What Should You Control When Quenching Slender Square and Rectangular Steel Tube?

Quenching a square or rectangular tube looks like a single operation from the outside - heat it, drop it in liquid, pull it out hard. In production, it is a sequence of separate decisions, and a mistake at any one of them shows up as either a bent tube, an uneven hardness reading, or a crack that was never visible until the part failed in service. This is a summary of the control points that actually decide the outcome, drawn from production practice rather than from the grade's mechanical property table.

A slender tube heated horizontally, whether in a chamber furnace or a bath, sags slightly under its own weight the moment it reaches temperature and loses stiffness - steel loses a significant fraction of its room-temperature elastic modulus well before it reaches typical hardening temperature, and a long unsupported span will bow measurably even under its own weight alone. Suspending the tube vertically in a salt bath furnace or a pit-type furnace removes that bending moment entirely - the tube hangs under axial tension rather than resting across a horizontal span, so there is no self-weight deflection to correct afterward.

This matters most for length-to-cross-section ratios where horizontal support points would otherwise sit far apart relative to the section's stiffness. A short, stocky section tolerates horizontal heating without much consequence; a long, thin-walled tube does not. The heating orientation decision should be made against the specific tube's slenderness, not applied as a blanket furnace policy.

Furnaces rarely run one part size at a time in production volume. When sections of different cross-section size share a furnace load, the smaller sections should sit toward the outer end of the furnace chamber, not buried in the interior - they reach temperature faster because they have less mass to heat through, and if they are positioned for easy removal, they can be pulled and quenched as soon as they reach temperature without disturbing the larger pieces still heating.

The practical consequence of getting this wrong is straightforward: if small and large parts are mixed without a removal plan, either the small parts overheat while waiting for the large ones, or the large parts get pulled early and quenched under-temperature. Both large and small parts should be timed and removed separately, with the furnace load physically arranged to make that separation possible rather than improvised at the furnace door.

A furnace has a fixed heat output. Load it beyond what that output can sustain, and the furnace temperature dips when the charge is introduced - the furnace has to spend time recovering to setpoint before the parts actually begin heating at the intended rate. An oversized charge does not just take longer to heat; it also means the parts sit at a lower-than-intended temperature for longer, which changes the austenitizing behavior in ways that are difficult to correct after the fact.

The practical rule is to size the charge to the furnace's actual power rating, not its chamber volume - a furnace can often physically fit more material than it can heat at the intended rate. When a heavier charge is unavoidable, the heating time needs to be extended to compensate for the initial temperature dip, and that extension should be built into the process plan rather than discovered by pulling an under-temperature part.

Most steel grades specify a quenching temperature range rather than a single number, and the choice within that range should follow the quenching medium, not be fixed regardless of what the part is being quenched into. Water or brine quenching draws heat out fast enough that the lower end of the temperature range is normally sufficient - going to the top of the range on top of an already-severe quench increases the risk of cracking and distortion without adding meaningful hardness benefit. Oil or molten-salt quenching removes heat more gently, so the upper end of the range is generally used to make sure the slower cooling rate still reaches full hardening depth.

This is a medium-matched decision, and treating it as a single fixed number regardless of what liquid the part goes into is one of the more common sources of inconsistent hardness results between two batches processed on paper to the same specification.

Dual-medium quenching - water or brine first, followed by oil or molten salt - is used specifically to get the fast initial cooling rate needed to avoid a pearlite transformation, while avoiding the cracking risk of carrying that fast cooling rate all the way down through the martensite transformation range. The dwell time in the first medium follows the same temperature-and-section logic used for single-medium quenching; the transfer between media is the part that needs separate control.

That transfer should take roughly 0.5 to 2 seconds - long enough to move the part without dropping it, short enough that the surface does not have time to recover temperature (recalescence) before it enters the second medium. A slow transfer defeats the purpose of the first medium by letting the surface warm back up before the second cooling stage begins; too abrupt a transfer risks mechanical shock to a part that is already partway through a hardening transformation. The timeline below shows where this transfer sits relative to the martensite start (Ms) and finish (Mf) points that the whole sequence is built around.

Heating steel in open air at hardening temperature does two things to the surface simultaneously: oxygen reacts with the surface to form scale, and carbon diffuses out of the surface layer faster than it can be replenished, leaving a decarburized skin with lower carbon content - and therefore lower achievable hardness - than the core. For tube surfaces where this cannot be tolerated, heating should be done in a properly maintained salt bath furnace or a protective-atmosphere furnace, both of which limit the surface's exposure to free oxygen.

Where neither is available, an air-atmosphere resistance furnace can still be used, but only with protective measures in place - a pack coating, an anti-oxidation compound, or a sacrificial wrap - because heating in open air without any protection will decarburize the surface layer to a depth that depends on time and temperature, and that depth is difficult to predict precisely without a specific process trial.

