Medical Device Link.

DETECTION TECHNOLOGIES

A prototype magnetic biosensor system by Diagnostic Biosensors LLC (Minneapolis) and NVE Corp. (Eden Prairie, MN). Photo by Roni Ramos

Micro- and nanosized magnetic particles are commonly used in biochemical assays as a means to capture, concentrate, and manipulate target analytes. They are also increasingly used as detection labels for assay readout. These detection applications benefit from the large magnetic-field signature generated by the beads when magnetized by an applied magnetic field, and from the very low incidence of background magnetic signals. These features together make magnetic detection an attractive technology option for a range of bioanalytical applications.

Several detection platforms can perceive and quantify magnetic labels. Magnetic microchips, superconducting quantum interference devices (SQUID), magnetometers, scanning probes, and induction techniques have this capability. While each magnetic detection technology offers strengths in one area or another, the focus of this article is on the microchip-based devices.

Magnetic microchip detectors are tiny, inherently rugged, and very inexpensive when produced in large quantities. The magnetoresistive manufacturing technology for their volume production is well developed. These sensors are incorporated into applications that range from implantable medical devices and hearing aids to magnetic hard drives and automobiles.

In vitro diagnostics presents some exciting application opportunities for magnetic microchip–based detection. Recently developed coil-based detector systems have generated sensitive and quantitative results from strip assays. The biochemical sensitivity of these assays is limited more by the lateral-flow strip technology than by the detection technology. Rapid improvements along several fronts—for example, in membrane properties and uniformity, magnetic-bead performance, and substrate micromanufacture—makes it likely that IVD manufacturers will want to take full advantage of magnetic detector capability in the near future.

Magnetic microchips give product developers the potential to build on previous accomplishments by increasing the numerical precision and spatial resolution of the assay readout. With the provision of a highly multiplexed reader element, many spots and lanes can be quantified on a single strip. Furthermore, the ultralow cost of detector chips qualifies them for incorporation with the assay strip as a disposable item.

This article reviews existing magnetic bioassay detection applications and presents possible directions for new development in the detection of immobilized magnetic-assay labels. In addition, it examines concepts relating to the integration of magnetic detection in microfluidic systems. IVD applications in this latter area are further off in time owing to the additional layers of complexity introduced by sample preparation and microfluidics. Still, it is important to probe the limits of magnetic detection in flowing streams and to understand how micromagnetic detection and manipulation can be applied in point-of-use (POU) bioassay applications.

Advantages of Magnetic Microchips

It is inherently difficult to fairly compare a novel technology with proven ones. Nevertheless, some characteristics of magnetic microchip detectors do distinguish them as superior to existing technologies.

The combined qualities of their readers and detectors make magnetic microchips very well suited for POU and disposable assay formats, as already exemplified by such popular applications as the pregnancy test (where detection is visual) and glucose monitoring (where it is electrochemical). Both of these test applications are well served by the existing microchip technology. However, that technology will not be easily adapted to multiplexed assays or to assays requiring sensitivity and precision. Disposable magnetic-assay chips, therefore, are likely to be best used for small panels of immunoassays—in the range of 2–100 analytes—on a single device.

Electrochemical detectors are probably the best example of a type of disposable electronic sensor. From a microchip-and-circuit perspective, they present a very simple arrangement of electrodes in solution. However, the measured current is directly dependent on ionic concentrations in the solution. Without having separate reaction chambers, thus, the same assay conditions prevail for all analytes. It has so far been impractical to optimize a multianalyte POU immunoassay using this kind of detector. Magnetically labeled detection, in contrast, is readily compatible with virtually any sample buffer format.

Microscopic detection is the dominant format for laboratory-based microarrays. High-density optically read microarrays such as so-called gene chips usually employ a mechanical scanning system to move the highly sophisticated optics across the assay surface. This arrangement is extremely challenging to adapt to POU applications because of its mechanical complexity and the need to isolate the system from vibration and light.

