A gas pressure sintering furnace densifies ceramic parts under 1–10 MPa of nitrogen or argon rather than at atmospheric pressure. The gas does two jobs at once: it suppresses decomposition at 1900–2000 °C, and it collapses the last closed pores. HuanYu publishes a 2000 °C platform with Φ200 × 230 mm and Φ300 × 300 mm crucibles at up to 9.8 MPa gas pressure (catalogue p.34).
Specification summary
A gas pressure sintering furnace is specified by the pressure it can hold at temperature, not by the temperature it can reach. Four numbers decide whether a quotation fits your part: the maximum working pressure and the temperature at which that pressure is still available; the usable hot-zone diameter and height inside the crucible; the vacuum the vessel reaches before backfill, because that sets the oxygen and moisture your powder sees; and the cooling rate under pressure, because that sets the cycle time and the residual stress. Everything else — heating element, insulation, control zones, safety code — follows from those four. A furnace rated 2000 °C at atmospheric pressure is a different machine from one rated 2000 °C at 9.8 MPa, and only the second one can sinter silicon nitride.
Key facts
- ✓The published HuanYu pressure-sintering platform is rated 2000 °C with two crucible sizes, Φ200 × 230 mm and Φ300 × 300 mm, isostatic graphite heating, ±1 °C control accuracy, PID auto-tuning, ultimate vacuum ≤5 Pa, gas pressure ≤9.8 MPa with pressure fluctuation ≤0.2 MPa, nitrogen or argon atmosphere, and 80 kW / 130 kW power. Source: catalogue p.34 (气压烧结炉).
- ✓It is a batch (periodic) machine, published as suitable for silicon nitride components, phosphor powders and special-shaped ceramics densified in high-pressure nitrogen or argon. Source: catalogue p.34.
- ✓Pressure is the process variable, not an accessory. The catalogue publishes a permissible fluctuation of ≤0.2 MPa against a 9.8 MPa ceiling — roughly 2% — which is a process-control statement, not a mechanical one.
- ✓The vacuum figure is deliberately rough. Ultimate vacuum ≤5 Pa (catalogue p.34) is mechanical-pump territory. That is correct for a pressure vessel: it is purged and backfilled, not pumped to 10⁻³ Pa, and oxygen control comes from purge cycles and gas purity rather than from pump speed.
- ✓The site's own product page does not yet publish a pressure figure.
/products/vacuum-pressure-sintering-furnace-92/currently publishes1400 / 1600 / 1800 / 800 °C configurationsandVacuum / controlled atmosphere, with the generic family applications (vacuum heat treatment, brazing, sintering, metal melting). It reads as a family template and its temperature set does not match the catalogue's 2000 °C pressure-sintering record. Treat the catalogue page as the technical record and the web page as a publication defect to fix — see the notes at the end. - ✓Zero information layer exists for this cluster.
/resources/gas-pressure-sintering-furnace/,/resources/gas-pressure-sintering/,/resources/vacuum-pressure-sintering-furnace/and/resources/pressure-sintering-furnace/all return 404, and the 924-word/resources/vacuum-sintering-furnace-selection-guide/contains no instance ofgas pressure,pressure sintering,silicon nitrideorHIP. - ✓Search evidence.
vertical gas pressure sintering furnaces industrial component productionshows 24 impressions at average position 38.4 with zero clicks over six months, and the URL Google returns for it is/products/— the category index. No dedicated page competes for the query. - ✓Company: 60 product series, 500+ models, 10+ years of manufacture, GB/T 19001-2016 / ISO 9001:2015 certificate U25Q2SH8002817R3M (issued 2025-07-24, valid to 2028-08-08). Source:
/about/.
What gas pressure sintering does that pressureless sintering cannot
Two different problems are solved by the same gas, and confusing them is the most common specification error in this category.
