Progress in the Fabrication and Optimization of High-Energy Diamond X-Ray Refractive Lenses
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Micromachines |
| Published | 2026-06-01 |
| Authors | Hao Huang, Kang Du, Wenbin He, Weiwei Zhang, Xiaohong Yang |
| DOI | 10.3390/mi17060687 |
Diamond X-ray refractive lenses reached a focal spot of 52 nm × 51 nm at 14 keV in a 2025 synchrotron demonstration by Wang et al. The phase-shifting lens used femtosecond laser direct writing and integrated aberration compensation. Huang and colleagues cite this result as evidence of nanometer-scale focusing, but the review does not establish that every lens geometry retains it under high heat flux.
Focusing performance and prior benchmarks
Section titled “Focusing performance and prior benchmarks”An early compound refractive lens focused 14 keV X-rays to an 8 µm line width. A 2005 hard X-ray nanoprobe later reached a 47.55 nm² focal spot at 21 keV. For diamond planar lenses, FIB correction produced about 50 × 60 nm² under hard X-ray illumination, although the review does not specify its photon energy.
Batch-produced diamond lens arrays reached 150 nm under hard X-rays. Reverse-molded polycrystalline optics produced 2 to 4 µm spots without post-polishing. These results cover different lens geometries and photon energies, so their spot sizes are useful benchmarks rather than direct equivalents.
Material selection under high heat flux
Section titled “Material selection under high heat flux”X-ray refraction is weak, so CRLs stack paired parabolic profiles, while kinoform and phase-shifting lenses reduce absorption by altering the phase profile. Their material must combine low absorption, high thermal conductivity, low expansion, and radiation tolerance.
The review’s property table lists 2250 W/(m K) thermal conductivity for single-crystal diamond, compared with 200 W/(m K) for beryllium and 140 W/(m K) for silicon. Its thermal expansion coefficients are about 1.1 × 10^-6/K for diamond, 11.3 × 10^-6/K for beryllium, and 2.6 × 10^-6/K for silicon.
At 10 keV, the listed absorption coefficients are 8.33 cm^-1 for diamond, 1.2 cm^-1 for beryllium, and 78.94 cm^-1 for silicon. Diamond combines the highest tabulated conductivity and lowest expansion with lower absorption than silicon. Beryllium’s tabulated 10 keV absorption remains lower than diamond’s.
Fabrication choices for lens structures
Section titled “Fabrication choices for lens structures”| Technique | Precision | Removal rate | Main limitation |
|---|---|---|---|
| Femtosecond laser ablation | Sub-micrometer to micrometer | 3.64 × 10^5 µm³/s at 56.6 J/cm², 1 kHz, and 1 mm/s | Surface roughness and laser damage |
| FIB | Nanoscale | 0.4 to 0.6 µm³/nC with Ga+ | Low throughput and ion-induced damage |
| RIE | Nanoscale to sub-micrometer | 50 to 100 nm/min | Mask selectivity and deep-etch morphology |
Femtosecond processing uses 30 to 300 fs pulses and peak power up to 5 × 10^20 W/m². Its nonlinear absorption confines removal before heat diffuses into the lattice. At 216 fs, single-pulse thresholds were about 1 to 2 J/cm² at 1030, 515, and 343 nm. Incubation lowered multipulse thresholds to 0.03 to 0.20 J/cm².
Under typical high-precision conditions, FIB resolution was 2 to 2.8 nm FWHM, with removal near 1 nm/s. In the cited ion-beam figuring results, figure error fell from 8 nm RMS to 0.76 nm RMS, and sometimes 0.36 nm RMS. Tangential slope error fell from 411 nrad RMS to 97 nrad RMS.
The cost was ion damage. A 30 keV Ga+ beam produced an amorphous layer as deep as 44 nm and swelling up to 13 nm. In 80 µm diamond, oxygen-assisted PFIB controlled taper at about 5°. Xenon PFIB raised the removal rate to about 60 times that of conventional Ga+ FIB.
