Hi,

Let me explain in short. Lets assume you have a molecule A ( e.g Li2Sn molecules in the mentioned paper) which is getting adsorbed on a surface or binder (B) (e.g the PEI in the paper) to generate a final structure, lets say A--B ( Li2Sn+PEI in the paper), then the binding energy (B.E) is basically the energy difference between the optimized structures of initial (before adsorption) and final (after adsorption) state. In this case,

B.E = Energy (A) + Energy (B) -Energy (A--B)

The more positive the value is the stronger will be the adsorption. (Some people do use the expression with reversed sign, in that case your interpretation will be opposite, nothing else will change)

Now, for G09w calculation (I'm including here the instruction on how to do it completely via Gaussview, since it seems that you're familiar with it):

Step 1: Draw structure of molecule A and B in Gaussview. Try your best to draw as good as possible. You can use some tools (e.g: Clean tool with the brush icon) provided within Gview to clean up your drawn structure if needed.

Step 1: Optimize molecule A (Gaussian Calculation Setup >Job Type > Select Opt+ Freq Option in Gaussview) select the desired level of theory (e.g: B3LYP/6-31G(d+p)) from the Method Tab. Add Memory (RAM) and processors as required in the link 0 tab and submit the calculation.

Step 2. Do the same for Molecule B. After both calculation finished open the output (.log) file via Gview. If there was some error, Gview will show you the error while opening. Otherwise go to Results> Vibration and check that all the shown normal modes are positive and not negative, this ensures that your structure is optimized to a local minimum and not to some Saddle point (e.g: Transition State). Do this for both A and B. From the Results > Summary you can note the energy value for A and B.

Step 3. Now draw the expected adsorbed structure in Gview where A is adsorbed on B. Depending on your structure you may have different binding sites, in which case you have to prepare multiple guess geometries to obtain different adsorbed structures). In this step, rather than drawing from scratch you can use your previous optimized structures of A and B which will be better.

Step 4. Optimize this (A--B) geometry as earlier. Check the normal modes again. Now note down the Energy Value of A-B similarly from Results> Summary.

Now you have all the three energy values. Subtract them in accordance to the above equation. The Energy values will be in Hartree, change it to desired unit (kcal/mol or eV etc.) Done.

If you have any further queries, let me know via a message to my inbox.

Hi,

Let me explain in short. Lets assume you have a molecule A ( e.g Li2Sn molecules in the mentioned paper) which is getting adsorbed on a surface or binder (B) (e.g the PEI in the paper) to generate a final structure, lets say A--B ( Li2Sn+PEI in the paper), then the binding energy (B.E) is basically the energy difference between the optimized structures of initial (before adsorption) and final (after adsorption) state. In this case,

B.E = Energy (A) + Energy (B) -Energy (A--B)

The more positive the value is the stronger will be the adsorption. (Some people do use the expression with reversed sign, in that case your interpretation will be opposite, nothing else will change)

Now, for G09w calculation (I'm including here the instruction on how to do it completely via Gaussview, since it seems that you're familiar with it):

Step 1: Draw structure of molecule A and B in Gaussview. Try your best to draw as good as possible. You can use some tools (e.g: Clean tool with the brush icon) provided within Gview to clean up your drawn structure if needed.

Step 1: Optimize molecule A (Gaussian Calculation Setup >Job Type > Select Opt+ Freq Option in Gaussview) select the desired level of theory (e.g: B3LYP/6-31G(d+p)) from the Method Tab. Add Memory (RAM) and processors as required in the link 0 tab and submit the calculation.

Step 2. Do the same for Molecule B. After both calculation finished open the output (.log) file via Gview. If there was some error, Gview will show you the error while opening. Otherwise go to Results> Vibration and check that all the shown normal modes are positive and not negative, this ensures that your structure is optimized to a local minimum and not to some Saddle point (e.g: Transition State). Do this for both A and B. From the Results > Summary you can note the energy value for A and B.

Step 3. Now draw the expected adsorbed structure in Gview where A is adsorbed on B. Depending on your structure you may have different binding sites, in which case you have to prepare multiple guess geometries to obtain different adsorbed structures). In this step, rather than drawing from scratch you can use your previous optimized structures of A and B which will be better.

Step 4. Optimize this (A--B) geometry as earlier. Check the normal modes again. Now note down the Energy Value of A-B similarly from Results> Summary.

Now you have all the three energy values. Subtract them in accordance to the above equation. The Energy values will be in Hartree, change it to desired unit (kcal/mol or eV etc.) Done.

If you have any further queries, let me know via a message to my inbox.

