Investigating Material Failures:
Were the Titanic and Challenger Disasters Preventable?
  Dana Ashkenazi

In April 1912, only three days after sailing off on her maiden voyage, the luxury cruiser Titanic collided with an iceberg in the North Atlantic. Two hours and forty minutes later, it sank into the deeps, taking some 1,500 of its passengers with it. This figure made the Titanic disaster the worst in the history of navigation. The tragic story of its drowning has not been forgotten since, mainly because of its symbolic significance: at the time, it was considered an unsinkable technological marvel, and the disaster came to represent the shortcomings of modernity. It took 73 years before the submerged wreckage was discovered, thanks to a latter-day technological feat. About four months after this historic discovery, and precisely 73 seconds after taking off, the Challenger Space Shuttle exploded in midair with millions of shocked spectators worldwide watching the launch on live TV. The tragedy was made even more poignant by the fact that the crew included a teacher selected to be the first "ordinary human being" in space. Following this catastrophe, all space shuttle flights were grounded for about 18 months, and the U.S. government appointed the Rogers Commission to investigate its causes. Nobel laureate physician Richard Feynman was the most celebrated member of the commission. Using a simple experiment, broadcasted live, he demonstrated the key engineering failure that led to the disaster. Both disasters are among the costliest and most tragic in the history of transportation technology. Although separated by 74 years, both were caused by the same material failure: the tendency of extensible materials to become brittle at low temperatures.
Our Wonderful Material World
When we look around us, we cannot fail to be astonished by the huge variety of materials. Their critical importance in our lives is attested to by the fact that key periods in the prehistory of humankind are named after the materials used at the time – the Stone, Bronze and Iron Ages. All materials are made up of elements, each with its own particular type of atom. Different atomic bonds form different materials. The types of atoms (elements) of which a material is made up, their bonds and structural defects are all related to its mechanical properties, such as strength and ductility. Such properties are the domain of materials engineers. Based on interatomic bonds and the structural ordering of atoms in the material, the materials engineer decides whether a certain material may be used for certain applications, and what newly developed materials may yield improved results. In a solid material with a crystalline structure, the atoms are ordered periodically. The basic crystal unit, which recurs multiple times, is called a unit cell. The simplest unit cell is comprised of eight atoms at the cube points – this is called simple cubic (SC) ordering.

Image 1: Top: Apples ordered in a recurring pattern, an instance of the typical, ordered periodic structure of crystals. Bottom: A simple cubic unit cell and its recurrence in a simple, three-dimensional lattice. The image shows only the atoms facing us.

