Low-sodium alumina: Na₂O content, why it matters and how to specify it

Low-sodium (low-soda) alumina usually means less than 0.1% Na₂O, the line Huber Advanced Materials uses for its low-alkali calcined aluminas. Our standard alumina is a low-sodium grade: the raw alumina carries 0.03% sodium, and the yearly lab analyses of our 99.5% alumina tube read 0.052%, 0.043%, 0.054% Na₂O (2024–2026), well under that line. Sodium matters where alumina sits next to semiconductor oxides or has to insulate when hot. For platinum thermocouples the documented contaminants are iron and silica, so put those on the specification too.

This page is about chemistry. Density and thermal data by purity are in alumina density and properties by purity, and the C 799 class and its equivalents in C 799, C 610 and C 530.

Na₂O and the other impurities in our lab reports

Low sodium starts with the powder: the raw alumina we form and fire carries 0.03% sodium. Chinalco Testing Technology (Zhengzhou) Co., Ltd. has analysed one of our alumina tubes every year since 2024, and Na₂O stayed between 0.043% and 0.054%.

Our brochure table for 99–99.5% corundum gives SiO₂ <0.2% and Fe₂O₃ <0.05–0.1%. Our Na₂O figures are the yearly analyses of the 99.5% grade above. For scale, JUMO gives 3% alkali for KER 610 mullite (C 610), and Haldenwanger gives 3.0% for its C 610 material.

Why sodium matters

• Insulation at temperature. NGK Spark Plug, in its patent on alumina insulators, writes that the Na component in alumina has high ionic conductance and lowers insulation resistance, especially at 500 °C and above. It calls alumina with less than 0.1 wt% of the Na component low soda, 0.1–0.2% medium soda and anything above that regular soda.
• Semiconductor processing. A Chalmers University thesis on SiC transistors found mobile sodium in gate oxides grown in an alumina furnace. The sodium reduced device reliability and shifted the threshold voltage, and the thesis presents a way to remove it. If a tube, boat or insulator sits in an oxidation or diffusion furnace, ask for Na₂O on the analysis. The flat and round twin-bore insulators used in diffusion-furnace thermocouples are in insulators for profile and spike thermocouples.
• Raw material. Alumina powder makers sell low-soda grades. Huber lists low alkali content (Na₂O below 0.1%) next to good electrical insulation for its Martoxid line.

Platinum thermocouples: iron and silica are the documented risk

We found no source that names sodium as a poison for type S, R or B wire. The documented ones are iron and silica:

• NASA's 1970 report TN D-7027 cites earlier work in which the main cause of drift in noble-metal thermocouples in alumina insulators was iron transferred from the insulator, with the change tracking the impurity content of the insulator grade.
• In the same report, the alumina insulators had magnesium, silicon, sodium and iron as the main impurities. After up to 10,000 hours at 1,600 K, the noble-metal wires had picked up aluminium, iron, silicon and magnesium.
• In NASA's 1969 test (TN D-5287, 3,700 hours at 1,530 K), iron showed the largest increase at the junction, and the alumina insulators were the most likely source.
• JUMO warns that in a reducing atmosphere as little as 0.2% SiO₂ in the insulator or tube is enough to form brittle platinum silicides.

So for the ceramic protection tube or insulator of a platinum thermocouple, a Na₂O limit alone isn't enough: set Fe₂O₃ and SiO₂ limits as well. Which grade and class suit platinum wire is covered in the grades guide.

How to specify low-sodium alumina

Put these on the drawing or the request:

1. The maximum Na₂O as % by weight, for example "Na₂O ≤ 0.1%".
2. Maximum Fe₂O₃ and SiO₂ if the part goes near platinum thermocouple wire.
3. The analysis method. Our lab measures Na₂O to GB/T 6609.1-2018.
4. Whether a typical value, a yearly report or an analysis of your own lot is needed.
5. The Al₂O₃ grade (99.5%, C799, as standard) and the part: size, length and quantity.

We confirm the grade, dimensions and tolerances in each quotation. Our yearly lab reports are available on request. If you need another analysis, such as a test of your own lot, say so on the request and the quotation states what we can provide.

Questions buyers ask

What is low-sodium alumina?

Alumina with little sodium oxide, usually less than 0.1% Na₂O. Huber Advanced Materials uses that line for its low-alkali calcined aluminas, and NGK Spark Plug's patent US 6,632,381 calls alumina with less than 0.1 wt% of the Na component low soda, 0.1–0.2% medium soda and more than that regular soda.

How much Na₂O is in your 99.5% alumina?

It is our low-sodium grade: the raw alumina carries 0.03% sodium, and Chinalco Testing analyses a sample of our alumina tube every year: 0.052% Na₂O in 2024, 0.043% in 2025 and 0.054% in 2026. If you need a limit, put the maximum Na₂O and the analysis method on the request.

Why does sodium matter in alumina?

Sodium moves. NGK's patent says the Na component in alumina conducts ions and lowers insulation resistance, especially at 500 °C and above. In semiconductor work, a Chalmers University thesis found mobile sodium in oxides grown on SiC in an alumina furnace, which reduced device reliability and shifted the threshold voltage.

Does sodium poison platinum thermocouples?

None of the sources we use names sodium. NASA traced drift of platinum thermocouples in alumina insulators to iron picked up from the insulator, and JUMO warns that 0.2% SiO₂ is enough to form brittle platinum silicides in a reducing atmosphere. For type S, R and B sheaths, put Fe₂O₃ and SiO₂ limits on the request as well as Na₂O.

Sources

1. Huber Advanced Materials: Martoxid calcined aluminium oxides (low alkali content, Na₂O < 0.1%)
2. NGK Spark Plug: US patent 6,632,381 B2, Production process for alumina based insulator (2003), background on sodium and soda classes
3. Allerstam, Electrical properties of thermal oxides on SiC manufactured in an alumina furnace, Chalmers University thesis, 2006
4. G. E. Glawe and A. J. Szaniszlo, NASA TN D-7027 (1970): thermal emf change of noble- and refractory-metal thermocouples at 1600 K
5. A. J. Szaniszlo, NASA TN D-5287 (1969): thermal emf change for Pt-13Rh/Pt thermocouples at 1530 K for 3700 hours
6. JUMO data sheet 90.1000 (German), KER 710 and KER 610
7. Haldenwanger: Ceramics for Kilns brochure, page 9 (alumina Al₂O₃ content, DIN VDE 0335 type, water absorption, leakage rate, density)
8. Our own figures: the factory’s standard-size list, 2026 orders and test report, as recorded in our product catalog.