The table gives a comparison rate of 50 to 100 nm/min for RIE. Individual recipes span 10 to 700 nm/min as gas ratios and plasma power change. Deep etching still suffers from mask erosion and redeposition, so mask selection and cleaning remain central.
Hybrid routes and surface correction
Section titled “Hybrid routes and surface correction”Hybrid processing assigns bulk shaping to femtosecond laser and final correction to ion beam. One hybrid placed 27,000 microlenses on a 2.5 mm diameter lens. Ultraviolet femtosecond machining followed by hot mixed-acid treatment reduced roughness from 37.4 nm to 9.4 nm in processed regions. The same study reported 0.64 nm in regions that had not been laser processed.
Near-field speckle metrology can measure slope errors below 100 nrad on X-ray mirror surfaces. Surface treatment also targets laser damage. Selective oxidation at 700 to 800 °C removed about 50 nm of amorphous carbon and restored transmission above 98% of the theoretical value. Plasma-assisted polishing reached about 0.5 nm roughness. Gas-cluster ion smoothing reduced roughness from roughly 15 nm to 0.3 to 0.4 nm.
Reverse molding avoids subtractive damage at the final surface. Diamond was grown conformally by MPCVD on a smooth negative silicon template and released by dissolving the silicon. The inherited surface supported 2 to 4 µm focal spots without post-polishing.
Crystal quality and wavefront preservation
Section titled “Crystal quality and wavefront preservation”PCD offers larger-area growth and relatively lower manufacturing cost, but grain boundaries and orientation mismatch add scattering. Dynamic friction polishing reduced roughness from 29.5 nm to 0.9 nm over a 100 µm scale. Intragranular roughness reached 0.1 to 0.2 nm, although uneven removal at grain boundaries prevented uniform sub-nanometer finishing.
SCD avoids grain boundaries, but CVD homoepitaxial growth can retain threading dislocations. X-ray topography relates (111) etch pits to these dislocations. Boron and nitrogen co-doping modified the local lattice stress and reduced dislocation density in selected regions. Substrate damage and seed thickness variation remain important controls on epitaxial quality.
Stitched SCD extends the aperture, but the junctions reduce physical continuity. The review lists demonstrated mosaics at 30 × 30 × 1 mm³ and a 2-inch-scale stitched single-crystal process. Misalignment, height mismatch, step bunching, polycrystalline inclusions, and stress peaks remain at stitched interfaces.
Thermal stability and packaging
Section titled “Thermal stability and packaging”Material conductivity alone does not control the lens temperature. Contact resistance at the diamond and cooling substrate can create local gradients and wavefront distortion. Under 17.5 keV, 1 mJ XFEL pulses at 10 kHz, with fluence up to 10 J/cm², an optimized multilens system still reached about 300 °C.
Liquid-nitrogen cooling with second-order compensation kept thermal distortion below 15 pm RMS at 1 MHz. A diamond and copper composite reached 807 W/(m K), reduced thermal resistance by 39%, and lowered peak temperature by 12.3 °C under identical high-power conditions. A separate diamond composite anode raised the stated power limit to 73 W, compared with 37.9 W for tungsten.
Laser-machined PCD microchannels reached 800 µm depth, aspect ratios of 4:1 and up to 16.7, and a 0.9° sidewall taper. Their bottom roughness was about 186 nm, which shows why the channels need finishing when surface quality matters.
Remaining barriers and engineering implications
Section titled “Remaining barriers and engineering implications”Three barriers remain. Conventional SCD is typically below 10 × 10 mm², which restricts large-aperture optics. Stitching can extend area, but nitrogen-assisted lateral growth can introduce impurity risk. Optical fabrication must also control laser roughness, FIB damage, RIE mask selectivity, and long-term thermal drift.
For coherent nanofocusing, SCD is the stronger choice for smaller apertures because dislocation control governs wavefront fidelity. PCD remains useful for larger heat-load windows and some lens arrays where grain scattering is acceptable. A practical process uses laser or plasma etching for volume removal, followed by targeted FIB correction where the error budget justifies it.
