Medical Device Link.

IVD Technology Magazine
IVDT Article Index

Originally Published May 2000

In-line manufacturing for rapid-flow diagnostic devices

Thomas C. Tisone

Because market demand for immunoassays in rapid, lateral-flow formats is continuing to increase, IVD manufacturers are gradually making the transition from batch manufacturing to automated processing systems capable of producing greater volumes of such tests. At the same time, IVD companies are also seeking to satisfy a growing demand for the development of immunoassays capable of providing quantitative test results. Because achievement of quantitative test results often depends on the repeatability of the individual processes used in manufacturing a test, greater control over such processes is essential.

Together, these trends are placing new demands on manufacturing technologies. This article discusses new technologies and equipment for achieving higher throughput, exercising greater control over individual steps in the manufacturing process, and achieving the level of repeatability necessary for the creation of quantitative tests in lateral-flow formats.

Traditional Processing

A typical lateral-flow immunoassay device employs a number of different material layers and associated reagent locations (see Figure 1). The components of such a device and their typical dimensions include a plastic backing (60–90 mm wide x 300 mm long), a membrane (25 mm wide), sample and absorbent pads (15–20 mm wide), and a conjugate pad (8–12 mm wide). The cut test strip is in the range of 3–8 mm wide x 60–90 mm long. Processing of these components is typically accomplished as described below.

Figure 1. Schematic of a lateral-flow test device showing laminate structure and reagent placements.

Membrane. The membrane substrate is usually made of nitrocellulose or a similar material. A positive-displacement dispensing system is normally used to apply test and control reagents to this substrate in the form of a line. An impregnation process can then be used to block the membrane, which is then dried in a batch or conveyor oven (see Figure 2). During test development, membrane blocking can also be done dynamically by impregnating the sample pad with blocking reagents.

Figure 2. Typical equipment and process flow for (a) batch and (b) in-line processing of lateral-flow test strips.

Conjugate Pad. This component is typically a polyester or glass fiber material that is treated with a conjugate reagent. A typical process for treating a conjugate pad is to use impregnation followed by drying. In use, the liquid sample added to the test must redissolve the conjugate so that it will flow into the membrane.

Sample Pad. The sample pad is treated with chemicals such as buffers or salts, which, when redissolved, optimize the chemistry of the sample for reaction with the conjugate, test, and control reagents.

Wicking Pad. This layer is used in tests where blood plasma must be separated from whole blood. An impregnation process is usually used to treat this pad with reagents intended to condition the sample and promote cell separation.

Absorbent Pad. This material acts as a reservoir for collecting fluids that have flowed through the device. It is not normally treated.

Plastic Backing and Adhesive. This material serves as a structural member for the above layers.

Whether the above components begin in sheet, strip, or web format, all are usually treated with the applicable reagents and processes before they are laminated together. However, some types of processing enable manufacturers to apply reagents after the test components have been partly laminated. Whichever process is used, reagent application is a critical step in achieving high-throughput, controlled test manufacturing.

Materials. Because the materials described above can exhibit large variations in liquid-holding capacity, they can represent a major source of problems for achieving controlled processes. On the typical scale of such lateral-flow tests, material thickness and density can vary as much as 30 to 50%. Such problems are most evident with the wicking, sample, and conjugate pads, which have traditionally been treated using impregnation processes. If not performed correctly, drying of impregnated materials can lead to nonuniformities among the test strips produced by the manufacturing process.

Density and thickness variations in test membranes are typically controlled much more readily than are variations in other components. When applying test and control reagents onto the test membrane, variations in membrane tolerances can generally be eliminated by using dispensing technologies that take advantage of very rapid protein-binding kinetics. For conjugate pads, a significant reduction in processing variations has been achieved by replacing the impregnation process with a dispensing process.¹ To make use of such an approach, it has been necessary to develop specially formulated high-density conjugate solutions.

