPER OFF NOSYM ZMAT ANGSTROM C 24 SEMEDIRECT TABLE CI USING DEFAULT OPTIONS 88 0 1 1 203 H 1 1 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS TZVP RUNTYPE CI OPEN 1 1 ACTIVE 1 TO 52 END MRDCI DIRECT TABLE SELECT CNTRL 15 SPIN 2 SYMM 3 SINGLES ALL CONF 38 38 5 38 28 29 38 38 1 2 37 38 39 28 37 28 37 38 Orre 29 37 23 39 4 5 28 wewewrher N ewe WR WwW Ww NAN AD BA wnnawan END THRESH 10 10 ROOTS 1 CI NATORB CIVEC 1 VECTORS ATOMS ENTER 4 5 Let us now consider a Semi direct Table Cl calculation on the 7B state of H2CO again using default options available within the module A valid data sequence for performing such a calculation is shown below where we are still performing all the computation in a single job TITLE H2C0 2B1 TZVP DEFAULT MRDCI SETTINGS MULT 2 CHARGE 1 ZMAT ANGSTROM C 0 1 1 203 H 1 1 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS TZVP RUNTYPE CI MRDCI DIRECT SWAP 78 END ENTER 24 SEMI DIRECT TABLE CI USING DEFAULT OPTIONS 89 Considering the changes to the closed shell run the following points should be note
85. e using a 3 21G basis While the molecular symmetry is Cooy the symmetry adaptation and subsequent Cl will be conducted in the Ca point group The resolution of the Coov into the Ca orbital species is given in Table 2 An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 326 84723972 2 0000000 1 104 02836336 2 0000000 3 1 40 71695637 2 0000000 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 15 4 1 35 46377378 2 0000000 5 3 35 45608069 2 0000000 6 2 35 45608068 2 0000000 7 1 10 42193940 2 0000000 8 1 7 88512031 2 0000000 9 2 7 88222844 2 0000000 10 3 7 88222844 2 0000000 11 1 5 07729175 2 0000000 12 1 3 38247056 2 0000000 13 3 3 35978308 2 0000000 14 2 3 35978307 2 0000000 15 1 1 01099628 2 0000000 16 3 0 53702948 2 0000000 17 2 0 53702947 2 0000000 18 4 0 49640067 2 0000000 19 1 0 49640067 2 0000000 20 1 0 44715317 2 0000000 21 3 0 39988537 2 0000000 22 2 0 39988537 2 0000000 23 1 0 35127248 2 0000000 24 1 0 00023285 0 0000000 25 3 0 06300102 0 0000000 26 2 0 06300102 0 0000000 27 1 0 12855448 0 0000000 28 1 0 19287013 0 0000000 29 3 0 25729975 0 0000000 30 2 0 25729975 0 0000000 31 1 0 39720201 0 0000000 32 1 0 86197727 0 0000000 33 2 0 88942618 0 0000000 34 3 0 88942618 0 0000000 35 1 1 01877167 0 0000000 36 1 2 16694989 0 0000000 37 3 3 96181512 0 0000000 38 2 3 96181512 0 0000000 39 4 3 98212497 0 0000000 40 1
86. ect with no specific need to reference ADAPT Cl or DIAG activity we also assume that the TABLE data set is available to the job e the ROOT directive is specifying selection with respect to the first 3 roots of the zero order problem which we assume will correspond to the states of interest RESTART NEW TITLE RK H20 TZVP DIFFUSE S P TABLE CI 3M 3R SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM 0 H 1 0 951 H 1 0 951 2 104 5 END BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI MRDCI TRAN CORE DISC 1000 1 1000 18 SELECT CONF 0 12 17 23 12 CALCULATING THE A STATES OF HO 2 2 3 1 17 23 21718 12 23 THRE 10 10 ROOT 3 ENTER Job 3 The Final 12M 3R CI 37 An examination of the output from the initial Cl reveals that the dominant configurations have as expected been included We show below the final Cl vectors for each of the states not surprisingly the ground state is more accurate by virtue of its SCF MOs having been em ployed Augmenting the reference set to improve the description of the two excited states follows straightforwardly from the statistics below Description of the X 4 state EXTRAPOLATED ENERGY CSF NO 1 1 M O 118 118 M O 129 129 M O SUM OF MAIN REFERENCE C C C C CONFIGURATION 94952507 1 2 17 00002448 2 3 1 00016368 17 18 1 0 94971323 Description of the 1 4 state EXTRAPOLATED ENERGY CSF NO 1 1 M 0 71 71 0 1
87. erformed under control of the conventional transformation code of GAMESS UK Control of the transformation is now carried out under the ACTIVE and CORE directives described previously in Part 5 of the manual The following points should be noted 1 The MOs to be used in the transformation process will be taken from the section as nominated on the ENTER directive or the default section in effect if explicit section specification is omitted 17 DATA FOR SEMEDIRECT TABLE CI INTEGRAL TRANSFORMATION 54 2 The transformation may be bypassed under control of the BYPASS directive Such usage is typically associated with restarting Table Cl calculations 3 The freezing and discarding of orbitals in the transformation is controlled by a combination of the CORE and ACTIVE directives 4 Users of the old module will be familiar with the ordering of the MOs required within the Table Cl module namely in terms of irreducible representation IRrep and numbering within each IRrep While this ordering is still the preferred means of driving the Cl module and associated specification of the reference functions etc it is no longer necessary to adhere to this numbering scheme when specifying CORE and ACTIVE orbitals for the code will automatically generate the appropriate numbering of these MOS This is best illustrated by considering the same examples used at the beginning of the chapter 5 Note again that the input orbital set will be reordered by the tra
88. erties associated with NVEC eigenvectors of the secular problem are to be generated subsequent data fields should contain NVEC integers the integers specifying the numbering of the Cl eigenvectors on the FORTRAN interface FTNO36 If the CIVEC directive is omitted under PROP processing an analysis of the first Cl vector will be performed Example The data line CIVEC 1 3 may be used to analyse the first and third Cl eigenvector generated by the Cl sub module 22 DATA FOR SEMI DIRECT TABLE CI ONE ELECTRON PROPERTIES 81 22 3 AOPR The AOPR directive may be used to request printing of the property integrals in the basis function AO representation If specified the directive consists of a single data line with the character string AOPR in the first data field Subsequent data fields are used to specify those integrals to be printed Valid character strings include S T X Y Z XX YY ZZ XY XZ and YZ requesting in obvious notation printing of the components of the overlap kinetic energy dipole and quadrupole moments respectively Example AOPR X Y Z would result in printing of integrals of the x y and z components of the dipole moment 22 4 MOPR The MOPR directive may be used to request printing of the property integrals in the molecular orbital MO basis If specified the directive consists of a single data line with the character string MOPR in the first data field Subsequent data fields are used to specify those inte
89. es TEXT NROOT IROOT I I 1 NROOT using format A NROOT 1 I e TEXT and NROOT are defined as above e NROOT integers are read to the array IROOT defining the vectors of the zero order matrix to be used in selection We now provide some further notes on the directive e the ROOTS directive may be omitted when the energy contributions are calculated with reference to the lowest eigenstate of the root problem only Omission of the directive is thus equivalent to presenting the data line ROOTS 1 e The number of root eigenstates to be specified will depend on the number of states required in the final Cl Thus if NVEC roots of the final Cl matrix are to be subsequently generated in DIAG processing the user should ideally perform selection with respect to at least the corresponding NVEC roots of the root secular problem to ensure a consistent treatment of each of the required states The choice of the reference set will clearly prove crucial and should be such as to ensure a one to one correspondence between each of the final Cl vectors and a certain vector of the root problem Indeed the whole process of extrapolation to zero threshold is meaningless if this condition is not obeyed 6 8 THRESH This directive defines the threshold factors to be used in the process of configuration selection and consists of a single line read to variables TEXT TMIN TINC using format A 2F e TEXT should be set to the character string THRESH e
