Contentserver (1).pdf

Plasmonics (2016) 11:753–761 DOI 10.1007/s11468-015-0106-0

Probing the Localized Surface Plasmon Field of a Gold Nanoparticle-Based Fibre Optic Biosensor Reshma Bharadwaj 1,4 & Suparna Mukherji 1,2 & Soumyo Mukherji 1,3,4

Received: 17 July 2015 / Accepted: 28 September 2015 / Published online: 7 October 2015 # Springer Science+Business Media New York 2015

Abstract Gold nanoparticles (GNP) have been used in a variety of localized surface plasmon resonance (LSPR)-based optical sensor systems and in a variety of forms, such as colloidal suspensions, immobilized GNP on flat surfaces or optical fibres. A key parameter affecting the sensitivity of these systems is the effective depth of penetration of the surface plasmons. This study aims to determine the plasmon penetration depth in the case of an immobilized GNP-based LSPR optical biosensor. The optical biosensor used for experimentation is a U-bend fibre optic probe of 200-μm core diameter and 1.5-mm bend diameter on which GNP is immobilized. Formation of multilayered nanostructures on the immobilized GNP was used to investigate the field of the localized surface plasmons. Two multilayered nanostructures were explored in this study, viz. a polyelectrolyte multilayer formed by layerby-layer (LBL) deposition of oppositely charged polyelectrolytes and an immunoglobulin G (IgG) multilayer formed through sequential immobilization of two mutually specific antibodies. Measurement of LSPR absorbance change with Electronic supplementary material The online version of this article (doi:10.1007/s11468-015-0106-0) contains supplementary material, which is available to authorized users. * Soumyo Mukherji [email protected] 1

Centre for Research in Nanotechnology and Science, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

2

Centre for Environment Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

3

Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

4

Centre of Excellence in Nanoelectronics, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

deposition of each analyte layer was used to determine the plasmon penetration depth (dP) of the LSPR biosensor. Probing the plasmon field with an IgG multilayer gave rise to at least twofold higher dP compared to dP obtained from the polyelectrolyte multilayer. The effect of GNP size was also studied, and GNP of three diameters, viz. 18, 36 and 45 nm, were used. The 36-nm-diameter GNP exhibited the highest dP. The outcomes of this study may provide leads for optimization of LSPR-based sensors for various biosensing applications. Keywords Plasmon penetration depth . Localized surface plasmon resonance . U-bend fibre optic probe . Gold nanoparticles . Polyelectrolyte multilayer . Immunoglobulin G (IgG) layer

Introduction The localized surface plasmon resonance (LSPR) property exhibited by nanoparticles of noble metals like gold and silver is manifested by the presence of an extinction band in the UVvisible wavelength range. This extinction band is influenced by the size, shape and composition of the nanoparticles and refractive index (RI) of the surrounding dielectric medium [1]. An analyte binding to the nanoparticle surface causes small deviations in the local RI of the surrounding medium. This results in changes in the extinction property of the nanoparticle which is characterized by a shift in the plasmon resonance wavelength and/or change in optical absorption. However, the RI sensitivity of the nanoparticles is distance dependent [2]. The localized surface plasmon (LSP) electric field originates from the surface of metal nanoparticles and decays exponentially into the surrounding medium with a penetration depth, dP. While the dP of propagating surface plasmons or surface

754

plasmon resonance (SPR)-based sensors can be obtained from Maxwell’s equations, less information is available on the decay length properties of localized surface plasmon fields [3]. The dP of SPR-based sensors is known to be in the range of 150– 200 nm. In the case of LSPR-based sensors, the value of dP depends on the size, shape and composition of the nanoparticle as well as the medium properties. The penetration depth of the evanescent field also plays an important role in determining the sensitivity and analytical volume of the LSPR sensor [4]. The response (R) of an SPR/LSPR sensor in terms of shift in peak absorbance wavelength (or intensity) to varying analyte layer thickness can be described by the general Eq. (1) [5]:

