Laser Annealing at 405 nm: MRID Watches Perovskite Crystallize Live
No sample swap, no offline check — MRID anneals with a 405 nm laser and reads with a 532 nm laser, in situ, in real time.
Perovskite microwires (MWs) are having a moment in flexible optoelectronics — thinner, higher surface-to-volume ratio, and better photoelectric conversion than the equivalent perovskite film. But like any perovskite material, their performance lives and dies by crystallinity. A poorly crystallized wire, full of disordered domains and degradation byproducts, simply doesn’t perform, no matter how promising the chemistry is on paper.
Laser annealing — using a focused beam to rapidly heat and recrystallize a material — has been explored as a fix. It’s fast, localized, and doesn’t require an oven. The catch has always been knowing whether it worked. Most prior laser-annealing studies checked results with a solar simulator or a UV-Vis spectrophotometer, run separately, after the fact, on a sample that had already been pulled off the annealing stage. That means every annealing “recipe” — laser power, wavelength, exposure time, scan speed — had to be guessed at, applied, and only judged good or bad after the fact.
The Obstacle
Two separate problems compounded here. First, wavelength choice: earlier laser-annealing work on CH3NH3PbI3 perovskites leaned on near-infrared sources — 1064 nm Nd:YAG lasers, for instance — sitting far outside the perovskite’s own absorption band, meaning much of that laser energy simply isn’t absorbed where it’s needed. Second, and more fundamental: with annealing and characterization split across two separate instruments, there’s no way to watch the crystallization happen, only to check what state it landed in afterward, and to do so blind to how the process unfolded along the way.
The Laser Annealing Innovation
Xiaoming Chen, Zixian Wang, Ren-Jie Wu, Horng-Long Cheng, and Hsiang-Chen Chui (Dalian University of Technology and National Cheng Kung University) built a single optical system that does both jobs through the same objective lens, published in Photonics (Chen et al., 2021, DOI: 10.3390/photonics8020030).
A 405 nm violet diode laser — chosen specifically because it sits near the perovskite’s own absorption peak, unlike the off-resonance NIR lasers used previously — was combined into ProTrusTech’s MRID polarized Raman system as a dedicated annealing source. An electronic shutter controlled exposure time, a wheeled neutral-density filter controlled power, and a 2D precision motor-driven stage let the beam scan a precisely defined area. MRID’s own 532 nm Raman excitation laser, meanwhile, was kept at just 0.5 mW — low enough to read the crystal structure without perturbing it. Anneal with one beam, measure with the other, through the same optics, on the same stage, without moving the sample at all.
Key Results
Once the combined annealing-and-readout system was in place, the team first pinned down what the annealing laser was actually capable of — its focused spot size, delivered power, and the point where it starts to damage the microwire — before turning to the crystallization result itself: a Raman intensity ratio that rose measurably after a single annealing pass, reproduced across a scanned area and confirmed on two different substrates.
The Annealing Laser Is Precisely Characterized, Not Just Switched On
Focused through a 100X objective, the 405 nm beam measured 2.6 µm in diameter, and delivered 16 mW to the sample from a 36 mW source. At 6 kW/mm2 for a 10-second exposure, the beam was intense enough to visibly burn the microwire — establishing a clear upper bound. A thermal infrared camera pinned the estimated damage threshold at 53 kJ/mm2, corresponding to roughly 110°C on a glass-mounted sample, with the heating curve running measurably higher on glass than on flexible PET film.
One Raman Ratio Captures Crystallization — and It Rose From 0.81 to 1.07
The spectrum of MAPbI3 carries a band near 105 cm-1 associated with organic-cation motion within an intact inorganic cage — a marker of good crystal quality — alongside a band near 97 cm-1 linked to a PbI2 degradation byproduct. The I105/I97 intensity ratio became the paper’s crystallization quality coefficient, and at a controlled 17 mW annealing power and 5-second exposure, that ratio rose from 0.81 before annealing to 1.07 afterward — read directly off the same Raman system that did nothing but sit and watch, no sample transfer required.
The Same Result Scales From One Point to a Mapped Area
Scanning the 405 nm beam across a programmable 30 µm × 30 µm region (0.5 µm step, 0.5 µm/s scan speed) produced a full Raman map of the I105/I97 ratio before and after annealing — turning a single-point proof of concept into a demonstrated area-crystallization technique.
