The subject nature of hydrogen embrittlement is technical and complex; the contents of this booklet are intended for the metal finisher. Some simplification of metallurgical concepts have been made in order for an easier understanding of the subject matter for the intended reader.

The information contained herein is considered accurate, general information on the subject of hydrogen embrittlement as it exists in the metallurgical sciences today. Some contractor or agency requirements may differ or vary from the parameters discussed within. If areas of conflict arise always follow the guidelines set forth by your contracting agency or customer.

Hydrogen embrittlement is a metallurgical phenomenon that occurs in many different metals; however, high strength steel by far seems to have the highest sensitivity to embrittlement. In addition, high strength steel is more commonly seen by the metal finishing industry than other metals or alloys that have been known to suffer from hydrogen effects. Hence, hydrogen embrittlement of high strength steels dominates the file records of aircraft/aerospace components that have failed over these last 40 years. Therefore, we have chosen to limit this discussion to steels only.

Almost 75% of all the elements known to exist in the universe are metals, so it is easy to see why man has chosen to utilize metals so extensively in our civilization. One of the most important properties of metals is ductility. Ductility can be more commonly understood as the ability to deform under stress. Although this deformation or stretching under stress can sometimes cause problems in itself, it is still one of the advantages of metals compared to other structural materials such as ceramics, concrete, stone, etc.

Hydrogen embrittlement as a pure theoretical phenomenon is still argued about in the scientific community today. Suffice it to say that, as the name implies, a metallurgical interaction occurs between atomic hydrogen and the ferrous metallic atomic structure, and the ability of the steel to deform or stretch under load is inhibited. Therefore the steel becomes "brittle" under stress or load. In general terms, as the strength of the steel goes up, so does its susceptibility to hydrogen embrittlement.

To the mechanical metallurgist, the primary attribute of most metals are their tensile properties, or the ability to sustain stress in tension. When a metallurgist analyzes metals he often performs a tensile test in order to characterize and understand many different properties of the metal alloy under investigation. In Figure A you will see a simplified representation of what happens when a precise tensile test bar is pulled under a tension load.

On the vertical axis is the load or stress with units of pounds/psi. On the horizontal axes, the elongation or strain that the test bar undergoes while it is being pulled can be seen. Elongation and strain are similar words for the attribute of ductility. As you follow the curve upwards and to the right, as the load or stress increases on the test bar, the elongation or deformation on the test bar increases also.

You will note two distinct areas on the curve. The point where the curve stops being a straight line is called elastic deformation, and it is in this area that metallic springs operate. Aside from the effects of fatigue, a metal part can be cycled up and down in this region millions of times, always returning to the same length or dimensions after unloading. During the tensile test, any stresses or loads placed on the test bar in excess of this elastic limit will force the metal bar to permanently stretch or deform. As the test continues, you can see that the ductility or stretching is very pronounced out to the point of fracture. In a true engineering stress/strain test, the actual stress also increases up to the point of fracture, but for our purposes here, it is only important to understand the ability of metals to not only elastically deform, but plastically or permanently deform.

Now you may ask, "What does all this have to do with hydrogen embrittlement?" In high strength steel alloys, the presence of hydrogen tends to block ductility. Metallurgists have spent decades researching how hydrogen can stop or inhibit a metal's ability to deform. It is sufficient to say that, as shown in Figure B, hydrogen has limited the same metal as shown in Fig. A to a much lower elongation. Since we now know that there is a predictable relationship between stress/load and ductility/elongation, this embrittled sample will not take the same high stress/load as before without breaking.

The theory stated above is the traditional or classic explanation of hydrogen embrittlement, where large quantities of hydrogen have infiltrated into the steel. However, of far more concern today is embrittlement from very small quantities of hydrogen where traditional loss-of-ductility bend tests will not detect the condition. This atomic level embrittlement manifests itself at levels as low as 10 ppm of hydrogen. Although difficult to comprehend, numerous documented cases of embrittlement failures with hydrogen levels this low are known. This type of embrittlement occurs when hydrogen is concentrated or absorbed in certain areas of metallurgical instability. This concentrating action occurs via either residual or applied stress, which tends to "sweep" through the atomic structure, moving the infiltrated hydrogen atoms along with it. These concentrated areas of atomic hydrogen can coalesce into molecular type hydrogen, resulting in the formation of high localized partial pressures of the actual gas.

