Why the Manufacturing Route Sets the Quality Ceiling of A335 Pipe
ASTM A335/A335M covers seamless ferritic alloy-steel pipe for high-temperature service, and the standard can only fix what is measurable on a finished piece: product chemistry, tensile properties, hardness, heat-treatment condition and the results of specified tests. What no standard can fully fix is the route taken to reach those numbers. Two mills may ship P22 pipe that both satisfy the same chemistry table and the same 415 MPa minimum tensile strength, yet behave very differently after eight years at 540 C and 10 MPa, because one billet was pierced within a narrow temperature window and the other was rolled from an overheated charge held too long in the reheating furnace. A purchaser therefore has to read a mill test certificate as the output of a process rather than as an isolated set of numbers, and has to know which process steps leave fingerprints that can still be detected at final inspection.
This article follows the chain from steelmaking to final marking for chrome-molybdenum grades P5, P9, P11 and P22 and for the modified 9Cr grades P91 and P92, and explains where quality is created, where it is lost, and what evidence a buyer can ask for at each stage.
Steelmaking and Chemistry Windows That Limit Everything Downstream
For A335 grades the chemistry window is narrow and the consequences of drifting inside it are not symmetrical. In P22, carbon is specified between 0.05 and 0.15 percent, chromium between 1.90 and 2.60 percent and molybdenum between 0.87 and 1.13 percent. Running carbon at the top of the range buys tensile strength but reduces weldability and toughness in the heat-affected zone; running it at the bottom forces the mill to rely on heat treatment rather than composition to reach the 415 MPa minimum tensile strength, which raises the risk of an under-tempered product. Chromium below 1.90 percent erodes steam-side oxidation resistance, while chromium well above 2.60 percent promotes delta ferrite and reduces creep ductility. The carbon equivalent, calculated from carbon, manganese, chromium, molybdenum, vanadium, nickel and copper, is a practical purchasing screen: a low value indicates a clean, weldable heat, a high value warns of hardenability that will need careful tempering.
For P91 the window is tighter still. Carbon is held at 0.08 to 0.12 percent, chromium at 8.00 to 9.50 percent, molybdenum at 0.85 to 1.05 percent, vanadium at 0.18 to 0.25 percent, niobium at 0.06 to 0.10 percent and nitrogen at 0.030 to 0.070 percent. Nitrogen is not a residual here; it is an alloying element that forms fine vanadium and niobium carbonitrides during tempering and provides the dislocation pinning that makes the grade usable at 600 C. A heat with nitrogen at 0.025 percent is chemically out of specification, and a heat with nitrogen at 0.075 percent will show excessive hardness and poor toughness after the same tempering cycle. Phosphorus and sulfur are normally held at 0.020 percent maximum or lower, and sulfur is often reduced to 0.010 percent or below with calcium treatment so that inclusion shape control supports ultrasonic inspection.
Secondary metallurgy, vacuum degassing and calcium-silicon treatment decide inclusion population and hydrogen content. Hydrogen above roughly 2 ppm in the as-cast condition creates a real risk of internal hairline cracks during the first cooling after rolling, particularly in the 9Cr grades, so degassing and controlled slow cooling of the billet are process steps with a measurable metallurgical purpose rather than housekeeping.
Billet Reheating, Piercing and Hot Rolling
The rotary piercing process takes a solid round billet, reheats it to a temperature at which the alloy has adequate ductility, and pierces it over a plug to create a hollow shell that is then elongated on a mandrel mill, a plug mill or an assel mill and finally sized. For 2.25Cr and 5Cr grades the billet is typically discharged from the reheating furnace in the range of 1200 to 1280 C and pierced around 1150 to 1200 C, and the finishing pass is kept above the recrystallisation temperature so that the worked structure is uniform. Three parameters dominate the outcome: the temperature at the piercer, the temperature at the finishing stand, and the total reduction ratio from billet to finished wall.
If the piercing temperature is too high, the steel develops an overheated austenitic grain structure that cannot be fully refined by normalising, and the finished pipe shows coarse grains and poor upper-shelf toughness. If the finishing temperature falls too low, the deformation is not recrystallised, and the pipe carries a banded, partially worked structure with directional properties that show up later as poor flattening behaviour and inconsistent hardness around the circumference. Reduction ratio matters because the centreline of a pierced shell inherits casting segregation from the billet; a wall reduction below about 4:1 leaves the segregation band close to the bore, where it appears as centreline segregation on a transverse macro section and as bands of elevated hardness on a hardness traverse.
Wall thickness control is a rolling problem rather than a heat-treatment problem. Eccentricity created at the piercer, mandrel wear and stand misalignment produce a wall variation that heat treatment cannot correct; because A335 pipe normally carries a minus tolerance on wall rather than a plus/minus band, an eccentric pipe forces the mill either to roll a heavier nominal wall or to risk under-thickness at the thin side. A buyer who measures wall at eight points per cross-section and calculates the eccentricity ratio learns more about the rolling line than from any single mid-wall reading.
