- Email: [email protected]
Application of chromogenic reagents in surface plasmon resonance (SPR)
Biosensors and Bioelectronics 22 (2007) 1163–1167
Short communication
Application of chromogenic reagents in surface plasmon resonance (SPR) Jan Mavri a,b,∗ , Peter Raspor b , Mladen Franko a a
University of Nova Gorica, Laboratory for Environmental Research, Vipavska 13, P.O. Box 301, SI-5000 Nova Gorica, Slovenia b University of Ljubljna, Biotechnical Faculty, Jamnikarjeva 101, SI-1000 Ljubljana, Slovenia Received 9 January 2006; received in revised form 6 July 2006; accepted 13 July 2006 Available online 22 August 2006
Abstract In this paper, a new simple approach for sensitivity optimization in surface plasmon resonance (SPR) chemosensors based on colorimetric ligands is presented. A new design of SPR sensor with tunable analytical wavelength (λSPR ) was constructed for this purpose, to perform studies on the ligand absorbance spectra related sensitivity enhancement. Unlike commercial SPR sensors which operate at one λSPR , the new device can be used for sensitivity analysis at selected λSPR in the range 550–750 nm, offering the possibility to identify the highest sensitivity λSPR in regard to the spectral changes of the selected ligand. Measurements can be easily done in ligand bulk solutions without immobilization steps. Sensitivity enhancement analysis and optimization of λSPR on chromogenic reagents with hypsochromic shift in their absorption spectra are demonstrated in this contribution. Optimal selection of analytical wavelength, set at the absorbance peak of chromogenic reagent Eriochrome Black T (EBT) was observed to result in up to two times increased SPR sensitivity to Cd2+ compared to wavelengths selected in other parts of the ligand absorbance spectra, with a limit of detection (LOD) 0.2 ppm. The sensitivity enhancement at optimal λSPR was observed to be related to increased refractive index (n), drop in extinction coefficient (α) and simultaneous hypsochromic shift of the EBT absorbance spectra causing the λSPR to match the absorbance peak shoulder. © 2006 Elsevier B.V. All rights reserved. Keywords: Surface plasmon resonance; Sensors; Colorimetric reagents; Heavy metals; Sensitivity enhancement
1. Introduction Surface plasmon resonance is an optical analytical technique, which offers the possibility to measure very small changes in the n of media on the surface of thin metal films, caused by molecular adsorption. During the last 10 years, it was efficiently employed for (bio)sensing in numerous fields such as molecular biology, medicine, biotechnology, drug and food monitoring, environmental monitoring. Depending on the recognition concept used on the detection surface (analyte–ligand), a wide variety of chemical and biochemical sensors have been developed for, the detection of various biological and organic molecules (antigens, toxins, pesticides, proteins, DNA . . .) (Silin and Plant, 1997; Homola et al., 1999). The margins of physical conditions for SPR, described in the plasmon dispersion relation (PDR) are narrow and strongly
∗
Corresponding author. Tel.: +386 1 423 11 61; fax: +386 1 2574 092. E-mail address: [email protected] (J. Mavri). URL: http://www.p-ng.si, http://www.bf.uni-lj.si.
0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.07.018
influenced by the optical properties of the dielectric, dielectric constant (ε —real part), which makes the technique extremely sensitive to small changes in n of the measured sample. LODs down to 5 × 10−7 refractive index unit can be reached corresponding to the binding of 1 pg/mm2 of adsorbate on the detection surface (Rich and Myszka, 2000). Beside high sensitivity, real-time measurement of analyte interactions in complex biological media and no need of molecular markers are the main advantages of SPR. Despite all qualities one of the main drawbacks of SPR applications today is analysis limited to adsorbates with molecular weight >500 Da, which efficiently cause the change of n of the sensing layer. A new and promising area in this regard involves the use of chromogenic ion-chelating reagents as sensitivity enhancement compounds for heavy metal ion detection. A series of new SPR designs using synthetic metal ion selective dyes was published in the last few years, reaching LODs as low as 1 ppt for metal ions (Kim et al., 2000a,b; Hur et al., 2002; Hanning et al., 1999; Lee et al., 2001; Ashwell et al., 1999). In these reports, the absorbance spectrum of the ligand was recognized to strongly influence SPR sensitivity. It should be noted, that
1164
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
