Equipment Guide20 min readUpdated 2026-11-12

Molybdenum & Tungsten Wire Furnace: Sintering to 2300 °C

Why 2000–2300 °C, vacuum and metal shields decide molybdenum and tungsten wire furnace performance — plus the temperature records our own pages disagree on.

Short answer

If you buy one of these furnaces, three decisions set the price and the yield: hot-zone metal versus graphite, published maximum temperature versus the temperature your process actually needs, and how the load is fixtured. HuanYu publishes a tungsten-wire platform at 2000–2300 °C in three chamber sizes (Φ80×120, Φ160×160, Φ200×250 mm) with 24, 45 and 65 kW, ±1 °C published uniformity and a ≤20 °C/min ramp (catalog p.32). That platform is specified with either a graphite heater plus graphite felt, or a tungsten heater plus tungsten-molybdenum metal radiation shields — the heater choice decides the insulation, and the insulation decides whether your tungsten pickups carbon or not. Everything else — vacuum level, load mass, cooling time — follows from those three choices and should be written into the RFQ rather than left to the quotation.

Key facts

  • ✓Published temperature envelope: 2000–2300 °C on the tungsten-wire platform (catalog p.32); the site's tungsten page states "Up to 2300 °C" (/products/vacuum-tungsten-wire-furnace-68/).
  • ✓Published chamber sizes: Φ80×120, Φ160×160 and Φ200×250 mm — 0.60 L, 3.22 L and 7.85 L of usable hot-zone volume (catalog p.32; volumes derived).
  • ✓Published power, ramp, uniformity: 24 / 45 / 65 kW at AC 380 V three-phase; ≤20 °C/min; ±1 °C (catalog p.32) — the ±1 °C needs a stated measurement basis before you accept it as contractual.
  • ✓Two insulation systems on one platform: graphite felt on the graphite-heater version, tungsten-molybdenum metal radiation shields on the tungsten-heater version (catalog p.32).
  • ✓Derived wall power density: 597 / 373 / 296 kW per m² of hot-zone wall across the three sizes — roughly 6–11× the 49–63 kW/m² published for our 1300 °C three-chamber platform (catalog p.35).
  • ✓Company: 60 product series, 500+ machine configurations, 10+ years of furnace building, GB/T 19001-2016 / ISO 9001:2015 certified (certificate U25Q2SH8002817R3M, valid 2025-07-24 to 2028-08-08).

What is a molybdenum or tungsten wire furnace?

It is one vacuum chamber platform sold with two hot-zone material sets. On the graphite version the heater is graphite and the insulation is graphite felt; on the metal version the heater is a tungsten wire element and the insulation is a stack of tungsten-molybdenum radiation shields, inside a double-layer water-cooled stainless shell (catalog p.32). The workload — molybdenum wire, molybdenum strip, tungsten wire, or pressed and pre-sintered powder bars — hangs or rests inside that hot zone.

The name of the machine follows the workload, not the engineering: molybdenum wire furnace, tungsten wire furnace, molybdenum strip furnace, molybdenum sintering furnace are normally the same platform ordered with different fixturing, different shields and a different temperature record. That is why the class is hard to compare, and why the specification must follow the process step rather than the product name.

Three geometries are published (catalog p.32):

Model sizeHot-zone bore × heightUsable volume (derived)Wall area (derived)PowerPower per litre (derived)Power per m² wall (derived)
Small

Φ80 × 120 mm

0.60 L

0.040 m²

24 kW

39.8 kW/L

597 kW/m²

Medium

Φ160 × 160 mm

3.22 L

0.121 m²

45 kW

14.0 kW/L

373 kW/m²

Large

Φ200 × 250 mm

7.85 L

0.220 m²

65 kW

8.3 kW/L

296 kW/m²

All six volumes, areas and densities are arithmetic derivations from the published bore and height; they are not catalog values.

Why does refractory metal need vacuum rather than air?

Because the metal you are trying to anneal is the same metal that oxidises. Molybdenum and tungsten form volatile oxides once there is free oxygen and enough temperature, and the oxide does not stay as a protective film — it leaves, and the wire loses section and gains a rough, contaminated skin. Vacuum removes the oxidant; dry hydrogen removes it faster and also reduces oxide already present.

Two mechanisms sit behind most "bad batch" complaints, and neither is visible on a quotation:

  1. 01

    Residual moisture and outgassing. Water desorbs from fixtures, shields and the workload during heat-up. A furnace that pulls 10⁻³ Pa on an empty, clean, hot chamber can sit two decades worse during the first hour of a loaded run. That is why the pressure-rise rate matters more than the headline ultimate pressure.

