Titanium Hydrogen Embrittlement: Why a Compliant Mill Certificate Does Not Guarantee a Ductile Part

Product Overview

Reviewed and maintained by the Engineering Team, Shaanxi Qianyi Tuoda Technology Co., Ltd.

Titanium Hydrogen Embrittlement: Why a Compliant Mill Certificate Does Not Guarantee a Ductile Part

A mill certificate can show titanium at 150 ppm hydrogen, fully inside the commonly quoted Grade 2 ceiling of 0.015 wt%, and the finished component can still crack in service. The reason is simple and it is almost never written into a purchase order: the certificate describes the metal as it left the mill, while hydrogen is added — and removed — by nearly every fabrication step that follows. Acid pickling, a damp shielding gas line, a cathodic protection system set too negative, an electrowinning cell, a furnace with a wet atmosphere, and a hot reducing acid stream can each move hydrogen across that boundary. Once solubility is exceeded, the excess precipitates as brittle hydride platelets; elongation, impact toughness and fatigue resistance fall before any surface change is visible. This guide sets out the mechanism, the six entry routes, why the mill limit is the wrong acceptance criterion, what ASTM E1447 actually measures, where published numbers disagree, the practical limits of bake-out, an eight-condition failure list and a purchase-order template. It is deliberately not a cleaning guide: for iron contamination, nitric-to-HF ratio and pickling acceptance, use our titanium pickling and iron contamination guide. That article covers the cleaning process; this one covers what hydrogen does to the metal afterward. For the incoming-document side, the EN 10204 3.1 certificate page shows what a mill certificate does and does not assert.

At a Glance: The Hydrogen Balance Sheet

Read the second, seventh and eighth rows first. Most disputes we see are not about bad metal; they are about a buyer who treated a chemistry ceiling as a fabrication allowance, a laboratory that reported a number below its own validated scope, or a shop that tried to fix hydrogen with a low-temperature baking cycle copied from plated steel. All three are documentation errors rather than metallurgy errors, and all three are cheaper to close in the enquiry than in a failure investigation. The numerical limits in the first row are widely quoted in industry summaries of ASTM B265, B338 and B348, but they differ between product forms, grades and standard editions. Treat them as orientation only; the governing number is the one printed in the edition named on your own purchase order, confirmed against the certificate in front of you.

The Hydride Mechanism: Why Titanium Is Not Steel

Hydrogen is an interstitial element in titanium, but it is not simply dissolved forever. Alpha titanium has very limited room for hydrogen at room temperature; published solubility figures vary widely with oxygen content, alloying, temperature and the source of the phase-diagram data, from single-digit ppm to a few tens of ppm. We will not give you one authoritative number, because the literature does not offer one that is safe to put in a specification. What matters commercially is the consequence: the amount of hydrogen a mill is allowed to deliver can already be at or above the amount that stays comfortably in solution at ambient temperature. When local solubility is exceeded — by cooling, by stress, by a composition gradient, or by fresh hydrogen arriving from a process step — the excess forms a hydride phase. Those platelets are brittle, they interrupt slip, and they become crack initiators under impact or cyclic load.

The difference from steel is not academic; it changes what you can do about the problem. In plated high-strength steel, a few hours at roughly 190 to 230 °C is a recognised de-embrittlement treatment because diffusible hydrogen leaves the lattice. In titanium, heating first dissolves the hydride back into solid solution, which can make the part look better on a quick bend test while the hydrogen is still inside. Unless the cycle provides a genuine driving force for hydrogen to leave the metal — high vacuum, clean surfaces, sufficient temperature and time, and a verified low partial pressure — the hydride simply re-forms on cooling, sometimes in a more damaging orientation. That is why a low-temperature oven cycle borrowed from a steel plating specification is worse than useless here: it can consume schedule, alter surface condition, and still leave the hydrogen in place.

Three further consequences follow. First, hydrogen damage is reversible in form but not in consequence: hydride morphology can change with thermal history, but a crack that has already initiated does not heal. Second, the loss shows up first in ductility and fracture toughness rather than in tensile strength, so a standard tensile test on a certificate can look entirely normal on material that fails a slow-bend or impact check. Third, local hydrogen matters more than average hydrogen: a surface layer charged by pickling or a weld root charged by moisture can embrittle a component whose bulk average still passes. That last point is the reason sampling position and sample preparation deserve their own section below, and why a single number on a certificate can never be the whole acceptance story.

