ÐÓ°ÉÔ­´´

Syrup does not spring instantly to mind as the inspiration for designing
aircraft engines. But just as sugar dissolves in water to make syrup, many
metals will dissolve in other metals to form solid solutions: brass is a
familiar solid solution of copper and zinc. In brass, as in many metallic
alloys, the atoms of the two or more constituents are arranged randomly,
in no particular order. But with some combinations of metals, the different
types of atom arrange themselves in a regular pattern to produce a compound
called an ‘intermetallic’. Such compounds have been the focus of much recent
attention by materials scientists, who hope they will turn out to be useful
as structural materials resilient enough to withstand the most extreme
conditions.

This is just the latest twist in a long story: the search for strong
metals. As long ago as 1709, the Midlands ironmaster Abraham Darby was looking
for a way to improve iron when he hit upon the idea of smelting it with
coke instead of charcoal. Cast iron, the new and extraordinarily strong
material that this produced, revolutionised construction, beginning with
the first iron bridge at Coalbrookdale in Shropshire. Metallurgists have
been on the trail of new materials ever since. A big problem they face today
is that materials which are strong at room temperature often turn out to
behave quite differently when they become hotter.

What makes any metal useful as a structural material, apart from its
strength, is its ability to deform to a new shape without cracking – its
ductility. A material which can be stretched by more than about a tenth
of its length can be described as ductile. But strong materials – those
that can resist a large force without breaking – are not always ductile.
At normal temperatures, pure metals are ductile, but they have little strength;
alloys such as brass are not quite as ductile as pure metals, but they are
stronger. Intermetallics, also known as ordered alloys, are often strong
at normal temperatures, but have poor ductility – in other words, they are
brittle.

Ancient remedies

Intermetallics are not new. By AD 660 the Chinese were using an intermetallic
containing silver, tin and mercury to fill teeth. The silver-mercury and
silver-tin intermetallics, Ag2Hg3 and Ag3Sn, with small amounts of tin-mercury
(Sn2Hg) and copper-tin (Cu5Sn6), are still used as dental fillings. But
the real promise of intermetallics, as structural materials, was not recognised
until 1909, when the German metallurgist Alfred Wilm made intermetallic
particles of the aluminium-copper compound Al2Cu in a piece of aluminium,
by dissolving copper in molten aluminium and ageing it at room temperature.
He noticed that this material was stronger than aluminium. A decade later,
Al2Cu was incorporated into aluminium alloys used in aircraft for exactly
the same reason. By 1919, the German metallurgist Gustav Tamman had shown
that intermetallic compounds were distinct from disordered alloys by virtue
of the arrangement of their atoms. Intermetallic compounds are not always
beneficial. Heating stainless steels to between 500 °C and 800 °C
makes them brittle, because an iron-chromium intermetallic compound, called
the sigma phase, forms at these temperatures.

Today, intermetallics are best known for other properties. The soft
magnet Permendur, which is based on the iron-cobalt intermetallic FeCo,
is used for transformer cores, while the samarium-cobalt compounds SmCo5
and Sm2Co17 are hard magnets used in motors in electronic wristwatches.
A gold-coloured palladium-indium intermetallic, PdIn, could replace more
expensive gold crowns in dentistry, while the creation of a purple compound
from gold and aluminium in Au2Al may give watchmakers a decorative new material
for watch cases. Previously, Au2Al had been thought of only as a nuisance:
the formation of brittle Au2Al by interdiffusion where gold wires were
connected to aluminium films was such a problem in the semiconductor industry
in the 1960s that it was known as the ‘purple plague’.

During the past 15 years, metallurgists have renewed their interest
in the structural properties of intermetallics. Driving much of the research
is the need to find materials which can be used at temperatures higher than
1400 °C – the limit of the materials used at present for gas turbine
engines on aircraft. As well as having to withstand fiercely high temperatures,
these new materials must be lighter, stronger and more resistant to oxidation
than those in use now.

