SPR biosensors: simultaneously removing thermal and bulk-composition effects1

https://doi.org/10.1016/S0956-5663(98)00121-3Get rights and content

Abstract

Surface plasmon resonance (SPR) is an established method for sensing analytes by monitoring changes in the plasmon dispersion relation (PDR) at the interface of a thin metal film and a fluid. When SPR is used in sensors for specific analytes, the changes in the plasmon dispersion relation of interest are generated by the binding of analytes to receptors immobilized to the metal film. However, changes in the PDR can also be generated by changes in the index of refraction of the bulk solution containing the analytes via changes in composition or temperature. Thus, there exist inherent systematic errors in SPR based chemical sensing when temperature and/or concentration conditions are not carefully controlled. We have demonstrated the efficacy of a single, simple, and inexpensive method for simultaneously discriminating both effects from those of binding and/or debinding of analytes with a two-element SPR sensor array. Although two-element SPR arrays have been used before, their ability to simultaneously discriminate out both thermal and bulk-composition (TAB) effects (in both SPR spectroscopic and spectrophotometric schemes) has not been previously addressed. We show an example of a relatively inexpensive SPR biosensor instrument using a compensator element and its stability over a period of days. This demonstration has implications for the development of reliable SPR based chemical sensors for environmental and remote sensing applications.

Introduction

In the last 12 years, surface plasmon resonance (SPR) has been extensively developed into a useful technique in the fields of chemical and biosensors, with more than 100 publications describing applications for real-time specific interaction analysis (Liedberg et al., 1995). SPR is a phenomenon which involves variable absorption of light by the surface electron plasma of a metal film under specific resonance conditions governed by the plasmon dispersion relation (PDR). These resonance conditions vary with the index of refraction of the media in contact with the metal film and allow one to detect changes in the index of refraction to the sixth decimal place (Kunz et al., 1996). The resonance conditions can also be changed by adding a thin dielectric film (e.g. adsorbed molecules, monolayers, or multilayers), with an index of refraction different to that of the bulk solution, to the back of the metal film. SPR can be used as a relatively simple transduction mechanism in chemical or biosensors with excellent sensitivity to the adsorption of analytes (down to ∼500 pg/cm2 of adsorbed analyte) (Häussling et al., 1991, Sigal et al., 1997, Schwarz et al., 1991, Peterlinz and Georgiadis, 1996).

If a receptor is available for a certain analyte and it can be readily immobilized to the metal film, it will require no additional labeling if SPR is used as the (reagentless) transduction mechanism. This is because SPR is sensitive to changes in index of refraction, and therefore is sensitive to changes in the composition of layers on the back of the gold, allowing (for example) schemes immobilizing receptors in polymeric matrices (Liedberg et al., 1995, Löfas and Johnsson, 1990, Kooyman et al., 1990) or on self-assembled monolayers (SAMs) (Häussling et al., 1991, Sigal et al., 1997). This sensitivity allows one to study real-time binding kinetics of target analytes since the molecular surface density (and hence the local index of refraction) of a layer changes continuously as analytes bind to immobilized receptors (Stenberg et al., 1991, Malmqvist, 1993). This has made SPR the tool of choice for a number of real-time studies of reaction kinetics (Kunz et al., 1996, Häussling et al., 1991, Sigal et al., 1997, Peterlinz and Georgiadis, 1996, Stenberg et al., 1991, Malmqvist, 1993, DeBono et al., 1996, Fågerstam et al., 1992, Siirila and Bohn, 1991, Altschuh et al., 1992, Morgan et al., 1992, Rahn and Hallock, 1995, Spinke et al., 1993). Furthermore, the kinetic monitoring capability of SPR allows one to determine the equilibrium constants of equilibrium reactions (Kooyman et al., 1988).

SPR occurs when the surface-parallel component of a light propagation vector, klight, approaches the magnitude of the surface plasmon propagation vector, kspw (directed along the interface of a metal film and the fluid in contact with it) for a given optical frequency, ω. The magnitude of kspw for a given ω is governed by the plasmon dispersion relation (PDR). The PDR is a physical relationship which can be directly measured (Swalen et al., 1980). The magnitude of klight can be made larger than the magnitude of kspw by attaching the metal film to a prism (illuminating the metal film through the prism). The surface parallel component of klight can then be made to match kspw by varying the angle of incidence of the light upon the metal film (this is known as the Kretschmann–Raether coupling scheme) (Raether, 1988). The angle of incidence that matches the dispersion relationships is characteristically past the angle of total internal reflection for the prism and the bulk media in contact with the metal film. Thus, Kretschmann–Raether type SPR-based sensors are a subclass of attenuated total reflection (ATR) sensors, the characteristics of which have been documented elsewhere (Sprokel and Swalen, 1991).