Problem one is chemical. Silicon nitride is not a stable compound at sintering temperature in a one-atmosphere furnace. Above roughly 1800 °C it decomposes to silicon and nitrogen, and the nitrogen leaves with the exhaust. The material loses mass, the surface roughens, and the part that remains is not the part you pressed. Raising the nitrogen partial pressure around the part shifts that equilibrium: at a few megapascals of nitrogen, the practical ceiling moves above 2000 °C, which is where the density and the microstructure you actually want become reachable. This is why the process exists. A gas pressure sintering furnace is, in its original purpose, a device for buying temperature on a nitride without destroying it.
Problem two is physical. Once a compact has densified to the point where its pores are closed and isolated, those pores contain gas at whatever pressure surrounded the part. If that is one atmosphere, the pores can only shrink by solid-state diffusion, and some of them never do. Raising the external pressure to several megapascals adds an isostatic driving force for the final few percent of density, and at 1900–2000 °C the material creeps readily enough to respond to it. The result is that the last 2–5% of porosity closes in a hold of an hour or two instead of not closing at all.
The magnitude is worth stating, because it explains why the chemical role dominates. The capillary stress holding a pore open is roughly 2γ/r; for a 1 µm pore in a ceramic with about 1 J/m² surface energy that is on the order of 2 MPa (derived). A furnace at 9.8 MPa therefore adds a pressure comparable to, not overwhelmingly larger than, the force already resisting densification. Gas pressure sintering is not hot isostatic pressing. It works because the temperature is high and the pores are already few, not because the pressure is enormous.
That distinction is what separates the three commercial answers to "I need full density": gas pressure sintering at 1–10 MPa, hot isostatic pressing at 100–200 MPa, and uniaxial hot pressing (published on HuanYu's platform as up to 2300 °C with 5–200 T on catalogue p.34). The comparison table further down makes the trade explicit.
The closed-porosity gate: where most gas pressure cycles fail
The single rule that governs cycle design is that you may not pressurize until the porosity is closed. Everything else in the recipe is scheduling around that fact.
If open porosity remains when the gas pressure comes on, the pressurizing gas enters the pore network. On depressurization at the end of the cycle, that trapped gas has no path out and expands as the external pressure falls. The part swells, blisters or delaminates — usually not visibly until the door opens, which is why the defect is so expensive. The failure is not gradual; a batch that was 88% dense at the moment of pressurization will come out looking like a failed experiment regardless of what happened afterwards.
In practice the gate is crossed somewhere between 92% and 95% of theoretical density for most technical ceramics, which corresponds to the point at which the pore network pinches off into isolated pores (general engineering range, not a HuanYu published limit). Two consequences follow for the recipe:
The first sintering hold is the important one. Time at 1700–1850 °C under vacuum or a few tenths of a megapascal of nitrogen exists to close the pore network, not to densify to final. It is usually the longest single stage in the cycle and the one most often shortened by operators trying to raise throughput.
Pressurization is a ramp, not a switch. Bringing a vessel from 0.1 MPa to 9.8 MPa is fast mechanically and must be slow thermally: the gas entering a hot zone cools it locally, and the cold front moving through a load of stacked parts produces exactly the density gradient the process is supposed to remove. Published practice on the HuanYu platform constrains pressure to ±0.2 MPa (catalogue p.34); the rate at which you approach the set point is a recipe decision you should write down.
≤5 Pa, then backfill; two or three cycles
Vacuum, then N₂
Remove air and moisture before any heat; the published ≤5 Pa is the pump-down target, not a process vacuum
1000–1400 °C
Vacuum or 0.1 MPa N₂
Burn out residual binder and lubricant, desorb moisture from the load and insulation
1700–1850 °C
0.1–0.5 MPa N₂
Close the pore network — the gate. Duration is set by section thickness
1–9.8 MPa while ramping to peak
N₂ (nitrides) or Ar (oxides, carbides)
Suppress decomposition and collapse residual closed porosity
Peak temperature, 0.5–3 h
Same gas at working pressure
Final densification and microstructure development
Cool to ~1200 °C before venting
Gas at pressure, circulated
Avoid thermal shock and re-expansion of dissolved gas
Below 100–200 °C
Vent to atmosphere
Safe access; vessel and load at ambient
The published HuanYu platform states 2000 °C maximum and ≤9.8 MPa gas pressure (catalogue p.34). Those are ceilings, not a recipe: whether your part sees 1850 °C at 4 MPa or 1980 °C at 9 MPa depends on the additive system and the section thickness, and any supplier who quotes a "standard" profile before seeing your part is quoting someone else's process.
