Home

GAMESS-UK USER'S GUIDE and REFERENCE MANUAL Version

Contents

1. RESTART TRANSFORM TITLE H2C0 TZVP BASIS TRANSFORMATION SUPER OFF NOSYM ZMATRIX ANGSTROM N 1 1 203 1 1 099 2 121 8 1 1 099 2 121 8 3 180 0 mzmoa END BASIS TZVP RUNTYPE TRANSFORM PASS 2 2 TRACC 9 ENTER 5 Direct CI Calculations The Direct Cl module performs general multi reference singles and doubles configuration inter action Cl calculations The method direct or integral driven used in the package is described in 1 For optimal running the main memory allocation should be at least 3 times longer than the number of configuration state functions CSFs in the Cl expansion The module will exe cute in less memory but at the cost of increased disc input output and higher overall job cost The following data sets will be used by the program e Transformed Integral File Integrals over the molecular orbitals MOs will be read from this file The user may direct that this dataset be read from any file except ED5 or ED8 by means of the FFILE directive see above The default is to use ED6 e Direct Cl File The file ED5 is used to hold control information accommodate the partial Hamiltonian matrix elements and to store the update vectors created by the Davidson diagonalization procedure Twice the Cl expansion length will be added at each iterative cycle The DIAGMODE directive see below may be used to reduce the maximum size of the Davidson sub space thus limiting the ultimate size of ED5 possibly at the ex
2. 2 0 5 The trial wavefunction will consist of a linear combination of the first and second CSF in the Cl expansion with coefficients 0 5 and 0 5 respectively It will be subsequently normalized by the program 2 Select a subspace of CSFs and compute the eigenvector of the matrix corresponding to the subspace In this case the syntax is TEXT TEXTA TEXTB NCSF or TEXTC TEXTD ISTATE TEXTE where e TEXT should be the literal string TRIAL e TEXTA should be the literal string DIAG e TEXTB may be set to SELECT or to FIRST If the string SELECT is supplied a number of CSFs with the lowest energies will be selected If the string FIRST is supplied a number of CSFs that are the first in order will be selected Only a maximum of 100 CSFs can be selected SELECT is default e TEXTC may be set to ref or vac selecting the whole reference space or vacuum space respectively 6 DIRECT CI DATA INPUT 14 e TEXTD may be set to state in which case ISTATE is an integer specifying which eigenvector from the subspace matrix should be used The default value is 1 i e the eigenvector with the lowest eigenvalue will be selected e TEXTE may be set to the literal string PRINT If supplied the trial vector will be printed By default the trial vector will NOT be printed if PRINT is given twice the selected H matrix is also printed Example TRIAL DIAG SELECT 25 This will cause the program to select the 25 CSFs with the lowest e
3. 22 20202020 2202202020 22 20022020 22 20200220 22 20202002 with 10 orbitals in the internal space and 10 in the external space given that the inner shell orbitals have been frozen The complete date file for performing the GVB Cl would then appear as follows where the canonicalised GVB orbitals are restored from the default section section 5 of the Dumpfile RESTART NEW TITLE H2C0 3 21G 4PAIR GVB SUPER OFF NOSYM BYPASS SCF ADAPT OFF ZMATRIX ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 11 099 2 121 8 3 180 0 6 DIRECT CI DATA INPUT END RUNTYPE CI SCFTYPE GVB 4 CORE 1 2 END ACTIVE 3 TO 22 END DIRECT 12 10 10 CONF 22 22 22 22 22 ENTER NNNON OOONO NNONN OONOO NONNN ONOOO ONNNN NOOO 0 39 We show below the output from the resulting Cl calculation SUM SUM TOTAL ENERGY 113 4467990858 95233628 08236057 08236123 05407166 16888657 03606575 03606465 07796058 PRPRPRPRPrRP RE VACUUM VACUUM VACUUM VACUUM VACUUM VACUUM VACUUM VACUUM oooooococo o oO OOO Om Ome OF SQUARES OF VACUUM CSF CI COEFFICIENTS OF SQUARES OF 5 MAIN CI COEFFICIENTS 6 25 FP Geometry Optimisations oP WN PB 163 224 279 2220202020 2202202020 2220022020 2220200220 2220202002 2121202011 2120212011 2220201111 973137161834E 00 951957435390E 00 Energy only optimisation for direct Cl wavefunctions may be performed using a variant of the RUNTYP
4. TITLE xx H20 TZVP DIFFUSE S P DIRECT CI 16M X1A1x SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM 0 H 1 0 951 H 10 951 2 104 5 END 6 DIRECT CI DATA INPUT 49 BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI CORE 1 END ACTIVE 2 TO 11 13 TO 16 12 17 TO 35 END DIRECT 8 14 20 CONF 2222000000000 0 22210010000000 22121000000000 22120001000000 2220002000000 0 22120000100000 22110010000010 22021000001000 22021000000010 22210000000100 22200010000100 22111000000100 21210010010000 21210010000001 21121000010000 21121000000001 ENTER The following points should be noted e There are now 14 orbitals in the internal space and 20 in the external space e We again assume that the Transformed Integral File is retained from the above job enabling the integral transformation to be bypassed in the jobs below e We show below the data file for the final Cl on the 11A state that on the third state follows in straightforward fashion Final CI Treatment of the 1 A state RESTART CI TITLE x H20 TZVP DIFFUSE S P DIRECT CI 16M 1A1 SUPER OFF NOSYM BYPASS TRANSFORM ZMAT ANGSTROM 0 H 1 0 951 H 10 951 2 104 5 END 50 6 DIRECT CI DATA INPUT 2222000000000 0 22210010000000 22121000000000 22120001000000 2220002000000 0 22120000100000 22110010000010 22021000001000 22021000000010 22210000000100 22200010000100 22111000000100 21210010010000 21210010000001 21121000010000 211210
5. strings VAC N 1 and N 2 refer to CSFs with 0 1 or 2 electrons in the external space respec tively MIN MAX specify the minimum and maximum number of electrons to be allowed to populate the internal MOs defined by the IORBS parameters IORBS A sequence of internal MO indices terminated by the integer 0 This data may be carried over to subsequent lines if necessary and the character string TO may be used to shorten the data if desired These pa rameters refer to the reordered MO list if a REORDER directive see above has been presented Example REFGEN VAC 78 1 T040 N 1 7 8 1 T040 N 2 7 8 1 T040 Internal MOs 1 to 4 inclusive are allowed to carry either 7 or 8 electrons in vacuum N 1 and N 2 CSFs 6 DIRECT CI DATA INPUT 34 6 20 PRCONF This directive consists of one line read to variables TEXT IPR using format A e TEXT should be set to the character string PRCONF e IPR specifies that every IPR th occupation pattern generated by the configuration gener ator is to be printed If IPR 1 all occupation patterns will be printed The directive may be omitted when no occupation patterns generated by the configuration generator will be printed The main use of the PRCONF directive is to generate a detailed occupation pattern listing Example PRCONF 1 6 21 CEPA The CEPA directive allows the user to calculate the unlinked cluster correction to the final Cl energy This correction factor is more accurate than the Davidson
6. 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 3 98212497 0 0000000 41 1 4 08851360 0 0000000 42 1 24 51368240 0 0000000 Assume that we wish to freeze the first 14 inner shell orbitals 10220 30 40 11 50 60 2215 70 80 3r requiring the following data lines for the transformation CORE 1 TO 14 END ACTIVE 15 TO 42 END 6 DIRECT CI DATA INPUT 24 To perform an 18 electron valence Cl calculation based on the SCF configuration 9740416410075 741107 7 would require the following CONF data CONF 222222222 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 ACTIVE 15 TO 42 END DIRECT 18 9 19 CONF 222222222 ENTER The inclusion of a second reference configuration corresponding to the doubly excited configu
7. 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END RUNTYPE CI ENTER then the calculation undertaken will be based on the following 1 The format of the 2e integral file will be automatically set to the required SUPER OFF NOSYM triggered by the presence of the Cl runtype 2 Integral transformation will use the set of orbitals from section 1 the default section for output of the closed shell SCF eigenvectors All orbitals will be deemed ACTIVE in the transformation 3 The Direct Cl module is the default module loaded under RUNTYPE Cl control so that the DIRECT directive is not required 4 The division of the molecular orbital space into an internal and external space typically specified by the DIRECT directive is now handled automatically with the internal space comprising all doubly occupied SCF MOs orbitals the external space all SCF virtual MOs All electrons will be deemed active in the Cl 5 The SYMMETRY and SPIN of the Cl wavefunction are taken to be those of the SCF wavefunction 6 A single reference configuration will be employed just the SCF configuration the final configuration space will include all single and double excitations from this SCF reference configuration 7 The spinfree natural orbitals will be written to section 11 of the Dumpfile The full data specification corresponding to the defaults generated from the above data file is shown below the role of each of the directives will be described in l
