On the potential application of surface plasmon-based core-shell particles to study blood functional parameters

We investigated the application of gold nanoshell particles as optical sensors and contrast agents to study the blood functional parameters. Gold nanoshell particles with a core size of 1 μm that exhibit two prominent plasmonic peaks at 750 and 830 nm were developed. The peaks correspond to the wavelengths typically used to study the oxygen saturation of the blood. The plasmonic properties of gold nanoshells in media with various refractive indices were studied. Glucose samples with concentrations 0, 15, and 20%w/v in water were used. The 750 and 830 nm plasmonic peaks exhibit peak wavelength shifts of 63.77 ± 49.40 nm and 31.18 ± 20.94 nm per unit refractive index change. The optical properties of blood samples mixed with gold nanoshells were also measured. The optical absorption of blood samples increased by 7% at these wavelengths in the presence of the nanoshells. The plasmonic peaks at 750 and 830 nm showed a 3.57 ± 0.56 and 1.44 ± 0.55 percentage variation in absorbance for a 1% change in hematocrit. The enhanced optical absorption at these wavelengths suggests that these particles are effective optical sensors/contrast agents for multimodal optical and photoacoustic sensing and imaging.


Introduction
Surface plasmon (SP) represents a collective oscillation of free electrons at the metal-dielectric interface [1,2]. Metal colloids have several beneficial optical properties, including strong optical absorption, enhanced light scattering, and a large and fast third-order nonlinear optical susceptibility [3][4][5][6][7][8][9][10]. The excitation of SP depends on the size of the metal particle and the local microenvironment. For example, a solid Au nanoparticle of size 60 nm in water exhibits maximum surface plasmon excitation compared to other sizes [11]. However, applications of solid Au nanosphere particles are limited to the visible regime of the electromagnetic spectrum since its peak plasmonic absorption lies around 530 nm, which can be tuned to approximately 30 nm by varying either the size or dielectric property of the surrounding medium [12]. Metal nanoshell or core-shell (polymer core-metal shell) (CS) is another class of gold nanoparticles that exhibits a strong optical absorption spectrum in the near-infrared region of the electromagnetic spectrum [13]. It finds applications in bioimaging as the tissue has minimal optical absorption in the IR regime which paves the way for deep interrogation of biological tissues. Nanoshells can exhibit a strong optical extinction peak between 400 nm to 20 μm depending upon the ratio of the shell thickness to the core size [9]. It could find applications in other imaging/sensing approaches such as Surface-Enhanced Raman spectroscopy or scattering (SERS), photoacoustics, ultrasound, reflectance confocal microscopy, and OCT as the dielectric contrast of core-shell particles (CS) can be engineered and tailored to improve its optical and acoustic scattering coefficient [5,14,15]. Plasmonic metal nanoparticles efficiently convert light into heat when excited at the plasmon resonance wavelength. The enhanced polarizability at the plasmon resonance of these compositions could also be used in therapeutic applications [16][17][18]. It has been reported that CS (gold nanoshell) can exhibit a millionfold greater optical extinction compared to indocyanine green (ICG) which has been used for decades clinically for monitoring the circulation, liver function, and vascular imaging in ophthalmology and neurosurgery [14,[19][20][21][22][23][24]. CS particles can, in principle, be used as a substitute for ICG for these clinical applications. The optical extinction of gold nanoparticles is attributed mainly to the excitation of SP at the metal-dielectric interface. The excited SP would decay through radiative relaxation (as light) and non-radiative de-excitations (as heat) [4,[25][26][27][28][29]. As Au nanoparticles are well known for transferring light to heat, CS could be used as photoacoustic contrast agents in the infrared regime. Recently optoacoustic systems have been used for deep tissue imaging as the imaging approach offers a scalable spatial resolution [30][31][32][33][34][35]. The underlying principle behind PA imaging is the measurement of pressure transients that are generated due to the thermo-elastic expansion of the medium resulting from the absorption of the incident optical irradiation. As the excited surface plasmon is very sensitive to its neighboring dielectric medium, it can be used for surface plasmon-based sensor applications [1,2,36]. Gold nanoshells have been used as a whole blood immunoassay for point care detection [37]. Gold nanoshell aggregates exhibited enhanced plasmon absorption due to inter-plasmonic coupling between gold nanoparticles within the same sphere and intra-plasmonic couplings between gold nanoparticles from neighboring spheres [3]. Our recent numerical investigations of the SPR properties of 300 nm polystyrene-gold nanoshells in various dielectric media showed that the plasmon peak shifted by 764 ± 13 nm per unit refractive index change (RIU) [38]. Therefore, these nanoshells could be used as sensors to detect changes in the refractive index of biofluids (such as blood) either in-vitro or in-vivo using approaches based on photoacoustics.