In their properties intermetallics lie somewhere between metals and
ceramics. They can have much higher strength and resistance to oxidation
than metals but are in general not quite as light, strong at high temperatures
or oxidation-resistant as some ceramics. However, in the short term, intermetallic
compounds look more promising for structural applications than the extremely
brittle ceramics.

Metal cocktails

Turbine blades and discs are currently made from materials such as nickel-based
superalloys and titanium alloys. Intermetallic compounds of aluminium with
titanium and nickel, such as TiAl and NiAl, and of molybdenum with silicon,
such as MoSi2, have attracted much recent attention for these applications.
Unlike pure titanium or nickel or molybdenum, these materials have in-built
protection from attack by oxygen. When they are attacked, some of the aluminium
or silicon they contain turns into a protective coating of aluminium oxide
or silicon oxide.

The search for materials with all the advantages of these alloys, but
which could also withstand prolonged exposure to high temperatures, began
in the 1950s. By 1959, Ray Guard and Jack Westbrook of the General Electric
Company in Schenectady, New York, had discovered that the nickel-aluminium
compound Ni3Al doubles its strength on going from room temperature to 700
°C, but that at higher temperatures than this, its strength starts
to decrease. This material looked promising as its behaviour was the exact
opposite of metals, which gradually lose their strength as they become hotter.
Researchers have since discovered other compounds, such as Ni3Si, TiAl and
FeAl, which behave similarly. The work went on throughout the late 1950s
and early 1960s. Then the trail went cold.

The problem materials scientists faced seemed insurmountable. They could
find intermetallics which became much stronger as the temperature increased.
But they were all far too brittle at room temperature to be of any practical
use. The turning point came in 1977, when two Japanese researchers reported
that they had succeeded in improving the ductility of Ni3Al. Kiyoshi Aoki
and Osamu Izumi, working at Tohoku University, reported that by ‘doping’
Ni3Al – adding tiny amounts of boron – they could increase its ductility
to the point where it could be elongated by 35 per cent at room temperature
before fracturing. In 1982, a group led by Chain Liu at Oak Ridge National
Laboratory in Tennessee took this approach a stage further and succeeded
in stretching the same material by half its original length. The key to
Liu’s success was to control carefully the composition of the material,
so that the ratio of nickel to aluminium was 76:24, and then to add around
500 parts per million of boron to it.

Materials scientists, most of them in the US, have been puzzling over
why this should work, and two main theories have now emerged. But to understand
them, we first need to look more closely at what makes an ordered intermetallic
compound different from a metal or a disordered alloy. The main difference
lies in the way it deforms. When a force is first applied to a material
such as a metal its atoms are simply pulled apart. If the force is then
removed, the atoms spring back to their original positions. If the force
increases, two things can happen. If the material is ductile, it deforms;
if brittle, it cracks.

Metals, disordered alloys and ordered intermetallics are all able to
deform as irregularities in the relative positions of the atoms, called
dislocations, move through them. Under a large enough stress, these dislocations
move and the lines of atoms shift. In this way the material can change shape
without breaking.

In disordered alloys and metals, single dislocations can move fairly
easily because there is no order to disrupt (see Figure 1). This is why
these materials tend to be ductile. But in an ordered structure such as
an intermetallic, moving single dislocations could cause havoc. For this
reason, dislocations often move in pairs. This works because the region
between such pairs – called an antiphase domain boundary – contains lines
of atoms which are ordered just like the rest of the crystal, although
out of step by one atom spacing (see Figure 2). Within this region the
atoms are in a high energy state – having been forced out of their preferred
arrangement in which different types of atom alternate and into an arrangement
where atoms of the same type lie next to each other – but outside it the
order of the crystal is not disrupted. Intermetallics tend to be brittle
because it is harder to move a pair of dislocations, with its antiphase
domain boundary, than a single dislocation.

