Analytical and Bioanalytical Chemistry (2012), 403(6), 1477-1484.

TRENDS
Nanostructured substrates for portable and miniature SPR biosensors
Julien Breault-Turcot &Jean-Francois Masson
Received:16January 2012/Revised:21February 2012/Accepted:21March 2012/Published online:24April 2012#Springer-Verlag 2012
Abstract Surface plasmon resonance (SPR)biosensing has matured into a valuable analytical technique for measurements related to biomolecules,environmental contaminants,and the food industry.Contemporary SPR instruments are mainly suit-able for laboratory-based measurements.However,several point-of-measurement applications would benefit from simple,small,portable and inexpensive sensors to assess the health condition of a patient,potential environmental contamination,or food safety issues.This Trend article explores nanostruc-tured substrates for improving the sensitivity of classical SPR instruments and nanoparticle (NP)-based colorimetric sub-strates that may provide a solution to the development of point-of-measurement SPR techniques.Novel nanomaterials and methodology capable of enhancing the sensitivity of clas-sical SPR sensors are destined to improve the limits of detec-tion of miniature SPR instruments to the level required for most applications.In a different approach,paper or substrate-based SPR assays based on NPs,are a highly promising topic of research that may facilitate the widespread use of a novel class of miniature and portable SPR instruments.
Keywords Kretschmann SPR .Au nanoparticles .Portable SPR .Grating substrates .Enhanced sensitivity .Paper-based SPR sensors
State-of-the-art miniature SPR
Detection of low levels of environmental pollutants,warfare chemicals that may compromise security,pathogens in food,
and biomarkers for disease diagnosis are contemporary chal-lenges of analytical science [1].These challenges explain the growing interest in instrumental devices that can provide an analysis that is:1.highly sensitive and specific;2.simple and affordable;3.rapid;
4.at the point-of-measurement (portable device);and
5.
label-free.
Among the different analytical techniques currently available,surface plasmon resonance (SPR)sensors have been identified as a potential solution to these needs [2].SPR sensors rely on the interaction of an electromagnetic field with free electrons at the surface of a thin metallic film or a nanoparticle (NP)(usually gold or silver).The electro-magnetic field of light with a wavevector matching that of the surface plasmon polariton (SPP)leads to an absorption band in the reflectance or extinction spectrum,for thin metallic films or NPs,respectively.The position of the plasmonic band in terms of angle of incidence or wave-length is dependent on the refractive index (RI)of a dielec-tric medium near the metallic bel-free monitoring of binding events with SPR sensors relies on tracking changes of the plasmonic band caused by RI changes near the me-tallic film.SPR sensors typically feature binding of a bio-molecule to a bioreceptor located on the metallic film,leading to detection limits in the picomolar to nanomolar range.The combination of SPR sensors with immunodetec-tion techniques has been extensively studied and is now commonly used in biosensing [3].
Although the vast majority of SPR sensing is performed on laboratory-based instruments,miniature and portable SPR devices have recently been developed for on-site bio-detection measurements by SPR panies such
J.Breault-Turcot :J.-F.Masson (*)
Departement de chimie,Universite de Montreal,C.P.6128Succ.Centre-Ville,Montreal,Qc H3C 3J7,Canada e-mail:jf.masson@umontreal.ca
Anal Bioanal Chem (2012)403:1477–1484DOI 10.1007/s00216-012-5963-1
as Texas Instrument(Spreeta)[4,5],K-MAC(Biochip Analyzer,SPRmicro),Seattle Sensor Systems(SPIRIT) [6],and Möbius advanced technology(handheld SPR)cur-rently market portable instruments.The RI resolution of portable instruments is of the order of~2×10−6refractive index units(RIU),in comparison with~2×10−7RIU[6–8] for a high-performance laboratory SPR instrument.The RI resolution refers to the smallest change in RI that it is possible to monitor.Higher detection limits in bioassays is a consequence of the worse resolution of portable SPR instruments.Thus,increasing the sensitivity of portable instruments is a contemporary challenge to match the per-formance of laboratory-based instrumentation.