Here we investigate the application of 1 μm CS plasmonic particles for surface plasmon sensor (SPR) and imaging. Gold nanoshells of 1 μm size core are selected as they exhibit multiple plasmonic peaks in the IR regime due to the support of higher-order plasmonic modes. 1-μm CS plasmonic particles exhibit two broad peaks in the near-infrared regime at 750 nm and 830 nm, wavelengths that have been traditionally used to calculate blood oxygen saturation. The hemoglobin in blood exhibits relatively low absorption at 750 nm compared to 830 nm when oxygenated (and the opposite when deoxygenated). The absorption of light by the red blood cell (RBC) is mainly due to the high concentration of hemoglobin which could be either oxygenated (oxyhemoglobin) or deoxygenated (deoxyhemoglobin) depending on the oxygen saturation level (SO 2 ). Human whole blood consists of about 40%-45% erythrocytes and 55%-60% plasma. The refractive index of blood plasma varies between 1.358 and 1.344 in the visible wavelength [39][40][41][42]. The refractive index (RI) of the RBC depends on the concentration of hemoglobin (Hb) within the cell because the membrane has a RI value close to the plasma [43]. The hemoglobin concentration in a given volume of packed blood (the mean corpuscular hemoglobin concentration) is 320 to 360 g l −1 [39]. Hemoglobin solution with a concentration of 355.5 g l −1 exhibits RI between 1.413 to 1.438 for visible light [39]. The present study aims to explore the application of CS particles for blood saturation measurements by increasing the optical contrast at the above wavelengths of interest, as well as exploring the potential to detect changes in the surrounding medium.
Photoacoustic (PA) studies on Au nanoshells are not common. Most PA studies were based on Au shells with silica as the core [32,44]. So far, there have not been any studies on Au nanoshells with polystyrene cores for PA applications. Polystyrene Au nanoshells exhibit a plasmonic peak further in the NIR regime compared to silica Au nanoshells due to the high refractive index of polystyrene [36,45,46]. Additionally, high-frequency PA studies on 350 nm gold nanoshells showed almost half the PA value from a single RBC [47,48]. CS particle exhibited a PA amplitude of 1.27 ± 0.18 mV per fluence (mJ/cm 2 ) compared to 3.3 ± 0.3 mV per fluence (mJ/cm 2 ) for a single RBC in a photoacoustic system [47].
The SP excitation depends on the surrounding dielectric medium of the Au layer, which is described by the following dispersion relation of SP excitation, where ω is the angular frequency of the incident radiation, k o is the free space wavevector, and ε m and ε d are the dielectric functions of the metal and the dielectric medium, respectively. The plasmonic property of gold nanoshell is investigated by changing the surrounding dielectric medium. Metallic nanoparticles find numerous applications as sensors, such as the non-enzymatic electrocatalytic glucose sensor [49]. Metallic nanoparticles of various surface morphologies (nanospheres, nanorods, nanofibers, nanopores, and nanoflakes) have been used as sensors, but gold nanoshells have not been examined. Gold nanoshells could be used as they exhibit enhanced plasmonic excitation due to inter and intra-plasmonic couplings. Gold nanoshells supported two plasmon modes due to surface plasmon excitation at the core-metal and metal-surrounding dielectric medium interfaces. The peak wavelengths of these surface plasmon modes depend on the size of the core-shell particle. Our earlier study on 300 nm PS spheres showed it supported the plasmon mode at the wavelengths 670 and 760 nm respectively [38], These plasmon modes can be shifted to a higher wavelength range by increasing the size of the core-shell particle. In our study, we used 1 micron CS particles because they supported two plasmonic modes around 750 nm and 830 nm, the wavelengths typically used for the blood oxygen saturation study. The study on the SP excitation in 300 nm PS spheres showed it supported the plasmon mode at the wavelengths 670 and 760 nm well below the isosbestic point (805 nm) used for the blood oxygen saturation study [38]. Blood has relatively low absorption at the IR wavelength compared to the visible wavelength. The addition of CS particles would enhance the absorption property of the blood. The plasmonic properties of the prepared 1-micron CS particles were investigated before CS particles were used for the blood study. The plasmonic properties were studied by changing the surrounding dielectric medium of CS particles as an SP excitation of the gold nanoparticles depends on the neighbor dielectric medium (equation (1)). In this work, we prepare media of various dielectric properties to study the plasmonic behavior of gold nanoshells. The media were prepared by dissolving glucose of multiple concentrations in water. Then the application of CS as an SPR sensor to study blood functional parameters was investigated.

