arXiv subjects

Sebastián Roca-Jerat

Publications and source records attributed to Sebastián Roca-Jerat.

9 recordsLinked to original sources

arXiv2026

Polariton-polariton coherent coupling in a molecular spin-superconductor chip

The ability to establish coherent communication channels is key for scaling up quantum devices. Here, we engineer interactions between distant polaritons, hybrid spin-photon excitations formed at different lumped-element superconducting resonators within a chip. The chip consists of several resonator pairs, slightly detuned in frequency to make them addressable, capacitively coupled within each pair and inductively coupled to a common readout line. They interact locally with samples of PTMr and Tripak$^{-}$ organic free radicals, deposited onto their inductors, which provide model $S = 1/2$, $g \simeq 2$ spin ensembles. Frequency-dependent microwave transmission experiments, performed at very low temperatures, measure polariton frequencies as a function of magnetic field in different scenarios. When only one resonator within a pair hosts a molecular sample, the results evidence that spins couple remotely to the empty LER as well as to the local cavity mode. If both resonators interact with a spin ensemble, the magnetic field tunes the polariton frequencies relative to each other, on account of the different spin-photon interactions at each LER. When polaritons are brought into mutual resonance, an avoided level crossing emerges that gives direct spectroscopic evidence for a coherent polariton-polariton interaction mediated by the circuit. Pump-probe experiments reveal that the excitation of a polariton within a connected pair is felt, thus it can be read out, by the other one. These observations, backed by model calculations, illustrate the control and detection of distant photon-photon and spin-spin correlations and entanglement in a scalable modular chip.

quant-ph↗

arXiv2026

Optimizing magnetic coupling in lumped element superconducting resonators for molecular spin qubits

We engineer lumped-element superconducting resonators that maximize magnetic coupling to molecular spin qubits, achieving record single-spin couplings up to $100$ kHz and collective couplings exceeding $10$ MHz. The resonators interact with PTMr organic free radicals, model spin systems with $S=1/2$ and a quasi-isotropic $g \simeq 2$, dispersed in polymer matrices. The highest collective spin-photon coupling strengths are attained with resonators having large inductors, which therefore interact with most spins in the molecular ensemble. By contrast, the coupling of each individual spin $G_{1}$ is maximized in resonators having a minimum size inductor, made of a single wire. The same platform has been used to study spin relaxation and spin coherent dynamics in the dispersive regime, when spins are energetically detuned from the resonator. We find evidences for the Purcell effect, i.e. the photon induced relaxation of those spins that are most strongly coupled to the circuit. The rate of this process gives access to the distribution of single spin photon couplings in a given device. For resonators with a $50$ nm wide constriction at the inductor center, single maximum $G_{1}$ values reach $\sim 100$ kHz. Pumping the spins with strong pulses fed through an independent transmission line induces coherent Rabi oscillations. The spin excitation then proceeds via either direct resonant processes induced by the main pulse frequency or, in the case of square-shaped pulses, via the excitation of the cavity by sideband frequency components. The latter process measures the cavity mode hybridization with the spins and can be eliminated by using Gaussian shaped pulses. These results establish a scalable route toward integrated molecular-spin quantum processors.

quant-ph↗

arXiv2025

Accelerating Quantum Monte Carlo Calculations with Set-Equivariant Architectures and Transfer Learning

Machine-learning (ML) ansätze have greatly expanded the accuracy and reach of variational quantum Monte Carlo (QMC) calculations, in particular when exploring the manifold quantum phenomena exhibited by spin systems. However, the scalability of QMC is still compromised by several other bottlenecks, and specifically those related to the actual evaluation of observables based on random deviates that lies at the core of the approach. Here we show how the set-transformer architecture can be used to dramatically accelerate or even bypass that step, especially for time-consuming operators such as powers of the magnetization. We illustrate the procedure with a range of examples structured around quantum spin systems with long-range interactions, and comprising both regressions (to predict observables) and classifications (to detect phase transitions). Moreover, we show how transfer learning can be leveraged to reduce the training cost by reusing knowledge from different systems and smaller system sizes.

quant-ph↗

arXiv2025

Waveguide quantum electrodynamics at the onset of spin-spin correlations

We explore the competition between light-mediated and intrinsic matter-matter interactions in waveguide quantum electrodynamics. For this, we couple a superconducting transmission line to a model magnetic material, made of organic free radical molecules with a spin $S=1/2$ and a $g_{S}$ factor very close to that of a free electron. The microwave transmission has been measured in a wide range of temperatures ($0.013$ K $\leq T \leq 2$ K), magnetic fields ($0\leq B \leq 0.5$ T) and frequencies ($0 \leq ω/2 π\leq 14$ GHz). We find that molecules belonging to one of the two crystal sublattices form one-dimensional spin chains. Temperature then controls the intrinsic correlations along these chains in a continuous and monotonic way. In the paramagnetic region ($T > 0.7$ K), the microwave transmission shows evidences for the collective coupling of quasi-identical spins to the propagating photons, with coupling strengths that reach values close to the dissipation rates. As $T$ decreases, the growth of spin correlations, combined with the anisotropy in the spin-spin exchange constants, tend to suppress the collective spin-photon coupling. In this regime, the spin visibility in transmission reflects also a gradual change in the nature of the dominant spin excitations, from single spin flips to bosonic magnons.

