Computational Chemistry - Science topic

B3LYP (Becke, 3-parameter, Lee-Yang-Parr) is a popular hybrid functional that combines the local density approximation (LDA) with the gradient-corrected exchange and correlation functionals. It works well for a wide range of systems but can struggle with predicting properties related to charge transfer and long-range interactions. Also, B3LYP, while widely used and successful in many cases, does not explicitly include dispersion corrections (also known as van der Waals forces).

CAM-B3LYP, on the other hand, includes a correction for long-range interactions. The "CAM" stands for "Counterpoise Corrected for Adsorption Model," and it's particularly useful when dealing with molecules that involve noncovalent interactions, like dispersion forces. This correction accounts for the overestimation of interaction energies that can occur when the basis sets of two interacting fragments overlap.

In a nutshell, if you're working with systems where intermolecular interactions play a significant role, CAM-B3LYP might be the way to go. Otherwise, good old B3LYP should serve you well.

Hello,

By using the NBO 3.1 program, you need to add the keyword "alpha" even if you're studying a closed-shell system. As in the following example:

$nbo

$end

$CHOOSE

alpha

BOND S 1 2

D 1 2

S 1 6

S 1 7

S 2 3

S 2 8

S 3 4

D 3 4

S 3 9

S 4 5

S 4 10

S 5 6

D 5 6

S 5 11

S 6 12 END

$END

My best regards

Anusree Nagarajan Here are some further resources (books) for a more in-depth study of the concept:

1) Density Functional Theory: A Practical Introduction by David Sholl and Janice A. Steckel

2) Computational Materials Science: An Introduction by June Gunn Lee

3) Electronic Structure: Basic Theory and Practical Methods by Richard M. Martin

Hopefully, this will facilitate your detailed comprehension of the topic.

Vacuum padding is a method of creating a vacuum or reducing air pressure inside a container or packaging. It's commonly used to extend the shelf life of food, protect delicate items, save space, and shape materials in various industries. The process involves removing air and sealing the container to maintain reduced pressure.

You can run simultaneously both relax-scan and td-dft calculations.

First, you need to determine the thermodynamic cycle.

You should optimize the complex with both reactants in a favorable position / orientation for the possible proton transfer to occur, with an appropriate level of theory and basis set. The final position of the hydrogen could provide an answer to your problem: if it stays on the acid it is H-bonding, and if it is transferred to the basis it is an acid-base reaction.

your results can be more accurate if you take the solvent into account (at least as a polarizable continuum) and if you compensate for basis set superposition error (either using a fairly large basis set or using the counterpoise scheme).

Broken symmetry DFT calculations in Gd-dimer if both the Gd are in +3 oxidation state is absolutely fine since it has isotropic spin ground state, that is octet and orbital singlet. There are many references as for now I can recommend https://pubs.rsc.org/en/content/articlehtml/2021/dt/d1dt01770e

You could tell them that MD concerns itself with simulating how atoms and molecules move, while DFT concerns itself with the energy and properties that electrons determine for fixed atomic positions. Avoid terms like "dynamics", "functional" or "electronic structure". These might be too complicated for elementary school children. Use "motion", "energy" even "forces" instead.

Dear Kanski

Would you please explain about parameters setting in your simulations? for example rc and r in a protein-ligand simulation?

Adam Matěj Thank you for your response- The project I am working on is for students to better visualize the quantitative calculations by comparing it to qualitative calculations taught in inorganic chemistry courses.

I am comparing qualitative results calculated from group theory to quantitative results from a computational calculation for a planar D5h symmetric molecule.

In the qualitative calculation, the py basis functions were oriented pointing towards the z axis to emphasize the symmetry. I would like to reflect that in the quantitative calculation so I can make a cleaner comparison.

I am using AOMix to print the coefficients from the computational calculation, however at the moment since the basis functions are not oriented as desired, I must do some trigonometry which is rather tedious. I also use the visualization software iqmol to visualize the basis functions, so having them oriented to emphasize symmetry would be useful.

