​‌​​‌​

EN AW 2017A / UNS A92017 / AMS 4110 / AMS 4116

​‌​​‌​

EN AW 2017A is a type of aluminium alloy generally known for its high strength and good machinability. The alloy belongs to the aluminium-copper (Al-Cu) group and the designation represents the EN (European Norm) number of alloy 2017. It is notable in particular for its high strength and slow oxidation. Its corrosion resistance can be lower than that of other aluminium alloys, however, so it generally requires coating or a protective treatment.

It is a high strength aerospace material, and is mostly used in the production of structural parts.​‌​​‌​

Heat treatment: The properties of this alloy can be improved by heat treatment. There are some particular points to observe during the machining and forming of this alloy, however.

Heat treatment (T conditions): Solution annealing gives the material better machinability. For heavily worked material, the T4 or T6 heat treatment conditions are preferred, and these treatments raise the mechanical strength of the material.​‌​​‌​

T6 condition (for aluminium 2017A): The material is heated to 490-510 °C, quenched rapidly in water and then subjected to a treatment at around 60 °C. This treatment gives the alloy high strength and high hardness.

T4 condition (for aluminium 2017A): The alloy is heated to the solution temperature and then cooled rapidly. This treatment gives the alloy better formability, but its strength is somewhat lower than in the T6 condition.​‌​​‌​

Machinability: Because it is a high strength alloy, it is more difficult to machine than some other aluminium alloys. With the right machining methods, however, it can be formed successfully.

​‌​​‌​

What EN AW-2017A Is — Europe’s Duralumin, and Why It Is Still Sold Only in T4

EN AW-2017A (chemical symbol Al Cu4 MgSi(A) / material number 3.1325 / old DIN name AlCuMg1 / French AFNOR name A-U4G / US equivalent A92017, i.e. AA 2017 / 2017A) is a heat-treatable aluminium-copper-magnesium alloy. Nominally it carries 3.5–4.5 % Cu, 0.4–1.0 % Mg, 0.4–1.0 % Mn and 0.2–0.8 % Si. It is the direct descendant of the first structural aluminium family ever used — the alloys historically called duralumin — and it is still the best-selling bar, rod and plate of that family in Europe.​‌​​‌​

There is exactly one key to understanding 2017A, and it is copper. Copper gives the alloy two things at the same time: 380–400 MPa tensile strength after solution treatment and ageing, and one of the worst corrosion behaviours in the aluminium family. A single element explains both why the alloy exists and where it must not be used. 2017A is a strength and machining alloy. It is not a corrosion alloy, and it is certainly not a welding alloy.

The second decisive purchasing fact is this: 2017A is in practice sold only in the T4 family of tempers (T4, T451, T3, T351) — that is, solution treatment + quench + NATURAL ageing. None of the manufacturer datasheets or EN 485-2 / EN 754-2 / EN 755-2 extracts we consulted contained an artificially aged (T6/T651) mechanical property table for 2017A. That is not an accident: artificial ageing of 2017A lowers ductility and worsens intergranular and stress corrosion susceptibility, so the industry chose to leave this alloy in T4. If you see a datasheet quoting numbers for 2017A-T6, ask where those numbers come from.​‌​​‌​

Are 2017 and 2017A the same thing?​‌​​‌​

In practice they are the same alloy family and are used interchangeably, but they are not the identical chemical band. AA 2017 is a registered American alloy; EN AW-2017A is the European variant registered in EN 573-3. The sources we consulted give 2017A a silicon band with a lower limit: Si 0.20–0.80 % — meaning silicon here is not an impurity but a deliberate alloying element that supports Al₂Cu with Mg₂Si precipitation. Order chemistry against the EN 573-3 band; do not assume “2017 ≡ 2017A”. In aerospace supply chains it is normal for the two numbers to appear separately on certificates.

Standards by Product Form

​‌​​‌​

​‌​​‌​

2017A is an EN alloy, not an “ASTM alloy”. That matters for ordering and certification: the natural language for 2017A is EN 573-3 (chemistry) plus the product-specific EN standard. On the ASTM side AA 2017 is a registered alloy, but the sources we consulted did not independently confirm which ASTM product specifications list 2017/2017A in their alloy tables — so we do not fill those rows below with grade lists. Before offering 2017A to a buyer who requires an ASTM certificate, have it confirmed that the grade appears by name in the relevant ASTM scope.