Successful disposable POU technology enables the user to acquire and prepare the sample in nonideal locations, without the aid of instruments, and with no more than rudimentary training. When approaching a development project from this perspective, designers place emphasis on the sample preparation and handling aspects of the system rather than the detection element, which is often simply the naked eye. The minimization of mechanical complexity, along with instrument size, weight, and power, is valuable. In these areas, an integrated magnetic microchip detection system would offer significant advantages. Following a technical discussion of magnetic microchip detection technology, this article considers possible applications, reader design, and manufacturing issues.

How Magnetoresistive Sensors Work

The basis for the use of magnetic labels in biological assays is that they can be attached to the analyte through a biochemically specific conjugation. When this capability exists for a given analyte, the challenge is then to count the labels in a way that yields data that are meaningful for the assay result. An illustration of the magnetic-detection mechanism is provided by the simple example of detecting a single magnetic label bound to a magnetoresistive sensor sandwiching a captured analyte. This single-bead example is not practical for most actual assay applications, but it does supply a framework for understanding more-realistic situations involving, for instance, moving labels or the detection of large numbers of immobilized labels.

After examining single-bead detection in depth, the article looks at the detection of two-dimensional bead layers, as with surface-immobilization assays, and three-dimensional plugs, as in lateral-flow strips.

Single-Bead Detection. Magnetic beads become magnetized—that is, magnetization is induced in their volume—by the presence of an applied magnetic field, H. The direction of the magnetization is parallel to the applied field. The magnitude of magnetization is linearly proportional to the applied field (until the saturation field is attained, beyond which the magnetization remains constant). Thus, the magnetization is zero when the applied field is zero.

(Note: For simplicity, this discussion is limited to spherical paramagnetic beads in fields below saturation. Other beads are ferromagnetic or permanent magnets, but these are not very commonly used.)

Figure 1. (click to enlarge) A single magnetic bead bound to the sensor chip surface, as in an immunoassay.

The magnetic fields in the vicinity of a magnetic bead are the sum of the externally applied field, H_applied, and the stray fields from the bead, H_bead. H_applied usually originates from a magnetic-field source that is far enough away to be considered spatially uniform in the bead’s vicinity. But H_bead varies strongly both in magnitude and direction. Magnetic-field sources typically are the passage of electrical current through a wire or a coil of wire, a permanent magnet, or the presence of both. Each bead magnetized by these fields binds to a magnetoresistive sensor (see Figure 1).

The magnetic-detector technology that this article discusses exclusively is known as giant magnetoresistance (GMR). GMR is one of a class of integrated magnetic sensing technologies that are sometimes called spintronics. This class includes—roughly in order of technological maturity—Hall effect sensing, anisotropic magnetoresistance, GMR, magnetic tunnel junctions, and colossal magnetoresistance (which is better for cryogenic applications). To the extent all of these are instrumentalized as solid-state chip-based devices, they can be applied similarly to the ways GMR sensors can, as described here. Which is best for a given application depends on a manufacturer’s marketing and technical requirements.

A GMR detector is a multilayer metal film whose total thickness is in the range of 10–100 nm. Resistors are formed from the film using standard semiconductor-wafer-processing techniques such as photolithographic patterning and ion mill etching.

Patterned resistors are about 1 µm wide and of arbitrary length. The total resistance of a typical GMR resistor is about 10 Ω times the length-to-width ratio. For example, a 100 × 2-µm resistor has a nominal resistance of 100 µm/2 µm × 10 Ω/sq = 500 Ω. By design, the GMR sensor is sensitive primarily along one axis in three-dimensional space. This direction is usually in the plane of the film, but it can be either parallel or perpendicular to the long dimension of the resistor.

The electrical resistance of the GMR sensor is used in a circuit to infer the magnitude of the total magnetic field at the sensor location. A constant current is passed through the resistor, and the voltage measured; the resistance is the voltage divided by the current. This voltage is independent of the frequency of measurement, and exists at direct current.