Procurement of blood
Blood is collected by netCAD (Vancouver, Canada), the research division of Canadian Blood Services, under protocol 2013-001 which involves standard Canadian Blood Services collection and testing procedures of whole blood. Delivery was made overnight at 4°C, with continuous monitoring during shipment to ensure no temperature deviations occurred. The research ethics boards of both Toronto Metropolitan University and Canadian Blood Services have approved this procedure.

Blood sample preparation
The guidelines on blood handling were followed by the recommendations of the International Society for Clinical Hemorheology and the European Society for Clinical Hemorheology and Microcirculation [50]. The blood was centrifuged at room temperature at 2000 × g for 6 min to separate the plasma and its buffy coat. Isotonic phosphate-buffered saline (PBS) was used to wash RBCs twice. The centrifuged RBC was then dispersed in PBS at various hematocrit levels for the present studies.

Thiolation of polystyrene spheres (PS)
Thiolated PSs are prepared by treating carboxylated PS with AET in the presence of EDAC. AET is a thiol compound that, in reaction with a carboxyl group of polystyrene spheres, forms thiolate, which behaves as ligands in the presence of metal ions and forms a metal-thiolate complex. 1 gram of 1 μm size carboxylated PS is diluted in 1 ml of HPLC water. To this solution, 0.4 ml of 500 mM MES is added and stirred vigorously. MES has a pK a value of 6.15 at 20°C and acts as a buffering agent. PS procured from Polyscience, Inc. (Warrington, Pennsylvania, United States) is used for the present studies. The microspheres are available in aqueous suspension with a concentration of 2.5% (w/v). The amount of EDAC and AET used to thiolate PS depends on the surface charge density of the PS. The surface charges of PS range between 0.1 and 2.0 milliequivalent/gram. The Zeta potential of PS particles is −91.74 mV (table 1). EDAC of concentration 5 mM is used for the present studies. 5 ml of EDAC is added to the stirring solution. After half an hour of stirring, 15 mg of AET in 0.5 ml of HPLC is added, and the stirring is continued for another 3 h. The reaction mixture is centrifuged 4000 × g for 10 min to remove unreacted AET. The unreacted solution is decanted, and the precipitated supernatant is redispersed in 20 ml of HPLC water and centrifuged 4000 × g for 10 min. This step is repeated several times until the unreacted excess AET is removed completely.

Preparation of Au(THPC) nanoparticles
Au nanoparticles of size 10 nm are prepared by reducing chloroauric acid in the presence of THPC [36,51]. Au nanoparticles are prepared by successively charging water using NaOH, THPC, and chloroauric acid. First, 0.5 ml of 1M NaOH is added to 45 ml of HPLC water and vigorously stirred for 2 min, and then 12 μl THPC is added. After waiting for 10 min of vigorous stirring, 10 ml of 5 mM chloroauric acid is added swiftly. The stirring is continued until the reaction mixture becomes dark brown, indicating the formation of the Au nanoparticles.

