New Scientist vol 153 issue 2075
- 29 March 97, page 36
IT USED to be so easy to tell humans and computers apart—humans are carbon-based life forms, computers are silicon-based machines. But then came the idea of "DNA computers" which hold the promise of performing trillions of calculations simultaneously using the blueprint of life. Now the distinction is being blurred still further—this time by the possibility of using silicon, the stuff of computer chips, to build artificial body parts. And tiny silicon capsules could even monitor and adjust the balance of chemicals in our bodies, delivering drugs or hormones to keep us in tip-top health.
Until now, silicon has not been considered a promising material for biological applications because it is not compatible with our bodies, meaning that bony deposits and living cells refuse to grow on it. The good news is that materials researchers have potentially found a way to wire up your body to silicon-based medical devices, melding the living with the artificial. The trick is to pepper the silicon with tiny holes. Unlike "bulk" silicon, this porous silicon could be tolerated by living organisms—it could be biocompatible.
The pores in PS are very narrow, with diameters as small as between 1 and 2 nanometres, but several micrometres long. Because they can be packed quite closely together, the resulting piece of silicon has a sieve-like surface. PS is made by etching a silicon wafer with hydrofluoric acid in an electrochemical cell, with the wafer acting as the positive electrode. Etching the wafer riddles the surface with nanometre-size pits. Once these have formed, etching continues at the bottom of the pits but leaves the pristine silicon surface alone, making the pits grow deeper rather than wider.
This occurs because electrical resistance is lowest at the bottom of the pits and so current, and hence etching, is greater at these points. Continued etching creates a forest of branched silicon "nanowires" or "nanocolumns", the precise dimensions of which can be controlled by altering the current and the strength of the acid. This fine structure is thought to be responsible for some of the material's remarkable properties.
Throughout the 1990s, PS has been billed as "the next big thing" in optoelectronics, because of its photovoltaic and photoluminescent properties. It can be made to produce an electric current or glow simply by illuminating it with certain wavelengths of light, and it will also light up if a voltage is applied ("A glowing future for silicon", New Scientist, 10 April 1993, p 23).
But PS has been slow to fulfil its promise—it was not until November last year that a team managed to combine microelectronic circuitry with light-emitting PS on a single chip. And while attention has focused on PS in optoelectronics, nearly everyone seems to have overlooked the enormous potential of a biocompatible material into which electronics engineers can carve intricate circuits and sensor devices.
PS only becomes biocompatible after removal from the cell at the end of etching. A thin layer of hydrogen atoms, built up during the etching process from the acid, is gradually replaced by oxygen, transforming the surface into silicon oxide. PS then behaves like a passive substrate on which it is possible to grow cells, and it also seems to actively encourage hydroxyapatite, the main form of calcium phosphate found in human bone, to deposit on it. The secrets of these two abilities are likely to be very different.
In the case of cell growth, the answer may be straightforward. "The oxide is inert and doesn't poison the cells," says Sue Bayliss, whose research group at De Montfort University in Leicester has been studying PS. "I'm convinced that has got a lot to do with it."
It is less clear why PS encourages hydroxyapatite to grow. When calcium phosphate and other biological salts are deposited on it, "it's acting as an active substrate," says Bayliss, encouraging the salts to form. "The salts start up at certain sites on the nanostructured surface and then the hydroxyapatite-like material grows from those." She believes that defects of the right shape in the surface of PS could be the sites where seeds of hydroxyapatite are sown, but stresses that no one really knows what happens.
Leigh Canham, from the Defence Research Agency in Malvern, Worcestershire, heads a group looking at the potential of microstructured silicon as an active biomaterial. He has tested various silicon surfaces by leaving them for several days in simulated body fluid—a solution of salts designed to mimic those found in blood plasma. Rather than behaving like materials such as bulk silicon oxide, aluminium oxide and titanium, which are bioinert—they are ignored by the body—Canham has found that PS attracts a layer of hydroxyapatite in a similar way to more established "active" biomaterials, such as Bioglass or Ceravital.