Dimensional Tolerances. The functioning of a lateral-flow test is based on the unidirectional flow of the fluid sample through the different pad structures to the test membrane, and ultimately to collection in the absorbent pad. As the sample flows, it must redissolve dried reagents in each of the pads before entering the membrane. A typical flow response in the membrane is nonlinear and decreases rapidly with flow distance (see Figure 3).

Figure 3. Flow rate of liquid in a nitrocellulose membrane as a function of distance.

The effective concentrations and reaction rates of reagents at the test and control lines are dependent upon the flow of the sample in the test membrane. If reagent loading in the pads varies as a result of the materials issues described above, it can become a source of variation in the reagent concentration, and hence of variation in the reaction rate at the test and control lines. Similarly, the signal levels generated by the test and control lines are especially sensitive to their positioning within the device. For these reasons, the reagent loading of the various pads, the geometry of the lamination assembly, and the placement of test and control lines are critical factors in achieving controlled and quantitative test results.

High Throughput

Lateral-flow tests have historically been manufactured in relatively low volumes using a batch-processing approach, with a typical manufacturing line producing somewhere between 2 million and 10 million devices per year. The substrates used for batch processing can be in either a sheet or strip format, and are usually about 300 mm in length. In order to effectively handle the large number of materials involved in such manufacturing, batch processing is somewhat labor intensive. To increase production volumes using this approach, the manufacturer must dramatically increase the amount of floor space dedicated to equipment and operators.

For manufacturers seeking to increase production, the next logical step is to move toward in-line approaches that can reduce labor requirements and speed up processing. One of the most effective approaches to in-line processing is the use of webs and web-handling systems (see Figure 4). The objective is to design processes that can be carried out in parallel, taking advantage of the relatively rapid speed of web processing (in the range of 10–100 mm/sec). Such an approach can produce 5-mm-wide test strips at rates in the range of 2–20 parts/sec, which is equivalent to 12 million to 120 million parts per shift per year.

Figure 4. Layout of a reel-to-reel reagent-processing system incorporating units for dispensing, impregnation, and drying.

However, scaling up to such higher production volumes can also require changes in both the reagent application and lamination processes.

Reagent Processing. The batch approach to reagent application often includes impregnation dwell and drying times that are in the range of minutes to hours. Using web-based systems can dramatically reduce the time required for such reagent application. When in-line impregnation tanks and drying tunnels are used to process webs of about 100 mm width, for example, sample and conjugate pads can be produced at speeds in the range of 5–10 mm/sec. When such 100-mm webs are later slit to typical widths of 10 mm, this type of processing can achieve effective throughput in the same range as for dispensed processes.

However, because impregnation commonly results in the application of 10–100 times more liquid than does dispensed processing, impregnation remains a much slower method. With so much more liquid to be evaporated, reagent drying is the most critical step in impregnation processing. The time required for this step can be reduced by limiting the liquid volume absorbed into the substrate and by maximizing the drying temperature (see Figure 5).

Figure 5. Material temperature as a function of time during the drying process for impregnated substrates.

Drying is accomplished in two stages: evaporation and desorption. During the evaporation stage, the temperature of the substrate remains lower than the ambient air temperature because of the cooling associated with evaporation. During the transition to desorption, the temperature of the substrate begins to increase, reaching air temperature when the drying process is complete.

Because high temperatures can damage protein-based reagents, the general practice for batch processing is to maintain drying temperatures at less than 40°C over a period of many hours. This practice minimizes the problems of temperature gradients and water-loading variations that generally exist in batch-drying processes.

To keep pace with web speeds in the range of 10 mm/sec, however, an accelerated drying process is required. At such processing speeds, a drying tunnel of 72 in. can provide an effective drying time of three minutes, but complete reagent drying within that time requires higher temperatures than are usually permitted. By using an in-line approach, each portion of the web is exposed to the same drying conditions, making it easier for the manufacturer to dry the substrate according to specified parameters. Because exposure to high drying temperatures can thus be controlled and limited to a very short time, manufacturers can use temperatures in excess of 50°C without causing damage to proteins. To minimize the footprint required for such equipment, the necessary drying path can be constructed in the form of a vertical drying tower (see Figure 4).

