Stainless Steel Investment Casting: An Engineer's Reference

1. The process, step by step

Investment casting takes its name from the ceramic shell that is "invested" (coated) around a wax pattern. The sequence below describes the shell process used for stainless steel; block-mould variants (Shaw process, plaster investment) exist but are rare for steel.

1.1 Wax pattern

A metal die, usually aluminium, is machined to the part geometry plus the total shrinkage of wax and steel (typically 1.5 to 2.5 percent combined for austenitic stainless). Wax is injected under modest pressure, cooled and ejected. Soluble wax or ceramic cores form internal passages that cannot be pulled from the die. For prototypes or very low volumes, patterns can be 3D printed in wax or in a castable resin, avoiding the die cost at the expense of a rougher surface and less predictable dimensional behaviour.

1.2 Assembly and shell building

Patterns are attached to a wax runner system to form a tree. The tree is dipped in a slurry of fine refractory (zircon or alumino-silicate flour in a colloidal silica binder), stuccoed with coarser refractory grain, and dried. Six to ten coats build a shell a few millimetres thick. Each coat must dry under controlled humidity and temperature; rushing this stage is a common root cause of shell cracks and inclusions. The fine refractory flour in the primary coat is what gives investment castings their characteristic surface finish and detail.

1.3 Dewax and firing

The wax is removed in a steam autoclave. The shell is then fired, typically to 1000 °C or above, to burn out residual carbon and convert the silica binder to a rigid ceramic. Firing also leaves the shell dry and largely inert, which is why investment moulds can be poured hot and why they tolerate a wide range of mould temperatures.

1.4 Melting and pouring

Stainless steel is melted in a coreless induction furnace, normally in air for the standard corrosion-resistant grades. Chemistry is checked by optical emission spectrometry before pouring. The shell is poured while still hot from the firing furnace, usually at 900 to 1100 °C for stainless, so that thin sections fill before the metal freezes. Vacuum melting and pouring is reserved for reactive or high-integrity alloys and is not standard for the grades discussed here.

1.5 Knockout, cut-off and finishing

After cooling, the shell is broken away by vibration or high-pressure water. Castings are cut from the runner by abrasive disc or plasma, gate stubs are ground flush, and the parts are shot blasted. Ceramic cores are removed mechanically or by leaching in molten or aqueous caustic, a route that is practical for ferrous castings but not for most light alloys. Heat treatment, pickling and passivation, straightening, non-destructive testing and machining follow as specified.

2. Cast stainless grades and how they differ from wrought

The single most common specification error is calling up a wrought grade (304, 316L, 1.4404) on a casting drawing. Cast stainless steels have their own designations, their own chemistry and their own product standards, and a foundry quoting against "316L" has to guess which cast grade you intend.

2.1 Designation systems

ASTM (originally ACI) cast grades use a letter code: C for corrosion-resistant service, a second letter indicating the position of the alloy on the chromium-nickel diagram (A for low nickel through to N for high nickel), and a number giving the maximum carbon content in hundredths of a percent. CF8M is therefore a corrosion-resistant grade in the 19Cr-10Ni region with 0.08 percent maximum carbon and a molybdenum addition. EN 10283 uses the material number and a descriptive name beginning with GX (G for cast, X for high alloy) followed by the carbon in hundredths and the principal alloying elements.

Equivalences in Table 1 are working analogues, not interchangeable specifications. Composition limits, mechanical property minima and heat treatment requirements differ between the ASTM and EN documents even where the alloy is nominally the same. State one standard and one grade on the drawing.

2.2 Why the cast chemistry is different

Cast grades carry more silicon (typically up to 1.5 or 2.0 percent) for fluidity, and their chromium-to-nickel balance is deliberately set so that the "austenitic" grades solidify with some delta ferrite. The industry consensus is that 5 to 25 volume percent ferrite in a cast austenitic stainless steel improves strength, improves weldability and castability by reducing cracking, and improves resistance to stress corrosion and intergranular attack in certain environments. The result is that a CF8M casting is metallurgically a dual-phase material even though it is called austenitic. This matters for magnetic permeability (cast 316 is slightly magnetic), for corrosion in some reducing media, and for cryogenic toughness.

