Choosing between yttria-stabilized zirconia (YSZ) and scandia-stabilized zirconia (ScSZ) is not a purchasing decision that can be deferred to the sourcing team. It fixes the cell's operating temperature, its electrolyte thickness, its mechanical architecture, and — since April 2025 — its regulatory exposure. This guide compares the two on the axes an engineer actually designs against: ionic conductivity versus temperature, long-term degradation, mechanical and sintering behavior, and supply risk. The short version is that ScSZ buys conductivity at intermediate temperature and pays for it in phase stability and price, and that the honest answer for most programs is 8YSZ unless a specific temperature target forces the upgrade. Who supplies these materials is a separate question, covered in the global ScSZ supplier landscape.
What the Electrolyte Actually Has to Do
A solid-oxide electrolyte has to do four things at once, and only the first is what gets quoted on a datasheet.
- Conduct oxygen ions well. Ohmic loss through the electrolyte is often the dominant single loss term in a well-built cell.
- Conduct electrons essentially not at all. Any electronic leakage short-circuits the cell internally and costs open-circuit voltage.
- Stay gas-tight and phase-stable for tens of thousands of hours, with one face in air and the other in wet reducing fuel, across thermal cycles.
- Survive being made. It must sinter dense at a temperature compatible with the electrodes it is co-fired with, and hold together mechanically at the thickness the cell architecture demands.
The usual design criterion collapses the first requirement into an area-specific resistance (ASR) budget. A commonly cited target is an electrolyte ASR of about 0.15 Ω·cm², and because ASR = L/σ (thickness divided by conductivity), that single number ties conductivity and thickness together. At a conductivity of 0.02 S/cm, a 0.15 Ω·cm² budget allows roughly 30 µm of electrolyte. Drop the conductivity by a factor of four — which is roughly what happens to zirconia over a couple of hundred degrees — and the allowable thickness drops by a factor of four too.
That is the whole tension in one line. You can chase conductivity through the material, or you can chase it through thickness. ScSZ is the material answer; anode-supported thin-film architecture is the thickness answer; and the two interact, because a thinner electrolyte cannot carry the cell mechanically.
Ionic Conductivity Against Temperature: The Core Trade
Zirconia conducts oxygen ions through vacancies created when a trivalent cation substitutes for Zr⁴⁺. Scandium's advantage is geometric: Sc³⁺ (ionic radius ≈ 0.087 nm) is a much closer match to Zr⁴⁺ (≈ 0.084 nm) than Y³⁺ is, so the substituted lattice carries less internal strain, vacancies are less strongly trapped, and the migration barrier is lower. The reported activation energies bear this out — roughly 0.88–1.07 eV for 10Sc1CeSZ against 1.04 eV for 8YSZ in one joint impedance-spectroscopy study of both materials.
A lower activation energy matters more the colder you run. Both materials lose conductivity as temperature falls, but YSZ loses it faster, so the ScSZ advantage widens as you move from high-temperature (HT-SOFC, roughly 850–1000 °C) into intermediate-temperature (IT-SOFC, roughly 600–800 °C) operation. Review literature puts ScSZ at roughly 1.5–3× the conductivity of YSZ depending on composition and temperature.
Here is the comparison with measurement conditions attached, because a bare S/cm figure is not usable:
Two things in that table deserve emphasis, and they cut in opposite directions.
First, the paired comparisons are the only ones worth trusting. Reported 8YSZ conductivity at 800 °C spans roughly 0.02 to 0.134 S/cm across the literature, and almost all of that spread is densification, grain size, and grain-boundary chemistry rather than composition. Comparing a number from one paper against a number from another tells you very little. When you evaluate candidate powders, measure them yourself on identically processed pellets.
Second, and more usefully as a warning: in the study cited in the last two rows, 10Sc1CeSZ came in below 8YSZ at 600 °C, and the result varied by more than a factor of two depending on which starting powder was used. A material's intrinsic advantage can be entirely erased by processing. Paying a scandium premium for a powder you cannot sinter well is the most expensive way to build a mediocre electrolyte, which is why lot-level verification of particle size, surface area, and reproducible sintering behavior is not optional on the ScSZ track.
There is also a specific reason not to push plain ScSZ too far down in temperature. Scandia-stabilized zirconia undergoes a cubic-to-rhombohedral (β-phase) transition at around 650 °C that is peculiar to the Sc system and that reduces conductivity — precisely in the temperature window where you adopted ScSZ to gain conductivity. Suppressing that transition is the reason co-doped grades exist.
Long-Term Degradation: Where the Cost of Conductivity Gets Paid
Initial conductivity is the easy number. What decides whether a stack meets its warranty is the slope over thousands of hours, and here the ranking is not the same.
