Gaussian 03 IOps Reference


Nov. 15 2005

Overlay 1 IOp(1/5) IOp(1/6) IOp(1/7) IOp(1/8) IOp(1/9) IOp(1/10) IOp(1/11) IOp(1/12) IOp(1/13) IOp(1/14) IOp(1/15) IOp(1/16) IOp(1/17) IOp(1/18) IOp(1/19) IOp(1/20) IOp(1/21) IOp(1/22) IOp(1/23) IOp(1/24) IOp(1/25) IOp(1/26) IOp(1/27) IOp(1/28) IOp(1/29) IOp(1/30) IOp(1/32) IOp(1/33) IOp(1/34) IOp(1/35) 1 1 1 1 1 2 3 3 4 4 4 4 5 5 5 5 6 6 6 7 7 7 7 7 8 8 9 9 9 9 9

IOp(1/36) IOp(1/37) IOp(1/38) IOp(1/39) IOp(1/40) IOp(1/41) IOp(1/42) IOp(1/43) IOp(1/44) IOp(1/45) IOp(1/46) IOp(1/47) IOp(1/48) IOp(1/49) IOp(1/52) IOp(1/53) IOp(1/54) IOp(1/55) IOp(1/56) IOp(1/57) IOp(1/58) IOp(1/59) IOp(1/60) IOp(1/61) IOp(1/62) IOp(1/63) IOp(1/64) IOp(1/65) IOp(1/66) IOp(1/67) IOp(1/70) IOp(1/71) IOp(1/72)

10 10 10 10 11 11 11 11 12 12 12 13 13 13 13 13 14 14 14 14 15 15 15 15 15 16 16 16 17 17 17 17 17

102. 104) IOp(1/105) IOp(1/106) IOp(1/107) IOp(1/108) IOp(1/109) IOp(1/110) IOp(1/111) IOp(1/112) IOp(1/113) IOp(1/114) Overlay 2 ii 18 18 18 19 19 19 19 19 20 20 20 20 20 20 21 21 21 21 22 22 22 22 22 23 23 23 23 24 24 24 24 24 25 IOp(2/9) IOp(2/10) IOp(2/11) IOp(2/12) IOp(2/13) IOp(2/14) IOp(2/15) IOp(2/16) IOp(2/17) IOp(2/18) IOp(2/19) IOp(2/20) IOp(2/29) IOp(2/30) IOp(2/40) IOp(2/41) Overlay 3 IOp(3/5) IOp(3/6) IOp(3/7) IOp(3/8) IOp(3/9) IOp(3/10) IOp(3/11) IOp(3/12) IOp(3/13) IOp(3/14) IOp(3/15) IOp(3/16) IOp(3/17) IOp(3/18) IOp(3/19) IOp(3/20) 25 25 25 25 26 26 26 27 27 27 27 27 28 28 28 28 29 29 30 31 32 32 32 34 34 34 34 35 35 35 36 36 36 . 97. 95. 96. 103.IOp(1/73) IOp(1/74) IOp(1/75) IOp(1/76) IOp(1/77) IOp(1/78) IOp(1/79) IOp(1/80) IOp(1/81) IOp(1/82) IOp(1/83) IOp(1/84) IOp(1/85) IOp(1/86) IOp(1/87) IOp(1/88) IOp(1/89) IOp(1/90) IOp(1/91) IOp(1/92) IOp(1/94. 98) IOp(1/101.

IOp(3/21) IOp(3/22) IOp(3/23) IOp(3/24) IOp(3/25) IOp(3/26) IOp(3/27) IOp(3/28) IOp(3/29) IOp(3/30) IOp(3/31) IOp(3/32) IOp(3/33) IOp(3/34) IOp(3/35) IOp(3/36) IOp(3/37) IOp(3/38) IOp(3/39) IOp(3/40) IOp(3/41) IOp(3/42) IOp(3/43) IOp(3/44) IOp(3/45) IOp(3/46) IOp(3/47) IOp(3/48) IOp(3/49) IOp(3/51) IOp(3/52) IOp(3/53) IOp(3/54) 36 37 37 37 37 37 38 38 38 38 38 38 39 39 39 39 40 40 40 40 40 41 41 42 42 42 42 43 44 44 44 44 45 IOp(3/55) IOp(3/56) IOp(3/57) IOp(3/58) IOp(3/59) IOp(3/60) IOp(3/61) IOp(3/62) IOp(3/63) IOp(3/64) IOp(3/65) IOp(3/66) IOp(3/67) IOp(3/70) IOp(3/71) IOp(3/72) IOp(3/73) IOp(3/74) IOp(3/75) IOp(3/76) IOp(3/77) IOp(3/78) IOp(3/79) IOp(3/80) IOp(3/81) IOp(3/82) IOp(3/83) IOp(3/84) IOp(3/85) IOp(3/86) IOp(3/87) IOp(3/88) IOp(3/89) 45 45 45 45 45 46 46 46 46 46 46 47 47 47 47 48 48 48 49 49 50 50 50 50 51 51 51 51 51 52 52 52 52 iii .

IOp(3/90) IOp(3/91) IOp(3/92) IOp(3/93) IOp(3/94) IOp(3/95) IOp(3/96) IOp(3/97) IOp(3/98) IOp(3/99) IOp(3/100) IOp(3/101) IOp(3/102) IOp(3/103) IOp(3/104) IOp(3/105) IOp(3/106) Overlay 4 IOp(4/5) IOp(4/6) IOp(4/7) IOp(4/8) IOp(4/9) IOp(4/10) IOp(4/11) IOp(4/13) IOp(4/14) IOp(4/15) IOp(4/16) IOp(4/17) IOp(4/18) IOp(4/19) IOp(4/20) iv 53 53 53 53 54 54 54 54 54 54 55 55 55 55 55 55 56 57 57 57 58 58 58 59 59 60 60 60 60 60 61 61 61 IOp(4/21) IOp(4/22) IOp(4/23) IOp(4/24) IOp(4/25) IOp(4/26) IOp(4/28) IOp(4/29) IOp(4/31) IOp(4/33) IOp(4/34) IOp(4/35) IOp(4/36) IOp(4/37) IOp(4/38) IOp(4/43) IOp(4/44) IOp(4/45) IOp(4/46) IOp(4/47) IOp(4/48) IOp(60-62) IOp(4/63) IOp(4/64) IOp(4/65) IOp(4/66) IOp(4/67) IOp(4/68) IOp(4/69) IOp(4/71) IOp(4/72) IOp(4/80) IOp(4/81) 61 62 62 62 62 63 63 63 63 63 63 64 64 64 64 64 64 65 65 65 65 65 66 66 66 66 66 67 67 67 67 67 68 .

IOp(4/82) IOp(4/110) Overlay 5 IOp(5/5) IOp(5/6) IOp(5/7) IOp(5/8) IOp(5/9) IOp(5/10) IOp(5/11) IOp(5/12) IOp(5/13) IOp(5/14) IOp(5/15) IOp(5/16) IOp(5/17) IOp(5/18) IOp(5/19) IOp(5/20) IOp(5/21) IOp(5/22) IOp(5/23) IOp(5/24) IOp(5/25) IOp(5/26) IOp(5/27) IOp(5/28) IOp(5/29) IOp(5/30) IOp(5/31) IOp(5/32) IOp(5/33) IOp(5/34) 68 68 69 69 70 70 71 71 72 72 72 72 73 74 74 75 76 76 77 77 77 78 78 78 78 79 79 79 79 80 80 80 80 IOp(5/35) IOp(5/36) IOp(5/37) IOp(5/38) IOp(5/39) IOp(5/40) IOp(5/41) IOp(5/42) IOp(5/43) IOp(5/45) IOp(5/47) IOp(5/48) IOp(5/49) IOp(5/50) IOp(5/51) IOp(5/52) IOp(5/53) IOp(5/55) IOp(5/56) IOp(5/57) IOp(5/55-58) IOp(5/59) IOp(5/60-62) IOp(5/63) IOp(5/64) IOp(5/65) IOp(5/70) IOp(5/71) IOp(5/73) IOp(5/74) IOp(5/75-78) IOp(5/79) IOp(5/80) 81 81 81 81 81 82 82 83 83 83 83 83 84 84 84 84 85 85 85 85 85 86 86 86 86 87 87 87 87 87 88 88 88 v .

IOp(5/81) IOp(5/82) IOp(5/83) IOp(5/84) IOp(5/85) IOp(5/86) IOp(5/87) IOp(5/88) IOp(5/89) IOp(5/90) IOp(5/91) IOp(5/92) IOp(5/93) IOp(5/94) IOp(5/95) IOp(5/96) IOp(5/97) IOp(5/98) IOp(5/99) IOp(5/100) IOp(5/101) IOp(5/102) Overlay 6 IOp(6/15) IOp(6/16) IOp(6/17) IOp(6/18) IOp(6/19) IOp(6/20) IOp(6/21) IOp(6/22) IOp(6/23) IOp(6/24) vi 88 88 89 89 89 89 90 90 90 90 90 90 91 92 92 92 93 93 93 93 94 94 95 95 96 96 96 96 96 97 97 97 97 IOp(6/25) IOp(6/26) IOp(6/27) IOp(6/28) IOp(6/29) IOp(6/30) IOp(6/31) IOp(6/32) IOp(6/35) IOp(6/36) IOp(6/37) IOp(6/38) IOp(6/39) IOp(6/40) IOp(6/41) IOp(6/42) IOp(6/43) IOp(6/44) IOp(6/45) IOp(6/46) IOp(6/47) IOp(6/48) IOp(6/49) IOp(6/51) IOp(6/52) IOp(6/53) IOp(6/54) IOp(6/55) IOp(6/56) IOp(6/57) IOp(6/58) IOp(6/59) IOp(6/60-62) 98 98 98 98 98 99 99 99 99 100 100 100 101 101 101 102 102 102 102 102 103 103 103 103 103 104 104 104 104 104 104 105 105 .

IOp(6/63) IOp(6/64) IOp(6/65) IOp(6/70) IOp(6/71) IOp(6/72) IOp(6/73) IOp(6/74) IOp(6/75) IOp(6/76) IOp(6/77) IOp(6/78) IOp(6/79) IOp(6/80) IOp(6/81) IOp(6/82) Overlay 7 IOp(7/6) IOp(7/7) IOp(7/8) IOp(7/9) IOp(7/10) IOp(7/11) IOp(7/12) IOp(7/13) IOp(7/14) IOp(7/15) IOp(7/16) IOp(7/18) IOp(7/25) IOp(7/28) IOp(7/29) IOp(7/30) 105 105 105 105 106 106 106 106 106 106 107 107 107 107 107 107 109 109 109 109 109 110 110 110 110 111 111 111 111 112 112 112 112 IOp(7/31) IOp(7/32) IOp(7/40) IOp(7/41) IOp(7/42) IOp(7/43) IOp(7/44) IOp(7/45) IOp(7/52) IOp(7/53) IOp(7/60-62) IOp(7/63) IOp(7/64) IOp(7/65) IOp(7/70) IOp(7/71) IOp(7/72) IOp(7/74) IOp(7/75) IOp(7/76) IOp(7/77) IOp(7/87) Overlay 8 IOp(8/5) IOp(8/6) IOp(8/7) IOp(8/8) IOp(8/9) IOp(8/10) IOp(8/11) IOp(8/12) IOp(8/13) IOp(8/14) 112 113 113 113 113 113 114 114 114 114 114 115 115 115 115 115 115 116 116 116 116 116 117 117 117 117 117 117 118 119 119 119 119 vii .

IOp(8/15) IOp(8/16) IOp(8/17) IOp(8/18) IOp(8/19) IOp(8/20) IOp(8/22) IOp(8/27) IOp(8/28) IOp(8/29) IOp(8/30) IOp(8/31) IOp(8/32) IOp(8/35) IOp(8/36) IOp(8/38) IOp(8/39) IOp(8/40) IOp(8/41) IOp(8/42) IOp(8/43) IOp(8/44) IOp(8/45) IOp(8/46) IOp(8/47) Overlay 9 IOp(9/5) IOp(9/6) IOp(9/7) IOp(9/8) IOp(9/9) IOp(9/10) IOp(9/11) viii 119 120 120 120 120 120 121 121 121 121 122 122 122 122 122 123 123 123 123 123 124 124 124 124 125 126 126 126 126 127 127 127 128 IOp(9/12) IOp(9/13) IOp(9/14) IOp(9/15) IOp(9/16) IOp(9/17) IOp(9/18) IOp(9/19) IOp(9/20) IOp(9/21) IOp(9/22) IOp(9/23) IOp(9/25) IOp(9/26) IOp(9/27) IOp(9/28) IOp(9/30) IOp(9/31) IOp(9/36) IOp(9/37) IOp(9/38) IOp(9/40) IOp(9/41) IOp(9/42) IOp(9/43) IOp(9/44) IOp(9/45) IOp(9/46) IOp(9/47) IOp(9/48) IOp(9/49) IOp(9/60-62) IOp(9/70) 128 128 128 129 129 129 129 130 130 130 130 130 131 131 132 132 132 132 132 132 133 133 133 134 134 134 135 135 135 135 135 136 136 .

IOp(9/71) IOp(9/72) IOp(9/73) IOp(9/74) IOp(9/75) IOp(9/81) IOp(9/82) IOp(9/83) IOp(9/84) IOp(9/85) IOp(9/86) Overlay 10 IOp(10/5) IOp(10/6) IOp(10/7) IOp(10/8) IOp(10/9) IOp(10/10) IOp(10/11) IOp(10/13) IOp(10/14) IOp(10/15) IOp(10/16) IOp(10/17) IOp(10/18) IOp(10/19) IOp(10/20) IOp(10/21) IOp(10/22) IOp(10/28) IOp(10/29) IOp(10/30) IOp(10/31) 136 136 136 136 136 137 137 137 137 137 138 139 139 139 140 140 140 140 140 140 141 141 141 141 142 142 142 142 143 143 143 143 143 IOp(10/32) IOp(10/45) IOp(10/46) IOp(10/47) IOp(10/48) IOp(10/60-62) IOp(10/63) IOp(10/72) IOp(10/73) IOp(10/74) IOp(10/75) IOp(10/76) IOp(10/77) IOp(10/78) IOp(10/79) Overlay 11 IOp(11/5) IOp(11/6) IOp(11/7) IOp(11/8) IOp(11/9) IOp(11/10) IOp(11/11) IOp(11/12) IOp(11/13) IOp(11/14) IOp(11/15) IOp(11/16) IOp(11/17) IOp(11/18) IOp(11/19) IOp(11/20) IOp(11/21) 144 144 144 144 145 145 145 145 145 146 146 146 146 146 147 148 148 148 148 148 148 149 149 149 150 150 150 150 151 151 151 151 151 ix .

IOp(11/22) IOp(11/23) IOp(11/24) IOp(11/26) IOp(11/27) IOp(11/28) IOp(11/29) IOp(11/30) IOp(11/31) IOp(11/32) IOp(11/33) IOp(11/39) IOp(11/40) IOp(11/41) IOp(11/42) IOp(11/43) IOp(11/45) IOp(11/46) IOp(11/53) IOp(11/60-62) IOp(11/63) IOp(11/70) IOp(11/71) IOp(11/75) Overlay 9999 IOp(9999/5) IOp(9999/6) IOp(9999/7) IOp(9999/8) IOp(9999/9) IOp(9999/10) IOp(9999/11) IOp(9999/12) x 152 152 152 152 152 152 153 153 153 153 154 154 154 154 154 155 155 155 155 155 155 156 156 156 157 157 157 157 158 158 158 158 159 IOp(9999/13) IOp(9999/14) IOp(9999/15) IOp(9999/16) IOp(9999/17) IOp(9999/18) IOp(9999/33) 159 159 159 159 159 159 160 .

IOp(1/6) L102. L109. CONVX=4*CONVF -1 0 N ConvF = 1/600 HARTREE/BOHR OR RADIAN CONVF = 0. L109. RMS EST. L107. L109) = Min(40. L114) N NSTEP = N IOp(1/7) L103. L109. IOp(1/8) L103. L103. L114: MAXIMUM NUMBER OF STEPS (OR NUMBER OF STEPS FOR AN LST SCAN). DISPLACEMENT . L117: Convergence on electric field/charges -1 0 N Default value for optimizations: 10**-7.NVAR+10) (L103.LT. SCAN THE LST PATH. 10**-N. L113.Overlay 1 IOp(1/5) L103 MODE OF OPTIMIZATION 0 1 N FIND LOCAL MINIMUM FIND A SADDLE POINT FIND A STATIONARY POINT ON THE ENERGY SURFACE WITH N NEGATIVE EIGENVALUES OF THE 2ND DERIVATIVE MATRIX L107: MODE OF SEARCH 0 1 LOCATE THE MAXIMUM IN THE LST PATH. L105. L105. L112: MAXIMUM STEP SIZE ALLOWED DURING OPT. L112. L112) = Min(20. L105. CONFV.NVar+20) (L113. 10**-7 in L117. 0 NSTEP = Max(20. L112.NVAR+10) (L102. 1 .0003 HARTREE/BOHR OR RADIAN CONVF = N*10**-6 L116.LT. L114: CONVERGENCE ON THE FIRST DERIVATIVE AND ESTIMATED DISPLACEMENT FOR THE OPTIMIZATION RMS FIRST DERIVATIVE . Default value for single-points: 10**-5 in L116. L113.

Read or CalcFC). search for minima.1 BOHR OR RADIAN (L103. 0 1 2 00 10 20 Whether to update trust radius (DXMaxT. II . Default is III for Tomasi (interlocking spheres) and IV for general for the grad RHO or Vne trajectories. Estm or UnitFC). N DXMAXT = 0. L114). N*0.Do-Self Polarization. scale for transition states. Number of small steps per ABM step to be used in starting ABM and when "slow down" is needed Which approximation to make. L117: Whether to delete points which are too close together: 0 2 No .3 Bohr or Radian (L103. But No Compensation. Apprx.3 Bohr or Radian (L113.Do III and Allow Surface To "Relax" in Solution if no spheres Whether to evaluate densities using orbitals or density matrix. IV . Default is Yes for minima. I .Do Self-Polarization and Compensation. = 0. Which type of basin to use to partition the density isosurface. Default is 5. L121: Time step.0 DXMAXT = 0. eqns. no for TS. in ABM. Search. 1000 2000 3000 4000 N0000 10000 20000 Apprx.01 * N L117: General control. L107: WHETHER TO MAINTAIN SYMMETRY ALONG THE SEARCH PATH. No. Default is to use density. default Yes). default 0. = 0. Yes. Use MOs.1 IOp(1/9) L103: Use of Trust radius. Whether to scale or search the sphere when reducing the step size to the trust radius (Default Scale. = 0.2 Bohr or Radian (L105). 0 1 YES.Don't Do Self-Polarization or "Compensation" Apprx. Default is 4. III .0001 fs. 0 1 2 3 4 N0 N00 N000 surface.). Default is 4 GradVne GradRho Don't Use Basins. Use only the Center of NuclearCharge Use Interlocking Spheres Order of Adam's-Bashforth-Moulton (ABM) predictor-corrector method to use in solving diff. Apprx. max is 9. NO. Use density.

(not valid for L109). Second derivative matrix calculated analytically. Read cartesian forces and force constants from the checkpoint file are convert to internal Read cartesian forces followed by cartesian force constants (both in format 6F12. How close to get to the isosurface in search. only recognized by 103).001) Step size for ABM method in Trudge for isodensity method.000 + N*(0. 3 .0**-N L121: Whether to read in initial velocities: 0 1 2 3 Default (same as 1) Generate random initial velocity Read in initial cartesian velocity (Bohr/sec) Read in initial MW cartesian velocity (sqrt(amu)*Bohr/sec) IOp(1/10) L103.8) from input stream. L105. (5I3. 0 1 2 3 4 5 6 7 8 9 10 Use defaults (not valid for L109). L114: Input of initial Hessian: All values must be in atomic units (Hartree. Read ((FC(I. and radians).F20.I).J). 0 1 2 DEFAULT (TEST for z-matrix or cartesian TS but not for LST/QST or for minimum). coordinates. L113. followed by a blank line.J=1.J). End with a blank card.J. using a default criteria (0.NVAR) (8F10. IOp(1/11) L103: TEST OF CURVATURE. BOMB THE JOB IF THE SECOND SECOND DERIVATIVE MATRIX HAS THE WRONG NUMBER OF NEGATIVE EIGENVALUES.FC(I. using a (10**-N Angstroms) criteria.010 1.0D-6 (N=20) 2.05 Angstroms) Yes. Bohr.0) (L103 only). Use unit matrix throughout. Use semiempirical force constants. Estimate force constants using valence force field. TEST.1 -N 0 N Yes. Use unit matrix (default for L105.6) (L103 only). L112. Read from checkpoint file in internal coordinates. Approx 1. Read I. L109.I=1. L117: Scaling Factor for Determining Overlaps of VDW atoms -1 0 N Turn off scaling Default is 1. DON'T TEST.

05 (N=2) 0. BFGS (not in L103) BFGS. FMAXT = 0. D2Corr and L115. Powell (not in L103). safeguarding positive definateness (not inL103 or L115) D2Corr (New. L121: Multi-time step parameter (NDtrC. D2Corr (New if energy rises.1/N IOp(1/12) L103: OPTIMIZATION CONTROL PARAMETERS 0 1 USE DEFAULT VALUES READ IN NEW VALUES FOR ALL PARAMETERS (SEE INITBS) IOp(1/13) L103. D2Corr (BFGS) D2Corr (Bofill Powell+MS for transition states).1 * N 4 .NDtrP) 0 NN MM00 No multi-time stepping Iterate density constraints NN times per step Do gradient once every MM steps IOp(1/14) L103: Max. IOp(1/15) L103. D2Corr (Old. number of bad steps to allow before attempting a linear minimization (i.L114. 1 for minima). 7 for L103 TS.L113. only in L103 and L115). Powell for L113 and L114). 0 N Default (0 for TS.e.L109: ABORT IF DERIVATIVES TOO LARGE -1 or 0 N No force test at all. no quadratic step). otherwise BFGS).L115: Type of Hessian Update: 0 1 2 3 4 5 6 7 8 9 Default (9 for L103 minimization.linear only starts with the N+1st. use initial Hessian).. D2Corr (No update. Allow N -. only in L103 and L115).0 N 0.

0001 EIGMIN = 1.L113. Default (2000000). Do optimization in Z-matrix coordinates.IOp(1/16) L103.0 HARTREE / BOHR**2 OR RADIAN**2 EIGMAX = 0. will occur. Do full optimization in redundant internal coord. AND PROCESSING CONTINUES. Skip MM atoms in internal coordinate definitions and do microiterations the new way.L113.L114: MINIMUM ALLOWABLE MAGNITUDE OF THE EIGENVALUES OF THE SECOND DERIVATIVE MATRIX. Do full optimization in redundant internal coords with large molecular tools. THE SIZE OF THE EIGENVALUE IS REDUCED TO THE MAXIMUM. done in L402. 0 1 10 20 30 40 50 100 1000 2000 3000 10000 20000 100000 0000000 1000000 2000000 3000000 4000000 Proceed normally Second derivatives will be computed as directed on the variable definition cards. SIMMILAR TO IOp(16) 0 N EIGMIN = 0.L114: MAXIMUM ALLOWABLE MAGNITUDE OF THE EIGENVALUESOF THE SECOND DERIVATIVE MATRIX. Read the AddRedundant input section for each structure. in L103. IOp(1/19) L103: SEARCH SELECTION 5 . Skip MM atoms in internal coordinate definitions and do microiterations the old way.1 * N IOp(1/17) L103. in L120. Microiterations for pure MM. IF THE LIMIT IS EXCEEDED. Include MM atoms in internal coordinate definitions (no microiterations). No optimization Do optimization in cartesian coordinates. Do not define H-bonds Define H-bonds with no related coordinates (default) Define H-bonds and related coordinates Reduce the number of redundant internals Define all redundant internals Old definition of redundant internals. Do full optimization in pruned distance matrix coords. 0 N EIGMAX = 25. / N IOp(1/18) L103: Coordinate system.

L109.0 2 3 4 5 6 7 8 9 10 11 13 Default (same as 6). L108. EXPERT MODE: CERTAIN CUTOFFS USED TO CONTROL THE OPTIMIZATION WILL BE RELAXED. RFO if not.L114: Search Selection: 0 1 P-RFO OR RFO STEP ONLY (DEFAULT) P-RFO OR RFO STEP FOR "WRONG" HESSIAN OTHERWISE NEWTON-RAPHSON IOp(1/20) L101. Linear as usual. L106. L113. Newton-Raphson and linear. GDIIS and linear GDIIS only. VECTOR . RFO and linear. First-order simultaneous optimization. Quadratic if curvature is correct. RFO if not. 0 1 2 Yes. Newton-Raphson only. RFO without linear. EIGMAX AND EIGMIN.L113. STEEPEST DESCENT AND LINEAR ONLY WHEN ESSENTIAL. IOp(1/22) L107: Whether to reorder coordinates for maximum coincidence. Read in a re-ordering vector from the input. DXMAXT. THESE INCLUDE FMAXT.L114: EXPERT SWITCH. L115: KIND OF SEARCH: 0 1 6 BOTH DIRECTIONS AND GENERATE SEARCH VECTOR FORWARD DIRECTION AND GENERATE S. L110: INPUT UNITS 0 1 2 3 ANGSTROMS DEGREES BOHRS BOHRS DEGREES RADIANS ANGSTROMS RADIANS IOp(1/21) L103. 0 1 NORMAL MODE. Assume reactant order equals product order. LINEAR AND STEEPEST DESCENT. No linear search. Quadratic if curvature is correct.

gradient 1. 0 1 2 Default (Yes). Energy + Forces + Force constants IOp(1/24) Whether to round tetrahedral angles. VECTOR 8F10.6 IOp(1/23) L112: Derivative availability. Yes. Normal accuracy for HF (energy and gradient both 1. 0 1 2 Energy only. the field in /Gen/ must be cleared each time. 1000*I+100*J+10*K+L) 7 . VECTOR BOTH DIRECTIONS AND GENERATE S. VECTOR 8F10.d-4) Fine grid accuracy for DFT (Energy 1. Energy + Forces.2 3 4 5 6 7 BACKWARD DIRECTION AND GENERATE S.d-6) IOp(1/27) = IJKL (i. Gradient 10**-(MM). IOp(1/26) Accuracy of function being optimized: -NNMM Energy 10**-(NN). Yes. IOp(1/25) Wether SCRF is used with numerical polarizability: 0 1 No.e.001 degree.d-7. VECTOR 8F10.6 FORWARD DIRECTION AND READ S.d-5.6 BACKWARD DIRECTION AND READ S. gradient 1. round angles within 0. VECTOR FORWARD DIRECTION AND READ S. No. -1 0 1 2 3 Read in values Default (same as 1).d-7). VECTOR 8F10.6 BOTH DIRECTIONS AND READ S. Standard grid accuracy for DFT (Energy 1.