 R ¼ mΔη 1−exp −d=d p ð1Þ where m is the sensor sensitivity in terms of wavelength shift or intensity change per RIU (RIU=refractive index unit), Δη is refractive index difference between the analyte and the bulk medium, d is the thickness of the analyte layer (which is varied) and dP is the plasmon penetration depth. Biological analytes vary in size (nucleic acids ∼1–2 nm, antibodies ∼10–15 nm, cells 1–5 μm). If dP ≫d, then the nanoparticlebased biosensor will be vulnerable to noise due to changes in the bulk RI, non-specific adsorbates, etc. In contrast, if dP ≪d, the nanoparticle will not elicit a sufficiently large response when the analyte binds to its surface (depending on the shape and steric hindrance properties of the analyte). Thus, dP must be optimized according to the dimension of the analyte [6–8]. Adam et al. attempted to measure the decay length of a surface plasmon evanescent field of silver film (SPR sensor) using a photon scanning tunnelling microscope (PSTM) [9]. The microscope consisted of a tapered optical fibre which was brought close to the sample surface. The evanescent field was coupled with the optical fibre using the phenomenon of frustrated total internal reflection. The fibre was moved vertically from the surface to follow the exponential decay of intensity as a function of distance. Experimental measurements revealed a decay length of ∼200 nm for silver. Many attempts have also been made to experimentally determine the localized surface plasmon penetration depth of LSPR sensors, by studying the decay in the LSPR response with sequential deposition of material layers. The layer materials used in these studies include oppositely charged polyelectrolytes [4, 10–12], polymers like polymethyl methacrylate (PMMA) [13], alkanethiols [8] and aluminium oxide films [14]. In a more recent study, Read et al. suggested that the plasmon field should be probed with a biomaterial instead of polymers or oxide layers [15]. Since the plasmon penetration depth also depends on the polarizability and RI of the material in the plasmon field, polyelectrolytes may not provide an accurate estimation of dP for an LSPR biosensor. Read at al. described a technique using two mutually specific antibodies to build a multilayered immunoglobulin (IgG) structure which was termed as Bbiostack^.

Plasmonics (2016) 11:753–761

In our research group, a gold nanoparticle (GNP)-based Ubend LSPR fibre optic biosensor has been demonstrated for immunosensing applications [16]. The fibre optic sensor was fabricated by decladding a small portion of the core of a 200-μm-diameter multimode optical fibre. The fibre was bent into a U-shaped structure using a butane torch, and the exposed core region was coated with GNP. Probe geometry and surface coverage of GNP have been optimized in previous studies [17]. In this study, two methods to determine the plasmon penetration depth of a GNP-based U-bend fibre optic sensor have been explored. The first method uses deposition of oppositely charged polyelectrolyte multilayers. The polyelectrolytes (PE) u s e d f o r l a y e r- b y - l a y e r d e p o s i t i o n a r e c a t i o n i c poly(allylamine) hydrochloride (PAH) and anionic poly(sodium 4-styrenesulfonate) (PSS). In the second method, the plasmon field was probed using two mutually specific antibodies, viz. rabbit anti-goat IgG and goat anti-rabbit IgG. A schematic representation of the two methods is shown in Fig. 1. Results obtained from both methods were compared. Additionally, the effect of nanoparticle size was also taken into consideration in this report and three nanoparticle diameters, viz. 18, 36 and 45 nm, were included in our studies. This is probably one of the first studies wherein the effective plasmon penetration depth (dP) of an LSPR fibre optic probe has been quantified. Knowledge on dP of the LSPR probe is essential for a better understanding of sensor probe

Fig. 1 Schematic representation of methods for determining the plasmon penetration depth of a gold nanoparticle-coated U-bend fibre probe. a Deposition of oppositely charged polyelectrolytes (PAH and PSS). b Sequential immobilization of mutually specific antibodies (goat antirabbit IgG and rabbit anti-goat IgG). Relative sizes of nanoparticles and biomolecules are not drawn to scale

Plasmonics (2016) 11:753–761

behaviour and its limitations and thereby facilitates optimization of sensor structure for better sensitivity.