Image 2: The three unit cell structures typical of metals: body-centered cubic (BCC), face-centered cubic (FCC) and hexagonal close-packed (HCP). In the figure depicting the HCP unit cell, the atoms of the top and bottom basis planes are colored red, while the central plane is colored orange.
Most naturally occurring metals are made up of crystals with either of three unit cell structures: body-centered cubic (BCC), face-centered cubic (FCC) and hexagonal close-packed (HCP). In the BCC structure, there are eight atoms at the cube's points, and one in its center. In the FCC structure, there are eight atoms at the cube's points and six additional atoms at the centers of its faces, making it much denser. In the HCP structure, the lattice is made up of atomic layers, each made up of an atom surrounded by six others. The atoms in each layer occupy the slots among the atoms in the adjacent layer. Thus, in this structure each atom "touches" 12 neighboring atoms: six around it, three above it and three underneath it.
In the real world, materials are imperfect, with numerous defects in the crystalline structures. These defects are classified according to their dimension: point, line, planar and bulk defects. Point defects are on the scale of a single atom. The simplest point defect is called vacancy – where an atom is lacking at a certain point in the lattice. In the self-interstitial defect, an atom "manages" to push itself into a tiny space (penetration site), usually unoccupied by atoms. Pure metals containing a single type of atoms are nonexistent in nature, as crystalline structures always contain foreign atoms, or impurities. These may be either substitutional impurity atoms, in which "guests" substitute for atoms from the "host" lattice, or interstitial impurity atoms, in which a guest manages to occupy a penetration site.
Image 3: Four types of point defects in crystalline lattices: vacancy, self-interstitial, substitutional impurity and interstitial impurity.
 Line defects are unidimensional, affecting a row of atoms. The most common line defect in crystals is called dislocation. Plastic deformation is a situation where a material is transformed due to the impact of an external force, without fracturing. When a material is plastically deformed, its atomic planes slide over one another. The line separating two such atomic planes in a crystal is called a dislocation line. When a crystal contains an atomic plane which does not extend to the whole length and breadth of the lattice, but ends somewhere in the middle, the lattice is deformed, and the line created along the atoms at its edge is called edge dislocation. The motion of dislocations in the lattice involves the breakdown of interatomic bonds. Dislocations have a significant effect on material (mainly metal) properties. Metals are hardened and toughened mainly by hindering dislocation motion, so that it occurs on planes and in directions where atomic density is maximal. In materials with FCC unit cells, the slip planes are denser than in materials with BCC unit cells (such as steels), making them more ductile. Ductility is the extent of a given material's ability to become plastically deformed without fracturing.
Image 4: The analogy between the progress of a dislocation in a crystal (from right to left) and the motion of a caterpillar. The arrows indicate the direction of forces operating on the crystal, and the blue spheres represent the excess atomic plane forming the dislocation line. Such partial planes (a type of line defect) are always found in materials. Pushing an excess atomic plane into a crystal creates a line called edge dislocation, which disturbs the perfect crystalline structure.

Image 5: Ordering of atoms (red spheres) around an edge dislocation. The excess atomic plane is located above the dislocation line. The edge dislocation deforms the crystal structure. In this case, the dislocation line pushes into the plane of the page.
When a liquid cools and solidifies, the solidification process is gradual, beginning in impurity areas. Each such area constitutes a solidification point, producing a crystal. Consequently, different areas inside the material have different unit cell directions. These areas are called grains, and measure between a few microns and several millimeters, depending on the type of material, the way it has been produced and processes it has undergone. A series of grains with varying directionalities produces a multicrystalline material. The boundary between grains is a planar defect area. Atoms at grain boundaries are not arranged in a perfect crystalline ordering and contain impurities, which are pushed towards grain boundaries in the solidification process. As areas of material discontinuity, grain boundaries obstruct the motion of dislocations. The more grain boundaries there are in a given crystal, the harder and tougher, but less ductile the material. Finally, bulk defects are three dimensional. These include atomic clusters and voids (when a whole atomic group is missing).
Image 6: Grains and grain boundaries in a multicrystalline material. To the right, atomic structure discontinuity within grain boundaries (orange spheres); to the left, grains (in yellow) and grain boundaries (in black) as seen through an optical microscope. Each solid has an outer surface, where its continuous structure ends, and which may therefore be referred to as a planar defect.