Default. of the two redundant coordinates are taken as the redundant coords for the TS. Get Z-matrix from the checkpoint file. the first one is the reactant. default J=5) LST constraint in internals QST constraint in internals LST constraint in distance matrix space QST constraint in distance matrix space Control parameters for climbing phase of QST (e. N. Abort job if model builder generates a z-matrix with too many variables. same as -1.1*J*R + 0.Transition state searching using QST and redundant internal coordinates L= 0. The third one is the initial guess of the transition structure. The union without QST. By model builder. By model builder. structure) from the checkpoint file. R and P are used to guide the QST optimization of the IOp(1/28) L103: Number of translations and rotations to remove during redundant coordinate transformations: -2 -1 0 N 0. QSTRad = 0. the second one is the product.g. GET CARTESIAN COORDINATES ONLY FROM THE CHECKPOINT FILE. R and P are used to guide the QST optimization of the TS. Set all optimization flags to optimize. Read optimization flags in format 50L1 after the z-matrix. either initial guess of the minimizing structure or transition structure Input 2 structures. Normal (6 or 5 for linear molecules). Print details of the model building process. Do not abort job if model builder generates a z-matrix with too many variables. charge and multiplicity. The first one is reactant the second one is the product. default QSTrad = 0. The values of the TS coord are estimated by interpolating the sturcture of R and P. 7 10 000 100 200 1000 2000 8 Get all input (title. IOp(1/29) L101: SPECIFICATION OF NUCLEAR CENTERS 0 1 2 3 4 5 6 stream. model A. GET Z-MATRIX AND VARIABLES FROM THE CHECKPOINT FILE. Default (same as 100). L= 3 TS.05) Input 3 structures.1*(10-J)*P. BY Z-MATRIX BY DIRECT COORDINATE INPUT (must set IOp(29) in L202). but read new values for some variables from the input .1 L= 2 Input one structure.01*I. model B. K = 1-9 J = 1 J = 2 J = 3 J = 4 I = 0-9 Interpolation of initial guess of TS between R and P (TS=0.

Default. Mark Z-matrix constants as frozen variables rather than wired-in constants. 9 .3000 4000 5000 00000 10000 20000 100000 200000 Purge flags except the frozen variables. Generate a symbolic z-matrix using all Cartesians if none is present on the checkpoint file (a Same as one. but retain the redudant internal coordinate definitions. 0 1 YES NO IOp(1/32) TITLE CARD PUNCH CONTROL. Rebuild the coordinate system. IOp(1/30) L103: ARE THE READ-WRITE FILES TO BE UPDATED? THIS OPTION IS SET FOR THE LAST CALL TO 103 IN FREQUENCY CALCULATIONS IN ORDER TO PRESERVE THE VALUES OF THE VARIABLES FOR ARCHIVING. hack to make IRCs work with Cartesian input). (2+3) Purge all flags but keep the coordinate definition. PUNCH. It also suppresses error termination on large gradients. 0 1 DON'T PUNCH. Do not retain symbolic constants. IOp(1/33) L101: L102 L103 L106 L109 L110 L113 L114 0 1 OFF ON DEBUG PRINT IOp(1/34) L101 L102 L103: DEBUG + DUMP PRINT 0 1 OFF ON IOp(1/35) RESTART (L102-L112). same as 10000.

N0 000 100 200 Continuation of run. L110.01 Angstrom in L110. 0 N -1 -N>1 Use internal default (0.001 Angstroms in L106. 0.0001*N (angstroms in L106. SUPPRESS CHECKPOINTING. THIS IS THE LAST POINT AT WHICH ANALYTIC SECOND DERIVATIVES WILL BE DONE. In L103: Initial entry of guided optimization using N levels. L109.0 1 NORMAL OPTIMIZATION. electric field au in L111). free-format.005 A in L109. L108. In L106: differentiate wrt electric field. GET GEOMETRY. (L106. L111). and L105). In L106: differentiate wrt field and nuclear. WAVEFUNCTION. L110. IOp(1/38) Entry control option (currently only by L106.001 Use step-size of 0. L103. 0 1 NORMAL CHECKPOINT OF OPTIMIZATION. Use step-size of 0. 1 for others). L109. DELETE THE D2E FILE AND THE BUCKETS AND TRUNCATE THE READ/WRITE FILES. Read stepsize (up to 2 for L106. L111. and L112 but not L102. 0. In L106: differentiate wrt nuclear coordinates. 0. FROM THE CHECKPOINT FILE. IOp(1/37) D2E CLEANUP (obsolete) 0 1 NO CLEANUP. L109. FIRST POINT OF A RESTART. L110. au in L111). ET. IOp(1/36) CHECKPOINT. In L110 and L111: differentiate energy N times. 10 .0001*N atomic units everywhere. In L106: differentiate Nth derivatives once. Path step size in L115. Initial entry. IOp(1/39) Step size control for numerical differentiation. L107. 0 1 N>1 .

Read from checkpoint file. Initial 11 by a blank line. L117: Cutoff to be used in evaluating densities.0. . Just update. terminated by a blank line. The default. centers will be on atoms Read-in centers and radii on cards Force Merz-Kollman radii (Default) Force CHELP (Francl) recommended radii. Read in replacement radii for selected atom types as pairs (IAn. 0 1 2 3 Dipole Quadrupole Octapole Hexadecapole L117: How to define Radii 0 1 2 10 20 30 100 200 Default is 11 Use internally stored Radii.Rad). Read from input stream. -1 for the initial geometry IOp(1/42) L103. L115: Number of points along the reaction path in each direction.IOp(1/40) L113.0D-10 1. 0 N 1.Rad) or (Symbol. Default is 6. execpt for CalcAll.Rad). terminated Read in replacment radii for selected atoms as pairs (I. -1 0 N Pick up analytic second derivatives every time. IOp(1/41) Step number of optimization from which to take geometry. Force CHELPG (Breneman) recommended radii.0D-N IOp(1/43) L116: Extent of Reaction Field. Recalculation the Hessian every N steps. L116: Whether to read initial E-field: 0 1 2 Start with 0. L114: Hessian recalculation.

from the outset. Mass-weighted. 0 1 Do not read isotopes.1 N.0 au N. 12 .M au = NM/10 L121: Seed for random number generator (ISeed) -1 0 N Use system time initialize iseed (Note each run will give different results) Use default seed value (ISeed = 398465) Set random number seed to N IOp(1/45) Read isotopes in L115. points which clearly lie in another basin. 0 NM 0. Internal. The final radius is then automatically optimized separately for each atom. This parameter should be chosen with the parameter Cont in mind 0 NM 10.radius of spheres to be placed around attractors to "capture" the gradient trajectories. Read Isotopes.used in Trudge only to eliminate. Maximum distance between a nucleus and its portion of the isosurface . IOp(1/46) Order of multipoles in numerical SCRF: 0 1 2 3 Dipole Quadrupole Octapole Hexadecapole.M = NM/10 IOp(1/44) IRC Type 0 1 2 3 L117: Default (same as 3). Cartesian.

The default is MK charges. Default: unimolecular IOp(1/52) L101 and L120: Type of ONIOM calculation: 0/1 2 3 00 10 20 100 N000 One layer. IOp(1/53) L120: Action of each invocation of L120: 0 1 2 there. Default: 1 2/1 dont/do optimize product structure. Use atomic charge type N-1 during microiterations. normal calculation Two layers Three layers Default (20) Include electrostatics in model systems using MM charges.IOp(1/47) Number of redundant internal coordinates to allow for. IOp(1/48) IRCMax control. 0 N Default: 50000 N. Default: 1 3/2/1 dont/QST3/QST2 optimize TS structure (for QST input). 3 Restore point M from data on the rwf. 13 Do nothing Set up point MM on rwf from initial data Set up point MM on rwf from initial data and restore point MM on chk file if ONIOM data is present . IOp(1/49) Options to IRC path relaxation (IJKL) L K J I 2/1 dont/do optimize reactant structure. Default: 1 2/1 unimolecular/bimolecular reaction. No electrostatics included in the model systems Do full square for testing. 1 20 Do IRCMax Include zero-point energy.

L115: IRC optimization. use gradients to find the next geometry. NN = MaxLev**2 + 1 (currently 17) to restore real system. gradients. and hessian Restore point MM from RWF but do not create a new model system. gradient. IOp(1/55) L103: Options for GDIIS: ICos*1000+IChkC*100+IMix*10+Method form. for use with parameters in input stream. Yes. 0 1 Default. Amber. IOp(1/57) Whether to produce connectivity: 14 . Use displacements to find the next geometry. Amber allowing any symbol. grid. hessian) of point NN of the ONIOM Next point to do is MM. Dreiding/UFF. IOp(1/56) Set of atom type names to parse: 0 1 2 3 Accept any. Save necessary information (some rwf's. energy.4 5 6 7 NN0 MM000 Integrate energy Integrate energy and gradient Integrate energy. Calc Level High | | | Mid | | | Low S M L 1--3--6 system size 2--5--8 4--7--9* IOp(1/54) Whether to recover initial energy during IRCMax from chk file: 0 1 No.

2 for regular) New versions. 15 . IOp(1/59) Update of coordinates in L103 0 1 2 Default (1 for large opt. NN NN fragments.0 1 2 3 4 5 10 100 Default (4 if reading geom from chk file and connectivity is there. read from rwf file. Connectivity input is in terms of z-matrix entries. 0 1 2 Default (No). generate connectivity. IOp(1/60) Interpret extra integer and fp values in z-matrix as scan information. including dummy atoms. otherwise 3). Yes. No. Useful for treating the full system as having electrons only on the QM atoms. Normal: leave the rwf set up for the low-level calculation on the real system. Yes. No. NN < 50. Old version. read from input stream Yes. MOMM: leave the rwf set up for the real system. Yes. but with NBasis and NBsUse for the high-level calc on the model system. read from checkpoint file. IOp(1/62) Counterpoise control. Yes. Read modifications. IOp(1/58) IRCMax control in L115. IOp(1/61) How ONIOM should leave the rwf at the end of each geomtry: 0 1 2 Default (1).

Use soft and hard-wired. Use soft and hard-wired. then the default is to apply soft parameters with higher priority. Use the last when there are equivalent matches. Do not read modifications to parameter set. Dreiding. IOp(1/65) Control of which terms are included in MM. 2=Gradient.IOp(1/63) Step in counterpoise calculation: MNN NN = 0 1-NFrag M = order of derivatives (1=Energy. UFF. If IOp(67)=3. AMBER.2*NFrag -. corresponding to the 'classes' in FncInf. Supermolecule Fragments with ghost atoms NFrag+1 . 0 1 10 100 1000 10000 Do all (default) Non-bonded Stretching Bending Torsion Out-of-plane 16 . Read modifications to parameter set. soft has priority. abort when different parameters match to the same degree. MMFF (NYI). Use only soft. Lowest 2 digits then have no meaning. Use the first when there are equivalent matches.lone fragments IOp(1/64) Molecular mechanics force field selection: 0 1 2 3 4 5 6 7 000 100 200 300 0000 1000 00000 10000 20000 None. hard-wired has priority. Use only hard-wired. With soft parameters. Quartic fitting field (NYI). MM3 (NYI). MM2 (NYI).

17 . else 1. Copy from chk file. 0 1 2 3 Default: 2 if reading geom from chk file. IOp(1/67) Source of MM parameters. Default (100) Do for atoms which have charge specified or defaulted to 0. Pick up non-standard parameters from chk file. reading from input if requested by IOp(64). 1==> 221) Do QEq. Default (10) Do for atoms which were not explicitly typed. IOp(1/72) L103: Algorithm choice for microiterations. IOp(1/70) L118 Type of sampling (Nact) 0 1 2 3 4 Defalt (same as 3) Orthant sampling Microcanonical normal mode sampling Fixed normal mode energy Local mode sampling ( now only Nact = 0 or 3 OK ) IOp(1/71) Whether to print out input files for each structure along an IRC: 0 1 No. Do for all atoms regardless of initial charge.IOp(1/66) Whether to generate QEQ charges. Yes. or to use the values already there. Generate here. Don't do QEq. over-written the values in AtChMM. 0 1 2 00 10 20 000 100 200 Default (2. Do for all atoms regardless of typing.

L121: Lagrangian constrain method for ADMP (ICType) Half*Gamma*Tr[(P*P-P)**2] + Lambda*[Tr(P)-Ne] + Eta*Tr(P*P-P) 0 0) 1 2 3 4 5-7 8-11 Use Lambda and Eta only. (Gamma=0) Use Lambda, Eta, Gamma. Gamma = .2 Use Lambda, Eta, Gamma. Gamma = 1. Constraints for scalar Mass case: Use exact constraint Sum(ij)[Vij*(P**2-P)ij] Iterative Scheme same as 4. Different initial guesses. 7 is default for scalar mass case. Mass-weighting constraints. Documentation maybe found in DVelV1. 10 is default. Default Same as 7 if no Mass-Weighting (IOp(76) < 0) Same as 10 if Mass-Weighting (IOp(76) >

Constraints for tensorial Mass:

L103: NInit for microiterations. L121: Initial Kinetic energy of the Nuclei (EStrtC) 0 N>0 N<0 Default (.1 Hartree) N*micro-Hartree 0.0 Hartree

Charge scaling for charge embedding in ONIOM. IJKLMN 6th through 1st nearest neighbors of current layer scaled by I*0.2, J*0.2, etc. 0 ==> 5 (no scaling); all layers are scaled by at least as much as ones farther out. The default is 500. M L0 K00 Factor for charges one bond away from link atom Factor for charges two bonds away from link atom Factor for charges three bonds away from link atom IJ etc. The actual factors used are: 0: 1.0

1: 0.0 2: 0.2 3: 0.4 4: 0.6 5: 0.8 6-9: 1.0

ADMP control flag (ICntrl) 0 1 2 3 00 10 20 Standard ADMP Read converged density at every step Fix the nuclear coordinates Test time reversability (MaxStp must be even) Default (20). Read stopping parameters from input. Do not read stopping parameters.


+/- XXXXZYYYY = Ficticous electron mass (EMass) YYYY IOp(76)>0 IOp(76)<0 Z XXXX BoxMas=0 Default (1000) YYYY*.0001 AMU MW core functions more than valence functions. YYYY*.0001 AMU. Use uniform scaling for all basis functions (Note YYYY > 9999 makes no sense) Mass-weighting option. If IOp(76)<0, Z is meaningless. If PBC: Mass of Box Coordinates (BoxMas) = XXXX*.0001 AMU Box coordinates not propagated (default).

Initial Kinetic energy of the density matrix (EStrtP) (For UHF, Alpha and Beta each get half this energy) and Option Number to compute initial kinetic energy. Format of Input: XXYYYY (six digits) IWType = XX N = YYYY (For UHF, Alpha and Beta each get half this energy) 0 N>0 Default (0.0 Hartree) N*micro-Hartree IWType is used to figure out how the initial velocity is is computed (in gnvelp).

If XXYYYY < 0 : Initial velocity = 0.0 Hartee (i.e., currently same as N=0 above)

Sparse in L121 -N 0 1 Sparse here with cutoff 10**(-N), full elsewhere Use full matrices or spase based on standard settings. Use sparse fixed form

IRCMax convergence in L115 Stopping criteria in L118 and L121.

L106: 0/1/2 Cartesian/Normal mode/Internal coordinate differentiation. 2 is NYI. L118: .eq.1 to surpress the 5th order correction after surface hop has been made in Trajectory Surface Hopping calculations. Needs also IOp(10/80=1) Nuclear Kinetic Energy Thermostat Option. (Currently only Velocity scaling is implemented) 19

0 11XXXXX 1000000

No Thermostat. Velocity scaling, but only for the first XXXXX simulation steps. (This options is useful, Velocity scaling, all the way through the simulation.

if thermostating in only required during equilibration.

Nuclear KE thermostat in ADMP -- temperate is checked and scaled every IOp(81) steps.

Temperature for nuclear KE thermostat in L121.

Whether to read in frequencies for electric and magnetic perturbations. 0 1 2 Default (No). Yes. No.

Differentiation of frequency-dependent properties. 0 N No. Mask for which properties on file 721 will be differentiated.

Band gap calculation in PBC ADMP: 0 1 2 Default (No). Diagonalizae Fock matrix to get band gap, evolution, etc. No.

Printing for NMR for ONIOM. 0 1 20 Default (1). Print tensors and eigenvalues.

IOp(1/89) Maximum allowed deviation from average nuclear KE during ADMP. The temperature and pressure are read first. as required by IOp(80). IOp(1/90) To read in the velocity in cartesian coordinates Nuclear Kinetic Energy Thermostat Option. Read isotopes from rwf.L. IOp(1/87) ONIOM integration of density. Average energy (in microhartree) to be maintained during Simulation. K. Read isotopes from chk.M.etc: 0: SCF 1: MP first order 2: MP2 3: MP3 4: MP4 5: CI one-particle 6: CI 7: QCI/CC 8: Correct to second order IOp(1/88) Whether to read in atomic masses (isotopes): 0 1 2 3 4 Default (1 if geometry read from input. 4 if geometry read from chk) Use most abundant isotopes. Read isotopes from input. Integrate densities specified by following digits: Density to use from gridpoint 1 Density to use from gridpoint 2 etc. 0 1 2 K0 L00 M000 Do not integrate. in Kelvin. 21 .2 Print eigenvectors as well. Integrate current densities. for backwards compatibility.

N3. Regular non-coupled macro step. direct /w full Hessian incore. 102. 97. 95. 103. Quadratic micro-iterations. full diagonalization. 96. IOp(1/101. Quadratic micro-iterations. Coupled macro step. Quadratic micro-iterations. Do not use dynamic convergence criteria for the micro-iterations. 104) Phase control in L115 and L118: N1. direct /w MM Hessian incore. fully direct. Coupled macro step. Default(15). direct /w raw MM Hessian incore. Quadratic micro-iterations. Regular micro-iterations. full diagonalization. N2. IOp(1/92) Maximum allowed deviation from average nuclear KE specified in IOp(81). Coupled macro step. Also in microhartree. N4 IOp(1/105) Reaction direction 00 22 Default (Same as 10) . fully direct. 98) IOp(94): IOp(95): IOp(96): IOp(97): IOp(98): <0 0 1 2 3 4 5 10 20 30 40 50 Davidson control for quadratic micro-iterations (see MMOpt2) RFO/Davidson control for quadratic micro-iterations (see MMOpt2) Davidson control for coupled QM/MM macro step (see MMOpt2) RFO/Davidson control for coupled QM/MM macro step (see MMOpt2) Control of quadratic micro-iterations and coupled QM/MM quadratic macro step. IOp(1/94.IOp(1/91) Thermostat Option. direct /w prepared Hessian incore. Coupled macro step.

10 20 0 1 2 00 10 20 000 100 200 300 0000 1000 2000 Forward direction Reverse direction Damped-Velocity Verlet (DVV) options for Dynamic Reaction Path Following Default (Same as 2) Use DVV Do not use DVV Default (Same as 10) Follow the rxn path in the forward direction Follow the rxn path in the reverse direction Default (Same as 200) Time step correction not used Time step correction used but not to recalculate current DVV step Time step correction used and current DVV step recalculated Default (Same as 1000) Use DVV stopping criteria Do NOT use DVV stopping criteria IOp(1/106) Damping constant for DVV Dynamic Rxn Path following (v0) 0 N Default v0=0.003 (N=30) Error=N*0.0001 IOp(1/108) Gradient magnitude for DVV stopping criteria (Crit1) 0 N Default (N=15) N*0.0001 IOp(1/109) Force-Velocity angle for DVV stopping criteria (Crit2) 0 N Default (90 Degrees) Use N Degrees 23 .0001 IOp(1/107) Error tolerance for DVV time step correction (Error) 0 N Default Error=0.04 (N=400) v0 is set to N*0.

IOp(1/113) Pressure for thermochemistry. 24 . N/1000000. 0 N Default (standard temperature. 0 N Default (1 atomosphere. 0 N Default (1 unless specified by IOp in overlay 7 or read in). unless read in). N/1000 degrees. unless read in). N/1000 atmospheres. 0 1 2 -n Do not scale Scale with 1/NRA Scale with 1/Sqrt(NRA) Scale with 1/n (NRA = number of atoms in fragment) IOp(1/111) Step-size to use with steepest descent when L103 is having trouble: -N -1 0 N Scale up to RMS step of N/1000 if DXRMS is less. Effectively disables the scaling Default (50) Scale up or down to maximum change in a variable of N/1000 IOp(1/112) Temperature for thermochemistry.IOp(1/110) Scaling of rigid fragment steps during microiterations. IOp(1/114) Scale factor for harmonic frequencies for use in thermochemistry and harmonic vibration-rotation analysis.

001 DEGREE OF 109. Default: print only for small cases Do not print the cartesian coordinates in the input orientation Do print the cartesian coordinates in the input orientation IOp(2/10) TETRAHEDRAL ANGLE FIXING 0 1 2 Default (don't test). ANGLES WITHIN 0. Do not print the angle matrix. Print dihedral angles. using z-matrix connectivity if possible. using the z-matrix for connectivity info. Use cutoffs instead of the z-matrix for determining which angles to print. IOp(2/11) PRINTING OF Z-MATRIX AND RESULTANT COORDINATES.Overlay 2 IOp(2/9) Printing of distance and angle matrices. using a distance cutoff for connectivity info. Print the angle matrix. Do not print the distance matrix. DO NOT TEST FOR SUCH ANGLES.471 WILL BE SET TO ACOS(-1/3). Print dihedral angles. Print distance matrix. 0 1 2 00 10 20 30 000 100 200 300 0000 1000 2000 Default: same as 2. Default: same as 100. 0 1 2 Default (print if 50 atoms or less) Print Don't print IOp(2/12) CROWDING ABORT CONTROL 25 . Do not print dihedral angles. Default: same as 20.

26 . but is useful for debugging the derivative transformation routines.3E20. Default (same as 2).12). and will determine the However. Do not abort the run regardless of 0 distances. IN 'ATOMS' FORMAT (3E20. If internal coordinates are in use. 0 1 2 3 10 linear independance check. but then disable further use of symmetry. the molecule is not oriented. Abort the run for zero atomic distances only Abort the run if any atoms are within 0. same as 0 for molecules but turns on assignment of space group ops. This is not correct for the job if they are dependant. Note that symm is still called.5 A. and confirm that the new structure has framework group. this is not a full optimization). THE ATOMIC NUMBERS AND COORDINATES ARE PUNCHED IN FORMAT (I2. Do not perform the test. Don't even call symm. for PBC. YES. 0 1 2 NO YES. -1 0 1 Turns on symmetry.0 1 2 3 Default (same as 1).. 2 3 4 5 Bring the molecule to a symmetry orientation. with tight cutoffs. Perform the test.e. 100 Abort the job if the number of z-matrix variables is not exactly the number of degrees of freedom (i. symmetrize the resulting coordinates. Unconditionally turn symmetry off. SO THEY ARE NOT PUNCHED. IN FORMAT SUITABLE FOR COORD INPUT TO Gaussian. Leave symmetry in whatever state it is presently in. the same symmetry. test the variables for linear independance and abort the Compute nuclear forces as well as second derivatives for the test. Call Symm once with loose cutoffs. NOTE. then confirm symmetry Recover the previous symmetry operations from the rwf. ATOMS WILL NOT TAKE THE ATOMIC NUMBERS. IOp(2/13) PUNCH COORDINATES.12) IOp(2/14) Internal coordinate linear independance. IOp(2/15) SYMMETRY CONTROL. but do not abort the job.

0 N>0 N<0 Default (determined in the symmetry package. use same tolerance for orientation. Suppress re-orientation for PBC. Default (10) Do re-orientation for PBC. 0 1 2 3 Abort the job. IOp(2/17) Tolerance for distance comparisons in symmetry determination.d-8). currently 1. IOp(2/18) Tolerance for non-distance comparisons in symmetry determination. Keep going and leave symmetry on. Keep going and leave symmetry on. using the old symmetry. but get symmetry info from the chk. as Hollerith string. 0 N>0 N<0 Default (determined in the symmetry package. 10**-N. Turn on symmetry operations for PBC. use same tolerance for orientation. 10**N. IOp(2/16) action taken if the point group changes during an optimization.6 00 10 20 100 Same as 5. IOp(2/19) Largest allowed point group.d-7). currently 1. using the new symmetry. 10**-N. 27 . Keep going. 10**N. IOp(2/20) Number (1-3 for X-Z) of axis to help specify which subgroup of the type specified in IOp(19) to use.

0 1 2 3 Default (1) No. 0 1 No Yes. IOp(2/30) Read in vector of atom types (for debugging). N/1000 Hartree/Bohr**2 28 . Default (no constraint unless reading constraint from chk file). Yes.IOp(2/29) Update of coordinates from current Z-matrix. IOp(2/41) Force constants for Harmonic constraints. Yes. -2 -1 0 N Read in force constants for each cartesian coordinates No constraints. Yes. but remove z-matrix. Pick up structure from rwf698 on the chk file. Read in structure from input stream. format (50I2) IOp(2/40) Save (initial) structure and possible constraints in rwf 698: 0 1 2 3 Default (No).