Demystifying ductility

Could researchers fit this explanation of brittleness to what they knew
from experience: that doping intermetallics with boron makes them much
more ductile? The first theory, put forward in 1985 by Liu’s group, suggests
that the individual crystals which make up Ni3Al are bonded together only
weakly. As a result, a force perpendicular to the boundary will pull the
crystals apart before dislocations start to move. The boron atoms – which
researchers already know tend to collect at grain boundaries, where these
crystals meet – increase the density of electrons between atoms near the
boundaries. This increases the strength of the existing bonding, improving
the cohesive strength of the grain boundaries and making the material ductile
instead of brittle, since then the crystals cannot easily be pulled apart.

A year later, Erland Schulson, myself and our colleagues at Dartmouth
College put forward an alternative proposal. We suggested that transmitting
deformation through Ni3Al is difficult because atoms are ordered in the
region around the boundaries between the crystals that make up the material.
According to our theory, the doping process – adding boron – alters the
chemistry and bonding of these grain boundaries, so that they become disordered,
like those in metals. Then small, unpaired dislocations, which can move
easily, can exist in this region and the material becomes ductile.

Exploiting the boron effect

There is evidence to support both models. But whichever is the correct
one, the practical successes of making Ni3Al more ductile by adding boron
have prompted many metallurgists to try adding boron to other intermetallic
compounds to improve their ductility. They have found that this works for
iron-aluminium (FeAl) and nickel-silicon (Ni3Si) intermetallics but makes
little difference to the nickel-aluminium compound NiAl.

Claudette McKamey at Oak Ridge has developed the iron-aluminium compound
Fe3Al as a material for making the tubing and piping that carry exhaust
gases in power stations burning fossil fuels. Here, the biggest problem
designers face is corrosion: such materials must be resistant to attack
by oxygen and by the sulphur in the gases produced when fossil fuels burn.
Similar problems face designers of chemicals plants.

In the late 1980s this challenge spurred on two independent groups of
researchers, one at Oak Ridge and led by Warren Oliver, the other headed
by Takayuki Takasugi at Tohoku University in Japan, to develop the nickel-silicon
compound Ni3Si. This material has an astonishing resistance to attack, not
only by oxygen, but also by hot sulphuric acid, because in these conditions
it forms a protective silica coating. A material such as Ni3Si could eventually
replace more expensive but commonly used alloys such as stainless steels
or chromium-nickel-cobalt.

Beating corrosion

Another potential replacement for stainless steel is the iron-aluminium
intermetallic FeAl. Like Ni3Si, it is cheaper and far more resistant to
oxidation than stainless steel. The aluminium it contains readily forms
a protective layer of aluminium oxide (Al2O₃) when attacked by
oxidising agents. The snag is that FeAl is brittle in the atmosphere, because
hydrogen formed by the reaction of aluminium with water vapour encourages
cracks to grow. Liu’s group at Oak Ridge is searching for alloying elements
to prevent the hydrogen attack.

Corrosion is not the only problem being tackled with new intermetallics.
The nickel-titanium compound NiTi, for instance, is under study by several
groups for its ‘shape memory’ effect. This is the curious ability to ‘remember’
the shape it took on at low temperatures. If deformed to a new shape this
material will spring back to its original shape when it is heated. This
useful property is already being exploited in some car engines and in hydraulic
couplings in the Grumman F14 fighter aircraft. It is also used in greenhouses,
to open and close windows automatically, according to temperature.

The possible combinations of metals that could form intermetallics seems
endless. But perhaps the compound with most short-term promise is Ni3Al
containing boron and small amounts (a few atoms in every hundred) of other
elements such as chromium, hafnium and zirconium. With its great strength
and oxidation resistance at high temperatures, this material is already
being developed by several companies. Valley-Todeco of Sylmar, California
is using it to make the fasteners for aircraft; Hoskins of Hanbury, Michigan
uses it in heating elements for furnaces; Metallamics of Traverse City,
Michigan, uses it in dies for pressing permanent magnets; and Cummins of
Columbus, Ohio, uses it for rotors for turbochargers in diesel engines.
Resistance to physical erosion and chemical corrosion makes this material
a good candidate for gas, water and steam turbines.

