TSSP ExSci 2023 projects

The main component of the summer program is an active participation in the selected Science research project offered by the staff members of the Nicolaus Copernicus University, please see the topics and their descriptions below. Interested students are welcome to contact possible advisors for more details concerning the foreseen projects and discuss the dates that the project could be undertaken.

Mining software repositories: augumenting Python security vulnerabilities dataset

Software bugs occur in the development cycle of nearly all of software projects and can cause severe problems. Information about software bugs are often delivered by users, who submit bug reports containing details about encountered defects. The Common Vulnerabilities and Exposures (CVEs) system is one of reference methods for documenting publicly known information security weaknesses and exposures.

The goal of this project is to extend the Python CVE subset with all sorts of additional information and source code annotations. Such extended dataset can be used in a variety of applications, like bug localization, code completion etc. It can also be used to train deep neural networks and large language models as well as adversarial attacks on these networks.

During this project Mining Software Repositories (MSR) techniques will be used (specific type of data mining). MSR concentrates on analysing rich data sources used in software development, like version control repositories, bug tracking tools, mailing lists and other, with the goal to uncover interesting and usable information that can help to improve software development.

Good programming skills are required, but prior experience in MSR is not a must. We will provide tutorials on relevant topics at the beginning of the project. Tasks:

1. Learn basics of Data mining, Machine Learning and MSR techniques.
2. Learn basics of cluster computing.
3. Help with the augmentation of Python security vulnerabilities dataset.

Supervisor:	Piotr Przymus (eror[at]mat.umk.pl, piotr.przymus[at]gmail.com)
Time:	June – July 2023 (to be agreed)

^ ^ ^ ^ ^

Mimicking human game-play strategies

Hive is an abstract strategy tabletop game for two players, in which the goal is to surround one of the tiles of the opponent. The players can either move one of their tiles or place a new one next to the already placed ones corresponding to the game rules. The main goal of the project is to create a deep reinforcement learning solution for playing Hive.

The complexity of the game is similar to the complexity of chess in several aspects. One difference, however, is the lack of a game board. Hive has a virtual game board, which can theoretically be of infinite size – the source of the only limitation is the number of pieces in play. Another challenge is to prepare the agent with the lack of outstanding computational resources. The motivation behind the task is the fact that – at the time of writing this document – no agent for Hive has been implemented which could surpass the skills of a human expert.

In order to create an intelligent AI, investigating the already existing solutions for developing intelligent agents for Hive or other similar problems is of paramount importance. Furthermore, a suitable implementation of the Hive game itself is necessary for the elaboration and learning process of the AI itself. This documentation leads through the steppes of the planning, design and implementation phases of the development, as well as the decision being taken before and during the development.

The main task is to implement a framework, which can be used as an environment for the intelligent agent. Furthermore, the student using Reinforcement Learning should implement an agent, which could surpass the efficiency of an AI which steps randomly.

Supervisor:	Krzysztof Rykaczewski (krykaczewski[at]gmail.com)
Time:	September 2023)

^ ^ ^ ^ ^

Cold atoms optical frequency standards for science

Cold atomic experiments are composed of complex setups which require team work and broad knowledge for operation. During the four weeks internship student will be involved in current research activities connected with optical atomic clocks in CASTLE group (http://fizyka.umk.pl/~castle/). Optical atomic clocks with neutral cold atoms allow for absolute precision on the order of 10-18, unmatched by any other device. The main goal of the internship is to get experience in research team work that includes broad range of design, simulations, theoretical and experimental work with active and passive optical clocks, construction of an ultra- stable laser for strontium atomic clock, searching of dark matter with atomic clocks and studies of fundamental physics with atom-light interaction. Other activity is related to investigation of methods for transfer and cooling of ultracold atoms by using hollow-core fiber in order to develop a compact source of laser-cooled atoms that can be used for quantum computers and atomic sensors. All projects are conducted with a cooperation with the leading quantum metrology groups like NIST (USA), LNE-SYRTE (France), University of Amsterdam (the Netherlands) or University of Bonn (Germany). The specific tasks can include laser stabilization by using an optical frequency comb, build optical and vacuum systems, testing electronic drivers or simulations that will be adopted to student's interests and skills.

Supervisor:	Sławomir Bilicki (slawko[at]fizyka.umk.pl)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Ultra-stable laser for optical frequency standards and fundamental physics

Ultra-stable laser is the heart of optical atomic clock that determines its short term stability. Such a laser was also suggested as a tool for quantum key distribution and telecommunication applications, where its long coherence time allows to stabilize optical path length over long distance with nm scale.