Ferrite content is estimated from composition using the Schoefer diagram, which plots a chromium equivalent against a nickel equivalent; ASTM A800/A800M gives the standardised procedure. It is measured on the part with a magnetic ferrite meter (Feritscope) or by metallographic point count. If your application needs a ferrite range, say so on the drawing, because the foundry otherwise controls only to the chemistry limits of the grade.

2.3 Pitting resistance

For chloride service, grades are compared by pitting resistance equivalent number, PREN = %Cr + 3.3 %Mo + 16 %N. Broadly, Mo-free 23Cr grades sit in the mid twenties, 22Cr duplex grades in the low to mid thirties, 25Cr grades in the high thirties, and the 25Cr-7Ni-4Mo super duplex family above 40. CF8M sits at roughly 24 to 28 and CD3MN at 33 to 38, which is the usual reason a valve body moves from austenitic to duplex.

3. Metallurgy that decides whether the casting is sound

3.1 Solidification mode and hot tearing

Whether a stainless steel freezes first to austenite or to delta ferrite is governed by the ratio of chromium equivalent to nickel equivalent. Suutala's value of Creq/Nieq = 1.55 is the accepted critical ratio for shaped castings; above it, primary ferrite forms. This is not an academic point. Published hot tearing work on stainless steels found that steels solidifying through the ferrite-plus-austenite mode, roughly Creq/Nieq 1.5 to 2.0, were crack-free, whereas any component of primary-austenite or primary-ferrite freezing outside that window led to some cracking. The same work found that phosphorus and sulphur are strongly harmful in fully austenitic solidification but have little effect once some delta ferrite is present.

For the buyer, the practical consequences are: standard CF8M and CF3M chemistries are already balanced to freeze with ferrite and are forgiving; fully austenitic grades such as CN7M (Alloy 20), CK3MCuN (6Mo) and the 20Cr-20Ni-6Mo types are crack-prone and need tighter P and S limits and better gating; and requests to push a cast 316 grade towards zero ferrite (for permeability or cryogenic reasons) should be discussed with the foundry before the drawing is issued.

3.2 Sigma phase and other intermetallics

Chromium-molybdenum rich stainless steels can precipitate sigma phase at grain boundaries between roughly 600 and 950 °C. Sigma is hard and brittle and it removes chromium and molybdenum from the surrounding matrix, so it costs both toughness and corrosion resistance. It is one of the standard problems of cast stainless steels. The defence is a full solution anneal followed by a fast water quench, and avoidance of slow cooling through the sigma range after welding or stress relieving. Duplex and super duplex castings are the most sensitive; heavy sections in these grades need controlled section thickness and a quench that actually reaches the core.

3.3 Oxide films and "loose grain"

Molten high-chromium, high-molybdenum stainless steels carry a solid, dry oxide film on the liquid surface. Turbulent pouring folds this film into the melt as double films (bifilms) that end up at grain boundaries. In super duplex and 20Cr-20Ni-6Mo castings these show on radiographs as diffuse microshrinkage and on penetrant inspection as a red spider's web of cracks, a condition known in some foundries as the loose grain effect. This is the bifilm framework of John Campbell, whose broader criticism is that conventional investment casting trees, with a conical pouring cup and oversized runners built for wax strength rather than for flow control, are among the poorest filling systems in the industry, and that the small ingates typical of the process are what stops matters being worse. It is his particular framework rather than universal foundry practice, but the design principles that follow from it (offset pouring basin, tapered sprue, bottom filling, no free fall of metal) are increasingly applied by better investment foundries and are worth asking about when a part is critical.

3.4 Surface decarburisation and carbon pickup

Because a fired investment shell is porous and inert, atmospheric oxygen persists inside it during pouring. Steel investment castings poured in air therefore show a decarburised surface layer whose depth increases with mould temperature and casting modulus. For low-carbon stainless grades the effect on corrosion resistance is minor, but it does affect surface hardness on martensitic and PH grades and is one reason why a machining allowance is retained on wear faces of CA6NM and CB7Cu-1 parts. The reverse problem, carbon pickup from residual wax or from organic shell binders, is controlled by proper burnout and is checked by product analysis on the casting rather than on the ladle sample.