8. 6 DIRECT CI DATA INPUT 30 CONF 22 The complete data file for performing the SCF and subsequent Cl would then be as follows TIME 60 TITLE CAH2 3 21G 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 CORE 1 TO 9 END ACTIVE 10 TO 21 END DIRECT 4 2 10 CONF 2 2 ENTER 6 15 CONF CARDS The directive consists of s single line where the first 2 character strings are either CONF CARDS or CONF FILE or CONF ASCII On the same line sub directives may be specified FILE string A 44 charactrer string may be supplied specifying the file to read the configuration information from The file must be as generated by the CARDS CASSCF directive i e fiexed format like 1 222220000 9644755897 COEF value Specify a minimum absolute value for a coefficient of a configuration to be included Different spin paths are combined WEIGHT value Specify the minimum weight of a configurations including all its spin possibilities to be included DOC ndoc Specify the number of doubly occupied orbitals not in the configurations in the list to be prepended to each configuration NFRZ nfrz Specify the number of orbitals to be frozen and kept doubly occupied NFRZ and NDOC can cancel each other The reference configurations are printed so the working of this directive is easily checked 6 DIRECT CI DATA INPUT 31 6 16 REFGEN The first line should co
9. ENTER oro ooo 0 0 1 5 5 ALTERNATE The following points should be noted e the SCF and Transformation are BYPASS ed this assumes that the Transformed Integral file from the initial Cl has been saved e in this job we are describing an excited state the 11A so that the default diagonalisation controls will no longer prove satisfactory The specification of SHIFT ALTERNATE and VMIN is typical in such calculations The TRIAL directive is now specifying a starting vector with the second configuration in the CONF list as the dominant term Job 4 Direct CI Treatment of the 2 A state RESTART CI TITLE x H20 TZVP DIFFUSE S P DIRECT CI 3M 2A1 SUPER OFF NOSYM BYPASS TRANSFORM 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 RUNTYPE CI CORE 1 END ACTIVE 2 TO 35 END DIRECT 8 8 26 CONF 222 3 1 0 SHIFT 0 5 ALTERNATE 6 DIRECT CI DATA INPUT 46 VMIN ENTER The following points should be noted e the SCF and Transformation are again BYPASS ed e we are again describing an excited state the 2 A so that SHIFT ALTERNATE and VMIN are again specified The TRIAL directive is now specifying a starting vector with the third configuration in the CONF list as the dominant term The Final 16 Reference CI Jobs An examination of the output from the initial Cl calculations reveals that the dominant con figurations have as expected been incl
10. INPUT Al DIRECT 15 8 34 CONF 22222221 ENTER Note again that some care must be taken when reducing the orbital space in FP Cl optimisa tions In open shell calculations the Cl step will derive the orbital set at each point from the second section specified on the ENTER directive i e the energy ordered MOs If this ordering varies from point to point in the FP optimisation and symmetry is used in minimising the configuration space it is quite likely that this space will vary during successive points with disastrous consequences on the optimisation pathway As a general rule the user should only consider freezing or discarding orbitals that are well separated from those MOs included in the Cl space i e inner shell or inner shell complement MOs 6 26 Calculating the A states of H 0 To conclude our discussion of the Direct Cl module we work through a typical example of using the Direct Cl method in calculating the energetics and properties of the three low lying Ay 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 a number of separate jobs in which we 1 perform the initial SCF 2 carry out an initial Cl for each state where the reference set employed acts to provide at least a qualitative descripti
11. Phys Lett 140 1987 225 doi 10 1016 0009 2614 87 80448 7 7 P J A Ruttink J H van Lenthe R Zwaans and G C Groenenboom J Chem Phys 94 1991 7212 7220 doi 10 1063 1 460204 8 K Andersson P Malmqvist and B O Roos J Chem Phys 96 1992 1218 doi 10 1063 1 462209 9 P G Szalay R J Bartlett Chem Phys Lett 214 1993 481 doi 10 1016 0009 2614 93 85670 J 10 R Gdanitz R Ahlrichs Chem Phys Lett 143 1988 413 doi 10 1016 0009 2614 88 87388 3 11 L Fusti Molnar P G Szalay J Phys Chem 100 1996 6288 doi 10 1021 jp952840 j 12 P J A Ruttink J H van Lenthe P Todorov Mol Phys 103 2005 2497 doi 10 1080 00268970500180725
12. a single dataline in which the first data field should contain the character string EXCIT The second data field may also be read in A format and if so used should be set to one of the character strings OCTAL or BINARY If this second A format field is omitted the program takes OCTAL as default Subsequent data fields are read in l format and should contain either octal numbers valid range 0 to 7 or binary numbers valid range 0 to 1 according to the character string contained in the second data field The octal or binary integers may be continued onto subsequent lines if necessary We now explain the significance of these integers if OCTAL input mode is selected The first octal integer specifies the external excitation pattern for the reference CSFs where no internal excitations have been applied This octal number should be translated into a binary format such that OCTAL NUMBER BINARY NUMBER oo PRRFRROOOO RROOHRRoOo PROPORORO X OO0 UNro tl The left most binary integer defines a double external excitation where two electrons are pro moted from the internal to the external space If it is O or 1 the process is forbidden or allowed respectively The middle binary integer defines the single external excitation process If this is set to 1 the process is allowed if it is O single external excitations are forbidden The right most binary integer corresponds to a no external excitation process the reference CSFs being left as
13. and LOCK modes of the diagonalizer see DIAGMODE directive above do not ordinarily compute the variance and therefore do not output this quantity The PRINTVAR directive may be used to turn such printing on and is redundant if the VMIN option of the DIAGMODE directive is used since the variance is always printed in this case 6 DIRECT CI DATA INPUT 13 6 8 TRIAL This directive may be used to define the trial Cl wavefunction as a linear combination of CSFs There are two ways to accomplish this 1 Specify the coefficients of the CSFs in the input file In this case the first data line should contain the character string TRIAL in the first data field There may be up to 20 lines following this directive initiator each being read to variables ICSF CCSF using format LF e ICSF should be set to the index of a CSF in the Cl wavefunction Normally such an index will only be known after running the program once so that the configuration generator output can be studied The user should understand the order in which the program generates spin states belonging to the same space occupancy pattern see below before using this directive particularly where CSFs involving large numbers of singly occupied MOs are involved e CCSF should be set to the coefficient of the CSF in the trial wavefunction The trial wavefunction will be a linear combination of the indicated CSFs with coefficients given by the CCSF parameters Example TRIAL 1 0 5
14. or for all occupied orbitals ALL default The choice may be may by specifying the additional string DOC or ALL Additional options applicable to some or all of the variants mentioned above may be specified on the same card e For the approaches that use the modified correlation energy one may specify how this is calculated Choices are PROJECT where the expectation value of the Cl function projected onto the reference space is used as reference VARIA where the energy of the optimised reference space is employed or PAIRS computing the correlation energy as sum of pair energies this is project The default is VARIA which is Ahlrichs choice e Using the keywords SIN or NOSIN one may specify if single excitations should be shifted SIN or not The default is that singles are shifted except for single reference CEPA2 For the multi reference variant the option is dubious e Using the texstring IT one may specify after what iteration the CEPA mode may start using IT ITCEPA A Default is 3 e Using the textstring CRIT one may specify at which particular threshold determined by the TESTER in the diagonalisation phase of the Cl calculation the CEPA is switched on The format is CRIT CRITC A F Default of CRITC is 0 01 e PRINT requests intermediate printing within the CEPA mode e PAUL refers only to the CEPA 2 mode and invokes an unpublished EPV correction formula due to P J A Ruttinck e MICRO allows one
15. ration 9074016410075 741207 8 would require incorporating the 120 orbital into the internal space leading to 10 internal and 18 external MOs The CONF data would then appear as follows CONF 2222222220 22222202 and the overall data file RESTART NEW TITLE CUCL 3 21G BYPASS SCF ZMAT ANGSTROM CU CL 1 CUCL VARIABLES CUCL 2 093 END BASIS 3 21G RUNTYPE CI CORE 1 TO 14 END ACTIVE 15 TO 42 END DIRECT 18 10 18 CONF 2222 6 DIRECT CI DATA INPUT 25 where we have assuming the Mainfile and Dumpfile to have been retained by passed the SCF and modified the DIRECT CONF and VECTORS data 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 orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 1 1 68 77130710 2 0000000 2 1 6 12943325 2 0000000 3 2 4 23503117 2 0000000 4 3 4 23503117 2 0000000 5 4 4 23503117 2 0000000 6 1 0 73046864 2 0000000 T 4 0 48480821 2 0000000 8 3 0 48480821 2 0000000 9 2 0 48480821 2 0000000 10 2 0 16291387 0 0000000 11 3 0 16291387 0 0000000 12 4 0 16291387 0 0000000 13 1 0 25681257 0 0000000 14 1 0 33606346 0 0000000 15 3 0 37087856 0 0000000 16 2 0 37087856 0 0000000 17 4 0 37087856 0 0000000 18 1 0 79946861 0 0000000 19 1 0 7994686
