Two tube bundles came out of the same crude unit overhead condenser last spring. Same chloride load on the spec sheet, same 90 °C metal temperature, same tube sheet. One had been in service six years and looked like it had another five in it. The other perforated in eleven months.
The one that failed was the 625 bundle. It had been specified on a single rule — the 316L was pitting, so move up to the higher alloy and the pitting stops. The one that lasted had been specified after someone sat down with the chloride number, the deposit behavior and the crevice geometry and asked a narrower question: which failure mode is actually going to take this part out, and does the grade I am buying block it? INCONEL alloy 625's own published data answers composition, temperature range and application fit. It does not answer that narrower question for you, and neither does any other datasheet.
What Alloy 625's Composition Range Actually Buys
Start with what the number 625 stands behind. The grade is a nickel-chromium alloy, UNS N06625, W.Nr. 2.4856, and its published use case is built on high strength. That is the first thing the specification tells you, and it is not the corrosion promise most people read into it.
The composition band is where the real information sits. Nickel runs about 58 percent minimum, balance. Chromium sits at 20 to 23 percent, molybdenum at 8 to 10, niobium plus tantalum at 3.15 to 4.15, iron capped at 5, carbon at 0.10 max, with small deliberate additions of aluminum and titanium. Chromium builds the passive film. Molybdenum is the lever that decides whether a chloride stream pits a surface or leaves it alone — every point of moly raises the critical crevice temperature, and 8 to 10 percent is a two-point spread, not a rounding error. Niobium plus tantalum ties up carbon as a stable carbide, which is why welded 625 does not sensitize the way an unstabilized austenitic does, and it is also the main lever on yield strength.
Here is the part the datasheet layout encourages you to skip: that band is a window, and a mill can sell you a heat at either edge of it. Two heats of 625, one at 8.0 percent molybdenum and one at 10.0, both fully compliant, do not behave the same in a chloride stream at 60 °C under a gasket. The lean one pits. The rich one often does not. The same holds for strength: a heat landing at 3.2 percent Nb+Ta has room-temperature yield sitting near the bottom of the published range, which matters if your design margin was calculated on a typical value instead of a minimum.
So when you put supplier datasheets side by side, put the specification limit next to your duty and then ask for the actual heat chemistry. A 625 that meets the chemistry spec is 625. It is not necessarily the 625 your corrosion calculation assumed.
How Cold and How Hot 625 Is Allowed to Run
The published service range runs from cryogenic all the way to 1800 °F, 982 °C. That is one sentence, and people read a lot into it.
At the cold end the number is honest in a useful way. The alloy keeps its toughness down through liquid nitrogen temperatures and below, which is why it turns up in cryogenic piping where a ferritic steel goes brittle. Fewer surprises there, and less argument about them.
The hot end is where a single number stops being an answer. A continuous-service envelope is not a time-temperature curve for your part. The same plate that is perfectly strong at 600 °C will, after long exposure above roughly 650 °C, precipitate secondary phases — carbides and a niobium-rich delta phase — that pull ductility and room-temperature toughness down. Short-term strength at 900 °C and ten-year strength at 900 °C are different materials, mechanically. Wall thickness changes the answer again: a 3 mm thermowell and a 25 mm flange do not see the same thermal history at the same furnace temperature.
What the range does not carry is exposure time, stress level or section size. Published allowable stress values for 625 fall off steeply toward the top of that band and, for several product forms, stop well short of 982 °C. The datasheet gives you the alloy's envelope. Your duty gives you the point inside it, and the two are not the same document.
Where the Strength-and-Corrosion Combination Earns Its Place
The published application list for 625 is unusually specific, and it is specific for a reason. Bubble caps, tubing, reaction vessels, distillation columns, heat exchangers, transfer piping, valves — all of them appear as parts that need the combination of strength and corrosion resistance the alloy offers.
Read that list and a pattern shows up. Every item sits in one of two positions: the wall is thin because heat transfer or flow demands it and it still has to carry pressure, or the process stream carries chlorides at a temperature where 316L has already given up. Bubble caps are stamped thin and live inside a column whose overhead is wet and chloride-bearing; you cannot thicken them for corrosion allowance without wrecking the hydraulics. Distillation columns and transfer piping move hot chlorinated streams where a 316L line pits through at the welds. Heat exchangers stack thermal cycling on top of deposit-driven attack. Valves need seat and stem strength while a corrosive, moving fluid runs past them, and a soft corrosion-resistant alloy that yields under actuator load is not a valve.
