TWO years ago, on 20 October, the space shuttle Columbia lifted off from Cape
Canaveral, Florida, on a 16-day mission. On board were the usual high-tech
gizmos鈥攖remendous banks of communications and navigation equipment,
sophisticated life-support systems and safety mechanisms of every conceivable
variety. But the shuttle also carried a more mysterious device. Buried deep in
its hold, surrounded by lasers and video cameras, and fixed in a framework of
stainless steel supports, was a peculiar little jar of ordinary water. And in
the water, floating in the dark, were millions of tiny and precisely engineered
plastic beads.
Once the shuttle was in orbit, an astronaut had to crawl down into the hold,
stir the beads into a swirling confusion, and turn on the cameras and lasers.
You might think the US military was up to something funny鈥攖esting the core
of a new weapon, perhaps, or some weird computing device. But the truth is less
sinister. This experiment was designed by American and British scientists, its
aim in space to capture on film one of the most basic processes here on
Earth鈥攖he formation of crystals.
As we usually think of them鈥攊n grains of salt or sugar, sparkling
jewels or silicon chips鈥攃rystals are made of atoms or ions stacked in
regular arrays. And they are generally tough customers. They make up the stuff
of steel bridges and bones, and the teeth of tigers and diamond saws. But
microscopic plastic beads suspended in a liquid will also, under the right
conditions, spontaneously organise themselves into a regular array, exactly as
atoms do in a solid. What gradually formed in that jar on the space shuttle was
a strange kind of crystal so soft that it would be shattered by even the most
gentle shaking.
Advertisement
Scale models
Colloidal crystals, as they are called, may be the world鈥檚 feeblest crystals,
but in the past few years they have become a source of fascination for
physicists and engineers alike. On a scale ten thousand times larger than
ordinary crystals, they provide scientists with outsized toys to play with in
order to learn about the fundamental transformations of atomic matter. Colloidal
crystals also have some curious and distinctive optical properties that make
them glisten and sparkle like rainbow-hued gems. By exploiting these properties,
engineers are using them to build everything from pollution sensors to
components for a light-based information-processing technology for the 21st
century.
But one of the most surprising things about colloidal crystals is that they
exist at all鈥攆or there doesn鈥檛 seem to be anything holding them together.
In a simple crystal such as a piece of metal or a chunk of salt, attractive
forces between atoms or ions assemble them into arrays much like the stacks of
oranges on a greengrocer鈥檚 stall. Salt crystals, for example, are bound by the
attraction between positively charged sodium ions and negatively charged
chloride ions
(see Diagram).
In a colloidal crystal, however, the 鈥渁toms鈥� are nothing more than tiny beads
made from polymers such as polystyrene or latex. When suspended in solution,
very weak forces act between the beads. These so-called van der Waals forces
arise due to random fluctuations in the electron clouds of the constituent
atoms. On average, these interact to generate a weak attractive force, which
gives small colloidal particles a tendency to clump together.
But this clumping can be avoided, just by coating the beads with chemical
groups before putting them into the solution. These groups give each bead a net
electrical charge, which makes the beads repel one another. Depending on the
coating and the liquid in which the beads are suspended, the attracting van der
Waals and repelling electrostatic forces between beads can be tuned to give a
net attractive or repulsive force, or no force at all. In this 鈥渘o force鈥� case,
the spheres act like miniature billiard balls鈥攖hey don鈥檛 affect one
another at all unless they touch, and then they bounce off one another.
Frozen solid
You might think that a collection of such beads would swirl about
indefinitely in a disorganised cloud, like motes of dust in the air. But it
turns out that if the volume fraction of beads exceeds 0.494鈥攊f, in other
words, they occupy more than 49.4 per cent of the total volume鈥攖hen the
suspension begins to 鈥渇reeze鈥�. Tiny crystals of regularly stacked spheres appear
in the liquid and settle to the bottom. So what brings the particles
together?
The answer, in a word, is entropy. The entropy of a system is a measure of
the number of ways that its microscopic components can be arranged while leaving
it with the same large-scale properties. A glass of water, for example, has a
relatively large entropy because the precise positions and velocities of all its
molecules can be changed in innumerable ways while leaving the water鈥檚
temperature and pressure unchanged. A crystal has fewer choices, since its atoms
or molecules sit in a regular lattice. So a crystal has lower entropy than a
fluid. In general, entropy is a measure of disorder.
The second law of thermodynamics says that, for any spontaneous change, the
total entropy of the Universe must increase. So how then can a suspension of
noninteracting spheres change itself from a disordered fluid into an ordered
crystal? The secret is that not all of the system freezes: the colloidal crystal
coexists with a liquid component. As the fraction of the total volume occupied
by the beads increases, there comes a point when the system can increase its
entropy by packing some of the beads together into crystallised clumps, in which
the beads lie closer to one another and collectively occupy less volume than
before. This frees up room for the remaining, noncrystallised particles. When
the volume fraction is high enough, the entropy increase of these free-swimming
beads outweighs the entropy loss caused by the crystallisation. To increase the
overall entropy, crystallites begin to form.