Figure 2. (click to enlarge) Schematization of voltage versus Happlied in a GMR sensor with and without a bead on it. A constant current supplied to the sensor is assumed. The units are somewhat arbitrary, but the shape of the curve is representative.

The H_bead has the shape of a dipole field, much like the magnetic field of Earth (see Figure 1). Its magnitude drops off as the function (r/d)³, where d is the distance from the bead and r is the bead radius. Also, the field from the bead is not uniform in the plane of the sensor. Clearly, the magnetic-field effect of the bead is greatest in the region immediately beneath it, and decreases rapidly with movement away from that area underneath. The sensor’s resistance changes appreciably only in that region. Ignoring, for a moment, the nonuniformity of H_bead, one can see the effect of the bead on magnetic fields detected by the sensor. For some H_applied, there is an H_bead in the opposite direction that decreases the total field at the sensor (see Figure 2). This effect is measured as a voltage change by electronic instrumentation.

If the bead is much smaller than the overall size of the sensor, the total fractional resistance change will be much smaller than if the bead were the same size as or larger than the sensor. In fact, the total resistance change is roughly proportional to the fraction of the sensor area that the bead covers. This fact becomes very useful when more than one bead is being measured.

Figure 3. (click to enlarge) Schematic view of a GMR sensor (the blue serpentine) with some paramagnetic beads immobilized on top. The ordinary sensor surface (indicated by diagonal hatching) has been modified in one region (marked by cross-hatching) to capture beads. Wires (green) connected beneath the sensor run off to external circuitry.

Bead Layers. In most bioanalytical applications, detecting a single bead would not be useful. Rather, it is necessary to measure, within the detection region, the surface density of beads in a quantitative way so that the concentration of some analyte can be inferred. The area of the sensor can be extended almost at will by patterning it in a serpentine shape (see Figure 3). Also, large arrays of very small sensors can be created; such arrays offer both high precision and a large sensing area. Assay conditions must be set to match the analyte concentration range of interest so that, after the time interval of conjugation and labeling, the number of bound labels is somewhere between zero and full coverage. The total resistance change of the sensor then corresponds to the number of beads bound there.

Some early assays in this arrangement were demonstrated using GMR sensors under a 200-µm-diameter circular spot to detect 2.8-µm Dynal beads bound to the sensor through a specific binding of captured DNA.^1,2 The dynamic range of the detector stretched from 10 beads to about 5000 beads—two and a half orders of magnitude. This level of bead detection corresponded to a maximum DNA concentration of 100 fmol as measured by means of optical detection. An analogous result was obtained using beads approximately 60 nm in diameter in order to quantify the immunological interaction between surface-bound mouse IgG and a-mouse IgG coated on superparamagnetic particles.³

Spots and Plugs. Three-dimensional sample volumes are, of course, important in bioanalysis. GMR technology can also be used to measure the magnetic content of very small volumes such as those in a microfluidic channel or a lateral-flow membrane. Detecting a magnetic plug or volume works the same way as detecting a single bead or layer of beads in that the sensor detects changes in the local magnetic field due to the presence of magnetic material there. The complicating factor—remembering that the stray magnetic field from a bead diminishes as (r/d)³—is that not all of the magnetic material can be said to be situated at the same distance from the sensor. That the math is more complicated does not mean that the detector will not work; rather, it means that more effort in sensor calibration and data analysis is required.

Magnetic Labels in Lateral-Flow Strips

A complete system for detecting magnetic labels in lateral-flow strips has been developed by MagnaBioSciences LLC (San Diego).^4,5 This system has been used successfully for sensitive and quantitative work on E. coli, outperforming two reference systems for the application.⁶ Its physics of detection differs from that of the GMR system. The magnetic beads are paramagnetic, just as discussed above, but the detectors are loops of wire acting as inductors.