according to Kramers–Kronig’s relation (ε = n2 − α2 ), change in the extinction coefficient can influence the PDR too and therefore, be efficiently detected by SPR (Raether, 1988; Salamon et al., 1997). The first studies concerning the influence of light absorbing species on SPs started already in the 1970s. Measurements performed with dyes applied in thin layers on SPR detection surface revealed distortions in the PDR function at λSPR close to sharp absorption bands, causing the resonant conditions to shift toward larger incident angles (W¨ahling et al., 1978; W¨ahling, 1981). Practical use of this phenomenon was first presented on solutions of hexamethylindodicarbocyanine dyes in the end of 1990s. The sensitivity of SPR was shown to strongly depend on the absorbance spectra shape of the dye and the related α. Increased sensitivity was demonstrated theoretically and in practice at λSPR matching shoulders of sharp absorbance bands of dye solutions at higher wavelength side. Dye added to solvents not showing own absorbance was found to abruptly increase the n at such wavelength in contrast with no change at the absorbance maximum, resulting in up to four times higher sensitivities (Hanning et al., 1999; Johnston et al., 1995; Pockrand et al., 1978). It should be noted that in these experiments the sensitivity enhancement did not rely on changes of the position of absorption bands in the spectra of the dye. Considerably, higher sensitivity enhancement was obtained using metal ion selective dyes (squarylium and azacrown dyes) immobilized on the sensing surface, reaching LODs down to 10−11 M, for Ag2+ , Cu2+ and Li+ (Kim et al., 2000a,b; Lee et al., 2001). The response in these detectors resulted from the contemporaneous change in n and α of the sensing layer due to the interaction between the dye and the analyte. Considering Hanning’s work the dye was selected to match λSPR at shoulder of dye’s absorbance maximum. The sensitivity enhancement was related to the hypochromic spectral shift after interaction with metal ion and the corresponding drop in α at the used λSPR . No change in the position of the absorbance maximum was present in the mentioned cases for the used ligands, which is a frequent event in many other colorimetric reagents. Since the slope of the graph of the absorbance spectra (dA/dλ) as well as α, influence SPR sensitivity, correct selection of λSPR is important in the case of bathochromic or hypsochromic reagents. Despite considerably low LODs in these applications, relatively limited work was done to completely clarify the exact relation between spectral changes and SPR response, as consequence of limited choice of suitable reagents, and corresponding light sources (λSPR ) to match enhancement conditions. The aim of this work was therefore, to develop a new SPR device and a methodology to perform fast and simple analysis of SPR sensitivity dependence on absorbance spectra of colorimetric reagents specific for analytes with low molecular weight. We present a new SPR configuration with tunable λSPR which offers the possibility to optimize analytical conditions and can contribute to the spread of the technique on numerous other chromogenic reagents used in UV–vis spectrophotometry. As well we present its practical use in analysis of sensitivity enhancement with chromogenic reagents, showing hypsochromic absorbance shifts and determination of optimal
conditions (λSPR ). We propose in addition, a new mechanism of sensitivity enhancement resulting from increased n, drop in α (Kramer–Kronig’s relation) and contemporaneous matching of λSPR at optimal analytical conditions (peak shoulder) after spectral change of the chromogenic ligand. 2. Experimental 2.1. Materials To investigate the influence of n and α on the SPR detector sensitivity and the correlation between λSPR and the absorption spectra of reagents showing hypsocromic shifts, model dyes; EB (4,4 -bis[7-(1-amino-8-hydroxy-2,4-disulfo)naphthylazo]3,3 -bitolyl tetrasodium salt, Kodak 3873, Eastman, USA) and EBT (1-naphthalenesulfonic acid, 3-hydroxy-4-[(1-hydroxy-2naphthyl)azo]-7-nitro-, monosodium salt, Merck 3170, Darmstadt, Germany) were used, in 10 mM phosphate buffer solutions (NaH2 PO4 , A.C.S., Sigma–Aldrich, Munich, Germany) using two times deionised water (>18 M cm−1 ). Standard solutions with EB consisted of 190 mg l−1 EB in phosphate buffer, pH 8 and 12, with d-glucose (A.C.S Riedel-de Ha¨en, Germany) added at concentrations 3–15 g l−1 . EBT standard solutions were prepared from EBT (102.9 mg l−1 ) in phosphate buffer (pH 10), with Cd2+ (99.99% CdCl2 ·2.5H2 O, Acros, NJ, USA) added in the concentration range 0–5 mg l−1 . The absorbance spectra of samples were measured on a UV–vis spectrometer HP 8453 (Agilent, USA). 2.2. SPR detector setup The SPR device used was originally based on the Kretschmann’s configuration ATR set-up (Kretschmann, 1971) but had the advantage of selective λSPR tuning in the range 550–750 nm, 10 nm bandwidth (Fig. 1). The light source was a 60 W Xe lamp (Cermax, CVI Spectral Products, NM, USA) connected to a monochromator (DK 240 1/4 Meter, CVI Spectral Products, NM, USA) and controlled by a 200 MHz personal computer (PC). A multimodal 0.5 mm optical fibre with 0.21 numerical aperture combined with a set of achromatic lenses (Thorlabs, NJ, USA), was used to drive light from the
Fig. 1. SPR detector with tunable λSPR : (1) Xe lamp, (2) monochromator, (3) PC, 4-optical fibre, (5) achromatic lenses, (6) flow cell, (7) glas prism, (8) detection disc, (9) polariser, (10) CCD, (11) oscilloscope, (12) electronic trigger and (13) pump.
