Optimization of Plasmonic Enhancement of Fluorescence on Plastic

Langmuir 2008, 24, 11261-11267

11261

Optimization of Plasmonic Enhancement of Fluorescence on Plastic Substrates Robert I. Nooney, Ondrej Stranik, Colette McDonagh, and Brian D. MacCraith* Biomedical Diagnostics Institute, National Centre for Sensor Research, School of Physical Sciences, Dublin City UniVersity, GlasneVin, Dublin 9, Ireland ReceiVed December 18, 2007 In this work, we report on the uniform deposition of tailored plasmonic coatings on polymer substrates and on the distance dependence of the plasmonic enhancement of a fluorescent dye. Silver, gold, and silver/gold alloy nanoparticles (NPs) with a range of diameters were synthesized using chemical techniques and characterized using UV-vis absorption spectroscopy, transmission electron microscopy (TEM), and atomic force microscopy (AFM). Reproducible polyelectrolyte (PEL) layers, which were deposited on plastic microwell plates using a layer-by-layer technique, served as both a stable and uniform substrate for deposition of the NPs as well as providing spacer layers of known thickness between the NPs and the fluorescent dye. A maximum enhancement factor of ∼11 was measured for 60 nm diameter pure silver NPs, for a dye-NP separation of approximately 3 nm. A shift in the localized surface plasmon resonance (LSPR) wavelength as a function of the effective refractive index of the PEL layers was also observed, and the measured shifts show a similar trend with theoretical predictions. This work will contribute toward the rational design of optical biochip platforms based on plasmon-enhanced fluorescence.

Introduction Increasing the sensitivity of fluorescence-based immunoassays is a key area of research in the development of biosensor platforms for a range of applications, including point-of-care devices for the detection of disease-related biomarkers. The detected fluorescence can be significantly enhanced by exploiting the plasmonic enhancement which can occur when a metal nanoparticle (NP) is placed in the vicinity of a fluorescent label or dye. This effect is due to the localized surface plasmon resonance (LSPR) associated with the metal NP, which modifies the intensity of the electromagnetic field around the dye and which, under certain conditions, increases the detected fluorescence signal. The effect is dependent on many parameters such as NP composition, NP size and shape, NP-dye separation, and dye quantum efficiency. The interaction of an electromagnetic (EM) wave with a spherical particle has been solved analytically by Mie.1 For the special case where the particle is smaller than the wavelength of the incident light and where the dielectric constant is negative (as is the case of noble metal NPs), Mie theory predicts a resonance between the incident light and the particle, leading to a large increase in the extinction coefficient. Two distinct enhancement effects are possible: an increase in the excitation rate of the dye and an increase in the quantum efficiency of the dye. The first effect occurs because the excitation rate is directly proportional to the square of the electric field amplitude, and the maximum enhancement occurs when the resonance wavelength, λres, coincides with the dye absorption band.2 The second effect involves an increase in the quantum efficiency of the dye and is maximized when λres coincides with the dye emission band.2-4 The work reported here focuses on the excitation enhancement approach, and λres has been tuned accordingly. Previous work * Corresponding author. E-mail: [email protected]. Phone: 0035317005299. (1) Mie, G. Ann. Phys. 1908, 25, 377. (2) Stranik, O. Plasmonic Enhancement of Fluorescence for Biomedical Diagnostics. Ph.D. Thesis, Dublin City University, Dublin, Ireland, 2007. (3) Lakowicz, J. R.; Shen, B.; Gryczynski, Z.; D’Auria, S.; Gryczynski, I. Biochem. Biophys. Res. Commun. 2001, 286, 875. (4) Chen, Y.; Munechuka, K.; Ginger, D. S. Nano Lett. 2007, 7, 690.