Once a vertically suspended tube enters the quenching medium, it should be moved with small, controlled up-and-down strokes rather than allowed to swing side to side. Swinging motion creates uneven local flow of the quenching medium against different faces of the tube as it moves through the liquid, which cools those faces at different rates - and an uneven cooling rate around the cross-section is a direct contributor to distortion and to residual stress imbalance between the faces.

A controlled vertical stroke, by contrast, keeps the flow of quenching medium relatively even around the section's full perimeter as it works through the liquid, which is the entire point of moving the part at all - agitation is there to strip the vapor blanket that would otherwise form against a stationary surface and slow the cooling rate unevenly, not to move the part around the tank.

Some parts need higher hardness at one location - a wear face, a bearing seat area - than the bulk quench alone reliably delivers, typically because that section's mass or geometry pulls heat away from that specific zone more slowly during a plain immersion quench. The fix is to keep the whole part immersed as normal, but direct an additional spray of quenching liquid specifically at the zone that needs the faster local cooling rate, on top of the full immersion rather than instead of it.

This is a targeted supplement, not a substitute for correct overall quenching - the rest of the part still needs the standard immersion sequence to reach its own required hardness and to avoid a hardness mismatch severe enough to introduce its own stress concentration at the boundary between the sprayed zone and the rest of the section.

Salt bath furnaces heat through electrical resistance between electrodes immersed in the molten salt, and a part positioned too close to an electrode picks up localized overheating from the concentrated current path near that electrode - the heating is not perfectly uniform through the bath volume, and proximity to the electrode is where that non-uniformity is strongest. A working clearance of at least 50mm from the electrodes is the practical minimum to avoid that localized effect.

The same clearance logic applies to two other dimensions inside the furnace: the distance from the furnace wall, and the depth of immersion below the bath's liquid surface - both should also be held to at least 50mm. A part sitting close to the wall or too near the liquid surface is exposed to a different thermal environment than one properly centered in the bath, and that difference shows up as inconsistent hardness or distortion between otherwise identical parts processed in different positions within the same load.

Graded quenching - also called martempering - holds the part at a temperature just above the martensite start point rather than continuing straight through to room temperature, allowing the temperature to equalize through the section before the final transformation happens more uniformly. Raising the austenitizing quenching temperature somewhat for this process increases the stability of the retained austenite, reducing the tendency for that austenite to decompose into pearlite during the isothermal hold stage rather than transforming cleanly to martensite once the part continues cooling.

This is a deliberate departure from the medium-matched temperature selection described earlier - graded quenching is a distinct process with its own temperature logic, driven by austenite stability during the hold rather than by the cooling severity of a single medium.

Structural and plain carbon steels tolerate being charged directly into a furnace already at quenching temperature, or even 20 to 30degC above it, without a preheating stage - their alloy content is low enough that thermal shock and grain growth during rapid heating are not major concerns. High-carbon and high-alloy steels are a different case: charging them directly into a hot furnace risks uneven thermal expansion across the section and excessive grain growth in the higher-alloy matrix. These grades should be preheated to around 600degC first, then raised to the final quenching temperature in a second stage, giving the section time to equalize before it is pushed through the higher-temperature range where the risks are greater.

Treating every steel type with the same direct-charge practice is a common shortcut in smaller operations, and it is specifically the high-carbon and high-alloy grades where that shortcut creates problems that a structural carbon steel would never show.

Where a tube needs a deep hardened case - a thick section carrying wear load through most of its cross-section - a somewhat higher quenching temperature within the grade's allowable range supports a deeper hardening depth. Where only a shallow hardened layer is required, a lower temperature within the range is the better choice, since it limits the hardening effect to nearer the surface and reduces the core embrittlement and distortion risk that comes with austenitizing more of the section than the application actually needs.

This is a third, independent axis for temperature selection alongside the medium-matched logic and the graded-quenching logic described above - case-depth requirement, quenching medium, and process type (single-medium versus graded) can each push the temperature choice in a different direction, and the final selection has to reconcile all three against the part's actual service requirement.

The quenching medium itself needs to be clean and temperature-controlled to behave predictably. Oil, soap residue, or other contamination in the quenching water changes the cooling characteristics unpredictably - a contaminated bath can form an uneven vapor blanket against the part's surface, slowing the cooling rate in a way that is difficult to detect until the hardness result comes back inconsistent. Water temperature should generally stay below 40degC, and oil temperature below 80degC - both media lose cooling severity as they warm up, and a bath that has crept above its intended temperature through repeated use without cooling will quench progressively less aggressively than the process was designed around.

Contact now