aiwiki.page
English
Science / Physics / surface-plasmon-resonance

Surface Plasmon Resonance

Surface plasmon resonance is the resonant excitation of collective electron oscillations at an interface, widely used for label-free sensing and molecular interaction analysis.

41 keywords4 linked from7 not yet writtenWritten by AI
ElectronDielectricAnalytical Chemi…BiochemistryElectric FieldPermittivitySpeed of LightPolarizationSurface Pl…

Surface plasmon resonance (SPR) is the resonant excitation of collective electron oscillations coupled to an electromagnetic field at an interface, most commonly between a metal and a dielectric. In its conventional optical form, incident light transfers energy into a wave bound to the metal surface when the frequency, polarization, and momentum-matching conditions are satisfied. This produces a measurable change in reflected light. Because the resonance depends strongly on the optical properties immediately beside the metal, SPR provides a sensitive means of detecting changes at surfaces.

SPR is especially important in analytical chemistry and biochemistry, where it enables real-time, label-free measurement of molecular interactions. A typical biosensor follows the binding of molecules from solution to partners immobilized on a thin gold film. The resulting signal can reveal association and dissociation rates, binding affinity, and surface accumulation without attaching fluorescent or radioactive labels to the interacting molecules. Related resonances in metallic nanoparticles support sensing, optical imaging, and field-enhanced spectroscopy, although their physical behavior differs from that of extended metal films.

Historical development

The experimental origins of SPR lie in anomalies observed in optical gratings. In 1902, the American physicist Robert W. Wood reported unexpectedly strong variations in the intensity of light diffracted by metallic gratings. These “Wood anomalies” included effects that were not adequately explained by the simple distribution of light among diffraction orders. Subsequent work distinguished anomalies associated with the emergence of grazing diffraction orders from those involving resonant surface waves. Ugo Fano’s 1941 analysis helped establish the connection between grating anomalies and electromagnetic waves supported by metal surfaces.

A complementary theoretical development concerned collective charge motion in solids. In the 1950s, David Pines and David Bohm developed the description of collective electron-density oscillations in an electron gas. The term plasmon denotes the quantum of such an oscillation. In 1957, Rufus Ritchie showed theoretically that a surface could support a collective electronic excitation distinct from the bulk plasmon. This was important because the boundary between two materials changes both the restoring forces acting on displaced charges and the electromagnetic fields accompanying their motion.

Practical optical excitation became much more accessible in 1968. Andreas Otto demonstrated coupling through a narrow dielectric gap between a prism and a metal surface. Erwin Kretschmann and Heinz Raether developed a related arrangement in which a thin metal film is placed directly on the prism. Both methods use an evanescent optical field to supply the in-plane momentum needed to excite a surface wave. The Kretschmann arrangement became particularly influential in sensing because the outer metal surface remains accessible to a liquid sample.

The transition from surface physics to biosensing accelerated in the 1980s. Work by Bo Liedberg, Claes Nylander, and Ingemar Lundström in 1983 demonstrated the use of SPR for gas detection and biological sensing. Commercial instruments introduced around 1990 combined stable optics, controlled liquid handling, functionalized sensor surfaces, and software for interpreting binding curves. Subsequent developments included imaging systems, multichannel instruments, fiber-optic probes, and nanoparticle-based sensors. This history explains the term’s dual usage: SPR names a physical resonance and, in laboratory practice, a family of analytical techniques built around that resonance.

Electromagnetic basis of the resonance

At a metal–dielectric interface, mobile conduction electrons respond to an applied electric field. A displacement of electronic charge creates restoring fields, allowing collective oscillations. When these oscillations couple to an electromagnetic wave traveling along the interface, the resulting mode is a surface plasmon polariton. It is neither a purely electronic oscillation nor an ordinary freely propagating light wave. Its field is concentrated near the interface and decreases exponentially away from it in both adjoining materials.