Materials and Methods Materials Trisodium citrate (molecular weight (MW)=294 g/mol), 3aminopropyltriethoxy silane (3-APTES or APTES), phosphate-buffered saline (PBS) tablets, human immunoglobulin G (HIgG), goat anti-human immunoglobulin G (GaHIgG), bovine serum albumin (BSA) and glacial acetic acid were obtained from Merck Chemicals (India). Absolute ethanol was procured from Commercial Alcohols (Canada). Gold(III) chloride solution [30 % wt. of HAuCl4 (MW= 3 3 9 . 8 g / m o l ) i n d i l u t e h y d r o c h l o r i c a c i d ] , 11 mercaptoundecanoic acid, 1-ethyl-3-(3dimethylaminopropyl)-carbodiimide) (EDC), Nhydroxysuccinimide (NHS), ethanolamine, rabbit anti-goat immunoglobulin G (RaGIgG), goat anti-rabbit immunoglobulin G (GaRIgG), poly(allylamine) hydrochloride (PAH, MW∼50,000) and poly(sodium 4-styrenesulfonate) (PSS, MW∼70,000) were purchased from Sigma-Aldrich (USA). De-ionized (DI) water was used as solvent for the reagents. Synthesis of Gold Nanoparticles GNP of three different diameters, viz. 18, 36 and 45 nm, were synthesized using Turkevich’s citrate reduction method [18]. Freshly prepared aqua regia solution was used for cleaning the glassware used for synthesis (note: aqua regia solution is highly corrosive and must be handled only with appropriate personal protection equipment). The glassware were thoroughly rinsed with DI water. HAuCl4 was reduced with the help of trisodium citrate, and the particle diameter was varied by controlling the molar ratio of citrate to gold. The molar ratios (citrate to gold) taken were 3.5, 2.6 and 2.2, and GNP of mean diameters and standard deviations (mean ± SD) 18 nm (±1.8 nm), 36 nm (±4.4 nm) and 45 nm (±6.4 nm) were obtained respectively. The GNP were stored at 4 °C. The synthesized nanoparticles were characterized using UV-visible spectrometry and transmission electron microscopy (TEM). The TEM images are shown in Fig. S1 (online resources). The synthesis results have been summarized in Table S2 (online resources). The UV-visible absorption spectrum is shown in Fig. S3 (online resources). Fabrication of a U-Bend Fibre Optic Probe Multimode, high-OH silica core optical fibres of 200-μm core diameter and numerical aperture (NA) 0.37 were purchased from CeramOptec®, USA. The U-bend probes were

755

fabricated by bending a partially decladded straight optical fibre while being heated by a butane flame. The procedure to fabricate U-bend fibre probes is described in Fig. S4 (online resources). The bend diameters of the U-bend fibres were measured using an optical microscope (Carl Zeiss®). Optical microscopy images of the U-bend probes are provided in Figs. S5, S6 and S7 (online resources). Previous studies have shown the highest RI sensitivity in the case of U-bend LSPR probes having bend diameters in the range 1.4–1.6 mm [16]. Therefore, probes having bend diameters in the range 1.4– 1.6 mm were chosen in this study. The fibre probes were cleaned according to methods described in our previous studies [19]. The U-bend fibre probe surface was silanized with the help of 3-APTES to obtain amine functional groups on the surface. 3-APTES (1 % (v/v)) was prepared in a 5:2 ratio of ethanol/acetic acid solution, and the U-bend probes were immersed in this solution for 5 min followed by rinsing with absolute ethanol. Experimental Setup The U-bend fibre probe was mounted on a glass slide with an adhesive tape. The glass slide was fixed on a stage. A broadband white LED (450–700 nm) was used as a light source, and a microscope objective with NA=0.65 was used to couple light from the LED to the fibre. One end of the fibre was aligned with the beam spot with the help of X-Y positioners (Newport®, USA). The other end of the fibre was coupled to a fibre optic spectrometer (USB 4000, Ocean Optics®, USA) using bare fibre adapters and SMA 905 connectors (Thorlabs®, USA). The schematic representation of the experimental setup is shown in Fig. S8 (online resources). Binding of Gold Nanoparticles The binding of GNP to silanized U-bend probes was observed in real time. The baseline absorbance spectrum was obtained in DI water. The fibre probe was dipped in GNP solution, and change of absorbance versus time at the LSPR wavelength was recorded. When the absorbance value reached ∼2.5, further binding of GNP to the probe surface was stopped. This value of LSPR absorbance was found to be optimum since any further increase in absorbance led to a poor signal-tonoise ratio [17]. The time-resolved absorbance change at the LSPR wavelength and absorbance spectrum changes due to binding of the 18-, 36- and 45-nm-diameter GNP are shown in Fig. S9 (online resources). The magnitude of light absorption by the GNP is directly proportional to the size of the GNP [20, 21]. Thus, the absorbance change measured by the U-bend fibre probe due to binding of GNP is dependent on the size and surface density (nanoparticles/μm2) of GNP. Therefore, higher nanoparticle density (∼250 nanoparticles/μm2) was required for the smaller (∼18 nm) diameter GNP to achieve an

