An Optimized Graphene-Based Surface Plasmon Resonance Biosensor for Detecting SARS-CoV-2

Abstract

Graphene-enhanced surface plasmon resonance (SPR) biosensors offer promising advancements in viral detection, particularly for SARS-CoV-2. This study presents the design and optimization of a multilayer SPR biosensor incorporating silver, silicon nitride, single-layer graphene, and thiol-tethered ssDNA to achieve high sensitivity and specificity. Key metrics, including SPR angle shift (Δθ), sensitivity (S), detection accuracy (DA), and figure of merit (FoM), were assessed across SARS-CoV-2 concentrations from 150 to 525 mM. The optimized biosensor achieved a sensitivity of 315.91°/RIU at 275 mM and a maximum Δθ of 4.2° at 400 mM, demonstrating strong responsiveness to virus binding. The sensor maintained optimal accuracy and figure of merit at lower concentrations, with a linear sensitivity response up to 400 mM, after which surface saturation limited further responsiveness. These results highlight the suitability of the optimized biosensor for real-time, point-of-care SARS-CoV-2 detection, particularly at low viral loads, supporting its potential in early diagnostics and epidemiological monitoring.

1. Introduction

Biosensors are analytical devices that integrate a biological recognition element with a physical transducer to convert a biological interaction into a measurable signal [1]. These devices are categorized based on their transduction mechanism, including electrochemical [2], optical [3], piezoelectric [3], and thermal [4] types. Optical biosensors, particularly those based on surface plasmon resonance (SPR), have gained significant attention for their ability to perform real-time, label-free detection with high sensitivity [5]. SPR biosensors operate by detecting shifts in the refractive index near the sensor surface, which arise from molecular interactions [6]. This capability makes SPR sensors highly effective for applications in pathogen detection [7], drug discovery [8], and biomolecular interaction [9] studies.

The theory of SPR is based on the excitation of surface plasmons [10], or collective oscillations of free electrons, at the interface of a metal (often silver or gold) and a dielectric. When polarized light strikes the metal–dielectric interface at a specific angle, it transfers energy to the surface plasmons, resulting in a reduction in reflected light intensity at that angle. The resonance angle is sensitive to changes in the refractive index near the sensor surface [11], allowing SPR biosensors to monitor molecular binding events quantitatively and in real time.

To enhance SPR sensor performance, multilayer architectures incorporating additional materials between the metal layer and the biological recognition layer are often employed. In these configurations, a metal layer, typically silver or gold, is deposited on a prism to initiate SPR. Additional layers, such as dielectrics like silicon nitride (Si₃N₄) [12] and two-dimensional (2D) materials like graphene [13], can be incorporated to improve optical characteristics and sensitivity. The metal layer facilitates plasmon excitation, the dielectric modifies the evanescent field to optimize analyte interaction, and the functionalized top layer provides specific binding sites for the target biomolecule.

Two-dimensional materials, especially graphene, have shown great potential in biosensing applications. Graphene’s atomic-scale thickness, high surface area, excellent conductivity, and tunable chemical properties make it an ideal candidate for enhancing SPR sensors [14]. Graphene can amplify the SPR signal through increased light-matter interaction at the metal–dielectric interface, which improves sensitivity and lowers the detection limit. Moreover, graphene’s surface can be functionalized with various biomolecules, allowing for selective binding and detection of specific targets, including viral particles and proteins [15].

In this context, SARS-CoV-2, the causal agent of COVID-19 [16], is a highly transmissible virus characterized by distinct viral proteins and RNA, which serve as biomarkers for detection. Rapid and accurate SARS-CoV-2 detection is critical for managing outbreaks, requiring diagnostic tools that offer sensitivity, specificity, and rapid response. Traditional detection methods, such as RT-PCR [17], have limitations in terms of time and scalability, whereas biosensors, especially those based on SPR, present a promising alternative due to their capacity for real-time detection without extensive sample preparation [18].

In this study, we investigate a multilayer SPR biosensor incorporating graphene and silicon nitride to enable sensitive and specific detection of SARS-CoV-2 across varying concentrations (i.e., 150, 275, 400, and 525 mM [19]). The multilayer architecture is designed to optimize sensor performance by harnessing the unique properties of these materials. Graphene, with its high electron mobility and large surface area, enhances biosensor sensitivity, while silicon nitride contributes structural stability and optical transparency. This biosensor aims to facilitate rapid and accurate SARS-CoV-2 detection, presenting a potential tool for point-of-care diagnostics and epidemiological surveillance.

