TSSP ExSci 2026 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.

Painlevé II transcendents and applications to PDEs

The Painlevé II equation arises as a self-similar reduction of the modified Korteweg–de Vries equation and has important applications in many fields of mathematics, including fluid mechanics, random matrix theory, and the theory of liquid crystals. Our aim is to study the second Painlevé equation using methods of complex analysis and techniques developed in the theory of integrable systems. In particular, we will focus on distinguished families of Painlevé II transcendents that are free of poles on the real line, such as the Ablowitz–Segur solutions. We will investigate their analytic properties, including the evaluation of total integrals, asymptotic behavior in appropriate sectors of the complex plane, and connection formulas between different asymptotic regimes. These will lead to applications in differential geometry. The plan of the work is as follows:

• establish and describe the connection of the Painlevé II equation with Riemann-Hilbert problem
• introduce and characterize real and imaginary Ablowitz-Segur solutions
• analyzing total integrals and the asymptotic behavior of the solutions in sectors of the complex plane
• explore applications of the solutions in a selected problem of differential geometry.

Supervisor:	Piotr Kokocki (pkokocki[at]mat.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Open Science in Practice: Replicability study of MSR Mining Challenge papers

The project focuses on replicating research presented at the Mining Challenge track at the Mining Software Repositories conferences, with the goal to assess whether the study can be repeated, and whether original finding holds when the study is repeated. Per conference Open Science policy results present in each publication should be reproducible and replicable, with available public packages.

The MSR 2025 challenge dataset consists of a Neo4j graph database containing Maven Central dependency graph. Research publications submitted to this challenge extracted useful information from this graph to gather insights into various aspects of software engineering and dependency management.

The MSR 2026 challenge dataset consists of AI-agent and human interactions in pull requests, from popular GitHub repositories. Research publications submitted to this challenge captured unique characteristics of AI-agent authored code.

Summer Project Objectives:

• Replicate experiments outlined in literature and related software replication packages available on Figshare or Zenodo platforms, or in a software repository on GitHub or other forge.
• Gather information if experiments cannot be replicated and how to fix the replication packages if necessary.
• Categorize different approaches to ensuring replicability, and different modes of failure (if any).
• Contribute to preliminary research by preparing a replicability report.

Mining Software Repositories (MSR) techniques related to software dependency analysis, code repository mining, etc., will complement this exploration.

While programming skills are essential, prior experience in MSR or other areas relevant to challenges is not mandatory. Tutorials at the project's commencement will equip participants with the necessary knowledge to actively contribute to this topic.

Supervisor:	Jakub Narębski (jnareb[at]gmail.com)
Co-supervisor:	Piotr Przymus (eror[at]mat.umk.pl , piotr.przymus[at]gmail.com)
Time:	July 2026

^ ^ ^ ^ ^

How bug fixes propagate in transitive dependencies

Modern software is built on long chains of third-party libraries. When a bug is fixed in one library, the fix reaches end users only after it is adopted and released by downstream packages. This project studies how such bug fixes propagate through transitive dependencies and what factors slow down or accelerate this process.

The main goal is to build a data-driven view of fix propagation in a large ecosystem. We will mine public software development data-version control history, release metadata, issue trackers, and package dependency information-to reconstruct dependency graphs and track when fixes appear and when they are incorporated by dependent projects.

Summer project objectives and tasks:

• Literature replication: reproduce an existing study on dependency propagation using published code and datasets (or re-implement the pipeline when needed), and verify key results.
• Pipeline & dataset construction: build an end-to-end pipeline that (a) extracts dependency graphs, (b) detects or links bug-fix events (e.g., commits, issues, releases), and (c) measures propagation delays across dependency levels.
• Exploratory analysis: quantify distributions of propagation times and relate them to graph properties (e.g., depth, centrality), project characteristics (activity, size), and release practices.
• Reporting & reproducibility: deliver a documented codebase, a cleaned dataset, and a short report summarizing findings and limitations.