Other theories show the hydrogen to act as a grain boundary surfactant that reduces the surface film energies at the grain boundaries, promoting dislocation slip movement, and eventually microcracks within the steel. These microcracks tend to grow quite rapidly upon formation, since the Kt factor or stress intensity factor at the crack tip is astronomically high. Fracture via this type of embrittlement manifests itself by not only ductility loss, but more importantly by the actual loss, via microcracking, of load supporting or cross sectional areas within the part. For example, a part may start out with one square inch of cross sectional area on the outside, but at time of fracture an actual load bearing area 10-20% lower than this may be present.

The facts are plain. The hydrogen has inhibited the metal's ability to deform, and as a result the metal will break or fracture at a much lower load or stress than anticipated. It is this lower breaking strength that makes hydrogen embrittlement so detrimental in nature. Design engineers rely on the capability of metals and alloys to carry the load or stress for which they design. However, after the part is no longer a "blueprint" but has been manufactured, it becomes quite sensitive to the processing that takes place in the metal finishing industry.

The purpose of this document is this: to put down black and white some of the things that you as a metal finisher should and should not do to high strength steels to avoid hydrogen embrittlement. Although most of the problems the world has seen with hydrogen embrittlement have occured with aircraft/aerospace parts, the part doesn't have to "fly" in order to "die." Hundreds of human lives have been lost over the years because of hydrogen embrittlement. The effects of hydrogen on metals is serious, deadly serious.

Hydrogen can enter steel from a multitude of places, starting from the original steel making operation. Hydrogen can also enter steel from subsequent casting and forging operations. Even grinding operations after final heat treatment can induce hydrogen absorption especially if sparking occurs in a moist environment. However, for the metal finisher, the two most important sources of hydrogen damage occur from acid type cleaning, and of course, actual electroplating operations.

In Figure C, a relative scale of hydrogen absorption is shown for some common acids. Also you can see that when current is applied, as in cathodic acid pickling, large quantities of hydrogen are liberated from the work piece.

Hydrogen is released by the reaction of any active metal with an acid. Typical examples of this are HCl or H₂SO₄ with steel.

Numerous methods have been developed over the years to test for the presence of hydrogen damage. An actual test for hydrogen contact is usually beneficial, as documented embrittlement of steels has occurred in samples that tested as low as a few ppm. The mobility of hydrogen within the steel, especially at elevated temperature is great; consequently analysis for hydrogen content results in just a "background" count. Hydrogen damage behaves in similar fashion to fatigue damage, in that hydrogen tends to migrate to areas of microstructural instability or defects within the steel, concentrating itself to the point that micro-crack development begins and culminates in rapid, catastrophic fracture. Fatigue damage in metallic parts is somewhat similar, as stress distribution within a loaded part is usually not uniform with the resulting stress concentrations, micro-crack developments, etc.

The majority of tests developed over the years have a common theme; a mechanical test either with stressed bars or rings. The net results is a sustained load over time, utilizing either true tensile loads or bending loads. Since bend tests are a combination of both tension and compression, this type of mechanical test does not tend to be as useful as true uniaxial tensile test. We stated that the test is a sustained load type. This means that the sample is held under a constant load for an extended period of time. The metallurgical reason for this is that hydrogen damage is usually delayed in nature, resulting from the introduction and migration of hydrogen through the steel over time. A physical transport of hydrogen over measurable distances occurs, sometimes approaching fractions of an inch. This hydrogen transport phenomenon is not simply an atomic position change.

Typical test sample configurations include the presence of a sharp notch within the test section or gauge length of the bar. This notch serves a number of purposes.

1. It concentrates stresses in one local place so we can "force" the failure to a specific, repetitive place.
2. It simulates numerous metallurgical and engineering conditions that promotes embrittlement reactions.
3. It accelerates the time to fracture if embrittlement has occured.

It is interesting to note once again this factor called time. Some metallurgists insist that after a part has been subjected to excessive amounts of hydrogen, that the embrittlement condition has already occurred. Others insist that hydrogen in and of itself does not necessarily mean damage has occurred.

There are three essential ingredients for actual damage. These are 1] hydrogen, 2] stress, and 3] time. If sufficient hydrogen has infiltrated into the steel, and if sufficient stress has been applied to the part, then it may just be a matter of time until fracture occurs. The stress we talk about is actually a stress gradient within the steel, causing a sweeping action of hydrogen atoms towards areas of metallurgical instability. Stress can be either residual or applied. The residual stress is stress already within the steel, resulting from heat treatment, cold work, or machining/grinding operations. Heat treat stress relief operations are beneficial in lowering this residual element; however, for our purposes here, a complete removal of residuals is not possible due to temperature constraints. Most of us have heard stories, or even experiencing "popping" of embrittled parts while on the shelf, were high residual stresses have forced failure while the parts are in storage. Applied stresses are of course the loads applied to actual parts during test or service.