756

Plasmonics (2016) 11:753–761

absorbance change of 2.5, whereas comparatively lower nanoparticle density (∼150–200 nanoparticles/μm2) was required for the 36- and 45-nm-diameter GNP. Representative scanning electron microscopy (SEM) images of the immobilized GNP are shown in Fig. S10 (online resources). Deposition of Polyelectrolyte Layers Equimolar solutions (3 μM) of PAH and PSS were prepared in 1 M NaCl solution. The GNP-coated fibres were first immersed in PAH solution for 15 min, and the probe was rinsed with 1 M NaCl solution, followed by rinsing with DI water. The probe was later dipped in PSS solution. The negatively charged PSS molecules attach themselves to the positively charged PAH layer. This step was followed by the same rinsing protocol as stated above for PAH. The procedure was repeated 12 times to obtain alternating layers of PAH and PSS. A total of 24 polymer layers were deposited. This procedure was followed for all the three gold nanoparticle sizes considered in this study. The absorbance changes were recorded during the deposition of each layer. Antibody Immobilization on GNP-Coated Probes The GNP-coated fibres were dipped in 10 mM solution of 11mercaptoundecanoic acid (11-MUA) prepared in absolute ethanol for 12 h. This helps in the formation of carboxyl groups on the GNP surface. The carboxylated GNP surface was activated with 0.4 M EDC:0.1 M NHS in a 1:1 ratio by volume (prepared in DI water) for 1 h. The fibre probes were rinsed with DI water. The probes were incubated in 100 μg/ml solution of rabbit anti-goat immunoglobulin G (RaGIgG), prepared in PBS (pH 7.4). The real-time absorbance change at 540 nm due to binding of RaGIgG is shown in Fig. 2b. The residual sites were blocked by incubating the fibres in 2 mg/ml solution of BSA (prepared in PBS, pH 7.4) for 30 min. The unreacted NHS esters were deactivated by incubating the antibody-immobilized probe in 1 M ethanolamine solution prepared in borate buffer (pH 8.5) for 30 min.

Results and Discussion Polyelectrolyte Multilayer Deposition of the polyelectrolyte (PE) layers was monitored in real time by measuring the absorbance change at peak wavelength as shown in Fig. 2a. For all three GNP sizes, the absorbance at the resonance wavelength increased with deposition of each PE layer as shown in Fig. 3. Increase in the absorbance starts to level off after about 19–20 layers. GNP of 36-nm mean diameter showed increase in absorbance even after 20

Fig. 2 a Time-resolved absorbance changes at 540 nm due to layer-bylayer deposition of polyelectrolytes: PAH and PSS (first 10 layers), starting with PAH. b Binding of rabbit anti-goat IgG to the EDC/NHSactivated LSPR probe. The GNP size considered here is 36 nm

polymer layers. However, the 18- and 45-nm-diameter GNP showed saturation after 19–20 layers. The binding kinetics of the PE films were quite different from receptor-ligand binding kinetics of biomolecules (shown in Fig. 2b). Equation (1) could not be fitted to the results (Fig. 3) obtained from PE films. Moreover, the absorbance response signal became noisy as more number of polymer layers were deposited. The thickness of a PAH-PSS layer in 1 M NaCl is reported to be ∼4 nm [22]. These observations indicate that the LSPR-based U-bend fibre probe can detect changes in RI up to a distance of 40 nm irrespective of particle size. The 36-nm-diameter GNP showed a slightly higher penetration depth compared to the other two GNP. The obtained values of dP for the 36-nm-diameter (dP ∼44 nm) and the 45nm-diameter (dP ∼40 nm) GNP were in close agreement with results obtained by other groups [4]. However, dP (∼40 nm) for the 18-nm-diameter GNP was higher than those obtained from previous investigations [4, 13]. The RI of the polyelectrolytes PAH and PSS is reported to be 1.468 and 1.484 respectively. The RI of the PE multilayer is

Plasmonics (2016) 11:753–761

Fig. 3 Absorbance change at peak wavelength versus number of polyelectrolyte layers for GNP of diameter 18, 36 and 45 nm

a weighted average of the RI of the individual layers and has been experimentally found to be around 1.477 [23]. It is evident from the high absorbance values seen in our experiments that the RI of the multilayered PE film is quite high. The higher RI of PSS compared to PAH can also be clearly seen from Fig. 2a, where the higher absorbance change was obtained during PSS deposition vis-à-vis PAH deposition. The plasmons of the GNP are excited evanescently using the optical fibre. Moreover, the U-bend geometry of the fibre probe provides enhanced penetration depth of the fibre optic evanescent field to excite the plasmons more efficiently. The RI of the silica core is 1.46 (at 532 nm). It can be speculated that high effective RI of the PE multilayer may have resulted in unguided modes of the optical excitation field. As a result, only slight absorbance change was observed when a higher number of PE layers >20 were deposited (Fig. 3). This prompted us to investigate the other factors (apart from GNP size) which might perturb the plasmon penetration depth.