So despite what you may think, the formation of a crystal from a liquid is
not just about atoms or molecules being bound together by attractive forces.
Entropy is not a force, but it can act like one. Colloidal crystals form even
when the chemistry of the bead coating and liquid is tuned to make a strongly
repelling net force. On account of entropy, crystallisation is inevitable.
These strange crystals provide scaled-up models for studying real crystals.
But to find them useful, scientists need to know the kinds of crystal structures
they form. For any collection of spheres, the densest possible packing is
produced by arranging them in hexagonally packed planes, with each sphere
surrounded by six others
(see Diagram).
A second layer sits atop the first
in such a way that its spheres hang above the gaps between spheres in the first
layer. But the third layer can be arranged in two ways relative to the first: it
can either sit directly above it, or slightly displaced to one side. In the
former arrangement, called hexagonal close packing (hcp), the stacking sequence
repeats every other layer. The latter arrangement gives a face-centred-cubic
(fcc) lattice, which repeats every third layer. Both arrangements have a volume
fraction of 0.74, which is the highest possible for identical spheres.
So which stacking sequence do colloidal crystals adopt? Experiments show that
they have trouble making up their minds. The spheres tend to crystallise neither
as pure hcp or pure fcc structures, but in an arrangement that has some regions
of fcc packing and others of a hybrid packing in which the three fcc layers
recur without any regular pattern. This is called random hexagonal close packing
(rhcp). It seems that tiny disturbances that occur while the fragile crystals
form, for example currents and eddies in the fluid solvent, are enough to
suppress any preference for a perfectly ordered structure.
This is where the space shuttle comes in. Paul Chaikin of Princeton
University and colleagues from Bristol University in Britain and NASA sent a
colloidal suspension into space on Columbia to discover how a colloidal crystal
grows in an ideal, undisturbed environment. For it is gravity, you see, that
messes things up.
In liquids on Earth, small temperature variations set up during normal
crystal growth cause tiny differences in liquid density, which, under the force
of gravity, cause the fluid to flow. Gravity also causes crystals, once formed,
to settle to the bottom of the vessel. Their motion through the fluid also
affects the growth process. The weightless environment of space suppresses such
effects, making it possible to study colloidal crystals as they would like to
grow, free from interference.
In Chaikin and his colleagues鈥� experiments, astronauts mixed suspensions with
volume fractions greater than the crystallisation threshold and left them to
crystallise. Last June, the group reported that their video cameras and lasers
had showed that for a volume fraction close to the freezing transition, their
crystals adopted purely the hybrid rhcp structure鈥攖here was no trace of
the fcc structure found in colloidal crystals formed on Earth.
This is important because it means that the randomness seen in crystals
formed here on Earth isn鈥檛 caused merely by gravitational disturbances, but
reflects an inherent preference of the crystals. And getting to grips with such
preferences is crucial if colloidal crystals are to be put to technological use.
Techniques for making tiny uniform polymer beads, between 0.1 and 1 micrometre
in diameter, were developed in the 1950s when they were intended for use in
paints and plastic coatings. But now scientists have invented a welter of much
more sophisticated applications, most of which centre around the fact that
colloidal crystals strongly scatter visible light. This raises the possibility
of using them to control and manipulate the flow of light.
Turn out the light
In the early 1990s scientists predicted that a regular, repeating array of
objects of any kind that scatter light can have a 鈥減hotonic band gap鈥�, meaning
that light having wavelength within a particular range will not pass through the
array. The forbidden wavelengths would be on the same scale as the spacing
between the scattering centres, because light scatters most strongly from
objects that have features about the same size as the light鈥檚 wavelength.
The existence of these photonic band gaps has been confirmed in 鈥減hotonic
crystals鈥� made by drilling a regular web of tiny holes in blocks of insulating
materials, or by arranging tiny pillars of material into regular lattices. The
lattice spacing in these structures is of the order of millimetres to
centimetres, so their band gaps occur in the microwave part of the
electromagnetic spectrum.
Far more useful would be materials with photonic band gaps in the visible and
near-infrared range, since they could be used to manipulate light for optical
telecommunications. Introducing a defect into the lattice鈥攕uch as a
missing row of scatterers鈥攃reates a waveguide. The light that falls
within the photonic band gap can travel along the fault, but can鈥檛 get out into
the surrounding material, where the band gap persists. Linear defects such as
this confine light like optical fibres. Individual, localised defects,
meanwhile, act as optical cavities, confining light in a kind of trap. Optical
cavities like these might make possible new types of miniaturised lasers.