As of my knowledge cutoff in September 2021, the ATB (Automated Topology Builder) server provided force field files for the Gromacs G54A7FF force field, which is an updated version of the Gromos 53a6 force field. The G54A7FF force field includes improvements and refinements over the original Gromos 53a6 force field.

If you specifically need the Gromos 53a6 force field files for your simulations, you may face challenges obtaining them directly from the ATB server since it primarily provides the G54A7FF force field files.

However, there are a few potential options to consider:

Please note that the availability of force field files on the ATB server might change over time, so it's essential to check the latest information and options available on the ATB website or contact their support team for the most up-to-date details.

Dear Max, you may try Constrained DFT in the case of charge localized in one fragment. Let's say we have complex A / A, with C-DFT you can optimize A / A+.

On the other hand, charge separation is different, let's say A+ / A-

In this case, I used to move a molecular orbital from one fragment to the other one, thus creating a hole (A+) in one fragment and an electron in the other (A-).

Hope this is useful.

Best, Pablo

Hi,

You can find all the related information in the log file.

In a successful optimization the following command will print the line:

grep "Otimization completed" log-file

If the optimized structure is a stationary one, then the following command will print that information:

grep "Stationary point found" log-file

After the frequency calculation the thermodynamic data is printed in the log file following the "Thermochemistry" line. Search for "Zero-point correction".

To get the optimized structure, you can use GAUSSIAN newzmat command:

newzmat -ichk -opdb -N 9999 chk-file out-file

Dear all, have you solved the problem? I have the same problem as Harry, could you please clarify how you solve this problem? Thanks much!

Aside from the already mentioned usage for computational chemistry, there is also potential for all experimental fields, but there it is crucial that the experimentators

While I'm not using AI on my data right now, I am eager to keep my data in an "AI-ready" structure so that when we decide to have a bot crawl through our data at some point, maybe a simpler AI of the year 2024 will already find something useful and we won't have to wait for a "careless operator handling" AI of the year 2030.

I guess you are asking for something practical, here an excellent guide:

J. B. Foresman and Æ Frisch, Exploring Chemistry with Electronic Structure Methods, 3rd ed., Gaussian, Inc.: Wallingford, CT, 2015.

Also, don't forget YouTube Channels by ADF:

or Schrodinger:

Or theory plus practice:

Essentials of Computational Chemistry. Theories and Models. Second Edition. Christopher J. Cramer.

See also Cramer's channel on YouTube:

or 'must read' theory:

Modern quantum chemistry introduction to advanced electronic structure theory /. Attila Szabo

or more fundamental courses in quantum mechanics: those given free on YouTube by 'Standford Online' or 'MIT OpenCourseWare'.

And finally, if you don't know where to start, you may ask ChatGPT for something you are interested to learn, we are already in the age of AI.

Best, Pablo

Dear Abdulrahman, you probably need to calculate the transition states for the electrophile attack on one carbon and on the other one. TS will tell to you which carbon reacts firts (kinetically faster). Other properties may help, as you mentioned, HOMO or atomic charges, but those values may only tell to you that the reaction is possible in both carbon atoms.

The advantage of vasp.xml is that everything is tagged and designed to be machine-readable. ASE will let you read the OUTCAR or the vasprun.xml. I think pymatgen will read either but can get a lot more info from the vasprun.xml. Looks like vaspkit can read both but I'm not sure what information it extracts from each. I'm a bit surprised that your library is very slow at reading the vasprun.xml--in my experience it usually takes less than 30 seconds.

Anyways, if you prefer to read from OUTCAR or OSZICAR, that should be fine. There are already libraries that do this, but it's easy to write your own function, as you say. Just make sure you're reading the version of the energy that you're interested in.

Sorry, actually never heard about it.

Conrad,

Thank you very much for your answer! The information was in the dftb_pin.hsd restart file, just like you said.