Preparation of Au nanoshell
Au nanoparticle-tagged PS is prepared by stirring 2 ml of thiolated PS with 25 ml of prepared Au(THPC) nanoparticles. Thiols are basic and easily attract electrophilic Au nanoparticles. The Zeta potential of Au(THPC) particles is −33.75 mV (table 1). Au-tagged PS are separated from the unreacted excess Au(THPC) nanoparticles by centrifuging the reaction solution at 1000 × g. The unreacted Au(THPC) that remains as a suspension above the CS is decanted from the CFG tube, and the brown precipitate is redistributed in fresh 15 ml of HPLC water. This process of centrifuge and redispersion in 15 ml of HPLC is repeated several times until all the unreacted Au(THPC) is removed completely. This process involves sparse coverage of Au nanoparticles around PS. The shell growth is continued using gold hydroxide as a precursor [36,52,53]. Gold hydroxide is prepared by dissolving 0.05 g of potassium carbonate in vigorously stirring 185 ml of HPLC and followed by the addition of 15 ml of 5 mM of HAuCl4. The stirring is continued until the yellow colour of the stirring solution becomes colourless. Au shell growth on Au(THPC) tagged PS is initiated by adding as prepared 50 ml of gold hydroxide solution in 3 ml of Au(THPC) tagged PS solution. A mixture of 50 μl of formaldehyde and 10 μl of ammonium hydroxide is added to this reaction. This results in the formation of Au(K2CO3) nanoparticles which eventually started conjugating with Au(THPC)-PS. The stirring is continued until the black precipitate of CS is observed. CS is formed in 4-5 h. The stirring is stopped, and the resulting solution is centrifuged to 1000 × g to remove CS particles from the unreacted chemical remnants. The centrifugation and subsequent dispersing of the precipitate in 20 ml of water is repeated until the precipitate contains only CS particles. The separated CS particles are finally dispersed in 10ml of HPLC water and kept as a stock solution for the studies. Totally 6 samples are prepared by repeating this process.
The plasmonic properties of the core-shell particles are investigated by dispersing the particles in dielectric media with different indices of refraction. Dielectric media are obtained by dissolving glucose of known concentration in water. The glucose concentrations chosen for the present study are 0 w v −1 %, 15 w v −1 %, and 20 w v −1 %. The concentrations chosen represent blood sugar levels measured in diabetic patients to investigate potential glucose sensor applications. Humans typically have glucose concentrations in the range of 4.4 to 6.1 mmol l −1 (79.2 to 110 mg dl −1 or 0.079 to 0.11 w v −1 %). The concentration of the core-shell particles is kept constant for all measurements.
Three samples (A, B, and C) containing the mixture of CS and blood were prepared. The transfusion of a 'unit' of blood will result in a 3% increase in the hematocrit or 1 g dl −1 of hemoglobin. It has generally been accepted that a 3% change in hematocrit is equivalent to a 1-'unit' loss of blood [54]. Blood of hematocrits 1.6%, 0.8%, and 0.4% were considered to determine the sensitivity of the gold nanoshells to hematocrit variations. These samples (A, B, and C) were prepared by mixing 1 ml of blood with hematocrit levels of 1.6%, 0.8%, and 0.4% to 2 ml of stock solution of the prepared CS particles (concentration~1% (w/v)). After mixing, samples A, B, and C contain blood with hematocrit 0.53%, 0.27%, and 0.13%, respectively.

Optical characterization
All optical absorption spectra were acquired with a UV-vis spectrophotometer (PerkinElmer Waltham, Massachusetts, USA), and the structure of CS was investigated using a Tecnai G2 transmission electron microscope (TEM). The ImageJ software was used for the size estimation. An Abbe refractometer was used to measure the refractive index of the glucose medium. The solutions with glucose concentrations of 15 w v −1 % and 20 w v −1 % exhibit refractive indices 1.3535 and 1.3588 at 24°C. Mie scattering theory was used to investigate the optical properties of the core-shell particles [55,56]. Mie scattering simulations were performed  Figure 1 shows the representative TEM image of PS spheres. Figure 1(b) is the histogram of the PS particle size distribution. The average size of PS particles is 931.95 ± 38.28 nm (table 1). Figure 2(a) is the photograph of the synthesized AuTHPC and Au(K2CO3) nanoparticle solution in the cuvette used for dressing thiolated PS spheres. Figures 2(b) and (c) are the representative TEM images of AuTHPC and Au(K2CO3) nanoparticles and their corresponding size distribution histograms (figures 2(d) and (e)). Table 1 gives the average size of various particles used to prepare CS particles.    Figure 4 represents the core-shell particles with completed Au nanoshells (the particle is covered by a layer of gold). The Au nanoshell is obtained by reducing Au(K2CO3) in the presence of Au(THPC) dressed PSs as the procedure described in the Materials and Methods section. The average size of obtained CS particles is 964 ± 54 nm. Figure 5 represents the optical extinction spectrum of Au(THPC) nanoparticles (black plot), PS spheres (blue plot), Au(THPC) dressed PS (red plot), and Au(K2CO3) nanoparticles (green plot).      represent the Mie scattering calculations of CS particles with Au shell thicknesses of 45 nm and 50 nm. The Mie scattering calculations for the shell thickness 45 and 50 nm showed the best fit to the experimentally obtained extinction data. The average thickness of the Au shell measured from the TEM image is 32 nm (table 2). The variation in the thickness between Mie calculation and experimental data could be the coarse nature of the nanoshell. The experimental CS particles exhibit two peaks at 750 nm and 830 nm, respectively (arrows in figure 6). The plasmonic peak at 750 and 850 nm are probably due to the SP excitation at the core-metal and metal-water interfaces [38]. The 830 nm peak was not prominent, probably due to weak coupling of the outermost plasmonic mode at the Au shell-water interface. The differences between the theoretical and experimental optical extinction features are likely due to the coarse nature of the Au nanoshell and the potential formation of nanoshell aggregation after preparation [37,46].

