English
Menu

Silver Bridge Collapse (1967): Brittle Fracture from an Eyebar Crack

On the evening of December 15, 1967, the U.S. 35 bridge over the Ohio River (the Silver Bridge) collapsed and 46 people died. The collapse began when one eyebar, a bar with an eye at each end that formed the bridge’s suspension chain, fractured at its eye. At the origin was a small crack, less than about 3 mm deep, that had grown by stress corrosion and related processes over about 40 years.

What happened

Date
December 15, 1967, at about 5 p.m. (EST)
Location
U.S. 35 bridge over the Ohio River between Point Pleasant, West Virginia, and Kanauga, Ohio, USA
Structure
Eyebar chain suspension bridge designed in 1927 and opened in 1928. Each chain link was a pair of heat-treated 2 × 12 in bars
Casualties
31 of the 37 vehicles on the bridge fell with it; 46 people died and 9 were injured
Origin
Lower limb of the eye of eyebar 330 at joint C13N of the north chain, Ohio side span
Investigation
U.S. National Transportation Safety Board (NTSB), report NTSB-HAR-71-1; metallurgical report of the National Bureau of Standards (NBS)

Sequence of events and the investigation’s findings

The collapse happened at about 5 p.m. According to the NTSB, once the north chain separated, the collapse spread across the whole bridge in about one minute.

The NTSB found that the collapse was caused by a cleavage fracture in the lower limb of the eye of eyebar 330 at joint C13N. The fracture resulted from a critical-size flaw that developed over the 40-year life of the structure through the joint action of stress corrosion and corrosion fatigue. The flaw was inaccessible to visual inspection and, the NTSB said, could not have been detected by any inspection method then known without disassembling the eyebar joint.

The chain of causes from the materials side

Silver Bridge: the chain of causes from the materials sideHeat-treated carbon steel eyebarAbout 0.6% C, quenched and temperedTensile stress at the eye’s boreA spot that could not be inspectedCrack grows by stress corrosion etc.Under about 3 mm deep in 40 yearsNear-freezing night, low toughnessThe crack reaches critical sizeCleavage fracture at the eyeNo obvious deformationChain separates; bridge collapses
Silver Bridge: the chain of causes from the materials side

The eyebar steel was a carbon steel with about 0.6% carbon and no significant alloying element other than manganese, water-quenched and tempered at about 1150–1200 °F (about 620–650 °C).

In the National Bureau of Standards (NBS) metallurgical examination, the fracture of the lower side of the eye started from pre-existing cracks at the surface of the hole, less than 1/8 in (about 3.2 mm) deep. All the crack origins were in severely corroded areas, and many other cracks nearby that were not part of the fracture were filled with oxide, showing that they had been held open by tensile stress for a long time. The lower side of the eye broke without appreciable plastic deformation, and the upper side then failed by excessive load with extensive bending.

University teaching material (the Open University) explains that the eyebar steel had low toughness at the near-freezing temperatures on the night of the disaster, making it susceptible to brittle fracture; with low toughness, even a small crack is critical. Each link was a pair of bars, and when one broke, the remaining bar slipped off the pin: the non-redundant structure collapsed completely.

On interpretations of the cause: the NTSB attributes the crack growth to “the joint action of stress corrosion and corrosion fatigue”, while Part I of the National Bureau of Standards report finds the cracks similar to those caused by stress corrosion in other steels and says the evidence suggests stress corrosion rather than fatigue.

The clues that revealed the cause

Open the guide with this case’s clues

Opening the guide with these clues ranks brittle fracture, the final mode of failure, at the top, followed by hydrogen embrittlement and stress corrosion cracking. External clues alone cannot separate these two; only examining the origin crack itself reveals the mechanism. In this accident, the National Bureau of Standards examination found that the evidence suggested stress corrosion rather than fatigue.

Related failure modes: Brittle fracture / Stress corrosion cracking

What changed afterwards: design, rules and inspection

Lessons

Check your understanding

Which statement about stress corrosion cracking is correct?

