arXiv subjects

Pan Yue

Publications and source records attributed to Pan Yue.

4 recordsLinked to original sources

arXiv2026

Reconfigurable Terahertz Multi-Harmonic Dual-Combs

Dual-comb spectroscopy, constructed from two frequency combs with slightly different repetition frequencies, enables real-time and high-precision measurements without mechanical scanning. In the terahertz spectral range, dual-comb techniques provide a powerful tool for high-resolution and rapid spectroscopy. Quantum cascade lasers (QCLs), owing to their compact footprint and favorable size, weight, and power, have emerged as promising sources for THz dual-comb systems. In QCL frequency combs, the repetition frequency generated through intrinsic four-wave mixing is typically equal to the cavity round-trip frequency, corresponding to the fundamental comb. Although this repetition frequency can be tuned via current and temperature control, the accessible tuning range remains limited. Recently, harmonic frequency combs, whose repetition frequencies are integer multiples of the cavity round-trip frequency, have attracted increasing attention, offering enhanced single-line signal-to-noise ratios and providing direct insight into the strong nonlinearity of QCLs. Here, we demonstrate a reconfigurable multi-harmonic dual-comb system realized on a single self-detected THz QCL platform. By precisely controlling the driving current and thermal conditions, we achieve and switch between multiple dual-comb configurations, including fundamental-fundamental, fundamental-second-harmonic, second-harmonic-second-harmonic, and second-harmonic-third-harmonic dual-combs. These results establish harmonic order as an additional degree of freedom for dual-comb operation within a single laser system. The demonstrated platform enables high-resolution and high-sensitivity measurements while significantly simplifying the system architecture. This reconfigurable multi-band dual-comb approach based on a single QCL opens new opportunities for compact THz spectroscopy and frequency metrology.

physics.optics↗

arXiv2026

Growth and hydrostatic-pressure study of a type-II superconductor Bi$_2$Ta$_3$S$_6$ single crystal

We report the growth and physical properties of single-crystalline Bi$_2$Ta$_3$S$_6$ crystallizing in $P6_3/mcm$ space group, which comprises alternating Ta-S layers and Bi layers with each Bi atom connected with adjacent S atoms. Temperature-dependent electrical resistivity measurements reveal a superconducting transition at 0.84 K, with upper critical field 231 Oe under an out-of-plane magnetic field. The magnetization measurements confirm its nature as a type-II superconductor, with anisotropic Ginzburg-Landau parameter $\kappa_{ab}$ = 7.67 and $\kappa_c$ = 4.50. Hall measurements indicate the dominant carriers as hole. Hydrostatic pressure is applied, under which both the superconducting transition temperature and upper critical field increase sharply under low pressure before undergoing slight suppression under higher pressure. Density functional theory calculations reveal non-trivial topological surface states on (100) surface in Bi$_2$Ta$_3$S$_6$, which may offer a new avenue for exploring potential topological superconductivity in layered transition metal dichalcogenides.

cond-mat.supr-con↗

arXiv2025

Grammar-Guided Evolutionary Search for Discrete Prompt Optimisation

Prompt engineering has proven to be a crucial step in leveraging pretrained large language models (LLMs) in solving various real-world tasks. Numerous solutions have been proposed that seek to automate prompt engineering by using the model itself to edit prompts. However, the majority of state-of-the-art approaches are evaluated on tasks that require minimal prompt templates and on very large and highly capable LLMs. In contrast, solving complex tasks that require detailed information to be included in the prompt increases the amount of text that needs to be optimised. Furthermore, smaller models have been shown to be more sensitive to prompt design. To address these challenges, we propose an evolutionary search approach to automated discrete prompt optimisation consisting of two phases. In the first phase, grammar-guided genetic programming is invoked to synthesise prompt-creating programmes by searching the space of programmes populated by function compositions of syntactic, dictionary-based and LLM-based prompt-editing functions. In the second phase, local search is applied to explore the neighbourhoods of best-performing programmes in an attempt to further fine-tune their performance. Our approach outperforms three state-of-the-art prompt optimisation approaches, PromptWizard, OPRO, and RL-Prompt, on three relatively small general-purpose LLMs in four domain-specific challenging tasks. We also illustrate several examples where these benchmark methods suffer relatively severe performance degradation, while our approach improves performance in almost all task-model combinations, only incurring minimal degradation when it does not.

cs.CL↗

arXiv2016

When is $a^{n} + 1$ the sum of two squares?

Using Fermat's two squares theorem and properties of cyclotomic polynomials, we prove assertions about when numbers of the form $a^{n}+1$ can be expressed as the sum of two integer squares. We prove that $a^n + 1$ is the sum of two squares for all $n \in \mathbb{N}$ if and only if $a$ is a perfect square. We also prove that for $a\equiv 0,1,2\pmod{4},$ if $a^{n} + 1$ is the sum of two squares, then $a^δ + 1$ is the sum of two squares for all $δ| n, \ δ>1$. Using Aurifeuillian factorization, we show that if $a$ is a prime and $a\equiv 1 \pmod{4}$, then there are either zero or infinitely many odd $n$ such that $a^n+1$ is the sum of two squares. When $a\equiv 3\pmod{4},$ we define $m$ to be the least positive integer such that $\frac{a+1}{m}$ is the sum of two squares, and prove that if $a^n+1$ is the sum of two squares for any odd integer $n,$ then $m | n$, and both $a^m+1$ and $\frac{n}{m}$ are sums of two squares.

math.NT↗