Home

ACES II Release 2.5.0 User Manual

Contents

1. 24 6 10 12 0UT 000 DUMP 000 1ELGRAD 00 0 2 02 a 7 25 7 File Formats 26 sl NEES 26 TEL Wile anatomia et as Dl a tee 26 Pole ESCAPES e sides Slds alga AS AO 27 TLS lr AAA 28 7 1 4 Molecular orientation Ze e ds Ne bie A a A 29 7 1 5 Dummy and ghost atoms Sa tad 2 erated 2 SA eee eo 30 7 1 6 Cartesian coordinates 2 ow a oe Bee ee Bo Be eo bk 30 7 1 7 Internal coordinates sock ate Bede A oe eS EIR SAM 31 7 1 8 Z matrix A IAE 32 E FACES namelist EE 34 7 1 10 Line item basis ECP definitions ENEE ENEE EE 36 a GENBAS ZMAT BAS ee LS a da EE 36 Go ECPDATA can a a il A Se een Zeen a Ae eisen 2 37 TAs GUESS A de heels A a EE el AR ZA 39 8 Keywords 41 Bl SPACES Wameligts lt lt cur 656 4 wit E AA ee AA E 41 H e EE 42 8 1 2 System molecular control 2 rasa 43 bie A ER 44 RIA E e A a E ie E R mi R ee 0 44 O lt A A EE 45 O ASIA A e o rl ps ne ok 45 A 5 2 ER A e ie e ee A 46 E a a EE EES 47 Bal SOS General aero as aes T ep ee hE tee coal oh poets 48 8 1 10 SCR Orbital control h4 k ar a ee a e el Eo Ee eS 48 8 1 11 SCF iteration control is se aii Sete ae tee EE pi ey 50 8 1 12 SCF reference adjustments eE 2 24 a Go eee GS ee Bot ee e 51 8 1 13 Post SCF file options ooa aa 54 8 1 14 Post SCF calculations sp ia Gerke E et rr die 56 8 1 15 Excited states general EE 57 8 1 16 Excited states properties et EL e er ait Er as 58 8 1 17 Excited states affinities ete 4 2 a 0S aaa e NET
2. 12 6 Methods for calculating excitation energies 12 7 Methods for calculating electron attachment energies 12 8 Time dependent Hartree Fock methods 12 9 HF DFT method iz ls eta tA ase eek 12 10Integral packages ai ae ah a Ae age a A Other Keywords A 1 Experimental obsolete and unused keywords A 2 Kohn Sham DFT namelists PED SNOT AL ao eel sl as R scent ee PEP DS DEER RA B Standard Basis Sets and ECPs B 1 Basis sets in GENBAS B 1 1 STO 2G 3G 6G 3G 0 B 1 2 MINI MINI SCALED MIDI MIDI 87 87 88 88 89 89 91 91 92 94 95 95 96 96 97 97 98 99 101 101 101 102 102 102 102 103 103 103 103 104 B13 321G E E EE 107 B 1 4 4 31G EE si e A E A IE 108 B 1 5 6 31G 6 31G 6 31G 6 31 G 6 31 G 6 3144 G etc 109 B 1 6 6 311G 6 311G 6 311G 6 311 G 6 3114 G ete 109 B 1 7 SV DZ TZ SVP DZP CHIPMAN A hos Gk RRA ie rt A 110 B 1 8 CC PVxZ PCVxZ PWCVxZ PVxZ_DK AUG CC PVxZ etc 111 B 1 9 GAMESS VTZ PVTZ e ads ee Soot EEN 112 B 1 10 AHLRICHS VDZ PVDZ VTZ TZV POL 113 B 1 11 PARPRIDG EAL 2 3 4 alle Ghat gie hue hoes ea A A 113 BALI Ae ste a e hr Fel tra Mass ME E hee Fan 113 B 1 13 SADLEJ PVTZ previously PBS 206000 Bata bo Bay A 113 B 1 14 WACHTERSEFS 0 a e A 114 B 1 15 ROOS lt ADZP lt ATZP 2 AEN AR e A e Ee 114 B 1 16 NASA AMES 10d amp a RO ee EO oS 115 B 1 17 B
3. 9 geit 1 27 13g 1 6 6 Gm 1 27 Wetz 1 6 1 6 DTIE 9 Lo TZ TZ CL CL DTTE 9 H Bet Bet 91 91 Ty 44 1UE DTE 9 D LT 1 7 Teig 1 6 1 6 z 9 PAS 16 9 e Les Les 1 6 1 6 Z O F TE 9 TZ LT LT gl el z ot E e Les Les 1 6 1 6 Z D 1 I 61 Lei I G Wei S xD TE 9 1 7 I 61 Lei Lo Wei Z DTE 9 6 Z DTE 9 Di 6 e e el e e 6 6 z DTE P e 6 6 Z ASDEZ Y e 6 6 g ASDTZ D LT Wel We el el z C GE Cp e L L T i z DH TZE Z E DIZ E D 1 6 I 61 6 6 Z a L 6 6 z DTZ E D D I 61 e D D I cT 61 D e 7 IGN Z IAIN D TI e L 6 0 6 9 g Z dATVOS ININ 0 6 9 g 7 ININ D Wen Wen o g Deors 6 6 o 0 6 6 G g DE OLS 6 6 o o D7OLS vo a uv to s a fis 1v on vw AN o In o a ag T aH H T s dg St dz SZ ST SVENAD Ul SJUSUIOTO Sp YSNoIY ST 10 SUOCTJOUNJ sIseq OY P9I9BAFJUOI Jo JoquINU 201 1 AQL 119 SOL MUTZL2Ad DO DE oct E oct a M ZOAdDO e e x EY chef ST 02 eg MUTZLAdDO S SCH 7 Yor ATANT OY 6 e e G SE Wei ZSA0d 00 DAV D D S Vs S T ZOAOd D00 DNV z e S 08 6ET gt ZLAOA DO DAW Wes 24 0T 69 z 9 ZA AO TF DO DAV E e E E eet PAFOO DAY z We 06 997 2 GOT GAN 06 047 Be y Be ZOAd OO DNV e e e CG 357981 ont ES ZLAFOO DNV S SC 2 6 l lt
4. 30 7 1 7 Internal coordinates The specification by internal coordinates is known as the Z matrix Centers of the nuclei are expressed relative to previously defined centers by means of distances and angles The specification includes a length a bond angle and a dihedral angle The number associated with each atom is governed by its position in the Z matrix The essentials of Z matrix construction can be illustrated by considering a Z matrix for a system of four atoms ABCD not linked in any particular order Arbitrary ABCD molecule gt B C 2 CAB D 2 DCB 1 TAU B1A C1A D 23 C The first line in the Z matrix contains the atomic symbol of one of the atoms say A The second line specifies the position of a second atom say B relative to the first atom Suppose that Bis a distance AB from A The second line then contains the atomic symbol B followed by the number 1 A is atom number 1 and a parameter label AB B 1 AB For the specification of the third atom a distance and an angle are needed We may use the distance between atoms A and C and the angle CAB or we may use the distance between atoms B and C and the angle CBA In the first case the third line would have the form C 1 AC 2 CAB while in the second case it would have the form C 2 BC 1 CBA Finally there is a line specifying the position of D relative to the other atoms This line must contain a distance a bond angle and a dihedral angle and could
5. JODA will orient water in the yz plane In this example the y axis in the Ca frame should be the z axis in the C frame so SUBGRPAXIS Z When the true Abelian subgroup is either C2 or Dan the orientation is not well defined and it might be necessary to run xjoda twice If SUBGROUP 0 in the first pass then the reference orientation for the true Abelian subgroup can be determined and the appropriate value of SUBGRPAXIS can be selected NOREORI switch OFF Forces the program not to change the internal orientation of the molecule specified by the input Cartesian coordinates To function correctly this keyword also requires turning off symmetry SYMMETRY 0FPF 8 1 3 System debug PRINT integer 0 Controls the amount of printing in all member executables except JODA A value of 1 will produce a modest amount of additional output over the default value which includes some useful information such as SCF eigenvectors JODA PRINT integer 0 Controls the amount of debug printing performed by JODA The higher the value the more information is printed Values of 25 or higher generally do not produce anything of interest to the general user and values greater than 999 will dump the core vector to disk 8 1 4 I O subsystem FILE RECSIZ special 1W Sets the length of the physical records used in direct access file I O This value should always be chosen as a multiple of 512 bytes and is parsed like the MEMORY_SIZE key word
6. The well tempered formula is a modification of the even tempered formula originally recommended by Reeves and explored by Ruendenberg and coworkers for complete references see Chem Phys Lett 212 260 1993 In the well tempered formula there are only four parameters which need to be optimized for each atom in the periodic table The values of exponential parameters are shared between the angular symmetries thus reducing the effort required in the optimization of basis sets Usually the higher angular subspaces use only a subset of the complete pool of exponential parameters defined by N in consequence the well tempered basis sets contain a hidden parameter the pattern of sharing the exponents between the s p d and f symmetries Savings in the computing time may be achieved especially in the evaluation of energy derivatives with the integral codes which utilize the sharing of the exponents between the atomic angular symmetries e g GAMESS HONDO or Gaussian KS Users can split off individual diffuse primitives for added flexibility 1 S Huzinaga B Miguel Chem Phys Lett 175 289 1990 2 S Huzinaga M Klobukowski Chem Phys Lett 212 260 1993 B 1 13 SADLEJ PVTZ previously PBS These are double zeta plus diffuse basis sets developed by Sadlej for the calculation of electrical properties They seem to do a good job of predicting dipole moments and polarizabilities and are also useful in excited state calcu
7. cd while test jodadone eq 0 do PRUN xgemini s xvmol amp amp xvmol2ja PRUN xp_vscf PRUN xp_scfgrd PRUN xgemini s xjoda cd rootdir jodadone xa2proc jareq i JODADONE 1 tail 1 cd done 94 11 Troubleshooting 11 1 Common mistakes The following tips should help users to detect and avoid errors For large calculations it is strongly recommended that users run with trial input files small basis set high convergence tolerances etc before running the actual system e The ACES2 namelist is not terminated properly If it begins with an open parenthesis then it must end with a close parenthesis otherwise it must end with a blank line Even if the namelist is terminated with a close parenthesis there must be a blank line after the namelist and the end of file mark e The BASIS SPECIAL option is used but the order of the basis sets does not corre spond to the order of atoms in the Z matrix The code does not check this and will not crash For example one is allowed to put a C basis set on a CL atom In fact ghost atoms would not be possible without this feature e All atomic symbols should be in upper case For example Cl has to be entered as CL Actually this is not true anymore but older versions of the code would simply give incorrect results If any user finds case sensitivity in the coordinate matrix parsing aces2 qtp ufl edu should be notified with the failing ZMAT e L
8. 5 27 8 jarec JAREC and its quiet variant JAREQ will show the formatted contents of a JOBARC record It requires three arguments data type record name case sensitive and dimensions Data types can be i integer d double f float r real ad array of doubles and ai array of integers The first four types are all one dimensional vectors and ad and ai are two dimensional arrays For the arrays the dimension string can be of the form R C RxC 20 and RXC in which R is the number of rows and C is the number of columns The following results were taken after an H202 DZP SCF geometry optimization gt xa2proc jareq i NATOMS 1 4 gt xa2proc jareq d TOTENERG 1 0 150816196754E 03 gt xa2proc jareq ad GRADIENT 3x4 0 000079292458 0 000079292458 O 000034723378 0 000034723378 O 000006762679 0 000006762679 O 000088780697 0 000088780697 O 000000000000 O 000000000000 O 000000000000 O 000000000000 5 27 9 xyz This module prints the Cartesian coordinates of the current geometry This could be used in a script that automatically runs a vibrational frequency calculation after a geometry optimization 5 27 10 test The TEST module is used by the automated regression test suite Its use is beyond the scope of this manual but interested users are encouraged to examine the files in the ACES II test directory 5 28 gemini xgemini is used to manage scratch directories in parallel ca
9. maximum memory to allocate Most member executables will not allocate more than the amount specified but some will attempt to allocate extra bits on the order of a few megabytes that are of little consequence compared to the main memory heap There is special parsing logic for the corresponding value string The tail of the string is checked for case insensitive units and optional prefixes Units can be W integer words or B bytes and prefixes can be K M or G kilo mega and giga respectively The number part is read with the atof C function NOTE The prefixes scale the units by 21 1024 not 1000 and the default unit is integer words 4 bytes for 32 bit binaries and 8 bytes for 64 bit binaries RESTART switch ON 42 Controls the creation updating and reading of a checkpoint directory Currently only coarse grain restarts are available which means geometry optimizations and finite difference calculations can be restarted from the last checkpoint Fine grain restarts are not available yet but will also checkpoint SCF T and Lambda iterations GRAD_CALC handle AUTO Specifies whether to calculate a gradient analytically or numerically This should be used only to override analytical gradients since the program will attempt to use them whenever possible DERIV_LEV handle 1 Specifies whether or not derivatives of the energy are to be calculated and if so then whether first or second NONE De
10. 103 4572 1995 7 A Wilson T van Mourik T H Dunning Jr J Mol Struct THEOCHEM 388 339 1997 8 T van Mourik A K Wilson T H Dunning Jr Mol Phys 96 529 1999 9 A K Wilson D E Woon K A Peterson T H Dunning Jr J Chem Phys 110 7667 1999 10 T van Mourik T H Dunning Jr Int J Quantum Chem 76 205 2000 11 J Koput K A Peterson J Phys Chem A 106 9595 2002 12 K A Peterson T H Dunning Jr J Chem Phys 117 10548 2002 13 W A de Jong R J Harrison D A Dixon J Chem Phys to be published B 1 9 GAMESS VTZ PVTZ This basis appears in the program GAMESS and is composed of bits and pieces from a variety of sources lt uses three functions to represent each valence AO and two functions for each inner shell AO 1 T H Dunning Jr J Chem Phys 55 716 1971 2 A D McClean G S Chandler J Chem Phys 72 5639 1980 112 B 1 10 AHLRICHS VDZ PVDZ VTZ TZV POL These basis sets were obtained by optimizing the exponents and contraction coefficients in groundstate ROHF calculations 1 A Schafer H Horn R Ahlrichs J Chem Phys 97 2571 1992 2 A Sch fer C Huber R Ahlrichs J Chem Phys 100 5829 1994 B 1 11 PARTRIDGE 1 2 3 4 These are groundstate RHF energy optimized uncontracted basis sets 1 H Partridge J Chem Phys 87 6643 1987 2 H Partridge J Chem Phys 90 1043 1989 B 1 12 WTBS
11. 1M F Guest and V R Saunders Mol Phys 28 819 1974 73 FeC14 sextet FE 0 0000 0 0000 0 0000 CL 0 0000 3 4826 2 2358 CL 0 0000 3 4826 2 2358 CL 3 4826 0 0000 2 2358 CL 3 4826 0 0000 2 2358 ACES2 REF ROHF PRINT 1 CALC SCF MULT 6 CHARGE 1 BASIS SPECIAL UNITS BOHR COORDINATES CARTESIAN SPHERICAL ON ECP 0N DAMP_TYP DAVIDSON DAMP_TOL 5 LSHF_A1 50 LSHF_B1 50 OCCUPATION 7 6 6 6 5 5 5 5 FE SBKJC CL SBKJC CL SBKJC CL SBKJC CL SBKJC FE SBKJC CL SBKJC CL SBKJC CL SBKJC CL SBKJC Another tip for ROHF convergence which the developers strongly recommend is starting the SCF from UHF orbitals from a closely related closed shell system UHF is necessary since ROHF reads both o and 7 orbitals 9 1 9 Hartree Fock stability analysis The Hartree Fock procedure at convergence guarantees the resulting wavefunction is a stationary point in the space of orbital rotations mixing of occupied and virtual orbitals In the majority of cases this stationary point is also a minimum De all orbital rotations increase the energy but the SCF procedure does not guarantee this In some cases the second derivative of the energy with respect to one or more orbital rotations is zero or neg ative indicating rotations which will leave the energy unchanged or lower it The ACES II program system has the ability to test RHF and UHF wavefunctions for some of the most common instabilities controlled by the HFSTA
12. 4 Namelists Any number of blank lines may pad the namelists In general a namelist is delimited by an asterisk immediately followed by the case sensitive name of the namelist ACES2 INTGRT Only the ACES2 namelist is required for every calculation Some programs or features recognize other namelists but the rules for parsing the strings are specific for each list Rules for parsing the ACES2 namelist can be found further down 5 Line item basis set and ECP data assignments If either BASIS SPECIAL the default or ECP ON off by default then there must be one blank line between the ACES2 namelist and the assignment block If both blocks are required then there must be one blank line between the ACES2 namelist and the basis set block followed by another blank line and the ECP block 6 Footer Any unrecognized text non namelist following the internal coordinate definitions is ignored If there is no footer then ZMAT must terminate with a blank line 7 1 2 Examples Line 1 RHF CCSD property calc of C6 in DZP Line 2 Line 3 Line 4 Line 5 Line 6 Line 7 Line 8 Line 9 Line 10 Line 11 Line 12 Line 13 Line 14 Line 15 Line 16 ACES2 CALC CCSD PROPS FIRST_ORDER BASIS DZP Line 17 CO CO CO CO CO CH G i Ge KD K K K K KD SR RE H KA KA k KK KA KA D D e e YS e NX O MH W HAHAHAHA Ps Il dl 1 33 90 60 Herds o 2d Line 1 Line 2 Line 3 Line 4 Line 5 Line 6 Line 7
13. 51 LSHF_B1 51 MAKERHF 54 VAN STEP 65 MEMORY SIZE 42 44 MULTIPLICITY 45 NEGEVAL 66 NEWVRT 49 NONHF 47 NOREORI 44 NT2EOMEE 58 NTOP_TAMP 56 OCCUPATION 49 49 OPT MAXCYC 66 OPT_METHOD 65 ORBITALS 47 ORDER RLE see CC EXPORDER 129 PERT_ORB 47 POINTS 68 POLYRATE 103 PRINT 44 PROGRAM 57 PROPS 63 PRP INTS 64 PSI 103 QRHF GENERAL 52 QRHF ORBITAL 53 QRHF SPIN 53 RAMAN 67 RDO 103 REFERENCE 47 RESET_FLAGS 103 RESRAMAN 62 RESTART 42 RLE see CC_EXTRAPOL ROT_EVEC 51 RPP see SCF EXTRAP RPP LATEST see SCF EXPSTAR RPP_ORDER see SCF EXPORDER SAVE_INTS 54 SCF_CONV 50 SCF EXPORDER 50 SCF_EXPSTAR 51 SCF_EXTRAP 50 SCF_MAXCYC 50 SCF PRINT 48 SCF_TYPE 48 103 SINGLE_STORE 54 SOLVENT 103 SPHERICAL 45 STP SIZ CTL 65 SUBGROUP 43 SUBGRPAXIS 43 SYMMETRY 43 TAMP_SUM 56 KS 103 TDHF 63 TRANS INV 67 TREAT PERT 64 TURBOMOLE 103 UNITS 45 UNO_CHARGE 53 UNO_MULT 54 UNO_REF 53 VIBRATION 67 VTRAN 55 XFIELD 64 XFORM_TOL 56 YFIELD 64 ZETA_CONV 62 ZETA MAXCYC 62 ZETA_TYPE 62 ZFIELD 64 INTGRT CUTOFF 105 ENEGRID 104 ENETYPE 105 EXACT_EX 104 FUNC 105 FUZZYITER 104 KSPOT 105 NUMACC 104 PARTPOLY 104 PARTTYP 104 POTGRID 104 POTRADPTS 104 POTTYPE 105 RADLIMIT 104 RADSCAL 104 RADTYP 104 TDKS 104 VSCF 130
14. Analytical Hessians e Independent particle models include RHF UHF and ROHF 13 4 Quickstart Guide At the bare minimum the user must provide an input file named ZMAT and a basis set file named GENBAS The main executable xaces2 is a driver program for other batch based member executables such as xjoda xvscf and xvcc Since xaces2 uses the system function to run these programs the executable path must be set by the user s regular shell environment For example if xaces2 does not appear in the default login environment then running xaces2 directly will likely result in error since the first attempt to run xjoda will fail with xjoda not found The ACES II program system uses many storage files and the developers initially rec ommend running the program in a directory that contains only ZMAT and GENBAS As expe rience with the program increases users will become comfortable with naming files that do not conflict with those used by ACES II gt ls GENBAS gt cat lt lt EOF gt ZMAT H2 H HIR R 0 7356 ACES2 BASIS DZP CALC SCF EOF gt xaces2 gt out The file named out now contains the RHF SCF results of H in the DZP basis set provided GENBAS contains the DZP basis set for H Since SCF is the default calculation level the CALC keyword is not necessary The minimum requirements for ZMAT are 1 line title molecular system in internal or Cartesian coordinates blank line ACES2 namelist declarin
15. Jensen P J rgensen J Olsen and P R Taylor HONDO GAMESS M W Schmidt K K Baldridge J A Boatz S T Elbert M S Gordon J J Jensen S Koseki N Matsunaga K A Nguyen S Su T L Windus M Dupuis J A Montgomery catalog should use the following citation in addition to the ACES II citation 1 2 Basis sets were obtained from the Extensible Computational Chemistry Envi ronment Basis Set Database Version 1 0 as developed and distributed by the Molecular Science Computing Facility Environmental and Molecular Sciences Laboratory which is part of the Pacific Northwest Laboratory P O Box 999 Richland Washington 99352 USA and funded by the U S Department of En ergy The Pacific Northwest Laboratory is a multi program laboratory operated by Battelle Memorial Institute for the U S Department Energy under contract DE AC06 76RLO 1830 Contact Karen Schuchardt for further information Specific authors Two determinant CCSD TD CCSD energies P G Szalay and A Balkova TD CCSD analytical derivatives P G Szalay Dropped molecular orbitals in analytical derivative calculations for RHF UHF and ROHF references K K Baeck Equation of motion CCSD calculation of dynamic polarizabilities including parti tioned scheme J F Stanton S A Perera and M Nooijen e Equation of motion CCSD calculation of NMR spin spin coupling constants including partitioned scheme S A Perera and M Nooijen e Partitione
16. SCF reference wavefunctions For KS SCF the numerical integrator xintgrt is used to calculate the functional energy and xintpack is used for OEP calculations The parallel HF SCF program xp_vscf should be used for GAMESS direct integrals FOCK AO DIRECT ON INTEGRALS GAMESS 5 9 dirmp2 and p dirmp2 xdirmp2 calculates the MBPT 2 energy using the GAMESS direct integral package xp_dirmp2 does this in parallel 5 10 vtran xvtran is responsible for the 4 index AO MO integral transformations after the SCF calculation 5 11 tdhf xtdhf performs time dependent Hartree Fock calculations 16 5 12 intpre xintprc sorts the two electron integrals into five basic types OOOO OOOV OOVV OVVV and VVVV in which O and V stand for occupied and virtual orbitals respectively It also calculates the MBPT 2 energy 5 13 vec vec5 t and vcc5q These programs calculate the CC energy by solving the T amplitude equations and calculating all non iterative contributions xvcc also calculates finite order perturbation theory energies by manipulating the CC iteration logic 5 14 mrcc The xmrcc program uses a different programming environment than the rest of ACES II This program implements many EOM related excited state theories like IP EOM DIP EOM STEOM etc 5 15 fno xfno generates an improved set of orbitals based on the MBPT 2 density natural or bitals which can be truncated during the more costly coupled cluster calcu
17. The path variable is the full name of the directory to use for archiving SAVEDIR u home yau RESTART 28 This MUST be unique for all jobs running simultaneously If multiple jobs are using the same save directory then they are clobbering each other s files and the last job to archive its files is the one that wins The default action is to create a directory named SAVEDIR in the run directory 7 1 4 Molecular orientation The orientation of the molecule in Cartesian space is related to its point group Two orientations are used extensively in the ACES II program system the standard or com putational orientation which is a standard orientation for the computational point group and a canonical orientation which is the standard orientation for the full point group Note that in some cases the two orientations are identical All calculations are performed in the computational orientation therefore orbital symmetries should be specified according to this Cartesian axis system For the most part the canonical orientation is used internally for tasks such as determining irreducible representations or other properties related to the full point group The standard orientation for each point group follows Cy Rotation axis along z Dy Rotation axes coincident with Cartesian axes and the highest order axis along z Cs Plane of symmetry is ry Sy Sy axis along z Cny Cy axis along z z is dy Cnn Cy axis al
18. This can be set to SEGMENTED or GENERAL NOTE Even for truly segmented basis sets both integral and integral derivative programs run significantly faster in GENERAL mode ECP switch OFF Controls whether effective core potentials a kind of pseudopotential are used ON or not OFF ECP ON requires BASIS SPECIAL and specification of the ECP data sets in a file named ECPDATA 8 1 7 Integrals INTEGRALS handle VMOL Specifies which integral package to use This is not a very robust keyword and should always be set to VMOL unless otherwise directed to change it Other values include SEWARD and GAMESS but these options will not work unless the binaries were linked to the appropriate libraries which require third party license agreements INTGRL_TOL tol 14 Sets the tolerance for storing the two electron integrals such that integrals having an absolute value greater than 107 will be stored on disk 46 DIRECT switch OFF Controls whether two particle AO integrals are calculated on the fly or calculated once and stored in a file This cannot be used with VMOL integrals 8 1 8 Reference REFERENCE handle RHF Specifies the type of SCF reference to use Supported references include spin restricted Hartree Fock RHF spin unrestricted Hartree Fock UHF and spin restricted open shell Hartree Fock ROHF These apply to Kohn Sham DFT references as well e g RHF is spin restricted Kohn Sham R
19. ZGAd 00 0NV e S S AA TH D Z BA lt Sz gt ZSAOMd 67 gt 0S G6T S ZOADMA DO 04 107 S 03 P01 gt S ZLADMA DO S DE Ger Det S ZIADMADO 01769 1 61 Z9A0d 00 6 e S 0 63 GOT SUL Gee 5 VOA e e a Gee e 99 e 0S9 G61 Sor lt ZDADA S i Eur Z S ZLAOd 00 D See Deet e 9 6p ZT AOT D DEL 0T 69 oy D I 6T 2 Weg dl T 61 SEET ZOAd OO e z 36 gt 289 SADO D 6g 04 017 eg Z Messe S Ca SE GPA HEUL T SE 901 aN oe E ZLAd DD 9E QPT e SE SI py SE Een LGT a5 91778 i 26 I T S 6E eg 7 NVINdIHO oF 1 61 ml z S Aza E gt Y s ESCH S dAS S D S 9T er z G Ce 6 VG Ier C z 2 g S Za o PI gt OL s Z AS Oo e lt OT S A z i nfofal aa IT TH H 15 sieg D z DIN VN aN O sz ST tofs a fis av dz vo 3 uv E sg sy SVANAD Ul suog sy y3noIyy S T QLL 10 SUOTJOUNJ Steg OV P9J98IJUO9 JO Joquinu out 3409 1 atd 120 e e Lp e Es e or II OTDI e e Se e e We e e 9 I OTOI re ES Hait oe ze zer 2 09 Sr oe s Ire gWOTN09 SHOIY THV rer Tor Cer CC oe Ce Tor re 1 ENOTAOO0 NONHA rer Tor 6g reg 1 6 Ier ADNVHOXA ZV SSAVDG e Tor 6g reg 6 Ter ENOTAOO ZV SSAVDA rer Tor 6g Ce 1 6 T H NVHOXH IV SSNV A rer Tor Ce re 1 61 T ENOTAOO TV SSA
20. and NONHF OFF Keywords of type string are self explanatory however they are used to read in fixed format arrays until we define a general expression for array notation Hopefully the formats of keywords that accept arrays are clearly described in the manual Integers are also self explanatory and can be used in place of handles for those keywords of such type As mentioned before a tol type is an integer that corresponds to the negative power of 10 Some keywords are defined with unusual units and these are appropriately identified in the manual Currently reals apply only to memory related keywords These are identified as type special because the parser will scale the real by units suffixed to the value string 8 1 1 System general CALCLEVEL handle SCF Specifies the general level of theory to be used throughout the program Acceptable values are SCF MBPT 2 MBPT 3 SDQ MBPT 4 MBPT 4 LCCD LCCSD UCCSD 4 CCD UCC 4 CCSD CCSD T CCSD TQ CCSDT 1 CCSDT 1b CCSDT 2 CCSDT 3 CCSDT 4 CCSDT LCCSDT CCD ST CCD QCISD T CCSD T QCISD CID CISD QCISD TQ CCSD TQ CCSD TQ CCSDT Q CCSDT Q CC5SD T CCSD T CC3 CCSDT T1T2 CCSDTQ 1 CCSDTQF 1 CCSDTQ 2 CCSDTQ 3 CCSDTQ ACCSD and HFDFT Emphasized calculation levels in italics are not implemented currently but are planned for future versions of the program CALC HFDFT has been obsoleted by SCF_TYPE KS MEMORY SIZE special 15000000W Sets the