16. the Dumpfile Omission is equivalent to CORE END 4 RUNTYPE and Restarting the Transformation In most applications the transformation module will be run as part of either Direct Cl or Green s function calculations under control of RUNTYPE Cl RUNTYP GF or RUNTYPE TDA speci fication In some circumstances it may be necessary to generate the transformed integrals only and a specific RUNTYPE code named TRANSFORM has been provided for this purpose Note that TRANSFORM processing includes both the SCF step and subsequent integral transforma tion with restarts possible in both steps For such restarting the Mainfile Secondary Mainfile and Transformed Integral file should have been permanent files in the startup job exactly the same files being presented to the restart job and if used identical MFILE SFILE and FFILE directives should be used in the startup and restart jobs The Sortfile need not be preserved between jobs Example In this example we are transforming the integrals from a TZVP calculation on H2CO The first data file represents the startup job the second the restart data assuming the processing in the startup job did not complete in the allocated time The Startup Job TITLE H2C0 TZVP BASIS TRANSFORMATION SUPER OFF NOSYM 5 DIRECT CI CALCULATIONS 6 ZMATRIX ANGSTROM Cc 0 1 1 203 H 11 099 2 121 8 H 11 099 2 121 8 3 180 0 END BASIS TZVP RUNTYPE TRANSFORM PASS 2 2 TRACC 9 ENTER The Restart Job
17. they are If set to 0 or 1 the reference CSFs will be eliminated from or retained in the final list of CSFs for the Cl Note that CSFs may also be eliminated from the Cl list because they are of the wrong spin space symmetry Example 1 EXCIT OCTAL 7 This excitation mask will cause the reference CSFs and the single and double external excita tions generated from them to be included in the final Cl list of CSFs Example 2 EXCIT OCTAL 5 6 DIRECT CI DATA INPUT 18 Will cause all the reference CSFs and all double external excitations generated therefrom to be added to the list of CSFs for the Cl Example 3 EXCIT OCTAL 4 Will cause the double external excitations of the reference states to appear in the final list of CSFs Note the reference and single external excited CSFs are excluded from the Cl list If a second octal integer is defined the reference CSFs have undergone a single internal ex citation process That is a transfer of an electron from one internal MO to another This second integer defines the external excitation mask on these newly constructed internal CSFs A third octal integer defines the external excitation pattern after a double internal excitation Additional octal integers may be presented up to a maximum to 21 integers Thus it is possi ble to define a state which is up to 20 fold internally excited and up to doubly externally excited Example 4 EXCIT OCTAL 7 3 1 corresponds to a Cl list containing the r
18. to control the CEPA micro iterations when the Cl vector and corre lation energy and shifts are updated without a matrix vector product in between The format is MICRO MCYC CRIT A F MCYC is the maximum number of micro iterations and CRIT the relative convergence criterion Defaults are 3 and 0 01 The directive can be invoked without parameters which will result in the following and crash for a multi reference case CEPA 1 SIN IT 3 CRIT 0 01 MICRO 3 0 01 6 22 MP The MP directive allows one to perform multi reference Mgller Plesset calculations 4 5 6 To generate the reference wavefunction one should run an MCSCF calculation on the required state first to obtain the correct orbitals and use the TRIAL DIAG directive to rebuild the MCSCF wavefunction Details on this process will be given in an example in the Direct Cl Multi reference MP section of Part 2 Once the reference wavefunction has been constructed the MP directive should be used to control the perturbation theory applied to it The syntax is 6 DIRECT CI DATA INPUT 36 TEXT TEXTA TEXTB TEXTC or IMODEL TEXTD ME where e TEXT should specify the character string MP e TEXTA specifies the order of perturbation theory required One can choose from 2 and ER e TEXTB should specify the character string MODEL to choose the form of the zeroth order Hamiltonian This Hamiltonian can be specified by name using RUTTINK to select HU S D minima tat
19. treated within the perfect pairing approximation The sequence of calculations included e performing the closed shell SCF calculation e localising the set of valence SCF orbitals e performing the GVB calculation using the set of LMO input under control of the VEC TORS option NOGEN An examination of the GVB output reveals the following orbital assignments with orbitals 5 7 9 and 11 corresponding to the strongly occupied orbitals and orbitals 6 8 10 and 12 to the weakly occupied orbitals of the GVB pairs 1 1 20 48204464 2 0000000 2 1 11 25090140 2 0000000 3 1 1 12106354 2 0000000 4 1 0 55607328 2 0000000 6 DIRECT CI DATA INPUT 38 5 1 1 42285552 1 9843054 6 1 0 02615928 0 0156946 7 1 1 42285519 1 9843052 8 1 0 02615932 0 0156948 9 1 1 93008469 1 9904691 10 1 0 02325080 0 0095309 11 1 1 07958389 1 9032736 12 1 0 09164423 0 0967264 13 1 0 62048723 0 0000000 14 1 0 87018377 0 0000000 15 1 0 87310154 0 0000000 16 1 0 92519148 0 0000000 17 1 1 03015198 0 0000000 18 1 1 38072957 0 0000000 19 1 1 79683989 0 0000000 20 1 1 81440429 0 0000000 21 1 1 97615975 0 0000000 22 1 3 26852884 0 0000000 Assume that we wish to freeze the Ols and Cls orbitals thus CORE 1 2 END ACTIVE 3 TO 22 END To perform a 12 electron valence Cl calculation based on the leading term in the GVB expansion together with those doubly excited configurations corresponding to each GVB pair would require the following CONF data CONF
20. 00 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 15 6 DIRECT CI DATA INPUT 43 25 1 2 72651832 0 0000000 26 2 2 73832720 0 0000000 27 1 3 07664215 0 0000000 28 3 3 26840142 0 0000000 29 2 3 54616570 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 Assuming that we wish to freeze the O1s inner shell orbitals and discard the inner shell comple ment orbital the following data lines should be presented in the transformation CORE 1 END ACTIVE 2 TO 35 END 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 re ordered sequence numbers allowing for the effective removal of the two a orbitals are 5 6 7 and 8 respectively To perform a balanced valence Cl treatment of the three states of interest will require a three root 8 electron reference set based on the SCF configurations of the ground and excited Rydberg states involving the single excitations 1b to 2b and 3a to 4a This specification will require the following CONF data In contrast to the Table Cl module see Part 6 where the three state Cl may be performed in a single job the Direct Cl treatment w
21. 000 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 29436539 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 1 3 71753400 0 0000000 30 5 4 09917273 0 0000000 Assume that we wish to freeze the two Nis inner shell orbitals thus CORE 1 2 END ACTIVE 3 TO 30 END To perform a 10 electron valence Cl calculation based on the SCF configuration 29 29214 20920730417 6 DIRECT CI DATA INPUT and associated 7 to 7 excitations da do sa lng 172 y 20 20 30 lana p172 y 20207307 lm mea 20220230212 172 da de sa iria dry would require the following CONF data CONF 22222 00 22202 02 22202 20 22220 20 22220 02 22211 11 28 10 11 12 13 14 with an internal space of 7 orbitals an external space of 21 note again that the ordering of the virtual MOs is such that no reordering is required within the ACTIVE directive The complete data file for performing the SCF and
22. 00000001 BASIS TZVP 0 TZVP H So 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI CORE 1 END ACTIVE 2 TO 11 13 TO 16 12 17 TO 35 END DIRECT 8 14 20 CONF SHIFT 0 5 ALTERNATE TRIAL 1 VMIN ENTER REFERENCES 51 References 1 P E M Siegbahn J Chem Phys 72 1980 1647 doi 10 1063 1 439365 V R Saunders and J H van Lenthe Mol Phys 48 1983 923 doi 10 1080 00268978300100661 2 M Yoshimine J Comp Phys 11 1973 449 P S Bagus B Liu A D McLean and M Yoshimine Energy Structure and Reactivity edited by D W Smith and W B McRae Wiley 1973 130 G H F Diercksen Theor Chim Acta 33 1974 1 doi 10 1007 BF00527620 S T Elbert Numerical Algorithms in Chemistry Algebraic Methods LBL 8158 edited by C Moler and Shavitt Lawrence Berkeley Laboratory University of California Berkeley 1978 1298 3 C Zirz and R Ahlrichs in Electron Correlation Proceedings of the Daresbury Study Weekend eds M F Guest and S Wilson Daresbury Laboratory Report DL SCI R14 1980 83 4 H J J van Dam J H van Lenthe and P J A Ruttink Exact size consistency of multi reference Mgller Plesset perturbation theory Int J Quant Chem 72 1999 549 558 5 H J J van Dam J H van Lenthe and P Pulay The size consistency of multi reference Mgller Plesset perturbation theory Mol Phys 93 1998 431 4309 doi 10 1080 002689798169122 6 K Wolinski H L Sellers and P Pulay Chem