90. f reference configurations and appropriate thresholds e Cl generation of the Cl Hamiltonian based on the set of selected configurations from SELECT and integrals from TRAN e DIAG calculation of one or more Cl eigenfunctions of the Hamiltonian generated under Cl The remaining modules are optional and may be used to analyse one or more of the Cl eigen vectors e NATORB to generate the spin free natural orbitals for one or more of the calculated Cl eigenvectors e PROP to compute various 1 electron properties of the Cl wavefunctions Note that the natural orbitals generated above may be routed to the Dumpfile and examined by the other analysis modules of GAMESS UK in a subsequent job 2 DIRECTIVES CONTROLLING CONVENTIONAL TABLE CI CALCULATIONS 2 e TM to compute the transition moments between nominated Cl eigenvectors In addition to the Mainfile Dumpfile and Scratchfile the following data sets will be used by the program e The Tablefile A dataset normally assigned using the local file name LFN TABLE will be used as a source of pattern symbolic matrix elements in the SELECT and Cl phases of the Table Cl procedure The space requirements of the Tablefile are about 2 MBytes e The Sortfile A dataset normally assigned using the LFN SORT will be used as a scratchfile in the generation of symmetry adapted and transformed integrals The maximum space requirements of the Sortfile are about twice that of the Mainfile although thi
91. f the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS BR oN on e UNBE R and the following orbital assignments from the converged closed shell SCF 4 DATA FOR CONVENTIONAL TABLE CI TRANSFORMATION T 1 1 15 66716423 2 0000000 2 5 15 66241865 2 0000000 3 1 1 51005217 2 0000000 4 5 0 76128176 2 0000000 5 1 0 63704931 2 0000000 6 3 0 63448705 2 0000000 7 2 0 63448705 2 0000000 8 6 0 17408343 0 0000000 9 7 0 17408343 0 0000000 10 5 0 30302673 0 0000000 11 3 0 38747796 0 0000000 12 2 0 38747796 0 0000000 13 1 0 42599317 0 0000000 14 1 0 49515007 0 0000000 15 7 0 57046706 0 0000000 16 6 0 57046706 0 0000000 17 5 0 92638361 0 0000000 18 5 1 12927523 0 0000000 19 1 1 83974927 0 0000000 20 3 2 01160646 0 0000000 21 2 2 01160646 0 0000000 22 5 2 01683641 0 0000000 23 4 2 08974396 0 0000000 24 1 2 08974396 0 0000000 25 7 2 16718057 0 0000000 26 6 2 16718057 0 0000000 27 1 2 16945852 0 0000000 28 3 2 20023738 0 0000000 29 2 2 20023738 0 0000000 30 8 2 67650248 0 0000000 31 5 2 67650248 0 0000000 32 5 2 98166352 0 0000000 33 1 3 57616884 0 0000000 34 7 3 58056531 0 0000000 35 6 3 58056531 0 0000000 36 5 4 37707106 0 0000000 37 1 6 02043977 0 0000000 38 5 6 12816913 0 0000000 39 1 35 89673292 0 0000000 40 5 35 91926868 0 0000000 Thus the inner shell Nis orbitals the log and lo transform as ay and bj respectively in Do symmetry with IRrep numbers 1 and 5 Both correspond
92. g a maximum of 20 MRDCI iterations to derive these roots The initial Cl will be based on the lowest 5 roots of the Cl Hamiltonian ROOTS 5 derived from the default set of generated main configurations CORE 20000000 TITLE H2C0 TZVP D SPD MRDCI TREATMENT OF THE 1A1 STATES HARMONIC ZMAT ANGSTROM C 0 1 1 203 H 1 1 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS TZVP 0 TZVP C TZVP H So 1 0 0 02 PO 1 0 0 02 29 ITERATIVE MRDCI CALCULATIONS 29 3 2 29 3 3 DO 1 0 0 02 END CORE 1 2 END ACTIVE 3 TO 57 END RUNTYPE CI MRDCI DIRECT SYMMETRY 1 SPIN 1 CNTRL 12 THRESH 5 5 ROOTS 5 ITERATE MAXI 20 SROOT 0 10 MAXROOT 10 DROOT WEIGHT 0 005 NATORB BYPASS ENTER Calculating the Excited states of the Formyl Radical The 2A States 119 Data for performing an iterative MRDCI calculation on the eight lowest 2A states of the formyl radical is given below The calculation is initiated with an SCF calculation on the 2A ground state under control of the OPEN directive CORE 20000000 TIME 300 TITLE HCO DZP BOND SP HARMONIC MULT 2 ZMAT G BQ 1 RCO2 X 21 0 1 90 0 O 2 RCO2 3 90 0 1 180 0 X 1 1 0 2 90 0 3 0 0 H 1 RCH 5 40 0 4 180 0 VARIABLES RCO2 1 125 RCH 2 076 END BASIS DZP H S BQ 1 0 0 02 P BQ 1 0 0 02 D BQ 1 0 0 02 DZP C DZP 0 0 0 29 ITERATIVE MRDCI CALCULATIONS 120 END OPEN 1 1 CORE 1 2 END ACTIVE 3 TO 42 END RUNTYPE CI MRDCI DIRECT SY
93. grals to be printed Valid character strings include S T X Y Z XX YY ZZ XY XZ and YZ requesting in obvious notation printing of the components of the overlap kinetic energy dipole and quadrupole moments respectively Example MOPR XX YY ZZ would result in printing of integrals of the diagonal components of the quadrupole moment 22 5 Configuration Data Lines In addition to evaluating the properties of a given Cl vector the module will also look to eval uating the corresponding properties of a nominated single configuration typically the leading term in the Cl vector the idea here of course is to provide a guide to the effect of the Cl treatment on the property with the nominated CSF being typically the corresponding SCF configuration Thus the final data for the properties module comprises a sequence of NVEC data lines each line a sequence of integers defining the single configuration for the Cl vector under consideration The format of these lines is identical to that of the CONF data used in nominating the reference functions and in most instances will be a repeat of that data Example Consider the valence Cl calculation on PH3 described in example 1 of the CONF directive Considering just the Cl data 23 DATA FOR SEMI DIRECT TABLE CI TRANSITION MOMENTS 82 MRDCI DIRECT TABLE SELECT CNTRL 8 SPIN 1 SYMMETRY 1 SINGLES 1 CONF 0123 15 013 4 15 0123 16 END CI NATORB then the first data line of the CONF dire
94. he 7By state of H2CO again using default options available within the module A valid data sequence for performing such a calculation is shown below where we are still performing all the computation in a single job TITLE H2C0 2B2 TZVP DEFAULT MRDCI SETTINGS 113 06446075 MULT 2 CHARGE 1 ZMAT ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS TZVP RUNTYPE CI MRDCI DIRECT ENTER Considering the changes to the closed shell run the following points should be noted e The set of vectors used in the Table Cl transformation will be restored from the default eigenvector section of the Dumpfile given that no section is specified on the ENTER 24 SEMEDIRECT TABLE CI USING DEFAULT OPTIONS 87 directive This will be section 5 the section used to store the energy ordered canonicalised orbitals from the open shell SCF calculation e The symmetry and spin of the Cl wavefunction will be deduced from the preceding SCF calculation i e a Cl wavefunction of Bz symmetry corresponding to SYMMETRY 3 and a doublet Cl wavefunction corresponding to SPIN 2 e The number of active electrons in the Cl will be set to be those involved in the SCF calculation i e CNTRL 15 e Singly excited configurations with respect to each of the default reference configurations SINGLES ALL will be included regardless of their computed energy lowerings e The set of reference configurations to be employed will follow t
95. he same algorithm used in the closed shell case above i e the SCF configuration plus those generated from this configuration by including i for each symmetry IRREP the doubly excited configuration arising from excitation of the highest occupied DOMO of that symmetry to the lowest virtual orbital VMO of the same symmetry and ii the lowest singly excited configu ration again arising from the highest occupied DOMO to the lowest VMO of the same symmetry In the present example this will correspond to the SCF configuration the dou ble and single excitation arising from the DOMO 5a to VMO 6a1 the double and single excitation arising from the DOMO 1b to VMO 2b and the double and single excitation arising from the DOMO 1ba to VMO 3b3 Note that the DOMO involved in the latter configurations is now the lbg given that the 2b is now singly occupied and again the absence of excitations involving az MOs given the absence of such orbitals involved in the occupied manifold This again results in a total reference set of 7 functions as shown thus in the job output numbers of open shells and corresponding main configurations 1 38 1 2 3 4 5 28 37 si SCF configuration 1 33 1 2 3 4 6 28 37 he 5al gt 6al double 3 5 6 38 1 2 3 4 28 37 Esa 5al gt 6al single 1 388 1 2 3 4 5 29 37 da 1b1 gt 2b1 double 3 28 29 38 1 2 3 4 5 37 a 1b1 gt 2b1 single 1 388 1 2 3 4 5 28 39 ie 1b2 gt 3b2 double 3 37 38 39 1 2 3 4 5 28 ae 1b2 gt 3b2 single Th
96. he strings NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output This option should be set to generate a print of the natural orbital coefficient array s FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass NATORB processing Such usage is typically associated with restarting Table Cl calculations 9 2 Natural Orbital Data CIVEC The CIVEC directive is used to specify those eigenvectors of the Cl matrix to be analysed The directive consists of a single data line with the character string CIVEC in the first data field If natural orbitals associated with NVEC eigenvectors of the secular problem are to be generated subsequent data fields should contain NVEC integers the integers specifying the numbering of the Cl eigenvectors on the FORTRAN interface FTNO36 as generated by the DIAG sub module If the CIVEC directive is omitted under NATORB processing the natural orbitals of the first Cl vector will be generated 10 DATA FOR CONVENTIONAL TABLE CI ONE ELECTRON PROPERTIES 29 Example CIVEC 1 3 The above data line may be used to generate natural orbitals from the first and third Cl eigenvector generated by the DIAG sub module 9 3 Natural Orbital Data PUTQ The PUTQ directive may be used to route spin free natural orbitals to the Dumpfile and
