The new and rapidly expanding field of nanoscale medical diagnostics requires the development of advanced sensor technologies to provide the selectivity and sensitivity needed to measure concentrations of target biomarker molecules at subnanomolar concentrations. These demands for higher sensitivity and, hence, lower detection limits have driven the development of novel approaches to molecular detection. Over the past 20 years, there have been only minor advancements in the technologies used in instruments designed for molecular spectroscopy, including those based on the Raman effect. Commonly, slit or pinhole apertures spatially define the light entering the section of a spectrometer that separates and detects wavelengths. In modern instruments, scanning monochromators largely have been replaced with fixed dispersion gratings coupled with either linear or two-dimensional CCD array detectors. To achieve useful spectral resolution, the entrance aperture is usually fixed at around 50 to 100 μm. This small entrance aperture severely limits light-collection efficiency and, as a result, sensitivity. Recently, analytical spectrometers using multimodal multiplex spectroscopy technology have been introduced. These spectrometers employ wide-area, binary-encoded apertures and computational transforms to greatly improve sensitivity without compromising spectral resolution. Figure 1. A schematic illustrates the optical layout of a typical slit-based spectrometer (left), accompanied by a typical CCD detector image and spectrum that it can record (right). In a slit-based dispersive instrument, the spectral width of a resolution element is directly proportional to the width of the slit (Figure 1). This relationship provides a challenge to spectroscopy in real-world conditions. To achieve a reasonable spectral resolution, the input slit must be narrow. However, real-world light sources are so diffuse in nature that only a small percentage of the total light can enter the narrow slit of the spectrometer. If the light source is weak, then the detector can be so starved of photons that no usable spectral measurement is possible. In these traditional designs, there is an inherent trade-off between resolution and light throughput. In contrast, an instrument based on multimodal multiplex spectroscopy technology samples several thousand optical channels simultaneously through a coded aperture instead of through a single-channel slit (Figure 2). Then mathematical algorithms precisely reconstruct the intensity-versus-wavelength information. Using this technology, both resolution and throughput can be maintained and optimized. The most dramatic performance advantage is realized when measuring weak, scattered or diffuse sources. Figure 2. This schematic shows the optical layout of a multimodal, multiplexed spectrometer (left), along with a typical CCD detector image and spectrum (right). In particular, for a Raman spectroscopy system, theory predicts the throughput of an encoded mask to be about 800 times that of a fiber-based system and 12 times that of a slit system with equivalent resolution. To validate this theory, data were collected using a spectrometer comprising identical optical and electronic components, with the only difference between experiments being the change in the entrance aperture. The spectra of toluene — comparing pinhole, slit and multimodal multiplex spectroscopy approaches — are shown in Figure 3. The multimodal, multiplex spectroscopy technique provides high sensitivity and resolution. Figure 3. The Raman spectra of toluene with pinhole (left), slit (middle) and multimodal, multiplexed (right) configurations show how the latter technique improves the sensitivity and reproducibility of the data. Lateral flow immunoassay Influenza virus causes significant illness and mortality worldwide. According to the Centers for Disease Control and Prevention in Atlanta, influenza infections result in more than 114,000 hospitalizations and 36,000 deaths annually and, together with pneumonia, represent the sixth-leading cause of death in the US. A rapid and accurate means of detecting the virus is, therefore, very important. Current commercial methods for detecting influenza include tissue culture isolation, which requires two to 14 days to complete, or ELISA measurements, which require two to six hours. There also are various rapid but less accurate lateral flow immunoassay tests. In a traditional lateral flow immunoassay, the sample to be analyzed is deposited on one end of a cellulose strip and the results are read at the other. As the sample flows through the strip, an antibody containing a spectroscopic tag typically attaches to the biomarker molecule. As the labeled biomarker moves farther down the strip, it reaches a zone where it becomes attached to another receptor that is fixed in place at the test/read point on the strip. The test is then analyzed to determine the quantity of bound biomarkers by quantifying the amount of spectroscopic tag that is present. Common measurement techniques include