-Joshua

Calculating the Fermi energy level and work function for ZIF-8 MOF in Materials Studio requires the following steps:

1. Open the ZIF-8 MOF structure in Materials Studio.

2. Perform a geometry optimization to ensure that the structure is relaxed to its minimum energy configuration.

3. Set up a DFT (Density Functional Theory) calculation using the Materials Studio CASTEP module.

4. Define the k-point sampling scheme for the Brillouin zone. A gamma-point sampling scheme (1x1x1) should be sufficient for the ZIF-8 MOF.

5. Select a suitable exchange-correlation functional. Commonly used functionals include PBE, PW91, and BLYP.

6. Specify the basis set. A plane-wave basis set with a cutoff energy of 500 eV should be sufficient.

7. Set up a SCF (Self-Consistent Field) calculation to obtain the electronic ground state.

8. After the SCF calculation is complete, view the electronic density of states (DOS) using the Materials Studio VNL module.

9. Identify the Fermi energy level from the DOS plot, which is the energy level at which there is a gap between the filled and unfilled states.

10. Calculate the work function by subtracting the Fermi energy level from the vacuum level. The vacuum level can be obtained from the DOS plot by identifying the energy level at which there is a significant drop in the electronic density.

Note that these calculations are computationally intensive and may require access to high-performance computing resources. Additionally, the accuracy of the results can be influenced by the choice of exchange-correlation functional, k-point sampling, and basis set.

Thank you all

As suggested by Alexey, I think the Stoichiometry of reactants could be an issue. Let me check.

Dear Dipta Mahanta,

But you cannot do this:

"nbnd: Used 100 (in bands) instead of 544 (which is default) for faster calculation and for testing purpose"

nbnd is the number of bands to be calculated; you might want to increase it, not decrease it. Indeed, it corresponds to the number of states you make available to the electrons in your materials. QE will start allocating the electrons to each band and then probably crash because there are not enough.

Indeed, in your output, you can read

"number of electrons = 640.00

number of Kohn-Sham states= 100"

How can you fit 640 electron in 100 states? The code is going to crash somewhere.

Moreover,

"ecutrho: 400 (scf) to 700 (bands)"

"ecutwfc: 50 (scf) to 100 (bands)"

These two changes do not make much sense. Remember that the quality of the SCF calculation is in the SCF step and parameters. The calculations="bands" is a NON-SCF type of calculation -- it starts by reading the SCF output and builds from there interpolating the missing points. For example, the density, the central piece of information in DFT, in the non-SCF calculation is not changed.

My advise for a band calculation is to copy the SCF input file and then simply change calculation="bands" and the k-point mesh to what you need.

Finally, why do you want to run a band calculation with a supercell? The result ought to be identical to the case with a simpler cell, once you have unfolded the bands. In this respect, the band calculation offers no new insight into the physics of the supercell.

Best regards,

Roberto

In general, the program failed to localize the TS. I don't use qst3 or qst2, it is much more input preparation to do and it does not have any advantage over opt=ts when you have a good approximation to the TS.

I suggest, what I always do, first to scan the PES:

#P DFT/basis Opt=ModRedundant

0 1

xyz

B 7 21 S 12 0.10

In this example, the bond "B" betwen atoms "7" and "21" is going to be scanned "S" through "12" steps of "0.10" angstroms step size.

Only if you obtain a smooth curve of the scan you'll know you are in the correct pathway, in that case select the structure at the maximum point of the curve and do a TS calculation directly:

#P DFT/basis Opt=(ts,noeigentest,calcfc) freq

If you don't see a smooth curve, then finding the TS becomes readly challenging, or you should look for another way to do the scan PES (other bonds, angles or even dihedrals).

Best, Pablo

Alexey N. Masliy thank you for this recommendation I already tried this and my cluster have already converge thank you so much.

Just an addition to Ismail Badran answer, sometimes geometry optimisation doesn't end in a minimum but in saddle point of 1st order, where the gradient vanishes too. The more complex the structure, the more likely this may happen.

To get a clearer information, which group might be affected, have a look at the normal modes/vibrations and then follow the normal mode.

If you don't have to stick with gaussion, try using the orca quantum chemical software, which has this outlined approach implemented.

