Home | About
Us | Services | Publications | Links | Contact
Us | Accreditation | Q & A Forums
Print This Document
Hydrogen Embrittlement in our aerospace industry is a metallurgical phenomenon in high strength structural alloys. Very simply, the ductility-strength properties of the alloy are lowered to a level below what the design engineer anticipated in his component design. The resultant reduction in ductility inherently lowers the load carrying capability of the metallic component, hence failure by sudden brittle fracture can occur. Dozens if not hundreds of aircraft accidents, many fatal, have resulted through the years from hydrogen embrittlement. The focus of this paper is centered on aircraft-aerospace steel alloys and the history of their embrittlement problems. This is not a research paper. The solid state physics of embrittlement mechanisms are not intended to be discussed here. Rather, a general overview of the fundamental metallurgical interactions are given, and discussions on inspection methodology are presented. A historical overview of the events, organizations, people, and sadly some serious accidents that have furthered the understanding of hydrogen embrittlement over these past 50 years are set forth.
Historically speaking, the birth of the hydrogen embrittlement dragon encompasses far more than a 9 month gestation period. The aircraft industry most probably saw the first hydrogen embrittlement failures in the late 1940’s during the beginning of the “golden age” of aircraft development. Man’s quest for “…higher - faster - farther…” led to intense efforts at weight reduction of air vehicles. Weight reduction invariably comes from pushing the strength envelope of the structural alloys used. However, little if any effort was made at the time to investigate the potential for hidden side effects of this quest for “…more from less…” . Low alloy steels such as 4130 or chrome-moly steel had been extensively used in the decade before WW II, but the majority of the applications used those alloys in the lower strength normalized heat treat condition. For the non-metallurgists, this equates to roughly 90,000 -125,000 psi tensile strength, a condition pretty much immune to the damaging effects of atomic hydrogen. It is well known that susceptibility to hydrogen embrittlement increases with the heat treated strength level of the steel alloy. This relationship has been shown by some to be quasi-exponential, i.e. double the strength – quadruple the hydrogen embrittlement susceptibility. Attention is directed to the words, “…heat treated…” defined by metallurgical phase transformations during hardening type heat treatments into martensitic microstructures. Austenitic, ferritic, and for the most part banitic structures are not associated with hydrogen embrittlement.
Low alloy steels, based on familiarity, were specified on a routine basis in the late 1940’s at higher and higher strength levels, eventually approaching 200ksi for anemic 4130 low alloy steel. 4140, 4340, and other “enriched” grades took over more applications in the early 1950’s, and they too soon fell victim to the designers’ zeal to get more bang for the buck, i.e weight reduction through higher strengths. As unit strengths went up, cross sectional areas went down and in many instances the load bearing or load transfer areas went down also. In addition, corrosion resistance of low alloy steels goes down with increasing strength levels, so more ideas for corrosion protection against mother nature were needed. Alas, extreme point loadings soon became associated with extreme wear. The process engineers were called in to provide ideas to solve these issues of wear and corrosion resistance. Cadmium, nickel and chrome plate began appearing on some of these very high strength steel components in the mid 1950’s. Invariably, hydrogen evolved from electrolytic plating processes began wreaking havoc. Regretfully, a well documented catalog of hydrogen embrittlement failures during these early years does not exist. This historical perspective is lost to the ages, as most of the process engineers, chemists and metallurgists who participated in this “golden age” of aviation have since passed away.