by the authors5,6 has demonstrated that it is possible to tune λres across the visible spectrum by altering NP size, shape, and composition. In recent years, much work has been published on the topic of plasmonic enhancement of fluorescence, and many experimental configurations have been used. For example, Ku¨hn et al. studied the interaction of a single gold NP with a single oriented dye molecule using near-field scanning optical microscopy (NSOM).7 In this highly controlled system, a gold NP was bound to the NSOM cantilever tip and moved over the dye molecule to investigate the behavior of the dye emission and lifetime as a function of dye-NP separation. However, for biochip applications, it is more relevant to examine the enhancement behavior in a planar configuration incorporating a monolayer of dye molecules and NPs. In the work reported here, a simple and very reproducible technique was used to fabricate a dye-NP monolayer system with controlled separation. Reproducible and uniform NP coatings have been achieved by using a polyelectrolyte (PEL) layer-by-layer deposition technique8 to produce initially a uniform positively charged surface to which the negatively charged NPs bind. Moreover, the dependence of fluorescence enhancement on the dye-NP separation was investigated by using the charged PEL layers as stackable spacers. The dye used in this study was bis(2,2′-bipyridine)-(5-isothiocyanato-phenanthroline) ruthenium bis(hexafluorophosphate), which was conjugated to a PEL (hereafter referred to as Ru-PEL) in order to ensure more reproducible deposition. In this work, the NP-dye system was deposited onto disposable 96-well plastic microplates. The paper sets out to (i) identify the optimal NP composition for maximum fluorescence enhancement, (ii) develop the optimum technique for reproducible deposition of the optimized NPs on (5) Stranik, O.; McEvoy, H. M.; McDonagh, C.; MacCraith, B. D. Sens. Actuators, B 2005, 107, 148. (6) Stranik, O.; Nooney, R.; McDonagh, C.; MacCraith, B. D. Plasmonics 2006, 2, 15. (7) Ku¨hn, S.; Håkanson, U.; Rogobete, L.; Sanoghdar, V. Phys. ReV. Lett. 2006, 97, Art. No. 017402. (8) Schmitt, J.; Decher, G.; Dressick, W. J.; Brandow, S. L.; Geer, R. E.; Shashidhar, R.; Calvert, J. M. AdV. Mater. 1997, 9, 661.

10.1021/la801631w CCC: $40.75  2008 American Chemical Society Published on Web 09/05/2008

11262 Langmuir, Vol. 24, No. 19, 2008

Nooney et al.

a planar substrate, and (iii) establish the optimum dye-NP separation in order to achieve maximum enhancement in a planar configuration. The main motivation for this work is the development of a high-sensitivity diagnostic biochip based on plasmon-enhanced fluorescence. In demonstrating a reproducible deposition technique which is compatible with planar biochip platforms and in achieving a significant enhancement factor of ∼11, this paper establishes the strategic route to achieving a generically important analytical objective. The paper also includes a detailed investigation on the affect of metal NPs on the sensitivity of fluorescence and limit of detection (LOD).