In contrast to impregnation, dispensing processes for applying test and control reagents can be quite rapid. Replacing impregnation processes with the dispensing of conjugates can greatly reduce the volume of liquid required, hence allowing very rapid application and drying times.¹ Because both application and drying times can be quite fast, dispensing processes for test, control, and conjugate reagents can easily be scaled to web speeds in the range of 10–100 mm/sec. Typical drying times are in the range of one to two minutes.

Lamination. In the batch mode, lamination is basically a manual process using tooling and fixtures. One of the strengths of the lateral-flow format is the simplicity of its mechanical assembly, which involves the lamination of a membrane strip to a single adhesive layer on a plastic backing. An experienced operator can create one or two such laminated master strips per minute, equivalent to a throughput of 1–2 parts/sec, or 6 million to 12 million parts per shift per year. In reality, such production rates may not be achievable because of the potential for operator fatigue.

In-line approaches using combinations of pretreated web or strip stock can provide lamination speeds in the range of 50–100 mm/sec. This is equivalent to 5–10 parts/sec, or 30 million to 60 million parts per shift per year. Figure 6 illustrates an in-line lamination machine using web materials with reagents already applied to produce a laminated master strip cut to 300–500 mm. The laminate is cut to a master strip to avoid subjecting it to bending stresses that would cause damage.

Figure 6. Layout of a lamination system with combined web and strip feed. System output is a fully laminated master strip.

Economics of Manufacture. For production rates in excess of 2 million units per year, the use of in-line processing methods can increase output per employee by a factor of 10 or more. Such processes can reduce materials and reagent use by 10% or more, while also improving manufacturing yields and product quality. However, the transition to in-line processing also requires the manufacturer to underwrite the costs of equipment, which increases product costs by the amounts associated with capital amortization.

Controlled Processes

Manufacturers seeking to establish very tightly controlled manufacturing processes often face significant challenges caused by the substrates and reagents in use or by the processing technologies themselves. As in the case of increased throughput, increased control can require manufacturers to pay special attention to their reagent application and lamination processes.

Reagent Processing. When batch processing is employed, reagent application through impregnation is subject to several major problems. First, the substrates can become distorted during drying. Second, depending on the orientation of the material in the drying oven, reagents can become segregated or redistributed as a result of surface tension and gravity effects during the drying of strip materials. And finally, changes in the chemistry of the reagent bath over time can lead to different absorption rates for the various chemistries in the bath, resulting in nonuniform coating of the substrates.

The last of these problems can be solved through in-line processing using a small-volume reagent bath that is continuously replenished with fresh solution (see Figure 7). Another in-line solution is to use saturation dispensing, which adds a constant volume of reagent per unit length. The saturation dispensing method can result in microscopic gradients in reagent concentration caused by chromatographic effects as the solution flows transversely in the web. However, such gradients may not be significant for lateral-flow tests, since the direction of sample flow is normal to the web direction.

Figure 7. Close-up of the impregnation station for the reagent-processing system shown in Figure 4.

Lamination. The essential requirement of the lamination process is for the materials to be capable of sustaining process- and use-related stresses while maintaining their specified functionality. To exercise increased control over lamination, the primary challenges are maintaining relative dimensional tolerances and constant material tensions so that every test will function within a specified coefficient of variance (CV). Regardless of whether batch or in-line processing is employed, the lamination process is subject to three major variables: material tolerances, material mechanics, and material stresses.