(VARIOUS IMPLEMENTATIONS MAY OMIT SECOND ROW ATOMS.6-31G.52111) for second row.6-311++G(2df.Si 6-31+G(df.Ar CBS basis #4 -. IOp(6) specifies type.) SEE IOp(6) FOR DETERMINATION OF THE NUMBER OF GAUSSIANS IN THE INNER SHELL. type selected by IOp(6) (=0-4 for V{D. whether intended for use in expanding AOs (IOp(5)) or in expanding the density (IOp(82)).T. USE IOp(8) TO SELECT 5D/6D. IOp(3/6)=5 for MTsmall basis set. EPR-III basis sets. Literature citations in CEPPot.6-31+G(d.6-31G (VALENCE FUNCTIONS ONLY) MINIMAL STO-NG EXTENDED LP-N1G (VALENCE BASIS FOR CORELESS HARTREE-FOCK PSEUDOPOTENTIALS) EXTENDED 6-311G (UMP2 FROZEN CORE OPTIMIZED) BASIS for first row.5.5-31G. G3large basis set. The same numbers are used for all basis sets. Literature citations in SDDPot.9p)-->(631111. Type selected by IOp(6) for H-Ar. MacLean-Chandler (12s. Type selected by IOp(6).6-31+g(d.Large APNO basis set CBS basis #6 -.6}Z) and augmented if IOp(7)=10. EPR-II basis sets. MIDI! basis sets. rest LANL1MB. IOp(6) selects options. B-Ne 2MWB.2p) on H . Cl CBS basis #5 -. Ahlrichs TZV basis sets. Stevens/Basch/Krauss/Jasien/Cundari ECP basis sets for H-Lu.Core correlation basis set Dunning cc basis sets. GENERAL--SEE ROUTINE GenBas FOR INPUT INSTRUCTIONS. Dunning/Caltech basis sets. He 6-311+G(2df) on Li . 17 18 19 20 21 22 23 24 25 26 Stuttgart/Dresden ECP basis sets.Be 2SDF. Coreless: Li. 6 7 8 9 10 11 12 13 14 15 16 LANL ECP basis sets.Ne 6-311+g(3d2f) on Na . S. 29 .p) on H . 0 1 2 3 4 MINIMAL STO-2G TO STO-6G EXTENDED 4-31G. use daggers if any polarization CBS basis #3 -. G3MP2large basis set.Ne 6-311++g(3d2f) on Na . CBS basis #1 -. Ahlrichs SV basis sets.p) on P.p) on H. 5 SPLIT VALENCE N-21G (OR NN-21G) BASIS FOR FIRST OR SECOND ROW ATOMS. UGBS basis set.Overlay 3 IOp(3/5) TYPE OF BASIS SET.Q.Ar CBS basis #2 -.

Same as 1. If non-standard ECPs are in use. and Goddard. N-21G. IOp(3/6) NUMBER OF GAUSSIAN FUNCTIONS N STO-NG. LANL2 ECP (where available. they are read along with the basis set information. DEFAULT OPTIONS IOp(6)=0 IF IOp(5)=0: N=3 STO-3G IF IOp(5)=1: N=4 4-31G IF IOp(5)=2: N=3 STO-3G (VALENCE) IF IOp(5)=3: N=3 IF IOp(5)=5: N=3 WHEN IOp(5)=7 (GENERAL BASES). LANL2 ECP (where available. J.STO-NG-VALENCE. where one wants to modify the basis differently during different steps. . Chem. Chem. THIS OPTION IS USED TO CONTROL WHERE THE BASIS IS TAKEN FROM: 0 1 2 3 4 1x READ GENERAL BASIS FROM THE INPUT STREAM.AL LP-41G FOR OTHER ROW1 AND TWO ATOMS. CEP-31G. NOTE IF IOp(5)=3 AND IOp(6)=8 . Dunning valence double-zeta.B. DZ. DZ.N-31G. 85. When IOp(5)=9 (CEP basis) this option selects the type (H-Ar only): 0 1 30 CEP-4G.LP-N1G. MBS. See Rappe. 1662 (1981) and J. When IOp(5)=6 (LANL basis and potentials) this selects the type: 0 1 2 3 0 1 2 LANL1 ECP. 85. SHC ECP on second row). Smedley. 3546 (1981). When IOp(5)=8 (Dunning bases) this option selects the type: Dunning full double-zeta. This option is useful when doing general basis geometry optimizations or properties using a wavefunction on the checkpoint file. for density basis (generated here from 1) Same as 2.NA.BE. MBS. WAG basis (Dunning VDZ on first row. otherwise LANL1). Read the general basis from the checkpoint file. selected by IOp(6) Auto-generated. LP-31G FOR LI.27 28 DGauss basis sets. for density basis (generated here from 2) Read from the alternate file and remove functions/ECPs for inactive atoms. Phys. otherwise LANL1). READ THE GENERAL BASIS FROM THE RW-FILES AND MERGE WITH THE COORDINATES IN BLANK COMMON TO PRODUCE THE CURRENT BASIS. LANL1 ECP. Used for counterpoise calculations. useful only for density basis sets.MG. Phys.

d. ONE SET OF P-FUNCTIONS AND ONE SET OF D-FUNCTIONS ON HYDROGENS. Augment non-hydrogens only (cc basis sets only). 2d. on heavy atoms (scaled up/down by a factor of 2 from the standard single D value). 2 D-funcs. When IOp(5)=17 (Stuttgart/Dresden ECP bases) this option selects the type according to: 6 7 0 1 2 3 SDD SDD for Z > 18. 31 . P-FUNCTIONS ON HYDROGENS. Three sets of p-functions. D-FUNCTIONS ON HEAVY ATOMS (2ND ROW ONLY FOR 3-21G). Three sets of d functions and two sets of f-functions. A DIFFUSE S FUNCTION ON HYDROGENS. along with LP-31G potential. 0 1 2 3 4 5 6 7 8 9 10 20 100 200 300 400 500 600 700 1000 NONE. Three sets of p-functions and one set of d-functions. Three sets of d functions. VSTO-4G basis for 1st row. CBS-Q d(f). TWO SETS OF P-FUNCTIONS AND ONE SET OF D-FUNCTIONS ON HYDROGENS. When IOp(5)=26 (Coreless basis) this selects the choice of basis (the same ECPs are used regardless): Default (3) Primitives which match the ECPs. Three sets of d functions and one set of f-functions. When IOp(5)=7 (DGauss basis sets): 1 2 3 4 5 DGDZVP DZVP2 DGTZVP DGA1 (fitting basis) DGA2 (fitting basis) IOp(3/7) DIFFUSE AND POLARIZATION FUNCTIONS.2d on 2nd and later atoms.p) -.p polarization basis Tight d for VnZ+1 (W1 theory) A SET OF DIFFUSE SP FUNCTIONS ON HEAVY ATOMS. 1d on 1st row atoms. TWO SETS OF D-FUNCTIONS AND ONE SET OF F-FUNCTIONS ON HEAVY ATOMS.d. ONE SET OF D-FUNCTIONS AND ONE SET OF F-FUNCTIONS ON HEAVY ATOMS (indicates an extra tight 2df with ccp basis sets. 2 SETS OF P-FUNCTIONS ON HYDROGENS.2 CEP-121G. Functions from extended Huckel theory. D95 and no ECP otherwise.

325*(IA-1). He=1. the values for H-Ar can be determined by Slater's rules: H=1. D. Na-Ar=(0. 6D/7F 5D/7F 5D 5D 5D 5D 5D/7F IOp(3/9) Where 308 should store dipole velocity integrals. For VSTO-nG. FORCE 6D.95)/3 None. Add ghost atoms to /B/ so that every shell is on a separate center. Store in RWF N. FORCE 7F.IOp(3/8) SELECTION OF PURE/CARTESIAN FUNCTIONS. Li-Ne=0. 0 SELECTION DETERMINED BY THE BASIS N-31G N-311G N-21G* STO-NG* LP-N1G* LP-N1G** GENERAL BASIS 1 2 10 20 FORCE 5D.7. Write over the dipole length integrals (518). 0 -1 N Usual place (572). IOp(3/10) Modification of internally stored bases (default 12000): 0 1 10 100 1000 2000 10000 20000 100000 200000 300000 DEFAULTS STO-NG STANDARD SCALE-FACTORS. F. Split S=P AO basis shells into separate S and P shells.2. 32 .65*I-4. Massage the data in Common /B/ and Common /Mol/. Do not split AO S=P=D Uncontract the AO basis. shells. Do not split S=P AO shells. Uncontract the density basis Uncontract both basis sets. Read in general basis data in addition to setting up a standard basis. Split S=P=D= AO shells into S=P. FORCE 10F.

15 FIRST ROW ATOMS ATOM B C N O F 1S 1.00 3SP* 1.40 2SP 3SP 0.55 2.00 2SP 1.99 0.36 4.00 1.48 3.00 1.00 3SP 0.43 17.00 1.31 5.24 1.88 3.PHYS.33 J.01 1.90 4.00 1.79 6.69 3. 2686 (1963) OUTER SHELL HAS BEEN SELECTED ON THE BASIS INNER SHELLS ARE BEST ATOM VALUES OF NUMEROUS OPTIMIZATION STUDIES ON VARIED SMALL MOLECULES.47 16.98 1.0 FOR LI-AR (INNER AND OUTER) STANDARD POLARIZATION EXPONENTS FOR N-31G* AND N-31G** BASES 33 .ATOM H HE LI BE B C N O F NE NA MG AL SI P S CL A 1S 1.00 SECOND ROW ATOMS ATOM P S CL 1S 1.98 0.00 1.72 1.75 1.12 1.68 5.15 1.03 1.53 14.80 1.00 2SP 1. 38.10 2.66 8.99 1.95 2.74 1.05 2.67 6.50 1.67 7.61 11.98 0.59 12.98 1. N-31G (ALSO N-31G* AND N-31G**) STANDARD SCALE-FACTORS HYDROGEN H 1S 1.00 2SP* 1.00 1.64 10.20 1S* 1.CHEM.02 1.56 13.65 9.04 0.69 2.70 1.90 2.26 6.25 2.83 5.68 4.00 0.70 1.01 LP-N1G SCALE=1.00 1.50 15.75 1.00 1.

8 STANDARD POLARIZATION EXPONENTS FOR STO-NG* BASIS. ZINDO/1. SUITABLE FOR USE WITH THE OPEN SHELL (UHF) SCF. CNDO/2. but not yet accepted by them. IOp(3/13) Nuclear center whose Fermi contact terms are to be added to the core hamiltonian. 0 -1x -6 No. CAN ONLY BE USED WITH THE CLOSED SHELL SCF. The magnitude is specified by IOp(3/15). Read coefficients of field.6 0. ATOM NA. starting with electric field. SCRF calculation -. RAFFENETTI '3' INTEGRAL FORMAT. blank terminated.1 0.2 0. ZINDO/S. IOp(3/14) Addition of electrostatic integrals to core hamiltonian. to account for sparkles. 34 . RAFFENETTI '1' INTEGRAL FORMAT IS USED.09 0. PROCEDURES.39 IOp(3/11) CONTROL OF TWO-ELECTRON INTEGRAL STORAGE FORMAT. up through 34 elements (hexadecapoles) in free format. 0 1 2 3 9 REGULAR INTEGRAL FORMAT IS USED. translation vectors and d functions properly: 1 2 3 MNDO/AM1. INDO/2. SUITABLE FOR USE WITH OPEN SHELL RHF SCF AND THE POST-SCF USE ILSW TO DECIDE BETWEEN RAFFENETTI 1 AND 2. IOp(3/12) Flag for semi-empirical runs.multiply moments by fudge factor for charged species. MG AL-CL VALUE 0.4 0.ATOM H LI BE B C-NE VALUE 1. RAFFENETTI '2' INTEGRAL FORMAT.

xyz field second derivatives. Read components of moments off rwf 521 on chk file. based on IOp(3/5). ECPs if defined with the basis set. no ECP otherwise. followed by xx. Read components of electric field only from /Gen/.zzz. Default. and hexadecapole perturbations).10). LANL2 potentials.zzzz.yyz.zzzy. xxy.yyyy.yy. yyyx. The nuclear repulsion energy is also modified appropriately. unused READ IN FROM CARDS (SEE PINPUT FOR DETAILS) Dresden/Stuttgart potentials .xyy. Read components of moments off rwf 521. zzxy field third derivatives in format (3D20. quadrupole. octopole. xxx.xxyz. USE INTERNALLY STORED 'CORELESS HARTREE-FOCK' Goddard/Smedley SECE/SHC potentials. 1-34 Just component number n in the above order with magnitude given by IOp(3/15).zz.SDD combination Dresden/Stuttgart potentials .yyxz. Yes.xxz. and the electric field is stored in Gen(2-4). and xxxx.y.z components of electric field.yyzz. Dresden/Stuttgart potentials . IOp(3/17) SPECIFICATION OF PSEUDOPOTENTIALS -1 0 1 2 3 4 5 6-7 8 9 10 11 Read potential in old format. N N * 0. read if general basis. IOp(3/15) Magnitude of electric field.xxzz.xy. No.-5 -4 -3 -2 -1 Read components of electric field only from /Gen/ on checkpoint file. xxxz.xz.yyy.xzz. Yes. D95V. LANL1 potentials.xxxy. IOp(3/16) Pseudopotential option 0 1 2 Default.xxyy. yzz.yz electric field gradient.0001.SDD for Z > 18.zzzx.yyyz. (These correspond to dipole. read 12 cards with x.SDF 35 . Stevens/Basch/Krauss CEP potentials.

MWB (second set) Dresden/Stuttgart potentials . 55296 words on Cray). PRINT DON'T PRINT IOp(3/19) SPECIFICATION OF SUBSTITUTION POTENTIAL TYPE 0 N DONT USE ANY SUBSTITUTION POTENTIALS REPLACE THE STANDARD POTENTIAL OF THIS RUN (EG. IOp(3/21) Size of buffers for integral derivative file.CHF).MHF (first set) Dresden/Stuttgart potentials .MDF Dresden/Stuttgart potentials . N integer words. 0 N Default (3200 integer words).MHF (second set) Dresden/Stuttgart potentials .SHF Dresden/Stuttgart potentials .12 13 14 15 16 17 18 19 20 Dresden/Stuttgart potentials . IOp(3/18) PRINTING OF PSEUDOPOTENTIALS 0 1 2 PRINT ONLY WHEN INPUT IS FROM CARDS or if GFPrint was specified. Alternative potentials for coreless basis. 16384 integer words on VAX.MWB (third set) Pseudopotentials for all coreless basis. WITH A SUBSTITUTION POTENTIAL OF TYPE N WHEREVER SUCH A SUBSTITUTION POTENTIAL EXISTS. N integer words. No longer used. IOp(3/20) Size of buffers for integral file. 36 .MWB (first set) Dresden/Stuttgart potentials . 0 N Default (Machine dependant.

Use looser cutoffs in L314.IOp(3/22) CONTROL OF THE PRE-CUTOFF IN THE TWO-ELECTRON D-INTEGRAL PROGRAM. We are re-using integrals produced earlier in the current calculation. Sleazy. -2 -1 0 1 >0 Use integrals from a previous job read /IBF/ from the checkpoint file. IOp(3/23) Disable use of certain basis functions. Direct SCF. DO ALL INTEGRALS AS WELL AS POSSIBLE in L311. 37 . USE OLD very inaccurate CUTOFFS IN LINK 311. 0 1 NO PRE-CUTOFF. Used only in L312. DO ALL INTEGRALS AS WELL AS POSSIBLE in L314. Print as GenBas input. TEST. terminated by a blank line. Read in a list of basis function numbers in Format (10I5). use the /IBF/ already WE ARE NOT USING TWO-ELECTRON INTEGRALS. TEST. IOp(3/24) Printing of gaussian function table. 0 1 2 10 20 DEFAULT. IOp(3/25) NUMBER OF LAST TWO-ELECTRON INTEGRAL LINK. INTEGRALS ARE COMPUTED TO 10**-10 ACCURACY. 0 1 Use all basis functions. PRE-CUTOFFS DESIGNED FOR THE 6-31G* BASIS. on the RWF. and set their dialgonal core Hamiltonian elements to +100. Print in more readable format. Print old-fashioned table. IOp(3/26) ACCURACY OPTION. Print shell coordinates. STO-3G. 0 1 10 100 1000 Default (don't print). LINK NUMBER.0.

THE SET2E WILL INTERROGATE ILSW TO SEE IF THE SYMMETRY RW-FILES EXIST. 0 N DISCARD INTEGRALS WITH MAGNITUDE LESS THAN 10**-10. d and higher elsewhere. Yes. TWO-ELECTRON INTEGRAL SYMMETRY IS TURNED ON. SYMMETRY HAS BEEN TURNED OFF ELSEWHERE. IOp(3/30) CONTROL OF TWO-ELECTRON INTEGRAL SYMMETRY. All integrals with d's -. HOWEVER. IOp(3/32) Whether to check the eigenvalues of the overlap matrix: 0 1 2 38 Default (4). .IOp(3/27) HANDLING OF SMALL TWO-ELECTRON INTEGRALS. No. IF THEY DON'T. IOp(3/29) Accuracy in L302: 0 N Default (10**-12).L311 does nothing. 0 1 2 Default. use IsAlg. SP integrals in link 311. IOp(3/28) Special SP code control. 0 1 TWO-ELECTRON INTEGRAL SYMMETRY IS TURNED OFF. AND SET2E WILL ALSO TURN IT OFF HERE. DISCARD INTEGRALS WITH MAGNITUDE LESS THAN 10**-N. IOp(3/31) USE OF SYMMETRY IN COMPUTING GRADIENT (Obsolete). 10**-N. NOTE.

0 1 2 3 NO DUMP. Hexadecapole. 39 . PRINT ONE-ELECTRON INTEGRALS. THE INTEGRALS ARE PRINTED (STANDARD FORMAT). COMBINATION OF 1 AND 3. IOp(3/33) INTEGRAL PACKAGE PRINTING. using SVD to reduce expansion space. IOp(3/34) DUMP OPTION. Default (same as 20). ADDITIONALLY. and reduce expansion space if linear dependence is found (NYI). PRINT TWO-ELECTRON INTEGRALS IN DEBUG FORMAT. Yes. Compute absolute overlap over contracted functions. 0 1 3 4 5 6 NO INTEGRALS ARE PRINTED. CONTROL WORDS PRINTED (AS USUAL). Yes. PRINT TWO-ELECTRON INTEGRALS IN STANDARD FORMAT. IOp(3/36) Maximum order of multipoles to compute in L303: -1 0 1 2 3 4 00 10 20 30 None Default (dipole). COMMON/B/ IS DUMPED AT THE BEGINNING OF EACH INTEGRAL LINK. IOp(3/35) LAST INTEGRAL DERIVATIVE LINK (No longer used in overlay 3). Octopole. Compute absolute overlap over both contracted and over primitive functions. Dipole. and use Schmidt orthogonalization to reduce expansion space. Quadrupole. IS THE NUMBER OF THE LAST TWO-ELECTRON INTEGRAL DERIVATIVE PROGRAM. COMBINATION OF 1 AND 4.3 4 5 Yes. ADDITIONALLY. 0 N WHATEVER LINK STARTS WRITING THE INTEGRAL DERIVATIVE FILE SHOULD ALSO CLOSE IT. Do not compute absolute overlaps.

same as 1. 0 1 Initialize. Explicit spdf code. Yes.IOp(3/37) Whether to sort integrals in L320. PRISM. neglect three center one-electron integrals. IOp(3/38) Algorithm for 1e integrals: 0 1 2 00 10 20 Default in 302. neglect 2e integrals with diatomic differential overlap. IOp(3/39) Initialization of force and force constant rwfs. neglect 1e integrals with diatomic differential overlap. neglect three center two-electron integrals as well. PRISM. IOp(3/40) Neglect of integrals: 0 1 2 3 10 20 30 keep all integrals. Default in 308. IOp(3/41) Various semi-empirical methods. neglect four center integrals. 0 1 2 Default (No). Do only overlap and not other 1e integrals. same as 1. Leave alone. No. Rys. 0 40 No NDDO .

Default parameters (same as 5).J=1. IOp(3/43) Solvent charge distribution to add to Hamiltonian: 0 1 2 3 4 None Read charges and DBFs from input stream in input orientation Read from RWF.18) (Format 3D20.J=1. Use Slater's rules scale factors. Read the integrals sequentially.12). 1100000 Do Harris functional scaling atomic densities for current charge and multiplicity. Read parameters for atomic numbers 1-18 in the order Scale (D20. or read-in parameters). geometric mean for NDDO/1 Arithmetic mean in NDDO. followed by ((HDiag(J. 1500000 J as usual but NDDO for K. 1200000 First-order XC. Do CNDO/2. 41 . 1300000 Second-order XC (NYI).I=1. Use STO-3G scale factors. Read from Chk. Read parameters from chk. Do ZINDO/S.12).3). Geometric mean in NDDO. 1400000 Regular SCF with separate K.1 00 10 20 000 100 200 300 400 500 0000 1000 00000 10000 20000 100000 200000 300000 400000 NDDO Default use of NDDO beta parameters (arithmetic mean for indo parameters.I).3). Original INDO/2 Beta and HDiag Parameters. GNDDO/1 parametrization.I). 1000000 Do Harris functional. for testing. Load all the integrals into memory.I=1. Do INDO/2.18) Read parameters from rwf. Default (unit overlap matrix). IOp(3/42) How to form NDDO core hamiltonian in L317: 0 1 2 Default (same as 1). followed by ((Beta(J. Use the unit matrix for the overlap. Same as 1. Use the real overlap matrix. Do ZINDO/1.

just orthogonalize functions on the same center. Force units of Bohr for coordinates. If negative. IOp(3/45) Transformation matrix in L318. only 1e. IOp(3/46) Whether to abort the job if badbas detects an error: 0 1 2 Default (yes). Use unit matrix (for debugging).T and S unchanged. no. IOp(3/44) integral rejection using L318. 0 1 2 use S**-1/2. yes IOp(3/47) Flags for use in PRISM and CalDFT throughout the program. NDDO approximation -. neglect four center transformed integrals. Force use of only the OS path for all calculations. neglect four center and 3 center (ab|ac) integrals. neglect four center and three center (0. Use expanded matrix logic for PBC exact exchange.0) integrals. .no (ab|xx) and no NDDO on 2e and V ints only -.5 10 20 Read charges and DBFs from input stream in standard orientation Force units of Angstroms for coordinates. the perturbation is computed separately and stored in the third and fourth matrices in the core Hamiltonian rwf. Order of multipoles in SCRF for L303. 0 1 2 3 4 5 6 keep all integrals. Do not transform 2e integrals. -1 Bit flags: 0 1 42 If bit 0 is set (use AllowP array) then read in a list of allowed paths.

Default is to use replicated Turn off Schwartz during FMM/NFx calculations. Use Mura radial grid instead of Euler-2 grid. Do not symmetry reduce grid points on unique atoms. No longer used. Force use of FoFCou. matrices only on Fujitsu and NEC. Forbid use of gather/scatter digestion even for small numbers of density matrices. Force all near field in FMM. Make all atoms large in XC quadrature. 43 . even if not doing FMM. Make all shells large in XC quadrature. Do allocation for parallel 2e integrals but run sequentially. Trace input and output using Linda/subprocess. Reserved for more control of scatter/gather. Turn on use of precomputed XC weights. currently only in PrismC. Turn off use of Sqrt(P) in density-based cutoffs. Do allocation for parallel XC but run sequentially. Turn off vFMM. Reverse normal choice of diagonal/canonical sampling in Prism and PrmRaf.2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Reverse choice of whether to precompute distance matrix during numerical quadrature. Force square loops. currently only affects CalDFT. Turn off MP-based cutoffs in FF part of NFx. Turn on angular offsets in XC grid generation. Do nuclear contribution in FoFCou even for non-PBC Do not use special Coulomb algorithm in FoFCou. Force single matrix code in CPKS. The default is diagonal only on vector machines. Forbid use of FoFCou. Reverse normal choice of Scat20 vs. IOp(3/48) Options for FMM: RRLLNNTTWW RR: LL: NN: TT: WW: Range (default 2) LMax (default from tolerance) Number of levels (default 8) Tolerance (default 18) IWS (default 2). replicated Fock matrices. Skip consistency checks for XC quadrature Do not do extra work to use cutoffs better.

Set by PsmSet to indicate whether the NAtoms test for defaulting FMM was passed. Default UseUAB/Use 256. if 128 set. Split primitives for better boxification. Apply symmetry to derivative distributions (NYI). so L302 does not do ECP ints. Turn off parallelism in FMM (does not use parallel logic). IOp(3/51) Parameters for NF exchange and box length (MMMMNN): 00 NN 0000xx MMMMxx no NFx NFx range NN (R+n with n=NN-1).IOp(3/49) More options for FMM: 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 Indicates whether FMM can be used by FoFCou. No Cutoffs 10**-N. 0 1 Default: if needed. Uncontract all shell pairs. 44 . Convert to sparse storage under FoFCou for testing. based on geometry (but minimum for molecules 3. Old routines will be used. IOp(3/53) Accuracy in ECP integral evaluation: 0 -1 N Default. IOp(3/52) Turn off normal evaluation of ECP integrals. Set up for parallel FMM but run loops sequentially. UseUAB. Force FMM on. Default box length.0 Bohr). Do not save as many multipole expansions as possible in memory. ECP integrals are evaulated in L302. Turn on FMM print. Box length MMMM/1000 Bohr. Do not default to FMM.

IOp(3/59) Threshold for throwing avay eigenvectors of S: 0 N Default (10**-4) 10**-N. IOp(3/58) Cholesky control options. of core electrons for Stuttgart/Dresden ECP's. cutoff 10**(-N) IOp(3/56) Cutoff for intermediate matrices during sparse operations: 0 N 100 times smaller than storage cutoff. No Yes. cutoff 5 x 10 ** (N+100) Yes. 45 .IOp(3/54) Type of core density to use with ECPs: -1 0 1 2 None Default (None) Non-relativistic Relativistic IOp(3/55) Use of sparse storage: N<-100 -3 -2 -1 0 N Yes. 10**(-N). IOp(3/57) No. default accuracy (10**-10). intermediate accuracy (5x10**-7) Yes. crude accuracy (5x10**-5) Yes.

IOp(3/64) Set value for ILSW derivative flag. 0 1 No. Nu are basis functions on the same atom. Orthogonalize and remove primitives with 0 or small coefficients. IOp(3/65) Number of k-points: -1 46 Just Gamma point. IOp(3/61) Sparse Semiempirical Hamiltonian Cutoffs in L302: XX XX00 F(Mu. IOp(3/63) Debug option to test point charge FMM. Leave alone.IOp(3/60) Control of orthogonalization and simplification of ccp basis sets. . 0 1 2 Default (1). Mu. Orthogonalize and remove primitives with 0 coefficients.Nu) cutoff). (Defaults to 15 angstroms). -2 -1 0 N Set to zero Set to -1. Lambda are basis functions on different atoms. 10**-(N). Yes. (Defaults to no F(Mu.Nu) atom--atom cutoff criterion (angstroms) Mu.Lambda) atom--atom cutoff criterion (angstroms). set to N. IOp(3/62) Maximum allowed error in S over orthogonalized basis functions: 0 N Default (10**-9). F(Mu. Only active if IOp(3/39)=0.