The size of portable SPR instruments is significantly smaller than that of tabletop and lab-based instruments. Currently,commercial and portable SPR devices are approx-imately the size of a lunch box.These portable instruments are integrated with the optical and electronic components needed for a SPR sensing device:a light source,optics,a fluidic cell,a detector,and electronics for data acquisition and processing.
A peristaltic pump is commonly added to deliver the sample to the biosensor chip,although direct injection from a syringe can also be used.Most of these commercial portable SPR devices have a multichannel fluidic cell in which one channel can be used as a reference to reduce interference from non-specific binding and bulk refractive index changes.The mul-tichannel fluidic cell is suited to analysis of several targets in a single injection of the sample.To reduce the effect of temper-ature fluctuation on the analysis,most of these systems are equipped with a temperature-regulated system.These features integrated to the portable and miniature SPR systems facilitate measurements and minimize the input of the user.However, these features must be compatible with the low power con-sumption required for battery-operated systems.
In the last decade,portable SPR devices have been used for detection of analytes ranging from small to large molecules or bacteria[6].Because of the small change in effective refrac-tive index of small molecules,a competitive immunoassay is usually performed to enhance the SPR response and detect these molecules(Fig.1A).Competition for a ligand adsorbed on a solid substrate is established between the analyte in the sample and a bioreceptor in solution.Without establishing competition with a tagged molecule,the signal generated by the analyte solely is not sufficient for quantitative measure-ments.Successful examples of competitive assays for small molecules include the analysis of pesticides(2-hydroxybi-phenyl)[9],antibiotics(fluoroquinone and sulfamethoxazole) [7,10],disease biomarkers(cortisol)[5],and toxic com-pounds(benzo[a]pyrene and2,4-dichlorophenoxyacetic acid) [9,11].
Although small molecules are still somewhat challenging to detect,larger molecules can be detected directly by ap-plying the solution for analysis to a surface functionalized with a molecular receptor(Fig.1B),for example single-strand DNA,antibody,protein.Proteins indicative of dis-eases(prostate-specific antigen and troponin I)[12,13]and biological warfare agents(Staphylococcus aureus enterotox-in)[4,14]have been detected with small and portable SPR instruments.Low nanomolar or high picomolar detection limits have been reported for most of these small molecules or biomolecules(Fig.2).Although these examples highlight the possibilities with portable and miniature SPR instru-ments,better sensitivity is needed to further improve detec-tion limits.The advent in recent years of novel materials for SPR sensing chips with increased sensitivity should foster the development of several applications with better detec-tion limits than currently possible.
Enhancing sensitivity with nanostructured substrates The development of sensing templates with high sensitivity and low limits of detection similar to laboratory-based instru-ments is a challenge to overcome for the miniaturization or the development of portable SPR instrument.The performance of different SPR sensing templates is often compared with the RI resolution in RIU and the sensitivity in“Units of measure-ment”/RIU that is achieved by a particular variant of SPR instrumentation.The resolution is generally obtained by cal-culating the noise of the output measurement(intensity,angle, wavelength,or phase—represented byσnoise)divided by the sensitivity of the sensor in units of output measurement per refractive index unit(δY/δη).Thus,the refractive index reso-lution(σRI)of the SPR signal is a function of the properties of the optical noise of the SPR response(instrument-related)and the plasmonic properties of the metallic substrate(related to the SPR sensing chip).
σRI¼σnoise d Y d)
=
ðÞ
=ð1ÞIt has been demonstrated that contemporary SPR instru-mentation,including portable SPR instruments,is reaching the theoretical limits imposed by optical limitations of the measurement system[2].While these systems all rely on a thin gold film as the plasmonic material,novel materials of greater sensitivity are currently being developed.In addi-tion,improved noise reduction algorithms for data treatment are necessary to reduce the noise level of the SPR sensor-grams;this will also contribute to improving the limits of detection of SPR sensors.
