2988 OPTICS LETTERS / Vol. 30, No. 22 / November 15, 2005

Investigation of dual-channel fiber-optic surfaceplasmon resonance sensing for biological

applications

Wei Peng, Soame Banerji, Yoon-Chang Kim, and Karl S. BookshArizona Applied NanoSensors, Department of Chemistry and Biochemistry, Arizona State University,

Tempe, Arizona 85287-1604

Received May 31, 2005; revised manuscript received July 21, 2005; accepted July 21, 2005

A dual-channel fiber-optic sensor based on surface plasmon resonance (SPR) for self-referencing refractive-index measurements has been proposed. Most applications of fiber-optic SPR sensors are designed to mea-sure the refractive index of a liquid or gas sample by measuring the signal from a single surface, the sen-sitivity and stability of which is easily affected by the fluctuation of external environmental conditions. Wehave designed a dual-channel fiber-optic surface sensor with two independent SPR signals from two areas ofthe same probe. A prototype sensor was fabricated and characterized. The preliminary experimental resultsdemonstrate the characteristic responses of both SPR signals from two channels that independently corre-spond to the refractive index changes in the liquid samples with which they are in contact. The design couldbe extended to a multichannel sensor with further developments. The experimental results confirmed thatone channel can be used as a reference sensor that could compensate for unexpected changes in bulk refrac-tion or temperature and develop this sensor as a practicable high-sensitivity biosensing device. © 2005 Op-tical Society of America

OCIS codes: 240.6680, 240.6690, 060.2370, 220.0220.

In the past decades, optical surface plasmon reso-nance (SPR) has been widely investigated for chemi-cal and biological sensing applications because of itscapability for real-time measurement with high sen-sitivity and label-free quantization of biochemicalcompounds.1,2 Compared to conventionalKretschman prism-based SPR sensing devices,3–5

fiber-optic SPR sensors have recently drawn consid-erable attention because of their fundamentally sim-pler structure, lower cost, smaller sample volume,and suitability for remote-sensing applications.6–10

Multimode fiber-optic SPR theory has been exten-sively described.8 With fiber-optic sensors, the clad-ding and the buffer are removed from a small sectionof the fiber to expose the core. This can be accom-plished by mechanical or chemical stripping alongthe side of the fiber to make a flat-tipped sensor forwhich the sensing area is on the side of the fiberoptic.6,7 Alternatively, a fiber-optic SPR probe can beconstructed by polishing through the cladding at thebuffer at a predetermined pair of complementaryangles to make a beveled tip sensor for which thesensing area is at the probe tip.8–10 In contrast toprism-based SPR sensors, fiber-optic SPR sensorscan function only in the constant angle, multiplewavelength regime. Hence low-numerical-aperturefibers are employed to limit the range of anglespropagating along the fiber. Multimode fiber-optic ge-ometry, however, allows the sensors to be employedas a compact dip probe for in situ analyses.

Currently, most fiber-optic SPR devices are usedfor single chemical parameter detection. As the re-quirements of high-sensitivity, multianalyte detec-tion increase for biological and environmental moni-toring, multichannel SPR sensors have attractedmore and more attention. Nenninger et al.11 reporteda light pipe configuration SPR biosensor with two

sensor surfaces for dual-channel monitoring of

0146-9592/05/222988-3/$15.00 ©

antibody–antigen binding. A second channel is usedto compensate for changes in the refractive index ofthe bulk solution caused by analyte concentration ortemperature difference. Berger et al.12 presented aprism-based SPR device containing multisensingchannels with recognition elements for specific ana-lytes in liquid phase. All these reported SPR devicesare based on bulk optical devices and need flow cellsfor liquid analyses that are not suitable for in situ re-mote sensing in biological environments such asfiber-optic SPR sensors.

Single-channel SPR systems lack the compensat-ing ability to prevent the unexpected effects that arecaused by the changes of instrumental or externaltesting conditions, which can also generate signifi-cant changes in the surface plasmon. Thus there areserious limitations on situations in which SPR can beeffectively used as the transduction mechanism forchemical biosensors in real-world applications. Inthis Letter we present a dual-channel fiber-optic SPRdevice with two sensing regions that respond to re-fractive index (RI) changes independently. This probeis similar to a reported novel tetra-tapered dual-channel fiber-optic SPR device that can be used forboth vapor and liquid analyses.13 However, the maindifference is that the new design is aimed at dual-channel measurements in one medium where bulkrefractive index compensation is required or wheretwo different analytes are to be monitored. This ge-ometry would be useful for in situ monitoring withthe chemical, biochemical, and biomedical applica-tions in which SPR sensors have been readilyemployed.7–9,14,15

