IRC calculation issue

Hi,

After obtaining a transition state (TS) from an FSM calculation in Q-Chem 6.4, I proceeded to run an IRC calculation to connect the two ends of the reaction pathway.

While the IRC does proceed, I am encountering an issue where the trajectory appears to get stuck near the TS region. It does not smoothly propagate away toward the reactant and product minima as expected.

I used the following input:

$molecule
0 1
  C      -1.3584668423     0.4909721018     0.0005096194
  H      -1.7661285679     1.4818306420    -0.2041847896
  H      -1.6922307384     0.1172297313     0.9717143176
  H      -1.6467427918    -0.2031601338    -0.7897262709
  O       0.0620917673     0.6879653187     0.0203866595
  C       0.7353311179    -0.5787327527     0.5556515705
  N       0.9826508724    -0.7585598344    -0.6160347207
$end

$rem
JOBTYPE ts
method wb97x-v
BASIS aug-cc-pVTZ
scf_max_cycles  600
geom_opt_max_cycles     600
scf_algorithm robust
INTEGRAL_SYMMETRY false
XC_GRID 3
POINT_GROUP_SYMMETRY false
MEM_TOTAL 80000
$end

@@@

$molecule
read
$end

$rem
JOBTYPE frequency
method wb97x-v
BASIS aug-cc-pVTZ
scf_max_cycles  600
scf_algorithm robust
INTEGRAL_SYMMETRY false
XC_GRID 3
POINT_GROUP_SYMMETRY false
MEM_TOTAL 80000
$end

@@@

$molecule
read
$end

$rem
JOBTYPE rpath
BASIS aug-cc-pVTZ
METHOD wb97x-v
SCF_GUESS read
RPATH_MAX_CYCLES 60
RPATH_MAX_STEPSIZE 300
XC_GRID 3
GEOM_OPT_HESSIAN  READ
MEM_TOTAL 80000
$end

Output file:

 ============================================
      FORWARD RXN PATH (direction = +1)      
 ============================================
 7
 IRC step # 1, E = -207.812070
    C    -1.361313    -0.191027     0.185180
    H    -2.192082    -0.077267    -0.512539
    H    -1.530787     0.381234     1.100665
    H    -1.212534    -1.245259     0.421614
    O    -0.231333     0.343934    -0.517737
    C     0.941615     0.488909     0.455588
    N     1.329179    -0.513923    -0.101779
 7
 IRC step # 2, E = -207.812070
    C    -1.361311    -0.191022     0.185180
    H    -2.192086    -0.077268    -0.512539
    H    -1.530788     0.381232     1.100668
    H    -1.212536    -1.245256     0.421613
    O    -0.231334     0.343934    -0.517728
    C     0.941616     0.488911     0.455573
    N     1.329186    -0.513931    -0.101775
 7
 IRC step # 3, E = -207.812070
    C    -1.361305    -0.191006     0.185177
    H    -2.192100    -0.077271    -0.512537
    H    -1.530787     0.381220     1.100681
    H    -1.212552    -1.245242     0.421607
    O    -0.231328     0.343913    -0.517719
    C     0.941623     0.488927     0.455556
    N     1.329195    -0.513941    -0.101773
 ============================================
      BACKWARD RXN PATH (direction = -1)     
 ============================================
 7
 IRC step # -1, E = -207.812070
    C    -1.361313    -0.191027     0.185180
    H    -2.192082    -0.077267    -0.512539
    H    -1.530787     0.381234     1.100665
    H    -1.212534    -1.245259     0.421614
    O    -0.231333     0.343934    -0.517737
    C     0.941615     0.488909     0.455588
    N     1.329179    -0.513923    -0.101779
 7
 IRC step # -2, E = -207.812070
    C    -1.361310    -0.191021     0.185179
    H    -2.192087    -0.077269    -0.512539
    H    -1.530787     0.381231     1.100668
    H    -1.212538    -1.245255     0.421613
    O    -0.231332     0.343929    -0.517730
    C     0.941618     0.488910     0.455577
    N     1.329182    -0.513926    -0.101776
 7
 IRC step # -3, E = -207.812070
    C    -1.361307    -0.191013     0.185177
    H    -2.192094    -0.077270    -0.512538
    H    -1.530786     0.381225     1.100675
    H    -1.212546    -1.245248     0.421609
    O    -0.231328     0.343918    -0.517725
    C     0.941620     0.488923     0.455570
    N     1.329187    -0.513935    -0.101777
 ============================================
 Total job time:  1426.85s(wall), 16658.43s(cpu) 
 Sun Mar 29 13:24:39 2026

Any suggestions on how to resolve this issue would be greatly appreciated.

Is there an option to start with an exact Hessian (JOBTYPE = FREQ) and then read it in, as in a TS search? For the latter, that’s usually the best option at the first sign of trouble.