Process Step, Controlling Parameter and Inspection Evidence
| Process stage | Controlling parameter | What it sets | Typical verification |
|---|---|---|---|
| Steelmaking and refining | Carbon, Cr, Mo, V, Nb, N windows; P and S maxima; hydrogen below 2 ppm | Hardenability, creep strength, inclusion cleanliness, cracking risk | Heat analysis per ASTM A751, product analysis, sulfur print |
| Billet reheating | Discharge temperature 1200 to 1280 C, soaking time, furnace atmosphere | Piercing ductility, decarburisation depth, grain coarsening risk | Furnace records, decarburisation check on finished wall |
| Piercing and elongation | Piercer temperature, mandrel condition, reduction ratio above 4:1 | Concentricity, centreline soundness, wall uniformity | Wall thickness survey, transverse macro etch |
| Normalising | 900 to 960 C for 1 to 2.25Cr grades; 1040 to 1080 C for P91 and P92 | Grain size, dissolution of carbides, homogenisation | Furnace chart, grain size to ASTM E112, hardness |
| Tempering | 650 to 700 C for 1 to 2.25Cr; 730 to 780 C for P91 and P92 | Toughness, hardness, residual stress relief | Furnace chart, hardness to ASTM E18, tensile test |
| Sizing and straightening | Diameter and ovality tolerance, straightness, cold work level | Dimensional fit-up, residual stress, yield strength | Laser gauge records, straightness check |
| Final inspection | UT coverage, hydrostatic or electromagnetic test, marking | Detection of longitudinal and transverse defects | ASTM E213, ISO 10893 series, ASTM E570 |
Heat Treatment: Normalising, Tempering and Cooling Rate
A335 pipe is supplied in the normalised and tempered condition for the 1Cr to 2.25Cr grades and for P91 and P92, and the heat-treatment cycle is where most certificates are won or lost. Normalising dissolves the carbides formed during rolling, removes the banding inherited from segregation and produces a fresh austenite grain structure that transforms to a controlled ferrite and pearlite or bainite on air cooling. Tempering then precipitates the alloy carbides that provide creep strength while restoring toughness. The two steps must be treated as a pair: a good normalise followed by an under-temper leaves a hard, brittle pipe, while an over-temper sacrifices the creep rupture strength that the grade was selected for.
Cooling rate after normalising is the parameter most often ignored in purchasing documents. Heavy wall pipe cools more slowly at mid-wall than at the surface, and for the 9Cr grades this difference can produce a surface layer of martensite over a softer, partially transformed core. That condition passes a surface hardness check and fails a cross-sectional hardness traverse. Purchasers of P91 and P92 heavy wall pipe should therefore require hardness measured at mid-wall as well as at the surface, and should require that the normalising furnace charge be limited by wall thickness rather than by weight, because a mixed charge of thin and heavy pipe forces the furnace operator to choose one holding time for two different thermal masses.
Furnace calibration, uniformity surveys, and the recording of charge thermocouples are the documentary evidence for this step. A certificate that states normalised and tempered without stating the temperature range, holding time and the fact that the record is traceable to the heat number gives the buyer almost nothing.
Microstructure, Grain Size and Hardness Verification
Because the process determines microstructure, and microstructure determines service behaviour, the most informative acceptance tests on A335 pipe are metallographic and mechanical rather than dimensional. Grain size is measured to ASTM E112; fine and uniform grain is what converts strength into toughness, and a coarse band indicates either an overheated billet or an incomplete normalise. For the 9Cr grades the microstructure should be fully tempered martensite with no delta ferrite stringers and no untempered martensite; delta ferrite above about one percent is a warning sign of an imbalanced chemistry or a normalising temperature above the recommended range.
Hardness is the cheapest and most sensitive proxy for tempering. A335 P91 and P92 pipe is normally limited to 250 HBW maximum, and a reading above that limit almost always means the tempering temperature was low, the holding time was short, or the pipe entered the tempering furnace before it had cooled below the martensite finish temperature. Tensile testing to ASTM A370 confirms the specified minimums, which are 415 MPa tensile with 205 MPa yield for P11, P22, P5 and P9, and 585 MPa tensile with 415 MPa yield for P91, rising to 620 MPa and 440 MPa for P92. Yield strength well above the minimum is not automatically good news in a creep-resistant grade: it often points to under-tempering and predicts poor stress-rupture ductility.
Hardness and tensile data should be read together across the heat. A heat showing a narrow hardness spread and tensile results clustered just above the minimum indicates a stable, repeatable heat-treatment cycle; a heat with a wide scatter indicates inconsistent furnace loading, and that heat is the one to inspect more heavily.
