Science fiction has long predicted the union of humans and computers. Many children in the 1960s and 1970s were avid fans of ‘The Six Million Dollar Man’ TV series and the ‘Star Wars’ films, both of which featured characters with electronic body parts. Some of those children grew up to become today's scientists, and are now working to turn science fiction into science fact. Preliminary research into designing electronic devices that could be implanted in the body and controlled by the brain is under way. This could lead to the possibility of ‘bionic’ limb replacements and electronic sensing devices for viewing images (to replace damaged eyes), hearing sounds (to replace damaged ears) and checking body chemistry (to monitor pain, disease or drug dosage).
All these devices need to be directly controlled by the brain and so must be linked with the human nervous system. It is this interface between the biological system and the electronic system that poses the major challenge for scientists in this area. One of the most promising materials for tackling the problem of biocompatibility is porous silicon, a form of silicon that is tolerated by the body's immune system.
Silicon as a Biomaterial
Devices based on bulk silicon semiconductors have been available for in vitro (outside the body) biosensing applications for several years. However, this form of silicon is not biocompatible and so far this has prevented its use in vivo (inside the body). Bulk silicon-based integrated circuits need ‘packaging’ in a biocompatible material if they are to be used in and linked to living tissues. By contrast, nanostructured porous silicon (PS) has properties that make it a very promising biomaterial, in particular for sensing devices that need to be linked to the biological system.
Such devices would almost certainly use semiconductor-based technologies. The two basic types of components built with semiconductor materials are transistors and laser diodes. Transistors are silicon-based and laser diodes are gallium arsenide-based. From an electronics point of view, it would better to make transistors and laser diodes from the same material. From the point of view of developing biologically compatible devices, the need for new materials is more pressing. Gallium arsenide is toxic to biological systems and although the toxicity of bulk silicon is unproven, the material is nonetheless poorly biocompatible.
The search for an efficient, luminescent semiconductor that could play the roles of both laser diode and transistor took a big step forward when porous silicon demonstrated these properties. The material is also biocompatible, and so it is hoped that the use of porous silicon will speed up the development of biologically interfaced devices. The material could prove to be the bridge that allows signals and information to be transmitted between a semiconductor device and a biological system.
Production of Porous Silicon
Porous silicon was discovered by accident. It was produced by non-uniform etching during the electropolishing of silicon with an electrolyte containing hydrofluoric acid. The etching resulted in a system of disordered pores with nanocrystals remaining in the inter-pore regions. Porous silicon is still manufactured by electrochemical etching of silicon in hydrofluoric acid (HF) solutions. Aqueous HF is unsuitable for the etching process because the silicon surface is hydrophobic. The porous layer can be made more structurally uniform if an ethanoic solution is used - this increases the wettability of the silicon and allows better surface penetration by the acid. Ethanoic etch solutions also reduce the formation of hydrogen gas bubbles as ethanol acts as a surfactant and prevents bubbles sticking to the silicon surface.
Developing an etching cell to allow the maximum control of the reproducibility, porosity and thickness of porous silicon is a major concern of researchers in the field. However, understanding how to control PS morphology is extremely difficult owing to the large number of contributing factors. This means that PS is an inherently disordered material with poorly specified properties. Its luminescent properties are explained by a quantum confinement model. This suggests that the enhanced and blue-shifted luminescent emissions result from excitation recombination in quantum confined nano-structures within the PS skeleton.
Characteristics of Porous Silicon
PS is classified as microporous (with a pore size greater than 50nm), mesoporous (pore size of 5-50nm) or nanoporous (pore size less than 5nm). Transmission electron microscopy (TEM) is used to directly image PS layers and accurate pore size distribution information is obtained from gas adsorption isotherms generated at low temperature. The material is usually mono-crystalline but evidence for amorphous regions has been found using x-ray scattering, x-ray absorption fine structure, Raman spectroscopy and electron microscopy. These techniques show that the amount of amorphous layer varies with oxidation, ageing and post-anodisation treatment, yet the gross microscopic structure remains that of silicon nanocrystals embedded in an amorphous web-like matrix.
The silicon nanocrystals in PS that emits visible light vary in size from 10-15Å. Raman spectroscopy gives indirect information about the microstructure of PS and has shown that the nanocrystals alter the selection rules relating to the interaction of optical phonons with incident photons. This broadens the associated Raman peak and gives rise to the photoluminescent properties seen in porous but not bulk silicon.