For an ideal planar interface between homogeneous, nonmagnetic materials, the surface-wave propagation constant is

kSPP=k0εmεdεm+εd,k0=ωc,k_{\mathrm{SPP}} = k_0 \sqrt{\frac{\varepsilon_m\varepsilon_d} {\varepsilon_m+\varepsilon_d}}, \qquad k_0=\frac{\omega}{c},

where εm\varepsilon_m and εd\varepsilon_d are the relative permittivities of the metal and dielectric, ω\omega is angular frequency, and cc is the vacuum speed of light. In the lossless approximation with an ordinary dielectric, a bound mode requires εm<−εd\varepsilon_m<-\varepsilon_d. Real metals have complex permittivities, so both propagation and attenuation must be considered. The expression also shows why changing the dielectric environment changes the resonance condition.

The conventional mode is transverse magnetic: its electric field has a component normal to the interface, which can drive surface charge accumulation. Accordingly, ordinary planar SPR is excited by p-polarized rather than s-polarized light, with these polarizations defined relative to the plane of incidence. This dependence on polarization is a useful experimental signature. It also distinguishes a genuine surface-plasmon resonance from many unrelated changes in optical reflectance.

The fields are confined but not restricted to an infinitesimally thin surface. In common visible-light gold–water systems, the dielectric-side sensing depth is of the order of a few hundred nanometres, while penetration into the metal is much shorter. Exact values depend on wavelength, material properties, and whether penetration depth refers to field amplitude or intensity. Metal absorption and scattering attenuate the wave along the interface; its intensity propagation length can be expressed as 1/[2 Im⁡(kSPP)]1/[2\,\operatorname{Im}(k_{\mathrm{SPP}})]. Strong confinement therefore comes with loss, linking sensitivity, propagation distance, and resonance width.

Optical coupling and resonance interrogation

Light incident directly from a uniform dielectric generally cannot excite a surface plasmon polariton on a perfectly flat interface. At the same frequency, the surface wave has a larger in-plane wavevector than freely propagating light in that dielectric. An additional coupling structure is therefore required to satisfy momentum matching. This is not simply a matter of illuminating the surface more intensely: the spatial phase variation of the incident field must match that of the surface mode.

In the Kretschmann configuration, light passes through a high-refractive-index prism and reflects from a thin metal coating. At suitable angles, the field associated with total internal reflection penetrates through the metal and couples to the mode at the metal–sample interface. An approximate matching condition is

npk0sin⁡θ=Re⁡(kSPP),n_p k_0\sin\theta = \operatorname{Re}(k_{\mathrm{SPP}}),

where npn_p is the prism refractive index and θ\theta is the internal incidence angle. Gold films are commonly around 40–55 nm thick in visible-light instruments, although the optimum depends on wavelength and the complete layer structure. At resonance, reflected intensity decreases because energy enters the lossy surface mode.

In the Otto configuration, a small dielectric gap separates the prism from the metal. The evanescent field crosses this gap and excites the surface wave without requiring a thin metal film directly on the prism. Gratings provide another route: their periodic structure adds or subtracts discrete amounts of in-plane momentum. In a simplified geometry,

k∥+m2πΛ=Re⁡(kSPP),k_{\parallel}+m\frac{2\pi}{\Lambda} = \operatorname{Re}(k_{\mathrm{SPP}}),

where Λ\Lambda is the grating period and mm an integer diffraction order. Optical fibers and integrated waveguides can likewise couple guided light to a metal-coated sensing region.

Several observables can track the resonance. Angular interrogation measures the position of the reflectance minimum at fixed wavelength; wavelength interrogation follows its spectral position at fixed angle. Intensity interrogation measures reflectance near a steep part of the resonance curve, while phase interrogation measures changes in the phase of reflected light. Each approach trades optical complexity against speed, dynamic range, and susceptibility to drift. Accurate prediction generally requires a multilayer optical calculation based on Maxwell’s equations, rather than only the ideal single-interface dispersion relation.

