Orderly arrays of microscopic balls can make light behave in unusual - and often useful - ways, Philip Ball reports.

Orderly arrays of microscopic balls can make light behave in unusual - and often useful - ways, Philip Ball reports

Geoff Ozin likes talking balls. He is convinced that the future is in balls, and his office walls are covered with pictures of them. They sit in little boxes, stacked neatly like marbles. They fill up long trenches, they cover flat surfaces, they lie in arrays of pits. They look pretty and pleasing in their geometric orderliness. 

Then you see the scale bars on these photos, and realise that the balls are less than a micrometre across. A hundred of these spheres laid end to end like peas in a pod will just about span the width of a human hair. 

Ozin’s group at the University of Toronto is one of several that are attempting to cajole these microscopic balls into arrangements that could transform information technology. They believe that regular, crystalline arrays of microspheres will make it possible to mould and shape light, guiding optical signals around a silicon chip just as conventional chip circuitry directs the flow of electrons. 

That would allow the light pulses currently used to send information down fibre-optic cables to be brought seamlessly onto microchips, where the signals can be processed and stored. Then there would no longer be any need for the cumbersome interconversion of electrical signals and light whenever computers interact across a network. It could all be done with photons: photonically. 

This could make information technology faster and more compact. But little balls that control light could find a spectacular variety of other uses too. Ozin’s group has recently used them to make a prototype ’photonic ink’ that can be switched electronically between different colours. Others are using microsphere arrays to make delicate chemical sensors, capable of spotting small concentrations of pollutants such as lead or medically important biochemicals such as glucose. They might be used in switchable windows that darken when the light gets too bright. 

Under control 
Nature got wise to photonic arrays long ago. The iridescent blue colour of some butterfly wings is produced by a grid of tiny ridges on the wing scales, spaced just 1mm or so apart. This acts as a diffraction grating that selects a particular colour by interference. The same effect creates the blues and greens of a peacock’s tail feathers. 

To find a natural analogue of the microsphere arrays now being developed for technological ends, we need look no further than the mineral opal, which is formed by the precipitation and sedimentation of ?m-sized particles of silica. Being all of similar size, these particles stack together in an orderly manner to produce a so-called colloidal crystal with the pearly, rainbow-like appearance that we call opalescence. 

Curiously, researchers didn’t at first think of making optical materials by mimicking opal. They thought it would be much harder to pattern materials like this at the microscopic scale. 

But when Eli Yablonovitch of Bellcore in New Jersey and Sajeev John of the University of Toronto conceived of photonic crystals in 1987, it wasn’t exactly a diffraction grating that they had in mind. They realised that Bragg scattering - the interference of light scattered from a periodic array of objects - can make such an array impervious to light. 

Imagine shining a torch through a forest of trees. If it is a microscopic forest, so that the spacing between trees is of the same order as the wavelengths of light, and if moreover the forest is an artificial plantation in which all the trees stand in equally spaced rows, then Bragg scattering can prevent light in a certain wavelength band (related to the ’tree’ spacing) from penetrating the forest at all. The beam is stopped at the forest’s edge. 

This array of pillars is a two-dimensional photonic crystal: its periodicity is two-dimensional, and it exerts its light-filtering behaviour on beams propagating parallel to the ground - that is, in the plane of the lattice. The band of excluded wavelengths is called a photonic band gap. 

A crystalline arrangement of microspheres presents a three-dimensional array of scatterers, which can in principle banish light that is travelling in any direction. Just a few layers of microspheres are enough to set up a good photonic band gap - that is, to erect a firm barrier to light - so photonic structures like this can be very small. Ozin is developing methods for patterning microspheres on silicon surfaces in order to make photonic ’opal chips’. In 2000 he and postdoctoral student San Ming Yang described a procedure for marshalling microspheres into orderly arrays in channels just a few ?m wide on silicon wafers. 

Ozin and Yang began by creating V-shaped troughs to hold the spheres. They imprinted a pattern of organic molecules on the silicon surface using a technique called microcontact printing. The ’inked’ areas are resistant to etching, while exposed areas are eaten away to produce the grooves. This is a cheaper and simpler process than the optical lithography used for patterning silicon chips for commercial microelectronics. 

Then the researchers closed off the tops of the grooves with a block of polymer, and allowed capillary forces to suck a suspension of microspheres into the channels. The spheres fill up the troughs layer by layer with the [100] crystallographic planes parallel to the silicon surface. 

Seen from above, the spheres pack into square arrays. This is important. The width, depth (that is, impermeability) and orientation of the photonic band gap in a close-packed array of spheres depend on how the balls are stacked. Some researchers have tried to control this stacking by using lithography to make a kind of ’egg-box’ surface that nucleates the crystals in a particular crystallographic orientation. But Ozin can exercise control simply by changing the geometry of his etched grooves. Under the etching conditions he and Yang used, the channels have a slope of about 70?, into which the face-centred cubic arrangements fits perfectly with the [100] plane uppermost. 