Cold Drawing, Sizing and Finishing Operations
Smaller sizes and tighter tolerances in A335 are frequently produced by cold drawing from a hot-rolled hollow, followed by a final normalise and temper. Cold drawing improves dimensional accuracy and surface finish but leaves 30 to 60 percent of induced cold work in the wall, which raises yield strength and destroys the creep ductility that the grade was purchased for. The final heat treatment after drawing is therefore not optional: a cold-drawn P22 tube that was only stress relieved will show a yield strength far above specification and a hardness that no longer reflects the intended microstructure. Purchasers of cold-drawn or cold-finished A335 should require that the certificate states normalised and tempered after the last cold reduction, not merely stress relieved.
Straightening, end facing and bevelling introduce local cold work and residual stress at the pipe ends. For thick-wall chrome-molybdenum pipe this is the origin of many end-area hardness excursions, and for that reason hardness testing is best performed away from the extreme ends and after any end re-work rather than on a raw as-straightened surface.
Process Defects and Their Inspection Signatures
Each stage of the route leaves a characteristic defect that a specific test will find. Piercing and rolling defects are longitudinal: internal scabs, laps, seams and, in severe cases, internal fissures from centreline segregation, all of which appear as longitudinal indications on ultrasonic inspection to ASTM E213 and as linear indications on a transverse macro or a penetrant check of the bore. Heat-treatment defects are volumetric and metallurgical rather than geometric: coarse grains, decarburisation, untempered martensite and hardness gradients, which are found by metallography, hardness traverses and tensile tests rather than by ultrasonic testing.
Hydrostatic testing confirms pressure integrity but will not reveal a coarse grain structure or an under-tempered wall, and ultrasonic testing will not reveal a chemistry error. This is why A335 acceptance combines a mechanical test, a hardness test, a nondestructive test and a document review rather than relying on any single gate. When a defect appears, the location of the defect usually identifies the responsible process step, and that is the practical link between an inspection finding and a corrective action at the mill.
What a Purchaser Should Verify on the Certificate
A useful certificate review follows the process sequence. Confirm that the heat analysis matches the grade chemistry for every specified element and not only for the alloying elements, that the product analysis is reported to ASTM A751 and remains within the permitted product tolerances, that the heat treatment is stated as normalised and tempered with temperatures and that the pipe was not subsequently welded or hot formed without re-treatment, and that the mechanical results are traceable to the specific heat and lot rather than to a generic mill average. Confirm the nondestructive test method, the acceptance level and the reference standard used, and confirm that dimensions are reported as measured values against the applicable tolerance rather than as nominal figures.
Marking is the final process output. Grade, heat number, size, wall, specification and the method of manufacture should be legible and should match the documentation, because a mismatch between the stencil and the certificate indicates a handling error somewhere in the line that is usually more serious than the marking itself. Where a project requires impact testing, additional hardness limits or a specific grain size, those requirements must be written into the order and confirmed on the certificate, because A335 default testing will not include them.
FAQ
Q: Why is A335 pipe always supplied normalised and tempered rather than as-rolled?
The as-rolled structure from a piercer and mandrel mill is banded and directionally worked, with segregation inherited from the billet centreline. Normalising dissolves the carbides and reforms the grain structure, and tempering precipitates the alloy carbides that provide creep strength. As-rolled A335 pipe would show wide hardness scatter and poor creep ductility, which is why the standard specifies the heat-treated condition.
Q: Which element has the largest effect on the high-temperature strength of P22 pipe?
Molybdenum, specified at 0.87 to 1.13 percent, provides the solid-solution strengthening and carbide stability that allow P22 to serve at elevated temperature, while chromium at 1.90 to 2.60 percent provides oxidation and corrosion resistance. Carbon controls carbide volume and therefore strength, but excessive carbon costs toughness and weldability, which is why the range is limited to 0.05 to 0.15 percent.
Q: Why is mid-wall hardness required for heavy wall P91 pipe?
The hardenability of the 9Cr composition means that surface cooling is fast enough to form martensite while the mid-wall cools more slowly, and if the tempering cycle is sized for the surface rather than the wall, the core remains under-tempered. Surface hardness then looks acceptable while the mid-wall exceeds the 250 HBW limit. A cross-sectional traverse exposes the condition.
Q: Does a high yield strength result prove better quality?
No. In creep-resisting grades, a yield strength significantly above the specified minimum usually indicates under-tempering, which reduces stress-rupture ductility and increases the risk of brittle behaviour in service. Conformance to the minimum with a narrow scatter across the heat is a better indicator of a controlled process than an unusually high single result.
Q: Can ultrasonic testing replace the mechanical tests?
No. Ultrasonic testing to ASTM E213 detects internal and surface-breaking geometric defects such as seams, laps and fissures, but it cannot detect a chemistry error, a coarse grain structure, decarburisation or untempered martensite. Mechanical and metallographic testing address the metallurgical condition, and both groups of tests are needed for a complete acceptance.
Q: What changes if the pipe is cold drawn instead of hot rolled?
Cold drawing improves dimensional accuracy and surface finish but introduces severe cold work, so the pipe must be normalised and tempered after the final cold reduction. Stress relieving alone leaves yield strength above specification and does not restore the tempered microstructure, so the certificate should state the full normalise and temper cycle following the last cold pass.