97. ia e A e RA A a 195 INGLES rr ca 190 CONF ceci AAA 197 RUOTS ar aia ras ERS EEE 19 8 THRESH ok oad da ARA me a de i 20 Data for Semi direct Table Cl Eigen Solution PORE whee ty a Pedersen A 20 2 ACCURACY sois EASA pra Cee De EHLERS OU A hh ae ed ee eee SP ee Ge Gre Stee eee es 21 Data for Semi direct Table Cl Natural Orbitals 21 1 Natural Orbital Data NATORB 2s 2 66 2 2 eee ws iii 32 32 33 41 46 51 52 53 53 53 57 58 58 58 59 59 59 60 60 75 76 17 77 77 78 78 CONTENTS lv 21 2 Natural Orbital Datai CWEC oa Sce t eea dl ee eB ee 79 21 3 Natural Orbital Data PUTO gt 2ac 86644454 20 ta aa 79 22 Data for Semi direct Table Cl One electron Properties 80 22 PROP cocos daa oa he bee e kei ebb eee A 80 Do NEC e e aide Ee Eh oe A a a 80 PUSE ssa ma da eB sy See OA wet a ta Et ct 81 2224 MOPR a aie Geek te bee Hed ke a oe ew arb aos eh a cee e we ok ook 81 22 5 Configuration Data Lines lt r sa sw raia ratoa ios ee we ae ena 81 23 Data for Semi direct Table Cl Transition Moments 82 A O E 83 24 Semi direct Table Cl Using Default Options 84 24 1 Calculations on the Formaldehyde Ground State 84 24 2 Calculations on the Formaldehyde Cations 0 86 25 Memory Specification for the Semi direct Table Cl Module 90 26 Calculating the A states of H20 92 27 Iterative Natural Orbital Calculations 100 28 Table Cl Calculations Using MCSCF Orbi
98. iagonalisation does not converge Setting IFLAG 1 will cause a detailed print of the Cl vectors corresponding to each root This directive may be omitted when the defaults PTHR 0 05 and PTHRCC 0 002 will be taken 21 Data for Semi direct Table CT Natural Orbitals 21 1 Natural Orbital Data NATORB The NATORB directive is used to request Natural Orbital NO generation and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string NATORB e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module 21 DATA FOR SEMI DIRECT TABLE CI NATURAL ORBITALS 79 IPRINT to produce an intermediate level of output This option should be set to generate a print of the natural orbital coefficient array s FPRINT or DEBUG to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass NATORB processing Such usage is typically associated with restarting Table Cl calculations 21 2 Natural Orbital Data CIVEC The CIVEC directive is used to specify those eigenvectors of the Cl matrix to be analysed The directive consists of a single data line with the character string CIVEC in the first data field If natural or
99. ify the number of root eigenstates NROOT and the sequence numbers of these vectors within the matrix of zero order eigenvectors IROOT I I 1 NROOT Two formats may be used in this specification 1 If the lowest NROOT vectors are to be used then the data line is read to the variables TEXT NROOT using format A e TEXT is set to the character string ROOTS 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 76 e NROOT is an integer specifying the number of roots to be used where the sequence numbers of the roots will be 1 NROOT 2 If the NROOT vectors to be used are not the lowest in the root eigenvector matrix then the sequence numbers within this matrix must be specified The data line is then read to the variables TEXT NROOT IROOT I I 1 NROOT using format A NROOT 1 I e TEXT and NROOT are defined as above e NROOT integers are read to the array IROOT defining the vectors of the zero order matrix to be used in selection We now provide some further notes on the directive e the ROOTS directive may be omitted when the energy contributions are calculated with reference to the lowest eigenstate of the root problem only Omission of the directive is thus equivalent to presenting the data line ROOTS 1 e The number of root eigenstates to be specified will depend on the number of states required in the final Cl Thus if NVEC roots of the final Cl matrix are to be subsequently generated in DIAG processing the user should i
100. ith the SUPER OFF NOSYM specification in effect The list of generated two electron integrals is routed to the FORTRAN interface FTNO21 the one electron integrals to the Dumpfile 4 DATA FOR CONVENTIONAL TABLE CI TRANSFORMATION 3 3 1 ADAPT The ADAPT directive is used to control the symmetry adapted integral generator and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string ADAPT e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress output from the module IPRINT to produce an intermediate level of output charactering for example the symmetry adapted functions FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass generation of the symmetry adapted integrals Such usage is typically associated with restarting Table Cl calculations Example ADAPT BYPASS is a valid data line to bypass the module in a Table Cl restart job 4 Data for Conventional Table CI Transformation 4 1 TRAN The TRAN directive is used to control the integral transformation module and comprises one or more data lines The first data line is read to the variables TEXT ISECV TEXTF TEXTC TEXTD and TEXTB using format A I 4A e TEXT
101. ithin that IRrep and not within the total orbital manifold see below 2 If the DISCARD keyword has been presented two additional data lines are now required to define the number Line 1 and the sequence numbers Line 2 of the orbitals to be discarded Line 1 is read in l format to the variables NODISC 1 I 1 NIRREP where NODISC I specifies the number of orbitals of irreducible representation IRrep that are to be discarded Line 2 is also read in l format and specifies the sequence numbers of the discarded orbitals in the spirit of the re ordered sequence above Thus the first NODISC 1 integers specify 4 DATA FOR CONVENTIONAL TABLE CI TRANSFORMATION 5 the discarded orbitals of IRrep the next NODISC 2 integers the discarded orbitals of IRrep2 and so on until all Rreps with more than zero orbitals have been specified Note that the integer specification within each Rrep again refers to the relative ordering within that IRrep and not within the total orbital manifold see below Example 1 In this example we wish to perform a valence Cl calculation on the H2CO molecule using a DZ basis of 24 gtos looking to freeze both the oxygen and carbon Is orbitals and to discard the inner shell complement orbitals An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 14 2 4 3 6 1 1 20 58952765 2 0000000 2 1 11 35779935 2 0000000 3 1 1 43525479 2 0
102. k ARK ROOT 3 RK x C C CONFIGURATION M 0 78360 2 3 1 17 0 00618 2 4 1 17 0 05594 2 5 1 17 M 0 07077 17 18 1 2 0 00236 3 5 1 17 0 00344 3 6 1 17 0 00494 3 7 1 17 0 00132 1 2 3 5 0 00191 2 3 17 18 0 00872 2 3 17 19 0 00639 2 3 23 25 0 00600 2 3 23 27 SUM OF MAIN REFERENCE C C 0 854543160127212 75 88256073 23 23 23 23 23 23 23 17 PRP Re 23 23 23 17 17 97 Taking as the criterion for inclusion a weight of 0 005 the final 12 reference set Cl is shown below We have assumed that the FORTRAN interface FTNO31 plus the TABLE data base table ci has been saved from the second job enabling us to bypass the transformation and data base generation RESTART CI TITLE H20 TZVP DIFFUSE S P TABLE CI 12M 3R SUPER OFF NOSYM BYPASS SCF TRAN ZMAT ANGSTROM 0 H 1 0 951 H 1 0 951 2 104 5 END BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI ACTIVE 2 TO 35 END CORE 1 END 26 CALCULATING THE A STATES OF H30 98 MRDCI DIRECT TABLE BYPASS SELECT CNTRL 8 CONF 0 12 17 23 2 2 3 1 17 23 2 2 4 1 17 23 2o25 1 17 23 2 17 18 12 23 2 18 19 12 23 4 17 18 23 25 1 2 4 17 18 23 27 1 2 4 2 71718 1 23 4 2 31719 1 23 4 2 32325 1 17 4 2 32327 1 17 END THRE 10 10 ROOT 3 ENTER Job 4 The Analysis Assuming that the diagonalisation interface FTN036 had been saved above then the final analysis job is straightforward again bypassing of the various sub modules invol
103. llowing table where we assume that we wish to freeze the nine Ca inner shell orbitals To perform an 4 electron valence Cl calculation based on the SCF configuration 29 2 49 30 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos Og 1 11 3 8 1 8 Tus 2 5 2 3 9 11 Tuy 3 5 2 3 12 14 gry 4 1 0 1 15 Ou 5 7 2 5 16 20 Tgau 6 2 0 2 21 22 Toy T 2 0 2 23 24 and associated to o excitations 29 2 504304 2472 40 40 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 75 would require the following CONF data CONF O 1 16 0 2 16 0117 END The complete data file for performing the SCF and subsequent Cl in which all singles are retained would then be as follows TITLE CAH2 6 31G 3M 1R SUPER OFF NOSYM ZMAT ANGS CA X 1 1 0 H 1 CAH 2 90 0 H 1 CAH 2 90 0 3 THETA VARIABLES CAH 2 148 THETA 180 0 END BASIS 6 31G RUNTYPE CI CORE 1 TO 9 END MRDCI DIRECT TABLE SELECT CNTRL 4 SPIN 1 SYMMETRY 1 SINGLES ALL CONF O 1 16 0 2 16 0117 END THRESH 2 0 2 0 CI NATORB IPRIN ENTER 19 7 ROOTS The ROOTS directive is used to specify those eigenvectors of the root secular problem to be used in the process of selection with the energy contributions of the configurations computed with respect to the nominated vectors The directive consists of a single data line with the character string ROOTS in the first data field Subsequent data comprises integer variables used to spec