47 . No. Old logic for NRecip=N. Read setting from checkpoint and modify them by reading from the input stream.N -N About N points. 0 1 2 3 4 5 0100 1000 2000 2100 2200 2300 3000 4000 10000 Default (1) Use defaults. with read-in values. C-PCM. IOp(3/70) SCRF flag. Generate COSMOTHERMO output. Yes. Read setting from the input stream. Set to N. IOp(3/67) Electric-field dependent functions: 0 1 2 3 Default (on if already present in basis read from rwf or chk. SCI-PCM. with standard values. IVC-PCM Cramer/Truhlar solvation model. IEF-PCM. IOp(3/71) IDeriv level flag (for SCRF setup). IOp(3/66) Over-ride setting of NThInc in lineary dependence cutoff: -1 0 N 0 Don't change. Read from rwf. Flag for macroiterations (IPCM). D-PCM. otherwise off). Yes. Onsager. Read setting from checkpoint.

Becke 3 with Perdew 86 correlation. Lee-Yang-Parr correlation. HCTH93. IOp(3/74) Type of exchange and correlation potentials: -24 -23 -22 -21 -20 -19 -18 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 00 00 01 02 03 04 05 06 48 O3LYP. B1LYP. B98. mPW91PW91. Default. PBE1PBE mPW3PBE. Becke3 with Perdew 91 correlation. BA1PBE. Becke3 using VWN/LYP for correlation. B97-1.5 HF + 0. PBE3PBE. Vosko-Wilk-Nusair method 5 correlation. No correlation. mPW1PBE. Becke "Half and Half": 0. Perdew 81 + Perdew 86 correlation. mPW1LYP. Perdew 81 correlation. BA3PBE. VWN 80 (LSD) correlation VWN 80 (LSD) + Perdew 86 correlation . B1B95.IOp(3/72) Solvent type flag (for SCRF setup). LG1LYP. IOp(3/73) ONIOM system flag (for SCRF setup).5 LSD Do only coulomb part. HCTH147. same as 100. skip exchange-correlation. HCTH407. B97-2. Becke "Half and Half" with LYP/VWN correlation.

and 402 is BLYP. LG exchange.0 for DFT and HF.0 and 1.7) Becke 1988 exchange. 205 is Local Spin Density. Note that Becke actually used Perdew correlation rather than LYP.07 08 09 10 11 18 19 20 100 200 300 400 500 600 700 800 900 1000 1100 1200 OS1 correlation PW91 PBE VSXC Bc96 VWN5+P86 LYP+VWN5 for scaling KCIS Hartree-Fock exchange. B97-1 coefficients. HCTH coefficints. IOp(3/75) Number of radial and angular points in numerical integration for DFT: 0 IIIJJJ Use a special grid designed for efficiency (default). with a=0.72 c=0. JJJ angular points. 200 is Hartree-Fock-Slater. Hartree-Fock-Slater exchange (Alpha = 2/3). IOp(3/76) Mixing of HF and DFT. same as B1B95. B97-2 coefficients.5 Xc + Corr Coefficients of 0 and 0 (no exchange). B98 coefficients. 49 . mPW91PW91 coefficients. respectively.81. Becke "Half and Half" 0.5 HF + 0. -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 O3LYP coefficients. Coefficients of 0. PW91 exchange Gill 96 exchange PW86 exchange mPW exchange PBE exchange BA exchange VSXC exchange So 100 is Hartree-Fock. III radial points. X-alpha exchange (alpha= 0. Becke 3 coefficients: aLSD + (1-a)HF + b(dBx) + VWN + c(LYP-VWN).8 b=0.

Negative to turn off screening of basis functions and grid points. 0 -1 -2 -3 -M<-3 N Default (SS weights) Use SS weights. DFT or mixed in accord with IOp(3/76) Mixture of MMMM/10000 DFT exchange and NNNNN/10000 HF exchange. 0 for both. IOp(3/80) Range for microbatching in DFT. Default (coefficients of 1 and zero as determined by IOp(3/42) IOp(3/79) Range cutoff in Becke weights. Use Becke weights with cutoff N Bohr. Sign is applied to the local term. IOp(3/77) Mixing of local and non-local exchange: -1 0 MMMMMNNNNN MMMMM/10000 non-local plus NNNNN/10000 local. 50 . Default (coefficients of 1 and zero as determined by IOp(3/42) IOp(3/78) Mixing of local and non-local correlation: -1 0 MMMMMNNNNN MMMMM/10000 non-local plus NNNNN/10000 local. 1 to set up for CAS-MP2 or 2 to do spin-orbit calculation. Use SS weights with XCal = M/1000. 1000000000 turns of microbatching logic. Use Savin weights. In L510. Use Becke weights with default cutoff of 30 au. 0 for both. Sign is applied to the local term.0 MMMMMNNNNN Default: pure HF.

None (static limit). Keep all functions with angular momentum up to MaxTyp+1. IOp(3/85) Pure vs. Cartesian. N Discard functions with L>=N. 0 1 2 Default (pure for read-in basis). IOp(3/83) Equivalent of IOp(3/6) for density basis. Do not split shells Split F and higher shells away from S=P=D. For auto-generated basis sets: -1 0 momentum. but returning zero for imaginary response contributions. 0 1 2 3 Default (same as 1). For auto-generated basis sets: 0 1 2 10 20 Default (22) Use all products of AOs. 51 . Gross-Kohn form. for debugging. -1 0 N None.IOp(3/81) Frequency-dependence (if any) for XC functional. Also static limit. Same numbering of basis sets as for AO basis. Keep all generated functions. including 7=General basis. Use only AO primitives squared in fititing basis. IOp(3/82) Fitting density basis set for Coulomb in DFT. where MaxTyp is the highest AO angular IOp(3/84) Equivalent of IOp(3/7) for density basis. See comments for IOp(3/5) and IOp(3/6) 28=Generate automatically from AO basis. Cartesian functions in density basis. Pure. Default (-1).

IOp(3/88) Modification of internally stored density basis. solving iterative with stored A. Solve iteratively with no preconditioning Solve iteratively with diagonal preconditioning. 0 1 2 3 4 5 6 7 1xx 2xx 3xx 4xx 52 Default (102 if a fitting set has been included and pure DFT is being used. 0 1 10 100 1000 2000 10000 20000 None. . . Use fits. Form inverse matrix once. solving iterative with direct products.. IOp(3/87) Discard density basis functions based on angular momentum: 0 N No. with A formed to generate preconditioning. F. Iterative. Use fits.IOp(3/86) Discard basis functions based on angular momentum: 0 N No. density shells into S=P. Form A' over neutral distributions via multiplies by A.. Solve iteratively with symmetric block-diagonal preconditioning. D. Add ghost atoms to /B/ so that every shell is on a separate center. shells. Do not use density fits. Split S=P=D=. This is also done if requested Split S=P density basis shells into separate S and P shells.. Massage the data in Common /B/ and Common /Mol/. IOp(3/89) Set up for density fitting. Do not split density S=P=D. Use fits. Discard basis functions with angular momentum >= N... Discard density basis functions with angular momentum >= N. forming Z = modified A^-1. in IOp(3/10). Do not split S=P density shells. no formation of A.. 1 otherwise). Form A' over neutral distributions via direct products. Read in general basis data in addition to setting up a standard basis.

DBFs have J-normalization. default MM=09. for debugging. DBF coefficients are for raw primitives. default NN=06. Douglass-Kroll-Hess 0th order. 53 . RESC.5xx 6xx 1xxx Solve iteratively with non-symmetric block-diagonal preconditioning. DBFs have normalization as for AOs. AO coefficients are for raw primitives. unless scalar relativistic) Point nuclei Single s Gaussians using formula of Quiney et. Put all functions into a single block in forming the preconditioning matrix. IOp(3/91) Scalar relativistic core Hamiltonian: 0 1 2 3 4 Default (1) Non-relativistic. AOs have J-normalization. al Very tight single s Gaussians. Same as 2 but exponents are 100x smaller. Include nuclear charge distributions in DBF set. IOp(3/90) Thresholds for density fitting MMNN 10**(-MM) on iterative solution. AOs have normalization as for AOs. for debugging. IOp(3/93) Nuclear charge distribution: 0 1 2 3 4 10x Default (1. Douglass-Kroll-Hess 2nd order. 10**(-NN) on generalized inverse. Solve non-iterative using precomputed A'^-1. IOp(3/92) Whether read-in basis sets are in terms of normalized primitives? 0 1 2 3 10 20 30 Default (12).

-N 0 N At least N cells in each direction. 0 N -M default (100). At least N. IOp(3/96) Number of PBC cells for DFT: 0 N As many as look significant.Mxxx Use method M to handle nuclear charges during density fitting. IOp(3/94) Range of PBC cells in Bohr. N Bohr. IOp(3/97) Number of PBC cells for exact exchange: 0 N As many as look significant. Multiply usual range by M. based on number of cells having overlap with cell 0. 0 N Default. IOp(3/99) Whether to set up precomputed quadrature grid in L302: 0 54 Default (2 if doing DFT. -1 otherwise). At least N. At least N cells total. . IOp(3/95) Minimum number of PBC cells. No more than N matrices. Based on range estimate (IOp(3/94)). IOp(3/98) Maximum number of density matrices in PBC.

Default is Max((1000.NDBF+100). IOp(3/100) Minimum Number of PBC cells for PBC-MP2 0 N Same as for HF exchange. and point coordinates. IOp(3/103) Maximum number of vectors allowed in expansion space during iterative density fitting. Default is Max(NDBF/2. IOp(3/104) Maximum number of iterations during iterative density fitting.1000). weights. IOp(3/102) Number of density fittings solutions to save from previous SCF iterations. No more than N total. Reuse. Negative to use projected equations rather than least-squares. N. storing grid parameters. IOp(3/105) Re-use of PBC cell data. Default is 6 (using 5 previous solutions plus the current right-hand side to generate the initial guess). storing grid parameters and weights. storing only grid parameter Yes. 0 1 Default (re-use if present).-1 1 2 3 No Yes. IOp(3/101) Maximum range of cells -N N No more than N in each direction No limit. 55 . Yes.

IOp(3/106) Override default number of atoms threshold for turning on FMM (for debugging). 56 . This number is scaled up appropriately if symmetry is in use.2 3 Do not reuse. 0 N Default (60) N atoms for the C1 case. Read from chk file. to compensate for the loss of some symmetry with FMM.

2.1 corresponding to 1. Suppress orthogonalization. EVEN WHEN BASES ARE IDENTICAL. Renormalize and orthogonalize intermediate SCF results which are on the RWF.5 * V. Default (1 for PBC without alter.4 above. Note that variable IGuess here has 4. in which case Read guess from the checkpoint file. Check MOs for othornormality. Don't check MOs for othornormality.3. Re-use Fock matrices instead of orbitals. IOp(4/6) FORCED PROJECTION WHEN GUESS IS READ FROM CARDS (401). FORCE PROJECTED GUESS. Default MO checking (yes). chosen via IOp(11).3. Huckel guess (only valid for NDDO-type methods). EVEN WHEN BASES ARE IDENTICAL. 100 1000 00000 10000 20000 Convert Guess=Check to Guess=Restart or to generating guess depending on what if anything is Use the simultaneous optimization recipe: S**-0. 57 . 9 it. Re-use orbitals not Fock matrices. Read generalized density specified by IOp(38) from the chk file & generate natural orbitals from Huckel is used. otherwise 2). 0 1 2 00 10 20 000 100 200 FORCE PROJECTED GUESS.2. on the checkpoint file. SUPPRESS PROJECTION. Read intermediate SCF results which are on the checkpoint file. Projected ZDO guess. IGuess values of 10-14 are generated internally and are the sparse versions of 0 and 5-8.Overlay 4 IOp(4/5) Type of guess: 0 1 2 3 4 5 6 7 8 it. Default orthogonalization (perform) Schmidt orthogonalize guess orbitals. Guess from model Hamiltonian. Read generalized density specified by IOp(38) from the rwf file & generate natural orbitals from Default. Renormalize and orthogonalize the coefficients which are currently on the read-write files. This uses the Harris functional unless atoms heavier than Xe are present.




SCF symmetry control (401). 0 1 2 3 4 5 10 20 100 200 1000 58 Default, same as 104 Read groups of irreducable representations to combine in the SCF. These are read before any orbitals and before alteration commands. Use no symmetry in the SCF. Pick up the symmetry mixing information from the Alteration read-write file. use the full abelian point group, as represented by the symmetry adapted basis functions produced by link 301. Initial guess orbital symmetries are assigned. (Use symmetry in SCF if possible, but do not assign initial guess abelian symmetries). Localize all occupied orbitals together and all virtual orbitals together Localize the orbitals within the selected or defaulted symmetry. Assign orbital symmetries for printing in full symmetry. Do not assign orbital symmetries in full symmetry. Force the guess orbitals to have the Abelian symmetry. This option can cause the symmetry

adapted basis function common blocks to be modified.

Orbitals to mix to form complex guess (401). 0 1 Mix the HOMO with the LUMO. Read from cards (2I3) pairs of integers indicating which pairs of orbitals are to be mixed. Reading


Type of Guess (401): For iterative ZDO Guess: -1 0 1 2 3 4 5 6 Force old path using old Huckel. Best available (6,4 in order of preference). Old Huckel. CNDO. INDO. New Huckel. Iterative extended Huckel. Harris, converted to IGuess=3 and IZDO=3 here.

For unprojected single diagonalization guess: 0 1 2 3 000 100 200 300 400 500 1000 n0000 MMMM00000 Use functional MMMM Default (same as 1). Use bare core matrix. Dress core Hamiltonian with QEq-based density. Use Harris Functional. Default, same as 2. Use SG1 and 10^-6 accuracy in Harris guess Use FineGrid and 10^-8 in Harris functional. Use UltraFine and 10^-8 in Harris functional. Use user's IRadAn and 10^-8 in Harris functional. Use (199,974) and 10^-12 in Harris functional. Save energy in Gen(43) for Harris functional. Force IDoV=n in HarFok.



Reading of specific orbitals (401). 0 1 No. Yes. For alpha orbitals, read one card with the format for the orbitals, followed by zero or

more sets of IVec (I5) -- vector to replace. If IVec is -1, all NBasis vectors follow. (Vector(I),I=1,NBasis) -- vector in the specified format. Input is terminated by IVec=0. For beta orbitals, the same format as for alpha is used. Note that if alter is also specified, the replacements are read before the corresponding alterations (thus the order is alpha orbitals, alpha alterations, beta orbitals, beta alterations).

Spin-state for initial guess (401). 0 N UHF). Use multiplicity in /Mol/. Use multiplicity N. This is useful for generating guesses for open-shell singlets or unusual

spin states involving orthogonal orbitals by treating them as high-spin in the guess (which only does

Whether to translate basis functions of read in guess (401). 0 1 2 3 Default (same as 2). Use the basis functions as is. Translate to the current atomic coordinates. Translate to the current atomic coordinates, and determine an overall rotation to provide to the read- in orbitals.

Number of open-shell orbitals (not electrons) in 402. 0 N 60 #open electrons. N. Number of electrons in the CAS space.

3/4 point on unless Pulay or Camp-King. INDO. No pseudo-diagonalization. L405: Number of orbitals in the CAS space. Use determinant basis with Sz=b/2. AM1. IOp(4/21) SCF type (402). L405: Trucation level for excitations -. Use pseudo-diagonalization. Pulay. IOp(4/19) L402: Spin change in CI (default based on multiplicity). Just list configurations.default full CAS. CNDO. Use slater determinants. Flags for MCSCF (L405): 1 10 100 1000 Read options from input stream. MINDO/3. 61 . No 3/4. No Pulay (DIIS). Camp-King. 0 1 2 3 4 5 Default (AM1). 3/4. IOp(4/20) Type of model (402): (This is also tested in 401 to see whether atomic number greater than 102 are special flags). no Camp-King.IOp(4/18) Number of orbitals in CI in 402. use pseudodiagonalization). No Camp-King. Default is number of open shells. MNDO. 0 1 2 10 20 100 200 1000 2000 Default (no Pulay.

MNDO. ROHF (NYI). Yes. IOp(4/22) Derivatives? (402). IOp(4/24) Whether to update orbitals. General CI. NDiag in L405.AM1). (For Opt=MNDOFC). Update second force array instead of first. Do 1st derivatives analytically if possible. and ILSW on the rwf (402). (For Opt=MNDOFC). 0 1 2 3 10 Default (don't update). Update. . Update. IOp(4/25) Wavefunction (402). Write formatted file of symbolic matrix elements. N. with N microstates read in.MNDO. Don't update. 403). 2nd derivatives. (For straight semiempirical calculations). Single determinant.AM1). eigenvalues. More flags for MCSCF: IFlag2 IOp(4/23) Number of iterations (402. using specified orbitals. /Mol/. Closed-shell 1/3 CI (only for MINDO3. 0 N Default. NRow in L405.10 binary switches in L405. but don't convert from Lowdin orbitals. Biradical 1/2 CI (only for MINDO3.10000 100000 Write unformatted file (NDATA) of symbolic matrix elements. 0 1 2 3 4 5 -N 62 Default (Same as 1). RHF/UHF from IOp(5). 0 1 2 12 100 1 No. Restart 2nd derivatives. General CI.

0 1 NO DUMP.05 au (according to ZDO) of the HOMO & virtuals within 0. IOp(4/34) DUMP OPTION. IOp(4/31) Root to solve for in CI (402) (Default is 1). default 10**-7). mix all equally. PRINT THE MO COEFFICIENTS. Equal occupations of the lowest N virtuals and high N occupieds. 0 1 2 NO PRINTING.15 IOp(4/28) SCF Convergence (10**-N. mix valence occupieds with 0. PRINT EVERYTHING. IOp(4/33) PRINTING OF GUESS. IOp(4/29) NC in L405. TURN ON ALL POSSIBLE PRINTING. -2 -1 N Yes.IOp(4/26) Whether to mix orbitals in generated guess density: 0 -3 au. No Yes. Yes. mix valence orbitals and an equal number of virtuals. 63 .

Also L510 must use Lanczos. 0 1 For normal route. since there can be many thousands in a large CAS. Eij's calculated here and stored on disk.IOp(4/35) Overlap matrix. Old S parameters. The configurations will not be listed unless see below. IOp(4/37) Selection of old MNDO parameters in L402: 0 1 2 Defaults. with all matels calculated here and stored on disk. 2 As option 1. For direct route. Configs printed as normal. Yes. IOp(4/44) 1 Prepare input for Mp2 implies IOp(21)=10 Slater Det. Copy on disk is used. IOp(4/38) Generalized density to use for natural orbitals: N Density number N. This type of computation can only be done in a CAS comp. Old Si parameters. 0 1 No. This will be the only way to print configs in a direct matel calc. reformat ZIndo integrals and wfn into rwf. but all configurations are printed. A flag is automatically sent to L510 to tell it to compute the remaining atels directly. Overlap assumed to be unity. 0 1 2 Default (copy on disk is used). option generates data for use in MC-SCF generation of zeroth order H note: for b=0 ie no unpaired spins forces use of Clifford Algebra Spinors 64 . IOp(4/43) IDiEij = Switch for direct matel calculation. IOp(4/36) ZIndo reformating.

implicitly. Use diagonal guess density. and IRanGd. Default (-1 if sparse is turned on) Yes. IOp(4/48) Whether to do (sparse) Conjugate Gradient methods in 402: 0 1 2 No. Yes. IRanWt. Use threshold 10**-N. 65 . IOp(60-62) Over-ride standard values of IRadAn. normal CAS calculation. but is smaller subsequently. Yes. L510 will execute normally. Use Lewis dot structure guess density. use the Lewis dot structure to generate a sparse guess directly. and sets up L510 to compute extra matels etc. L405 performs a CAS on the inner space. IOp(4/46) CI basis in CASSCF: 1 2 3 10 Hartree-Waller functions for singlets Hartree-Waller functions for triplets Slater Determinants Write SME on disk IOp(4/47) Convert to sparse storage after generating guess. This occupies as much space as a full CAS in this link.instead of simple determinants IOp(4/45) Ipairs = number of GVB pairs in GVBCAS. There are n pairs: 2*n extra orbitals and electrons will be added into the active space later. This is a normal GVBCAS calculation. No pairs. -1 0 N No. This is the GVBCAS test mode. 0 n Default. -n There are n pairs: 2*n orbitals and electrons of the specified CAS are to be considered to be GVB type orbitals when generating configs / matels.

10**-N. initial threshold 5 x 10 ** (N+100) IOp(4/66) Dielectric constant to be used in MM calculations. Yes. initial threshold 5x10-5. r**-1 Coulomb. IOp(4/65) Tighten the zero thresholds as the SCF calculation proceeds. No variable thresholds. Eps = N / 1000. IOp(4/64) Cutoff for MM non-bonded term. .0. Turn on bond stretches. initial threshold 10**(-N) Yes. 0 N Default (no cutoff). 0 N Eps = 1. Turn on angle bending. Turn on non-bonded terms. 0 1 N N N<-100 Default: Yes. IOp(4/67) Whether to use QEq to assign MM charges. Turn on all terms. 2 otherwise.IOp(4/63) Flags for which terms to include in MM energy: 0 1 2 10 100 1000 10000 100000 Default (111111) Turn on all terms. r**-2 Coulomb. 1==> 221) Do QEq. 0 1 66 Default (211 if UFF. Turn on inversions/improper torsions Turn on torsions. 10**-N.

0 IJKLMNOP List of up to 8 irrep numbers to include. Default (10 for PBC. Do the extra guess regardless. IOp(4/69) Whether to do a new additional guess in addition to reading orbitals from the rwf: 0 1 2 3 00 10 20 Default: yes if no Guess=Alter. Save the Harris guess as an initial Fock matrix. Do for all atoms regardless of typing.2 00 10 20 000 100 200 Don't do QEq. 10**(-N). Just generate orbitals from the Harris guess. and not a small geometry step. 20 otherwise). Default (200) Do for atoms which have charge specified or defaulted to 0. All IOp(4/80) The maximum conjugate gradient step size (MMNN) 67 . Do for all atoms regardless of initial charge. IOp(4/68) Convergence criterion for microiterations in L402: 0 N Default. Default (20) Do for atoms which were not explicitly typed. Harris guess. IOp(4/71) Write out AM1 integrals in 402 0/1 No/Yes. IOp(4/72) Irreps to keep in MCSCF CI-wavefunction. Do the extra guess and store as the initial Fock matrix. Do not do the extra guess.

No preconditioning. Don't use CG DIIS Use CG DIIS. Lambda are basis functions on different atoms. F(Mu.Nu) cutoff).Lambda) atom--atom cutoff criterion (angstroms) Mu. (defaults to 3 cycles).0000 MMNN No maximum step size Step size of MM. IOp(4/110) Scaling of rigid fragment steps during microiterations. Maximum Number of purification cycles per CG iteration.NN IOp(4/81) Sparse SCF Parameters MM NN00 PP0000 Maximum number of SCF DIIS cycles. (defaults to 4 CG cycles). IOp(4/82) Conjugate-Gradient Parameters MM NN00 00000 10000 000000 100000 0000000 1000000 Maximum Number of CG cycles per SCF iteration. MM=01 turns DIIS off) F(Mu.Nu) atom--atom cutoff criterion (angstroms) Mu. (MM=00 defaults to 20 cycles. 68 . Nu are basis functions on the same atom. (defaults to 15 angstroms). Polak-Ribiere CG minimization Fletcher-Reeves CG minimization Use diagonal preconditioning in Conjugate-Gradient. (defaults to no F(Mu.

Default -. HGP sp.only in L502. NNNNN=ICntrl with values as below. Force FoFDir. Force Rys only. Rys f. L510: Direct MCSCF control (L510). Do not compute operator matrices. Use option NNNNN in control of 2e integral calculation.Overlay 5 IOp(5/5) Direct SCF control (L502. No cutoffs. BraKet only. The default for integrals or second derivatives. Sleazy (10**-6) Cutoffs.: ICntrl = Algorithm control: 0 1 2 3 4 5 6 7 8 10 20 30 40 100 200 Default for MCSCF is (1522). Cutoffs for high accuracy. Force HGP only. Compute 2e integrals and store in-core. BraKet up to L=8. forbidding incore Force conventional Something obsolete use option NNNNN in control of 2e integral calculation.(How to Obtain the Integrals) 0 1 2 3 4 5 6 NNNNNx Incore or Direct(FoFDir) according to available Memory. Read the integrals off disk.incremental Fock matrix formation only for direct SCF. Form delta-F each iteration -. Rys f (no sp done here at all). HGP d. Read the integrals off disk OR Incore (Acording to Memory) Compute 2e integrals(Using FoFDir). The default for first derivatives. Compute 2e integrals and forbid in-core. Compute SCF Fock matrices. Compute 2e integrals (Use TrnDir + FoFDir = 4 Can. HGP spd. Form full Fock matrix every time. 69 . HGP df (no sp done here at all). Compute 2e integrals. Rys df (for debugging). Calcs). rest not done here. Cutoffs for 10**-10 accuracy. 0 1 2 3 4 NNNNNx 0000000 1000000 2000000 Default (same as 1). L508).

300 400 500 600 700

Compute CIS operators Compute WTilda terms. GVB: DA==>FJ,FK. Compute regular integrals and load into R03, in canonical form if IOpcl=0 and square form if Compute raffenetti integrals: IOpCl=0 IOpCl=1 IOpCl=2 IOpCl=3 IOpCl=4 1000 2000 3000 4000 Load R1. Load R1 and R2. Load R1 and R03. Load R1, R2, and R03. Load R2 and R03. Do not compute forces. Compute forces. Make derivative Fock matrices Make derivative Fock matrices and form contributions to polarizability derivatives (ie 6 sets of forces will be returned in FXYZ, and 3 extra sets of densities must be supplied in PA,B). 5000 10000 Compute forces using including CIS 2PDM terms. Compute second derivatives.

IOpcl=1. NMatS is used as the dimension of R0 if IOpcl=1.

0000000 1000000 2000000

Default -- incremental Fock matrix formation only for direct SCF. Form full Fock matrix every time. Form delta-F each iteration -- only in L502.