colorimetry, chemiluminescence and fluorescence. It also is possible to measure the concentration of the biomarker using Raman spectroscopy, which, although traditionally not a trace analysis technique, is very sensitive to low concentrations when combined with surface-enhanced Raman spectroscopy. Surface-enhanced Raman effects occur when molecules are in proximity to certain metallic surfaces. The metal causes orders-of-magnitude enhancement in the Raman signal. In an experiment to test the ability of Raman spectroscopy to improve lateral flow immunoassays, the surface-enhanced particles were provided by Oxonica plc of Yarnton, UK. The particles, which consist of a gold nanoparticle core that is coated with the Raman probe molecule (that is, the reporter), are further encased in a glass layer that stabilizes the Raman probe molecules. In the Raman-based lateral flow immunoassay, the antigen bound to the surface of the glass layer and the Raman signal was recorded from the reporter molecule when the biomarker/antigen was bound at the measurement point of the test strip (Figure 4). Figure 4. This schematic shows the steps involved in a typical lateral flow assay, including sample deposition, tagging and detection. The test solution that includes the biomarker was applied to the end of the test strip where it then bound to the labeled nanoparticles. The labeled biomarker and the unattached tags flowed down the strip. At the test point, the tagged biomarkers bound to receptors fixed on the strip, and the reporter molecules were read. The unbound reporter molecules (the background) were read at the control point. The difference between the two signals was used to calculate the concentration of the biomarker and, therefore, the presence or absence of an infection. Another advantage of using the multimodal multiplex method with Raman spectroscopy for bioassay measurements versus conventional detection approaches is the spectral contrast that enables the measurement of multiple Raman reporter molecules — hence, multiple biomarkers — simultaneously. This is possible because of the narrow features of the Raman emission spectra. Besides providing higher sensitivity through increased spectral contrast, the multimodal, multiplex, wide-area-aperture approach also is suited to large-area sampling. Typical Raman spectrometers use an illumination spot size of ~100 to 150 μm. Many samples are not homogeneous over such a small area and, as a result, the measured spectrum may not be representative of the bulk of the sample. For example, a surface-enhanced Raman spectroscopy substrate was measured with a conventional Raman spectrometer that had a spot size of ~100 μm and with a multimodal, multiplex system with a sample spot size of 2 mm. A series of spectra was collected across the sample, using 500-μm steps (Figure 5). Using conventional Raman spectroscopy, the spectral intensities show a wide variation because of the surface variations. On the other hand, the much larger illumination spot size of the multimodal, multiplex system resulted in more reproducible spectra. Figure 5. In a spectrographic analysis of a sample surface covered with nanoparticles (inset), the Raman spectra acquired using a typical slit-based system (top) offer less sensitivity than those acquired using the multimodal, multiplexed system (bottom). These results also demonstrate the improvement in signal-to-noise that results from using the multimodal, multiplex approach. Such uniform detection is a critical part of obtaining accurate, reproducible bioassay measurements. Using three Raman reporter nanoparticles, it is possible to measure more than one virus at a time. For example, influenza A, influenza B and respiratory syncytial virus are detectable in a single 15-minute assay with a detection limit of approximately 10 ng/ml — an order of magnitude better than current lateral flow immunoassays using colorimetric detection. In summary, Raman spectrometry enhanced with multimodal, multiplex technology provides approximately 12 times the throughput of a conventional slit-based system with equivalent resolution. This translates into a signal-to-noise advantage of more than 3 1/2 times for equivalent measurement periods. In addition, the wide-area aperture is suited to large sample spot illumination, which yields measurements that are more representative of the bulk of the sample being analyzed. This greatly enhanced sensitivity and larger area sampling make the multimodal, multiplex Raman spectrometry system ideal for surface-enhanced bioassay measurements. Acknowledgments The authors thank Scott Norton and William Doering at Oxonica for providing the samples used in these experiments and for their technical help. Meet the authors Michael Fuller (e-mail: mfuller@centice.com) is senior director of product development, Prasant Potuluri (e-mail: potuluri@centice.com) is a cofounder and chief technology officer, and Michael Sullivan (e-mail: msullivan@centice.com) is a cofounder and vice-president of business development of Centice Corp. in Morrisville, N.C.