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. 103. 96. 98) IOp(1/101. 95.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.

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 .

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

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

0 1 2 DEFAULT (TEST for z-matrix or cartesian TS but not for LST/QST or for minimum). (not valid for L109).0) (L103 only). Estimate force constants using valence force field. TEST.1 -N 0 N Yes.001) Step size for ABM method in Trudge for isodensity method.J).I). L113.0**-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. L105. 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. (5I3. coordinates.J).8) from input stream. Bohr. and radians). Approx 1. 0 1 2 3 4 5 6 7 8 9 10 Use defaults (not valid for L109). Use semiempirical force constants.NVAR) (8F10. L109. BOMB THE JOB IF THE SECOND SECOND DERIVATIVE MATRIX HAS THE WRONG NUMBER OF NEGATIVE EIGENVALUES.J. Use unit matrix (default for L105. Use unit matrix throughout.6) (L103 only). L117: Scaling Factor for Determining Overlaps of VDW atoms -1 0 N Turn off scaling Default is 1. Read from checkpoint file in internal coordinates. End with a blank card. L112.F20. followed by a blank line.000 + N*(0. IOp(1/11) L103: TEST OF CURVATURE.010 1.I=1. Read ((FC(I. Second derivative matrix calculated analytically.FC(I. using a default criteria (0.0D-6 (N=20) 2. using a (10**-N Angstroms) criteria.05 Angstroms) Yes. Read I. only recognized by 103). L114: Input of initial Hessian: All values must be in atomic units (Hartree. DON'T TEST.J=1. 3 . How close to get to the isosurface in search.

L113. only in L103 and L115).L109: ABORT IF DERIVATIVES TOO LARGE -1 or 0 N No force test at all. Allow N -.1 * N 4 . D2Corr (BFGS) D2Corr (Bofill Powell+MS for transition states). L121: Multi-time step parameter (NDtrC. Powell for L113 and L114). D2Corr and L115.linear only starts with the N+1st. 0 N Default (0 for TS.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. FMAXT = 0. no quadratic step)..L115: Type of Hessian Update: 0 1 2 3 4 5 6 7 8 9 Default (9 for L103 minimization. D2Corr (No update.0 N 0.1/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. use initial Hessian). 1 for minima). only in L103 and L115). D2Corr (New if energy rises. D2Corr (Old. BFGS (not in L103) BFGS. Powell (not in L103). number of bad steps to allow before attempting a linear minimization (i. 7 for L103 TS. otherwise BFGS). IOp(1/15) L103.e.L114. safeguarding positive definateness (not inL103 or L115) D2Corr (New.05 (N=2) 0.

Do full optimization in redundant internal coord. IF THE LIMIT IS EXCEEDED. THE SIZE OF THE EIGENVALUE IS REDUCED TO THE MAXIMUM.0001 EIGMIN = 1. Skip MM atoms in internal coordinate definitions and do microiterations the new way.L113. / N IOp(1/18) L103: Coordinate system.L114: MAXIMUM ALLOWABLE MAGNITUDE OF THE EIGENVALUESOF THE SECOND DERIVATIVE MATRIX. Default (2000000). in L103. Skip MM atoms in internal coordinate definitions and do microiterations the old way.L113.1 * N IOp(1/17) L103. 0 N EIGMAX = 25.L114: MINIMUM ALLOWABLE MAGNITUDE OF THE EIGENVALUES OF THE SECOND DERIVATIVE MATRIX. Do full optimization in pruned distance matrix coords. in L120. Read the AddRedundant input section for each structure.IOp(1/16) L103. Do full optimization in redundant internal coords with large molecular tools. 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. 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.0 HARTREE / BOHR**2 OR RADIAN**2 EIGMAX = 0. IOp(1/19) L103: SEARCH SELECTION 5 . Include MM atoms in internal coordinate definitions (no microiterations). SIMMILAR TO IOp(16) 0 N EIGMIN = 0. Do optimization in Z-matrix coordinates. AND PROCESSING CONTINUES. No optimization Do optimization in cartesian coordinates. Microiterations for pure MM. done in L402. will occur.

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

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

1*(10-J)*P. The third one is the initial guess of the transition structure. but read new values for some variables from the input . BY Z-MATRIX BY DIRECT COORDINATE INPUT (must set IOp(29) in L202). R and P are used to guide the QST optimization of the TS. Default. structure) from the checkpoint file. default QSTrad = 0.Transition state searching using QST and redundant internal coordinates L= 0.01*I. Abort job if model builder generates a z-matrix with too many variables. Default (same as 100). model A. Set all optimization flags to optimize. either initial guess of the minimizing structure or transition structure Input 2 structures. 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. The first one is reactant the second one is the product. model B. The values of the TS coord are estimated by interpolating the sturcture of R and P. GET CARTESIAN COORDINATES ONLY FROM THE CHECKPOINT FILE. Do not abort job if model builder generates a z-matrix with too many variables. The union without QST. charge and multiplicity. 7 10 000 100 200 1000 2000 8 Get all input (title. By model builder. of the two redundant coordinates are taken as the redundant coords for the TS. the first one is the reactant. the second one is the product.1*J*R + 0. Read optimization flags in format 50L1 after the z-matrix. IOp(1/29) L101: SPECIFICATION OF NUCLEAR CENTERS 0 1 2 3 4 5 6 stream. GET Z-MATRIX AND VARIABLES FROM THE CHECKPOINT FILE. L= 3 TS. Get Z-matrix from the checkpoint file. By model builder. Normal (6 or 5 for linear molecules). 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.g.1 L= 2 Input one structure.05) Input 3 structures. same as -1. N. 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. Print details of the model building process.

0 1 DON'T PUNCH. Mark Z-matrix constants as frozen variables rather than wired-in constants. (2+3) Purge all flags but keep the coordinate definition. Rebuild the coordinate system. 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). 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. Generate a symbolic z-matrix using all Cartesians if none is present on the checkpoint file (a Same as one. PUNCH. 0 1 YES NO IOp(1/32) TITLE CARD PUNCH CONTROL. but retain the redudant internal coordinate definitions. 9 . Default. same as 10000. hack to make IRCs work with Cartesian input). It also suppresses error termination on large gradients. Do not retain symbolic constants.3000 4000 5000 00000 10000 20000 100000 200000 Purge flags except the frozen variables.