Sensitivity is a function of the plasmonic material(metal composition and structure)and a few experimental condi-tions(wavelength,incident angle,and refractive index mea-sured).Of these variables,modifying the structure of the plasmonic material is most likely to improve SPR sensing. Nanostructuring the surface of a gold film can lead to the
1478J.Breault-Turcot,J.Masson
excitation of the grating modes of SPR,waveguide modes,long-range surface plasmons,and the co-excitation of local-ized and propagating surface plasmons [15].These plas-monic materials lead to enhanced sensitivity or lower detection limits by improving the signal-to-noise ratio of the measurement and will be the focus of this Trend article.A recent review provides a detailed account of other meth-ods for enhancing sensitivity [15].
Several of these plasmonic materials exploit grating or grating-like structures to enhance the performance of SPR sensing.The anomalous low intensity in the diffraction spectrum on a metallic grating of light at a specific wave-length was the first literature account,reported by Wood more than a century ago,of the observation of a SP on a grating [16].Later,it was explained that the component of the diffracted light,whose wavevector corresponds to that of the surface plasmon polariton (SPP),is effective in exciting the surface plasmon (SP)on a metallic grating [17].Light can be coupled into grating modes of a SPR substrate con-sisting of an Au grating on an Au film or a corrugated Au film (Fig.3A )by either direct illumination or by use of attenuated total reflection in prism-coupling geometry.These two configurations are compatible with small,porta-ble,and low-cost SPR instrumentation.
A first example exploits direct illumination of the SP modes supported on a corrugated Au grating [8].The sen-sitivity was determined at 620nm/RIU,lower than for thin Au films,but expected for direct illumination of gratings.A disadvantage of this configuration is the size of this SPR configuration,which is limited by the focal length needed to achieve the spectral resolution required for the analysis.Higher spectral resolution is needed to improve the accuracy of the position of the plasmonic band,which dictates the noise level of the measurement.The version developed by Piliarik et al.has an approximately 15cm×15cm footprint with an RI limit of detection of 3×10−7RIU [8].This limit of detection (LOD)rivals that of a commercial instrument because of a sharp SPR band and the associated excellent noise level (0.0002nm)on the sensorgram.The excellent RI resolution,the limited number of optical parts and the pos-sibility of multiplexing to several channels are advantages of this configuration.
Although direct illumination of a grating may have lim-itations,because of the length of the optical path required for high resolution,classical Kretschmann configuration of SPR is also suited to excitation of SP on metallic gratings supported on a thin Au film and may be more suitable for the design of small instruments.In addition,this
excitation
Fig.1Detection strategies for small molecules or large
biomolecules in SPR sensing.(A )A competitive assay features competition between an analyte labelled with a large molecule,for example an antibody,with the unlabeled analyte to be quantified.(B )The antibody immobilized on the surface captures the analyte in direct immunoassay
3
2
1
P e
s t i c i
d e
A n
t i b i o t i D i s e a s B
i o m a r k e T o x i c c o
m p o u n d Small molecule D N
A P
r o t e i n Biomolecule
L o g L O D (p M )
Fig.2Limits of detection (LOD)of different compounds analyzed by use of portable SPR devices
Substrates for portable and miniature SPR biosensors 1479
mode is advantageous because of indirect illumination of the Au film,which is of interest for analysis of highly scattering media,for example biological fluids.The SPP of the grating network and the thin Au film can interfere and lead to higher intensity for resonant conditions [17].Perturbation of the propagating SP by the grating is at the origin of the enhanced sensitivity observed with these struc-tures.A theoretical investigation of the SPR response with these structures predicted enhancement of the sensitivity by a factor of 6to 7compared with thin-film SPR sensing [18].The enhancement observed for nanograting plasmonic sub-strates was reported with the angle-scanning configuration of SPR.It is also important to note that the enhancement occurs in a specific refractive index range which depends on the features of the grating.Sites of higher enhancement exist on the surface of metallic gratings.For example,selective immobilization of antibodies on the mesa of the nano-grating further enhanced the detection of anti-TNF-α[19].Accurate positioning of the bioreceptor in the regions of highest electromagnetic field is a current challenge to fully achieve the potential of nanostructured surfaces.