21. time dependent Hartree Fock TDHF calculations The TDHF code can solve the general order TDHF problem for closed shell RHF wavefunctions The TDHF keyword must be set to ON and additional parameters controlling the TDHF calculation are included in a namelist which is located at the end of the ZMAT file The variables in the namelist are as follows IOPDA 1 prints the density matrix 0 default does not IOPEV 1 prints the MO coefficients 76 IOPU 1 prints the U matrix IOPFE 1 prints the Fock matrix IOPPR 1 prints the property integrals For subsequent options 0 means do not do this type of calculation otherwise do it The default is to do the calculation so if a parameter is not specified the program will perform that calculation For the polarizability a IDALPH calculates o For the first hyperpolarizability IOR calculates optical rectification IEOPE calculates electrooptic Pockels effect ISHG calculates second harmonic generation For the second hyperpolarizability y IOKE calculates the optical Kerr effect IDCOR calculates the DC electric field induced OR IIDRI calculates the intensity dependent refractive index IDCSHG calculates the DC electric field induced second harmonic generation ITHG calculates the third harmonic generation The method of solution of the TDHF equations is controlled by the parameter NITER the default value is 20 If NITER 0 an iterative method is used NITER 1 uses a non ite
22. use direct integrals and communicate over the Message Passing Interface MPI The SCF gradients in xp_scfgrd use direct integral derivatives in the same fashion The parallel finite displacement algorithm is handled by the main driver program xp_aces2 and distributes displacements to each MPI task All of these programs require each MPI task to have its own set of files which can be managed before and after the parallel run by xgemini ACES III the fully parallel successor to ACES II will be able to compute HF SCF MBPT 2 and CCSD energies and gradients using a message passing protocol either MPI or shmem Despite the new internal architecture it will be compatible with the current input files ZMAT and GENBAS but it will not require separate file sets for each task The following scripts illustrate each parallel capability The sections that follow describe in further detail how some of these capabilities work bin sh Script 1 parallel SCF and MBPT 2 energies mpirun np N xgemini i s t shared GRANK xjoda amp amp xvmol amp amp xvmol2ja mpirun np N xp_vscf mpirun np N xp_dirmp2 OMIT THIS LINE FOR SCF ENERGIES ONLY mpirun np N xgemini x s bin sh Script 2 parallel finite differences mpirun np N xgemini i s t shared GRANK mpirun np N xp_aces2 mpirun np N xgemini x s bin sh Script 3 parallel SCF geometry optimizations mpirun np N xgemini i s t shared OGRANKO PRUN mpirun np N
23. 10 hartree bohr While this initialization and the subsequent numerical updates work satisfactorily for most small molecules there can be oc casional problems In these cases one might wish to use an alternative initial force constant matrix particularly one obtained by ACES II at the same or another level of theory There are a number of ways in which one might do this First the EVAL_HESS keyword can be used If nonzero the value associated with this keyword directs ACES II to calculate the SCF Hessian matrix prior to the first optimization step and then every N steps thereafter where N is the value of EVAL_HESS By setting N to a sufficiently large value larger than OPT MAXCYC then the Hessian will never be recalculated and the optimization will begin with the SCF Hessian However the strategy based on EVAL_HESS is not sufficient for all purposes For ex ample one might wish to use a Hessian which is evaluated at the correlated level This is not possible with EVAL_HESS since it will direct ACES II to calculate only the SCF Hessian regardless of the calculation type Alternatively one might adopt the economi cal strategy of using a Hessian which is evaluated at a low level of theory such as SCF with the STO 3G basis set This is the recommended approach for transition state searches and all optimizations for which the default Hessian is inadequate In any event it is quite straightforward to use another set of force constants to begin
24. 4 GridEngine S SHELL N JOBNAME o STDOUT_FILE 126 e STDERR_FILE 1 COMPLEX pe PE EN NPES 127 Index ACES2 ABCDFULL 55 ABCDTYPE 54 55 56 ACC_SYM 58 AO_LADDERS 54 BASIS 45 BRUCK_CONV 52 BRUECKNER 52 CACHE_RECS 44 CALCLEVEL 42 CC_CONV 56 CC_EXPORDER 56 CC_EXTRAPOL 56 CC_MAXCYC 56 CCR12 103 CHARGE 45 CHECK SYM 48 CONTRACTION 46 CONVERGENCE 66 COORDINATES 45 CPHF_CONVER 64 CPHF_MAXCYC 64 CURVILINEAR 65 DAMP TOL 50 DAMP_TYP 50 DAMPSCE 50 DEA_CALC 59 DEA SYM 59 DENSITY 57 DERIV_LEV 43 DIP_CALC 61 DIP_SYM 62 DIRECT 46 DOHBAR 57 DROPMO 51 128 EA_CALC 59 EA SYM 59 ECP 46 EE_SEARCH 60 EE_SYM 60 EIGENVECTOR 65 EOM MAXC 57 EOM PRJCT 58 EOMPROP 63 EOMREF 57 EOMXFIELD 103 EOMYFIELD 103 EOMZFIELD 103 ESTATE_MAXC 57 ESTATE PROP 58 ESTATE SYM 60 ESTATE_TOL 57 EVAL_HESS 66 EXCITE 57 EXTERNAL 68 FD_IRREPS 67 FD_PROJECT 68 FD_STEPSIZE 67 FD USEGROUP 68 FILE_RECSIZ 44 FILE_STRIPE 103 FOCK 48 FUNCTIONAL 103 GAMMA ABCD 55 GENBAS1 45 GENBAS_2 46 GENBAS 3 46 GEOM OPT 65 GLOBAL_MEM 103 GRAD_CALC 43 GUESS 48 HBARABCD 55 HBARABCI 55 HESS UPDATE 66 HF2 FILE 54 55 HFSTABILITY 51 57 IMEM SIZE 58 INCORE 44 INIT_HESSIAN 66 INSERTF 103 INTEGRALS 46 INTGRL_TOL 46 IP_CALC 61 IP_SEARCH 61 IP_SYM 61 JODA PRINT 44 KS POT 103 KUCHARSKI 103 LINDEP_TOL 45 LOCK ORBOCC 49 LSHF A1
25. ACES II does not check to make sure that the atom to which the basis set belongs corresponds to the atomic designation in the corresponding row of the Z matrix No entries are made for dummy atoms The format of the basis set names in GENBAS is XX BASNAM where XX is the atomic symbol of the atom in capital letters and BASNAM is the name of the basis For new users it is probably best to search GENBAS for XX XX being the atomic symbol since this will show all of the available basis sets for that atom A description of the format of GENBAS and its contents are given in the next section 7 2 GENBAS ZMAT BAS The following fixed format is used to store basis sets in the GENBAS file Note that lines 5 7 and 8 define numbers of shells NS contractions NC and exponents NE Line Fortran format Description 1 A80 blank line 2 A80 name of the basis set 3 A80 comment line 4 A80 blank line 5 I3 the number of shells in the basis set NS 6 NS I5 angular momentum for each shell L 7 NS I5 number of contracted basis functions for each shell NC 8 NS I5 number of exponents for that shell NE 9 A80 blank line 10 NE F14 7 exponents for the first shell 11 A80 blank line 12 NC F10 7 1X contraction coefficients for the first shell 13 A80 blank line It is necessary that the shells are grouped by angular momentum with the s shell s first followed by p shell s etc otherwise the input file written for the V
26. Line 8 Line 9 RHF MBPT 2 property calc of Formaldehyde in DZP 0 0 0 0 0 1 22 C 0 0 0 0 0 0 H 0 0 0 873489539 0 545816842 H 0 0 0 873489539 0 545816842 ACES2 CALC MBPT 2 PROPS FIRST_ORDER BASIS DZP COORD CARTESIAN 7 1 3 File directives Users can control the locations of files with file directives in the header By default all files used by ACES II JOBARC JAINDX IIII MOINTS etc are kept in the directory in which the xaces2 program is invoked Any file except ZMAT may be relocated with FILE some path to the FILE in which FILE is the name of the particular file The directory path does not have to be absolute For example if a user wants to calculate isotopic shifts after a vibrational frequency calculation it is useful to keep JOBARC and JAINDX in a safe place not in the current directory For water the following input could be used Line 1 Line 2 Line 3 Line 4 Line 5 Line 6 Line 7 Line 8 Line 9 Line 10 Line 11 Line 12 Line 13 Line 14 4 JOBARC u usr safe JOBARC 4 JAINDX u usr safe JAINDX RHF SCF vib freq calc for Water in DZP DO OO CO 1R 2R1A R 0 957 A 104 51 ACES2 VIB EXACT BASIS DZP After this job executes the files JOBARC and JAINDX are stored in the directory u usr safe Then the user can simply change to this directory create the appropriate ISOMASS file and run xjoda directly Coarse grain restart capabilities depend on a file named SAVEDIR
27. Phys 52 2769 1970 3 J B Collins P R Schleyer J S Binkley J A Pople J Chem Phys 64 5142 1976 4 W J Pietro B A Levy W J Hehre R F Stewart J Am Chem Soc 19 2225 1980 5 W J Pietro E S Blurock R F Hout Jr W J Hehre D J DeFrees R F Stewart Inorg Chem 20 3650 1981 6 W J Pietro W J Hehre J Comp Chem 4 241 1983 B 1 2 MINI MINI SCALED MIDI MIDI The MINI minimal basis sets use 3 Gaussians to expand each AQ Exponents and contraction coefficients were determined by minimizing the energy for the restricted Hartree Fock ground state In the paper of Tatewaki and Huzinaga H Tatewaki S Huzinaga J Comp Chem 1 205 1980 these were called the MINI 1 basis sets zi Main group elements in this set have their valence orbitals scaled by factors determined by John Deisz of North Dakota State University These scale factors were determined by minimizing the energy of a set of one and two heavy atom molecules at the experimental geometry The actual value is 106 an average over 20 30 molecules zi The MIDI basis set is derived from the MINI minimal basis sets by floating the outermost valence primitives and renormalizing the remain 2 Gaussians The MINI minimal basis sets use 3 Gaussians to expand each AO Exponents and contraction coefficients were determined by minimizing the energy for the restricted Hartree Fock ground state zi The MIDI basis sets were designed
28. Properties PROPS handle OFF Specifies whether properties other than the energy geometry and vibrational frequen cies are calculated Acceptable values are OFF no properties are calculated FIRST_ORDER dipole moment quadrupole moment electrical field gradients spin densities etc and the approximate perturbative relativistic correction to the energy by Cowan Griffin SECOND_ORDER frequency independent polarizabilities commonly known as the CPHF polarizabilities provided CALC SCF EOM_NLO both frequency dependent and independent polarizabilities at the CCSD level using EOM CC see EOMPROP J_SO paramagnetic and diamagnetic spin orbit contributions to the NMR spin coupling constant J_SD spin dipole contribution to the NMR spin coupling constant J_FC Fermi contact contribution to the NMR spin coupling constant JSC_ALL gt all four contributions NMR NMR chemical shifts limited to SCF and MBPT 2 with SPHERICAL OFF Note that options J SO J SD J FC and JSC_ALL require CALC CCSD and that options J_SO J_FC and JSC_ALL require REF UHF see EOMPROP EOMPROP handle CILIKE Specifies the method of calculating EOM CCSD second order properties polarizabil ities and spin spin coupling constants CILIKE uses a Cl like formula which is not rigorously size extensive LINEAR removes unlinked diagrams from the Cl like formula QUADRATIC uses a size extensive quadratic form
29. RHF is limited to singlets but UHF can do singlets and triplets using the N N N N notation like DEA SYM EE_SEARCH handle LOWEST Specifies the character of the states to calculate If EE SYM is not specified the program attempts to determine all states of the given character otherwise it uses the symmetry constraints imposed by EE SYM The values are LOWEST find the lowest energy roots regardless of character CORE find excitations from core orbitals using guess vectors from a projected TDA matrix LUMO find excitations to the LUMO and HOMO find excitations to the HOMO ESTATE_SYM string of 1D array 0 Specifies the number of excited states which are to be determined in each irreducible representation of the computational subgroup The program attempts to find all of the lowest roots but this is not guaranteed since the eigenvalue problem is not solved by direct matrix diagonalization rather by an iterative modified Davidson algorithm For excited state gradient calculations TDA only only one root can be specified so only one non zero entry in the string is allowed and that entry must be set to one The format used for this keyword is identical to that used in the OCCUPATION keyword For example for a computational subgroup having four symmetry species the string 60 ESTATE_SYM 3 1 0 2 specifies that 6 total roots should be searched for three in the first block one in the second block and two in th
30. SCF xvscf creates a formatted file called AOBASMOS Before the frequency calculation the OCCUPATION keyword must be removed from ACES2 the option GUESS READ_AO_MOS must be set and the AOBASMOS file must be renamed OLDAOMOS At each point in the frequency calculation the MOs are read from 72 OLDAOMOS and transformed to the current symmetry The occupation is determined in the current point group and since the displacements from the reference geometry are small the initial guess is usually very good It is also possible to use the final AOBASMOS file from a geometry optimization or transition state search as the OLDAOMOS file in a frequency calculation The GUESS READ_AO_MOS option can be used in other situations but it might not work well when the geometry changes significantly The reason for this is that a transformation matrix relating the different geometries must be calculated and this is only approximated well by a rotation if the geometries are close 9 1 8 Improving SCF convergence VSCF contains a number of options for accelerating and controlling the SCF convergence By default the first few iterations proceed by repeated diagonalization of appropriate Fock matrices Once either a certain number of iterations have been performed specified by SCF_EXPSTART or an initial convergence criterion has been met the DIIS convergence extrapolation procedure of Pulay begins If convergence difficulties are experienced with the
31. X 4R1 2 A 1 TO H 3 RHT 13 A 2 T H 4 RHT 14 A 2 T R1 1 207 R 1 05183 RX 0 08546 RHX 1 48313 RHT 0 569 A 90 T 180 T60 60 T12 120 TO 0 TM2 120 TM6 60 ACES2 BASIS DZP CALC 1 OPT_METHOD MANR EVAL_HESS 3 MAX_STEP 750 CONV 5 This ZMAT file specifies a geometry optimization for the D3q isomer of beryllium boro hydride BeB2Hg Compared with the water example above which uses defaults for the optimization keywords this optimization input adds some features The MANR optimiza tion algorithm is used The SCF Hessian will be reevaluated every three cycles and the maximum step length is set to 750 millibohr The convergence criterion CONV is set to 5 which means the optimization will continue until the root mean square force is below 1075 81 rather than the default of 107 atomic units An MBPT 2 calculation is specified by the CALC keyword This has been specified using the number corresponding to MBPT 2 al though our recommendation is to use the name The default values for CHARGE MULT REF and OCCUPATION are used The geometry optimization request automatically turns on the necessary gradient options 9 2 3 Full optimization of Cartesian coordinates RIC H2 geom opt H 0 0 0 H 0 0 0 5 ACES2 basis DZP geom_opt full This feature is controlled by the GEOM_OPT keyword The RIC analysis includes in formation about atomic radii If the input geometry is far from equilibrium then
32. an optimization First one simply runs a harmonic frequency calculation at the desired level of theory and saves the file named FCMINT This formatted file contains the full internal coordinate force constant matrix When xjoda performs a geometry optimization it checks the active working area for the presence of FCMINT no special keywords or commands are needed to do this If the file exists xjoda uses those force constants to initialize the Hessian matrix While the geometry and even the point group symmetry specified by ZMAT in the har monic frequency calculation need not be the same as that used in the first step of the geometry optimization the Z matriz connectivity must be identical If one attempts to use 84 an entirely different Z matrix then the definitions of internal coordinates are no longer the same and chaos may ensue While this point may seem to be unimportant this situation occurs relatively frequently Suppose the equilibrium geometry for a transition state is as sumed to be planar and the user attempts to locate a stationary point However when the harmonic frequencies are calculated two modes are found to have imaginary frequencies an a mode in plane and an a mode with the a mode corresponding approximately to the reaction coordinate of interest As a result the true transition state geometry does not contain a plane of symmetry and another search must be performed in a reduced symmetry To this end
33. and units must be one string no spaces 19 For example a user might be nursing a large calculation by running each executable by hand and backing up the files between them If he or she discovers ACES needs more memory then the user changes the MEMSIZE value in ZMAT and runs xa2proc mem on the backup files 5 27 3 zerorec This module flushes records in the JOBARC file with zeroes 5 27 4 rmfiles The RMFILES module will delete the five list storage files MOINTS GAMLAM MOABCD DERINT and DERGAM More importantly it will reset the appropriate pointers and counters so the next AME that attempts to initialize the I O subsystem will not crash 5 27 5 parfd The PARFD module is used to export and import finite difference information It was created as a proof of concept program to demonstrate JODA s ability to operate in a parallel finite difference calculation Its main capabilities are incrementing the displacement data parfd update dumping the data to standard output parfd dump and importing data from other calculations parfd load file An example of manual parallel finite differences can be found in Section 10 3 2 page 92 5 27 6 molden and hyperchem These modules create Molden and HyperChem input files respectively that contain geometry wavefunction and vibrational frequency information 5 27 7 jasum and iosum These modules print summary information of the JOBARC records and storage lists re spectively
34. angle Although most users would definitely use the first Z matrix the idea of using just chemical bonds as internuclear distances can be dangerous For example in a regular hexagonal ring a little reflection will show that one cannot include all six inter vertex distances in the Z matrix If only five are specified such as in the Z matrix below the 1 6 distance is missing the internal coordinate gradient cannot have the full symmetry of the molecule and the first step of a geometry optimization will break the molecular symmetry This results in extremely slow convergence and significantly increased CPU time due to the reduced molecular symmetry A poor Z matrix for hexagonal C6 C C 1 R C 2 R 1 C3R2A1T C4R3A2T C5R4A3T R 1 33 A 120 T 0 In these situations it is always best to use dummy atoms which are merely mechanisms to reference a point in space for other coordinates One or more dummy atoms are needed in essentially all Z matrices for molecules with high symmetry As an example of their use a good Z matrix for the Ce ring is shown below Don t be afraid to use dummy atoms A better Z matrix for hexagonal C6 X X 1 RX C2R1A C2R1A3T C2R1A4T C2R1A5T C2R1A6T C2R1AT7T RX 1 0 R 1 33 A 90 T 60 33 7 1 9 ACES2 namelist The ACES2 namelist is a keyword value pair listing with a few tweaks Every keyword has an internally declared type which controls what form the value word s may take Curre
35. as a modification of the original Huzinaga MIDI basis which give good geometries and charge densities 1 J Andzelm M Klobukowski E Radzio Andzelm Y Sakai H Tatewaki in Gaussian Basis Sets for Molecular Calculations edited by S Huzinaga Elsevier Amsterdam 1984 2 R E Easton D J Giesen A Welch C J Cramer D G Truhlar Theor Chim Acta 93 281 1996 3 J Li C J Cramer D G Truhlar submitted to Theor Chem Acc B 1 3 3 21G 3 21G 3 21 G 3 21 G This is a small split valence basis set developed by Pople and coworkers It uses a minimal basis or single zeta description for the core orbitals and a double zeta description for the valence orbitals This and other sets of this type such as 4 31G and 6 31G are often termed double zeta valence This is not strictly accurate since the s and p exponents are constrained to be equal which they are not in a true double zeta set although the two sets give results of similar quality The 3 21G basis set contains the same number of Gaussian primitives as the STO 3G basis but the valence electrons are described with two functions per AO instead of one In most cases the 3 21G basis set gives results which are as good as the more expensive 4 31G and 6 31G sets zi The 3 21G basis set adds a single set of 6 term d functions to elements Na Ar to account for the participation of d functions in second row bonds The authors suggest that the 3 21G basis n
36. if the SCF virtual orbitals are to be replaced by a set determined from an N 1 potential This is an orthogonal transformation within the virtual space As long as appropriate fi and fab are included the energies of standard single reference methods are unchanged However for TD CC methods this keyword mixes one of the occupied orbitals with the virtual space and the results are changed It is anticipated that TD CCSD calculations will be improved by the use of this keyword but this has yet to be demonstrated NEWVRT is useful for interpreting results of EOM EE calculations since excitations tend to be more easily identified as involving one occupied and one virtual orbital OFF does not rotate the virtual space and ON rotates the virtual space 8 1 11 SCF iteration control SCF_CONV tol 7 Sets the convergence criterion for the SCF equations Equations are considered con verged when the maximum change in density matrix elements is less than 107 SCF_MAXCYC integer 150 Sets the maximum number of SCF iterations DAMP_TYP handle NONE Specifies what type of damping is used during the SCF iterations The value can be NONE no damping DAVIDSON Davidson s empirical dynamical damping scheme and STATIC a fixed damping parameter Note that RPP convergence extrapolation is not turned on until the damp factor has gone below DAMP_TOL and the energy change is below 0 05 a u DAMP_TYP DAVIDSON is recommended even
37. is shown for demonstration purposes only bin ksh echo Do not run this script as is amp amp exit 1 pick a function alias loop lastraman for analytical gradients w raman intensities alias loop lastgrad for analytical gradients 1st order grid alias loop lastener for numerical gradients 2nd order grid test n 1 amp amp procs 1 procs 1 test procs 1t 1 amp amp exit 1 92 alias xj xjoda rank rank procs procs if true Hartree Fock alias scf xvscf alias dint xvdint else DFT alias scf xvscf_ks amp amp xintgrt alias dint xvdint amp amp xvksdint fi function lastener lastgeom 0 while test lastgeom eq 0 do xa2proc rmfiles xj amp amp xvmol amp amp xvmol2ja amp amp scf return 1 xvtran amp amp xintprc ER xvcc return 1 lastgeom xa2proc jareq i LASTGEOM 1 tail 1 done function lastgrad lastgeom 0 while test lastgeom eq 0 do xa2proc rmfiles xj amp amp xvmol amp amp xvmol2ja amp amp scf return 1 xvtran amp amp xintprc E xvcc amp amp xlambda return 1 xdens amp amp xanti amp amp xbcktrn return 1 dint return 1 lastgeom xa2proc jareq i LASTGEOM 1 tail 1 done function lastraman lastgeom 0 while test lastgeom eq 0 do xa2proc rmfiles xj amp amp zumol amp amp xvmol2ja amp amp xvprops amp amp scf return 1 xvtran 48 xintprc amp amp xvcc amp amp xlambda
38. is that the user requests a restricted open shell Hartree Fock reference The program automatically turns on the NON HF option so the appropriate non Hartree Fock terms are included in the coupled cluster equations It also automatically sets ORBITAL SEMICANONICAL which is needed to evaluate triple excitations non iteratively as in the CCSD T method 9 1 4 QRHF CCSD T energy H20 QRHF CCSD T energy calculation DZP basis set H O 1 ROH H 2 ROH 1 HOH ROH 0 95 HOH 104 5 ACES2 REF RHF CALC CCSD T BASIS DZP QRHF_GENERAL 1 ORBITAL SEMICANONICAL OCCUPATION 3 1 1 0 3 1 1 0 This is another way in which one can calculate the energy of an open shell molecule Here the energy of the 7A state of the water cation is being calculated by the QRHF CCSD T method The orbitals however are not from an SCF calculation on H20 Rather they are orbitals from the neutral molecule In general QRHF means orbitals are taken from a convenient closed shell system and are then used in an open shell system An important difference between this example and the previous examples is that the CHARGE MULT REF and OCCUPATION keywords do not refer to the system being studied SOT but instead refer to the system from which the orbitals are obtained H20 QRHF_GENERAL 1 along with the default values for QRHF_ORBITAL and QRHF SPIN means the reference function for the correlated calculation is to be formed by removing a P spin electron from
39. it would certainly be useful to use FCMINT obtained in the frequency calculation to start the search but one must be careful that the molecular configuration used in the reduced symmetry has exactly the same connectivity scheme as that used in the search for the planar structure 9 3 Frequency calculations 9 3 1 Numerical frequencies from analytical gradients N4 finite difference frequency calculation N X1R N 2 R 1 TDA N2R 1 TDA 3 T N2R 1 TDA 4 T R 0 945 TDA 110 T 120 ACES2 VIB FINDIF BASIS TZ2P CALC MBPT 2 This ZMAT file directs ACES II to perform a harmonic frequency calculation for N4 and to compute the force constants by numerical differentiation of analytic gradients The TZ2P basis set is selected Note that the TDA parameter is used in the Z matrix Although the specified value in the parameter input section is not the exact tetrahedral angle the program will convert TDA to the correct value 109 4712206 degrees internally Since finite differences are used no symmetry specific keywords can be specified so the orbital symmetry specification is omitted This is because the determination of the force constants requires several gradient calculations at geometries with different symmetries The FINDIF option automatically turns on the requisite gradient and property options Note that no asterisks may appear in the Z matrix in a frequency calculation This is a common error since frequency calculations are usu
40. obsolete and unused keywords ACES II is an evolving program with multiple contributors As such certain areas of the code are highly volatile and keywords relating to those areas cannot be documented with 100 accuracy For completeness this section lists all remaining keywords in the ACES2 namelist that are not listed in Section 8 page 41 but they should not be used Furthermore some programs are developed independently using ACES II protocols but they are not distributed with adequate documentation Their namelists are also defined here but are not documented SOLVENT integer 0 Sets the dielectric constant used to determine the orbitals A cavity size may be specified as well by creating a file named radius which is read by the SCF code This contains the cavity radius in A If the file is not present the program uses a value calculated from 0 5 A plus half of the longest internuclear distance TURBOMOLE switch OFF POLYRATE switch OFF CCR12 switch OFF KUCHARSKI switch OFF KS_POT Obsoleted by the VSCF namelist FUNCTIONAL Obsoleted by the INTGRT namelist Unused EOMXFIELD EOMYFIELD EOMZFIELD RDO FILE_STRIPE RESET FLAGS PSI INSERTF GLOBAL_MEM A 2 Kohn Sham DFT namelists A 2 1 VSCF KS switch ON Controls the type of SCF KS ON performs a Kohn Sham DFT calculation with nu merical integration and KS OFF performs the standard analytical Hartree Fock SCF with the KS SCF