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

L115: Number of points along the reaction path in each direction. 0 N 1. terminated Read in replacment radii for selected atoms as pairs (I.0D-10 1. -1 for the initial geometry IOp(1/42) L103. terminated by a blank line. Recalculation the Hessian every N steps. IOp(1/41) Step number of optimization from which to take geometry.Rad) or (Symbol. Default is 6. L114: Hessian recalculation. Force CHELPG (Breneman) recommended radii. execpt for CalcAll.Rad). 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. Just update.IOp(1/40) L113.0. -1 0 N Pick up analytic second derivatives every time. The default. Read from input stream. 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. L116: Whether to read initial E-field: 0 1 2 Start with 0. Read from checkpoint file.Rad). Initial 11 by a blank line.0D-N IOp(1/43) L116: Extent of Reaction Field. L117: Cutoff to be used in evaluating densities.

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. Internal.1 N. 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. 12 .radius of spheres to be placed around attractors to "capture" the gradient trajectories. from the outset.0 au N. 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. 0 1 Do not read isotopes.used in Trudge only to eliminate.M = NM/10 IOp(1/44) IRC Type 0 1 2 3 L117: Default (same as 3). 0 NM 0. Cartesian. Read Isotopes. points which clearly lie in another basin. Mass-weighted.

No electrostatics included in the model systems Do full square for testing. Default: 1 2/1 dont/do optimize product structure. 0 N Default: 50000 N. 3 Restore point M from data on the rwf. IOp(1/49) Options to IRC path relaxation (IJKL) L K J I 2/1 dont/do optimize reactant structure. Default: unimolecular IOp(1/52) L101 and L120: Type of ONIOM calculation: 0/1 2 3 00 10 20 100 N000 One layer. Default: 1 3/2/1 dont/QST3/QST2 optimize TS structure (for QST input). 1 20 Do IRCMax Include zero-point energy. Use atomic charge type N-1 during microiterations. 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 . The default is MK charges. IOp(1/48) IRCMax control. normal calculation Two layers Three layers Default (20) Include electrostatics in model systems using MM charges. IOp(1/53) L120: Action of each invocation of L120: 0 1 2 there. Default: 1 2/1 unimolecular/bimolecular reaction.IOp(1/47) Number of redundant internal coordinates to allow for.

gradients. 0 1 Default. hessian) of point NN of the ONIOM Next point to do is MM. Dreiding/UFF. Yes. 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. Amber allowing any symbol. IOp(1/55) L103: Options for GDIIS: ICos*1000+IChkC*100+IMix*10+Method form. and hessian Restore point MM from RWF but do not create a new model system. Amber.4 5 6 7 NN0 MM000 Integrate energy Integrate energy and gradient Integrate energy. energy. L115: IRC optimization. gradient. IOp(1/57) Whether to produce connectivity: 14 . for use with parameters in input stream. IOp(1/56) Set of atom type names to parse: 0 1 2 3 Accept any. use gradients to find the next geometry. Use displacements to find the next geometry. NN = MaxLev**2 + 1 (currently 17) to restore real system. Save necessary information (some rwf's. grid.

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

Quartic fitting field (NYI). UFF. MM3 (NYI).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. Do not read modifications to parameter set. Read modifications to parameter set. soft has priority. IOp(1/65) Control of which terms are included in MM. Use soft and hard-wired. If IOp(67)=3. Use only soft. Use the last when there are equivalent matches. hard-wired has priority. Lowest 2 digits then have no meaning. 2=Gradient. abort when different parameters match to the same degree. Use the first when there are equivalent matches. corresponding to the 'classes' in FncInf. Supermolecule Fragments with ghost atoms NFrag+1 . MM2 (NYI). then the default is to apply soft parameters with higher priority. With soft parameters. Dreiding. 0 1 10 100 1000 10000 Do all (default) Non-bonded Stretching Bending Torsion Out-of-plane 16 . Use only hard-wired. MMFF (NYI). AMBER. Use soft and hard-wired.2*NFrag -.IOp(1/63) Step in counterpoise calculation: MNN NN = 0 1-NFrag M = order of derivatives (1=Energy.

over-written the values in AtChMM. Yes. Don't do QEq. Default (10) Do for atoms which were not explicitly typed. Copy from chk file. else 1. Default (100) Do for atoms which have charge specified or defaulted to 0. Generate here. 0 1 2 3 Default: 2 if reading geom from chk file. 0 1 2 00 10 20 000 100 200 Default (2. Do for all atoms regardless of initial charge. IOp(1/67) Source of MM parameters. Pick up non-standard parameters from chk file. 1==> 221) Do QEq.IOp(1/66) Whether to generate QEQ charges. 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. reading from input if requested by IOp(64). IOp(1/72) L103: Algorithm choice for microiterations. Do for all atoms regardless of typing. 17 . or to use the values already there.

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.

M. IOp(1/89) Maximum allowed deviation from average nuclear KE during ADMP.L. Read isotopes from rwf. Average energy (in microhartree) to be maintained during Simulation. Integrate current densities. as required by IOp(80).2 Print eigenvectors as well. 4 if geometry read from chk) Use most abundant isotopes. IOp(1/90) To read in the velocity in cartesian coordinates Nuclear Kinetic Energy Thermostat Option. Read isotopes from input. 21 .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. for backwards compatibility. 0 1 2 K0 L00 M000 Do not integrate. Read isotopes from chk. IOp(1/87) ONIOM integration of density. Integrate densities specified by following digits: Density to use from gridpoint 1 Density to use from gridpoint 2 etc. in Kelvin. The temperature and pressure are read first. K.

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

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=0.003 (N=30) Error=N*0.0001 IOp(1/108) Gradient magnitude for DVV stopping criteria (Crit1) 0 N Default (N=15) N*0.04 (N=400) v0 is set to N*0.10 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.

0 N Default (1 unless specified by IOp in overlay 7 or read in). N/1000000. unless read in). 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. N/1000 atmospheres. 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. 0 N Default (1 atomosphere. N/1000 degrees.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. unless read in). 24 . 0 N Default (standard temperature. IOp(1/113) Pressure for thermochemistry.

