Introduction

Surface plasmons (SPs) are electronic dilatational waves transmitting on a metal surface that are generated by the interaction between freely vibrating electrons and photons existing on that metal surface. A feature of the propagation of SP waves is that they are highly sensitive to the variation of refractive index of the dielectric adjacent to the metal. The most conventional approach to stimulating the excitation of surface plasmon is the Kretschmann configuration, in which a prism is coupled with a thin metal film [1]. SPR is greatly applied in biological science, gas detection, and environmental monitoring thanks to its many advantages such as smaller size, low-cost components, highly controllable, multiple channels and better reflectivity and sensitivity [2]. In order to improve the sensitivity of these kinds of sensors, over the past decade, many researchers have investigated SPR in aspects of, for example, improving the materials [3,4,5], changing the geometrical shapes [6, 7], or increasing the number of the channels [8, 9].

It is well known that graphene can be used to improve the sensitivity of conventional SPR structure. Graphene is a two-dimensional carbon material which has favorable strength, malleability, electrical conductivity, thermal conductivity, and optical properties and therefore, has been widely applied in biological science, electronic information, and aerospace [10,11,12]. There are several advantages for using graphene-based SPR substrate for sensing. First, they can induce large field enhancement at the substrate interface. Second, graphene has a large surface area, which makes it have better surface contact. Finally, graphene surface specific detection of aromatic compounds via polyimide (PI) layer force, which will help challenge experimental studies and other interactions with protein DNA in extremely dilute conditions [13,14,15]. The SPR sensor based on graphene-on-metal has also been researched in order to improve the sensing performance [16].

On the other hand, the transparent conductive ITO has also attracted great attention since it can be used as the functional material for modulators and SPR with the advantages of tunable permittivity, low capacitance, and high transmittance [17,18,19,20,21]. The transparency of ITO film can be as high as 95% for visible lights and 80% for infrared lights. As proven by Byun K M et al., ITO is able to enhance SPR’s sensitivity in the visible range [17]. Gupta B D et al. utilized ITO to prepare SPR ammonia gas and PH sensors, and demonstrated that the sensitivity and the operating range of the sensor is highly depended on ITO’s thickness [18, 19]. A. K. Mishra et al. designed a highly sensitive SPR sensor working in the infrared band and proved that ITO offers a positive effect on enhancing SPR’s sensitivity [20].

It should be noted that most of the previous sensor designs focus on angular or intensity interrogation. However, angular interrogation requires high precision instruments to track the variation of the resonance angle, which is not that cost-effective while the intensity noise of light source can significantly limit the achievable low detection limit of the intensity-interrogation-based sensor. Such a drawback can be gained by passing from angular/intensity interrogation to phase interrogation [22,23,24], since generally phase noises of laser sources are orders of magnitude lower compared to the intensity ones. In phase interrogation, the p-polarized “probe” beam and s-polarized “reference” beam experience different phase shift; the differential phase shift thus can be employed as a sensing parameter.

As mentioned above, though both graphene and ITO have been proven to possess the capability of sensitivity enhancement, how the performances can be affected by combining these two materials in a single SPR sensor has not been revealed, especially their impacts on phase sensitivity. In this paper, we propose a SPR biosensor with Ag-ITO-graphene hybrid structure and investigate its reflectivity and phase sensitivity at visible wavelength. It is found that ITO and graphene can be combined to further enhance the sensitivity.

Sensor Configuration and Investigation Method

In the proposed SPR sensor, the Kretschmann configuration is considered, in which a thin Ag film coated with glass slide is attached to the base of an equilateral prism made of high refractive index glass through an index matching fluid. The SPR sensor with Ag-ITO-graphene structure is shown in Fig. 1. This structure contains seven layers, namely Prism, BK7 glass slide, Ag, ITO, graphene, and analyte, among which BK7 and Ag are bind together via the titanium adhesion layer. The He-Ne laser of wavelength 632.8 nm is fixed as p-polarized light.