41. parscfopt sh mpirun np N xgemini x s 87 10 2 Running xgemini xgemini is a tool for managing private scratch directories from a central location Fach MPI task of xgemini or the parallel AMEs must start in this directory which means it must be globally visible on distributed computers Henceforth this directory will be called WORKDIR Depending on the machine architecture the two choices the user must make are 1 where will the private scratch directories be created and 2 how will the AMEs get to them These decisions will affect performance if local disks are available on each node and will determine how a parallel job should be restarted When the user runs a parallel member executable usually xp_aces2 each parallel task must start in WORKDIR which usually contains ZMAT and GENBAS Before reading any files each task attempts to cd into a directory called nodename rank The tasks on a machine named crunch would look for crunch 0 crunch 1 etc If they cannot find these directories then they try to cd into directories called shared rank shared 0 shared 1 etc If those are not found either then the tasks start running in the current directory 10 2 1 Local scratch directories For the following example assume that the parallel computer has compute nodes called compnode0 compnodel etc and that the user named smith is allowed to create directories in local tmp The following command will create a directo
42. popu lated or depopulated For calculations involving more than one removal or addition of electrons values are separated by forward slashes and correspond one for one to the QRHF_GENERAL value array For example specifying QRHF_GENERAL 2 4 and QRHF_ORBITAL 3 2 means that an electron will be added to the third lowest virtual in symmetry block 2 and another will be removed from the second highest oc cupied orbital in symmetry block 4 Examples in Sections 9 1 4 and pages 70 and further illustrate the QRHF input options QRHF SPIN string of 1D array Specifies the spins of QRHF occupations by overriding the remove from 8 and add to a behavior An element value of 1 means o spin and 2 means spin NOTE This option allows one to construct low spin determinants which generally are unsuitable for single reference coupled cluster calculations An important exception is the open shell singlet coupled cluster method REFERENCE TWODET UNO_REF switch OFF Controls the use of the new UNO reference state The ACES II calculation is initial ized by an open shell SCF calculation specified in the usual way either UHF or RHF The resulting a and 8 density matrices from the SCF calculation are then added into a spatial density matrix and this is diagonalized to yield unrestricted natural orbitals Pulay s UNOs These orbitals are used to create a new reference determinant The UNOs can be used in conjunction with QRHF and NEWVRT
43. program xvscf_ks The ACES2 keyword SCF_TYPE must be set to KS in addition to this switch which defaults to ON The developers advise against 103 using xvscf_ks for routine HF SCF calculations and for this reason the keyword is listed in italics i e users should ignore it A 2 2 INTGRT POTRADPTS integer 50 RADTYP handle Handy Handy Gauss Legendre PARTPOLY handle bsrad equal bsrad dynamic RADSCAL handle Slater none Slater PARTTYP handle fuzzy rigid fuzzy FUZZYITER integer 4 RADLIMIT real 3 0 NUMACC switch ON EXACT_EX switch OFF TDKS switch OFF ENEGRID integer 4 104 POTGRID integer 4 ENETYPE handle lebedev POTTYPE handle lebedev FUNC string none This keyword can accept multiple types of strings The regular expression is exch coeff corr coeff This means one exchange or two exchange and correlation functionals can be used to evaluate the total SCF energy and each can have an optional coefficient default is 1 0 To use B3LYP the format must be func b3lyp with no coefficients or additional correlation functionals The following exchange and hybrid functionals are available lda LDA Slater Xalpha exchange becke Becke exchange hf Exact Nonlocal Exchange exchange pbeex Perdew Burke Ernzerhof exchange pw9l_ex Perdew Wang 91 exchange b3lyp Becke III LYP hybrid The following correlation fu
44. return 1 xdens amp amp xanti amp amp xbcktrn return 1 dint amp amp xcphf return 1 lastgeom xa2proc jareq i LASTGEOM 1 tail 1 done clear out old FD data rm rf shared fd out rank procs 1 while test frank ge 0 do this is the xgemini portion mkdir shared frank cd shared rank ln s ZMAT ln s GENBAS this is the meat of the routine loop gt rank out exit 1 xa2proc parfd updump gt gt fd out update and print the FD data reset and cycle to the next process 93 cd let rank 1 done cd shared 0 xa2proc parfd load fd out load the FD data from the other procs xjoda procs procs run the final xjoda cd For geometry optimizations every process will have to load the FD data and the entire procedure will have to loop over coordinates until convergence when the integer record JODADONE is 1 Needless to say this is not an exercise for even intermediate users of ACES II but the capability exists should the need arise 10 3 3 SCF geometry optimizations In Script 3 from the Overview section page 87 the core of the commands are contained in another shell script parscfopt sh That script contains the loop logic to iterate until xjoda has found the minimum energy geometry bin sh parscfopt sh test d shared 0 amp amp s s s rootdir ls d 0 PRUN xgemini s xjoda cd rootdir jodadone xa2proc jareq i JODADONE 1 tail 1
45. s and their associated basis sets are designed to replace all but the outermost electrons in an atom For example for K the 1s 2s 2p 3s and 3p are considered core 1 W J Stevens H Basch M Krauss J Chem Phys 81 6026 1984 2 W J Stevens M Krauss H Basch P G Jasien Can J Chem 70 612 1992 3 T R Cundari W J Stevens J Chem Phys 98 5555 1993 B 2 4 CRENBL CRENBS The ECPs for Cf Es Fm and Md are described in J Chem Phys 106 5133 1997 E These ECPs are sometimes referred to as shape consistent because they maintain the shape of the atomic orbitals in the valence region zi For example for copper only the 4s and 3d orbitals are included in the valence space 1 T H Dunning Jr P J Hay in Methods of Electronic Structure Theory Vol 3 edited by H F Schaefer III Plenum Press 1977 2 L F Pacios P A Christiansen J Chem Phys 82 2664 1985 3 M M Hurley et al J Chem Phys 84 6840 1986 4 L A LaJohn et al J Chem Phys 87 2812 1987 5 R B Ross W C Ermler P A Christiansen et al J Chem Phys 93 6654 1990 117 6 W C Ermler R B Ross P A Christiansen Int J Quant Chem 40 829 1991 7 C S Nash B E Bursten W C Ermler J Chem Phys 1997 B 2 5 ST RLC RSC97 Stuttgart Relativistic Large and Small Core ECP Basis Sets The ST RSC97 basis sets and ECPs correspond to Revision Fri Jun 27 1997 of the Stuttgart Dresd
46. the bond analysis may not be accurate and the calculation will become unstable but it might not crash 9 2 4 Transition state search CH20 gt H2 CO transition state search DZP basis U C 1 Rx H 2 R2x 1 Ax H 3 RHHx 2 Ax 1 T R 1 254 R2 1 08 RHH 1 01 A 133 5 T 0 ACES2 OPT_METHOD EVFTS MAX_STEP 150 BASIS DZP CALCLEVEL MBPT 2 OCCUPATION 7 1 As implied by the job title this ZMAT file specifies an initial geometry and basis set for a transition state search In the Z matrix the parameters designated R R2 RHH and A will be optimized while parameter T will be fixed at 0 degrees This obviously constrains the structure to be planar The ACES2 namelist parameters tell ACES II that 1 this is a transition state search using the Cerjan Miller eigenvector following method 2 the maximum allowed step length is 150 millibohr this corresponds to the absolute step length 1 norm since the default value of SCALE_ON has not been overridden 3 the DZP basis set is to be used 82 4 the calculation type is MBPT 2 and 5 the occupation appropriate for the C point group is 7 1 The value EVFTS automatically turns on the necessary gradient options In transition searches it is necessary to provide an initial estimate of the Hessian This is usually done by saving the FCMINT file from a frequency calculation and copying this to the workspace at the beginning of a transition state search Another examp
47. the equal sign is passed to a number parser so trailing text like a comment is ignored Angles must be entered in degrees and bond angles as distinct from dihedral angles of 0 and 180 are not allowed since these lead to a singularity in the transformation between Cartesian and internal coordinates and do not allow dihedral angles to be defined This of course does not mean that ACES II is unable to handle linear molecules such as carbon dioxide Rather for linear molecules dummy atoms must be used in the Z matrix to avoid problematic bond angles To facilitate construction of Z matrices for highly symmetric molecules certain variable names have been reserved for specific values To use these parameters which are listed below the user must specify a value in the parameter input section but it need not be correct since it will be converted to the exact value internally TDA Specifies the tetrahedral angle Cos Les 109 4712 IHA Specifies the icosahedral angle Cos v3 De 63 4349 7 1 8 Z matrix analyzer A unique feature of ACES II is the Z matrix analyzer which is capable of detecting subtle and obvious deficiencies in the definition of internal coordinates This is particularly important for geometry optimizations in which the construction of the Z matrix and the choice of parameters to be optimized are of vital importance The analyzer inspects the internal coordinates and carries ou
48. uses rotationally projected coordinates while OFF re tains the rotational degrees of freedom At a stationary point on the potential surface both options will give equivalent harmonic force fields but OFF should be used at non stationary points FD_USEGROUP handle FULL Specifies the point group to be used in generating the symmetry adapted vibrational coordinates FULL specifies the full molecular point group COMP specifies the Abelian subgroup used in the electronic structure calculation POINTS handle DOUBLE Specifies either single SINGLE or double sided DOUBLE numerical differen tiation in the finite difference evaluation of the Hessian Two sided numerical differ entiation is considerably more accurate than the single sided method and its use is strongly recommended for production work 8 1 28 External interfaces EXTERNAL handle NONE Specifies what type of external file interface for third party programs to create Current interfaces include HYPERCHEM and MOLDEN A2PROC help This is not a keyword per se but is used by ACES II to generate the external interfaces and it can be used to generate such interfaces after a calculation has finished 68 9 Examples 9 1 Single point calculations 9 1 1 RHF CCSD T energy H20 CCSD T energy calculation DZP basis set H O 1 ROH H 2 ROH 1 HOH ROH 0 95 HOH 104 5 ACES2 CALC CCSD T BASIS DZP This illustrates a single point C
49. within SAVEDIR but will not delete that top level directory This behavior could change in the future such that if xjoda created SAVEDIR then it will delete it as well 9 2 6 Initializing the Hessian with FCMINT in a geometry search All of the geometry optimization algorithms incorporated into ACES II are based on the Newton Raphson method in which step directions and sizes are related to the first and 83 second derivatives of the molecular potential energy However in almost all practical calcu lations the exact second derivative matrix is not evaluated but rather approximated As the calculation progresses well established numerical methods are used to estimate the elements of the Hessian matrix based on all previous optimization steps After a large number of steps have been taken one may safely assume that the totally symmetric Hessian used to form the step is a reasonable approximation to the correct Hessian However in the early stages of the optimization there is not a sufficient amount of available information to accurately estimate the Hessian and problems may ensue By default ACES II geometry optimizations begin with a very crude estimate of the Hessian in which all force constants for bonded interac tions as specified by the Z matrix connectivity are set to 1 hartree bohr all bending force constants corresponding to bond angles in the Z matrix are set to 0 25 hartree bohr and all torsional force constants are set to 0
50. 0 7 also for trans metals d set eta0 2 p A0 0 78 beta 3 15 gamma 0 91 gt only for comparison f set eta0 2 d A0 1 095 if 1f auxfunct for transition metals the 3 f functions have been determined with eta0 2 d A0 beta 2 5 gamma 0 64 115 taken from optimization of Mn03F g set where possible determined by optimization of the atoms else eta0 2x d A0 for Mn g s have been taken without modification from the optimized 3 1g set of Fe d functions have been determined by optimization of atoms where possible for group V elements they have been determined by interpolation contraction coefficients for p and d functions have mostly been determined for the hydrides zi numbers of functions used main group elements 1 period 3 1 1 1 1 1s 3p 3d 1f 2 period 8 1 1 1 15 4 1 1p 4 1d 1f 1g8 As 4p transition metals 10 2 1 1 1 1 1 s 4p 2 1 1d 3f 3 1g 1 K Eichkorn O Treutler H Ohm M Haser R Ahlrichs Chem Phys Lett 240 283 1995 2 K Eichkorn F Weigend O Treutler R Ahlrichs Theor Chim Acc 97 119 1997 B 1 20 IGLO II III These basis sets are intended for computing NMR chemical shifts IGLO Individual Gauge for Localized Orbitals 1 W Kutzelnigg U Fleischer M Schindler in The IGLO Method Ab Initio Calcula tion and Interpretation of NMR Chemical Shifts and Magnetic Susceptibilities vol 23 Springer Verlag Heidelberg 1990 B 2 ECP sets in ECPDATA B 2 1 HAY WADT MB HA
51. 1 Instability classification RHF gt UHF with broken symmetry Eigenvalue 0 2166464610 Instability classification RHF gt RHF with broken symmetry There are 2 instabilities within irrep 7 Eigenvalue 0 3586792381 Instability classification RHF gt UHF with broken symmetry Eigenvalue 0 2166464610 Instability classification RHF gt RHF with broken symmetry There are 0 instabilities within irrep 8 It is sometimes desirable to obtain solutions that correspond to following instabilities to lower energy stationary points in the orbital rotation space The keywords HFSTABIL ITY FOLLOW and ROT_EVEC are provided for these situations 75 When HFSTABILITY FOLLOW is set the stability analysis is performed as it would be for HESTABILITY 0N and then the orbital rotation corresponding to the chosen in stability is applied to the SCF eigenvectors The SCF calculation is repeated with these rotated vectors as the starting guess This is not strictly eigenvector following nor is it direct minimization SCF but in practice the procedure is quite effective By default the lowest eigenvalue of the totally symmetric irrep is followed Other eigenvectors can be followed by explicitly specifying them with the ROT_EVEC parameter Only instabilities that transform as the totally symmetric irrep irrep 1 can be fol lowed since all others reduce the symmetry of the wavefunction Following other instabilities is possible pro
52. 280022 NN 37 202 30523682 2 5 53380013 154 84190369 2 10 20059967 9 21743488 1 2 66059995 3 18838096 O 32 17929840 P D 19 07518959 2 3 69499993 63 05695343 2 4 45380020 127 18070221 2 6 17630005 158 41213989 2 8 83930016 5 66128206 1 14 67029953 5 39882612 O 30 43350029 x CU CRENBS ECP from CRENBS ECP 18 CORE ELECTRONS x NCORE 18 LMAX 2 D 0 17287099 2 0 47510001 1 67033803 2 1 42369997 8 22487736 2 4 03959990 23 39514351 2 12 44390011 11 42614460 1 38 74039841 S D 6 22903204 2 0 81150001 48 31729889 2 1 10739994 113 06871033 2 1 38030005 65 06185150 2 1 65380001 19 98191833 1 2 90939999 3 21701002 O 10 47480011 P D 35 69552994 2 0 73470002 103 46006775 2 0 83590001 151 97230530 2 1 03230000 87 70287323 2 1 25300002 20 80924225 1 2 32179999 4 79295111 0 9 20419979 The first line contains a single star The name of the data group starts with the element symbol followed by the ECP name It is useful to give a comment introduced by the hash symbol in the next line to indicate the origin of the ECP The actual ECP data are given between two lines with asterisks The first line specifies the number of core electrons described by the ECP NCORE and the maximum angular momentum number of the projector operators LMAX in integers s 0 p 1 d 2 These are followed by the description of the effective core potential which consists of the angular momentum numbers and by the anal
53. 31G adds hydrogen p polarization functions to the 6 31G basis zi The 6 31 G basis set adds a diffuse s p shell to elements Li Cl and a single diffuse s to hydrogen zi The 6 31 G basis set adds a diffuse s p shell to elements Li Cl and a single diffuse s to hydrogen zi 6 31G 3DF 3PD extends the 6 31G basis by adding three sets of 5 component d s and one set of 7 component f s to second and third period elements and three sets of p s and one set of d s to hydrogen 1 W J Hehre R Ditchfield J A Pople J Chem Phys 56 2257 1972 2 P C Hariharan J A Pople Theor Chim Acta 28 213 1973 3 J D Dill J A Pople J Chem Phys 62 2921 1975 4 M M Francl W J Pietro W J Hehre J S Binkley M S Gordon D J DeFrees J A Pople J Chem Phys 77 3654 1982 5 M J Frisch J A Pople J S Binkley J Chem Phys 80 3265 1984 6 V Rassolov J A Pople M Ratner T L Windus J Chem Phys 109 1223 1998 B 1 6 6 311G 6 311G 6 311G 6 311 G 6 311 G etc 6 311G is a split valence set with a triple zeta description of the valence orbitals and a minimal basis set description of the core orbitals 6 311G supplements 6 311G with po larization functions This should be used with the SPHERICAL ON option 5d functions Diffuse and polarization references follow from the 6 31G series 109 This basis has increased flexibility in the valence region relative to the
54. 3dxy 3dxz 3dyz atomic orbitals This is not the default in Gaussian which keeps the 3d space active zi The correlation consistent core valence basis sets cc pCVxZ extend the ideas of the original cc pVxZ sets by including extra functions designed for core core and core valence correlation The weighted core valence basis sets 111 cc pwCVxZ weight the KL intershell correlation energy at the expense of the KK correlation zi The extra diffuse nonpolarization functions aug were optimized in Hartree Fock calculations on the lowest state of the anion zi The DK contraction coefficients were generated by performing atomic ROHF calculations using the atomic code ATSCF The version used is an adaption by Bernd Hess of a code by R Pavani G C Lie and E Clementi unpublished The Hamiltonian applied is the so called Douglas Kroll Hess Hamiltonian with the speed of light 137 0360 a u The contraction coefficients were generated for the ground state of each atom Atomic symmetry was imposed in the generation of the contraction coefficients 1 T H Dunning Jr J Chem Phys 90 1007 1989 2 R A Kendall T H Dunning Jr R J Harrison J Chem Phys 96 6796 1992 3 D E Woon T H Dunning Jr J Chem Phys 98 1358 1993 4 D E Woon T H Dunning Jr J Chem Phys 100 2975 1994 5 K A Peterson D E Woon T H Dunning Jr J Chem Phys 100 7410 1994 6 DE Woon T H Dunning Jr J Chem Phys
55. 5 XFORM TOL tol 11 Sets the tolerance for storing transformed integrals Integrals less than 107 are ne glected and not stored on disk NTOP _ TAMP integer 15 Sets the N largest T amplitudes to print for each spin case and excitation level TAMP SUM integer 0 56 Sets how often the largest T amplitudes are printed 0 only prints amplitudes at the beginning and end of the run 1 prints amplitudes after every iteration 2 prints amplitudes after every other iteration etc DENSITY handle RELAXED Specifies whether the relaxed density matrix is computed for correlated wavefunctions RESPONSE skips the orbital relaxation terms that contribute to the density matrix This keyword should only be used by advanced users and ACES II developers who are testing new theories DOHBAR switch OFF Controls what action is taken by the linear response program 0N calculates and saves the full effective Hamiltonian OFF solves the lambda linear response equa tions 8 1 15 Excited states general EXCITE handle NONE Specifies the type of excited state calculation to perform Available values are NONE TDA RPA EOMEE CIS CIS D P EOMEE EOM BWPT2 and STEOM This keyword needs a lot of attention ESTATE TOL tol 5 Sets the tolerance used in converging EOM CC excited state calculations The iterative diagonalization continues until the RMS residual falls below 107 ESTATE_MAX
56. 6 31G basis because it uses three functions to represent each valence AU The original 6 311G set is extended to second row elements using the McLean Chandler basis sets zi Note These basis sets use 5 component d functions zi 6 311 G 2D 2P contains two sets of polarization functions on every atom 6 311G 2DF 2PD contains three sets of polarization functions on every atom 6 311 G 3DF 3PD is the largest Pople style basis set which has been published Note This basis set uses 5 component d and 7 component f functions 1 R Krishnan J S Binkley R Seeger J A Pople J Chem Phys 72 650 1980 2 A D McLean G S Chandler J Chem Phys 72 5639 1980 3 J P Blaudeau M P McGrath L A Curtiss L Radom J Chem Phys 107 5016 1997 4 L A Curtiss M P McGrath J P Blaudeau N E Davis R C Binning Jr L Radom J Chem Phys 103 6104 1995 5 M N Glukhovtsev A Pross M P McGrath L Radom J Chem Phys 103 1878 1995 6 T Clark J Chandrasekhar P v R Schleyer J Comp Chem 4 294 1983 B 1 7 SV DZ TZ SVP DZP CHIPMAN DZ is the well known Dunning double zeta contraction of Huzinaga s 9s5p primitive gaus sian basis set for first row atoms DZP is the DZ set augmented with the polarization functions recommended by Redmon Purvis and Bartlett These were determined from correlated calculations The Dunning Hay split valence basis sets are moderate sized se
57. 91 for more information 5 2 joda Along with parsing ZMAT building the keyword environment and initializing the ACES II file set JODA is responsible for everything between single point calculations Geometry optimizations numerical finite differences and restart capabilities are all handled by this program 5 3 mopac ACES II has a modified version of MOPAC version 5 that can generate an initial guess Hessian for geometry optimizations 5 4 vmol xvmol calculates the one and two electron AO integrals over Gaussian basis functions VMOL was written by J Aml f and P R Taylor and was modified to include an option for effective core potentials 15 5 5 vmol2ja xvmol2ja creates most of the transformation matrices needed to switch between internal VMOL and external ZMAT ordering of atoms and atomic orbitals It also creates the Cartesian Spherical orbital transformations 5 6 vprops xvprops evaluates one electron integrals needed for the calculation of various first order properties such as dipole moment quadrupole moment electrical field gradients or spin densities It originates from POLYATOM and was interfaced to the VMOL integral program by P R Taylor 5 7 nddo xnddo calculates the NDDO density which can be used as an initial guess for the SCF algorithm 5 8 vscf p vscf vscf_ks intgrt and intpack These programs are responsible for generating the Hartree Fock xvscf and Kohn Sham xvscf_ks
58. ACES II Release 2 5 0 User Manual Quantum Theory Project P O Box 118435 University of Florida Gainesville FL 32611 January 31 2006 Contents Contents 1 1 Authors 1T Offcial ACES NEE e a a ata 7 1 2 Spec authors pd a A A Sl A A a ore Pte 7 2 Preface 9 3 Introduction 10 3 1 Overview of capabilities of ACES Il 11 4 Quickstart Guide 14 5 Program Structure 15 Dell AGES2 and P AGES a chess elt a aa ds teen e bats 15 E Te REECH 15 O MOPAG se oS shes A o a 15 Ds MOL a ee een e Ea AY Tel AY Sah 15 Dd NTE ie Gee EE 16 DO IVBROPS ir di as ol ee sho 16 OS NEIE ond TA eee ES tn ee E en E 16 5 8 VSCF P VSOR VSCF_KS INTGRT and INTPACK oda e EEN e 16 5 9 DIRMP2 and P DIRMP2 ts 20 ar db a a Y 16 DOC VIPRANAD e t R ier hie e e ot ante tes een ors a A 16 A RN 16 DLUZANTBROS y at Da 4 8 TR E RS das 6 Seca Nae aer Laa Dans 17 E VCCoT and VOCI s reia ia a ts 17 IMAMRCOO a a we R aw Bae A et eet da 17 DO RNO a A A A A RA A TAT L TR 17 DIO LAMBDA sira a a A ee ete N Al A Se e eet Ae 17 517 VEA and BE Gl oS eet e bl a ES SR 17 SS M aer A a O D ei hn A 17 EEN E ene kee AS en A A Ae im A 18 E a a es gee dts ts a Me na A ay 18 DL CANTE tg dt EE EE 18 REESEN o a a eeh O A a R 18 5 23 VDINT VKSDINT SCFGRD and PSCFGRD 0008 5 18 A A a a a era a AE a a A A A A AN r 18 E a o o pi at Soe ae Go E EEN 18 OLOT AS Ves ad td e he E di EE 19 EE A A ta 19 I2 IOERDIRTVS 2 to rr Gerke de
59. AUSCHLICHERSANO ia ll es MIA eae 115 B 1 18 DGAUSS DZVP DZVP2 TZVP etc DEMON COULOMB 115 B 1 19 AHLRICHS COULOMB Seepage ee ark CARN eA A 115 B L20 GEO MME y ES da a 116 BL ECP sets I ECRDATA E e rt oth a Ae e A 116 B 2 1 HAY WADT MB HAY WADT VDZ 116 B 2 2 LANL2DZ LANL2DZDP y uan a 116 SA O EE ER AE EE 117 B 2 4 CRENBL CRENBS AS AR ee ee 117 B20 T 0 S CI a ia A AER a 118 Bat Basi seu tables a at a A Ue Pal OSG ene i 118 C Queue Scripts 126 Gal ACES EE script body cire ati E a A dee Lee DA E A 126 Coz oad Leveler Li a de le ale Ree oc ee ad 126 Caos ME Sud ai is a ete 126 Cul Grid Engine a a E ae EE EE de 126 Index 128 1 1 1 Users of ACES II should use the following citation when referencing the program system The basis set distributed with the ACES II program system is obtained through Pacific Northwest National Laboratory PNNL and any calculation that uses basis sets from this Authors Official ACES II citation ACES II is a program product of the Quantum Theory Project University of Florida Authors J F Stanton J Gauss S A Perera J D Watts A D Yau M Nooijen N Oliphant P G Szalay W J Lauderdale S R Gwaltney S Beck A Balkov D E Bernholdt K K Baeck P Rozyczko H Sekino C Huber J Pittner W Cencek D Taylor and R J Bartlett Integral packages included are VMOL J Alml f and P R Taylor VPROPS P Taylor ABA CUS T Helgaker H J Aa
60. BILITY and ROT_EVEC keywords Using HFSTABILITY ON will perform a stability analysis after the two electron inte gral transformation and processing step Use of HFSTABILITY ON is compatible with continuing on to correlated calculations in the same job Stability analysis is accomplished by forming the orbital rotation Hessian and diagonalizing it The eigenvalues of this matrix and their associated eigenvectors indicate the number and type of instabilities present Each negative eigenvalue indicates an instability and its magnitude indicates the sever ity Analysis of the eigenvector corresponding to each instability reveals its nature The 74 direct product of the symmetry irreps of the orbitals involved in the rotation determines the symmetry of the instability Only for instabilities whose direct product is the totally symmetric irrep irrep 1 will the wavefunction maintain the symmetry of the molecular framework Any other result means the instability leads at least initially to a symmetry broken wavefunction In some cases symmetry breaking instabilities arise from the presence of lower energy electronic states different occupations and the rotation specified by the corresponding eigenvector will correspond to changing the occupation to relieve the instabil ity For an RHF wavefunction the UHF orbital rotation Hessian is constructed This allows detection of instabilities in which the wavefunction would prefer to be UHF by comparis
