QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: mantun on May 8, 2014, 04:41

Title: complex band structure
Post by: mantun on May 8, 2014, 04:41
I'm a user of ATK13.8. After i compute the complex band structure, i only care about the most relevant band in  the imaginary part (that is the line link bottom of CB and top of VB). How can i extrat the data i need?
Title: Re: complex band structure
Post by: kstokbro on May 9, 2014, 13:43
check
http://www.quantumwise.com/documents/tutorials/latest/ComplexBandstructure/index.html/chap.sibulk.html#sect1.sibulk.analyze
there is a script with an if statement,
            # Discard rapidly decaying modes which clutter the plot
            if ki>-0.3:
                e += [energy]
                kvr += [kr]
                kvi += [ki]
change the cutoff value, to select different set
Title: Re: complex band structure
Post by: mantun on May 10, 2014, 09:59
check
http://www.quantumwise.com/documents/tutorials/latest/ComplexBandstructure/index.html/chap.sibulk.html#sect1.sibulk.analyze
there is a script with an if statement,
            # Discard rapidly decaying modes which clutter the plot
            if ki>-0.3:
                e += [energy]
                kvr += [kr]
                kvi += [ki]
change the cutoff value, to select different set


I understand what you mean, that is change the '-0.3' in the if statement. But in my situation, i want to extract the most slowly decaying modes and plot it by myself. Because i have three different system, i want to compare their slowly decaying modes in a picture. Thank you in advance!
Title: Re: complex band structure
Post by: Anders Blom on May 10, 2014, 19:07
You can just take the minimal value, actually the largest, so max[kvi] then, after the loop.
Title: Re: complex band structure
Post by: mantun on May 12, 2014, 03:06
You can just take the minimal value, actually the largest, so max[kvi] then, after the loop.
Can you tell me how to do with the scripts in detail. I don't know python very well. Thanks a lot!
Title: Re: complex band structure
Post by: Anders Blom on May 12, 2014, 09:23
it also depends a lot on whether there are other lines crossing the one you are interested in or not, and e.g. which "k real" value the lines have. Making a programming solution yourself for this would be straightforward, but if you are unable to write the code, it's very difficult for someone else to write a general code that works in all cases.

Take the Si example in the tutorial (Fig. 3) - if we just want the line connecting the CB min at G and the VB max at G (the red line), we have to somehow exclude the line which originates at the Delta minimum (blue). Besides, in the code we are talking about (referenced above), the goal is to make a 3D plot - I'm not sure want that, you just want the data points. So the code has to be modified to print out the data in a format you can import elsewhere. All this is easy if you have control of the data and code, but very difficult to make generic.

You may be better off just exporting all the data, and picking out the points by hand.
Title: Re: complex band structure
Post by: mantun on May 13, 2014, 09:04
it also depends a lot on whether there are other lines crossing the one you are interested in or not, and e.g. which "k real" value the lines have. Making a programming solution yourself for this would be straightforward, but if you are unable to write the code, it's very difficult for someone else to write a general code that works in all cases.

Take the Si example in the tutorial (Fig. 3) - if we just want the line connecting the CB min at G and the VB max at G (the red line), we have to somehow exclude the line which originates at the Delta minimum (blue). Besides, in the code we are talking about (referenced above), the goal is to make a 3D plot - I'm not sure want that, you just want the data points. So the code has to be modified to print out the data in a format you can import elsewhere. All this is easy if you have control of the data and code, but very difficult to make generic.

You may be better off just exporting all the data, and picking out the points by hand.

I get what you mean. I have to extrat the data and pick by hand. Thank you all the same
Title: Re: complex band structure
Post by: Anders Blom on May 13, 2014, 10:31
In this case, it's probably just as fast as writing code for it.
Title: Re: complex band structure
Post by: mantun on May 13, 2014, 14:06
In this case, it's probably just as fast as writing code for it.
Yeah, i can't agree any more! I have made it today by hand.
Title: Re: complex band structure
Post by: GJK on May 13, 2014, 14:17
Hi,

What is meant by complex band structure.Kindly provide me with theory(link) and its significance.

When and where to use?

Title: Re: complex band structure
Post by: Anders Blom on May 13, 2014, 23:42
A well written reference on this topic is Journal of Magnetism and Magnetic Materials 240 (2002) 108–113.
You can also find a lot of good information by using Google ;)