Equipped with this basic knowledge on the structure of materials, we will now try to understand what are ductile and brittle materials, and the conditions under which a ductile material may become brittle. Fracture Toughness During World War II, significant brittle fractures were formed in many of the Liberty class cargo vessels used by the US navy to deliver supplies to Europe. Some of them actually broke in two in mid-ocean, while others suffered such severe fractures that they had to be decommissioned. The Liberty ships were mass-produced in the US between 1941 and 1945, using a welding technology that was then in its infancy. Moreover, the ship welders were not experienced professionals. More than 2,700 such vessels were built during the war, and over 1,500 suffered severe damages, many of them in the North Sea, where temperatures were very low. There were no previous warnings about these failures, caused by brittle fractures in the ships' steel frames. Subsequent inquiries have found that although made of ductile steel, the ships' hulls became brittle at low temperatures. This finding reopened the question of why the Titanic sank. Researchers who compared the cases of the Titanic and the Liberty ships concluded that despite being ductile under ordinary conditions, materials such as steel are liable to become brittle at very low temperature. This change in material properties is called ductile-brittle transition. Thus, the frequent failures of the Liberty ships during the Second World War brought a new engineering area into being – fracture mechanics.
Fracture mechanics seeks to predict where cracks may be expected to occur and to determine the circumstances under which these cracks might develop into fractures and ways of stopping their propagation. We are all familiar with crack propagation in daily life: a broken glass, a crack in the wall, cracks in dried earth and the cracked windowpanes of vehicles involved in accidents. Many structures, both those made of brittle materials such as concrete and those made of ductile materials such as steel, might fail under certain conditions. It is crucial to understand the characteristics of cracks, since they might cause structural failure. Fracture mechanics relies on mathematical tools, empirical methods and computer simulation to predict crack propagation. Using strain calculus (force per unit area), the propagation of cracks in a strained structure may be predicted, as well as the conditions under which structural failure will occur.
Fracture toughness measures a material's ability to resist fractures. In general, tough and low-ductility materials will also have low fracture toughness; such materials are called brittle. Glass is the ultimate brittle material. On the other end of the brittleness-ductility scale, rubber is the ultimate ductile material. The mechanic behavior of most materials lies somewhere in-between.
Materials are classified as brittle or ductile based on the following assessments:
The material's characteristics during a tensile test.
The degree to which the stretched body's cross-section area narrows after the tensile test.
The material's characteristics during an impact test, that is, a test measuring the effect of temperature on the amount of energy required to cause a fracture.
 The way the fractured body's fracture surface looks after the tensile and impact tests.
Before discussing ductility and brittleness at greater length in order to understand their relevance to the Titanic and Challenger disasters, we will describe the tensile and impact tests, how they are conducted in a laboratory, and the information that can be gleaned from those tests.

Image 7: Right: A steel tensile specimen, with a "neck". Left: Engineering stress-strain curve typical of metals.

