Chrome Oxide for Sputtering Targets and Varistors: Purity at the ppm Scale
Two very different components — a plasma-deposited coating and a ceramic surge protector — share one raw material and one obsession: chemical purity measured in parts per million.
Cr₂O₃ Sputtering Targets
Chromium oxide coatings are the industrial standard for wear resistance: anilox rolls, pump seals, wire guides, textile machinery. The coatings are typically reactively sputtered or deposited from sintered Cr₂O₃ targets, and film quality is a direct photograph of powder quality.
| Impurity in powder | Effect in film |
|---|---|
| Alkali (Na, K) | Vapor-phase contamination, poor adhesion at the interface |
| Iron | Micro-arc defects, localized colour shift |
| Magnetic particles | Arcing on the target surface, pinholes in the film |
EG-99H at ≤50 ppm Na₂O, ≤100 ppm Fe and ≤5 ppm magnetic impurities is the feed stock that lets a target house guarantee its own product.
ZnO Varistors: Chemistry at the Grain Boundary
A ZnO varistor is a polycrystalline ceramic whose non-linear I-V curve lives almost entirely in a few nanometers of grain-boundary chemistry. The classic formulation — ZnO plus Bi₂O₃, Sb₂O₃, MnO, CoO and Cr₂O₃ — builds a double Schottky barrier at every grain boundary. Chromium stabilizes the boundary phase and suppresses the electrical aging that shifts clamping voltage over the surge lifetime.
Too much sodium or iron in that boundary chemistry, and the non-linearity coefficient α degrades; the device becomes a mediocre resistor. The doping window is small — which is exactly why varistor makers buy electronic-grade Cr₂O₃ and not pigment.
Two Grades, Two Jobs
- EG-99 — the volume electronic grade for varistor doping and thick-film work: ≥99.2 %, Na₂O ≤100 ppm.
- EG-99H — the clean-room grade for targets and premium dielectrics: ≥99.5 %, Na₂O ≤50 ppm, iron and magnetics tightened to target-house levels.
CONTINUE READING
Sodium is the target killer
Ask a target house which single impurity it fears most and the answer is usually sodium. Fifty parts per million of Na₂O in the powder sounds trivial — in most industries it is a rounding error — but inside a magnetron it behaves like a slow-motion saboteur. Alkali metals sputter efficiently, migrate along grain boundaries at operating temperature, and outgas into the plasma. Once in the gas phase, sodium is driven by the plasma potential straight onto the substrate, where it poisons the film-substrate interface. The visible symptoms arrive later: adhesion failure at the interface, haze and colour drift in Cr₂O₃ wear films, and — in dielectric work — leakage currents that end the component's life early.
The failure is not subtle. A 300 mm class target run deposits onto hundreds of wafers before anyone sees the defect pattern, and every one of them is scrap. In anilox-roll coating the same mechanism shows up as a roll that wears prematurely: a ten-thousand-euro component condemned by a few grams of sodium spread through a 30 kg powder batch.
There is a second, mechanical channel for sodium damage: Na₂O is hygroscopic. Sodium-bearing powder pulls moisture out of the air, and trapped moisture flashes into gas during sintering or sputtering — pressure spikes, spits, micro-arcs. The 50 ppm sodium spec is therefore really two specs: an interface-chemistry spec and an outgassing spec.
Context that explains why the discipline exists: pigment-grade Cr₂O₃ runs 500-3000 ppm Na₂O — ten to sixty times over the electronic limit — and nobody cares, because pigment goes into concrete and paint. The same oxide in a target is a disaster. There is no partial credit in ppm work. EG-99H is specified at Na₂O ≤50 ppm because that is the number below which a target house can guarantee its own product, and every lot is verified by ICP-OES before it ships.
The EG-99H impurity budget
| Parameter | EG-99H limit | Failure it prevents | Source controlled |
|---|---|---|---|
| Cr₂O₃ | ≥99.5 % | foreign phases in film and grain boundary | raw oxide purity |
| Na₂O | ≤50 ppm | interface adhesion loss, outgassing bursts | ore alkalis, process water |
| Fe | ≤100 ppm | micro-arc defects, colour shift | mill wear, raw ore |
| Magnetic impurities | ≤5 ppm | target arcing, pinholes | tramp metal from equipment |
| D50 | 0.3-0.8 μm | green-body density, sintering uniformity | classification control |
Arcing, pinholes and the 5 ppm magnetic spec
Arcing is the failure mode that makes target engineers superstitious, and it has a measurable root cause: conductive defects on an insulating surface. Cr₂O₃ is an electrical insulator, so a chromium-oxide target is sputtered with RF or pulsed DC, and every conductive speck on the target face becomes a charge-concentration site. Magnetic impurities — iron, nickel and stainless-steel wear particles shed by mills and handling equipment — are exactly such specks. Under the plasma they charge, discharge, and punch micro-arcs into the target surface. Each arc spits material, pits the target, and seeds the next arc. The film inherits the damage as pinholes, particles and local thickness defects; the chamber inherits flakes.
The ≤5 ppm magnetic impurity spec in EG-99H exists to starve this mechanism at the powder stage. The test is direct: a dispersed sample passes a rare-earth magnet, the captured fraction is weighed, and the number goes on the certificate. Five parts per million is a brutal bar — one stainless bolt worn down inside a mill and spread through a tonne of powder is enough to fail it — which is why the spec also drives plant design: ceramic-lined mills, demagnetised stainless contact surfaces, and magnetic traps at every transfer point.