The basic design and manufacture of flat-tippedfiber-optic SPR probes and the fiber-optic SPR sensorsystem were described previously.7 Figure 1 is a briefillustration of the fiber-optic SPR sensor system used

in this study. A bifurcated 0.22 numerical-aperture

2005 Optical Society of America

November 15, 2005 / Vol. 30, No. 22 / OPTICS LETTERS 2989

200 �m optical fiber jumper was used to guide thelight from the dc regulated quartz–tungsten–halogen(QTH) lamp (Oriel) to the probe and back to the de-tector. A hot filter was employed to block light greaterthan 850 nm. Light from the QTH lamp was focusedon the optical fiber with a subminiature version Afiber-optic coupling lens (Thor Labs). For the initialtests, returned light from 502 to 950 nm was dis-persed by a Kaiser Hollospec spectrograph (KaiserOptical Systems, Inc.) and collected on a PrincetonInstruments CCD camera (Roper Scientific) with aresolution of 0.438 nm. For the small quantitativestudy, a Chromex 500IS imaging spectrometer fromBruker Optics, Inc., with a 300 groove/mm gratingprovided 86.9 nm of spectral coverage. Dispersedlight was collected by a 1024�1024 pixel Andor CCDcamera with a resolution of 0.085 nm.

The system employed for this study deviates fromprevious fiber-optic sensors in having an increasednumber of sensing regions on the probe tip. Here two5 mm sensing regions were placed approximately5 mm apart on a 400 �m core, 0.39 numerical-aperture Thorlabs TECS buffered optical fiber. Forboth regions the core and the cladding were removed.We sputtered 5 nm of chrome and 50 nm of gold ontoeach sensing region. Similarly, a gold mirror (5 nm ofchrome, 200 nm of gold) was sputtered onto the fibertip. The sensing area nearest the distil end wascoated with a polymer matrix to shift the SPR signalapproximately 100 nm. This strategy was previouslyemployed with prism-based SPR sensors.16 Poly(ally-lamine hydrochloride) was chosen as the coating ma-terial for its good water solubility and low toxicity.17

The lower half of the SPR probe was immersed in a50 mM solution of dithiobis(succinimidyl propionate)in dimethyl sulfoxide for 30 min. and then placed for2 h in a solution of 10 mM poly(allylamine hydrochlo-ride) and 10 mM sodium hydroxide in water with ep-ichlorohydrin added as a cross linker. This solutionformed a polymer layer upon the sensing surface,causing the SPR minima to shift by approximately100 nm. When the light was incident into the fiber,the two SPR sensing areas produced two SPR signalsat different wavelengths.

The dual-channel SPR probe was tested. The char-acteristic SPR spectra for aqueous samples were ob-tained by normalizing the spectra to an air referencespectrum collected for the sensor before analyses.Figure 2 shows the water SPR spectrum detected bythis sensor probe. It shows that two SPR sensing ar-eas fabricated along the same sensor probe have twoSPR dips located with the first bare-gold SPR sensingarea at 573.9 nm and the second polymerized SPRsensing channel displaying a dip at 680.9 nm.

The ability of both channels in the dual-channelsensor to track RI changes across a wide range of so-lutions and experimental conditions was demon-strated with a series of six RI standards that in-cluded methanol (1.3280), water (1.3325), 4:1 water–isopropanol (1.3350), 2:1 water–isopropanol (1.3420),1:2 water–isopropanol (1.3440), and pure isopropanol(1.3450). The minimum of each spectrum ��R� was de-

termined by least-squares fitting of a second-order

polynomial about each localized resonance. Plottingthe calculated �R for each sensing region yielded cali-bration curves for both channels (Fig. 3) that are ap-proximately linear across the span of RIs tested. Thecorresponding linear regression equations are

�RI1 = 956.448n − 419.765,

�RI2 = 849.369n − 531.956, �1�

where �RI1 and �RI2 are the two SPR coupling wave-lengths and n is the refractive index of the sample.The 10% increase in calibration slope for the polymer(redshifted) region is expected because the change inwavelength with respect to a change in RI increasesas �R shifts to the red.

The root-mean-square errors of these two calibra-tions are 1.084 and 0.341 nm, which correspond to RIstability of the order of 10−3 to 10−4 RI after self-referencing. Whereas in most cases a much more pre-cise control of the determined RI is needed, root-mean-square errors of this magnitude are commonfor wide calibration ranges and are justified by the2�10−4 precision on the Abbe refractometer em-ployed to analyze the RI standards. The resolution ofthe Chromex 500IS spectrometer that we used forthis test is 0.085 nm; thus the smallest detectablechanges with the polymerized gold SPR and the bare-gold SPR channel for refractive-index changes are

Fig. 1. Schematic of the flat-tipped dual-channel fiber-optic SPR sensor system.

