Guangdong Juyao Intelligent Technology Co., Ltd. · Laser Software & Process R&D
23°07′N 113°15′E · GUANGZHOU · PURE STATIC MONOGRAPH
← Back to news
Research Watch2026.11.01· 3 min read

Research Watch: Ultrashort laser pulses slice diamond along its difficult plane — an 8×8 mm CVD diamond split cleanly on {100}

Diamond is excellent at everything except being sliced into thin wafers — and the plane it most wants to be cut along happens to be the hardest to crack. The Chiba team’s answer is not to “cut harder” but to “lay signposts for the crack inside the crystal first”: ultrashort pulses write a grid of modified points in the bulk, then let the crack follow the signposts for the final leg. At 8×8 mm, parted with a light tap, that route has reached a usable size for the first time.

JUYAO editorial visual: tweezers hold a transparent faceted diamond crystal whose interior carries a glowing grid of modification points, a thin laser beam enters from the side, and a slice has separated below along the glowing plane
COVERJUYAO editorial visual (concept schematic, not a measured photograph or a paper figure).
EXECUTIVE SUMMARY

On 1 November, Professor Hirofumi Hidai’s team at Chiba University, Japan, published a study in Diamond and Related Materials: first comparing how three single-crystal diamonds — one grown by high-pressure high-temperature (HPHT) synthesis and two by chemical vapour deposition (CVD), with different nitrogen impurity levels — respond to ultrashort laser pulses, then turning the findings into a large-area slicing demonstration. Irradiating samples with focused pulses from 233 femtoseconds to 30 picoseconds, the team found CVD diamonds could be modified with much shorter pulses than the HPHT diamond; in the CVD samples the shortest pulses left tiny, separated marks, while pulses of about 10 picoseconds or longer produced more continuous modified regions; nitrogen content alone could not explain the differences between samples, pointing instead to microscopic defects tied to each crystal’s growth history. The team then applied 10 picosecond pulses to an 8 mm × 8 mm CVD diamond, writing a grid of modified regions inside it and progressively adding points midway between existing ones; after four such rounds, the diamond split cleanly along the {100} plane with a light tap from tweezers — an area about 710 times larger than the team’s earlier 300 µm-square HPHT result. The {100} plane is the easiest to polish, making it especially important for practical use.

01 / The problem: the hardest material makes the hardest wafers

Diamond’s high thermal conductivity, wide bandgap and transparency in visible light make it a candidate material for next-generation, high-performance technologies: pulling heat away from semiconductor devices, high-power electronics and demanding optical applications. But using those properties first requires turning diamond into thin, precisely shaped substrates — and its extreme hardness makes mechanical machining a struggle.

Laser slicing works “from the inside out”: focused lasers create microscopic structural modifications directly inside the crystal, and the resulting internal stresses promote and guide crack propagation to separate the material, without grinding or cutting in from the outside. That route has proven feasible in recent years, but a key question hung open: do diamonds grown by different methods respond to laser treatment in the same way? Without that answer, process parameters cannot be transferred from one material to another.

02 / The comparison: HPHT and CVD diamonds respond quite differently

The team compared three single-crystal diamonds — one HPHT, two CVD, each with a different level of nitrogen impurity — sweeping focused laser pulse durations from 233 femtoseconds to 30 picoseconds, then examining the size, shape and continuity of the modified regions inside each sample by optical microscopy, spectroscopy and image analysis.

The results split cleanly: the two CVD diamonds could be modified using much shorter pulses than the HPHT diamond, and within the CVD samples the shortest pulses produced tiny, separated marks while pulses of about 10 picoseconds or longer produced more continuous modified regions. More notably, nitrogen content alone did not explain the differences between samples — the team infers that microscopic defects associated with each crystal’s individual growth history may play an important role, and that CVD diamonds also tended to show less stable laser-induced modified regions with greater morphological variation. For process engineers the subtext is plain: which material you buy decides which pulse you use — there is no universal “diamond cutting parameter table”.

03 / The demonstration and JUYAO take: from 300 µm to 8 mm, a 710× larger area

Building on those findings, the team applied 10 picosecond pulses to an 8 mm × 8 mm CVD diamond, writing a grid of modified regions inside it — rather than attempting a complete crack network at once, progressively adding points midway between existing ones. After four such rounds, the diamond split cleanly along the {100} plane with a light tap from tweezers. The {100} plane is the easiest to polish, which is why it matters for practical use. The area comparison given in the report is the point: the 8×8 mm slice is approximately 710 times larger than the team’s earlier 300 µm-square HPHT result — the step from “laboratory demonstration” toward “usable substrate size”.

The authors point to high-power laser optical windows for lithography systems, heat spreaders for advanced electronic devices such as AI chips, and next-generation diamond power semiconductors, and argue the process steps are potentially scalable into future mass-production flows that would cut diamond-substrate manufacturing costs and improve material utilisation. For laser-processing readers, the value here is methodological transfer: once “internal modification plus guided cracking” with ultrashort pulses scales in size on so extreme a material as diamond, similarly brittle, hard crystals gain a reference process paradigm. The boundary is equally plain, though: this demonstration is a single-sample laboratory result, with yield, surface quality and production throughput undisclosed.

PRIMARY REFERENCES & LINKS

  • Chiba University report: Ultrashort laser pulses can slice lab-grown diamond along a useful crystal plane
    phys.org (Chiba University report) · 2 November 2026

    Report read in full; used to verify the publication date, authors and affiliation, the three-sample comparison design, the pulse-duration range, the 10 picosecond grid writing and four-round densification demonstration, the 8×8 mm area and 710× comparison, and the application outlook.

    Open source link ↗
  • Ikeno et al.: Ultrashort-pulse laser-induced internal modification in HPHT and CVD diamond specimens and {100} laser slicing of single-crystal CVD diamond
    Diamond and Related Materials · 1 November 2026

    Journal publication record (DOI link), used to verify the paper title and publication date. Quantitative parameters beyond the report are not cited.

    Open source link ↗