3.5 Nitrogen

Nitrogen is a deliberate addition in duplex grades (typically 0.1 to 0.3 percent) and a contaminant in austenitics. Above its solubility limit at the freezing point it produces gas porosity, and in 18/8 steels it becomes severe at around 0.3 percent. Melt practice, not the casting process, controls this, and it is why duplex castings are best sourced from foundries that pour these grades routinely.

4. Heat treatment

Heat treatment is part of the material specification, not an optional extra, and the condition must be stated on the drawing or purchase order.

Two failure modes to watch. First, an austenitic casting supplied "as cast" to save cost will fail an intergranular corrosion test (ASTM A262 Practice E, or EN ISO 3651-2, which EN 10213 clause 7.4 and EN 10283 provide for by agreement on austenitic and duplex grades) because of carbide precipitation during cooling in the shell. Second, a heavy duplex section that is solution annealed but cooled in air rather than quenched will contain sigma phase and fail impact tests. If you need evidence, ask for a microstructure report or an A262 test certificate rather than relying on the furnace chart alone.

After heat treatment, castings are pickled to remove scale and then passivated, normally in nitric or citric acid to ASTM A967 or ASTM A380. Passivation does not add a coating; it removes free iron and lets the chromium oxide film re-form. Electropolishing is an option where hygiene or fatigue performance justify it.

5. Design rules for investment cast stainless

5.1 Size and wall thickness

Routine parts run from a few grams to around 30 to 50 kg; specialist foundries go beyond 100 kg but tooling, shell handling and yield all become more difficult. Minimum wall thickness for stainless is about 1.5 mm over short distances (the SFSA Steel Castings Handbook cites 1.5 mm walls tapering to under 1 mm as common in steel investment castings) and 2.5 to 3 mm as a general design value. Thin walls fill better when they are fed from a thicker adjacent section rather than from an isolated ingate. Keep wall thickness as uniform as practical; where a change is unavoidable, blend with a taper or a generous radius so that the thicker section can feed the thinner one as it freezes.

5.2 Draft, radii and undercuts

Draft is not needed on the casting itself, because the wax pattern shrinks away from the die and the shell is destroyed. A nominal 0.5° on deep internal features helps wax ejection and die life. Internal and external radii of 1 to 2 mm are preferred; sharp internal corners concentrate stress in the wax, the shell and the casting. Undercuts are possible using soluble wax inserts, ceramic cores or collapsible die sections, each at a cost, so a design review with the foundry before the die is cut pays for itself.

5.3 Holes and cored passages

Blind holes are practical to a depth of about one to one and a half times the diameter; through holes to about three times the diameter for small bores. Long, thin ceramic cores deflect during pouring and are the main source of wall thickness variation in valve bodies and manifolds. Holes below about 2 mm diameter are usually drilled rather than cast.

5.4 Tolerances

Dimensional tolerances for investment castings are specified in ISO 8062-3 (dimensional casting tolerance grades, DCTG). For investment castings ISO 8062-3:2023 Table A.1 gives DCTG 4 to 6 for parts up to 100 mm overall, DCTG 4 to 8 from 100 to 400 mm and DCTG 4 to 9 above that, on the basis of long series production with the tooling fully developed. In practice that means a total tolerance of 0.3 to 0.6 mm (about ±0.15 to ±0.3 mm) on a 25 mm feature and 0.4 to 0.8 mm on a 100 mm feature (Table 7). Tighter figures, down to about ±0.1 mm on small features, are held by good foundries but should be agreed rather than assumed. Wall thickness is toleranced one grade coarser than the general DCTG unless stated otherwise (clause 9). Dimensions across the die parting line, long thin features prone to wax distortion, and features formed by cores are looser. On larger parts, distortion of the wax pattern before investment is the main limit on accuracy and straightening jigs for the wax are sometimes needed. Geometric tolerances (flatness, concentricity) should be quoted separately, and any feature tighter than the process capability should be shown as a machined feature with a machining allowance of 0.5 to 1.5 mm depending on size (ISO 8062-3 gives RMAG E for investment castings, about 0.4 mm below 40 mm rising to 1.4 mm at 250 mm).