Experimental Section Materials. Hydrogen tetrachloroaurate, silver nitrate, aniline, sodium citrate, poly(ethyleneimine) (PEI, molecular weight, Mw 750 000 g), polystyrene sulfonate (PSS, Mw 250 000 g), poly(allyamine) hydrochloride (PAH, Mw 70 000 g), sodium chloride, bis(2,2′-bipyridine)-5-isothiocyanato-phenanthroline-ruthenium bishexafluorophosphate (Ru-ITC), dimethyldichlorosilane, and chloroform were all purchased from Sigma-Aldrich. Deionized water >18 MΩ was used from a Millipore academic system. Transparent polystyrene 96-well microplates were purchased from Greiner BioOne, Germany. All glassware used in the synthesis of NPs was thoroughly cleaned with an aqua regia solution and then silanized by immersion in a 2% dimethyldichlorosilane solution of chloroform. Glass slides were purchased from Menzel Glaser, Germany. Synthesis of NPs. Three different NP compositions were used in this study: pure silver, pure gold, and a gold/silver alloy composition. The alloy composition, which corresponded to the optimum λres in order to match to the Ru-complex absorption, has already been established in a previous publication,4 and this is the composition used here. Gold NPs, 15 ( 1 nm in diameter (Figure 1A), were prepared by reducing hydrogen tetrachloroaurate with sodium citrate following the method of Turkevich et al.9 In brief, 0.01 wt % hydrogen tetrachloroaurate dissolved in 100 mL of deionized water was heated to boiling and 2 mL of 1 wt % sodium citrate was added with rapid stirring. The solution was refluxed for a further 20 min in which time the solution turned from a light blue (nucleation) to a burgundy color. The solution was centrifuged at 9000 rpm for 2 h, doubly concentrated in deionized water, and stored in a polystyrene bottle at 4 °C. The final concentration of gold particles was estimated to be 6 × 1012 particles/mL, and the suspension exhibited an absorption band at 520 nm with a full width at half-maximum of 40 nm. Gold/silVer alloy NPs, 45 nm ( 7 nm in diameter (Figure 1B), were prepared using a two-step procedure. In the first step, small (20 nm in diameter) alloy NPs were prepared by the coreduction of silver nitrate and hydrogen tetrachloroaurate using sodium citrate.10 In brief, a solution of 20.3 mmol of silver nitrate and 5.08 mmol of hydrogen tetrachloroaurate in 95 mL of deionized water was raised to boiling and 2 mL of 1 wt % sodium citrate was added with rapid stirring. The solution was stirred for 30 min during which time the solution turned from a light blue to a light red to a dark yellow color. The final concentration of colloid was estimated to be 3.2 × 1010 particles/mL with a molar ratio of four silver atoms to every gold atom. In the second step, small alloy particles were used to seed the growth of larger NPs. Citrate-reduced gold and silver atoms precipitate onto the small NPs in preference to the formation of new nucleating sites. In brief, 23.5 mmol of silver nitrate and 5.88 mmol of hydrogen tetrachloroaurate were dissolved in 75 mL of deionized water and heated to boiling. To this solution was added simultaneously 50 mL of boiling seed NP solution and 2.5 mL of 1 wt % sodium citrate with rapid stirring. The solution was stirred for a further 30 min during which time the colloid darkened slightly in color. The solution was centrifuged at 4500 rpm for 2 h, doubly concentrated in deionized water, and stored in a polystyrene bottle at 4 °C. The (9) Turkevich, J.; Stevenson, P. C.; Hiller, J. Discuss. Faraday Soc. 1951, 58, 55. (10) Link, S.; Wang, Z. L.; El-Sayed, M. A. J. Phys. Chem. B 1999, 103, 3529.

Figure 1. Transmission electron micrographs of (A) gold (L ) 15 ( 1 nm), (B) gold/silver alloy (L ) 47 ( 5 nm), and (C) silver (L ) 60 ( 11 nm) nanoparticles.

resultant colloid had an estimated concentration of 2.4 × 1010 particles/mL and exhibited a plasmon absorption peak at 441 nm. Pure silVer NPs, 60 nm ( 10 nm in diameter (Figure 1C), were prepared by reducing silver nitrate with sodium citrate in the presence of aniline.11 Briefly, 4 mL of AgNO3 (0.02 M) and 4 mL of aniline (0.02 M) were dissolved in 104 mL of deionized water and purged with nitrogen. The solution was heated to boiling, and 8 mL of 1 wt % sodium citrate was added with rapid stirring. The solution was refluxed for 30 min during which time the solution changed color from clear to dark brown. The solution was centrifuged at 4500 rpm for 2 h and doubly concentrated in deionized water. The colloid was stored in a polystyrene bottle at 4 °C. The final concentration of colloid was estimated to be 1.6 × 1013 particles/mL, and the suspension exhibited a plasmon absorption peak at 440 nm. Layer-by-Layer Assembly of PELs. Transparent microwell plates with a polystyrene substrate were used as a convenient experimental platform. To activate each well for deposition of (11) Tan, Y.; Li, Y. F.; Zhu, D. J. Colloid Interface Sci. 2003, 258, 244.

