FAQ

How easy is it to use Quantemol software?

With every new version of the Quantemol software we continue to improve the interface and make it even more use friendly. Very simple inputs are used for all applications and outputs are usually presented in a convenient text or graphical format. We have the special option of manual tuning of the program for advanced users.

We can organise a web-demonstration of Quantemol applications during which you can find out more about Quantemol and see how the programs actually work. After you buy a license we will arrange an introductory training session for users. We are always in touch with clients and ready to answer any kinds of questions.

How can I order a demo-version of Quantemol-software?

To order a web demonstration please e-mail sales@quantemol.com or give us a call on +44 (0)203 549 5841.

What is the difference between expert license, team license and site license?
  • “Expert licence”: Tbuseosinglworstatiob1users
  • “Team License”: Includeutworkstationoclusteusebteautuseriongroupcabinstalled
  • “Site licence”: Coverut2useracrosdifferengroupwithiainstitution/company

Quantemol-EC

How easy is it to use QEC software?

QEC is designed to be highly user-friendly, and provides an easy to use graphical user interface (GUI) through which users are prompted to enter quantities required for their calculations. A tutorial on how to run a QEC calculation may be found here.

What quantities does QEC calculate?

QEC calculates numerous observables, including:
– Total elastic scattering cross sections
– Momentum transfer cross sections
– Electron-impact electronic, rotational, and vibrational excitation cross sections
– High-energy electron-impact ionization cross sections
– Dissociative electron attachment cross sections
– Differential cross sections
– Scattering reaction rates

What types of molecule can QEC treat?

The code can treat both closed-shell and open-shell molecules, and can also treat positively-charged molecular ions. The code can handle molecules of various sizes. To date, calculations have been carried out on molecules containing up to 25 atoms, as well as molecules with up to 150 electrons.

What parallelism does QEC use?

QEC interfaces with the UKRmol+ suite of codes, which use both MPI and OpenMP parallelism throughout the calculation.

How long does a typical QEC calculation take?

The runtime for a calculation depends on the level of detail retained in the calculation, and the type of machine on which the calculation is run. For example, a calculation using Hartree-Fock Theory and a medium-sized basis set may take only a few minutes. Calculations using a higher-level of detail, for example complete-active-space calculations with a large number of active electrons may take several hours even when MPI or OpenMP parallelism is used. An estimate of the time remaining at each stage of the calculation is displayed in the GUI, so that the user is aware of the length of time to expect for the calculation to complete.

What is the most computationally demanding part of a QEC calculation?

The calculation of vibrational excitation cross sections is typically the most time-consuming part of a QEC calculation. This is because it involves an R-matrix calculation for each of the 3N-6 vibrational modes of an N-atom molecule. An estimate of the length of time required to perform this part of the calculation is displayed in the GUI during the calculation.

Can calculations be re-run or restarted using the setup from a previous calculation?

Yes, QEC stores the parameters used for each calculation in a pickle file. If the user wishes to re-run a previous calculation with, for example, a different basis set, then the pickle file of a previous calculation can be loaded (using the menu option File > Open > Previous QEC run) . When this is done, all other parameters from the previous calculation are automatically entered into the GUI without having to be re-entered by the user

How to obtain momentum transfer cross section in QEC?
The momentum transfer cross section is obtained from the fully differential cross section sigma(E, theta, phi). This differential cross section is then multiplied by (1-cos(theta)) and a volume element factor of sin(theta), and then integrated over theta and phi to give the momentum transfer cross section
Can QEC calculate rotational excitation and de-excitation processes?
Currently the code only calculates excitation processes. However, de-excitation cross sections can be derived from the excitation cross sections using the principle of detailed balance.
Can QEC calculate ionization CS from not only ground state but also excited states? And is multiple ionisation possible?
Currently the code assumes ionisation from the ground state. However it should be possible to modify the BEB expression to account for ionisation from excited states. This would probably involve shifting reducing the binding energy by the excitation energy in the formula. Multiple ionisation is not possible in the current code.
Can QEC calculate CS for generating ion pairs?
The code cannot treat multiple ion pairs, one of the assumptions in the method for dissociative electron attachment is that one particle remains neutral.
How high can QEC calculate with BEf?
Typically the BEf cross section is calculated out to energies of 3000 eV.
If we calculate CF₄, using the same settings for CH₄ , how long would the computation take? Additionally, if we use the default active orbitals for CF₄ (assuming at least 15 orbitals), would the calculation be feasible?

For CF₄, calculations using the SE and SEP methods typically take 5-10 minutes with standard basis sets like cc-pVTZ. However, increasing the complexity of the SEP calculation can extend the runtime to around an hour. Calculations using the close-coupling (CC) method generally take longer. In a sample CC calculation for CF₄ with 10 active orbitals, the process took approximately 2 hours on a 28-core workstation. The runtime depends on the number of active orbitals—reducing them can speed up the calculation but may compromise accuracy, requiring convergence tests to find the optimal balance. Conversely, increasing the number of active orbitals raises both runtime and memory demands, potentially making larger calculations less practical.

How much additional time is required when using HFSCF instead of MCSCF?

Calculations with MCSCF orbitals (using the CC method) are significantly more time-consuming than those with HFSCF orbitals (using the SE and SEP methods)

Is it practical to use HF orbitals in industrial applications?

In industrial applications, HFSCF orbitals are highly practical, especially when combined with the SEP method. This approach leverages a beneficial mathematical property of HFSCF orbitals to capture target polarisation effects accurately is crucial for calculating resonances involved in vibrational excitation and dissociative electron attachment. However, for determining electronic excitation cross sections, MCSCF orbitals are necessary.

What is the major bottleneck in calculating the cross section? In DFT calculations, the diagonalization process scales as O(n³). Are there any other processes (besides DFT) that take even longer?

The primary bottleneck in CC calculations is the Hamiltonian construction and diagonalization. In a test calculation, this step took 102 minutes out of a total runtime of 116 minutes. Generally, this process is the most time-consuming part of CC calculations, often accounting for 50% to 90% of the total runtime.

Quantemol-VT

Does the model take into account differences due to use of different rf frequencies for plasma generation? How about use of Faraday shields?

Yes the model will treat the differences between rf frequencies and Faraday shields can be included.

How does it help to predict charging damage?

Ion and electron fluxes onto surfaces can be predicted. This combined with an understanding of surface conductivity will give a usable analysis of possible charge damage.

 

What kind of output do you get? Can you predict deposition rate, thickness and film property uniformity (e.g. index of refraction & optical loss), film composition, control of particulate contamination?

Outputs include 2D visualisations of all simulated parameters (such as species temperatures and concentrations, electromagnetic fields etc). Outputs also include surface fluxes and surface site fractions for any simulated species. Output can also include deposition of etch rates. Outputs such as thickness and film properties would need to be extrapolated from calculation results.

Typically, how long does it take to build a new reactor configuration? What skills are required, how user-friendly is it?

We have a user friendly graphical interface that allows the user to paint a 2D representation of the tool. Given an existing simulated chemistry and a description of the physics being modeled a new or adapted geometry can be created in say 10 mins.

What process pressure range can be modeled?

Quantemol-VT can handle pressures up to about 10 Torr and as low as about 5 mTorr.

What is the biggest mesh size possible to model in Quantemol-VT?

Quantemol-VT itself does not limit the mesh size, it is however restricted by the computer’s memory. In the design phase, you are limited to 132 columns, but the mesh can be expanded later by an arbitrary integer