Strong programming skills are required; prior experience in MSR or data mining is not necessary. Short tutorials and starter notebooks will be provided to help participants contribute quickly.

Supervisor:	Krzysztof Rykaczewski (krykaczewski[at]gmail.com)
Co-supervisor:	Piotr Przymus (eror[at]mat.umk.pl , piotr.przymus[at]gmail.com)
Time:	July 2026

^ ^ ^ ^ ^

Tracing stellar properties in Gaia microlensing events using high-resolution spectroscopy

The aim of this study is to analyse high-resolution spectroscopic data for selected gravitational microlensing events observed by the Gaia space mission. Gaia provides precise astrometric and photometric measurements that enable the detection and characterization of dark and compact microlensing events (e.g. black holes, neutron stars, white dwarfs) across the Milky Way, offering a unique, homogeneous dataset for such studies. In order to classify the events properly, one needs to manage spectroscopic analysis of selected objects. High-resolution spectroscopy plays also a key role in accurate determination of their physical and atmospheric parameters. These parameters are essential for constraining microlensing models, reducing degeneracies in the lens-source configuration and improving estimates of lens properties. As a result, the spectral type, astrophysical parameters and spectroscopic distances to the source stars are expected to be determined. The spectra for the analysis were obtained with the world-class telescopes, like 8-m Very Large Telescope (VLT) or 10-m South African Large Telescope (SALT). A subset of microlensing events will be selected for a detailed study of the absorption lines present in their spectra. The analysis will be based on the iSpec environment (Blanco-Cuaresma 2019, https://arxiv.org/abs/1902.09558). Therefore, proficiency in Python programming is highly required.

Supervisor:	Paweł Zieliński (pzielinski[at]umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Quantum technologies for optical clocks

Quantum technologies for optical clocks are a key frontier in next-generation time and frequency metrology. Optical clocks based on ultracold atoms already reach fractional frequency uncertainties at the 10⁻¹⁸ level, and further advances rely on improved quantum control and robust system engineering. This project focuses on quantum-technology building blocks that enable high-performance yet portable optical clocks. Core themes include schemes for continuous production of ultracold atoms suitable for clock interrogation, spatial transfer of atoms using hollow-core optical fibers, and cooling or manipulation in frequency-modulated dipole traps. In addition, concepts of ultra-stable and compact laser systems and miniaturized cooling architectures will be explored with an eye towards integration into transportable platforms. The developed modules are intended not only to support transportable optical clocks, but also to benefit other atom-based quantum sensors and devices. An important objective is to identify architectures and operating regimes that allow robust, long-term, and potentially continuous operation without degrading clock performance. Depending on the participant’s background, the work may involve theoretical modeling, numerical simulations, and/or close interaction with ongoing experiments in the laboratory. Specific tasks will be adapted to the student’s skills and interests, providing an opportunity to contribute to the design and understanding of building blocks for future portable quantum-enhanced optical clocks.

Supervisor:	Sławomir Bilicki (slawko[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

It’s a trap! - making (trapping) cold atoms for active optical atomic clock

Ultracold atoms are used in the most sensitive and accurate measurements achievable. In particular, optical atomic clocks with cold atoms are the most accurate devices ever built by humans. They provide an excellent platform for fundamental physics research as well as powerful tools for practical applications. However, despite their exceptional performance, optical clocks are limited by the optical frequency oscillator—specifically, the clock laser used to probe the ultra-narrow optical atomic transition. The clock laser is essential because optical clocks are passive devices (Rev. Mod. Phys. 87, 637 (2015)), requiring an external laser field to interrogate the extremely narrow atomic transition, known as the clock transition.
A new optical clock design has been proposed: a continuous active clock based on the phenomenon of superradiance. Such a clock has not yet been demonstrated and requires excellent control over a cold-atom ensemble.
The proposed research program of internship focuse research towards constructing a novel superradiant continuous optical clock setup, with particular emphasis on cold-atom formation and manipulation processes. In the limited time frame of the internship two main goals are to gain experience in collaborative research within a research team and to learn about the operation and construction of optical atomic clocks.
The internship will emphasize primarily experimental aspects, including assembly of the experimental setup, laser stabilization and control, ultra-high-vacuum systems, and control electronics. Additional topics include optical path-length stabilization, use of optical frequency combs, and the assembly and characterization of high-Q optical cavities.