The main two goals of proposed internship are to gain experience in research team work and learn about optical atomic clock operation and construction. The involved student will mostly work with ultra-stable laser part of the clock. The internship can be focused either on:

• more theoretical aspects, looking for better way to know our Universe using optical atomic clock and ultra-stable laser as a tool (i.e. gravitational wave detection: https://arxiv.org/abs/2301.08200, dark matter search: Science advances vol. 40, no. 12 aau4869);
• simulations and calculations, i.e. of the best operating conditions or finding the most efficient operation algorithms; - experimental with assembly of the ultra-stable laser setup, connecting it to optical atomic clock and integrating it into a fibre network connecting a few cities in Poland;
• combination of all above.

Specific task may involve ultra-high vacuum technology assembly, laser stabilization and laser line narrowing, electronics, optical path length stabilization, optical frequency comb operation, high-Q optical cavities, heat transfer simulation, vibration non sensitive support points calculation, investigation of fundamental noise limits in optical cavity, data analysis for dark matter search.

Supervisor:	Marcin Bober (bober[at]fizyka.umk.pl)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Theoretical time-resolved X-ray absorption spectroscopy of thiophene

Thiophene is the quintessential building block molecule for materials used in organic photovoltaics (OPVs) and optoelectronics. As the photo-induced dynamics of organic molecules can strongly affect the stability of the corresponding devices, it is important to be able to track the excited state processes involved. With its sensitivity to subtle changes in the chemical environment, X-ray spectroscopy can be the perfect tool to track photo-induced dynamics and de-excitation processes. The proposed project aims to apply theoretical spectroscopy tools to determine the excited state dynamics in thiophene and identify the spectroscopic fingerprints of possible de-excitation pathways.

Goal: Identify the time-resolved (TR) X-ray absorption spectroscopy (XAS) signatures of two de-excitation mechanisms in thiophene.

Tasks:

1. Calculate the UV-vis absorption spectrum of thiophene using time-dependent density functional theory (TDDFT).
2. Determine natural transition orbitals, attachment and detachment densities and characterize the nature of the lowest four singlet excited states.
3. Construct the excited state potential energy surface (PES) by interpolation.
4. Determine excited-state dynamics trajectories and identify the ring-opening and ring-puckering de-excitation mechanisms [1].
5. Calculate XAS spectra along different excited state trajectories and identify potential fingerprints in XAS of the ring-opening and ring-puckering de-excitation mechanisms.

The theoretical calculations will be performed using the eChem platform [2] in the VeloxChem quantum chemistry software [3].

References:

[1] A. Prlj et. al, Phys. Chem. Chem. Phys. 17 (2015), 14719.
[2] https://kthpanor.github.io/echem/docs/title.html
[3] https://veloxchem.org/docs/int

Supervisor:	Iulia Emilia Brumboiu (iubr[at]umk.pl)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Charge transfer process in small diamine cations

Charge transfer is a fundamental process presented in various biological reactions (photosynthesis) and many modern technological applications (photovoltaics). However, it is tough to investigate one process in such complicated systems where many different reactions coincide. To do so, much simpler molecules, such as diamine cations, can be employed to help understand the migration of a charge better.

The study will focus on investigating the potential energy surface describing the charge transfer process between different states of diamine cation using different DFT functionals.

Supervisor:	Marta Gałyńska (marta.galynska[at]gmail.com)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Printed nanometallic structures for plasmon - enhanced smartphone fluorescence microscopy

In the project we focus on fabrication of substrates with printed nanometallic structures with controlled morphology and location. Next, the surface of nanoparticles will be properly functionalized to facilitate specific attachment of desired molecular targets. Assignment of their presence will be possible by correlating images with the pattern of printed nanostructures. The method of laser-induced photochemical deposition of printed nanometallic structures has a number of advantages. It does not require advanced apparatus or specific reagents, allows for controlling the morphology and geometry of fabricated structures, it is also fast, cheap and scalable. It is based on reduction of metal salt using laser beam with appropriate wavelength. The nanoparticles exhibit chemically active Surface suitable for functional group attachment, required for obtaining selectivity of detection. Unique benefit of nanometallic structures is related to plasmonic enhancement, which yields strong increase of emission intensity of fluorophores placed in their vicinity. Particular aspects to be studied within the TSSP project:

• Printing nanometallic structures
• Surface modification of nanometallic structures
• Conjugation between nanometallic structures and photoactive proteins
• Experiments using wide-field fluorescence imaging microscopy
• Creating smartphone fluorescence microscopy
• Fluorescence detection of photoactive proteins in real-time down to the single molecule level

Supervisor:	Sebastian Maćkowski (mackowski[at]fizyka.umk.pl)
Co-supervisor:	Karolina Sulowska (sulowska[at]doktorant.umk.pl)
Time:	August 2023 (to be agreed)

^ ^ ^ ^ ^

Water bridging in the molecular dynamics simulations of proteins – new Python tool development

The research goal of the project is to develop a new computational tool in Python programming language which will detect water bridges within protein structures in molecular dynamics (MD) simulations and in the crystal structures.

