Processing on a chip

According to Derek Craston and Simon Cowen, advances in microengineering are resulting in smaller, cheaper and faster instruments that promise to revolutionise the way we carry out analyses.

The development of large and powerful instrumentation has profoundly influenced developments in many areas of chemistry, perhaps none more so than chemical measurement. Over the past few decades analysts have gradually shifted their reliance on wet chemical processes to procedures that use advanced instruments for separating and detecting different molecules. Few can doubt the significance of developments in gas and liquid chromatography and mass spectrometry, which have led to improvements in measuring speed, accuracy, reliability and detection limits. Nevertheless, important as these laboratory instruments have become, they fail to meet all the requirements of the analytical community; they are largely non-portable, unable to supply real-time information, and are economically impractical for widespread distribution. Applications in process control, for monitoring factory sites, rivers and reservoirs, and in clinical diagnostics, are not adequately served by conventional instrumentation.

So far the quest for small, fast and inexpensive systems has favoured relatively simple systems - generally chemical or biochemical sensors, where the key is to design detectors that respond selectively to individual compounds. Sensors such as those used to detect blood glucose levels are now enjoying commercial success, though the number reaching the market is still far short of early expectations.

While sensor research will no doubt continue to flourish, a complementary technology is emerging for manufacturing devices that look and act like sensors, but operate along the same principles as laboratory instruments.

Downsizing equipment
At the Laboratory of the Government Chemist (LGC) we first became interested in micro-total analysis systems (?TAS) about four years ago. It was at about this time that our central research unit was set up to tackle issues such as the need for rapid screening methods that could be used to detect the presence of specific chemicals in the laboratory or onsite. Although sensors could have fulfilled this need, they were not applicable to all of LGC’s operations. LGC is active in forensic science, food and environmental analysis, consumer protection, and occupational hygiene, with analysts routinely performing many hundreds of different analyses - each of which would require its own batch of sensors. We began, therefore, to seek an alternative approach - one that could lead to systems capable of taking many different measurements, rather than detecting a single species. In the end, our search led us back to laboratory instruments, which are generic in their applications; for example, high performance liquid chromatography with a UV detector can be used to measure many things, including a number of drugs, vitamins, proteins and pesticides. Also, laboratory analysis is reliable, mainly because the instruments are easy to calibrate, and the samples can be pre-treated before analysis to remove any components that might interfere during separation and detection. In contrast, calibrating and removing interferences is difficult with commercial sensors, which usually make one-step, in situ measurements with little or no sample pre-treatment.

Adopting a laboratory-style approach to on-site screening seemed sensible, but posed the question of how we could achieve this without compromising the fast response times, low cost, small size and easy operation offered by sensors. A survey of the literature showed that not only was this goal achievable, but we were a good 10 years behind in our thinking. In the late 1970s researchers at Stanford University published a paper describing an entire gas chromatograph (GC) on the surface of a silicon wafer. Although this never became commercial, a number of portable and hand-held GC systems have become available in recent years, most of which use microengineered parts such as injectors. They are used for field analyses on air, water and soil samples, and for identifying substances at crime scenes.

Encouraged by this early work, we have been attempting, with some success, to produce part of a miniaturised, portable laboratory for on-chip micro-liquid chromatography (Fig 1). Our work is just a small part of the recent explosion of interest in ?TAS, which can be attributed in part to developments in microengineering, that allow us to fabricate relatively complex structures on the surfaces of substrates.

Fig 1. A miniature liquid chromatography column 1 m long fits on a chip measuring 20 mm by 20 mm

Light work
Microengineering emerged in the 1970s as an offshoot of the microelectronics industry, which had already established many of the processes for making ?TAS components. In particular, photolithography - used to define the complex circuitry present in components such as microprocessors and operational amplifiers - is used to produce the ?m-size patterns of miniature channels and chambers in ?TAS. We can produce any desired pattern on a surface by photolithography by using appropriately designed masks. By building up surface layers through film deposition (another common microengineering technique) whole structures can be built on top of and under the surface to provide some pseudo-three dimensionality. Solid-state bonding procedures such as glass-glass, silicon-silicon and anodic bonding can be used to produce enclosed structures on a chip. With careful design and choice of materials, we can now produce ?TAS versions of almost anything - from simple channels to more complex gear wheels and levers, and even working valves and pumps  - all on chips just a few square millimetres in size. We can also make miniaturised detectors, such as electrodes, fibre optics and various types of sensors. Putting all these components together with integrated electronics allows us to reproduce a series of relatively complex instruments in miniature.