Plasmonic Enhancement of Fluorescence polyelectrolyte, the plate was oxygen-plasma-treated for 5 min under vacuum using a Harrick plasma chamber (model PDC-200). In the first instance, five preliminary layers of polyelectrolyte were deposited onto each well to generate a uniform surface of amine groups for attachment of NPs. Five is the minimum number of preliminary layers required; any less leads to a surface with incomplete polyelectrolyte coverage and mixed charge. In brief, solutions of PEI, PSS, and PAH were prepared at 2 mg mL-1 in 0.5 M NaCl aqueous solution and dissolved by sonication at room temperature for 25 min. Polyelectrolyte solutions were deposited into each well at volumes of 150 µL with incubation times of 15 min using the following sequence PEI/PSS/PAH/PSS/PAH. The wells were rinsed with deionized water before the addition of each new PEL layer. It should be noted that PELs are highly versatile and can be coated onto both plastic and glass surfaces and, after the preliminary layers, provide a generic binding chemistry which is substrate-independent. Deposition of NPs. Each NP solution was deposited into 16 wells of the microtiter plate at well volumes of 150 µL. The plate was sealed with Parafilm to prevent evaporation and incubated for 12 h. Two rows of wells were coated with NPs of each type: gold/silver alloy NPs, pure gold NPs, and pure silver NPs. After rinsing, absorption measurements were performed under dry conditions and with 150 µL of water added. Each set of 16 NP-coated wells was further divided into four sets of wells, and a series of (PAH/PSS)x spacer layers was deposited on top using the same procedure as that for the deposition of the preliminary layers, where x ) 0, 2, 6, and 12 bilayers. Finally, solutions were deposited into each well at volumes of 150 µL with a concentration of 2.8 × 10-5 M ruthenium complex, 2.8 × 10-4 M PAH, and 0.5 M NaCl. Preparation and Deposition of Ru-PEL Conjugate. The Ru-PEL conjugate was prepared by mixing 1 × 10-4 M PAH dissolved in 10 mL of sodium hydrogen carbonate buffer at pH of 9.2 with 0.5 × 10-5 M Ru-ITC dissolved in 5 mL of DMF for 2 h. The Ru-PEL conjugate was recovered by centrifugation at 4000 rpm using a 3000MW centrifugal device from Millipore, Ireland. The ruthenium PAH solution (Ru-PEL) bound electrostatically to the PSS layer. The Ru-PEL solution was aged for 15 min, and then each well was rinsed thoroughly with deionized water. The quantum efficiency of the Ru-PEL complex was calculated to be 0.031 with reference to pure tris(2,2′-bipyridyl)dichlororuthenium(II) hexahydrate with a literature value of 0.042.12 The extinction coefficient determined using absorption spectroscopy was 12 335 M-1 cm-2. The excitation and emission maxima for the Ru-PEL were 464 and 606 nm, respectively. Experiments were also performed in a reverse configuration where the Ru-PEL was deposited directly over the polyelectrolyte preliminary layers. A series of polyelectrolyte spacer layers were then deposited over the Ru-PEL layer to the same thicknesses as the standard configuration. Finally, 60 nm silver NPs were deposited over the top polyelectrolyte layer using the method described above. The reverse configuration was used to determine the enhancement of fluorescence without a change in surface area. Characterization Techniques. Prior to coating onto the microplates, the UV-vis extinction spectra of the as-synthesized NP colloids were measured with a Cary 50 scan UV-vis spectrophotometer (Varian Ltd.) in transmission mode. Atomic force microscopy (AFM) measurements were performed on a “Dimensions 3100 AFM” from Digital Instruments. Analysis was performed in tapping mode using silicon tips purchased from Veeco. AFM images were analyzed using freeware software WSxM from Nanotec Electronica. The thickness of the PEL layers was measured on a white light ellipsometer (DeltaPsi 2, Jobin Yvon, Horiba). The data were acquired in reflection mode with an angle of incidence of 60°. The measurements were performed over the wavelengths of 300-800 nm with 5 nm steps. The acquisition time for each wavelength was 1 s, and the data were averaged over five measurements. A program supplied with the ellipsometer was used to fit the layer parameters (thickness, refractive index) so that the theoretically predicted ellipsometric coefficients (12) van Houten, J.; Watts, J. J. Am. Chem. Soc. 1976, 98, 4853.

Langmuir, Vol. 24, No. 19, 2008 11263

Figure 2. UV-vis absorbance of pure silver NPs (O), gold/silver alloy NPs (dashes), and pure gold NPs (solid line) coated onto polyelectrolyteactivated microplate wells analyzed at room temperature in air. Fortyeight wells were examined in total, 16 for each NP.

