I've only seen this warning reported twice, both times I don't think a solution was presented.
I ran IRC in Gaussian over a TS found at wb97xd/3-21+g with no problems at all. But when I take the TS up to wb97xd/6-31+g(d) and rerun IRC I am presented with the following:
---------------------------------------------
Pt 1 Step number 1 out of a maximum of 300
Modified Bulirsch-Stoer Extrapolation Cycles:
EPS = 0.000010000000000
Maximum DWI energy std dev = 0.000012714 at pt 1
Maximum DWI gradient std dev = 2.234578259 at pt 29
WARNING: Bulirsch-Stoer Method is not Converging
CORRECTOR INTEGRATION CONVERGENCE:
Recorrection delta-x convergence threshold: 0.010000
Delta-x Convergence NOT Met
------------------------------------
I have tried a number of approached:
1) EulerPC rather than HPC
2) Reducing the step size to 6 (0.06 bohr)
3) Increasing the step size to 30 (0.30 bohr)
4) tight and int=ultrafine
5) use=L115 (this won't ever run).
All of these variations present with the same warning, and I am unable to move off a single point on the reaction coordinate. It is probably worth noting, at 3-21+g the vibration is approx -500, but at 6-31+g(d), it drops to approx -150.
Any thoughts/suggestions are appreciated.
Thanks
Anthony
- Badges
- Science topic