However, some are still with us, reminding us “kids” daily of how things once were. This author is honored to have had a decade’s long friendship with an individual who was in the thick of things during the early days of hydrogen embrittlement panics. W.L. (Bill) Aves was a process engineer at Chance Vought Aircraft, Dallas Texas during the 1950’s. An interesting, if not unique aircraft, the F-8 Crusader, was developed by Vought beginning in 1952. The aircraft first flew in 1955 and later went on to become the first aircraft to exceed 1000 mph. It was the winner of the Thompson and Collier Trophies for its design, development and speed, but it suffered from probably the first documented catastrophic hydrogen embrittlement failure. The aircraft had a unique design feature, that of a variable incidence or tilting wing useful for U.S. Navy carrier landings. This was in addition to span-wise folding wings for storage purposes. The wing tilt mechanism consisted of an elaborate hydraulic actuator and associated structural links and other components made from 4340 low alloy steel heat treated to 260-280 ksi and subsequently chrome and cadmium plated. Does the alloy and heat treat level sound familiar to those of us knowledgeable with hydrogen embrittlement testing? The Navy recorded several land based flight test accidents where hydrogen embrittlement was found to be the root cause.
Apparently, many aircraft companies struggled with the hydrogen embrittlement dragon for a number of years during the mid to late 1950’s. Eventually enough accidents were documented that some seriously concerned engineers proposed a conference. Indeed, a historical conference symposium on hydrogen embrittlement was held in late July 1960 in Los Angeles in conjunction with the American Electroplaters Society. Landmark presentations dealing with the metallurgy of hydrogen embrittlement failures, its detection, testing and control were given and Bill Aves was a participant in this conference. Shortly after this history making conference, the “..triplets..”, QQ-C-320 on chrome plate. QQ-N-290 on nickel plate, and QQ-P-416 on cadmium plate were re-written to address, for the first time, testing and control of hydrogen embrittlement in high strength steels. Worst case scenario testing was adopted then, a concept still sound and valid 46 years later.
On August 27, 1969 a commercial air-taxi helicopter carrying 21 people broke up in flight over Los Angeles California. The fracture of a main rotor blade spindle and the ensuing disintegration in flight of the helicopter resulted in the deaths of all 21 people on board. Accident investigation findings (Reference: ICAO Report # 107-AN/81) found that 2 of 4 causes of the accident were factors associated with the nickel plating used to repair the spindle and associated improper shot peening before plating. Whether it be aircraft components of new manufacture or components that have undergone repair-refurbishment by metal finishing techniques, detrimental effects of hydrogen embrittlement can and do surface with deadly results.
It’s good to change gears here and talk about the most important metallurgical conditions and testing philosophies associated with hydrogen embrittlement. As we spoke about earlier, hydrogen embrittlement is damage caused by atomic hydrogen. It is manifested as a reduction in ductility in the elastic range thereby causing a reduction in yield and ultimate strength. The damage can be immediate or delayed, and failures can occur before or after an embrittled part enters service. Embrittled parts have been known to “crack or pop on the shelf” from hydrogen driven by residual stress, even before being assembled. If ductility is lowered from embrittlement, the usefulness of metals goes away; i.e. embrittled parts now behave like ceramics. The bottom line is that a hydrogen embrittled part will not deliver the strength or performance the design engineer wanted. Premature failures can happen, sometimes with catastrophic results.
High strength martensitic steels are the most prone to embrittlement due to the high transformation stress of the martensitic lattice or microstructure. Other steels such as the PH grade stainless steels can sometimes be embrittled, and some austenitic grades such as 18-8 or the 300 series can also be embrittled if fully work hardened to the martensitic phase. Titanium alloys can also be embrittled, but the phenomenon is different as the mechanism tends along the lines of the formation of brittle ceramic titanium hydrides along the grain boundaries. Aluminum and nickel alloys are not prone to embrittlement, but some super alloys such as MP-35 etc. have been reported as susceptible. (Authors Note: These purported hydrogen embrittlement failures of multi phase alloys have been challenged by numerous metallurgists.)
