What is the thermal shock resistance of alumina?

Low for a ceramic. In makers' quench tests on small test pieces, alumina loses its strength after a sudden drop of about 150 to 300 °C: Morgan gives 170–300 °C and Kyocera 150 °C for 99% alumina dropped into water from 550 °C. Silicon nitride survives about 800–900 °C. These are lab coupon tests, not safe steps for a real tube.

How fast can I heat an alumina tube?

Furnace makers give 1 to 5 °C per minute. Carbolite Gero limits ceramic work tubes to 5 °C/min and gives 400 ÷ inner diameter in mm as the maximum heating or cooling rate in °C/min; Nabertherm gives 300 K/h below 120 mm OD and 200 K/h above; Deltech says 100 °C per hour for any ceramic process tube. Bigger and thicker tubes belong at the slow end.

How fast should a ceramic protection tube go into a hot furnace?

Slowly, and slower for bigger tubes. In the DIN 43724 thermal shock test, as tabled by Morgan Haldenwanger, 10 mm OD sheath tubes go into the furnace at 100 cm/min, 15 mm tubes at 50 cm/min and 24–26 mm tubes at 1 cm/min. WIKA says C 799 tubes must be pre-heated and then inserted slowly.

Why do alumina crucibles crack?

The makers warn about three causes: heating or cooling too fast, putting a hot crucible on a cold surface, and a charge that expands differently from the crucible. CoorsTek limits high-alumina labware to 150 °C per hour; Almath gives 5 °C/min for larger crucibles and says to stand them on a plate of the same material.

Is mullite better than alumina for thermal shock?

Yes. The DIN 43724 sheath-tube table rates C 610 mullite medium to good and porous C 530 very good, against medium for C 799 alumina. Mullite expands less (6.0 against 8.6 × 10⁻⁶/K from 20 to 1,000 °C in Haldenwanger's table) and is less stiff. The price is a lower temperature limit and less purity, which matters for platinum thermocouples.

Alumina Thermal Shock: Why Tubes and Crucibles Crack

High-purity alumina is a poor ceramic for thermal shock. In makers' quench tests it loses its strength after a sudden drop of 150–300 °C, and tube makers rate it below mullite and porous ceramics. So alumina tubes and crucibles are heated at 1–5 °C per minute and pushed into hot furnaces slowly, and the bigger and thicker the part, the slower it goes.

The usual ways to break one are the ones the makers warn about: a cold tube pushed into a furnace at full temperature, a work tube ramped too fast, a hot crucible set down on a cold surface. Below are the published numbers with the source next to each, then what to change for each kind of part.

Why alumina cracks

When a ceramic cools fast, its surface cools before the inside does. The surface wants to shrink and the core holds it, so the surface ends up in tension. Metals would stretch; alumina cracks. Heating fast does the same thing the other way round, with the tension inside.

JUMO puts it in one line: resistance to temperature shock rises with thermal conductivity and strength and falls with thermal expansion. Alumina conducts heat well and is strong, but it expands a lot and is very stiff, so the stress builds fast. JUMO adds that thin-walled tubes are better than thick ones, and that even a hairline crack can let gases poison the thermocouple and make it drift.

Quench tests: temperature drop a ceramic survives
Material	ΔT (°C)	Source
Alumina	170–300	Morgan Technical Ceramics
99% alumina, dropped into water from 550 °C	150	Kyocera
Zirconia	300–500	Kyocera; Morgan
Silicon carbide	400	Kyocera; Morgan
Silicon nitride	800–900	Kyocera; Morgan

Small test pieces quenched in the lab, not a safe temperature step for a real tube. Morgan notes that the result also depends on the part's geometry.

Alumina, mullite and porous tubes

This is why thermocouple makers still use mullite (C 610) and porous C 530 where the temperature allows. The DIN 43724 sheath-tube table rates C 799 alumina "medium", C 610 "medium to good" and C 530 "very good". Haldenwanger's property table shows the reason:

Properties behind thermal shock, by DIN EN 60672 class
Class	Material	Expansion 20–1,000 °C (10⁻⁶/K)	Conductivity (W/m·K)	Young's modulus (GPa)	Flexural strength (MPa)	DIN 43724 shock rating
C 799	High-purity alumina	8.6	25	300–380	300	medium
C 610	Mullite	6.0	2	100	120	medium to good
C 530	Porous aluminium silicate	5.7	1.4	60	45	very good

Haldenwanger, Ceramics for Kilns, page 9 (test pieces; conductivity at 20–100 °C). Shock rating from the DIN 43724 table in Morgan Haldenwanger's tube brochure. Other makers' materials, not our parts.