ASME Code Acceptance and Service Temperature — the Honest Answer

​‌​​‌​

There is no published ASME code temperature table for 2017A, because the alloy itself is not accepted as a material in the ASME pressure vessel code. Rather than invent numbers for this section, we give the verifiable temperature information we have and state its limits plainly.​‌​​‌​

Product Forms With NO Standard — the Commercially Valuable Section​‌​​‌​

Knowing which product form of an alloy has no standard usually earns more money than knowing which ones do. For 2017A the list starts with welding consumables.

Chemical Composition

The band below follows EN 573-3. One row carries a genuine contradiction between sources and we are not hiding it.​‌​​‌​

What the chemistry directly causes​‌​​‌​

1. 4 % copper = no fusion welding. In the Al-Cu system this composition creates a wide brittle range during solidification, and the weld metal cracks as it freezes. 2. 4 % copper = poor corrosion. Al₂Cu particles are cathodic to the matrix; the aluminium around them dissolves preferentially. That is the mechanism behind pitting, intergranular corrosion and exfoliation. 3. 4 % copper = good machining. The same hard particles break chips; 2017A machines noticeably more comfortably than the 5xxx series. 4. 4 % copper = no decorative anodising. A copper-bearing surface gives a dull, yellowish, inconsistent colour in the anodic film; the manufacturer rating for decorative anodising is 5 (unsuited). 5. 4 % copper = no food contact (DIN EN 602).

Mechanical Properties​‌​​‌​

The values below are standard minima unless noted otherwise. Hardness figures are typical values for information, not acceptance criteria — do not reject an aluminium plate on Brinell. The Rm values are given as minima for T4/T451.​‌​​‌​

Minimum versus typical: the Rp0.2 and A values above are minima and are the inspection criteria. Measured values on a real plate will typically exceed them. But the number used in a design is the minimum, not the typical. We found one manufacturer page quoting Rm 450 MPa / Rp0.2 400 MPa / 120 HB for 2017A: those figures match none of the EN 485-2 or EN 755-2 tables and must not be used in design.​‌​​‌​

Physical Properties

Electrical conductivity measurement is a quality tool on 2017A. A correctly solution-treated and correctly quenched 2017A has low conductivity (copper in solution). In a slowly cooled or overaged part the copper precipitates and conductivity rises. This is why aerospace uses eddy-current conductivity measurement as the non-destructive way to check whether the heat treatment was done correctly. High conductivity is not good news here.

Heat Treatment and NATURAL AGEING — the Most Misunderstood Topic on This Alloy​‌​​‌​

Natural ageing never fully stops — and that is a practical problem​‌​​‌​

The textbook says “5–8 days”, and that is the time after which the strength becomes engineering-stable. But the precipitation process does not end there. In the Al-Cu-Mg system, room-temperature precipitation continues for years, decelerating logarithmically. This has four practical consequences, and all four show up on the shop floor:

1. The rivet problem — the most concrete proof of natural ageing. 2017 is an aerospace rivet alloy. Immediately after quenching the rivet is soft and can be driven; within hours it hardens and then cracks when driven. Aerospace practice has historically been to keep rivets of this alloy refrigerated (“ice-box” rivets) or to re-solution-treat them immediately before driving. That proves natural ageing is not an abstract metallurgy topic but a daily production constraint. 2. Stock changes while it sits. A 2017A-T4 bar that has been on the rack for two years can be measurably harder than one delivered three months ago. The difference in chip behaviour and burr between two batches run on the same program is usually exactly this. 3. Hardness test “discrepancies”. The small gap between the HB on the certificate and the HB you measure today is usually not a non-conformance — it is time. 4. The forming window after quenching. Difficult bends are made in the W condition right after solution treatment; that window is measured in hours, not days.​‌​​‌​