The MTC Trap: 150 ppm Is a Mill Limit, Not a Service Limit

A product standard such as ASTM B265 for plate or ASTM B348 for bar sets a maximum hydrogen content for the material in its delivered condition. Industry summaries commonly quote 0.015 wt% (150 ppm) for unalloyed Grade 2, roughly 0.0125 wt% (125 ppm) for Grade 5 and about 0.010 wt% (100 ppm) for Grade 23 ELI. Those figures are useful for reading a certificate, but they are not interchangeable: the applicable number depends on the product form, the grade, and the edition of the standard named in the contract. We state them as commonly cited orientation values, not as a table to copy into a design calculation. Ask the mill to quote the exact clause and edition on the certificate, and check it against your own PO.

The one-sentence version: a hydrogen line on an EN 10204 3.1 certificate is evidence about the parent heat at the time of manufacture. It is not a measurement of the component in your hand after cutting, pickling, forming, welding, heat treatment or six months of cathodic protection. If hydrogen is a design concern, the acceptance point has to be re-created after the last hydrogen-generating operation, not read off the incoming paperwork. The equipment arrival acceptance guide covers the incoming-document check; it does not replace a post-fabrication hydrogen determination.

There is a second, subtler trap. A certificate that shows a value near the limit is compliant, but a buyer who plans any hydrogen-generating process has already spent the entire allowance before starting. A specification for a welded or pickled component therefore needs a budget: an incoming maximum well below the product-standard ceiling, plus a defined maximum increase after fabrication, plus a named re-test point. This is the same logic used for hydrogen control in high-strength fasteners, and the titanium fastener assembly guide shows the mechanical consequence when ductility is quietly lost in a loaded joint. Where a project writes only "hydrogen per ASTM B265" on the PO, the supplier has done nothing wrong by shipping compliant material — and the buyer has no contractual basis for a dispute when the finished part is brittle.

A third trap is methodological. Published descriptions of mill certificate practice and of ASTM E1447 place the routine quantitative scope of the inert gas fusion method in the range of approximately 9 to 320 mg/kg hydrogen. If a project specification demands a single-digit ppm maximum, a laboratory cannot simply print that number on a routine E1447 report and call it compliant; the method extension needs validation, and the report should say so. Ask which edition of E1447 was used, what calibration and blank corrections were applied, and whether the reported value sits inside the validated range. Do not confuse this with the mechanical hydrogen-embrittlement tests used for plated steel fasteners: those are different test families with different acceptance logic, and importing their language into a titanium PO is one of the most common specification errors we see.

Where Hydrogen Enters: Six Sources From Melt to Site

Hydrogen is not a defect that appears once at the mill. It is a balance that shifts at every step where titanium sees moisture, acid, a reducing environment, or a cathodic potential. Listing the six routes in one table matters because most investigations look only at the last one. A part that fails after pickling is often blamed entirely on the pickle bath, when the incoming stock was already near its ceiling and the bath simply pushed it over; a part that fails in seawater is often blamed on the grade, when an impressed-current system was the actual hydrogen source.

Routes one and two are the supplier domain. Sponge quality, vacuum arc remelting practice and the atmosphere during forging and annealing decide the starting hydrogen. This is the number the certificate reports, and it is the only one the mill can be held to. Ask for the hydrogen value on the certificate rather than a generic statement of compliance, and for critical orders ask whether the heat was multiply melted. Route three is the fabrication step most often underestimated. Nitric-hydrofluoric pickling is a controlled dissolution process; when the hydrofluoric fraction is too high, immersion too long, or rinsing incomplete, hydrogen is generated at the metal surface faster than it can leave. The ratio matters more than the absolute concentration, and the correct practice is set out in the pickling and contamination guide.

Route four overlaps with weld shielding. Titanium above roughly 400 °C absorbs oxygen, nitrogen and hydrogen readily, so moisture anywhere in the gas train, on the filler wire, or in the joint preparation becomes an interstitial problem. Our GTAW shielding guide covers the argon purity, dew point and three-zone protection in detail; the early weld failure review shows what those lapses look like on real tube welds. Here the relevant point is narrower: a weld can be visually perfect, with an acceptable silver or light-straw tint, and still have introduced hydrogen along the fusion line if the gas was wet. Colour is an oxygen-contamination indicator, not a hydrogen test.