Convergence (RMS density except in L506 (SQCDF), L508(rms rotation gradient), and L510 (Energy)). 0 N 10**-8, except 10**-7 for PBC. 10**-N.


Maximum number of iterations. 0 128, except 512 in L503 and L508.

L510: MAXIMUM NUMBER OF ITERATIONS TO BE DONE (MaxIt) 0 N -1 MaxIt=64 (Default Value) MaxIt=N It does only a CI calculation. Options other than the standard SCF ones:


SELECTION OF THE PROCEDURE OF DIRECT MINIMIZATION (L503). 0 1 2 4 5 STEEPEST DESCENT WITH SEARCH PARAMETERS DEFAULT STEEPEST DESCENT WITH SEARCH PARAMETERS READ (SEE BELOW) CLASSICAL SCF (ROOTHAAN'S METHOD OF REPEATED DIAGONALIZATION CONJUGATE GRADIENTS WITH SEARCH PARAMETERS DEFAULT CONJUGATE GRADIENTS WITH SEARCH PARAMETERS READ: MAX. NUMBER OF SEARCH POINTS (I1) MIN. NUMBER OF SEARCH POINTS (I1) INITIAL STEPSIZE, TAU (G18.5) SCALING FACTOR FOR SUBSEQ. TAU (G20.5) Q (G20.5) Search method (L508). 0 1 2 3 00 10 20 000 100 200 Default (123). Steepest descent. Scaled steepest descent. Quadratic convergence (after rotation gradient is sufficiently small). Default linear search (full search). Do a full linear search to locate a minimum. Do a linear search only if the energy goes up after the initial step. Default handling of wrong curvature (switch direction). Reverse direction if curvature in NR step direction is wrong. Take pure NR steps, even if curvature is wrong.

Flags for L510: 1 10 100 1000 10000 1000 IRdF2, read damping coefficients. IFrzCI, freeze CI coefficients after 1st iteration. Read unformatted symbolic matrix elements from NDATA instead of rwf. Read in damping factors from cards. Use Levy damping. Read Fock matrix restriction matrix.


Number of pair iterations (L504, L506). -1 None; coefficients are frozen at initial values (L504: causes coefficients to be read in in order 71

11 12 22). 0 L510: 5.

IVShft Level shifting: -N -2 -1 0 N Dynamic level shifting to achieve a gap of -0.001*N Dynamic level shift to a default goal (same as -200) No level shifting. Default: -200 for diagonalization calculations, -1 for sparse diagonalization replacements, and if energy DIIS is turned on. Shift by 0.001*N


Whether to allocate only two N**2 arrays for RHF: 0 1 2 Default (No). Yes. No.

Number of GVB pairs (L506). If non zero, the number of orbitals in each pair is read in format (30I2). Each pair consists of the highest available occupied from the guess (after high spin orbs are accounted for) and the lowest available virtuals. If <0, pair coefficients are read; otherwise standard initial values are used.

Action on convergence failure (L502): 0 1 2 72 Default (2). Continue the run even on non-convergence. The ILSW flag for convergence failure is set. Terminate on non-convergence.

use IOp(17)). Turn the current RHF run into a uhf run at the end of this link. Skip core-valence Fock matrix elements. AND THE SYSTEM RW-FILES FOR THE APPROPRIATE DENSITY MATRICES ARE UPDATED (USEFUL IF ONE WANTS A POPULATION ANALYSIS). Number of rows in an initial transformation of MO.-For testing purposes. CONTROL OF ANNIHILATION OF SPIN CONTAMINANTS (L502). Just recompute band structure from stored real-space Fock matrix. first cycle: use initial AO densities. CONTINGENT ON SPACE. 73 . State average density matrices. 0 1 2 ON BESSEL CRITERION ON STRONGER INDIVIDUAL-OVERLAP CRITERION OFF None. Do not use GEW routines even for CP. Also redoes the fit at the end even if using fits during SCF. 0 1 2 CALCULATION IS PERFORMED (PROVIDED OF COURSE THAT ENOUGH SPACE EXISTS IN THE RW-FILES). Fit the converged density even if fitting is not in use during the SCF. REORDERING OF THE ORBITALS (MAINTAINING CONTINUITY OF THE WAVEFUNCTION ALONG THE SEARCH PATH. CALCULATION IS PERFORMED. L503). (Obsolete. 200 1000 2000 10000 ADMP. (More input from cards See below) IOp(5/14) Special functions in L502: 0 1 10 20 100 matrices. IRdNLp. CALCULATION IS BYPASSED. Skip valence-virtual Fock matrix elements. Use Generalized energy-weighted density routines regardless. Terminate after computing the 2e terms at the first iteration.L510: MCSCF flags: 2 10 100 XX000 Generate MOs using UHF natural orbitals. later cycles: transform the density from L121 before calculating the energy and Fock L510: Flags for MCSCF: 1 10 100 1000 10000 100000 Skip valence-valence Fock matrix elements. Use full diagonalization method rather than Lanczos.turned on automatically in FoFDirINFC Number of Frozen Core Orbitals IRdNT.(Calculates directly Nact*(Nact+1)/2 Fock matrices by contracting the AO integrals with the Density matrices. Skip core-valence Fock matrix elements. ADMP.

No. 4 for sparse). 0 1 NO YES Selection of OCBSE vectors (L506). Pseudo-diagonalization whenever possible. Force formation of the Fock matrix using full storage. INHIBIT PERFORMANCE OF MINIMIZATION OF ALTERNATE WAVEFUNCTION PROVIDED BY SECOND ORDER PROCEDURES (L503). Sign Matrix Method. CEM. SNRDMS. Yes. 0 1 DONE TAU IS KEPT FIXED IOp(5/16) Diagonalization method (L502): 0 -N 1 2 3 4 5 6 7 8 1xx 2xx Default (1 for full matrices. KyDiag. By orbital least change. By energy least change. Force formation of the Fock matrix using sparse storage. Lanczos starting vector in L510: -1 74 Read in eigenvector. Default (use Abelian symmetry in diagonalization). CONTROLS THE AUTOADJUSTMENT OF TAU (L503).(ILzVec=-1) See below . Do not use Abelian symmetry in diagonalization. Use Abelian symmetry in diagonalization. 0 1 2 3 00 10 20 Default (1 for L502. Pseudo-diagonalization with real diagonalization every Nth cycle. Yes. keep overall wavefunction the same as the initial guess. but doing the minimal amount of orbital switching to accomplish this. DiagD. PDM. keep occupation of each irrep the same as the initial guess. 2 for L501 and L506). 0 1 2 By eigenvalue. CGDMS.IOp(5/15) Apply Abelian symmetry constraints on orbitals.

choose between replicating integrals and symmetrizing the Fock matrix based on whether the current density matrix is symmetric.LE. Lowdin orthogonalize C+O and V. Use of symmetry (in L502 and L508) and linear equation convergence (in L508): 0 1 2 3 10 20 30 40 100 200 0000 1000 2000 Default (1032 for 502.LE. L510: MCSCF flags. in preparation for CPMCSCF.J)). Choose LinEq convergence based on orbital gradient. Bad for 1st order method.O. Force the density matrix to have full symmetry at every iteration. 0 1 Virtuals obtained by diagonalization of hamiltonians. Don't use natural orbitals each iteration.V by separate Lowdin. Same as 30 in 502 but 20 in 508. SET TO ZERO IF Abs(F(I. 1012 for 508).D-5 FUZZY AND DEGEN READ IN (2D20. 2nd order iteration using RFO type step + level shift and prepare for non-direct CPMCSCF 75 . 2nd order iteration using RFO type step + level shift. then Schmidt.0 IOp(5/17) CONDITION OFF-DIAGONAL TERMS OF THE Fock MATRIX (L503). Don't orthogonalize. Default (1000) If the density matrices pass the symmetry test. VECTORS 0 1 FUZZY=1. replicate integrals so that density matrices and wavefunctions need not If 2E symmetry is on. 2nd order iteration at end. Tighten convergence by an extra factor of 10. Warning!! should be used only for large jobs where Hessian does not fit in memory. Use full 2nd order convergence.D-10. If 2E symmetry is on. DELETE COUPLING TERMS BETWEEN ALMOST DEGENERATE (DELTA E . DEGEN) M. be symmetric. symmetrize them to ensure that they are exactly Do not symmetrize the density matrices. If 2E symmetry is on. Prepare for CPMCSCF(FREQ): Direct method with no storage of Hessian.0 C(N) = 1.FUZZY. DEGEN=2. Just schmidt. then Schmidt. Virtuals obtained by Schmidt orthogonalization to occupieds. Always use tight convergence.0 N C(1) = 1. symmetrize Fock matrices and require proper density matrix symmetry. Force the density matrix to have full symmetry at the first iteration.O.14) Selection of virtual orbitals (L506). symmetric. 0 1 2 5 10 100 200 300 400 500 Orthogonalize C.

Read CI vector and use it every iteration.(the weights N/100 new density. L502. L508): -1 0 1 2 3 4 1x 2x 3x 76 Choose the best given amount of memory available. scaled steepest descent is used): 0 N Default (1. . Force Lower-triangular in-memory storage.(IRdCIV) Use full diagonalization method rather than Lanczos. 30000000 Do SA and prepare for Gradient of Energy difference.8) 20000000 Do SA and prepare for SA-CPMCSCF. do not convert to in-core if direct and enough memory for in-core is available. 40000000 Do SA and prepare for SA Second Derivative Computation (terms involving 2nd order orbital rotation derivatives not included) IOp(5/18) L502: Mixing when doing damping: -3 -2 -1 0 N MO damping at all iterations. (100-N)/100 old density. Default (-1 unless reoptimizing during Stable=Opt). Force allocation for 1 or 2 buffer case conventional case (VV. SPAN=5.d-2).J)).J)). 10000000 Use State Average density matrices. IOp(5/19) Over-ride integral storage control (L501.SPAN ARE CONSIDERED TO BE ZERO 0 1 STHRS=1. Forbid in-core: force re-reading of integrals even if they fit in 2 buffers if conventional. Save generated integrals on disk (file 610). L506.10000 100000 1000000 Attempt to control root flipping in CI.IBuf2E).D-7 STHRS AND SPAN READ IN (2D20. Force Square in-memory storage. L503: CUTOFF CRITERIA IN SYMMETRY DETERMINATION OF M. Force computation of raff 1 and 2 integrals even for RHF.O.14) Damping (L506) Maximum rotation gradient for Newton-Raphson in L508 (above this value. Turn off damping.S. 2 if possible. SYMMETRY IS DETERMINED IF LARGEST OFF-DIAGONAL M.O.STHRS ELEMENTS Abs(F(I. Do not save integrals (same as 0x). FOCK-MATRIX ELEMENT Abs(F(I.GE. otherwise 1.D-4. 10**-N.LE. Dynamic selection of density damping based on band gap and DIIS error.

L504). (READ ONE CARD WITH START. Yes.001) Damp if error > 10**-N For MO damping: 0 N Default. IOp(5/21) DIIS error for density damping. CONTINUE RUN. 0 1 2 Default (no). 0 Default (1042) for calculations using diagonalization (2) for calculations using sparse diagonalization replacements. N Rotations are turned on when SQCDF is below 10**(-N). IRdNLp. L502. Maximum N/1000 virtual component. do a final unextrapolated diagonalization after convergence is reached. Extrapolation control in L506. Orbital rotation control (L506). maximum virtual mixing for MO damping: For density damping: 0 N Default (Damp if error > 0.T. 0 1 NO YES IOp(5/20) Final non-DIIS iteration (L501. 0 1 ABORT RUN VIA LNK1E. IOp(5/22) Use of DIIS extrapolation (L501. 77 . L502.PRINT F(1). no more than 1/3 virtual component for any occupied at each iteration. ACTION IF OTEST DETECTS PROBLEMS (L503). MCSCF flags: 2 10 Generate MOs using UHF natural orbitals. L504). No. just quit when extrapolated convergence is reached.END 2I2) (L503).

Yes. Regular DIIS Energy-based mixing Energy DIIS when DIIS error has increased significantly or is above threshold Energy DIIS when DIIS error has increased significantly. Yes. 3 78 3rd root of CAS(2. mixture of energy and Use energy DIIS when commutator gives huge coefficients. 0 1 2 Read from input stream. IOp(5/24) Orbital freezing (L506). Apply rotations sequentially. Read from chk. IOp(5/26) Type of calculation (L504). with Fermi broadening as well. high spin and first natural orbitals. Yes. Read from rwf. L509). weight of energy DIIS in method 4. 0 1 Default (exponentiate rotation angles). used (DIIS error/10^-N) for Use print level M in DIIS. IOp(5/25) Rotation application (L506). Orbital mixing control in L506. 0 1 Optimize all orbitals. otherwise. so that pair and hamiltonian information can be reused (L506.2) . Switch from energy to commutator when error is 10^(-N) in method 3. Yes. deciding on the fly between the two forms. Freeze all closed. with "FON" version of Fermi broadening. with "pFON" version of Fermi broadening.1 2 3 4 5 10 20 30 40 1xx Nxxx Mxxxx No. commutator. Optimize only 2nd and higher naturals. IOp(5/23) Flag for later points of an optimization.

If zero. if mulitplicity 1). IOp(5/30) Whether to symmetrize final orbitals using abelian symmetry operations (L502. GVB as CAS(2. IOp(5/27) Whether to do closed-shell calculation in L502. It will never use Raffenetti for SCF. 79 .2) GVB(1/2) Orthogonal open-shell singlet. closed for Multip=1. 0 1 2 Default (Yes. Integrals with degree of contraction greater than or equal to N are done are regular integrals. Force closed shell. If -1. -1 0 1 N All integrals done as Raffenetti. All integrals are done as regular integrals. an open-shell singlet is assumed. error if Multip>1. Yes.2). 0 1 Default (Yes). not needed in L506). IOp(5/28) L510: Root of CI to use in MCSCF (IState) 0 -1 N Defaults to Istate=1 Read IState from cards (see below) IState = N IOp(5/29) Use of rafinetti integrals during direct SCF. the unpaired orbitals are assumed to be high spin. Default: let FoFDir decide. Number of hamiltonians to read in (L506). Closed/Open control for L511: 0 1 2 Default.2 1 0 -1 -2 Excited singlet as 2nd root of CAS(2. Force UHF. ROHF Triplet (a debugging option). No Yes (used for RHF direct SCF). L505.

N. IOp(5/32) Sleazy SCF (L502. 80 . Do iterations with sleazy XC grid. SAME AS IOp(33)=3. doing an inexpensive pass 0 and then full accuracy in pass 1. then one iteration with next grid up.2 No. BUT ALL MATRIX TRANSACTIONS ARE PRINTED (BEWARE!!! MUCH OUTPUT EVEN ON SMALL MOLECULES. IOp(5/33) PRINT 0 1 2 3 4 IOp(33) PRINT OPTION. ONLY SUMMARY RESULTS ARE PRINTED (WITH POSSIBLE CONTROL FROM THE 'NO. THE EIGENVALUES AND THE M.PRINT' OPTION). Force use of general DBF logic. use loose integral cutoffs. SAME AS IOp(33)=1. REGULAR SYSTEM DEFAULTS APPLY HERE. Force use of 1c instead of general DBF logic. No longer used. SAME AS IOp(33)=2. Yes. The default is CoarseGrid for iterations and SG1 for final energy.) IOp(5/34) DUMP OPTION. 0=normal 1=Linear approximation to Xc. BUT ADDITIONALLY THE DENSITY MATRIX IS PRINTED. Use general DBF logic only if the DBF rwf is present. Foc formation. Decide between 1 and 4 based on details of the calculation. BUT AT THE END OF EACH ITERATION. O. IOp(5/31) How many vectors to form at a time during microiterations in L508 (NYI) and L509: 0 N Default (3 in L509). Yes. L510): 0 1 2 3 4 5 6 00 N00 I000 00000 10000 20000 Default (No). convergence on either energy or density and always do incremental No. COEFFICIENTS ARE PRINTED AT THE END OF THE SCF. No longer used. Thresholds similar to DGauss for convergence and integrals. Use approximation I.

do integrals at same accuracy as convergence until final iteration. VarAcc forbidded because of guess=read. VarAcc allowed. do integrals 3 digits more accurately than current convergence. L502).IOp(5/35) Whether basis is orthonormal (L501. Yes. IOp(5/36) Whether to checkpoint after every SCF cycle. 0 1 2 Default (checkpoint only if direct). No. accurately. IOp(5/39) New On-Fly symbolic matrix element generator. IOp(5/38) Whether to vary integral cutoffs during direct SCF: 0 1 2 3 4 5 6 Default (same as 1). No. -1 0 N Do not do incremental Fock formation. then 2 digits more Converge to 10**-5 with integrals good to 10**-6 first. Yes. decide based on details of problems. then full convergence. Yes. Don't checkpoint. Checkpoint. Every Nth cycle. IOp(5/37) Frequency at which to do full Fock formation instead of incremental (L502). 0 1 2 Default (No). must be set. 1 2 3 Hartree-Waller functions for singlets Hartree-Waller functions for triplets Slater Determinants 81 . allows different default actions for PBC. Default (every 20 for direct). REQUIRE 'NOFULLDIAG' Remember: the first digit indicating the type of function to be used.

xx0 integral yy000 100000 200000 300000 1000000 2000000 Use cutoff = 10**(-xx) on integral value to exclude contributions. IOp(5/41) Whether to converge on maximum density change as well or instead of RMS: 0 N -1 -2 N0 in L510. Maximum allowed change is 10**N larger than RMS. L510: xx yy00 zz0000 k000000 Davidson options. WARNING !!! ratio (zz+k-1)/k must be equal to n. increment field in Gen(2-4) No. Converge only on RMS density change. Multipoles of order N. Default=60 Number of vectors provided in input BEWARE !!! Davidson will execute zz updating per iteration. Update surface and restart DIIS when within 10**-2 of convergence.50) Maximum dimension of iterative subspace. only used now for Onsager and control of details of SCIPCM -N 0 N 00000 10000 20000 30000 40000 50000 Multipoles of order N..e. Option xx is used also by Lanczos if IOp(39)=10000n or 20000n Maximum dimension of reduced Hamiltonian used as guess Default=Min(NSec. Update surface every iteration in pass 1 only. Update surface on pass 2 iterations only. Default is DON'T EXCLUDE any integral Use sum of the first IState roots of a Reduced Hamiltonian as guess for Lanczos Use IState-th root of a Reduced Hamiltonian as guess for Lanczos Save first IState roots on disk for Davidson (this option will automatically call Davidson instead Print S**2 Print S**2 and its orbital components of Lanczos) IOp(5/40) Use of reaction field. Maximum allowed changed is same as RMS (i. Converge on energy to 10**(N)*RMS-density-accuracy Also control of iterative diagonalization . value must be 0<=ll<=20 Default=0 82 22. store field in Gen(2-4) Default (same as 10000). Default is DON'T EXCLUDE any Use cutoff = 10**(-yy) on the product Integr*DenMat. Default=1 (no reduction) ll0000000 Davidson iteration at which to scale back the number of vectors WARNING !!! For overflow reasons. Same as 3. convergence only on maximum). Update surface every iteration. Default=IVEC Reduction factor between number of guess vectors provided and number of vectors wanted at the end (1<=k<=9). but re-use 1e matrix instead of surface terms. number specified in nroot=n.

IOp(5/42) Number of orbitals to localize in L510 1 n Localize all active orbitals Localize first n (strongly occupied!) orbitals IOp(5/43) L509: Whether 5th order terms are treated explicitly 0 1 2 Default: set to 1 All 5th order terms are treated implicitly (Debug option) 5th order GG and WG terms are explicitly computed in L715 L510: DFT corrections to MCSCF on last iteration 0 1 2 No Yes.7: the displacement in geometry in internals in Angs.8 f10. 1 2 Prepare data for Mp2 (l906 obsolete) Compute transition spin density and SO coupling IOp(5/48) Options to be passed to CalDFT: 83 . Uses MC-SCF density to compute B88 + LYP energy (These are hard-wired since they were the only choices that gave sensible results) Replaces diagonal elements of MC-SCF CI with B88 + LYP energy IOp(5/45) Numerical Derivative Coupling calculation(for testing) 0 1 No Yes (Needs NonStd root and two cards in input stream): i3 the other vector which coupled with iVec. IOp(5/47) In L510. 1 to set up for CAS-MP2 or 2 to do spin-orbit calculation. If negative reads the vector from rwf. 10 Include the CSF contribution to the orbs for the DerCpl. If positive reads vector from input 4f20.

L510: Option for using reorthogonalization procedure in Lanczos 0 1 No Yes IOp(5/49) Use of sparse storage and Conjugate Gradient optimization instead of N**2 memory and diagonalization. 0 1 Yes canonicalize. Square storage (only in Fock formation if CG). L510: Option for using lanczos in CPMCSCF calculations 0 1 2 No Yes Use lanczos except for the last iteration IOp(5/50) L510: Option for setting the maximum number of lanczos iterations in Direct CPMCSCF IOp(5/51) L510: Canonicalize MC-SCF orbitals by diagonalization of Core and Virt Fock operators. L510: configuration cutoff for mp2 0 i 84 . -1 0 N None Default. or 22 if sparse is set in ILSW).N Control flag for CalDFT is N. Linear storage (only in Fock formation if diagonalization). also none.) IOp(5/52) Amount of memory to allocate to stashing integrals. N words.1 Float(1/i) . 0 1 2 10 20 Default (11. Speeds up convergence in CP-MCSCF Do not Canonicalize (turn this on to maintain compatability with previous versions of code. Diagonalization Conjugate gradient.

075 When set.5). in link 1003 a normal frequency calculation is performed when the optimization is in 85 . L510: See below IOp(5/57) First iteration at which to level shift and do FON.1 unless doing stable=opt. N N/1000 Bohr.001 . which is usefull when a trajectory is started at a degenerate region. L510: see below IOp(5/56) A0 for Onsager SCRF.) When IOp(55).lt.5-0. N of each. The state averaging will be switched off (graduately) when the degenerate region is left. Default is 0. default parameters Dielectric constant |N|/1000 IOp(5/55) How many HOMOs and LUMOs to solve for after CG: 0 N None.050 Threshhold for switching off.5). (Or reversed. L510: See below IOp(5/55-58) L510: Switching on a State Averaged calculation graduately. Usefull for optimizations or trajectory calculations where only a part of the surface is (nearly) degenerate.0-1. IOp(56) IOp(57) IOp(58) Threshhold for switching on. the number of steps for switching on/off is -IOp(55). Default is 0. Energy difference larger than IOp(57)*0. Energy difference smaller than IOp(56)*0. IOp(55) Number of steps over which the SA coefficients are brought from (0.0) to (0. In that case.IOp(5/53) PCM input and solvent type.0.5-0. then start after instability searches.001 . the calculation is started with coefficients (0. 0 Default . when the SA calculation graduately is switch off. N>0 N<0 Solvent type N.

0001 (checks Offdiagonal element).5/0. use N. IRanWt. 0 1 Use global default.8 0.1/0.7 0.Threshold for a determining an adiabatic hop threshold=n*0. instead of a State Averaged second derivatives calculation. only used with IOp(97)=11 IOp(5/60-62) Override standard values of IRadAn. Yes. IOp(5/63) Whether to do FMM.a region of (0.0).8 0. for a hypothetical example with nroot=2: cycle 4 5 6 7 8 old SA 0.2/0.6 0.2/0.2/0.7 0. 86 . and IRanGd. this can cause jobs sensitive to SA to fail to converge The new method checks the gap each cycle and will decide to increase or decrease the SA based on that result.0-1.4/0. 1: 2: Normal computation when (0.1 Egap <limit <limit >limit >limit >limit Note: trajectory option set in overlay 10: IOp(10/99)=n -. regardless of weather that continues to decrease after the switch point.9 0.3/0.0) (default) Allways SA second derivatives THESE OPTIONS MUST BE SET IDENTICAL FOR OVERLAY 10! (only IOp(10/55) needs to be set in overlay 10) IOp(5/59) IOp(59) 0 1 (Default) No action Turn on new more flexible SA options: The old method switches SA on when you reach certain energy gap.5 crash new SA 0.0/0. Turn off FMM here regardless.3/0.0-1. IOp(5/64) Over-ride default value of FMFlags 0 N No.8 0.

-2 -1 0 N None Start at a high temperature (limited only by DIIS error). Yes.10 steps FON / 20 steps PFON N steps IOp(5/73) Options for ADMP: 0 1 2 Default (2) Use Lowdin basis for CP orthonormal transform. 0 N No. IOp(5/70) Maximum Initial temperature for FON (non-PBC). Default (3000K = 10 milliHartrees for non-PBC.IOp(5/65) Over-ride NFx parameter. IOp(5/74) Type of k-point integration: 0 1 2 3 9 10 20 90 100 200 Default (911). or temperature for broadening (PBC and IOp(74)=[1-4]xx). Occupy lowest NE at each k point regardless of the energies Improved LT with quadratic corrections Original LT method No concern for corrections Smearing Marzari method I Smearing Marzari method II 87 . temperature broadening for metals. Use LT method (interpolation) Occupy entire points (used together with broadening) Full points for insulators. 6000K for PBC) N degrees IOp(5/71) Number of steps to apply simulated annealing (L502): 0 N Default -. use N. Use Cholesky basis for CP orthonormal transform.

IOp(5/82) C. Convergence criterion 0 N 88 Defaults to 10**(-7) 10**(-N) . alpha orbitals. 0 Default. IOp(5/80) The maximum conjugate gradient step size -1 0 MMNN No maximum step size Default maximum (. IOp(5/79) Range around Fermi level where temperature distribution will be applied if broadening is turned on for PBC.8) Step size of MM. No preconditioning.NN IOp(5/81) Conjugate-Gradient Parameters MM NN00 00000 10000 000000 100000 0000000 1000000 Maximum Number of CG cycles per SCF iteration. (defaults to 3 cycles). Maximum Number of purification cycles per CG iteration. Don't use CG DIIS Use CG DIIS Polak-Ribiere CG minimization Fletcher-Reeves CG minimization Use diagonal preconditioning in Conjugate-Gradient.G. (defaults to 4 CG cycles).300 400 500 900 First order Hermite-Gaussian of Paxton and Methfessel Gaussian smearing Classical Fermi-Dirac broadening No broadening (this will be Gaussian broadening with small T) IOp(5/75-78) Number of alpha electrons. and beta orbitals for fractional occupation. beta electrons. a value will be chosen in ZInLT1.

