The most difficult specification to verify on large-aperture UV substrates is internal uniformity. Surface roughness, surface figure, and even transmission on sampled cross-sections can all be measured—but "how defects distribute through the thickness direction" is a question almost no one is willing to answer by destructively sectioning a large boule. Yet when the distribution is non-uniform, it gets amplified in a large-aperture optical path into wavefront distortion and localized absorption: the same substrate can behave like two different materials at center and edge.

1. Defects Appear Only Where Irradiated—and They Are a Bulk Effect

Reconstructing two-dimensional and three-dimensional distributions using EPR imaging (EPRI, 10 G/cm gradient field, 40 mm field of view) reveals that both Ge-E′ generated under UV irradiation and Ge-E′H formed after lamp-off are confined within the irradiated volume. When UV light is passed through a physical grating before entering the sample, the reconstructed image clearly resolves periodic intensity modulation—demonstrating that the defect distribution can be both "written" by patterned illumination and read out by EPRI. Three-dimensional reconstruction, slicing, and orthogonal projections further confirm that the activated region penetrates the full sample thickness: polishing off the surface layer cannot eliminate color centers already formed in the bulk.
 
[Figure 1] Original Figures 4(a)–(d): 2D spatial mapping of Ge-E' in the ZY plane, mapping of the dark state Ge-E'H, spatially modulated UV irradiation pattern through the grating, and 3D volumetric reconstruction.
 
This has two implications for substrate applications. First, localized illumination creates a performance separation between "irradiated" and "non-irradiated" zones within the same substrate; optical-path design should be evaluated against the actual beam footprint rather than an averaged whole-substrate value. Second, EPRI provides a non-destructive bulk diagnostic suitable for incoming-material inspection and lot-to-lot comparison of large optical substrates—especially when destructive sampling of the full-aperture boule is not feasible.

2. Two Color Centers: One Lasts 10 Days, the Other 3 Months

Long-term tracking cleanly separates the stability of the two defect species. The Ge-E′ signal disappears completely within 10 days under room-temperature dark storage, being gradually consumed by diffusion-limited hydrogen trapping. In contrast, the 119 G Ge-E′H doublet persists for months under ambient conditions as a relatively stable intermediate state (Figure 2). During aging, the doublet intensity slowly declines due to structural rearrangement of deep traps: hydrogen is either deeply captured into EPR-silent covalent Ge–H bonds (hydrogen passivation) or fully escapes the trap, leaving dangling bonds that recombine into diamagnetic GeODCs.
 
 
 
[Figure 2] EPR spectra during thermal bleaching at 600°C for 1 hour after long-term dark aging (3 days, 10 days, and 3 months). Figure caption: The Ge-E' signal disappears after approximately 10 days, while the 119G doublet persists for several months; after 600°C/1 h, all signals return to baseline.
 
This lifetime gap is directly relevant to substrate inventory and turnover. Substrates in the as-annealed, as-polished "clean" condition perform differently from those that have been in storage for months. For batches requiring high consistency, inventory duration and storage conditions should be recorded, and a baseline spectrum re-acquired before deployment if necessary.

3. Full Signal Erasure at 600 °C / 1 h: An Executable Rework Window

A 600 °C soak for one hour completely eliminates all paramagnetic signals, returning the glass network to its initial ground state. This defines a clear rework and regeneration window: the energy is sufficient to cross activation barriers for defect bleaching and network reconstruction, while remaining well below the softening point of fused silica (~1730 °C). The practical value for large-aperture substrates is that material misused, overdosed, or aged through long service can be restored rather than scrapped—though repeated annealing requires separate evaluation of surface-figure change and residual stress.
Combining the kinetic and spectroscopic data yields a complete energy-landscape diagram (Figure 3): UV irradiation pushes the system from the diamagnetic ground state (Ge–H and GeODCs) to a high-energy state containing Ge-E′ and highly mobile transient hydrogen; upon removal of the light source, hydrogen is captured by active precursor sites via thermally activated hopping or interstitial H₂ diffusion, and the system relaxes into the metastable Ge-E′H state; aging or annealing crosses the barrier to complete bleaching and reconstruction back to the ground state. The same diagram also explains the 20-fold acceleration reported in Blog 2—starting from the metastable state (hydrogen detrapping) requires far less energy than breaking bonds from the ground state.
 
[Figure 3] Schematic diagram of the potential energy levels during the evolution of Ge-related paramagnetic defects. Caption: The complete energy path from the ground state to the excited state to the metastable state Ge-E'H → aging or annealing, explaining the memory effect and thermal repair.
 
The applicability boundary is equally clear: photon energies below the UV range are insufficient to cross the activation barriers for defect photolysis and charge detrapping. Unlike high-energy ionizing radiation that indiscriminately damages the intrinsic lattice, UV-induced color centers depend on pre-existing defect precursors, and the memory effect additionally requires an adequate structural hydrogen reservoir. This is precisely why JGS1 (lacking Ge trapping centers) and JGS3 (hydrogen-deficient) do not exhibit this behavior, while JGS2 uniquely does—grade selection must be tied to the operating wavelength band.

4. Quality-Control Checklist for Large-Aperture UV Substrates

1. EPR baseline spectrum before UV irradiation, recording whether Ge-E′ is present and its initial intensity.

2. EPRI volume-distribution map after UV irradiation, confirming whether defects are confined to the designed optical path region and whether the distribution is uniform through the thickness.

3. Dual-time-point aging data (days and months), using the 119 G doublet intensity to assess the hydrogen reservoir size and metastable-state fraction.

4. Annealing-process verification: spectral recovery after 600 °C / 1 h bleaching, as a rework feasibility criterion.

5. Cross-validation with 200–800 nm transmission and 3670 cm⁻¹ hydroxyl data to avoid mis-grading based on a single indicator.

 
 
 

5. Supply and Customization

Jingmu Optoelectronics supplies UV-grade fused silica substrates, windows, and optical blanks in all grades. Dimension, thickness, flatness, and polishing grade can be customized per drawing, with support for laser cutting and contoured machining. For customers operating under long-term UV exposure, we provide material-grade documentation, UV-Vis transmission and hydroxyl-content data, and can coordinate third-party EPR / EPRI characterization to help establish incoming-inspection baselines and annealing rework criteria.