  1. AIt tends to occur where shot peening has introduced compressive residual stress at the surface
  2. BIt occurs only when the applied stress exceeds the yield stress and the material deforms plastically
  3. CIt occurs when a susceptible material is under tensile stress in a specific corrosive environment
Show answer and explanation

Answer: C. It occurs when a susceptible material is under tensile stress in a specific corrosive environment

Stress corrosion cracking is the growth of cracks, even at stresses far below the yield stress, when a susceptible material under tensile stress (from external loads or from residual stresses due to welding or working) is exposed to a specific corrosive environment. Classic examples are austenitic stainless steels in hot chloride environments and brass in ammonia-containing environments (season cracking). It is dangerous because the part fractures suddenly with almost no visible corrosion. Removing any one of the three conditions prevents it, so residual stresses are reduced by stress-relief annealing or compressive residual stress is introduced at the surface by shot peening. Compressive stress does not open cracks, so the choice saying it occurs readily where there is compressive residual stress is backward. The choice saying it cannot occur below the yield stress is also wrong.

Which material shows a distinct ductile-to-brittle transition at low temperatures?

  1. AAustenitic stainless steel (FCC)
  2. BAluminum alloys (FCC)
  3. CFerritic carbon steel (BCC)
Show answer and explanation

Answer: C. Ferritic carbon steel (BCC)

In BCC metals, the stress needed to move dislocations (the Peierls stress) rises sharply as temperature falls, so at low temperatures cleavage fracture occurs before yielding and the absorbed energy in the Charpy impact test drops abruptly over a certain temperature range. This is the ductile-to-brittle transition, and ferritic carbon steel is the classic example. FCC austenitic stainless steels and aluminum alloys have many slip systems and dislocations remain mobile at low temperatures, so they show no distinct transition and are used for cryogenic applications such as liquefied gas tanks. The transition temperature is lowered by grain refinement and raised by neutron irradiation and similar damage.

A steel plate with a fracture toughness K_IC = 50 MPa·√m carries a tensile stress of 200 MPa. Using K = Yσ√(πa) with a geometry factor Y = 1.0, which value is closest to the critical crack length a at which fracture occurs?

  1. AAbout 20 mm
  2. BAbout 62.5 mm
  3. CAbout 80 mm
Show answer and explanation

Answer: A. About 20 mm

Setting K = K_IC gives a = (1/π)(K_IC/Yσ)². K_IC/σ = 50/200 = 0.25 √m, so a = 0.25²/π = 0.0625/3.14 ≈ 0.0199 m, about 20 mm. About 62.5 mm forgets to divide by π, and about 80 mm is the mistake of taking 0.25/π without squaring. The critical crack length is inversely proportional to the square of the stress, so halving the stress allows a crack four times as long. Comparing the crack sizes that nondestructive inspection can detect against this critical value is the basis of damage-tolerant design.

When the fracture toughness of plates of the same material but different thickness is measured, what happens to the apparent fracture toughness K_c as the thickness increases?

  1. AIt keeps increasing, since thicker plates have larger plastic zones
  2. BIt is a material constant, so it is the same for any thickness
  3. CIt decreases as constraint grows, approaching the plane-strain value K_IC
Show answer and explanation

Answer: C. It decreases as constraint grows, approaching the plane-strain value K_IC

In a thin plate, constraint through the thickness is weak (close to plane stress), so the plastic zone at the crack tip spreads widely and absorbs much energy, giving a high apparent fracture toughness K_c. As the plate thickens, deformation through the thickness is constrained and the plate approaches plane strain; the plastic zone shrinks and K_c decreases, leveling off above a certain thickness at a nearly constant minimum, K_IC. This is why K_IC can be treated as a material constant and is used in design as a conservative value. Test standards require a thickness, proportional to (K_IC/yield stress)², large enough for plane strain to hold.

Related terms

Stress corrosion cracking Fracture toughness Stress intensity factor Tempering Yield before fracture

Main sources

  1. NTSB, Highway Accident Report NTSB-HAR-71-1: Collapse of U.S. 35 Highway Bridge, Point Pleasant, West Virginia, December 15, 1967 (summary, cause and recommendations)
  2. J. A. Bennett, NBS Report 9981: Metallurgical Examination and Mechanical Tests of Material from the Point Pleasant, W. Va. Bridge, Part I (U.S. National Bureau of Standards)
  3. West Virginia Department of Transportation: Silver Bridge
  4. The Open University, OpenLearn: The Silver Bridge disaster: the collapse
  5. FHWA, Happy 50th Anniversary, National Bridge Inspection Standards

Dates, numbers of people, dimensions, temperatures and other values were checked against the official reports above and sources that cite them.

Other case studies

Back to the failure analysis guide