Analysts in miniature

---

Abstract

With RSC's new journal Lab on a Chip being launched in September, JEM takes a look at how microtechnology is changing the face of chemical analysis.

---

Miniaturisation and automation are revolutionising virtually every aspect of society. The current generation of computers is totally unrecognisable from its predecessors of 20–30 years ago and the average laptop, palmtop or mobile phone today contains more computing power than was used to put man on the moon.

Slowly but surely these developments are finding their way into the chemical laboratory too. Since the 1980s we have seen the deployment of laboratory information management systems (LIMS), connecting the analytical instruments in a lab to one or more workstations or personal computers.¹ In the 1990s these LIMS began to be integrated with in-house software systems (“ERP systems” in the IT jargon). More recently single site systems have been linked together, using the internet and other technologies, into powerful multi-site networks.

While laboratory automation helped streamline laboratory processes, particularly in terms of information flows, it had relatively little impact on the analytical processes themselves. All this is changing rapidly however. The convergence of microelectronics, nanotechnology and materials science is opening the way to a new generation of miniature analytical devices, the so-called lab-on-a-chip (LOC). This era of micro-chemistry will not only bring major changes in laboratory practices but will also open up whole new areas in chemistry itself.

The naming of parts

“Lab-on-a-chip” is a confusing and abused term used rather indiscriminately to cover a wide range of micro-scale systems and devices. It is now generally understood to mean a device integrating chemical reaction and analysis functionalities.² LOC goes much further than merely the integration of chemical sensors with microelectronics: it involves the capability to analyse the composition of chemical samples online using one or more miniaturised analysis techniques. Far from simply shrinking down traditional benchtop techniques, the LOC approach requires innovative research in many fields, from chemistry and biology to optics, materials, fluid mechanics and microfabrication.

Two main sub-classes can be distinguished: microfluidics and microarrays.² Microfluidics are fluid control devices in which fluids flow in or along tiny channels under conditions of laminar (non-turbulent) flow. These are the engines of many analytical LOC systems. Microarrays, on the other hand, are specialised devices generally associated with DNA analysis, sometimes known as “biochips”. Containing up to several hundred thousand immobilised DNA fragments, microarrays provide a systematic way to evaluate gene expression and are becoming standard research tools in molecular biology and clinical diagnostics. While potential elements for LOC devices, DNA arrays occupy a specialised but important market niche.

An alternative terminology, consistent with the concept of LOC as involving integrated chemical synthesis and analysis, is micro-total analysis systems (µTAS).³ This entails, for example, the movement of samples and reagents within the device, either to mix them together as part of a chemical process or to deliver them to analytical units to generate information. Using microfabrication techniques, it is now possible to make highly specialised miniaturised components, such as chromatographic and electrophoretic separation columns, polymerase chain reaction vessels, and pumps and valves, for use on chips just a few centimetres square.

Taking a wider reference, LOC can be seen as a specialised form of microsystems technology (MST). Similar to the microelectronic integrated circuits used in computers and other electronic devices, a microsystem is a miniaturised device comprising multiple elements, such as sensors, signal processors and actuators.² MSTs are already used commercially in many mechanical and chemical applications. Examples include the accelerometers used to trigger automotive airbags, the mechanical devices used in inkjet printer heads, and microchip-based chemical sensing devices used in portable blood analysers.

Why miniaturise?

Lab-on-a-chip technology has many advantages compared with conventional bench-scale systems.^1–4 Foremost amongst these are the high throughput and the use of very small chemical assays.

The performance of high-throughput screening programmes in areas such as pharmaceuticals and healthcare is a function of speed and parallelism.⁴ The ability to carry out a large number of analyses in parallel on the same chip using small quantities of materials makes microarray-type devices a powerful tool for DNA-based diagnostics and genotyping. Similar approaches are beginning to be applied in toxicology and eco-toxicology.