Supervisor:	Marcin Bober (bober[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Vibrational Spectroscopy of Solvated Molecules

Many organic molecules of technological or biological importance are studied experimentally in solution. This means that simple computational models which include the single molecule in vacuum are typically not enough to explain measured spectroscopic signals of solvated molecules, as for example infrared (IR) spectroscopy. A simple and widely-used approach to include solvation effects is to use the conductor-like polarizable continuum model (CPCM), where the influence of the solvent on the solute is modeled using a set of Gaussian charge distributions on the surface of a cavity enclosing the solute molecule. The goal of the project is to implement the analytical equations required to compute the molecular Hessian [1, 2], i.e., the second order derivative of the energy with respect to atom coordinates, including solvent effects described using density functional theory (DFT) and CPCM [3, 4]. The implementation will be carried out in the VeloxChem program [5] and will be used to determine IR spectra in solution of a series of organic molecules with applications in organic optoelectronics.

References:
[1] M. Hodecker, P. Norman, I. E. Brumboiu, Chem. Methods, 5, e202500033 (2025)
[2] A. Pausch, J. Chem. Theory Comput. 20, 8, 3169– 3183 (2024)
[3] F. Liu et al., J. Chem. Theory Comput. 11, 3131 (2015)
[4] J. Liu and W. Liang, J. Chem. Phys. 138, 024101 (2013)
[5] Z. Rikevicius et al., WIREs Comput. Mol. Sci. 10, e1457 (2020).

Supervisor:	Iulia Emilia Brumboiu (iubr[at]umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

New approach to enhance the efficiency of organic solar cells using azo dyes

The development of new organic materials for photovoltaic applications is crucial due to environmental needs and socio-economic interests to reduce fossil fuel use. Azo dyes, as photosensitive materials, have unique photophysical and photochemical properties. Particularly their ability to undergo reversible trans–cis photoisomerization upon UV-Vis light makes them promising materials for photovoltaic applications. It is well known that by modifying the substituents in azobenzene core, it is possible to tune the HOMO/LUMO energy levels, allowing for better adjustment to electron donors or acceptors in organic solar cells (OSCs).
This project focuses on exploring novel azo dye materials and integrating them into photovoltaic technologies. The expected outcomes could significantly expand the role of azo dyes in solar energy applications, promoting more efficient and scalable renewable energy solutions.
This project aims to study novel photoswitchable azobenzene-type molecules to enhance OSC efficiency. Azo compounds will be integrated into the active layer of the OSC device to improve its performance. Additionally, nanoscale patterning will be used to increase light scattering and optical path length. A thorough investigation of new molecular architectures will reveal photoactivation mechanisms critical for optimizing their photovoltaic properties in OSCs.

Supervisor:	Beata Derkowska (beata[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Ab-initio investigation of the magnetic interactions in the van der Waals material GdGaI

GdGaI [1] is a recently synthesized layered van der Waals (vdW) magnetic system that brings the heavy 4f elements of the periodic table into the 2D limit, with potential for 2D spintronics. Each vdW slab consists of two identical layers of Gd atoms with large magnetic moments of 7 μB arranged on a 2D triangular lattice. The ground state is insulating and hosts a complex, non-collinear and non-coplanar triple-q magnetic order with skyrmion-like topology. This behavior is linked to the 4f-5d inter-orbital Hund’s exchange interaction, as recently confirmed by a model Hamiltonian study [2]. In this project, we will calculate the inter-site exchange (Jij) and on-site anisotropy (KU) using the LMTO-based DFT code RSPt. By employing the LKAG approach [3], we will efficiently extract exchange parameters from several reference states. Comparing these results will help us understand the coupling between electronic and magnetic degrees of freedom and assess if the magnetism in this system follows a Heisenberg-type model. Along with Jij, we will calculate KU using the force theorem in the presence of SOC. These calculated parameters will help us determine if the observed non-collinear magnetic order is purely due to the geometric frustration on the triangular lattice or if other interactions, such as second-neighbor exchange or Dzyaloshinskii-Moriya interaction (DMI), contribute to it.