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
monochromator to the flow cell with 10 l inner volume. Light from the optical fiber was focused through a BK 7 glass prism (Optosigma, CA, USA) on the detection disc. The detection discs were prepared in lab using microscope cover slides, coated with 48 nm of silver (99.99%, Sigma–Aldrich, Germany) by vacuum evaporation, according to procedures from the literature (Jordan et al., 1994). The detection disc was fixed on the prism with matching liquid (nd = 1.517, Neolab, Heidelberg, Germany) and mounted with the metal surface toward the flow cell. A Glan-Thompson polarizer (Elan, St. Petersburg, Russia) set after the prism was used to selectively pass only p-polarised light. Reflected light was further collimated on the CCD video camera light detector (Sony, B/W, 1/3 CCD, Sony, USA). The signals from the CCD were averaged (512 signals) on a digital oscilloscope (Wave-runner LTT 344, Lecroy, USA), triggered with an in-lab constructed electronic processor (LM1881, National Semiconductor, USA) and connected to the PC. The averaged signals were stored and processed on the computer using Microcal Origin 6 software (OriginLab Corporation, Northampton, USA). The signal on the oscilloscope—CCD photodiode voltage/pixel had a characteristic curved form with a minimum corresponding to light incidence angles matching resonant conditions of SPs. The optimal angle of resonance (θ SP ) (expressed in pixels) was determined from the signal interpolation Gauss function minimum. An HPLC pump (LC10 Ai, Schimadzu, Japan) was used to pump samples in the flow cell. 2.3. Measurement conditions The SPR sensitivity to n or metal ion concentration was determined using standard solutions of metal ions with colorimetric reagent or with known concentrations of glucose, used as solvent n modifier. All measurements were done in flow conditions at constant flow rate 0.5 ml min−1 . Five replicate recordings over 15 min were taken for each sample few minutes after sample front reached the flow cell. θ SP was linearly proportional to the n of the sample and therefore, sensitivity was determined from the slope (m) of the linear regression function of θ SP versus analyte concentration. The lower LOD was calculated from the standard deviation (S.D.) of the calculated θ SP from replicates, using the formula LOD = 3 × S.D./m. 3. Results and discussion Initial SPR sensitivity measurements with pure glucose water solutions resulted in increasing sensitivity for approximately 11% (0.72 ± 0.02 pixel l g−1 ) when decreasing the λSPR from 600 to 680 nm, supporting the theory on PDR (Raether, 1988). The LOD for glucose concentration was 0.10 ± 0.05 g l−1 . Solutions of EB at pH 8 show a characteristic absorbance maximum at λmax = 603 nm (blue), shifting to lower wavelengths, 565 nm at pH 12 (red), with a contemporaneous drop in absorbance. This effect was exploited in experiments to measure and compare the sensitivity of the detector at λSPR , set to match distinct parts of the EB absorbance spectra. Concretely, at pH 12 they were: left peak shoulder at 585 nm, peak apex at 605 nm, right peak
1165
Fig. 2. Surface plasmon resonance (SPR) sensitivity at various analytical wavelengths for Evans Blue (EB) samples in 10 mM NaH2 PO4 with corresponding overlaying EB absorbance spectra at pH 8 (- - - - -) and 12 (. . . .).
shoulder at 625 nm and two wavelengths matching parts closer to the right side of the peak base, 655 and 690 nm, respectively. Comparing the graphical representation of SPR sensitivity in pixel l g−1 as function of λSPR at both pH values, increased sensitivity can be observed when the absorbance drops in the EB spectra at pH 12 (Fig. 2). Considering the Kramer–Kronig’s relation these results confirm the expected increase in SPR sensitivity due to the decrease in absorbance and related α of the medium. The highest difference in the sensitivity between both pH values can be observed at 605 nm, resulting in a sensitivity ratio 1.28 ± 0.02 (for pH 12/ pH 8 samples). Observing the EB absorbance spectra at 605 nm, a shift in the position of the absorbance maximum toward lower wavelengths is evident versus λSPR . The hypsochromic shift related to the pH change caused the λSPR to apparently move from the peak apex (pH 8) toward its shoulder at higher wavelength side of the peak (pH 12), in relation to the dye absorbance spectra, which in compliance with the literature represents the part of the spectra where n is most influenced by α changes, the most sensitive part (Hanning et al., 1999). The observed sensitivity enhancement in experiments with EB was after that verified on EBT, to estimate practical application of this event for the detection of Cd2+ in water solutions. EBT is a colorimetric ligand used in UV–vis spectrophotometry for the determination of various cations, such as Ca2+ , Mg2+ , Al3+ , Ga3+ , In3+ (Snell, 1881; Gettar et al., 1999). The complexation of EBT with Cd2+ at pH 10, typically results in a hypsochromic shift in its absorbance spectra, with a simultaneous drop in absorbance. The EBT absorbance spectrum exhibits a peak at 618 nm with a saddle at approximately 650 nm. At highest concentration of Cd2+ added the absorbance maximum shifts to 578 nm (Fig. 3). In these experiments, the λSPR were therefore, chosen, respectively, at: 620 nm—peak, 650 nm—shoulder/saddle and 670 nm—shoulder. The measured absorbance at selected λSPR and the related α was calculated to drop for ∼1.8, 2.5 and 2.8 times, respectively at highest Cd2+ concentration. The results of SPR sensitivity analysis for the detection of Cd2+ with EBT are presented in Fig. 3.
1166
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
Fig. 3. SPR sensitivity to EBT solutions with Cd2+ at various analytical wavelengths and absorbance spectra hypsochromic shift of EBT (18.1 mg l−1 ) in 10 mM NaH2 PO4 (pH 10), complexed with Cd2+ .