Microfluidic devices share many of these benefits.^3,5,6 Manufacturing costs are low since units are produced in large numbers and/or use inexpensive materials. Operating and maintenance characteristics are also attractive. The devices have small footprints, compact design and low power consumption: some units are even disposable. Consumption of sample and reagents is much reduced, as are waste streams. LOC technologies allow designers to create small, portable, rugged, low-cost and easy-to-use diagnostic instruments offering high levels of precision and accuracy.

Current commercial LOC systems focus on five main applications: analytical systems for DNA sequencing in genome research; high-throughput drug screening; devices for point-of-care clinical analyses; analytical systems to defend against biological and chemical weapons; and microreactor systems for highly toxic compounds. Devices suitable for analysis of complex samples and mixtures are now entering the market (see below). In the longer term, LOC systems are likely to be used in a wide range of applications within the chemicals industry and other sectors. This could include, for example, low-throughput production processes, sophisticated biological analyses, and environmental monitoring.

One of the most exciting aspects of LOCs is that they are yielding entirely new phenomena and new insights within chemistry itself.^2,3 Some physical phenomena do not scale linearly and as spatial dimensions are reduced new micro-scale phenomena come into play. Thus, in monolithic microdevices new methods and techniques become available which were not possible in macrosystems. Interesting areas are rapid heating/cooling, rapid mixing/separation, large heat dissipation, and integrated sensors. These in turn will open up new applications in unforeseen areas.

LOC processes and components

LOC research aims to draw on the knowledge accumulated in relation to microelectronics and mechanical MSTs in developing the microfluidic devices, chemical sensors and other elements needed for chemical microsystems. Early efforts in this field focused on micropumps and valves to manipulate fluids inside a microfabricated structure. While basic tools for fluid manipulation are now available, experience with chemistry in microfluidic manifolds is still lacking. It is this that will be the focus of the next stage of LOC development and is arousing the interest of many chemists and analytical scientists.

LOC systems typically consist of a piece of silicon, glass, quartz or plastic around 2–3 cm square etched with grooves and chambers with cross-sections as low as 50 µm.^2,5 Silicon devices benefit from the mass production techniques already used in the semiconductor industry. However as a semiconductor, silicon is not suitable for the high voltages used in many microchemical techniques. Plastics offer an attractive alternative, as they are much cheaper, easier to work with and transparent. Polymer-based chips are rapidly becoming the norm as advances are made in production techniques such as hot embossing, injection moulding and laser ablation.

Early fluid control devices were based largely on mechanical principles.⁵ Recently, however, electro-kinetic (non-mechanical) principles for fluid pumping and switching have been successfully developed. This in turn has led to a general trend to perform all fluid manipulation (valving, filtering, separation, mixing) by electrical means. Despite the early successes of such systems, however, LOC development is not limited to electro-driven devices. It is increasingly apparent that the role of surfaces and their chemistry is of crucial importance, especially in the further miniaturisation step to sub-μm dimensions (“nanofluidics”).

A large number of functional elements need to be developed and integrated as the basis for LOC solutions applicable to the sort of analytical problems encountered, for example, in environmental chemistry.^3,5 These functional elements include: channels and fluidic connections; pumping, dosing and injection devices; reactors, mixers and valves; physical filters, sorters; heaters and coolers; physical and chemical sensors; separation and extraction media; light sources, waveguides, detectors and optical filters; integrated electronics and electrical connections; and feedback and control loops.

The interface between the chip and the outside world remains one of the key barriers in LOC development.⁵ Various means of introducing samples into a system have been tried, including traditional pressure-driven flow, injection through a capillary attached to the chip, and multi-channel micropipetters.

Towards real-time analysis

In microfluidic devices, fluids normally flow through channels without turbulence. With laminar flow, adjacent layers of miscible fluids move alongside each other without transverse convective mixing and mix by diffusion only. This mixing is very rapid and is one of the main advantages of microfluidic systems.⁶

Most microfluidic chips drive the fluids electrokinetically, utilising electroosmotic flow.^2,5,6 Electroosmosis is a macroscopic phenomenon that results from an electrical layer formed between ions in the fluid and surface electrical charges on the channel walls. Application of an electrical field causes the bulk solution to move towards one of the electrodes built into the device. The direction and speed of the flow is determined by the magnitude and frequency of the applied voltage. Computer-controlled ac fields are used to achieve fluid flow rates of the order of nanolitres per second.