23. 04976149 2 DOUBLET o 12 89 0 05925473 2 DOUBLET o 14 89 0 05097226 1 DOUBLET o 13 124 22210000 0 10693810 1 DOUBLET o 13 125 22200010 0 08035163 1 DOUBLET o 10 153 21210010 0 07987526 1 DOUBLET oO 15 153 SUM OF SQUARES OF VACUUM CSF CI COEFFICIENTS 0 909083899548E 00 SUM OF SQUARES OF 3 MAIN CI COEFFICIENTS 0 903281844886E 00 Description of the 2 A state TOTAL ENERGY 75 8837522007 FEO OO OR A kkk kk COEFFICIENT INTERNAL EXTERNAL EXTERNAL INTERNAL SPIN SPIN MOS CONF OCC FE OOO I A RA RR a aK kkk SPIN COUPLING REFERS TO REORDERED ORBITALS FO OO A A A aR a a kK 0 26477982 1 VACUUM 0 0 2 22210010 0 88348347 1 VACUUM 0 0 3 22121000 0 08640544 1 VACUUM 0 0 6 22120001 0 04120749 1 VACUUM 0 0 9 22111010 0 03104762 1 VACUUM 0 0 15 22022000 0 23849192 1 DOUBLET 0 9 87 22120000 0 04891462 1 DOUBLET 0 9 91 22021000 0 05841301 1 DOUBLET o 12 91 0 07064985 1 DOUBLET o 14 91 0 03000718 1 DOUBLET o 13 125 22200010 0 09148369 1 DOUBLET o 13 126 22111000 0 07828915 1 DOUBLET o 10 154 21121000 0 07660116 1 DOUBLET oO 15 154 SUM OF SQUARES OF VACUUM CSF CI COEFFICIENTS 0 863430795611E 00 SUM OF SQUARES OF 3 MAIN CI COEFFICIENTS 0 850807372991E 00 Taking as the criterion for inclusion a coefficient of 0 05 the final reference set to be employed is constructed based on both e the appropriate external MOs in the above coefficient lists which now must be assigned internal orbital statu
24. 1 0 0000000 20 4 0 86232544 0 0000000 21 3 0 86232544 0 0000000 22 2 0 86232544 0 0000000 23 1 1 23833149 0 0000000 24 4 1 44033091 0 0000000 25 3 1 44033091 0 0000000 26 2 1 44033091 0 0000000 27 1 3 13181655 0 0000000 6 DIRECT CI DATA INPUT 26 Assume that we wish to freeze the first 5 silicon inner shell orbitals requiring the following CORE and ACTIVE directives CORE 1 TO 5 END ACTIVE 6 TO 27 END To perform a 8 electron valence Cl calculation based on the SCF configuration would require the following CONF data CONF 2222 The complete data file for performing the SCF and subsequent Cl would then be as follows TITLE SIH4 6 31G DIRECT 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 CORE 1 TO 5 END ACTIVE 6 TO 27 END DIRECT 8 4 18 CONF 2222 ENTER Example 4 In this example we wish to perform a valence Cl calculation on the Nz molecule using a 4 31G basis While the molecular symmetry is Deon the symmetry adaptation and subsequent Cl will be conducted in the D point group The resolution of the Dan 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 6 DIRECT CI DATA INPUT 1 1 15 65951533 2 0000000 2 5 15 65474750 2 0000000 3 1 1 50615941 2 0000
25. 2 0000000 6 DIRECT CI DATA INPUT 21 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 1 1 45058658 0 0000000 25 1 3 78674557 0 0000000 Assume that we wish to freeze the five inner shell orbitals 1024301044 3 requiring the following data lines for the transformation CORE 1 TO 5 END ACTIVE 6 TO 25 END To perform an 8 electron valence Cl calculation involving the SCF configuration and two de generate 1e to 2e doubly excited configurations 57284920174 4 and 506030274 5 would require the following CONF data CONF 22 NOO DA 2022 2 2202 0 where there are six orbitals in the internal space the four doubly occupied valence SCF MOs and the two components of the 2e virtual orbital Note that no re ordering of the MOs is required since the 2e orbitals are the two lowest unoccupied VMOs The complete data file for performing the SCF and subsequent Cl would then be as follows 6 DIRECT CI DATA INPUT 22 TITLE PH3 6 31G VALENCE CI 3M 1R ZMAT P H 1 RPH H 1 RPH 2 THETA H 1 RPH 2 THETA 3 THETA 1 VARIABLES RPH 2 685 T
26. CONTENTS i Computing for Science CFS Ltd CCLRC Daresbury Laboratory Generalised Atomic and Molecular Electronic Structure System A _ 22z _ 3 _ __ __ lA zz GAMESS UK USER S GUIDE and REFERENCE MANUAL Version 8 0 June 2008 PART 5 INTEGRAL TRANSFORMATION DIRECT CI 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 2 Integral Transformation 1 3 Directives Controlling Integral Transformation 3 1 TRACCE 2 ee S PADS e oe ae oe es Un ete Goce ng he Re A hay nh oa ct et ab ete oe es YE Oo IBQURE insano rata Be oS and ene Be wa ap RS BY CORE opoe y wire arabe Die a eM ee ee ee ae a S g Ae N N N 4 RUNTYPE and Restarting the Transformation 5 CONTENTS 5 Direct Cl Calculations 5 Direct Cl Default CISD Calculations 004 5 1 1 Closed shell Systems 2 2 ee 5 1 2 Open shell Systems 2 2 e 6 Direct Cl Data Input 6 1 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 6 10 6 11 6 12 6 13 6 14 6 15 6 16 6 17 6 18 6 19 6 20 6 21 6 22 6 23 6 24 6 25 6 26 MECO 2 ga ade SE a ee Re ea Se aei Es a MG ae ee oe ee ee a gi Bers 6 hohe DREAD EHS a DESEO ERO CASGEN AM Ge eek ee we eS a ER a ER ad bee oY S SCREEN e lt Gee 25 ba weed eo bee be kha ede a de RES
27. E OPTIMIZE directive The data line RUNTYPE OPTIMIZE CI requests use of the Fletcher Powell FP optimiser with subsequent data used to characterize the direct Cl wavefunction to be employed in the energy calculation We illustrate such usage below for the case of a direct Cl calculation on the H2CO cation performing the calculation in several steps The first two steps carry out an RHF open shell geometry optimisation and the third the corresponding Cl optimisation Runs I and Il The SCF Optimisation TITLE H2C0 TZVP CLOSED SHELL STARTUP ZMATRIX ANGSTROM C 6 DIRECT CI DATA INPUT D 1 CO H 1 CH 2 HCO H 1 CH 2 HCO 3 180 0 VARIABLES co 1 203 CH 1 099 HCO 121 8 END BASIS TZVP ENTER 40 The first step is merely used to generate a suitable set of MOS for initiating the SCF geometry optimisation on the ion below Run III RESTART NEW TITLE H2C0 TZVP GEOMETRY OPTIMISATION SCF MULT 2 CHARGE 1 ZMATRIX ANGSTROM C D 1 CO H 1 CH 2 HCO H 1 CH 2 HCO 3 180 0 VARIABLES co 1 203 CH 1 099 HCO 121 8 END BASIS TZVP RUNTYPE OPTIMIZE OPEN 1 1 ENTER The Direct CI Calculation RESTART NEW TITLE H2C0 TZVP CI 1M AT RHF GEOMETRY MULT 2 CHARGE 1 ZMATRIX ANGSTROM C D 1 CO H 1 CH 2 HCO H 1 CH 2 HCO 3 180 0 VARIABLES CO 1 2063534 HESS CH 1 0876643 HESS HCO 117 8503752 HESS END BASIS TZVP RUNTYPE OPTIMIZE CI OPEN 1 1 192378 681619 709462 6 DIRECT CI DATA
28. FGEN 13142324 will produce a reference space consisting of all possible CSFs that can be generated by distribut ing 4 electrons in 4 MOs 6 17 CASGEN The CASGEN directive is on 1 card and is meant to produce a selective Complete Active Space reference function More than one CASGEN directives may be given Each one working with respect to the previous one The first word on the card is the textstring CASGEN A The following may be specified on the same card e DOC NDOC Al Specifies number of orbitals to remain doubly occupied e SDOC ND1 ND2 A I I As DOC but per symmetry The number of integers is equal to the number of representations DOC and SDOC are mutually exclusive e NORB NORBC A Number of orbitals taken into account e MAXEX MAX A 1 Maximum excitation level with respect to the current reference set e EXCIT A Use current excitation allowance for the newly generated configurations only Often sensible if restrictions like DOC are employed The default is the excitation mask 7b no internal excitations which is appropriate for a real CAS e NOSYM A Do not select the configurations on symmetry default e SYM A Select configurations on symmetry e SPIN A Select configurations on spin symmetry 6 DIRECT CI DATA INPUT 33 e PRINT A Print the generated reference configurations To generate a 4 orbital casscf wavefunction for water having the 3 inner orbitals doubly occupied in the reference fu
29. HETA 93 83 END BASIS 6 31G RUNTYPE CI CORE 1 TO 5 END ACTIVE 6 TO 25 END DIRECT 8 6 14 CONF 2222 2022 2202 ENTER ONO NOO Note that the SUPER OFF NOSYM constraint on the two electron integral file generated during the initial SCF will be automatically imposed by virtue of the nominated RUNTYPE The SUPER directive need only be presented in an initial SCF calculation that is driven under RUNTYP SCF control when the user plans to subsequently access this file in a separate Cl step under control of the BYPASS directive 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 Cy point group The resolution of the Coy into the Ca orbital species is given in Table 1 An examination of the SCF output reveals the following orbital analysis IRREP NO OF SYMMETRY ADAPTED BASIS FUNCTIONS 6 DIRECT CI DATA INPUT 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 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
30. Partially transformed 2 electron integrals are output to a dataset referred to as the Secondary Mainfile and the user may direct this dataset to any of the files EDO ED19 and MTO MT19 The LFN ED4 is used in default the SFILE directive described in Part 3 may be used to re assign Secondary Mainfile output Care should be taken if the user assigns the Secondary Mainfile to the same file as the Mainfile The Secondary Mainfile should not be allowed to overwrite the Mainfile except where the integrals sort has been completed in one pass NPASS 1 The Mainfile should not be over written if it is wished to perform a 2 index transformation of a Fock operator Generally it is advisable to route the Secondary Mainfile to a different file rather than to use the Mainfile e Transformed Integral file Fully transformed 2 electron integrals over the molecular orbitals are output to datasets referred to as the Transformed Integral File written in default to ED6 The user may re direct this dataset to any file EDO ED19 and MTO MT19 using the FFILE directive of Part 3 The Transformed Integral file may overwrite the Mainfile but should not be allowed to overwrite the Secondary Mainfile except where the second integrals sort has been completed in one pass NPASS2 1 Again overwriting of the Mainfile is not allowed if a Fock operator is to be transformed It is advisable to route the Transformed Integral file to a different file than is allocated to either the Mai