The obtained results revealed the highest SPR sensitivity to Cd2+ concentration at 650 nm, followed by 620 and 670 nm wavelengths. However, observing the response linearity this was found to be most optimal at 620 nm, with LOD for Cd2+ concentration 0.20 ± 0.05 mg l−1 , resulting in up to two times higher sensitivity compared to measurements at 670 nm. To confirm that the SPR sensitivity enhancement was related to the complexation of Cd2+ with EBT, measurements were in addition done with Cd2+ solutions in 10 mM NaH2 PO4 , pH 10, at same analytical conditions, but without EBT. The sensitivity at 620 nm for Cd2+ solutions with EBT added resulted to be higher than in the case of Cd2+ solutions, 4.26 ± 0.05 and 6.35 ± 0.12 pixel l mg−1 , respectively. Considering their ratio (1.49 ± 0.03) an estimation of the sensitivity enhancement due to the complexation of the colorimetric reagent with Cd2+ was ∼49.1%. This data is in good correlation with experiments published by Hanning, where 10–20 times higher concentrations of dye used, resulted in up to 10 times higher sensitivity enhancements (Hanning et al., 1999). Similarly, as in EB measurements, the EBT hypsochromic shift in the spectra due to the complexation with Cd2+ caused a change in the position of λSPR in relation to the absorbance peak of the spectra. After the hypsochromic spectral change, the λSPR matches again the absorbance peak shoulder at higher wavelength side. At the same time a strong decrease in absorbance can be observed at 620 nm. The SPR response is supposed to result from increased n and drop in the absorbance coefficient and related α, in good agreement with other studies (Ashwell et al., 1999). The sensitivity enhancement can be attributed to the change in the shape and position of the PDR function (W¨ahling, 1981; W¨ahling et al., 1978), as result of EBT complexation with metal ions and the consequent absorbance spectra shift. Slightly higher value in SPR sensitivity calculated at 650 nm matching the saddle in the EBT absorbance spectra is presumed as consequence of a larger decrease in absorbance at the given wavelength (∼30% more than at 620 nm). The nonlinear detector’s response can be attributed in addition to changing contribution of dA/dλ and α to sensitivity at different Cd2+ concentrations which is difficult to determine in practice. This
fact confirms the importance of the possibility of appropriate (optimal) λSPR selection. Depending on the colorimetric reagent used, this is most easily determined empirically—with the proposed SPR detector with tunable λSPR . The LOD value for Cd2+ concentration is evidently higher compared to similar published SPR experiments on metal ion detection, reaching 1 × 10−12 M concentrations. Lower LODs in the later cases can be in great part attributed to the utilization of immobilized ligands, where the recognition effect occurred condensed on the sensing surface, the most sensitive part of SPR detectors (Kim et al., 2000a,b; Lee et al., 2001). It is, however, necessary to emphasize, that our experiments were primarily oriented in the investigation of SPR sensitivity dependence on the absorbance spectra characteristics of chromogenic reagents with shifts in the position of their absorbance maximum and for this purpose the development of appropriate analytical equipment and methodology. The use of ligands in solutions in our experimental work represents an efficient alternative to complex immobilization procedures making possible fast and simple preliminary determination of optimal λSPR in respect to the optical properties of the selected pair, analyte–ligand. 4. Conclusions A new SPR sensor with tunable λSPR was developed for simple and fast analysis of sensitivity enhancement related to the absorbance spectra of colorimetric reagents. Measurements on colorimetric reagents with hypsochromic spectral shift, EB and EBT resulted in highest sensitivity enhancement, when λSPR was selected to match the absorbance maximum of the reagent and coincided with its absorbance peak shoulder, with lower absorbance at higher wavelength side after spectral change. According to the results and in compliance with the theory, the sensitivity enhancement can be attributed to the contemporaneous increase of n, decrease of α and related absorbance coefficient as consequence of the absorbance spectra shift of the reagent, causing the change in the shape and position of PDR and related resonant conditions. Up to two times increased sensitivity to Cd2+ ions was obtained at absorbance maximum λSPR ,
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
using EBT compared to wavelengths in other parts of the spectra. The sensitivities reached using bulk reagent solutions are considerably lower compared to applications with immobilized reagents. However, the presented new detector and analytical methodology offer the possibility to perform simple study on numerous other conventional colorimetric reagents in preliminary phase, for future applications in SPR chemosensors. Acknowledgements The authors would like to thank the Ministry of Higher Education, Science and Technology, Slovenia, for the financial support and Dr. Egon Pavlica for his help in the technical part of the project. References Ashwell, G.J., Skjonnemand, K., Roberts, M.P.S., Allen, D.W., Li, X., Sworakowski, J., Chyla, A., Bienkowski, M., 1999. Clloids Surf. A: Physicochem. Eng. Aspects 155, 43–46. Gettar, R.T., Gautier, A.E., Servant, R.E., et al., 1999. J. Chromatogr. A 855, 111–119.
1167
Hanning, A., Roeraade, J., Delrow, J.J., Jorgenson, R.C., 1999. Sens. Actuators B 54, 25–36. Homola, J., Yee, S.S., Gauglitz, G., 1999. Sens. Actuators B 54, 3–15. Hur, Y., Ock, K., Kim, K., Jin, S., et al., 2002. Anal. Chim. Acta 460, 133– 139. Johnston, S.K., Karlsen, S.R., Jung, C.C., et al., 1995. Mater. Chem. Phys. 42, 242–246. Jordan, C.E., Frey, B.L., Kornguth, S., Corn, R.M., 1994. Langmuir 10, 3642–3648. Kim, S., Han, S., Jang, G., et al., 2000a. Dyes Pigments 44, 169–173. Kim, S., Han, S., Kim, J., et al., 2000b. Dyes Pigments 44, 55–61. Kretschmann, E., 1971. Ztg. Phys. 241, 313–314. Lee, S., Kang, S., Kim, D., 2001. Dyes Pigments 49, 109–115. Pockrand, I., Swalen, J.D., Santo, R., Brillant, A., Philpott, M.R., 1978. Chem. Phys. 69, 4001. Raether, H., 1988. Surface Plasmons on Smooth and Rough Surfaces and on Gratings. Springer–Verlag, Berlin Heidelberg. Rich, R.L., Myszka, D.G., 2000. Curr. Opin. Biotechnol. 11, 54–61. Salamon, Z., Angus Macleod, H., Tollin, G., 1997. Biochim. Biophys. Acta 1331, 117–129. Silin, V., Plant, A., 1997. TIBTECH (September (15)), 353–359. Snell, F.D., 1881. Photometric Fluorimetric Methods Analysis. J.Wiley & Sons, New York. W¨ahling, G., M¨obius, D., Raether, H., 1978. Z. Naturfosch. 33a, 907–909. W¨ahling, G., 1981. Z. Naturfosch. 36a, 588–594.