Nitrogen or argon: the gas choice is a materials decision
The gas is not interchangeable, and the choice usually follows from the powder rather than from the furnace.
Si₃N₄, sialon, AlN, nitride-bonded composites
Suppresses nitride decomposition at 1800–2000 °C; chemically compatible with the matrix
Nitrides some metals and can react with oxide additives; higher gas purity matters because any oxygen in the supply is an oxidant at 2000 °C
SiC, ZrO₂, Al₂O₃, spinel/AlON/YAG transparent ceramics, WC-Co
Fully inert; denser than nitrogen, so forced cooling is more effective per unit volume
No decomposition suppression, so nitrides cannot be run at the same temperature; roughly an order of magnitude more expensive per cubic metre
Graded nitride/oxide composites, process development
Tunes nitrogen activity independently of total pressure
Needs a mixing panel and a gas analysis record; rarely worth it in production
Not at pressure
—
A hydrogen partial pressure inside a 9.8 MPa vessel is a safety case no standard furnace is built for. Reduction, if needed, happens in a separate vacuum or partial-pressure step
Two supply facts are worth costing before the furnace is ordered. Gas demand per cycle scales with vessel volume times pressure: a 100 L vessel taken to 9.8 MPa consumes about 1 Nm³ of gas per fill before purge losses, and a 300 L vessel about 3 Nm³ (derived from published vessel/pressure arithmetic). Purge cycles multiply that. Over a few hundred cycles a year, that arithmetic decides whether you buy a liquid dewar, a cylinder bundle or a small on-site generator — and whether nitrogen, at a fraction of argon's price, is a hard requirement rather than a preference.
Purity is a process spec. At 2000 °C an oxygen impurity measured in ppm at the inlet is an oxidant at the part. Ask for the supply purity, the point at which it is measured, and the moisture dew point — and expect the answer to be about the gas at the load, not at the regulator.
Published HuanYu configurations
Every value below is published either on the HuanYu site or in the printed catalogue with a page reference. Where the published record gives no value, the cell says so rather than inferring it from a neighbouring model.
2000 °C
Φ200 × 230 mm and Φ300 × 300 mm
N₂ or Ar; gas pressure ≤9.8 MPa; pressure fluctuation ≤0.2 MPa; ultimate vacuum ≤5 Pa
Isostatic graphite heater, ±1 °C, PID auto-tuning, 80 kW / 130 kW
Catalogue p.34
1400 / 1600 / 1800 / 800 °C configurations
Not published
Vacuum / controlled atmosphere — no pressure figure published
Not published
/products/vacuum-pressure-sintering-furnace-92/
Up to 2300 °C; catalogue p.34 gives 1000–2300 °C
Φ160 × 160 mm and Φ250 × 250 mm; 5–200 T
Vacuum, ultimate 10 Pa / 5×10⁻³ Pa / 5×10⁻⁴ Pa selectable; 0.67 Pa/h rise rate
Graphite or molybdenum wire heating, ±1 °C, ≤20 °C/min, 80 / 130 kW
/products/vacuum-hot-pressing-furnace-70/, catalogue p.34
2000 °C
Not published
Pressure vessel, working pressure not published
Not published
—
Read across the first two rows and the publication problem is visible. The catalogue describes a 2000 °C pressure-sintering machine with a 9.8 MPa vessel; the web page describing the same family publishes four temperature configurations and no pressure at all, and its application list is the generic vacuum-furnace sentence. A buyer who reads only the website cannot tell that the platform exists, which is consistent with the search evidence: Google answers vertical gas pressure sintering furnaces industrial component production with the /products/ index rather than with any product page.