The internal surface area of a PS layer varies from 200-600ml per square centimetre of external surface. The surface contains impurities from the air and the etching process that affect the optical and electrical properties of the material and could potentially affect biological systems such as living cells in contact with the surface. Common impurities include hydrogen, fluorine and oxygen. The levels of hydrogen and fluorine decrease over time as they are replaced with hydroxyl groups on hydrolysis by atmospheric water. As much as 1% oxygen is normally adsorbed within minutes of air drying. Over a few days Si-O-Si, O-Si-H and O3-Si-H groups are formed. These oxides at the PS surface are thought to play a crucial role in the biocompatibility of the material.
Biocompatibility is the ability of a material to interface with a natural substance without provoking a natural response. The human body typically responds to contact with synthetic materials by depositing proteins and cells from body fluids at the surface of the materials. This can cause infection and biological rejection of devices manufactured from non-compatible materials. The majority of today's medical devices are made from materials such as PVC, polypropylene, polycarbonate, fluorinated plastics and stainless steel. These materials are ‘tolerated’ by the human body and are described as ‘bioinert’.
An effective biomaterial must bond to living tissue - in other words, it has to be ‘bioactive’. The success of any medical implant depends on the behaviour of cells in the vicinity of the interface between the host and the biomaterial used in the device. All biomaterials have morphological, chemical and electrical surface characteristics that influence the response of cells to the implant. The initial event is the adsorption of a layer of protein on to the biomaterial. Generally, uncontrolled adsorption of a lot of proteins is undesirable in a biocompatible material.
The absorption of human serum albumin (HSA) and fibrinogen has been measured for porous silicon. Hydration of the porous surface significantly decreases the adsorption of HSA but increases the amount deeper in the porous film. Hydration does not affect the adsorption of fibrinogen, a protein essential in blood clotting processes. Another important test is the in vitro deposition of hydroxyapatite onto the surface of a biomaterial from a simulated body fluid. This has become a standard indicator of potential bioactivity for materials for bone implantation. Porous silicon is reactive towards hydroxyapatite formation - this was the first indication of its potential as a biomaterial.
The possibility of toxic effects from porous silicon also needs to be considered. Silicon is essential in biological systems as it affects both morphological development and metabolic processes. However, despite its importance little is known about the biological processes that handle silicon at the molecular level. What is known is that silicon-induced toxicity may occur if a system is exposed to more silicon than is needed physiologically.
Many studies have been carried out on the possible toxicity of implanted silicone, but there are few positive reports about the toxicity of silicon or its compounds. At Leicester we have demonstrated that cells seeded into a plate containing silicon substrates grow to form a complete continuous sheet -confluence - on the plastic surrounding the silicon material in the predicted time. We have also shown that it is possible to culture cells on a PS substrate and that these cells are viable in terms of structure and metabolism. The PS wafers were not fatally toxic to the cells over a period of 10 days and presented an acceptable surface for growth. The effect of any substances released from the PS substrates was minimal. However, these results must be quantified, and further studies are in progress to confirm the non-toxicity of PS.
If PS proves to be non-toxic then nanoporous silicon will offer a distinct advantage over other semiconductors such as gallium arsenide and indium arsenide. Arsenic and indium induce apoptosis (programmed cell death) in rat cells in vitro. In addition, cells are sensitive to topological, chemical and electrical properties of substrates on which they are grown. Cells cultivated on microstructures made by semiconductor technology grow normally on silicon surfaces covered with microelectrode arrays, as well as on microperforated silicon membranes with square pores made by anisotropic etching. The pores have edges 5, 10 or 20 µm long at the top and 1.2, 6.2 or 16.2µm at the bottom. The cells spread over the 5 and 10 µm pores, but mostly failed to cover the 20 µm ones. The size of the pores in PS therefore does not present a problem to cell growth.
The ability to culture mammalian cells directly onto PS, coupled with the material’s apparent lack of toxicity, offers exciting possibilities for the future of biologically interfaced sensing. This could involve the development of biologically interfaced neural networks, or electronic sensing with signals being directly sent from a living system to a PS device. Another benefit is that the optical and optoelectronic properties of PS could allow it to be linked to a data logger by optical fibres. This would remove the risk of electromagnetic forces influencing the responses of cells.
In this way, porous silicon has the potential to produce devices for replacing damaged tissues in the ear, eye, skin or nasal cavity. Such devices could, for example, receive optical information and convert this to a biological signal that would be passed into neural tissue as a substitute ‘sight’ sensation. Alternatively, PS could be used to build environmental or pharmaceutical sensing systems. Optical signals to and from the material could be used to sense wavelength shifts, which would correspond to changes in attached cells caused by the presence of chemicals or drugs.