Water bridges are hydrogen bonds that are formed when water molecules bridge between two or more protein residues, creating a hydrogen-bonding network that helps stabilize protein-protein and protein-ligand interactions. Developed code will help to estimate conserved waters and protein regions that are forming additional interactions mediated by water molecules. Moreover, the code will become an integral part of the ESty (Energetics & Stability) module, developed by the author of the project, which is available in the ProDy API; a commonly used package by the scientific community (> 2.3 downloads).

The following tasks will be performed within the framework of the project:

1. Familiarization with the ProDy package to enhance the understanding of the development environment. The knowledge will be utilized to develop the code effectively.
2. Develop Python code using ProDy to identify water bridges in (i) crystal structures and (ii) molecular dynamics simulations of protein structures.
3. Create supporting functions to provide statistical analysis.
4. Test the developed code on various systems to validate its efficacy.
5. Incorporate additional modifications to adapt the newly developed code to the ProDy package.
6. Gain proficiency in using GitHub to share the code with ProDy developers.

Supervisor:	Karolina Mikulska-Rumińska (karolamik[at]fizyka.umk.pl)
Time:	10 July - 06 August 2023

^ ^ ^ ^ ^

Enhanced Wide-Field OCT Imaging for Comprehensive Eye Examination

Optical Coherence Tomography (OCT) is a widely used imaging method in ophthalmology that allows for the visualization of the internal structures of the eye. However, traditional OCT scans can capture a small part of the eye at a time, limiting their ability to provide a big picture of possible eye disease detection. To overcome this limitation, we propose a project that aims to stitch multiple OCT scans together to generate a panoramic, enhanced Wide-Field OCT eye preview.

The project involves the development of a software program preferably in Python scripting language that will automatically align and stitch together multiple OCT scans, taken at different angles and depths, into a single panoramic view.

To achieve this, the software will first preprocess the individual OCT scans to correct for any image distortions or artifacts that may be present. It will then use image processing techniques, such as feature detection and registration, to align and stitch together the individual scans into a seamless panoramic image. Preferably data processing tool for this task will be the Computer Vision library (OpenCV).

In conclusion, the proposed project aims to develop a software program that can stitch multiple OCT scans together to generate a panoramic eye full-length preview (from the anterior to the eye posterior). It will provide more information including a detailed view of the retina, optic nerve, vitreous body, etc. This new, synthesized tool will provide a more comprehensive view of the entire eye, allowing ophthalmologists better visualize and identify any abnormalities or changes in the eye’s internal structures.

Supervisor:	Daniel Rumiński (drdr[at]fizyka.umk.pl)
Time:	10 July - 06 August 2023

^ ^ ^ ^ ^

Optical response of polar system in the vicinity of metallic nanoparticles

The project is focused on calculations of the emission spectrum of a polar system (e.g., molecule, quantum dot) enhanced by a neighboring nanoparticle. The unique feature of polar systems is the permanent dipole moment (PDM) associated with their eigenstates which, upon interaction with light, gives rise to new phenomena. Since these phenomena may be weak, we exploit metallic nanoparticles to enhance the associated signals.

Our model contains a two-level system (TLS) with different values of the PDMs in both states. It is illuminated by a strong, classical, and monochromatic beam in the visible frequency regime.

Until now, investigations revealed that TLSs with PDMs scatter light not only in the optical regime but additionally in a low-frequency (near THz) range. However, this additional radiation’s frequency and power depend on the intensity of the driving field and is usually relatively small.

Metallic nanoparticles are capable of confining electromagnetic energy into small volumes of space and hence, of locally increasing electric field intensity. They can thus enhance the weak PDM signal.

This project aims to develop the analytical description of scattered low-frequency radiation from the TLS with PDMs in the vicinity of a metallic nanosphere.