Investigation of SPR sensitivity as a function of refractive index
The SPR sensitivity of the core-shell is investigated by monitoring the plasmon peak wavelength shift as a function of the refractive index change of the surrounding medium. The first sample contains CS in water. The second sample is a mixture of CS and 15 w v −1 % glucose in water. The third sample contains CS and 20 w v −1 % glucose in water. All three samples have similar CS concentrations. Figure 7 shows the measured optical extinction of CS particles in water, 15 w v −1 % and 20 w v −1 % glucose. The extinction spectra show a dominant peak around 750 nm and a relatively weak peak around 830 nm. The peak at 830 nm can be observed by differentiating the spectrum around 850 nm.  Figure 7(c) represents the obtained differential spectra. These plasmon peaks are found to be broadened and red-shifted in media with increased glucose concentration. We investigate the plasmonic properties of each peak separately in the following sections.
3.1.1. Blue-end plasmon peak (750 nm) In figure 7(b), all spectra are normalized with respect to 702 nm to observe better the peak shift in the 750 nm plasmon peaks. The plasmon peak is red-shifted and broadened. The FWHM of the peak and its peak shift are obtained by the multi-Gaussian fit (blue curves) on the experimental data (black data points) [58,59]. The convolution (the red curve) of three Gaussian fits provided the best fit to the experimental value. Figures 8(a)-(c) represent the multi-Gaussian fits on the 750 nm plasmon peak for different glucose concentrations. The solid red line represents the convolution of three Gaussian fits (blue lines) superimposed on the experimental data. The FWHM of the 750 nm peak was obtained by adding the FWHM of three Gaussian fits. The peak shift was obtained from the first Gaussian fit around 740 nm. Figure 9(a) represents the plot of plasmon peak shift as a function of glucose concentration. The solid line represents the linear fit on the experimental data (solid circles). It exhibits the slope of 0.099 ± 0.06 nm peak wavelength shift per w/v% of glucose concentration. The R-Squared value (R2-value) of the fit is 0.81. The plasmon peak is also broadened with increased glucose concentration.  The linear fit has a slope of 1.26 ± 0.39 nm per w/v% of glucose concentration. The broadening of the plasmon peak was due to plasmon losses associated with the changes in the surrounding dielectric medium.
We also investigated the dependence of SP on the surrounding dielectric medium. An Abbe refractometer is used to measure the refractive index of the glucose medium. The glucose media of concentrations 15 w v −1 % and 20 w v −1 % exhibit refractive indices of 1.3535 and 1.3588 at 24°C. Figure 9(b) represents the 750 nm plasmon peak shift as a function of the refractive index. The plasmon peak is red-shifted as the medium's refractive index (RI) increases. Plasmon shifts of 63.77 ± 49.40 nm per refractive index unit (RIU) change are measured. The R-Squared value (R2-value) of the fit is 0.72. The plasmon peak also broadens with an increase in RI ( figure 10(b)). The broadening of the peak is obtained by summing the FWHM of three Gaussian fits. The plasmon peak 750 nm exhibited a broadening of about 813.20 ± 28.77 nm per refractive index unit (RIU) change. The R-Squared value (R2-value) of the fit is 0.91. The wider error bars are likely due to the precipitation and aggregation of core-shell particles in the glucose medium because of less density. Since the 750 nm peak corresponds to the plasmon mode excited inside a CS at the core and metal interface, the aggregation of CS particles would influence a 750 nm plasmon excitation. This aggregation of CS particles can be prevented by stabilizing CS particles in the liquid medium [60]. 