∆, Ψ correspond to the measured coefficients ∆, Ψ. Each polyelectrolyte bilayer was determined to be approximately 3 ( 0.25 nm thick (see the Supporting Information). Transmission electron microscopy (TEM) micrographs were obtained using a Hitachi 7000 transmission electron microscope operated at 100 kV. Images were captured digitally using a Megaview 2 CCD camera. Specimens were prepared by dropping aqueous solutions of the NPs onto a Formvar carbon-coated copper grid. Fluorescence excitation measurements were performed on a Safire (Tecan) microplate reader. The emission wavelength was set to match the peak of the emission band of the Ru-PEL at 610 nm. The excitation wavelength was varied from 350 to 550 nm to cover both the LSPR of the NPs and the absorption band of the dye. Self-quenching of the dye was not expected because of the large Stokes shift.

Results and Discussion 1. Reproducibility of NP Deposition. Extinction spectra of NP-coated microplate wells are shown in Figure 2. Two extinction bands were observed for pure silver NPs in air at 375 and 415 nm, respectively, with corresponding absorbance maxima of 0.351 and 0.357 AU. The second band relates to the dipole plasmon resonance.13 From previous work14 it is assumed that the first band, at 375 nm, derives from the quadrupole resonance in the NP. An important observation from this figure is the low standard deviation in extinction maxima across the 16 wells analyzed, measured to be of less than 3%, which highlights the high reproducibility of NP deposition from well to well. During subsequent washing steps and analysis there is a slight shift of the dipole plasmon resonance band from 415 to 420 nm. For gold/silver alloy NPs, the dipole resonance band was observed at 430 nm, with extinction maxima of 0.228 AU and a standard deviation of ∼2.6%. For pure gold NPs a weak band was observed at 520 nm with absorbance of 0.067 and a standard deviation of ∼1.5%. These results compare favorably with work reported on the deposition of gold NPs onto glass substrates from colloidal suspensions, where a standard deviation of 2.8% was observed.15 The intensity of the extinction maxima of the NP-coated wells is dependent on the fractional surface coverage of NPs bound to the surface. An AFM image of silver NPs coated onto a PELcoated glass slide with a similar extinction spectrum to the wells studied herein is shown in Figure 3. Several AFM pictures were taken at different positions on the surface, and the same results (13) Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles; Wiley-VCH: Berlin, 1998. (14) Malynych, S.; Chumanov, G. J. Am. Chem. Soc. 2003, 125, 2896. (15) Grabar, K. C.; Smith, P. C.; Musick, M. D.; Davis, J. A.; Walter, D. G.; Jackson, M. A.; Guthrie, A. P.; Natan, M. J. J. Am. Chem. Soc. 1996, 118, 1148.

11264 Langmuir, Vol. 24, No. 19, 2008

Nooney et al.

Figure 3. AFM image of pure silver NPs examined using tapping mode with 4.2 × 109 particles/cm2 and 12% total surface coverage.

Figure 4. Excitation fluorescence spectra of a layer of fluorescent dye, Ru-PEL, deposited over pure silver NPs with polyelectrolyte bilayers placed between the dye and NPs to regulate the degree of overlap of the electromagnetic field (symbols: PEL only, (dashed line); pure Ru-PEL, solid thin line; one bilayer, O; one bilayers, ∆; four bilayers, solid thick line; eight bilayers, +).

Table 1. Diameter and Corresponding λres for Each Metal NPa metal/alloy

diameter (nm)

λres/absorbance* (nm)

silver gold/silver alloy gold Ru-PEL

60 ( 10 45 ( 7 15 ( 1

415 430 520 412-470*

a The λres were calculated from the average absorbance over 16 wells for each NP. Measurements were performed on PEL-coated microplates in an air medium. The broad absorbance peak of the Ru-PEL dissolved in water is also given.