61. C integer 20 Sets the maximum number of iterative diagonalization steps for each root in excited state calculations EOM_MAXCYC integer 50 Sets the number of iterations in EOM excited state calculations If it has the value N and NROOT roots have been requested for a given symmetry then the program will allow up to N xN ROOT iterations to find all of the requested roots for that symmetry This keyword is only used by the iterative solution of the HF Stability analysis and has nothing to do with EOM PROGRAM handle DEFAULT 97 EOMREF handle NONE IMEM_SIZE special 3000000W EOM_PRJCT handle NONE NT3EOMEE handle NONE ACC_SYM handle NONE 8 1 16 Excited states properties ESTATE_PROP handle OFF Specifies whether any excited state one electron properties are to be calculated but it only applies to EOM CC calculations Proper use of this keyword might require some fairly advanced knowledge of quantum chemistry and the available options are discussed here The values are OFF no properties or transition moments are calculated EXPECTATION transition moments and dipole strengths are calculated along with selected one electron properties which are evaluated as expectation values UNRELAXED gt selected one electron properties are calculated in an approximation that neglects relaxation of molecular orbitals RESPONSE selected one electron properties are calculated
62. C MBPT gradi ent calculations In addition to integral derivatives with respect to geometrical perturbations it calculates one and two electron integrals required for chemical shift calculations within the GIAO scheme For the most part xvdint calculates the gradient in ACES II For KS reference wavefunctions xvksdint calculates the contribution from the functional deriva tive xscfgrd calculates the HF SCF gradient using the GAMESS direct integral derivative package and xp_scfgrd does this in parallel 5 24 cphf xcphf solves the coupled perturbed Hartree Fock equations either for geometric displace ments or for electric or magnetic field components as perturbations 5 25 nmr xnmr calculates the paramagnetic contribution to NMR chemical shifts at correlated levels 18 5 26 asv An ACES State Variable ASV is a runtime variable that controls the calculation and users can affect ASV initialization with keywords in the ACES2 namelist xasv is not a member executable that other AMEs use but rather a tool for the user to examine the validity of keyword value pairs For example if a user wants to check if CALC LCCD is valid then the user would have to create a valid ZMAT file and run xjoda to guarantee it passes keyword parsing Alternatively the user can run xasv CALC LCCD at the shell prompt without the hassle of managing files 5 27 a2proc This program was initially created to reduce the clutter of member executabl
63. CSD T energy calculation on the water molecule with the DZP basis set This input is particularly simple since all of the keywords default to the closed shell singlet state of the neutral molecule In other words CHARGE 0 MULTIPLICITY 1 and REFERENCE RHF can all be omitted The OCCUPATION string is not specified which means the program will determine this itself By default all electrons are correlated 9 1 2 UHF CCSD T energy H20 UHF CCSD T energy calculation Mixed basis set H O 1 ROH H 2 ROH 1 HOH ROH 0 95 HOH 104 5 ACES2 REF UHF CALC CCSD T OCCUPATION 3 1 1 0 2 1 1 0 H DZ 0 DZP H DZ This illustrates a single point CCSD T energy calculation on the 7A state of the water cation with a special basis set For this system one cannot rely on the defaults that were used in the previous example It is necessary to specify REFERENCE and either OCCU PATION or both CHARGE and MULTIPLICITY Since we want a particular state we use the OCCUPATION keyword The ACES2 namelist is valid even without the initial and final parentheses however if an open parenthesis is used to start the namelist then it must terminate with a close parenthesis 69 9 1 3 ROHF CCSD T energy H20 ROHF CCSD T energy calculation DZP basis set H O 1 ROH H 2 ROH 1 HOH ROH 0 95 HOH 104 5 ACES2 REF ROHF CALC CCSD T BASIS DZP OCCUPATION 3 1 1 0 2 1 1 0 This is essentially identical to the previous example The difference
64. EF TWODET two determinant refer ence for open shell singlet CC calculations is available but the orbitals are obtained from a closed shell state and are not optimum open shell singlet or low spin triplet SCF orbitals The orbital occupancy for subsequent CCSD calculations is specified by the various QRHF keywords NONHE switch OFF Signals the correlation energy code if the reference wavefunction satisfies the Hartree Fock equations E 0 Usually there is no need to set this keyword since standard non HF reference functions QRHF and ROHF set this flag internally however if you input a set of orbitals to the correlation energy code directly then it might be necessary to set NONHF ON ORBITALS handle 1 Specifies semicanonical orbitals Fiz Fiz 0 in non HF calculations Semi canonical orbitals are obtained by diagonalizing the occupied occupied 77 and virtual virtual ab blocks of the spin orbital Fock matrix and can be advantageously exploited in certain post SCF calculations This is particularly true for ROHF MBPT and non iterative triple excitation corrections There is no default value for this parameter and considerable logic is used to determine the orbital type in post SCF non HF cal culations if the keyword is not defined It is strongly recommended that this key word not be used by anyone who is not thoroughly familiar with non HF CC MBPT methods since the logic used to set the default value
65. I 8 Sr CHG ZAZINVT 7 e e SS ZL AY H g ZI e e SC Et aIN LAVM AVH sjyos siseq GOW st 8g ezez tos Io 3 0 Ce 13 96 rop Ion aINOTAOO SHOIY THY 0700 0700 oo aINOTAOO NOWAG s oo ADNVHOXA 2V SSAVOA e oo ENOTNOD ZV SSAVDA 68 0700 oo ADNVHOXH TV SSAVIDA 68 0700 oo ENNOTNOD TV SSAVDA e s 6 Zd AZG SSAVOG Ce eee 6 9 JAAZA SSAVDA s s s 2 601 ONV AAHOTTHOSAVE s s 0169 e Gout e oreo e SHINV VSVN e 97 611 AZIV SOOH s mon AZAV SOOH s s 2 09 AtSYALHOVM Weg e Op e e ZLAd ATAVS Wert 1 61 1 97 SALM us gu a ua as sv ao vo NZ no IN 00 aq NN T wo aA I os tee 8g dy PE gan UI syuotuo e SG THO pg 10 SUOTJOUN J siseq OY poprsyuoo jo Joqunu oy 17u02 z AQEL 124 s 8e Lt e Cer L6OSHU LS 8 zm et 8 e OTH LS e e 01 SENAHO zg To rer TENHYHO 8 8 re rg OLMES e e Eer e e JAUZAZINV1 zI 8 0 re ZAZINVT zI e e re ZdA LAVM AVH Z e e TE SIL YM AY H 9109 siseq GOW st g9 ERS 11 19 Mer tiros Eg Sp ti 6p 83 16 v3 26 amp 16 got v3 26 vo 96 1 16 13 96 so ENOTAOO SHOITTHV e op op aWOTNOO NOWAC e op op HONVHOXH TV SSAVDA e op op ENOTAOO TV SSAVDA e Cep Cer JAZA SSAVDA e e tr e e ZLAd LATAVS 0 6 9 o 9 SALM N N 2719 e N xD TTE 9 s e S 19 e
66. If BA SIS SPECIAL then each atom in coordinate order must be specified after the ACES2 namelist SPHERICAL switch OFF Controls whether 0N spherical harmonic 5d 7f 9g or OFF Cartesian 6d 10f 15g basis functions are used LINDEP_TOL tol 5 Sets the tolerance for linear dependencies in the basis set The basis set is considered linearly dependent and eigenvectors of the overlap matrix are neglected if the associated eigenvalues are less than 107 45 GENBAS 1 integer 0 This keyword applies to first row elements H and HE and specifies the number of contracted Gaussian functions per shell There is usually no need to use this keyword but it can be useful for using a subset of the functions in a particular entry in the GENBAS file particularly for generally contracted basis sets For example if entry H BASIS in the GENBAS file contains 7 contracted s functions 4 p functions and one d function then setting GENBAS_1 730 would eliminate the last p function and the d function The default for this keyword is to use the unaltered entry in GENBAS GENBAS_2 integer 0 This keyword performs the same function as GENBAS_1 but applies only to second row atoms GENBAS_3 integer 0 This keyword performs the same function as GENBAS_1 and GENBAS_2 but applies only to third row atoms CONTRACTION handle GENERAL Specifies the contraction scheme used by the integral and integral derivative programs
67. JAI Press p 139 1991 e J Gauss and D Cremer Adv Quant Chem 23 205 1992 a General MBPT CC gradient theory 99 e E A Salter G W Trucks and R J Bartlett J Chem Phys 90 1752 1989 b Implementation for closed and open shell CCSD e J Gauss J F Stanton and R J Bartlett J Chem Phys 95 2623 1991 c QRHF CCSD gradients e J Gauss J F Stanton and R J Bartlett J Chem Phys 95 2639 1991 d ROHF CCSD gradients e J Gauss W J Lauderdale J F Stanton J D Watts and R J Bartlett Chem Phys Lett 182 207 1991 e QCISD gradients e J Gauss and D Cremer Chem Phys Lett 150 280 1988 e J Gauss J F Stanton and R J Bartlett J Chem Phys 95 2623 1991 f MBPT 4 gradients for closed and open shells e J Gauss and D Cremer Chem Phys Lett 138 131 1987 153 303 1988 e G W Trucks J D Watts E A Salter and R J Bartlett Chem Phys Lett 153 490 1988 e J D Watts G W Trucks and R J Bartlett Chem Phys Lett 164 502 1989 e J D Watts J Gauss and R J Bartlett Chem Phys Lett 200 1 1992 g CCSD T CCSD CCSD T gradients for closed and open shells e J D Watts J Gauss and R J Bartlett Chem Phys Lett 200 1 1992 e J D Watts J Gauss and R J Bartlett J Chem Phys 98 8718 1993 h QCISD T gradients e J Gauss and D Cremer Chem Phys Lett 163 549 1990 e J D Watts J Gauss and R J Bartlett Chem Phys Lett 200 1 1992 i UCC 4 gradi
68. L handle RFA Specifies the behavior if negative eigenvalues are encountered in the totally symmetric Hessian during an NR or MANR search If NEGEVAL ABORT then the job will terminate with an error message If NEGEVAL ABSVAL then the program will just switch the eigenvalue to its absolute value and keep plugging away this is strongly discouraged If NEGEVAL RFA then OPT_METHOD is switched to RFA internally and the optimization is continued 8 1 24 Geometry optimization iteration control CONVERGENCE tol 4 Sets the convergence criterion for geometry optimizations The optimization terminates when the RMS gradient is below 107 Hartree Bohr OPT MAXCYC integer 50 Sets the maximum number of optimization cycles INIT_HESSIAN handle SPECIAL Specifies the initial Hessian for geometry optimizations SPECIAL generates an approximate Hessian internally FCMINT imports an approximate Hessian from the FCMINT file MOPAC generates the FCMINT file by running MOPAC and EXACT uses analytical second derivatives to generate an exact Hessian limited to SCF HESS_UPDATE handle POWELL or BOFILL Specifies the algorithm used to update the Hessian NONE uses the same Hessian throughout the optimization POWELL uses the Powell update BFGS uses the Broyden Fletcher Goldfarb Shanno update MS uses the Murtagh Sargent update BOFILL uses the Bofill mixture of Powell and Murtagh Saregent and PSB uses the Powell symmetri
69. MOL integral program will be incorrect Lines 10 through 13 repeat NS number of times An example of the boron PVTZ basis set entry is included below 36 B PVTZ JFS DUNNING CORRELATION CONSISTENT BASIS FROM FTP 4 0 1 2 3 4 3 2 1 10 5 2 1 5473 0000000 820 9000000 186 8000000 52 8300000 17 0800000 5 9990000 2 2080000 0 5879000 0 2415000 0 0861000 0 0005550 0 0001120 0 0000000 0 0000000 0 0042910 0 0008680 0 0000000 0 0000000 0 0219490 0 0044840 0 0000000 0 0000000 0 0844410 0 0176830 0 0000000 0 0000000 0 2385570 0 0536390 0 0000000 0 0000000 0 4350720 0 1190050 0 0000000 0 0000000 0 3419550 0 1658240 0 0000000 0 0000000 0 0368560 0 1201070 1 0000000 0 0000000 0 0095450 0 5959810 0 0000000 0 0000000 0 0023680 0 4110210 0 0000000 1 0000000 12 0500000 2 6130000 0 7475000 0 2385000 0 0769800 0 0131180 0 0000000 0 0000000 0 0798960 0 0000000 0 0000000 0 2772750 0 0000000 0 0000000 0 5042700 1 0000000 0 0000000 0 3536800 0 0000000 1 0000000 0 6610000 0 1990000 1 0000000 0 0000000 0 0000000 1 0000000 0 4900000 1 0000000 7 3 ECPDATA The parameters of the effective core potentials are specified in a format similar to GENBAS For example entries for copper are given below CU CRENBL ECP from CRENBL ECP 10 CORE ELECTRONS x NCORE 10 LMAX 2 4 73227501 2 16 30159950 34 06355667 2 49 98759842 90 69224548 2 173 02969360 10 26460838 1 651 10559082 S D 87 13957977 209 05120850 3 70869994 4 51
70. MR chemical shift calculations e J Gauss Chem Phys Lett 191 614 1992 J Chem Phys 99 3629 1993 12 6 Methods for calculating excitation energies a Tamm Dancoff CI singles approximation b Random Phase approximation RPA c Equation of Motion Coupled Cluster EOM CC methods e J F Stanton and R J Bartlett J Chem Phys 98 7029 1993 101 12 7 Methods for calculating electron attachment energies The electron affinity equation of motion coupled cluster method e M Nooijen and R J Bartlett J Chem Phys 102 3629 1995 12 8 Time dependent Hartree Fock methods e H Sekino and R J Bartlett J Chem Phys 85 976 1986 e H Sekino and R J Bartlett Int J Quantum Chem 43 119 1992 12 9 HF DFT method e P M W Gill B G Johnson and J A Pople Int J Quantum Chem Symp 26 319 1992 e G E Scuseria J Chem Phys 97 7528 1992 e N Oliphant and R J Bartlett J Chem Phys 100 6550 1994 12 10 Integral packages The direct integrals are obtained by Rys quadrature 1 using an implementation ex tended to spdfg and L shells taken from GAMESS 2 1 J Rys M Dupuis H F King J Comput Chem 4 154 157 1983 2 M W Schmidt K K Baldridge J A Boatz S T Elbert M S Gordon J J Jensen S Koseki N Matsunaga K A Nguyen S Su T L Windus M Dupuis J A Mont gomery J Comput Chem 14 1347 1363 1993 102 A Other Keywords A 1 Experimental
71. RENCE RHF CALC CCSD BASIS TZ2P CHARGE 1 MULT 1 SPHERICAL ON DROPMO 1 2 EA_CALC EA_EOMCC EA_SYM 5 3 0 1 79 3 The calculation of high spin triplet states for systems with a closed shell ground state Take as a CC reference the high spin doublet ground state of the positive ion and add an extra alpha electron by specifying EA SYM e g EA SYM 4 3 0 2 0 0 0 0 This yields high spin triplet excited states of the neutral In addition the closed shell ground state can be obtained by adding a beta electron to the proper symmetry block e g EASY M 4 3 0 2 1 0 0 0 This has the advantage that the proper excitation energies of the system are tabulated by the program Singlet and low spin triplet excited state energies can also be obtained by adding a beta electron however such calculations do not yield satisfactory results due to spin contamination of the resulting EA_EOMCC states The following input yields triplet excited states for the beryllium atom The SCF calculation is on the closed shell neutral system and the QRHF option is used to create the positive Be ion BE Atom Excitation spectrum QRHF Reference BE ACES2 REFERENCE RHF CALC CCSD BASIS WMR SPHERICAL ON QRHF_G 1 EA_CALC EA_EOMCC EA_SYM 8 4 0 4 0 0 0 0 0 0 0 0 0 0 0 0 WARNING In all EOMCC calculations it is highly recommended that the reference state transforms according to a one dimensional representation of the true molecular point group Otherwis
72. The default value 1 determines a size based on the value of MEMORY_SIZE CACHE RECS integer 1 Sets the number of physical records of length FILE_RECSIZE held in the I O cache The maximum number of records is 128 and the default value 1 determines a number based on the value of MEMORY SIZE INCORE handle NONE 44 This keyword can be used to reduce disk I O usually by a significant amount Ac ceptable values are NONE NOABCD NOABCI NOABIJ ALL and T NONE and ALL load no and all lists into core memory respectively NOABCD NOABCI and NOABIJ load all lists that are at most the size of Ab C7 Ab Ij and Ij Ka molecular integral lists respectively These options are implemented in VCC LAMBDA DENS ANTI BCKTRN and VEE The T option only applies to VCC which will load T and T amplitudes into memory All of the other executables will treat T as NONE including LAMBDA although this should change soon 8 1 5 Chemical system COORDINATES handle AUTO Specifies whether the molecular system is defined in INTERNAL Z matrix or CARTE SIAN xyz coordinates UNITS handle ANGSTROM Specifies the units used for molecular geometries The value can be ANGSTROM or BOHR CHARGE integer 0 Sets the molecular charge MULTIPLICTY integer 1 Sets the spin multiplicity 25 1 8 1 6 Basis set BASIS string SPECIAL Defines the name of the basis set to use for all atoms in the molecule
73. VDA e Le e e 1 61 e e 9 dAZL SSAVDG s 1 61 e e Ier 1 21 E zdAZC SSAVOG ze Ire 1 61 1 61 Ter We Z JAAZU SSAVDA e e e e s ONV YAHOTTHOSNVA e LT e TA e ds e e Ire SHINV VSVN e 6 69 6 69 6 99 6 99 Cor AZIV SOOY Ce 6 6 Sos Sos 6 AZAV SOOY e e e e e s AFSHHLHOVM E N PTC AS N 29 amp 9 N 0 ZLAd HIAVS OT 6 9 E S T e SLL s e e e e e HHDOTALU Vd TZ El 90 99 1g ST e E HADATHIHVA ro 19 zs er 9T e cA CIN VA ES e 8h mm rr e IESSE E 1 9 19 9 C C TOd SHOTYTHV er D IT rr E e AZI SHOTWTHV CZ zz er er 9 C ZLA SHOTH THV rr 1 61 oT We 6 9 g ZdAd SHOIN THV I ET L 6 Z ZAA SHOIN THV s e e 1 07 We e e 9 ZLAA SSANVD e TG TZ D vr e s ZLA SSHNVO vo a uv to s a fis 1v ow vw AN d o njolal ag T aH H ats sy de SC dz SZ ST SVANAD Ul suog sp YSno1yy s JO SUOTJOUNJ Step OY Payoe1juos jo Jequinu oy 3409 9 qeL 121 8 D 6ISHULS Sea TI H Tp LL OTY LS SENAO 9T 9T TANAYUO o0 00 BIER s TAT 1 81 s dUZUZ INV I 6 ZU C INN ZIA LUVM AVH eje o e AR AO IN LAVM AVH 9109 siseq GOW Vo Sv uv 109 s de d rel tv om SVANAD Ul suog sp y3nOIyy s se VN HN A dz o Nnjo la ag SZ TI JO SUOTJOUNJ Step OY Payoe1juos jo Joquinu aq 3409 f equ HH ST 398 sise
74. Y WADT VDZ The Hay Wadt n 1 effective core potentials include an extra shell of electrons beyond what is traditionally available in effective core potentials For example on potassium the 3s and 3p electrons are not subsumed into the core as they would be in other ECP s This set is derived from the exponents and contraction coefficients given in the Hay Wadt paper and were obtained directly from P J Hay Any discrepancy between the numbers contained in these basis sets and the JCP paper are attributable to the file obtained from Jeff at Los Alamos National Laboratory The elements beyond Kr include the 1 electron Darwin and mass velocity relativistic corrections in their definitions 1 P J Hay W R Wadt J Chem Phys 82 299 1985 B 2 2 LANL2DZ LANL2DZDP This basis set family is a collection of all electron and ECP sets designed to mimic the LANL2DZ family in Gaussian 9x 116 1 T H Dunning Jr P J Hay in Methods of Electronic Structure Theory Vol 2 edited by H F Schaefer III Plenum Press 1977 2 P J Hay W R Wadt J Chem Phys 82 270 1985 3 P J Hay W R Wadt J Chem Phys 82 284 1985 4 P J Hay W R Wadt J Chem Phys 82 299 1985 5 J V Ortiz P J Hay R L Martin JACS 114 2736 1992 6 C E Check T O Faust J M Bailey B J Wright T M Gilbert L S Sunderlin J Phys Chem A 105 8111 2001 B 2 3 SBKJC The Stevens Basch Krauss Jasien Cundari 21G relativistic ECP
75. acters but it is usually included as a reminder of the options The input is as follows NIRREP is the order of the computational point group or 4 in the present case and NSPIN is the number of spins 1 for RHF 2 for UHF and ROHF Line 1 A80 A title Next NSPIN NIRREP 13 The alpha beta occupation vector Next NIRREP NSPIN 4 13 Pairs of orbitals to be swapped in each spatial symmetry block for each spin sym metry Two numbers are needed to specify each pair therefore no more than two interchanges may be made for a given symmetry block and spin Next NSPIN NIRREP 13 Symmetry block occupation lock flags Orbital occupation proceeds in the direction of 39 minimum change by monitoring CZ n S C Zero is unlocked a positive integer is locked Next NSPIN NIRREP 13 Print flags for alpha beta initial guess A positive integer prints the guess for that symmetry block while a zero does not Next 1 2 13 Stopping parameters Set the first value to a positive integer to stop the SCF after computing the initial guess Next 1 13 I O parameter If set to a positive integer then the initial guess MOS are read from OLDMOS Next 1 13 UHF creation parameter which copies the alpha MOs to the beta MOs This is only meaningful if the guess is read from OLDMOS This allows a user for example to start a UHF calculation with an RHF closed shell set of orbitals A positive integer reads only the alp
76. allows field strengths E gt 1078 to be used YFIELD integer 0 Sets the Y component of an external electric field See above ZFIELD integer 0 Sets the Z component of an external electric field See above PRP_INTS handle PARTIAL Specifies the types of property integrals that are computed Setting this to FULL or PARTIAL computes the full set or sub set respectively This works in conjunction with the PROPS keyword and the defaults are set automatically 64 8 1 22 Geometry optimization general GEOM_OPT handle NONE Specifies the scope of coordinate optimizations This keyword is automatically set to PARTIAL if a geometry optimization is implied with asterisks in the internal coordi nate Z matrix Setting GEOM_OPT FULL will optimize all coordinates regardless of asterisks in the Z matrix COORD INTERNAL and will optimize a Cartesian input geometry using Redundant Internal Coordinates RICs CURVILINEAR handle OFF Specifies whether or not the Hessian matrix is transformed nonlinearly to curvilinear internal coordinates OFF turns the transformation off if analytical force constants are not available while it is always performed if CURVILINEAR ON NO uncondi tionally turns the transformation off This keyword is set automatically 8 1 23 Geometry optimization stepping algorithm OPT METHOD handle AUTO Specifies the geometry optimization strategy AUTO NR uses a straightforward Newt
77. ally preceded by geometry optimizations which require the asterisks 85 9 3 2 Numerical frequencies from energies CCSD TQf NUMERICAL VIBRATION CALCULATION FOR N2 N NIR R 1 116 ACES2 CALC CCSD TQf BASIS DZP VIB FINDIF GRAD_CALC NUMERICAL This example specifies a finite difference frequency calculation for N using energy points from the CCSD TQf method This method of calculating frequencies is applicable to all types of energy calculations not just CCSD TQf 9 3 3 Isotopic shift ACES II provides a straightforward way to calculate changes in harmonic vibrational frequencies and infrared intensities due to isotopic substitutions This is accomplished with a free format file called ISOMASS which contains the desired atomic masses in their ZMAT order excluding dummy atoms For example if a user wants to calculate the 0 180 isotopic shift for the vibrational frequencies of water then the following ZMAT and ISOMASS files may be used ZMAT Water frequency calculation ACES2 CALC SCF BASIS DZP VIB EXACT ISOMASS 18 0 1 00797 1 00797 86 10 Parallelization 10 1 Overview ACES II can perform the following calculations in parallel 1 HF SCF and MBPT 2 single point energies 2 HF SCF analytical geometry optimizations and 3 all finite displacement methods numerical geometry optimizations and vibrational fre quency calculations The SCF and MBPT 2 energy programs xp vscf and xp_dirmp2
78. almqvist Persson and Roos are based on atomic valence SD CI s However in contrast to the ANO contractions of Almlof and Taylor these contractions were based not just on the atomic ground states but on the positive and negative ions and electric field polarized atom as well ek Additional low lying electronic states were included for some atoms For example the singlet D state of oxygen was included in the averaging process ek The Gaussian primitive set for first row elements is based on the van Duijneveldt 13s 8p set augmented with a extra shell of diffuse functions The second row sets were derived in an analogous fashion from a 16s 11p set The primitives for the transition metals were based on the 20s 12p 9d set set of H Partridge 1 H Partridge J Chem Phys 90 1043 1989 2 P O Widmark P A Malmqvist B O Roos Theor Chim Acta 77 291 1990 3 P O Widmark B J Persson B Roos Theor Chim Acta 79 419 1991 4 R Pou Amerigo M Merchan I Nebot Gil P O Widmark B Roos Theor Chim Acta 92 149 1995 114 B 1 16 NASA AMES The atomic natural orbital basis sets of Almlof Taylor and coworkers are based on atomic valence SD CI densities or densities derived from state averaged MR SD CI calculations The aim was to derive generally contracted sets with large numbers of Gaussian primitives without significant loss in the SCF or correlation energy Choices of what mixture of s p d f g funct
79. als This works in conjunction with the ABCDTYPE key word When ABCDTYPE AOBASIS the ABCD integrals are not stored on disk and GAMMA_ABCD is set to DIRECT When ABCD integrals are stored to disk GAMMA_ABCD must be set to DISK 8 1 14 Post SCF calculations CC_CONV tol 7 Sets the convergence criterion for the CC and Lambda equations Equations are con sidered converged when the maximum change in amplitudes is less than 107 CC_MAXCYC integer 50 Sets the maximum number of CC or Lambda iterations CC_EXTRAPOL previously RLE handle DIIS Specifies the type of convergence acceleration used during the CC iterations STAN DARD uses the RLE method of Bartlett and Purvis with periodic extrapolation of the solution vector DIIS uses the DHS approach of Pulay NOJACOBI uses the RLE method with continuous extrapolation and OFF uses no convergence acceleration In general DIIS works well for a wide range of molecular systems while RLE can be better for certain cases NOJACOBI might offer advantages in cases where the reduced subspace becomes singular too rapidly NOJACOBI requires some additional disk stor age which might be disadvantageous for very large calculations OFF is generally a bad idea for CC calculations but might be preferred by some CI calculations CC_EXPORDER previously ORDER_RLE integer 5 Sets the maximum number of iterates to include in the R matrix used by RLE and DIIS The maximum value allowed is 2
80. as analytic first deriva tives of the energy Except for EOMCC calculations on two electron systems which are exact proper ties obtained by the three approaches will not be equivalent The default value for this keyword is slightly complicated For TDA calculations the default is EXPEC TATION since the evaluation of transition moments involves only a negligible amount of additional computation relative to the evaluation of the excitation energies For EOMCC the default is OFF since evaluation of any transition moments or properties requires approximately twice the computational time ESTATE_PROP RESPONSE is not available for EOMCC calculations Transition moments and dipole strengths are evaluated by default for all values of ESTATE_PROP other than OFF 58 8 1 17 Excited states affinities EA_CALC handle NONE Specifies the method for calculating electron affinities of a closed shell parent state The values are NONE MBPT 2 strict second order perturbation theory SO_DYSON second order Green s Function iterating the MBPT 2 formula OVGF P_EOMEA partitioned EA EOM the 2p1h 2p1h block is treated as diagonal EA EOMCC H bar is diagonalized over 1p and 2p1h configurations COMBO OS_CCSD a state selective multireference CC method in which both orbitals and cluster amplitudes are optimized for the final state of interest This last method requires additional input in the mrcc_gen namelist Analytical gradien