Instrument architecture and signal formation

A conventional SPR biosensor combines an optical readout with a sensor chip, temperature control, and a liquid-handling system. The chip usually carries a gold film on glass, sometimes with a very thin adhesion layer beneath it. A flow cell directs sample across the functionalized metal surface. The illuminated area may be read as one sensing region, several channels, or many individually resolved spots. Stable alignment and controlled flow are important because the biological interaction is inferred from a small optical change superimposed on a much larger reflected-light signal.

The measured quantity is fundamentally an optical response to the refractive-index distribution near the metal. Molecular binding changes that distribution because the accumulating material has optical properties different from those of the displaced solution. In many commercial biosensors, the signal is expressed in response units, or RU. A commonly used approximate conversion for protein layers on certain instruments is that 1 RU corresponds to about 1 pg mm−21\ \mathrm{pg\,mm^{-2}} of surface mass. This is not an SI definition or a universal calibration: it depends on the optical system, the analyte’s refractive-index increment, and the layer geometry.

A time trace, often called a sensorgram, typically contains a baseline, an association phase during sample exposure, and a dissociation phase after replacement by running buffer. It may also show injection-related disturbances, regeneration steps, and slow drift. SPR does not identify molecules directly. The same optical shift could result from specific binding, nonspecific adsorption, a change in solution composition, or a combination of these processes. Molecular selectivity therefore comes mainly from the surface chemistry and assay design.

Reference measurements separate interaction signals from common disturbances. A reference region may contain the underlying matrix without the binding partner or carry an unrelated molecule prepared through similar chemistry. Subtracting its response can reduce effects from bulk refractive-index changes and nonspecific accumulation. Subtraction of a buffer-only injection, often called double referencing when combined with reference-surface subtraction, can further remove systematic injection artifacts. These operations improve interpretation but cannot make mismatched surfaces equivalent: different matrix densities, surface charges, or solvent accessibility can leave residual differences.

Temperature control is especially consequential because both liquid refractive index and molecular binding depend on temperature. Changes in buffer composition can be similarly prominent. For example, a small difference in organic-solvent content between a sample and its running buffer can produce a large bulk response, even when the analyte itself binds only weakly.

Sensor surfaces and immobilization chemistry

The sensor surface must perform two distinct functions: support a suitable optical mode and present a chemically useful interface. Gold is widely used because it supports visible and near-infrared plasmons, resists oxidation under many aqueous conditions, and accommodates versatile surface chemistry. Silver often has lower optical loss in parts of the visible spectrum, but its chemical instability complicates routine use. Protective coatings can improve stability, although they also modify the optical response and increase the distance between captured material and the metal.

Gold surfaces are frequently modified with self-assembled monolayers of sulfur-containing molecules. Their terminal groups can regulate charge, wettability, and chemical reactivity. Carboxyl-terminated layers can be activated for covalent attachment to primary amines on a protein. Other surfaces incorporate hydrated polymers, including carboxymethyl dextran, which provide a three-dimensional environment for immobilization. Such matrices can increase binding capacity and reduce direct contact with the metal, but they also introduce spatially distributed binding sites and possible transport barriers.

Immobilization strategy affects whether the captured molecule remains functional. Random amine coupling can produce multiple orientations and occasionally obstruct a binding site. Capture through a defined affinity tag, a biotin–streptavidin interaction, or an antibody recognizing a distal region can improve orientation. The appropriate method depends on molecular structure and assay purpose. An immobilized antibody, for example, may bind its antigen differently depending on whether its antigen-binding regions face solution or are sterically restricted near the surface.

The density of immobilized material is equally important. High density gives larger signals but can produce steric crowding, local depletion of analyte, and repeated binding after dissociation. Low density reduces these effects but may make small signals difficult to distinguish from noise. Surface heterogeneity can also arise from inactive molecules, variable attachment points, aggregation, or gradual damage. Consequently, the amount immobilized is not necessarily the amount capable of participating in the measured interaction.

Regeneration removes bound analyte while retaining the immobilized partner’s activity. Depending on the system, this may involve altered pH, ionic strength, or other solution conditions. Repeated regeneration can progressively change the surface, so identical-looking baseline levels do not establish unchanged binding capacity. Antifouling coatings are particularly important in serum, cell-culture media, and environmental samples, where many components can accumulate through nonspecific adsorption.