Limit is N cycles.1). IOp(5/85) Over-riding of maximum cycles for XQC. Level shift only if the HOMO-LUMO gap is zero or insignificant (>-0. 89 . up to twice the goal No longer used. -1 0 N Default for first step (128). Level shift as much as necessary for HOMO>LUMO. Reduce DIIS space whenever energy is above the lowest energy. Turn off level shift after energy rises. Just continue as usual if energy goes up. 0 1 No. Yes. Reduce DIIS space when energy rises from previous cycle. Turn off 3 and 4 point extrapolation if DIIS is on. Level shift to a maximum of the Goal. Reduce DIIS space when energy goes above the lowest energy. Level shift only if the HOMO-LUMO gap is zero. Turn on dynamic level shift from the beginning Turn on dynamic level shift only after FON is over. IOp(5/86) Strategy options 000000 0 1 2 3 4 10 20 100 200 1000 2000 3000 4000 5000 6000 N0000 100000 Default (101100).1) Level shift only if the HOMO-LUMO gap is zero or insignificant (>-0. No. Restart using Fock matrix in L502. Level shift to a maximum of 2*Goal. Keep level shift after energy rises. then 4).IOp(5/83) Maximum SCF DIIS vectors 0 N Default (20). Use SCF DIIS with N vectors IOp(5/84) Restart in L509. Default (1 except during Stable=Opt.

Typically 10^-10 for molecules and 10^-12 for periodic systems. IOp(5/90) Whether to generate sparse guess here. 1 2 Yes. do preliminary AM1 calculation. set in FoFDir/FoFCou/CalDSu based on accuracy part of IOp(5). do preliminary AM1 calculation and compare with guess from previous step in geometry optimization. if DIIS is on. IOp(5/88) L510: controls the amount of printing. variable=kPrint 0 1 2 3 Print only summary information Print the a(t) vector and probability for each csf Print almost everything for debugging Print everything for debugging warning! this is a lot of stuff and you will only be able to do a few cycles IOp(5/89) Linearly dependent basis control for PBC. separate Coulomb and NFx exchange for PBC). IOp(5/91) Control option for chebyshev sparse control. The energy is only checked after FON has been turned IOp(5/87) Accuracy criterion in Fock matrix formation: 0 N Default. . IOp(5/92) Whether to use FoFDir or FoFCou for exact exchange: 0 90 Default: normal processing based on FMM for non-PBC. 10**-N. Yes. Retain 3 and 4 point extrap. this and ZFormV should be moved to L302.200000 off.

N iterations.125 in GS How long to wait before checking for a hop after going through a hop. variable=IWait. threshold for a hop up (after a hop down) = 0.01) or the lower state (=(1-x)*0. which determines what type of gradient calculation to do: xxN Choose the basis variable=mBasis 0 1 2 3 xNx (Default) same as 1 Use a(t) basis Use mcscf basis orthogonal to a(1) and a(2) Use currnet mcscf vectors to check the code must be used with mHTest=2 91 Do with diagonalisationof Ecc variable=mTDGrd: . threshold =0. where: Zyy yy Z xxx Threshold for a hop down determined by probability of being on the upper state (=x*0. FoFCou for Coulomb.25 in GS) and if Zyy=125. value=xxx.01) Determines a variable Ulimt which lies between 01 and 99 Determines if this probability is halved after the first hop default Zyy=25. separate FoFCou/NFx for exchange.1 fempto seconds If IOp(97)=23. L510: has different meaning depending on if you are using IOp(97)=22 or IOp(97=23) If IOp(97)=22. L510: flag hopping controls starting and stoping options (x=0 or x=1) variable=iBack for first hop being up: xxxx0 xxxx1 variable=iStNow xxx0x xxx1x variable=iSpace xx0xx xx1xx variable=iEnd x0xxx xNxxx 0xxxx Nxxxx Use full space CI basis Use reduced space in projection of alpha number of cycles to carry on before stopping the timedep code after exiting the hopping region Default stop 6 iterations later Stop time dep on cycle N after exit (if IOp(92) is negative then stop immediately) No effect Stop time dep on cycle N allowable values 1-9 Use Energy gap criteria to start timedep Start timedep imediately Hopping down (forward) Hopping up (backward) variable=iFcTD to stop the time dependent code on cyle Z IOp(5/93) Number of initial iterations for which damping is allowed: 0 N Default (10).1 2 FoFCou for Coulomb. wait time is xxx*0.75 in ES (or 0. Value = xxxzyy. separate FoFDir for exchange.

0 1 2 N00 0 1 2 information (Default) same as 1 Only check Ecc if MCSCF energies are almost degenerate Force check on Ecc by diagonalising it (Default) no testing Testing construction of ecc Testing construction of the ecc or exx portion of hessian using the mcscf vector 3 Calculate but do not use TD gradient Testing options. and for pass 0 grid in L502. The CAS active space is subdivided into three RAS active subspaces. Use grid N. Ras3 orbitals are empty. default xxx=4 ie deltaE<0. 0 N Use default. and the maximum number of electrons in the Ras3 space: zzyyxxww. variable=mHTest IOp(5/94) PCM/ONIOM calculation 0 1 2 Standard PCM calculation PCM/ONIOM calcn.05 switch timdep off only used with IOp(97)=22 and IOp(97)=23 IOp(5/95) Option for using Davidson in RFO calculations 0 1 Yes No use Lanczos (not recommended) IOp(5/96) Over-ride IRadAn for CPHF-like step in L509. Ras2 and Ras3. on the real system PCM/ONIOM calcn. on the model system L510: Threshold for turning propagation method on and off yyy xxx The first three digits determine the energy gap for turning off The last three digits determine the energy gap for turning on Threshold =xxx*0. We also need to define the maximum number of holes in the Ras1 space (ie the number of electrons that can be excited out of the Ras1 subspace. Ras1 orbitals are doubly occupied. default yyy=5 ie deltaE>0.04 switch timdep on.01 (checks energy gap). In the reference space. where ww 92 Number of Ras1 orbitals . Ras1. RAS control in L510.

No. 22 23 Propagation of the wavefunction. 0 1 2 10 11 21 Default (Yes. Hopping threshold during trajectories with L510. Needs option 80 and option 99. Vector following or Root following hop alone (Needs option 80) Make a hop based on the secular equation (adiabatic hop) this option also includes a hop decision Debugging Option: Propagation of the wavefunction is switched on. but the hop is determined by based on the vector following method (diabatic hop) (Needs option 99 and 80) the diabatic or adiabatic criteria (whichever determines a hop first). for detailed control of the propagation conditions see options 92-96. hop is based on the probability of being in a specific state Propagation of the wavefunction but no hopping. the molecule continues in a mixed state (Needs (Needs also option 80) (detailed control is determined by options 92-96) option 71 and 80) IOp(5/98) Whether to save eigenvalues and orbitals at all k-points. No.2. 0 Default: 32.xx yy zz Maximum number of holes in Ras1 Number of Ras3 orbitals Maximum number of electrons in Ras3 IOp(5/97) Whether to update precomputed grid data with timing information. No. Yes.12. 93 . IOp(5/99) Grid for numerical k-integration in FT-LT method. Yes. Yes.8 for 1.3d IOp(5/100) Tight convergence during CGDMS: 0 1 2 Default (No). 0 1 2 Default (No). if available).

L510 Notes These options must be set in multiple links: L1003 iop(97) iop(55-59) iop(80) yes yes yes yes yes no no no yes L510 L118 These options must be set for the following links: IOp l118 l510 l1003 55-58 no yes yes 80 yes no yes 97 no yes yes 98 no no yes 99 no no yes 94 .IOp(5/101) SDif test on numerical accuracy of PBC diagonalization. 0 -1 N>0 Default (10) No test Abort of SDif is larger than N. N. IOp(5/102) Maximum number of configurations for CAS-MP2: 0 N Default (1000).

95 .12). IF MORE THAN ONE OF THESE IS REQUESTED. Read in a set of cards specifying a grid of points at which the electric potential will be computed.number of increments and vector.Y-coord.X2.Z-coord. The potential (NEFG=3). Only one card is needed: N. N1 records will be written to unit KTape.XY-EFG. Three cards are required: KTape.NEFG.ZO N1. THE LISTS ARE IN SEPARATE INPUT SECTIONS IN THE ORDER LISTED BELOW. EVALUATE THE PROPERTIES ONLY AT EACH ATOMIC CENTER.Z1 N2. Y AND Z COORDINATES IN ANGSTROMS (FREE FORMAT).X1. Read in centers at which to evaluate the potential from the rwf.output unit and coordinates of one corner of grid.Y-field. STARTING AT EACH POINT.Z-field. If KTape is 0. Two forms of specifications are allowed: A.Z2 -. -. it defaults to 51.ZZ-EFG.LTape. READ ADDITIONAL CENTERS. or potential. B.Potential X-field. field. Thus if NEFG=3 for each point there will be 4 cards written per point.number of increments and vector.KTape The coordinates of N points in Angstroms will be read unit LTape in format (3F20. 0 1 2 4 NO ADDITIONAL CENTERS. Read grid. ONE CARD PER CENTER WITH THE X.XO. LOCATED THE NEAREST STATIONARY POINT IN THE ELECTRIC. with N2 values in each record. and field gradient (NEFG=1) will be computed and written along with the coordinates to unit KTape in format (4F20.12). Evenly spaced rectangular grid. Constrain the dipole in fitting charges.XZ-EFG YZ-EFG Note that either form of grid should be specified with respect to the standard orientation of the molecule. READ IN COORDINATES AS FOR 1.Y1. potential and field (NEFG=2).XX-EFG YY-EFG.Overlay 6 IOp(6/15) SPECIFICATION OF ADDITIONAL CENTERS.Y2. 8 16 32 128 Do potential-derived charges. containing: X-coord. An arbitrary list of points. -.YO. POTENTIAL. do not default cube.

Read in replacement radii for selected atom types as pairs (IAn. 0 1 Use full accuracy in calculations at specific points. Exact is still in standard orientation. COMPUTE ONLY THE CONTRIBUTION OF SHELL N.Rad) or (Symbol. but use sleazy cutoffs in mapping a grid Do all points to full accuracy. COMPUTE ONLY THE ELECTRONIC CONTRIBUTION. one per line. IOp(6/20) How to do electrostatic-potential derived charges: 0 -1 1 2 3 00 10 20 30 100 96 Default (1). Merz-Kollman point selection CHELP point selection. so rotate.Rad). 0 1 2 -N COMPUTE ALL CONTRIBUTIONS TO SELECTED PROPERTIES. IOp(6/18) Whether to update dipole rwf 0/1 yes/no. Default radii are those defined with the selected method. Read a list of points at which to fit. Force CHELP (Francl) recommended radii. same as 1. IOp(6/19) Whether to rotate exact polarizability before comparing with approximate (which will be calculated in the standard orientation). 0 1 2 Default. CHELPG point selection. This is like IOp(9) in L9999. Force CHELPG (Breneman) recommended radii. COMPUTE ONLY THE NUCLEAR CONTRIBUTION. use as-is. of points. IOp(6/17) DEBUGGING CONTROL (L602). Exact is already in z-matrix orientation. Force Merz-Kollman radii.IOp(6/16) Cutoffs in L602. terminated .

or the HF density if the one for the current method Use the density matrix for method N (see Link 1 for the numbering scheme).by a blank line. Fit united atoms (heavy atoms only) rather than all atoms. IOp(6/23) Density values to evaluate over grid in L604: 0 1 2 3 Default (same as 3).chk file. Use only active atoms in the fit. Yes. L602. Optimize density basis set. 97 . is not available. L604): -1x +1x -5 -4 -3 -2 -1 N. IOp(6/21) Operation of L603: 0 1 2 Default (same as 2). IOp(6/22) Selection of density matrix (currently only in L601.0 Read density matrices from . Density values and gradients. Density for the excited state given by IOp(29). Default (Yes). Read density matrices from . Read in density basis functions and compute populations. All available transition terminated by a blank line. IOp(6/24) Frozen core: -N 0 1 2 Freeze N orbitals. Density values. gradients and divergence. No. 200 1000 10000 Read in replacment radii for selected atoms as pairs (I.chk file.Rad). Use the density matrix for the current method. Use all available density matrices. Density values. Transition density between the states given by IOp(29) and IOp(30).

Total. Do bonding Mulliken Populations. O(N**4)). Don't do bonding Mulliken Populations. Don't do Mulliken populations. IOp(6/26) Which density to use in L602 and L604: 0 1 2 3 4 Default (same as 1). quantum mechanically (no self terms . classically (including self terms . but do not mark. IOp(6/27) Choice of population analysis: 0 1 2 10 20 100 Default (12).IOp(6/25) Whether to compute coulomb self-energy in L601: 0 1 2 O(N**5)). and only available for HF. Beta. Spin. No. Yes: mark as well. IOp(6/28) Mark SCF density as current density. IOp(6/29) Excited state to use if requested by IOp(22). Yes. Do Minimal population analysis. Yes. 98 .requires 2e integrals.requires 2e integrals. Do Mulliken populations. 0 1 No: save SCF density. Alpha.

Compute kinetic energies and multipole moments of AIMs (IDoPrp). using alpha and beta separately for UHF. L606: naming of RPAC interface file. 10. Store only alpha NOs.01*IDcInt are interpreted as lone pair MOs rather than bond MOs (1.9. Yes. using total density. where MItLoc ITlLoc IDcInt IPrLoc MItLoc*NOrb*(NOrb-1)/2 is the maximum number of iterations in localization of (spin)orbitals (1. Store only beta NOs. Localized (spin)orbitals with atomic occupancies less than 0. IOp(6/31) Whether to determine natural orbitals from densities: 0 1 2 3 4 5 No.. Compute atomic overlap matrices (IDoAOM). Compute charges of AIMs (IDoAtC). This precludes calculations of atomic property derivatives with respect to nuclear displacements. L605.IOp(6/30) 2nd excited state for transition density: 0 N Transition denstiy between state IOp(29) and g. Name this file 'rpac. attractor interaction lines. default 9). ring points. default 6).. and cage points.. Use spin density.. 99 . 0 1 Make this a scratch file. Yes.11' IOp(6/35) WHAT TO DO: 0 1 2 4 10 100 Determine attractors. Compute energies of electrostatic interactions between AIMs (IDoPot). IOp(6/32) CONTROL PARAMETERS FOR COVBON in L609 (NOT TO BE CHANGED UNDER MOST CIRCUMSTANCES): 10000*MItLoc+1000*ITlLoc+100*IDcInt+IPrLoc.**(-ITlLoc) is the convergence criterion for (spin)orbital localization (1. Determine zero-flux surfaces (IDoZrF). Transition denstiy between state IOp(29) and state N.s. 1: Do not print the atomic occupancies..9.. default 10). 0: Print the atomic occupancies of localized (spin)orbitals (default).99.

Length of Fourier expansion for the trial path (1..999. and IEpsIn=100. IOp(6/38) CONTROL PARAMETERS FOR ZRFLUX AND OIGAPI in L609 (NOT TO BE CHANGED UNDER MOST CIRCUMSTANCES): 100 . IOp(6/36) CONTROL PARAMETERS FOR NEGLECT OF ORBITALS AND PRIMITIVES in L609: 10000*INoZer+100*IPrNDe+IPrNAt. If necessary. zero occupancies.. default 2).. IPrNAt=5. 1: Do not ignore (spin)orbitals with Neglect primitive contributions below 10. (IHwAug=1). default 10). This option sets IPrNDe=5..**(-IPrNAt) in integrations over atomic basins (099. IOp(6/37) CONTROL PARAMETERS FOR ATINLI. Neglect primitive contributions below 10.99. If necessary.. Compute atomic orbitals in molecule (IDoAOs).200 400 1000 Compute other atomic matrix elements (IDoAMa).. augment valence electron densities with relativistic core contributions. Use numerical instead of analtyic integration. default 100).99. default 7). augment valence electron densities with nonrelativistic core contributions Abort if pseudopotentials have been used (IHwAug=3). where INoZer IPrNDe IPrNAt 0: Ignore (spin)orbitals with zero occupancies (default).. Include zero-flux surface relaxation terms in all atomic matrix elements (IDoSRe) Compute derivatives of atomic properties with respect to electric field (IDoSeP). 99. which is a default anyway (IHwAug=0).9. No other properties can be calculated. Note that IDoSRe should be set to 1 in order to obtain correct results! Also note that analytical polarizabilities have to be available but force constants have to be absent! 2000 10000 20000 100000 200000 400000 1000000 2000000 3000000 Compute derivatives of atomic properties with respect to nuclear displacements as well (IDoNuD). default 20). Number of grid points in critical point search (1..**(-IPrNDe) in evaluations of electron density and its derivatives (0 default 7). Use numerical instead of analtyic integration and use reduced cutoffs. Maximum value of the control sum (1. Reduce accuracy so atomic charges can be computed more rapidly (IQuick). AND CAGPNT in L609 (NOT TO BE CHANGED UNDER MOST CIRCUMSTANCES): 1000000*MxBpIt+100000*SBpMax+1000*NGrd+LookUp. Note that analytical force constants have to be available! Compute localized orbitals and bond orders (IDoLoc). RNGPNT. where MxBpIt SBpMax NGrd LookUp Maximum number of iterations in trial path determination (1.

. default 6).9.9..99.9. default 5). must be >=4. IOp(6/41) Number of layers in esp charge fit. 0 1 2 Default NBO analysis . N layers.. IOp(6/40) Control of link 607...*IPraIn is the cut-off for zero-flux surfaces (1. whose energy can then be computed by one of the SCF links. This link must have been invoked with IOp(40) = 0 or 1 prior to invoking it with IOp(40)=2...9.9.**(-IGueDi) is the initial displacement from the critical point in the Runge-Kutta integrations (1. 0 N Default (4). default 6)..0001*IEpsIn is the target for integration error (1... default 2). 0.. IRScal is the scaling factor in the nonlinear transformation used in the intersection search 10.. Read input data to control NBO analysis.100000*INStRK+10000*IHowFa+1000*IGueDi+100*IPraIn+10*IRScal+IRtFSe INStRK IHowFa IGueDi IPraIn IRScal IRtFSe 10*INStRK is the number of steps in the Runge-Kutta integrations along gradient paths (1.9.*IRtFSe is the safety factor used in the intersection search (1..9.. default 2).. Delete selected elements of NBO Fock matrix and form a new density.9... 5+INInCh is the initial number of sampling points in the intersection search (1.. default 5). default 2).... default 2). 10. default (1. 101 . IHowFa is the maximum distance in the Runge-Kutta integrations along gradient paths (1.9.**(-5-IToler) is the tolerance for the intersection search (1.. 10*INInGr is the initial number of grid points in theta and phi in the adaptive integration subroutine (1. IOp(6/39) More CONTROL PARAMETERS FOR ZRFLUX AND OIGAPI in L609 (NOT TO BE CHANGED UNDER MOST CIRCUMSTANCES): 1000000*IToler+100000*INInGr+10000*INInCh+1000*IEpsSf+10*IEpsIn+INTrig IToler INInGr INInCh 2). default 2). 3 Read the deletion energy produced by a previous run with IOp(40)=2 and print it. 10*INTrig is the number of sine and cosine functions in the trial function for surface sheets 10. (1.. IEpsSf IEpsIn INTrig IEpsSf is the safety factor used for patches with surface faults in the adaptive integration subroutine (1. default 2).9.don't read input. 10..9. default 2).

0 N Default (1). IOp(6/46) Threshold for molecular volume integration.4/Sqrt(#layers)) 0. 102 . 50 is recommended. points per unit area.IOp(6/42) Density of points per unit area in esp fit. IOp(6/45) Number of points per bohr**3 for Monte-Carlo calaulation of molar volume -1 0 N Read from input Default (20) N points . 0 N Default (0. N*10**-4.for tight accuracy. IOp(6/44) Type of calculation in L604: 0 1 2 3 10 Default. 0 -1 N Default . IOp(6/43) Increment between layers in MK charge fit.10**-3 Read from input. same as 2.01*N. Compute the molar volume Evaluate the density over a cube of points Evaluate MO's over a cube of points Skip header information in cube file.

Yes. IOp(6/52) Number of radial integration points in L609: 0 N Default (100). Fine grid. all possible IPs and EAs. Yes. on non-SCF densities. N.for debugging. IOp(6/48) Use of cutoffs 0 N Default (10**-6 accuracy for cubes. Coarse grid. 3 points/Bohr. 103 . No. 12 points/Bohr. on non-SCF densities. IOp(6/51) Whether to apply Extended Koopman's Theorem (EKT): 0 N -1 -2 Default (No). 6 points/Bohr. N*0. 10**-N IOp(6/49) Approximate number of points per side in cube in l602/l604: 0 N -1 -2 -3 -4 Default (80) N points Read from cards.01 . -N>4 Grid using 1000 / N points/Bohr.IOp(6/47) Scale factor to apply to van der Waals radii for the box size during volume integration: 0 N Default. up to N IPs and EAs. 1 digit better than desired acuracy for volumes). Medium grid.

IOp(6/54) Maximum number of domains. 0 Default (99001). Yes. 0 N Default (100000). IOp(6/55) Number of inner angular points in numerical integration in L609: -1 0 N 0 (no inner sphere) 302 N point Lebedev grid (see AngQad). Must be an unpruned grid. IOp(6/57) Whether to generate data over a grid using the total SCF density: 0 1 2 3 No. Yes. 104 . Yes. IOp(6/56) Whether to read in density matrix from input stream in L608. N. also read in name for input file with a different grid and compare.IOp(6/53) Distribution of radial points in L609: 0 N Default (cubic) Polynomial of order N. 0 1 No. read in name for output file. IOp(6/58) Grid to use in generating tables of density and potential. Output in the form of data statements.

IOp(6/59) Approximations to Exc -1 0 1 2 3 Test superposition of atomic densities using L608: Do correct energies. Yes. Default (1 pmm). of all tensor components. IRanWt. Do test as usual. Yes. IOp(6/63) Suppress number of electrons test in XC quadrature in L608 (for debugging with small grids): 0 1 2 Default (do test). Do correct energies and 0th order approximation Do correct energies and 0th-1st order approximations Do correct energies and 0th-2nd order approximations IOp(6/60-62) Over-ride standard values of IRadAn. and IRanGd. IOp(6/65) Threshold for printing of NCS contributions. N/1000 ppm IOp(6/70) Control of L610. Suppress test. of isotropic value. -1 0 N Zero. IOp(6/64) Natural Chemical Shielding Analysis: 0 1 2 3 No. 105 . of diagonal tensor elements and isotropic value. Yes.

reading isotopes IOp(6/73) Whether to save orbitals from NBO: 0 1 2 3 10 Default (No). Save NBOs in place of regular MOs. Yes. IOp(6/76) Threshold for linear dependence in L607.D-6). IOp(6/72) Whether to read isotopes for hyperfine interractions and do hyperfine terms in L602: 0 1 2 3 4 Default (1). Yes. 10**(-N). IOp(6/75) model for CM2 charges. 0 N 106 Default (1.IOp(6/71) XC functional in L610. Yes. IOp(6/74) Whether to use Gaussian connectivity in choosing Lewis structure for NBO. Suppress re-orthogonalization. Save NLMOs in place of regular MOs. Use. regardless of other terms. Save NLMO occupieds and NBO virtuals. 0 1 2 Default (use if present and choose is selected in NBO input). NMR data is available. and other terms are being computed No. Don't use. if open-shell. .

Yes. No. IOp(6/81) Print kinetic energy of orbitals? 0 1 2 3 Default (yes. if doing other orbital results). 107 .IOp(6/77) Restraint in charge fitting in L602: 0 -1 N None. 0 1 2 Default (2). Yes. IOp(6/79) Whether to calculate Hirshfeld charges.d-4 N * 10^-5. 0 1 2 Default (No). 0 1 2 Default (No). for the top 5 occupieds and lowest 5 virtuals. for all orbitals. Yes. Use density. Yes. IOp(6/78) Use MOs instead of density in AtmTab. IOp(6/80) Whether to calculate Lowdin charges and Mayer bond orders. IOp(6/82) Tensors for hyperfine spectra. 2. Use MOs. No. No.

0 1 tensors. 2 Default. 108 . compute if there are 100 or fewer atoms Compute QEq tensors and for open-shell systems compute isotropic and anisotropic splitting Do not compute tensors.

only harmonic thermochemistry.91844) If M=1. If M=3. Print higher precision normal modes. 1/1.Overlay 7 IOp(7/6) operation of link 705 (NYI). 0/1 yes/no. with most common isotopes. Yes. Do not the ecp contribution to the SCF forces. 0 1 2 10 20 Default (12). 109 . If M=2. No. Do not form 1e derivative matrices. IOp(7/7) USE OF INTERNAL COORDINATES.12. IOp(7/9) Whether to rotate derivatives back to the z-matrix orientation. Increment the 1e derivative matrices with ecp terms. Yes. Print normal mode displacements in redunant internals.91671. do hindered rotor analysis. Form the ecp contribution to the SCF forces. but neglect first derivatives in conversion of second derivatives to internal coordinates. Read hindered rotor parameters from input. IOp(7/8) Harmonic frequency calculation: 0 1 2 3 10 20 30 Nxx Mxxx No. 2=CBS4=0. Default scale factor is #N (1=HF. with read-in isotopes. 3=CBSQ=0. Print both HP modes and internal displacements. 0 1 2 YES NO Yes.

old vector Rys DF. Rys f. Use generalized density number N for both the one-electron integral derivatives and the corresponding 2PDM terms. FoFDir Rys df. Use external 2PDM. PRISM df. Berny SP. Use both HF and external 2PDM. Berny SP. FoFDir HGP df. FoFDir: HGP spdf. do both. Old vector Rys SPDF. Rys df. Generate 2PDM from CIS square 1PDM (for debugging) . FoFDir: HGP spd.IOp(7/10) First/second derivative control. 0 N Usual SCF density. FoFDir: Berny SP. IOp(7/11) control of integral derivative algorithm: 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Default use IsAlg to decide. Berny SP. do only second derivatives. Berny SP. FoFDir: Rys spdf. Use HF 2PDM. Berny SP. FoFDir: PRISM spdf. Scalar Rys SPDF. FoFDir: HGP sp. FoFDir HGP d Rys f. IOp(7/13) Contraction with two-particle density matrices: 0 1 2 3 4 110 Default (same as 1). FoFDir Rys df (same as 7). IOp(7/12) Selection of density matrix. Berny SP. Scalar Rys DF. 0 1 2 do only first derivatives.