Routes five and six are where buyers have the least control and the most leverage. Cathodic protection is applied to protect steel, and titanium in the same circuit is a hydrogen-evolving cathode; electrowinning cathode plates are deliberately polarised in exactly the same way. For an electrochemical view of the assembly, the titanium cathode plate guide describes the electrowinning hardware, while the Gr2 vs Gr7 seawater comparison covers grade selection in the chloride service where such systems are common. Neither replaces hydrogen surveillance. If a titanium component sits in a polarised circuit, the potential window belongs on the drawing, the reference electrode must be named, and a periodic hydrogen determination belongs in the maintenance plan.

Finally, reducing acids deserve a specific warning. Where the passive film is attacked, the corrosion reaction itself evolves hydrogen at the surface and some of it is absorbed. The HCl grade-limit guide sets out the concentration and temperature boundaries by grade, and the hydrometallurgy acid-leaching plate page shows the plant context. A grade chosen from an iso-corrosion curve solves general metal loss; it does not by itself close the hydrogen balance. Where the process is hot, deaerated and reducing, ask for both a corrosion assessment and a hydrogen assessment rather than assuming one covers the other.

Pickling, Welding and Cathodic Protection: The Three Post-Delivery Sources

These three deserve a separate chapter because they are the operations that happen after the certificate is signed, and therefore after most buyers stop thinking about hydrogen. Each has its own specialist article on this site. This section does not repeat those guides; it explains what each one does to the hydrogen balance and what evidence you should demand when it is finished.

Pickling: the hydrogen increment is the acceptance criterion

A pickled surface can be clean, bright, free of iron and still be hydrogen-charged. The useful measurement is not the absolute hydrogen after the bath but the increment between before and after, taken on samples from the same heat and prepared the same way. Our pickling guide discusses a maximum hydrogen pickup on the order of 0.002 wt% (20 ppm) and the controls behind it: nitric-to-HF ratio, bath temperature, immersion time, inhibitor practice and thorough rinsing. Treat any such figure as a project acceptance value confirmed against the applicable standard and the bath qualification record, not as a universal constant. Where a supplier cannot produce a before-and-after pair, they have shown you a clean part, not a controlled process. Passivation of fasteners follows the same logic; the passivated fastener page covers the product side.

Welding: cleanliness is a hydrogen control, not just an appearance control

Shielding-gas purity, dew point, filler storage and joint cleaning are usually specified to prevent oxygen and nitrogen embrittlement and discolouration. They also control hydrogen. Moisture in a gas hose, condensation on cold filler wire, cutting fluid left in a machined groove, or a humid shop after a weekend shutdown can all supply the same hydrogen that an acid bath would. Because the weld bead colour gives no information about hydrogen, the only defensible evidence is the gas and consumable records plus, on critical work, a hydrogen determination on the completed joint. Qualification of the procedure is a separate question, covered in the WPS and NDT qualification guide; a qualified WPS proves the shop can make a sound weld, not that the gas bottle was dry on a particular Tuesday.

Cathodic and electrolytic polarisation: protection for steel, charging for titanium

When a mixed assembly is protected, titanium is normally the cathode and the steel is consumed instead. Push the potential too negative and hydrogen evolves on the titanium itself, where it is absorbed and forms hydrides. In electrowinning the same thing is intentional: the titanium cathode plate evolves hydrogen as part of the process, which is why cathode plates have a finite service life and are inspected for embrittlement. Published negative potential limits disagree and depend entirely on which reference electrode is named, so we do not print a single safe number here. The executable actions are: name the reference electrode on the drawing, set the window from the project specification, log potential with a reference cell at the titanium surface, and test hydrogen if over-protection is suspected. For parts in such circuits, the titanium coil page and the cathode plate guide show the hardware that most often carries this risk.

The common thread is that all three operations move the hydrogen balance after the mill has closed its file. If the purchase order contains only an incoming chemistry line, there is no contractual moment at which the final condition is checked. Adding one post-process re-test point, with a numeric limit and a named method, closes that gap for very little cost. That single line is the difference between a materials argument and an acceptance test.

