You’re running a batch of magnesium fluoride for excimer laser windows, and the transmission at 193 nm comes back at 88% instead of the spec’d 92%. The synthesis conditions were textbook—hydrofluoric acid in stoichiometric excess, 400°C reaction temperature, two-hour hold. The raw MgO passed incoming inspection for purity at 98.5%. But the problem wasn’t the MgO purity number. It was what that number didn’t tell you.

We saw this pattern across six production batches last quarter. Every time the MgO feedstock had a calcium oxide (CaO) content above 0.4 wt%, the final magnesium fluoride showed a transmission drop of 3–5% at 193 nm. Below 0.3% CaO, the transmission held steady at 91.5–92.5%. That’s a narrow window, and it cost us two weeks of rework before we traced it back.

The Calcium Incorporation Mechanism

Calcium substitutes for magnesium in the MgF₂ lattice because the ionic radii are close—Ca²⁺ at 1.00 Å versus Mg²⁺ at 0.72 Å. The catch is that CaF₂ has a different refractive index and a slightly different band gap. In a polycrystalline MgF₂ window, even 0.1 wt% CaO in the starting oxide translates to roughly 0.05 wt% calcium in the final fluoride, assuming complete conversion. That’s enough to create scattering centers at deep UV wavelengths.

We confirmed this by running ICP-OES on both the MgO feed and the finished MgF₂ from the same batch. The calcium mass balance closed within 3%, meaning nearly all the calcium from the oxide ends up in the fluoride. It doesn’t wash out during the HF reaction or the subsequent sintering step.

In practice, this means your MgO supplier’s certificate of analysis needs to list calcium, not just total purity. A 98.5% MgO with 0.6% CaO will perform worse for optical-grade MgF₂ than a 97.5% MgO with 0.1% CaO. I’ve seen procurement reject the lower-purity material because the number looked worse. That’s a mistake.

Processing Window: Reaction Temperature and Calcium Mobility

We ran a designed experiment varying the MgO-to-HF reaction temperature from 350°C to 500°C, using a single MgO source with 0.35% CaO. At 350°C, the conversion to MgF₂ was incomplete—about 94%—and the residual oxide left the material hazy. At 500°C, conversion hit 99.5%, but the calcium had enough thermal energy to diffuse and form discrete CaF₂ domains visible under SEM at 5000x magnification.

The sweet spot was 400–430°C. At these temperatures, conversion was above 99%, and the calcium remained dispersed as a solid solution rather than precipitating out. The transmission at 193 nm for material made at 420°C was 91.8%. At 480°C, it dropped to 89.2%.

The lesson: if you’re chasing transmission, don’t just crank the temperature to maximize conversion. You’re trading one problem (residual oxide) for another (calcium clustering). The processing window is tighter than most datasheets suggest.

Specification Gap: What Your MgO Cert Doesn’t Tell You

I pulled certificates from three different MgO suppliers over the past year. All guaranteed ≥98% MgO. Only one listed calcium content. The others listed “acid-insoluble matter” and “loss on ignition” but nothing specific to CaO. That’s a problem because the standard test method for MgO purity (titration with EDTA) measures total magnesium—it doesn’t distinguish calcium from magnesium. A batch could have 0.8% CaO and still pass a 98.5% purity spec if the calcium is counted as magnesium.

We now require our MgO supplier to provide CaO content via XRF or ICP-OES, with a spec of ≤0.3% CaO for optical-grade material. We also run our own XRF on every incoming drum as a verification step. It adds 15 minutes per batch. It’s saved us from scrapping three batches of MgF₂ in the last six months. That’s about $12,000 in avoided rework costs.

Technical Specifications & Performance Data

The following data comes from internal QC records for MgO feedstock used in optical-grade MgF₂ synthesis at our facility.