8YSZ ages, and not gently. After 1,000 hours of annealing at 1000 °C, 8YSZ conductivity has been reported to fall by roughly 40%, while 10YSZ over the same exposure stayed nearly constant. The mechanism is that 8 mol% yttria sits close to the cubic/tetragonal boundary, so the fully stabilized cubic phase slowly decomposes at high temperature. Against a commonly cited design target of no more than about 0.1% degradation per 1,000 hours, a 40% loss is not a rounding error — it is a reason many HT designs run over-stabilized compositions or accept the aging and size the stack for end-of-life performance rather than beginning-of-life.
ScSZ has the same disease in a different form. Analysis of the 10Sc1CeSZ composition concludes it is slightly under-doped and not fully stabilized, leaving it prone to slow cubic-to-tetragonal decomposition at operating temperature. The observed behavior is consistent with that: the surface of a 10Sc1CeSZ electrolyte has been reported to transform from cubic through tetragonal to monoclinic during operation at 900 °C for 100 hours. Patent literature on phase-stable doped-zirconia compositions argues for at least 2 mol% of additional stabilizing oxide — ceria, yttria, and/or ytterbia — to suppress the aging decomposition, which tells you where the commercial grades are heading.
The cerium in 10Sc1CeSZ is doing two jobs, and one of them has a catch. The 1 mol% CeO₂ raises the tendency to form the high-symmetry, high-conductivity phases at the expense of low-symmetry ones, and co-dopant studies find ceria-doped compositions among the highest-conductivity variants across 400–1000 °C while fully suppressing the cubic-to-rhombohedral transformation. The catch is the anode side: in a reducing atmosphere, Ce⁴⁺ partially reduces to Ce³⁺, and that redox activity has been associated with lower fracture strength and reduced stability. So the additive that fixes the phase-transition problem introduces a fuel-side chemistry problem. If your design runs at high fuel utilization or sees redox cycling, that trade needs testing rather than assuming.
The practical conclusion: neither material is stable enough to be designed around a single beginning-of-life conductivity number. Qualify on aged samples, and treat the composition on the datasheet — particularly total stabilizer content — as a degradation-rate parameter, not just a conductivity parameter.
Mechanical and Manufacturing Reality
The mechanical properties of these electrolytes are frequently overestimated by engineers whose mental model of zirconia comes from dental or structural 3Y-TZP. That material is a different animal: the fully stabilized cubic grades used as electrolytes have no transformation-toughening mechanism, and they are correspondingly weak.
Three design consequences follow.
ScSZ does not buy you mechanical margin. A flat-plate SOFC built with an 11ScSZ electrolyte was reported to have mechanical strength comparable to a conventional 8YSZ electrolyte. Scandia is a conductivity upgrade, not a strength upgrade, so if your cell is cracking, changing dopant will not fix it.
Architecture follows from the ASR budget, and it is a fork in the road. Electrolyte-supported designs put the load on the electrolyte and therefore need it thick — reported support layers run from roughly 75 µm up to several hundred micrometres, with 200–350 µm a common band. That thickness only meets a 0.15 Ω·cm² budget at high conductivity, which is why electrolyte-supported cells are HT designs: at 850 °C, an 0.108 S/cm ScSZ would support about 160 µm within budget, and a 78 mS/cm 8YSZ about 120 µm. Anode-supported designs move the load to a porous Ni/zirconia substrate and thin the electrolyte to roughly 5–15 µm (3 µm has been demonstrated by magnetron sputtering), which is what makes 600–800 °C operation reachable — at the cost of a fragile, camber-prone thin layer that must be co-fired without pinholes.
Sintering is where ScSZ is genuinely easier, which is the one process advantage it has. Conventional YSZ electrolytes are typically densified at 1400–1450 °C. ScSZ reaches near-full density at around 1400 °C conventionally, and with fine (nanoscale) powders or sintering aids has been densified at temperatures as low as 900 °C. YSZ/ScSZ composites have been co-sintered successfully at 1400 °C for 5 hours. A lower densification temperature widens the co-firing window with electrodes and reduces cation interdiffusion at interfaces — a real benefit, though one that depends entirely on powder quality, which brings the argument back to lot verification.
Supply Risk as a Design Input
For most of the history of this materials choice, supply was a procurement footnote. It is now a design parameter, because both stabilizers sit under the same Chinese export-control regime and neither is comfortable.