The voltage induced in a loop of wire is proportional to its area and to the frequency and magnitude of the oscillating magnetic applied field. No signal is detected in a loop or coil if the field does not change over time. The bioanalytical volume being measured in the case of a lateral-flow strip is on the order of 1 × 3 × 0.4 mm, the first dimension being the width of the capture stripe on the strip, the second the strip width, and the third the strip thickness. The pickup coils for this system must be well matched to the sample dimensions; curve fitting and mathematical adjustments have to account for the geometrical realities of a production device.

Figure 4. (click to enlarge) The sensor region of a multisite lateral-flow strip assay with integrated detector, viewed from above. There are seven lanes of flow, each containing two tests and one control spot along with a reference sensor for each spot.

Multiplexed Assays. It may be desirable to have many assays performed in the same device, potentially saving time and money. This multiplexing means breaking up the single test and control bars across the strip into several discrete spots (see Figure 4). Lateral-flow strip technology appears to be ready to support multispot arrays, but there are not many detectors ready for use. Some proposed designs for multispot lateral-flow strips with integrated detectors based on GMR technology follow.

The ideal detector for such an array of test and control spots as is shown in Figure 4 would have the GMR sensor size match the spot size, and have the sensors be as close as possible to the spots. The spots should be situated so that the magnetic signature from one spot is picked up primarily by the appropriate detector. Supposing the flow membrane thickness is 100 µm, the spots could be sized about 80 µm in diameter and spaced at 200 µm center to center.

Figure 5. (click to enlarge) Close-up cross-sectional view of the many beads bound in the control spot of the lateral-flow strip in a GMR-based assay. The GMR sensors are embedded in the flow membrane backing material.

The magnetic detector array can be a separate surface upon which the flow strip is pressed for the purpose of reading. Or, the detector array can be incorporated into the disposable lateral-flow strip just as it is in microfluidic systems (see Figure 5). The latter design involves a more complex strip-manufacturing process, but it makes possible a highly precise quantitative multiplexed assay.

Detection Signal. Should a certain spot have a 1% volume packing of 250-nm-diameter magnetic beads (amounting to about one bead per cubic micron), the total number of beads in that spot volume would be about 1 × 10⁶ (because 100 µm × 100 µm × 100 µm = 1 × 10⁶ µm³).

Clearly, the detector and beads then are no longer separated by a distance less than one bead radius. It is better in this case to think of the detector as measuring the aggregate signal from all of the beads rather than counting them individually. In this configuration, detection and quantification of a 0.01 to 1.0% volume fraction of the magnetic beads in the spot could be expected—a dynamic range of two decades.

There are no published data from experiments using GMR sensors for detecting plugs like this in lateral-flow membranes. However, results for detecting moving plugs in a microfluidic system do exist.

Detection in a Microfluidic System

In most of the early demonstrations of GMR in bioanalytical applications, researchers assumed that sample handling would be done with a lab-on-a-chip device. Here, rather than the passive wicking action of lateral-flow strips, active pumping or the influence of a vacuum moves the test specimen. Well-defined reservoirs and microfluidic channels replace the spots and strips. Consequently, labs-on-a-chip have the ability to perform a greater variety of functions, and to perform them much more precisely, but at the expense of introducing greater product complexity.

Ferrofluid plugs of magnetic nanoparticles (10-nm magnetite particles in aqueous suspension) diluted to 1.2% magnetite volume-to-volume with a Tris-buffered fluorescein solution for fluorescence imaging were detected by a GMR sensor beneath a microfluidic flow channel in one investigation.⁷ Plug dimensions were 13 × 18 × 85 µm for a total volume of 20 ± 3 pl. The first two dimensions were the channel width and depth, respectively, and the third was the plug length. The plug contained about 5 × 10⁸ 10-nm particles. The individual particles were distant from the sensor by many times more than the length of their diameter.

While this experimental demonstration was not entirely analogous to the lateral-flow-strip example, it did show that the GMR sensor is capable of detecting and analyzing composite plugs nearby, a function rather different from counting single beads.