104. ls generated in the previous step is fairly straightforward The following points should be noted e The MCSCF data presented in the preceding step must remain as part of the input data set with that computation now BYPASS ed e Specification of the input orbital set must be driven through section specification on the TRAN directive without such specification the default MCSCF sections will be used which are not appropriate as input to the subsequent Cl In the data below this is achieved through the data line TRAN 10 where the section specified is just that nominated on the CANONICAL directive e With no frozen or discarded orbital the orbital indices specified on the CONF directive follow in obvious fashion from the list of IRREPs given above We are performing a simple 16 electron 3 reference calculation deriving just the first root and using a 2 micro hartree threshold RESTART TITLE H2CO MCSCF 10E IN 9 M O MRDCI FROM MCSCF NOS SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS DZP 0 DZP C DZP H FC 11 0 FO 1 0 1 0 END RUNTYPE CI SCFTYPE MCSCF MCSCF ORBITAL 14 TABLE CI CALCULATIONS USING MCSCF ORBITALS 50 COR1 COR1 COR1 DOC1 DOC3 DOC1 DOC2 DOC3 VOC2 VOC1 UDC3 UDCA END PRINT ORBITALS VIRTUALS NATORB CANONICAL 10 FOCK DENSITY FOCK MRDCI ADAPT TRAN 10 SELECT SYMMETRY 1 SPIN 1 CNTRL 16 SINGLES 1 CONF 012345 29 42 4
105. may be omitted when NTRIAL will be set to the maximum allowed value of 80 8 6 PRINT The PRINT directive may be used to control the printing of Cl coefficients and weights through out the extrapolation passes and in the final analysis This directive consists of a single data line read to variables TEXT PTHR PTHRCC IFLAG using format A 2F I 9 DATA FOR CONVENTIONAL TABLE CI NATURAL ORBITALS 28 e TEXT should be set to the character string PRINT e Cl coefficients less than PTHR in absolute magnitude will not be printed during the extrapolation passes e Cl weights coefficients less than PTHRCC in absolute magnitude will not be printed in the final analysis of the Cl wavefunctions e IFLAG may be used to control the printing of the Cl wavefunctions in the event that the diagonalisation does not converge Setting IFLAG 1 will cause a detailed print of the Cl vectors corresponding to each root This directive may be omitted when the defaults PTHR 0 05 and PTHRCC 0 002 will be taken 9 Data for Conventional Table CI Natural Orbitals 9 1 Natural Orbital Data NATORB The NATORB directive is used to request Natural Orbital NO generation and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string NATORB e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include t
106. n initial Cl where the reference set employed comprises just the SCF configu ration using the SCF MOs of interest 3 based on the output from the initial Cl we augment the reference set to include the leading secondary configuration generating the resulting natural orbitals 4 carry out the 2 reference Cl based on the natural orbitals generated in the previous step We now consider various aspects of each job in turn Job 1 The SCF TITLE ETHYLENE DZ GROUND STATE SCF SUPER OFF NOSYM ZMATRIX ANGSTROM C C 11 4 H 1 1 1 2 120 0 H 1 1 1 2 120 0 3 180 0 H 2 1 1 1 120 0 3 0 0 H 2 1 1 1 120 0 3 180 0 END BASIS DZ ENTER The only points to note here is the use of the SUPER directive in suppressing skeletonisation and use of the default eigenvector section section 1 for storage of the closed shell eigenvectors Job 2 The Initial 1M 1R CI An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 13 ITERATIVE NATURAL ORBITAL CALCULATIONS and the following orbital assignments characterising the closed shell SCF configuration 1a51b3 247205 103 305103 103 1 1 2 5 3 1 4 5 5 3 6 1 7 7 8 2 9 6 10 5 11 1 12 3 13 5 14 1 15 3 16 7 17 2 18 6 19 7 20 1 21 5 22 5 23 1 24 3 25 5 26 7 27 1 28 5 25533463 25413119 02052567 79195744 64152856 57038928 51438205 36388693 13190687 25441028 25558269 33918097
107. n of the Doon Species into the Do Species Orbital IRrep Doon Do Sequence No Og ag 1 6bg 22 y2 Tua b3u 2 Tu y boy 3 g zy big 4 Ou biu 5 Ou 22 y2 Tg z bag 6 Toy b3g 7 u zy au 8 The SINGLES directive may be omitted when the program will use the energy lowerings as the sole criteria for including configurations in the final Cl Example Presenting the data line SINGLES 1 in a Table Cl calculation of a closed shell system where the SCF configuration is the first in the CONF list will lead to the inclusion of all single excitations with respect to the SCF function in the final Cl Such inclusion leads of course to a marked improvement in the quality of one electron properties computed from the Cl wavefunction 6 6 CONF The CONF directive is used to specify the reference CSFs for the Cl expansion The first line of the CONF directive is set to the character string CONF Each subsequent line defines a 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 12 reference CSF by specifying the sequence numbers of the component active orbitals in l format A given reference CSF is defined by 1 the number of open shell orbitals NOPEN NOPEN includes any unpaired orbitals to gether with those non identical spin coupled pairs open to substitution 2 NOPEN integers specifying the sequence numbers of these orbitals 3 the NELEC NOPEN 2 sequence numbers of the doubly occupied orbitals i e the iden tically
108. nd THRESHF to 0 001 8 DATA FOR CONVENTIONAL TABLE CI DIAGONALISATION 27 Example Assuming the default selection thresholds Tmin 10 Tinc 10 and the default number of extrapolation passes NEXTRP 2 then presenting the data line ACCURACY 0 005 0 0005 will result in a diagonalisation threshold of 0 005 for the two extrapolation passes corresponding to solving the secular problem at selection thresholds of 20 and 30 microhartree and a threshold of 0 0005 for the final secular problem that corresponding to the 10 microhartree selection 8 4 MAXD This directive consists of one line read to variables TEXT MAXD using format A l e TEXT should be set to the character string MAXD e MAXD specifies the maximum number of iterative cycles to be carried out by the Davidson diagonalizer The directive may be omitted when MAXD will take the default value 50 8 5 TRIAL This directive may be used to define a trial Cl wavefunction by diagonalising a sub Hamiltonian obtained by sampling the diagonal elements of the Cl Hamiltonian and selecting the lowest energy terms The directive may be used to define the number of such elements to be included in the sub Hamiltonian and is read to the variables TEXT NTRIAL using format A l e TEXT should be set to the character string TRIAL e NTRIAL should be set to the number of lowest energy configurations in the Cl list to be used in constructing the sub Hamiltonian The TRIAL directive
109. nsformation module such that e Rreps having zero orbitals are discarded and e orbitals of common Rrep are grouped together these groups being arranged in order of increasing IRrep number and e orbitals of common Rrep are ordered according to their relative disposition in the input orbital set e g by eigenvalue ordering if SCF MOs 6 Generation of the Transformed Integral Interface to FTN031 is carried out on completion of the integral transformation using the integrals written to the Transformed Integral File ED6 This conversion is triggered by the appearance of the MRDCI DIRECT data line in the job input any attempt to use the Transformed Integral File produced say during a run of the Direct Cl module see Part 5 as input to a subsequent semi direct Table Cl calculation will lead to an error condition as the generation of FTNO31 would have not have attempted by the previous run of the transformation module Example 1 In this example we wish to perform a valence Cl calculation on the H2CO molecule using a DZ basis of 24 gtos looking to freeze both the oxygen and carbon 1s orbitals and to discard the inner shell complement orbitals An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 14 2 4 3 6 with the following orbital assignments from the closed shell SCF 17 DATA FOR SEMEDIRECT TABLE CI INTEGRAL TRANSFORMATION 0000000 0000000