In this scheme, the magnitude of kSPW (a function of the PDR and the frequency of the excitation light) is typically measured by monitoring the reflectance profiles, or the reflectance of the metal film as a function of angle of incidence of a monochromatic light ray. Each profile has a characteristic dip with a minimum located at θSPR (the angle where the surface parallel component of klight exactly matches kSPW) (Raether, 1988). This corresponds to a maximum coupling of light into surface plasmons at the interface of the metal film and the fluid in contact with it.

The PDR (and thus θSPW) can be affected by adsorbing molecules to the back of the metal film (Raether, 1988). The PDR can also be changed by thermal and bulk-composition (TAB) effects. These include variations in the total solute concentration or the temperature of the solution in contact with the metal film, since such variations will change the index of refraction of the solution (Sigal et al., 1997, Bass, 1995).

Changes in the PDR are typically monitored over time in two ways:

  • 1.

    minima tracking method (SPR spectroscopy) wherein a device collects several reflectance profiles over time, locating and recording the angles of minimum reflectance (θSPR) (Rahn and Hallock, 1995) and

  • 2.

    fixed angle method (SPR spectrophotometry) wherein the reflectance at one particular angle on the sloping sides of the reflectance profile is monitored over time (as the PDR changes, the reflectance increases or decreases) (Mayo and Hallock, 1989).

Additional background literature dealing with SPR and its application to chemical sensing is widely available (Barker, 1973, Raether, 1988, Salamon et al., 1997, Sambles et al., 1991).

Thermal and bulk-composition effects can generate significant changes in the PDR, which may obscure those caused by specific binding and/or debinding of the subject analytes. This places serious limitations on the situations where SPR can be effectively used as the transduction mechanism for chemical biosensors. Hence, simple, reliable, and effective techniques to compensate for these phenomena are of crucial importance in creating high sensitivity SPR transduction-based instruments.

It has been shown with the commercially available BIACORE SPR sensor (Pharmacia Inc.) that bulk-composition effects can be subtracted out from short term binding measurements by comparing the changes in the angle of minimum reflectance obtained during the binding experiment to those obtained when the same solution is exposed to an inert (i.e. one to which analytes or other solutes do not adsorb) surface on a gold film (Sigal et al., 1997, Shen et al., 1996, Persson et al., 1997). It has also been shown that the changes in the PDR caused by temperature changes can be subtracted out using a high-quality temperature sensor to remove thermal effects over a long period of time when the consistency of the solution is kept constant and no binding or debinding occurs (Melendez et al., 1996, Melendez et al., 1997). It would seem, then, that both of these schemes would be needed for a highly sensitive chemical sensor immune to TAB effects and capable of long term measurements in environments where temperature and concentration conditions are not controlled. However, as we demonstrate here, a separate temperature transduction mechanism is totally unnecessary since it has been established that the changes in the PDR due to temperature changes allow SPR to be used as an accurate and precise temperature transducer (Chadwick and Gal, 1993). This simplifies the design of new TAB compensating SPR biosensors significantly.

In this report we describe the fabrication and use of a two-element SPR sensor configuration to illustrate the simultaneous removal of both thermal and bulk-composition effects from SPR-based biosensors. The first element, the active element, will measure association and/or dissociation of analytes as well as the undesired TAB effects. The second element, the reference element, is treated with a coating that resists adsorption of the analytes being studied (and other solutes as well) and thereby only measures TAB effects. The TAB effects can then be subtracted away to leave only data relevant to association and dissociation processes, provided that three requirements are met:

  • 1.

    the two elements should be spatially close to each other and exposed to the same liquid to minimize thermal differentials between them;

  • 2.

    the reference element must be relatively inert (compared to the active element) to its environment and resistant to non-specific adsorption of solutes; and

  • 3.

    the SPR signals of the two elements must be empirically related in a well-defined manner when exposed to TAB effects in the absence of binding or dissociation.