5.5 Surface finish

As-cast surfaces are typically Ra 3.2 µm, with Ra 1.6 µm achievable on small, well-controlled parts. Gate witness areas will be ground and are visible; identify on the drawing any faces where a gate stub is not acceptable. If a feature is cosmetically critical, say so, because the foundry chooses gate positions for soundness first.

5.6 Marking

Cast-in identification (grade, heat number cavity, logo, part number) is nearly free if it is in the wax die and forms part of traceability under EN 10204 and PED. Raised characters cast better than recessed ones.

6. Defects, inspection and acceptance

6.1 Defects typical of the process

Requirements and the defects that matter differ between processes, so the notes below apply to shell investment casting in air; sand and die casting have different populations. Ceramic inclusions arise from primary-coat spalling or shell cracks during dewax and appear as surface pits or subsurface non-metallics on radiographs. Positive metal (fins, "flash") follows shell cracks and is ground off. Misruns and cold shuts occur in thin sections poured too cold or with too little metallostatic head. Shrinkage porosity concentrates at isolated heavy sections that the runner cannot feed; the fix is design or gating, not process parameters. Hot tears follow the solidification mode discussed in section 3.1 and appear at constrained junctions and at the gate. Gas porosity indicates melt practice problems, especially nitrogen in austenitics. Surface decarburisation and carbon pickup were covered in section 3.4. Distortion, particularly of flat plates and thin rings, arises from wax handling, shell restraint during cooling and quench distortion after solution annealing; straightening is normal practice and should be allowed for in the specification.

6.2 Chemical and mechanical testing

Product standards require a ladle analysis per heat and, for pressure-containing grades, mechanical tests on separately cast test bars heat treated with the castings (ASTM A351, EN 10213). Test blocks, cast-on or separately cast to EN 1559-2, represent the melt and the heat treatment; EN 10213 clause 7.2.2.3 states that the room temperature yield and tensile values then also apply to the casting up to the maximum wall thickness given in its Table 3, but this is a specification convention, not a measurement of the thick section. Where the section matters, for example duplex over 25 mm, specify test coupons attached to or cut from a casting. Hardness on the casting is a cheap check on heat treatment for martensitic and PH grades. Positive material identification by handheld XRF is a standard incoming check and catches grade mix-ups, though it does not read carbon or nitrogen.

6.3 Non-destructive testing

Visual inspection is normally to ISO 11971 or MSS SP-55. Liquid penetrant (EN 1371-2, which is the part written for investment castings since EN 1371-1 excludes them; ISO 4987; ASTM E165 with acceptance to ASTM A903 or a named severity level) is the workhorse for stainless because the material is non-magnetic or only weakly magnetic; magnetic particle testing (EN 1369) applies to martensitic grades. Radiography is specified to EN 12681 or ISO 4993 and judged against reference radiographs: ASTM E192 is written specifically for investment castings, with ASTM E446 and E186 for heavier sections, and ISO 4993 itself points to the ASTM reference radiographs for acceptance. Ultrasonic testing is rarely applied to investment castings because of their size and geometry.

The critical part of any NDT specification is the acceptance level, not the test method. EN 10213 is explicit about this: clause 7.3.3.5 requires every order to state the NDT method, the severity level for each method, the areas of the casting to be tested and the percentage of castings inspected, and allows different criteria for different zones. The same logic applies whichever acceptance document you use, whether the EN severity levels, ASTM A903 or the ISO reference radiograph grades: classify the casting by the consequence of its failure, which sets how many castings are examined, then zone the drawing and state a quality level for each zone. A drawing that is silent on both leaves the foundry to choose, and it will choose the level that suits its scrap rate. Stated up front, the foundry can price the inspection and the yield that goes with it.

