Full chemical composition of ASTM A516 Grade 70, including carbon limits by thickness, heat and product analysis ranges, and the role of each element.

ASTM A516 Grade 70 is a plain carbon steel. It contains no chromium, nickel or molybdenum, and only five elements are controlled by the specification: carbon, manganese, phosphorus, sulphur and silicon.
What makes the composition interesting is how tightly those five are held. Carbon changes with plate thickness. Manganese and silicon are given as ranges with a minimum rather than a simple ceiling. And the standard allows manganese to rise if carbon comes down, which catches out buyers reviewing certificates.
This blog gives the full composition limits from the current specification, explains why carbon varies with thickness, covers the difference between heat and product analysis, describes what each element does in the steel, and sets out what to check on a mill test certificate.
These are the limits from ASTM A516/A516M, the current edition being the 2025 revision.
| Element | Heat analysis | Product analysis |
|---|---|---|
| Carbon | 0.27 to 0.31% max, by thickness | Same as heat analysis |
| Manganese, 12.5 mm and under | 0.85–1.20% | 0.79–1.30% |
| Manganese, over 12.5 mm | 0.85–1.20% | 0.79–1.30% |
| Phosphorus | 0.025% max | 0.025% max |
| Sulphur | 0.025% max | 0.025% max |
| Silicon | 0.15–0.40% | 0.13–0.45% |
Two things to note straight away. Manganese and silicon are given as ranges with a minimum, not simply as maximum limits. The specification will not allow a plate with too little manganese, because manganese contributes toughness as well as strength. Phosphorus and sulphur limits apply to both heat and product analysis, with no widened tolerance.
Carbon is the only element in Grade 70 whose limit varies with thickness.

| Plate thickness | Carbon, maximum |
|---|---|
| 12.5 mm and under | 0.27% |
| Over 12.5 to 50 mm | 0.28% |
| Over 50 to 100 mm | 0.30% |
| Over 100 to 200 mm | 0.31% |
| Over 200 mm | 0.31% |
The reason is physical. Thicker plate cools more slowly as it comes off the rolling mill, so it develops a coarser structure and less strength from rolling alone. Allowing slightly more carbon compensates, so a 150 mm plate can meet the same 485 MPa tensile minimum as a 10 mm plate.
The trade-off is weldability. More carbon means a higher carbon equivalent, which increases the risk of cracking in the heat-affected zone, so preheat becomes more important on thicker sections. This is why a thick Grade 70 plate needs more care during fabrication than a thin one even though both meet the same specification.
This provision causes more unnecessary certificate queries than any other, so it is worth understanding.
The standard permits manganese to exceed 1.20% where carbon is reduced below its maximum. For every 0.01 percentage point that carbon falls, manganese may rise by 0.06 percentage points, up to a ceiling of 1.50% on heat analysis and 1.60% on product analysis.
So a certificate showing 1.42% manganese is not automatically wrong. If the carbon figure is correspondingly low, the plate conforms.
The metallurgy behind it is sound. Both elements add strength, but carbon does more damage to weldability and toughness than manganese does. Allowing a mill to substitute manganese for carbon gives the same strength in a steel that welds better and stays tougher.
Certificates report one of two analysis types, and confusing them leads to plates being rejected that were never non-conforming.
Taken from the molten steel during casting. This is what the mill controls, and it is what appears on most mill test certificates.
Taken from the finished plate. As steel solidifies, elements segregate slightly, so a reading taken from the rolled plate can differ from the ladle. The specification allows a wider tolerance to account for this.

| Element | Heat analysis | Product analysis |
|---|---|---|
| Manganese | 0.85–1.20% | 0.79–1.30% |
| Silicon | 0.15–0.40% | 0.13–0.45% |
| Carbon | By thickness | Same limits |
| Phosphorus | 0.025% max | 0.025% max |
| Sulphur | 0.025% max | 0.025% max |
Product analysis is not carried out as standard. It is a supplementary requirement, listed as S2, and appears on the certificate only if the purchase order asked for it.
Understanding the role of each element makes the limits easier to remember.
Carbon is the main source of strength. It is also the main enemy of weldability and toughness, which is why the specification caps it and allows manganese to be traded against it.
Manganese adds strength and improves toughness. It also combines with sulphur to reduce its harmful effects. Because it does less damage to weldability than carbon, it is the preferred route to strength in a pressure vessel steel, which is why the standard sets a minimum of 0.85% rather than just a ceiling.
Silicon is a deoxidiser. Molten steel dissolves oxygen, and if that oxygen stays in, the plate solidifies with gas porosity. Silicon removes it, which is part of what makes the steel fully killed. The specification requires a minimum of 0.15%, so silicon is not optional.
Phosphorus is an impurity from the iron ore and scrap. It increases strength slightly but makes the steel brittle at low temperature, so it is held to 0.025%.
Sulphur is also an impurity. It forms manganese sulphide inclusions that stretch out during rolling into elongated stringers, which reduce ductility through the plate thickness and provide a path for hydrogen cracks. It is held to the same 0.025% limit.
Talk to our team about sizes, stock and mill test certificates.
The composition table alone does not describe what makes A516 a pressure vessel steel. Two manufacturing requirements matter just as much.