Mullite conducts heat about twelve times worse than alumina, which on its own would make it worse. But it expands about 30% less and is three times less stiff, so the same temperature difference produces far less stress. The catch: C 610 is rated 1,400 °C continuous against 1,600 °C for C 799, and it isn't pure enough to protect platinum thermocouples at high temperature. The trade-offs are in C 799 vs C 610 vs C 530 and alumina maximum temperature.

Protection tubes: how fast to put them in

WIKA says C 799 tubes are only partly resistant to temperature changes, that a shock can easily cause stress cracks, and that they must be pre-heated and then inserted slowly, as should other ceramic tubes depending on the temperatures. WIKA gives no speed. The only published speeds we found are the ones in the DIN 43724 thermal shock test: the tube goes into a furnace already at its maximum continuous temperature, stays 20 minutes and comes out at the same rate, and must show no damage.

Insertion rate in the DIN 43724 thermal shock test
Class	OD × ID (mm)	Insertion rate (cm/min)	Our size
C 799	10×6	100	10×6 mm
C 799	15×10	50	15×10 mm
C 799	24×18	1	24×18 mm
C 610	10×7	100	10×7 mm
C 610	15×11	50	15×11 mm
C 610	24×19	1	–
C 530	26×18	1	–

As tabled by Morgan Haldenwanger (reference values only). DIN 43724 has been withdrawn; DIN Media recommends DIN EN 50446. Our sizes are matched by dimensions only.

Look at the jump. A 10 mm tube goes in at a metre a minute; a 24 mm tube at one centimetre, a hundred times slower. That's the diameter and wall effect in numbers. If you fit 24×18 tubes into a furnace that is already hot, plan for a slow push or a pre-heat, not a quick one.

Dry the tube first. Thermo-Kinetics notes that ceramics hold moisture at room temperature and that trapped moisture can destroy a tube under thermal shock; it pre-heats ceramic tubes to about 400 °C to drive it off.
Pull it out at the same slow rate; the DIN test withdraws the tube at the speed it went in.
Support horizontal tubes. WIKA warns that bending forces on horizontal C 799 tubes are damaging, and a tube already under bending stress has less margin for thermal stress.
Where shocks can't be avoided, put a porous C 530 outer tube over the gas-tight alumina inner tube, as WIKA and the DIN guideline describe.

Furnace work tubes: heating rate

Furnace makers don't agree on one number, but they all sit between 1 and 5 °C per minute for alumina and mullite work tubes:

Published heating and cooling limits for ceramic tubes and labware
Maker	Applies to	Limit
Carbolite Gero	Ceramic work tubes (TF data sheet)	5 °C/min
Carbolite Gero	Ceramic work tubes, heating or cooling (TZF manual)	400 ÷ ID in mm, °C/min
Nabertherm	C 799 and C 610 work tubes below 120 mm OD / from 120 mm	300 K/h / 200 K/h (5 / 3.3 K/min)
Deltech	Any ceramic process tube	100 °C/h (1.7 °C/min)
Deltech (McDanel sheet)	Ceramic furnace tubes, to reduce the risk	60 °C/h (1 °C/min)
CoorsTek	High-alumina labware, heating or cooling	150 °C/h (2.5 °C/min)
Almath	Larger alumina crucibles, heating and cooling	5 °C/min

Each maker's figure for its own furnaces or ware. Where two figures differ, the slower one is the safe choice.

Carbolite's rule scales with size: 400 divided by the inner diameter in mm gives °C per minute. For our standard furnace tubes, ID 20 to 70 mm, it gives 20.0 down to 5.7 °C/min, so Carbolite's flat 5 °C/min limit is the tighter of the two for every size we list. The rule only drops below 5 °C/min above 80 mm ID.