Millions of holes
But making such devices would require drilling millions of perfectly spaced
microscopic holes into a material. And to get the band gap in the visible part
of the spectrum, the holes would have to be no more than one-thousandth of a
millimetre apart鈥攁 formidable challenge. But now researchers have realised
that colloidal crystals might provide photonic-band-gap materials that
assemble themselves, since they contain periodic arrays on just the right
scale.
In 1996 Inanc Tarhan and George Watson of the University of Delaware in
Newark showed that silica spheres packed into an fcc structure have a photonic
band gap at visible wavelengths. Tarhan and Watson achieved the fcc structure by
careful selection of the bead size, surface charge and volume fraction.
Generally, however, researchers want to be able to encourage colloidal crystals
to adopt a wide range of packing arrangements, so that the photonic properties
can be tailored to order. Finally, last January Alfons van Blaaderen from the
Institute for Atomic and Molecular Physics in Amsterdam, along with Pierre
Wiltzius and Rene Ruel from Lucent Technologies in New Jersey, found a way to do
it.
The group used a technique that has been familiar to materials scientists for
over a decade: epitaxial crystal growth. This involves growing a crystal on an
underlying substrate with a periodic structure, which acts as a template.
Epitaxial growth is widely used to ensure that thin films of semiconducting
materials grown from a vapour form well-ordered crystalline layers. Van
Blaaderen and his colleagues realised that this idea can be scaled up to control
the growth of a colloidal crystal from a suspension. They made a patterned
template by coating a glass plate with a polymer film, then cutting holes in the
film using an electron beam. The holes were arranged with a structure and
spacing that corresponded to the positions of spheres on one facet of an fcc
crystal. The result was a microscopic 鈥渆ggbox鈥� ready to take a load of spherical
silica 鈥渆ggs鈥�
(see Diagram).
To ensure that the crystals formed only on this template and not within the
bulk of the suspension itself, the researchers used a volume fraction well below
that at which crystallisation would normally occur. The template was enough to
initiate the growth. They found that, instead of forming the hybrid rhcp/fcc
crystals that grow under normal conditions, the crystal that formed on the
template had a perfect fcc structure: the order imposed on the first layer was
sufficient to ensure that subsequent layers maintained the fcc arrangement.
Channel hopping
The researchers used other templates to obtain different structures. In
particular, they were able to grow channels of specific crystal structure by
using two templates separated by a gap. If the gap was large
enough, the center crystal had a different structure from that at the
sides. The controlled fabrication of such channels will be important for making
optical waveguides.
At the University of Pittsburgh, Sanford Asher and his team have other uses
for colloidal crystals in mind: as switches and sensors. Because the crystals
diffract light at specific wavelengths determined by the spacing between
spheres, the colour of the crystals can be varied by changing this spacing.
Asher has devised a way to reversibly expand and shrink a colloidal crystal by
embedding it in a polymer gel that swells and contracts in response to certain
stimuli.
Gels like this are examples of smart materials鈥攖hey respond to their
environment鈥攁nd are also being developed for use as synthetic valves and
muscles in medicine and robotics. Smart gels are networks of chainlike polymer
molecules which switch between a collapsed state, in the shrunken gel, and an
extended state in the swollen gel. The transition between collapsed and extended
polymer chains can be induced by changes in the temperature, acidity or ionic
strength of the solvent that permeates the gel network.
Asher鈥檚 idea was to create a colloidal crystal within a gel and then let the
volume changes of the gel alter the lattice spacing of the crystal. His group
has made polymerised colloidal crystalline arrays by first forming a colloidal
crystal and then polymerising the components of the gel around it. They showed
last year that temperature-induced changes in the gel volume would alter the
crystal鈥檚 colour by changing the wavelength of light scattered at a certain
angle.
The king of sensors
And in the 23 October issue of Nature, Asher and his colleague John
Holtz describe a chemical-sensing material that uses the same principle to
detect the presence of specific metal ions in solution. It works like this. They
attached chemical groups called crown ethers to the chains of the polymer gel.
Crown ethers are rather like rings into which only metal ions of the right size
can fit. When the crown ethers capture metal ions, they change the ionic
strength of the solution, making the gel swell, and altering the colour of the
colloidal crystal. So for crown ethers that selectively bind lead ions, for
example, the colour change is seen only in solutions of lead salts and not for
iron, copper or cadmium. These new materials might therefore be useful for
monitoring heavy metal pollution.
The pair have also made a colour-change sensor that detects glucose, by
attaching the enzyme glucose oxidase to the gel. The chemical reaction between
the enzyme and glucose molecules induces swelling in the gel. Monitoring glucose
concentrations is extremely important in the treatment of diabetes, though it
may be too early to say whether this composite material will lead to sensors as
good as the glucose biosensors already in clinical use. With so many
applications, there can be no doubt that these materials have a lot more to
offer than you might expect from the world鈥檚 feeblest crystals.