Also, Few related Q/A from RG:

https://www.researchgate.net/post/How_to_calculate_binding_energy_using_Gaussian_09

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

https://www.researchgate.net/post/How_does_one_determine_the_interaction_energy_in_an_A-B_complex_using_DFT_calculation_in_G09

Thanks!!

Turbasu Sengupta , Your answer is best for me.

How can we get band gap.calculation by using gaussian..

Abhijit Dutta The concept of band gap in mostly reserved for bulk solid states (Periodic system) and I'll suggest to use periodic DFT calulation (like with VASP and Quantum Espresso etc) for that. Gaussian do have a Periodic Boundary Condition (PBC) mode through which calculation of periodic systems is possible, however most researcher prefer VASP, QE, Siesta etc over Gaussian for evaluating bulk properties. Tutorials (for PBC) are available online, e.g: http://scuseria.rice.edu/gau/PBC-Guide.pdf

Gaussian, however, can provide you the so called HOMO-LUMO gap for non-periodic systems in case that serves your purpose.

Thanks Turbasu Sengupta .

Each optimisation calculation produces one "Heat of formation" = HF energy, listed few times, last time in archive block of Gaussian. It should not be confused with Hartree-Fock, as long as calculation is DFT.

One run - one energy. To get Binding energy you need at least 3 energies.

Further one can correct each energy to ZPE (from frequency calculation), be aware that ZPE and ZeroPoint energy listed twice, once in Hartree and once in kcal/mol, while Heat of formation is given in Hartrees. Normally this changes result pretty slightly.

Even further one can take termochemistry and enthropy factor in to account.

I don't think you need to optimize the structures of A and B separately. Just extract A and B structures from the optimized structure of A-B complex. And calculate the single point energy of them. Or define the A and B fragments in the complex structure and take use of counterpoise=2 in your input file in Gaussian.

- Optimize the molecule AB with the the keyword opt+frequency for the minimum energy

- Find out the minimum energy A and B separately using the same keyword as above.

- Now subtract the energy as

Binding energy of AB =AB - (A+B)

Hello guys, i read the answer for everyone's input.

My question how do i know if my adsorbed molecule will form a bond with the catalyst surface (i.e whether i am calculating adsorption energy or binding energy (physisorption vs chemisorption). Also when i put my adsorbed/binding molecule to the catalyst's surface in gaussview, how far should i place them?

Hello guys, i read the answer for everyone's input. My question how do i know if my adsorbed molecule will form a bond with the catalyst surface (i.e whether i am calculating adsorption energy or binding energy (physisorption vs chemisorption). Also when i put my adsorbed/binding molecule to the catalyst's surface in gaussview, how far should i place them?

1) DFT and any other quantum/chemical methods do not operate with "bond" concept. Thus, there is conceptually no difference between chemisorption and physisorption. You can claim that it is chemisorption if your visualiser (based on some default atomic radiuses) shows the bond of your interest.

2) how far - if you expect and want to find chemisorbed state - put absorbate on the distance that corresponds to typical bondlength of particular atom pair. These values are normally easy to find in web. For example there are some for carbon. So you could give distance around 2 Angstroms if you adsorb system through carbon or oxygen to some metal surface. If you want or expect to find vdW minimum - take distance over 3.3 A. May be 4.

3) If you are working with periodic systems - Gaussian is really far not the best software for doing it. Its very good software for many applications, but for periodic DFT there are way more appropriate solutions.

How to get required energy value from the opt freq log file? There are many energy values like Sum of electronic and thermal Free Energies, nuclear repulsion energy, etc. Which energy value we should take to calculate binding energy?

Mhejabeen Sayed Binding energy is relative energy, so you need 3 values: moleculeA, moleculeB, complex. These are 3 individual calculations.

Which energy to take - depends on method and purpose. The easiest is just "total energy" without any corrections. For DFT it can be found in the archive block of output after "HF=" (where HF means Heat of Formation) or after the last "SCF Done: E(". Keep in mind that MP2 or CCSD outputs will also contain "HF" in archive block, but there it is "Hartree Fock" and it is NOT the value of interest.

First regularly applied correction is ZPE, which is mentioned in the archive block twice, once as ZPE and once as ZeroPoint, in different units. After that one can put entropy correction on a particular temperature. This also has to be performed manually, based on Total S written in Thermochemistry block in Cal/Mol-Kelvin. Also there one can get data for enthalpy correction to a given temperature, how to do this described in gaussian manual https://gaussian.com/thermo/ , and it is rarely performed to my knowledge. Speaking about corrections to binding energy (but not about which energies to extract from the output) - it might be reasonable to use counterpoise correction. https://gaussian.com/counterpoise/

Dmitry Sharapa: Thank you somuch. I have few more queries, how to calculate HOMO LUMO Energy gap and how to obtain molecular orbital diagram by using gaussian.