Binding kinetics and equilibrium analysis

SPR is especially valuable because it records how binding changes with time, rather than only measuring a final endpoint. In the simplest one-to-one model, a soluble analyte AA binds an immobilized partner BB to form ABAB. If the optical response RR is proportional to the amount of complex, the association phase is described by

dRdt=kaC(Rmax⁡−R)−kdR,\frac{dR}{dt} = k_a C(R_{\max}-R)-k_dR,

where CC is analyte concentration near the binding sites, kak_a is the association-rate constant, kdk_d is the dissociation-rate constant, and Rmax⁡R_{\max} is the response at complete occupation of accessible sites. Typical units are M−1 s−1\mathrm{M^{-1}\,s^{-1}} for kak_a and s−1\mathrm{s^{-1}} for kdk_d.

For constant CC, an initially unoccupied surface follows an exponential approach to equilibrium with observed rate constant kaC+kdk_aC+k_d. During dissociation into analyte-free buffer, the same model gives

R(t)=R(t0)exp⁡[−kd(t−t0)].R(t)=R(t_0)\exp[-k_d(t-t_0)].

The equilibrium dissociation constant is

KD=kdka,K_D=\frac{k_d}{k_a},

and the equilibrium response is

Req=Rmax⁡CKD+C.R_{\mathrm{eq}} = R_{\max}\frac{C}{K_D+C}.

These expressions connect chemical kinetics with chemical equilibrium. They apply only when the assumed reaction scheme and experimental conditions adequately describe the system.

Kinetic analysis commonly fits multiple concentration series simultaneously, sharing parameters such as rate constants across curves. Equilibrium analysis instead uses responses measured after sufficient equilibration. The two approaches provide complementary information: equilibrium measurements estimate affinity, whereas time-resolved analysis can distinguish interactions with similar affinity but different association and dissociation rates. A complex with slow dissociation may persist much longer than another complex with the same KDK_D, provided its association rate is correspondingly slower.

Biological interactions often violate the one-to-one model. Multivalent antibodies can engage several surface-bound partners; proteins can change conformation after binding; several populations of sites can coexist. The resulting effective stability may reflect avidity rather than the intrinsic affinity of one binding site. More elaborate models can represent these mechanisms, but additional adjustable parameters do not by themselves establish a mechanism. Independent concentration measurements, alternative immobilization orientations, competition experiments, and orthogonal methods help distinguish plausible interpretations.

Affinity values also depend on solution conditions. Temperature, salt composition, pH, and competing molecules can alter both kak_a and kdk_d. An SPR-derived KDK_D therefore characterizes a specified interaction under specified conditions, not an unchanging property of a protein pair.

Transport effects, artifacts, and measurement uncertainty

The concentration entering a flow cell need not equal the concentration immediately beside the sensor surface. Analyte reaches the interface through convection and diffusion, then is consumed locally by binding. If surface capture is sufficiently rapid, a depleted region develops near the sensor. This is called mass-transport limitation. Under these conditions, the observed association curve partly reports delivery to the surface rather than only the intrinsic chemical association rate.

Transport limitation is influenced by flow-cell geometry, flow rate, analyte diffusivity, and binding-site density. Increasing flow often improves delivery, but the improvement can be modest because transport through the near-surface boundary layer does not increase linearly with volumetric flow. A hydrated polymer matrix can introduce additional diffusion constraints. During dissociation, released molecules may bind again before escaping into bulk solution. This rebinding slows the apparent decay and can make the inferred dissociation rate smaller than the intrinsic value.

Bulk refractive-index changes constitute another major class of interference. Differences in salt, sugar, glycerol, or dimethyl sulfoxide content can shift the signal without changing surface occupancy. In small-molecule studies, solvent correction is often necessary because the optical effect of a solvent mismatch can exceed the binding response. Nonspecific adsorption, aggregation, gradual sample degradation, carryover, and injection dispersion can generate additional features that superficially resemble molecular kinetics.

