Science Case:

Recent advances in nanofabrication are the driving force for making quantum technologies miniature and scalable. Especially for photonic quantum networks, integrated, narrowband single photon sources, optical nonlinearities and detection at the single photon level are key. The first two devices can be implemented by strongly coupling a quantum emitter to the light field of nanophotonic waveguides and cavities, that can then transport the output states to a tailored single photon detectors. Such systems that can be operated at room temperature have thus far remained elusive.

Hypotheses and Objectives:

However, the advantageous properties of quantum emitters in hexagonal boron nitride (hBN), that can show emission of indistinguishable photons at room temperature, show a promising route to the first chip-based integrated single photon source and optical nonlinearity. As hBN is a CMOS compatible material, all components of the system can be set up within an integrated, potentially mass producible device. In order to implement such a system, several challenges need to be tackled: the co-design of appropriate nanophotonic waveguides, cavities and electrodes for Stark tuning of the quantum emitters; the integration of these components with the emitters; the implementation of detectors with the required temporal resolution; the separation of excitation and fluorescence light; the analysis of the photon statistics via field programmable gate arrays (FPGAs) or directly in combination with the detector using mixed signal ASIC design.

Approaches and Methods:

• Nanophotonic components and integration with quantum emitters: For a strong-light matter interaction, appropriate waveguides and cavities based on SiN on SiO2 will be designed and fabricated. The quantum emitters in hBN will be obtained from Graphene Supermarket and from a collaborator from the University of Cambridge, UK (Prof. Stephan Hofmann). Here, the expertise of co-supervisor S. Skoff will be essential, who has also the required expertise of interfacing quantum emitters in hBN with such components.
• Tuning electrodes: Electrodes that will allow for the emission wavelength tuning of the quantum emitters will be designed and then fabricated via electron beam lithography on the nanophotonic chips.
• Avalanche photodiode arrays: To detect and analyze the generated photons and any nonlinear effects in an integrated device, efficient and fast on-chip detectors are required. We will employ a tailored Avalanche Photodiode (APD) array design in an integrated HV CMOS process, with integrated amplification and system read out. For the appropriate layout and performance optimization, we will use numerical and EDA tools such as COMSOL, TCAD, Quantum-ATK and Cadence. Fabrication will be done using commercial partners available through the Europractice Network.
• Optical filtering: As APDs are not wavelength sensitive, optical filter structures need to be implemented in order to ensure that the decoupling of excitation and fluorescence light is efficient. Here, first efforts will be made by using commercial Opto-CMOS options but on chip semiconductor heterostructures will also be investigated. Here an existing cooperation with Dr. Detz from CEITEC in Brno, (CZ) will be extended.
• Analysis: Analyzing the photon statistics of the light-matter interface will be crucial to determine the efficiency of the device and ensure that single photons are indeed interacting with single quantum emitters. For processing the signals efficiently field programmable gate arrays (FPGAs) will be used, that enable fast signal processing without the need to store all of the huge amount of raw data generated. This part will greatly benefit from exchange with the members of the consortium working on High Energy Physics, as they are experts in signal processing.

Supervision:

C. Koller (FHWN, Supervisor), S.M. Skoff (TU Wien, Co-supervisor), M. Schlauff (FHWN, Co-supervisor)

Science Case:

Cryogenic TES-based detectors are versatile for dark matter searches and neutrino physics. We employ them in three experiments: NUCLEUS, CRESST and COSINUS.NUCLEUS will perform precision tests of the Standard Model of particle physics by measuring the coherent neutrino-nucleus scattering at a nuclear power reactor in Chooz, France. The CRESST experiment is specialized on low-threshold dark matter detectors and leads the world-wide search for dark matter particles lighter than 1 GeV. COSINUS uses NaI-based cryogenic detectors to clarify the dark matter claim by DAMA/LIBRA in a material- and model-independent way. COSINUS was inaugurated in 04/2024 at the LNGS underground laboratory in Italy, it will start dark matter data taking in early 2025.

Hypotheses and Objectives:

All these three projects share the same data acquisition (named VDAQ), which is currently developed in the joint group of TU Wien and HEPHY. For purposes of cross-checks, the data are analyzed with different software packages in the collaborations, one of the main two packages (named CAIT) is also developed in the TU Wien/HEPHY group.  This Ph.D. project focuses on the COSINUS dark matter search, but its development may also be used by CRESST and NUCLEUS. The main objective of the Ph.D. work is to enhance the sensitivity of TES-based cryogenic detectors by contributions to data acquisition and raw data analysis. The versatile data acquisition, VDAQ, is an integrated system for the operation of the TES detectors and the acquisition of the resulting signals. As the name implies, it is far more versatile than existing systems which shall be fully exploited with this Ph.D. project.