Over the last ten years, a variety of microfluidics-based LOC systems have been developed and commercialised by manufacturers such as Aclara, Caliper Technologies, Cepheid, and Orchid Biosciences.⁶ These generally work well for highly predictable and homogenous samples common in applications such as genetic testing and drug development. Chemical analysis of complex, heterogeneous samples, such as encountered in clinical testing or environmental monitoring, is much more demanding and systems able to deal with these sort of real-world environments are only just coming to the market (Box 1).

Box 1: Sandia shows the way in portable chemical analysis

The µChemLab^™, developed by the US Department of Energy's Sandia National Laboratories, is a powerful and portable chemical analysis system incorporating lab-on-a-chip technologies (Fig. 1).

Fig. 1 Components of the µChemLab^™ Gas Analysis Subsystem. From left to right: the SAW detector (an array of four), the sample collector/preconcentrator, and a 1 m gas chromatography column.

The system creates a chemical signature, or fingerprint, by separating constituents with chromatography, a process of moving the mixture through separation channels containing a variety of materials. The materials retain constituents to different extents, so they appear sequentially in separate batches at the end of the channel. From an initial mixture injected manually into a reservoir, the system distributes a fraction of a droplet (about 0.1 nl) to channels between 10 and 100 µm wide at pressures exceeding 9000 psi. Within 1 min or so the results are shown on a small display screen. Detection sensitivities of 10–100 ppb have been demonstrated for analyses of chemical explosives and protein biotoxins.

This hand-held chemical analysis device includes a compact power source and solid-state relays that regulate and switch energy drawn from camera batteries. Built-in lasers and photodiodes, fabricated in a tiny piece of semiconductor, read results in each of three channels. Results are analysed by an internal microprocessor, which also automates the separation. To program the microprocessor, users toggle through a menu of commands using four buttons on a touch pad.

Sandia researchers foresee automated field-portable systems such as µChemLab^™ being used in a variety of situations where results are needed in real time. Examples include land mine detection, food quality monitoring, emergency management of chemical or biological hazards, pollution monitoring, bedside medical diagnostics, optimising industrial processes, and screening for new drugs.

For example, chemical analyses often require the separation of particles from soluble components by filtration or centrifugation, neither of which are suitable for integrated microsystems.⁷ Hence, microfiltration devices that can be integrated on a chip are important elements in extending the range of applications for LOC systems. Microfluidics-based separation and detection technologies offer promising solutions (Box 2). These have already been demonstrated for a variety of analytes including enzymes, proteins, electrolytes and heavy metals.

Box 2: The disposable analyser

Microfluidic chips exploit the fact that most fluids show laminar behaviour in miniature flow structures at channel cross-sections below 0.5 mm. Under these conditions, different layers of miscible fluids and particles can flow next to each other in a microchannel without any mixing other than by diffusion. Adjacent flowing solutions form diffusion interaction zones between them. Since small particles diffuse faster than larger ones, this allows separation of particles by size. It is possible to design fluidic microchips in which separations, chemical reactions, and calibration-free analytical measurements can be performed directly in very small quantities of complex samples such as contaminated environmental samples.

In the T-Sensor^™ developed by Micronics Inc., a sample, a reagent, and a control solution flow in parallel in a microfluidic channel under hydrostatic pressure. Typically, a fluorescent or absorption indicator is added to the reagent. The sample and control analyte molecules then diffuse into the reagent stream, reacting with the indicator and forming visible diffusion interaction zones. The width and intensity (absorption, fluorescence intensity, chemiluminescence, etc.) of these diffusion interfaces at a particular location in the channel and for a particular flow speed are proportional to the concentration of the analyte.