And as Ismail Badran mentioned, try use some larger basis set and include dispersion correction. Van-der-Waals forces should not be neglected even for small systems.

Historically, Pople and crew were most interested in biochemical systems: HCNO mostly. Metals are poorly represented in 3-21 or 6-31 bases. More recently, people have become more interested in metal enzymes and the like and a significant effort went into the development of bases like the def2 set, that can support heavier metals efficiently. So, it's not uncommon to see people mix and match, using 6-311+g* for HCNO and some def2 triple zeta basis for metal atoms that are not included in the Pople suite. You can always customize your basis choices, just be ready to defend any particularly odd choices.

Hello Shivam Kansara , really sorry for the late reply. We had restarted the calculations, however as it was large calculation, it shall take few more days to complete.

Dear Anandhu Gopan,

The easiest way to compute a pKa value in Gaussian is to use the isodesmic method. This allows you to avoid any complex thermodynamic cycles. In this method you use a reference with known pKa (e.g. formic acid) to compute the pKa of interest. Please be also aware that the implicit solvation models implemented in Gaussian are far from perfect. Fortunately it fails in a predictable manner so you can scale the obtained raw pKa value in order to obtain a reasonable guess. Using this method you can get pKa values with an accuracy of roughly 1 pKa unit with minor computational costs.

A detailed description of the method, its accuracy and suitable scaling relations can be found here:

Best wishes

Michael

The question doesn't make sense. The Bose-Einstein distribution is an equilibrium distribution-since it refers to a temperature, which doesn't make sense when things depend on time.

What it might mean is just a trivial application of the formula for the average occupation number, that depends on two variables.

For MD simulations and other compute-intensive tasks, it is helpful to have a CPU with a relatively high core/thread count and clock speed (frequency). Currently, one of the best CPU chips for high-end workstations is the AMD Threadripper Pro 5975wx, which has 32 cores/64 threads, a base frequency of 3.6GHz, and a turbo-boost frequency of 4.5GHz. This should be paired with 1-4 graphics cards capable of substantial acceleration via GPU computing (either OpenCL or CUDA). Among the best GPU accelerator cards are the Nvidia RTX 4090 and Nvidia RTX A6000 or RTX 6000 Ada. Recommended CPU memory is 64 - 512GB 3200 MHz RAM. The boot disk should be an M.2 1-2TB SSD. For fast read-write speeds of data, there are data-center class SSDs in the 2-16TB range with M.2 or U.2 interfaces. Longer-term data storage could consist of a NAS with spinning HDDs. All of this would not be cheap -- the computer would cost upwards of 20k USD and a few k more for the external NAS storage unit.

in GMX_MMPSA its recommended to use PB model with charmm force field. since you prepared your system using charmm36 you should calculate the binding free energy by MM-PBSA and set the PBRadii=7.

Mohammad Mirghotbi

So, you check your system core covers the software. If protein has large residues so have to increase your core.

When you use PCM, the point group symmetry is switched off internally.

This is limitation of PCM in GAMESS.

Did you get some help regarding your query? I also need to create CG file for RNA.

No idea

OpenMM sounds optimal for that, you can just load a pdb define a forcefield and do a minimization. Afterwardas the system has an energy. Basically the quickstart code that they have listed here (bottom of the page):

I hope this is what you were looking for.

Best regards,

Ben

thank you Corentin Lefebvre ! will check the free software out.

Bojidarka B. Ivanova thanks! i have already budgeted for the gaussian software.

I respectfully disagree. You must know what the reactants and products are (regardless of how many steps may be involved in between). The enthalpy of reaction is based on THAT enthalpy difference. For example, the Fischer-Tropsch or Water Gas Shift reactions involve many catalytic steps. However, the enthalpy of reaction refers to the starting materials and the final product of interest. We don't need to know anything about the intermediate steps.

Hi Priyanka Garg

Did you get how to calculate the d-band center? I am also looking for the same. Any help will be highly appreciated.

Thanks in advance.