Oxygen is removed from the melt before casting, so the plate solidifies without gas voids. Silicon is part of how this is achieved, but the requirement is stated separately in the standard.
Fine grain is the single largest contributor to resistance against brittle fracture, and it is the reason A516 exists as a specification distinct from ordinary carbon steel plate.
A516 covers four grades, and chemistry is what separates them.
Carbon, maximum, by thickness
| Thickness | Grade 55 | Grade 60 | Grade 65 | Grade 70 |
|---|---|---|---|---|
| 12.5 mm and under | 0.18% | 0.21% | 0.24% | 0.27% |
| Over 12.5 to 50 mm | 0.20% | 0.23% | 0.26% | 0.28% |
| Over 50 to 100 mm | 0.22% | 0.25% | 0.28% | 0.30% |
| Over 100 to 200 mm | 0.24% | 0.27% | 0.29% | 0.31% |
| Over 200 mm | 0.26% | 0.27% | 0.29% | 0.31% |
Carbon rises steadily with grade, which is how each grade reaches its higher tensile strength.
Manganese also differs at the thin end. On plate of 12.5 mm and under, Grade 55 allows 0.60 to 0.90% while Grades 65 and 70 require 0.85 to 1.20%. Grade 60 plate in that thickness may use either range. Above 12.5 mm, Grades 60, 65 and 70 all require 0.85 to 1.20%.
Phosphorus, sulphur and silicon limits are identical across all four grades.
Carbon equivalent is not a requirement of A516, but most mills report it and your welding engineer will want it.
It expresses the combined effect of carbon and the other elements on weldability as a single number. The higher it is, the greater the risk of hydrogen-induced cracking in the heat-affected zone, and the more preheat matters.
For A516 Grade 70 it typically falls between 0.43 and 0.45. Because carbon is allowed to rise with thickness, the carbon equivalent tends to climb on heavier plate, which is another reason preheat becomes important above roughly 40 mm.
If the certificate does not report it, it can be calculated from the reported chemistry.
When the chemistry section of a certificate arrives, work through it in this order.

Carbon 0.27 to 0.31% maximum depending on thickness, manganese 0.85 to 1.20%, phosphorus 0.025% maximum, sulphur 0.025% maximum, and silicon 0.15 to 0.40%, all on heat analysis. Manganese and silicon have slightly wider ranges on product analysis, and manganese may go higher where carbon is reduced.
The maximum depends on plate thickness. It is 0.27% for plate 12.5 mm and under, 0.28% up to 50 mm, 0.30% up to 100 mm, and 0.31% above that. Thicker plate is allowed more carbon because it cools more slowly during rolling and needs the extra carbon to reach the same strength.
No, it is a plain carbon steel. It contains no deliberate additions of chromium, nickel, molybdenum or vanadium. Its performance comes from tightly controlled carbon and manganese, low impurities, and the requirement for killed steel with a fine grain structure.
Almost certainly because of the trade-off provision in the standard, which permits manganese to rise as carbon falls, up to 1.50% on heat analysis and 1.60% on product analysis. Check the carbon figure on the same certificate before raising a query. It may also be a product analysis, which allows up to 1.30% in any case.
Heat analysis is taken from the molten steel during casting and is what the mill controls to. Product analysis is taken from the finished plate, where slight segregation during solidification shifts the readings, so the specification allows a wider tolerance. Product analysis is a supplementary requirement and is carried out only when ordered.
Both are impurities that make steel more likely to fail in a brittle way. Sulphur is the more damaging of the two, because it forms inclusions that stretch into stringers during rolling, reducing ductility through the plate thickness and giving hydrogen cracks a path to follow. Both are capped at 0.025%.
Not as standard. Carbon equivalent is not part of the base specification, though most mills report it and it typically runs 0.43 to 0.45. Certain supplementary requirements do impose a carbon equivalent limit for specific services, and clients working in demanding environments often add one to the purchase order.
Yes, identical. SA-516 is the ASME adoption of ASTM A516 and carries the same chemical requirements. The difference between the two designations concerns which body publishes the standard and whether the material may be used on a code-stamped vessel, not what is in the steel.
ASTM A516 Grade 70 is controlled on five elements. Carbon rises from 0.27% to 0.31% as plate gets thicker, manganese sits between 0.85 and 1.20% with room to go higher when carbon comes down, silicon runs 0.15 to 0.40%, and phosphorus and sulphur are both held to 0.025%.
Two points cause most of the confusion in practice. Carbon must always be compared against the correct thickness band, and manganese above 1.20% may be perfectly legitimate if the carbon has been reduced to allow it.
The composition table also does not tell the whole story. The requirements for killed steel and a fine grain structure sit outside it, and they are as much a part of what makes this a pressure vessel plate as the chemistry itself.