Ozin and Yang were also able to fill isolated etched pits with microspheres by spinning the silicon wafer to spread the suspension evenly over the surface and let the balls settle into the pyramid-shaped holes. They can alter how the spheres pack merely by changing the proportions of the pits: like marbles, the spheres seek the best fit to the shape of their confining walls. Younan Xia and colleagues at the University of Washington in Seattle have shown recently that a silicon wafer covered with a regular array of square pyramidal pits provides a template for growing thick colloidal crystals oriented in the [100] plane. 

To turn these crystalline assemblies into useful photonic structures, however, the researchers must find a way to introduce defects in a controlled manner so that they can open up transparent channels and cavities. 

Ozin and colleagues have used a laser to ’write’ defects into a photonic crystal made from amorphoue silicon. Laser heating allows the silicon to crystallise locally, producing a refractive index difference between the amorphous and (nano)crystalline phases. In this way, they have been able to create defect channels about 1?m wide in the photonic crystal lattice. 

Inside out 
Methods for making uniformly sized microspheres are well developed for silica, but there are good reasons to want to make colloidal crystals from other materials too. Silicon is particularly attractive, since this would make the structures compatible with traditional microelectronics. Germanium is also a desirable target material since it has a large refractive index. The larger the difference in refractive index between the scatterers and the matrix (air) of the photonic array, the wider the photonic band gap. 

Ozin and others have developed a neat trick to convert silica colloidal crystals into photonic crystals of a range of other materials, including silicon and germanium. They use the self-assembled silica array as a mould, casting the other substance in the spaces between the spheres. Etching away the silica using hydrofluoric acid then reveals an interconnected, ordered array of solid ’bubbles’ of the new material: an ’inverse opal’, which can also display a photonic band gap. 

Removing the silica microspheres requires that they be accessible to the etchant, and not totally encased in the secondary material. To ensure this, the colloidal crystal may be sintered - heated to 1000?C or so - which welds the microspheres together via narrow necks where they touch. A gentler method of growing necks involves room-temperature chemical vapour deposition of silica onto the packed microspheres. Welding the spheres in this way can also prevent large-area colloidal crystals from cracking as they dry. 

Ozin, together with Sajeev John and collaborators in Spain, has used this approach to make a silicon inverse opal. They deposited silicon in the interstices of a silica colloidal crystal using chemical vapour deposition, infiltrating the array with disilane gas. Once the silica was removed, the voids in the silicon were close to 1?m across, and the researchers calculated that the resulting material should have a band gap at a wavelength of around 1.5?m: the infrared wavelength used for optical telecommunications. By confining the silica template within straight channels several ?m across, Ozin’s group has cast free-standing inverse-opal fibres of silicon several hundreds of ?m long, which could act as the basis for a new and highly efficient form of optical fibre. 

Meanwhile, Pierre Wiltzius, now at the University of Illinois at Urbana-Champaign, and coworkers have made inverse opals of cadmium sulphide and cadmium selenide - materials with a high refractive index - by using purely ’wet’ chemical synthesis. They use electrodeposition from solution, rather than vapour deposition, to fill the gaps of a colloidal crystal of silica microspheres with the other materials. 

Crystal sensors 
These photonic crystals are strongly coloured thanks to the same diffraction effects that decorate butterfly wings. Sanford Asher at the University of Pittsburgh is using these diffraction colours to turn colloidal photonic crystals into chemical sensors and optical switches. 

The colours depend on the spacing between the microspheres. Asher realised that if this spacing could be changed in a reversible way by some external influence, he’d have a colour-changing sensor. Asher’s group began to develop these photonic crystal sensors in 1997. They made colloidal crystals of polystyrene microspheres which they infused with acrylamide monomers that could be photopolymerised to form a crosslinked polymer gel. 

Polyacrylamide gels absorb water - they are hydrogels - and the volume of the gel depends on how much solvent it contains. This in turn may depend, for example, on the temperature, pH or ionic strength of the solution. Changes in these conditions cause the smart hydrogel material to swell or shrink. This swelling in turn increases the spacing between the colloidal polystyrene microspheres embedded in the hydrogel matrix and induces a red-shift in the light that the material diffracts. 

To demonstrate the principle, Asher’s team attached crown ether groups to the polyacrylamide chains, capable of latching onto lead ions. When this happens, the polymer chains become charged and repel one another, and the gel network swells, changing the material’s colour. If the lead is removed, the material shrinks again. The resulting colour change is visible to the naked eye even for lead concentrations as low as 20?M. 

Asher and his coworkers have used the same principle to make a variety of other colour-change sensors by appending different ’capture’ groups to the hydrogel network. They have been able to detect metal ions such as copper, zinc, cobalt and nickel at concentrations of less than 1?M. Recently, they attached boronic acid groups to the polymer chains, which bind selectively to sugars such as glucose. A cheap and convenient sensor for glucose would be very useful for monitoring blood sugar levels in diabetes patients. To make a clinically viable device, such a sensor would have to satisfy some pretty exacting demands. However, Asher and his team say that they have now developed a photonic crystal sensor that meets all the necessary criteria. 