81. by xvtran and store partially transformed two electron integrals for use in xintprc Usually they are deleted by xintprc unless a particular post SCF option requires them such as ABCDTYPE MULTIPASS 6 10 8 IUHF This file contains the RHF UHF flag Its use is limited and it should disappear in a future release 6 10 9 TGUESS LGUESS These files store the coupled cluster T and A amplitudes respectively for restart pur poses 6 10 10 VPOUT This file contains the first order property integrals 6 10 11 GAMESS LOG MP2 LOG DIRGRD LOG All of these files are generated by the GAMESS interface in VSCF DIRMP2 and SCFGRD They are strictly log files and might contain error messages if a program crashes 24 6 10 12 0UT 000 DUMP 000 1ELGRAD 000 All of these files are generated by the GAMESS interface in VSCF DIRMP2 and SCFGRD and are tagged with the MPI rank of each process They are strictly output files and might contain error messages if a program crashes 25 7 File Formats 7 1 ZMAT 7 1 1 File anatomy This specification pertains to a particular version of JODA Older versions might require more rigid input formats but the general structure of ZMAT should never change Every line in ZMAT can be at most 80 characters long The parser does not check the length of each line therefore there are no guarantees that the rules of ZMAT parsing will apply once the line length has been exceeded 1 Header Three types of li
82. c Broyden update POWELL and BOFILL are the defaults for minimum energy and transition state searches respectively EVAL_HESS integer 0 Sets the cycle interval for recomputing the Hessian For correlated calculations the Hessian is evaluated only at the SCF level 66 8 1 25 Geometry optimization integral derivatives TRANS INV handle USE Specifies whether or not translational invariance is exploited in derivative calculations USE uses translational invariance while IGNORE does not 8 1 26 Frequencies and other 2 order properties VIBRATION handle NO Specifies the method for calculating vibrational frequencies EXACT SCF only performs normal mode analysis on an analytic force constant matrix and computes rotationally projected frequencies and infrared intensities FINDIF signals ACES II to compute the force constant matrix by finite difference of analytically computed gradients or energies using symmetry adapted mass weighted Cartesian coordinates RAMAN switch OFF Controls raman intensities and depolarization ratios of numerical vibrational frequency calculations VIB FINDIF The Raman intensities are limited to the SCF and CCSD level via EOM CCSD and as a result RAMAN keyword must be used in conjunction with CALC SCF or CALC CCSD We recommend setting EOMPROP QUADRATIC for raman intensities at the CCSD level NOTE This method requires an extra input file named frequency which is de
83. d g iga The scaling factor is 1024 not 1000 Doubles are not used currently even though we have this capability there are no key words with values of this type Examples would be doublel 1 double2 2 d0 double3 3 5e 6 Version 2 3 introduced environment variable awareness for keyword values It is now possible to enter a value as VARNAME and have xjoda pull the value from the shell environment The most practical application of this would be to loop over variables in a shell script that define various keywords Here is an example gt cat lt lt EOF gt ZMAT envvar test job H1R R 0 7 ACES2 calc CALC basis BASIS EOF gt for CALC in scf ccsd gt do export CALC for BASIS in DZP TZP TZ2P do export BASIS clean or other cleaning script xaces2 gt CALC BASIS out done N N N N N 35 7 1 10 Line item basis ECP definitions If the BASIS keyword is set to SPECIAL the default then the basis set specification will be read directly after the ACES2 keyword list One blank line must separate the last line of the keyword list from the beginning of the basis set input section Each entry must be placed on an individual line and the ordering of atoms must follow the Z matrix ordering exactly The names will then be used to find the definitions in the basis set file either GENBAS or ZMAT BAS in the current directory If a basis set is not found ACES II exits with an error condition
84. d equation of motion CCSD calculations of excitation energies S R Gwalt ney and M Nooijen e Equation of motion CCSD gradient calculations for excited states J F Stanton and J Gauss 2 Preface This manual is maintained as a set of BTFX documents under CVS control along with the ACES II source code Every version of the code has an accompanying manual and while the user interface is relatively stable it is not guaranteed that the manual of one version is compatible with binary executables of a different version Any errors in the manual should be reported to aces2 qtp ufl edu The goal of this manual is to provide beginners and experts alike with enough information to run any calculation that the software allows A reference section is provided for further reading on the theories and algorithms implemented in the program 3 Introduction ACES II is a set of programs that performs ab initio quantum chemistry calculations The package has a high degree of flexibility and supports many kinds of calculations at a number of levels of theory The major strength of the program system is using many body methods to treat electron correlation These approaches broadly categorized as many body perturbation theory MBPT and the coupled cluster CC approximation offer a reliable treatment of correlation and have the attractive property of size extensivity which means the energies scale properly with the size of the system As a result of
85. d with this keyword must be separated by forward slashes and be positive or parenthesized negative The absolute value of each element specifies the symmetry block involved in the addition or removal of electrons The numerical ordering of the symmetry blocks is consistent with then OCCUPATION keyword the symmetries of the integrals By default the electrons of each irrep are added to the lowest unoccupied o orbital in the symmetry block and removed from the highest occupied orbital Different orbitals and spins can be specified with the QRHF_ORBITAL and QRHF_SPIN keywords 52 NOTE Gradients and property calculations are currently available only for cases in volving addition or removal of electrons Mixed cases involving both processes are not available Gradients and properties are available for open shell singlet CCSD wave functions for the case that the open shell orbitals have different symmetries 22 QRHF_ ORBITAL string of 1D array 1 By default in QRHF calculations electrons are removed from the highest occupied 7 orbital in a symmetry block symmetry block HOMO while electrons are added to the lowest unoccupied a orbital within a symmetry block symmetry block LUMO The purpose of the QRHF_ORBITAL keyword is to allow additional flexibility in choosing which orbitals will have their occupation numbers altered The value of this keyword gives the index with respect to the default orbital for the orbital which will be
86. dat 59 8 1 18 Excited states electronic absorption 60 8 1 19 Excited states ionizations 61 8 1 20 Excited states gradients les a ek ea eo A Ss 62 al UP POP CRUIES 9 H fy ho e i Sa he thnk YE ee Ae hee Gace A hn Gee ye Canis er Sh 63 8 1 22 Geometry optimization general 65 8 1 23 Geometry optimization stepping algorithm 65 8 1 24 Geometry optimization iteration control 66 8 1 25 Geometry optimization integral derivatives 67 8 1 26 Frequencies and other 2 4 order properties 67 8 1 27 Finite displacements d i te duc i Ants ne ete Gh wien a A ee 67 8 1 28 External interfaces 4 ete e eb ts i Om Ss ot em ace tte 68 9 Examples 69 9 1 Single point calculations 4 4 2 Leen de eke ala ek ii te oe 69 gi RAR GOSD en eee da we 69 9 1 2 UHE CCSD T energy ds A A E a 69 9 1 3 ROME CCSD T energy ere ia ind 70 9 1 4 QRHF CCSD T energy A o 70 9 1 5 Effective core potentials 2 0 0 ee ee 71 9 1 6 Initial guessing with OLDMOS 71 9 1 7 Initial guessing with OLDAOMOS e 72 9 1 8 Improving SCF convergence a 73 9 1 9 Hartree Fock stability analysis 2 2 26 eee Ee 1 74 9 1 10 Time dependent Hartree Fock 76 9 1 11 EOM CCSD excitation energy aes doe ee el a ee EE 78 9 1 12 EOM CCSD electron attachment energy 78 9 2 Geometry optimizations goa eke he a i eve es A 80 9 2 1 Full optimization of interna
87. default scheme the user has a number of options In RHF closed shell and UHF open shell calculations difficult cases will often converge with the use of a dynamical damping algo rithm due to E R Davidson This is achieved through the option DAMP_TYP DAVIDSON Damping serves to prevent excessively large oscillations in the early iterations Once the SCF convergence appears to be sufficiently smooth the damp factor is smaller than DAMP_TOL and the energy difference is sufficiently small the program reverts to repeated diagonaliza tion and DIIS extrapolation For ROHF calculations the level shifting technique can be particularly useful in addition to the damping algorithm In this scheme one adds a positive number a to all diagonal elements of the singly occupied orbital block of the Fock matrix and a positive number a 8 to the diagonal elements of the virtual orbital block of the MO basis Fock matrix a and P are set by the LSHF_A1 and LSHF_B1 keywords If one wishes to use a value 0 2 a u for the level shifters LSHF_A1 and LSHF_B1 should be set to 20 This is a reasonable value for most systems Larger values of level shifters are sometimes necessary for transition metal systems especially when the singly occupied orbitals lie below some of the double occupied orbitals An example is provided by the following FeCl input excitation energies of this system were studied by N Oliphant and R J Bartlett J Am Chem Soc 116 4091 1994
88. e DTIE 9 e e et e e DTIE 9 e Iris ee Cep DIZE e e Cor e e Sg e Ir e s IAIN e e 67 67 DE OLS va so 4X I aL as ns NI ao ov da Hu na oL ow an T Z T EE s9 dq PV SVENAD U SyUSUIETA sg YSNoIY pp 10 SUOCTJOUNJ sIseq OY popoRIyWOS Jo JoquInu AL 9ABL 125 C Queue Scripts While nothing prevents ACES II programs from running interactively or in the back ground many computing facilities require large batch jobs to be executed by automated queueing systems Documenting every scheduler that users might encounter is far beyond the scope of this manual but the following sections list the most common options that the developers have used C 1 ACES II script body bin sh zmt ZMAT_FILE out 0UT_FILE genbas GENBAS_FILE workdir WORK_DIRECTORY test d workdir amp amp rmwd 0 rmwd 1 test rmwd eq 1 amp amp mkdir p workdir cd workdir cp zmt ZMAT cp genbas GENBAS xaces2 gt out cd test rmwd ne 0 amp amp rm rf workdir C 2 LoadLeveler output STDOUT_FILE error STDERR_FILE class CLASS job_type JOB_TYPE node_usage USAGE_TYPE node NNODES total_tasks NTASKS network MPI css0 shared us Requirements OPT1 VALUE1 E OPT2 VALUE2 wall_clock_limit HH MM SS C 3 LSF BSUB P ACCOUNT BSUB J JOBNAME BSUB o STDOUT_FILE BSUB e STDERR_FILE BSUB q QUEUE W TIME BSUB n NPES C
89. e the likely outcome is inaccurate results which in addition are very hard to interpret due to symmetry breaking 9 2 Geometry optimizations 9 2 1 Full optimization of internal coordinates H20 CCSD optimization H O 1 ROH H 2 ROH 1 HOH ROH 0 95 HOH 104 5 ACES2 CALC CCSD BASIS DZP This specifies a geometry optimization of the water molecule with the CCSD method and the DZP basis set The input is exactly the same as the RHF single point energy example except that an asterisk is placed after each variable to be optimized both bond lengths 80 and the bond angle Since all of the parameters are being optimized it is also sufficient to use GEOM_OPT FULL and leave the asterisks out of the Z matrix Defaults for all of the optimization keywords e g METHOD and CONVERGENCE are used The presence of the asterisks promotes the GEOM_OPT keyword from NONE to PARTIAL which turns on all appropriate derivative keywords Geometry optimizations by default are performed using analytical gradients If analytical gradients are not available one can do geometry optimizations from energies using GRAD_CALC NUMERICAL 9 2 2 Partial optimization of internal coordinates Beryllium borohydride D3d structure geometry optimization X BE 1 R1 B 2 R 1 A B 2 R 1 A3T X 2 RX 1 AST X 2 RX 1 A4T H 5 RHX 2 A1T H 5 RHX 2 A 1 T60 H 5 RHX 2 A 1 TM6 H 6 RHX 2 A 1 T12 H 6 RHX 2 A 1 TM2 H 6 RHX 2 A 1 TO X 3R1 2 A 1 TO
90. e fourth block This keyword is the same as HE SYM but EE_SYM is read when PROGRAM ACES3 and ESTATE SYM is read when PROGRA M ACES 2 8 1 19 Excited states ionizations IP_CALC handle NONE Specifies the method for calculating ionization potentials of the closed shell parent state The values are NONE MBPT 2 strict second order perturbation the ory SO_DYSON second order Green s Function iterating the MBPT 2 formula OVGF P_EOMIP partitioned IP EOM the 2h1p 2h1p block is treated as diago nal IP_EOMCC H bar is diagonalized over 1h and 2h1p configurations COMBO and OS_CCSD a state selective multireference CC method in which both orbitals and cluster amplitudes are optimized for the final ionized state of interest This last method requires additional input in the mrcc_gen namelist Analytical gradients are available for IP EOMCC and P EOMIP and require additional input though the mrcc_gen namelist Properties and transition properties of the IP states can be re quested by additional input in MRCC namelists ip_calc section IP_SYM string of 1D array 0 Specifies the number of states to calculate by an IP calculation A string e g 4 2 2 0 has to be provided that indicates the number of doublet states in each symmetry block The example string requests 4 doublet states of Al symmetry 2 doublet states in symmetry blocks 2 and 3 and 0 in block 4 The number of IP states to be calculated can also be
91. e func tion analysis codes such as Natural Bond Orbital NBO and semi emperical programs such as HyperChem NDDO and MOPAC The ACES II distribution comes with these capabil ities however ACES II developers are not responsible for distributing or maintaining the external programs Having independent licenses for external programs along with ACES II will allow users to take full advantage of this functionality Since ACES IT is the product of an academic research group and not a software company we are unable to guarantee that all results obtained with it are correct Although we have made great progress in removing serious errors from the codes problems may still occur and should be reported to aces2 qtp ufl edu Any suggestions for improving the input or output wish lists for features or other comments may also be sent to this address 3 1 Overview of capabilities of ACES II The general capabilities of ACES II to determine single point energies and properties analytical gradients and analytical hessians are as follows 11 Single point energy calculations e Independent particle models include RHF UHF and ROHF e Correlation methods utilizing RHF and UHF reference determinants include MBPT 2 MBPT 3 SDQ MBPT 4 MBPT 4 CCD CCSD CCSD T CCSD TQ CCSD CCSD TQ CCSDT 1 CCSDT 2 CCSDT 3 QCISD QCISD T QCISD TQ UCCS 4 UCCSD 4 CID and CISD e Correlation methods that can use ROHF reference determi
92. elist 6 5 NEWMOS OLDMOS These files contain the MOs in the symmetry adapted AO basis of the SCF wavefunc tion NEWMOS is created by the SCF program after the last iteration and the user can copy this file to OLDMOS to initialize the MOs of a later calculation on the same molecule and basis set The example in Section 9 1 6 page 71 illustrates this capability 6 6 AOBASMOS OLDAOMOS These files contain the MOs in the AO basis before symmetry adaptation AOBASMOS and OLDAOMOS work the same as NEWMOS and OLDMOS except that OLDAOMOS can be used to initialize SCF orbitals in vibrational frequency calculations in which the point group symmetry could change for each displacement 22 6 7 FCMINT FCM FCMSCR and FCMFINAL FCMINT contains the full internal coordinate force constant matrix The other files FCM FCMSCR and FCMFINAL correspond to the symmetrized mass weighted and analytical force constant matrices respectively The example in Section 9 2 6 page 83 shows how FCMINT can be used to initialize the Hessian matrix in a geometry search 6 8 frequency This file is a simple ASCII free format file that specifies the frequency or frequencies at which dynamic polarizabilities are computed 6 9 ISOMASS Vibrational frequencies can be calculated with standard atomic masses or user supplied masses usually of isotopes If a file named ISOMASS is found then the vibrational frequency logic in JODA replaces the atomic ma
93. en groups 1 2 P Fuentealba H Preuss H Stoll L v Szentpaly Chem Phys Lett 89 418 1982 P Fuentealba L v Szentpaly H Preuss H Stoll J Phys B 18 1287 1985 M Dolg U Wedig H Stoll H Preuss J Chem Phys 86 866 1987 G Igel Mann H Stoll H Preuss Mol Phys 65 1321 1988 W Kuechle M Dolg H Stoll H Preuss Mol Phys 74 1245 1991 M Kaupp P v R Schleyer H Stoll H Preuss J Chem Phys 94 1360 1991 A Bergner M Dolg W Kuechle H Stoll H Preuss Mol Phys 80 1431 1993 M Dolg H Stoll H Preuss R M Pitzer J Phys Chem 97 5852 1993 B 3 Basis set tables The tables on the following pages list the numbers of generally contracted AO basis functions for each element in each set The negative subscript shows the number of redundant Cartesian AOs For example oxygen in the CC PVQZ basis set listed as 70 15 has 70 Cartesian AOs but only 55 if spherical harmonics are used with SPHERICAL ON in the ACES2 namelist 118 e 9 8 9 EG Dor Dor 1 81 er LO AGE DF FTTE 9 9728 9700 NGC Ier dd 4d7 DTTE 9 AE LA d Co gei gel 67 67 6 OL dz an 3 CO H H H Lei We We 9 L xx D TTE 9 H H H 1 61 EZ EZ OFTIL 9 E E S ES 9 Dal
94. ents for closed and open shells e J D Watts G W Trucks and R J Bartlett Chem Phys Lett 157 359 1989 e J D Watts G W Trucks and R J Bartlett Chem Phys Lett 164 502 1989 j Two determinant CCSD TD CCSD analytical gradients for open shell singlet states e P G Szalay and R J Bartlett J Chem Phys 101 4936 1994 100 12 4 Analytical second derivatives for MBPT CC methods a MBPT 2 second derivatives for closed shells e N C Handy R D Amos J F Gaw J E Rice E D Simandiras T J Lee R J Harrison W D Laidig G B Fitzgerald and R J Bartlett in Geometrical Derivatives of Energy Surfaces and Molecular Properties edited by P Jorgensen and J Simons Reidel Dordrecht 1986 e N C Handy R D Amos J F Gaw J E Rice E D Simandiras Chem Phys Letters 120 151 1985 e R J Harrison G B Fitzgerald W D Laidig R J Bartlett Chem Phys Lett 124 291 1986 b MBPT 2 second derivatives for open shells UHF and ROHF e J F Stanton J Gauss and R J Bartlett Chem Phys Lett 195 194 1992 e J Gauss J F Stanton and R J Bartlett J Chem Phys 97 7825 1992 125 NMR chemical shift calculations a GIAO SCF NMR chemical shift calculations e R Ditchfield Mol Phys 27 789 1974 e K Wolinski J F Hinton and P Pulay J Am Chem Soc 112 8251 1990 e M Haser R Ahlrichs H P Baron P Weis and H Horn Theoret Chim Acta 83 455 1992 b GIAO MBPT 2 N
95. er molecule using the TZ2P basis set The program will attempt to find the lowest root in each of the four symmetry species of the Ca point group and oscillator strengths will be evaluated for all excited states 9 1 12 EOM CCSD electron attachment energy EOM CC electron attachment calculations yield energy differences between an N electron reference state and one or more electronic states of the N 1 electron system obtained by adding an electron The keywords such as REFERENCE CHARGE MULTIPLICITY and OCCUPATION define the electronic state of the N electron system If EA CALC is set to EA EOMCC then energies of N 1 electron states are calculated The states are speci fied by the EA SYM keyword as a string of NIRREP REFERENCE RHF or 2 NIRREP REFERENCE UHF or ROHF integers NIRREP is the number of irreducible represen tations in the computational point group The string of numbers specifies the numbers of 78 No N 1 electron states of a given symmetry and the spin of the additional electron in each 1 electron state For closed shell systems only the alpha roots have to be specified EA SYM 3 2 0 2 for example For open shell systems one can either attach an electron of alpha spin or one of beta spin leading to different states of the N 1 electron system The different spin blocks are separated by a slash as in EA SYM 3 2 0 2 0 1 0 4 This keyword does not have to be specified in an EA EOMCC calculat
96. erator ghost atoms serve as a center for basis functions This feature is particularly useful for calculations that determine the basis set superposition error BSSE and has several other applications such as describing lone pair electrons of a molecule by functions that are not centered at any of the molecular nuclei Symmetry can be used in such calculations but is restricted to the symmetry of the supermolecule comprised of the real and ghost atoms The additional basis functions do not necessarily form a complete set of symmetry adapted functions within the point group of the supermolecule This is different from the use of dummy atoms which do not affect the symmetry of the calculation Currently only single point energy calculations are possible with ghost atoms In ad dition the basis set must be supplied explicitly with BASIS SPECIAL and the line item basis set definitions after the ACES2 namelist 7 1 6 Cartesian coordinates The format is straightforward Each line defines one atom with the atomic symbol and the values of the x y and z coordinates in free format The coordinates may be given in either atomic units or Angstroms Older versions of xjoda require COORDINATE CARTESIAN for this to work but any version after 2 5 0 attempts to figure it out automatically since the first line of a Z matrix has only one word If the Cartesian coordinates are specified in atomic units then the keyword UNITS BOHR must be used
97. ertainly doable and because the synchronization is handled through ASCII files the calculations can even be performed on different architectures however it is not pretty to coordinate and if the user is running an optimization with numerical gradients then he or she should expect a lot of intervention before the coordinates converge In addition to complicated shell scripting the user must also know the exact order of ACES member executables that are required for each single point calculation The serial xjoda binary can accept two command line flags procs and rank that instruct it to set up single point calculations on a subset of the displacements Once the last displacement is complete but before the final xjoda the user must update each local file set with the data from all of the other file sets For vibrational frequencies this can be done on a single file set but for geometry optimizations each file set must be updated before taking the next step The two programs that are of primary concern are xjoda and xa2proc Every time xjoda is called the number of processes and the rank must be supplied on the command line xa2proc contains a module that will update print and load the data needed from each file set The following Korn shell script will run through every AME for every virtual process and collate the results of a vibrational frequency calculation This example does the same thing as the serial caces2 program on one CPU and
98. es in the ACES II program system Its two main purposes are to gather many small single use programs and to provide interfaces to external programs like Molden and HyperChem xa2proc help will show the list of available modules and the arguments that each one expects 5 27 1 clrdirty During an optimization or frequency calculation with RESTART ON the default the ACES II file set is tagged with a dirty flag Immediately before a call to xjoda xaces2 will clear the dirty flag thus signaling xjoda to backup the files If the dirty flag is not clear then xjoda will assume the calculation has crashed and restore the previous file set instead of saving the current set Users must clear the dirty flag manually with xa2proc clrdirty if they are running each AME separately otherwise ACES II will loop over the same geometry forever it will not even increment the step counter and stop after a certain number of steps 5 27 2 mem A user can alter the MEMORY_SIZE state variable of a STATIC ACES II file set with the MEM module If no AMEs are using the JOBARC and JAINDX files then xa2proc mem amount will change the value that each AME uses to allocate memory This change will remain in effect until the next run of xjoda which will reset it to whatever value is in the ZMAT file amount is a double precision number optionally followed by a unit Valid case insensitive units are B KB MB GB W KW MW and GW The number
99. files during INTPRC by reading the HF2 files created by VTRAN multiple times note that this option requires more CPU processing time and the setting of HF2_FILE SAVE AOBASIS uses an AO based algorithm to evaluate all terms involving the VVVV MO integrals Again use of this option results in considerably longer CPU times but significantly reduces the amount of disk storage more so for UHF and ROHF references than RHF AO LADDERS handle SINGLEPASS Specifies the algorithm used by ACES II when ABCDTYPE AOBASIS MUL TIPASS uses an approach where the AO integral file is read a number of times to maximize vectorization and is usually the optimal strategy on vector supercomputers SINGLEPASS determines the contributions with only a single pass through the AO integral files at the cost of significantly reduced vectorization 54 SAVE INTS switch OFF Controls deletion of the AO integrals file s after the AO to MO integral transformation If SAVE_INTS ON then these files are not deleted Any method that requires AO integrals during a post SCF calculation will automatically switch on this keyword VTRAN handle FULL PARTIAL Specifies what type of integral transformation is to be performed in the transforma tion program VTRAN FULL PARTIAL allows the program to choose the appro priate type of transformation while FULL forces a full integral transformation and PARTIAL skips the VVVV integrals This keyword is set automa
100. g 122 e e Iren V ADUAIULUVA e e 8611 Sot luet 8 611 TFET 8 611 E ADUAIULAVA e 01 631 vom rer leet STOTI Tiger Ol HeT HEL 67811 Sat Z ADUIULUVd DLEET 6 6TT 6 0TT 87 LOT 6 0TT 8 poT 6 0TI TADATALA Vd e Io TOd SHOINTHV e Cep KCC AZL SHOIYTHV e Lac Kess ZLA SHOINTHV e Co 9 ZdAd SHOINTHY e 6 Ez ZCA SHOIMTHV e 98 OFT e M ZEAd 00 91 78 MOT ZA S Ser s MO ZA e 67 e MO ZAAd 00 e 92 961 e Z amp Ad 00 DNV N 92611 N ZDAJ D9 DNV S 01 69 ZLAd D0 DNV e Ces e ZAAd 00 DNV e 9 OFT e ZGAd DO 9T p8 DA e 9 6h e ZLAd OO0 e 67 e ZAAd 00 e Lp e see DITE9 e Lp e OTTE 9 e Tp e DTTE 9 e e Cer xx D189 e e Cer DTE 9 e e 67 DIE 9 TLE Les Go IZ e e 67 e e IAIN e e ICI lt Cer T 61 D9OLS Les Ier or D OLS Les e H e DZ OLS us gu ux ua S sv ap vo NZ no IN 00 aq NN T uo A IL Os aoe sg T PE SVENAD U SJU9TIA O SG YINOITI PE 103 SUOCTJOUNJ sIseq OY P9JIBIFJUO Jo JoquINU AL Z SRL 123 ee a e 6g zer L6OSHU LS 8 zer TT 8 6 OTH LS e e Ka SENHHO 07 oe 9 19 TANTHO 8 Ce TE OLMAS e e et e e JZE TN YT Z