were obtained. We estimated 4.2 × 109 particles/cm2 with a 12% total surface coverage. This is comparable with previous results for the deposition of 13 nm diameter gold NPs onto an organosilane surface with a total surface coverage of 13%.15 The NPs appear larger in the image as deconvolution was not carried out to correct for the tip size. The average NP center-to-center distance for our silver NP coated slide is 154 nm, and assuming an average particle size of 60 nm, this implies an interparticle separation distance of 94 nm. The NPs are bound to the surface via either electrostatic interaction between the negatively charged NPs and the positively charged PEL or the formation of a dative bond to the lone pair of electrons on the amine of the PEL. Electrostatic repulsion between adjacent NPs is the most likely reason why more NPs do not occupy the interstitial spaces. 2. Distance Dependence of Fluorescence Enhancement. The differing optical properties of the three metal NP types used in this work are shown in Table 1. For pure silver and silver/gold alloy NPs, the λres values are close to the absorption peak of the Ru-PEL. The diameters of the pure silver and silver/gold alloy NPs were also chosen for maximization of plasmon-enhanced fluorescence. The λres of pure gold NPs is far way from the absorption peak of the Ru-PEL and was used here as a negative control. The excitation spectra of Ru-PEL deposited onto pure silver NP-coated microplate wells and the dependence on the spacer distance between dye and NP is shown in Figure 4. Ru-PEL coated onto negatively charged PEL preliminary layers without NPs (thin solid line in the figure) gave a highly reproducible fluorescence maximum of 6398 ( 2.3% units at an excitation wavelength of 460 nm. At this wavelength, there is considerable overlap with λres of the NPs, which constitutes the optimum condition for plasmonic enhancement. As stated in the Introduction, a dye in the proximity of a NP senses an altered EM field, and its fluorescence properties change. The enhancement of

fluorescence is a product of the enhancement of excitation and of emission. In order to maximize the enhancement of excitation, we matched the LSPR wavelength of the NP to the dye excitation band. However, it is likely that both excitation and emission enhancements contribute to the overall enhancement factor, with excitation enhancement being the dominant contributor. We have not attempted to separate the relative enhancement contributions from these two processes as we are unable to make assumptions regarding the behavior of the dye nonradiative rates in the presence of the NPs. From theory and experiment a dye very close to the surface of a metal NP is expected to experience quenching from nonradiative energy transfer back into the metal.7 However, this effect drops off very quickly. After deposition of Ru-PEL directly on to the negatively charged NPs, the fluorescence signal increased significantly to 40 591 ( 1.6% (shown as circles in Figure 4). The Ru-PEL in direct contact with the metal surface shows enhanced fluorescence for two possible reasons. First, the Ru-PEL has an intrinsic thickness of approximately 1.5 nm, and we assume that the fluorophores are distributed evenly throughout the polyelectrolyte layer. Fluorophores furthest from the surface are less likely to be quenched. Second, the Ru-PEL coats the entire substrate surface, and we only have a metal surface coverage of 12%. Therefore, one can expect a considerable number of fluorophores in the interstitial spaces between the NPs to be enhanced. On the addition of a single bilayer of PEL (3 nm in thickness) the signal increased to its maximum value (60 736 ( 3.5%). This increase is a combination of the reduced quenching effect and modified EM field and therefore represents an optimum NP to fluorophore spacer distance for this twodimensional detection system. The relative enhancement, Ren, of fluorescence of Ru-PEL arising from the proximity of metal NPs was calculated using the following equation:

Ren )

FRu-PEL,NP - FNP FRu-PEL - FPEL

where FRu-PEL,NP is the enhancement of Ru-PEL in close proximity to NPs, FNP is the background fluorescence of NPs, FRu-PEL is the fluorescence of Ru-PEL only, and FPEL is the background fluorescence of the polyelectrolyte layers. With the use of this equation, the relative enhancement of fluorescence of the Ru-NP system compared to that without NPs is approximately 16 times. On increasing the PEL spacer layer to four bilayers (12 nm in thickness), for the work presented here,

Plasmonic Enhancement of Fluorescence

Figure 5. Relative enhancement of fluorescence of Ru-PEL coated over silver (b), alloy (9), and gold (2) NPs with polyelectrolyte bilayers placed between the dye and NPs to regulate the degree of overlap of the electric field (the dashed line is an enhancement factor of 1).

