Dear administor:
Recently,I have used mcsqs to calculated the BCC and FCC construction about Fe-Cr.
In the FCC, Fe=0.8, Cr=0.2,the mcsqs -n, n determined by 0.8*0.2=16/100, as FCC has 6 pairs,100/6 is not a integer, so 16/100=1/25=6/150.A a result value of n is 150/6=25.In this process, I got " Objective_function= Perfect_match" in best correlation.
However,In the BCC,Fe=0.9,Cr=0.1,using the same method to calculate the n is 25. But an error occured to me,"Impossible to match point correlations due to incompatible supercell size.
".I know the n must be changed.so I change the n to 10,20,80.But as an unknewn reason,the mcsqs dose not update for a day.
Finally, whether the method determined the -n is right or not? Should I wait for the calclulation finished by itself?
other:it’s my command below
corrdump -l=rndstr.in -ro -noe -nop -clus -2=5
mcsqs -n=*( * is the value 25,10,20 or 80)
Some versions of ATAT had a bug with bcc supercells (the code hung while trying to find the most symmetric way to built the cell). This is now fixed, please download the "Beta version of the latest update".
Your calculation of the number of atoms needed is not quite right. For bcc, the nb of nn is 8, so you have 4 pairs per cell and you want a composition with 1/10, to cell would have to be a multiple of 20 atoms (lcm of 4 and 10).
Think you for your attention.Your reply helps me a lot.
But I have another question.For the FCC,using the method displayed above, nn = 6,and Cr composition is 1/5. So It should be 30,but using n=30 or 60, "Objective_function= Perfect_match" didn’t appear and the code still hung.My versions are 3.44 and 3.36.
Sencond question, nn value is determined by the number of nearest neighbor atom, however ,how to get the value 1/10 above? Is it determined by the minimum concentration or the denominator of the simplest fraction? Giving examples,for Fe=0.8,Fe=0.15,Ni=0.05.Should it be 20?Fe=0.8,Cr=0.11,Ni=0.09,should it be 100?Fe=0.8,Cr=0.11,Ni=0.05,Co=0.04,Should it be 100?
Looking forward to your reply.
First, I just want to clarify that it is non necessary to obtain the "perfect match" message. In fact, most of the time, you would want to include more clusters and let the code try to find the best possible solution even if a perfect match is not possible. In this case, the code is in an infinite loop on purpose and you just stop it when the best solution so far seems adequate for your purpose… or has not changed in a few hours. If the code has gone to the point of at least generating one possible sqs, it has not hung.
The correct rule for the minimum size depends on the smallest common denominator of all the concentrations.
Fe=0.8,Fe=0.15,Ni=0.05 would indeed be 20.
Fe=0.8,Cr=0.11,Ni=0.09,would indeed be 100.
Fe=0.8,Cr=0.11,Ni=0.05,Co=0.04, would indeed be 100
This is so that the code does not stop immediately because the point correlation can’t be matched.
To at least match a nn pair correlation, then you need to make sure the nb of nn pair per cell is a mutiple of the above numbers.
Thank you! Thank you! Think you! I can go on my calculation with your guidance.