T3 · T351 · T4 · T451 — what each one means

T4: solution treated + naturally aged. T3: solution treated + cold worked + naturally aged — you see it on drawn products (EN 754-2); the cold work lifts the strength slightly. The trailing “51” (T351, T451): stress relieved by stretching. The trailing “510” and “511” (T4510, T4511): stress relieved by stretching in extrusion — 510 = no further straightening after stretching, 511 = minor straightening permitted. Practical machining rule: if you are cutting an asymmetric part out of thick plate, specify a stress-relieved temper — T451 / T351 / T4511. In non-stress-relieved thick plate the part bananas after the first roughing pass.​‌​​‌​

Welding — 2017A IS NOT WELDABLE

There is no nuance in this section. The manufacturer datasheets we consulted rate 2017A gas welding 5, TIG 5, MIG 5 — on the scale used, 5 = unsuited. Another supplier states it in plain words: “traditional welding is not possible”. 2017A is not joined by fusion welding processes.​‌​​‌​

Why it does not weld — three separate mechanisms​‌​​‌​

1. Solidification (hot) cracking. Al-Cu compositions around 4 % copper have one of the widest brittle temperature ranges during freezing: as the weld pool solidifies, liquid film still remains at the grain boundaries when shrinkage stress arrives, and it cracks. This is not a workmanship problem; it is the composition itself. 2. HAZ liquation and strength loss. Welding heat exceeds the solution treatment temperature; the heat-affected zone goes through an uncontrolled heat treatment, low-melting phases at grain boundaries liquate, and the zone both weakens and opens to cracking. 3. Collapse of corrosion resistance. One supplier puts it exactly this way: “traditional welding is not possible as the corrosion resistance characteristics would be affected”. Welding heat produces a copper-rich precipitate network at grain boundaries; the result is a narrow band along the weld that is open to intergranular corrosion.

What is done in practice: 2017A parts are riveted, bolted, spot welded or bonded. If a welded aluminium structure is required, the alloy is changed — for welded chassis, tanks and boats EN AW-5083 and EN AW-5754 are the right addresses, and for weldable structural profiles EN AW-6082. There is no middle road called “we only weld a little”.​‌​​‌​

Machining

This is 2017A’s strongest side. In the heat-treated condition the manufacturer machinability rating is 1–2 (very good / good); in the soft annealed condition it is 4 (poor). In other words, machine 2017A hard, not soft — a rule that looks counter-intuitive but is correct for aluminium: soft aluminium smears and builds up on the edge (BUE); hard 2017A breaks clean chips.​‌​​‌​

Distortion is the real issue. The hard part of machining 2017A is not cutting — it is residual stress. Quenched thick plate carries serious residual stress between surface and core. Removing material from one side unbalances that stress and the part moves. Three countermeasures: (1) specify a stress-relieved temper — T451 / T351 / T4511; (2) a symmetric cutting plan — remove material evenly from both faces; (3) rest the part between roughing and finishing, with a separate setup if possible. Without these three, no tool will save the part.​‌​​‌​

Corrosion — WHERE IT FAILS

​‌​​‌​

This section describes where 2017A cannot be sold, and it should be read honestly. The manufacturer rating tables are unambiguous: normal atmosphere / weather: 4 (poor), seawater: 4–5 (poor to unsuited). For comparison, on the same scale EN AW-5754 scores 1 in atmosphere and 1–2 in seawater; EN AW-5083 scores 1 in both. That is not a nuance, it is a class difference.​‌​​‌​

Mechanism — why it is this bad

Aluminium’s corrosion resistance rests on a self-repairing surface Al₂O₃ film. On 2017A that film is still there — the problem is underneath it. In the matrix and at the grain boundaries sit Al₂Cu (θ) and Al₂CuMg (S) particles. These are cathodic to the surrounding aluminium: in the presence of an electrolyte the aluminium around the particle becomes the anode and dissolves. In other words, the alloy contains microscopic galvanic cells inside itself. The consequences:​‌​​‌​