110. of different symmetry to the first then their vectors will almost certainly reside on a different interface which we will assume reside on FTNO37 i e IF TNE should be set to 37 ISECE defines the position of the first of the excited state vectors on the interface defined by IFTNE Typically if all the states involved are of the same symmetry residing on the same data set IFTNX IFTNE then the first excited state vector will be located second on the data set i e ISECE 2 When the first and excited states are of different symmetry then different data sets will be involved and the first of the excited state vectors will be the first on IFTNE NSTATE defines the number of excited state vectors involved and is usually equal to the number of transition moment calculations to be performed Thus if we wished to calculate the transition moment between the two lowest states of H20 then NSTATE would equal 1 and the TM data would appear as follows TM 36 1 36 2 1 24 SEMEDIRECT TABLE CI USING DEFAULT OPTIONS 84 24 Semi direct Table CI Using Default Options In order to simplify the process of configuration specification and data preparation the semi direct module now provides a set of default options that require little or no data input While these defaults are not expected to cover most in depth requirements they do provide a starting point for users and a route to subsequent more extensive calculations To illustrate this default
111. of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass DIAG processing Such usage is typically associated with restarting Table Cl calculations 8 2 EXTRAP This directive consists of one line read to variables TEXT MAXE using format A l e TEXT should be set to the character string EXTRAP e MAXE specifies the maximum number of extrapolation cycles to be carried out by the Davidson diagonalizer The directive may be omitted when MAXE will take the default value 2 83 ACCURACY This directive may be used to define the diagonalisation thresholds for the extrapolation passes and for the final secular problem at the threshold Tmin and consists of a single line read to variables TEXT THRESHO THRESHF using format A 2F e TEXT should be set to the character string ACCURACY or DTHRESH e THRESHO On the NEXTRP extrapolation passes the diagonalization is converged to a threshold THRESHO e THRESHF On the final diagonalisation solving the secular problem corresponding to the threshold Tmin the diagonalization is converged to a threshold THRESHF The THRESH directive may be omitted when THRESHO will be set to 0 005 a
112. onalisation 1 DIAG oe srona RA Oe ESTARE car Sent eS A A A AR RAE A Or ACCURACY cirio A wed OF MANI lt lt ipod E e A e a aD A RA A oS ee eS EN e a is ee ar a Se A A e A e 9 Data for Conventional Table Cl Natural Orbitals 9 1 Natural Orbital Data NATORB o e ce coco soe yaana g a e a p e 9 2 Natural Orbital Datas CIVEC s s s ome saor 8604 na g a Ee ES 9 3 Natural Orbital Data PUTQ aoaaa e o 10 Data for Conventional Table Cl One electron Properties IOI PROP A roe Swe ee be Se PORES eee eee we SS oS ESS e ke eS De we Eh ER BS MEA or a oe eh oe be ee A BES ee eee ee E a WA MOPR o ae Beets BE A ye KER A EA AA eS 10 5 Configuration Data Lines s c sa a 8 A ae ee eran eR ei 00 o Oo oo O 10 11 23 24 25 25 25 26 26 26 27 27 27 28 28 28 29 CONTENTS 11 Data for Conventional Table Cl Transition Moments 12 Calculating the A states of H20 13 Iterative Natural Orbital Calculations 14 Table Cl Calculations Using MCSCF Orbitals 15 The Semi direct Table Cl Module 15 1 Sub Module Structure of Semi direct Table Cl 02 2 16 Directives Controlling Semi direct Table Cl Calculations UN TARO e doo he ir a De eS a ee EE Ss 17 Data for Semi direct Table Cl Integral Transformation 18 Data for Semi direct Table Cl Data base Generation E TABLE Is e e de o Gee o 19 Data for Semi direct Table Cl Selection LA SELEC A SANO 19 2 CNTRE AAN IIS SPIN e i II EA 194 SYMMETRY sa go
113. onducted in the C point group An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS and the following orbital assignments characterising the closed shell SCF configuration 1a 2a 1e13a24a72e45a 15 or in the C symmetry representation 1a22a1a2304a 56020270 16 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 62 1 1 79 93661395 2 0000000 2 1 7 48916431 2 0000000 3 nh 5 38319410 2 0000000 4 2 5 38319405 2 0000000 5 1 5 38149104 2 0000000 6 1 0 85610769 2 0000000 7 1 0 52191424 2 0000000 8 2 0 52191424 2 0000000 9 1 0 38579686 2 0000000 10 1 0 16819544 0 0000000 11 2 0 16819544 0 0000000 12 1 0 26587776 0 0000000 13 1 0 46072690 0 0000000 14 2 0 46072690 0 0000000 15 1 0 47871033 0 0000000 16 1 0 56106989 0 0000000 17 1 0 89229884 0 0000000 18 2 0 89229885 0 0000000 19 2 0 91131383 0 0000000 20 1 0 91131383 0 0000000 21 1 0 93118300 0 0000000 22 1 1 17900613 0 0000000 23 2 1 45058658 0 0000000 24 al 1 45058658 0 0000000 25 1 3 78674557 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the five inner shell orbitals 1a 2010 30 44 17 IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos a 1 18 4 14 1 14 a 2 7 1 6 15 20 To perform an 8 electron valence Cl calculation involving the SCF configuration and two de generate 1e
114. ormat 3A e TEXT should be set to the character string TABLE e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress output from the module IPRINT to produce an intermediate level of output FPRINT or DEBUG to produce output suitable for debugging purposes e TEXTB is a further optional parameter that should be set to the string BYPASS if the user wishes to bypass generation of the Table Cl data base Such usage assumes that the data set table ci is resident in the directory in which the calculation is preceding having been generated there is some previous run of the direct Cl module Such usage is typically associated with restarting Table Cl calculations 19 Data for Semi direct Table CI Selection Data for the configuration selection module is initiated with the SELECT directive followed by those directives characterising the symmetry of the state s of interest and reference configura tions CNTRL SPIN SYMMETRY CONF etc and terminated by data ROOTS THRESH controlling the process of selection 19 1 SELECT The SELECT directive is used to control the configuration selection module and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string SELECT e TEXTF is an optional parameter that may be used to control
115. rbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 6 DATA FOR CONVENTIONAL TABLE CI SELECTION and the following orbital assignments from the converged closed shell SCF BPR FOMOMOOAN OOP WNEH e e w N NNDNNNRR R CO 6 WNrFrFODADAAN DW N p ES e PNWORENWAREPEPANWHRHEAWNNWAHREABAWNHRE 77130710 12943325 23503117 23503117 23503117 73046864 48480821 48480821 48480821 16291387 16291387 16291387 25681257 33606346 37087856 37087856 37087856 79946861 79946861 86232544 86232544 86232544 23833149 44033091 44033091 44033091 13181655 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 17 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the first 5 silicon inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos az 1 9 2 7 1 7 by 2 6 1 5 8 12 ba 3 6 1 5 13 17 a2 4 6 1 5 18 22 To perform a 8 electron valence Cl calculation based on the SCF configuration would require the following CONF data CONF 018 13 18 The complete data file for performing the SCF and subsequent Cl would
116. ree NOs derived from the NVEC Cl vectors nominated by the CIVEC directive are to be placed Example 22 DATA FOR SEMI DIRECT TABLE CI ONE ELECTRON PROPERTIES 80 PUTQ AOS 100 120 The spin free NOs in the basis set representation are output to sections 100 and 120 respectively of the Dumpfile A section setting of 0 on the PUTQ directive will act to suppress natural orbital output to the Dumpfile 22 Data for Semi direct Table CI One electron Properties 22 1 PROP The PROP directive is used to request the computation of one electron properties and comprises a single data line read to the variables TEXT TEXTF and TEXTB using format 3A e TEXT should be set to the character string PROP e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module IPRINT to produce an intermediate level of output FPRINT or DEBUG to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass PROP processing Such usage is typically associated with restarting Table Cl calculations 22 2 CIVEC The CIVEC directive is used to specify those eigenvectors of the Cl matrix to be analysed The directive consists of a single data line with the character string CIVEC in the first data field If the prop
