Detecting Magnetic Phase Transitions in Ion-Irradiated CrSBr Through Resonant Raman Scattering

Abstract: Controlling magnetic phases and accurately determining their transition temperatures are essential for the development of low-dimensional magnetic materials. Here, we demonstrate that He$<sup>+$ ion irradiation provides a versatile route for engineering magnetic phases in layered CrSBr and establish temperature-dependent polarization-resolved Raman spectroscopy as a sensitive optical probe for identifying irradiation-induced magnetic phase transitions. We reveal that the magnetic response of the modified CrSBr is governed by both irradiation dose and crystal thickness. The temperature evolution of the Raman tensor elements resolves the antiferromagnetic transition in pristine CrSBr at T$_N \approx 132$ K as well as irradiation-induced magnetic transitions at T$_C \approx 105-110$ K and T$_D \approx 40$ K corresponding to ferromagnetic and defect-related magnetic phase transitions. Complementary magneto-optical measurements confirm the progressive suppression of antiferromagnetic order and the emergence of new defect-engineered magnetic phases, including pure ferromagnetic behavior at high irradiation doses. These findings establish irradiated CrSBr as a platform for controllable magnetic phase engineering while demonstrating polarization-resolved Raman spectroscopy as a rapid, non-destructive, and broadly applicable method for probing magnetic phase transitions in van der Waals magnets.

• The paper demonstrates that polarization-resolved resonant Raman spectroscopy detects magnetic transitions through discontinuities in Raman tensor elements, revealing pristine Néel order near 132 K and irradiation-induced transitions near 105–110 K and 40 K.
• He⁺ irradiation progressively weakens interlayer antiferromagnetism in CrSBr, producing modified antiferromagnetic states at low doses and full ferromagnetism at a double dose in bulk regions, as confirmed by magneto-optical measurements.
• Conventional phonon frequency and linewidth analysis shows no transition anomalies, while the Raman tensor response remains sensitive to dose- and thickness-dependent phase changes, although the microscopic origin of the 40 K defect-related transition remains unresolved.

Overview

This paper demonstrates two coupled results: He$^+$ ion irradiation can be used to engineer magnetic phases in the layered antiferromagnet CrSBr, and temperature-dependent polarization-resolved Raman spectroscopy—specifically, tracking discontinuities in the Raman tensor elements rather than phonon frequencies or linewidths—provides a sensitive, non-destructive optical probe of both intrinsic and irradiation-induced magnetic phase transitions (2608.18909). The authors identify the Néel transition in pristine CrSBr at $T_N \approx 132$ K and irradiation-induced transitions at $T_C \approx 105$–110 K (paramagnetic to modified AFM/FM) and $T_D \approx 40$ K (defect-related), with magneto-optical spectroscopy confirming progressive suppression of AFM order and emergence of ferromagnetic behavior at high doses.

Irradiation scheme and defect physics

CrSBr orders ferromagnetically within each layer and antiferromagnetically between layers via Br-$p$-orbital-mediated super-superexchange; monolayer CrSBr is intrinsically ferromagnetic. He$^+$ irradiation introduces Frenkel pairs (vacancy plus interstitial) for all three constituent species, with interstitials occupying van der Waals gap positions where they form additional covalent bonds between adjacent layers. For Cr interstitials, first-nearest-neighbor interlayer exchange becomes mediated by hybridization with inner Br atoms, weakening and eventually overcoming the AFM interlayer coupling. This microscopic picture underpins the observed dose-dependent phase control.

Samples were irradiated with combined 8 keV and 1 keV He$^+$ beams at fluences yielding uniform displacement-per-atom (DPA) levels of 0.01 ("single-dose") and 0.02 ("double-dose"). SRIM simulations show a key structural asymmetry: DPA is significantly lower in the first few atomic layers than deeper in the crystal, because high-energy ions near the surface undergo fewer nuclear collisions. This depth profile directly explains the thickness dependence of all subsequent results.

Raman tensor anomalies as transition markers

Raman measurements were performed in co-polarized geometry with 4° angular resolution from 4 to 200 K under near-resonant 1.96 eV excitation, on few-layer (8–9 nm) and bulk (30–90 nm) regions. The angular profiles were fit to extract the Raman tensor amplitudes $a$ and $b$ along the crystallographic axes, including a complex phase $\phi_{ab}$ accounting for virtual/real intermediate-state interference.

The central experimental finding is that the $T_N \approx 132$0 component of the Raman tensor exhibits sharp discontinuities at magnetic transition temperatures: a dip at $T_N \approx 132$1 K for pristine samples, and dips at $T_N \approx 132$2–110 K and $T_N \approx 132$3 K for irradiated samples. Notably, signatures of $T_N \approx 132$4 and $T_N \approx 132$5 were never observed simultaneously in the same sample, indicating that irradiation fully replaces the intrinsic AFM ordering rather than coexisting with it.