How does hydrogen get in? Hydrogen is the smallest size element and in ferrous alloys hydrogen will travel interstitially, or in-between the iron atoms themselves in the iron-carbon lattice. Its easy to see that the permeability of hydrogen is tremendous. In electroplating, it enters at the surface of the steel during the plating process. Hydrogen diffusion obeys Fick’s Law; that is, it always travels from highest to lowest concentration – from the surface of the steel inward. Metallurgically, hydrogen embrittlement is still being debated after almost 50 years. The two prevailing schools of thought today are:
By far high strength steels are the most susceptible to hydrogen embrittlement due to the martensitic transformation stresses and, quite simply, high strength steels are attractive to aircraft design engineers. As stated prior, susceptibility to embrittlement goes up with higher strength levels. Hydrogen can enter steels from many, many sources. It can be present from the original steel making operations, from later casting and forgings steps, from grinding operations, from electrolytic plating and from contact with reducing acids to name a few. However, metal finishing seems by far to be the major contributor to hydrogen embrittlement.
Assuming the worst has happened, how does the aircraft industry detect embrittlement? First, hydrogen must exist, and second, a driving force must be present. The phenomenon is diffusion controlled, as hydrogen must enter the steel and must be forced to areas of structural instability. If hydrogen levels are low – embrittlement may not occur. Likewise, if the driving force is low – embrittlement may not occur, but both are needed before embrittlement happens. Diffusion is time dependent and therefore time is also very critical for its testing. The scientific community has concluded over several decades that short duration, short cut tests can miss potential embrittlement. Analytically and empirically speaking, all accepted mechanical tests for embrittlement have fallen into a 150 - 200 hour test period. Embrittlement also requires a driving force or stress. Residual stresses from cold work, grinding, shrink fits, high strength levels, and altered surface treatments such as carburizing, induction hardening etc. microstructurally will drive hydrogen inward. Residual stresses are the cause for “shelf popping” - parts cracking while sitting on the shelf. For embrittlement testing, the stress is applied as a long term external or sustained load. Hydrogen embrittlement testing is many times called sustained load testing - SLT.
The history of hydrogen embrittlement testing goes back to shortly before the Los Angeles Electroplaters meeting of July 1960. Numerous engineers throughout the aircraft industry had developed mechanical methods of testing 4-5 years prior to the Los Angeles meeting. Early results, now proven over the past 50+ years, show the need for a notch and sustained loads. 4340 low alloy air-melted steel, heat treated to the highest condition of 260-280 KSI (Rc51-53) and notched with a standardized Kt factor of approximately 3 is the accepted protocol. The notch was found to accelerate the test by an order of magnitude (10) and forces the failure to a predictable, reproducible site. I have always referred to this test as a Draconian Test, i.e. a test where an alloy very susceptible to embrittlement is used in combination with a high heat treat level and an elevated stress concentration level from a notch. The philosophy simply is that a plating processes passing this test should produce parts free from embrittlement. Sustained load embrittlement tests have a track record of millions of data points over these past 50+ years. Most probably this data set has never been exceeded by any other metal finishing test.