Defaults to 1. Form derivative 2PDM from CIS and HF derivative density matrices. The types of derivatives are Leave the external 2PDM on the disk instead of deleting it. DO AS ACCURATELY AS POSSIBLE in L708. 0 IJK 1000 Default (1st order nuclear and electric field). Magnetic Field Ith order. Contract with external 2PDM derivatives. while 6-7 imply use of the generalized density derivatives in L701. 10. Use sleazier cutoffs in L708. IOp(7/14) State for CIS gradients. The types of derivatives are given by IOp(15). and 11 are valid in overlay 7. DO AS ACCURATELY AS POSSIBLE in L702. Only 1. Nuclear Kth order. DO AS ACCURATELY AS POSSIBLE in L703. Normal (6 or 5 for linear molecules). IOp(7/16) Number of translations and rotations to remove during redundant coordinate transformations: -2 -1 0 N 0. IOp(7/15) The nature of the perturbation(s). N.5 6 7 10 Generate 2PDM from CIS square 1PDM and use HF/Z 2PDM as well. given by IOp(15) 0-5 imply use of the generalized density in L701. Use sleazier cutoffs in L703. 111 . Generate simulated density derivatives. USE THE ORIGINAL 'BERNY' VALUES in L702. same as -1. Electric field Jth order. IOp(7/18) Derivative accuracy option: 0 1 2 10 20 100 200 Compute to 10**(-8) accuracy. Default.

Second. same as 2. Default method for initial force constants IOp(7/30) USE OF SYMMETRY IN OVERLAY 7: 0 1 USE (SUBJECT TO AVAILABILITY). Use FXYZ in conversion of forces constants to internal coordinates (PSCF opt with HF 2nd deriv). Normal + Generate estimated initial force constants. Compute HF forces from D2E file and increment both FX and FXYZ (non-O11 PSCF grad and HF freq). Read derivatives from checkpoint file (in Z-matrix orientation). IOp(7/31) Handling of forces contributions. DON'T USE. 0 1 00 10 Just use the forces in IRWFX. Third. 0 1 2 10 First. Normal NUCLEAR REPULSION ONLY (USEFUL FOR TESTING). IOp(7/28) SKIP OPTION TO DEFER INTEGRAL EVALUATION TO L703. Use FX in conversion of force constants to internal coordinates. 0 2 COMPUTE AS NORMAL. PSCF freq=numer). (HF freq. 0 1 2 6 00 Normal.IOp(7/25) Type of derivatives available. 112 . DO ALL GRADIENT INTEGRALS IN L703 IOp(7/29) MODE OF USE OF L716.

Punch energy.12 format.false. IOp(7/41) NDDO flag. Neglect three center two-electron integrals as well. Punch energy. Punch energy.true. and derivatives in redundant internals only in compressed form. IOp(7/40) Neglect of integrals (only option 1 works in Overlay 7): 0 1 2 3 10 20 30 Keep all integrals. 0 1 . Do punch only if second derivatives are available. Apply NDDO approximation. Neglect three center one-electron integrals. and derivatives in cartesians and redundant internals. Neglect four center integrals. Neglect 1e integrals with diatomic differential overlap. Force constants and third derivatives are punched in 4E20. forces and lower triangular force constants in format 6F12.IOp(7/32) PUNCH OPTION. Do only overlap and not other 1e integrals.8. IOp(7/42) 1PDM: 0 N Use SCF total density. IOp(7/43) 2nd order simultaneous optimization flag. .12 format in compressed form. Punch nuclear coordinate derivatives. 0 1 Evaluate integrals correctly. coordinates. Neglect 2e integrals with diatomic differential overlap. coordinates.16. 0 1 2 3 4 5 1x None. Punch energy in format D24. first and second derivatives in both cartesian and internal coordinates. (other 2nd derivative options must also be set appropriately) 113 . Use generalized density N. Forces are punched in 3D20. one card per atom.

Compute optimimum lambdas.true. Project using forces if the RMS force is larger than 1. IOp(7/45) Controlling the projection of the reaction path. No. with forces done the usual way for CIS or MP2 2nd derivatives.2 . (debugging option: compute fifth order WG and GG terms in L715) IOp(7/44) Handling of an applied electric field. Update forces for the self-consistent reaction field (SCRF) method Update forces for a uniform electric field. Project using the Newton-Raphson step. and IRanGd. IOp(7/60-62) IOp(60-62) Over-ride standard values of IRadAn.d-6. Yes. IRanWt. IOp(7/52) Whether ECP integrals should be done in L701 as usual. -1 0 1 2 Do not add electric field terms to forces. 0 1 Yes. 0 1 2 3 Do not project. Update forces for a uniform electric field. The point is a stationary point. IOp(7/53) Convert forces over shells to field-dependent dipole and forces over atoms (for debugging): 0 1 10 No. Project the reaction path and compute 3N-7 frequencies. 114 .

0 1 2 3 Default (1). -1 0 1 No.IOp(7/63) Whether to do FMM. if ref structure is present and has non-zero force constants). Lines Lorenzians Both IOp(7/65) Harmonic constraints with respect to initial structure during geometry optimization. Yes. Default (Yes. 0 1 Use global default. Currently lots of hacks to determine where we are in the process instead of different values of this option. No. IOp(7/70) Do vibro-rotational analysis: 0 1 2 Default (No). Yes. IOp(7/72) Read additional parameters for anharmonic computations 0 No 115 . IOp(7/71) Do vibrational 2nd order perturbation: 0 1 No Yes. Turn off FMM here regardless. IOp(7/64) Type of simulated spectrum in output.

Force FoFCou. 116 . Copy fit density over real density and do P(Fit)*Jx*P(Fit). 10^-10 for molecules. Clear real density and do -1/2 P(Fit)*Jx*P(Fit). The threshold is automatically lowered for each mode until 90% of the absolute displacements are included. IOp(7/76) Over-ride use of FoFCou in L703: 0 1 2 Normal processing. IOp(7/75) Threshold for printing redundant internal contributions to normal mode displacements: 0 N -1 Default (10%) 10**-N Zero (all printed). Turn off 1c logic for 1c DBF case. Prohibit FoFCou. Yes.1 Yes IOp(7/74) Non-equilibrium PCM gradients: 0 1 No. 10**(-N). Omit subtraction and do P(Fit)*Jx*P. IOp(7/77) Debuging options for DBFs: 0 1 2 3 4 Normal processing. 10^-12 for PBC. IOp(7/87) Accuracy in FoFDir/FoFCou/CalDSu: 0 N Default.

No.(IK/KL). BD: (IJ/AB). 0 1 2 Default (No). No. BUCKETS FOR STABILITY: (IA/JB).(IA/JB). Full transformation if this is consistent with MaxDisk. -1 0 Yes. BUCKETS FOR SEMI-DIRECT MP4DQ. IOp(8/7) SCF convergence test. IOp(8/6) Bucket selection. BUT INCLUDES (IA/BC). 0 1 2 3 4 5 6 7 BUCKETS FOR MP2: (IA/JB). THE COMPLETE SET OF TRANSFORMED INTEGRALS.Overlay 8 IOp(8/5) Whether to pseudocanonicalize ROHF orbitals. QCISD. otherwise same as 4. BUCKETS FOR CID OR MP3: (IJ/AB).(IJ/KL). Yes. IOp(8/8) Whether to delete MO integrals in L811. Do not test SCF convergence (mainly used for testing). IOp(8/9) Debug control (L802): 117 .(IJ/AB). CISD. CISD or MP4SDQ or MP4SDTQ. otherwise same as 3.(IJ/KA). Full transformation if this is consistent with MaxDisk. 0 1 Test that SCF has convergd.(IA/JB).

(Upper triangle of symmetry compressed integrals for IOp(6)=5 or 4 only!) Reorder MOs. Accumulate MP2 force constant terms in direct fashion Write the MO basis first derivative ERI's to disk Force fully in-Core algorithm (L804 only). L811): 0 1 2 3 10 20 30 100 200 1000 2000 3000 4000 00000 10000 20000 Operate normally. Direct Transformation Control (L804. Force semi-direct mode 2.. Force output bucket in Core antisymmetrization. The window is recovered from rwf 569.virt-rep2.. reordered in the order of representations like occ-rep1. Force N orbitals per pass.. The window is recovered from file 569 on the checkpoint file. Eigenvalues and symmetry assignment vectors will be put in correspondence with vectors.) 30000 100000 Symmetry compress transformed integrals only if RHF. 1 <= N <= 89 selects frozen-core type N: The largest noble gas core is frozen. core virtuals are also frozen. Do symmetry compress transformed integrals (buckets) (This will cause windowed MOs. Generate and test RInt3 array (L804). A negative value for the end deletes the top virtuals. The window is specified by IOp(37-38). a card is read in indicating the start and the end.. Default (10000) Do not symmetry compress transformed integrals.. For basis sets with double-zeta cores. The largest noble gas core and main group d's are frozen.. Read a list of orbitals to freeze. Default (200). Force semi-direct mode 3 if IOp(6)=3. Force semi-direct mode 1. The next to the largest noble gas core is frozen. eigenvalues and symmetry assignment vectors according to ther representations IOp(8/10) Window is selected as follows: -N 0 N 1 2 3 4 90 91 92 93 94 000 118 Use the top N occupieds and lowest N virtuals. Force semi-direct transformation. Force sorting for output bucks. G2 frozen-core: the largest noble gas core and main group d orbitals are frozen. except that the outer sp electrons of 3rd row and later alkalai and alkalai earth elements are retained.0 -N Operate normally. same as 4. virt-rep1. Force semi-direct mode 4 if IOp(6)=3. Default. Force transformed integrals in Core algorithm. If IOp(37) is 0. Use all MOs. .occ-rep2.

119 .0 or the smallest difference between occupied and virtual orbital energies is less than 0. Used to select output mode. 0 1 Output to Gaussian system buckets. and number of electrons test. same as 10. e. Use overlap with atomic core orbitals from Core Ham to choose core orbitals. Kill the job if there is no correlation energy. orbital energy. Read 'old-style' drt input cards. IOp(8/11) MO coefficient. Also.10x 20x 30x Use orbital energies to choose core orbitals. IOp(8/13) Control of output. 0 1 2 Default. IOp(8/14) Control of drt input. if there is only 1 electron or 1 virtual spin-orbital..g. Kill the job if any mo coefficients are greater than 1000.K over deleted orbitals and add to Core-Hamiltonian. Calculate 2 J . Use overlap with atomic core orbitals from Harris to choose core orbitals. kill a frozen-core job if there there is significant core-valence mixing in the canonical orbitals 00 10 20 Default.001. IOp(8/12) Calculation of frozen-Core contributions. Output transformed integrals for DRT-CI calculation. Suppress such a test (CPHF may still be done for such a case). same as 2 Just print a warning message. 0 1 No. IOp(8/15) Control of DRT output. This is done when IOpCl=0 or when IOpCl=1 and the calculation is rohf or gvb. 0 1 Take necessary input from Gaussian data structures.

0 1 2 Write DRT output to RW-files. Canonical occupieds. 0 N Default excitation level = 2. Do both. Uij. Non-canonical.x.x/2 IOp(8/19) The nature of the perturbation(s) in L811: 0 IJK Default (1st order nuclear and electric field). Nuclear Kth order. Electric field Jth order. Uij. Write DRT output to Fortran unit 'drttap'. . Non-canonical. Non-canonical.x = -1/2 Sij.x.x = -Sab. (only if integral derivative file is being written) Excitation level for SDGUGA-CI.x = (Fij. Uij. IOp(8/16) Maximum number of orbitals per pass in L811.x = -1/2 Sij.x Uji. Uij.x = -Sij.x . Magnetic Field Ith order.EjSij. Uij. IOp(8/18) Which type of derivative transformation to do in L811: 0 1 2 3 4 5 6 Default. except canonical in frozen-active blocks. IOp(8/20) Which terms to include in L811: 0 120 Default (same as 11). Uab.x = -Sij. IOp(8/17) Specification of integral block size for GUGA CI programs.x = 0. Integral Block Size = N. Canonical. 0 N Default let program decide.x) / (Ei-Ej) Note that this blows up for degenerate orbitals and is intended primarily for debugging.x/2 Canonical virtuals. same as 3. Excitation level = N.

e. This only applies if the integral derivatives are not stored. N words. but use multi-batch logic.e. This will determine the number of times AO integrals and derivatives are evaluated unless overridden by IOp(31). transformation).. Do one batch.1 10 MO derivative times integral term. Do as many in a batch as can be overlapped with sorting space for half transformed integrals. -1 0 N Use as much as needed for maximum efficiency. N triplets. set NOA=NOB=NROrb in No.. MO times integral derivative term. set NOA=NOB=N for purposes of transformation). IOp(8/27) Maximum amount of disk to use in L804 and L811: 0 N Unlimited. L805. Use an amount which is similar to the maximum disk usage in other parts of the MP2 frequency . Default (-1) N evaluations and hence N coarse tiled batches (1.. IOp(8/28) Hack number of occupieds for full ci using links 921 or 922: -1 0 N Transform all orbitals (after freezing Core) as occupieds (i.. Transform N orbitals (after frozen Core) as occupieds (i. independant of MAXDISK.6 are the currently implemented options) 121 Use as much as desired. Look there for more information. -3 -2 code. IOp(8/29) Maximum number of perturbations per batch in L811: (only applies if integral deriv file is written) -3 -2 -1 0 1 N Do not use batching logic. IOp(8/22) These options control the in-Core post-SCF link. Requested disk usage. subject to the limit imposed by MAXDISK (IOp(27)). Default (same as 1). Do a single atom at a time (minimum disk usage).

0 N Let the program choose (make it as large as possible) maximum fine tile batch size. . 0 1 NO YES IOp(8/32) Whether to do CI in the interacting space only. Set up for full but zero Core MOs. No. Interacting only. only for no-Ix routines). 0 1 2 00 122 Default (Yes). Yes. (Yes if Ix is not stored. Explicit control of the fine tile batch size (largely a debugging option.=j a<=b) for closed-shell AA and AB consistent with integrals. order g i. Set up /Orb/ for a full window but then blank the wavefunction coefficients in L804. IOp(8/35) Output format for closed-shell and debugging control: (only for when integral derivative file is written) 0 1 2 10 Default (consistent with integrals for open-shell. 0 1 2 Default (all spin-eigenfunctions). i<=jab alpha-beta only for closed-shell). up to 9. IOp(8/36) Whether to update force constants with the MP2 product of MP2 integral derivatives term (only applies if integral derivative file is not written). All. IOp(8/31) PERFORM PRIMITIVE POST-SCF OPERATIONS (NOT CURRENTLY FUNCTIONAL). Default on whether to make Poo and Pvv for MP2. Set up /Orb/ as indicated by IOp(10). 0 1 -1 Default. Test window. no otherwise).IOp(8/30) Type of window. Do extra debugging computations. Store only the unique AB integral derivatives (gO2V2/4.

Population method. 0 1 2 3 Default. No. 0 No 123 . but save MOs. transformation. Use ROMP2 approach. No for the 1st geometry of opt. IOp(8/41) Transformation of spin-orbitals (alpha only) within occupied and unoccupied orbital subspaces by minimum orbital-deformation (MOD) method. If C1. transforming only alpha orbitals. IOp(8/40) Handling of ROHF window: 0 1 2 Default (2). IOp(8/38) Integers specifying the window to use. Yes. use unity as rotation. Take reference MOs from disk if available. No. Rotate MOs to compute the displaced overlap. No localization. IOp(8/39) Localized orbital method adopted in SAC/SAC-CI. yes otherwise.10 20 Yes. Boys method. 0 1 2 3 00 10 20 Default. forming pseudo-canonical alpha and beta orbitals and doing UHF Treat as RHF. No. Force not to rotate MOs. IOp(8/42) Whether to reorder MOs during potential surface exploration. Force to rotate MOs. Boys + population method.

N-th occupied band at the k-point for which the hoco is highest (by default) or at k-point N-th virtual band at the k-point for which the luco is lowest (by default) or at k-point specified specified by IOp(44) by IOp(44) IOp(8/46) Indicates special case of non-HF calculation.1 2 00 10 000 100 0000 1000 Yes.MOs are canonical HF orbitals. Input orbitals are not canonical HF and pseudocanonical orbitals must be generated here for the 124 . 0 N Use the k-point for which the hoco is highest and luco is lowest. IOp(8/45) Type of quasiparticle job: 0 -1 1 -N N Band gap. Default . (for SAC-CI single point calculation) Use orbital energies in ordering Don't use orbital energies in ordering Use second moments in ordering Don't use second moments in ordering Use dipole moments in ordering Don't use dipole moments in ordering IOp(8/43) Number of Laplace points to use. Yes. Electron affinity. IOp(8/44) K-point specification for MP2 band correction. 0 1 post-SCF. N -N Use N points for MP2. Use N points and set up for band gap correction. Ionization potential. K-point at which to apply correction.

IOp(8/47) Whether 804/811 should generate results compressed over active atoms: 0 1 2 3 Default (2). No difference from 2 for overlay 8. Full list. but blank contributions from inactive atoms. 125 . Active atoms. Full list.


0 No. UMP43.4) for first iteration Use DD[1-3]R and UMP4xR (closed-shell) on 1st iteration Original code for 2nd and later iterations. The defaults are 22 for RCI. L914: State of interest: 0 N WE ARE NOT DOING GRADIENTS. UMp41U. 10**-6 wfn. 10**-8 energy. default 1 in 907 and 10 in 919. fast and R where possible). 10**-6 wfn. IOp(9/10) Test flag in link 902 Whether to do "fake" frozen-core (i. Use DD[1-3]R and UMP4xR (closed-shell). NO IOp(9/8) L902: Constraint on output wavefunction for stability calculations (see link 902). Slava routines. 10**-5 wfn.3. Original code (DD1. Original routines. same as 1. 0 Default: L913 single point: 10**-7 energy. 11 for UCI. 42 for RQCI. WITH EMP4(SDQ) OR EMP4(DQ) IF SINGLES ARE NOT AVAILABLE. with a full transformation). L913 gradient: L914 gradient: N 10**-N. UMP42.2.3.. DD4RQ (closed-shell). Only active in L914.e. Default. L913: Whether to use fast routines: 000 1 2 10 20 30 40 000 100 200 Default (no Slava. Use DD1. wavefunction for L914). UMP43. 127 . UMP42. DD4UQ Use DD1R. UMP41.2. Number of roots in 907 and 919. Term and method selection for debugging in 906. L914 single point:n: 10**-4 wfn. UMP41R. follow /Orb/.2 7 YES. and 31 for UQCI. FP OR CIS-MP2 WE ARE INTERESTED IN THE NTH EXCITED STATE IOp(9/9) Convergence criterion (on energy for L913.

QCISD(T). Force direct computation of contributions. and store full /Orb/ back on disk. Force direct computation of contributions. note number of frozen core and virtual and reset /Orb/ for full. Force use of sort for intermediates. IOp(9/13) Symmetry constraint of output wavefunction from stable=opt: 0/1 yes/no. Read EMin. Spin projection control in L913: 0 1 2 Default (1) Do basic projection. Read window of MOs to refine in the same format as 801. Yes. Same as 2. IOp(9/11) Flags for Green's function calculations: 0 1 2 00 10 100 1000 10000 Normal use of MO integrals. and pole strength warning level on one line. Normal production of intermediates (in-core if possible). Yes. Test flag in l902. but with two ranges on the same line Force N**3 algorithm in GFSCMA. BD(T). EMax. but save the amplitudes. . Link 909 only. Include triples? IOp(9/12) Test flag in l902.1 2 3 For AO usage (NYI here). for open-shell. Fourth and fifth order singles and triples . Fourth-order triples (NYI). IOp(9/14) Non-iterative corrections: ICNonI 0 1 2 3 128 No.

2N**3 words for derivatives). IOp(9/15) Type of derivative information generated: 0 1 2 None. -2 -1 0 M Force a single integral evaluation (two for UMP2) using disk-based algorithm. suppress in-core storage. Do gradient in L913. Do Lagrangian in L906. IN THE FORMATION OF A NEW WAVE FUNCTION. 0 USE DE DEPENDING ON THE METHOD USED. but also force use of the fully out-of-core algorithm Try to minimize integral evaluations. Use AO integral algorithm (L914 only). requires 2OVN words of memory for E2. IOp(9/18) ITERATION SCHEME: DE= (IN A(S)=W(S)/(DE-DELTA(S)) I. Default (same as -3) Use disk storage for partially transformed integrals handling M occupieds at once. default: in-core if possible. Try to minimize integral evaluations as for -3. Functional to use in L914. L913. using the fully out of core allgorithm. but also force use of the fully out-of-core Force a single integral evaluation as for -2. FOR METHOD = 129 . L914: Control of in-core integrals for W(Tilda): -6 -3 0 1 Force in-core storage. in Tran4D. Do AO derivatives and Lagrangian in L906. otherwise spilling algorithm in Tran4D. Force the fully in-core algorithm. IOp(9/17) Auto-adjustment of tau in L918. but do E4T as well. using fully direct methods if possible. IOp(9/16) L906: Control of (Semi-) Direct MP2: -N -6 -5 -4 -3 to disk.E.4 Same as 2. included Z-amplitudes if necessary. Do a maximum of (-N-6) occupieds per pass. Force in-memory algorithm (fully direct MP2. (IOp(5)).

No (used in HF second derivative calculations).10. Do not extrapolate. 1. Use DIIS. Use A as guess for Z. IOp(9/22) Conversion factor in L919. 130 .GT. NOTE THAT FOR PERTURBATION METHODS (METHOD=2. Reset RLE for Z iterations. 0. Use scaled A as guess for Z.1->.5) DE IS NOT REALLY NEEDED SINCE THE WAVE FUNCTION FORMED NEVER GETS USED. ALWAYS.3. IOp(9/21) Guess for eigenvector of y-matrix in link 902. 1 2 W(0)/A0.4. Use BFGS.->DE = 0. ALWAYS. IOp(9/20) Whether to update the total energy with the MP2 energy in L901. 0 1 Yes. -1 0 N Read in factor in format D20. 10**-N. Default of 10**-8. IOp(9/19) EXTRAPOLATION. IOp(9/23) Localization of orbitals in L919. 0 1 2 3 4 5 00 10 100 Default: CI using old extrapolation.0. FOR METHOD .DE = W(0)/A0. QCISD using RLE. Use RLE. Use old extrapolation for CI.

AT EACH CYCLE YES. SUM(S) A(S)**2 = 1 (ALL S) NOTE: PERTURBATION THEORETICAL RESULTS ARE VALID WITH NORM=0 ONLY 131 . IOp(9/25) PRINT PAIR CONTRIBUTION AND WEIGHT TO CORRELATION ENERGY 0 1 2 3 4 NO YES. Maximum order of perturbation theory in L921 and L922. AT THE END OF CI YES. AT ONE CYCLE GIVEN BY INPUT (I3) YES. Localize both. Localize virtuals. Triplet states. Localize occupieds. Rettrup-Davidson RPA. Default (same as 10). AT FIRST CYCLE AND AT END IOp(9/26) NORMALIZATION OF THE WAVEFUNCTION 0 1 NORMALIZED TO A(0)=1. Read in configurations. Jorgensen-Linderberg Hermetian RPA. In-core method. The functional is given by IOp(17). Correction to CIS in L914: 0 -2 -1 1 2 No CIS-DFT (in primitive energy code) CIS-MP2 (in primitive in-core program) CIS-MP2 (in MO Basis disk routine) CIS-DFT (in production code). Choose configurations by simple truncation.0 1 2 3 00 10 20 000 100 0000 1000 00000 10000 None. Singlet states. Out-of-core method.

Do not form the CI one-particle density matrix. IOp(9/28) PRINTING OF DOMINANT CONFIGURATIONS. MP2 energy). Same for DIIS extrapolation. IOp(9/36) Compute the T1 Diagnostic of T.0001*N. 132 . The maximum dimension is 25. Lee IOp(9/37) Maximum dimension for the QCISD extrapolation.1 and above). Print all coefficients every iteration. For BFGS extrapolation. Compute the CI one-particle density matrix. 0 -3 -2 -1 N Default (print coefficients 0. N words. Scan the 'A' vector and print all coefficients. IOp(9/30) Calculation of the one-particle density matrices: 00 1 2 10 20 Default (21 for CI. Compute the density correct to second order (NOT the same as the density corresponding to the Do not compute the density correct to second order. 22 otherwise). IOp(9/31) Print vectors and matrices in 902 and 918 0/1 no/yes. Do not print coefficients. Scan the 'A' vector and print all coefficients having coefficients greater than 0. force fully direct method by default.default size is ten.J.IOp(9/27) Maximum amount of disk to use in L906: -1 0 N No disk. Use as much disk as needed for a single pass.

For DIIS extrapolation. OR NUMBER OF STATES TO PRINT OUT INFORMATION FOR WHEN USING DODIAG: 0 N -N DEFAULT TO 2 LOWEST N STATES READ IN PRINCIPLE COMPONENT OF N GUESSES (DAVIDSON). THRESHOLD FOR PRINTING EIGENVECTOR COMPONENTS in L914: 0 N ITHR = 1 ITHR = N WHERE THRESHOLD = GFLOAT(10)**(-ITHR) IOp(9/41) L914: NUMBER OF STATES TO SEEK WHEN USING DAVIDSON. the only dimension is IOp(39) L913: Type of convergence test 0 .IOp(9/38) Minimum dimension for the QCISD extrapolation.Default: energy and gradient. The maximum dimension is eight. 1 .Converge on energy only 2 .Converge on gradient only Convergence on gradient is for extrapolated CI and QCISD procedures. FORMAT I5 ON LAST CARD BEFORE EOF 133 .Converge on energy and gradient 3 . For BFGS extrapolation. CASSCF. the previous job indicates wavefunction not just expansion vectors has converged). L914: Pick out guesses from restart file or othogonalize guesses to the states already on restart file (IOp49 must be set to 1 or 2 for this option to be valid) 0 N -1 Just take guess from restart file Make N additional orthogonal guesses to those present Read which N states to use (free format integers) *** WARNING: The states on the restart file MUST be orthogonal to the convergence requested (ie. the smallest dimension one and the default is three. IOp(9/40) Reference wavefunction for MP2 in L906: 0 1 2 Default (HF). HF.

a restart. 333.BB AB BA AA. 222. then don't. AO basis.O. MO Mapped to 3. DEFAULT IS: 3 (RHF REFERENCE STATE) (UHF REFERENCE STATE) MATRIX AA.IOp(9/42) METHOD AND MATRIX BLOCKS TO WORK ON in L914 (See below) -NNN 1 2 3 0 333 BITS 1 NYI NYI 2 20 200 3 30 300 Mapped directly to NNN below. BASIS --|-> FORCE DODIAG TO FIND ALL ROOTS --|-> FORCE DAVIDSON IN M. or 20 as appropriate.BB AB BA AB BA METHOD -|-> FORCE DAVIDSON IN A. In-core. and transitions densities among all excited 134 . If this is Force Davidson to half the number of states at iteration 2. Mapped to 2. BASIS -- IOp(9/43) How to handle subsequent Davidson Iterations in L914: 0 1 2 If this is not a restart.O. or 30 as appropriate. transition densities from ground.BB 10 100 AA. Force Davidson not to half the number of states at iteration 2. IOp(9/44) Density matrix control for filling RWF 633 in L914: 0 1 2 3 states Same as 2 Do densities of each excited state Do densities and transition densities from ground Do densities. then half the number of states at the second iteration.