Another type of test associated with hydrogen embrittlement is the Lawrence meter or gauge. Although this is not a mechanical type test, it is useful in establishment trends for specific metal finishing operations. The concept is also known as the hydrogen porosity test. Regardless of nomenclature, the principle behind it is that a steel vacuum tube is prepared so a zero baseline of hydrogen content is established. Then the tube is placed into the desired plating tank and electroplated per normal parameters. Then the tube is taken back to the laboratory where, under controlled conditions, it is baked out. During this bake out operation, the electrical properties through the tube or probe are measured and plotted over time. The net result of this type of test is a measure of the efficiency of the baking operation as compared to a baseline, unplated part. Highly porous metallic coatings will "out gas" more efficiently, resulting in lower residual hydrogen within the parts. Many variations on this theme have been developed over these last 35 years resulting in important information for the metal finishing community. However, it can be said that all Lawerence or Porosity gauge concepts must be correlated back to an actual sustained load mechanical test. Structural aircraft components do not perform their function because of electrical properties through a vacuum tube; but rather due to metallic alloys shouldering loads and stresses of flight. Optimal benefits are obtained from a combination of Lawrence tests correlated to mechanical sustained load tests.

1. Lowering or eliminating hydrogen generation.
2. Removing damage levels of hydrogen after metal finishing.

Preventative maintenance is always the best route, and because of that fact metallurgists have spent decades researching ways of preventing or minimizing hydrogen generation in the first place. Specific methods are detailed in the following pages under each specific plating process. In general, however, electrolytic and non-electrolytic processes that maximize plating efficiency and minimize hydrogen generation at the cathode are pursued. Examples of this are Ti-Cadmium, non cyanide cad, electroless nickel, and vapor deposited metallic coatings to name a few.

The second method of hydrogen control, that of removal after the fact, is an equally important tool in the control of hydrogen embrittlement. Embrittlement relief, or baking as it is commonly referred to, is a powerful method in eliminating hydrogen before embrittlement damage can occur. It is both good news and bad news that the concept of baking is so important.

The good news is that given the proper plating properties, baking is an incredibly efficient method for "relief of embrittlement." The metallurgical process of hydrogen diffusion and transport out of the part is simple, well understood and repetitive. The actual costs to operate a baking oven are low compared to all other operations and functions in a metal finishing facility. Good baking habits are probably the most important step in the control of hydrogen embrittlement.

However, the bad news is that sometimes we tend to forget the "preventative measures" needed in the first place. It is easy to fall back and assume that good baking will correct bad plating. There are many situations where poor plating practices can very well prevent efficient bake out of hydrogen afterwards. A sound combination of process control and good baking habits is always the best approach. Several large aircraft/aerospace contractors do not require periodic embrittlement testing because of their tight plating parameters and stringent baking requirements.

Almost 71% of documented aircraft embrittlement failures over the last 30 years have been attributed to the baking operation, that being:

• A missed or omitted bake
• Extended delay from plating to baking
• Insufficient baking temperature
• Short baking times

Baking or embrittlement relief quite simply is a diffusion process. Diffusion is defined as the movement of atoms within a solution. This solution can be either a gas, liquid, or solid. For embrittlement relief, the solution is of course a solid, or the steel part being baked. The essential characteristics for diffusion are energy or heat, time, and the diffusion or concentration gradient.

For embrittlement relief, temperature is the most important characteristic since it defines the heat energy or driving force for the exiting hydrogen.

Time is the second most important characteristic during baking. The longer the time at temperature, the more hydrogen diffuses through the steel part. Diffusion of hydrogen during baking does not involve any chemical reactions whatsoever, only the movement of hydrogen atoms through the steel atomic structure. Therefore, there is no beginning or end of any type of chemical reaction. Some have thought in the past that the term baking meant that it involved some type of "cooking" so only short periods of time were necessary for the "cake to get baked." This is far from the truth; the longer a plated part is baked, the better the job of hydrogen removal. The U.S. Navy completed tests several years ago showing that for bright cadmium plated parts, an optimal bake cycle of 96 hours minimum is necessary. Research work performed about 25 years ago by Batelle Laboratories showed baking cycles of 100 hours were best in removing hydrogen from plated parts. Time requirements for baking are usually given as minimums in most specifications such that baking time extensions are allowed virtually across the board.