quant-ph↗

arXiv2025

A genetic algorithm to generate maximally orthogonal frames in complex space

A frame is a generalization of a basis of a vector space to a redundant overspanning set whose vectors are linearly dependent. Frames find applications in signal processing and quantum information theory. We present a genetic algorithm that can generate maximally orthogonal frames of arbitrary size $n$ in $d$-dimensional complex space. First, we formalize the concept of maximally orthogonal frame and demonstrate that it depends on the choice of an energy function to weigh the different pairwise overlaps between vectors. Then, we discuss the relation between different energy functions and well-known frame varieties such as tight and Grassmannian frames and complex projective $p$-designs. Obtaining maximally orthogonal frames poses a global non-convex minimization problem. We discuss the relation with established numerical problems such as the Thomson problem and the problem of finding optimal packings in complex projective space. To tackle the minimization, we design a hybrid genetic algorithm that features local optimization of the parents. To assess the performance of the algorithm, we propose two visualization techniques that allow us to analyze the coherence and uniformity of high-dimensional frames. The genetic algorithm is able to produce highly-symmetric universal frames, such as equiangular tight frames, symmetric, informationally complete, positive operator-valued measurements (SIC-POVMs) and maximal sets of mutually unbiased bases, for configurations of up to $d=6$ and $n=36$, with runtimes of the order of several minutes on a regular desktop computer for the largest configurations.

physics.comp-ph↗

arXiv2025

Qudit Machine Learning

We present a comprehensive investigation into the learning capabilities of a simple d-level system (qudit). Our study is specialized for classification tasks using real-world databases, specifically the Iris, breast cancer, and MNIST datasets. We explore various learning models in the metric learning framework, along with different encoding strategies. In particular, we employ data re-uploading techniques and maximally orthogonal states to accommodate input data within low-dimensional systems. Our findings reveal optimal strategies, indicating that when the dimension of input feature data and the number of classes are not significantly larger than the qudit's dimension, our results show favorable comparisons against the best classical models. This trend holds true even for small quantum systems, with dimensions d<5 and utilizing algorithms with a few layers (L=1,2). However, for high-dimensional data such as MNIST, we adopt a hybrid approach involving dimensional reduction through a convolutional neural network. In this context, we observe that small quantum systems often act as bottlenecks, resulting in lower accuracy compared to their classical counterparts.

quant-ph↗

arXiv2025

First- and second-order quantum phase transitions in the long-range unfrustrated antiferromagnetic Ising chain

We study the ground-state phase diagram of an unfrustrated antiferromagnetic Ising chain with longitudinal and transverse fields in the full range of interactions: from all-to-all to nearest-neighbors. First, we solve the model analytically in the strong long-range regime, confirming in the process that a mean-field treatment is exact for this model. We compute the order parameter and the correlations and show that the model exhibits a tricritical point where the phase transition changes from first to second order. This is in contrast with the nearest-neighbor limit where the phase transition is known to be second order. To understand how the order of the phase transition changes from one limit to the other, we tackle the analytically-intractable interaction ranges numerically, using a variational quantum Monte Carlo method with a neural-network-based ansatz, the visual transformer. We show how the first-order phase transition shrinks with decreasing interaction range and establish approximate boundaries in the interaction range for which the first-order phase transition is present. Finally, we establish that the key ingredient to stabilize a first-order phase transition and a tricritical point is the presence of ferromagnetic interactions between spins of the same sublattice on top of antiferromagnetic interactions between spins of different sublattices. Tunable-range unfrustrated antiferromagnetic interactions are just one way to implement such staggered interactions.

quant-ph↗

arXiv2024

Transformer Wave Function for Quantum Long-Range models

We employ a neural-network architecture based on the Vision Transformer (ViT) architecture to find the ground states of quantum long-range models, specifically the transverse-field Ising model for spin-1/2 chains across different interaction regimes. Harnessing the transformer's capacity to capture long-range correlations, we compute the full phase diagram and critical properties of the model, in both the ferromagnetic and antiferromagnetic cases. Our findings show that the ViT maintains high accuracy across the full phase diagram. We compare these results with previous numerical studies in the literature and, in particular, show that the ViT has a superior performance than a restricted-Boltzmann-machine-like ansatz.

quant-ph↗

arXiv2023

Circuit Complexity through phase transitions: consequences in quantum state preparation

In this paper, we analyze the circuit complexity for preparing ground states of quantum many-body systems. In particular, how this complexity grows as the ground state approaches a quantum phase transition. We discuss different definitions of complexity, namely the one following the Fubini-Study metric or the Nielsen complexity. We also explore different models: Ising, ZZXZ or Dicke. In addition, different forms of state preparation are investigated: analytic or exact diagonalization techniques, adiabatic algorithms (with and without shortcuts), and Quantum Variational Eigensolvers. We find that the divergence (or lack thereof) of the complexity near a phase transition depends on the non-local character of the operations used to reach the ground state. For Fubini-Study based complexity, we extract the universal properties and their critical exponents. In practical algorithms, we find that the complexity depends crucially on whether or not the system passes close to a quantum critical point when preparing the state. For both VQE and Adiabatic algorithms, we provide explicit expressions and bound the growth of complexity with respect to the system size and the execution time, respectively.

quant-ph↗