Wavelength scanning SPR in the Kretschmann configura-tion has been simulated on grooved nano-gratings.A narrow (circa 10nm)100nm-deep groove grating (Fig.3B )in gold or silver resulted in sensitivity of the order of 500nm/RIU [20],comparable with direct excitation of grating modes.Although this sensitivity is relatively low,the differential reflectance of these films was improved with the grooved nanograting,which would be the advantage of this nanostructured sub-strate.Another example of nanostructured surfaces exploits the optical diffraction of arrays of nanowires (Fig.3C )sepa-rated by micrometer distances that can be interrogated in total internal reflection —phase imaging [21]or transmission dif-fraction measurements [22].These nanowires support a local-ized surface plasmon resonance (LSPR)mode that could be exploited in a biosensing assay [22].The extremely sharp bands of the diffraction orders should provide low noise levels and good detection limits.As demonstrated above,nanograt-ings are an interesting substrate to improve SPR sensors.However,long optical paths are usually required to improve the resolution of diffraction modes and achieve low detection limits.For SPR sensors to benefit from the diffraction properties of the grating in a small and portable instrument will require careful optical design.
Microhole arrays (Fig.3D )are structures consisting of hexagonally patterned holes embedded in an Au film with a hole diameter of the order of 500nm to 2μm and period-icity of the order of a few microns [23].Microhole arrays share properties of gratings,as they can also be viewed as thin stripes of Au repeating at a rotational angle of 60°.Refractive index sensitivity was enhanced by at least a factor of 2in comparison with thin Au film with microhole arrays.Detection of IgG was also enhanced by approximately the same factor [23].The excitation of the SP on microhole arrays with a Dove prism SPR [24]led to a refractive index resolu-tion of 1.5×10−6RIU,equivalent to thin film SPR sensors;the comparison was performed on the same instrumentation [23].The slightly worse noise level observed for microhole arrays is explained by the asymmetry of the peak for this plasmonic substrate,which is very similar to the peak shape simulated for nanograting,as predicted by Alleyne et al.[18].It is likely that the scattering of the SP is more important at longer wave-length,which causes the asymmetry of the peak.Thus,the peak is broader on the low-energy branch.Improving the data-analysis algorithm to improve detection limits will be one challenge of this methodology.
Conventional data-analysis methods assume a sym-metrical peak shape and may fail to accurately predict the SPR wavelength of asymmetrical peaks,for example those observed with nanostructured plasmonic sub-strates.Chemometric data analysis,for example principle-component analysis,must be validated to capture changes in the spectra and contribute to denoising the data.For example,a pre-processing step involving singular value decomposition and reconstitution of the data with the very first components reduced the noise level of SPR data by one order of magnitude [24].Harnessing the full power of chemometric data analysis will be essential to further reduce the noise in SPR data and to improve detection limits of SPR sensing.An alternative data-analysis method may involve phase imaging of the SPR wavelength [21].The phase change observed near the reso-nant conditions of the SPR signal is striking,and led to precise determination of the phase change and to 100-fold improve-ment in detection limit [21].Better data-analysis tools
in
Fig.3Different grating structures for SPR sensing.(A )Corrugated Au film or nanograting on an Au film;this structure is excited by prism-coupling and direct illumination.(B )Grooved nanograting structure for excitation in the prism-coupling configuration.(C )Nano-wire gratings are suitable for direct illumination or prism-coupling.(D )Microhole arrays are excited in the prism-coupling configuration
1480J.Breault-Turcot,J.Masson
conjunction with novel plasmonic substrates are poised to improve SPR sensing.