MOFCOM Announcement No. 18 of April 4, 2025 placed seven medium and heavy rare earths — scandium and yttrium among them — under per-shipment licensing. So the naive hedge of "specify YSZ to avoid the scandium problem" does not work: yttrium is controlled too, under classification 1C908, and with no concentration-floor threshold, which makes a low-content defense weak for YSZ powder. The mechanics of the yttrium side, including the HS-code mismatch that appears at the Korean border, are in our Korea import guide for YSZ.
What differs is the severity, and it differs in both directions.
- Scandium is structurally scarce. Global output is on the order of 40 tonnes a year, recovered as a byproduct rather than mined directly, and Sc₂O₃ moved from about $1,200/kg before the controls to $3,500–4,370/kg after. Because 10Sc1CeSZ is roughly 11 wt% Sc₂O₃, both the powder and its feedstock fall squarely in scope. The full picture is in how China's scandium export controls reshaped the SOFC supply chain.
- Yttrium is not scarce, but it has been repriced violently. Y₂O₃ went from under $8/kg before the controls to roughly $1,100/kg by May 2026, with shipments outside China down about 50% and a split dual-price structure between European and Chinese domestic levels. The demand and pricing context is in our 2026 YSZ market outlook.
The design-relevant asymmetry is not the loading, and putting both on one basis shows why: 8 mol% Y₂O₃ is about 13.7 wt% of the powder, while 10 mol% Sc₂O₃ is about 11.0 wt% — so 8YSZ actually carries slightly more stabilizer by weight. The asymmetry is entirely on the supply side: yttria is abundant but currently expensive, while scandia is expensive and genuinely scarce. Yttrium pricing is a regulatory artifact that can unwind; scandium's 40 t/y output is a physical constraint that cannot unwind quickly. A design that can only meet its performance target with ScSZ has taken on a supply dependency that no amount of contracting fully removes. That is a legitimate reason to spend engineering effort on making 8YSZ work — thinner electrolyte, better electrodes, higher operating temperature — before committing to scandium.
Note also that neither of these is an embargo. Both are licensing regimes, and shipments with documented civilian end use have room to clear. But license lead time becomes part of your program schedule, and it is not a variable you control.
A Decision Framework
Working through the above, the choice usually resolves on operating temperature first and supply risk second.
Specify 8YSZ when:
- You are running at 850–1000 °C, where 8YSZ conductivity comfortably meets the ASR budget at a manufacturable thickness.
- You can build anode-supported cells with a 5–15 µm electrolyte, which recovers at intermediate temperature much of what the material gives up.
- Program economics or supply resilience outweigh a last increment of power density.
- You need the deepest supplier pool. YSZ is made at scale for dental, thermal-barrier, and structural markets as well as SOFC, so alternate sources exist.
ScSZ earns its cost when:
- The target is genuinely intermediate-temperature operation with a thick, self-supporting electrolyte — an ASR budget that 8YSZ cannot meet at the thickness your architecture requires.
- Lowering operating temperature is worth more than the powder premium: cheaper interconnects and seals, slower degradation elsewhere in the stack, faster start-up.
- You have the qualification budget to prove out phase stability, because the 10Sc1CeSZ aging and Ce redox questions above are real and composition-specific.
What to prototype, in order:
- Measure conductivity on your own pellets, at your operating temperature, in the relevant atmosphere, on identically processed samples of every candidate powder. Do not carry literature values into a design calculation.
- Age before you decide. Hold samples at operating temperature for as long as the schedule allows and re-measure. A material that starts 30% ahead and ages 40% is behind one that starts even and holds.
- Test the fuel side specifically if you are evaluating a ceria co-doped grade — reducing atmosphere, redox cycles, and post-exposure strength, not just conductivity.
- Run the ASR arithmetic backwards from your thickness limit rather than forwards from the material. If the required thickness is below what you can co-fire reliably, the answer is a different architecture, not a different dopant.
- Price the supply risk explicitly. Verify license track record and lead time for each candidate as a schedule input, and keep a second qualified source per material.
Grade-level context on the YSZ side — the 3Y-TZP versus 8YSZ split and why the same label covers two different materials — is in our 2026 YSZ market outlook, and the sourcing side of both materials, including the specification of 10Sc1CeSZ and regional supplier options, is in the global ScSZ supplier landscape. Our SOFC materials capabilities are summarized on the SOFC materials page.
Frequently Asked Questions
Is ScSZ always more conductive than YSZ?
No, and that is a common misreading. Intrinsically, ScSZ is reported at roughly 1.5–3× YSZ conductivity, with the advantage widening as temperature falls because scandia's activation energy is lower. But the measured advantage depends heavily on processing — in one study 10Sc1CeSZ ranged from 0.019 to 0.046 S/cm at 600 °C depending on the starting powder, while 8YSZ measured 0.059 S/cm in the same work. Plain ScSZ also has a cubic-to-rhombohedral transition near 650 °C that reduces conductivity. Trust paired measurements on identically processed samples, not cross-paper comparisons.