One deliberate omission. The catalogue does not pair 80 kW and 130 kW to the two crucible sizes. Taking the obvious pairing gives roughly 11 kW/L on Φ200 × 230 mm (7.2 L) and 6 kW/L on Φ300 × 300 mm (21.2 L) — both derived, and both plausible for a 2000 °C graphite hot zone. The reverse pairing would give 3.8 kW/L and 18 kW/L, which is a different machine class. Because the published record does not state the pairing, this guide states neither figure as fact.
Another deliberate omission. Rise rate appears as 0.67 Pa/h on catalogue p.34, on a spread where the pressure-sintering and hot-pressing tables are set in two columns. The value is credible for both platforms and is published for the plate, but this guide attributes it to the page rather than to the individual model, and it should be confirmed against the factory specification sheet before it appears in a contract.
Gas pressure sintering, hot pressing and HIP: three ways to apply pressure
Buyers arrive at this category from three directions, and the three processes are not substitutes.
Inert gas, 1–10 MPa (HuanYu publishes ≤9.8 MPa)
Mechanical, 5–200 T through a graphite die (HuanYu p.34)
Inert gas, 100–200 MPa
Good — isostatic, no die, stacked parts
Poor — simple discs, blocks, plates; pressure axis only
Good, but capsules or pre-sintered parts are needed
2000 °C published; 1850–2000 °C typical for Si₃N₄
1000–2300 °C published
1000–2000 °C, material-dependent
One cycle, 8–24 h
One cycle, but a die per part geometry
Usually a second operation after sintering
Batch, high load density, complex geometry in one load
Batch, one or few parts per die
Batch, often toll-processing
Silicon nitride bearing balls, cutting inserts, turbocharger rotors, substrates; net-shape parts that must not be machined after sintering
Research samples, functionally graded discs, materials that will not densify without mechanical pressure
Aerospace and medical critical parts, transparency, closing porosity in pre-sintered carbide
Bloating when pressurised before closed porosity
Density gradient along the pressing axis, die reaction
Capsule failure, surface contamination, cost
The practical decision rule: if the part has to come out at final shape, gas pressure sintering is the only one of the three that will do it at production cost. Hot pressing needs a die the shape of the part, and a die that survives 2000 °C in nitrogen is a consumable with a short life. HIP gives better properties but is a second operation on a part that has already been sintered, and encapsulating a complex ceramic is often the hardest step in the whole route.
There is a fourth route worth naming because it is usually cheaper: pressureless sintering to closed porosity followed by HIP only if the acceptance test fails. It is cheaper per part when the yield is high and much more expensive when it is not, which is why it is common in carbide and rare in structural nitride.
Cooling under pressure: the half of the cycle nobody quotes
Furnace quotations describe heating. Cycle time is decided by cooling, and in a pressure vessel cooling is a design feature rather than a consequence.
A vacuum furnace cools by radiation, which is slow and gets slower as the load mass rises. A gas pressure sintering furnace can cool by forced convection: a blower circulates the working gas through an external heat exchanger and back into the hot zone, and because the vessel is already at several megapascals the gas has enough density and heat capacity to carry real power. Published platform practice is to keep the load under pressure until it is cool enough that depressurization cannot cause re-expansion or thermal shock — in the 1000–1200 °C region for most ceramics (general engineering range).
Three consequences for specification:
- ✓Cooling rate is a number you should demand. Ask for the cooling time from peak to 1200 °C with a stated load mass, not for a rate in °C/min measured on an empty chamber. An empty-chamber figure is a property of the blower; a loaded figure is a property of the machine.
- ✓Cooling rate is also a quality variable. Fast cooling under pressure raises residual stress in thick sections; slow cooling costs hours per cycle. Silicon nitride tolerates thermal shock well, but a 40 mm cross-section is not a 4 mm one.
- ✓Depressurization is a thermal step with a mechanical deadline. Venting early is the fastest way to produce bloated parts from an otherwise correct cycle, and the interlock that prevents it is worth more than any control feature on the spec sheet.
Gas pressure sintering troubleshooting: symptom, cause, fix
Pressure applied before the pore network closed, or first hold too short for the section
Density vs position in the load; log of pressure against temperature
Extend the 1700–1850 °C hold; delay pressurization until closed porosity is confirmed on a coupon
Open porosity at pressurization; gas trapped in the part
Was the part ≥92–95% dense when pressure came on?