These zones can be visually interpreted, by comparing width and colour to a chart similar to a test strip for qualitative or semi-quantitative assays. Alternatively, they can be monitored with optical systems such as CCD cameras, linear diode arrays, or scanning lasers. Through determination of the ratio of an optical property exhibited in these two zones, an essentially calibration-free concentration measurement can be derived.

These devices lend themselves to applications such as ultra-low-cost disposable qualitative and semi-quantitative clinical and environmental assays for home, office, and field use, and as sample- or reagent-preparation tools that produce processed liquids for downstream analysis or synthesis.

Power consumption represents another important parameter for real world systems. For LOC devices to be viable as part of portable, or even disposable, analytical instruments then power consumption must be reduced to levels compatible with standard off-the-shelf batteries.^8,9 Further miniaturisation and integration of electronic and fluidic components can be expected to yield power savings. Some researchers favour more radical approaches based on passive microfluidic components that function without any moving parts or external power sources.¹⁰ These elements, which include particle separators, valves, detectors, mixers and diluters, utilise power sources such as gravity, air pressure or simple manual actions. Micronic's credit-card sized T-Sensor is one example of a hydrostatically driven disposable detector (Box 2).^6,7,10

The field is moving very quickly at both research and commercial levels. Sandia National Laboratories, for example, has recently licensed its microfluidics technology to Waters Corporation, a leading supplier of analytical instrumentation.¹¹ Waters will use the technology as the basis for miniaturised high performance liquid chromatography (HPLC) instruments and integrate them into its mass spectrometry products. Sandia's sister facility, Oak Ridge National Laboratory (ORNL) also has significant expertise in this field. It recently announced a “point-and-shoot” portable instrument suitable for environmental analysis.¹² ORNL claims this is the first battery-operated portable device with tunable filters and performance comparable to that of laboratory-scale instruments.

An array of possibilities

DNA microarrays also offer important opportunities in the environmental arena. The identification of hazards from chemicals found in the environment represents a major challenge for researchers, regulators and the chemical industry. Identification of the causes of diseases or adverse reactions is an essential first step to protecting human health. While many compounds are harmless, determining which chemicals contribute to or influence environmentally related diseases and the nature of the risks requires major effort.¹³

The rodent bioassays traditionally used to identify potentially hazardous substances are slow, expensive and, with changing attitudes to animal testing, socially controversial.¹⁴ Such assays are also not totally reliable, since they take no account of factors such as mechanisms of action, dose–response relationships and chemical interactions in determining the risk of chemicals to human populations. Nevertheless, in the interests of minimising both costs and data requirements they have become the staple of the mass screening programmes used in chemical regulation.

Defining and understanding the expression profile of given genotypes is essential to understanding the effects from acute or chronic exposure to environmental contaminants or other stimuli. Complementary DNA (cDNA) technology is helping toxicologists to assess how organisms function in response to chemical exposure by improving their understanding of the underlying molecular mechanisms.

A cDNA microarray comprises a large number of genes or expressed sequence tags in a condensed array on glass slides.¹⁵ The DNA microarrays are used to directly compare the gene expression profiles of two RNA samples that are simultaneously hybridised to the chip. Over the last five years, DNA arrays have been developed and refined to the stage where this is now a relatively accessible and affordable technology. They vary in design from membrane-based filters with a few hundred cDNAs, to glass-based chips with tens of thousands of genetic elements. The latest generation of devices, such as BioTrove's Living Chip^™ technology have 100 000 or more channels, providing platforms for massively parallel screening.¹⁶ This will allow expression analysis of mammalian DNA on a genome-wide scale, similar to that already accomplished in some lower organisms (e.g., yeast (S. Cerevisiae) and E. coli.).¹⁷ These whole-genome arrays should be powerful tools for identifying and characterising environmental pollutants (Box 3).¹⁸

Box 3: Microarray for environmental toxicology

In the US, researchers at the National Institute of Environmental Health Sciences (NIEHS) have developed a microarray tailored to the needs of environmental toxicology. The ToxChip^™ is a cDNA array containing around 2100 human genes known to respond to different types of toxic injury. These include, for example, oxidative stress genes, estrogen-responsive genes, oncogenes, tumour suppressor genes, and peroxisome proliferators.