Short communication
Application of chromogenic reagents in surface plasmon resonance (SPR) Jan Mavri a,b,∗ , Peter Raspor b , Mladen Franko a a
University of Nova Gorica, Laboratory for Environmental Research, Vipavska 13, P.O. Box 301, SI-5000 Nova Gorica, Slovenia b University of Ljubljna, Biotechnical Faculty, Jamnikarjeva 101, SI-1000 Ljubljana, Slovenia Received 9 January 2006; received in revised form 6 July 2006; accepted 13 July 2006 Available online 22 August 2006
Abstract In this paper, a new simple approach for sensitivity optimization in surface plasmon resonance (SPR) chemosensors based on colorimetric ligands is presented. A new design of SPR sensor with tunable analytical wavelength (λSPR ) was constructed for this purpose, to perform studies on the ligand absorbance spectra related sensitivity enhancement. Unlike commercial SPR sensors which operate at one λSPR , the new device can be used for sensitivity analysis at selected λSPR in the range 550–750 nm, offering the possibility to identify the highest sensitivity λSPR in regard to the spectral changes of the selected ligand. Measurements can be easily done in ligand bulk solutions without immobilization steps. Sensitivity enhancement analysis and optimization of λSPR on chromogenic reagents with hypsochromic shift in their absorption spectra are demonstrated in this contribution. Optimal selection of analytical wavelength, set at the absorbance peak of chromogenic reagent Eriochrome Black T (EBT) was observed to result in up to two times increased SPR sensitivity to Cd2+ compared to wavelengths selected in other parts of the ligand absorbance spectra, with a limit of detection (LOD) 0.2 ppm. The sensitivity enhancement at optimal λSPR was observed to be related to increased refractive index (n), drop in extinction coefficient (α) and simultaneous hypsochromic shift of the EBT absorbance spectra causing the λSPR to match the absorbance peak shoulder. © 2006 Elsevier B.V. All rights reserved. Keywords: Surface plasmon resonance; Sensors; Colorimetric reagents; Heavy metals; Sensitivity enhancement
1. Introduction Surface plasmon resonance is an optical analytical technique, which offers the possibility to measure very small changes in the n of media on the surface of thin metal films, caused by molecular adsorption. During the last 10 years, it was efficiently employed for (bio)sensing in numerous fields such as molecular biology, medicine, biotechnology, drug and food monitoring, environmental monitoring. Depending on the recognition concept used on the detection surface (analyte–ligand), a wide variety of chemical and biochemical sensors have been developed for, the detection of various biological and organic molecules (antigens, toxins, pesticides, proteins, DNA . . .) (Silin and Plant, 1997; Homola et al., 1999). The margins of physical conditions for SPR, described in the plasmon dispersion relation (PDR) are narrow and strongly
∗
Corresponding author. Tel.: +386 1 423 11 61; fax: +386 1 2574 092. E-mail address: [email protected] (J. Mavri). URL: http://www.p-ng.si, http://www.bf.uni-lj.si.
0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.07.018
influenced by the optical properties of the dielectric, dielectric constant (ε —real part), which makes the technique extremely sensitive to small changes in n of the measured sample. LODs down to 5 × 10−7 refractive index unit can be reached corresponding to the binding of 1 pg/mm2 of adsorbate on the detection surface (Rich and Myszka, 2000). Beside high sensitivity, real-time measurement of analyte interactions in complex biological media and no need of molecular markers are the main advantages of SPR. Despite all qualities one of the main drawbacks of SPR applications today is analysis limited to adsorbates with molecular weight >500 Da, which efficiently cause the change of n of the sensing layer. A new and promising area in this regard involves the use of chromogenic ion-chelating reagents as sensitivity enhancement compounds for heavy metal ion detection. A series of new SPR designs using synthetic metal ion selective dyes was published in the last few years, reaching LODs as low as 1 ppt for metal ions (Kim et al., 2000a,b; Hur et al., 2002; Hanning et al., 1999; Lee et al., 2001; Ashwell et al., 1999). In these reports, the absorbance spectrum of the ligand was recognized to strongly influence SPR sensitivity. It should be noted, that
1164
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