The process also holds up across substrates: the same annealing-and-Raman-readout sequence was demonstrated on perovskite microwires coated on both glass and PET film, with the heating curve differing predictably between the two — a direct data point for anyone considering this for flexible, wearable perovskite devices rather than only rigid ones.
Taken together, the beam is characterized, the improvement is quantified with a single Raman ratio, and that improvement holds whether measured at one point, across a mapped area, or on a different substrate entirely.
Why ProTrusTech Is Featuring This Research
While MRID case study has been about measurement, only one of MRID’s two lasers is actually reading anything in this case.
The 405 nm inducing laser sits in a fixed position in the optical path, dedicated to in situ annealing or photothermal work. The second laser slot — the one doing the spectral acquisition, running at 532 nm in this paper — is swappable, so a lab can drop in a 633 nm module instead if their sample needs a different Raman excitation wavelength (to dodge fluorescence background, or to hit a different resonance condition), without ever touching or realigning the fixed 405 nm induction path.
Anneal and measure stay decoupled by design: one fixed laser doing the inducing, one swappable laser doing the reading.
That decoupling is also why this works at all without the kind of microsecond-level synchronization a pump-probe experiment would need. Thermal annealing is a one-way, cumulative process — the crystal structure doesn’t relax back once it’s improved — so the 405 nm and 532 nm beams don’t need to be triggered in lockstep. They just need to share a beam path and a stage, which is exactly what this optical layout, and MRID’s built-in architecture, provide.
Engineering the Solution
The practical elegance here is that nothing about this setup requires exotic hardware. A diode laser, an electronic shutter, a neutral-density filter wheel, and a 2D motor stage — all fairly standard optical bench components — layered onto an existing polarized Raman microscope.
The result is a closed loop: anneal a spot or scan an area, read the Raman signal immediately, adjust laser power or exposure time based on what the I105/I97 ratio says, and repeat — all without ever unmounting the sample. That loop is the actual innovation, more than any single laser parameter.
Featured Instruments:
Why This Matters for Thin-Film Crystallization
The generalizable claim in this paper’s own conclusion is broader than perovskite microwires specifically: this same in situ Raman-monitored laser annealing scheme is proposed as extensible to other laser-based material processing more broadly, and the authors specifically flag elliptical beam shaping combined with online Raman as a path toward large-area processing.
For anyone developing flexible or wearable perovskite electronics, or other organic-inorganic hybrid optoelectronic materials, that combination — real-time feedback plus scalable beam geometry — is what turns a lab demonstration into something closer to a production-line inspection step.
What’s Next
- Quantifying “how much” crystallization improved, not just that it did. The authors are explicit about this limitation in their own conclusion: the I105/I97 ratio shows direction and relative improvement, but not yet an absolute crystallinity percentage — that quantification work is described as still ongoing.
- Elliptical beam shaping for large-area processing. The paper flags elliptical laser beams combined with online Raman spectroscopy as the proposed route to scaling this from a 30 µm × 30 µm demonstration area up to production-relevant large-area processing.
- Extension to other laser-based material fabrication. The authors frame the in situ annealing-plus-Raman-readout scheme as a general design pattern, not one limited to perovskite microwires specifically.
If you’re developing laser annealing, thermal processing, or crystallization workflows, for perovskites or any other material where you need to know the result before you move the sample, we’d like to hear what you’re working on.
Key Takeaways
- A 405 nm violet annealing laser, chosen to sit near the perovskite’s own absorption peak, was combined into the same beam path as ProTrusTech’s MRID polarized Raman system, so crystallization could be induced and read out without moving the sample.
- The Raman intensity ratio I105/I97 (crystal-quality marker over degradation marker) rose from 0.81 to 1.07 after a single annealing pass at 17 mW for 5 seconds, and the same improvement was demonstrated as a full spatial map across a 30 µm × 30 µm scanned area.
- The annealing laser’s operating envelope was precisely characterized — 2.6 µm beam diameter, 16 mW delivered power, an estimated 53 kJ/mm2 / 110°C damage threshold — with the process shown working on both rigid glass and flexible PET substrates.
- MRID’s role here was a fixed 405 nm inducing laser paired with a swappable spectral-acquisition laser (532 nm in this paper, interchangeable with 633 nm), letting the annealing hardware stay untouched while the measurement side adapts to the sample.