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

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

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

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

2p) on H .Overlay 3 IOp(3/5) TYPE OF BASIS SET. rest LANL1MB. MacLean-Chandler (12s.52111) for second row. G3large basis set.Q.Be 2SDF.Large APNO basis set CBS basis #6 -. Ahlrichs SV basis sets. 5 SPLIT VALENCE N-21G (OR NN-21G) BASIS FOR FIRST OR SECOND ROW ATOMS.6-311++G(2df. EPR-III basis sets. Coreless: Li. He 6-311+G(2df) on Li . 0 1 2 3 4 MINIMAL STO-2G TO STO-6G EXTENDED 4-31G. use daggers if any polarization CBS basis #3 -.6-31+G(d. Ahlrichs TZV basis sets. UGBS basis set. IOp(6) selects options. IOp(6) specifies type. MIDI! basis sets.p) on H .p) on H. 29 . Dunning/Caltech basis sets. EPR-II basis sets. IOp(3/6)=5 for MTsmall basis set. S. 17 18 19 20 21 22 23 24 25 26 Stuttgart/Dresden ECP basis sets. 6 7 8 9 10 11 12 13 14 15 16 LANL ECP basis sets.6}Z) and augmented if IOp(7)=10. Literature citations in SDDPot. type selected by IOp(6) (=0-4 for V{D. USE IOp(8) TO SELECT 5D/6D. Type selected by IOp(6) for H-Ar.p) on P.T. Stevens/Basch/Krauss/Jasien/Cundari ECP basis sets for H-Lu.) SEE IOp(6) FOR DETERMINATION OF THE NUMBER OF GAUSSIANS IN THE INNER SHELL.Ar CBS basis #4 -. (VARIOUS IMPLEMENTATIONS MAY OMIT SECOND ROW ATOMS.Si 6-31+G(df. B-Ne 2MWB. Cl CBS basis #5 -. G3MP2large basis set.5.6-31+g(d.Ne 6-311+g(3d2f) on Na . GENERAL--SEE ROUTINE GenBas FOR INPUT INSTRUCTIONS.Ne 6-311++g(3d2f) on Na . Type selected by IOp(6). CBS basis #1 -. The same numbers are used for all basis sets.Ar CBS basis #2 -. whether intended for use in expanding AOs (IOp(5)) or in expanding the density (IOp(82)). Literature citations in CEPPot.6-31G.9p)-->(631111.5-31G.Core correlation basis set Dunning cc basis sets.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.

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

p) -. ONE SET OF P-FUNCTIONS AND ONE SET OF D-FUNCTIONS ON HYDROGENS. 0 1 2 3 4 5 6 7 8 9 10 20 100 200 300 400 500 600 700 1000 NONE. VSTO-4G basis for 1st row. TWO SETS OF P-FUNCTIONS AND ONE SET OF D-FUNCTIONS ON HYDROGENS. Functions from extended Huckel theory. D95 and no ECP otherwise. Augment non-hydrogens only (cc basis sets only). 31 . 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. 2 SETS OF P-FUNCTIONS ON HYDROGENS. 2 D-funcs. Three sets of d functions. ONE SET OF D-FUNCTIONS AND ONE SET OF F-FUNCTIONS ON HEAVY ATOMS (indicates an extra tight 2df with ccp basis sets. P-FUNCTIONS ON HYDROGENS. along with LP-31G potential. CBS-Q d(f). 1d on 1st row atoms. 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. Three sets of p-functions and one set of d-functions. Three sets of d functions and two sets of f-functions. Three sets of d functions and one set of f-functions.p polarization basis Tight d for VnZ+1 (W1 theory) A SET OF DIFFUSE SP FUNCTIONS ON HEAVY ATOMS. on heavy atoms (scaled up/down by a factor of 2 from the standard single D value).d. 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.d.2 CEP-121G. D-FUNCTIONS ON HEAVY ATOMS (2ND ROW ONLY FOR 3-21G). A DIFFUSE S FUNCTION ON HYDROGENS. 2d. Three sets of p-functions.2d on 2nd and later atoms. TWO SETS OF D-FUNCTIONS AND ONE SET OF F-FUNCTIONS ON HEAVY ATOMS.

Write over the dipole length integrals (518). FORCE 10F. Add ghost atoms to /B/ so that every shell is on a separate center. Do not split S=P AO shells. Do not split AO S=P=D Uncontract the AO basis. D. F. FORCE 6D. Split S=P AO basis shells into separate S and P shells. 0 -1 N Usual place (572). 32 .95)/3 None. shells. Split S=P=D= AO shells into S=P.7. Na-Ar=(0. Store in RWF N. 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.2. Uncontract the density basis Uncontract both basis sets.325*(IA-1).IOp(3/8) SELECTION OF PURE/CARTESIAN FUNCTIONS. Massage the data in Common /B/ and Common /Mol/.65*I-4. Li-Ne=0. 6D/7F 5D/7F 5D 5D 5D 5D 5D/7F IOp(3/9) Where 308 should store dipole velocity integrals. Read in general basis data in addition to setting up a standard basis. 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. the values for H-Ar can be determined by Slater's rules: H=1. He=1. FORCE 7F. For VSTO-nG.

43 17.25 2.50 1.80 1.00 1.00 1.75 1.40 2SP 3SP 0.88 3.61 11.56 13.10 2.00 2SP 1.00 SECOND ROW ATOMS ATOM P S CL 1S 1.24 1.69 3.03 1.68 5.CHEM. 2686 (1963) OUTER SHELL HAS BEEN SELECTED ON THE BASIS INNER SHELLS ARE BEST ATOM VALUES OF NUMEROUS OPTIMIZATION STUDIES ON VARIED SMALL MOLECULES.98 1.PHYS.90 2.05 2.36 4.79 6.01 1.00 1.0 FOR LI-AR (INNER AND OUTER) STANDARD POLARIZATION EXPONENTS FOR N-31G* AND N-31G** BASES 33 .98 1.12 1.98 0.66 8.50 15.00 1.00 1.53 14.ATOM H HE LI BE B C N O F NE NA MG AL SI P S CL A 1S 1.90 4. 38.00 1.55 2.75 1.00 3SP 0.70 1.47 16.69 2.70 1.31 5.02 1.65 9.99 1.95 2.00 2SP 1.15 1. N-31G (ALSO N-31G* AND N-31G**) STANDARD SCALE-FACTORS HYDROGEN H 1S 1.83 5.67 6.67 7.00 0.00 1.74 1.15 FIRST ROW ATOMS ATOM B C N O F 1S 1.04 0.20 1S* 1.00 3SP* 1.00 1.64 10.33 J.48 3.01 LP-N1G SCALE=1.00 2SP* 1.72 1.59 12.26 6.68 4.99 0.98 0.