117. s will be much reduced in high symmetry e In contrast to the other post Hartree Fock modules of GAMESS UK the Table Cl routines make extensive use of unformatted sequential FORTRAN data sets or interfaces The data set reference numbers and associated LFNs of these files have been given in Table 6 of Part 2 2 Directives Controlling Conventional Table CI Calculations Data input characterising conventional Table Cl calculation commences with the MRDCI data line and is typically followed by a sequence of directives terminated by presenting a valid Class 2 directive such as VECTORS or ENTER An overview of the data structure has been given in Part 2 we provide additional detail on the directives associated with each sub module below 2 1 MRDCI The Table Cl data initiator consists of a single line containing the character string MRDCI in the first data field It acts to transfer control to those routines responsible for inputing all data relevant to the MRDCI calculation Termination of this data is achieved by presenting a valid Class 2 directive that is not recognised by the Table Cl input routines for example VECTORS or ENTER 3 Data for Conventional Table CI Symmetry Adaptation The Symmetry Adaptation module generates the list of symmetry adapted 1 and 2 electron integrals using as input the full list of integrals in the basis function representation It is assumed that this latter list is not skeletonised but has been generated w
118. should be set to the character string TRAN e ISECV is an optional integer parameter used to specify the section number on the Dumpfile wherein lies the set of eigenvectors to be used as the molecular orbital coefficient array in the integral transformation If ISECV is omitted the MOs will be either be taken from the section nominated on the ENTER directive or from the default eigenvector section deemed to be in effect through the associated SCFTYPE Note that examples of restoring orbitals from both MCSCF and CASSCF calculations are given below e TEXTF is an optional parameter that may be used to control the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress output from the module 4 DATA FOR CONVENTIONAL TABLE CI TRANSFORMATION 4 IPRINT to produce an intermediate level of output FPRINT to produce output suitable for debugging purposes e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass generation of the transformed integrals Such usage is typically associated with restarting Table Cl calculations e TEXTC is an optional parameter that should be set to one of the strings CORE or FREEZE if orbitals are to be frozen in the transformation e TEXTD is an optional parameter that should be set to one of the strings DISCARD or DELETE if orbitals are to be discarded in the transformation Additional data lines for the
119. spin coupled orbitals where the sequence numbers refers to the symmetry ordered orbitals performed at the outset of processing Within the set of open and doubly occupied orbitals the MOs are presented in groups of common Rrep with the groups presented in order of increasing IRrep sequence number Note that all reference function nominated by CONF must be of the same symmetry as that nominated on the SYMMETRY directive A few examples will help clarify this order of presentation Example 1 Consider performing a valence Cl calculation on the PH3 molecule using a 6 31G basis While the molecular symmetry is Cz the symmetry adaptation and subsequent Cl will be conducted in the C point group An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS and the following orbital assignments characterising the closed shell SCF configuration 1a 2a71e43a24a72e45a 1 or in the C symmetry representation 1a 2a 1a 3a 4a 5a 6a 20 Ta 2 M O IRREP ORBITAL ENERGY ORBITAL OCCUPANCY al 1 79 93661395 2 0000000 2 1 7 48916431 2 0000000 3 1 5 38319410 2 0000000 4 2 5 38319405 2 0000000 6 DATA FOR CONVENTIONAL TABLE CI SELECTION 13 5 1 5 38149104 6 1 0 85610769 7 1 0 52191424 8 2 0 52191424 9 1 0 38579686 10 1 0 16819544 11 2 0 16819544 12 1 0 26587776 13 1 0 46072690 14 2 0 46072690 15 1 0 47871033 16 1 0 56106989 17 1 0 89229884 18 2 0 89229885 19
120. states This iterative treatment is requested and controlled by the user through a number of options specified by the ITERATE directive Before detailing these options we would point to the following aspect of ITERATE usage e When specifying higher angular functions in the basis set to describe either Rydberg or Polarisation functions the user is strongly recommended to use the HARMONIC directive to conduct all MRDCI calculations in a spherical harmonic rather than cartesian basis e As described previously the memory requirements of the semi direct module may be sig nificantly greater than those associated with the conventional algorithm As the iterative cycles of the ITERATE algorithm proceed each energy calculation will become more demanding in memory as the number of associated reference configurations and overall size of the selected configuration space increases While the default memory allocations may prove adequate at the outset of ITERATE processing they are unlikely to prove so throughout and the user should use the MEMORY pre directive to request at least 20 MWords in calculations with say more than 20 active electrons The user should try and avoid the onset of multi passing of the eigenstate generation a consequence of running with restrictive memory in the later stages of the iterative processing 29 1 The ITERATE Directive and Associated Options The role of the ITERATE directive is twofold i to trigger a sequence of i
121. ta driven allocation is achieved under control of the CORE directive which should be presented immediately after the MRDCI DIRECT data line The directive consists of a single line with the character string CORE in the first data field Subsequent data fields are read in pairs to variables TEXTM IMEMM in format A I each such pair indicating the parameter which is to be modified followed by an integer defining the revised value Valid data fields are thus e TEXTM may be set to one of the character strings NTEINT IOTM NEDIM or MDI 26 CALCULATING THE A STATES OF H30 92 e IMEMM is an integer parameter used to specify the required valued to be used in deter mining the memory Example Presenting the data line MRDCI DIRECT CORE NTEINT 5000000 will act to increase the memory allocation for holding the transformed two electron integrals 26 Calculating the A states of H O To clarify our discussion of the Semi direct Table Cl module we work through a typical example of using the Table Cl method in calculating the energetics and properties of the three low lying A states of the H2O molecule The basis set employed is the TZVP triple zeta plus polarisation set this is augmented with a diffuse s and p orbital on the oxygen to provide a reasonable description of the known Rydberg character of the states of interest The computation is split into four separate jobs in which we 1 perform the initial SCF 2 carry
122. tals 105 29 Iterative MRDCI Calculations 110 29 1 The ITERATE Directive and Associated Options 111 29 1 1 The MAXITER Directive 111 29 1 2 The WEIGHT Directive a a a aa 112 29 13 The C2 Directive sa cee aor tasa bee dae aaa 112 29 1 4 The Ethylene Ground state Wavefunction 112 29 2 The Algorithm for Controlling Multi root Calculations o o aoaaa 114 29 2 1 The MAXROOT Directive aoaaa a a a 115 29 2 2 The SROOT Directive gt ss 62 2 ew eee tsi dasari erat 115 29 2 3 The DROOT Directive 2 26 cores keena ea 116 29 2 4 The RETAIN Directive aooaa be a ee we 116 CONTENTS v 29 3 Examples of Excited State Generation o oo a e 118 29 3 1 Calculating the tA states of Formaldehyde 118 29 3 2 Calculating the Excited States of the Formyl Radical 119 2933 The AA StS once ke hee em ER Se RAE CRAKS 119 TA The A States coco boo eb boo RES See eke be ae 120 1 INTRODUCTION 1 1 Introduction In this chapter we describe the data requirements of Table Cl a conventional configuration driven Cl module featuring configuration selection and energy extrapolation The methods used in the package are described in 1 Note that Version 6 2 of GAMESS UK also contains a new more efficient semi direct version of the Table Cl module that is capable of performing significantly larger calculations At this stage both old
123. terative MRDCI calculations rather a single calculation and ii to provide a mechanism for overriding the default MRDCI settings The latter is achieved by specifying the ITERATE options described below on one or more data lines each containing the character string ITERATE in the first data field the user may present as many data lines as desired in specifying these options providing the mechanism for presenting long option lists over several lines Note that the ITERATE data lines should be the last of the MRDCI options presented being typically followed by e g the VECTORS or ENTER directive 29 1 1 The MAXITER Directive This directive may be used to specify the maximum number of iterative MRDCI calculations to be undertaken The directive consists of two data fields read to the variables TEXT MXITER using format A I e TEXT should be set to the character string MAXITER e MXITER is an integer used to specify the maximum number of MRDCI calculations to be performed The directive may be omitted when MXITER will be set to the default value of 8 Example 29 ITERATIVE MRDCI CALCULATIONS 112 ITERATE MAXITER 20 29 1 2 The WEIGHT Directive This directive may be used to establish the criterion whereby a secondary configuration will be elevated to the status of a reference function in all subsequent iterative calculations The directive consists of two data fields read to the variables TEXT CWEIGHT using format A F e TEXT sho
