! EXAM29.  surface scan
!  The scan is done over a 3x3 grid centered on the SCF
!  transition state for the endothermic SN2 type reaction 
!      F- + NH2OH -> F-NH2-OH anion -> FNH2 + OH-
!
!  Groups 1 and 2 are F and OH, and their distance from
!  the N is varied antisymmetrically, which is more or
!  less what the IRC should be like.  The results seem to
!  indicate that the MP2/3-21G saddle point would shift
!  further into the product channel, since the higher
!  MP2 energies occur at shorter r(NF) and longer r(NO):
!
!  FINAL E= -229.0368324512, E(MP2)= -229.3873300818
!  FINAL E= -229.0356378315, E(MP2)= -229.3866641810
!  FINAL E= -229.0309266214, E(MP2)= -229.3822093947
!  FINAL E= -229.0372146613, E(MP2)= -229.3923234590
!  FINAL E= -229.0385440206, E(MP2)= -229.3936484220
!  FINAL E= -229.0367369459, E(MP2)= -229.3913681166
!  FINAL E= -229.0328601039, E(MP2)= -229.3918932844
!  FINAL E= -229.0364643856, E(MP2)= -229.3948324033
!  FINAL E= -229.0372478139, E(MP2)= -229.3943496933
!
!  A more conclusive way to tell this would be to compute
!  single point MP2 energies along the SCF IRC, since the
!  true reaction path always curves, and thus does not lie
!  along rectangular grid points.
!
 $contrl scftyp=rhf runtyp=surface
         icharg=-1 coord=zmt mplevl=2 $end
 $system timlim=1 $end
 $surf   ivec1(1)=2,1 igrp1=1       
         ivec2(1)=2,5 igrp2(1)=5,6  
         disp1= 0.10 ndisp1=3 orig1=-0.10
         disp2=-0.10 ndisp2=3 orig2= 0.10 $end
 $basis  gbasis=n21 ngauss=3 $end
 $guess  guess=huckel $end
 $data
F-NH2-OH exchange (inspired by J.Phys.Chem. 1994,98,7942-4)
Cs

F
N 1 rNF
H 2 rNH   1  aFNH 
H 2 rNH   1  aFNH   3 aHNH  +1
O 2 rNO   3  aONH   4 aONH  -1
H 5 rOH   2  aHON   1 180.0

rNF=1.7125469
rNH=0.9966981
rNO=1.9359887
rOH=0.9828978
aFNH=90.18493
aONH=79.34339
aHON=100.78851
aHNH=108.57000
 $end

Link to the log file of this example.
exam29 Log File