Different position of the dihedral angle or different dihedral angle will lead to different configuration of the same molecule, each configration will have different energy (Potential energy- which acquired due to the relative position of each atom in the molecule). Hence, the values will not be same for all dihedral angle. Among these dihedral angle, just see which dihedral angle produces lowest minimum energy- that is the dihedral angle we should take into account - only that will give the most stable configuration.

Dear Sir,

Thank you for your kind reply.

Whatever the pair of dihedrals used for scanning the PES I would suggest a relaxed PES calculation. The computational time will increase but this way you can avoid locating an erroneus global minimum. You might use a semiempirical method (e.g PM6) for this purpose.

I agree with both responses already given by Sengeny and Vasile.  In particular, Vasile makes a very good point in stating that you should try a relaxed PES scan where you allow the other coordinates (bonds and bond angles) to optimize as you scan the rotational coordinates.  Some deformation of the bonds and bond angles may have to occur to find the true minimum energy path (MEP).  Also doing this scan initially with a semi-empirical method to first find an MEP and then re-using the coordinates points of the scan with an optimization with higher level method is something that can also could be done to see if the true MEP has been found.

Fully agree with Jamin. Re-optimizing the semiempirically obtained minima at a higher level of theory is compulsory for getting a reliable energetics of them.

Dear vasile chis and jamil perry,

Thank you for your reply. I have got come idea from your answers. Generally what I do is I will write input file and do my pes scan by selecting particular dihedrals. Kindly tell me how to do relaxed PES scans.

A similar thread can be found on ResearchGate at

https://www.researchgate.net/post/dihedral_scan_in_Gaussian

Hope it helps!

Dear  young friend!

  You have not mentioned any details on your software and keywords used. My experience is that PES scan procedures usually use SCF restarts from previous geometries (with finished SCF). This causes serious problems especially by changing the symmetry group. If you use various SCF restarts you might obtain various results (e.g. energies), especially in the case of poor SCF convergency. This is most probably your problem.  Thus I propose to switch-off the SCF restart by PES scan calculations.

  Best wishes

                      mb

Dear martin,

The keywords I used is 

# b3lyp/6-31g** scan

This is what am using for scan after selecting the required as 0.0 36 10.0.  Kindly give me some suggestions that I would like to make in my keyword section.

For a relaxed PES at PM6 level the route section would be:

# PM6 Opt=modredundant

After the geometry specification leave a blank line and then input the dihedrals to be scanned like in the following example:

D * A1 A2 *

D * A3 A4 *

D An A1 A2 Am S 23 15.0

D Aj A3 A4 Ak S 23 15.0

Here, the two dihedrals defined by the An A1 A2 Am (A1 A2 bond is contained in both planes) and Aj A3 A4 Ak ( A3 A4 bond is contained in both planes) atoms will be scanned. All the other geometrical parameter will be optimized for each pair of the two scanned dihedrals.

A note of caution: this style is OK for G09. For another version of Gaussian the input could be slightly changed.

Good luck!

Dear young friend!

  I suppose that you use Gaussian software. In order to prevent the SCF restart from previous calculations in PES scan you must add the keyword

     Guess=Always

 Best wishes

                   mb

Dear vascile chis,

I have tried your option and the result is I have got optimized structure instead of scan results. Whats the problem with it. I hAve selected the two dihedrals using modredundant option and has given scan. My input line is

# B3lyp/6-31g** opt=modredundant test.

Kindly clarify me or send me any input output file fully writen for this type of scan as an example for me.

Which version of Gaussian are you using?

Did you add at the end of geometry specification (after a blank line) the scanned dihedrals?

If you did so something is still wrong with your job. Please give your input file.

Dear vasila chis,

Kindly give me your mail id. I will send u your file.