PWHT of Thick Ferrous Valve Castings

Steel crystalizes when it solidifies

Because steel crystal structures change, it’s important to know the upper and lower critical temperatures to fully understand post weld heat treat (PWHT).

Just like water, steel can be melted and frozen repeatedly. Steel can be heated to form a face-centered cubic crystal structure and then cooled to reform a body-centered cubic structure. The ability to do this is called polymorphism. Iron-based steel alloys can be cycled up and down in temperature to grow and regrow crystals of face-centered cubic or austenite crystals (hotter than the upper critical temperature) and body-centered cubic or ferrite crystals (below the lower critical temperature).

Heat-treat grows new crystals to make steel strong and tough

Carbon is a small atom compared to iron (Fe). As a result, carbon is the most influential element on iron. Because the carbon atom is small, it can fit in between the iron atoms in the face-centered cubic FCC structure called austenite. Two percent carbon can fit into the austenite!

Table 1 shows the three solid phases of iron (Fe), their various names and temperatures where they exist.

Cast steel typically has a carbon content of 0.25%. At 1700˚F, when the steel’s crystal structure is FCC, the 0.25% carbon fits nicely into the crystal structure. Upon cooling to room temperature, the FCC crystals are consumed by newly grown BCC crystals that only hold 0.007% carbon.

When ferrite crystals grow from austenite, any excess carbon (>0.007%) will form iron carbides. The resulting layers of ferrite and iron carbide create a striped (lamellar) structure. This striped structure is named pearlite. It is a tiny composite material of soft ferrite and hard iron carbide. The microstructure of carbon steel is a mixture of ferrite and pearlite.

For example, a carbon steel valve casting typically has a 0.25% carbon (see Figure 2). Steel can be heated to the lower critical, 1333°F (723˚C). At this temperature, the ferrite (BCC) begins to grow austenite (FCC) crystals. FCC crystals grow off of and consume the original BCC crystal. The FCC consumes the BCC as the temperature rises. Once the temperature exceeds the upper critical (this temperature varies with carbon content—see Figure 3), all of the crystals—even the iron carbide dissolves—are incorporated in FCC crystals. This is an austenitize cycle.

Once the steel is austenite, it can be cooled at various rates to make the steel hard (fast cooling) or soft (air cooling). Alloy steels can be cooled fast enough to form a metastable crystal called martensite. This crystal is a body-centered tetragonal, BCT, which is very hard. The hard metal can then be tempered (heated to an intermediate temperature below the lower critical temperature) to soften the martensite. The hotter the temper cycle, the softer the metal.

Low-temperature PWHT cycles will not form austenite. They are essentially temper cycles. They soften heat-affected zones (HAZ) and welds that may have cooled fast enough to form martensite. Since fast cooling indicates that stress was developed during cooling, a PWHT cycle is often called a stress relief. A PWHT is a cycle performed after welding to soften the HAZ.

If during the PWHT the temperature were to exceed the lower critical and create even a few austenite crystals, the heat treatment of the steel would be compromised. The steel must then be fully re-austenitized and tempered to bring back the mechanical properties. If weld metal is in the casting, a procedure qualification record (PQR) will need to show the weld metal can endure and flourish after being re-austenitized and tempered.

Weld consumables are designed for many situations. Most are designed to form mechanical properties upon solidification and/or after PWHT. A few consumables are non-heat treatable. This means the weldment loses its strength if a phase change occurs. Avoid going over the A1 temperature for these weld consumables.

PWHT methods

PWHT can be low-temperature cycles that occur below the lower critical or alternately full austenitize and temper dependent upon specification. Table 2 is a guide to PWHT for different steel grades.

Monitoring oven temperature

Some foundries use gas-fired furnaces, and refer to them as ovens to keep them from being confused with melting furnaces. Heat-treating is like baking because the product is more useful afterward. As in baking, temperature is important. When heat-treating castings, we need to know the temperature of the metal being heat-treated, not just the oven’s air temperature.

A typical heat-treat oven is equipped with two thermocouples that protrude from the oven wall. They register the air temperature, control the cycle heating and act as a safety device that will shut down the oven if the temperature exceeds a predetermined level. Neither thermocouple is monitoring the temperature of the casting.

A casting’s temperature is monitored by attached thermocouple(s) (TC) on its surface. Typically, a TC is attached to the thickest section. Depending on the size and shape of the casting more TC maybe added.