When implementing tightly controlled processes, the dimensions and straightness of the substrate materials must be consistent with the tolerances established as part of the device design. Moreover, such tolerances must be maintained after reagent treatments. The applying and drying of reagents can significantly alter the dimensions of materials and distort their shapes. Compared with full impregnation, small-volume dispensing has less of an effect on the dimensions and shapes of substrates. Because in-line processing exerts a constant tension on the substrate throughout the reagent application and drying steps, it can help to maintain the shape and dimensional uniformity of the substrate along the web length. Using wide webs followed by slitting to smaller widths can also help to control the dimensional stability and shape of materials during lamination processes.

The mechanical properties of the substrates used for a test must be capable of meeting the design tolerances established for the device. Moreover, such materials must be chosen not only for their ability to function in the test device, but also for their ability to be processed. Many materials that function well from the point of view of a test's chemistry cannot be processed according to the required dimensional tolerances. In many instances, such as in the case of an unbacked nitrocellulose membrane, the problem is that the edges of the material are not mechanically well defined. One method of dealing with an unbacked membrane is to first laminate it to the plastic backing, and then to carry out the reagent dispensing and additional lamination steps. In this way, tolerances of pad placement relative to the test and control line positions can be better controlled using the plastic backing as a mechanical reference. Other considerations include the ability of materials to withstand the bending and shear stresses associated with the lamination and cutting processes. Again, many materials that function well chemically cannot withstand manufacturing processes and maintain test functionality. Fine fiberglass mat materials used for conjugate or sample pads are examples of materials that fail under bending stresses. Before selecting a material to be used in product development, the manufacturer should test it for manufacturability using the processes that will be employed in actual production.

The third consideration in lamination is that of material stresses, particularly stresses on the membrane as a result of web tension generated by mechanical handling of the materials. Such stresses can alter the effective porosity of a material and hence the flow rate of sample through that material. When compared with batch processing of short strips of materials, in-line lamination using web formats can provide more-effective control over material stresses. Web-handling equipment is generally designed to laminate materials under a condition of constant tension, which can be adjusted as necessary to avoid over-stressing the substrates (see Figure 6).

Quantitation

The development of lateral-flow tests from simple yes/no and threshold formats to true quantitative tests is an active area of R&D. To be useful, most test analytes require some level of quantitative readout. One of the key factors necessary for achieving quantitative test formats is the development of controlled and quantitative manufacturing processes.

Reagent Processing. When developing quantitative tests, manufacturers must go beyond merely controlling their reagent processing to achieving uniform and quantitated distribution of reagents on the various test substrates. Such precision cannot be accomplished using impregnation processing because that method is subject to variations caused by differences in the liquid-holding capacities of the materials or by reagent redistribution during drying.

The best approach to achieving truly quantitative reagent application seems to be through dispensed processing using a positive-displacement system. Three such platforms include a drop-on-demand ink jet (BioJet), an aerosol dispensing system (AirJet), and a contact or near-contact syringe needle.^2–6 Each of these systems can be programmed to deliver a specified volume per unit length, and can deliver a dispensed line with a CV of less than 5% (see Figure 8). The BioJet and syringe needle can dispense narrow lines, with profiles determined respectively by the drop size or syringe orifice size. The AirJet delivers droplets in the picoliter range with an angular distribution; line width is controlled by varying the height above the substrate, the aerosol pressure, and the flow rate of the liquid being dispensed. Typical line widths for an AirJet are in the range of 0.5 to 10 mm, but jets with wider angles can produce line widths up to 25 mm.

Figure 8. Dispensed lines of conjugate on a fiberglass sheet: (a) lines dispensed using an ink jet and (b) lines dispensed using an aerosol system. In both examples, the liquid volume dispensed is 2 µl/cm of gold conjugate concentrated to an optical density of 100.

By using dispensed processing, manufacturers can achieve quantitative delivery of reagents. Such precision enables the company to reduce the liquid volume applied to test substrates, thereby speeding up drying processes and reducing the risk of material changes associated with drying. Thus applied, the reagent should appear on the substrate as a uniform band, with penetration through the pad so that it offers a uniform front to the flow of the test sample. In addition, the dried reagent should be readily redissolved by the test sample.