0 1 2 3 9 REGULAR INTEGRAL FORMAT IS USED. ZINDO/S.multiply moments by fudge factor for charged species. Read coefficients of field. 34 .09 0. translation vectors and d functions properly: 1 2 3 MNDO/AM1. RAFFENETTI '3' INTEGRAL FORMAT. 0 -1x -6 No. to account for sparkles. CAN ONLY BE USED WITH THE CLOSED SHELL SCF. IOp(3/14) Addition of electrostatic integrals to core hamiltonian. ATOM NA.39 IOp(3/11) CONTROL OF TWO-ELECTRON INTEGRAL STORAGE FORMAT.8 STANDARD POLARIZATION EXPONENTS FOR STO-NG* BASIS. IOp(3/12) Flag for semi-empirical runs.6 0. PROCEDURES.2 0.1 0. IOp(3/13) Nuclear center whose Fermi contact terms are to be added to the core hamiltonian. RAFFENETTI '1' INTEGRAL FORMAT IS USED. The magnitude is specified by IOp(3/15). MG AL-CL VALUE 0. SCRF calculation -. starting with electric field. ZINDO/1. blank terminated. INDO/2.ATOM H LI BE B C-NE VALUE 1. RAFFENETTI '2' INTEGRAL FORMAT. SUITABLE FOR USE WITH OPEN SHELL RHF SCF AND THE POST-SCF USE ILSW TO DECIDE BETWEEN RAFFENETTI 1 AND 2. CNDO/2. SUITABLE FOR USE WITH THE OPEN SHELL (UHF) SCF.4 0. up through 34 elements (hexadecapoles) in free format. but not yet accepted by them.

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

IOp(3/21) Size of buffers for integral derivative file.MWB (third set) Pseudopotentials for all coreless basis.MWB (first set) Dresden/Stuttgart potentials . 36 . 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.MWB (second set) Dresden/Stuttgart potentials . IOp(3/20) Size of buffers for integral file. No longer used.SHF Dresden/Stuttgart potentials . Alternative potentials for coreless basis. 55296 words on Cray). WITH A SUBSTITUTION POTENTIAL OF TYPE N WHEREVER SUCH A SUBSTITUTION POTENTIAL EXISTS.MHF (second set) Dresden/Stuttgart potentials . N integer words. 0 N Default (Machine dependant. IOp(3/18) PRINTING OF PSEUDOPOTENTIALS 0 1 2 PRINT ONLY WHEN INPUT IS FROM CARDS or if GFPrint was specified. 16384 integer words on VAX. N integer words.MDF Dresden/Stuttgart potentials .MHF (first set) Dresden/Stuttgart potentials .CHF).12 13 14 15 16 17 18 19 20 Dresden/Stuttgart potentials . 0 N Default (3200 integer words).

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

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

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

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

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.12). GNDDO/1 parametrization. Use Slater's rules scale factors. Original INDO/2 Beta and HDiag Parameters. geometric mean for NDDO/1 Arithmetic mean in NDDO. Do ZINDO/S.I). for testing. 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. Same as 1. 1100000 Do Harris functional scaling atomic densities for current charge and multiplicity.I).3). Load all the integrals into memory. Read the integrals sequentially. Do CNDO/2. Use the real overlap matrix. Default (unit overlap matrix). 1200000 First-order XC. IOp(3/42) How to form NDDO core hamiltonian in L317: 0 1 2 Default (same as 1). Do INDO/2. Read parameters from chk. Use the unit matrix for the overlap.I=1.J=1. Read parameters for atomic numbers 1-18 in the order Scale (D20. 1500000 J as usual but NDDO for K. Geometric mean in NDDO. or read-in parameters).18) Read parameters from rwf.J=1. 1400000 Regular SCF with separate K.I=1.12). Read from Chk. Do ZINDO/1. Use STO-3G scale factors. 1000000 Do Harris functional. Default parameters (same as 5).18) (Format 3D20. 41 . 1300000 Second-order XC (NYI). followed by ((Beta(J. followed by ((HDiag(J.3).

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

Make all atoms large in XC quadrature. Forbid use of gather/scatter digestion even for small numbers of density matrices. even if not doing FMM. 43 . 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). The default is diagonal only on vector machines. Force use of FoFCou. Reverse normal choice of Scat20 vs. Trace input and output using Linda/subprocess. Do allocation for parallel XC but run sequentially. 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. Reverse normal choice of diagonal/canonical sampling in Prism and PrmRaf. Turn off vFMM. Turn off MP-based cutoffs in FF part of NFx. No longer used. Skip consistency checks for XC quadrature Do not do extra work to use cutoffs better. Use Mura radial grid instead of Euler-2 grid. Reserved for more control of scatter/gather. Force square loops. Force all near field in FMM. currently only affects CalDFT.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 single matrix code in CPKS. Forbid use of FoFCou. matrices only on Fujitsu and NEC. Do not symmetry reduce grid points on unique atoms. Turn off use of Sqrt(P) in density-based cutoffs. replicated Fock matrices. currently only in PrismC. Default is to use replicated Turn off Schwartz during FMM/NFx calculations. Do allocation for parallel 2e integrals but run sequentially. Turn on use of precomputed XC weights. Make all shells large in XC quadrature.

Convert to sparse storage under FoFCou for testing. Box length MMMM/1000 Bohr. if 128 set. Apply symmetry to derivative distributions (NYI). 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). Force FMM on. IOp(3/53) Accuracy in ECP integral evaluation: 0 -1 N Default. Do not save as many multipole expansions as possible in memory. so L302 does not do ECP ints. 44 . Uncontract all shell pairs. Split primitives for better boxification. based on geometry (but minimum for molecules 3.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. Set up for parallel FMM but run loops sequentially. Turn off parallelism in FMM (does not use parallel logic). Turn on FMM print. Do not default to FMM. No Cutoffs 10**-N. Default UseUAB/Use 256. Set by PsmSet to indicate whether the NAtoms test for defaulting FMM was passed. ECP integrals are evaulated in L302.0 Bohr). 0 1 Default: if needed. IOp(3/52) Turn off normal evaluation of ECP integrals. Old routines will be used. Default box length. UseUAB.

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. cutoff 10**(-N) IOp(3/56) Cutoff for intermediate matrices during sparse operations: 0 N 100 times smaller than storage cutoff. 45 . No Yes. intermediate accuracy (5x10**-7) Yes. 10**(-N). of core electrons for Stuttgart/Dresden ECP's. cutoff 5 x 10 ** (N+100) Yes. default accuracy (10**-10). crude accuracy (5x10**-5) Yes. IOp(3/59) Threshold for throwing avay eigenvectors of S: 0 N Default (10**-4) 10**-N. IOp(3/58) Cholesky control options. IOp(3/57) No.

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

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

Becke3 using VWN/LYP for correlation. B1LYP. VWN 80 (LSD) correlation VWN 80 (LSD) + Perdew 86 correlation . PBE1PBE mPW3PBE. PBE3PBE. Default. HCTH93. BA3PBE. mPW1PBE. B97-1. mPW91PW91. Becke3 with Perdew 91 correlation. HCTH407. B97-2. Perdew 81 + Perdew 86 correlation. Becke 3 with Perdew 86 correlation. Becke "Half and Half": 0. Lee-Yang-Parr correlation.IOp(3/72) Solvent type flag (for SCRF setup). B98. Perdew 81 correlation. B1B95. HCTH147. mPW1LYP. IOp(3/73) ONIOM system flag (for SCRF setup). No correlation.5 LSD Do only coulomb part. Vosko-Wilk-Nusair method 5 correlation.5 HF + 0. skip exchange-correlation. LG1LYP. 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. BA1PBE. same as 100. Becke "Half and Half" with LYP/VWN correlation.