The last characteristic for the diffusion of hydrogen from steel is the concentration gradient. This is in essence the potential energy within the steel part that can be attributed to the presence of hydrogen. One can see that as the localized concentration of hydrogen goes up, so does its desire, or "potential," for dispersion throughout the part. There has been significant work performed through the years in investigating the concept of a "threshold" level of hydrogen; above which eventually hydrogen damage occurs and below which no damage occurs. The majority of this research work has been confirmed, and it is probably safe to say that the diffusion of small amounts of hydrogen throughout and away from the steel part contributes to the efficient baking of the component; i.e. the term "bake-out cycle" in essence involves both out-gassing and diffusion within the steel.

To review again the three essential steps in baking or hydrogen diffusion, we need, first of all, energy or heat. This is measured as temperature. Secondly, we need time, of which sufficiently long periods are required to allow the migrating hydrogen to move. And thirdly, we need a concentration gradient, of which hydrogen concentrations move quickly throughout the atomic structure.

As was said previously, energy/heat/temperature is the most important parameter. Never short cut the baking temperature. Normally most specifications call for a 375 degree F. bake temperature plus or minus 25 degrees F. If the baking oven is set at 350 degrees F. with the thought that money is saved on electricity, one actually removes about half the hydrogen possible by baking at 400 degrees F. The actual mathematical relationship between temperature and diffusion rate is exponential; a little extra heat goes a long way in efficient baking. A word of caution; excessive temperatures can cause metallurgical changes in the alloy parts. NEVER exceed the allowed baking temperature range. Over-tempering or softening of the steel can occur especially if a carburized, induction hardened or similar treated surface is present on the part.

A final but very important point for the metal finisher: Know your parts, i.e. what it the strength/hardness level of the steel part you are plating? Don't assume it's a soft, low strength steel. Ask questions of your customer. Make sure you know any specific or unusual aspects about your parts.

General Comments:

Cyanide based cadmium processes tend to liberate large quantities of hydrogen at the cathode. Non cyanide baths tend to have higher cathode efficiency and therefore less hydrogen evolution. Brighteners tend to make the coating less porous and therefore more difficult for hydrogen escape during baking.

• Non cyanide baths
• Plating thickness

  Avoid over-plating both in thickness and surface area coverage.

• Low brightener contents

  This produces a more porous coating for easier bake-out. Many companies have arrived at the optimum brightener content by performing salt spray tests vs. brightener content. By starting with no brightener and then working up to successivly larger amounts, a relationship between acceptable salt spray corrosion performance and hydrogen embrittlement performance can be established. The best balance appears to be with brightener contents just at the point where salt spray corrosion results start coming up marginal. This establishes the brightener content at the minimum level to satisfy corrosion properties, but yet keeps the potential hydrogen situation at a minimum.

• High cathode efficiency/high current density

  Lower hydrogen damage occurs with high current density at the cathode. As an example, ASTM F519 ANNEX shows two types of cadmium baths. The optimum bath is non-brightened, with a current density of 60 amps/square foot. A highly embrittling bath has brighteners along with a low current density of 10 amps/square foot and a long plating time.

• Fast transfer times into the baking oven

  We recommend a one hour maximum delay.

• Longer bake times at the maximum permissable temperatures.

  We recommend 23 hours minimum.

• Surface preparation - non acid methods best

• Cyanide baths



• Over-plating in thickness and area

  Thick plating not only makes the bake-out more difficult, but also generates more hydrogen during the actual plating operation due to the longer plating time involved.

• Low current density at the cathode

  This results in longer plating times.

• Acid pickling prior to plating

  This is a hydrogen generating process in itself, and its effects can be additive on top of the actual plating generation of hydrogen.

• Cathodic acid cleaning/pickling

  A real problem!

• Baking delays into the oven



• Short baking times and low oven temperatures

  The U.S. Navy has shown that three hour baking times can actually be counter-productive in the out-gassing of hydrogen. They actually recommend a 96 hour minimum baking cycle.

• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

CHROME Top

Due to its low cathode efficiency, chrome plating tends to be one of the worst processes in the generation of hydrogen during plating. However, due to the somewhat porous nature of the coating, it is one of the easiest to bake out during the embrittlement relief cycle.

• Proper bath chemistry (CrO₃ / SO₄ ratio)



• Proper current density and bath temperature



• Avoid over-plating both in thickness and surface area coverage



• Surface preparation - non acid methods are best



• Fast transfer times into baking oven

  We recommend a one hour maximum baking delay.




• Longer bake times at the maximum permissible temperature

  We recommend 23 hours minimum.