Tasks for the participant:

Week 1. Introduction to the time evolution of a TLS with and without PDMs. Calculations will be based on the Bloch equations.

Week 2. Introduction to light scattering by spherical nanoparticles. Results will be obtained based on the Lorentz-Mie equations.

Week 3-4. Derivation of analytical form of the power spectrum of the TLS with PDMs driven by an enhanced laser field for different points around the nanosphere.

Supervisor:	Karolina Słowik (karolina[at]fizyka.umk.pl)
Co-supervisor:	Piotr Gładysz (glad[at]doktorant.umk.pl)
Time:	August 2023 (to be agreed)

^ ^ ^ ^ ^

Investigation of second-order correlation energy expression in the context of Density Functional Theory

Kohn-Sham (KS) density functional theory (DFT) is the most used electronic structure computational approach for molecular and solid-state systems. Its accuracy depends on the choice of the approximation for the exchange-correlation (XC) functional which, at the highest-rung of the Jacob’s ladder, involves all the occupied and virtual KS orbitals as well as the eigenvalues i.e., double-hybrid (DH) functionals. This type of functional was successfully applied in systems e.g., to model weakly interacting molecular systems, or to calculate properties such as bond-length alternations or charge- transfer excitations.

The current density and functional driven error analysis have shown that the performance of DH functionals is mainly governed by functional-driven error. Within this project, we will try to construct a new DH functional by optimizing the functional parameters to lower this error.

Potential candidates should be well-motivated to pursue scientific work. Good mathematical skills, and basic knowledge of quantum chemistry and the Linux operating system will be more than welcome.

+ preparation of computational setup
+ analysis of the results

Supervisor:	Szymon Śmiga (szsmiga[at]fizyka.umk.pl)
Co-supervisor:	Aditi Singh (aditisingh4812[at]doktorant.umk.pl)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Quantum chemistry methods in soft confinement regimes

Quantum confinement substantially alters the electronic structure of quantum systems (e.g., atoms, molecules, and solids) as compared to their corresponding free-state counterparts. This is exhibited in the changes in electronic energy levels, electronic shell filling, and orbitals which, in consequence, affect their physical as well as chemical properties such as energetics, reactivity, response properties, etc. Therefore, the chemistry of confinement systems may drastically change. One of the most crucial points when analyzing quantum confined systems is the choice of an accurate model and quantum chemical method, which allows describing the changes in the electronic wavefunction correctly due to the effect of confinement. Within the project, we will study the performance of various wave function theory (WFT) and density functional theory (DFT) methods in a soft quantum confinement regime simulated by harmonic oscillator potential. Based on that, we will create a database of confined chemical systems, which will be used to assess to construct new, improved quantum chemistry methods.

Potential candidates should be well-motivated to pursue scientific work. Good mathematical skills and basic knowledge of quantum chemistry and the Linux operating system will be more than welcome.

+ preparation of computational setup
+ analysis of the results

Supervisor:	Szymon Śmiga (szsmiga[at]fizyka.umk.pl)
Co-supervisor:	Gabriel Chirchir (503507[at]doktorant.umk.pl)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Derivation and computational implementation of regularization procedures for calculating the leading relativistic and QED corrections for D states of atoms

Some of the relativistic and QED operators include singular terms; mass velocity operator, Darvin operator, and Dirac delta function operator are exactly such terms. The expectation values of singular operators usually converge much slower with the number of the basis functions used to expand the wave function than the expectation values of non-singular operators. The slow convergence is mainly due to the local character of the singular operators, i.e. their expectation values depend on the accuracy of small fragments of the wave function rather than on the overall accuracy of the wave function. For approximate wave functions, local errors may be considerably more significant than the global error. Thus, in the calculations of the expectation values of local/singular operators with such wave functions, the error can be larger than the error of the expectation values calculated for such global operators as, for example, the Hamiltonian. This behavior, in general, may occur regardless of the basis set employed in the calculation.

To reduce the accuracy loss in the calculations of the expectation values of local/singular operators, it was proposed to replace these operators by equivalent operators but less singular and less local [1,2,3]. Drachman proposed the so-called regularization approach, to construct such replacement operators based on the work of Trivedi [3] that made use of an expectation value identity.

In the limit of the exact wave function, the original operators and the replacement operators give the same expectation values, while for approximate wave functions the expectation values of the replacement operators are usually much closer to the exact values than expectation values of the original operators.