Lamination. For quantitative tests, the major issue relating to lamination is the precision placement of parts relative to a reference such as a plastic backing edge (see Figure 9). When batch processing is used, placement precision is limited by the mechanical stability of the substrate edges. Placement of materials with good edge definition can be controlled to within tolerances of ±1 mm. However, the placement of dispensed reagents on such batch-processed materials can also vary, thereby increasing the potential for placement error relative to actual flow position on the membrane. When the placement errors from batch processing and reagent dispensing are combined, the effective placement error between the entry point of a sample into the membrane and the test and control lines can easily be as much as ±2 mm—a measure far too great to be acceptable for a quantitative device.

Figure 9. Schematic of a lateral-flow device showing the position of a dispensed conjugate line in the conjugate pad relative to the membrane and sample pads.

The placement tolerances of both reagents and lamination can be better controlled by employing an in-line processing system that makes use of adjustable mechanical guides and sensor-based tracking systems. Figure 11 shows an in-line dispensing module in which the dispensing heads are located on a servo-driven slide that operates in conjunction with a sensor system for detecting the edge of the substrate. When processing materials with well-defined edges, systems such as this can control the accuracy of reagent placement to within ±0.25 mm. This type of dispensing system can work well even with unbacked nitrocellulose membranes. Tracking systems similar to that used on the illustrated system are also used to maintain edge alignment during the rewinding of processed roll-stock and for in-line lamination.

Figure 11. Schematic of QC inspection for a master strip, showing defects and related inspection marking.

QC Inspection

Quality control (QC) is one of the most neglected areas in the manufacturing of lateral-flow test devices. Such neglect is due, in part, to the lateral-flow format itself, which makes it difficult for manufacturers to inspect, detect, and remove bad components during processing. Even when the manufacturer recognizes that a processing step has exceeded its assigned parameters, the resulting faulty portions or components often cannot be physically removed until the master strip or webs have been cut into individual parts.

For manufacturers, the difficulty of carrying out QC operations implies several related challenges, including how to inspect for process defects, how to track identified defects through the manufacturing process, and how to remove faulty components or bad sections of processed materials. Although such operations can be carried out manually—by cutting the master strips and then sorting out the good parts from the bad—such an approach adds processing steps and requires additional labor and equipment.

An alternative strategy for tracking bad parts is illustrated in Figure 12, which shows an uncut master strip that has been inspected for defects in reagent application and lamination processes. The presence of a defect is indicated using a master mark on a specified portion of the laminate, which can then be automatically read during the cutting process. As the part is passed through the cutter, faulty portions of the laminate are automatically detected and removed from the process. Methods of inspection and marking are discussed below.

Figure 12. An in-line system for reagent dispensing on sheet or strip substrates, combined with an automated QC inspection sensor and marking system.

Reagent Processing. When impregnation processing is used, the problems of inspecting and marking defects can be difficult to overcome. With impregnated substrates, defects tend to be related to gradients in the composition of the reagent rather than to the simple absence or misplacement of the reagent. For this reason, there is often no clear-cut correlation between a defect and an optical readout. Such complexities in carrying out QC on impregnated substrates are among the many reasons that the industry trend is toward the increased use of dispensing processes.

When dispensed processing is used, inspection and marking can be done either manually or with machine control. Inspection can be performed by means of an optical readout such as a dye or by an optical sensor that detects the presence or absence of reagents using the contrast between wet and dry areas as an indicator. When the sensor detects an area in which reagent is absent or has been misapplied, software in the inspection system triggers an ink jet to mark the exact position of the defect (see Figure 10). It is also possible to use combinations of sensors to measure the width of dispensed reagent lines, thereby providing an indirect measurement of the reagent volume dispensed.

Figure 10. Close-up of the dispensing station for the reagent-processing system shown in Figure 4. The station includes ink-jet and aerosol dispensers and automated edge tracking.