• Over-plating in thickness and area



• Acid pickling cleaning prior to plating

  This is a hydrogen generating process in itself and its effects can be additive on top of the actual plating generation of hydrogen.

• Baking delays into oven



• Short baking times and low oven temperature



• Some thin dense chrome processes i.e., less porous coating, resulting in more difficult hydrogen removal during baking



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

NICKEL Top

Normally, cathode efficiency ranges from 93 to 97% for most nickel plate processes. Due to this, low hydrogen evolution during the process can be expected. However, even minute amounts of hydrogen have been shown to initiate embrittlement failures; therefore, care in processing should be exercised in like fashion to other more embrittling processes.

• Avoid over-plating in both thickness and surface area coverage



• Non acid surface preparation best



• Fast transfer times into baking oven

  We recommend a one hour maximum delay

• Longer bake times at the maximum permissable temperature

• Over-plating in thickness and area



• Acid pickling and cleaning
(See notes on Chrome Plate)

• Bake oven transfer delays, short baking times and low oven temperatures



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

SILVER
Top

Although silver plating has generally been applied in the past to decorative or electronic applications, in recent years newer applications on structural components and bearings have been seen. These type components can be very high strength-high hardness steels and therefore more prone to embrittlement. Generally higher porosity, non-brightened coatings minimize embrittlement damage.

• Non acid surface preparation



• Avoid over-plating both in thickness and surface area.



• Low brightener contents

  (See comments for cadmium plate)




• Fast transfer times into the bake oven, and longer bake times at maximum permissable temperature.

• Over-plating



• Acid pickling/cleaning prior to plating



• Bad baking habits, i.e. transfer delays into bake, short bake times, temperatures too low



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

ZINC
Top

Hydrogen embrittlement is more likely to occur in cyanide zinc plating than in the plating of any other common coating including cadmium. Cathode efficiency is usually less than cyanide cadmium.

• Non cyanide baths



• Optimum cathode efficiency

  This includes correct bath temperature, cyanide to metal ratio, control of zinc metal and NaOH.

• Low brightener content

  (See comments for cadmium plate)

• Good baking habits

  Short transfer times into oven, long bake times at maximum permissable temperature

• Cyanide baths



• Poor control of cathode efficiency



• Highly brightened bath



• Poor baking habits



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

ELECTROLESS NICKEL
Top

Electroless nickel, or auto catalytic nickel plating, is truly one of the future growth areas for metal finishing. By its nature, it tends to be far less embrittling than electrolytic processes.

• Surface preparation

  Non acid methods are best

• Alkaline bath compositions



• Good baking habits

  Fast transfers into bake ovens, long bake times at maximum permissible temperature

• HCl acid pickle/clean



• Poor baking habits



• Over aged bath, depositing a coating in residual tension



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

PHOSPHATE COATINGS
Top

GENERAL COMMENTS:

Phosphate treatments generally do not present major embrittlement problems to the metal finishing industry. This is mainly due to the rather dilute nature of the phosphoric acid solutions used during the process and the fact that this is not an electrolytic or galvanic process. Hence only small quantities of H₂ are exuded during the process.

However, documented cases of embrittlement from phosphatizing are known, with the cause invariably traced to the pre-phosphatizing cleaning operation (concentrated acid pickle).

GOOD:

• Careful pre-coating surface preparation - minimize times in any concentrated acidic cleaners

  Abrasive blasting is best

• Elevated temperature bake

  This is recommended over the room temperature age option

POTENTIAL PROBLEMS:

• Concentrated acid pickles for pre-cleaning



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

COPPER PLATE
Top

GENERAL COMMENTS:

The majority of copper plating for functional uses is accomplished in either the cyanide alkaline bath or the pyrophosphate alkaline bath. Although both bath types are highly alkaline in pH, hydrogen is still evolved during the process. Cathode efficiency tends to be high, especially in the high concentration sodium and potassium cyanide baths. Since deposition rates are high, time is minimized in the bath and therefore the tendency for hydrogen embrittlement is lowered. Few hydrogen embrittlement failures have been documented due to copper plating.

GOOD:

• Abrasive blast type cleaning



• Fast transfer times into baking oven



• Longer baking times at the maximum permissible temperature.

POTENTIAL PROBLEMS:

• Avoid concentrated acid pickles for cleaning



• Avoid bake oven transfer delays, short baking times and low oven temperatures



• Omission of pre-plate stress relief treatments, if required (high hardness-strength parts)

Figure A



Figure B



Figure C