At what temperature should I switch from YSZ to ScSZ?
There is no single crossover temperature, and any specific figure quoted without stating thickness and architecture is misleading. The decision is set by your ASR budget: work out the electrolyte thickness your architecture requires, divide by the conductivity you measure at your operating temperature, and see whether the result fits your budget. Broadly, 8YSZ is comfortable in the 850–1000 °C band, and ScSZ becomes worth its cost when the target is intermediate temperature with an electrolyte too thick for 8YSZ to carry.
Does the cerium in 10Sc1CeSZ solve the stability problem?
Partly, and it introduces another. The 1 mol% CeO₂ promotes high-symmetry, high-conductivity phases and co-dopant studies find ceria among the more effective suppressors of the cubic-to-rhombohedral transformation. But the 10Sc1Ce composition is still assessed as slightly under-doped and prone to slow cubic-to-tetragonal decomposition, and in reducing atmosphere the Ce⁴⁺-to-Ce³⁺ transition has been associated with lower fracture strength. Patent literature argues for at least 2 mol% additional stabilizing oxide to properly suppress aging.
Is ScSZ mechanically stronger than YSZ?
No. An 11ScSZ electrolyte has been reported as comparable in mechanical strength to a conventional 8YSZ electrolyte. Both are fully stabilized cubic grades with no transformation toughening, with reported flexural strength around 100–180 MPa — far below the 400–1000 MPa range reported for 3 mol% tetragonal YSZ. Mechanical margin comes from cell architecture, not from the dopant.
Does choosing YSZ avoid the export-control problem?
No. China's April 2025 Announcement No. 18 controls yttrium as well as scandium, and the yttrium classification (1C908) carries no concentration floor, so YSZ powder has a high likelihood of falling in scope. What differs is the nature of the exposure: yttrium is abundant but currently repriced by regulation, while scandium is physically scarce at roughly 40 tonnes a year of global output. Both regimes are per-shipment licensing rather than embargo.
Can I use the same cell architecture for both materials?
Not necessarily, and this is worth checking before treating them as drop-in alternatives. The two differ in sintering behavior — ScSZ densifies at or below the 1400–1450 °C typical of YSZ, and with fine powders or sintering aids considerably lower — which changes the co-firing window with the electrodes. Conductivity differences also change the thickness that meets your ASR budget. Requalify the co-fire schedule and the thickness target together rather than substituting the powder.
References (Public Sources)
- Solid State Ionics — "Characteristic and challenges of scandia stabilized zirconia as solid oxide fuel cell material: in-depth review" (2023)
- Materials Advances (RSC) — "Highly conductive and stable electrolytes for solid oxide electrolysis and fuel cells" (2025)
- Powder Metallurgy and Metal Ceramics — joint impedance spectroscopy analysis of 10Sc1CeSZ and 8YSZ solid electrolytes
- Journal of the European Ceramic Society / Ceramics International — 8YSZ ionic conductivity across conventional and non-conventional sintering routes; effects of sintering on ceria-doped scandia-stabilized zirconia
- Solid State Ionics — conductivity aging studies on 1M10ScSZ (M⁴⁺ = Ce, Hf); XRD analysis of conductivity degradation and phase stability in SOFC zirconia electrolytes; stability of Sc₂O₃/CeO₂ co-doped ZrO₂ during SOFC operation
- Scripta Materialia — scandia-stabilized zirconia-ceria (xSc1CeSZ, 5 < x < 11) structure and conductivity for IT-SOFCs
- Journal of Physical Chemistry C — scandia as an oxygen-vacancy stabilizer in Sc–Y co-doped zirconia
- Materials & Design; Solid State Ionics — flexural strength and mechanical properties of 8YSZ and 3YSZ/8YSZ composite electrolytes
- US patents on phase-stable doped-zirconia electrolyte compositions with low degradation (additional-stabilizer content)
- Ionics — tutorial review on solid oxide fuel cells: fundamentals, materials, and applications
- China MOFCOM Announcement No. 18 (April 4, 2025)
- lanthanides.io — scandium oxide pricing; Rare Earth Exchanges and The Oregon Group — yttrium price movements
Between these two materials the expensive failure is not the wrong choice but a silent substitution — a changed stabilizer level or dopant package that surfaces months later in aging data. Nami Tech Solutions (NTS) writes the specification so that cannot arrive unannounced — stabilizer and mol%, named dopants, measurement method — and reads each production-lot COA against it before the lot ships.