The next step up in QC measurement is to use an in-line vision inspection system that measures the width of the reagent line, its position, and its intensity profile.⁷ Vision inspection systems offer a more-quantitative approach to QC inspection. Depending on the visual resolution required and the speed of the manufacturing process, such systems may use more than one camera.

Both optical and vision inspection systems can be adapted for use with either batch (x-y motion) or in-line web systems.

Lamination. QC for laminated substrates can also be accomplished using sensor arrays, vision inspection systems, and automated marking methods. In this application, such systems inspect the placement of the various materials during lamination. Vision inspection systems generally offer better resolution and sensitivity.

Defects should be marked in a standard position on the membrane. This practice ensures that defect marks can be detected during later processing steps, so that faulty parts can be rejected.

Hybrid Systems

At the high end of the equipment spectrum are hybrid systems that combine dispensing and lamination into a single process, working at speeds up to 100 mm/sec (see Figure 13). Such systems enable manufacturers to exercise greater control over the relative positioning of both substrates and reagents.

Figure 13. Layout of an in-line hybrid processing system. Such systems laminate adhesive and membrane to strips of bare plastic backing, and then dispense test and control reagents.

The use of hybrid systems represents a new paradigm for the manufacture of lateral-flow tests because it allows for the dispensing of different reagents on different substrates while maintaining high relative positional accuracy (in the range of ±0.1 mm). Thus the placement of the conjugate line can be very accurate relative to the test and control lines and independent of the limitations of material dimensional tolerances. The same is true for the placement of dispensed reagents on sample and wicking pads.

Conclusion

IVD manufacturers are currently using both batch and in-line processing for the manufacture of lateral-flow tests (see Table I). As demand for such tests continues to grow, companies will increasingly be seeking to employ processing technologies that enable them to produce greater volumes of tests. At the same time, product developers are looking for ways to better control their processes, so that the lateral-flow format can be used for quantitative tests.

Table I. Comparison of the key characteristics of batch and in-line manufacturing strategies for lateral-flow test devices.

Regardless of the processing method used to manufacture such tests, manufacturers should ensure that they select reagents, substrates, and other materials that are compatible with those processes. Correct selection of materials and components is the essential basis for developing controlled manufacturing processes.

Advancing technologies are offering manufacturers significant alternatives for improving their manufacturing methods. Both batch and in-line processing can be improved by replacing impregnation with quantitative dispensing of reagents. Reagent dispensing and lamination can be performed using continuous in-line processes that now offer considerable advantages in throughput, quality, and cost. And new hybrid systems are showing potential for additional improvements in the development of more-quantitative process technologies.

Together, these major trends in the manufacture of lateral-flow tests suggest a continually evolving technology that IVD manufacturers should be keeping abreast of. Staying informed about the latest processing technologies can help manufacturers gain production efficiencies that can ultimately improve their bottom line.

References

1. J Colanduoni, unpublished research, Arista Biologicals Inc., 115 Research Drive, Bethlehem, PA 18015.

2. TC Tisone, Dispensing Systems for Miniaturized Diagnostics, IVD Technology 4, no. 3 (1998): 40–46.

3. TC Tisone, Precision-metered solenoid valve dispenser, U.S. Pat. 5,743,960, April 28, 1998.

4. TC Tisone, Precision-metered aerosol dispensing apparatus, U.S. Pat. 5,738,728, April 14, 1998.

5. TC Tisone, Method of dispensing a liquid reagent, U.S. Pat. 5,741,554, April 21, 1998.

6. TC Tisone, Dispensing apparatus having improved dynamic range, U.S. Pat. 5,916,524, June 29, 1999.

7. TC Tisone et al., Image Analysis for Rapid-Flow Diagnostics, IVD Technology 5, no. 5 (1999): 52–58.

Thomas C. Tisone is vice president for R&D and engineering at BioDot Inc. (Irvine, CA)