  2. 02

    Carbon transfer from graphite. With graphite felt and a carbon heater in the same hot zone as molybdenum above roughly 1600 °C, any residual oxygen or moisture can shuttle carbon into the metal as Mo₂C at grain boundaries. The wire then measures correctly on the micrometer and breaks on the coiler. Metal shields remove the carbon source; they do not remove the need for a leak-tight chamber.

That is why the platform publishes two insulation options rather than one. The choice is a process decision, not an upgrade, and the comparison table below sets out the trade. For a sense of what "clean" means on a production furnace, our HTCC platform publishes oxygen at ≤0.1 ppm with dew-point control from −60 to +60 °C (catalog pp.27–28) — a different machine, but the same standard of housekeeping.

What temperature does each process step actually need?

The single most common specification error in this equipment class is buying the platform's headline number instead of the process window. The published envelope is 2000–2300 °C (catalog p.32). Your process probably needs less, and one step may need more.

Process stepTypical metal temperature — general engineering range, not our published recordCovered by the published 2000–2300 °C envelope?
Vacuum bake-out / degassing of the empty hot zone

800–1200 °C

Yes, with margin

Stress relief of rolled molybdenum strip

900–1100 °C

Yes

Recrystallization anneal of molybdenum wire

1100–1400 °C

Yes

Recrystallization anneal of tungsten wire

1300–1600 °C

Yes

Sintering of molybdenum powder bars

1800–2000 °C

Yes, lower half of envelope

Sintering of tungsten powder bars

2200–2400 °C (usually in dry hydrogen)

Only to 2300 °C — top of the band falls outside

The last row decides projects. If your route is tungsten powder-bar sintering at the upper end, the published envelope does not reach it, and the honest answer is that the requirement must be confirmed against the actual charge, the atmosphere and the intended density before any quotation. Our vacuum hydrogen platform publishes 1700 °C (catalog p.30) — a sintering and metallization machine, not a tungsten sintering machine. Numbers are not interchangeable across platforms, and treating them as if they were is how a project ends up one platform short.

Two numbers in that envelope need interpretation. The ≤20 °C/min ramp (catalog p.32) is what the hot zone can sustain without cracking shields or over-driving the element; thin wire can usually be ramped faster, thick powder bars cannot, because the centre lags the surface and the lag is what creates the density gradient. The ±1 °C uniformity (catalog p.32) is a strong claim at 2300 °C and should be tied to a measurement basis in the contract: empty or loaded, at what distance from the element, with which thermocouple type, and over what soak length. A uniformity figure without those four qualifiers is marketing, not a specification.

Graphite felt or tungsten-molybdenum shields?

CriterionGraphite heater + graphite feltTungsten heater + W-Mo metal shields
Published on the same platform

Yes (catalog p.32)

Yes (catalog p.32)

Carbon pickup risk on Mo/W workload

Present above ~1600 °C when moisture is present

Removed at source

Thermal mass and cool-down

Higher mass, slower cool-down, more stored heat

Lower mass, faster cool-down

Outgassing after a wet or dirty cycle

Felt absorbs and re-releases moisture; long pump-down

Shields bake out quickly, mechanically cleanable

Mechanical robustness

Felt is fragile at the door seal and element feedthroughs

Shields deform if bumped but do not crumble

Consumable cost

Felt and graphite heater are lower cost, replaced more often

Metal is higher first cost, replaced after sag or erosion

Best fit

Lower half of the envelope, graphite-compatible loads, cost-driven cycles

Refractory-metal workload, cleanliness-critical cycles, fast turn-around

The rule from the published configurations: the heater and the insulation are a matched pair, not two independent options. A tungsten heater with graphite felt keeps the contamination; a graphite heater with metal shields keeps the thermal mass — both are possible, both are almost always wrong.

What hot-zone size and power do you need?

Start from the load, not the chamber. Wire and strip are surface-dominated: what limits the batch is exposed surface, not mass. A 7.85 L chamber of tightly coiled 0.5 mm wire holds more kilograms than the same chamber of 2 mm rod while annealing far worse, because the inner turns never see the hot zone.

Three sizing rules that hold across the published range:

  • ✓Load envelope against bore. Keep the loaded fixture under roughly 70–80 % of the bore so the shields still see the element rather than the workload. Beyond that the outer turns run cold and the inner turns run hot.
  • ✓Power per litre falls as the chamber grows. 39.8 kW/L at 0.60 L against 8.3 kW/L at 7.85 L (derived). A large chamber quoted at small-chamber power density will miss its ramp at temperature; a small chamber quoted at large-chamber density will run its element near the limit continuously.
  • ✓Ask for the ramp at temperature. 20 °C/min from 800 °C to 2000 °C is the number that matters; the same figure quoted from cold is easy and useless.