The physics of sustained load testing - SLT - is based on stress and time. It accelerates detection of embrittlement by going immediately to 75% of Notched Ultimate Tensile Strength, (NUTS) in conjunction with extended dwell time at load. It provides the best opportunity to catch potential embrittlement that may be missed with bend or shortened quickie tests. The philosophy of this type test is very similar to aircraft wing certification tests. Wings are tested at max flight level loads, Max limit loads (similar to embrittlement tests) and finally proof loads to failure. Since cracks will form and propagate only under tensile (tension) stress, most embrittlement tests are tensile tests. Probably 90% of all embrittlement tests are pure sustained load notched tensile tests. A notch held under tension creates a tri-axial state of stress, further enhancing the conditions for embrittlement to manifest itself. Embrittlement bend tests can be less sensitive and reproducible than pure tension tests. Bend samples inherently have 3 stress zones – Tensile, Neutral, & Compression. Once again, cracks open only under tension, hence only 50% or less of the notch cross sectional area is available for detection of embrittlement. Actual chemical content tests for hydrogen have not been found useful in the past. Hydrogen contents as low as 5 ppm have been shown to induce embrittlement and contents this low are in the experimental error range of the test methods employed. In addition, residual and applied stresses known to initiate embrittlement mechanisms tend to sweep the atomic lattice or structure of hydrogen, concentrating atomic hydrogen into embrittlement initiation sites with subsequent fracture. The fracture liberation of localized hydrogen will skew any later chemical testing results. Hydrogen embrittlement always produces intergranular cracking, or cracking along grain boundaries. If a failure investigation does not find intergranular cracking – you do not have hydrogen embrittlement. However, investigators should be cautioned that intergranular cracking can also be caused by temper embrittlement (precipitation of epsilon carbides on prior austenitic grain boundaries) and also by stress corrosion cracking (SCC). SCC is many times a precursor to hydrogen embrittlement, being the initiator or catalyst for hydrogen generation. Porosity type gauges are sometimes used to measure the porosity of various types of platings, but this is only an embrittlement potential test. It only tests the ability of a coating to out-gas hydrogen during a subsequent baking operation. It does not test if the baking operation has been done, nor does it test any diminished mechanical properties of a potentially embrittled part.
DESIGN and PROCESSING CONSIDERATIONS – The History of Problems
Metallurgically speaking, the higher the strength of the alloy, the more prone it is to hydrogen embrittlement. The subtle upward creep of the strength of aircraft parts during the 1950’s and 1960’s resulted in some real shocks to aircraft designers. Accepted design guidelines show that hardened and tempered (martensitic) alloy steels above 150KSI, carburized steels, ball bearing alloys, and spring steels-piano wire should be considered susceptible to hydrogen embrittlement. Other alloys such as the original martensitic grades of stainless (400 series) are also susceptible. Appropriate measures should be taken to reduce hydrogen generation during actual plating and always perform the subsequent embrittlement relief or baking operations. The origin of the term “Baking” is unknown, although references can be found in metallurgical literature dating to the early 1940’s. Although hydrogen embrittlement relief is the more technical term for this moderately elevated temperature step (usually 375 deg. F), baking seems to be the universally recognized description. From the available literature of the 1950’s, it appears 375 deg. F is based on the time honored heat treating practice of going 25 deg. below any previous tempering temperature. Since 4340 alloy steel parts heat treated to 260-280 ksi back in the 1950’s were normally tempered at 400-425 deg. F the aircraft industry settled on 375 deg. F as a bake temperature set point. There is no magical metallurgical reason for the 375 bake temperature other than that. As stated earlier, baking is a diffusion controlled phenomenon governed by Fick’s Law, an exponential relationship relating concentration gradient, temperature and time. The avoidance and control of hydrogen embrittlement clearly begins with prudent designs. Engineers with little metallurgical training can tend to ask for the impracticable – ultra high strength alloys with low tolerance to embrittlement, risky manufacturing processes or low toughness. Control of hydrogen embrittlement continues by utilizing metal finishing methods that lower or eliminate the inherent generation of hydrogen. Stay away from reducing acids such as HCL or H2S04. Maximize the plating efficiency at the cathode. Minimize plating time – maximize current density. Use electroless process when possible. Use vapor deposited coating when prudent.
The last line of defense against hydrogen embrittlement is baking. Never shortcut baking time or temperatures. Since Fick’s Law is a temperature dominant exponential relationship, bake temperature is very, very important. As an example of the power of temperature, look at the widely specified bake temperature of 375 deg. F +/- 25 deg. You will bake out twice as much hydrogen at 400 deg. vs. 350 deg. Time is of course very important also. Time is a linear relationship in Fick’s Law – double the time - double the bake out of hydrogen. Quick transfers into the bake oven from the final plating operation is critical. You do not want residual stresses within the parts to begin their nasty trait of sweeping and concentrating hydrogen to potential areas of fracture. Usually these transfer times are set at a 4 hour maximum time delay. Some aircraft companies are more cautious and specify a one hour maximum delay. Don’t confuse bake delay times with bake times. Delay or transfer times are 4 hours maximum. Actual bake times of 23 hours or longer are common. You will always see a benefit with extended bake times. Some SAE-AMS specifications call for 96 hour bake times for ultra high strength ball bearing steels and TD Chrome. Baking is a simple but powerful step. Almost 71% of documented aircraft embrittlement failures over these past 40 years appear to have been attributed to the baking operation - a missed or omitted bake, extended delay from plate to bake, insufficient bake temperature or short bake times.