So you want to investigate

2H₂O -> OH^-+H₃O⁺

This will of course only happen at elevated temperatures. If you do plain DFT, you will get 0K physics at which water is as frozen as frozen can be, so you will have to perform molecular dynamics. If you have the computational power to do DFT-MD you can do that, of course; if that's out of specs for you, you could start with a semiempiric method. The larger you choose your cell, the more complex moieties you will observe probably.

Using GROMACS to perform molecular dynamics simulation is a very good choice. GROMACS is quite fast, flexible and meantime freely available.

Manuals: AMBER, Gromacs, LAMMPS.

Here is a link on that

Isothermal compressibility defined in gromacs is not the fluid property, its an indirect way of putting the fictitious mass for the P-R barostat. While its recommended to use the compre value close to the compre of the fluid, its not a necessary requirement. If you are obtaining the right density and other thermodynamics correctly with the default (say water compressibility), its okay to go with it further. All you need is to make sure that the given parameters describe the physics and thermodynamics of the fluid correctly.

Check the ram and disk space. Its not an error but shortage of either (RAM or storage).

Just as Jurgen Weippert stated, one of the main application of Wiener index is that it have been proven to be correlated.

The other point is, because it is based on graph theory, it can be used to distinguished between isomers, which is pretty important.

The main application is chemistry though, is in QSAR as a molecular descriptor.

Thinh Quoc Le Truong Hi

I hope you doing well

This error is showing in .dpf file. So you should generate again .dpf file or check notepad file of .dpf and change the terms which will be shown in attach pic.

Hi Nikolaos Kourkoumelis, thanks for your answer. Indeed, unfortunately I didn't find the code

Dear Haris Saeed ,

Concerning computational chemistry it depends on what you want to study: chemical reactivity/properties, solids etc ...

I could suggest you at first to have an access to a QM software. Orca, NWChem, Psi4, Gamess are completely free (I have a preference for Orca personnally). Many good tutorials are available online and even on Youtube (transition state research, IR/Raman calculations, excited-states, ESP scan etc ... I found all of them detailed on videos).

Another Software is Gaussian but unfortunately not free.

For visualisation, Molden (Gmolden), Avogadro, VMD, IBOView are completely free. You could also have a look on Multiwfn which is a wonderful software and whose the documentation is very pedagogical!

Of course, a theoretical background will help you, there are plenty excellent books dealing with quantum chemistry:

- Quantum Chemistry by Andrew Cooksy (not so much DFT inside nevertheless).

- A Chemist's guide to density functional theory by W. Koch and M. C. Holthausen.

- Essential of computational chemistry by Cramer.

- Molecular Modelling, principles and applications by Andrew R. Leach (this book talks also about molecular mechanics if you are interested in).

Concerning DFT, online, you could have a look of DFT poll for having an idea of which functional you could use, depending on the study you want to carry out. The choice of the basis-set can be guided by the Basis-set Exchange website.

I hope these little information would help you!

hi Gautam Bhardwaj, thanks for yout answer. I used qsar models too for the prediction of activity, using fingerprints and molecular descriptors. But the idea here is really to have the fragments, the AA they are interacting and the type of interaction. The methods of fragmentation I know are only smarts-based or rules-based, and I'm looking for a structure-based methods

What is your dcd, out , pressure frequency (steps) in the nmd file?

If you increase the frequency than your file size will decrese.

While it is not confirmed by Gaussian, and is not in their internal manual either: I have confirmed via test structures and cross checking of different values that this set of integers is indeed a matrix of atom indices in sets of 4, indicating the atoms involved with the degree of freedom, so the example I showed earlier becomes:

(1 2 0 0) bond length between atoms 1 and 2

(1 3 0 0) bond length between atoms 1 and 3

...

(2 1 3 0) bond angle between atoms 2, 1, and 3

(2 1 4 0) bond angle between atoms 2, 1, and 4

...

(4 1 5 6) dihedral angle between atoms 4, 1, 5, and 6

with 0 as null values indicating no atom there.