49 . B98 coefficients.81.0 for DFT and HF.5 HF + 0. Coefficients of 0. III radial points. mPW91PW91 coefficients. Hartree-Fock-Slater exchange (Alpha = 2/3). JJJ angular points. LG exchange. 200 is Hartree-Fock-Slater. B97-2 coefficients. Becke "Half and Half" 0.0 and 1.72 c=0. IOp(3/75) Number of radial and angular points in numerical integration for DFT: 0 IIIJJJ Use a special grid designed for efficiency (default). same as B1B95. with a=0. respectively.8 b=0. and 402 is BLYP.7) Becke 1988 exchange.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. PW91 exchange Gill 96 exchange PW86 exchange mPW exchange PBE exchange BA exchange VSXC exchange So 100 is Hartree-Fock.5 Xc + Corr Coefficients of 0 and 0 (no exchange). Becke 3 coefficients: aLSD + (1-a)HF + b(dBx) + VWN + c(LYP-VWN). B97-1 coefficients. IOp(3/76) Mixing of HF and DFT. Note that Becke actually used Perdew correlation rather than LYP. HCTH coefficints. 205 is Local Spin Density. X-alpha exchange (alpha= 0. -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 O3LYP coefficients.

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

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

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

DBFs have normalization as for AOs. Put all functions into a single block in forming the preconditioning matrix. IOp(3/90) Thresholds for density fitting MMNN 10**(-MM) on iterative solution. for debugging. Douglass-Kroll-Hess 2nd order. IOp(3/91) Scalar relativistic core Hamiltonian: 0 1 2 3 4 Default (1) Non-relativistic. Solve non-iterative using precomputed A'^-1.5xx 6xx 1xxx Solve iteratively with non-symmetric block-diagonal preconditioning. Same as 2 but exponents are 100x smaller. DBFs have J-normalization. AOs have normalization as for AOs. IOp(3/93) Nuclear charge distribution: 0 1 2 3 4 10x Default (1. RESC. AOs have J-normalization. 53 . al Very tight single s Gaussians. 10**(-NN) on generalized inverse. IOp(3/92) Whether read-in basis sets are in terms of normalized primitives? 0 1 2 3 10 20 30 Default (12). DBF coefficients are for raw primitives. AO coefficients are for raw primitives. default MM=09. Douglass-Kroll-Hess 0th order. unless scalar relativistic) Point nuclei Single s Gaussians using formula of Quiney et. Include nuclear charge distributions in DBF set. for debugging. default NN=06.

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

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

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

Re-use Fock matrices instead of orbitals.4 above. otherwise 2). EVEN WHEN BASES ARE IDENTICAL. Check MOs for othornormality. Note that variable IGuess here has 4. Don't check MOs for othornormality. chosen via IOp(11). Renormalize and orthogonalize intermediate SCF results which are on the RWF. Renormalize and orthogonalize the coefficients which are currently on the read-write files.2. Huckel guess (only valid for NDDO-type methods). Default MO checking (yes). 57 . Read generalized density specified by IOp(38) from the chk file & generate natural orbitals from Huckel is used. IOp(4/6) FORCED PROJECTION WHEN GUESS IS READ FROM CARDS (401).2. 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. Default orthogonalization (perform) Schmidt orthogonalize guess orbitals. IGuess values of 10-14 are generated internally and are the sparse versions of 0 and 5-8. Suppress orthogonalization. in which case Read guess from the checkpoint file. Guess from model Hamiltonian. Default (1 for PBC without alter.1 corresponding to 1.3.5 * V. 9 it. SUPPRESS PROJECTION.Overlay 4 IOp(4/5) Type of guess: 0 1 2 3 4 5 6 7 8 it. on the checkpoint file. Read intermediate SCF results which are on the checkpoint file. Projected ZDO guess. 0 1 2 00 10 20 000 100 200 FORCE PROJECTED GUESS. EVEN WHEN BASES ARE IDENTICAL. FORCE PROJECTED GUESS. This uses the Harris functional unless atoms heavier than Xe are present. Re-use orbitals not Fock matrices.3. Read generalized density specified by IOp(38) from the rwf file & generate natural orbitals from Default.




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.

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

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

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

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

Yes. This is the GVBCAS test mode. Use Lewis dot structure guess density. This is a normal GVBCAS calculation. No pairs. 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. L405 performs a CAS on the inner space. IOp(60-62) Over-ride standard values of IRadAn. but is smaller subsequently. There are n pairs: 2*n extra orbitals and electrons will be added into the active space later. Use diagonal guess density. Yes. IOp(4/48) Whether to do (sparse) Conjugate Gradient methods in 402: 0 1 2 No.instead of simple determinants IOp(4/45) Ipairs = number of GVB pairs in GVBCAS. normal CAS calculation. 65 . use the Lewis dot structure to generate a sparse guess directly. IRanWt. Default (-1 if sparse is turned on) Yes. -1 0 N No. 0 n Default. and sets up L510 to compute extra matels etc. and IRanGd. L510 will execute normally. -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. Use threshold 10**-N. implicitly.

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

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

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

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

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.

Just recompute band structure from stored real-space Fock matrix. Skip valence-virtual Fock matrix elements. Terminate after computing the 2e terms at the first iteration. 200 1000 2000 10000 ADMP. 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. CONTINGENT ON SPACE. Fit the converged density even if fitting is not in use during the SCF.turned on automatically in FoFDirINFC Number of Frozen Core Orbitals IRdNT. L503). REORDERING OF THE ORBITALS (MAINTAINING CONTINUITY OF THE WAVEFUNCTION ALONG THE SEARCH PATH.(Calculates directly Nact*(Nact+1)/2 Fock matrices by contracting the AO integrals with the Density matrices. use IOp(17)). CALCULATION IS BYPASSED. CONTROL OF ANNIHILATION OF SPIN CONTAMINANTS (L502). Number of rows in an initial transformation of MO. 73 . (More input from cards See below) IOp(5/14) Special functions in L502: 0 1 10 20 100 matrices. (Obsolete. Use full diagonalization method rather than Lanczos. 0 1 2 CALCULATION IS PERFORMED (PROVIDED OF COURSE THAT ENOUGH SPACE EXISTS IN THE RW-FILES). IRdNLp. Skip core-valence Fock matrix elements. State average density matrices. 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). Turn the current RHF run into a uhf run at the end of this link. Use Generalized energy-weighted density routines regardless. first cycle: use initial AO densities.L510: MCSCF flags: 2 10 100 XX000 Generate MOs using UHF natural orbitals.-For testing purposes. CALCULATION IS PERFORMED. ADMP. 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.