3.1.2. SP sensitivity of red plasmon peak (830 nm) Next, we investigate the refractive index sensitivity of the 830 nm plasmonic peak. As the peak is not well pronounced, the first derivative is used to analyze the peak as in [36]. Figure 7(c) shows the first derivative extinction spectra of the 830 nm plasmonic peak. The peak wavelength is obtained from the Gaussian fit. A red shift in the plasmon peak with increasing glucose concentration is observed. Figure 11 represents the 830 nm plasmon resonance shift as a function of (a) glucose concentration and (b) refractive index. The CS particles exhibited plasmon resonance shifts of 0.046 ± 0.02 nm (the R-Squared value (R2-value) of the fit is 0.81) and 31.18 ± 20.94 nm (the R2-value is 0.72) per w/v % of glucose and RIU, respectively. The FWHM of the plasmon peak also increased with an increase in glucose concentration. The change in the FWHM of the plasmon peak is due to the dampening of the excited surface plasmon due to the surrounding dielectric medium. Figure 12 plots the FWHM of the red-end plasmon peak as a function of glucose concentration. A broadening of 0.34 ± 0.27 nm per (w/v %) glucose concentration (the R2-value of the fit is 0.43) is measured. Figure 12(b) represents the study of FWHM of plasmon peak as a function of refractive index. A resonance peak shift of 217.56 ± 196.1 nm RIU −1 (the R2-value of the fit is 0.3) is measured.

Application in blood
The application of CS to study the functional properties of blood was investigated. Hirch et al used gold/silica nanoshell (CS) as an immunoassay to detect analytes within complex biological media [37]. They could detect a sub-nanogram concentration of immunoglobulins per milliliter in saline, serum, and whole blood within 10-20 min.
The CS of size 1 μm was used as it exhibited broad plasmonic peaks around 750 nm and 830 nm where the hemoglobin within an RBC possesses significant absorption variation depending on its oxygenation state. The oxyhemoglobin has significantly lower absorption at 750 nm (518 cm −1 M −1 ) than deoxyhemoglobin (1405.24 cm −1 M −1 ), while at 850 nm, oxyhemoglobin exhibits slightly higher absorption (1058 cm −1 M −1 ) than deoxyhemoglobin (691.32 cm −1 M −1 ) [61]. The absorption of light by the red blood cell (RBC) is primarily determined by its hemoglobin concentration. The refractive index (RI) of the RBC depends on its density and optical absorption. For example, a hemoglobin solution with a concentration of 355.5 g l −1 has a RI between 1.413 and 1.438 for visible light [39]. The blood functional properties were investigated based on the refractive index properties of the whole blood.
Three blood samples of hematocrit 1.6%, 0.8%, and 0.4% were prepared in phosphate buffer solution (PBS), as described in section 2.2. Then three mixtures (A, B, and C) containing CS and blood were prepared.   figure 13(b)) exhibits a nearly exponential decay extinction profile from 600 nm to 900 nm. The addition of CS increased the extinction spectrum of the RBCs with strong peaks at around 750 nm and 850 nm. Figure 13(b) provides a closer view of normalized extinction spectra around 800 nm. The plasmon peak shifts with an increase in hematocrit. The peak absorbance is obtained by a Gaussian fit on the experimental data.
3.2.1. Blue-end plasmon peak (750 nm) The graph in figure 14(a) shows the blue-end plasmon peak shift as a function of hematocrit. The plasmon peak exhibits wavelength shifts of −15.35 ± 2.82 nm per unit hematocrit change (the R2-value of the fit is 0.88). Figure 14(b) represents the percent variation in absorbance as a function of hematocrit. The absorbance was found to decrease with increasing hematocrit. The solid dot represents the average of 5 measurements. The linear fit (solid blue line) to the experimental data gives a 1.44 ± 0.55 percentage decrease in absorbance per hematocrit change (the R2-value of the fit is 0.78).

3.2.2.
Red-end plasmon peak (830 nm) Next, the dependence of 830 nm of the plasmonic peak on hematocrit was investigated. The first derivative in the extinction spectrum is used to find the plasmon peak shift.  Figure 15(a) shows the red-end plasmon peak wavelength plot versus hematocrit. CS particles exhibited a red shift in peak with increased hematocrit of about 1.7 ± 0.27 nm per hematocrit change (the R2-value of the fit is 0.93). The plasmon peak was found to decrease with the increase in hematocrit. Figure 15(b) represents the plot of percentage variation in absorbance as a function of hematocrit. CS particles exhibited a 3.57 ± 0.56 percentage decrease in absorbance per increase in hematocrit (the R2-value of the fit is 0.95).