The Pittsburgh group has adapted these photonic crystal hydrogels to make materials that can be switched rapidly from essentially transparent to opaque. This makes them potentially useful as optical switches in photonic information technology, and as optical limiters that shut out light if it exceeds a certain threshold. 

The researchers made a colloidal crystal from uniformly-sized microparticles of a smart polymer: poly(N-isopropylacrylamide) (PNIPAM). This material shrinks abruptly from a water-swollen state when the temperature reaches 32?C. In a macroscopic lump of gel this transition can take several minutes, but in individual microparticles it happens in micro- or even nanoseconds. 

A flash of bright light can suffice to shrink the gel and cause it to become opaque, transmitting just 0.1-10 per cent of the incident light. Much of this change happens within 1?s, which may be fast enough for practical optical switching. 

Photonic ink 
Ozin’s team hopes to use colour changes like this to make full-colour ’electronic ink’, which can be printed on ’electronic paper’ to produce colour pictures that can be switched at the push of a button. Most flat-screen colour displays use the standard trichromatic colour mixing of television screens to produce different colours. By contrast, Ozin’s ’photonic ink’ would change its intrinsic colour in each pixel. It is basically a photonic crystal embedded in a polymer matrix that can swell by taking up more solvent. 

Crucially, the polymer gel’s solvent content can be controlled electrically by tuning the redox state of the polymer chains. Ozin has collaborated with Toronto colleague Ian Manners, who has developed metal-containing polymers made from ferrocene. The charge state of the iron atoms in the chains can be varied between 0 and +1, and the oxidation state of the metallopolymer as a whole can be adjusted smoothly by changing the applied potential. 

Altering this redox state reversibly changes the amount of solvent the metallopolymer absorbs, allowing the diffraction peak to be tuned rapidly right across the visible spectrum. In principle, each voltage-controlled pixel of a ’P-ink’ screen could be tuned to a different colour. 

So photonic colloidal crystals promise to bring ’soft’, cheap processing techniques and chemical tunability to technologies such as optoelectronics and visual displays, which have traditionally been associated with hard materials created in specialised, high-tech environments. Ozin has visions of an all-optical computer with circuits and screens made by ’bucket chemistry’. If you think that sounds like a lot of balls, so does he. 

Source: Chemistry in Britain


Philip Ball

A hole different... 

One way of making a two-dimensional photonic crystal is to pepper a slab of material with a honeycomb of holes. A lattice of holes punched through a slab of transparent material will do the same job as a lattice of pillars in air. What matters is that there is a marked difference in refractive index between the scattering ’objects’ (pillars or holes) and the matrix in which they are embedded. 

At Bellcore in New Jersey Eli Yablonovitch has made a three-dimensional array of scatterers by drilling a series of holes through a block of material at three different angles. The resulting photonic crystal blocks electromagnetic radiation in any direction. 

Because the periodicity of Yablonovitch’s Swiss-cheese material is of millimetre dimensions, its photonic band gap lies in the microwave range. Photonic crystals this big can be useful for microwave technology, but the ability to control visible and near-infrared light would open up the field to much more valuable applications in optoelectronic information technology. 

Researchers realised that drilling holes was not the only way to make periodic arrays of scatterers - they could also stack spheres or grids of parallel bars. But once these objects get very small, it becomes harder to manipulate them into the right geometrical arrangement. Making optical photonic crystals appeared to be technologically very challenging. 

In fact, nothing could be simpler. Colloid scientists have perfected methods for making spherical microparticles of uniform size from polystyrene and silica. If left to settle under gravity from a liquid suspension, such microspheres will accumulate layer by layer in close-packed, orderly arrays called colloidal crystals. There is no need for any direct manipulation at all: the crystals simply self-assemble spontaneously. 

So a periodic array can be prepared quickly and cheaply using ’wet’ chemistry, without any expensive lithographic or micromachining processes. Colloidal crystals, long regarded as a curiosity or a humble surrogate for studying atomic-scale crystallisation, are now coming into their own at the forefront of photonic information technology. 

How might these photonic crystals be used? Waveguides for channelling light - a kind of microscopic optical fibre - can be prepared by threading defects through a photonic crystal, such as a missing row of scatterers. A light pulse can penetrate the defect while being excluded from the photonic crystal all around, rather as light passes down the core of an optical fibre confined by its cladding. But because the confinement does not rely on simple reflection, the light in a defective photonic crystal can be guided around much tighter bends than in an optical fibre. 

A defect surrounded on all sides by a perfect photonic crystal constitutes an optical cavity, where light can be pooled. Cavities like this are used in lasers. Again, because the confinement is so good, miniaturised lasers made this way might have very low ’turn-on’ power thresholds, reducing power consumption and heating in photonic integrated circuits.  

Contact and Further Information

Philip Ball
Nature consultant editor
4-6 Crinan Street, London N1 9XW