Grades and Microstructures: Gr2, Gr5, Gr12 and Gr23 ELI Do Not Respond Alike

Hydrogen solubility and hydride habit depend on phase constitution, not just on chemistry. Alpha alloys, near-alpha alloys and alpha-beta alloys distribute hydrogen differently, and the beta phase generally accommodates more hydrogen in solution. That is why a single grade-independent hydrogen limit is a poor specification, and why two parts with the same measured bulk hydrogen can behave differently under impact.

Commercially pure grades (Gr1 to Gr4)

Unalloyed alpha titanium is the usual chemical-process choice and the most common material in our stock. Its hydrogen ceiling is the familiar 0.015 wt% figure in many industry summaries, and its hydride platelets form readily once local solubility is exceeded. Oxygen and iron content rise from Grade 1 to Grade 4, raising strength and slightly changing tolerance to interstitial embrittlement. For plate and bar selection by end use, the ASTM B265 plate guide and the ASTM B348 bar guide give the ordering framework; hydrogen control is an additional line on top of it, not a substitute for choosing the right grade.

Grade 5 and Grade 23 ELI

Ti-6Al-4V is an alpha-beta alloy. The beta phase holds more hydrogen, so the relationship between total hydrogen and observed embrittlement is not the same as in CP titanium; this is one reason implant and aerospace specifications set their own, tighter limits and their own test regimes. Grade 23 ELI is the low-interstitial variant with reduced oxygen, iron and hydrogen ceilings, and it is a different certification chain, not a marketing label. The Gr5 vs Gr23 comparison explains why ELI is selected for fracture toughness rather than corrosion resistance. A buyer who specifies industrial Grade 5 and cites an implant hydrogen number has mixed two standards systems in one PO; auditors will notice.

Gr7, Gr12 and other corrosion grades

Palladium-containing Grade 7 and the Mo-Ni Grade 12 are chosen for reducing-acid and hot-chloride performance, not for hydrogen behaviour. Their value is in keeping the passive film intact, which indirectly reduces hydrogen evolution from active corrosion. Where crevice corrosion or reducing acid is the real threat, the Gr2 vs Gr7 seawater guide and the HCl limits guide are the correct starting points. Do not pay a palladium premium to solve a hydrogen problem that comes from a wet shielding line or an over-negative protection system: the money belongs in the process control instead.

Nickel and zirconium neighbours

Hydrogen damage is not unique to titanium. Nickel alloys and zirconium have their own hydride and embrittlement behaviours, and the analytical method family overlaps. ASTM E1447 is widely described as applying to reactive metals including both titanium and zirconium, which is convenient for a plant that runs both material systems but also a source of crossed specifications. For a sense of how another alloy family handles delivery condition and heat treatment, the Monel K-500 age-hardening guide is a useful parallel: in both cases the certificate identifies the material, while a separate test is needed to identify the condition.

How to Write Hydrogen Into a Purchase Order

Most hydrogen disputes are won or lost in the wording of the PO, not in the laboratory. A clause that says "low hydrogen" or "per ASTM B265" is unenforceable after fabrication because it names neither a number, nor a method, nor a moment in time. The template below is deliberately verbose. Copy it, delete the lines that do not apply, and keep the edition references current with your own standards library.

Hydrogen control clause (copy and edit)
Grade and product standard (with edition): ________
Maximum hydrogen at delivery: ____ ppm by weight, measured per ASTM E1447 (edition ____) or agreed equivalent ____
Maximum hydrogen increase after all pickling / welding / heat treatment: ____ ppm, measured on the same heat with identical sample preparation
Sampling: location ____ (surface / mid-thickness / weld metal / HAZ), sample mass ____, cutting method (low-heat, no coolant residue), cleaning and drying procedure ____
Laboratory report must state: method and edition, calibration reference material, blank value, duplicate results, and confirmation that the result lies inside the validated quantitative range
Re-test hold point: after ____ (final pickling / final weld / final heat treatment), witness status ____
Prohibited without written approval: unapproved acid pickling, electroplating, chemical milling, steel-style low-temperature bake-out, and any furnace cycle with an uncontrolled atmosphere
If vacuum degassing is performed: furnace batch record, temperature and vacuum log, soak time, section thickness, and a post-cycle hydrogen re-test are required
Documentation: EN 10204 3.1 tied to the parent heat, OES chemistry, tensile results, hydrogen report, and a fabrication route summary attached to the same dossier

Three lines in that template do most of the work. The increment line turns a vague expectation into a measurable change caused by fabrication. The hold point line names the moment after which the number must still be true. The validated range line prevents a laboratory from reporting a single-digit ppm value using an unvalidated routine method. Together they convert hydrogen from a material property into an acceptance process, which is what a QA department can actually audit.

