QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Dipankar Saha on October 29, 2015, 17:08

Title: Rigid v/s. Fixed
Post by: Dipankar Saha on October 29, 2015, 17:08

Hi,
I  just  wanted to know__

While assigning constraint to a set of  "Tagged" atoms, we find the options "rigid" and "fixed" ...../ How these two are different ?

Thanks_
Dipankar

Title: Re: Rigid v/s. Fixed
Post by: Umberto Martinez on October 29, 2015, 17:15

Fixed means that atoms are kept frozen during the optimization. Usual constrain.
In ATK 2015 a new constrain method has been added, the RigidBody constrain where all atoms within a rigid body undergo the same translational displacement, which is determined by the force on the center-of-mass of the rigid body atoms.
See more details here:
http://www.quantumwise.com/documents/manuals/latest/ReferenceManual/index.html/ref.rigidbody.html

a good example of the rigid body usage is when relaxing an interface as shown here:
http://quantumwise.com/publications/tutorials/item/878-geometry-optimization-of-interfaces

Title: Re: Rigid v/s. Fixed
Post by: Dipankar Saha on October 29, 2015, 20:22

Ohhh...that's a nice tutorial .... / Let me go through it.....

Anyways, thank you Umberto .....!!!  :)

Title: Re: Rigid v/s. Fixed
Post by: Dipankar Saha on October 30, 2015, 07:12

Dr. Umberto Martinez,

What is the purpose of  switching-off the rotational motions ?? Is this only to save time?

Best_
Dipankar

Title: Re: Rigid v/s. Fixed
Post by: Julian Schneider on October 30, 2015, 10:00

The reason for this is that in order to enable rotational motions of the rigid bodies, we would have to switch from normal Newton-dynamics to Newton-Euler-dynamics, which would require to implement completely different equations of motion. Although this is planned for some time in the future, right now only translational motions are allowed, which can be implemented into our regular dynamics and optimization framework.
Are you considering a system which explicitly requires rotational motions of rigid bodies, as well?

Title: Re: Rigid v/s. Fixed
Post by: Dipankar Saha on October 30, 2015, 15:30

No no..., I was not dealing with any system where rotational motion is explicitly required...!! But, the concept that_ "all atoms within a rigid body undergo the same translational displacement" could be helpful in my case..... (where I want to make an interface  between two planes horizontally/ As well as, I need to capture  the typical nature of the interfacing region without hampering the periodicity of those individual planes far from the interface.../ Perhaps, in that case, I will not be reaching  the global minimum.... )
Anyways, thanks a lot Julian for your reply.... :)

Regards_
Dipankar