Parameter Standard Grade MgO Optical Grade MgO (Our Spec) Test Method
MgO content (wt%) ≥98.0 ≥98.5 Titration (EDTA)
CaO content (wt%) ≤0.8 ≤0.3 XRF / ICP-OES
Fe₂O₃ content (wt%) ≤0.05 ≤0.02 ICP-OES
SiO₂ content (wt%) ≤0.10 ≤0.05 ICP-OES
Loss on ignition (wt%) ≤2.0 ≤1.0 ASTM D7348
BET surface area (m²/g) 10–30 15–25 ASTM D3663
Median particle size D50 (µm) 5–15 3–8 Laser diffraction (ISO 13320)

Note the CaO spec. That 0.3% limit came directly from our production data: batches using MgO with CaO >0.4% consistently failed the 193 nm transmission test. Batches at 0.2–0.3% passed with margin. Below 0.2% gave no additional improvement—the transmission leveled out at 92.3%. So 0.3% is our practical upper limit, not an arbitrary safety factor.

Root Cause Breakdown: Why High CaO Kills Transmission

It’s not just about scattering. Calcium fluoride has a band gap of about 10.2 eV, compared to magnesium fluoride’s 11.8 eV. That difference means CaF₂ domains absorb slightly more at 193 nm (6.4 eV). The absorption isn’t huge, but in a 3 mm thick window, every 0.1% CaF₂ adds roughly 0.3% absorption loss. Combined with Mie scattering from CaF₂ particles in the 100–500 nm range—which we confirmed by SEM-EDS—the total loss adds up fast.

We also saw that high-CaO MgO produced MgF₂ with lower green density after cold pressing. The calcium domains acted as internal stress concentrators, causing microcracking during the sintering ramp. This dropped the sintered density from 98.5% of theoretical to 96.2%, which further increased scattering. So the calcium problem cascades: it affects chemistry during reaction, microstructure during pressing, and final optical properties.

Frequently Asked Questions

Q: How does calcium oxide content in MgO affect the dielectric strength of MgF₂ films?
A: We haven’t tested this directly for thin films, but for bulk MgF₂, CaO levels above 0.4% reduce the dielectric breakdown voltage by about 15%—from 18 kV/mm down to 15.3 kV/mm. The mechanism is likely the same: CaF₂ domains create weak interfaces that fail earlier under field stress. If you’re depositing MgF₂ films for capacitor or insulator applications, I’d keep CaO below 0.2% in the starting oxide.

Q: Can you remove calcium from MgO before the fluorination reaction?
A: Not economically at scale. Washing with dilute acid can leach some calcium, but it also attacks the MgO. You lose 5–8% of your yield per wash cycle, and the wastewater treatment costs add up. We tested it. It’s cheaper to buy MgO with low calcium from a supplier who controls it at the mining stage. Zhongyao Chemical, for instance, sources from a deposit where the natural CaO content runs 0.15–0.25%. That’s the kind of supply chain spec you want.

Q: Does the MgO particle size distribution affect calcium incorporation in MgF₂?
A: Yes, but indirectly. Fine MgO (<5 µm D50) reacts faster with HF, which can lead to localized hot spots if the reaction exotherm isn’t managed. Those hot spots drive calcium diffusion and clustering. Coarser MgO (10–20 µm) reacts more slowly and evenly, giving the calcium less thermal drive to segregate. We run our optical-grade MgO at 3–8 µm D50 with a narrow span—that balances reaction rate against thermal control. Below 3 µm, we saw inconsistent transmission batch to batch.

Q: What’s the fastest way to screen incoming MgO for calcium content on the production floor?
A: XRF is the most reliable, but if you don’t have one, you can do a quick wet chemistry test: dissolve 5 g of MgO in 50 mL of 1:1 HCl, filter, then add ammonium oxalate solution. A white precipitate (calcium oxalate) forms within 10 minutes if CaO exceeds about 0.5%. It’s semi-quantitative, but it’s enough to flag bad batches before they hit the reactor. We use this as a go/no-go check while the XRF sample is queued.