There is also a commercial benefit to writing the clause this way. A supplier who holds stock can quote an incoming value from the certificate immediately, but can only quote a post-fabrication value if the shop controls its pickling and shielding. Asking for both separates a stockist with a controlled process from a trader forwarding certificates. We are comfortable with the second question: we cut non-standard sizes from held stock, work from customer drawings, and can tie the hydrogen report to the same heat as the dimensional and tensile records. For document structure, the 3.1 certificate page and the arrival acceptance guide show the rest of the dossier this clause has to sit inside.

One caution on standards editions. ASTM B265, B338 and B348 are revised, and implant standards F67 and F136 have their own composition tables. A PO that names a standard without an edition can be satisfied against whatever version the supplier holds, which may not be the one your design basis used. The same applies to E1447: name the edition, and require the report to say which detection mode (thermal conductivity or infrared) was used. Where a Chinese supply chain is involved, GB/T 4698 series methods and GB/T 3620.1 grade limits may appear on the certificate; reconcile the method and units explicitly instead of assuming the two systems report the same quantity.

Testing: What ASTM E1447 Measures, and What It Cannot

ASTM E1447 determines total hydrogen in reactive metals by inert gas fusion: a small clean sample is melted in a graphite crucible under helium or argon, the released hydrogen is carried to a thermal conductivity or infrared detector, and the result is reported in ppm or weight percent. It is fast, widely accepted and the right tool for a conformance check. It is also a bulk total on a gram-scale sample, which means it cannot see where the hydrogen is, whether it has formed hydrides, or whether a surface layer is charged while the core is clean.

The four questions a hydrogen report has to answer

Was the sample representative? A chip taken from the middle of a 40 mm plate says nothing about a pickled surface layer. Was the sample clean? Oil, cutting fluid, moisture and oxide on chips can bias a result upward, which is why low-heat cutting, degreasing, drying and a documented preparation method are part of the test rather than housekeeping. Was the value inside the validated range? Published scope statements place routine quantitative E1447 work at roughly 9 to 320 mg/kg; below that, a laboratory needs a documented validation (often referencing ASTM E2857 for method extension) rather than a routine certificate line. Was there a blank and a calibration? A report without blank correction, a certified reference material and duplicate results is a number, not a measurement.

What the report cannot tell you

E1447 cannot distinguish diffusible hydrogen from hydrogen already tied up in hydrides, cannot map hydride orientation, and cannot predict impact toughness at a given temperature. If the failure mode under investigation is delayed cracking or a ductility loss, the hydrogen number is one input alongside metallography for hydride platelets, a slow-bend or impact test on material from the same location, and a fractographic review. ISO 17081 is sometimes cited for hydrogen permeation and uptake measurements; that is a different experiment answering a different question, and it should not be substituted for a conformance determination without agreement. Where a buyer asks for "a hydrogen test" without saying which of these they mean, the laboratory will quite reasonably choose the cheapest one.

Reading a disagreement between two laboratories

When an MTC value and a post-fabrication value disagree by a margin that cannot be explained by the process, check the sample preparation before accusing the material. Surface contamination, moisture on chips, an inadequate blank, a different edition of the method, a result outside the validated range, and a sample taken from a different location are all more common than a genuinely segregated heat. Keep a retained sample from the same lot, ask both laboratories for their preparation procedure and calibration records, and repeat with a third laboratory using a single agreed sampling plan. This is the same discipline used for disputed chemistry in other alloy systems, and it is why the PO template above makes sample preparation a contractual line rather than a footnote.

For context on where a hydrogen report fits into the wider quality file, the WPS and NDT guide covers weld qualification and the arrival acceptance guide covers the receiving checklist. Neither includes hydrogen by default, which is precisely why the clause has to be added explicitly on orders where hydrogen is a design concern.

