Failure of 17-4 PH Stainless Steel Bolts on a Titan Space Launch Vehicle

Louis Raymond, Manager, Metallurgy Research Section, Aerospace Corp. and Ernest G. Kendall, head, Metallurgy and Ceramics Dept., Aerospace Corp.


From: L. Raymond and E.G. Kendall, Failure of 17-4 PH Stainless Steel Bolts on a Titan Space Launch Vehicle, Metal Progress Vol 93 (No 1) Jan 1968 as published in Source Book in Failure Analysis, American Society for Metals, 1974, p 105–107

Abstract: Several broke at the shank, and failure was attributed to stress-corrosion cracking. But results could not be duplicated in laboratory with salt-solution immersion tests until the real culprit was established: the secondary effect of galvanic coupling—hydrogen embrittlement.

Keywords: Bolts; Galvanic corrosion; Spacecraft

Material: 17-4 PH (Precipitation-hardening stainless steel), UNS S17400

Failure type: Hydrogen damage and embrittlement


Figure 1

Fig. 1  Success of the mission depends greatly on the reliability of high-strength stainless steel fasteners.

Analysis of service failures is complicated because several mechanisms operate simultaneously or sequentially. For example, a structural component in a space launch vehicle can be simultaneously subjected to stress, a marine environment, and complex fatigue loads and frequencies.
A series of bolts on thrust control valves for a Titan vehicle failed at Cape Kennedy (Fig. 2). They were made of 17-4 PH in the H900 condition — a martensitic, age-hardening alloy. In the H900 condition (solution treated and aged 1 hr at 900 F), the material had nominal strengths of 180,000 psi yield and 200,000 psi tensile. The alloy's composition: 0.07 C max, 15.5 to 17.5 Cr, 3.00 to 5.00 Ni, 3.00 to 5.00 Cu, 0.15 to 0.45 Cb+Ta, and trace elements of 1.00 Mn max, 1.00 Si max, 0.04 P max, and 0.03 S max.
Figure 2

Fig. 2  Thrust control valve bolts failed because of hydrogen embrittlement.

Preventing Future Failures

Laboratory findings indicate that the most fool-proof method of preventing such failures without sacrificing any notched yield strength is to age the bolts for one more hour at 1,000 F. Anodized aluminum in the flange interface will minimize corrosion — it will also inhibit galvanic corrosion only if the bolts are inserted carefully and the anodized layer is not abraded.
Failure analysis established that the composition was within specification, as were hardness and tensile properties. Fracture was intergranular; there was no evidence of grinding, burning, mechanical damage, or other microstructural conditions.
The significant observation was that the failure of the bolt was located in the vicinity of a 7075-T6 aluminum alloy flange which corroded — evidenced by white, chalky corrosion products. Rust spots (pits) were observed along the shank about $\genfrac{}{}{0.1ex}{}{1}{2}$ in. from the head. At this point, the bolt emerges from the flange. The formation of rust spots could be prevented by covering bolts with a suspension of graphite in grease (Lox-Safe).
It became apparent that the presence of aluminum accelerated the stress-corrosion cracking of the stainless. Aluminum is anodic to stainless in the galvanic series — it can be used as the sacrificial anode in restricting the corrosion of the steel. In fact, the 7075-T6 alloy has the same potential as cadmium, and it should protect stainless in the same manner as cadmium plating.
The only possible explanation of the failures was hydrogen evolving at the cathode (the stainless steel bolt) and causing hydrogen embrittlement.

Role of Aluminum

In a laboratory test, 7075-T6 was coupled to one of each of a series of duplicate test samples loaded with a stress ring and exposed to salt solution. The specimens broke overnight. To verify results, the remaining uncoupled specimens were exposed to the salt-solution immersion cycle for two more weeks. No failures or rust spots were observed. The aluminum alloy was then coupled to these specimens, and they failed within 24 hr.
Test results established that the critical nature of the failures was related to the aluminum, not the alloy itself. Other tests eliminated the possibility of crevice corrosion introducing pits which lead to stress-corrosion cracking. Serious pitting can be initiated in stainless beneath substances lying on the surface, independent of the chemistry of the object in question. Once pitting starts, serious localized corrosion attack can set up within the stainless. Once attack starts, failure can occur rapidly under the influence of stress.