757

loosely bound analytes. The probe was subjected to RaGIgG followed by a similar rinsing protocol, and subsequently alternating layers of GaRIgG-RaGIgG were deposited. Figures 4, 5 and 6 show typical real-time absorbance change and the LSPR absorbance spectrum change due to sequential IgG binding, for 18, 36 and 45 nm GNP respectively. The time-resolved absorbance changes from Figs. 4, 5 and 6 reveal that the U-bend LSPR sensor showed increase in absorbance with each antibody layer. Absorbance response after each binding step of the IgG layer started decreasing after six layers, which may be expected due to exponential decay in the evanescent field strength of the LSPR field. In few cases, e.g., Fig. 6b, a saturated or steady-state response was not achieved by the LSPR probe. A steady-state response from the LSPR biosensor is dependent on the availability of active binding sites on the probe surface [16]. The antibodies used in this study were polyclonal antibodies. Therefore, it was quite possible that multiple epitopes were recognized by the antibodies and large numbers of analyte binding sites were available at each binding step [15]. However, a poor signal-to-

Formation of an IgG Multilayer Probing the plasmon field with a protein like an antibody (RI of pure antibody ∼1.42) is expected to give a different value of plasmon penetration depth, due to lower RI of proteins. Furthermore, since this approach emulates a biosensor environment more accurately, it might provide us a more realistic picture of the plasmon penetration depth of an LSPR biosensor [15]. Rabbit anti-goat immunoglobulin G (RaGIgG) immobilized LSPR probes were used for this study. Goat anti-rabbit immunoglobulin G (GaRIgG) solution of 100 μg/ml concentration was prepared in PBS (pH 7.4). The baseline spectrum was obtained in PBS and the probe was incubated in GaRIgG solution to obtain a primary response. The absorbance changes at the LSPR wavelengths were observed in real time. Once saturation was obtained, the probe was rinsed in PBS to remove

Fig. 4 U-bend probe coated with the 18-nm-diameter GNP: a timeresolved absorbance changes at 530 nm with binding of each IgG layer and b LSPR absorbance spectrum change due to sequential binding of each IgG layer (5-point moving average used to smoothen data)

758

Plasmonics (2016) 11:753–761

Fig. 5 U-bend probe coated with the 36-nm-diameter GNP: a timeresolved absorbance changes at 535 nm with binding of each IgG layer and b LSPR absorbance spectrum change due to sequential binding of each IgG layer (5-point moving average used to smoothen data)

noise ratio (SNR) was observed from the probes, for higher number (>10) of IgG layers. The number of IgG layers was therefore limited to 9–10 layers in this study, in order to maintain consistent SNR from the LSPR probes. To determine the penetration depth (dP) from these results, Eq. (1) was modified as follows: y ¼ að1−expð−x=bÞÞ

ð2Þ

y is the LSPR biosensor response in terms of absorbance, a is a constant, x is taken as the IgG layer number and hence b gives the decay step. Non-linear regression (on replicates) was performed using the Levenberg-Marquardt algorithm (statistical package: Statistica version 8). Equation (2) was fitted to the experimentally obtained absorbance values (from various probes) with increasing IgG layers, and a and b were determined (Fig. 7). The results have been summarized in Table 1. The molecular weight of an IgG molecule is 150 kD, and the molecular dimensions of an intact IgG molecule (determined by X-ray diffraction) is reported to be around 15 nm× 10 nm×4 nm [24]. We have taken the antibody length as 12 nm in our calculations. A total of 11 antibody layers were deposited (including the immobilization layer), but maximum

Fig. 6 U-bend probe coated with the 45-nm-diameter GNP: a timeresolved absorbance changes at 540 nm with binding of each IgG layer and b LSPR absorbance spectrum change due to sequential binding of each IgG layer (unprocessed data)

absorbance response was observed in the first six layers which correspond to a minimum dP of ∼80 nm. This value of dP was at least two times the value obtained by using polyelectrolyte layers. The highest value of dP was obtained for the 36-nmdiameter GNP which showed absorbance response for up to eight layers corresponding to a minimum dP of ∼100 nm. From Table 1, we can infer that 36 nm GNP exhibited ∼1.6 times higher decay length than 18 nm GNP. The penetration depths obtained from the three GNP were statistically analysed using one-way ANOVA. It was concluded that a significant difference exists between the dP of 36 nm GNP and 18 nm GNP (P value=0.001) as well as between the dP of 36 nm GNP and 45 nm GNP (P value=0.025). However, there appears to be no significant difference between the dP obtained from 45 and 18 nm GNP (P value >0.5). The mutual specificity and polyclonal nature of the antibody pair used in this study was essential for maintaining a consistent step length [15]. A similar set of experiments were performed with another polyclonal antigen-antibody pair, viz. human immunoglobulin G (HIgG) and goat anti-human