Fig. 2. SPR spectrum from the dual-channel flat-tipped

fiber-optic sensor probe.

2990 OPTICS LETTERS / Vol. 30, No. 22 / November 15, 2005

8.889�10−5 and 1.0�10−4 separately. Consequently,over this wide calibration range a self-referenced sta-bility of 4 parts in 103 is possible.

Just as for other fiber-optic sensors, usually tem-perature fluctuations are a main factor in determin-ing the SPR probe’s signal stability. The dual-channelstructure of this sensor makes it possible to fabricatea temperature-compensating sensor probe. Byplacing the sensor probe described above in atemperature-controllable water container, we testedthe temperature dependence of both channels over a35°C range spanning room temperature and physi-ological temperature (Fig. 4). The linear temperaturecalibration is

�T1 = − 0.637T + 734.760,

�T2 = − 0.249T + 607.237, �2�

where �T1 and �T2 are the SPR coupling wavelengthsof the two SPR channels at temperature T. With theChromex 500IS spectrometer the temperature reso-lution of the bare-gold coated and polymer-recoated

Fig. 3. Refractive-index calibration curves for the dual-channel SPR probe.

Fig. 4. Temperature dependence testing for the dual-channel SPR sensing probe.

SPR channels was 0.133°C and 0.341°C separately.As with the RI calibration, the resolution would beimproved at the expense of dynamic range by use of amore finely ruled grating in the spectrometer.

By calibrating each sensing region to temperatureand background RI, one can eliminate thetemperature–matrix effect on the analyte-coated re-gion. There are numerous analyte-responsive coat-ings, including antibodies, aptomers, and molecularimprinted polymerizers, that can be placed on thesensing region18,19 and are compatible with self-compensating, highly selective chemical sensing.Further investigation is in process to develop thissensor as a practicable self-compensating SPR bio-sensor probe for biological monitoring that will elimi-nate unexpected environmental condition changes aswell as bulk refractive-index variations in less-controlled environments.

The authors acknowledge financial support fromthe National Science Foundation (grantOCE0119999). W. Peng’s e-mail address iswei.peng@asu.eduReferences

1. H. Raether, Surface Plasmons on Smooth and RoughSurfaces and on Gratings (Springer-Verlag, 1988).

2. J. Homola, S. S. Yee, and G. Gauglitz, Sens. ActuatorsB 54, 3 (1999).

3. C. P. Cahill, K. S. Johnston, and S. S. Yee, Sens.Actuators B 45, 161 (1997).

4. J. Bowen, L. J. Noe, B. P. Sullivan, K. Morris, V.Martin, and G. Donnelly, Appl. Spectrosc. 57, 906(2003).

5. P. Tobiska, O. Hugon, A. Trouillet, and H. Gagnaire,Sens. Actuators B 74, 168 (2001).

6. R. Slavik, J. Homola, and J. Ctyroky, Sens. ActuatorsB 51, 311 (1998).

7. J. F. Masson, M. Barnhart, T. M. Battaglia, G. E.Morris, R. A. Nieman, P. J. Young, C. L. Lorson, and K.S. Booksh, Analyst (Cambridge, U.K.) 129, 855 (2004).

8. L. A. Obando and K. S. Booksh, Anal. Chem. 71, 5116(1999).

9. L. A. Obando, D. J. Gentleman, J. R. Holloway, and K.S. Booksh, Sens. Actuators B 100, 439 (2004).

10. A. K. Sharma and B. D. Gupta, Sens. Actuators B 100,423 (2004).

11. G. G. Nenninger, J. B. Clendenning, C. E. Furlong, andS. S. Yee, Sens. Actuators B 51, 38 (1998).

12. C. H. Berger, T. A. M. Beumer, R. P. H. Kooyman, andJ. Greve, Anal. Chem. 70, 703 (1998).

13. Y. C. Kim, W. Peng, S. Banerji, and K. S. Booksh, Opt.Lett. 30, 2218 (2005).

14. Y. C. Kim, S. Banerji, J. F. Masson, W. Peng, and K. S.Booksh, Analyst (Cambridge, U.K.) 130, 834 (2005).

15. J. F. Masson, L. A. Obando, S. Beaudoin, and K. S.Booksh, Talanta 62, 865 (2004).

16. P. Parmpi and P. Kofinas, Biomaterials 25, 1969 (2004).17. P. Kofinas and R. E. Cohen, Biomaterials 18, 1361

(1997).18. A. L. Jenkins, R. Yin, and J. L. Jensen, Analyst

(Cambridge, U.K.) 126, 798 (2001).19. Z. Zhang, H. Li, H. Liao, L. Nie, and S. Yao, Anal. Lett.

38, 203 (2005).