Hydrogenation Tests

To establish the susceptibility of the 17-4 PH fasteners to hydrogen embrittlement, V-notched bolts were charged cathodically with hydrogen for 15 min in a 4% sulfuric acid electrolyte at 0.02 amp per sq in. The specimens were immediately cleaned and loaded to 10,000 lb in a tensile testing machine. Failure took place in about $4\genfrac{}{}{0.1ex}{}{1}{2}$ hr in the bolts that were heated at 900 F for 1 hr. The results established the susceptibility of 17-4 PH to hydrogen embrittlement.

Simulation of Bolt Failures

Actual bolts removed from the installation at Cape Kennedy were also used as specimens which were loaded with 10,000 lb for 1,000 hr to establish “fail” or “no fail” criteria. Environmental cycle consisted of immersion in a 3.5% NaCl solution for 10 min of every hour. When bolts aged 1 hr at 900 F were coupled to the aluminum alloy, they failed in about $4\genfrac{}{}{0.1ex}{}{1}{2}$ hr — the same failure time recorded for the hydrogen-charged specimens statically loaded in air. It meant that the absorption of hydrogen in the stressed bolt coupled to the aluminum alloy was comparable in severity to that caused by electrolytic charging.
Bolts aged 4 hr at 900 F had longer times to failure, averaging from three to four days. Bolts aged 1 hr at 1,000 F did not fall in tests up to 1,000 hr.
The aluminum alloy, after failure is simulated in laboratory tests, was white and chalky. The stainless steel had surface rust spots in the area covered by the aluminum alloy. These test results duplicated the appearance of failures in service.
A simple method of eliminating hydrogen embrittlement in the 17-4 PH bolts is to age them an additional 1hr at 1,000 F. Heat treatments which promote toughness also add resistance to hydrogen embrittlement. The almost linear relationship between hardness and toughness is shown in Fig. 3. Each point on the graph represents the data obtained by using various combinations of time and temperature. Therefore, it can be concluded that the impact energy is a function of the hardness independent of the specific variable of the aging treatment (Fig. 4). Such behavior is not common to bolts made from low-alloy steels. It can be explained on the basis that the notched-yield to the unnotched yield strength ratio increases as the hardness of the 17-4 PH decreases.
Figure 3

Fig. 3  The relationship between hardness and toughness of 17-4 PH is almost linear.

Figure 4

Fig. 4  Behavior of bolts with different aging treatments was studied under varying loads (see table). Permanent plastic deformation occurs at a load slightly above 11,000 lb regardless of heat treatment. Failure was in the threaded portion.

Insulation is another way of eliminating hydrogen embrittlement. When a suspension of graphite in grease (Lox-Safe) was applied at the stainless-aluminum interface, there were no failures after 1,000-hr exposure — again consistent with the observations made at Cape Kennedy. The grease was an insulator between the dissimilar metal couples, and it eliminated the galvanic cell reaction. Even bolts heat treated to maximum hardness did not fail as long as the grease remained intact.
Nickel-cadmium plating gave inadequate protection because the layer of cadmium was porous and allowed the diffusion of hydrogen. Cadmium plating for protection against stress-corrosion cracking may not inhibit hydrogen embrittlement.
Anodizing the aluminum effectively inhibits hydrogen embrittlement. The aluminum contact with stainless steel is only slightly attacked electrolytically because the flow of current is limited by polarization. The effect was very pronounced when aluminum pieces were galvanic couples to the stainless in the laboratory tests. The aluminum pieces were given a commercial anodizing treatment by the chromic-acid process. Anodized aluminum pieces showed no sign of general corrosion (the chalky white residue), while untreated pieces were attacked extensively under identical conditions of environment and time.

Related Information

W.J. Jensen, Failures of Mechanical Fasteners, Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 1986, p 529–549
Hydrogen Damage and Embrittlement, Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 2002, p 809–822