Plasmonics (2016) 11:753–761

759 Table 1 GNP

Comparison of plasmon penetration depths of different-sized

GNP size (nm)

Decay step b

Confidence interval for b (level of confidence 95 %)

Plasmon penetration depth dP =12(b+1) (nm)

18 36

5.6 9.3

4.7–6.6 7.3–11.3

79.2 (±6.0) 123.6 (±12.0)

45

6.2

4.5–7.9

86.4 (±9.6)

Values of localized surface plasmon penetration depth determined from the decay step b obtained for the three different GNP diameters. Error is given in standard error (n≥4)

concentrations of HIgG and GaHIgG were used. The response from GaHIgG was higher than that from HIgG, implying that an equal number of epitopes were not available for binding on each antibody surface. Thus, low mutual specificity of the IgG pair was indicated, and therefore, the step length of the IgG molecules and hence penetration depth cannot be ascertained in this case. Surface Plasmon Resonance Studies In order to corroborate the method of estimating dP using Eq. (2), a similar set of sequential IgG binding experiments were performed on a commercially available surface plasmon resonance (SPR) system. Since analytical equations for calculating the plasmon decay length are available for propagating surface plasmons, Eq. (3) was used to determine the theoretical SPR decay length and the experimentally obtained penetration depth was compared with the theoretical value. The experiments were performed on a dual-channel SPR instrument (Autolab Twingle®, 670-nm operating wavelength). The gold coated glass disc was modified with 11-MUA and activated with EDC/NHS using methods similar to that used

Fig. 7 Absorbance change versus IgG layer number (replicate experimental data shown) and non-linear regression results for a the 18nm-diameter GNP (n=4), b the 36-nm-diameter GNP (n=5) and c the 45nm-diameter GNP (n=4). Errors in a and b are given in standard error

immunoglobulin G (GaHIgG). The mutual specificity of HIgG and GaHIgG was unknown. HIgG was immobilized on the 36 nm diameter GNP coated fibre using methods followed in the BAntibody Immobilization on GNP-Coated Probes^ section. The HIgG-immobilized probe was subjected to alternating layers of GaHIgG and HIgG. The real-time absorbance change at 540 nm and LSPR spectral change are shown in Figs. 8a and S11 (online resources) respectively. It can be observed from inset Fig. 8b that the absorbance change for each IgG binding step was unequal even though equal

Fig. 8 U-bend probe coated with the 36-nm-diameter GNP: a time resolved absorbance changes at 540 nm due to binding of alternating IgG layers of GaHIgG and HIgG and inset graph b shows response in the initial 3.5 h (G: GaHIgG, H: HIgG)

760

for the GNP-coated fibre probes. The gold disc was immobilized with rabbit anti-goat IgG by injecting 10 μg/ml of RaGIgG in channel 1 and channel 2 of the SPR sensor. The residual sites were blocked using BSA. A sensogram showing angular shift in the SPR wavelength during various immobilization steps is plotted in Fig. S12 (online resources). Channel 1 of the SPR sensor was subjected to alternating layers of GaRIgG and RaGIgG of 10 μg/ml concentration. Channel 2 was used as reference and PBS (pH 7.4) was circulated in the channel. The real-time angular shift with deposition of each IgG layer is shown in Fig. 9a. The angular shift in the plasmon wavelength did not increase significantly after about 13 layers of antibody deposition. By fitting Eq. (2) to the experimental data in Fig. 9b, a decay step of 13.1 was obtained. This corresponds to a practical plasmon decay length of around 168 nm (including the immobilization layer). The decay length ld of an SPR sensor is given by Eq. (3) [25]: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   λ  εgold þ η2  ð3Þ ld ¼  η4 2π  where εgold is the real part of the dielectric permittivity of gold at that wavelength and η is the refractive index of the antibody. The value for dielectric permittivity (εgold) at 670 nm was

Fig. 9 a Real-time shift in the SPR angle with sequential binding of IgG layers and b angular value versus IgG layer number

Plasmonics (2016) 11:753–761

taken as −14.4 [26], and η was taken as 1.42 [27]. The value of ld was obtained as 186 nm. This matches fairly well with the experimentally obtained value, after taking into account inhomogeneities in the immobilization layer. Thus, we can assume that the step length of the IgG molecules has been reasonably preserved in the multilayered immunoassay in the SPR biosensor. We can further conjecture that the multilayered IgG film in the LSPR biosensor is also similarly structured and the experimentally obtained value of LSP dP is a close approximation of the true value.