Fig. 1
figure 1

Schematic diagram of the proposed SPR biosensor with Ag-ITO-graphene hybrid structure

Full size image

The first layer is the SF11 glass prism and its refractive index (n1) is calculated through the following relation [25]:

$$ {n}_1=\sqrt{\frac{1.73759695{\lambda}^2}{\lambda^2-0.013188707}+\frac{0.313747346{\lambda}^2}{\lambda^2-0.0623068142}+\frac{1.89878101{\lambda}^2}{\lambda^2-155.23629}+1} $$
(1)

where λ is the wavelength of incident light. The second layer is the BK7 glass slide and its refractive index (n2) is determined by the following relation [26]:

$$ {n}_2=\sqrt{\frac{1.03961212{\lambda}^2}{\lambda^2-0.00600069867}+\frac{0.231792344{\lambda}^2}{\lambda^2-0.0200179144}+\frac{1.01046945{\lambda}^2}{\lambda^2-103.560653}+1} $$
(2)

The complex refractive index of the titanium adhesion layer (n3) at 632.8 nm is obtained from the experimental measurement data by Palik [25]. The fourth layer is the Ag thin film and its complex refractive index (n4) is calculated through the Drude dispersive model [27, 28].

$$ {n}_3=\sqrt{1-\frac{\gamma_1{\lambda}^2}{\gamma_2^2\left({\gamma}_1+ i\lambda \right)}} $$
(3)

where γ2 = 0.14541 μm is the wavelength corresponding to the bulk plasma frequency and γ1 = 17.6140 μm denotes the collision wavelength and is related to the losses.

The fifth layer of ITO film’s optical properties can be quantitatively described by using the classical Drude free-electron theory [29].

$$ {n}_5=\sqrt{3.8-\frac{\gamma_c{\lambda}^2}{\gamma_p^2\left({\gamma}_c+ i\lambda \right)}} $$
(4)

where the γp = 0.56497 μm and γc = 11.21076 μm.

The seventh layer of model is graphene and its complex refractive index (n6) in the visible range is given as [30]

$$ {n}_6=3.0+i\frac{C_1}{3}\lambda $$
(4)

where the constant C1 ≈ 5.446 μm−1 is implied by the opacity measurement by Nair [31]. The thickness of monolayer graphene is 0.34 nm. The sensing medium for initial calibration is deionized (DI) water and its refractive index (n7) is 1.333 [32]. The refractive index change of the sensing medium induced by the adsorption of biomolecules on the surface of monolayer graphene is denoted by Δnbio. Thus, based on the above parameters and equations, the refractive indices of the seven layers used in our SPR modeling at 632.8 nm are respectively: n1 = 1.7786, n2 = 1.5151, n3 = 2.1526 + i3 × 2.9241, n4 = 0.1350 + i4 × 3.9850, n5 = 1.8580 + i5 × 0.0580, n6 = 3 + i6 × 1.1487, n7 = 1.3330 + Δnbio.

In order to investigate the sensing performance, we have applied the transfer matrix method (TMM) and Fresnel equations based on an N-layer model to carry our analysis in details. For each layer, the dielectric constant (ε = n2) and the thickness (d) of the layer are decisive factors, and all layers are optically isotropic and nonmagnetic. In the electromagnetic field, tangential direction of the first boundary Z1 is Z = Z1 = 0. In consideration to the last boundary Zn−1 tangential field, we have the control point of the first boundary Z1 [29, 30]:

$$ \left[\frac{H_1}{W_1}\right]=M\left[\frac{H_{N-1}}{W_{N-1}}\right] $$
(5)

where H1 and H2 are tangential components of the first and the nth layer of the electric field, respectively; W1 and WN−1 are tangential components of the first and the nth layer of the boundary magnetic field, respectively; and M represents the combination of the layer structure of the transfer matrix (TM). The interference matrix of the N-layer structure can be expressed by

$$ M=\prod \limits_{k=2}^{N-1}{M}_k=\left[\begin{array}{l}{M}_{11}\\ {}{M}_{21}\end{array}\begin{array}{l}{M}_{12}\\ {}{M}_{22}\end{array}\right] $$
(6)

where Mk is expresses as

$$ {M}_k=\left[\begin{array}{cc}\cos {\alpha}_k& \left(-i\sin {\alpha}_k\right)/{p}_k\\ {}-i{p}_k\sin {\alpha}_k& \cos {\alpha}_k\end{array}\right] $$
(7)