The anomaly magnitude increases with irradiation dose but decreases with sample thickness. The authors attribute the latter to the DPA depth profile: although bulk flakes experience stronger modification of exchange interactions, their vibrational response may be partially suppressed by the higher defect density deeper in the crystal. This explanation is plausible but not independently verified, and rests on qualitative SRIM estimates using an assumed 25 eV displacement energy for which no experimental values exist for CrSBr.

A methodologically important negative result: phonon energies decrease monotonically with temperature and show no discernible anomalies at any transition temperature, in contrast to previous reports that used Raman shifts to track spin-phonon coupling (2608.18909). The paper also notes that second-order Raman features between 400 and 800 cm$T_N \approx 132$6, previously associated with intralayer FM order, are strongly suppressed by irradiation and cannot be used for phase analysis. The sensitivity of this technique therefore resides specifically in the resonant Raman tensor response, not in conventional Raman shift analysis.

Magneto-optical confirmation

Photoluminescence measurements at 4.7 K with field along the $T_N \approx 132$7-axis corroborate the Raman findings:

Sample	Saturation field $T_N \approx 132$8	PL energy shift	Interpretation
Pristine (5 nm)	±0.37–0.41 T	~13 meV	AFM ground state
Single-dose, few-layer	±0.19–0.20 T	~4 meV	Modified AFM
Single-dose, bulk	±0.09–0.11 T	3–4 meV	Strongly modified AFM
Double-dose, bulk	none observed	—	Fully ferromagnetic

The saturation fields and PL shifts of single-dose samples are more than a factor of two smaller than pristine values, indicating substantial but incomplete conversion of the AFM state. The double-dose bulk region shows no field-induced PL signature, consistent with complete ferromagnetic ordering, while the double-dose few-layer region retains a barely discernible field response—directly consistent with the lower surface DPA. The zero-field PL energy of the double-dose few-layer region is blueshifted ~20 meV relative to bulk, reflecting thickness- and dose-dependent electronic structure modification. Power-law exponents $T_N \approx 132$9–1.021 confirm linear PL intensity dependence without saturation up to 1000 μW, ruling out excitation-power artifacts.

Detection mechanism

The physical basis for the Raman sensitivity is that magnetic ordering modifies the intermediate electronic states entering the resonant Raman amplitude—excitonic energies, damping, and electron-phonon matrix elements all change across a transition. Under near-resonant excitation, the FM transition breaks time-reversal symmetry and produces exchange-field-induced Zeeman-like splitting of excitonic resonances, so the Raman tensor components acquire characteristic discontinuities at $T_C \approx 105$0, $T_C \approx 105$1, and $T_C \approx 105$2. The mode-selective response—all three A$T_C \approx 105$3 modes respond at $T_C \approx 105$4, only A$T_C \approx 105$5 and A$T_C \approx 105$6 at $T_C \approx 105$7, and A$T_C \approx 105$8 only in the strongest-anomaly sample—is attributed to selective coupling of different modes to different electronic states, modified by irradiation-induced changes in band structure. This mode selectivity supports the interpretation but also means the technique's applicability depends on favorable exciton-phonon resonance conditions specific to each material.

Limitations and open questions

Several caveats are stated explicitly. The nature of the $T_C \approx 105$9 K transition remains unresolved: candidate explanations include ferromagnetic ordering of dilute defects, spin freezing from suppressed spin fluctuations, or a crossover from XY to triaxial anisotropy. The authors note that because exchange interactions act over atomic length scales shorter than typical defect separations, the very existence of a collective defect-driven transition at this dilute concentration is difficult to reconcile, leaving the question open. Second, the assignment of the broad feature near A$T_D \approx 40$0 is ambiguous—it could be defect-related, or the B$T_D \approx 40$1 mode, since a similar peak appears in pristine samples below $T_D \approx 40$2. Third, the DPA calculations are explicitly qualitative upper-limit estimates given the assumed displacement energies. Finally, the proposed mechanism for reduced Raman anomalies in thick flakes (vibrational suppression by high defect density) is inferred rather than directly measured.

Conclusion

The paper establishes that low-dose He$T_D \approx 40$3 irradiation converts layered CrSBr from an AFM semiconductor into dose- and thickness-tunable magnetic states culminating in full ferromagnetism, and that polarization-resolved Raman tensor analysis under resonant excitation detects the associated transitions at temperatures where conventional Raman shift analysis fails. The main open question left by the work is the microscopic origin of the 40 K defect-related transition, which will require probes sensitive to local defect magnetism to resolve.