 Image 8: Engineering stress-strain curve of a ductile material and a brittle material. The area under the curve represents the energy absorbed up to fracture, or the energy required to fracture the material. The greater the size of the area under the curve the more fracture resistant the material.
Tensile Test
Among all mechanical tests, the tensile test provides more information than any other single test. This test uses a model (a chunk of material with standard shape) which is inserted to a special tensing device and stretched until it fractures. During the test, the force applied and the lengthening of the model are measured. Based on these measurements, a graph called engineering stress-strain curve is plotted. The graph represents engineering rather than "real" data, since stress and strain are calculated relative to the model's initial dimensions, rather than its dimensions at each moment. Stress is defined as force per unit area – in this case, the model's cross section. Strain is defined as the change in length relative to the model's initial proportions as a result of the load applied. Engineering stress-strain curves inform us about the material's behavior and properties.
Typical ranges and points on the graph include the following: elastic range, yield stress, plastic range, ultimate tensile strength (UTS) and stress to fracture. In the elastic range, only elastic (recoverable) strains are formed, which disappear completely once the force applied to the model is removed. In other words, once the load is no longer applied, the model will revert to its original shape. Yield stress is the stress required for the material to shift from the elastic to the plastic range. When lengthening goes beyond the elastic range, strain is no longer proportional to stress, and plastic deformation occurs. On the atomic level, plastic deformation is caused by dislocation slide. Interatomic bonds are severed due to dislocation motion in the material, and new bonds are formed. In the plastic range, when the stress is removed, the material will not revert to its initial state. Next, Ultimate tensile strength is the extreme point in which the model begins to form a "neck", or a local area which becomes gradually smaller (the cross section narrows) until a fracture is formed. This phenomenon occurs only in ductile materials. Finally, stress to fracture is the point at which the model fails, or fractures. Fracture toughness may be assessed according to the size of the area below the stress-strain curve. This area represents the amount of energy absorbed by the material before it fractures. The greater this area (the higher the energy required for fracture) the more ductile (or fracture resistant) the material, because more energy must be applied in order to fracture it.
Assessing fracture toughness in terms of the area underneath the stress-strain curve is appropriate for laboratory tests, but laboratory conditions do not resemble real-life usage conditions, under which the section in question may be exposed to forces loaded at high velocities and under varying temperatures. In order to cope with this problem, an impact test must also be conducted.
Impact Test
Many materials are used under a variety of temperatures, usually within the range of between a few and several hundreds of degrees above zero. Impact test is conducted within a broad range of temperatures. As opposed to the relatively slow loading rate of tensile tests, in impact tests the model is tested under rapid loading. The impact device is equipped with a heavy pendulum, which revolves around a hinge located at the top of the device. The friction between the hinge and the pendulum is negligible. At the bottom of the device, the tester places a standard model in which a groove is cut. During the test, the pendulum is lowered until it collides with the model held in a vice and fractures it in a single blow. After this collision, it continues its motion and rises to a certain elevation determined by its residual kinetic energy. When the pendulum is raised prior to the actual experiment, is potential energy is E1. Once the pendulum is released and freefalls, it accumulates kinetic energy, which reaches its maximum value at the lowest point, at which the potential energy in relation to the model reaches zero. At this point, the pendulum collides with the model, fractures it, and goes on to rise to a certain elevation which is lower than its starting point. At this point, its potential energy is E2. The difference between the initial and final potential energies – E1 - E2 – represents the energy required to fracture the model. This amount is called fracture energy and denoted E.
The main objective of impact tests is to select materials with high fracture toughness. Recurrent impact tests of models made of a given material under various temperatures provide its fracture energy values at those temperatures. In materials with a BCC structure – steels mostly – an extreme change in fracture energy is observed within a certain temperature range. Above a certain temperature, the material is ductile, requiring relatively high energy to fracture it; below this point, the material becomes brittle and requires very low energy to fracture. This material behavior is called ductile-to-brittle transition. The temperature at which this phenomenon occurs is called ductile-to-brittle transition temperature. Thus, in BCC steels, under conditions of sudden load and low temperatures, brittle fracture without deformation occurs. Ductile-to-brittle transitions have also been observed in polymers, due to elasticity loss. On the other hand, in FCC-structured crystalline materials, the fracture energy changes with temperatures in a gradual fashion. Brittle fracture is a type of cleavage fracture resulting from the separation of two parallel atomic layers, until the cohesion force (which attracts atoms to one another) falls to zero. In a brittle fracture, we "tear" two atomic layers apart. This fracture occurs at the end of the elastic range of the stress-strain curve. It occurs suddenly, without any warning, so that failure is immediate. Naturally enough, this type of failure is particularly dangerous from the engineering perspective. Danger is of course maximal when engineers assume they've used a ductile material, but this material becomes surprisingly brittle under certain conditions. Such circumstances are a recipe for disaster. Ductile fracture, on the other hand, is characterized by considerable plastic strain as the crack propagates. A material about to fracture this way "delivers prior notice". An ideal ductile fracture is created by sliding two atomic layers one over the other (shear stress), until both layers are fully separated. In the stress-strain curve, the ductile fracture will appear only after the maximal tensile stress, with necking observed in the process.

Image 9: Impact test. Right: Impact test machine. Left: Impact test specimen.