according to Kramers–Kronig’s relation (ε = n2 − α2 ), change in the extinction coefficient can influence the PDR too and therefore, be efficiently detected by SPR (Raether, 1988; Salamon et al., 1997). The first studies concerning the influence of light absorbing species on SPs started already in the 1970s. Measurements performed with dyes applied in thin layers on SPR detection surface revealed distortions in the PDR function at λSPR close to sharp absorption bands, causing the resonant conditions to shift toward larger incident angles (W¨ahling et al., 1978; W¨ahling, 1981). Practical use of this phenomenon was first presented on solutions of hexamethylindodicarbocyanine dyes in the end of 1990s. The sensitivity of SPR was shown to strongly depend on the absorbance spectra shape of the dye and the related α. Increased sensitivity was demonstrated theoretically and in practice at λSPR matching shoulders of sharp absorbance bands of dye solutions at higher wavelength side. Dye added to solvents not showing own absorbance was found to abruptly increase the n at such wavelength in contrast with no change at the absorbance maximum, resulting in up to four times higher sensitivities (Hanning et al., 1999; Johnston et al., 1995; Pockrand et al., 1978). It should be noted that in these experiments the sensitivity enhancement did not rely on changes of the position of absorption bands in the spectra of the dye. Considerably, higher sensitivity enhancement was obtained using metal ion selective dyes (squarylium and azacrown dyes) immobilized on the sensing surface, reaching LODs down to 10−11 M, for Ag2+ , Cu2+ and Li+ (Kim et al., 2000a,b; Lee et al., 2001). The response in these detectors resulted from the contemporaneous change in n and α of the sensing layer due to the interaction between the dye and the analyte. Considering Hanning’s work the dye was selected to match λSPR at shoulder of dye’s absorbance maximum. The sensitivity enhancement was related to the hypochromic spectral shift after interaction with metal ion and the corresponding drop in α at the used λSPR . No change in the position of the absorbance maximum was present in the mentioned cases for the used ligands, which is a frequent event in many other colorimetric reagents. Since the slope of the graph of the absorbance spectra (dA/dλ) as well as α, influence SPR sensitivity, correct selection of λSPR is important in the case of bathochromic or hypsochromic reagents. Despite considerably low LODs in these applications, relatively limited work was done to completely clarify the exact relation between spectral changes and SPR response, as consequence of limited choice of suitable reagents, and corresponding light sources (λSPR ) to match enhancement conditions. The aim of this work was therefore, to develop a new SPR device and a methodology to perform fast and simple analysis of SPR sensitivity dependence on absorbance spectra of colorimetric reagents specific for analytes with low molecular weight. We present a new SPR configuration with tunable λSPR which offers the possibility to optimize analytical conditions and can contribute to the spread of the technique on numerous other chromogenic reagents used in UV–vis spectrophotometry. As well we present its practical use in analysis of sensitivity enhancement with chromogenic reagents, showing hypsochromic absorbance shifts and determination of optimal
conditions (λSPR ). We propose in addition, a new mechanism of sensitivity enhancement resulting from increased n, drop in α (Kramer–Kronig’s relation) and contemporaneous matching of λSPR at optimal analytical conditions (peak shoulder) after spectral change of the chromogenic ligand. 2. Experimental 2.1. Materials To investigate the influence of n and α on the SPR detector sensitivity and the correlation between λSPR and the absorption spectra of reagents showing hypsocromic shifts, model dyes; EB (4,4 -bis[7-(1-amino-8-hydroxy-2,4-disulfo)naphthylazo]3,3 -bitolyl tetrasodium salt, Kodak 3873, Eastman, USA) and EBT (1-naphthalenesulfonic acid, 3-hydroxy-4-[(1-hydroxy-2naphthyl)azo]-7-nitro-, monosodium salt, Merck 3170, Darmstadt, Germany) were used, in 10 mM phosphate buffer solutions (NaH2 PO4 , A.C.S., Sigma–Aldrich, Munich, Germany) using two times deionised water (>18 M cm−1 ). Standard solutions with EB consisted of 190 mg l−1 EB in phosphate buffer, pH 8 and 12, with d-glucose (A.C.S Riedel-de Ha¨en, Germany) added at concentrations 3–15 g l−1 . EBT standard solutions were prepared from EBT (102.9 mg l−1 ) in phosphate buffer (pH 10), with Cd2+ (99.99% CdCl2 ·2.5H2 O, Acros, NJ, USA) added in the concentration range 0–5 mg l−1 . The absorbance spectra of samples were measured on a UV–vis spectrometer HP 8453 (Agilent, USA). 2.2. SPR detector setup The SPR device used was originally based on the Kretschmann’s configuration ATR set-up (Kretschmann, 1971) but had the advantage of selective λSPR tuning in the range 550–750 nm, 10 nm bandwidth (Fig. 1). The light source was a 60 W Xe lamp (Cermax, CVI Spectral Products, NM, USA) connected to a monochromator (DK 240 1/4 Meter, CVI Spectral Products, NM, USA) and controlled by a 200 MHz personal computer (PC). A multimodal 0.5 mm optical fibre with 0.21 numerical aperture combined with a set of achromatic lenses (Thorlabs, NJ, USA), was used to drive light from the
Fig. 1. SPR detector with tunable λSPR : (1) Xe lamp, (2) monochromator, (3) PC, 4-optical fibre, (5) achromatic lenses, (6) flow cell, (7) glas prism, (8) detection disc, (9) polariser, (10) CCD, (11) oscilloscope, (12) electronic trigger and (13) pump.