124. the Nz molecule using a TZVP basis While the molecular symmetry is Don the symmetry adaptation and subsequent Cl will be conducted in the Dg point group The resolution of the Daon into the Dg orbital species has been given in Table 2 An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 15 66716423 2 0000000 2 5 15 66241865 2 0000000 3 1 1 51005217 2 0000000 4 5 0 76128176 2 0000000 5 1 0 63704931 2 0000000 6 3 0 63448705 2 0000000 7 2 0 63448705 2 0000000 8 6 0 17408343 0 0000000 9 7 0 17408343 0 0000000 10 5 0 30302673 0 0000000 11 3 0 38747796 0 0000000 12 2 0 38747796 0 0000000 13 1 0 42599317 0 0000000 14 1 0 49515007 0 0000000 15 7 0 57046706 0 0000000 16 6 0 57046706 0 0000000 17 5 0 92638361 0 0000000 18 5 1 12927523 0 0000000 19 1 1 83974927 0 0000000 20 3 2 01160646 0 0000000 21 2 2 01160646 0 0000000 18 DATA FOR SEMEDIRECT TABLE CI DATA BASE GENERATION 57 22 5 2 01683641 0 0000000 23 4 2 08974396 0 0000000 24 1 2 08974396 0 0000000 25 7 2 16718057 0 0000000 26 6 2 16718057 0 0000000 27 1 2 16945852 0 0000000 28 3 2 20023738 0 0000000 29 2 2 20023738 0 0000000 30 8 2 67650248 0 0000000 31 5 2 67650248 0 0000000 32 5 2 98166352 0 0000000 33 1 3 57616884 0 0000000 34 7 3 58056531 0 0000000 35 6 3 58056531 0 0000000 36 5 4 37707106 0 0000000 37 1 6 02043977 0 0000000 38 5 6 12816913 0 0000000 39 1 35 89673292 0 0000000 40 5 35 9192686
125. the quantity of printed output produced by the module Valid settings include the strings NOPRINT to suppress the major part of the output from the module in particular all details of the perturbative energy lowerings associated with the initial set of configurations IPRINT to produce an intermediate level of output FPRINT or DEBUG to produce output suitable for debugging purposes This in cludes the energy lowerings associated with the complete configuration list e TEXTB is an optional parameter that should be set to the string BYPASS if the user wishes to bypass SELECT processing Such usage is typically associated with restarting Table Cl calculations 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 59 192 CNTRL This directive consists of one line read to variables TEXT NELEC using format A e TEXT should be set to the character string CNTRL or NELEC e NELEC is used to specify the total number of active electrons in the Cl calculation Note that any inner shell electrons frozen out under control of the CORE directive should not be included 193 SPIN This directive consists of one line read to variables TEXT NSPIN using format A e TEXT should be set to the character string SPIN e NSPIN is used to specify the spin degeneracy of the Cl wavefunction of the electronic eigenstate s of interest using the values 1 2 3 etc for singlet doublet triplet states etc respectively It is also possible to us
126. tially the user must specify certain of the SELECT data fields the fundamental decision taken here was that he she should NOT have to specify any CONF data i e any explicit configuration date in the entire process The consequences of this decision are twofold 1 the following SELECT directives CNTRL SYMMETRY SPIN must be presented together typically with THRESH and ROOTS 2 the user is responsible for ensuring that the set of input orbitals restored under control of the VECTORS directive are consistent with the specifications given under SYMMETRY and SPIN We shall clarify this requirement below note that the code will check for this consistency and abort if it is not obeyed The initial calculation will be carried out using an internally constructed set of main config urations from which NROOT states will be generated as specified by the ROOTS directive Experience suggests that setting NROOT to 5 is generally quite reasonable Subsequent cal culations will iteratively generate a number of higher states as specified under control of the MAXROOT SROOT DROOT and RETAIN directives 29 ITERATIVE MRDCI CALCULATIONS 115 29 2 1 The MAXROOT Directive This directive may be used to specify the number of states to be obtained from the sequence of iterative MRDCI calculations The directive consists of two data fields read to the variables TEXT MXSTATE using format A l e TEXT should be set to the character string MAXROOT e M
127. try is Dooh the symmetry adaptation and subsequent Cl will be conducted in the Dg point group The resolution of the Daon into the Dg orbital species is given in Table 2 An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 15 65953319 2 0000000 2 5 15 65476539 2 0000000 3 1 1 50616872 2 0000000 4 5 0 75782843 2 0000000 5 1 0 63245667 2 0000000 6 2 0 63136574 2 0000000 7 3 0 63136574 2 0000000 8 7 0 20154120 0 0000000 9 6 0 20154120 0 0000000 10 5 0 63882720 0 0000000 11 1 0 82490877 0 0000000 12 2 0 89633728 0 0000000 13 3 0 89633728 0 0000000 14 1 0 91811776 0 0000000 15 6 1 10035435 0 0000000 16 7 1 10035436 0 0000000 17 5 1 17624961 0 0000000 18 5 1 66993831 0 0000000 19 4 1 70516525 0 0000000 20 1 1 70516525 0 0000000 21 2 1 91000364 0 0000000 22 3 1 91000364 0 0000000 23 8 2 29434948 0 0000000 24 5 2 29434948 0 0000000 25 1 2 84353563 0 0000000 26 6 3 00847612 0 0000000 27 7 3 00847612 0 0000000 19 DATA FOR SEMEDIRECT TABLE CI SELECTION 69 28 5 3 37444027 0 0000000 29 1 3 71749475 0 0000000 30 5 4 09916047 0 0000000 Based on the above output the CONF data lines may be deduced from the following table where we assume that we wish to freeze the two Nls inner shell orbitals IRrep IRrep No of Basis Frozen Active Sequence No Functions MOs MOs Nos Og 1 8 1 7 1 7 Tug 2 3 0 3 8 10 Tuy 3 3 0 3 11 13 gzy 4 1 0 1 14 Tu 5 8 1 7 15
128. uence numbers 6 7 8 and 9 respectively The symmetry re ordered sequence numbers allowing for the effective removal of the two a orbitals are 3 24 18 and 5 respectively To perform a three root 8 electron valence Cl calculation based on the SCF configurations of the ground and excited Rydberg states involving the single excitations 1b to 2b and 3a to 4a would require the following CONF data CONF 0 12 2 2 3 1 17 23 17 23 21718 12 23 END The following data will perform this three root Cl where 26 CALCULATING THE A STATES OF H30 95 the SCF computation is BYPASS ed e the freezing and discarding of the two a MOs is accomplished using the CORE and ACTIVE directives e the default sub module specifications are in effect with no specific need to reference Cl or NATORB activity we assume that the table ci data set is not available to the job and is to be generated in this run e the ROOTS directive is specifying selection with respect to the first 3 roots of the zero order problem which we assume will correspond to the states of interest RESTART NEW TITLE H20 TZVP DIFFUSE S P TABLE CI 3M 3R SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM 0 H 1 0 951 H 10 951 2 104 5 END BASIS TZVP O TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI ACTIVE 2 TO 35 END CORE 1 END MRDCI DIRECT SELECT CNTRL 8 CONF 0 12 17 23 2 2 3 1 17 23 21718 12 23 END THRE 10 10 ROOTS 3 ENTER Jo
129. uld be set to the character string WEIGHT e All configurations with a Cl weight coefficients greater than CWEIGHT in magnitude in any of the derived Cl wavefunctions will in all subsequent MRDCI iterations be treated as a reference function The directive may be omitted when CWEIGHT will be set to the default value of 0 005 Example ITERATE WEIGHT 0 003 29 1 3 The C 2 Directive Limited to the treatment of a single Cl wavefunction this directive may be used to define the required level of accuracy of the final wavefunction as reflected by the sum of the weights C 2 of the main reference configurations The directive consists of two data fields read to the variables TEXT WEIGHTM using format A F e TEXT should be set to the character string C 2 e The reference set will continue to be augmented with secondary configurations until the final value of C 2 coefficients for the Cl wavefunction equals or exceeds the value specified by WEIGHTM This is accomplished by reducing the default level of CWEIGHT for the secondary coefficients in consecutive MRDCI iterations The directive may be omitted when WEIGHTM will be set to the default value of 0 95 Example ITERATE C 2 0 95 29 1 4 The Ethylene Ground state Wavefunction TITLE ETHYLENE CI GROUND STATE ITERATE to C 2 0 95 29 ITERATIVE MRDCI CALCULATIONS 113 ZMATRIX ANGSTROM 4 1 2 120 0 1 2 120 0 3 180 0 1 1 120 0 3 0 0 1 1 120 0 3 180 0 Tama NNRPRP