Corrosion cells
"In the case of PS, the mechanisms aren't understood by anyone," he says. "But we know PS provides a more chemically active surface than bulk silicon oxide." Canham suspects that minute "corrosion cells" are set up on the surface where atoms are lost, perhaps containing trapped electrons to give a negative charge. These could attract the positively charged calcium ions needed to get hydroxyapatite deposition started.
Canham's recent experiments support his idea that such electrochemistry is the key. He found that the rate of hydroxyapatite deposition can be altered by giving the PS an electrical charge. With a negative charge, a coating of calcium phosphate was created in a few hours—much faster than the natural rate of bone-building—while a positive charge slowed the process compared to uncharged PS.
Whatever gets them going, the structure of the hydroxyapatite deposits that accumulate on PS bears a close resemblance to natural bone. According to Bayliss, this could make PS an ideal material for coating orthopaedic implants. "Natural hydroxyapatite is disordered, but [synthetic versions are] highly crystalline, and not very like what you get in the body," she says. "The material you get on the PS surface is much more like the substance in the body."
Synthetic hydroxyapatite coatings are already applied to many artificial joints such as hips and knees to encourage the patients' bone to bond with the implants. Such a bond ensures that the implant does not come loose from the bone it is supposed to be reinforcing.
Coating an implant with PS might provide an even better bond. By applying a negative electrical charge to the surface of the implant, doctors might be able to promote bonding between the PS and the living bone, rather than simply providing a passive substrate for tissue growth such as artificial hydroxyapatite, which can't easily be electrically charged.
PS scientists are relative newcomers to the field of biocompatible materials, however, and established researchers are sceptical about the potential benefits of the new material. Larry Hench of Imperial College, London, has been studying the mechanisms behind mineralisation for over thirty years, and does not see the PS work as anything new. "The field of electrical stimulation of bone is pretty advanced [already]," he says.
Blocked signals
Ironically, Hench thinks that the tendency of PS to accumulate a bony covering could hamper attempts to exploit its electronic properties inside the body in small efficient electrical devices. For example, if you wanted to develop a PS-coated chip implant to electrically stimulate nerve cells, hydroxyapatite deposits might form over the electrical connection sites, blocking signals. "How could you get signals to and from the chip?" asks Hench.
But Canham has already established that applying different electrical potentials to PS affects the rate at which hydroxyapatite deposits. And as it is already possible to create very fine patterns on the surface of silicon, he thinks it should be possible to build chips that have tightly controlled regions of electrical potential across their surface. These would only accumulate hydroxyapatite where necessary, and leave certain areas free to make electrical connections. "It should be possible to, say, keep a sensor surface clear, or perhaps keep hydroxyapatite formation to one small area in order to fix the chip in place," says Canham.
An alternative method of creating pores could be used to etch patterns of the required precision. It involves making nanoscale protrusions on the surface of silicon—creating pores from the bottom up rather than from the top down. The De Montfort team is using the interference pattern generated by two lasers to control the deposition of a stream of silicon atoms on a silicon base. In this way, the researchers build up a regular nanostructure of bumps. "The atoms only land close to the minima in the [lasers' interference] pattern," says Bayliss. She believes such a tightly controlled technique could be useful in creating material similar to PS on a complex chip. The equipment is more expensive than that for electrochemical etching, but once it was up and running, the technique would be relatively cheap.
Whichever way it is made, given its biocompatibility and its potential for intricate electrical manipulation, PS could make a big impact in a variety of applications. In today's smart implants, such as pacemakers or cochlear implants—which are used to improve the hearing of profoundly deaf people—any electronic components have to be kept away from direct contact with body fluids. They are often encased in boxes made of titanium alloy. By using established chipmaking techniques, PS itself could act as the packaging material, incorporated on the same chip as the electronics, leading to a much smaller device.
Once again Hench injects a note of caution. "It still doesn't get over the problem of how to get signals safely to interface with the body," he says. He points out that pacemakers and cochlear implants already function successfully in patients, delivering signals to the nerves via standard electrodes. "No one has yet come up with a superior way of stimulating nerve tissues," he argues. But none of the PS researchers is suggesting that these smart, miniaturised implants are short-term prospects, although the potential for future applications is clear.