The regularization approach of Drachman is particularly useful and effective in calculating expectation values of the operators representing the leading relativistic and QED corrections with wave functions expanded in terms of explicitly correlated all-particle Gaussian functions (ECGs) [4]. For ECGs, the application of the regularization method is particularly important, as these functions do not strictly satisfy the Kato conditions. The implementation of the regularization method in the calculation of the leading relativistic corrections for atomic D states is the aim of this project.

Recently we have implemented algorithms for calculating the leading relativistic corrections for D states of atoms using the wave functions expanded in terms of explicitly correlated Gaussian basis functions and obtained in the calculations where the Born-Oppenheimer approximation was not assumed. The algorithms for calculating the corrections have been derived using the standard procedure applied before for calculating the relativistic corrections for S and P atomic states. The implementation of the algorithms has shown slow convergence of the values of the corrections with the number basis functions. To remedy the problem, the above-described regularization method will be implemented in this project to obtain better converged values of the corrections for the atomic D states.

The implementation will involve the derivation of the regularization algorithm for the Dirac delta function dependent on a inter-particle distance. Once implemented in the computer code that is used to calculate the leading relativistic corrections, some application work will start. It will involve calculating inter-state transition energies for such atomic systems as lithium, beryllium, boron, and carbon, as well as their ions.

1. Hiller J., Sucher J. and Feinberg, G., Phys. Rev. A, 18 (1978)
2. Drachman R. J. and Sucher J., Phys. Rev. A, 20 (1979)
3. Trivedi H. P., J. Phys. B, 13 (1980)
4. Pachucki K, Cencek W., Komasa J., J. Chem. Phys., 122 (2005)

Supervisor:	Monika Stanke (ms[at]umk.pl)
Time:	July 2023 (to be agreed)

^ ^ ^ ^ ^

Searching a needle in a haystack - What are the differences in the interaction of different KATP channel isoforms with ligands?

ATP-activated potassium channels (KATP) are complex protein structures found throughout our bodies. Depending on their location, they differ in the choice of components from which they are constructed. Despite their considerable similarity, individual KATP isoforms differ in their sensitivity to their natural ligands ATP and MgATP. We still need to figure out why and we will do this with the help of numerical modelling.

The project's goal will be to directly compare all three major channel isoforms-pancreatic, cardiac and vascular-under the same ligand conditions. The participant in the summer research program will learn to build models of molecular systems based on existing structures and homology models. We will use bioinformatics tools to compare sequences and identify essential motifs and ligands binding sites. Then, we will use classical and enhanced molecular dynamics methods to determine hot spots in the ligand binding in the systems and understand their structural differences.

Supervisor:	Katarzyna Walczewska-Szewc (kszewc[at]fizyka.umk.pl)
Time:	17 July - 15 September 2023 (to be agreed)

^ ^ ^ ^ ^

Fiber-coupled tapered amplifier

The project aims to assemble a tapered amplifier (TA) which will be used to amplify a continuous-wave laser beam coming from an external cavity diode laser. TA will be coupled via optical fibers to the rest of the already operating cold Rb and Hg atoms experimental system. All necessary components required for the project will be purchased in advance and delivered to the student. The student will select appropriate lenses to collimate the laser beam and install the Peltier module for thermal stabilization. The next task will be to couple both input and output beams to fibers and characterize the amplitude and frequency noise induced by the TA. All activities related to the project will be performed by the student under the guidance of the project supervisor.

Supervisor:	Marcin Witkowski (marcin_w[at]fizyka.umk.pl)
Time:	June 2023 (to be agreed)

^ ^ ^ ^ ^

Spectroscopy with high-finesse optical cavities

Cavity ring-down spectroscopy (CRDS) is one of the most sensitive spectroscopic techniques and has been recognized as an ideal tool for providing reference data for satellites monitoring the Earth’s atmosphere. The use of a high-finesse optical cavity can extend the optical pathlength to several kilometers and allows precise measurements of weak absorption spectra important in atmospheric studies (O2, CO, CO2, etc.). Recently developed in our group two novel techniques namely CMWS (cavity mode width spectroscopy) and CMDS (cavity mode dispersion spectroscopy) enables to overcome the limitations of the CRDS in the high absorption regime while retaining its precision and accuracy.

Depending on the willingness, skills and interests, participant may be involved in various activities such as:

• development of existing spectrometers,
• development of the experimental control software,
• participation in measurements,
• data analysis including advanced spectral line shape models (e.g. Hartmann-Tran profile) and multispectrum fitting technique.

Supervisor:	Szymon Wójtewicz (szymon[at]umk.pl)
Time:	3 - 28 July 2023

^ ^ ^ ^ ^