Why does the smallest chamber need the most power per litre?

Because heat loss at these temperatures is dominated by radiation from the hot-zone wall, and the wall area does not shrink as fast as the volume does. Our three published sizes make the effect unusually visible:

  • ✓Volume rises 13.0× from the small to the large chamber (0.60 → 7.85 L).
  • ✓Wall area rises 5.5× (0.040 → 0.220 m²).
  • ✓Published power rises only 2.7× (24 → 65 kW).

So power tracks wall area far more closely than volume: power per litre collapses from 39.8 to 8.3 kW/L while power per m² of wall falls only from 597 to 296 kW/m². Our 1300 °C three-chamber platform shows the same relationship — 35 / 80 / 140 kW against 0.66 / 1.28 / 2.88 m² gives 53 / 63 / 49 kW per m² (catalog p.35, derived). Compared on that one metric, a 2000–2300 °C hot zone needs 6–11× the wall power density of a 1300 °C hot zone of similar geometry, consistent with radiative loss scaling as the fourth power of absolute temperature: (2573 K / 1573 K)⁴ ≈ 7.2.

The commercial consequence: do not size a 2300 °C furnace by scaling a 1300 °C one. The extrapolation understates power, element loading and cooling-water demand, and it quietly understates the price. Size from the published high-temperature set.

How should wire and strip be loaded?

Fixture design decides whether the published ±1 °C ever reaches the workload. Three arrangements cover most refractory-metal work:

  • ✓Vertical hanging. Wire hangs from a molybdenum hook bar, free at the bottom. Default for wire annealing: keeps the strand straight, avoids contact marks, lets it contract without buckling. The limit is creep — at the top of the envelope a long strand sags under its own weight.
  • ✓Horizontal on a molybdenum or ceramic setter. For strip and short rod. Contact points shadow the underside, so the strip is flipped for a second pass or the run is long enough for conduction to even out the shadow.
  • ✓Loose batch in a molybdenum basket. Fastest to load, worst for uniformity. Fine for degassing and stress relief, not for a recrystallization anneal that will be inspected by grain size.

Two hard rules. No graphite fixtures inside a metal-shield hot zone — the fixture reintroduces the carbon source the shields were bought to remove. And no mixed loads — molybdenum and tungsten in the same batch will not reach the same grain structure at the same setpoint, and the batch record will be unusable.

What vacuum level and pressure-rise rate should you accept?

The catalog publishes no ultimate pressure for the tungsten-wire platform (catalog p.32) — only temperature, sizes, power, ramp and uniformity. That gap is the first thing to close in an RFQ. Adjacent platforms give the workable benchmark:

Published platformUltimate pressurePressure-rise rateSource
High-vacuum annealing / brazing, 1300 °C, three chambers

≤2 × 10⁻⁴ Pa (provisional, OCR not re-verified)

≤0.67 Pa/h

catalog p.35

Vacuum hot press, 1000–2300 °C

10 Pa / 5 × 10⁻³ Pa / 5 × 10⁻⁴ Pa, any one selectable

0.67 Pa/h

catalog p.34

Carbon tube sintering, 2000 °C (ZKT)

6.67 × 10⁻³ Pa

≤2 Pa/h

catalog pp.19–20

3D-printing vacuum/atmosphere furnace

≤5 Pa

≤1 Pa/h

catalog p.25

Two readings matter. First, the same catalog quotes 0.67 Pa/h and 2 Pa/h, a three-fold difference, which tells you the number describes the hot zone and the pumping set rather than a universal standard of workmanship. Second, Pa/h alone is not comparable between chambers of different size — the same leak produces a larger pressure rise per hour in a small chamber. Convert to a leak-rate-like figure before comparing quotations:

  • ✓Φ80 × 120 mm chamber, 0.60 L: 0.67 Pa/h is ≈ 0.0004 Pa·m³/h
  • ✓Φ160 × 160 mm chamber, 3.22 L: ≈ 0.0022 Pa·m³/h
  • ✓Φ200 × 250 mm chamber, 7.85 L: ≈ 0.0053 Pa·m³/h

(All derived; the conversion is pressure rise × chamber volume.) Ask every supplier for the pressure-rise rate and the volume that produced it. A supplier quoting Pa/h without a volume is quoting a number you cannot use — the point we make in the manufacturer audit.

