simulations under a periodic electric field - possible bug

From: Oleksii Zdorevskyi (
Date: Fri Mar 05 2021 - 11:11:29 CST

Dear NAMD developers,

I am performing all-atom simulations of DNA dodecamer neutralized by
potassium ions in a water box under an external periodic electric field.

When the frequency of the electric field is high enough, the simulation
works fine.
That is, under 1Ghz (eFieldFreq is 0.001 1/ps), the ionic current changes
its sign every quarter of the period (in this case every 0.25 ns).

But when reducing the frequency by one order of magnitude, namely to 0.1
GHz (eFieldFreq is 0.0001 1/ps), the current must change its sign every
2.5 ns. However, during the whole 10-ns simulation the system experiences
a constant ionic current. It is seen from visual examinations as well as
from the explicit calculation of ionic current within a Tcl script.

Could you give me an idea if NAMD has some limits to eFieldFreq? That is,
could it be that when this parameter is, for instance 0.0001 (1/ps), the
program just equals it to zero?

In both cases (for frequencies 1Ghz and 0.1 GHz) the initial electric
field vector was:

eField -0.522 0 0

If it is necessary, I can provide you with all the configuration files and
simulation results.

Unfortunately, I have not found any documentation about the implementation
of a periodic field in NAMD. I have seen only this post
however, it does not discuss any limitations of the eFieldFreq parameter.

Best regards,

This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:10 CST