Restrain a cluster in z-axis

[Jay Shah]

Hi,

I have made a group with atoms in a cluster, i want to apply bias potential so that it forms a particular angel with the z axis.
What's the best way to do it and is there a way where i can manually make a virtual atom. 

Thanks, 

Jay
 

[Gareth Tribello]

Hello

This is an interesting idea, but it is not entirely clear what you want to do.  You can calculate the angle between two vectors.  From your message, it is clear that one of these vectors is the z-axis.  How precisely are you planning on defining a vector from the group of atoms in the cluster?

If you want to make a virtual atom at a particular fixed point in space you use FIXEDATOM

I hope this helps

Gareth

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/plumed-users/95299f21-51d6-45ab-86ee-09aab2546fe7n%40googlegroups.com.

[Jay Shah]

Thanks Gareth, it helps. 

I am defining the one vector as a principal axis of the cluster and the another vector as z-axis. I am unsure about how to define the bias potential to make that angle zero. 

[Gareth Tribello]

Hi

You can do this with PLUMED.  Here is how:

# Find the gyration tensor of the cluster - in doing the command this way, we are giving a weight of one to each of the atoms

g: GYRATION_TENSOR ATOMS=<atoms>

# Diagonalize the gyration tensor to find the principal eigenvector - the output from this command is a 3-dimensional vector called d.vecs-1

d: DIAGONALIZE ARG=g VECTORS=1

# Calculate the length of the principal eigenvector

v2: CUSTOM ARG=d.vecs-1 FUNC=x*x PERIODIC=NO

mag2: SUM ARG=v2 PERIODIC=NO

mag: CUSTOM ARG=mag2 FUNC=sqrt(x) PERIODIC=NO

# Now get the z-component of the principal eigenvector, as the dot product between the z-axis and the eigenvector is basically the z-component of the eigenvector

zcomp: SELECT_COMPONENTS ARG=d.vecs-1 COMPONENTS=3

# And now calculate the angle

ang: CUSTOM ARG=zcomp,mag FUNC=acos(x/y) PERIODIC=NO    

It is probably not even necessary to calculate the length of the eigenvector as I think that DIAGONALIZE returns normalised eigenvectors.  I would double check that, but if I am right you can just do:

# Find the gyration tensor of the cluster - in doing the command this way, we are giving a weight of one to each of the atoms

g: GYRATION_TENSOR ATOMS=<atoms>

# Diagonalize the gyration tensor to find the principal eigenvector - the output from this command is a 3-dimensional vector called d.vecs-1

d: DIAGONALIZE ARG=g VECTORS=1

# Now get the z-component of the principal eigenvector, as the dot product between the z-axis and the eigenvector is basically the z-component of the eigenvector

zcomp: SELECT_COMPONENTS ARG=d.vecs-1 COMPONENTS=3

# And now calculate the angle

ang: CUSTOM ARG=zcomp FUNC=acos(x) PERIODIC=NO 

Notice, this will only work with v2.10 or a higher version.  I would recommend using the master version of the code, which you can get with the command:

git clone https://github.com/plumed/plumed2.git

You then compile and build plumed in the usual way.

I hope this helps

Gareth

To view this discussion visit https://groups.google.com/d/msgid/plumed-users/89448359-b7f4-4a5a-9df1-183b91553920n%40googlegroups.com.

[Jay Shah]

This is really helpful. Thanks a lot Gareth !!
Best, 

Jay