Patterning the sensing and reference elements next to each other can meet the proximity requirement of the two elements. For biosensor applications, the reference element can be rendered inert by coating it with an oligo(ethylene glycol)-terminated self-assembled monolayer (SAM). Such SAMs have been shown to resist the non-specific adsorption of proteins (Pale-Grosdemange et al., 1991, Prime and Whitesides, 1993, Lopez et al., 1993). There is strong evidence that these SAMs can also resist the adsorption of cellular materials and a wide range of soluble components (e.g. lipids, polysaccharides) in water (Lopez et al., 1993, Ista et al., 1996), making them attractive for biosensing uses in physiological and environmental applications. The relationship between the two signals is addressed by empirical calibration of the two elements to TAB effects (see below).

Kooyman et al. (1991)proposed the use of a two-element SPR sensor configuration for the removal of thermal or bulk-composition (not both simultaneously) effects in a paper which used a reference element to compensate for degradation of the silver film in an SPR experiment. They did not, however, demonstrate the use of a reference element for such applications, nor did they acknowledge the crucial operational requirements and details that we describe here.

The BIACORE commercial instrument is a multichannel SPR spectroscopic array sensor (Jönsson et al., 1991, Sigal et al., 1997, Altschuh et al., 1992, Fågerstam et al., 1992, Malmqvist, 1993, Shen et al., 1996, Persson et al., 1997). One element of the array has been used to subtract out bulk-composition effects. However, the BIACORE has an advanced microfluidic system and temperature can be held constant (Shen et al., 1996). So while the BIACORE instrument is well capable of using a reference element to screen out both temperature and bulk-composition effects, this has not been demonstrated to our knowledge. Furthermore, some bulk-composition compensator users (for example, Sigal et al., 1997, Shen et al., 1996, Persson et al., 1997) have made the assumption that bulk-composition effects are identical in magnitude across all elements. We show here that this is not necessarily so.

Furthermore, we illustrate how a TAB compensator can allow a relatively inexpensive, custom-built device to obtain very precise long term results. This device and other new SPR sensor platforms do not require temperature stabilization elements or thermocouple-based compensators. We also show that such new instruments can use SPR spectroscopy or the simpler SPR spectrophotometry.

Section snippets

Sensor design

The experimental setup used is shown in Fig. 1. An intensity-stabilized, polarized, red He–Ne laser (wavelength 632.8 nm, Melles-Griot Inc.) is projected through a beam splitting periscope, consisting of a 50/50 non-polarizing cube beam-splitter and a metallic mirror. The two parallel beams exit with a separation of about 1 cm in a plane that is normal to the optics table. The beams then pass through different regions of a dual-frequency chopper (Stanford Research Inc. model SR540) so that they

Demonstration of TAB compensation

In Fig. 2 we show a typical progression of the θSPR over a long period of time (as recorded with the sensor's minima-tracking mode) when the sample was in contact with pure DI water. The periodic fluctuations on both signals are due to temperature fluctuations in the laboratory (∼5–10°C) corresponding to the rising and falling of the sun throughout the day. Since the fluctuations in the active element's signal are paralleled in that of the reference element, the TAB effects can be subtracted

Conclusions

We have demonstrated how one inert reference element can be easily fabricated and used to simultaneously compensate for both thermal and bulk-composition effects that generate systematic errors in the measurement of binding/debinding of analytes to receptors over time with SPR techniques. Provided that a suitable inert reference element coating is available, the TAB compensator would allow for real-time in situ applications of SPR outside the laboratory, and gives researchers an alternative to

Acknowledgements

This work was funded by ONR Multidisciplinary University Research Initiative Grant N00014-95-1-1315, by ONR Grant N00014-95-1-0901, by ONR Defense University Research Instrumentation Program Grant N00014-95-1-0255, and by NSF Grant HRD-9450475.

References (40)

  • T. Schwarz et al.

    Detection of nucleic acid hybridization using surface plasmon resonance

    Tibtech

    (1991)
  • E. Stenberg et al.

    Quantitative determination of surface concentration of protein with surface plasmon resonance using radiolabeled proteins

    Journal of Colloid and Interface Science

    (1991)
  • D. Altschuh et al.