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

Don't orthogonalize. SET TO ZERO IF Abs(F(I.0 N C(1) = 1. symmetrize them to ensure that they are exactly Do not symmetrize the density matrices. Use full 2nd order convergence. 0 1 Virtuals obtained by diagonalization of hamiltonians. Force the density matrix to have full symmetry at the first iteration. DEGEN) M. in preparation for CPMCSCF. Don't use natural orbitals each iteration. Just schmidt.J)).14) Selection of virtual orbitals (L506). symmetrize Fock matrices and require proper density matrix symmetry. Warning!! should be used only for large jobs where Hessian does not fit in memory.O.FUZZY. choose between replicating integrals and symmetrizing the Fock matrix based on whether the current density matrix is symmetric. replicate integrals so that density matrices and wavefunctions need not If 2E symmetry is on. 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. Prepare for CPMCSCF(FREQ): Direct method with no storage of Hessian. 2nd order iteration using RFO type step + level shift. then Schmidt. 0 1 2 5 10 100 200 300 400 500 Orthogonalize C. Tighten convergence by an extra factor of 10. 1012 for 508). If 2E symmetry is on. Choose LinEq convergence based on orbital gradient. L510: MCSCF flags.V by separate Lowdin.D-10. symmetric. DELETE COUPLING TERMS BETWEEN ALMOST DEGENERATE (DELTA E .D-5 FUZZY AND DEGEN READ IN (2D20.O. Default (1000) If the density matrices pass the symmetry test. then Schmidt. Force the density matrix to have full symmetry at every iteration. be symmetric.LE. Lowdin orthogonalize C+O and V. DEGEN=2.0 IOp(5/17) CONDITION OFF-DIAGONAL TERMS OF THE Fock MATRIX (L503). 2nd order iteration at end. Always use tight convergence. If 2E symmetry is on.LE. Bad for 1st order method. Virtuals obtained by Schmidt orthogonalization to occupieds. Same as 30 in 502 but 20 in 508.0 C(N) = 1. 2nd order iteration using RFO type step + level shift and prepare for non-direct CPMCSCF 75 . VECTORS 0 1 FUZZY=1.

LE.D-4.J)).(the weights 8F10. Force Lower-triangular in-memory L506. N/100 new density. L503: CUTOFF CRITERIA IN SYMMETRY DETERMINATION OF M. 10000000 Use State Average density matrices.d-2). .SPAN ARE CONSIDERED TO BE ZERO 0 1 STHRS=1. (100-N)/100 old density. L502. Turn off damping. otherwise 1. SYMMETRY IS DETERMINED IF LARGEST OFF-DIAGONAL M. IOp(5/19) Over-ride integral storage control (L501. Save generated integrals on disk (file 610).8) 20000000 Do SA and prepare for SA-CPMCSCF.O.(IRdCIV) Use full diagonalization method rather than Lanczos. Do not save integrals (same as 0x).14) Damping (L506) Maximum rotation gradient for Newton-Raphson in L508 (above this value. Force allocation for 1 or 2 buffer case conventional case (VV.IBuf2E).S. scaled steepest descent is used): 0 N Default (1.J)).D-7 STHRS AND SPAN READ IN (2D20. Default (-1 unless reoptimizing during Stable=Opt). Forbid in-core: force re-reading of integrals even if they fit in 2 buffers if conventional.10000 100000 1000000 Attempt to control root flipping in CI. 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.STHRS ELEMENTS Abs(F(I.GE. Force Square in-memory storage. L508): -1 0 1 2 3 4 1x 2x 3x 76 Choose the best given amount of memory available.O. 30000000 Do SA and prepare for Gradient of Energy difference. FOCK-MATRIX ELEMENT Abs(F(I. 2 if possible. SPAN=5. Read CI vector and use it every iteration. 10**-N. Force computation of raff 1 and 2 integrals even for RHF. Dynamic selection of density damping based on band gap and DIIS error. do not convert to in-core if direct and enough memory for in-core is available.

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

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

L505. the unpaired orbitals are assumed to be high spin. Yes. ROHF Triplet (a debugging option). Force UHF. not needed in L506). -1 0 1 N All integrals done as Raffenetti. If -1.2) GVB(1/2) Orthogonal open-shell singlet. Closed/Open control for L511: 0 1 2 Default. Number of hamiltonians to read in (L506). IOp(5/30) Whether to symmetrize final orbitals using abelian symmetry operations (L502. IOp(5/27) Whether to do closed-shell calculation in L502. Default: let FoFDir decide. It will never use Raffenetti for SCF. 0 1 Default (Yes). Integrals with degree of contraction greater than or equal to N are done are regular integrals. if mulitplicity 1). 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. closed for Multip=1. All integrals are done as regular integrals.2 1 0 -1 -2 Excited singlet as 2nd root of CAS(2. 0 1 2 Default (Yes. an open-shell singlet is assumed. error if Multip>1.2). If zero. GVB as CAS(2. 79 . No Yes (used for RHF direct SCF). Force closed shell.

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

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

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

8 f10. If negative reads the vector from rwf. If positive reads vector from input 4f20.7: the displacement in geometry in internals in Angs.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. 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. 10 Include the CSF contribution to the orbs for the DerCpl.