For small SPR instruments,integration of a simple sample-preparation step should possible by use of micro-fluidics,but it would increase the complexity of the setup and lengthen analysis time.Extensive and complex sample preparation may not be possible with portable SPR systems, because of the lack of available infrastructure,time issues, and the absence of qualified personnel to run the analysis.It would thus be beneficial to detect biomolecules directly in crude biofluids.Thus,to enable analysis of analytes in crude biofluids(for example serum,blood,urine,or saliva),use of low biofouling surface chemistry is judicious to minimize the background response of the molecules present in the biofluid.Secondary signal amplification is another option to further improve the limits of detection of portable SPR assays.This secondary amplification scheme must be inte-grated in a manner to minimize the complexity of the instrument,to maintain the advantages of portable SPR systems.Amplification of the SPR signal and novel approaches for low biofouling surface chemistry were the object of a recent review[25].
(Colorimetric)paper and test-strip-based SPR sensors Although miniature and portable SPR instruments are not extremely complex,a dedicated instrument is still required to perform measurements on the SPR sensing chip.In some instances,for example remote and underdeveloped locations without access to modern laboratories or even electricity,it would be desirable to develop a simple SPR system of simplicity similar to that of the glucometer.One possibility involves use of a colorimetric test-strip detection scheme based on the SPR phenomenon.A paradigm change in analytical sciences would result from reading an assay with the naked eye or by simple picture processing with an application on smart phones or a small battery-operated instrument.This approach has been re-popularized as paper-based diagnostics for traditional colorimetric organic reagents[26].
The optical phenomenon of localized surface plasmons (LSP),which are excited by direct illumination of light,has attracted much attention in the past decade[27].The bright coloration of LSP nanostructures and the possibility of designing bioassays leading to color changes leads to col-orimetric sensors based on plasmonic nanostructures sup-porting LSP.Indeed,Au NP can be functionalized with bioreceptors,facilitating the design of bioassays based on affinity interactions between a bioreceptor and a ligand to analyze.These characteristics of Au NPs are strong indica-tors of the viability of this approach for the design of small and simple SPR-based sensors.
Two different detection schemes can be envisioned with plasmonic NPs.The first strategy uses the strong absorptivity of the Au NPs to design a bright colorimetric probe.The second scheme exploits the color changes induced by capture of the analyte on the Au NP or the aggregation triggered by the biorecognition event occurring on the Au NP.These colori-metric changes are quantitative with regard to the concentra-tion of the analyte.The readout is simply the absorbance or wavelength shift induced by the biorecognition event.Other surface-enhanced spectroscopy can be exploited with NP-based bioassays,for example surface-enhanced Raman scat-tering(SERS)[28]or surface plasmon coupled fluorescence [29].The relative complexity of the contemporary instrumen-tation associated with these techniques is currently prohibitive for paper or test-strip-based detection with miniature and portable instruments.However,these techniques will definite-ly affect analytical science in the near future.This section will focus on the implementation of colorimetric tests based on the colorimetric properties of Au NPs in paper or test-strip-based detection schemes.
Biosensors have been designed on the basis of the color-imetric or plasmonic properties of Au NPs immobilized on a anic substrates,for example paper[30],plastic [31],or nitro-cellulose[32],are suitable for immobilization of Au NPs by deposition of a sol–gel loaded with reactants on paper,hydrophobic interactions of BSA with nitrocellu-lose(analogous to blotting techniques for proteins),and polyethyleneterephthalate(PET)-activated surfaces with Au NP chemically immobilized on the PET substrate. Advantages of organic substrates include flexibility,robust-ness,low cost,and lightness[31].Another approach uses glass substrates,which serve to deposit Au NP via grafting of an antigen to the glass substrate(Fig.4).Au NP labelled with secondary antibodies or streptavidin induce the color change in the presence of the analyte at the surface of the glass slide[33,34].Well-established surface chemistry and adequate transparency for transmission measurements are advantages of glass.The increase in scattering of Au NP and wavelength shift from the absorption of Au NP on glass can lead to a functional biosensor[35].Use of a flatbed scanner as detection method has also been demonstrated for reading bioassays on glass[33,34].In one example,BSA was detected at1ng mL−1with the paper-based biosensor, which is close to the detection limit of ELISA assays re-cently reported in the literature(0.38ng mL−1)[36]or with commercial ELISA kits at0.25ng mL−1[37].IgG was also detected with a colorimetric immunoassay with a detection limit of the order of1ng mL−1(low pmol L−1LOD),which is similar to other techniques reporting limits of detection in the low picomolar regime[38].Contemporary techniques still mostly rely on detection with a spectrophotometer,but these examples validate the possibility of using the“naked-eye”or a simple tool to detect color changes in these assays.