Fix the gate before changing anything else; strip and re-sinter only if the part never went above the gate
Si₃N₄ decomposition — peak temperature above what the nitrogen pressure supports
Peak temperature vs pressure at that moment; gas purity
Raise nitrogen pressure at peak, lower peak temperature, or both
Carbon from graphite insulation or setters; oil backstreaming from the pump
Insulation condition, pump oil, setter material
Replace aged felt, service the pump, use BN-coated or ceramic setters
Reaction between the part and graphite at temperature in nitrogen
Setter material and contact area
Switch to a non-reactive setter or a boron nitride coating; reduce contact area
Uneven gas flow, cold zone near the inlet, leak at a feedthrough
Flow path, inlet position, leak-up rate with the pump isolated
Re-route the inlet, add a diffuser, fix the seal
Exaggerated grain growth from over-temperature or long hold; additive system mismatched
Grain size and β-fraction by XRD/SEM; hold time
Cut peak temperature or hold; revisit the sintering additive package rather than the furnace
Support and creep — the part sags under its own weight at 1900 °C
Setter flatness and support spacing
Use a conforming setter, reduce span, or sinter vertically
Regulator or valve leak, seal degradation, gas supply starving at high flow
Leak-up test at working pressure; supply pressure at the inlet
Service the seal, resize the supply; a dewar frosting is a supply problem, not a furnace problem
Feedthrough leak, pressure effect on the thermocouple, shield contact
Compare against a second sensor; inspect the feedthrough
Replace the feedthrough, relocate the sensor out of the gas jet, recalibrate
Code requirement, not a furnace fault
Nameplate design pressure and inspection interval
Schedule it; a pressure vessel out of inspection is out of service
The first three rows account for most failed batches. Note that only one of them is a hardware problem: the rest are recipe and sequencing faults that a pressure-logged cycle would have shown before the door opened.
What a pressure vessel changes about procurement
A 9.8 MPa vessel is not a bigger vacuum chamber, and treating it as one is how projects slip.
Safety and code. The chamber is a pressure vessel and will be regulated as one — ASME Section VIII in the United States, PED 2014/68/EU in Europe, and an equivalent registration elsewhere (general regulatory statement; confirm the applicable code and notified body for your jurisdiction). Expect a design pressure above the working maximum, a pressure-relief device sized for the worst credible gas input, a documented hydrostatic or pneumatic test, and a nameplate that governs your inspection interval. Ask for the code stamp and the relief device calculation with the quotation, not after delivery.
Gas supply and venting. The fill arithmetic above sizes your supply. Venting is the other half: a 9.8 MPa vent is loud, cold, and must be routed to a safe location with the oxygen-displacement case considered for any enclosed space. Nitrogen asphyxiation is the safety issue in this category, not fire.
Floor and foundations. Vessel mass plus load plus insulation is a point load in a way a sheet-metal vacuum furnace is not. Ask for the loaded mass, not the empty mass.
Instrumentation through the wall. Every thermocouple, heater feed and gas line penetrates a pressure boundary. Each penetration is a leak path and a maintenance item, so a machine with one consolidated feedthrough is cheaper to keep leak-tight over ten years.
Load handling. Published crucible sizes of Φ200 × 230 mm and Φ300 × 300 mm (catalogue p.34) sound generous until the insulation, heater and support furniture are subtracted — a Φ300 mm crucible with graphite heating and rigid insulation leaves a usable diameter well under 300 mm. Specify the usable envelope in writing.
How to write a gas pressure sintering furnace specification
Ten lines, in the order a supplier needs them. Items 3, 5 and 8 are the ones that change the price.