31. T direc tive below The internal MOs normally correspond to that set capable of producing a qualitatively correct wavefunction Notice that NINT 2 must be greater than or equal to NELEC NEXT specifies the number of external MOs Such MOs will be unoccupied in all reference CSF Single and double excitations from the internal to the external MOs cause the latter to contribute to the Cl expansion Notice that NINT NEXT must be less than or equal to the number of active MOs as specified under control of the ACTIVE directive If less than then some active MOs will not take part in the Cl and in the absence of a REORDER directive see below these will be the highest indexed active MOs 6 2 THRESH This directive consists of a single line read to variables TEXT C K using format A F l e TEXT should be set to the character string THRESH 6 DIRECT CI DATA INPUT 11 e C K The diagonalization is converged to a threshold T such that T C 10 K If K is not set it will be given the value 0 The lowest value to which T may be set is 1E 8 and this minimum will be selected if the user attempts to set a smaller T value The THRESH directive may be omitted when T will be set to 3E 4 Example THRESH 2 5 THRESH 2E 5 THRESH 0 00002 are equivalent causing T to be set to 2E 5 6 3 MAXCYC This directive consists of one line read to variables TEXT MAXC using format A l e TEXT should be set to the character string MAXCYC e MAXC s
32. TPR Similarly the parameters N 1 and N 2 control printing of those CSFs with one or two electrons in the external space respectively This directive may be omitted when the defaults NPR 100 and TPR 1E 7 will be taken Example VPRINT 50 0 02 VAC will cause all vacuum state coefficients to be printed All coefficients greater in absolute mag nitude than 0 02 will be printed unless there are more than 50 of these in which case only the largest 50 will be printed 6 11 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 6 DIRECT CI DATA INPUT 16 e NSPIN is used to specify the spin degeneracy of the Cl wavefunction 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 QUINTET SEXTET SEPTET OCTET and NONET to specify NSPIN The SPIN directive may be omitted when the program will set NSPIN to 1 or 2 if NELEC is even or odd respectively Example SPIN 4 SPIN QUARTET are equivalent the wavefunction will be four fold spin degenerate 6 12 NATORB The NATORB directive consists of a single dataline read to variables TEXT KSPACE KSPIN ATEXT using format A 21 A e TEXT should be set to the character string NATORB e KSPACE is an integer between 0 and 350 inclusive specifying the section number of the Dumpfile w
33. TRICT ecos a ae oe eee a ee he a ee eS Me II a e pa 2 Bia kb ae map me Gug cade wand ate d Ee es Spin Functions set e cee de ed a e Usine GB UrbilalS 4 eani sa eas A o ee ee e aoe a a E i FP Geometry Optimisations gt lt lt s ssc icedra ccad iria betra Calculating the YAy states of H2O oaoa aaa a O N N O 1 INTRODUCTION 1 1 Introduction In this chapter we turn our attention to the post Hartree Fock modules within GAMESS UK considering initially the integral transformation routines and associated data input and then the Direct Cl module Note that the transformation module acts not only as a precursor to Direct Cl but finds more widespread usage in for example both OVGF and TDA Green s function calculations and in the semi direct Table Cl module 2 Integral Transformation Before detailing the directives associated with the transformation of 1 and 2 electron integrals over atomic orbitals to the corresponding set over molecular orbitals we describe briefly the data sets used Note that the algorithm employed is basically that due to Yoshimine 2 In addition to the Mainfile Dumpfile and Scratchfile the following files will be used with the associated space requirements considered below e Sortfile A dataset assigned to SORT will be used as a scratchfile in sorting operations The space requirements are slightly more than twice the length of the Mainfile if a single pass sorting is adopted e Secondary Mainfile
34. USEHOLDER 1 8 6 23 Spin Functions It may be necessary for the user to understand the nature and order of spin functions associated with a given occupation pattern The program makes use of Yamanouchi Kotani genealogi cal spin functions the coupling order being such that higher indexed MOs are coupled before 6 DIRECT CI DATA INPUT 37 lower indexed MOs The ordering of the MOs is as defined using the ACTIVE directive of the transformation module see above Use the digits 0 and 1 to denote down and up spin coupling respectively Proceeding from the highest to the lowest indexed singly occupied MO write down the digitized representation of the possible spin functions the digits being written from left to right The resultant binary number defines the lexical ordering of the members of the spin canonical set the higher the number the higher the lexical index Example Consider 5 doubly occupied MOs coupled to a doublet The possible spin functions in digitized representation are in order of increasing lexical index e 10101 Spin function 1 e 10110 Spin function 2 e 11001 Spin function 3 e 11010 Spin function 4 e 11100 Spin function 5 6 24 Using GVB Orbitals In this section we briefly outline specification of a multi reference direct Cl based on a GVB 1 PP wavefunction Consider again the example of section 4 4 2 where a 4 pair GVB PP calculation on H2CO is described in which the two C H bonds and two C O orbitals are
35. ater sections TITLE H2CO 3 21G CISD DIRECT CI CALCULATION SUPER OFF NOSYM ZMATRIX ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 11 099 2 121 8 3 180 0 END RUNTYPE CI DIRECT 16 8 14 CONF 22222222 NATORB 11 O PRINT ENTER 5 DIRECT CI CALCULATIONS 9 5 1 2 Open shell Systems Let us now consider a Direct Cl calculation on the B state of H2CO again using default options available within the module A valid data sequence for performing such a calculation is shown below TITLE H2C0 2B2 3 21G DEFAULT CISD DIRECT CI OPTION MULT 2 CHARGE 1 ZMAT ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 11 099 2 121 8 3 180 0 END RUNTYPE CI ENTER As with the closed shell run above no explicit data is required to define the nature of the Cl calculation In practice the defaults adopted correspond to the following 1 The Cl will be based on the high spin open shell RHF calculation 2 The set of vectors used in the transformation will be the energy ordered SCF orbitals from section 5 of the Dumpfile the default section in the absence of section specification on the ENTER directive 3 The symmetry and spin of the Cl wavefunction will be deduced from the preceding SCF calculation i e a doublet Cl wavefunction of B2 symmetry corresponding to SPIN 2 4 The number of active electrons in the Cl will be set to be those involved in the SCF calculation i e 15 5 The reference configuration to be employed will be just t
36. correction factor as it is done in an iterative way and the correction shift may be different for different n 2 states Both the classical CEPA variants 0 1 2 for closed shell single determinant reference states and Multi Reference variants are provided This directive consists of a single data line read to the variables TEXT CEPA variant A A followed by optional options TEXT should specify the character string CEPA and CEPA variant is the text string specifying which CEPA or approximated Coupled Cluster approach is requested Options are e 0 CEPA 0 closed shell shift is correlation energy Cf 3 e 1 CEPA 1 closed shell recommended default Cf 3 e 2 CEPA 2 closed shell not invariant for mixing of occupied orbitals Cf 3 e MRO Straight multi reference variant of CEPA 0 e ACPF Averaged Coupled Pair Functional Shift is modified correlation energy 10 The correlation energy is default the difference with the variationally determined reference function e AQCC Averaged Quadratic Coupled Cluster Shift is modified correlation energy 9 The correlation energy is default the difference with the variationally determined reference function e MRD Multi Reference CEPA taking Variationally Included VI terms into account 7 6 DIRECT CI DATA INPUT 35 e MR1 Multi Reference CEPA taking VI and EPV terms into account 11 12 The EPV terms may be determined for the inactive space DOC
37. eference CSFs plus all single and double excitations internally and externally If the BINARY option is chosen then the full binary patterns of each equivalent octal integer must be given Example 5 EXCIT OCTAL 7 3 1 EXCIT BINARY 111 011 001 are equivalent It is possible to give different excitation patterns to different reference CSFs Example 6 EXCIT OCTAL 7 3 1 CONF EXCIT OCTAL 50 1 CONF The first set of reference CSFs are associated with the excitation pattern 7 3 1 while the second set have an excitation pattern 5 0 1 If the EXCIT directive is not invoked the excitation pattern will default to the setting 7 3 1 corresponding to all single and double excitations An EXCIT directive presented without parameters will cause restoration of the 7 3 1 default 6 DIRECT CI DATA INPUT Table 1 Resolution of the Cow Species into the C2 Species Orbital IRrep Esso Coy Sequence No o ay 1 r2 y2 Ta b 2 Ty ba 3 Oxy a2 4 Table 2 Resolution of the Doon Species into the D p Species Orbital IRrep Dooh Do Sequence No Og ag 1 g z2 y2 Tua b3u 2 Tuy boy 3 g zy big 4 Ou biu 5 u z2 y2 Tg z bag 6 Toy b3g 7 6 DIRECT CI DATA INPUT 20 6 14 CONF The CONF directive is used to specify the reference CSFs for the Cl expansion CONF may be presented more than once in the data input usually in conjunction with a different excitation pattern see the EXCIT directive above acti