Conclusions In this study, we have evaluated two methods to determine the LSP penetration depth (dP) of a GNP-based fibre optic biosensor. We observed significant differences in the value of dP obtained from the two methods. The minimum value of dP when the plasmon field was probed by polyelectrolytes was 40 nm, whereas when the plasmon field was probed using IgG, the minimum dP was 80 nm. This may be attributed to the possibility that the material (polyelectrolyte or protein) significantly influences the decay length of the plasmon evanescent field. It is more pertinent to use a protein like an immunoglobulin to probe the plasmon field as it conforms to a typical biosensor environment more accurately. The size of GNP was also found to affect dP, and the 36nm-diameter GNP was found to have higher dP than the 18nm-diameter GNP. The 45-nm-diameter GNP however did not give higher dP than the 36-nm-diameter GNP. The 45nm-diameter GNP solution exhibited the highest absorption bandwidth (Table S2, online resources). Larger diameter (>40 nm) GNP are known to exhibit significant line broadening of the plasmon resonance peak due to multipolar excitations [20, 28]. Excess broadening of the plasmon peak may cause LSPR absorbance change, caused by RI change, to be indistinguishable. This may have resulted in lower (effective) dP of the 45-nm-diameter GNP compared to the 36-nmdiameter GNP. The higher dP of the 36-nm-diameter GNP was also in close agreement with studies reported by Nath and Chilkoti, wherein the 39-nm-diameter GNP showed the highest penetration depth [4]. The dP was determined from the experimental data using empirical equations, which were adapted from established equations, for modelling the response from SPR and LSPR biosensors. The approach was validated with experiments performed using a commercially available SPR instrument. The LSPR fibre biosensor can thus be tailored according to the biological analyte of interest. For detection of small proteins, GNP of a smaller diameter (18 nm or less) can be used. Larger diameter GNP (∼36 nm) may be used for detecting analytes as large as viruses (size∼100 nm). Future studies may be directed to study the influence of other parameters

Plasmonics (2016) 11:753–761

such as fibre core diameter, nanoparticle shape, composition, etc. on the plasmon penetration depth. To the best of our knowledge, this is the first comprehensive study investigating the plasmon penetration depth of an LSPR-based fibre optic biosensor. Acknowledgments The Centre for Research in Nanotechnology and Science (CRNTS) and the Sophisticated Analytical Instrument Facility (SAIF) of IIT Bombay are acknowledged for providing TEM and SEM facilities. The Department of Biotechnology (DBT, Govt. of India) is acknowledged for financial support. Reshma Bharadwaj thanks Ms. Sutapa Chandra for her help in setting up the SPR experiments.

References 1. 2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

Mayer KM, Hafner JH (2011) Localized surface plasmon resonance sensors. Chem Rev 111:3828–3857. doi:10.1021/cr100313v Willets KA, Van Duyne RP (2007) Localized surface plasmon resonance spectroscopy and sensing. Annu Rev Phys Chem 58:267– 297. doi:10.1146/annurev.physchem.58.032806.104607 Jung LS, Jung LS, Campbell CT et al (1998) Quantitative interpretation of the response of surface plasmon resonance sensors to adsorbed films. Langmuir 14:5636–5648. doi:10.1021/la971228b Nath N, Chilkoti A (2004) Label-free biosensing by surface plasmon resonance of nanoparticles on glass: optimization of nanoparticle size. Anal Chem 76:5370–5378. doi:10.1021/ac049741z Malinsky MD, Kelly KL, Schatz GC, Van Duyne RP (2001) Chain length dependence and sensing capabilities of the localized surface plasmon resonance of silver nanoparticles chemically modified with alkanethiol self-assembled monolayers. J Am Chem Soc 123:1471–1482. doi:10.1021/ja003312a Kedem O, Vaskevich A, Rubinstein I (2011) Improved sensitivity of localized surface plasmon resonance transducers using reflection measurements. J Phys Chem Lett 2:1223–1226. doi:10.1021/ jz200482f Kedem O, Vaskevich A, Rubinstein I (2014) Critical issues in localized plasmon sensing. J Phys Chem C 118:8227–8244. doi:10. 1021/jp409954s Haes AJ, Zou S, Schatz GC, Van Duyne RP (2004) A nanoscale optical biosensor: the long range distance dependence of the localized surface plasmon resonance of noble metal nanoparticles. J Phys Chem B 108:109–116. doi:10.1021/jp0361327 Adam PMP, Salomon L, de Fornel F, Goudonnet JP (1993) Determination of the spatial extension of the surface-plasmon evanescent field of a silver film with a photon scanning tunneling microscope. Phys Rev B 48:2680–2683. doi:10.1103/PhysRevB. 48.2680 Schmitt J, Mächtle P, Eck D et al (1999) Preparation and optical properties of colloidal gold monolayers. Langmuir 15:3256–3266. doi:10.1021/la981078k Marinakos S, Chen S, Chilkoti A (2007) Plasmonic detection of a model analyte in serum by a gold nanorod sensor. Anal Chem 79: 5278–5283