where Pk and αk are written as

$$ {p}_k=\sqrt{\left(\frac{\nu_k}{\varepsilon_k}\right)}\cos {\theta}_k=\frac{\sqrt{\varepsilon_k-{n}_1^2{\sin}^2{\theta}_1}}{\varepsilon_k} $$
(8)
$$ {\alpha}_k=\frac{2\pi {d}_k}{\lambda}\sqrt{\varepsilon_k-{n}_1^2{\sin}^2{\theta}_1} $$
(9)

where εk and dk are the dielectric constant and the thickness of the kth layer, respectively, and θ1 and λ are incident angle and wavelength, as shown in Fig. 1. The matrix of the total reflection polarized light (γp) can be expressed as

$$ {\gamma}_p=\frac{\left({M}_{11}+{M}_{12}{p}_N\right){p}_1-\left({M}_{21}+{M}_{22}{p}_N\right)}{\left({M}_{11}+{M}_{12}{p}_N\right){p}_1+\left({M}_{21}+{M}_{22}{p}_N\right)} $$
(10)

where p1 represents the relation between the prism (at the first layer) and the dielectric constant and pN represents the relation between the analyte (at the nth layer) and the dielectric constant. Finally, the reflectance (Rp) for p-polarized wave is obtained:

$$ {\phi}_p=\arg \left({\gamma}_p\right) $$
(11)
$$ {R}_p={\left|{\gamma}_p\right|}^2 $$
(12)

The SPR phase sensitivity (Sp) is defined as

$$ {S}_p=\frac{\triangle {\phi}_p}{\triangle {n}_{\mathrm{bio}}} $$
(13)

In order to highlight the enhancement provided by graphene -ITO hybrid structure, in the following sections, we optimize the SPR sensor with conventional metal-dielectric structure, metal-dielectric-graphene structure, and metal-dielectric-graphene-ITO hybrid structure progressively and compare their sensitivity.

Analysis and Discussion

Metal-Dielectric Structure

In this structure, T5 (ITO) and T6 (graphene) layers of the Ag-ITO-graphene hybrid structure are removed, where the Tn represents nth layer. In the following analysis, the thicknesses of the first three layers are fixed at d1 = 200 nm, d2 = 100 nm, and d3 = 2.5 nm, respectively. The sensing medium of water refractive index variation (Δnbio) is 0.0001. We optimize the thicknesses of Ag on SPR’s performance, namely the reflectivity and the phase sensitivity. The reflectivity and the phase sensitivity are investigated by changing the thickness of Ag. Figure 2a, b shows the variation of reflectivity and phase with respect to angle of incidence for different thickness of Ag. The phase sensitivity with respect to Ag thickness is shown in Fig. 2c. A peak value of ~ 1.736 × 105°/RIU can be achieved with Ag thickness of 51 nm.

Fig. 2
figure 2

Variation of a reflectivity and b phase with respect to angle of incidence, c phase sensitive with respect to Ag thickness (mm)

Full size image

From Fig.2a, Ag with different thickness puts up different reflectivity characteristics in SPR with increasing incident angle. When the incident angle is 52.72°, the reflectivity is the optimal value. When the thickness of Ag increases from 10 to 50 nm, the reflectivity decreases from 0.757 to 8.237 × 10−3 a.u. When the thickness of Ag is 51 nm, the reflectivity has touch bottom 7.759 × 10−4 a.u. Nevertheless, when the thickness of Ag increases from 52 to 70 nm, the reflectivity increases from 4.235 × 10−3 to 0.337 a.u. SPR has the characteristics of phase change due to the change of reflectivity [33]. From Fig.2b, the change of phase reaches the maximum when the thickness of Ag is 51 nm. In the metal-dielectric structure, the thickness of Ag is the best in 51 nm. From Fig.2c, when the thickness of Ag increases from 10 to 50 nm, the phase sensitivity decreases from 1029.2 to 158°/RIU. When the thickness of Ag is 51 nm, the phase sensitivity reaches 1.736 × 105°/RIU. However, when the thickness of Ag increases from 52 to 70 nm, the phase sensitivity decreases from 4.488 × 104 to 4.309 × 103°/RIU. Therefore, in the metal-dielectric structure, when the thickness of Ag is 51 nm, the phase sensitivity is the highest.