6.4 Pressure equipment and certification

Castings for pressure equipment sold in the UK or EU fall under the Pressure Equipment (Safety) Regulations 2016 or PED 2014/68/EU. Material must be to a harmonised standard (EN 10213 for cast steel), to a European Approval for Materials, or by particular material appraisal. Certification is to EN 10204: type 3.1 (foundry's own inspection department) is normal, and under PED Annex I section 4.3 a 3.1 certificate is only acceptable for pressure-retaining material if the foundry's quality assurance system has been certified by a competent body established in the EU, so ask to see that certificate and check its scope covers the grade, product form and size range. Type 3.2 (countersigned by an independent inspector or the purchaser's representative) is required for some category III and IV equipment and by many end users in oil and gas. Ask for the certificate format at RFQ stage; adding 3.2 witness testing after the castings are poured is expensive and sometimes impossible.

7. Choosing investment casting, and specifying it

Investment casting competes with machining from bar, with sand casting, and increasingly with metal additive manufacturing. It wins on cost when the part has internal passages, thin walls or complex external shape that would generate a lot of swarf, and when the quantity justifies a wax die (from a few hundred pieces per year for a large part to a few thousand for small ones). Sand casting wins above roughly 50 kg, or where the surface finish and tolerance of investment casting are not needed. Additive manufacturing wins below about 50 pieces or where the geometry cannot be cored.

A complete specification for a stainless investment casting contains: the cast grade and product standard (for example CF3M to ASTM A351, or 1.4409 to EN 10213); the heat treatment condition; the ferrite range if it matters; tolerance grade to ISO 8062-3 and any machined features with allowance; surface finish and gate-free faces; NDT method, extent and acceptance level by zone; test bar requirements; certification type to EN 10204; passivation standard; and marking. Our earlier article on what to include in a casting RFQ covers the commercial side.

Frequently asked questions

Is cast 316 the same as wrought 316?

No. The cast equivalent is CF8M (ASTM A351/A743) or GX5CrNiMo19-11-2, 1.4408 (EN 10283). Cast grades carry more silicon and are balanced to solidify with 5 to 25 percent ferrite, which improves castability and resistance to hot tearing. Corrosion performance is similar, but the grade name, chemistry and test standards differ, so specify the cast grade.

What tolerances can stainless steel investment casting hold?

ISO 8062-3:2023 gives DCTG 4 to 6 for investment castings up to 100 mm overall (4 to 8 up to 400 mm), which means about ±0.15 to ±0.3 mm on a 25 mm feature. Tighter tolerances can be agreed with the foundry. Features across the die parting line or affected by wax distortion are looser, and critical features are normally machined.

What surface finish does an investment cast stainless part have?

As-cast surfaces are typically Ra 3.2 µm, with Ra 1.6 µm achievable on well-controlled small parts after blasting.

Do stainless steel investment castings need heat treatment?

Austenitic grades should be solution annealed above about 1040 °C and water quenched. CA6NM is normalised and tempered. CB7Cu-1 is solution treated and aged. Duplex grades are solution annealed and quenched to avoid sigma phase. Specify the condition on the drawing.

Which standards cover stainless steel investment castings?

Materials: ASTM A351, A743, A744, A747, A890, A995, A487; EN 10283 and EN 10213; ISO 11972. Tolerances: ISO 8062-3. Testing: EN 12681, ISO 4993 or ASTM E192 (radiography), EN 1371-2, ISO 4987 or ASTM E165 (penetrant), ASTM A800 (ferrite), EN ISO 3651-2 or ASTM A262 (intergranular corrosion). Certification: EN 10204.

Sources and further reading

Sources: Campbell, J., Complete Casting Handbook, 2nd ed., Butterworth-Heinemann, 2020 (paraphrased, chiefly chapters 4, 6, 8 and 16). Steel Founders’ Society of America, Steel Castings Handbook, 6th ed., 1995, Supplements 2, 3 and 8. BS EN 10283:2019, BS EN 10213:2007+A1:2016, BS EN ISO 8062-3:2023, BS ISO 4987:2020, BS ISO 4993:2024. Standard numbers are for identification; always work from the current edition.

Related pages and guides