101. g a basis set to use blank line For more information on ZMAT GENBAS and most of the other files used and created by ACES II Section 6 page 22 describes the overall list of files used by the program and Section 7 page 26 shows the input formats for certain user files 14 5 Program Structure The ACES II program system is a collection of programs that work together to perform the user s calculation An ACES Member Executable AME is referenced by the name of its source code e g JODA and the name of its binary executable e g xjoda Most users will only interact with the driver program xaces2 but it is strongly recommended that users familiarize themselves with xjoda since that program reads the input file and initializes the ACES II file set 5 1 aces2 and p_aces2 xaces2 is the main program that drives the ACES II program system After an initial call to xjoda it determines the proper calling sequence of programs based on the calculation level and various other keywords In principle this program is not necessary if the user knows the exact calling sequence of member executables and the calculation does not involve dropped MO gradients xp_aces2 is a parallel version of xaces2 that should be used ONLY for calculations that perform numerical finite differences geometry optimizations with GRAD_CALC NUMER ICAL or vibrational frequencies with VIB FINDIF See the examples of parallel calcula tions in Section 10 3 page
102. gmented contractions which can be used efficiently in programs that don t support general contractions za The DZ set of contracted basis functions for B F was derived from the Huzinaga 9s 5p primitive set using contraction coefficients taken directly from atomic calculations A given primitive is not allowed to appear in more than one contracted function ie these are segmented contractions zz The DZ Al Cl set of functions was derived from the Huzinaga 11s 7p primitive set In these contractions one of the s primitives and one of p primitives are used twice ek 110 The DZ neon basis is unpublished It was derived from the SV contraction of Dunning and Hay zi The TZ set of contracted basis functions for B F was derived from the Huzinaga 10s 6p primitive set using contraction coefficients taken directly from atomic calculations zi CHIPMAN for first row elements was developed as a modest sized alternative to more extended basis sets which are therefore more costly to use The exponents began with the 9s 5p 4s energy optimized sets of Huzinaga and Dunning s DZ contraction coefficients To these were added a set of diffuse s p functions for elements B F and a tighter inner s function on hydrogen Finally a set of two d type polarization functions were added These contractions were designed to mimic numerical spin density results obtained with an MCSCF procedure zi CHIPMAN set uses 5 co
103. ha orbitals and a zero reads both sets Next 1 13 A flag that forces xvscf to read GUESS every time an SCF calculation is performed This applies only to calculations that run multiple SCF calculations geometry opti mizations HF stability analyses etc Set to 0 for just the first time otherwise set to a positive integer If it is set to 0 GUESS is deleted after it has been read 40 8 Keywords 8 1 ACES2 namelist The user can control the behavior of an ACES II job through the use of keywords in the ACES2 namelist In some cases the value for a keyword can be specified by an integer or by a character string In our opinion the latter is preferable as it makes the input file more readable All possible keywords in the ACES2 namelist are discussed below As there are a lot of keywords we have grouped them according to the general flow of a calculation Keyword Conventions Keywords in the ACES2 namelist are matched to an initial substring of the actual keyword in xjoda For example the full keyword is CALCLEVEL but the unique substring is CALC so CALC CALCULATION and CALCLEVEL may all be used to set the calculation level The underlined substring in the following keyword definitions is what is used to match the keyword As another example the memory keyword is defined as MEMORY_SIZE but any keyword in the ACES2 namelist starting with MEM will be used to set the memory size Some keywords should not be set by the user w
104. have the form D 3 CD 2 DCB 1 TAU with TAU being the angle between the BCD and ABC planes More formally in the plane perpendicular to the BC axis TAU is the angle needed to rotate the projection of the D lt C vector into the projection of the B A vector Clockwise is defined to be positive and values must range from 180 to 180 For a system with more than four atoms the fifth and subsequent lines follow the same pattern as the fourth line of the example given above i e they also contain a length angle and dihedral angle and the numbers of three previously specified centers It should be emphasized that this is a minimal description that would work for molecules such as NH3 and H203 but would not work for C2H2 The latter requires dummy atoms These and several other tips for forming Z matrices are discussed below All variable names like CD BCD and TAU in the previous example are limited to five characters An asterisk x immediately after the variable name implies an optimization No numbers are allowed inside the Z matrix all internal coordinates must be given a symbol even if the value is not going to be optimized 31 Z matrix Parameters After a blank line following the Z matrix the values of all unique internal coordinates those with different names are specified as follows PNM VALUE where PNM is a variable name and VALUE is the value assigned to that coordinate The first non blank string after
105. hich correspond to RHF solutions will be considered It is important to understand that following non symmetric eigenvectors lowers the symmetry of the wavefunction and that following RHF UHF stabilities leads to a UHF solution To converge the SCF roots associated with such instabilities one must run the calculation in reduced symmetry and as a closed shell UHF case respectively If not set to 0 the program follows the vector with the N lowest eigenvalue having the proper symmetry totally symmetric and spin RHF RHF or UHF UHF properties BRUECKNER switch OFF Controls whether Brueckner orbitals are to be determined for the specified CC method BRUCK_CONV tol 4 Sets the convergence criterion in Brueckner iterations The calculation is considered converged when the largest single excitation amplitude falls below 1077 QRHF_GENERAL string of 1D array 0 The presence of this keyword turns on QRHF logic and the value of each element specifies which irrep is created positive or annihilated negative In the QRHF scheme an RHF closed shell SCF calculation is performed with the OCCUPATION keyword or the CHARGE and MULTIPLICITY keywords The correlated calculation is then performed on an open shell reference generated from this closed shell state by removing adding or exciting electrons Any number of electrons may be added and all of the electrons may be removed from the SCF reference The array elements associate
106. ich are still in an experimental stage Analytical gradients are available for TDA and STEOM and require additional input though the mrcc_gen namelist Properties and transition properties of the DIP states can be requested by additional input in MRCC namelists dip_calc section DIP_SYM string of 1D array 0 Specifies the number of states to be calculated by a DIP calculation A string e g 4 2 2 0 2 1 1 1 has to be provided that indicates the number of singlet states in each symmetry block followed by the number of triplet states in each symmetry block The example string requests 4 singlet states of Al symmetry 2 singlet states in symmetry blocks 2 and 3 and 0 in block 4 In addition 2 triplet states will be calculated in block 1 and 1 in blocks 2 3 and 4 each The triplet vector and the forward slash are optional 8 1 20 Excited states gradients ZETA_TYPE handle DIS Specifies the algorithm used to solve linear equations zeta equations in excited state gradient theory POPLE uses Pople s orthogonal subspace approach and DIIS uses Pulay s DUS approach ZETA_CONV tol 12 Sets the convergence criterion for the iterative solution of the Zeta equations and Z vector equations The solutions are considered to be converged when the error falls below 1071 ZETA_MAXCYC integer 50 Sets the maximum number of cycles allowed for the solution of the Zeta equations RESRAMAN switch OFF 62 8 1 21
107. ins occ 1 1 2 2 1 1 2 2 e 2 irreps 1 spin ip sym 1 0 e 8 irreps 3 spin pairs ee sym 1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 Sets are merely one dimensional arrays of values The set delimiters are the same as the matrix ones except that the dash specifies a range of values and the forward slash separates single values Currently the only keywords that accept this type of string are DROPMO FD_IRREPS ESTATE_SYM QRHF_GEN QRHF_ORB and QRHF_SPIN Here are some examples 34 e drop orbitals 1 2 3 10 11 and 12 dropmo 1 3 10 12 e compute frequencies of modes that transform as irreps 1 3 and 4 fd_irreps 1 3 4 e add an electron to the third lowest virtual of irrep 2 and remove an elec tron from the highest occupied of irrep 4 qrhf gen 2 4 qrhf orb 3 1 NOTE This syntax is very different from previous versions Some value strings may be allowed by any version but might mean entirely different things For example DROPMO 1 31 used to mean dropping orbitals 1 and 31 from the correlated calculation If that string was parsed by the new xjoda VTRAN would attempt to remove every orbital from 1 to 31 long Integers are parsed in a fairly straightforward manner For example print 1 and charge 1 There is one special state variable that recognizes units appended to the value and that is MEMORY Recognized units are b ytes and w ords with SI prefixes k ilo m ega an
108. ion If EA_SYM is not specified but EA CALC EA_EOMCC the program tries to find the ground state of the N 1 electron system internally The EA EOMCC program is recommended for three types of applications 1 The calculation of electron affinities Only EA CALC EA_EOMCC needs to be spec ified If the ground state symmetry of the N 1 electron system is known then EA_SYM can be specified The following input yields the electron affinity of the sodium atom NA atom NA ACES2 REFERENCE UHF CALC CCSD BASIS DZP MULTIPLICITY 2 SPHERICAL ON EA_CALC EA_EOMCC The keyword EA_SYM 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 might have been specified such that only the closed shell 3s state of the sodium anion is calculated and no other possibilities are considered for the symmetry of the anion ground state The calculation of excitation spectra for systems with an odd number of electrons Take as a reference a closed shell configuration of the system with one less electron Specify EA SYM to obtain a number of roots of desired symmetry The excitation spectrum is then calculated The following example specifies the input for calculation of the excitation spectrum of MgF The two core orbitals are excluded from the correlation treatment both in CCSD and in EA EOMCC 5 roots are found in symmetry block 1 X and A symmetries 3 in block 2 II symmetry and one in block 4 MgF Excitation Spectrum MG FIR R 1 752 ACES2 REFE
109. ions to group together were based on the natural orbital occupation numbers 1 J Almlof P R Taylor J Chem Phys 86 4070 1987 2 C W Bauschlicher Jr S R Langhoff Kormornicki Theor Chim Acta 77 263 1990 3 C W Bauschlicher Jr P R Taylor Theor Chim Acta 86 13 1993 B 1 17 BAUSCHLICHER ANO These ANO sets were derived from CI density matrices which were averaged over the 3dn 4s2 and 3dn 1 4s1 states Symmetry and equivalence restrictions were imposed throughout the calculations ek The s p d primitive exponents came from H Partridge 1 H Partridge J Chem Phys 90 1043 1989 2 C W Bauschlicher Jr Theor Chim Acta 92 183 1995 B 1 18 DGAUSS DZVP DZVP2 TZVP etc DEMON COULOMB DGauss basis sets provided courtesy of Cray Research Inc ek These local spin density basis sets were developed by Nathalie Godbout and Jan Andzelm They were taken from a file provided by Cray Research Inc The file format was by M Kohout and the basis sets were originally entered by M Kohout and Ilene Carpenter 1 N Godbout D R Salahub J Andzelm E Wimmer Can J Chem 70 560 1992 B 1 19 AHLRICHS COULOMB s and d functions are optimized for the atoms p f g functions have been determined with the formula eta i 1 eta i beta 1 gamma i n 1 2 beta gamma etaO are taken of the result of the optimization of BH A1H GaH zi p set eta0 s A0 p A0 0 865 beta 2 4 gamma
110. is sound STANDARD uses the orbitals obtained in the reference function calculation without modification and SEMICANONICAL forces a transformation to semicanonical orbitals PERT ORB handle UNKNOWN 47 Specifies whether the gradient formulation assumes that the perturbed orbitals are not those in which the Fock matrix is diagonal STANDARD CANONICAL means that the perturbed orbitals are assumed to be canonical This keyword must be set to CANONICAL in gradient calculations with methods that include triple excitations De MBPT 4 CCSD T CCSD T and QCISD T 8 1 9 SCF general SCE TYPE handle HF Switches the SCF code between the standard Hartree Fock program HF and the Kohn Sham DFT program KS If SCF_TYPE KS then the KS DFT program gen erates the SCF reference and requires separate namelists viz VSCF and INTGRT FOCK handle PK Specifies the algorithm for constructing the Fock matrix in SCF calculations PK uses the PK supermatrix approach while AO constructs the matrix directly from the AO integrals In general PK is faster but results in considerable use of disk space when out of core algorithms are required CHECK SYM handle OVERRIDE Specifies the action taken when the density matrix does not transform as the totally symmetric irreducible representation of the full molecular point group NORMAL terminates the run if the molecule is nonlinear while OVERRIDE allows the job to conti
111. ithout careful consideration of what the program will do Either the keyword controls experimental pathways through the program or the program uses considerable logic to determine the correct default behavior These keywords are listed in ztalics instead of bold Value Conventions Value strings are parsed differently depending on how the corresponding keyword is de fined in JODA Currently there are handles strings integers and reals For clarity this manual also mentions switches and tols tolerances A switch is a handle with only two values ON OFF and a tol is an integer N that corresponds to a value of 10 In ACES II handles are character strings that map onto integers For example the CALC keyword has 41 possible values The code might do one thing if CALC 0 SCF and another if CALC 10 CCSD Keywords of type handle can accept the handle string or the integer as a value so CALC CCSD does the same thing as CALC 10 in the namelist Ultimately the internal integer is irrelevant to the user and could change any time therefore all keywords are described according to their handles only Some keywords only recognize switch like handles TRUE FALSE ON OFF 1 0 etc and are listed as type switch If these keywords appear in the namelist without a value string then they will be set to ON Similarly they can be negated by prefixing them with an exclamation mark For example ACES2 RESTART NONHF will register as RESTART ON 41
112. keywords to create more complicated vacuum states for post Hartree Fock calculations NOTE The UNO keywords together with MAKERHF can be used to create a refer ence determinant that is used primarily in connection to certain types of multireference calculations run in MRCC 53 UNO_CHARGE integer 0 Sets the charge of the final reference state UNO_MULT integer 1 Sets the spin multiplicity of the final reference state The orbitals are occupied in order of their natural occupation number Often this keyword is used to create a closed shell reference vital if the MRCC module is used In this case the program can proceed as an RHF calculation MAKERHF switch OFF MAKERHF ON instructs the post SCF logic to proceed as an RHF calculation and must be specified in an MRCC calculation 8 1 13 Post SCF file options SINGLE STORE switch OFF Controls storing permutations of commonly resorted two particle quantities like MO integrals and T amplitudes If a calculation is running out of disk space then setting SINGLE_STORE ON might allow the calculation to finish ABCDTYPE handle STANDARD Specifies the way that VVVV molecular orbital integrals having four virtual MOs are handled in post SCF calculations STANDARD uses a technique that mini mizes CPU time but makes liberal use of disk storage particularly during the integral processing program INTPRC MULTIPASS avoids creating intermediate sort
113. l coordinates 80 9 2 2 Partial optimization of internal coordinates 81 9 2 3 Full optimization of Cartesian coordinates 82 9 2 4 Transition state search Ze aan ee eae ea as Bh eas 82 9 2 5 Restarting an optimization or frequency calculation 83 9 2 6 Initializing the Hessian with FCMINT in a geometry search 83 9 3 Frequency calculations EE EES 85 9 3 1 Numerical frequencies from analytical gradients 85 9 3 2 Numerical frequencies from energies a o 86 9 3 3 Isotopic shift e A a e ea ES e 86 10 Parallelization 10 1 Overview paa a Se ete Ee ee 10 2 Running xgemini id a EES 10 2 1 Local scratch directories 10 2 2 Shared scratch directories 10 2 3 Command line flags and pattern macros TY 3 Exampl s soia zs oho a die he dde 10 3 1 Parallel finite differences with MPI automatic 10 3 2 Parallel finite differences with scripts manual 10 3 3 SCF geometry optimizations 11 Troubleshooting 11 1 Common mistakes Aere bg tae ase 11 2 Basic program restrictions 11 3 Suggestions for reducing resources 12 References 12 1 Many body perturbation theory MBPT 12 2 Coupled cluster CC theory 12 3 Analytical gradients for MBPT CC methods 12 4 Analytical second derivatives for MBPT CC methods 12 5 NMR chemical shift calculations
114. lations in which the first member of a Rydberg series is usually recovered These medium size basis sets are designed to reproduce molecular electric properties especially polarizabilities They were developed using the basis set polarization method Orbital exponents and contraction coefficients were not energy optimized They were obtained from an analysis of the assumed form of the electric field dependence of Gaussians 113 1 A J Sadlej Collec Czech Chem Commun 53 1995 1988 2 A J Sadlej M Urban J Mol Struct THEOCHEM 234 147 1991 3 A J Sadlej Theor Chim Acta 79 123 1992 4 A J Sadlej Theor Chim Acta 81 45 1992 5 A J Sadlej Theor Chim Acta 81 339 1992 B 1 14 WACHTERS F The f functions are contracted 3f 1f You can split off the last f for a 3 2f contraction The contraction coefficients are from a three term fit to an STO with exponents of 1 6 Sc to 4 8 Cu in steps of 0 4 It is easy to replace 3f with 1f using Stewards fits For Sc and Ti the 3s and 3p orbitals sometimes mix with ligands which are bound to the metal You can get around this by changing the 3p contraction to 3311111 and correlate the metal 3s and 3p 1 A J H Wachters JBM Tech Rept RJ584 1969 2 A J H Wachters J Chem Phys 52 1033 1970 3 C W Bauschlicher Jr S R Langhoff L A Barnes J Chem Phys 91 2399 1989 B 1 15 ROOS ADZP ATZP The ANO basis sets of Widmark M
115. lations without substantial loss of accuracy In practice only the unoccupied i e virtual orbitals are truncated 5 16 lambda xlambda solves the A equations to determine the response of the CC amplitudes to a given perturbation 5 17 vea and vee xvea calculates electron attachment energies by the EOM CC method xvee calculates excitation energies transition moments and excited state density matrices for TDA EOM CC methods Unlike MRCC they both use the standard ACES II programming environment 5 18 vcceh xvcceh calculates EOM CCSD polarizability and NMR spin spin coupling constants 17 5 19 dens xdens calculates the one and two particle correlated density matrices in the MO basis 5 20 props xprops computes all of the first order properties dipole moments electric field gradients electric quadrupole moments electrostatic potentials spin densities for open shell molecules etc It also computes the scalar relativistic corrections and the Mulliken population anal ysis 5 21 anti xanti sorts and de antisymmetrizes the two particle density matrix 5 22 bcktrn xbcktrn performs the MO gt AO transformation of the density matrices for direct con traction with the integral derivatives in the AO basis 5 23 vdint vksdint scfgrd and p_scfgrd VDINT is a heavily modified version of the integral derivative program ABACUS written by T Helgaker P Jorgensen H Aa Jensen and P R Taylor suitable for C
116. lculations Parallel AMEs xp aces2 finite differences xp_vscf xp_scfgrd SCF energies and gradients and xp_dirmp2 MBPT 2 energies all operate under the premise that each MPI task has its own ACES II file set to modify To prevent the tasks from clobbering each other s files xgemini can create destroy and manipulate private scratch directories for each task 21 6 File Structure 6 1 ZMAT ZMAT is the primary user interface to ACES II and it must exist in the run directory 6 2 GENBAS ZMAT BAS The files named GENBAS and ZMAT BAS contain the basis set definitions that the program can use In practice GENBAS is a large file and xjoda can spend most of its time scanning the file for the basis set definitions ZMAT BAS is created by xjoda to cache the relevant basis sets from GENBAS in other words if xjoda sees ZMAT BAS then it will try to read the basis information from there If a definition is missing from ZMAT BAS then xjoda will crash just as if GENBAS was missing the definition Appendix B 1 page 106 lists the contents of the standard GENBAS file 6 3 ECPDATA This file contains the data for effective core potentials Standard sets can be found in Appendix B 2 page 116 6 4 GUESS The GUESS file is used to control the placement of electrons and to manipulate orbitals when the SCF initial guess is read from an external file named OLDMOS This can be used only with GUESS READ_SO_MOS in the ACES2 nam
117. le details this procedure Like all optimizations transition states searches use analytical gradients by default If these are not available then the user must use GRAD_CALC NUMERICAL 9 2 5 Restarting an optimization or frequency calculation A savedir SAVEDIR restartable H2 geom opt H H 1 Rx R 1 0 ACES2 basis DZP restart This example reinforces the default restart behavior If restarting the calculation is not necessary then use RESTART OFF or RESTART in the ACES2 namelist All of the restart capability is built into xjoda and is controlled by two values the SAVEDIR directive and the RESTART flag The SAVEDIR directive is case insensitive and the path may be absolute or relative Each time xjoda is called it copies the optimization history files to SAVEDIR If xjoda is called and the files are contaminated meaning the previous job was interrupted or the dirty flag was not cleared then it will restore the history files from SAVEDIR and continue on as if nothing had happened This applies to both geometry optimizations and VIB FINDIF frequency calculations WARNING Restarts are only implemented for coarse grain checkpointing If other ACES II files are in the directory interesting things might happen For the best results calculations should be restarted from a clean directory i e only the original user input files should be present When the optimization is finished xjoda will delete the CURRENT and OLD directories
118. ls to a small formatted file called NEWMOS These are expressed in terms of the symmetry adapted AO basis functions 71 and are printed by spin and symmetry block To run an ROHF calculation starting from these UHF orbitals 1 copy NEWMOS to OLDMOS in a new directory 2 copy in ZMAT and change REF UHF to REF ROHF 3 add GUESS READ_ SO MO to the ACES2 namelist 4 copy in ZMAT BAS or GENBAS 5 run xaces2 The UHF job cd usr var tmp yau scr rm cp yau nh2 uhf zmt ZMAT cp yau GENBAS GENBAS xaces2 gt yau nh2 uhf out cp NEWMOS yau nh2 uhf mos The ROHF job cd usr var tmp yau scr rm cp yau nh2 rohf zmt ZMAT cp yau nh2 uhf mos OLDMOS cp yau GENBAS GENBAS xaces2 gt yau nh2 rohf out cp NEWMOS yau nh2 rohf mos 9 1 7 Initial guessing with OLDAOMOS The main use of this option is in finite difference vibrational frequency calculations In these calculations it is not possible to use the OCCUPATION keyword since the symmetry can change at various displacements Therefore the SCF program is not always able to converge to the correct electronic state at each geometry and sometimes the frequency calculation cannot be completed The GUESS READ_AO_MOS option is intended to solve this problem Users begin by performing a single point SCF calculation at the geometry at which vibrational frequencies are to be calculated In this calculation the occupation can be set to obtain the correct electronic state After the
119. lytical gradients are available for TDA and STEOM and require additional input though the mrcc_gen namelist Properties and transition properties of the DEA states can be requested by additional input in MRCC namelists dea_calc section DEA SYM string of 1D array 0 59 Specifies the number of states to be calculated by a DEA calculation A string e g 4 2 2 0 2 1 1 1 has to be provided that indicates the number of singlet states in each symmetry block followed by the number of triplet states in each symmetry block The example string requests 4 singlet states of Al symmetry 2 singlet states in symmetry blocks 2 and 3 and 0 in block 4 In addition 2 triplet states will be calculated in block l and 1 triplet in blocks 2 3 and 4 each The triplet vector and the forward slash are optional 8 1 18 Excited states electronic absorption EE_SYM string of 1D array 0 Specifies the number of states to be calculated by an EE calculation A string e g 4 2 2 0 has to be provided that indicates the number of singlet states in each sym metry block The example string requests 4 singlet states of Al symmetry 2 states in symmetry blocks 2 and 3 and 0 in block 4 The number of EE states to calculate can also be specified by listing an energy range using ee_low and ee high keywords in the mrcc_gen namelist This can be particularly relevant in vibrational frequency calculations in which the symmetry changes at different geometries