Diagonalization Conjugate gradient. L510: configuration cutoff for mp2 0 i 84 . Square storage (only in Fock formation if CG).1 Float(1/i) .N Control flag for CalDFT is N. 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. 0 1 Yes canonicalize. or 22 if sparse is set in ILSW). N words. 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: 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.) IOp(5/52) Amount of memory to allocate to stashing integrals. also none. Linear storage (only in Fock formation if diagonalization). -1 0 N None Default.

in link 1003 a normal frequency calculation is performed when the optimization is in 85 . N N/1000 Bohr. N>0 N<0 Solvent type N. Default is 0.5-0.IOp(5/53) PCM input and solvent type. the number of steps for switching on/off is -IOp(55). Usefull for optimizations or trajectory calculations where only a part of the surface is (nearly) degenerate. 0 Default . (Or reversed. when the SA calculation graduately is switch off. L510: see below IOp(5/56) A0 for Onsager SCRF.050 Threshhold for switching off.5). which is usefull when a trajectory is started at a degenerate region. IOp(56) IOp(57) IOp(58) Threshhold for switching on. N of each.0-1.5-0. then start after instability searches. IOp(55) Number of steps over which the SA coefficients are brought from (0.0) to (0. default parameters Dielectric constant |N|/1000 IOp(5/55) How many HOMOs and LUMOs to solve for after CG: 0 N None. Energy difference larger than IOp(57)*0. L510: See below IOp(5/55-58) L510: Switching on a State Averaged calculation graduately. In that When set. the calculation is started with coefficients (0. L510: See below IOp(5/57) First iteration at which to level shift and do FON.001 .1 unless doing stable=opt. Default is 0.) When IOp(55). Energy difference smaller than IOp(56)*0.5).001 .0. The state averaging will be switched off (graduately) when the degenerate region is left.

use N. Yes.0).9 0.0-1.6 0. 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. IOp(5/64) Over-ride default value of FMFlags 0 N No.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.7 0.5/0.8 0. regardless of weather that continues to decrease after the switch point.2/0.3/0. IOp(5/63) Whether to do FMM. 0 1 Use global default.1/0. only used with IOp(97)=11 IOp(5/60-62) Override standard values of IRadAn.8 0.Threshold for a determining an adiabatic hop threshold=n*0.5 crash new SA 0. instead of a State Averaged second derivatives calculation. 86 . for a hypothetical example with nroot=2: cycle 4 5 6 7 8 old SA 0.3/0.7 0.0-1.a region of (0.4/0. 1: 2: Normal computation when (0.0/0.8 0.2/0.2/0. and IRanGd. Turn off FMM here regardless.0001 (checks Offdiagonal element). IRanWt.1 Egap <limit <limit >limit >limit >limit Note: trajectory option set in overlay 10: IOp(10/99)=n -.

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

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. and beta orbitals for fractional occupation. (defaults to 4 CG cycles). alpha orbitals.G. 0 Default. (defaults to 3 cycles). Convergence criterion 0 N 88 Defaults to 10**(-7) 10**(-N) . No preconditioning. IOp(5/80) The maximum conjugate gradient step size -1 0 MMNN No maximum step size Default maximum (.8) Step size of MM. IOp(5/82) C. IOp(5/79) Range around Fermi level where temperature distribution will be applied if broadening is turned on for PBC. beta electrons.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.NN IOp(5/81) Conjugate-Gradient Parameters MM NN00 00000 10000 000000 100000 0000000 1000000 Maximum Number of CG cycles per SCF iteration. a value will be chosen in ZInLT1.

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

do preliminary AM1 calculation. set in FoFDir/FoFCou/CalDSu based on accuracy part of IOp(5). The energy is only checked after FON has been turned IOp(5/87) Accuracy criterion in Fock matrix formation: 0 N Default. separate Coulomb and NFx exchange for PBC). do preliminary AM1 calculation and compare with guess from previous step in geometry optimization. IOp(5/88) L510: controls the amount of printing. . if DIIS is on. IOp(5/92) Whether to use FoFDir or FoFCou for exact exchange: 0 90 Default: normal processing based on FMM for non-PBC. 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. this and ZFormV should be moved to L302. 1 2 Yes. IOp(5/90) Whether to generate sparse guess here.200000 off. IOp(5/91) Control option for chebyshev sparse control. 10**-N. Retain 3 and 4 point extrap. Yes. Typically 10^-10 for molecules and 10^-12 for periodic systems.

FoFCou for Coulomb. separate FoFDir for exchange.1 fempto seconds If IOp(97)=23.25 in GS) and if Zyy=125. variable=IWait. 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: .01) or the lower state (=(1-x)*0.01) 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 2 FoFCou for Coulomb. L510: has different meaning depending on if you are using IOp(97)=22 or IOp(97=23) If IOp(97)=22.125 in GS How long to wait before checking for a hop after going through a hop. threshold for a hop up (after a hop down) = 0. N iterations. threshold =0. wait time is xxx*0. value=xxx. where: Zyy yy Z xxx Threshold for a hop down determined by probability of being on the upper state (=x*0. Value = xxxzyy. 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).75 in ES (or 0.

RAS control in L510. 0 N Use default.01 (checks energy gap).04 switch timdep on. 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. variable=mHTest IOp(5/94) PCM/ONIOM calculation 0 1 2 Standard PCM calculation 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. default yyy=5 ie deltaE>0. Ras1 orbitals are doubly occupied. on the real system PCM/ONIOM calcn. and the maximum number of electrons in the Ras3 space: zzyyxxww. default xxx=4 ie deltaE<0. In the reference space.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. where ww 92 Number of Ras1 orbitals . Ras3 orbitals are empty. Ras1. The CAS active space is subdivided into three RAS active subspaces. and for pass 0 grid in L502.0 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. Ras2 and Ras3. Use grid N.

IOp(5/99) Grid for numerical k-integration in FT-LT method. if available). Yes. Needs option 80 and option 99. Hopping threshold during trajectories with L510. Yes.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.2. 0 1 2 10 11 21 Default (Yes. 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. 0 Default: 32.8 for 1.12. 93 . Yes. 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. No. 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). hop is based on the probability of being in a specific state Propagation of the wavefunction but no hopping. for detailed control of the propagation conditions see options 92-96. No.3d IOp(5/100) Tight convergence during CGDMS: 0 1 2 Default (No). 0 1 2 Default (No). 22 23 Propagation of the wavefunction.

N. 0 -1 N>0 Default (10) No test Abort of SDif is larger than N. IOp(5/102) Maximum number of configurations for CAS-MP2: 0 N Default (1000). 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.

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

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

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

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

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

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

.99.9. default 2). IHowFa is the maximum distance in the Runge-Kutta integrations along gradient paths (1.9. default 6).0001*IEpsIn is the target for integration error (1..9.9....*IPraIn is the cut-off for zero-flux surfaces (1. default 2).9..don't read input.. default 2).. This link must have been invoked with IOp(40) = 0 or 1 prior to invoking it with IOp(40)=2. 10. 3 Read the deletion energy produced by a previous run with IOp(40)=2 and print it. 10...9...**(-IGueDi) is the initial displacement from the critical point in the Runge-Kutta integrations (1.. IOp(6/40) Control of link 607. 10*INInGr is the initial number of grid points in theta and phi in the adaptive integration subroutine (1.9.. 0.9.9.. IEpsSf IEpsIn INTrig IEpsSf is the safety factor used for patches with surface faults in the adaptive integration subroutine (1.. Read input data to control NBO analysis.9. 0 N Default (4). N layers. IOp(6/41) Number of layers in esp charge fit. default 2). 5+INInCh is the initial number of sampling points in the intersection search (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).*IRtFSe is the safety factor used in the intersection search (1. default 2). must be >=4. default 5). IRScal is the scaling factor in the nonlinear transformation used in the intersection search 10. whose energy can then be computed by one of the SCF links. default (1..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.. Delete selected elements of NBO Fock matrix and form a new density. (1.....9..**(-5-IToler) is the tolerance for the intersection search (1. 0 1 2 Default NBO analysis . 101 . default 2). default 6). default 2). 10*INTrig is the number of sine and cosine functions in the trial function for surface sheets 10.. default 5).