Metal-Graphene Hybrid Structure

Graphene is known as a functional material to enhance the sensing performance. For Ag-graphene structure, T5 (ITO) is removed and different layers of graphene are added to evaluate their contributions. When the thickness of Ag changes from 10 to 60 nm, the different layers of graphene have diverse peaks. The thickness of graphene of monolayer, bilayer, 3-layer, 4-layer, and 5-layer are 0.34 nm, 0.68 nm, 1.02 nm, 1.36 nm, and 1.70 nm, respectively. When the incident angle is 52.91°, the reflectivity is the optimal value. From Fig. 3, when the thickness of Ag increases from 10 to 48 nm, the phase sensitivity of graphene in the same layer is decreasing. In addition, with the increase of the number of graphene layers, the phase sensitivity is also decreasing. When the grapheme is monolayer and the thickness of Ag is 49 nm, a peak value of ~ 2.358 × 105°/RIU can be achieved. When the thickness of Ag increases from 50 to 60 nm, the phase sensitivity of graphene in the same layer decreases from 6.879 × 104 to 7.521 × 103°/RIU, and with the increase of the number of graphene layers, the phase sensitivity is getting lower and lower. Therefore, in metal-graphene hybrid structure, when the thickness of Ag is 49 nm and the graphene is monolayer, the phase sensitivity is the highest.

Fig. 3
figure 3

Variation phase sensitivity for diverse layers of grapheme and different thickness of Ag (nm)

Full size image

Metal-Graphene-ITO Hybrid Structure

According to the above analysis, it can be found that though graphene is able to improve the phase sensitivity, the enhancement is not that prominent. For comparison, the Ag-graphene-ITO hybrid structure is investigated. First, the different thickness of Ag and ITO are analyzed when the graphene is monolayer. Figure 4 shows the phase sensitivity for different thickness of Ag and ITO. When the incident angle is 49.45°, the reflectivity is the optimal value. When the thickness of ITO and Ag are 100 and 5 nm, the phase sensitivity is 1.683 × 103°/RIU. With the increase of the thickness of Ag, the phase sensitivity decreases gradually, as far as the thickness of Ag is 35 nm, the phase sensitivity is 386.96°/RIU. When the thickness of ITO and Ag are 150 and 15 nm, the phase sensitivity reaches the highest for 3.526 × 104°/RIU. When the thickness of Ag is 20 nm and the thickness of ITO is 220 nm, the phase sensitivity is maximum of 1.707 × 106°/RIU. When the thickness of ITO and Ag are 221 and 20 nm, the highest of phase sensitivity is obtained by 7.43 × 105°/RIU. When the thickness of ITO is 250 nm and 300 nm, the maximum phase sensitivity of Ag at 30 nm is 2.135 × 104°/RIU and 2.133 × 103°/RIU, respectively. Then, the influence of the number of graphene layers on the resonance spectra is explored. With the increase of graphene layer number, the phase sensitivity is gradually reduced, when the thickness of Ag is 20 nm and the thickness of ITO is 220 nm, which is showed in Fig. 5. Therefore, when the thickness of Ag, graphene, and ITO are 20 nm, monolayer and 220 nm respectively, a peak value of ~ 1.707 × 106°/RIU can be achieved.

Fig. 4
figure 4

Variation phase sensitivity for different thicknesses of Ag and ITO (nm)

Full size image
Fig. 5
figure 5

Variation phase sensitivity for different layers of graphene

Full size image

For comparison, the optimized structure parameters and corresponding sensing performances of the three structures are listed in Table 1.

Table 1 Three structures sensing performances of reflectivity and phase sensitivity
Full size table

Conclusion and Discussion

In this paper, three SPR biosensors based on Ag, Ag-graphene, and Ag-ITO-graphene are optimized and compared. The SPR sensor based on Ag-ITO-graphene structure with phase sensitivity of 1.707 × 106°/RIU shows better performance when compared to the other two. It is therefore reasonable to conclude that ITO can be used to further enhance the performance of SPR biosensor, especially the phase sensitivity. Hence, we believe that the proposed sensor is of a great potential in biomolecule sensing applications.