Vacuum Annealing and Hydrogen Removal: What Can Actually Be Undone

Hydrogen can be removed from titanium, but not by baking it in an ordinary oven. Degassing relies on a chemical potential gradient: low hydrogen partial pressure above a clean surface, enough temperature for diffusion, and enough time for the section thickness. Industry descriptions of vacuum degassing for titanium commonly place the cycle in the 600 to 800 °C range under high vacuum, with time set by section size. We quote that as an orientation range, not a recipe: the parameters have to be qualified for the specific alloy, part geometry and furnace, and the result has to be proven by a hydrogen determination afterwards. A cycle copied from a blog, without a vacuum log and a post-cycle test, is an anneal, not a degas.

Why the steel bake-out fails here

Plated high-strength steel is commonly baked at roughly 190 to 230 °C to relieve diffusible hydrogen. Applying the same instruction to titanium is a specification error with two consequences. At that temperature the titanium hydrides largely remain, diffusion is slow relative to the section, and there is no low-partial-pressure sink for hydrogen to leave into; an air furnace may also grow an oxygen-rich surface layer while doing it. Worse, a quick bend test performed while the part is still warm can look improved because hydrides have temporarily dissolved back into solution. The part then re-precipitates them during cooling. Any PO that says "bake per standard plating practice" on a titanium line should be rejected at review, not negotiated on the shop floor.

What a real degas cycle costs you

A high-vacuum, high-temperature cycle is not free of side effects. It can change the microstructure and therefore the mechanical properties that the original product standard certified; it can distort thin or already-machined sections; it can alter surface finish and dimensional tolerance; and on a finished assembly it is usually impossible because of dissimilar components, seals or welds. For critical parts, the cycle needs its own qualification: pre- and post-cycle hydrogen, tensile or hardness confirmation, dimensional re-inspection, and a furnace batch record tied to the heat. If the part has already cracked, degassing does not repair the crack. In many cases replacement from controlled stock is cheaper and faster than attempting recovery.

Prevention beats recovery

The practical hierarchy is: buy material with a hydrogen value comfortably below the ceiling; keep pickling ratio, time and rinsing under control; keep shielding gas and filler dry; set and log cathodic potential; avoid unapproved plating and chemical milling; and re-test after the last hydrogen-generating step. Every one of those controls is cheaper than a vacuum cycle and does not put the delivered mechanical properties at risk. On our side, this means that when an enquiry flags hydrogen as a design concern, we quote against a specific heat with its certificate value, cut to the drawing from held stock, and state which thermal or chemical operations are in scope before any work starts. What we will not do is apply an unqualified furnace cycle to make a disputed number look better.

For the surrounding documentation, the 3.1 certificate guide explains heat traceability and the pickling guide explains the process controls that prevent the problem in the first place. A hydrogen report is only as useful as the route summary that accompanies it.

Limitations and Failure Conditions: Eight Ways a Compliant Part Becomes Brittle

These eight conditions are organised by operation and service situation rather than by metallurgical category, because that is how they arrive in an RFQ or a failure report. Use them as a review checklist against the fabrication route and the process data sheet before the purchase order is issued. None of them is fixed by choosing a stronger grade, and every one of them is cheaper to close at specification stage than during a shutdown.

Acid pickling with the wrong nitric-to-HF ratio

The classic shop-floor route. Excess hydrofluoric acid, excessive immersion, a warm bath and incomplete rinsing generate atomic hydrogen at a freshly active surface. The part leaves the tank bright and apparently clean, passes a visual and ferroxyl check, and fails a slow bend or impact test days later. The correct controls are bath ratio rather than concentration alone, time and temperature limits, a qualified inhibitor, thorough rinsing and a before-and-after hydrogen increment on the same heat. Free-iron contamination and hydrogen charging are separate failure paths with separate tests, which is why a clean iron test does not close this item. The pickling and iron contamination guide covers the cleaning standard; the hydrogen increment belongs on the PO.