It's not only better versions of today's implants that might one day be available. For example, researchers from MIT and the Harvard Medical School have begun work on an artificial retina chip that would sit at the back of the eye ("Sight for sore eyes", New Scientist, 19 August 1995, p 38) This is exactly the kind of device that could exploit PS as a signal-carrying biocompatible substrate linked to the retina.
As well as biocompatibility, PS has another crucial property—a massive surface area of up to 800 square metres per gram. Bayliss and other researchers are hoping to take advantage of this property to make a range of chemical sensors, for industrial monitoring applications and for implants to monitor chemicals in the bloodstream. These could be incredibly sensitive if the whole of the available surface was coated with enzymes or other molecules able to "recognise" a specific target. The pores would have to be made large enough to accommodate these molecules, however, which would reduce the total surface area per gram of material.
A team from the University of Lund in Sweden has built just such a sensor implant coated with the enzyme glucose oxidase to detect glucose in the blood of a diabetic. When an enzyme molecule encounters a glucose molecule, a reaction occurs and an electron is released at the surface of the sensor. This can travel back through the underlying PS to an electronic monitor. The bigger the coated surface area, the more electrons are likely to be released for a given concentration of glucose molecules, and the larger the current that will be sent back to the monitor. The current detected will give a measurement of the concentration of glucose molecules.
A major advantage of this technique is its adaptability—pick a different enzyme to coat the surface and you have got a completely different sensor. By coating different areas of a chip with different enzymes and separating their electrical responses, a single implant could monitor several important chemicals at once.
Permanent sensor implants are some way off, however. The main obstacle is the stability of the enzyme coatings, which can break down after only a few hours. Initially, such implants are far more likely to be used as short-term, disposable sensors for monitoring patients during operations. For example, levels of blood gases could be monitored by a set of chips designed to last 5-6 hours—long enough for most operations.
Living conditions
In industrial applications, PS sensors could take off much faster. The De Montfort team is looking at ways of monitoring the conditions under which cells are grown in the biotechnology industry—these have to be carefully controlled to prevent the cells from dying. Cells attached to the surface of PS could act as monitors, checking that conditions for the growth of other cells are just right. Bayliss and her team have successfully grown chinese hamster ovary cells on PS. "CHO cells are surface-growing cells that don't grow on silicon but do grow on silicon oxide," she says.
CHO cells are widely used in biotechnology because they can be engineered to produce important pharmaceutical compounds, such as the anticancer drug interferon and factor 8, the blood clotting agent that haemophiliacs lack. At the moment, measurements in the reaction vessels in which the cells are grown, for example the concentration of glucose or the pH of the growing medium, have to be taken periodically. Continuous monitoring of the electrical activity of CHO cells growing on a PS chip would provide more precise feedback on conditions. If reactor conditions become less than ideal, then the electrical activity of the CHO cells would drop, alerting staff to the problem.
Back in the human body, cells attached to a PS substrate inside a PS "cage" could be used as sophisticated drug delivery implants. According to Canham, the pore size of a PS capsule could be tightly controlled so that it acts as a filter, allowing only certain molecules in and out. In theory, at least, this might enable biological molecules to be trapped in a PS implant and used to produce drugs for delivery directly into the bloodstream through the pores. Similarly, nutrients for the cells would be small enough to enter the implant, but antibodies and immune cells would be too large to get in. "I haven't really considered specific applications at this stage," says Canham, "but I don't see why biologically complex things couldn't be contained within a chip, provided that the pore size can be engineered, and over different [pore] lengths."
So by combining a complex PS monitor with several drug or hormone-producing implants, our bodies might one day play host to a silicon pharmacy which would maintain levels of vital chemicals in our bodies if our natural systems have broken down. The marriage of carbon-based life and silicon-based machine would have begun in earnest.
| Chemical counter implant |
) | ||
|
Michelle Knott