The authors propose hybrid workflows and active curvature control using piezoelectric micro-actuators or thermal gradient arrays. These are proposed directions, not demonstrated performance in the reviewed lens tests.
Materials and support
Section titled “Materials and support”The paper calls for single-crystal or polycrystalline diamond lens substrates with controlled crystallinity before femtosecond-laser, ion-beam, or plasma structuring. The diamond materials catalog offers MPCVD-grown single-crystal and polycrystalline plates and wafers up to 125 mm, with substrates up to 10 mm. For Ti or Cu metallization of the paper’s thermal interfaces, 6CCVD offers engineering support to assess material and process requirements.
View Original Abstract
The extreme thermal loads encountered in fourth-generation synchrotron radiation sources and X-ray free-electron lasers (XFEL) impose stringent requirements on X-ray optical components. Conventional materials such as beryllium and silicon increasingly exhibit limitations under high-energy conditions, including insufficient thermal conductivity, limited radiation stability, and significant absorption losses, rendering them inadequate for next-generation high-energy X-ray optics. In this context, single-crystal diamond, with its high thermal conductivity, low absorption coefficient, and excellent mechanical strength and radiation resistance, has emerged as a promising candidate for high-energy X-ray refractive optics. This review systematically summarizes recent advances in the fabrication and performance optimization of diamond X-ray refractive lenses for high-energy applications. Starting from the evolving demands of modern synchrotron radiation facilities and XFEL, the fundamental requirements for materials and structural design in high-energy X-ray optics are analyzed. Through comparisons with representative materials, the advantages of diamond in thermal management and transmission performance are highlighted. Major micro- and nanofabrication techniques, including femtosecond laser processing, focused ion beam milling, and plasma etching, are comprehensively reviewed, with emphasis on their respective characteristics in terms of processing efficiency, precision control, and damage introduction. The emerging trend of hybrid fabrication strategies is also discussed. Furthermore, the effects of surface roughness, subsurface damage, and crystal defects on wavefront quality and focusing performance are examined, along with corresponding post-processing and surface correction methods. Finally, current challenges related to large-size single-crystal growth, high-precision low-damage fabrication, and long-term operational stability are discussed, and future development directions for diamond-based X-ray refractive optical components are outlined.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
Boron Doped Single Crystal Diamond Plate
View Specs →
Single Crystal Cvd Diamond Plate
View Specs →References we cover
Section titled “References we cover”- Imaging Heat Transport in Suspended Diamond Nanostructures with Integrated Spin Defect Thermometers
- Synergistic Control of Crystal Planes and Defects in CVD Single-Crystal Diamond - A Review of Growth Mechanisms and Frontier Applications
- Enhanced Lateral Growth of Homoepitaxial (001) Diamond by Microwave Plasma Chemical Vapor Deposition with Nitrogen Addition
- Optimization of Diamond Polishing Process for Sub-Nanometer Roughness Using Ar/O2/SF6 Plasma
- Studies of Dislocations in Type Ib, Type IIa HPHT and CVD Single Crystal Diamonds
Original source
Section titled “Original source”References
Section titled “References”- 2021 - New Era of Synchrotron Radiation: Fourth-Generation Storage Ring [Crossref]
- 2019 - Coherence Properties of the High-Energy Fourth-Generation X-Ray Synchrotron Sources [Crossref]
- 2022 - 4th Generation Synchrotron Source Boosts Crystalline Imaging at the Nanoscale [Crossref]
- 2008 - Coherence Properties of the Radiation from X-Ray Free Electron Laser [Crossref]
- 1994 - A Review of Thermo-Mechanical Considerations of High Temperature Materials for Synchrotron Applications [Crossref]
- 2021 - Beam Heating from a Fourth-Generation Synchrotron Source [Crossref]
- 2017 - Next-Generation Materials for Future Synchrotron and Free-Electron Laser Sources [Crossref]
- 2022 - Active and Adaptive X-Ray Optics at Diamond Light Source
- 2010 - In Situ Measurements of Thermal Distortions in Synchrotron Optics under High Heat Load [Crossref]