38. he molecular orbitals must be doubly occupied and no partial occupied orbitals can be factored out in this manner The first line is read to TEXT NSECT in format A I e TEXT should be set to the character string CORE although ONELEC is also acceptable e NSECT specifies the section number on the Dumpfile where the transformed 1 electron integrals are to be placed If omitted the integrals are routed to section 466 If specified NSECT must lie between 1 and 350 Subsequent lines specify the frozen doubly occupied orbitals as a sequence of integers corre sponding to the MO ordering that came from the SCF module The sequence can be abbreviated using the string TO and is terminated by a line containing the string END in the first datafield Example 1 CORE 200 12 3456780910 END This example routes the transformed 1 electron integrals to section 200 of the Dumpfile Molec ular Orbitals 1 to 10 have been declared to be doubly occupied and frozen Example 2 CORE 200 1 TO 10 END 4 RUNTYPE AND RESTARTING THE TRANSFORMATION 5 This example shows the use of the string TO to shorten the data input and is equivalent to example 1 Example 3 CORE 1 TO 20 END This example assumes default routing of the transformed 1 electron integrals to section 466 MOs 1 to 20 will be frozen Example 4 The CORE directive can be omitted when no MOs will be frozen and the transformed 1 electron integrals will be routed to section 466 of
39. he open shell SCF configuration The internal space comprises the doubly plus singly occupied SCF orbitals with the external space comprising the SCF virtual orbitals All electrons will be deemed active in the Cl 6 The spinfree natural orbitals will be written to section 11 and the spin natural orbitals to section 12 of the Dumpfile The full data specification corresponding to the defaults generated from the above data file is shown below the role of each of the directives will again be described in later sections TITLE H2C0 2B2 3 21G CISD DIRECT CI CALCULATION SUPER OFF NOSYM CHARGE 1 MULT 2 6 DIRECT CI DATA INPUT ZMATRIX ANGSTROM C 0 1 1 203 H 11 099 2 121 8 H 1 1 099 2 121 8 3 180 0 END RUNTYPE CI OPEN 1 1 DIRECT 15 8 14 SPIN DOUBLET CONF 22222221 NATORB 11 12 PRINT ENTER 6 Direct CI Data Input 6 1 DIRECT 10 DIRECT consists of one line read to variables TEXT NELEC NINT NEXT using format A 31 e TEXT should be set to the character string DIRECT e NELEC specifies the number of electrons in the Cl calculation Notice that any inner shell electrons frozen out using the CORE directive of the transformation module should not be included NINT specifies the number of internal MOs These will be used to construct reference CSFs If a MO is not occupied in any reference CSF it should not ordinarily be classified as internal unless high levels of internal excitation are contemplated see EXCI
40. here the spin free NOs are to be placed If KSPACE 0 spin free NOs will not be routed to the Dumpfile e KSPIN is an integer between 0 and 350 inclusive specifying the section number of the Dumpfile where the spin NOs are to be placed If KSPIN 0 spin NOs will not be routed to the Dumpfile Notice that spin NOs will not be produced for singlet wavefunctions see SPIN directive above because they have an occupation number of zero in this case e ATEXT may be set to the character string PRINT when the NOs will be printed If ATEXT is omitted the NOs will not be sent to the output Example NATORB 12 14 PRINT The spin free and spin NOs are output to sections 12 and 14 respectively of the Dumpfile and routed to the output Note In the absence of the NATORB directive Version 6 3 onwards both spin free and if appropriate spin natural orbitals will be generated in default and routed to sections 10 and 11 of the Dumpfile This default thus corresponds to presenting the data line NATORB 10 11 PRINT so that the NATORB directive now need only be presented to override these defaults 6 DIRECT CI DATA INPUT 17 6 13 EXCIT This directive is used to define the excitation pattern allowed for a set of reference CSFs defined using the CONF directive see below The EXCIT directive may be used more than once in the data input to allow the user to specify different excitation patterns for different reference CSFs and normally consists of
41. igenvalues The matrix in this basis is solved for the eigenvector to obtain the trial vector Default The TRIAL directive may be omitted when the trial wavefunction will be selected from the reference space 6 9 JACDAV The JACDAV directive sets the controls of the Jacobi Davidson preconditioning method The syntax of this directive is JACDAV SUBDIR SUBDIR where JACDAV is a literal string acting as the directive initiator and SUBDIR is a valid subdirective Each subdirective consists of a literal string eventually followed by an integer real or string argument The supported subdirectives OFF ON SHIFT THRESH MAXCYC and PRINT are detailed below 1 OFF This is a literal string it switches the Jacobi Davidson preconditioner off By default the preconditioner is switched on 2 ON This is a literal string that switches the Jacobi Davidson preconditioner on This is the default 3 SHIFT DYNAMIC RSHIFT This subdirective sets the level shifter for the pre conditioner Here SHIFT is a literal string that is followed either by the literal string DYNAMIC or a real value for RSHIFT If a real value RSHIFT is supplied then that value will be used for the level shifter If the string DYNAMIC is supplied the level shifter will be automatically adjusted to force convergence The latter is the default 4 THRESH RTHRSH This subdirective sets the convergence threshold for the precon ditioner Here THRESH is a literal s
42. ill require three separate jobs each job looking to describe a specific state This is achieved in the case of the excited states through use of the TRIAL directive which identifies the particular state under investigation The data files for these three jobs are given below Job 2 Direct CI Treatment of the X A RESTART NEW TITLE x H20 TZVP DIFFUSE S P DIRECT CI 3M X1A1 SUPER OFF NOSYM BYPASS SCF ZMAT ANGSTROM 0 H 1 0 951 6 DIRECT CI DATA INPUT 44 H 10 951 2 104 5 END BASIS TZVP O TZVP H Ss 0 1 0 0 02 PO 1 0 0 02 END RUNTYPE CI CORE 1 END ACTIVE 2 TO 35 END DIRECT 8 8 26 CONF The following points should be noted e the SCF computation is BYPASS ed e the CORE and ACTIVE directives act to freeze and discard the two a MOs e the parameters on the DIRECT data line reflect the number of active electrons 8 number of internal orbitals 8 and the number of external orbitals 26 e in this job we are describing the X A so that the default diagonalisation controls will prove satisfactory Job 3 Direct CI Treatment of the 1 A state RESTART CI TITLE H20 TZVP DIFFUSE S P DIRECT CI 3M 1A1 SUPER OFF NOSYM BYPASS TRANSFORM 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 CORE 1 END 6 DIRECT CI DATA INPUT 45 ACTIVE 2 TO 35 END DIRECT 8 8 26 CONF 2222 2221 2212 SHIFT 0 TRIAL 2 1 0 VMIN
43. ion classes in MRCEPA 7 PULAY to select Ay in 4 in honour of his work on MRMP methods 6 ANDERSSON to select the Pulay zeroth order Hamiltonian but removing the exci tations within the reference space The name was chosen in honour of her work on CASPT2 8 in 4 in honour of his introduction of the exci Alternatively the model may be specified by a number where the units stand for 1 using a projector operator on the combined single and double excitation space 2 same as 1 but eliminating all single excitations 3 using projector operators onto the space of single and doubly excitations separately the decades stand for 0 no modification 1 removing all excitations within the reference space the sign stands for no modification removing all parts of the zeroth order Hamiltonian that connect different excitation classes i e RUTTINK can also be entered as 1 PULAY as 1 and ANDERSSON as 11 e TEXTD can be used to specify algorithm to orthogonalise the set of single and double excitations One can choose from SCHMIDT for modified Gramm Schmidt orthogo nalisation LOWDIN for orthogonalising by diagonalisation of the overlap matrix or HOUSEHOLDER for applying the House Holder method The floating point number M and the integer E specify the orthogonalisation accuracy M x 107 The default settings can be specified by either of the 2 following MP MP 2 RUTTINK HO
44. nction and doing a Single Double Cl from this reference function one specifies EXCIT OCTAL 7 3 1 CASGEN DOC 3 NORB 7 EXCIT 6 18 SCREEN This directive consists of one line whose first data field should contain the character string SCREEN If this directive is presented all reference configurations will be checked to see if they are of the same spin space symmetry as that of the required Cl wavefunction If they are not they will be eliminated from the reference space and will take no part in the excitation process Normally this directive may only be required if a REFGEN directive is presented since reference states of undesired properties will not usually be presented under control of the CONF directive 6 19 RESTRICT This directive provides a means of selectively eliminating CSFs from the Cl expansion by speci fying a minimum and maximum number of electrons which user specified sets of internal MOs may carry The first line should contain the character string RESTRICT in the first data field Subsequent lines are read to variables TEXT MIN MAX IORBS i i 1 m using for mat A 2I ml TEXT should be set to one of the character strings REF VAC N 1 or N 2 If REF is chosen the restrictions will apply to the reference CSFs should such a state fail to comply with the restriction applied it will take no part in the excitation process used to generate the Cl list and will most probably be used when REFGEN directives are used The character