Material system and sintering additives, with the target phase assemblage
Sets the temperature ceiling and whether the gas must be nitrogen
Required final density and the acceptance method (Archimedes, image analysis, both)
Density is the acceptance criterion, so the method belongs in the contract
Peak temperature and the pressure that must be available at that temperature — as one pair of numbers
A 2000 °C rating at 0.1 MPa is not a 2000 °C rating at 9.8 MPa
Maximum and minimum working pressure, allowable fluctuation, pressurization and depressurization rates
Process control and the vessel's relief sizing both follow from this
Usable hot-zone envelope with heater, insulation and setters installed, in mm
Nominal crucible size is not usable volume
Vacuum level before backfill, plus leak-up rate at working pressure
Oxygen and moisture at the part are set here
Gas: composition, purity, dew point, and supply mode (dewar, bundle, generator)
Running cost and the purity the process actually requires
Cooling time from peak to 1200 °C at a stated load mass, and the cooling gas path
Cycle time and residual stress
Applicable pressure-vessel code, relief device, inspection interval, and documentation package
Compliance is not negotiable and is not retrofittable
Cycle data logging: temperature, pressure, gas flow, vacuum, alarm history, export format
Every defect in this article is diagnosed from a trend, not from a memory
Frequently asked questions
What is a gas pressure sintering furnace used for?
Densifying ceramic parts that cannot reach full density at atmospheric pressure — most commonly silicon nitride, and also sialon, aluminium nitride, silicon carbide, and transparent ceramics such as spinel or AlON. HuanYu's published application record for the 2000 °C platform names silicon nitride components, phosphor powders and special-shaped ceramics (catalogue p.34).
What pressure does a gas pressure sintering furnace need?
Typically 1–10 MPa of nitrogen for silicon nitride (general engineering range); HuanYu publishes a ceiling of ≤9.8 MPa with ≤0.2 MPa allowable fluctuation (catalogue p.34). The pressure you actually need is the pressure that keeps the nitride stable at your peak temperature — not the maximum the vessel can take.
Why nitrogen rather than argon for silicon nitride?
Because the job is partly chemical. Silicon nitride decomposes above roughly 1800 °C at atmospheric pressure, and nitrogen overpressure suppresses that decomposition, which is what makes 1900–2000 °C usable. Argon is inert and gives no such suppression, so nitride sintering in argon is limited to lower temperatures. Argon is the right choice for oxides, carbides and transparent ceramics, where there is no decomposition to suppress.
What is the difference between gas pressure sintering and HIP?
Pressure magnitude, and therefore cost and shape handling. GPS runs at 1–10 MPa in a single sintering cycle and needs no capsule; HIP runs at 100–200 MPa and either encapsulates the part or treats a pre-sintered one. GPS is a production sintering process; HIP is a densification or property-upgrade step, usually second.
Can I sinter silicon nitride at 2000 °C in an ordinary vacuum furnace?
No. At 2000 °C and low pressure the material decomposes faster than it densifies. This is the reason the category exists, and it is why a furnace with a 2000 °C rating and no pressure capability is not a substitute at any price.
Does the furnace need high vacuum before pressurizing?
No, and expecting it is a common misunderstanding. The published ≤5 Pa (catalogue p.34) is mechanical-pump vacuum, which is normal and sufficient for a pressure vessel that is purged and backfilled. Oxygen control comes from purge cycles, gas purity and leak-tightness — not from reaching 10⁻³ Pa.
What density can gas pressure sintering reach?
Above 99% of theoretical for silicon nitride with a well-matched additive system, and above 99.5% for carbide and some oxides (general engineering range). Reaching the number is routine; reaching it without exaggerated grain growth, and with a repeatable Weibull modulus, is the actual engineering problem.
How long is a cycle?
Roughly 8–24 hours depending on section thickness, hold times and cooling (general engineering range). Cooling under pressure is usually the largest single block of time, and it is the stage where forced gas circulation earns its price.
What kills a gas pressure cycle most often?
Pressurizing before the porosity is closed. The part looks correct until the vessel vents, and then it swells. The fix is always upstream: a longer first hold, or a denser pre-sinter, or both.
How do I compare two quotations that look identical?
Put the temperature and the pressure in one sentence — "2000 °C at 8 MPa, held for 2 h" — and ask both suppliers to confirm it in writing, together with the usable hot-zone envelope and the cooling time at your load mass. Quotations that quote temperature and pressure separately are hiding which one is available at the same time. ---