the fluorescence dropped but not significantly (53 540 ( 1%). On extending the PEL spacer layers to eight (16 nm in thickness) the emission fluorescence signal dropped significantly (23 991 ( 1.5%). However, there was still a significant enhancement of 5.5. Malicka et al. studied the distance dependence of fluorophores, Cy3 and Cy5, in close proximity to silver metal islands approximately 500 nm by 100 nm in size.16 Bovine serum albumin conjugated to biotin with an avidin linker was used as a spacer between the dye and silver particles where each spacer layer was approximately 9 nm in diameter. On increasing the number of spacer layers from one to six, Malicka et al. observed a significant decrease in enhancement from 12 to 2 times but no maximum in enhancement. The minimum spacer layer distance is 3 times greater than the diameter of a PEL bilayer and explains why no maximum for enhancement against spacer layer thickness was observed. Ray et al. studied the distance dependence of enhancement of fluorescence of sulforhodamine B over silver island films using PEL bilayers to control the spacer distance.17 A maximum enhancement at a distance of between 8 and 9 nm from the silver island film surface was observed. Silver metal island are larger than the NPs used herein, and sulforhodamine B has different optical properties to Ru-PEL. The position of the maximum enhancement is dependent on physical properties and should be determined independently for each experimental setup used. A comparison of the relative enhancement of fluorescence of Ru-PEL coated over pure silver NPs, gold/silver alloy, and gold NPs plotted against PEL spacer layer thickness is shown in Figure 5. The largest enhancement is for pure silver with a polyelectrolyte spacer thickness of 3 nm as discussed above. The gold/silver alloy enhancement versus spacer distance follows a similar pattern peaking at 3 nm, but the enhancement value of 3 is significantly less. In previous work we studied the enhancement of fluorescence of ruthenium(II) tris(4,7-diphenyl1,10-phenanthroline) dichloride (Ru(dpp)32+) which was electrostatically linked to a silica-coated gold/silver alloy NP in the liquid phase.4 The alloy NP used in that study had exactly the same molar ratio of silver to gold (4:1) as the NP used in this work and a similar particle size (54 nm in diameter NP plus a 4.8 nm silica shell). The maximum enhancement of 3.5 reported from that study is very similar to the value obtained here. The lower enhancement factor obtained for the alloy NPs compared (16) Malicka, J.; Gryczynski, I.; Gryczynski, Z.; Lakowicz, J. R. Anal. Biochem. 2003, 315, 57. (17) Ray, K.; Badugu, R.; Lakowicz, J. R. Chem. Mater. 2007, 19, 5902.

Langmuir, Vol. 24, No. 19, 2008 11265

to pure silver NPs is attributed to the different dielectric properties of gold and silver. The imaginary part of the dielectric constant is higher for gold/silver than for silver18 resulting in a broader and less intense plasmon resonance peak for the alloy compared to pure silver.13 The pure gold NPs exhibit no enhancement in this investigation. This is related in part to the mismatch between λres of the NPs and the excitation wavelength of 460 nm for the enhancement experiments and also to the small NP.4 In evaluating the maximum enhancement achieved, it is useful to compare these data with the work of others. In a study using rhodamine red in the proximity of silver NPs on glass substrates, Pan et al. observed a relative enhancement varying between 6 and 20 times depending on the excitation wavelength used.19 With the use of electron beam lithography Pompa et al. fabricated highly ordered gold nanopatterns onto planar substrates.20 They spin-coated a blend of CdSe/ZnS nanocrystals dispersed in poly(methyl methacrylate) over this surfaces and obtained enhancement factors varying between 21 and 33 depending on the type of nanopattern. Both of these independent studies showed higher enhancement factors than the work presented herein; however, the degree of enhancement is highly dependent on the physical properties of both fluorophore and NP which are different in all three cases. A proportion of the observed enhancement is due to the increase in dye coverage arising from the increased surface area in the presence of the NPs.21 On the basis of 12% coverage of NPs, the increase in surface area is 48%. Hence, the measured enhancement factor of 16 is renormalized by a factor of 1.48 to yield a value of 11. In order to eliminate the surface area effect, the experimental fluorescence enhancement was also measured in the reverse configuration where the NPs were placed on top of the dye layer. This yielded a maximum enhancement of 12 at a separation distance of 3 nm, which agrees well with the corrected value above. Hence, we deduce that our measured fluorescence enhancements are largely due to the plasmonic effect. An increase in sensitivity and a reduction in LOD are key goals in the enhancement of diagnostic devices. In this investigation, both sensitivity and LOD were evaluated for experimental configurations with and without NPs. In order to measure sensitivity, fluorescence experiments were carried out at different dye concentrations using a fresh batch of silver NPs (see Figure 6). For all experiments, reproducible results (standard deviation