What has to be clean is the workload surface at temperature, governed by dew point and leak rate together, not by the gauge reading at the end of pump-down. When a run comes out discoloured, the useful question is not "what was the vacuum?" but "what was the pressure rise after a 30-minute soak at 1200 °C with the load in?"

Troubleshooting: what actually goes wrong

SymptomMost likely causeFirst checkCorrective action
Wire discoloured or dull grey after anneal

Water vapour or oxygen ingress during heat-up

Pressure-rise rate with a hot, empty chamber

Extend the bake-out, re-seat door and feedthrough seals, check dew point on gas inlet

Wire brittle, breaks on coiling despite correct elongation test

Carbon pickup forming Mo₂C at grain boundaries; or recrystallization too coarse from an over-long soak

Graphite fixtures or felt in the hot zone; soak time versus batch record

Switch to W-Mo shields, remove graphite fixtures, shorten the soak

Recrystallization uneven along the strand

Cold ends from hanging too close to the shield stack, or excessive coil density

Load map versus hot-zone map at temperature

Shorten the hanging run, open the coil, raise the soak, reduce the load

Element sag or early failure

Continuous operation near the published ceiling without margin

Logged setpoint versus published envelope

Lower setpoint, increase element size, accept a longer ramp

Pump-down slows cycle by cycle

Felt absorbing moisture, or a slowly growing leak at the water-cooled shell

Rate-of-rise trend across batches, not a single value

Replace felt or bake the shields, pressure-test the shell

Arcing during heat-up

Volatile condensate bridging the element and shield at low pressure

Pressure trace in the 10–100 Pa band

Slow the ramp through the band, add a hold, clean shields

Thermocouple reads low at the top of the range

Type C or W-Re drift, or contamination of the thermocouple well

Cross-check against an optical reading or a second couple

Replace the couple, keep a spare pair, log hours

Uniformity passes empty but fails loaded

Fixture shadowing the shields, or load mass above the ramp capability

Survey with the production load in place

Redesign the fixture, split the batch

What our own pages disagree on

This is worth publishing rather than hiding, because the disagreement is exactly what a buyer would hit while comparing.

Four HuanYu pages describe what a buyer would reasonably call the same equipment class, and they publish four different temperature records:

PagePublished temperaturePublished vacuum, size, power
/products/vacuum-tungsten-wire-furnace-68/

"Up to 2300 °C"

None published

/products/vacuum-molybdenum-strip-furnace-69/

"1300 / 1200 / 600 °C configurations"

None published

/products/vacuum-molybdenum-wire-sintering-furnace/

"1500 °C" for all three models (FF-400, FF-200, VMF-32-15)

None published

/products/vacuum-molybdenum-wire-furnace/

Not published — "Configuration specific"

None published

The catalog's tungsten-wire platform is 2000–2300 °C (catalog p.32). So the site currently offers 600, 1200, 1300, 1500 and 2300 °C as the top temperature of a refractory-metal furnace, and none of the four pages publishes a single vacuum value, chamber size or power figure. A 600 °C ceiling is not a molybdenum furnace specification — it is a template artefact.

Two consequences. First, treat any of those numbers as a placeholder until a datasheet confirms it — the specification that counts carries the chamber size, heater and shield material, ultimate pressure, pressure-rise rate, power and uniformity basis. A product page is a starting point, not a record. Second, the inconsistency is itself a signal: pages that copy a temperature string across models and leave vacuum and power blank were generated from a template, not from a test report. When you audit any supplier, ours included, the fastest filter is to ask for the pressure-rise rate and the chamber volume together.

The 1300 °C brazing value was corrected from 1550 °C to 1300 °C to match the catalog during a previous review — the standard we are trying to hold: where a page and the catalog disagree, the catalog wins and the page moves. The four pages above are queued for the same treatment.

How to write the RFQ for a refractory-metal furnace

Ten lines, in this order. Each one produces an answer that can be compared across suppliers.

  1. 01

    Process step and acceptance criterion — sintering a powder bar, recrystallizing wire, stress-relieving strip; grain size, elongation, density or sag test.

  2. 02

    Workload geometry and mass per batch — wire diameter or strip thickness, strand length, total kilograms, fixture type.

  3. 03

    Maximum temperature and soak time — your setpoint plus margin, not the platform ceiling.

  4. 04

    Ramp rate at temperature — °C/min between two stated temperatures, not from cold.

  5. 05

    Atmosphere — vacuum only, or vacuum plus a partial pressure of argon or dry hydrogen; if hydrogen, the safety requirement.