Hope this helps someone in the future!

Hello Sean Nations Could you please share in detail ? I am facing same issue. I am using following band.conf

ATOM_NAME = Si O #Your system

CELL_FILENAME = POSCAR-Unit

DIM = 2 2 2 #SUPERCELL size in step 3

BAND = 0.0 0.0 0.0 0.5 0.0 0.0 0.333333333333 0.333333333333 0.0 0.0 0.0 0.0 #full symmetry point

BAND_LABELS = G M K G

FORCE_CONSTANTS = READ

Good Morning dr, I would like to ask can we apply electrical field parameters in gaussian 09 or the gaussian 16 software

Hi,

in theory, you can't. In practice, if you know which residue to modify, you can add all missing parameters to the topology if 1) no gaps exist in atom numbering, and 2) you will have to fine-tune partial charges to address changes.In your case is easy: change O ->N, and add 1 hydrogen (with correspondent partial charge bonds and angles, probably you will not need dihedrals in this case). Herein, to avoid gaps/renumbering the added hydrogen will be the last atom in your topology (while keeping the residue number to which it belongs). You can take those informations 1) from another GLN residue, or 2) directly from the forcefield parameters.

Dear Jürgen Weippert

By unstable, I mean that the SCF Done is coming out too low than the expected. The target energy is coming from a the Ethanal+OH/O2 system that has been studied in literature that we are trying to emulate. The unit of measurement is in Hartrees

Dr. Soumyadip Ray

In group theory for solid-state physics, a set of operators can form a group.

A basis is defined as a set of functions, each one of which, when operated on by one of the operators, is converted into a linear combination of the set of the functions (QM operators).

The most important for us are those that form an irreducible group representation because it has a number of properties, described in the group orthogonality theorem

An irreducible representation of a group is an "irrep" i.e., it has no nontrivial invariant subspaces as in math, and important for us is that physics uses irreducible representations to understand the rest (is like a basis of vectors).

In addition, I guess, that an example is needed, to comment more on this issue, a hydrogen atom solution in this form or the quadrupole moment decomposition discussion for example.

See for example:

Revo (https://physics.stackexchange.com/users/4521/revo), Irreducible representation in physics, URL (version: 2012-03-02): https://physics.stackexchange.com/q/21756

Best Regards.

Dear Yuvam,

Pressure in LAMMPS (and in MD in general, as far as I know) is calculated based on the virial, roughly speaking the forces between the particles. A negative pressure simply means that on average, the force between the atoms is attractive, i.e. the interatomic distances are, on average, larger than the equilibrium distance defined by the potential energy function.

Your system's dynamics are governed by the nve fix, which enforces the microcanonical ensemble, keeping the volume and the total energy of the system fixed. This ensemble is the default MD case without any additional modifications.

If you want to pressurize your system, you have to scale the box size during the simulation. Fix press/berendsen, that is commented out in your input, is one way to accomplish this. It is a simple fix that increases or decreases the box volume so that the pressure changes towards the desired value. This will not drain the heat away from the system, so your temperature will "float". For that, you would need a thermostat like temp/berendsen, which scales the velocities of atoms to change the temperature toward the desired value.

If you want to simulate a proper isothermal-isobaric (NPT) ensemble (in an atmospheric simulation, you probably do?), you will need to use the Nosé-Hoover thermostat+barostat instead of the Berendsen ones. This can be accomplished by using the npt fix instead of nve. Please see the manual page https://docs.lammps.org/fix_nh.html for details.

(The temperature increase you observe in your simulation is due to the potential energy transforming to kinetic energy. With the atoms initially stationary, but far away from each other, gradually start moving due to the attractive forces, and find more optimal (closer to equilibrium) positions with respect to each other. However, with the fixed box volume, the average distance is still "too large", and the system is likely too hot to condense into a droplet.)

OK, if you decide to use chemcraft with the output that includes the keyword pop=full and gfinput, the surfaces are visualized as follows:

Tools > Orbitals > Render molecular orbitals > select the orbitals you want

and ready!