The meaning of sensitivity must also be specified. Refractive-index sensitivity may be reported as angular or wavelength shift per refractive-index unit. Detection limits additionally depend on noise, drift, sampling time, reference subtraction, and the statistical criterion used. A high optical sensitivity does not automatically produce a low analyte-concentration detection limit, because that limit also depends on capture affinity, molecular mass, matrix interference, and assay duration. Comparisons between platforms require consistent definitions.

Reliable parameter estimation includes examination of residuals, reproducibility across preparations, and sensitivity to the chosen model. A small numerical fitting error may underestimate measurement uncertainty if analyte concentrations are inaccurate or the model is systematically wrong. Analyte activity is especially important: a nominal protein concentration can include molecules unable to bind. Conversely, an apparently excellent fit can arise because several correlated parameters compensate for one another. Instrumental calibration, biological controls, and model checking address different parts of the uncertainty budget and are not interchangeable.

Biochemical and analytical applications

A principal application of SPR is the characterization of molecular recognition. Antibody–antigen, receptor–ligand, protein–protein, protein–nucleic-acid, and host–guest interactions can all be investigated when one partner can be presented at a surface without unacceptable loss of function. Measurements can establish concentration-dependent binding, compare related molecules, test competition, and distinguish rapid from persistent interactions. The method is particularly useful when the time course itself carries information that a single endpoint assay would conceal.

In drug discovery, SPR is used to characterize compounds binding purified biological targets. Small-molecule measurements are demanding because, for equal numbers of occupied sites, their low molar mass produces much smaller responses than binding by a large protein. High-quality surfaces, solvent matching, and controls for nonspecific binding are therefore central. Fragment-based screening uses relatively small compounds as starting points for developing larger binders; SPR can contribute direct binding evidence, while structural methods such as X-ray crystallography or nuclear magnetic resonance provide different information.

Antibody characterization includes affinity measurements, competition studies, and epitope binning. In binning experiments, the ability of antibodies to bind an antigen sequentially or competitively helps group them by overlapping or sterically interfering recognition behavior. Such groups do not necessarily identify exact amino-acid contacts. Immobilization geometry, conformational changes, and large antibody dimensions can all produce competition without identical molecular epitopes. Studies of multivalent molecules similarly distinguish functional binding behavior from a simple one-site affinity.

Nucleic-acid applications include DNA hybridization, sequence-dependent recognition, and interactions of proteins with DNA or RNA. Membrane-related studies use supported lipid layers, vesicle capture, or other membrane-mimicking preparations to investigate binding to a cell membrane environment. Enzyme studies can detect substrate or inhibitor binding, although a binding signal is not itself a direct measurement of catalytic turnover. When a reaction changes surface mass or refractive index, an assay may also report a consequence of catalysis.

Environmental and food-analysis applications use recognition layers to detect contaminants, toxins, pathogens, or specific chemical species. Direct formats measure captured analyte; competitive formats can be useful for small molecules whose direct optical contribution is weak. In complex samples, sample preparation and matrix controls often determine performance more strongly than the idealized optical sensitivity. SPR supplies an analytical signal, while validated identification or quantitative reporting requires the corresponding assay-specific evidence.

Localized resonances and nanoparticle platforms

A distinction is made between propagating SPR on extended interfaces and localized surface plasmon resonance (LSPR) in metallic nanoparticles or nanostructures. In a small particle, conduction electrons move collectively within a finite object rather than supporting a wave that travels indefinitely along a flat boundary. The resonance depends strongly on particle size, shape, material, surrounding dielectric, and interactions with neighboring structures. Gold nanospheres, nanorods, silver particles, and lithographically fabricated arrays therefore exhibit different resonance spectra.