Substrates for portable and miniature SPR biosensors1481
Amplifying the response of colorimetric assays is a current challenge in the design of naked eye and simple detection methods.Although instrument-based detection is suitable for laboratory settings,naked eye or simple colorimetric detection by picture acquisition would not work with the minute changes in coloration of current Au NP-based assays.Strategies must be developed to overcome this limitation.By establishing an Au NP dissolution assay in the presence of Pb 2+,Lee and Huang demonstrated a colorimetric sensor sensitive to low nmol L −1levels of metal ions [32].The striking color change resulting from dissolution of the Au NP makes it possible to detect lead with the naked eye.Plasmonic coupling is another efficient method for increasing the color change of colorimetric bioassays,as demonstrated by aggregation assays for DNA [39].For example,an aggregation assay for melamine reported large color changes at the micromolar level [40].Another demonstration involved an aggregation assay for DNA and adenosine,which was spotted on paper for detection [41].Although spotting the product of the assay on paper works well,the assay is not conducted directly on the paper substrate and thus requires manipulation of the sample and reagents before detection.
An alternative method that is interesting (but has not yet been reported to the best of our knowledge),would exploit the immobilization of Au NP on the paper substrate and amplification of the color change with a secondary antibody labelled with an Au NP.This approach is very similar to SERS bioassays [42].One of the challenges will be to develop surfaces with NPs deposited at a specific location on the substrate,ideally at a distance greater than the dis-tance required for plasmonic coupling.Photolithographic and nanofabrication techniques are possible solutions to the fabrication of highly regular surfaces.Currently,these
types of surfaces are relatively expensive to fabricate,which may limit their suitability in low-cost applications.Thus,techniques with low production cost must be developed to create these substrates.By using techniques capable of generating surfaces with controlled spacing of Au NP,plas-monic coupling will be exclusive to the presence of the analyte (via the secondary antibody labelled with an Au NP).Colorimetric detection of analytes by reading a test-strip,similar to a pregnancy test,would be possible for a variety of applications.Issues about the stability of immo-bilized antibodies and conditions in which the assays must be performed would have to be resolved.Despite develop-ments still to come,the sensitivity and simplicity of moni-toring a striking color change of a bioassay will be highly beneficial and should be increasingly investigated in the near future.
Outlook
SPR sensors are commonplace in the development of novel analytical strategies for detection of biomolecules,pollutants,and chemical agents.Although the market for SPR instru-ments is continuously growing,with several instruments be-ing introduced recently,the development of small,portable,and miniature SPR detection schemes remains to be popular-ized for routine applications.Current instruments remain bulky and expensive or the small instruments commercially available require a significant input from the user or lack analytical performance.In addition,the stringent requirement in terms of sensitivity,limit of detection,and selectivity for the applications envisaged are increasing the complexity of the development of portable and miniature SPR sensors.To
Wavelength
(nm)
A) Fabrication protocol
B) Detection protocol
Bioreceptor
Light
Fig.4(A )AuNP monolayer can be constructed on glass by use of silane chemistry.The AuNP can be further
functionalized by use of thiol chemistry.(B )Direct detection of binding by measurement of a wavelength shift and an
increase in absorbance resulting from increased scattering characterizes the response of SPR sensors based on AuNP.Adapted,with permission,from Anal Chem 74(3):504-509.Copyright 2002American Chemical Society ([35])
1482
J.Breault-Turcot,J.Masson。