Image 10: Ductile-to-brittle transition. Right: Ductile and brittle fractures as observed through the microscope (voids in the ductile fracture, as opposed to cracks in the brittle fracture). Left: Fracture energy as a function of temperature in materials with ductile-to-brittle transition.
Equipped with the proper conceptual tools, we may now return to the circumstances that led to the sinking of the Titanic and the Challenger disaster. Image: Right: Close-up of a model (in sky blue) used for an impact test. The Blue arrow indicates the direction in which the pendulum hammer impacts the model. Left: How an impact testing device operates.
Why Did the Titanic Sink?
The luxury cruiser Titanic embarked on her maiden voyage from Southampton to New York on April 10, 1912, with over 2,200 passengers and crew on board. This gigantic ship, built in Britain, was named after the mythological Titans, descendents of Uranus and Gaia, whom the Greeks believed ruled the universe until they were defeated by Zeus with the dawn of the Olympic era. This was certainly an apt designation, in view of the cruiser's dimensions – it measured 882'8" in length, 92" in breadth and 59" in height (from the water line to the boat deck), and could carry up to 3,500 passengers and crew. Powered by 29 steam boilers, the Titanic could reach a record speed of 23 knots (26.5 mph). It had squash courts and a swimming pool. Its luxury first-class suites housed famous artists, businessmen and journalists. Its ample cargo space was used to store the finest wines. Down below, in basic wooden cabins, it carried some 700 immigrants on their way to the land of opportunity. The Titanic's hull was made of steel, considered at the time to be the best of its kind and the most appropriate for shipbuilding. It had a double bottom and its hull was divided into 16 separate (supposedly) watertight compartments, the passageways among which were controlled directly from the bridge. Its designers' dream was to build a huge luxury cruiser – a floating fancy hotel – that could sail from Britain to the United States at maximum speed and safety. Although the Titanic was equipped with lifeboats for emergencies, they were enough for only about half the passengers and crew.
When the Titanic put out to sea, its captain was notified that icebergs were forecast along its planned route. Nevertheless, he chose not to reduce its speed, as he was determined to reach New York before the designated arrival time. Four days later, on April 14, 1912, at 23:40 GMT, a sailor on the bow spotted an iceberg. He reported to the navigator, who started turning the ship. Nevertheless, the iceberg struck the Titanic's side and breached the hull. After the collision, the order was given to deploy the lifeboats, but many of the passengers were reluctant to abandon the safe gigantic ship for the cold, stormy waters and the tiny lifeboats. As a result, the first lifeboats to leave the Titanic were half empty. Only later, when the ship began to tilt ever more ominously, reality sunk in and the lifeboats began to fill. Most survivors were first- and second-class passengers, since their compartments were closer to the boat deck. Due to the rift caused by the collision, the water which rushed into the Titanic pulled its bow downwards and made it tilt. At 2:20 on the morning of April 15, the hull finally caved in under the pressure and broke in two. Only two hours and forty minutes after colliding with the iceberg, the Titanic sunk. Due to the shortage of lifeboats, only about 700 of its passengers and crew escaped with their lives. The rest drowned or froze to death in the ice-cold Atlantic.
The committees of inquiry appointed in the aftermath of the disaster concluded that the ship was inappropriately navigated after the iceberg had been spotted. They reasoned that it was better to collide with the iceberg frontally rather than turn the ship and have the iceberg hit it sideways. Another early conclusion was that the lifeboats were inadequate in number.
In 1985, 73 years after the Titanic sank Dr. Robert Ballard made history when he discovered the remains of the sunken ship at the bottom of the ocean. An expert in marine geology and geophysics, Ballard managed to unlock many marine mysteries using state-of-the-art robots and engineering equipment. Ballard found the Titanic at a depth of about 12,000 feet under the ocean. The ship was broken into two sections, about 2,000 feet apart. Assorted steel fragments lay between them. After discovering the wreck, Ballard said that as a child, he loved reading Jules Verne's adventure stories, and dreamt of going out on adventures in a submarine like Captain Nemo. Undoubtedly, this is one dream that came true. Ballard's discovery enabled researchers to reexamine the circumstances that led to the disaster.
In 1998, Katherine Felkins, Phil Leighly and Alex Jankovic of Missouri-Rolla University published a seminal article which raised the possibility that the Titanic sunk because its designers had chosen the wrong steel. During a research trip in the North Atlantic n 1996, the three researchers brought up steel samples from the Titanic wreck, for metallurgic testing at the university laboratory. Their tests showed that the steel of which the Titanic had been made of contained excessive ratios of sulfur and phosphor, and that its sulfur content was too high in relation to its manganese content. Such steel becomes brittle at low temperatures. To be fair, this fact was unknown to the shipbuilders of that time, who made the optimal choice based on the information available to them. Models made of the Titanic's steel were then observed through an optical microscope and a scanning electron microscope (SEM) and compared to models made of a modern steel alloy – ASTM A36 – when impact tested under temperatures of between -67 and 354 degrees Fahrenheit. The tests found that the steel taken from the Titanic had undergone a ductile-brittle transition which made it unsuitable for navigating in the low sea temperatures (close to 28°F). The ductile-brittle transition temperature of modern steel alloys is significantly lower than that, and therefore more fracture resistant within the whole tested temperature range, and particularly under low temperatures such as those of the North Atlantic in April.