For a sphere much smaller than the optical wavelength in a uniform surrounding medium, the quasistatic polarizability is proportional to

α∝a3εm−εdεm+2εd,\alpha \propto a^3 \frac{\varepsilon_m-\varepsilon_d} {\varepsilon_m+2\varepsilon_d},

where aa is the sphere radius. A resonance occurs approximately when the real part of the denominator becomes small, often expressed as Re⁡(εm)≈−2εd\operatorname{Re}(\varepsilon_m)\approx-2\varepsilon_d when losses are modest. This simplified condition neglects retardation, radiation damping, and departures from spherical shape. Larger particles and complex structures require more complete electromagnetic calculations.

LSPR produces wavelength-dependent absorption and scattering, explaining the strong colors of many metal nanoparticle suspensions. A change in the nearby refractive index shifts the resonance, allowing surface binding to be detected spectrally. The sensing field is usually more tightly localized than the field of conventional propagating SPR, frequently extending only several to a few tens of nanometres from the particle surface. This can favor detection of material close to the particle while reducing sensitivity to more distant material. It can also make the thickness of linking layers especially important.

Interactions between nearby nanoparticles create additional effects. Narrow gaps can concentrate electric fields, while particle aggregation changes optical coupling and can produce visible color shifts. These effects underlie some nanoparticle-based assays, but their response may reflect aggregation geometry rather than simply the mass of captured molecules. Reproducible particle size, spacing, surface chemistry, and colloidal stability are therefore essential to quantitative interpretation.

Localized plasmons also support surface-enhanced Raman scattering, in which strong local fields enhance the weak Raman signals of nearby molecules. This is related to SPR through plasmonic field concentration, but it is a different measurement: conventional SPR typically follows an optical resonance shift, whereas Raman methods obtain vibrational information. Likewise, plasmon-assisted fluorescence depends on emitter–metal distance and competing enhancement and quenching processes, rather than on a single universal enhancement mechanism.

Imaging, hybrid methods, and physical limits

Surface plasmon resonance imaging extends the technique from individual sensing regions to spatially resolved measurements. A camera records reflected intensity over an area containing many spots or patterned structures. Arrays of immobilized molecules can then be monitored simultaneously, enabling multiplexed interaction studies and spatial analysis of adsorption. Many instruments operate at a fixed angle or wavelength on a steep portion of the resonance, although angle- and wavelength-resolved imaging approaches also exist.

Imaging introduces additional trade-offs. A propagating surface wave can spread optical information along the interface, so spatial resolution is affected by plasmon propagation as well as the imaging optics. Illumination uniformity, spot-to-spot surface preparation, and camera stability matter for quantitative comparisons. Cell-based SPR imaging measures changes in the region close to the substrate, including changes associated with adhesion and redistribution of cellular material. Its signal does not automatically identify a particular intracellular pathway or image the entire cell volume.

Hybrid measurements combine SPR with complementary information. Electrochemical SPR couples optical sensing to controlled electrode potential, enabling the study of adsorption, film growth, or electrochemically induced changes. Interpretation must account for potential-dependent optical properties and interfacial charge as well as molecular accumulation. Coupling to mass spectrometry can add chemical identification of captured species, whereas a quartz crystal microbalance responds to mechanically coupled interfacial material, including associated liquid. Differences between the two signals can help reveal hydration and viscoelastic behavior.

Long-range surface plasmon modes use thin metal films between approximately symmetric dielectric environments to reduce attenuation and narrow resonances. Waveguide-coupled structures, optical fibers, and integrated photonic devices seek smaller instruments or alternative sensing geometries. Their performance depends on the complete optical architecture: a narrower resonance is useful only if it can be measured stably, and tighter confinement can increase losses or reduce the volume sampled.

At nanometre and subnanometre separations, classical local-permittivity models can become inadequate. Nonlocal electronic response, electron spill-out, and quantum tunneling across narrow gaps can alter predicted resonances and limit the field enhancement expected from purely classical models. In routine biosensing, more immediate limits usually arise from surface heterogeneity, transport, drift, and sample complexity. These different regimes require different models: a detailed quantum description of a nanogap does not resolve an incorrectly specified binding reaction, just as a precise kinetic fit does not establish the optical properties of an unfamiliar nanostructure.