MOPAC6 is a semiempirical package for calculating physical properties of molecules. Its source code is written in FORTRAN 77 and is freely available. This page has been written with the goal of helping those interested to add the PDDG methods into MOPAC6 to run on a unix or linux machine. The instructions and source code presented have not been tested with any windows versions of MOPAC.

The simplest way to add the PDDG methods to MOPAC6 is to replace either PM3 or MNDO with PDDG/PM3 or PDDG/MNDO. However, the more robust approach which is discussed herein is to add PDDG/PM3 and PDDG/MNDO into MOPAC6 as seperate methods. By following this approach all prior functionality of MOPAC6 is retained with the addition of the PDDG/PM3 and PDDG/MNDO methods using the keywords, PDG (for PDDG/PM3) or MDG (for PDDG/MNDO).

Only five source files will need to be modified to achieve this goal. These are listed for downloading below.

1. Obtain a copy of the MOPAC6 source code. If needed the source code is available via the CCL at http://server.ccl.net/cca/software/SOURCES/FORTRAN/mopac6_sources/
   
   
2. Download the five modified files above (block.f, moldat.f, refer.f, rotate.f, and wrtkey.f) and use them to replace the original MOPAC6 versions of these files.
   
   
3. Compile MOPAC6. Generally this is done by entering the source code directory and typing make. For additional information see documentation included with the source code and/or the Makefile.
   
   
4. Change the Hamiltonian keyword in input files from AM1/PM3/MNDO to PDG/MDG.
   
   
5. Test against results presented in supplimentary information tables (available for download here) listing heats of formation for several hundred organic molecules
   
   

The added common blocks store the reoptimized parameters for PDDG/PM3 and PDDG/MNDO in a format very similiar to the storage of PM3 and MNDO parameters. The block storing PDDG/PM3 terms is shown below, notice the PAPDG, DAPDG, PBPDG, and DBPDG variables which are the important pairwise distance directed gaussian parameters.

      COMMON /PDDGP/  USSPDG(107), UPPPDG(107), UDDPDG(107), ZSPDG(107),
     1ZPPDG(107), ZDPDG(107), BETASG(107), BETAPG(107), BETADG(107),
     2ALPPDG(107), EISOLG(107), DDPDG(107), QQPDG(107), AMPDG(107),
     3ADPDG(107), AQPDG(107) ,GSSPDG(107), GSPPDG(107), GPPPDG(107),
     4GP2PDG(107), HSPPDG(107),POLVOG(107),PAPDG(107),PBPDG(107),
     5DAPDG(107),DBPDG(107)

In moldat.f, parameters in the PDDGP and PDDGM common blocks are read into arrays used in the calculation of molecular properties. The logical variables are necessary for catching the PDG and MDG Hamiltonian keywords in the input file (instead of AM1, PM3, or MNDO). Switching in PDDG parameters is accomplished by the following code added immediately following assignment for AM1, PM3, and MNDO:

       ELSEIF(LPDG) THEN
 *
 *    SWITCH IN PDDG/PM3 PARAMETERS
 *
          DO 32 I=1,107
             DO 22 J=1,10
                GUESS1(I,J)=GUESG1(I,J)
                GUESS2(I,J)=GUESG2(I,J)
    22       GUESS3(I,J)=GUESG3(I,J)
             POLVOL(I)=POLVOG(I)
             ZS(I)=ZSPDG(I)
             ZP(I)=ZPPDG(I)
             ZD(I)=ZDPDG(I)
             USS(I)=USSPDG(I)
             UPP(I)=UPPPDG(I)
             UDD(I)=UDDPDG(I)
             BETAS(I)=BETASG(I)
             BETAP(I)=BETAPG(I)
             BETAD(I)=BETADG(I)
             ALP(I)=ALPPDG(I)
             EISOL(I)=EISOLG(I)
             DD(I)=DDPDG(I)
             QQ(I)=QQPDG(I)
             AM(I)=AMPDG(I)
             AD(I)=ADPDG(I)
             AQ(I)=AQPDG(I)
             GSS(I)=GSSPDG(I)
             GPP(I)=GPPPDG(I)
             GSP(I)=GSPPDG(I)
             GP2(I)=GP2PDG(I)
             HSP(I)=HSPPDG(I)
             PREA(I)=PAPDG(I)
             PREB(I)=PBPDG(I)
             DA(I)=DAPDG(I)
             DB(I)=DBPDG(I)
    32    CONTINUE
       ELSEIF(LMDG) THEN
 *
 *    SWITCH IN PDDG/MNDO PARAMETERS
 *
          DO 33 I=1,107
             POLVOL(I)=POLVOH(I)
             ZS(I)=ZSMDG(I)
             ZP(I)=ZPMDG(I)
             ZD(I)=ZDMDG(I)
             USS(I)=USSMDG(I)
             UPP(I)=UPPMDG(I)
             UDD(I)=UDDMDG(I)
             BETAS(I)=BETASH(I)
             BETAP(I)=BETAPH(I)
             BETAD(I)=BETADH(I)
             ALP(I)=ALPMDG(I)
             EISOL(I)=EISOLH(I)
             DD(I)=DDMDG(I)
             QQ(I)=QQMDG(I)
             AM(I)=AMMDG(I)
             AD(I)=ADMDG(I)
             AQ(I)=AQMDG(I)
             GSS(I)=GSSMDG(I)
             GPP(I)=GPPMDG(I)
             GSP(I)=GSPMDG(I)
             GP2(I)=GP2MDG(I)
             HSP(I)=HSPMDG(I)
             PREA(I)=PAMDG(I)
             PREB(I)=PBMDG(I)
             DA(I)=DAMDG(I)
             DB(I)=DBMDG(I)
    33    CONTINUE

In refer.f the keywords PDG and MDG are mapped to new modes, 5 and 6, a reqirement for adding the PDDG methods to mopac

         IF (ITYPE.EQ.5.OR.ITYPE.EQ.6) THEN
 C
 C THE PDDG FUNCTION HAS BEEN ADDED FOR THE PDDG METHODS
 C
             QCORR = 0.0D0
             ZAF=TORE(NI)/(TORE(NI)+TORE(NJ))
             ZBF=TORE(NJ)/(TORE(NI)+TORE(NJ))
             QCORR =
      1 (ZAF*PREA(NI)+ZBF*PREA(NJ))*EXP(-10.0D0*(RIJ-DA(NI)-DA(NJ))**2)
      2 (ZAF*PREA(NI)+ZBF*PREB(NJ))*EXP(-10.0D0*(RIJ-DA(NI)-DB(NJ))**2)+
      3 (ZAF*PREB(NI)+ZBF*PREA(NJ))*EXP(-10.0D0*(RIJ-DB(NI)-DA(NJ))**2)+
      4 (ZAF*PREB(NI)+ZBF*PREB(NJ))*EXP(-10.0D0*(RIJ-DB(NI)-DB(NJ))**2)
          ELSE
             QCORR = 0.0D0
          ENDIF
          ENUC=ENUC+SCALE+QCORR

Hopefully this information will help in adding PDDG/PM3 and PDDG/MNDO to MOPAC6. For further assistance and aid please send email to Matt Repasky.