Image 11: The Titanic during its construction and embarkation on its maiden voyage on April 10, 1912.
What Caused the Challenger Disaster?
The Challenger disaster in January 28, 1986, was one of the worst in NASA's history: the space shuttle blew up in the air only 73 seconds after taking off from Cape Canaveral. Millions of spectators around the world watched the crash on live TV. Until that day, NASA prided itself on flying 55 successful manned shuttle missions. Media hype peaked, among other things for the fact that one of the Challenger's seven astronauts was a 37 school teacher named Christa McAuliffe, selected among thousands of volunteers to be "the first ordinary person in space". To a large extent, this had been a media gimmick to begin with. Christa was supposed to teach a lesson from space to students in the United States and the rest of the world, and later share her impressions from space with youngsters. Prior to the January 28 takeoff, the Challenger's launch had been postponed several times for technical reasons and due to stormy weather. On that fateful day, the shuttle exploded at 46,000 feet and broke to pieces in a cloud of fire and smoke, shocking thousands of spectators on the scene and millions worldwide. There were no survivors. The shuttle's debris fell into the Atlantic a few miles away from the launch site.

Image 12: The Challenger disaster in January 28, 1986.
The terrible failure was dramatically portrayed in the media, and led to the grounding of all shuttle missions for about 18 months. After the crash, Reagan appointed a presidential investigative commission headed by former Secretary of State William Rogers. One of the best-known members of this commission was the late Nobel laureate physicist Richard Feynman, considered one of the most brilliant scientists of the 20th century. Feynman was determined to seek the truth behind the disaster. The investigation, the omissions and oversights it exposed and its conclusions are described in his fascinating best-seller What Do You Care What Other People Think (1988). The commission's investigations found that engineers employed by Thiokol, a subcontractor hired by NASA to build the shuttle's booster rockets had warned prior to the launch of possible fuel leakage from certain parts installed in the boosters. Although their claims were checked, NASA Okayed the launch as planned. A Thiokol engineer named Al McDonald showed up uninvited to one of the commission's meetings and testified that Thiokol engineers had concluded that there was a relationship between low temperature and sealing problems in the shuttle's booster rockets. The night before the launch, company representatives approached NASA officials in an attempt to avoid launching below 53°F. At the morning of the launch, temperature at Cape Canaveral was an unusually low 32°F. NASA was astounded by the warning, and its executives claimed there was no clear evidence for any trouble. Under their pressure, Thiokol retracted its warning. Nevertheless, Mr. McDonald remained adamantly opposed to the decision, and kept on warning. Commission members were astonished his testimony, as it clearly showed that the o-rings' failure was facilitated by an organizational failure by NASA executives.
In view of the commission's findings, Feynman figured out the main cause for the disaster. Using a simple experiment, broadcast live on TV, Feynman demonstrated that the low temperature at the launch site was the culprit. He threw an o-ring identical (though much smaller, of course) to those installed in the shuttle into a glass of ice water. After several minutes in the frozen (32°F) water, he took it out and demonstrated that the rubber lost its elasticity and became brittle, just like the shuttle's o-rings. The main cause for the disaster was now evident to all: one of the o-rings installed in the Challenger's booster rockets became brittle. This o-ring seal, made of Viton rubber, shrunk due to the low temperature, creating cracks which enabled the fuel burning inside the rocket to leak. This fuel came into contact with another tank containing hydrogen and oxygen, causing the massive explosion. Image: Space shuttle Challenger on the ground, and its explosion shortly after takeoff.