45. nfile and the Secondary Mainfile The following approximation may prove useful in the considering the space requirements of the Secondary Mainfile Transformed Integral File and Sortfile Given NBASIS basis functions and NACT active orbitals and defining 3 DIRECTIVES CONTROLLING INTEGRAL TRANSFORMATION 2 Il NBASIS NBASIS 1 2 N NACT NACT 1 2 L Number of blocks in the Mainfile 1 block 512 words Then approximate space requirements are given by Secondary Mainfile S 2 L N M Transformed Integral file L N N M M Sortfile 2 L NPASS1 or S NPASS2 whichever is the greater 3 Directives Controlling Integral Transformation 3 1 TRACC The TRACC directive consists of a single line read to variables TEXT K using the format A e TEXT should be set to the character string TRACC e The integer K is used to compute the threshold factor ACC 107 and if the absolute value of a transformed 2 electron integral is less than ACC that integral will not be output to the Transformed Integral file A factor ACC1 ACC 107 is also computed and if the absolute value of a partially trans formed 2 electron integral is less than ACC1 that integral will not be output to the Secondary Mainfile The TRACC directive may be omitted when the default value K 10 is assumed The smaller the value of K the shorter will be the size of the Secondary and Transformed Integral file and the shorter the comp
46. ng on the reference CSFs The first line of the CONF directive is set to the character string CONF Each subsequent line specifies a reference CSF by giving the number of electrons 0 1 or 2 in each internal MO Thus each reference CSF is defined by NINT numbers the ordering of which should conform to the order of the internal MOs as specified under control of the ACTIVE directive If necessary CSF defining data may be carried over to further lines The CONF directive may also be used to determine a reference set based on CARDS data dumped by a preceeding CASSCF e g CONF CARDS see below We now illustrate CONF usage through a series of examples that will be subsequently used in Part 6 when describing configuration input for the Table Cl module Example 1 Consider performing a valence Cl calculation on the PH3 molecule using a 6 31G basis While the molecular symmetry is C3 the symmetry adaptation and subsequent direct 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 2a71e13a24a72e15a 1 or in the C symmetry representation 1a 22434 10 24a 5602070 2 1 1 79 93661395 2 0000000 2 1 7 48916431 2 0000000 3 1 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
47. nsist of the character string REFGEN in the first data field Subsequent lines are read using l format to paired integers IA and IC As many such lines as required may be presented lA and IC define annihilation and creation operators respectively which will be allowed to operate on the set of reference CSFs in existence at the time when REFGEN is called Thus at least one CONF directive must have been presented before using REFGEN and the IA IC integers refer to internal MOs The result of the annihilation creation process will be further reference CSFs whose excitation mask will be that of the most recently issued EXCIT directive Example 1 EXCIT OCTAL 7 3 1 CONF 2200 EXCIT OCTAL 5 REFGEN 13 24 will result in a further two reference states to give three in all of the form CSFS EXCITATION MASK ORIGIN 2200 7T 3 1 CONF directive 1210 5 REFGEN directive 2101 5 REFGEN directive Example 2 EXCIT OCTAL 7 3 1 CONF 2200 EXCIT OCTAL 5 REFGEN 1324 is equivalent to example 1 more than one IA IC pair may be given on a single line Example 3 EXCIT OCTAL 7 3 1 CONF 2200 EXCIT OCTAL 5 REFGEN 13 EXCIT OCTAL 3 REFGEN 24 6 DIRECT CI DATA INPUT 32 will produce 4 reference CSFs of the form CSFS EXCITATION MASK ORIGIN 3d CONF directive First REFGEN directive Second REFGEN directive Second REFGEN directive Example 4 EXCIT 7 3 1 CONF 2200 REFGEN 13142324 REFGEN 13142324 REFGEN 13142324 RE
48. on of the states of interest 3 based on the output from the initial Cls we augment the reference set to provide a quantitative description of the first three states We now consider various aspects of each job in turn Job 1 The SCF TITLE H20 TZVP DIFFUSE S P SCF SUPER OFF NOSYM 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 6 DIRECT CI DATA INPUT 1 0 0 02 END ENTER 42 The only point to note here is the use of the SUPER directive in suppressing skeletonisation given the user wishes to access this file in a subsequent Cl step under BYPASS control Jobs 2 4 The Initial CIs An examination of the SCF output reveals the following orbital analysis NO OF SYMMETRY ADAPTED and the following orbital assignments characterising the closed shell SCF configuration 1a72a71b33a 1b 56084959 35696939 72200122 58247942 50858566 02724259 04894440 05589681 06133571 20403420 22824210 53700802 56235022 58645643 66887228 74805617 07690608 88545053 92243836 12944874 20541910 34202871 39946430 697838310 OANDOABRWN KE NNNNNRPRPRP BRB BRB pl ps BPWNHRPOWDAN DA PUNEO BASIS FUNCTIONS WWRrRWNHPRPRPWRENPFPWWRPRRPENWHRENP WRB ONNNNFPRPRFPOOOVOCOCCOCOCOOOCAOC COO O OOOOO OOOO OOOO OOOO ON NNN N 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0000000 00000
49. pecifies the maximum number of iterative cycles to be carried out by the Davidson diagonalizer The directive may be omitted when MAXC will take the default value 50 6 4 SHIFT This directive consists of one line read to variables TEXT SHIF using format A F e TEXT should be set to the character string SHIFT e SHIF should be set to the desired value of the level shifter to be used in the Cl diagonal ization phase If the SHIFT directive is omitted the default SHIF 0 0 will be taken For ground states small values between 0 0 and 0 2 provide an optimal rate of convergence and usually there is little point in using the SHIFT directive For excited states the rate of convergence may sometimes be markedly improved by using a SHIF value of between 0 3 and 0 5 particularly if the ALTERNAT directive see below is used 6 5 ALTERNAT This directive consists of one line which should contain the character string ALTERNAT in the first data field If presented the directive causes the sign of the SHIF parameter see the SHIFT 6 DIRECT CI DATA INPUT 12 directive above to be altered at each iterative cycle of the Cl diagonalization and this may improve the convergence rate for excited states We do not recommend use of ALTERNAT except in cases where severe convergence problems are encountered 6 6 DIAGMODE This directive consists of a single data line read to variables TEXT ATEXT NDAVID using format 2A l e TEXT should be se
50. pense of an inferior rate of convergence If it is thought possible that the diagonalization procedure may not converge in one job ED5 must be retained to allow for restarts 5 DIRECT CI CALCULATIONS 7 e The P Sortfile A dataset normally assigned using the local file name LFN PSORT will be used as a scratchfile in a pre sort of the transformed molecular integrals The space requirements of the P Sortfile are about 1 5 times that of the Transformed integral File produced by the transformation module e the Sortfile A dataset normally assigned using the LFN SORT will be used as a scratchfile in a post sort of the transformed molecular integrals The maximum space requirements of the SORT FILE are about twice that of the Transformed integral File produced by the transformation module although this will be much reduced in high symmetry by an inverse factor approaching the order of the point group involved e Overflow File During the construction of the partial matrix elements a scratchfile allo cated as file ED8 may be required This is particularly likely to occur if a large number of reference CSFs are specified while its likelyhood is decreased if a large main memory allocation is used ED8 is usually of the order 1000 to 4000 blocks long if it is required Data input characterising the Cl calculation commences with the DIRECT data line and is typically followed by a sequence of directives terminated by presenting a valid Class 2 directi
51. put to the Transformed Integral File if all four orbitals are specified using the ACTIVE directive The first data field consists of the character string ACTIVE in the first data field Subsequent data lines are read to an array IACTIV I I 1 NACT using free l format The last data field presented should be the character string END Example 1 ACTIVE 10 11 12 13 14 15 16 17 18 20 END Molecular Orbitals 10 to 18 and 20 are made active and will be re indexed 1 to 10 respectively with the Direct Cl module for example referring to the orbitals in this re indexed convention Example 2 ACTIVE 10 TO 18 20 END This example shows the use of the string TO to abbreviate consecutive sequences of integers and is equivalent to example 1 3 DIRECTIVES CONTROLLING INTEGRAL TRANSFORMATION 4 3 4 CORE The CORE directive allows the user to e route a 2 index transformed Fock operator to a specific section of the Dumpfile e factor frozen doubly occupied orbitals such orbitals retain their double occupation in all configurations generated by the Direct Cl module for example into the Fock Operator F where F H 2 J R K R H denotes the usual one electron operator sum of kinetic and nuclear attraction R denotes the CORE shells density matrix and J and K are coulomb and exchange matrices constructed therefrom When employing the CORE directive in conjunction with factoring out molecular orbitals it must be remembered that t