How it is protected​‌​​‌​

1. Alclad. The classic solution for copper-bearing 2xxx sheet is cladding both faces with pure or low-alloy aluminium (alclad). The cladding acts both as a barrier and as cathodic protection: even when scratched, the cladding sacrifices itself. Caveat: alclad 2024 is common; alclad 2017A availability could not be verified in this study — confirm before ordering. Also, if you machine into the thickness of an alclad sheet you remove the protection; a machined part is not alclad. 2. Protective anodising. Rating 2 (good). Chromic or sulphuric acid anodising plus a paint primer is the aerospace standard. The decorative anodising rating is 5 (unsuited) — copper gives a dull, inconsistent colour. Never promise a “natural anodised decorative finish” on 2017A. 3. Paint and coating. Rating 3 (moderate); used together with a chromate-phosphate or chromate-free conversion coating primer. 4. Design. Remove water-trapping pockets, capillary gaps and hidden crevices; protect edges and hole walls too — that is where exfoliation starts.

Where it must not be used — the explicit list​‌​​‌​

Seawater and marine atmosphere (uncoated). Buried or permanently wet service. Welded structures. Food and beverage contact (not suitable per DIN EN 602). Decoratively anodised architectural surfaces. Uncoated exterior parts exposed to road salt. Welded pressure vessels. For every item on this list the right answer is to change the alloy — not to use thicker 2017A.

Frequently Asked Questions​‌​​‌​

Should I buy 2017A-T4 or 6082-T6? Both give roughly the same yield strength.

In almost every case, 6082-T6. Look at the numbers: 6082-T6/T651 plate at 3–6 mm gives Rm 310 MPa · Rp0.2 260 MPa; 2017A-T4/T451 plate in the same band gives Rm 390 MPa · Rp0.2 245 MPa. So 6082 leads on yield strength and 2017A leads on tensile strength. Because most designs are sized on yield strength, the strength argument usually does not justify 2017A. Meanwhile 6082 welds, resists corrosion and anodises; 2017A does none of the three. Choose 2017A only for these three reasons: (1) high tensile strength and fatigue are genuinely required, (2) chip breaking and tool life in volume machining are decisive, (3) the part is a rivet or fastener. If there is welding, outdoor exposure or a visible surface, move to EN AW-6082.​‌​​‌​

The certificate says T4 but the hardness reads higher than the certificate. Is this a non-conformance?

Most likely not — most likely it is time. 2017A is a naturally ageing alloy and the precipitation process continues for years at room temperature, decelerating but never fully stopping. The textbook “5–8 days” marks the point where the condition becomes engineering-stable; it does not mean the process halted. A bar that has been in stock a long time can read measurably harder than fresh material from the same cast. The acceptance criteria are not hardness but the Rp0.2, Rm and A minima in EN 485-2 / EN 754-2 / EN 755-2 — and those minima may be exceeded upward. The real non-conformance is this: elongation (A) falling below the minimum. If there is a dispute, run a tensile test; do not decide on hardness. Note: this effect hits production planning too — it is normal for two batches delivered two years apart to give different burr and dimensions on the same program.​‌​​‌​

Can we repair our 2017A part with TIG? It is only a small crack.

No. This has nothing to do with the welder’s skill. Three mechanisms act at once: (1) solidification cracking — an Al-Cu composition at 4 % copper leaves liquid film at grain boundaries as the weld pool freezes and it cracks under shrinkage stress; this is the composition itself and no parameter fixes it; (2) HAZ liquation and strength loss — welding heat exceeds the solution treatment temperature and the heat-affected zone undergoes an uncontrolled heat treatment; (3) collapse of corrosion resistance — in a supplier’s own words, “traditional welding is not possible as the corrosion resistance characteristics would be affected”. On top of that there is no suitable filler wire: no filler metal of 2017 chemistry exists in the AWS A5.10 list. What to do instead: replace the part, or make a mechanical repair — doubler plate, rivets, bolts, fasteners. And if the part lives inside a structure that will be welded, the wrong alloy was specified and the design should move to 5083 / 5754 / 6082.​‌​​‌​

What is the real difference between 2017A and 2024, and which should I stock?