Researchers, in the US, Japan and Germany, are also interested in the
rather brittle titanium-aluminium compound TiAl for applications such as
compressor discs in the relatively cool (between 1600 and 1000 °C) front
end of jet engines. TiAl is lighter, stronger and more resistant to oxidation
than the titanium alloys used now. However, it seems likely that this material
will make its debut in a more down-to-earth application: in 1990, Yukio
Nishiyana and co-workers at Kawasaki Heavy Industries in Akashi, Japan,
successfully made a turbocharger rotor out of TiAl.

Finally, compounds such as the oxidation-resistant aluminides and silicides
could find themselves in components under high stress where strong but brittle
load-carrying fibres or particles – typically ceramics such as silicon carbide
or alumina – are enmeshed in a so-called metal matrix composite. Usually,
this supporting matrix is made from an alloy, of aluminium, for example,
but several research groups in the US are investigating how these can be
replaced with intermetallics. Silicon carbide fibres with carbon cores,
embedded in a matrix of Ti3Al, for instance, could be used for turbine blades
or discs.

Intermetallics may look similar to existing alloys, but their properties
are far superior. As design engineers discover more about these strong new
materials they will begin to appear in the marketplace, and the performance
of car and aircraft engines, power stations and chemical plants could all
see dramatic improvements as a result.

Ian Baker is associate professor in the Thayer School of Engineering,
Dartmouth College, Hanover, New Hampshire.

* * *

HOW STRONGER METALS SHAPE UP

The solubility of one metal in another ranges from almost negligible
(lead and aluminium) to complete mutual solubility (silver and gold). Most
metals lie between these two extremes. Nickel, for example, can dissolve
up to 21 atomic per cent of aluminium (21 atoms in 100 are aluminium) at
1385 °C, although aluminium will dissolve only minute amounts, 0.11
atomic per cent at most, of nickel, at any temperature. An intermetallic
forms between two elements – aluminium and nickel, for example – because
atoms of these two elements have a greater affinity for each other than
for atoms of their own type.

These two elements form the inter-metallic compounds Al3Ni, Al3Ni2,
AlNi, Al3Ni5 and AlNi3, whose existence has long intrigued chemists. The
compounds Al3Ni2 and Al3Ni exist only with compositions which reflect the
exact aluminium-to-nickel ratios of 3:2 and 3:1, respectively, but the others
can exist over a wider range of compositions. NiAl, for example, can contain
from 31 to 58 atomic per cent aluminium; in other words, NiAl can itself
dissolve a substantial amount of both nickel and aluminium.

The crystal structures of most intermetallic compounds remain ordered
until they melt. A few can swap between an ordered and a disordered state,
however, depending on temperature. Cu3Au, for instance, is disordered above
390 °C, when each atom site in the face-centred cubic crystal structure
has a 75 per cent probability of being occupied by a copper atom and a 25
per cent probability of being occupied by a gold atom. At lower temperatures,
Cu3Au has an ordered face-centred cubic structure with gold atoms at the
corners and copper atoms occupying positions at the centre of each face.
Heating the crystal brings about the change from order to disorder; cooling
reverses the process.

Why should a compound take on an ordered state at one temperature but
a disordered state at another? At low temperatures, an intermetallic compound
such as Cu3Au forms because copper atoms ‘prefer’ to bond to gold atoms
rather than to other copper atoms. The energy released when the atoms bond
is what lies behind this preference. More energy is released when copper
atoms bond to gold atoms (in a three to one ratio) than when either of these
elements bond to their own type. But as the temperature increases, entropy
starts to take over. When it eventually outweighs the energy benefits of
bonding between copper and gold atoms the material becomes disordered.

For many such compounds, cooling does not bring about ordering everywhere
in the crystal. Instead, numerous small ordered regions form in much the
same way as sugar crystals grow in cooling water (and sink to the bottom
when they become too heavy to stay in suspension). These regions of order
all have the same crystal structure. But when they meet, the arrangement
of atoms in adjacent regions, or domains, may not line up. The line where
they meet is the antiphase domain boundary.