When heating thick sections, there is a lag time between the casting’s surface and the center of the section. Steel conducts heat quite well. The lag time was defined in an experiment in which thermocouples were attached to the surface and center of a 16-inch cube of carbon steel. The cube weighed about 1200 lbs. Temperature readings were taken every second as the cube was heated from ambient to 1700˚F. The lag in Figure 4 between the surface and centerline is about 40 minutes. This experiment was repeated four times in different ovens, but still using the 16-inch cube, with similar results.

The oven thermocouples that monitor the air temperature reached 1750˚F in under an hour while it took the surface of the 16-inch cube nearly 4 hours with the center following 40 minutes behind.

The most significant lag is between the casting and air during heating. The casting heated at a rate of about 350˚F/hour. By imposing a ramp-up rate of 350˚F/hour on heating cycles, the whole casting (including the center) is assured to reach temperature.

Time at temperature

A heat-treating rule of thumb is to run a cycle for 1 hour per inch of maximum metal thickness. If the 16-inch cube was run per this rule, using the oven’s TCs, the casting would have been at temperature for 12 hours (16 minus 4). This is more than enough time to austenitize the steel.

If the steel is only 2 inches thick, the temperature would be held two hours, and the rule would have failed to allow the casting to get to temperature. Therefore, to assure the casting is at temperature at least an hour, the oven TC needs to be at temperature at least 4 hours minimum.

When performing a low-temperature PWHT cycle, many standards require the part be PWHT based on weld thickness. For thinner sections (<4 inches) this rule may suffice. However, a casting having a 10-inch thick section and a ½ inch weld will need more than a half hour at temperature (per oven TC) for the casting (and weld) to reach temperature. As shown in Figure 4, a thick casting acts like a heat sink. PWHT cycles need to run a minimum of 4 hours, unless attached TC is used. Another way to assure that part gets to temperature is to impose a ramp-up rate. Using both a 4-hour minimum and 350˚F/hour ramp-up rate is even better when TC are not used.

PWHT cycles are often run using attached TC. A standard such as MIL-STD-278F requires steel to be heated and cooled at slow rates, for example, 100˚F/hour. The original purpose of these slow ramp-up and down cycles may be lost in time. If it was to assure that the steel reaches temperature, then attaching thermocouples would have been a faster way to perform the cycle. These slow ramp-up and down PWHT cycles can last for over a day for some castings. If the purpose of the slow rates is to reduce stress, then it misses the mark. Temperature is what lowers stress/hardness and the standard does not address a measurement for stress. Manufacturers are running these cycles per MIL-STD-278F, and yet the cycle’s purpose is unclear.

These types of standards need revision to consider clean steel and modern welding techniques. Castings are isotropic (crystals are equiaxed) while wrought products are anisotropic (crystals are stretched in one or more directions). This author believes castings do not have the same innate stresses as wrought products. Castings should not be held to the same stress relief alleviation as wrought products.

Ramp rates might create temper embrittlement

Most specifications that require slow ramp-up and down rate will have the rate monitored between 800˚F and the PWHT temperature. A few lower that origin temperature to 600˚F and at least one is 300˚F. The Blue Brittle zone, where temper embrittlement occurs, is roughly between 700˚F to 900˚F or 300˚F to 700˚F (sources disagree). By slowing the heating/cooling rate during this temperature range, temper embrittlement is allowed to occur in the steel. For alloys made specifically for improved toughness, these cycles are doing harm. Company- specific stress relieve cycles with slow ramp rates exist. Please review your specifications to avoid time spent in the temper embrittlement zone.

Here are some examples of specifications that require slow ramp rates: Mil-STD-278F; USC-56 (Section VIII of ASME Division I); and MIL-S-16993A CL.1

Heat treating for valves

Just as the material of a valve casting is determined based on its application, heat treating is also performed to enhance the material for its use. By carefully controlling the PWHT process, the casting can be welded to upgrade levels determined by the customer and specification.

Some popular valve alloys, such as C12A (P91), require very specific PWHT cycles. If done incorrectly, the casting is ruined and a WC9 (P22) casting would have been sufficient at a much lower cost. When C12A is specified, the foundry must be knowledgeable and experienced in PWHT of the alloy or risk scraping the casting.

Additionally, valve castings in duplex stainless steel or super austenitic (6% Molybdenum) also have very specific PWHT cycles of which the foundry should have extensive knowledge. All too often valves in these grades have the incorrect chemical analysis and/or PWHT procedures that cause the valve to receive a bad reputation, when in reality the problem is the casting.

Elaine Thomas is director of metallurgy at Bradken Tacoma. Reach her at Ethomas@bradken.com or call 253.279.2039.