120. mini will not recursively create directories up to the full path name The following table shows the complete list of macros and their substituted values Macro NODENAME CLOGNAMEO SIDO PID PPID GRANK GPROCS CHRANKO CHPROCSO Substitution the string at uname nodename as described by sys utsname h most likely uname n the string returned by getlogin r or cuserid most likely whoami the session ID of the xgemini process the ID of the xgemini process the ID of the xgemini parent process the MPI process rank in MPI_COMM_WORLD the number of MPI processes in MPI_COMM_WORLD the MPI process rank in MPI_COMM_HOST the number of MPI processes in MPI_COMM_HOST 90 10 3 Examples 10 3 1 Parallel finite differences with MPI automatic The two programs that are of primary concern in this exercise are xgemini and xp_aces2 Assume xgemini creates shared scratch directories WORKDIR gt 1s ZMAT GENBAS WORKDIR gt xgemini i s t local tmp smith GRANK WORKDIR gt Le F ZMAT GENBAS shared 00 shared 10 Before running xp_aces2 the user should determine if output tagging is available for parallel processes Without this the standard output stream of every process will be merged and it will be almost impossible to discern which task did what Furthermore since the operating system buffers I O streams the final lines of output might not contain the final answer IBM s parallel environmen
121. mponent d functions 1 T H Dunning Jr J Chem Phys 53 2823 1970 2 T H Dunning Jr J Chem Phys 55 716 1971 3 T H Dunning Jr P J Hay in Methods of Electronic Structure Theory Vol 3 edited by H F Schaefer III Plenum Press New York 1977 4 E Magnusson H F Schaefer II J Chem Phys 83 5721 1985 5 D Chipman Theor Chim Acta 76 73 1989 B 1 8 CC PVxZ PCVxZ PWCVxZ PVxZ_DK AUG CC PVxZ etc These are the correlation consistent basis sets with various additional functions like core valence PCVxZ weighted core valence PWCVxZ and augmentation AUG The DK variant is described in the following comments The basic idea behind the correlation consistent basis sets is that functions which contribute approximately the same amount of correlation energy should be grouped together when considering what mixture of s p d etc basis functions to use For hydrogen the polarization exponents were determined by optimizing them at the SD CI level for molecular hydrogen in its ground state The s p exponents for B Ne were optimized in atomic Hartree Fock calculations on the ground state The polarization exponents were optimized at the SD CI level The p contraction coefficients for Li Be Na and Mg were based on 2 P excited state calculations zi Note that the Ga Kr basis sets are intended for use with a 14 orbital frozen core i e 1s 2s 2px 2py 2pz 3s 3px 3py 3pz 3d z2 3d x2 y2
122. nants include MBPT 2 CCSD CCSDT CCSD T CCSDT 1 CCSDT 2 and CCSDT 3 e Correlation methods that can use QRHF or Brueckner orbital reference determinants include CCSD CCSDT CCSD T CCSDT 1 CCSDT 2 and CCSDT 3 e Two determinant CCSD calculations for open shell singlet state e Equation of motion CCSD calculation of dynamic polarizabilities including parti tioned scheme e Equation of motion CCSD calculation of NMR spin spin coupling constants including partitioned scheme e Partitioned equation of motion CCSD calculations of excitation energies e Kohn Sham DFT methods combined with a wide selection of density functionals Analytical gradients e Independent particle models include RHF UHF and ROHF e Correlation methods utilizing RHF and UHF reference determinants include MBPT 2 MBPT 3 SDQ MBPT 4 MBPT 4 CCD CCSD CCSD T CCSD CCSD T CCSDT 1 CCSDT 2 CCSDT 3 QCISD QCISD T UCC 4 UCCSD 4 CID and CISD e Correlation methods that can also utilize ROHF reference determinants include MBPT 2 CCSD and CCSD T e Correlation methods that can also utilize QRHF reference determinants include CCSD e Two determinant CCSD calculations for open shell singlet state based on QRHF or bitals 12 e EOM CCSD analytical gradients for excited states e TD CCSD analytical derivatives e Dropped core and or virtual orbitals in analytical derivative calculations for RHF UHF and ROHF references
123. nctionals are available vwn Vosko Wilk Nusair lyp Lee Yang Parr pbe_cor Perdew Burke Ernzerhof pw91_cor Perdew Wang 91 wl Wilson Levy wi Wilson Ivanov KSPOT string hf The KSPOT value string is processed like FUNC The available exchange potentials are Ida becke and hf The available correlation potentials are vwn and lyp CUTOFF tol 12 105 B Standard Basis Sets and ECPs B 1 Basis sets in GENBAS The following is a listing of the basis sets currently included in the standard GENBAS file Text in a fixed width font show some of the comments that are included with the basis set from the EMSL see Section 1 on page 7 All EMSL basis sets in GENBAS were synchronized on 29 January 2006 B 1 1 STO 2G 3G 6G 3G These are the well known minimal basis sets developed by Pople and coworkers in the early 1970s They are now largely obsolete except for some calculations on large molecules Their use is not recommended for modern calculations other than code testing and rough preliminary investigations The exponents and contraction coefficients for the STO nG basis sets were obtained by least squares fitting of Slater type AO s with scaling factors based on optimal values for a variety of molecules The s and p exponents for the valence shell were constrained to be equal 1 W J Hehre R F Stewart J A Pople J Chem Phys 51 2657 1969 2 W J Hehre R Ditchfield R F Stewart J A Pople J Chem
124. nctions They are also available for certain CCSD calculations based on quasi restricted Hartree Fock QRHF reference functions namely those for high spin doublet cases and two determinant CCSD TD CCSD calculations for open shell singlet states in which the open shell orbitals have different symmetries Efficient algorithms for geometry optimization and transition state searching have also been included and may be used at all levels of theory The analytical gradients employed during geometry optimizations and vibrational frequency calculations depend on their avail ability When analytical gradients are not available automated finite differencing proce dures can be used to compute the derivatives Analytic second derivatives have been imple mented for SCF using RHF UHF and ROHF reference functions In addition analytically evaluated NMR chemical shift tensors are available at the SCF and MBPT 2 levels using 10 gauge including atomic orbitals GIAOs to ensure exact gauge invariance Other features include the direct calculation of electronic excitation energies using the Tamm Dancoff or configuration interaction singles model CIS the random phase approximation RPA the equation of motion coupled cluster approach EOM CC and similarity transform equation of motion STEOM and molecular ionization potentials and electron affinities with EOM STEOM and Fock space coupled cluster methods Transition moments between ground and exci
125. nes are allowed in the header e blank space consisting of spaces and or tabs e comments first non blank character is a hash mark e file directives first non blank character is a percent sign The header can have an arbitrary number of lines but the first line that does not qualify as one of these three will be read as the job title 2 Job title The first non blank non comment non directive line is the job title It may consist of any character and may be at most 80 characters long 3 Molecular system The coordinate matrices must immediately follow the job title Every line must describe one atom dummy or otherwise The first blank or comment line terminates the coordinate matrix Atom descriptions may have hash delimited comments and the coordinate strings may be separated with spaces or tabs There are two types of coordinate matrices internal Z matrix and Cartesian Internal coordinate matrices have two parts separated by a blank line the Z matrix and the parameter definitions Cartesian coordinate matrices are the standard X Y Z format with no blank lines For LST and QST geometry search algorithms multiple XYZ and Z matrix parameter matrices can be supplied in the following order INITIAL TRANSITION FINAL The transition geometry is only used for QST Supplying a transition geometry for 26 LST will yield incorrect results since the parser will read the second set of coordinates as the final geometry
126. nge during a running calculation There might be some hacks that involve changing convergence tolerances mid stream but these are not documented supported or advised 11 3 Suggestions for reducing resources e ABCDTYPE AOBASIS This saves disk space for CC methods e SINGLE_STORE ON This saves disk space for correlated methods e DIRECT INTEGRALS GAMESS FOCK AO SYMMETRY OFF This save disk space for all calculations but only applies to SCF and MBPT 2 theories without DROPMO 96 12 References Of the many methods currently implemented in ACES II some are well established while others are new and descriptions have not yet been published It is the purpose of this section to list pertinent literature references which provide more information about the techniques and their implementations and should be cited when results from ACES II calculations are published Some references to basis sets included in the program are also included Guide to Correlated Methods e R J Bartlett and J F Stanton Applications of Post Hartree Fock Methods A Tuto rial in Reviews of Computational Chemistry 5 65 169 edited by K B Lipkowitz and D B Boyd VCH Publishers New York 1994 e R J Bartlett Coupled Cluster Theory An Overview of Recent Developments in Modern Electronic Structure Theory Part I edited by D R Yarkony World Scientific Publishing Co Singapore 1995 12 1 Many body perturbation theory MBPT Revie
127. ntly the known types are handle string long integer and double Handles are strings that map to an integer An example of this is the CALC state variable When the parser reads CALC SCF it scans the CALC lookup table for the case insensitive string that matches SCF Upon finding a match the value of CALC is set to the offset of SCF in the table In this case CALC is set to 0 zero since SCF is the first element Alternatively the namelist could have read CALC 0 and the effect would be the same Some keys take switch values ON and OFF For these cases the keyword may be specified without a value and the parser will assume the value is ON Similarly the negation operator may be used to turn the value to OFF An example is sym ecp meaning SYM ON ECP OFF Strings are character arrays that are treated differently depending on the keyword they define There are two types plain text strings and array strings plain text strings are only used by the BASIS keyword currently The exact case sensitive value string is used to find the basis set definition in GENBAS array strings matrices are used by the keywords OCCUPATION IP_SYM EA_SYM etc They are loosely defined as irrep by spin This means the parser expects to find spin columns of nirrep rows The row delimiter is a dash and the column delimiter is a forward slash Here are some examples e 4 irreps 2 sp
128. nue but prints a warning message Nonsymmetric density matrices result from calculations on electronically degenerate states or from broken symmetry SCF solu tions Often such calculations do not give meaningful results and inexperienced users are encouraged to use CHECK_SYM NORMAL in their calculations on nonlinear molecules For IT A states of linear molecules however meaningful calculations can still be performed even though the density matrix is not symmetric SCF_PRINT integer 0 This keyword is currently not used but exists for future compatibility 8 1 10 SCF orbital control GUESS handle MOREAD Specifies where the initial SCF eigenvectors are read from The HF SCF executable checks multiple places for pre existing orbitals but this keyword can override that logic 48 MOREAD JOBARC file CORE core Hamiltonian READ SO MOS gt OLDMOS file from NEWMOS READ_AO_MOS OLDAOMOS file from AOBASMOS Other options include NDDO WALT_PRJDEN MIN_BASIS and HUCKEL although some of these are still experimental and not fully supported OCCUPATION string of irrep by spin array Specifies the orbital occupancy of the reference wavefunction in terms of the occupation numbers per irrep per spin The occupancy is specified by NIRREP or 2 NIRREP integers that define the number of occupied orbitals of each symmetry type NIRREP is the number of irreducible representations in the computa
129. oe se Bote areas 19 dile de MEN ss e e e ee deg de Ee 19 E 1 ZB RORECOS at di tod A A DO Sd erga ot amp 20 D2 E A E DEENEN 20 O22 or PARED EENEG 20 5 27 6 MOLDEN and HYPERCHEM 4 422 4 e eee e a 20 52C JASUM and FOSUM o aa a he A Be Be BRS Bnd 20 9 2138 TAREA ek et SS ke a Ais AI Sodas A AS OR E 20 AS AMAR 21 E a de A A AS ek BORG A O RAS 21 0 29 GEMINIS N a ies elt der eres ce oe e hr 21 File Structure 22 Galo ANMAT e O50 E bre ta ted So matt dec e eh E A EE 22 6 2 GENBAS ZMAT BAS 2 aa 22 09 ECPDATA Net Bt E e Ad A e a a TTT 22 0AF GUESS ke A ee e ele eene EE EEN dE a ett Ee 22 6 5 NEWMOS OLDMOS 22 6 6 AOBASMOS OLDAOMOS 7 0 6 de st AAA a ae we ee 22 6 7 FCMINT FCM FCMSCR and FOHEINAL aoaaa aa a 23 DEIER es weg Sey oh a Khe A ik Ae ee ee we a Ee hh 23 6 9 TSOMASS A Sree tei ad ek oe oe eh a A R 23 E System ales Sanss a a el ek eh 23 6 10 1 JOBARC JAINDX lt o e mas cs 23 6 10 2 MOINTS GAMLAM MOABCD DERINT DERGAM 23 6 10 3 MOL IIII IIJJ IJIJ IJKL 23 6 10 4 OPTARC OPTARCBK 2 2 0 2 000000 eee ee 24 6 10 5 DIPOL DIPDER POLAR POLDER 24 OelLOLG GRD gt a 15 SA ay tikes oh tS he oh inthe eee aot ori aa SY Gls TaD NY so 24 6 10 7 HF2 HF2AA HF2AB HF2BB 24 610 8 MAD 2 a a te boo o eee io ob a AY is 24 6 10 9 TGUESS LGUESS e 24 GAOTOVPOUT ario Er ii ts el a Sl Ad io a eg 24 6 10 11GAMESS LOG MP2 LOG DIRGRD LOG
130. on of the a and P spin eigenvectors Zero eigenvalues will occur for degenerate electronic states and merely indicate the equiv alence of occupations within the computational point group Small numerical inaccuracies frequently result in nominally zero eigenvalues having small non zero values The sign of such eigenvalues will determine whether or not they are reported as instabilities Thus the program might not show the expected number of zero small eigenvalues Other instabilities such as the wavefunction becoming complex are not tested since complex wavefunctions are not presently supported in ACES II In the program output stability analysis is headed by the label RHFSTAB or UHF STAB depending on the value of the REF keyword The number of instabilities in each irrep is given along with their eigenvalue and classification For example RHFSTAB Performing stability analysis of RHF wavefunction Orbital rotation parameters will be evaluated for each symmetry block There are 0 instabilities within irrep 1 There are 1 instabilities within irrep 2 Eigenvalue 0 1411211526 Instability classification RHF gt UHF with broken symmetry There are 1 instabilities within irrep 3 Eigenvalue 0 1411211526 Instability classification RHF gt UHF with broken symmetry There are 0 instabilities within irrep 4 There are 0 instabilities within irrep 5 There are 2 instabilities within irrep 6 Eigenvalue 0 358679238
131. on Raphson minimum energy search RFA uses the Rational Function Approx imation minimum energy search and can be used when the initial structure is in a region where the number of negative Hessian eigenvalues is nonzero MANR uses a Morse adjusted Newton Raphson minimum energy search and is very efficient if the Hessian is available EVFTS uses Cerjan Miller eigenvector following for finding a transition state can be started in a region where the Hessian index the number of negative Hessian eigenvalues is not equal to one MAEVFTS uses a Morse adjusted eigenvector following for finding a transition state MAX STEP integer 300 Sets the largest step size in millibohr STP_SIZ_CTL handle TRUST_RADIUS Controls how the step size is scaled TRUST_RADIUS uses the dynamic scaling by Fletcher NORM uses the absolute step length and MAXIMUM uses the largest individual step in the internal coordinate space EIGENVECTOR integer 1 Sets which eigenvector of the totally symmetric part of the block factored Hessian is to be followed uphill in a transition state search Eigenvectors are indexed by their 65 eigenvalues the lowest eigenvalue is 1 the next lowest is 2 etc The default should always be used if you are not looking for a specific transition state which you know corresponds to motion along a different mode The value of EIGENVECTOR has no meaning if OPT_METHOD is not set to EVFTS or MAEVFTS NEGEVA
132. ong z ry is Op Dyn Cy axis along z one Ca axis along x Dya Son axis along z one Ca along z T Ca axis along q q q Ta S4 axis along z Tn Ca axis along q q q symmetry planes are xy TZ and yz O C axes along x y and z Op C4 axes along x y and z I C axis along z one C3 lies in the xz plane 29 In Cs axis along z xz is a symmetry plane For groups with ambiguities D2 Dan C2 there is no standard orientation at present and users might want to run xjoda once to show which orientation is used before assigning orbital symmetries For example water belongs to the C2 point group and the symme try plane containing the hydrogen atoms might be assigned to either the xz or yz planes leading to an ambiguity between the b and bs irreducible representations Eventually some criterion could be established that can define a standard orientation for these groups thereby alleviating this issue 7 1 5 Dummy and ghost atoms Dummy atoms represented with X are only useful for internal coordinates and define points in space They do not have basis functions and do not affect the symmetry of the molecule They are necessary to break angles of 180 and are used to define highly symmetric molecules with no atom at the center of mass Ghost atoms which are specified by the symbol GH have zero nuclear charge However while dummy atoms are invisible to the program outside the coordinate gen
133. ot be considered a full polarized basis set Note This basis uses 6 component d functions zi The 3 21 G basis set adds a diffuse s p shell to elements Li Cl and a single diffuse s to hydrogen These exponents were optimized for 8 small anions using the 3 21G basis set at the HF level of theory by Clark et al Frisch Pople and Binkley reoptimized the exponents at the MP4 level for both neutral and anionic systems 1 J S Binkley J A Pople W J Hehre J Am Chem Soc 102 939 1980 2 R Krishnam J S Binkley R Seeger J A Pople J Chem Phys 72 650 1980 107 3 W J Pietro M M Francl W J Hehre D J DeFrees J A Pople J S Binkley J Am Chem Soc 104 5039 1982 4 T Clark J Chandrasekhar P v R Schleyer J Comp Chem 4 294 1983 5 M S Gordon J S Binkley J A Pople W J Pietro W J Hehre J Am Chem Soc 104 2797 1983 6 K D Dobbs W J Hehre J Comput Chem 7 359 1986 7 K D Dobbs W J Hehre J Comput Chem 8 861 1987 8 K D Dobbs W J Hehre J Comput Chem 8 880 1987 9 P M W Gill B G Johnson J A Pople M J Frisch Chem Phys Lett 197 499 1992 10 E D Glendening D Feller J Phys Chem 99 3060 1995 B 1 4 4 31G 3 21GSP 4 22GSP Note For Li and Be Gaussian actually uses a 5 21G set as are given here He and Ne are unpublished basis sets taken from the Gaussian program zi The 3 21GSP basis set family was obtained by reop
134. ower case characters have been used instead of upper case This situation is difficult to pinpoint Most of the ACES2 namelist parsing is case insensitive along with the atomic symbols File directives in the header are still case sensitive as are basis set names e The OCCUPATION keyword takes precedent over CHARGE and MULTIPLICITY This can lead to confusion in open shell SCF calculations Here are some usage tips 1 If OCCUPATION has been specified then the CHARGE and MULTIPLICITY keywords are ignored It is however good practice to make these consistent with OCCUPATION 2 If the REFERENCE keyword is absent from an open shell calculation or is erro neously set to RHF then unpredictable things might happen ACES II does not have a default type of open shell SCF 3 The occupation specified in the GUESS file takes precedent over all keywords Again it is sensible to make them the same to avoid confusion 95 4 In older binaries specifying only the o occupation with OCCUPATION dropped all electrons This is trapped in the current version e An input file contains text beyond the 80th column e There is no title line and the first line of the Z matrix has been entered as the title e There were files from a previous ACES II calculation in the workspace In general one should clear the workspace of all previous ACES II files prior to copying in the new files 11 2 Basic program restrictions e ZMAT should not cha
135. rative method and values greater than 1 use the reduced linear equation method When NITER 1 more than one frequency can be solved for For the other methods of solution only one frequency can be considered in a single calculation The number of non zero frequencies is specified by NFREQ The NFREQ frequencies are listed one per line following the INPUTP namelist Static calculations are also performed along with the dynamic calculations To obtain only static results all parameters should be set to 0 The following is an example of a TDHF calculation on No TT N2 TDHF TEST Sadlej basis set d R 2 07434 ACES2 UNITS BOHR BASIS PBS TDHF 0N INPUTP IOPU 0 IOPFE 0 IOPEV 0 IOPPR 0 NITER 1 NFREQ 4 IDCSHG 1 IOKE 1 IDCOR 1 IIDRI 1 ITHG 1 IWRPA 1 END 0 072 0 0886 0 0934 0 0995 WARNING There does not seem to be a portable format for Fortran namelists The code that parses ZMAT contains an error handler that shows the expected format in the event a read error occurs For TDHF calculations in particular the developers suggest running a small test calculation to verify the proper formats are used 9 1 11 EOM CCSD excitation energy EOM CCSD excitation energies and transition moments for water H 01R H2R1A R 0 957 A 104 5 ACES2 BASIS TZ2P CALC CCSD EXCITE EOMEE ESTATE_SYM 1 1 1 1 ESTATE_PROP EXPECTATION This example specifies an equation of motion coupled cluster excitation energy calculation for the wat
136. rivatives not calculated FIRST First derivatives to be calculated SECOND Second derivatives to be calculated This is automatically set 8 1 2 System molecular control SYMMETRY handle ON Specifies which subgroup computational point group of the full point group is to be used in the energy and or gradient calculation OFF forces a no symmetry run in C1 ON runs the calculation in the largest self adjoint subgroup Da and its subgroups and FULL uses the full point group Currently ACES II does not support groups with degenerate representations so the FULL option has no value unless JODA is used to make input files for another program package SUBGROUP handle DEFAULT Specifies a lower computational point group symmetry to use provided it is a subgroup of the full group Acceptable values are DEFAULT C1 C2 CS CI C2V C2H D2 and D2H SUBGROUP C1 is equivalent to SYMMETRY OFF The DEFAULT option uses the highest order Abelian subgroup including the full group SUBGRPAXIS handle X This is a somewhat complicated keyword to use The value can be X Y or Z The use of the keyword is best described by example Suppose one is running a calculation on water and wishes to run it in the C point group with the special plane being the one that bisects the H O H bond angle SUBGRPAXIS specifies which Cartesian direction in the C2 frame becomes the special direction in the C frame Normally 43
137. ros for customizing the path and directory names but the simplest command to use for the job would be xgemini i s t global tmp smithdir GRANK Each task will replace the string AGRANKO will its global rank an integer from 0 to N 1 where N is the number of parallel tasks Creating scratch directories in global allows all nodes in the computer to see all scratch directories but the symbolic links in WORKDIR will still be named compnodeX Y unless the s flag is used If a parallel job is restarted process 0 might be running on compnode2 and it would expect to see compnode2 0 in WORKDIR However if the previous run had process 0 running on compnode0 then it would have created compnode0 0 and the restarted job will crash With s the symbolic links in WORKDIR are named shared rank so it does not matter which actual node created the scratch directory or the link to it If shared links are used then every call to zgemini must use the s flag The downside to remote scratch directories is that performance might decrease with high network traffic although this is highly dependent on the hardware architecture and system load at run time Some users might feel more comfortable limiting all activity to WORKDIR In this case the scratch directory pattern can be the same as the symbolic link For example xgemini i s t shared GRANK will create directories instead of symbolic links in WORKDIR The parallel AMEs will not know
138. ry called smithdir on every node xgemini i t local tmp smithdir Assuming all compute nodes took part in the parallel run then it is likely that the user would see the following WORKDIR directory listing ZMAT GENBAS compnode0 0 compnodel1 1 compnode2 2 Each compnodeX X is a symbolic link that points to local tmp smithdir and inside each of those local directories are symbolic links back to ZMAT and GENBAS The i flag is what instructs xgemini to create the directories and links It is usually the case that if local directory partitions are available then performing I O on them will yield better performance than reading and writing to a partition over the network The downside is that if a parallel job is stopped and needs to be restarted then the user must tell the parallel job scheduler like LoadLeveler or PBS to run the new job on the same nodes that the previous job ran on simply because those are the nodes with the data 88 10 2 2 Shared scratch directories If a global file system like GPFS or even NFS is used for the scratch directories then a task on any node can get to any scratch directory Assume user smith can create directories in global tmp Reusing the previous xgemini command line with local changed to global will not work because every parallel task will attempt to create global tmp smithdir One task will succeed but all of the others will fail and crash GEMINI has a rich set of pattern mac
139. scribed in Section 6 page 22 8 1 27 Finite displacements for numerical gradients and Hessians FD STEPSIZE integer 50 Sets the step length in 107 amu Bohr used in generating the force constant matrix by finite difference of Cartesian gradients The default is 0 005 amu Bohr FD IRREPS string of 1D array 0 Lists the symmetry types to be evaluated in a VIBRATION FINDIF calculation The numbers of the irreducible representations for which vibrational analysis is to be per formed are separated by forward slashes For example FD IRREP 1 3 4 computes the frequencies of modes transforming as the first third and fourth irreducible rep resentations If a symmetry is specified for which there are no vibrational modes the program will terminate The labels of the irreducible representations for this keyword 67 are not usually the same as those used in the symmetry adapted integrals Moreover for some point groups like those of linear molecules the two sets of labels refer to different subgroups There is still no straightforward way to determine what they will be without starting a calculation One JODA run will list the relevant irreducible repre sentations If all vibrational frequencies are desired this keyword need not be included The default is to calculate all irreps ED PROJECT switch ON Controls whether or not rotational degrees of freedom are projected out of symmetry adapted coordinates ON