The measurements presented suggest that CS particles could be used in-vitro to monitor changes in hematocrit. The hematocrit is a parameter monitored by physicians/surgeons in patients post-transfusion. Hemorrhaging patients with the transfusion of a 'unit' of blood will result in a 3% increase in the hematocrit or 1 g dl −1 of hemoglobin. The study using glucose as a dielectric medium suggested that CS particles can also be used to determine blood glucose levels. It can be used in the non-enzymatic determination of glucose in the blood by an electrocatalytic method. The role of gold nanoparticles in this method is to enhance the conductivity of the electrode, which is influenced by the presence of glucose during measurement [49]. As gold nanoshells would provide enhanced electron conductivity than other metallic nano morphologies by inter and intra-plasmonic coupling due to being spherical in nature, it could be used in an electrocatalytic method for the non-enzymatic determination of glucose in the blood. The gold nanoshell used in the present study is in the liquid medium resulting in precipitation and aggregation during the measurement. This aggregation is observed in the less dense glucose medium but not in the blood solution. The measurement can be improved by stabilizing the CSs particles or depositing/immobilizing them on the solid substrate for any in vitro or lab-on-a-chip or SERS application [60]. In addition, as both plasmonic peaks, 750 nm, and 830 nm, are sensitive to refractive index change, it can be used to study the blood oxygen saturation as the RI of hemoglobin in the oxy state is higher than the deoxy state in the visible region. Faber et al measured the refractive index of oxygenated and deoxygenated hemoglobin solutions of concentration 93 g l −1 at 800 nm were 1.392 and 1.388, respectively [39,41]. Nienke Bosschaart et al have calculated theoretically the real refractive index of oxygenized blood (SO 2 > 98%) and deoxygenized blood (SO 2 = 0%) using the Kramer-Kronig relation which was 1.39260 and 1.387945, respectively [40]. Thus, the proposed CS can be used to enhance the sensitivity of absorption-based optical modalities for a more efficient way of measuring oxygen saturation and hematocrit change in vitro. Besides, the variation in hematocrit level would influence the glucose measurement [62]. The future study will be on the hematocrit influence on glucose measurement.
Microspheres also have potential applications in whispering gallery mode (WGM) sensors and superresolution imaging [63][64][65][66]. The WGM sensor supports a standing light mode at its periphery, which depends on the surrounding dielectric medium. The SP excited in the Au particle is sensitive to the surrounding dielectric medium (equation (1)), and a change in the neighboring dielectric medium would influence the excitation of the WGM. The addition of Au particles to the WGM microsphere would enhance its sensitivity. These particles would find potential applications for in vitro studies like a lab-on-a-chip and a substrate-based SERS application.

Conclusion
CS particles of size 1 μm showed two well-pronounced plasmonic peaks at 750 nm and 830 nm. We have investigated using the CS particles as possible SPR sensors for various applications. Two plasmonic peaks, at 750 nm and 830 nm, were considered for the present studies. Investigation of the plasmonic peaks showed shifts in the peak wavelength and broadening with changes in the surrounding medium's refractive index (RI). The plasmonic peaks at 750 nm and 830 nm exhibit peak wavelength shifts of 63.77 ± 49.40 nm and 31.18 ± 20.94 nm per unit RI change when glucose was used to change the RI of the medium. The peaks also broadened by 813.20 ± 28.77 nm and 217.56 ± 196.1 nm per unit RI change. Then the application of CS as an SPR sensor to study blood functional parameters was investigated by probing optical extinction coefficients at the wavelengths between 750 nm and 850 nm (typically used to measure blood oxygenation). Addition of the CS to blood enhanced optical contrast in this wavelength regime. The 830 nm plasmonic peak shows a 3.57 ± 0.56 percentage variation in absorbance per hematocrit change, whereas the 750 nm plasmonic peak shows a 1.44 ± 0.55 percentage variation in absorbance per hematocrit change. Hence, the gold nanoshell can be used in vitro to determine the hematocrit concentration of blood in patients suffering from hemorrhage. Gold nanoshells precipitate and aggregate in a less-dense medium. Further, the CS can, in principle, be extrinsically administered as a contrast agent to study the blood oxygen level in pre-clinical experiments.