HYDROGEN EMBRITTLEMENT and FLIGHT SAFETY
A Flight Safety Part is a part whose failure would result in either a loss of structural integrity, a loss of control function, or the forced transfer of loads or control function to a backup system. Flight Safety parts normally are 100% inspected prior to release for assembly, buildup and flight. Hydrogen embrittlement must be treated as a process parameter capable of affecting Flight Safety. Whether it be new aircraft parts or rebuilt-overhauled parts, many times conflicts arise during manufacture. Sometimes specification requirements conflict with actual purchase orders. The results can be serious with service failures and product recalls. Metal finishing is usually the last manufacturing operation before final assembly. Any prior errors in heat treatment, shot peening , grinding or stress relief can manifest themselves during plating. Accidental carburization, abusive grinding burns or shot peening mistakes spell disaster during plating. AMS Committee “B” on Finishes-Processes-Fluids, Society of Automotive Engineers (SAE) has taken a lead in changing the way plating and processing specifications are written to mandate critical information flow-down to the metal finisher to minimize or prevent human errors leading up to metal finishing.
HYDROGEN EMBRITTLEMENT and IN - SERVICE INSPECTION
Quite simply, hydrogen embrittlement manifests itself as cracks, of which traditionally used NDT techniques such as magnetic particle, eddy current and ultrasonic methods can be used. Many times unique embrittlement testing and inspection of in-service parts are employed. These revolve around mechanical testing or the use of loads and stress to initiate or accelerate potential embrittlement cracks. Proof loads above expected flight level loads have been used. Acoustic emission techniques have been employed for detection of incipient cracking. New research methods utilizing radiation methods are under study. Interesting analytical experiments utilizing existing residual stresses from shrink fits, clamp up loads or transition stress profiles from carburized or shot peened surfaces have been used in tandem with “Father Time” to provide acceptable disposition of suspect parts. Hydrogen embrittlement rarely involves a singular part, “rogue elephant” situation. The leading cause of hydrogen embrittlement in the aerospace industry today is metal finishing, usually involving batch processing. If a single embrittlement failure surfaces, usually more will follow in a similar time frame. Although the exposure to catastrophic events is heightened, additional occurrences can and have alerted the industry to the potential for failures. Good failure analysis, promptly executed, many times brings to light additional factors initiating the hydrogen embrittlement.
In my career as a aircraft metallurgist, I have found that issues involving improper heat treat (high hardness, accidental carburization etc.) abusive grinding burns, excessive material removal from grinding, rolled thread stress profiles, EDM-ECM artifacts and many other manufacturing process irregularities will gang up and catalyze hydrogen embrittlement. Sometimes, but not very often due to painful costs, actual flight hardware is subjected to full up flight loads over a sustained period of time to establish freedom from embrittlement. In-service inspection of hydrogen embrittlement forces process engineers to look in the mirror - errors in plating can be the legacy of hydrogen embrittlement. Forgetting to bake the parts, baking for the wrong time, wrong temperature or having bake oven equipment failures can result in suspect parts. What can be done? Many times nothing can be done as irreversible damage occurred quickly with cracking already present. Or, the complexity of the parts or the assembly prevents re-work. Flight safety can be compromised here if face value acceptance of suspect parts occurs. Good Material Review engineers really earn their pay here. If in-process inspection reveals errors or mistakes in plating, we always recommend emergency damage control. Act quickly in getting parts into a bake oven, or the bake oven back on line. Document what happened while facts are fresh. Ongoing embrittlement stops increasing the moment the parts reach proper bake temperature. Implement a recovery plan by completing the proper bake cycle in its entirety – do not cut corners on either time or temperature. However, what about embrittlement damage that has already been established? Notify the customer or the cognizant engineering authority.