Welding with wet shielding gas or contaminated filler

Titanium above roughly 400 °C absorbs interstitials readily. Moisture in a gas hose, condensation on cold filler wire, oil or cutting fluid in a machined groove, and humid storage all feed hydrogen into the weld pool and heat-affected zone. The bead may show an acceptable silver or light-straw colour, because colour tracks oxygen contamination, not hydrogen. Require high-purity argon with a logged dew point, dry and cleaned consumables, three-zone shielding and a post-weld hydrogen determination on critical joints. The GTAW shielding guide and the early weld failure review show the protection discipline and the field symptoms; procedure qualification is covered separately in the WPS and NDT guide.

Cathodic protection driven beyond its potential window

Protecting a steel structure is correct; driving the potential too negative turns the titanium in the same circuit into a hydrogen-evolving cathode. Absorbed hydrogen precipitates as hydrides and ductility falls with no corrosion to photograph. Published negative limits disagree and are meaningless until the reference electrode is named, so the control is procedural: name Ag/AgCl, SCE or CSE on the drawing, set the window from the project CP specification, place reference cells at the titanium surface, log the trend, and run an ASTM E1447 determination if over-protection is suspected. A number copied from a datasheet without its reference electrode is not a limit. The Gr2 vs Gr7 seawater guide covers the chloride service where these systems are most common.

Titanium cathode plates in electrowinning cells

Here hydrogen evolution is the intended process, not an upset. A titanium cathode plate sits polarised in electrolyte for long campaigns, absorbs hydrogen at the working face, and gradually loses ductility until it cracks or distorts during stripping. This is a consumable-life problem, not a material defect: specify a replacement and inspection interval, track plate campaigns by heat, and sample hydrogen at end of life to validate the interval. The titanium cathode plate guide describes the copper-titanium composite lug hardware used in this duty. Do not attempt to solve a spent cathode plate with a low-temperature bake-out; replace it.

Hot reducing acids and hydrogen-bearing process streams

Where the passive film is attacked, the corrosion reaction itself evolves hydrogen at the metal surface and a fraction is absorbed. The risk rises with temperature, acid concentration, deaeration and the absence of oxidising ions, and it is not captured by an iso-corrosion curve that reports only metal loss. Grade 7 extends the useful envelope in hydrochloric acid but does not remove the hydrogen balance from the design. Start with the concentration and temperature limits in the HCl grade-limit guide, confirm with a coupon or loop test in the real stream, and add a hydrogen assessment where the duty is hot, reducing and continuous. The acid-leaching plate page shows the plant context.

Uncontrolled furnace atmosphere and hot working

Forging, rolling, stress relief and annealing in a wet or hydrogen-bearing atmosphere add hydrogen before the part ever reaches a machine shop. The resulting value may still be inside the product-standard ceiling while the entire fabrication allowance has already been consumed. Require vacuum or controlled inert atmospheres, dew-point logging and furnace batch records tied to the heat number, and treat any undocumented thermal cycle as a route change that has to be re-qualified. Where a supplier offers a low-cost re-anneal to fix a dimensional problem, ask for the atmosphere record before accepting it. The ASTM B348 bar guide covers the incoming bar specification this risk attaches to.

Machining, grinding and coolant or cleaning residues

Cutting fluid left in a groove, a chlorine- or sulfur-bearing extreme-pressure additive, a grinding burn, and solvent residues that are driven off during a later weld can all supply hydrogen or leave reactive contamination on the surface. This is a process-hygiene item rather than a grade-selection item. Specify low-heat cutting for test samples, degrease and dry before any thermal operation, and keep titanium machining tools and fixtures segregated from carbon steel to avoid the separate free-iron problem. The TC4 CNC blank page gives practical starting windows and coolant guidance for machined bar; the passivated fastener page covers the cleaned product side. Neither substitutes for a post-process hydrogen check on a critical part.

Mixed-metal assemblies, stray current and graphite coupling

Galvanic coupling to a more noble material such as graphite can make titanium the anode, while stray direct current from nearby equipment can polarise it without anyone calling it cathodic protection. Both change the surface electrochemistry in ways that are invisible on a certificate. Audit the assembly for conductive gaskets, bonding straps, earthing paths and nearby impressed-current systems, and measure insulation resistance after final assembly. Where polarisation cannot be excluded, add a named reference electrode, a potential window and periodic hydrogen testing to the maintenance plan. Our titanium coil page and the fastener assembly guide show the components most often involved in these retrofitted assemblies.