[1] arXiv: 2405.16781
[2] https://doi.org/10.1103/pt6l-xvlx
[3] Physical Review B 91, 125133 (2015)

Supervisor:	Igor Di Marco (igor.dimarco[at]umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

The influence of silver nanoparticles on the physical properties of thin azobenzene films

The main goal of this project is to investigate new silver nanostructured azobenzene thin films as photosensitive materials for optoelectronics and photonics. This nanohybrid materials will combine the light-responsive azo function (cis-trans isomerization) with silver nanoparticles (AgNPs). Student will be learning "green chemistry" techniques of the AgNPs synthesis. The physical properties of the resulting nanoparticles will be examined using UV-VIS spectroscopy and scanning electron microscopy (SEM). Thin azobenzene films doped with AgNPs will be studied for photoisomerization, photoinduced birefringence, and surface-relief gratings (SRG).

Supervisor:	Dorota Kowalska (dorota[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Exploring the Impact of the Treatment of Electron Density on Molecular Docking

Molecular docking is a crucial tool in the design of both new materials and drug molecules. Current methods use Newtonian mechanics. This is because both materials such as zeolites and protein targets are large systems and calculations upon these systems become computationally expensive for even lower levels of theory. The treatment of electron density is therefore reduced to forcefield models which offer only common interactions. The use of Quantum Mechanics would increase the available information regarding the electron densities of both the bound molecule and the binding site. In order to address the computational cost, this project will utilize QMMM layering methods.

This project aims to assess:
1.) The impact of the Quantum Methods on computational cost.
2.) Whether knowledge of the electron density landscape can aid in the design of targeted interacting molecules.

Tasks will include writing calculation input files, submitting the jobs to high performance computing resources and analyzing the output. The ideal student will have an interest in chemistry physics or computing and be looking to apply these interests to chemical and biological systems. Students will gain experience in using Computational Chemistry for cutting edge applications.

Supervisor:	James Mattock (james.mattock[at]umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

From Molecular Dynamics to Binding Modes: Python Tool for Ligand Pose Clustering in ProDy

This project brings together computational biology/biophysics and scientific software development. The goal is to design and integrate into the ProDy API (http://www.bahargroup.org/prody/) a Python-based tool that analyzes all-atom molecular dynamics (MD) simulations to identify and describe ligand binding modes—clusters of the most frequently observed ligand poses in the binding pocket—relevant to structure-based drug discovery workflows. The work is application-oriented: the student will develop code that extract the data about ligand (chemical compound) position into interpretable, quantitative descriptors of how a ligand occupies a protein binding pocket over time.

The tool will be built within the ProDy framework and developed in a modular, maintainable manner to facilitate integration into a broader protein–ligand interaction analysis framework. The pipeline will (i) read trajectory files using ProDy functionality, (ii) align MD frames to a protein reference, (iii) isolate ligand heavy atoms (or user-defined selections), and (iv) compute pose similarity measures—primarily ligand-to-ligand RMSD across frames. Based on these descriptors, the student will implement at least one clustering approach (e.g., hierarchical clustering or k-medoids) to identify distinct ligand binding modes and quantify their populations and structural diversity.

Beyond clustering, the tool will provide post-processing utilities: within-cluster compactness metrics, identification of representative poses (centroid/medoid frames), export of representative structures for visualization, and contact profiling that reports which residues most frequently interact with the ligand in each cluster.

The final deliverable will include tested code, documentation, and example workflows, prepared for collaborative development and potential contribution via GitHub as a ProDy-compatible module.