IOp(9/45) Debug option for comparing previous results in L914.IOp(41) Converge on Ci Amplitudes for N lowest states IOp(9/47) Control of Davidson Iterations in L914: 0 1 2 Usual Don't do any iterations (guess=print) Stop after first iteration IOp(9/48) RESTRICTION ON TYPES OF ROOTS (DAVIDSON RHF ONLY) 0 1 2 3 GUESS ONLY SINGLETS Same as 0 GUESS BOTH SINGLETS AND TRIPLETS GUESS ONLY TRIPLETS NOTE: A SINGLET GUESS MAY RESULT IN A TRIPLET ROOT IN EXTREME CASES (SMALL NUMBER OF ROOTS SOUGHT) IOp(9/49) INITIAL GUESS VECTORS 0 1 2 3 4 MAKE A GUESS BASED ON DIAGONAL ELEMENTS USE GUESS VECTORS ALREADY ON RWF USE GUESS VECTORS ALREADY ON CHK GENERATE GUESSES FROM CIS DENSITIES on CHK GENERATE GUESSES FROM CIS DENSITIES on RWF 135 . 0 1 Use Phycon to convert to eV's Use old conversion to eV's IOp(9/46) Control of Davidson Convergence in L914: <0 0 N Use Ortvec convergence only Converge on the number of roots .

Use form N (see IOp(88) in overlay 5).IOp(9/60-62) Over-ride standard values of IRadAn. No. IOp(9/75) Whether to save amplitudes and integrals in L906: 136 . IOp(9/74) Over-ride default choice of frequency dependence of the XC functional in L914: 0 N Use default value. IOp(9/70) 1 to force TDHF in L914. use equilibrium. Yes. 0 1 2 Default (No). no for stability). if doing excited states. No. IOp(9/72) Whether to computed frequency-dependant polarizabilities. IOp(9/73) Whether to do non-equilibrium solvation in L914: 0 1 2 Default (Yes. Yes. 0 1 No. IOp(9/71) Whether to do an extra iteration after Davidson convergence. IRanWt. Yes. and IRanGd.

Minimal population. Calculate the extrapolated value at N only. -1 0 1 2 3 4 5 10 100 No localization. IOp(9/81) Minimum number of Pair Natural Orbitals (PNO) to start the extrapolations from. 0 -N N Default . IOp(9/85) Flags for SAC-CI 137 .5 (assuming CBS-4 calculations. Get the lowest energy value between CBS(N) and CBS(NVirt). Yes. intended for visualization). i. 0 1 No. No localization. Save amplitudes and integrals.e. Save amplitudes. IOp(9/82) Convergence tolerance for CBS localization.0 1 2 Save only if doing second derivatives (SqS12 set). IOp(9/84) Save CBS localized orbitals to RWF (this will overwrite the SCF orbitals. Boys. Use 10**(-N) IOp(9/83) Localization Method. Do 2nd order. 6-31+G(d'.p')). NStart. Localize core even if not needed. 0 N Use the default. don't save (default). Default (4). save them. Population Boys+Population.

138 . Yes. No.IOp(9/86) Whether L906 should generate data compressed to active atoms during mp2 frequencies with ONIOM: 0 1 2 Default (2).

just set up here unless doing HF 2nd derivatives simultaneously. Set up for GIAO MP2 calculation. D E(CIS) / D R Calc. D E(CID) / D R CALC. Only implemented for closed-shell and UHF. 0 1 2 00 10 20 30 100 1000 10000 20000 No. Available for RHF and UHF.Overlay 10 IOp(10/5) CALCULATION OF FIRST DERIVATIVES OF POST-SCF ENERGIES. D E(CISD) / D R Calc. D E(MP2) / D R CALC. D E(BD) / D R Calc. Don't process any derivative terms here. Use new Px/Wx code and don't save S1 but do save F1. Default derivative processing . Do DFT 3rd derivatives. D E(CCSD/QCISD) / D R Calc. Z-Vector method. Compute dipole derivatives using only electric field CPHF and F(x) matrices. 139 . Partially coded but NYI for high-spin ROHF. D E(CCD) / D R Calc. No contributions to the force constants are done here). D E(MP3) / D R Calc. IOp(10/6) Calculation of the second derivatives of the SCF energy. do D2 E(SCF) / D R(I) D R(J) Setup For MP2 2nd Derivatives (i. Use old Px/Wx digestion code. Setup for external processing of W and Z. Do hyperpolarizabilities for second-harmonic generation.e. 0 1 2 3 4 5 6 7 8 9 00 10 20 30 000 100 200 NO CALC. save as little data as possible. Compute F1 and S1 derivative terms here. Test Z-Vector using full CPHF. Yes. Default: use new Px/Wx digestion code if possible. D E(MP4) /D R Default CPHF usage (Z-vector unless HF D2E) Full 3*NAtoms CPHF. Use new Px/Wx code but save both S1 and F1 over MOs.

0 N Default: 1.D-N. 0 1 No.D-9. Expand each variable in a separate expansion space. Solve all equations together. 0 -1 N Default invert directly if there is enough memory. 0 1 2 3 Default (same as 2). N. the secondary solution. 1.IOp(10/7) RMS CONVERGENCE ON C1(I. Yes. IOp(10/9) Whether to compute Born-Oppenheimer corrections. 0 IJKL Default (1st order nuclear and electric field). possibly reverting to the old (one variable at a time) method in Invert the A matrix directly. except 1. Magnetic Field Jth order. Nuclear Lth order. Nuclear magnetic moment Ith order. 140 . IOp(10/8) Selection of linear equation solution method. never invert directly. IOp(10/11) Largest matrix for direct inversion in LinEq2. The max element is tested against 10* this value. IOp(10/13) The nature of the perturbation(s). IOp(10/10) Control of CPMCSCF during avoided crossing/conical intersection searches.D-11 for Z-Vector CPHF. Electric field Kth order.A) contributions. Always use DIIS.

and S2 off the AO 2PDM. but use of electric field expansion vectors for nuclear coordinate CPHF can cause errors of up to 1 cm**-1 with current tolerances. reusing old expansion vectors for new B vectors can reduce accuracy. Don't update polarizability. This may be acceptable in the electric field second order CPHF. for debugging frozen-core with integrals over the full window.IOp(10/14) Whether to update dipole and polarizability derivatives. 141 . 0 N Use standard machine tolerance (MDCutO) on maximum and rms. C2. Update dipole. Default (Use old vectors if available). 0 1 2 3 00 10 20 Default (=1 if IOp(8)=1 and electric field only and no derivatives are being computed. Don't update dipole Update polarizability. Convert /Orb/ to full. which is used only for one term in polarizability derivatives and for which the accuracy requrirements are less stringent. Note that because of numerical instabilities in the simultaneous solution method. This option is normally used to pass 1st order electric field results to the second invokation of 1002 during frequency calculations. IOp(10/16) Convergence in secondary linear equations (only for simultaneous solution). IOp(10/15) What to do with expansion vectors from the linear equations.0). Pass vectors from 1st to 2nd order CPHF. Use old vectors if available. 0 1 2 10 20 100 Default (yes if IOp(5). Do C1.eq. Ignore old vectors. Save as 2. otherwise). Delete vectors at end of each CPHF. =2 Save vectors at end. S1. IOp(10/17) Frozen-core: 0 1 2 3 Default (use AO 2PDM for Lagrangian only if orbitals are frozen in /Orb/). but leave the full version of /Orb/ on the disk. but delete at end of link. Force 2nd order cphf for polarizability derivatives. Convergence is 10**(-N) for max and rms.

e.e. Only a single set of products AX are computed. IOp(10/19) Whether overlap (S1) terms must be included. 0 1 2 Default (yes).IOp(10/18) Whether to do correct or approximate CPHF. IOp(10/21) Whether to store Uai. Spq. Compute the 2e integrals when needed. Don't use integrals. 142 . 0 1 2 10 Default (No). 0 1 2 3 CPHF is done correctly. even if present. Simultaneous solution is implied. i. MO if possible. even if MO ones are available. force AO or direct. Use option MN in control of 2e integral calculation. Disables use of symmetry to reduce the size of the CPHF problem here. No.. Yes. Read the 2e integral files. The U-matrices are set to zero. and hence the U-matrices are set equal to the B-matrices (i. and full MO Fock matrix derivatives in permanent rwfs. Note that the appropriate rwf (588) must be present in any case. Save magnetic MO deriviatives. independent of convergence criteria. No. IOp(10/20) How to handle 2e integral contributions: 0 1 2 3 4 MNx Default (decide on the fly). The A-matrix is neglected. Force use of AO integrals. Yes. uncoupled Hartree-Fock is used).

recover ints if available on rwf 610. Uniform electric field (dipole) only. all as regular if there are Default: let FoFDir decide. Hexadecapole. -N 0 1 N All integrals done as Raffenetti if there are N or more matrices.IOp(10/22) Which multipole (electric field) perturbations to include? Only used if J part of IOp(13) is non-zero. Quadrupole (electric field gradient. IOp(10/31) Whether to use symmetry to reduce the number of CPHF equations: 0 1 2 Default (yes). Force in-core storage. 143 . less than N. 0 1 2 3 4 Default. Integrals with degree of contraction greater than or equal to N are done are regular integrals. IOp(10/30) In-core storage of 2e integrals: 0 1 2 Default . All integrals are done as regular integrals. all 6 cartesian if possible in direct calculation. IOp(10/28) State for CPMCSCF: 0 N Default (ground state). Force recomputation. Dipole (uniform electric field). Octopole. Nth excited state. IOp(10/29) Use of rafinetti integrals during direct SCF. No. Yes.

IOp(10/47) Whether to do spin-spin coupling constants. Whether to read D2E file in link 1003: 0 1 2 Default (No).the coordinates of which are read in (in Angstroms). Use GIAOs. No. IOp(10/45) Type of Gauge Transformations to perform to calculate the current distribution within the molecule. Use IGAIM method . Yes. 0 1 2 Default (No) Yes.IOp(10/32) Whether to apply interchange in link 1004: 0 1 2 Default (No).the gauge used to calculate the angular momentum perturbed wavefunctions. IOp(10/46) Whether to calculate dipole and rotational strengths (VCD). Use single gauge origin . Default (16 if doing magnetic CPHF). Yes. No. 0 1 2 3 No (Default) Yes No Do only optical rotational strength. Use single gauge origin . -1 0 1 2 4 8 16 None. and hence the molecule's other magnetic properties. No.gauge origin coincident with the nucleus of the integrated atomic regions. 144 . Use CSGT method.

compress if overlay 11 did). No. Use FMM if turned on globally. use more aggressive cutoffs in Xc integration. calculation is performed. use more aggressive cutoffs in integrals and FMM unless doing NFx. Don't compress. Update frequency-dependent property file if frequency-dep. Yes. Use local. N. IOp(10/63) Changing defaults. 0 1 2 3 Default (For nuclear. lower cutoffs suitable only for CPHF/CPKS. unless both electric and magnetic properties are requested). Turn off FMM here regardless. 0 N Default (1000). IRanWt.IOp(10/48) Whether to operate only over perturbations involving active atoms. IOp(10/73) Maximum number of CPHF cycles. read in frequencies. IOp(10/72) Whether to do frequency-dependent properties: 0 1 2 3 4 00 10 20 Default (No. and IRanGd. but blank contributions for inactive atoms. Update regardless. Yes. Yes. 0 1 2 10 20 Default: Use FMM if turned on globally. with formalism for frequency-dependent XC response. 145 . Don't compress. Do not update. Compress. Use global cutoffs. IOp(10/60-62) Over-ride standard values of IRadAn.

Use type N (see IOp(88) in overlay 5). 0 1 2 Default (No). Yes. IOp(10/77) Test CPHF results by checking the CPHF equations using the complete MO Fock and density derivatives. 146 . Default (20). MOD orbital derivatives. 0 1 2 Default (1). IOp(10/76) Over-ride general choice of exchange-correlation frequency dependence. Yes. IOp(10/75) Print during NMR.IOp(10/74) Whether to do non-equilibrium solvation. Canonical MO derivatives. Print eigenvectors as well. Solve with SimEqn. Solve using DiagD. IOp(10/78) Whether to solve CPHF equations for MOD method. 0 1 2 00 10 20 Default (1). Print tensors and eigenvalues. No. 0 1 2 Default: Only if frequency-dependent. No. 0 N Use global value for this job step.

restart point NN. for testing restarts. 147 . MNN Stop at pass M (default 1).IOp(10/79) Stop the link at selected points.

IOp(11/9) IDOUT: FIRST-DERIVATIVE OUTPUT OPTION. also contracted electric field density matrix derivatives to form the two-electron integral derivative contribution to the polarizability derivatives. I0 148 WHETHER OR NOT TO USE THE CONTENTS OF IRWFX. Yes. IOp(11/6) IFHFFX: WHETHER OR NOT TO CONTRACT INTEGRAL DERIVATIVES WITH HARTREE-FOCK DENSITY MATRIX TERMS TO PRODUCE HARTREE-FOCK TWO-ELECTRON CONTRIBUTION TO THE FORCES.Overlay 11 IOp(11/5) IFWRT: DERIVATIVE INTEGRAL WRITE OPTION. but generate and write out the HF 2PDM here for debugging purposes. 0 1 2 NO. 0 1 DO NOT PRODUCE A D2E FILE. PRODUCE A D2E FILE. 0 1 No. IOp(11/8) IFF1: WHETHER OR NOT TO COMPUTE F1 OVER AO'S. Yes. . YES. CONTAINS I2*100+I1*10+I0. 0 1 2 NO. IOp(11/7) IFTPDM: WHETHER OR NOT TO CONTRACT INTEGRAL DERIVATIVES WITH A 'READ-IN' TWO-PARTICAL DENSITY-MATRIX. YES. Yes.

0 1 I1 0 1 2 3 I2 0 1 NO. TAKE HF CONTRIBUTIONS FROM FX1 (A LA IFHFFX). PROCESSING OF TPDM CONTRIBUTIONS. YES. MERELY SET THE ARRAY TO ZEROES. Illegal here. Illegal here. Illegal here. Illegal here. TAKE HF CONTRIBUTIONS FROM F1 (A LA IFF1). use IsAlg to decide. IOp(11/10) Whether to compute Fock matrices. ADD IN CONTENTS OF FX2. IOp(11/12) Selection of 1PDM in L1102 and L1110: 0 N Usual SCF density. 149 . Yes. FoFDir: Prism spdf. Illegal here. (forms the 1/2(F-H) term in link 1110). Use generalized density number N for both the one-electron integral derivatives and the corresponding 2PDM terms. and SCF energy in L1110: 0 1 No. NONE. Illegal here. Illegal here. Lagrangian. IF NOT THERE. NONE. Form 1/2(F+H) term in link 1110. PROCESSING OF TWO-ELECTRON HARTREE-FOCK CONTRIBUTIONS. Illegal here. IOp(11/11) Control of integral derivative algorithm: 0 2 3 4 5 6 7 8 9 10 11 12 13 Default Illegal here. Scalar Rys SPDF. Illegal here.

Force O2V2 method. N**4 I/O algorithm. 150 0 IJK Default (1st order nuclear and electric field). Old gOV3 I/O algorithm. IOp(11/15) Controls output of derivatives to rw-files. Formation of Fx*T*T terms: default is to choose based on available memory. Formation of Ux*I*T Use Ix. default. Use (2g+O)V2 memory algorithm even if O2V2 memory is available. Use L(x) and Ux*I. i4 . Magnetic Field Ith order. Also compute dipole derivative integral contribution to the HF dipole derivatives. same as 1. Also compute HF contribution to the dipole i4*10000+i3*1000+i2*100+1i*10+i0 i0 .0 load fxyz from rw-files if it exists.eq.1 calculate nuclear contribution. i1 . Electric field Jth order.0 calculate one-electron contribution. 0 1 10 Default: compute dipole derivative matrices only. Force old N**5 I/O algorithm.0 controls output of 'old' format. i2 . Default Ix*T algorithm (1) Force new algorithm. IOp(11/14) The nature of the perturbation(s).IOp(11/13) Flags for L1112: 0 1 2 00 10 20 000 100 200 300 0000 1000 2000 Default for Ix==>Sx (same as 1).0 forces out-of-core algorithm IOp(11/16) Mode of operation of L1102. Force old algorithm. Nuclear Kth order. i3 .

Frozen-core in L1111: 0 1 2 3 10 Default (use AO 2PDM for Lagrangian only if orbitals are frozen in /Orb/). Do C1, C2, S1, and S2 off the AO 2PDM. Convert /Orb/ to full, for debugging frozen-core with integrals over the full window. Save as 2, but leave the full version of /Orb/ on the disk. Form the derivative integral contribution to the Lagrangian as well. This is stored on disk as RL(NBasis,NBasis,NAt3,IOpCl+1) in rwf 1001.

Save AO 2PDM from L1111. 0 N No. Save the AO 2PDM on rwf N. It is (NTT,NTT) and includes factors (2-Delta(ij))(2-Delta(kl)). It doesn't include any normalization factor.

Whether to delete MO integrals after 1112: 0 1 2 Default (Yes). Yes. No.

How to handle 2e integral contributions in L1112: 0 1 2 3 Default (same as 1). Read the 2e integral files, MO if possible. Compute the 2e integrals when needed. This link must have been built with the non-dummied version of FoFDir and associated integral routines. Force use of AO integrals, even if MO ones are available. MNx Use option MN in control of 2e integral calculation.

Size of buffers for integral derivative file. 0 N Default (Machine dependent; see DSet2E). N integer words.


In-core option in 1112.

Use of rafinetti integrals during direct term in L1112: -N 0 1 N All integrals done as Raffenetti if there are N or more matrices; all as regular if there are Default: let FoFDir decide. All integrals are done as regular integrals. Integrals with degree of contraction greater than or equal to N are done are regular integrals. less than N.

Output of 1102: 00 1 10 Default (01). Contract with density matrix to form dipole derivative contributions. Store dipole derivative matrices on disk.



Location or generation of MO 1 and 2 PDMs for L1111: -7 -6 -5 152 Compute QCISD 2PDM Compute CCD 2PDM Compute CIS 2PDM

-4 -3 -2 -1 0 N

Compute CISD 2PDM. Compute CID 2PDM. Compute MP2 2PDM. Compute HF DMs. Default (RWFs 626, 627, and 628). RWFS N (1PDM), N+1 (W), and N+2 (2PDM).

What to do: 1 10 100 200 it. 1000 10000 20000 Suppress writing alpha, beta, and spin density rwfs. Form and sort the 2PDM derivatives rather than the 2PDM. Generate replicated 2PDM copies for testing. Transform 1PDM and Lagrangian from MO to AO. Transform 2PDM from MO to AO. Sort AO 2PDM into shell order. If back transformation has not been requested, the double-length AO 2PDM is expected in file 1001. The sorted 2PDM is left in file 602. Form the contribution of the 2PDM to the forces right here. Note that if the 2PDM is also to be left behind, it will be over 6d/10f and have the HGP d and f scale factors in

What to compute using integrals or D2E file. 1 10 Energy. Gradient.

Whether to use symmetry in Rys integral derivatives in L1110: 0 1 Yes. No.

Whether to do 2PDM or just Lagrangian in L1111: 0 1 2 3 4 Compute Full Gradient Compute Full Gradient (Same as Default). Compute Density Only. Compute Density and W Only. Compute 2PDM only, no density or W. 153

Compute over the full list of atoms. Do only overlap and not other 1e integrals.IOp(11/33) IPRINT IOp(33) 0 1 2 NO PRINTING. PRINT COMPUTED FIRST-DERIVATIVES. IOp(11/40) Neglect of integrals (only option 1 works in Overlay 10): 0 1 2 3 10 20 30 Keep all integrals. but blank contributions for inactive atoms. Evaluate matrices in the NDDO approximation. Force compressed form. IOp(11/41) NDDO flag. . IOp(11/39) Compression of derivative matrices: 0 1 2 3 Default (2). Neglect three center two-electron integrals as well. Force expanded form. Neglect 1e integrals with diatomic differential overlap. Compute over the full list of atoms. Neglect three center one-electron integrals. Compute over active atoms only. PRINT OPTION. 0 1 Evaluate usual integrals. 0 1 2 154 Default: compressed. PRINT F1 MATRICES. Neglect 2e integrals with diatomic differential overlap. Neglect four center integrals. IOp(11/42) Compressed file formats.

IOp(11/63) Whether to do FMM. smallest possible number of passes. IOp(11/60-62) Over-ride standard values of IRadAn. IOp(11/53) Convert forces over shells to field-dependent dipole and forces over atoms (for debugging): 0 1 No. Yes. Do at least one pass. and IRanGd. IOp(11/45) Force NAt3 instead of NAt3+3 storage of matrices (for debugging): 0 1 No.3 Compressed Sx but separate H1 and F1. 0 1 Use global default. 155 . Yes. Turn off FMM here regardless. Do at least N passes. but using the out of core algorithms. 0 1 2 No. Convert 1PDM to canonical representation. IRanWt. IOp(11/46) Whether to include orbital rotation gradient terms for SAC-CI. N is 0/1/2 for default/in-core/out-of-core. 0 1 N Default. needed for non-canonical methods. Save gradients to disk. IOp(11/43) Batching in overlay 11. For Rys in L1110.

156 . Copy fitting density over real density. 0 1 2 Default (No). No. Only works using 1C shell pairs for the density basis and only with cartesian functions. 0 1 2 Default (1). Yes.IOp(11/70) Whether to allow cavity to move in PCM derivatives. Ignore fitting density and just process real density in L1110. Print eigenvectors as well. IOp(11/75) Print during NMR. IOp(11/71) Debugging option for DBF derivatives: 0 1 2 Normal processing. Print tensors and eigenvalues.

but don't remove extra data (i. No GVB2P5 trans file. SAVE THE PERMANENT INFORMATION (MOS. IOp(9999/6) Controls output of Fortran unformatted files for other programs: 0 1 00 10 100 200 300 1000 No PolyAtom output. GVB2P5 trans file to unit 14.. Yes. if a new version was not generated in this step). BASIS SET INFO ETC. PolyAtom output in working precision to Fortran unit 8. Archive data from the checkpoint file.E. Use natural orbitals in WFN file IOp(9999/7) Controls whether MOs are written to the polyatom integral tape in LANL style. DELETE THE RESTART INFORMATION IF THE RUN IS FINISHING NORMALLY (I.) ON THE CHECKPOINT FILE.e. IF THE ERROR TERMINATION ILSW BIT IS NOT SET). Save data on the chk file. Do not write anything to the checkpoint file.Overlay 9999 IOp(9999/5) CONTROLS HANDLING OF THE CHECKPOINT FILE: 0 THE RUN IS AN OPTIMIZATION OR FREQUENCY RUN. WFN file output WFN file output with magnetic orbital derivatives. 1 2 3 4 5 THE RUN IS NOT AN OPTIMIZATION. 157 . 0 1 No. Restart a multi-step job. WFN file output with GIAO magnetic orbital derivatives. recovering data from the checkpoint file and figuring out which job step to run next and whether it needs restart if an optimization or numerical frequency. SO BOTH THE PERMANENT AND RESTART FILES ARE IN THE CHECKPOINT FILE.

Old format.. pressure. No. An input deck for HONDO. IOp(9999/10) Controls punching of assorted information (i.12). and isotopes during multi-step energy calculations: 0 1 2 Default (same as 1). 0 1 2 4 8 16 32 64 128 256 512 1024 2048 Nothing. The natural orbitals generated by link 601. 158 . A GAMESS input deck.IOp(9999/8) Reading temperature. 0 1 2 Archive all as is. Output hyperfine tensors as input to Pickett's program (sent to the output file).3D20. formatted output to unit 7). This is independant of normal archiving to the main file. Use natural orbitals in WFN file. Yes. The molecular orbitals. A WFN file for PROAIMS. IOp(9999/11) Which type of database to update: 0 1 2 3 Default (3). Read a list of atoms to use in the Pickett input.e. New format. but rotates to z-matrix orientation first. Atomic numbers and coordinates in format (I3. in format suitable for guess=cards. orientation. use defaults. Derivatives (forces and force constants) in format (2X. except for archiving from the chk file. IOp(9999/9) Controls archiving of dipole moment and other electic field derivatives.3D20.12). These are in the Z-matrix The archive entry. Don't archive. Archive all. Both. Title. in the standard orientation.

Yes. 159 . IOp(9999/16) Treat the job as type (Info(7)) given by IOp(16). IOp(9999/17) Treat as MSJDon=IOp(17) step in a multi-step job. Yes. IOp(9999/13) Whether this is the end of the job step: 0 1 2 3 Default (Yes). and update rwfs. Go back to Link 1. Yes.IOp(9999/12) Flag for coordinate optimization: 0 1 No. remove /ZMat/ and /ZSubst/ from the rwf and chk files. IOp(9999/18) How many virtual orbitals to include in the WFN file. IOp(9999/14) Whether to attempt to express the final optimized structure in terms of the input z-matrix: 0 1 2 3 Yes if there are 20 or fewer atoms. No. No. Yes. IOp(9999/15) Act as though in multi-step job type IOp(15).

Include all virtual orbitals. 160 . IOp(9999/33) CONTROLS DEBUG PRINT: 0 1 NO DEBUG PRINT.0 -1 N Default (None). DEBUG PRINT. Include N virtual orbitals.

Sign up to vote on this title
UsefulNot useful