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
monochromator to the flow cell with 10 l inner volume. Light from the optical fiber was focused through a BK 7 glass prism (Optosigma, CA, USA) on the detection disc. The detection discs were prepared in lab using microscope cover slides, coated with 48 nm of silver (99.99%, Sigma–Aldrich, Germany) by vacuum evaporation, according to procedures from the literature (Jordan et al., 1994). The detection disc was fixed on the prism with matching liquid (nd = 1.517, Neolab, Heidelberg, Germany) and mounted with the metal surface toward the flow cell. A Glan-Thompson polarizer (Elan, St. Petersburg, Russia) set after the prism was used to selectively pass only p-polarised light. Reflected light was further collimated on the CCD video camera light detector (Sony, B/W, 1/3 CCD, Sony, USA). The signals from the CCD were averaged (512 signals) on a digital oscilloscope (Wave-runner LTT 344, Lecroy, USA), triggered with an in-lab constructed electronic processor (LM1881, National Semiconductor, USA) and connected to the PC. The averaged signals were stored and processed on the computer using Microcal Origin 6 software (OriginLab Corporation, Northampton, USA). The signal on the oscilloscope—CCD photodiode voltage/pixel had a characteristic curved form with a minimum corresponding to light incidence angles matching resonant conditions of SPs. The optimal angle of resonance (θ SP ) (expressed in pixels) was determined from the signal interpolation Gauss function minimum. An HPLC pump (LC10 Ai, Schimadzu, Japan) was used to pump samples in the flow cell. 2.3. Measurement conditions The SPR sensitivity to n or metal ion concentration was determined using standard solutions of metal ions with colorimetric reagent or with known concentrations of glucose, used as solvent n modifier. All measurements were done in flow conditions at constant flow rate 0.5 ml min−1 . Five replicate recordings over 15 min were taken for each sample few minutes after sample front reached the flow cell. θ SP was linearly proportional to the n of the sample and therefore, sensitivity was determined from the slope (m) of the linear regression function of θ SP versus analyte concentration. The lower LOD was calculated from the standard deviation (S.D.) of the calculated θ SP from replicates, using the formula LOD = 3 × S.D./m. 3. Results and discussion Initial SPR sensitivity measurements with pure glucose water solutions resulted in increasing sensitivity for approximately 11% (0.72 ± 0.02 pixel l g−1 ) when decreasing the λSPR from 600 to 680 nm, supporting the theory on PDR (Raether, 1988). The LOD for glucose concentration was 0.10 ± 0.05 g l−1 . Solutions of EB at pH 8 show a characteristic absorbance maximum at λmax = 603 nm (blue), shifting to lower wavelengths, 565 nm at pH 12 (red), with a contemporaneous drop in absorbance. This effect was exploited in experiments to measure and compare the sensitivity of the detector at λSPR , set to match distinct parts of the EB absorbance spectra. Concretely, at pH 12 they were: left peak shoulder at 585 nm, peak apex at 605 nm, right peak
1165
Fig. 2. Surface plasmon resonance (SPR) sensitivity at various analytical wavelengths for Evans Blue (EB) samples in 10 mM NaH2 PO4 with corresponding overlaying EB absorbance spectra at pH 8 (- - - - -) and 12 (. . . .).
shoulder at 625 nm and two wavelengths matching parts closer to the right side of the peak base, 655 and 690 nm, respectively. Comparing the graphical representation of SPR sensitivity in pixel l g−1 as function of λSPR at both pH values, increased sensitivity can be observed when the absorbance drops in the EB spectra at pH 12 (Fig. 2). Considering the Kramer–Kronig’s relation these results confirm the expected increase in SPR sensitivity due to the decrease in absorbance and related α of the medium. The highest difference in the sensitivity between both pH values can be observed at 605 nm, resulting in a sensitivity ratio 1.28 ± 0.02 (for pH 12/ pH 8 samples). Observing the EB absorbance spectra at 605 nm, a shift in the position of the absorbance maximum toward lower wavelengths is evident versus λSPR . The hypsochromic shift related to the pH change caused the λSPR to apparently move from the peak apex (pH 8) toward its shoulder at higher wavelength side of the peak (pH 12), in relation to the dye absorbance spectra, which in compliance with the literature represents the part of the spectra where n is most influenced by α changes, the most sensitive part (Hanning et al., 1999). The observed sensitivity enhancement in experiments with EB was after that verified on EBT, to estimate practical application of this event for the detection of Cd2+ in water solutions. EBT is a colorimetric ligand used in UV–vis spectrophotometry for the determination of various cations, such as Ca2+ , Mg2+ , Al3+ , Ga3+ , In3+ (Snell, 1881; Gettar et al., 1999). The complexation of EBT with Cd2+ at pH 10, typically results in a hypsochromic shift in its absorbance spectra, with a simultaneous drop in absorbance. The EBT absorbance spectrum exhibits a peak at 618 nm with a saddle at approximately 650 nm. At highest concentration of Cd2+ added the absorbance maximum shifts to 578 nm (Fig. 3). In these experiments, the λSPR were therefore, chosen, respectively, at: 620 nm—peak, 650 nm—shoulder/saddle and 670 nm—shoulder. The measured absorbance at selected λSPR and the related α was calculated to drop for ∼1.8, 2.5 and 2.8 times, respectively at highest Cd2+ concentration. The results of SPR sensitivity analysis for the detection of Cd2+ with EBT are presented in Fig. 3.
1166
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
Fig. 3. SPR sensitivity to EBT solutions with Cd2+ at various analytical wavelengths and absorbance spectra hypsochromic shift of EBT (18.1 mg l−1 ) in 10 mM NaH2 PO4 (pH 10), complexed with Cd2+ .