140. specified by listing an energy range using ip_low and ip_high keywords in the mrcc_gen namelist This can be particularly relevant in vibrational frequency calculations in which the symmetry changes at different geometries IP_SEARCH handle VALENCE Specifies the character of the IP states to zoom in on If IP_SYM is not specified the program attempts to determine all IP s of the given character otherwise it uses the symmetry constraints imposed by IP_SYM The values are VALENCE all valence IP s LOWEST only the ground state of the cation COREIP only the core ion ization potentials IPs larger than about 100 eV SHAKEUP shake up states and KOOPMANS all principle IP s having predominantly 1h character KOOPMANS type can be very difficult for inner valence ionization potentials and SHAKEUP is not implemented yet 61 DIP_CALC handle NONE Specifies the method for calculating double IP states of the closed shell parent state This type of calculation can be very useful to describe certain multireference situations like biradicals The values are NONE TDA the analogue of CIS in which a CI calculation is performed over 2 hole states EOMCC H bar is diagonalized over 2h and 3hlp calculations STEOM Similarity Transformed Equation of Motion OS_CCSD and SS_STEOM To run a meaningful DIP STEOM calculation one must also set IP CALC IP_EOMCC OS CCSD and SS STEOM are multireference CC methods wh
141. sses with those found in the file The content is free format ASCII and the order of the masses must match the non dummy centers in ZMAT 6 10 System files 6 10 1 JOBARC JAINDX The JOBARC file stores records named arrays for the ACES II program system The accompanying file JAINDX stores metadata about the records 6 10 2 MOINTS GAMLAM MOABCD DERINT DERGAM These files store lists of double precision arrays used by all of the post SCF member executables They will be VERY large for big molecules and basis sets 6 10 3 MOL IIII IIJJ IJIJ IJKL The MOL file is created by xjoda and stores the molecular system and basis set information used by xvmol and any AME that uses GAMESS integrals The IITI IIJJ IJIJ and IJKL files store the two electron integrals in the AO basis The one electron integrals are also stored in IIIT 23 6 10 4 OPTARC OPTARCBK The iteration history of geometry optimizations is stored in OPTARC OPTARCBK is a backup of the true OPTARC file which gets clobbered during geometry optimizations with numerical gradients 6 10 5 DIPOL DIPDER POLAR POLDER DIPOL and POLAR contain the dipole moments and polarizabilities respectively and DIPDER and POLDER contain their derivatives 6 10 6 GRD This file was intended for programs to extract gradients from ACES II With the func tionality in the ACESCORE library and xa2proc this file interface is obsolete 6 10 7 HF2 HF2AA HF2AB HF2BB These files are created
142. t a number of checks These include determination of whether coordinates given the same name are actually equivalent or coordinates having different names are equivalent whether a non zero gradient is possible with respect to modes which are not being optimized etc In addition it determines the number of degrees of freedom within the totally symmetric subspaces of nuclear configurations and compares this value with the number of independent coordinates which are being optimized If they are not equal a warning message is printed out For most Z matrices with poorly defined internal coordinates the analyzer prints out a number of warning messages but does not halt the ACES II execution sequence However for Z matrices which are particularly bad it will terminate the job All users are encouraged 32 to carefully inspect the output of the analyzer and to check that the full molecular point group printed out below the output from the analyzer is the one intended If warning messages are printed or if the symmetry is not what the user expects then reconstruct as necessary A rule of thumb for Z matrix construction is that each internal coordinate included in the Z matrix must be accompanied by all others which are equivalent to it by the symmetry of the molecule For example in water it is best to specify the molecular geometry by the two O H distances and the H O H bond angle rather than by the O H distance the H H distance and the H H O
143. t can tag each line with the global rank if labelio yes is used or if MP_LABELIO yes is set in the environment HP Compaq computers have a similar feature WORKDIR gt xp_aces2 labelio yes gt out WORKDIR gt grep 0 out 0 output from task 0 And finally WORKDIR gt xgemini x s WORKDIR gt ls ZMAT GENBAS out The user should not request more process elements than there are finite displacements If this happens then some instances of xjoda will see there are no displacements and bomb The surest way of checking this before submitting the job is by running xjoda on the input file The FD logic will print out a table like the following data from CH fully numerical vibrational frequencies in C4 91 VIBINF I Symmetries species for nuclear motions Irrep Label Total Vibrations Translations Rotations 1 A 15 00 9 00 3 00 3 00 Total number of calculations required 90 Number of single point energy calculations 90 Number of energy gradient calculations 0 For frequency calculations using numerical gradients the reference geometry is added to the set of displacements so the example above can be run with at most 91 process elements CH frequencies in CO with analytical gradients or geometry optimization with numerical gradients can use at most 18 process elements 10 3 2 Parallel finite differences with scripts manual Running a set of calculations in parallel by hand i e without MPI is c
144. ted states can be calculated for all of the methods as well as selected excited state properties The excited state geometry optimization and frequency calculations employ the analytical gradients capabilities available for the EOM and STEOM methods The programs collectively known as ACES II began development in early 1990 and the first version of the code was written by J F Stanton J Gauss J D Watts W J Lauderdale and R J Bartlett Program development is continuing and the capabilities as well as the contributors to the development of the ACES II program system are continually increas ing At present there are more than 30 member executables each of which performs a well defined function and communicates with the rest of ACES II through stored files The ACES II program system has been interfaced with external programs such as MOLCAS and GAMESS The primary function of the MOLCAS and GAMESS interfaces is to provide in tegral and integral derivative programs that are more efficient and have direct capabilities to complement the functionalities of the locally modified version of VMOL VPROPS and VDINT programs A complete replacement of these member executables is not yet feasible since the specialized integrals such as Gauge origin independent GIAO integrals and NMR spin spin coupling operator integrals are only available in VMOL VPROPS and VDINT ACES II can also generate interfaces to graphical programs such as MOLDEN and gOpenMol wav
145. the difference when they attempt to cd into the scratch directory 10 2 3 Command line flags and pattern macros As previously demonstrated xgemini can create the runtime environment with i can use shared node names with s and can create private scratch directories with almost any name the user requires with t and macro substitution In addition it can use arbitrary input and basis files with z and b can run serial commands in each directory all arguments that are not flags and can clean up after a parallel run with x The command line structure is as follows 89 xgemini h s ka z file b file t pattern x exec Flag z file b file t pattern exec Description print usage text use shared scratch directories useful for restarts initialize make scratch directories and symlinks link file to ZMAT default is ZMAT link file to GENBAS default is GENBAS pattern defines scratch directory paths clean remove scratch directories and symlinks terminate xgemini flag parsing pass exec arguments to a system command in each scratch directory The pattern that defines the scratch directories can be an absolute or relative path with respect to WORKDIR Another example of directory patterns is usr var tmp LOGNAME scr CGGRANKO which would become usr var tmp smith scr 0 for the root task of user smith The pattern cannot be a deep path in other words xge
146. the highest occupied orbital of symmetry 1 A4 As in the ROHF example the program automatically turns on the NON HF option so that the appropriate non Hartree Fock terms are included in the coupled cluster equations 70 9 1 5 Effective core potentials CRF6 SINGLE POINT ENERGY CALCULATION USING AN ECP FOR CR CR RMC RMC 2 W1 RMC 3 W1 2 T1 RMC 4 W1 3 T2 RMC 5 W1 4 T1 RMC 6 W1 5 T2 0 yyy yy KA KA KK KA pa ps RMC 1 676125 W1 90 T1 90 T2 270 ACES2 BASIS SPECIAL ECP ON CALCLEVEL SCF OCC 10 6 6 2 6 2 2 0 SPHERICAL 0N CR SBKJC F DZP F DZP F DZP F DZP F DZP The use of effective core potentials is specified in the ZMAT file by the keyword ECP ON The keyword BASIS must be set to BASIS SPECIAL The ECP names are listed below the last basis set separated by a blank line using the same format as the non standard basis set specification XX ECPNAM As for the GENBAS file the ECPDATA file may be searched for XX where XX is the atomic symbol to show what ECPs are available for that atom Atoms included in the ZMAT file without an ECP parameter set are marked by the nickname NONE 9 1 6 Initial guessing with OLDMOS Suppose the user wishes to run an ROHF calculation but the default ROHF procedure does not converge or converges to the wrong electronic state The user could try to start the ROHF calculation from a converged set of UHF orbitals First run a UHF calculation xvscf writes the UHF orbita
147. this property MBPT and CC methods are ideally suited for the study of chemical reactions While ACES II can perform Hartree Fock Self Consistent Field HF SCF and Kohn Sham Density Functional Theory KS DFT calculations the ACES II program system is not intended for large scale HF SCF or KS DFT calculations Two important features of the ACES II program system are its effective use of molec ular symmetry particularly in MBPT and CC calculations and the sophisticated gradient methods which are included in the program Even if only a few elements of symmetry are present in a molecular system differences in execution times required for calculations with symmetry and without can be dramatic The implementation of symmetry currently is lim ited to Da and its subgroups and the expected speedup due to symmetry utilization will be on the order of the square of the order of the computational point group for all steps except integral and integral derivative generation and integral sorts in which the speedup can be no greater than the order of the group Gradient techniques are implemented for SCF and the following correlated levels of the ory MBPT 2 MBPT 3 MBPT 4 CCD QCISD CCSD QCISD T and CCSD T for both restricted and unrestricted Hartree Fock RHF and UHF respectively reference func tions In addition for the MBPT 2 MBPT 3 CCSD and CCSD T methods gradients are available for restricted open shell Hartree Fock ROHF reference fu
148. though NONE is the default DAMP_TOL integer 10 If DAMP_TYP DAVIDSON then the cutoff is defined to be N x 0 01 so the default cutoff is 0 1 DAMPSCF integer 20 If DAMP_TYP STATIC then the static damping factor used in the SCF iterations is N x 0 01 so the default damping factor is 0 2 SCF_EXTRAP previously RPP switch ON Controls whether or not the reduced partitioning procedure is to be used to accelerate convergence of the SCF equations 50 SCF_EXPORDER previously RPP ORDER integer 6 Sets the number of density matrices to be used in the RPP convergence acceleration procedure SCF_EXPSTAR previously RPP_LATEST integer 8 Sets the latest iteration for initiation of the RPP convergence acceleration procedure RPP is switched on when the error falls below a certain threshold but in difficult cases in which the iterations are oscillatory it is necessary to force it on when the error is still large In these cases the RPP will begin on the iteration number specified by this parameter LSHEF AT integer 0 Sets the doubles singles level shifting parameter a in which a Vx0 01 This keyword is currently only meaningful in ROHF calculations LSHF B1 integer 0 Sets the singles virtuals level shifting parameter 3 in which 8 N 0 01 This keyword is currently only meaningful in ROHF calculations 8 1 12 SCF reference adjustments DROPMO string of 1D array 0 Lists
149. tically based on other relevant keywords HF2 FILE handle USE Specifies whether the HF2 file series including HF2AA HF2BB and HF2AB is deleted after the first stage of integral processing The default is to delete these files however when ABCDTYPE MULTIPASS these files must not be deleted and the program sets HF2_FILE SAVE ABCDFULL handle UNKNOWN This is a debug aid and is concerned with the storage of VVVV integrals and effective Hamiltonian elements It is only relevant to RHF calculations and it should never be necessary for users to set this keyword Possible values are UNKNOWN the program will choose an appropriate value ON full storage and OFF reduced storage HBARABCD handle UNKNOWN Controls the formation and storage of the VVVV block of the effective Hamiltonian It should never be necessary for users to set this keyword Possible values are UN KNOWN the program will choose an appropriate value ON full storage and OFF reduced storage HBARABCI handle UNKNOWN Controls the formation and storage of the VVVO block of the effective Hamiltonian It should never be necessary for users to set this keyword Possible values are UN KNOWN the program will choose an appropriate value ON full storage and OFF reduced storage GAMMA_ABCD handle DISK 59 Controls the evaluation of the two particle density elements when all four indices cor respond to virtual orbit
150. timizing the Gaussian exponents and contraction coefficients subject to the constraint that the valence s p shells share the same exponents The resulting energy lowerings were C 0 116 Eh N 0 151 Eh and O 0 197 Eh E The 4 22GSP basis set family was obtained by reoptimizing the Gaussian exponents and contraction coefficients subject to the constraint that the valence s p shells share the same exponents 1 R Ditchfield W J Hehre J A Pople J Chem Phys 54 724 1971 2 W J Hehre J A Pople J Chem Phys 56 4233 1972 3 W J Hehre W A Lathan J Chem Phys 56 5255 1972 4 A V Mitin G Hirsch R J Buenker Chem Phys Lett 259 151 1996 5 A V Mitin G Hirsch R J Buenker J Comp Chem 18 1200 1997 108 B 1 5 6 31G 6 31G 6 31G 6 31 G 6 31 G 6 31 G etc 6 31G supplements 6 31G with d polarization functions p functions for H and He It should be used with the SPHERICAL OFF option 6d functions since this is how the set was defined and developed The 6 31G basis set excludes p functions from H and He while retaining d functions on other atoms The diffuse exponents plus notation follow from the 3 21 G and 3 21 G references The 6 31G basis set uses 6 Gaussian primitives to expand the 1s core of second period elements Note He and Ne are unpublished basis sets taken from the Gaussian program Note This basis set uses 6 component d functions zi 6
151. tional point group If there are no orbitals of a particular symmetry type then a zero must be entered If the reference wavefunction is for an open shell system then two strings of NIRREP occupation numbers separated by a forward slash define the o and 7 sets of orbitals An example of the use of the OCCUPATION keyword for the water molecule would be OCCUPATION 3 1 1 0 For the 7A water cation an open shell system the keyword would be specified by OCCUPATION 3 1 1 0 2 1 1 0 It should be noted that the VMOL integral program orders the irreducible representa tions in a strange way which most users do not perceive to be a logical order Hence it is advised to run a single point SCF calculation to determine the initial number and ordering of the irreducible representations The occupation keyword may be omit ted in which case an initial orbital occupancy is determined by diagonalizing the core Hamiltonian In many cases SCF calculations that use the core Hamiltonian guess will converge to the lowest energy SCF solution but this is not guaranteed LOCK ORBOCC switch OFF Controls orbital occupancies among symmetry blocks in the SCF iterations ON locks the occupation to the OCCUPATION value array or the initial guess if OCCUPATION is undefined OFF permits the occupation to change NOTE If the OCCUPATION keyword is defined then LOCK_ORBOCC is switched on automatically NEWVRT switch OFF 49 Signals
152. ts J Gauss and R J Bartlett J Chem Phys 98 8718 1993 f CCSDT 2 and CCSDT 3 methods e J Noga R J Bartlett and M Urban Chem Phys Lett 134 146 1987 g CCSDT methods e J Noga and R J Bartlett J Chem Phys 86 7041 1987 e G E Scuseria and H F Schaefer Chem Phys Lett 152 382 1988 98 e J D Watts and R J Bartlett J Chem Phys 93 6104 1990 e J D Watts and R J Bartlett Int J Quant Chem Symp 27 51 1993 h ROHF and QRHF CC methods e M Rittby and R J Bartlett J Phys Chem 92 3033 1988 i Brueckner CC methods e R A Chiles and C E Dykstra J Chem Phys 74 4544 1981 e J F Stanton J Gauss and R J Bartlett J Chem Phys 97 5554 1992 j QCISD and QCISD T methods e J A Pople M Head Gordon and K Raghavachari J Chem Phys 87 5968 1987 k UCC methods e J D Watts G W Trucks and R J Bartlett Chem Phys Lett 157 359 1989 1 CCSD TQ CCSD CCSD TQ and QCISD TQ methods e R J Bartlett J D Watts S A Kucharski and J Noga Chem Phys Lett 165 513 1990 e K Raghavachari J A Pople E S Replogle and M Head Gordon J Phys Chem 94 5579 1990 m Two determinant CCSD method e A Balkova and R J Bartlett Chem Phys Lett 193 364 1992 12 3 Analytical gradients for MBPT CC methods Review article e R J Bartlett J F Stanton and J D Watts in Advances in Molecular Vibrations and Collision Dynamics Vol 1 ed J Bowman
153. ts are available for EA EOMCC and P EOMEA and require additional input though the mrcc_gen namelist Properties and transition properties of the EA states can be requested by additional input in MRCC namelists ea_calc section EA SYM string of 1D array 0 Specifies the number of states to be calculated by the EA CALC calculation A string e g 4 2 2 0 has to be provided that indicates the number of doublet states in each symmetry block The example string requests 4 doublet states of Al symmetry 2 doublet states in symmetry blocks 2 and 3 and 0 in block 4 The number of EA states to be calculated can also be specified by listing an energy range using ea_low and ea_high keywords in the mrcc_gen namelist This can be particularly relevant in vibrational frequency calculations in which the symmetry changes at different geometries DEA_CALC handle NONE Specifies the method for calculating double electron affinities after calculation of the closed shell parent state This type of calculation can be very useful to describe cer tain multireference situations like the C atom or the N20 molecule The values are NONE TDA the analogue of CIS in which a CI calculation is performed over 2 particle states EOMCC STEOM Similarity Transformed Equation of Motion OS_CCSD SS_STEOM DEA STEOM is only meaningful if EA CALC EA_ EOMCC OS CCSD and SS STEOM are multireference CC methods which are still in an ex perimental stage Ana
154. ula and COMBO com putes all approximations This keyword only has meaning when PROP EOM_NLO J_SO J_FC J_SD or JSC_ALL 63 TDHF switch OFF Controls whether a time dependent Hartree Fock calculation of nonlinear optical prop erties is to be performed This keyword can only be used for closed shell SCF calcula tions with no dropped MOs The nonlinear properties which are to be calculated are specified by a namelist which is put at the end of the ZMAT file A description of this namelist and an example can be found in Section 9 1 10 page 76 CPHF_CONVER tol 12 Sets the convergence criterion for the iterative solution of the CPHF and Z vector equations The solutions are considered to be converged when the error falls below Re CPHF_MAXCYC integer 64 Sets the maximum number of cycles allowed for the solution of the CPHF and Z vector equations TREAT PERT handle SIMULTANEOUS This keyword is used for certain types of correlated second derivative calculations presently only GIAO NMR shift calculations and directs ACES II to either treat all perturbations at once or treat them sequentially The latter approach results in less demand for physical disk space at the cost of increased cpu time Available options are SIMULTANEOUS and SEQUENTIAL XFIELD integer 0 Sets the X component of an external electric field The value must be specified as an integer and the field used by the program will be N x 1076 This
155. vided that they are brought into irrep 1 by reducing the symmetry of the computational point group Note that in C symmetry all instabilities will be in irrep 1 Likewise instabilities that take RHF wavefunctions into UHF ones must be followed using a UHF calculation REF UHF Following instabilities often leads to solutions that are lower in energy but lack the symmetry of the molecular framework are heavily spin contaminated in the UHF case or are otherwise non physical Instabilities in Hartree Fock wavefunctions is a very complex problem and must be treated with great care in order to obtain meaningful information Experience has shown in general that molecules with instabilities will have more than one and that solutions obtained by following an instability can have further instabilities It is impossible to predict exactly what will be uncovered by a stability analysis and the current drivers for ACES II are configured to handle only the simplest cases Examining a tree of instabilities or other specialized procedures will require the user to create an appropriate driver of their own which is usually specialized to the system under study Even though a surprising number of molecules will exhibit instabilities under various conditions not all instabilities indicate pathological cases or intractable problems 9 1 10 Time dependent Hartree Fock To provide properties such as frequency dependent polarizabilities ACES II provides
156. w article e R J Bartlett Ann Rev Phys Chem 32 359 1981 a Closed shell RHF MBPT for molecules e R J Bartlett and D M Silver Phys Rev A10 1927 1974 J Chem Phys 62 3258 1975 ibid 64 4578 1976 e R J Bartlett and I Shavitt Chem Phys Lett 50 190 1977 b Open shell UHF MBPT for molecules e R J Bartlett and G D Purvis III Int J Quantum Chem 14 561 1978 c Open shell ROHF MBPT for molecules e W J Lauderdale J F Stanton J Gauss J D Watts and R J Bartlett Chem Phys Lett 187 21 1991 J Chem Phys 97 6606 1992 97 12 2 Coupled cluster CC theory Review article e R J Bartlett J Phys Chem 93 1697 1989 a CCD method e R J Bartlett and G D Purvis III Int J Quantum Chem 14 561 1978 b CCSD method e G D Purvis III and R J Bartlett J Chem Phys 76 1910 1982 c CCSDT 1 method e Y S Lee S A Kucharski and R J Bartlett J Chem Phys 81 5906 1984 d CCSD T CCSD method e M Urban J Noga S J Cole and R J Bartlett J Chem Phys 83 4041 1985 e CCSD T method adds one HF case or two non HF case small terms to CCSD T CCSD e K Raghavachari G W Trucks J A Pople and M Head Gordon Chem Phys Lett 157 479 1989 e R J Bartlett J D Watts S A Kucharski and J Noga Chem Phys Lett 165 513 1990 e J Gauss W J Lauderdale J F Stanton J D Watts and R J Bartlett Chem Phys Lett 182 207 1991 e J D Wat
157. which molecular orbitals will be dropped from the post SCF calculation The orbitals are ordered by eigenvalue from the most stable negative energy to the most unstable largest positive energy regardless of irrep The delimiter is a forward slash that separates single orbitals and orbital ranges x y inclusive For example the string 1 5 55 62 64 will cause VTRAN to drop orbitals 1 2 3 4 5 55 62 63 and 64 For UHF calculations the appropriate orbitals are deleted from both spin cases HFSTABILITY handle OFF Checks the stability of RHF and UHF wavefunctions and optionally rotates the orbitals to a lower SCF solution There are three possible options for this keyword OFF does nothing and ON performs a stability check and returns the number of negative eigenvalues in the orbital rotation Hessian FOLLOW performs the stability check and then proceeds to rotate the SCF orbitals in the direction of a particular negative eigenvalue of the orbital rotation Hessian see the explanation of the ROT_EVEC keyword below after which the SCF is rerun 51 ROT_EVEC integer 0 Defines which eigenvector of the orbital rotation Hessian will be used to rotate the original SCF orbitals By default it will use the eigenvector with the lowest eigenvalue of irrep 1 the totally symmetric part of the block factored Hessian This choice often leads to the lowest energy SCF solution For RHF stability checks only those instabilities w
158. ytical representation of the operator The latter includes the coefficient Co the exponent Nm of r and the exponent ou of the gaussian Ur Sener rim For a detailed description see L R Kahn P Baybutt and D G Truhlar J Chem Phys 65 3826 1976 38 7 4 GUESS The GUESS file is used to control the placement of electrons and to manipulate orbitals in the SCF initial guess and can be used only with GUESS READ SO MOS orbitals come from OLDMOS All options in the GUESS file must be specified there are no defaults The following is an example of the GUESS file for a state of the water cation Line 1 H20 TEST Line 2 3 1 1 0 alpha occ Line 3 3 0 1 O beta occ Line 4 O 0 0 O alpha pairs to be swapped irrep 1 Line 5 O 0 0 O alpha irrep 2 Line 6 O 0 0 O alpha irrep 3 Line 7 O 0 0 O alpha irrep 4 Line 8 0 0 O O beta pairs to be swapped irrep 1 Line 9 O 0 0 O beta irrep 2 Line 10 O 0 O O beta irrep 3 Line 11 O 0 0 O beta irrep 4 Line 12 O 0 0 O alpha locking within each irrep Line 13 O O0 0 O beta locking within each irrep Line 14 O 0 0 O alpha printing of initial guess Line 15 O 0 0 O beta printing of initial guess Line 16 0 0 stopping parameters Line 17 1 read from OLDMOS Line 18 1 beta orbitals are copied from alpha orbitals Line 19 0 reuse GUESS for every SCF calculation This is a formatted file The trailing text on each line is not parsed as input as long as each line is at most 80 char

Download Pdf Manuals

Related Search

Related Contents

Samsung Galaxy S Duos Kasutusjuhend    télécharger la notice d`utilisation en français  Statistical Commission Background document Thirty  VIE LOCALE - Châteauponsac  Liste Des Pièces  User Manual for the RC:110A Reader Unit and the Intrinsically Safe    

Copyright © All rights reserved. Failed to retrieve file