Supervisor:	Karolina Mikulska-Rumińska (karolamik[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Precision Landings on Chemical Peaks: A New Distance Geometry Protocol for Transition State Discovery

Locating a Transition State (TS) on a rugged Potential Energy Surface (PES) is notoriously difficult: conventional searches are like climbing an unknown mountain from the bottom, where a slightly wrong start can trap you in valleys of equilibrium structures. This project reframes TS discovery as a precision landing—placing an initial guess directly in the barrier region, close to the desired first-order saddle point.

This one-month intensive project will develop a TS-search protocol based on Distance Geometry (DG) as implemented in RDKit. Instead of relying on simple reactant–product interpolation (often unreliable when large reorientation is needed), the approach explicitly analyzes bond-breaking and bond-forming events to construct chemically informed coordinates for the transition structure.

Inspired by the Schlegel group’s work on redundant internal coordinates and recent DG-based conformer generation advances, you will:

• Build a workflow to generate high-quality TS guesses with RDKit.
• Reduce dependence on computationally expensive full Hessian calculations.
• Accelerate convergence by targeting structures near the single negative eigenvalue region.

We seek a motivated student with a possible background in molecular modeling proficiency in Python (RDKit experience is a plus), and interest in high-impact computational chemistry research.

Supervisor:	Szymon Śmiga (szsmiga[at]fizyka.umk.pl)
Co-supervisor:	Amol Patil (amolpatil[at]umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Identifying ligand binding sites in proteins: ion channels and enzymes in focus

ATP-sensitive potassium (KATP) channels are vital metabolic sensors that link cellular metabolism to electrical activity. Recent high-resolution cryo-electron microscopy (cryo-EM) structures have revealed non-protein electron densities that appear to be cholesterol molecules bound at specific sites. However, whether these represent stable, functional binding sites or transient lipid interactions remains an open question. This project aims to bridge the gap between static structural data and dynamic molecular behavior. Using two different isoforms of the KATP channel, the student will investigate the stability and potential functional roles of these observed cholesterol-binding sites. By analyzing existing molecular dynamics (MD) trajectories, the participant will determine how long cholesterol molecules remain in these pockets and how their presence influences the local protein conformation.

The student will focus on "data mining" existing MD simulations to calculate lipid residence times and occupancy maps. A key objective is to compare two isoforms to see if these sites are structurally conserved. To gain a complete understanding of the MD workflow, the student will also set up a simplified system and run it on a high-performance computing (HPC) cluster. This hands-on experience will cover everything from system preparation to the initial stages of production and data analysis.

Strong Python skills are essential for processing trajectory data and creating custom visualizations. A background in biochemistry or biophysics is preferred. This project is ideal for a student interested in testing structural hypotheses and exploring the "hidden" dynamics of membrane proteins.

Supervisor:	Katarzyna Walczewska-Szewc (kszewc[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^

Automated Polarization Control System for Fiber-Optic Communication

This project focuses on developing an automated system to adjust the orientation of a half-wave plate (HWP) to achieve the correct polarization of laser light entering a polarization-maintaining (PM) fiber. The system will measure amplitude noise at the fiber's output and automatically adjust the HWP's angle, optimizing signal quality and minimizing noise.

In this project, the student will work closely with a supervisor to integrate various hardware components and develop a control system. The key tasks include using a light-sensitive element (such as a photodiode or camera) to monitor amplitude noise in the output signal, adjusting the angle of the HWP mounted on a controllable frame based on the noise measurements, and writing a control script in Python to communicate with both the photodetector (or camera) and the adjustable frame, automating the process of fine-tuning the HWP's position.

The project supervisor will provide guidance and support throughout the process. The student will gain valuable hands-on experience in optics, fiber-optic communication, automation, and real-time control systems while also developing programming skills.

Supervisor:	Marcin Witkowski (marcin_w[at]fizyka.umk.pl)
Time:	July 2026

^ ^ ^ ^ ^