Error Message

[ashokgrg19]

Hello,

I am getting this kind of error while  doing my simulation of Hydrogen Bombardment on Tungsten Matrix.
I don't why I am getting this error.  The attached file is the description of error.

Thanks,
Ashok

[Ulrik G. Vej-Hansen]

It looks like there is something wrong with the file when it tries to write to it. Could you provide the script that causes this? If possible, also the .nc file you are writing to. Which version are you using?

[ashokgrg19]

Hello,
Thanks for the response. I think, I fixed the problem.
However, I have another issue. I tried to see the trapping of hydrogen atom in tungsten matrix. For this, I bombarded hydrogen atom, with K.E around 100eV,  on tungsten matrix at 2000K. During my simulation, I found that hydrogen atom get transmitted through bottom layer of tungsten, although I made reservoir fixed. Is this because of thin bottom layer of tungsten matrix? The attached is my input file. My software version is 2016.4.
Thanks,
Ashok

[Ulrik G. Vej-Hansen]

It is only the tungsten atoms in the bottom layer which are fixed, so when you shoot in Hydrogen atoms at very high speed, you will need a much thicker tungsten layer to absorb them.

[ashokgrg19]

Thank you for the response.
But if I run that simulation for large number of atoms let say 108 Hydrogen atoms on reservoir, with making  total MD step around 500000, log interval of around 100, hydrogen atom deposition interval around 100. Then I still found the trapping of some number of Hydrogen atoms. How does that possible if thickness of the tungsten is not thick enough?

Another question for you, Is there any way to study how much percentage of hydrogen atom get trapped inside the tungsten? Currently, I am running simulation to see the trapping of Hydrogen in Tungsten matrix, but my hydrogen atom gets reused after they reflected back from tungsten matrix, which I don't want. Is there any way to make reflected hydrogen not get reused up, make them freeze at reservoir, and next deposition start with the new hydrogen atom, so that I can count how many number of atoms get trapped and how many atoms reflected from tungsten matrix among given number of hydrogen?
Thank you in advance.
Regards,
Ashok

[Ulrik G. Vej-Hansen]

My point was that your tungsten matrix is probably too thin to absorb all the hydrogen atoms when they are coming in at such a high energy, but this is a stochastic process, so it makes good sense that some are caught anyway. I understood your original question to be why some Hydrogen atoms were not absorbed.

If you do not run it any longer than 108*100=10800 steps, you can avoid the hydrogen atoms being re-used. You could also let it run until there are no more hydrogen atoms in the reservoir and simply count how many have been shot at the tungsten and how many are absorbed.

I see that you use a non-standard potentialset, so if you share it I can make sure that the exact forcefield does not make any difference in the behavior.

[ashokgrg19]

Thank you Ulrik.
Actually I already did the method you described in your email. To see more detail about the deposition, I did the simulation with only 2 Hydrogen atom on reservoir, I ran the simulation making 100 deposition interval with only 1 log interval, making 200 md steps. What I observed in this situation, on next deposition it reuse the same Hydrogen that it already used for the first deposition, not the new hydrogen atom. So, Is there anyway to fix this issue?
I tried to attach my potential file, but its size more than 400KB. This is the link where I found potential for Tungsten, Hydrogen, and Helium interaction.
https://www.ctcms.nist.gov/potentials/W.html, I used these potential files: potential-WHHe-EAM1.eam.alloy , potential-WHHe-EAM2.eam.alloy

Regards,
Ashok
,

[Ulrik G. Vej-Hansen]

But the Hydrogen atoms are identical, apart from the index number, so why should it matter which atom in the reservoir is used? As long as you can count exactly how many attempts you have made, and how many hydrogen atoms have been absorbed, it should give you the ratio between absorption and reflection.

Thanks for the link, I will try that one as well as the one we have built-in and see it changes anything.

Edit: It seems that the potential you are using is actually already included in ATK.