In chemistry the single decisive row is magnesium: 0.4–1.0 % in 2017A against 1.2–1.8 % in 2024. The copper bands are nearly identical (3.5–4.5 against 3.8–4.9). The extra magnesium strengthens the Al₂CuMg (S phase) precipitate and lifts the strength: in plate, 2024-T351: Rm ~435–440 MPa · Rp0.2 ~290 MPa, against 2017A-T4/T451: Rm ~390 MPa · Rp0.2 ~245–260 MPa. The price: 2024 has even worse corrosion behaviour (alclad is close to mandatory), lower ductility and formability, and higher cost and supply constraint. Neither is fusion weldable (both rate 5 for gas/TIG/MIG). Stocking decision: if you do aerospace structural sheet work, 2024 (and alclad) is unavoidable; for general machine building, tooling plate, machined mechanical parts and fasteners, 2017A is cheaper, easier to machine and more ductile. Do not substitute one number for the other — on a certificate they are separate alloys.​‌​​‌​

Common Datasheet Errors and Traps

1. “Rm 450 MPa, Rp0.2 400 MPa, 120 HB”. We found these figures on a datasheet for 2017A. They match none of the EN 485-2, EN 754-2 or EN 755-2 tables. The standard minima in plate are in the Rm 390 / Rp0.2 245–260 band. Do not use them in design.
2. “Electrical conductivity 33.5 MS/m”. Same page. That works out to roughly 58 % IACS, which is effectively pure aluminium territory. The verified band is 18–28 m/(Ω·mm²), roughly 31–48 % IACS. If you measure 58 % IACS on a 2017A part, what you are measuring is not 2017A.
3. “Good weldability”. Some commercial pages state this for 2017A. It is wrong. Manufacturer rating tables give gas/TIG/MIG a 5 = unsuited. The correct sentence is: resistance (spot) welding is suitable, fusion welding is not.
4. “Good corrosion resistance”. Some supplier pages use this phrase. The same source then says a few lines later that welding would damage the corrosion resistance. The rating tables are clear: 4 in atmosphere, 4–5 in seawater. 2017A’s corrosion resistance is poor relative to copper-free aluminium; the word “good” is defensible only against steel, which is not a useful comparison.
5. Treating T4 and T451 as the same. Same strength, different residual stress. The trailing 51 means stress relieved by stretching. If you are cutting an asymmetric part from thick plate, that difference decides whether the part bananas on the table.
6. Confusing T3 with T4. T3 includes cold work in between and appears on drawn products (EN 754-2); T4 appears on extrusions and plate. The values are not the same.
7. Looking for 2017A-T6. The EN tables we consulted contain no T6/T651 mechanical values for 2017A. If someone quotes standard values for 2017A-T6, ask for the source.
8. Mistaking “service temperature 135–145 °C” for a code limit. It is guidance from one manufacturer datasheet family, not a design limit. There is no ASME code temperature for 2017A.
9. Treating the solidification range as one number. Sources contradict between 512–650 °C and 555–640 °C. We have seen both; do not publish one of them as the single truth.
10. The chromium row. One source gives Cr max 0.10 %, a manufacturer gives 0.10–0.25 %. If you are writing a purchase specification, confirm that row against the current edition of EN 573-3.
11. Treating 2017 and 2017A as identical. Same family, not the same band. In aerospace supply chains they appear separately on certificates.
12. “Cast 2017A”. No such material exists; 2017A is a wrought alloy. Copper-bearing aluminium castings are a separate family (EN AC- series).
13. Counting alclad as protection on a machined part. The alclad layer is on the surface; machining into the thickness removes the protection. And alclad 2017A availability could not be verified in this study.
14. Promising decorative anodising. The protective anodising rating is 2 (good), the decorative anodising rating is 5 (unsuited). Copper gives a dull, inconsistent anodic colour.
15. Using it in food contact. Not suitable per DIN EN 602. On the same criterion, 5754 and 5083 are suitable.
16. Planning bends from thick plate. Elongation falls to 4 % at 150 mm and 2 % at 200 mm. Thick 2017A is not a forming material.
17. Assuming “thinner is stronger” in extrusions. In the EN 755-2 table the highest values are not at ≤25 mm but in the 25–75 mm band.
18. Using stainless fasteners. 2017A is active relative to stainless; in a wet environment the aluminium is consumed at the joint. Insulation or compatible fasteners are mandatory.​‌​​‌​

​‌​​‌​