761 12.

Tian L, Chen E, Gandra N et al (2012) Gold nanorods as plasmonic nanotransducers: distance-dependent refractive index sensitivity. Langmuir 28:17435–17442. doi:10.1021/la3034534 13. Okamoto T, Yamaguchi I, Kobayashi T (2000) Local plasmon sensor with gold colloid monolayers deposited upon glass substrates. Opt Lett 25:372–374. doi:10.1364/OL.25.000372 14. Whitney AV, Elam JW, Zou S et al (2005) Localized surface plasmon resonance nanosensor: a high-resolution distance-dependence study using atomic layer deposition. J Phys Chem B 109:20522– 20528. doi:10.1021/jp0540656 15. Read T, Olkhov RV, Shaw AM (2013) Measurement of the localised plasmon penetration depth for gold nanoparticles using a noninvasive bio-stacking method. Phys Chem Chem Phys 15:6122– 6127. doi:10.1039/c3cp50758k 16. Sai VVR, Kundu T, Mukherji S (2009) Novel U-bent fiber optic probe for localized surface plasmon resonance based biosensor. Biosens Bioelectron 24:2804–2809. doi:10.1016/j.bios.2009.02. 007 17. Satija J, Punjabi NS, Sai VVR, Mukherji S (2013) Optimal design for U-bent fiber-optic LSPR sensor probes. Plasmonics 9:251–260. doi:10.1007/s11468-013-9618-7 18. Turkevich J, Stevenson PC, Hillier J (1951) A study of the nucleation and growth processes in the synthesis of colloidal gold. Discuss Faraday Soc 11:55. doi:10.1039/df9511100055 19. Bharadwaj R, Mukherji S (2014) Gold nanoparticle coated U-bend fibre optic probe for localized surface plasmon resonance based detection of explosive vapours. Sensors Actuators B Chem 192: 804–811. doi:10.1016/j.snb.2013.11.026 20. Jain PK, Lee KS, El-Sayed IH, El-Sayed MA (2006) Calculated absorption and scattering properties of gold nanoparticles of different size, shape, and composition: applications in biological imaging and biomedicine. J Phys Chem B 110:7238–7248. doi:10.1021/ jp057170o 21. Lee K, El-Sayed MA (2006) Gold and silver nanoparticles in sensing and imaging: sensitivity of plasmon response to size, shape, and metal composition. J Phys Chem B 110:19220–19225. doi:10. 1021/jp062536y 22. Büscher K, Graf K, Ahrens H, Helm CA (2002) Influence of adsorption conditions on the structure of polyelectrolyte multilayers. Langmuir 18:3585–3591. doi:10.1021/la011682m 23. Wong JE, Rehfeldt F, Hänni P et al (2004) Swelling behavior of polyelectrolyte multilayers in saturated water vapor. Macromolecules 37:7285–7289. doi:10.1021/ma0351930 24. Sarma VR, Silverton EW, Davies DR, Terry WD (1971) The threedimensional structure at 6 a resolution of a human gamma Gl immunoglobulin molecule. J Biol Chem 25. Barnes WL (2006) Surface plasmon–polariton length scales: a route to sub-wavelength optics. J Opt A Pure Appl Opt 8:S87–S93. doi: 10.1088/1464-4258/8/4/S06 26. Johnson PB, Christy RW (1972) Optical constants of the noble metals. Phys Rev B 6:4370–4379. doi:10.1103/PhysRevB.6.4370 27. Vörös J (2004) The density and refractive index of adsorbing protein layers. Biophys J 87:553–561. doi:10.1529/biophysj.103. 030072 28. Kvasnička P, Homola J (2008) Optical sensors based on spectroscopy of localized surface plasmons on metallic nanoparticles: sensitivity considerations. Biointerphases 3:FD4. doi:10.1116/1. 2994687

Plasmonics is a copyright of Springer, 2016. All Rights Reserved.