Size isn't the only thing that matters. Where tubes crack, and what the makers say about it:

Big and thick tubes. Carbolite: large-diameter tubes are more susceptible than small ones. Deltech advises avoiding large, thick-walled tubes, because the strain from a temperature gradient grows with tube size.
Where the tube leaves the furnace. Deltech says gradients are worst where the tube passes through the furnace wall, and Carbolite that tubes extending beyond the heated zone are more at risk. End plugs or radiation shields cut the gradient.
Below 900 °C. Carbolite says ceramic tubes are more brittle below 900 °C, and that loads pushed in too quickly crack them. Use boats with feet and low thermal mass.
Clamping. Deltech: fix the tube at one end only, so it can grow and shrink freely.

More on furnace tubes, including vacuum limits and quartz, is in alumina vs quartz tubes for a tube furnace.

Crucibles

A crucible has the same material limits, plus one problem of its own: whatever is inside. CoorsTek limits high-alumina labware to 150 °C per hour, heating or cooling, and says to use an oven or hot plate as an intermediate step if you need to go faster, and never to set heated ware on a cold surface. Almath, a UK crucible maker, adds:

Pre-heat larger crucibles, and heat and cool them at no more than 5 °C/min; small ones usually take faster rates.
Stand the crucible on a plate of the same material.
Keep it at least 25 mm, better 50 mm, away from the heating elements.
Allow for the charge expanding differently from the crucible; don't pack it tight.
Don't stack crucibles inside each other.

Our crucibles, boats and lids, in 99%, 99.5% and 99.7% Al₂O₃ from 0.5 to 1500 ml, are on the crucibles and boats page.

What to tell us when you order

We don't publish a thermal shock figure for our own grades; no test value is on record for them, and a coupon number wouldn't tell you how a 1 m closed-end tube behaves anyway. What helps is the use. If tubes have been cracking, tell us the size, how the tube goes into the furnace (cold or hot, how fast), the furnace temperature, how it's mounted, and where it breaks: at the tip, at the furnace wall or along the length. We'll say whether a thinner wall, a smaller diameter or a different grade is worth trying.

Our standard protection tubes are 99.5% Al₂O₃, comparable to C 799, including the European sizes 10×6 mm, 15×10 mm and 24×18 mm. Standard sizes are usually in stock and ship in about 3 days for a small order, about 7 for a bulk one; lengths are cut to order. See the protection tubes and furnace tubes pages.

Questions buyers ask

What is the thermal shock resistance of alumina?: Low for a ceramic. In makers' quench tests on small test pieces, alumina loses its strength after a sudden drop of about 150 to 300 °C: Morgan gives 170–300 °C and Kyocera 150 °C for 99% alumina dropped into water from 550 °C. Silicon nitride survives about 800–900 °C. These are lab coupon tests, not safe steps for a real tube.
How fast can I heat an alumina tube?: Furnace makers give 1 to 5 °C per minute. Carbolite Gero limits ceramic work tubes to 5 °C/min and gives 400 ÷ inner diameter in mm as the maximum heating or cooling rate in °C/min; Nabertherm gives 300 K/h below 120 mm OD and 200 K/h above; Deltech says 100 °C per hour for any ceramic process tube. Bigger and thicker tubes belong at the slow end.
How fast should a ceramic protection tube go into a hot furnace?: Slowly, and slower for bigger tubes. In the DIN 43724 thermal shock test, as tabled by Morgan Haldenwanger, 10 mm OD sheath tubes go into the furnace at 100 cm/min, 15 mm tubes at 50 cm/min and 24–26 mm tubes at 1 cm/min. WIKA says C 799 tubes must be pre-heated and then inserted slowly.
Why do alumina crucibles crack?: The makers warn about three causes: heating or cooling too fast, putting a hot crucible on a cold surface, and a charge that expands differently from the crucible. CoorsTek limits high-alumina labware to 150 °C per hour; Almath gives 5 °C/min for larger crucibles and says to stand them on a plate of the same material.
Is mullite better than alumina for thermal shock?: Yes. The DIN 43724 sheath-tube table rates C 610 mullite medium to good and porous C 530 very good, against medium for C 799 alumina. Mullite expands less (6.0 against 8.6 × 10⁻⁶/K from 20 to 1,000 °C in Haldenwanger's table) and is less stiff. The price is a lower temperature limit and less purity, which matters for platinum thermocouples.

Alumina thermal shock: why ceramic tubes crack, and how to stop it