130. ves explicit mention of the TABLE and CI modules in addition to flagging the previous SELECT data lines with the BYPASS keyword We have routed the natural orbitals from the 3 Aj states to the Dumpfile using the PUTQ directive RESTART CI TITLE xk H20 TZVP DIFFUSE S P TABLE CI ANALYSIS SUPER OFF NOSYM BYPASS SCF TRAN ZMAT ANGSTROM 0 H 1 0 951 H 10 951 2 104 5 END BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END ACTIVE 2 TO 35 END CORE 1 END 26 CALCULATING THE A STATES OF H30 99 RUNTYPE CI MRDCI DIRECT TABLE BYPASS SELECT BYPASS CNTRL 8 CONF 0 12 17 23 2 2 3 1 17 23 2 2 4 1 17 23 2 2 5 1 17 23 21718 12 23 21819 12 23 4 17 182325 1 2 4 17 18 2327 1 2 4 2 71718 1 23 4 2 31719 1 23 4 2 32325 1 17 4 2 32327 1 17 END THRE 10 10 ROOT 3 CI BYPASS NATORB IPRIN CIVE 1 2 3 PUTQ AOS 50 51 52 PROP CIVE 1 2 3 0 12 17 23 21718 12 23 2 2 3 1 17 23 MOMENT 36 1 36 2 2 ENTER Description of the Output for MRDCI Moments The MRDCI module calculates the oscillator strength using both the dipole length formalism f r 2 3 lt VW rjw gt AE and the dipole velocity formalism HU 2 3 lt Y y W gt AE The most significant contributions due to individual molecular orbitals are printed out as a table containing the largest coefficients of the transition density matrix and the following correspond ing integrals lt vilz y gt lt bilyly gt lt bi21103
131. w semi direct version of the Table Cl module that is capable of performing significantly larger calculations The main differences as far as the user is concerned include the following e The original adapt and transformation modules of the Conventional Table Cl module have now been replaced by the standard 4 index transformation module of GAMESS UK e Semi direct Table Cl calculations require at least two reference configurations i e CISD calculations based on a single reference configuration are not possible with this module However we do not consider this to be a major disadvantage given that the process of configuration choice and specification has been simplified through the use of automated configuration generation see below e The original Cl and Diagonalisation modules have now been condensed into a single Cl module e The formal limits that apply to conventional calculations are significantly extended in the semi direct module There is now a limit of 800 000 selected configurations derived from an initial list of configurations generated by single plus double excitations from a 15 THE SEMEDIRECT TABLE CI MODULE 52 user specified list of reference functions the number of which may not exceed 256 The selection and extrapolation procedure may now be applied on up to thirty roots of a given secular problem e The memory requirements of the semi direct module may be significantly greater than those associated with the conventional
132. which point the iterative process will terminate The DROOT directive will typically appear together with the SROOT directive in controlling the energetics of this process clearly the user may have little interest in deriving the final Cl vector of an eigenstate whose zero order description is separated by a large energy gap from the zero order vector of the preceding state Rejecting such a solution and terminating the MRDCI process is controlled by the value specified by SROOT Example SROOT 0 20 DROOT Given the above data sequence the iterative MRDCI process will continue until either a MXSTATE eigenvectors of the Cl matrix have been obtained or b having obtained N roots of the Cl eigen vector the N 1 th root of the zero order problem lies more than 0 20 a u above the Nth root The following points should be noted e The strategy behind the iterative sequence follows from an appreciation of the quantities defined by the MAXROOT SROOT and DROOT directives In practice three criteria are used in deciding whether to continue the sequence of MRDCI calculations 1 has the number of main configurations requested using WEIGHT changed from the preceding pass 2 does the eigen value spectra of the zero order problem justify expanding the no of roots NROOTS to be used in selection i e ABS EIGVAL NROOTS 1 EIGVAL NROOTS lt ROOTDEL 3 based on 2 should we increase the no of eigenstates roots to be extracted from the D
133. will correspond to the SCF configuration the double and single excitation arising from the DOMO 5a to VMO 6a and the double and single excitation arising from the DOMO 2b3 to VMO 3b3 Note that there no exci tations involving a2 or b MOS given the absence of such orbitals in the doubly occupied manifold This now results in a total reference set of just 5 functions as shown thus in the job output numbers of open shells and corresponding main configurations 1 28 1 2 3 4 5 37 38 Ya SCF configuration 1 28 1 2 3 4 6 37 38 a 5a1 gt 6a1 double 3 5 6 28 1 2 3 4 37 38 Ei 5a1 gt 6a1 single 1 28 1 2 3 4 5 37 39 s 2b2 gt 3b2 double 3 28 38 39 1 2 3 4 5 37 a 2b2 gt 3b2 single The sequence of data lines defining the Semi direct Table Cl calculation is again terminated by the ENTER directive Note that the full data specification corresponding to the defaults generated from the above data file is as follows TITLE H2C0 2B1 TZVP EXPLICIT DATA FOR DEFAULTS MULT 2 CHARGE 1 SUPER OFF NOSYM 25 MEMORY SPECIFICATION FOR THE SEMI DIRECT TABLE CI MODULE 90 ZMAT ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END BASIS TZVP RUNTYPE CI OPEN 1 1 ACTIVE 1 TO 52 END MRDCI DIRECT TABLE SELECT CNTRL 15 SPIN 2 SYMM 2 SINGLES ALL CONF 28 28 5 28 28 38 39 37 38 37 38 37 38 37 39 4 5 37 N N wewer eOrr N 00 Pwr WwW Ww OPN PAD wauw ogo NS END THRESH 10 10 ROOTS 1 CI NA
134. xplicit zero order Hamilto nian Ho over the reference functions described above followed by perturbative selection of configurations with respect to the lowest root of Hp The minimum threshold to be used in selection Tmin is 10 micro Hartree with an increment of 10 uH to be used in defining the higher threshold case to be solved in the process of extrapolation 1 7 In default the module will having solved the secular problem for the lowest root of the Cl secular problem generate the spinfree natural orbitals from the associated Cl eigenfunction The sequence of data lines defining the Semi direct Table Cl calculation is terminated by the ENTER directive Note at this stage that the full data specification corresponding to the defaults generated from the above data file is as follows TITLE H2C0 TZVP EXPLICIT DATA FOR DEFAULT MRDCI SETTINGS SUPER OFF NOSYM ZMAT ANGSTROM N C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 24 SEMI DIRECT TABLE CI USING DEFAULT OPTIONS END BASIS TZVP RUNTYPE CI ACTIVE 1 TO 52 END MRDCI DIRECT TABLE SELECT CNTRL 16 SPIN 1 SYMM 1 SINGLES ALL CONF 0 1 2 3 4 5 28 37 38 0 1 2 3 4 6 28 37 38 2 5 6 1 2 3 4 28 37 38 0 1 2 3 4 5 29 37 38 2 28 29 1 2 3 4 5 37 38 0 1 2 3 4 5 28 37 39 2 38 39 1 2 3 4 5 28 37 END THRESH 10 10 ROOTS 1 CI NATORB CIVEC 1 ENTER 24 2 Calculations on the Formaldehyde Cations 86 Let us now consider a Semi direct Table Cl calculation on t
135. y to this behavior can be found from examining the nature of the configurations tagged as r retain in the output from the selection module Initial attempts to implement an ITERATE strategy revealed that such configurations often remain isolated from the selected set of reference configurations throughout the iterative cycles even though their associated energy remains lower than those of the final eigenstates of some of the higher Cl states The RETAIN directive may be used to establish the criterion whereby the appearance of con figurations taged r will trigger their inclusion as reference functions in the subsequent MRDCI iterations regardless of their computed weights in any of the current eigenstates The directive consists of three data fields read to the variables TEXT ERETAIN IRETAIN using format A F l e TEXT should be set to the character string RETAIN e Those configurations that are tagged r in the selection process will be retained as ref erence configurations in the next iterative cycle if their associated energy lies within ERETAIN au of the energy of the NROOTS root of the current zero order eigen problem i e E con figuration EIGV AL N ROOTS lt ERETAIN 39 e IRETAIN may be used to delay the onset of this selection criteria Specifying IRETAIN results in the criteria only coming into effect on iteration IRETAIN of the MRDCI iterative process Example RETAIN 0 10 The following points should b

GAMESS-UK USER'S GUIDE and REFERENCE MANUAL Version

Contents

Download Pdf Manuals

Related Search

Related Contents