Approaches and Methods:

• The student will work on the switch from a constant-current biasing of the TES to a modulated bias. One disadvantage of the currently used constant bias is that artificial electrical heater-pulses are needed to determine the operating point of the TES in its transition between normal and superconducting state. With a modulated bias, the TES resistance is continuously measured, making these heater-pulses obsolete, thereby reducing dead time. In addition, a bias modulation might shift the signal to frequency domains less affected by noise. An enhanced signal-to-noise ratio directly improves the sensitivity of the detectors and, thus, directly translates into more reach for the physics cases of the experiments. The hardware of the VDAQ is already able to send a modulated bias. The HEPHY moves to a new location in 05/2024 and foresees establishing a cryogenic laboratory, including a cryostat for the operation of TES-based detectors. This cryostat would become readily available for the Ph.D. student and the R&D of the modulated bias operation. Additional test facilities will be available at collaborating institutes, mainly MPI Munich (CRESST/COSINUS), TU Munich (CRESST/NUCLEUS) and LNGS (CRESST/COSINUS).
• The second pillar of the proposed Ph.D. topic will be raw data analysis in the frame of CAIT. We will record a continuous dead-time-free stream of the data. Then a software trigger is applied offline and the pulse height of each pulse is reconstructed. These tasks are typically done with a so-called optimum filter. This filter is optimal (in the frequency domain) for a fixed pulse shape (with variable amplitude) and fixed noise conditions. The Ph.D. student will extend this filter to work if these conditions are not met. The main potential extension would be to switch from a fixed to an adaptive filter (also called matrix filter) which allows for different baseline models and includes the reconstruction of pile-ups. Also for this pillar, the main objective is reducing dead time, extending the dynamic range and enhancing signal-to-noise for a direct improvement of the dark matter sensitivity of the COSINUS experiment.
• The Ph.D. student will have the chance to apply their methods to data from COSINUS and to publish the results in dedicated articles. The connection of working on acquisition and raw data analysis will lead to in-depth knowledge on the two systems, which is very beneficial for maximum output.
• The long-term perspective is to contribute to the emerging field of quantum sensing with superconducting sensors and our expertise in their long-term application in rare event searches. In this context, natural connections arise to the other Ph.D. topics with a great potential of knowledge transfer because of complementary expertise, experience and skill set.

Supervision:

F. Reindl (TU Wien, Supervisor), W. Treberspurg (FHWN, Co-Supervisor), H. Frais-Kölbl (FHWN, Co-supervisor)

Science case:

Semiconductor sensors with deep sub-electron noise provide the potential to exploit a non-excluded parameter space within direct detection experiments of light dark matter. The scattering process between dark matter candidates and electrons enables a high sensitivity even with detector masses of a few kg.

Hypotheses and Objectives:

Advanced RNDR-DEPFET detectors provide unique properties for dark matter searches w.r.t. the time resolution, which enables an enhanced background rejection. The high level of parallelization and single electron resolution – even at moderate temperatures (app. -40°C) – enables a detector integration in compact systems with a simplified, modular instrument. The high potential of such an instrument for satellite-based applications (e.g. to probe DM candidates in space or for direct detection of exo-planets) shall be investigated. In contrast to CCDs based technologies, DEPFETs are active pixel sensors without significant charge transfer. This results in an improved radiation hardness, which shall be assessed with irradiation studies of prototypes at the MedAustron facility. In this context, displacement damage mechanism on the crystal lattice will be studied, which are a promising candidate to be established as event signature for dark matter searches. 

Approaches and Methods:

• Characterization of advanced RNDR-DEPFETs: Single pixel sensors will be operated and tested to characterize specific features and the impact on DM searches. Those structures include gated pixels to blind the charge collection during readout or to redirect the signal to an additional readout node. Irradiation tests at MedAustron serve to investigate the impact of damage mechanism on the crystal lattice – with deep temperature facilities – on the instrument performance, as well as its potential for another event signature of dark matter. The commissioning and analysis of detectors is supported by the semiconductor laboratory of the Max-Planck Society, which fabricates and designs the sensors.
• Data taking and analysis: Kilo-pixel sensors will be operated to record data within a surface run and analyse them. An evaluation of the background signals concludes in an optimization of the shielding and operation of the detector. The background studies and the calculation of exclusion regions is done in close cooperation with HEPHY experts and PhD colleagues within the doctoral school.
• Model employment of advanced RNDR-DEPFETs: Based on the sensor and background characterization, the achievable improvement of an instrument, that employs advanced RNDR-DEPFETs in a kilo-pixel sensor will be modelled. This esp. includes the operation modes w.r.t. the time resolution. A fast readout, up to full-parallel operation is expected to requires advanced features but enables an improved operation even at non-optimized background environments.