4/Sqrt(#layers)) 0. 0 -1 N Default .01*N. same as 2. 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 . IOp(6/46) Threshold for molecular volume integration.IOp(6/42) Density of points per unit area in esp fit. IOp(6/43) Increment between layers in MK charge fit. 0 N Default (1).10**-3 Read from input. 102 .for tight accuracy. 0 N Default (0. 50 is recommended. N*10**-4. points per unit area. IOp(6/44) Type of calculation in L604: 0 1 2 3 10 Default. 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.

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

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

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

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

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

2 Default. 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.0 1 tensors. 108 .

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

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

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

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

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

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

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

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). IOp(7/77) Debuging options for DBFs: 0 1 2 3 4 Normal processing. Force FoFCou. Prohibit FoFCou. IOp(7/87) Accuracy in FoFDir/FoFCou/CalDSu: 0 N Default. Turn off 1c logic for 1c DBF case. IOp(7/75) Threshold for printing redundant internal contributions to normal mode displacements: 0 N -1 Default (10%) 10**-N Zero (all printed). 116 . Yes. Omit subtraction and do P(Fit)*Jx*P. 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. 10^-12 for PBC.1 Yes IOp(7/74) Non-equilibrium PCM gradients: 0 1 No. 10**(-N).

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

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

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

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

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

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

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

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. Yes. K-point at which to apply correction. Use N points and set up for band gap correction. 0 N Use the k-point for which the hoco is highest and luco is lowest. Default .MOs are canonical HF orbitals. Input orbitals are not canonical HF and pseudocanonical orbitals must be generated here for the 124 . Electron affinity. (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. IOp(8/45) Type of quasiparticle job: 0 -1 1 -N N Band gap. N -N Use N points for MP2. IOp(8/44) K-point specification for MP2 band correction. Ionization potential.1 2 00 10 000 100 0000 1000 Yes. 0 1 post-SCF.

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


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

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

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

QCISD using RLE.10. IOp(9/23) Localization of orbitals in L919. ALWAYS. Use BFGS. IOp(9/19) EXTRAPOLATION. Use RLE.DE = W(0)/A0. 0.1->. Do not extrapolate. ALWAYS.5) DE IS NOT REALLY NEEDED SINCE THE WAVE FUNCTION FORMED NEVER GETS USED. 130 .GT.->DE = 0. No (used in HF second derivative calculations). -1 0 N Read in factor in format D20. 1 2 W(0)/A0. IOp(9/21) Guess for eigenvector of y-matrix in link 902. IOp(9/20) Whether to update the total energy with the MP2 energy in L901. 1. NOTE THAT FOR PERTURBATION METHODS (METHOD=2. IOp(9/22) Conversion factor in L919. 0 1 2 3 4 5 00 10 100 Default: CI using old extrapolation. 10**-N.3.4. 0 1 Yes. Use scaled A as guess for Z. Use old extrapolation for CI. Reset RLE for Z iterations. Default of 10**-8. Use A as guess for Z. FOR METHOD .0. Use DIIS.

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

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

CASSCF. the only dimension is IOp(39) L913: Type of convergence test 0 .Default: energy and gradient.IOp(9/38) Minimum dimension for the QCISD 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). For BFGS extrapolation. the smallest dimension one and the default is three. IOp(9/40) Reference wavefunction for MP2 in L906: 0 1 2 Default (HF).Converge on energy and gradient 3 . FORMAT I5 ON LAST CARD BEFORE EOF 133 . For DIIS extrapolation. 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.Converge on energy only 2 . The maximum dimension is eight. HF. 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. 1 .Converge on gradient only Convergence on gradient is for extrapolated CI and QCISD procedures. the previous job indicates wavefunction not just expansion vectors has converged).

or 20 as appropriate.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. AO basis. or 30 as appropriate. and transitions densities among all excited 134 .O. BASIS -- IOp(9/43) How to handle subsequent Davidson Iterations in L914: 0 1 2 If this is not a restart.O. BASIS --|-> FORCE DODIAG TO FIND ALL ROOTS --|-> FORCE DAVIDSON IN M. transition densities from ground. a restart. then half the number of states at the second iteration.BB AB BA AA. then don't. DEFAULT IS: 3 (RHF REFERENCE STATE) (UHF REFERENCE STATE) MATRIX AA.BB AB BA AB BA METHOD -|-> FORCE DAVIDSON IN A. 333.BB 10 100 AA. MO Mapped to 3. 222. 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. Force Davidson not to half the number of states at iteration 2. In-core. If this is Force Davidson to half the number of states at iteration 2. Mapped to 2.

IOp(9/45) Debug option for comparing previous results in L914. 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 .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 .

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

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

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

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

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

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

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

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

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

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

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

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

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

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

Force old algorithm. default. IOp(11/14) The nature of the perturbation(s). N**4 I/O algorithm. Also compute HF contribution to the dipole Old gOV3 I/O algorithm. i1 . IOp(11/15) Controls output of derivatives to rw-files.0 forces out-of-core algorithm IOp(11/16) Mode of operation of L1102. Force O2V2 Force old N**5 I/O algorithm. Nuclear Kth order.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). Default Ix*T algorithm (1) Force new algorithm. Magnetic Field Ith order. Formation of Ux*I*T terms. i3 .ne. 0 IJK Default (1st order nuclear and electric field). i2 . Formation of Fx*T*T terms: default is to choose based on available memory.0 load fxyz from rw-files if it exists. i4 . 0 1 10 Default: compute dipole derivative matrices only. same as 1.0 calculate one-electron contribution. Also compute dipole derivative integral contribution to the HF dipole derivatives. 150 . Use Ix. Use (2g+O)V2 memory algorithm even if O2V2 memory is available.1 calculate nuclear contribution. Use L(x) and Ux*I.0 controls output of 'old' format. i4*10000+i3*1000+i2*100+1i*10+i0 i0 .ne. Electric field Jth order.

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

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

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

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

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

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

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

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