With the ToxChip^™, experiments that previously took weeks or months to perform can now be undertaken in a few hours. The researchers look for a common set of changes in gene expression that form a signature for that type of chemical. Having established the signatures for various classes of pollutant, these can then be compared to the gene expression profile of unknown agents in the same model system. A match suggests a presumed mechanism of action for the pollutant under test.

To manufacture the ToxChip^™ large quantities of synthetic genes are cloned through the polymerase chain reaction. Microarrays of up to 10000 clones are then etched onto a glass substrate or microchip using high-speed robotics. Version 2.0 of the ToxChip^™ will contain around 12000 individual genes. Chips for other organisms applicable to toxicological studies, such as rats, mice, frogs and yeast, are either available or under development.

The application of genetics knowledge to environmental medicine is known as toxicogenomics.¹⁹ This emerging scientific discipline promises to bring major benefits for chemical risk assessment, dramatically reducing the time and expense of chemical screening programmes and leading to quicker and more rigorous decision-making. Work in this field is concentrated in the US, where the Environmental Genome Project aims towards an understanding of the role of genetic variations (polymorphisms) in susceptibility to environmental agents.^20,21

The potential to analyse the expression of thousands of genes in one experiment allows researchers to tackle important toxicological questions in new ways. For example, microarray technology may be used to identify toxic substances individually or in mixtures and to determine whether toxic effects occur at low doses. It can also be used to identify susceptible tissues and cell types, and extrapolate effects from one species to another. Assuming that exposures to different classes of pollutants result in distinct patterns of altered gene expression, microarray technology can be used to categorise and classify these effects by finding chemical-specific expression signatures in exposed and control samples. Arrays may also prove useful for monitoring genetic variability and its relationship to chemical susceptibility in human populations.

References

1. A. Kimery, Lab On A Chip, Military Medical Technology Online. www.mmt-kmi.com/Archives/3_5_MMT/3_5_art5.cfm.
2. Chem. Eng. News, 1999, 77, 27..
3. Mikroelekronik Centret, Danish Technical University, www.mic.dtu.dk/research/microtas .
4. “Timely Toxicology”, in Environ. Health Perspect., 1999, 107..
5. Chemical Processes in Lab-on-a-chip, WordsWork Consulting Inc, 1998..
6. B. H. Weigl, Science, 1998, 283, 346.
7. A. E. Kamholz, B. H. Weigl, B. A. Finlayson and P. Yager, Anal. Chem., 1999, 71(23), 5340.
8. A Chemical Analysis Lab On A Chip, µChemLab^™ fact sheet, Sandia National Laboratories, see www.sandia.gov.
9. Sandia National Laboratories Press Release 26th March 2000, see www.sandia.gov.
10. B. H. Weigl, in Proceedings of Micro Total Analysis Systems 2000, ed. A van den Beerg et al., Kluwer, 2000..
11. Sandia National Laboratories Press Release, March 2001, see www.sandia.gov.
12. Oak Ridge National Laboratory Press Release, May 2001, see www.ornl.gov.
13. For an overview of current practices in environmental screening programmes, see, J. Environ. Monit., 2001, 3, 33N.
14. NIEHS Microarray Centre website, see www.niehs.nih.gov/microarray/.
15. Xenobiotica, 2000, 30, 155.
16. BioTrove Inc. Press Release, June 2001, see www.biotrove.com.
17. Environ. Health Perspect., 1999, 107, 681.
18. For the latest information on DNA microarray research and technology, see www.gene-chips.com.
19. For genetics research as applied to toxicology, see National Center for Toxicogenomics www.niehs.nih.gov/nct/and the Chemical Industry Institute of Toxicology www.ciit.org/toxicogenomics/homepage.html.
20. For information on the Environmental Genome Project, see www.niehs.nih.gov/envgenom/concept.htm.
21. For a broader view on environmental microarray activities in the US, see the Environmental Protection Agency Microarray Consortium, www.epa.gov/nheerl/epamac.

---