The obtained results revealed the highest SPR sensitivity to Cd2+ concentration at 650 nm, followed by 620 and 670 nm wavelengths. However, observing the response linearity this was found to be most optimal at 620 nm, with LOD for Cd2+ concentration 0.20 ± 0.05 mg l−1 , resulting in up to two times higher sensitivity compared to measurements at 670 nm. To confirm that the SPR sensitivity enhancement was related to the complexation of Cd2+ with EBT, measurements were in addition done with Cd2+ solutions in 10 mM NaH2 PO4 , pH 10, at same analytical conditions, but without EBT. The sensitivity at 620 nm for Cd2+ solutions with EBT added resulted to be higher than in the case of Cd2+ solutions, 4.26 ± 0.05 and 6.35 ± 0.12 pixel l mg−1 , respectively. Considering their ratio (1.49 ± 0.03) an estimation of the sensitivity enhancement due to the complexation of the colorimetric reagent with Cd2+ was ∼49.1%. This data is in good correlation with experiments published by Hanning, where 10–20 times higher concentrations of dye used, resulted in up to 10 times higher sensitivity enhancements (Hanning et al., 1999). Similarly, as in EB measurements, the EBT hypsochromic shift in the spectra due to the complexation with Cd2+ caused a change in the position of λSPR in relation to the absorbance peak of the spectra. After the hypsochromic spectral change, the λSPR matches again the absorbance peak shoulder at higher wavelength side. At the same time a strong decrease in absorbance can be observed at 620 nm. The SPR response is supposed to result from increased n and drop in the absorbance coefficient and related α, in good agreement with other studies (Ashwell et al., 1999). The sensitivity enhancement can be attributed to the change in the shape and position of the PDR function (W¨ahling, 1981; W¨ahling et al., 1978), as result of EBT complexation with metal ions and the consequent absorbance spectra shift. Slightly higher value in SPR sensitivity calculated at 650 nm matching the saddle in the EBT absorbance spectra is presumed as consequence of a larger decrease in absorbance at the given wavelength (∼30% more than at 620 nm). The nonlinear detector’s response can be attributed in addition to changing contribution of dA/dλ and α to sensitivity at different Cd2+ concentrations which is difficult to determine in practice. This
fact confirms the importance of the possibility of appropriate (optimal) λSPR selection. Depending on the colorimetric reagent used, this is most easily determined empirically—with the proposed SPR detector with tunable λSPR . The LOD value for Cd2+ concentration is evidently higher compared to similar published SPR experiments on metal ion detection, reaching 1 × 10−12 M concentrations. Lower LODs in the later cases can be in great part attributed to the utilization of immobilized ligands, where the recognition effect occurred condensed on the sensing surface, the most sensitive part of SPR detectors (Kim et al., 2000a,b; Lee et al., 2001). It is, however, necessary to emphasize, that our experiments were primarily oriented in the investigation of SPR sensitivity dependence on the absorbance spectra characteristics of chromogenic reagents with shifts in the position of their absorbance maximum and for this purpose the development of appropriate analytical equipment and methodology. The use of ligands in solutions in our experimental work represents an efficient alternative to complex immobilization procedures making possible fast and simple preliminary determination of optimal λSPR in respect to the optical properties of the selected pair, analyte–ligand. 4. Conclusions A new SPR sensor with tunable λSPR was developed for simple and fast analysis of sensitivity enhancement related to the absorbance spectra of colorimetric reagents. Measurements on colorimetric reagents with hypsochromic spectral shift, EB and EBT resulted in highest sensitivity enhancement, when λSPR was selected to match the absorbance maximum of the reagent and coincided with its absorbance peak shoulder, with lower absorbance at higher wavelength side after spectral change. According to the results and in compliance with the theory, the sensitivity enhancement can be attributed to the contemporaneous increase of n, decrease of α and related absorbance coefficient as consequence of the absorbance spectra shift of the reagent, causing the change in the shape and position of PDR and related resonant conditions. Up to two times increased sensitivity to Cd2+ ions was obtained at absorbance maximum λSPR ,
J. Mavri et al. / Biosensors and Bioelectronics 22 (2007) 1163–1167
using EBT compared to wavelengths in other parts of the spectra. The sensitivities reached using bulk reagent solutions are considerably lower compared to applications with immobilized reagents. However, the presented new detector and analytical methodology offer the possibility to perform simple study on numerous other conventional colorimetric reagents in preliminary phase, for future applications in SPR chemosensors. Acknowledgements The authors would like to thank the Ministry of Higher Education, Science and Technology, Slovenia, for the financial support and Dr. Egon Pavlica for his help in the technical part of the project. References Ashwell, G.J., Skjonnemand, K., Roberts, M.P.S., Allen, D.W., Li, X., Sworakowski, J., Chyla, A., Bienkowski, M., 1999. Clloids Surf. A: Physicochem. Eng. Aspects 155, 43–46. Gettar, R.T., Gautier, A.E., Servant, R.E., et al., 1999. J. Chromatogr. A 855, 111–119.
1167
Hanning, A., Roeraade, J., Delrow, J.J., Jorgenson, R.C., 1999. Sens. Actuators B 54, 25–36. Homola, J., Yee, S.S., Gauglitz, G., 1999. Sens. Actuators B 54, 3–15. Hur, Y., Ock, K., Kim, K., Jin, S., et al., 2002. Anal. Chim. Acta 460, 133– 139. Johnston, S.K., Karlsen, S.R., Jung, C.C., et al., 1995. Mater. Chem. Phys. 42, 242–246. Jordan, C.E., Frey, B.L., Kornguth, S., Corn, R.M., 1994. Langmuir 10, 3642–3648. Kim, S., Han, S., Jang, G., et al., 2000a. Dyes Pigments 44, 169–173. Kim, S., Han, S., Kim, J., et al., 2000b. Dyes Pigments 44, 55–61. Kretschmann, E., 1971. Ztg. Phys. 241, 313–314. Lee, S., Kang, S., Kim, D., 2001. Dyes Pigments 49, 109–115. Pockrand, I., Swalen, J.D., Santo, R., Brillant, A., Philpott, M.R., 1978. Chem. Phys. 69, 4001. Raether, H., 1988. Surface Plasmons on Smooth and Rough Surfaces and on Gratings. Springer–Verlag, Berlin Heidelberg. Rich, R.L., Myszka, D.G., 2000. Curr. Opin. Biotechnol. 11, 54–61. Salamon, Z., Angus Macleod, H., Tollin, G., 1997. Biochim. Biophys. Acta 1331, 117–129. Silin, V., Plant, A., 1997. TIBTECH (September (15)), 353–359. Snell, F.D., 1881. Photometric Fluorimetric Methods Analysis. J.Wiley & Sons, New York. W¨ahling, G., M¨obius, D., Raether, H., 1978. Z. Naturfosch. 33a, 907–909. W¨ahling, G., 1981. Z. Naturfosch. 36a, 588–594.