Application Map: Which Orders Actually Need a Hydrogen Clause

Where the standard product limit is enough: general chemical-process plate and tube for ambient, non-reducing duty; structural bar and machined parts with no pickling, no welding and no polarisation; and stock cut-to-size orders that leave our warehouse in the mill condition. For these, a certificate showing hydrogen inside the applicable product standard is a reasonable stopping point. Adding a low-hydrogen clause to every order raises cost without lowering risk, and it trains suppliers to issue boilerplate reports that nobody reads.

Where a full hydrogen clause is warranted: implant and aerospace chains, any part that is pickled or chemically milled, critical welds in pressure equipment, titanium in a protected or electrolytic circuit, hot reducing-acid service, thin sections and fasteners loaded in tension, and any component whose failure would be a safety event. For these, the incoming maximum, the permitted increment, the sampling plan and the post-process hold point all belong in the contract. The WPS and NDT qualification guide and the arrival acceptance guide show the surrounding weld and receiving documentation; hydrogen is the line those documents usually omit.

On cost and schedule: a hydrogen determination is a small, fast laboratory test on a prepared sample, while a disputed batch, a re-melt or a scrapped welded assembly is not. The expensive part is rarely the test; it is defining the sampling and hold point early enough that the test can be taken before the part disappears into an assembly. We are a Baoji (China Titanium Valley) stockist-supplier holding roughly 800 tonnes of titanium and nickel stock, including about 400 tonnes of plate and about 200 tonnes of bar and ingot. We cut non-standard sizes, work from customer drawings, certify to ASTM / ASME / GB, export-pack, and dispatch confirmed in-stock orders the same day. What we will not do is quote a low-hydrogen requirement without asking which fabrication steps follow delivery, because that answer determines whether the clause is a certificate line or an acceptance test.

Summary: Choose a Standard Limit if, Choose a Low-Hydrogen Clause if, Choose Replacement if

Choose a plain product-standard hydrogen limit if:

• The material ships in the mill condition with no pickling, welding, plating or polarisation after delivery.
• Service is ambient, non-reducing, and the part is not impact-critical or loaded in sustained tension.
• The applicable standard edition is named on the PO and the certificate reports a numeric hydrogen value.
• A normal receiving inspection against the 3.1 certificate is the agreed acceptance point.

Choose a full low-hydrogen specification with a re-test hold point if:

• Any acid pickling, chemical milling, welding, plating or high-temperature cycle follows delivery.
• The component sits in a cathodically protected, electrolytic or stray-current environment.
• Duty is a hot reducing acid, a hydrogen-bearing gas stream, or an implant or aerospace chain.
• You can name the method and edition, the sampling location and preparation, the permitted increment, and the post-process witness point.

Choose replacement from controlled stock rather than a recovery cycle if:

• Cracking, severe distortion or a confirmed hydride network is already present.
• The part is a finished assembly containing seals, dissimilar metals or machined tolerances that a vacuum cycle would disturb.
• No furnace batch record, vacuum log or pre-cycle hydrogen value exists to qualify a degas cycle against.
• Schedule matters more than the salvage value of the component.

The whole subject reduces to one sentence: the certificate describes the metal the mill shipped, and hydrogen is decided by everything that happens after that. When you send an enquiry, the six most useful inputs are the grade and standard edition, the numeric hydrogen value on the offered heat, the full fabrication route including any pickling or welding, the service medium and maximum metal temperature, whether cathodic protection or electrolysis is present and with which reference electrode, and the point at which acceptance will be measured. Those six lines decide whether you need a certificate line or a test plan, and they remove the round trip that delays shutdown orders. We hold roughly 800 tonnes of titanium stock in Baoji (China Titanium Valley), cut to non-standard sizes, work from customer drawings, certify to ASTM / ASME / GB, export-pack, and dispatch confirmed in-stock orders the same day. Send us the fabrication route with the grade and we will confirm the heat value, the realistic test plan and the date together. Request a quote →

Frequently Asked Questions

Is a compliant mill certificate enough to rule out hydrogen embrittlement?

No. An EN 10204 3.1 certificate reports the chemistry and tests tied to the parent heat at the time of manufacture, including hydrogen if the product standard requires it. It says nothing about hydrogen added b