Supervision:

W. Treberspurg (FHWN, Supervisor), J. Schieck (TU Wien, Co-supervisor), C. Scharlemann (FHWN, Co-supervisor)

Literature:

• Bähr, A., Kluck, H., Ninkovic, J. et al. DEPFET detectors for direct detection of MeV Dark Matter particles. Eur. Phys. J. C 77, 905 (2017). https://doi.org/10.1140/epjc/s10052-017-5474-5
• W. Treberspurg, A. Bähr, H. Kluck, P. Lechner, J. Ninkovic, J. Treis, H. Shi, J. Schieck, "Performance of a kilo-pixel RNDR-DEPFET detector," Proc. SPIE 12191, X-Ray, Optical, and Infrared Detectors for Astronomy X, 1219119 (2022); https://doi.org/10.1117/12.2629248

Science Case:

Experiments at CERN's Large Hadron Collider (LHC) probe Nature at the fundamental level. At the Compact Muon Solenoid (CMS) experiment, particles collide at a rate of 40 MHz – much more often than can be recorded. For this reason, a sophisticated set of filters, the so-called trigger, is deployed, selecting only events of interest from the stream of events and discarding the others. The two-stage trigger system first applies relatively simple rules at the hardware level, then more complicated ones at the software level. At the end, collision events are recorded at a rate of about 1kHz. 

Machine learning (ML) has started to play an important role in these steps, but more powerful algorithms are usually too slow, and compromises must be made.

Hypotheses and Objectives:

Recent advancements in machine learning and neural networks employ low-precision architectures that drastically simplify the computations needed for general nonlinear multivariate algorithms. Order-of-magnitude reductions in the power consumption of large-language models (LLMs) and graph-neural networks (GNNs) have been demonstrated with almost no loss in the predictive power of the trained model. The simplicity of the underlying binary and ternary computations makes FPGAs the ideal framework for capitalizing on the increased power efficiency in the context of high-throughput applications. 

Approaches and Methods:

In this thesis, the successful applicant will develop advanced binary- and ternary-precision machine-learning algorithms to be used for event and object reconstruction in the selection and reconstruction of data recorded by the CMS experiment at the LHC, including the critical stage of online event selection where FPGA applications play a decisive role.

• With the HEPHY ML group, appropriate algorithms will be selected and trained, and tools for benchmarking and validation will be developed.
• Within the HEPHY CMS group, the relevant data sets will be identified, the ML data representation will be defined, and the performance will be compared to the existing online selection algorithms.
• The model will be tested with an FPGA emulator framework, deployed on FPGAs, and if it is sufficiently performant, it will become part of CMS’ future data-taking.

Supervision:

Claudia-Elisabeth Wulz (TU Wien, Supervisor), Robert Schöfbeck (TU Wien, Co-Supervisor), Claudius Krause (HEPHY, Co-Supervisor), Helmut Frais-Kölbl (FHWN, Co-Supervisor) 

Our Project:

We want to investigate how the control of atomic interactions can improve the performance of quantum sensing devices such as clocks or trapped atom interferometers. We chose Caesium as our matter wave source, since the atomic interactions can be tuned with static magnetic or microwave fields. We want to develop a detector with a high quantum efficiency in the near-IR, in order to measure atom numbers accurately, ideally with single atom resolution, and so infer the properties of the prepared quantum states.

Our goals:

Typical experimental setups to perform such scientific projects are highly-specialized, one-of-a-kind apparatuses that allow one to carry out a specific experiment. The cold-atom community is following this approach successfully with ever-increasing control over complex quantum systems, but future applications in sensing and metrology call for integrability, scalability, and potentially parallel operation. We aim at designing a detector, that has versatile applications in quantum metrology but can as well be used to investigate quantum systems that are of fundamental interest.

Our Methods:

• We will start from existing detectors at FWHN and to find the favorable solution for fluorescence and absorption imaging of Caesium atoms at 852 nm. First, we will pick the best type of detector, with respect to quantum efficiency, detection area and readout rate. The detector will be integrated with optical trapping for fluorescence detection under an atom chip.
• We will collaborate with the other members of Scies4free on readout electronics and statistical data analysis. We will investigate if and how the new detector can enhance the performance of quantum sensing, by comparing the detected photonic signal to the expected atom number and its quantum statistics.
• Once the detector is fully integrated with our experimental setup, we will perform fluorescence detection of optically trapped Caesium atom. The results will be gained by measuring collision and loss rates, which both require atom number counting.

Supervision:

T. Schumm (TU-Wien, Supervisor), S. Manz (TU Wien, Co-supervisor), C. Koller (FHWN, Co-supervisor)