Michael W Barsoum & Tamer El-Raghy. American Scientist. Volume 89, Issue 4. Jul/Aug 2001.
If one could increase the average engine temperature of the world’s jets by one degree Celsius, the fuel savings alone would be worth around $1 billion per year. Using a similar strategy to increase the efficiency of our vehicle fleet by 3 miles per gallon would save roughly a million barrels of oil a day. These stunning figures reflect a simple fact of thermodynamics: The efficiency of any fuel-burning engine is directly proportional to its operating temperature. A jet engine made from a material 50 percent lighter in weight and able to run 200 to 300 degrees hotter could have a staggering economic impact. Today’s jet engines are not running hotter for a simple reason: No material exists that can take that kind of heat while spinning furiously. The same goes for the internal-combustion engine; if an automobile engine could be built with a temperature-tolerant material, its radiator, water pump and cooling water could be thrown away. Such an efficient, lighter and higher-temperature engine would squeeze more miles from every gallon of fuel.
Roughly three decades ago, most materials scientists appreciated that above 1,000 degrees Celsius even the best metal alloys turn into taffy and become useless for load-bearing applications. Building hotter engines would require employing another class of materials, namely ceramics. This realization ignited a worldwide research and development race that to date has cost billions of dollars. And although there has been undeniable progress, the wide use of ceramics as structural materials in the aerospace, automotive and chemical industries remains elusive. It is fair to say that had ceramics not been so hard or brittle, they would have been more useful. The glibness of this statement becomes apparent, however, when it is appreciated that properties come in bundles: Hardness, brittleness and poor machinability are inexorably linked to good high-temperature properties. One cannot get one without the other, or so it was believed.
Kink bands are common in nature and are familiar to most people. They typically form when layered materials are loaded parallel to the layers. To form a kink band, stand a deck of cards on end and carefully add weight on top. For small loads, the cards will buckle elastically (that is, if the load is removed, the cards snap back to their original shape). However, at a maximal load the cards will buckle irreversibly, forming one or more kink bands where the layers bend sharply and remain bent. Kink bands have been invoked to explain deformation in wood, polymer fibers, polymer composites, phone books and geologic formations, among others. In crystalline solids, kink bands are formed by creating edge dislocations in the center of what eventually becomes the kink band. These dislocations then move in opposite directions and eventually arrange themselves in walls, which define the kink boundaries.
In most cases, the formation of kink bands in a material results in a decrease in mechanical integrity. The fundamental and crucial difference between the kink bands that form in the MAX phases and those in other materials is that even after the initiation of damage, the ternary phases can still carry substantial loads. This so-called damage tolerance was initially quite puzzling and highly unusual. For most materials, once damage is initiated, the point of damage becomes weaker, resulting in further damage, which leads to further weakening and rapid failure. For example, when a graphite-reinforced tennis racquet or ski develops a crack, it is time to replace it.
Working with Igor Levin of the National Institute of Standards and Technology, and the transmission-electronmicroscopy expert in our group, Leonid Farber, we were able to shed some light on the problem. Simple geometry dictates that for a kink band to form, part of the crystal that is forming the kink band must be able to glide relative to the part that is not. Furthermore, the layers of the two parts must detach from each other. Going back to the card-deck analogy: For the deck to kink, the cards have to simultaneously be able to slide relative to each other and to delaminate. Had the cards been glued together, or clamped normal to the applied load, they would not have kinked.
Containing the Damage
This analysis explains how the kink bands form, but it does not totally explain why the MAX phases are so extraordinarily damage-tolerant. The transmission electron micrographs have helped reveal the answer. For a delamination crack to extend in one direction, it has to move all the dislocations in the kink boundary ahead of it-an energetically very expensive proposition. Extension in the other direction, however, is easier, since there are no dislocations in its path. The end result is that the crack is not only asymmetrical, but, more important, it also is stopped by the kink boundaries. We like to think of the kink boundaries as containers or reflectors of damage. After the damage is initiated, the damaged area locally “hardens,” rather than weakens, and the damage migrates to another region.
These materials blend two powerful ideas that have emerged in materials science. The first is that of nanoengineered solids, in which reducing the scale of the microstructure is believed to result in unusual and unexpected benefits. In the past, nanolayered materials were typically fabricated by laying those layers down one at a time, using the very expensive technique known as molecular beam epitaxy. In addition to being quite costly, layers formed in this way had another major disadvantage in that they could not be used at elevated temperatures, since that resulted in their intermixing. The MAX phases benefit from the special properties arising from nanolayering but do not have this disadvantage.
The second idea is that of biomimetics, the attempt to capture the properties of the splendid designs that have evolved over millions of years in nature. A good material for engineers to mimic is the abalone shell, which is mainly composed of a brittle calcium carbonate and yet is quite tough. This toughness arises from a submicron polymer film that lies between the layers of the calcium carbonate. The polymer essentially behaves like the crack-stopping ligaments mentioned above. The microstructural similarities between the fractured surfaces of an abalone shell and some of the micrographs we obtained in Ti3SiC2 are noteworthy. The layering in abalone, however, is at a coarser scale than the layered compounds discussed herein. Another crucial distinction is that the properties of abalone are optimized for normal temperatures at the earth’s surface. Heating an abalone shell to a few hundred degrees destroys the polymer and concomitant toughness.
A metal’s machinability is inextricably and inversely linked to its performance at higher temperature. The same mechanisms that render metals useful at higher temperatures (namely, obstacles in the way of the dislocations) are the ones that make them difficult to machine. Superalloys, difficult and expensive to machine, are a case in point. Ceramics are even more difficult to machine because they are harder and more brittle. One of the joys of working with these new ternary compounds is their ease of machinability. Regular high-speed tool bits can be used with no lubrication or cooling. They are the only carbides or nitrides that can be “sliced or diced” with a manual hacksaw. It is interesting that they do not machine by plastic deformation, as in the case of metals, but rather by the breaking off of tiny microscopic flakes. A good analogy here is that of scooping ice cream versus shaving ice; the former is like machining metals, the latter like working with the MAX compounds.
The fact that these solids machine readily is of technological importance for three reasons. First, easy machinability allows for the fabrication of relatively cheap prototypes, which in turn means it will be possible to test the materials for a variety of applications. The second reason has to do with tight tolerances. In many applications, such as engines, the tolerances required are more stringent than can be fabricated, and a post-fabrication machining step is typically required. The third has to do with joining. Joining ceramics to other materials is a nontrivial exercise fraught with pitfalls. Machinability allows for alternative solutions. For example, the tubes used by the petrochemical and chemical industries can be quite long (about 12 meters) and pose a serious processing challenge to any traditional ceramic material that currently exists. Machinability greatly simplifies the problem: 1- or 2 meter lengths of tubes can be extruded and their ends threaded and screwed together, end to end, to any desired length.