Discussion: Could the Disasters Have Been Averted?
The preceding analysis of the Titanic and Challenger disasters has shown that at low temperatures, construction materials underwent a ductile-brittle transition which led to crack propagation and fatal failure.
In the case of the Titanic, recent research has found that high sulfur and phosphor content in its steel hull was responsible to its brittleness in the cold waters of the Atlantic. In the case of the challenger, the presidential investigative commission found that a critical o-ring which turned brittle and failed due to the low temperature at the launch site led to a fatal fuel leak. Importantly, the type of steel of which the Titanic was made of was not the only cause of its sinking. The ship sank due to a tragic series of contributing circumstances, including the collision with a gigantic iceberg at high velocity, low sea-water temperature, ductile-brittle transition of its steel hull, and faulty planning of the walls insulating its supposedly 16 watertight compartments. To these we must add the vanity of the Titanic's designers and crewmembers, all of whom were absolutely convinced it was unsinkable. Interestingly, the Olympic, launched in 1910 by the same shipyard, with a similar design and the same steel hull, served for over 20 years as one of three mammoth luxury liners built by White Star Line (including the Titanic itself).
Could a similar tragedy occur today?
Probably not: today we have advanced sonars, navigation aids and means of communication allowing crews to spot icebergs at a great distance and avoid such collisions. Even if a present-day ship had collided with such an iceberg, it is safe to assume the damage would not have been so great, since today's ships are designed with greater care, using engineering software and based on rigorous standards. Above all, our understanding of materials has advanced significantly. Today, we have a much deeper understanding of the relationships between materials' structure and composition and their properties and failure potential. The steels used nowadays are more ductile than the alloy used to build the Titanic's hull, with ductile-brittle transitions at much lower temperatures.
As for the Challenger, it was first launched into deep space in April 1983 and flew an additional eight low-orbit missions before exploding on its tenth launch. It was the first of two space shuttles to crash while on a manned mission – the second was the Columbia, on February 2003, with similar tragic results. The Challenger disaster led to a significant slowdown in manned space exploration programs. Like the Titanic disaster, it too was caused by a ductile-brittle transition, this time of a polymer (rubber). Similarly to the Titanic disaster, the explosion of the Challenger was also contributed to by human factors. According to investigative commission member Richard Feynman, NASA executives ignored warnings by engineers and took no steps to avert the failure. In his book, Feynman writes that when advanced technologies are at stake, no political considerations must twist our sense of reality and scientific truth. Thus, despite the important lessons learned from the Challenger catastrophe and the continued development of new materials, similar disasters will still occur in the future, if engineers do not take the limitations of their materials into account and politicians fail to heed their warnings.
Further Reading
Feynman, Richard P., What Do You Care What Other People Think?, 1988, W W Norton (1988). 

The article was first published at Galileo 103 (2007) (Hebrew).
I wish to thank Rafi Ashkenazi for his astute comments.

Dr. Dana Ashkenazi received her Ph.D in Mechanical Engineering from Tel-Aviv University and her first and second degrees in Materials Science and Engineering from Ben-Gurion University.She conducts research and lectures in the field of Materials Science and Engineering and also has great interest in popular science, art and human culture.