Powder quality sets the ceiling; target fabrication decides what is reached. Sintered Cr₂O₃ targets fired at 1600-1700 °C in air reach 96-98 % of theoretical density, and every residual pore is a preferred arc site. Impurities concentrate at grain boundaries during sintering — sodium there, iron there — turning a 99.5 % pure powder into a target whose boundaries behave like a lower-purity material. The PSD spec (D50 0.3-0.8 μm) is part of the same logic: a narrow, fine distribution packs green bodies uniformly, sinters evenly, and leaves fewer density defects than a broad one.
The cost model behind all of it: one arcing-scarred target costs thousands in rework or replacement, and the chamber clean after a heavy arc event is 4-8 hours of downtime on equipment worth hundreds per hour. The powder line item is 1-3 % of the target's bill of materials. Arcing is a powder problem that shows up as a machine problem — which is precisely why it is solved before the powder ships.
Buying electronic-grade Cr₂O₃ like a component
Serious target houses qualify powder the way they qualify a bearing: three lots, full documentation, and no shortcuts. The qualification protocol starts with an ICP-MS scan across 60+ elements — not the five-element ICP the pigment trade runs — followed by laser-diffraction PSD, magnetic impurity count, moisture and tap density. Ten-point lot sampling guards against stratification. Acceptance windows are pre-agreed and enforced lot by lot; a lot that misses the Na₂O window by 10 ppm is rejected outright, because the target house has already sold its own purity promise to its customer.
Documentation is half the product. Every EG-99H lot ships with a CoA naming the methods — ICP-OES for sodium and iron, magnetic separation for the magnetic count, wet laser diffraction for PSD — plus lot number, production date and the ore-source traceability chain. Retained samples run five years. When a target fails in the field eighteen months after the powder shipped, the target house and its supplier meet in front of the retained samples, not in front of lawyers.
Supply discipline matters as much as chemistry. Electronic grade is made on a segregated line: dedicated mills that never see pigment material, deionised process water, clean-room packing at ISO 8, double polyethylene bags with a vacuum outer, lot-labelled to the pallet. Cross-contamination is the industry's quiet killer — a one-percent pigment crossover would blow the sodium budget of an entire campaign — so segregation is audited as hard as the chemistry.
The same discipline serves the varistor side of the business. Zinc-oxide varistor makers buy EG-99 for grain-boundary chemistry: sodium and iron at the boundary degrade the double Schottky barrier, flatten the non-linearity coefficient α and accelerate electrical aging, so the doping window is small and the powder must not add noise. Different grade, same logic: in electronic ceramics, the powder is a component. Buy it like one.
The qualification record is the part most buyers under-specify. Before the first production lot, ask for the vendor's internal lot-release data across the last twelve months — Na₂O scatter, magnetic counts, D50 drift — not the three best certificates. A supplier whose sodium runs a flat 30-45 ppm for a year is a process; one that oscillates between 20 and 90 ppm is a roulette wheel wearing a lab coat. The same logic applies to packaging: a lot that arrives with a broken vacuum bag has already started absorbing moisture, and no CoA can un-absorb it. Inspect the bag, then the paper.
For the varistor buyer, the practical translation of the same discipline is a doping trial: two batches, one control and one candidate, ten devices each, measure α and clamp-voltage drift after a 1000-hour accelerated aging run. The powder that wins is not the cheaper one — it is the one whose grain-boundary chemistry does not move. That is a purchase decision made with data instead of a price list.
Target houses do not buy powder; they buy insurance against arcing. The powder is 1-3 % of the target's material cost, but it decides whether the target reaches end of life or dies in a shower of micro-arcs — so the cheapest quote is usually the most expensive purchase. Our rule is simple, and we state it plainly: an electronic lot that fails the magnetic test never ships. No exceptions, no concessions. EG-99H is a discipline, not a certificate — segregated line, deionised water, ceramic-lined mills, clean-room packing — and when a customer audits us, we walk them through the plant floor, not the brochure. That is what ppm purity actually costs, and what it is worth.
FAQ
EG-99 or EG-99H for sputtering targets? EG-99H — the Na₂O ≤50 ppm and magnetic ≤5 ppm limits exist specifically for target houses; EG-99 (Na₂O ≤100 ppm, D50 0.6-1.5 μm) covers varistor doping and thick-film work. Can I use pigment-grade Cr₂O₃ and save money? Pigment grade runs 500-3000 ppm Na₂O and carries magnetic tramp from pigment mills; one arcing event costs more than years of powder savings. Don't. How is ≤5 ppm magnetic content measured? A dispersed sample passes a rare-earth magnet and the captured fraction is weighed; the result ships on every EG-99H certificate. What particle size do target manufacturers need? D50 0.3-0.8 μm for EG-99H, with the full PSD certified per lot — narrow and fine packs green bodies uniformly and sinters to higher density. Does sodium matter in ZnO varistors too? Yes — sodium migrates to grain boundaries, degrades the double Schottky barrier and accelerates electrical aging; varistor makers buy EG-99 for the same reason targets buy EG-99H.