    Determination of kinetic constants for the interaction between a monoclonal-antibody and peptides using surface plasmon resonance

    Biochemistry

    (1992)
  • A.S. Barker

    Optical measurements of surface plasmons in gold

    Physical Review B

    (1973)
  • Bass, M., 1995. Handbook of Optics, Vol. 1. McGraw-Hill, New York, 4321...
  • B. Chadwick et al.

    An optical–temperature sensor using surface plasmons

    Japanese Journal of Applied Physics

    (1993)
  • R.F. DeBono et al.

    Self-assembly of short and long-chain n-alkyl thiols onto gold surfaces: a real-time study using surface plasmon resonance techniques

    Canadian Journal of Chemistry

    (1996)
  • L. Häussling et al.

    Biotin-functionalized self-assembled monolayers on gold: surface plasmon optical studies of specific recognition reactions

    Langmuir

    (1991)
  • L.K. Ista et al.

    Attachment of bacteria to model solid-surfaces: oligo(ethelyne glycol) surfaces inhibit bacterial attachment

    FEMS Microb. Lett.

    (1996)
  • U. Jönsson et al.

    Real-time biospecific interaction analysis using surface-plasmon resonance and a sensor chip technology

    Biotechniques

    (1991)
  • Cited by (36)

    • Photothermal effects induced by surface plasmon resonance at graphene/gold nanointerfaces: A multiscale modeling study

      2019, Biosensors and Bioelectronics
      Citation Excerpt :

      Therefore, the much lower power light sources (∼mW/mm2) were recommended to be used in order to reduce the photothermal effects (Galvez et al., 2017). Temperature correction methods were also proposed, such as temperature stabilization, compensation, and multi-channel design methods, to attenuate the influences of the photothermal phenomenon (Naimushin et al., 2003a; O'Brien et al., 1999; Taylor et al., 2006; Zhao et al., 2015). Variation in optical properties by temperature related changes are expected to influence Plasmon resonance curves, signifying that the sensitivity and detection accuracy of SPR platforms require a detailed relook into estimation of their sensor performance.

    • Self-referencing fiber optic particle plasmon resonance sensing system for real-time biological monitoring

      2016, Talanta
      Citation Excerpt :

      Thus, it is necessary to accurately separate or compensate the systematic errors in order to produce meaningful data. Such systematic errors are typically corrected through the use of a reference sensor that is monitored in parallel with the analysis sensor [4,6,7]. Recently, biosensors based on particle plasmon resonance (PPR), also known as localized surface plasmon resonance (LSPR), have received more and more attention because of the high sensitivity of the PPR spectrum of noble metal nanoparticles to adsorbate-induced changes in the local refractive index (RI) of the surrounding environment [3,8–10].

    • Self-referencing SPR-sensor based on integral measurements of light intensity reflected by arbitrarily distributed sensing and referencing spots

      2015, Sensors and Actuators, B: Chemical
      Citation Excerpt :

      A common approach is based on introduction of a dedicated referencing channel, spatially separated from the sensing one and coated by a material with a decreased affinity to the analyte of interest. Subtracting the signal of the referencing channel from the signal of the sensing channel, possible changes of the volume refractive index [7,8] can be compensated. The overall efficacy of such spatial referencing is increased by scaling down mean distance between the sensing and referencing spots into microscopic range while maintaining macroscopic value of total area of the sensing and referencing spots; this approach named as distributed referencing was suggested in [9].

    • Mimicking nature's noses: From receptor deorphaning to olfactory biosensing

      2011, Progress in Neurobiology
      Citation Excerpt :

      Resonance changes measured using current Biacore models, such as the T100, have been used to monitor compounds with molecular weights between 157 and 341 g mol−1 (Papalia et al., 2006), and fragments of drug compounds down to 100 g mol−1 (de Kloe et al., 2010). However, when monitoring interactions of such small molecules using SPR, factors such as temperature have been shown to contribute to inaccuracies in the measurement (Moreira et al., 2008; O’Brien et al., 1999; Xiao et al., 2010). It is factors such as these which need to be overcome in the development of a truly portable biosensor device using SPR for low molecular weight molecules such as odourants.

    View all citing articles on Scopus
    1

    This paper was presented at the Fifth World Congress on Biosensors, Berlin, Germany, 3–5 June 1998.

    View full text