Wood's anomalies

Between 1902 and 1912 R.W. Wood (1868-1955) at Johns Hopkins University (Baltimore, USA) noticed that when he shone polarized light onto a metal-backed diffraction grating, a pattern of unusual dark and light bands appeared in the reflected light (1),(2). Although he speculated about how the light, gratings and metal interacted, a clear answer to the phenomenon was not provided.

The first theoretical treatment of these anomalies was made by Lord Rayleigh in 1907 (3). He based his "dynamical theory of the grating" on an expansion of the scattered electromagnetic field in terms of outgoing waves only. With this assumption, he found that the scattered field was singular at wavelengths for which one of the spectral orders emerged from the grating at the grazing angle. He then observed that these wavelengths, which have come to be called the Rayleigh wavelengths λR, correspond to the Wood anomalies. Furthermore, these singularities appeared only when the electric field was polarized perpendicular to the rulings, and thus accounted for the S anomalies; for P polarization, his theory predicted a normal behaviour near λR (4). Wood's later papers, (2),(5) however, suggest that also P anomalies could sometimes be observed. Palmer (6),(7) very clearly demonstrated that P anomalies did exist in deeply ruled gratings. Thus, anomalies of both the S and P type were obtainable, but P anomalies were found only on gratings with deep grooves (4). Theoretical analysis undertaken by Fano in 1941 led to the conclusion that these anomalies were associated with surface waves (surface plasmon) supported by the (grating) network (8).

In the fifties more experimentation was done on electron energy losses in gasses and on thin foils (9),(10). Pines and Bohm suggested (11),(12),(13) that the energy losses were due to the excitation of conducting electrons creating plasma oscillations or plasmons. Further research (10) revealed that the energy loss resulted from excitation of a surface plasma oscillation in which, part of the restoring electric field extended beyond the specimen boundary. Therefore, the presence of any film or contaminant on the specimen surface affects the surface plasma oscillation. This effect was later described in terms of excitation of electromagnetic ‘evanescent’ waves at the surface of the metal, and in the 1970s evanescent waves were described as a means to study ultra-thin metal films and coatings (14).

Surface plasmons

In the late sixties, optical excitation of surface plasmons by means of attenuated total reflection was demonstrated by Kretschmann and Raether (15),(16) and Otto (17).

There have been two major approaches to optical excitation of surface plasma waves: attenuated total reflection in prism coupler-based structures and diffraction at gratings. The application of surface plasma waves excited by the attenuated total reflection method for sensing has been pioneered by Nylander and Liedberg (18). Particularly because of its relative simplicity, this method has been widely applied for characterization of thin films (19),(20) and (bio)chemical sensing (21),(22),(23).

The use of diffraction grating-based systems for SPR sensing has been advocated first by Cullen et al. (24). The grating-based surface plasmon resonance (SPR) sensors have been studied as an alternative to prism-based systems (25),(26).

In the 1980s, SPR and related techniques exploiting evanescent waves were applied to the interrogation of thin films and biological and chemical interactions (19),(21),(27),(28),(29). These techniques allow the user to study the interaction between immobilized receptors and analytes in solution, in real time and without labelling of the analyte. By observing binding rates and binding levels, there are different ways to provide information on the specificity, kinetics and affinity of the interaction, or the concentration of the analyte.

SPR biosensors

In 1980, Pharmacia became interested in SPR and began investigating the possibilities of the technique. In 1984, Pharmacia founded the company Pharmacia Biosensor AB to develop, produce and market a functional SPR-machine. The development of appropriate sensor surfaces by Pharmacia Biosensor (30),(31) and the fabrication of the silicon microfluidic cartridge brought an easy-to-use SPR-machine closer to becoming a reality (32).

In a short period, many publications from Pharmacia Biosensor described the new hydrogel of dextran (30),(31), the correlation between the SPR signal and the RIA assay (33),(34) and gave a description of the BIAcore machine (33),(35). BIAcore instruments make use of a wedge-shaped laser beam and a diode array for detection, which results in no moving parts in the detection unit.

In 1990 the first BIAcore was sold (36). In 1994 a simplified machine, the BIAlite was released. With this machine, the sample handling was manual instead of computer controlled. The development of different, more sensitive and specialized machines gave us the BIAcore X, 2000, 3000 and Q for quality control. Other developments involved the way the liquid was handled.

Typically, SPR machines use microfluidic channels with valves to address the sample to different sensor spots. In 2005 the first machine (BIAcore A100) with dynamic addressing was released (37). The four flow channels in the microfluidic cartridge are much wider and have five detection spots in each channel. The spots can be hydrodynamically addressed by changing the flow rate from two inlet channels, one with sample and one with buffer. The latest introduction of Biacore is the 8K/8K+ system with 16 flow cells in 8 channels. This instrument is positioned as a high-throughput, high sensitive instrument for screening and characterization of small molecules in drug discovery programs.

Over the years, different manufacturers developed other SPR systems. The Spreeta Evaluation module of Sensata Technologies, which can be used to make an in-house system, is probably the simplest SPR measuring device. The ProteOn XPR36 system from Bio-Rad uses a crisscross 6 x 6-interaction array capable of simultaneous measurements of 36 interactions. The Sierra SPR-32 instrument from Bruker has eight flow channels with each four detection spots which are individual addressable. Some machines make use of a cuvette system (e.g. the IBIS Biosensor, IAsys Biosensor) in which binding of large cells is possible. Other machines use a resonant mirror to determine the resonance angle (e.g. IAsys Biosensor, Horiba SPRi). The SPRi systems make an image of the reflectivity of the sensor surface. This set-up makes it possible to monitor many interactions (up to several hundreds) at the same time.

Multi-Parametric SPR

Multi-Parametric Surface plasmon resonance (MP-SPR) was introduced to the market in 2011 by BioNavis from Finland. The instrument, targeted at Life Sciences was the first commercial instrument that offered measurements with multiple wavelengths. The sensor slide range was extended extensively with Ag, Pt, Cu, Al, TiO₂, SiO₂, CMD, Ni²⁺ sensors. In 2012 a special model for the Material Science market was introduced. In addition, the novel technique of Selectively Amplified SPR was introduced for signal enhancement of tricky assays. A year later an extensive range of flow-cells was introduced such as the standard two channel, electrochemical, High Chemical Resistance and one for measurements in gasses. For the first time, a commercial SPR instrument succeeds to measure transcellular and paracellular drug-cell interactions in a label-free manner (38).

Localized SPR

Other types of SPR – the so called Localized SPR (LSPR) – are marketed by for instance Nicoya since 2016. LSPR is induced on nanoparticles as opposed to traditional SPR which uses a planar surface. Instruments using LSPR are generally smaller (benchtop) and less susceptible to temperature and bulk refractive changes. In 2020 Nicoya introduced the 16 channel Alto instrument which is using digital microfluidics to move the sample over the detection area.

Evaluation software

With the new instruments, the machine control software has greatly improved by taking the scientist by the hand in performing experiments. In addition, the software that analyses the sensorgrams is much better nowadays. For instance, most programs have routines for automatic cleaning and aligning of the curves. From analysing curve by curve, to global analysis of one single dataset, the new software also makes it possible to analyse several datasets consisting of several different analytes in different concentrations at the same time.
However, manufacturers provide software that is dedicated to their machine. The buyer is greatly dependent on the options of the software and it is difficult to reanalyse results with different software programs.