  6. 06

    Ultimate pressure, pressure-rise rate, and the chamber volume that produced them — all three together, or the number is unusable.

  7. 07

    Uniformity and its measurement basis — loaded or empty, survey length, thermocouple type, number of points.

  8. 08

    Heater and insulation material — graphite plus felt, or tungsten plus W-Mo shields; state whether graphite fixtures will be present.

  9. 09

    Cool-down time from soak to unloading — a cycle-time number that belongs in the quotation, not in the commissioning report.

  10. 10

    Documentation — datasheet per unit, calibration certificates, heater and shield material certificates, and the ISO 9001 certificate number for the build site.

The document set is the same one we audit in the manufacturer audit guide.

Frequently asked questions

What is the difference between a molybdenum furnace and a tungsten furnace?

The heater and shield material, and therefore the ceiling and the contamination risk. The published tungsten-wire platform reaches 2000–2300 °C with a tungsten element and W-Mo shields (catalog p.32). A molybdenum-heated version of the same chamber runs lower and costs less; the body, pumping set and controls are the same architecture. "Molybdenum furnace" also loosely means a furnace with a molybdenum hot zone, which is a different question — that one is covered in the molybdenum hot zone guide.

Can one furnace sinter both molybdenum and tungsten?

Yes, if the published envelope covers your tungsten setpoint and the hot zone is metal-shielded. The constraint is the top of the tungsten sintering band — general practice sits at roughly 2200–2400 °C, and the published platform tops out at 2300 °C. Batch them separately: mixed loads do not reach the same grain structure at one setpoint.

Why is my molybdenum wire brittle after annealing?

Three candidates, in order of frequency: oxygen or water ingress during heat-up; carbon pickup from graphite felt or graphite fixtures; an over-long soak producing unnecessarily coarse recrystallized grain. Check the pressure-rise rate with a hot, empty chamber first, then look for carbon in the hot zone, then cut the soak.

Graphite felt or metal shields for a refractory-metal workload?

Metal shields, when the workload is the refractory metal and carbon pickup would scrap the batch. Graphite felt stays right for graphite-compatible loads in the lower half of the envelope, where lower cost and higher thermal mass are advantages. The platform publishes both; it is a process decision, not an upgrade (catalog p.32).

What vacuum level do I need at 2000 °C?

There is no single number, and the platform's ultimate pressure is not published (catalog p.32). What you can compare is pressure-rise rate against chamber volume: our 1300 °C platform publishes ≤0.67 Pa/h (catalog p.35), our carbon-tube platform ≤2 Pa/h at 6.67 × 10⁻³ Pa (catalog pp.19–20). Convert both to Pa·m³/h before comparing, and require the volume to be stated.

What pressure-rise rate should I accept?

The rate that keeps your surface clean at temperature, quoted with the chamber volume. Published values across our platforms run 0.67 to 2 Pa/h (catalog pp.19–20, p.34, p.35); for a 0.60 L chamber, 0.67 Pa/h converts to roughly 0.0004 Pa·m³/h, and for 7.85 L to roughly 0.0053 Pa·m³/h (derived). The small chamber is the demanding case.

How much wire can I load per batch?

Not a published figure, and it should not be quoted as one, because it depends on fixture geometry rather than chamber volume. Keep the loaded fixture under roughly 70–80 % of the bore, hang strands rather than coil them densely, and validate with a loaded uniformity survey. A supplier who quotes kilograms without asking your strand length and fixture design is guessing.

Do I need hydrogen instead of vacuum?

Hydrogen reduces oxide already on the surface and transfers heat better; vacuum avoids the safety engineering. Our vacuum hydrogen platform publishes 1700 °C with multiple explosion-protection measures (catalog p.30) — a metallization and sintering machine. A 2300 °C hydrogen route is a different design and must be specified as such.

What uniformity can I expect at 2300 °C?

Published: ±1 °C (catalog p.32). Before accepting it as contractual, fix the measurement basis — loaded or empty, number and position of survey points, thermocouple type, soak length. At 2300 °C an empty-chamber survey is not evidence of loaded performance, and the fixture usually matters more than the controller.

Why do your pages show 600, 1500 and 2300 °C for the same kind of furnace?

Because four product pages were generated from templates and only the tungsten page reflects the catalog's 2000–2300 °C platform (catalog p.32). The 600 °C figure on the strip page and the 1500 °C figure on the wire-sintering page are placeholders, not specifications. Ask for a datasheet; treat this page's key-facts table as the record until those pages are corrected.

Need help applying this?

Bring the guide into a real equipment conversation.

Discuss your process