Dear Keshav,

You should run single point calculations for all optimized structures (GM, TS and P). An example of the line to be included in the input file is:

# B3LYP/6-311++G(d,p) SP SCF=Tight SCRF=(iefPCM,Read,Solvent=Water).

The meaning of the above mentioned line:

This keyword requests that a calculation be performed in the presence of a solvent (for example water) by placing the solute in a cavity within the solvent reaction field.

The Polarizable Continuum Model (PCM) using the integral equation formalism variant (IEFPCM) is the default SCRF method. This method creates the solute cavity via a set of overlapping spheres. It was initially devised by Tomasi and coworkers and Pascual-Ahuir and coworkers.

http://www.gaussian.com/g_tech/g_ur/k_scrf.htm

Dear Keshav,

You should be aware that your minima and transition state structures obtained in the gas phase will not be the corresponding structures within your PCM model. Although structural differences are typically very small, they will have an effect on energies.

You mention that you have included four explicit water molecules in your gas phase calculations and so you should also be careful that you are not modelling part of the immediate coordination environment explicitly and part implicitly.

Finally, if you are interested in thermochemical properties, then bear in mind that any calculated vibrational frequencies will be invalid if you move away from the bottom of your potential energy surface (as you will do if you perform SPE calculations with PCM at the gas phase geometry).

I hope that this helps,

Andy

Like Andrew suggested the minima corresponding to gas-phase and PCM model (and any other solution model) are definitely not the same. If you're a purist you should definitely redo all the calculations using PCM model to get new minima (checked by frequency calculations) and treat the gas-phase calculations just as a good starting point for PCM ones.

On the other hand there are many researchers (inlcuding me) that just do single-point calculations in solvent to get the energy (or free energy) of solvation. The problem is that implicit-solvent methods tend to be more computationally demanding (but just barely), have much worse convergence (like a lot) and often end the geometry optimization in a wrong stationary point. For large systems (~100 atoms) this can be particularly difficult. So we claim that the gas-phase minimum is close to the PCM minimum (which is also a model and not a great one) and do single-point calculations.

This is usually OK, for many systems the results are in good agreement with the experimental data (perhaps due to error cancellation, but still). The reviewers oare also usually OK with it as long as the system is large.

In your case, though, the system is kind of small, so redoing optimization in PCM shouldn't be that time-consuming. Also, as Andrew said you already have some solvent molecules explicitely in your system so you should be careful with that.

Thank you Rafik Karaman, Andrew Kerridge and Bartosz Trzaskowski for your valuable suggestions.

Dear Keshav,

be aware that PCMs were developed to simulate liquid solution, but not ice at close to zero K. These situations differ dramatically, since the equilibration of the solvent molecules one finds in liquid solution that is implicitly included in the PCM description is absent in frozen water. Hence, it is very suggestive to include a large number (at least the first "shell") of explicit water molecules and in addition use a PCM for the electrostatics of bulk ice. However, I'm not sure how well a PCM works for bulk ice in space. It should at least be considered that the central PCM parameter (dielectric constant, epsilon) exhibits a strong temperature dependency (increases with decreasing temperature).

Finally, let me point out one issue apart from your question that has not been discussed so far: Your method is a poor choice for the kind of problem you are treating. You should at least include some type of dispersion correction, even if you are aiming for a qualitatively correct picture. Furthermore, for such a small system, why not use ab-initio methods like SCS-MP2 instead of the poor B3LYP?

In the attached work by Grimme, you can find an extensive benchmark of DFT and MP2, which includes systems very similar to yours (ice clusters).

Cheers,

Jan

http://pubs.rsc.org/en/Content/ArticleLanding/2011/CP/c0cp02984j#!divAbstract

Thank you Jan-Michael Mewes. I am planning to include dispersion correction in calculations. But I am not sure which dispersion corrected functional should be use. The gaussian 09 version i am using only has WB97XD. it did not show any other functional like B3LYP-D3,MO6-2X in calculation setup drop-down menu.I read many literature which use DFT for reaction modelling that's why I use DFT.(for example the work by David E Woon University of Illinois, Urbana-Champaign). For PCM model , I will use dielectric constant=78.5 equal to that of liquid water as i did not find any experimental or theoretical work which says anything about dielectric constant of water ice at cryogenic temperature Also thank you for the paper . your answer and the paper both are quite illuminating.

The concept of the dielectric constant is usually used for liquid systems. Nevertheless, one can measure epsilon also for solids like ice. Yet, I dont know how PCMs work in such a case.

For water at 0°C, epsilon is rather 90 than 80 (first picture here https://en.wikipedia.org/wiki/Relative_permittivity). In this work (http://pubs.acs.org/doi/pdf/10.1021/jp105975c) epsilon of water/ice is measured down to a temperature of -120°C (Figure 5). However, if you take 78.5, 80.4 or even 180 doesn't really matter too much for the results, just be consistent and more importantly, pay attention to the explicit "solvent".

Gaussian 09 supports a large number of functionals, just write the input file yourself instead of using a GUI. You can find them here: http://www.gaussian.com/g_tech/g_ur/k_dft.htm. omega-B97XD is already a much better choice, which is specifically mentioned in the paper I cited. That's maybe the reason why its in your drop down menu :)

Regards,

Jan

Thank you Jan-Michael Mewes. Your answer is very halpful

Dear all,

 I have performed a QM calculation with M062X/6-31++G(d,p) level of theory in gas phase. Further, I have carried out a single point calculation in aqueous phase with SMD solvation model. I have performed frequency calculation in gas phase as well as in aqueous phase. I have calculated the free energy differences for gas phase and aqueous phase (∆G=GAB-GA-GB). I have a doubt whether the free obtained from the single point frequency calculation in aqueous phase is correct/reliable free energy or not? Please suggest me.

Thanks in advance,

Kalyanashis Jana