That is why 625 lands on those parts and not on a cool, dry, chloride-free storage tank. Each one has to hold pressure in a thin wall while the same surface faces chlorides, and those two demands do not often arrive together.
Why 625 Still Pits Under Your Deposits
Most people in a reliability meeting assume that the moment a 316L part pits, moving up to 625 removes the problem, and that the higher alloy number keeps them safe across the whole temperature range. Actually the corrosion data behind the 625 datasheet is generated on clean, flowing, aerated surfaces in a controlled cell. Your heat exchanger tube is none of those things.
Put a sludge layer, a scale band or a gasket crevice on the surface and the local chemistry changes. Chloride migrates into the trapped liquid, oxygen differentials set up, and the concentration under the deposit climbs far above the bulk value, especially when the metal temperature is high enough to evaporate the trapped water. A ten- to hundred-fold local chloride increase is enough to push a 625 surface out of its passive region. Crevice corrosion and pitting then show up on an alloy the datasheet lists as resistant to both. Stagnant seawater in a dead leg is the same story with a biological twist: 625 is not immune to microbiologically influenced corrosion, and a molybdenum addition does not disinfect anything.
The failure modes are predictable once you accept the mechanism. A 625 overhead condenser with a fouling seawater side perforates tube walls in months, not years. A transfer line with a low point that never drains leaks at that low point. A flanged joint with a spiral-wound gasket crevice-corrodes in the gap long before the pipe body shows anything. None of this contradicts the datasheet. The datasheet was simply never describing your surface.
This is the condition that overturns the high-grade-equals-safe reasoning: if the deposit cannot be removed, or the dead leg cannot be drained, or the cleaning interval cannot be held, then the corrosion resistance you paid for is not in contact with the corroding liquid. The strength advantage is still real and still there. The corrosion premium is not being delivered.
When a Cheaper Grade Beats Alloy 625
The second assumption that costs money is that a higher alloy number delivers proportionally better service life. In fact 625 occupies one specific point in the family — a strength-plus-chloride-resistance point — and it is not the top of a corrosion ladder. Move off that point in either direction and you are paying for capability you never use.
A mid-temperature organic acid duty with no chlorides is a 316L job, and 625 there is a large material cost multiplier with nothing to show for it. A reducing acid with chlorides and no structural load is a C-276 job: more molybdenum and tungsten, higher pitting and crevice resistance than 625, at a higher price and with lower yield strength — so if the part also carries load, C-276 forces a thicker wall and can end up costing more than the 625 you were trying to avoid. Many sulfuric and phosphoric acid duties at moderate temperature are covered by alloy 825, which trades some of 625's pitting resistance for good chloride stress corrosion cracking resistance at a lower alloy cost. Over-specifying 625 brings its own problems, too: it welds readily, but the niobium in the weld pool segregates and heavy sections need care, and above 650 °C you inherit the embrittlement question from the temperature range rather than escaping it.
So the selection rule is narrow, and it belongs in the specification rather than in a habit: use 625 when the same part must carry structural load and resist chloride attack, and when the deposits on that part can be controlled or cleaned on a defined cycle. If either half of that is missing, pick the grade that matches the failure mode you actually expect — and expect 316L, 825 or C-276 to win on cost for a lot of duties that currently get 625 by default.
Which brings back the bundle that was still running after six years, and answers the fair question left over: is the premium ever worth paying? It is, and the conditions are specific. That specification did not say "alloy 625." It said alloy 625 with a molybdenum heat chemistry in the upper half of the range, a maximum metal temperature of 90 °C, and a chemical cleaning interval of six months written into the operating procedure. The bundle that failed had the same grade stamped on the tube sheet and none of those three numbers anywhere in the file. Put the wall thickness, the maximum metal temperature and the cleaning interval into the purchase specification, and ask for the heat chemistry on the mill certificate before the tubes are cut. The grade number on its own was never the insurance policy it looked like.