Microengineering processes can be applied to most materials, but silicon is used most often because of its low cost, chemical inertness and high mechanical strength. Using silicon also allows us to incorporate the electronics that control instrumental functions on the same wafer, to produce a fully integrated package. And, because all microengineering processes are amenable to mass production, we can manufacture ?TAS at a fraction of the cost of conventional equipment, making these instruments disposable and easily replaced.

Fig 2. Reducing the column radius increases the resolution between two-component mixtures with the same elution times

Recent examples of ?TAS include analytical systems based on capillary electrophoresis or chromatography, and flow injection analysis systems that use specific sensors or methods such as laser-induced fluorescence for on-line detection. Interestingly, ?TAS systems should often out-perform their benchtop counterparts. For example, theoretical modelling of chromatographic and electrophoretic separations indicates improvements in separation efficiency and speed as the column dimensions are reduced. For a two-component mixture travelling along a coated chromatography column. Fig 2 shows that if the two elution times are the same, the smaller the column radius the better the resolution between the components. This is essentially because the smaller column dimensions allow greater interaction between the solutes and the coating on the column walls. Because the column is narrower there is less variation in these interactions, and solute particles spend a similar length of time in the stationary phase. This results in less sample spreading and sharper detection peaks.

A team of Swiss researchers has achieved high-resolution electrophoretic separations of amino acids in less than a minute with a ?TAS. This compares with an analysis time of several minutes for conventional capillary electrophoresis instruments, and makes the rate of analysis similar to that achieved by sensors. Besides being useful in the laboratory, portable micro;TAS will be capable of providing a wide range of chemical information from relatively inaccessible locations, inside pipes and boreholes for example. Before realising this potential, however, we must find a more reliable method of producing ?TAS, and need to resolve problems such as particles blocking the microstructures within the measuring environment, and how to obtain representative data from nanolitre-sized sample volumes. Improved pumps and valves might also be required to allow operation at higher pressures, and the development of miniature voltage and current sources should be useful for micro-electrophoresis.

Speeding ahead
Although separation and flow injection analysis are the most popular systems for investigating the efficiency of ?TAS, miniature instruments are by no means limited to these techniques. For example, the possibility of using miniature chambers and flow channels as microreactors, where chemicals react under controlled conditions, opens up intriguing possibilities.

Researchers have been quick to apply ?TAS to the increasingly useful polymerase chain reaction (PCR) for amplifying tiny amounts of DNA. Performing the reaction in a microreactor accelerates the speed of thermal cycling in this process, owing to the efficient heat dissipation from the reactor. Combining the microreactor with on-chip electrophoresis will lead to complete analysis systems such as those described by researchers at the Lawrence Livermore Laboratory in the US.7 Similary, J. Michael Ramsey and colleagues at Oak Ridge National Laboratory in the US have recently described a device for integrated DNA fragment analysis.

The possibility of combining analysis and reaction chemistry on a chip is not restricted to PCR. Another advantage of on-chip combination, besides speed, is that it allows us to make, analyse and characterise tiny amounts of products. This is not always possible when synthesis and analysis are separate, because the physical transfer of solutions is only practical with volumes that can be handled easily. Thus, miniaturisation could be very important in synthesising valuable commodities, or screening a wide range of compounds - eg in combinatorial libraries - for desirable chemical or biochemical properties.

The commercial prospects of such ’total systems’ are compelling and their development is attracting considerable research activity. Their likely impact might be compared with the way that microelectronics have revolutionised the computer industry. Thanks to microelectronics small PCs with dramatically improved processing powers and memory capacity have now become commonplace at home and in the workplace. Microengineering looks set to have a similar impact on the analytical sciences, leading to cheap and disposable miniature instruments for use in and outside the laboratory.

Source: Chemistry in Britain

Acknowledgements

Derek Craston is head of strategic research and Simon Cowen is a researcher at LGC