Microchip-Based Magnetic Applications

Most magnetic beads are designed and used primarily for capturing and concentrating analytes. This utility made it natural for assay designers to adapt macroscale magnetic force systems to miniature bioanalytical systems.

Figure 6. (click to enlarge) Directing and counting magnetic beads in a microfluidic stream.

Electric current passing through a wire generates magnetic fields that are directed around that wire. This means that wires on a circuit chip generate magnetic fields that impinge upon a magnetic bead on or near the chip. When combined with other field sources, such on-chip wires can be used to magnetize beads and attract or repel them. The magnetic forces are sufficient to enable the sorting of magnetic objects in flow streams.^8,9 Combining the ability to direct magnetic particles in flow with a means to detect them downstream presents the opportunity to count magnetically labeled analytes as they flow by (see Figure 6). Beads are attracted to a current-carrying guidewire, which draws them down to the center of the channel bottom as, in the diagram, they flow from the area of cross section A to cross section B. They then are easily detected by the downstream sensor.

Sorting and detection by means of this technique have been demonstrated in investigations toward the development of a magnetic handheld cytometer.^10,11 Since single beads can be detected in a flow, it is possible to make ultraminiature flow detection systems for molecular analysis using magnetic-label carriers. Details on bead detection design and demonstration are available in the scientific literature, which includes review articles on magnetoresistive-based sensing¹² and microscale sorting.¹³

System Production Requirements

The key to making integrated sensor chips cheaply is to make them very small, and to use manufacturing, packaging, and testing processes that are well suited for the sensor. The raw fabrication cost of a microchip is on the order of $0.10/mm². A chip size of 1 mm² is ideal. This tiny chip, then, must be integrated successfully with a lateral-flow strip that is about 100 mm² in order for the costs to work out. The additional costs of testing the sensor and mounting it on an electrical interconnection substrate are just as important. In fact, they can dominate the total cost if special handling is required.

In keeping with the goal of simplicity and smallness, the approach to reader design should be to minimize the volume of the disposable detector chip. This in turn minimizes the size and cost of the magnetic-field source. Further, the power required for the reader to generate the applied magnetic field is proportional to the volume being energized. Preliminary analysis of designs for magnetic reader instruments indicates that the whole device can be about the size and shape of a universal serial bus (USB) flash memory stick. This thumb-sized reader would be powered by the USB port itself.

Potential Applications

The most likely early applications for this technology will be those in which the user is interested in maximizing convenience while minimizing cost. This is not going to be the case in a clinical environment where laboratory staff can send samples off to a central processing facility. Rather, the POU technology saves the user from making the trip to the clinic or laboratory in the first place.

Military and Homeland Security Department applications probably have the most immediate need for magnetic bioassay detection technology. Funding for these government areas is driving much of the technology development and will determine what the earliest practical uses will be. But the general-marketplace applications will be largely decided by consumers making choices. For example, systems that enable owners to test their pets for vaccination levels, diseases, and other conditions may prove to be attractive if enough functionality can be loaded in a single disposable device. Human personal home health-testing products may become feasible, though the safety and regulatory hurdles are much higher for such uses as these.

Mark Tondra is chief manager of Diagnostic Biosensors LLC (Minneapolis). He can be reached at mark@ diagnosticbiosensors.com.

Conclusion

Microchip-based magnetic detectors have been configured to detect and quantify a wide range of magnetic bioassay labels, and can detect single labels, or plugs containing millions of labels. These sensors, when combined with integrated magnetic means for manipulation, have the potential to advance several areas of IVD technology.

Many of the results reported to date have demonstrated the capabilities of microfluidic lab-on-a-chip systems. However, the complexity of such systems, along with the immaturity of the manufacturing processes available for them, has slowed their commercialization. Lateral-flow-strip technology, in contrast, has a well-established commercial infrastructure and manufacturing base. Thus, integrating magnetic-microchip technology directly into flow membrane technology may prove to be a faster route to combining electronics with in vitro diagnostics. The very low cost per sensor in high-volume production could allow for disposable sensor electronics.

---