52. s e the internal configurations specified in the ci vector output above 6 DIRECT CI DATA INPUT 48 This information is now provided directly in printing the final Cl vector it may also be derived from the print of the occupation patterns assuming the data line PRCONF 1 had been presented in the 3 reference job VACUUM STATES CONF 85 STATES 127 1 1 22220000 2 1 22210010 3 1 22121000 6 1 22120001 12 1 22200020 DOUBLET STATES CONF 112 STATES 1738 SYM 1 ONE CONF 38 EXTERNAL DIMENSION 12 87 1 22120000 89 2 22110010 90 2 12121000 91 1 22021000 SYM 2 TWO CONF 28 EXTERNAL DIMENSION 4 124 1 22210000 125 1 22200010 126 2 22111000 SYM 3 THREE CONF 28 EXTERNAL DIMENSION 8 153 2 21210010 154 2 21121000 The final 16 reference Cl jobs are shown below Since the ordering of the orbitals has now changed we repeat the transformation with a revised ACTIVE list designed to incorporate the external orbitals referenced in the coefficient lists above into the internal space Note that the external MO indexing within the Cl module does not take into account the inner shell frozen in the transformation adding one to the indices referenced in the Cl will provide the integers for specification in the revised ACTIVE data e 9 10 11 13 14 15 and 16 thus ACTIVE 2 TO 11 13 TO 16 12 17 TO 35 END Final CI Treatment of the X A RESTART NEW
53. subsequent Cl would then be as follows TITLE N2 4 31G ZMAT ANGS N N 1 NN VARIABLES NN 1 05 END BASIS 4 31G RUNTYPE CI CORE 1 2 END ACTIVE 3 TO 30 END DIRECT 10 7 21 CONF 22222 00 22202 02 22202 20 22220 20 22220 02 22211 11 NATORB 10 O PRINT ENTER Example 5 6 DIRECT CI DATA INPUT 29 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 Dg 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 20 1 1 32404002 0 0000000 21 5 1 45900204 0 0000000 Assume that we wish to freeze the nine Ca inner shell orbitals thus CORE 1 TO 9 END ACTIVE 10 TO 21 END To perform a 4 electron valence Cl calculation based on the SCF configuration would simply require the following CONF data
54. t to the character string DIAGMODE e ATEXT should be set to one of the character strings EMIN VMIN or LOCK EMIN causes the program to unconditionally minimize the total energy and is normally the best option for ground states VMIN causes the program to minimize the variance sum of squares of residuals in the secular problem and is usually the best option if convergence to an excited state is required LOCK causes the program to seek a solution to the Cl problem looking most like the trial wavefunction see TRIAL directive below and is therefore another way of trying to converge onto an excited state e NDAVID specifies the maximum size of the sub space to be used in the Davidson di agonalization procedure If omitted the default NDAVID 50 will be taken It may be necessary to use a smaller sub space in order to limit the size of ED5 as explained above The largest possible sub space is 50 attempts to set NDAVID larger than 50 will cause the program to use this maximum The DIAGMODE directive may be omitted when the defaults ATEXT EMIN and NDAVID 50 will be taken It is possible to omit the TEXT parameter of this directive Example DIAGMODE VMIN 50 VMIN 50 VMIN all have an equivalent effect causing the variance minimization option to be selected with a maximum sub space of 50 6 7 PRINTVAR This directive consists of a single line whose first field should contain one of the character strings PRINTVAR or VARPRINT The EMIN
55. tring and RTHRSH is a real value By default the threshold is set to half the Cl threshold at time of calling JACDAV 6 DIRECT CI DATA INPUT 15 5 MAXCYC IMAXC This subdirective sets the maximum number of cycles for the preconditioner Here MAXCYC is a literal string and IMAXC is an integer value By default IMAXC is set to 100 6 PRINT This subdirective controls the output level of the preconditioner Here PRINT is a literal string Every time this string is supplied the output level is increased causing the preconditioner to generate a more detailed output By default the printing is off NOTE Because the default threshold is determined from the Cl threshold at time of calling JACDAV interchanging the order of the THRESH directive and the JACDAV directive may lead to different convergence behaviour 6 10 VPRINT This directive consists of a single data line read to variables TEXT NPR TPR ATEXT BTEXT CTEXT using format A I F 3A and is used to control printing of the Cl wavefunction e TEXT should be set to the character string VPRINT e NPR specifies the maximum number of Cl coefficients to be printed e Cl coefficients less than TPR in absolute magnitude will not be printed e ATEXT BTEXT CTEXT may each be set to one of the character strings VAC N 1 or N 2 If VAC is specified Cl coefficients of all the vacuum states those with no electrons in the external MO space will be printed irrespective of the values of NPR and
56. uded 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 employed Augmenting the reference set to improve the description of the two excited states follows straightforwardly from the statistics below Description of the X A state TOTAL ENERGY 76 2725934815 SO A k kk kkk kkk kk COEFFICIENT INTERNAL EXTERNAL EXTERNAL INTERNAL SPIN SPIN MOS CONF OCC FEO OO A A Rk a a kk SPIN COUPLING REFERS TO REORDERED ORBITALS GE OOO A A a a a a kak 0 97419222 1 VACUUM 0 0 1 22220000 0 03737671 1 SINGLET 13 13 201 22200000 0 03086972 1 SINGLET 10 15 202 20220000 0 03143494 1 SINGLET 15 15 202 0 03223099 1 SINGLET 14 15 218 21120000 SUM OF SQUARES OF VACUUM CSF CI COEFFICIENTS 0 949366016991E 00 SUM OF SQUARES OF 3 MAIN CI COEFFICIENTS 0 949241228068E 00 Description of the 1 4 state TOTAL ENERGY 75 9018772961 FOO A kkk RK kk kk COEFFICIENT INTERNAL EXTERNAL EXTERNAL INTERNAL SPIN SPIN MOS CONF OCC FO OG A A aR a a kkk SPIN COUPLING REFERS TO REORDERED ORBITALS FO OO OR RA a RK a aK Kk 6 DIRECT CI DATA INPUT 47 0 91631377 1 VACUUM 0 0 2 22210010 0 25186711 1 VACUUM 0 0 3 22121000 0 05309578 1 VACUUM 0 0 12 22200020 0 08694410 1 DOUBLET 0 9 87 22120000 0 03713591 1 DOUBLET 0 9 88 12210010 0 04438742 1 DOUBLET o 12 89 22110010 0 04467943 1 DOUBLET o 14 89 0 03994355 2 DOUBLET 0 9 89 0
57. utation time in the second phase of the 4 index transformation 3 2 PASS The PASS directive consists of a single dataline read to variables TEXT NPASS1 NPASS2 using format A 21 e TEXT should be set to the character string PASS e NPASS1 is an integer specifying the minimum number of passes of the Mainfile in the first phase of the 4 index transformation e NPASS2 is an integer specifying the minimum number of passes of the Secondary Mainfile in the second phase of the 4 index transformation 3 DIRECTIVES CONTROLLING INTEGRAL TRANSFORMATION 3 In the absence of a PASS directive the module will calculate the minimum number of passes required which will depend on the basis set size and the amount of memory available to the module The user should note the following e The size of the Sortfile is usually inversely proportional to either NPASS1 or NPASS2 To reduce the size of the Sortfile multi passing of the Mainfile and Secondary Mainfile must be employed e The program forms a dump enabling a restart of the 4 index transformation process at the end of each pass The more passes the shorter time interval between dumps Example PASS 3 2 Specifies a 3 and 2 pass sort of the Mainfile and Secondary Mainfile respectively 3 3 ACTIVE This directive specifies those members of the molecular orbital set which are deemed active in the integral transformation so that integrals of the form lt ij kl gt will be computed and out
58. ve such as VECTORS or ENTER The directives may be presented in any order although in some cases directives are inter related and care should be taken when presenting them since the order in which the directives are presented is in such cases often significant this is particularly the case for the EXCIT CONF REFGEN and CASGEN directives Before describing each of the directives in detail we cater for those users who wish to fast forward through the directive descriptions by outlining how to perform default single reference CISD calculations This provides a set of default attributes that bypasses the requirement for explicit data specification while of somewhat limited applicability it does provide a starting point for users and a route to subsequent more extensive calculations 5 1 Direct CI Default CISD Calculations In order to simplify the process of configuration specification and data preparation the Direct Cl module now provides a set of default options that require little or no data input To illustrate this default working of the module we consider below a number of example calculations based on those that will be described in more detail in the subsequent sections 5 1 1 Closed shell Systems A Direct Cl calculation is to performed on the formaldehyde molecule Given the following data sequence TITLE H2C0 3 21G DEFAULT DIRECT CI CISD OPTION ZMAT ANGSTROM C 0 1 1 203 5 DIRECT CI CALCULATIONS 8 H

Download Pdf Manuals

Related Search

Related Contents

CABEZAL GBR POWER4100 MANUAL  the Inspiration Quick Start tutorial  User Guide for Cisco Unity Connection  PDFマニュアル  HERO3+ Silver  平成14年(ワ)第11150号 意匠権侵害差止等請求事件 口頭    4th July 2007 to 6th August 2007  KS – 7715/7717/7515/7517 Fan Free Touch Terminal User's Manual    

Copyright © All rights reserved. Failed to retrieve file