Propose a recovery plan if plausible, after documenting all the facts. Potential dispositions of flight hardware have been to Scrap out the parts or use as is or bake the parts for an additional time at normal temperature or bake the parts for an additional time at an elevated temperature or strip – bake – NDT – replate. Sometimes actual parts are mandated to be mechanically tested, but this is expensive and time consuming. We proposed to the industry several decades ago to test standard notched tensile samples that have seen the identical plating or baking error. Using separate test samples “ Replicate the problem – Test the problem”. This has become a widely used method today to disposition suspect parts.
A short review is needed of the process control testing common in metal finishing. Process control tests are tests intended to control the plating process over time. The interval may be daily, weekly or monthly type tests. These tests utilize standardized test samples and standardized test loads. The purpose is to determine trends or patterns over time. Process control tests should not necessarily be used for lot buy off or for actual part acceptance, as they can be processed on different days, different plating conditions (current density, pre-clean methods etc.) However, acceptance tests are different as they represent actual parts. Hydrogen embrittlement test samples can be wired or attached to actual parts. Flight safety serialized parts or components can be cross referenced now that the best simulation of actual plating conditions are present. Specific involvement of a process engineer is likely with hydrogen embrittlement acceptance testing. The aircraft industry is faced with a daunting task here, as in-service inspection of aircraft, pertaining to hydrogen embrittlement, generally relies on traditional NDT techniques after problems surface. Tight monitoring and control of processes that can generate hydrogen beforehand are absolutely critical.
Hydrogen embrittlement is a relatively young but serious issue for the aircraft industry. Its appearance parallels the development of the flying machine itself. Numerous aircraft accidents have been caused by embrittlement failures. Avoidance involves a thorough understanding of hydrogen generating reactions. Control involves both avoiding hydrogen generation and hydrogen elimination referred to as baking. In-service inspection of aircraft components for hydrogen embrittlement relies on time tested NDT techniques and, many times, unique methods and thinking to determine acceptance for flight. Vigilance is required for continued flight safety. The physics of hydrogen embrittlement is omnipresent – one mistake is all it takes for disaster.
Call or contact us with your testing or processing questions.
Omega Research Inc.
Copyright 2002 - 2008, Omega Research Inc. All Rights Reserved
Text, Content, Concepts and Information Protected under the Copyright Laws of the United States.
This information MAY NOT BE reproduced in whole or in part, stored in a retrieval system, photocopied, recorded or transmitted in any form or means electronic, mechanical, photographic or otherwise. This Copyrighted material is intended solely for the clients of Omega Research Inc. and under no conditions may the Text, Content, Concept and/or Information shown be used for financial gain, profit, business or professional enhancement by any individual party.
The reader is cautioned that the subject nature of metal finishing and hydrogen embrittlement is technical and complex. Some simplification of metallurgical concepts may be presented in order for an easier understanding of the subject matter for the intended reader.
The information contained in this website should be considered general information on the subjects of metal finishing, hydrogen embrittlement and associated problems, as they exist in the metallurgical sciences today. Some contractor, agency or specification 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, customer or specification. The reader is solely responsible for determining the usefulness of the information presented, as it pertains to the readers specific product or application.
Omega Research Inc. assumes no liability either explicit or implied as to the suitability, accuracy, or applicability thereof. Any reader choosing to utilize the information from this website does so by their own discretion and accord, and as such without recourse against Omega Research Inc.
Copyright | Please Read Our Disclaimer | Help |