2. Materials and Methods

2.1. Biosensor Architecture and Initial Parameters

The architecture of the proposed biosensor and the initial configuration parameters are presented in Figure 1 and Tables S1 and S2. This multilayer SPR biosensor is designed to detect SARS-CoV-2 by leveraging the unique properties of various materials in a layered configuration. The performance of this sensor is enhanced by incorporating graphene and silicon nitride layers between the metal substrate (silver) and the functionalization layer, which specifically targets SARS-CoV-2 viral components. Figure 1 points out the difference between the basic [20] and multilayer configurations. The basic SPR sensor configuration (Figure 1, left) consists solely of a BK-7 prism and a silver (Ag) layer, with water acting as the sensing medium. The resonance angle (θ) in this simple configuration is highly sensitive to changes in the refractive index of the medium but lacks specificity and sensitivity for biomolecular interactions. In contrast, the proposed multilayer configuration (Figure 1, right) integrates additional layers, i.e., silicon nitride (Si₃N₄), graphene, and a thiol-tethered single-stranded DNA (ssDNA) layer. This architecture is designed to enhance the interaction between the evanescent field generated by SPR and the target analyte (SARS-CoV-2), thereby improving detection sensitivity and specificity [21].

Table S1 lists the systems analyzed in this study, covering a range of configurations from the basic SPR sensor to more complex multilayer systems. The simplest configurations (Sys₀ and Sys₁) consist of the prism/silver structure with water (H₂O) and phosphate-buffered saline (PBS) solution, respectively, as sensing media. These configurations serve as baseline systems for comparison, allowing for the evaluation of performance improvements introduced by multilayer design. The remaining configurations (Sys₂ to Sys₉) include sequential additions of Si₃N₄, graphene, and ssDNA, building towards the complete multilayer configuration (Sys₉) of prism/silver/silicon nitride/graphene/ssDNA/PBS. The inclusion of PBS in the medium is essential, as it reflects the experimental preparation conditions for SARS-CoV-2 at different concentrations [19]. It is expected that PBS will serve to maintain pH stability and simulate physiological conditions, which are critical for maintaining viral structure during detection.

Table S2 provides the initial parameters used to configure each layer of the biosensor, with values sourced from prior research [22,23,24,25]. The BK-7 prism (refractive index 1.5151) serves as the substrate for SPR excitation, while the silver layer (thickness 55.0 nm, refractive index 0.056253 + 4.2760i) is optimized to facilitate strong plasmon resonance. Silicon nitride (Si₃N₄), with a refractive index of 2.0394 and a thickness of 5.0 nm, is applied as a dielectric layer to enhance the evanescent field without excessive damping. The graphene layer, deposited at 0.34 nm thickness, exhibits a refractive index of 2.7611 + 1.6987i, which enhances light-matter interactions at the sensor interface. Finally, the functional layer, comprising thiol-tethered ssDNA with a refractive index of 1.462 and a thickness of 3.20 nm, is designed to capture SARS-CoV-2 with high specificity through molecular recognition interactions [26].

2.2. Modeling Approach

We have carried out a numerical analysis to calculate the reflectance curve using the Transfer Matrix Method (TMM) and Fresnel equation; details about the mathematical model can be seen in [20,25]. The proposed biosensors are constructed by sequentially stacking multiple layers in parallel (see Figure 1). The thickness of each layer is measured along the perpendicular axis (z-axis). The interface boundary conditions of the first and final layers are represented as $Z=Z_1=0$ and $Z=Z_{n-1}$, respectively.

The transfer matrix describes the relationship between the tangential components of the electric and magnetic fields as

$$\left[\begin{array}{c}{\mathrm{E}}_1\\ {\mathrm{H}}_1\end{array}\right]=\mathrm{M}\left[\begin{array}{c}{\mathrm{E}}_{\mathrm{N}-1}\\ {\mathrm{H}}_{\mathrm{N}-1}\end{array}\right]$$

(1)

where ${\mathrm{E}}_1$, ${\mathrm{H}}_1$, ${\mathrm{E}}_{\mathrm{N}-1}$, and ${\mathrm{H}}_{\mathrm{N}-1}$ correspond to the tangential components of electric and magnetic fields at the first and last layer interfaces, respectively. $\mathrm{M}$ represents the characteristic matrix of the $N$-layer structure with elements $M_{ij}$:

$$\mathrm{M}=\prod _{k=2}^{N-1}{M}_k=\left[\begin{array}{cc}{M}_{11}& M_{12}\\ M_{21}& M_{22}\end{array}\right]$$

(2)

And $M_k$ is expressed as

$$M_k=\left[\begin{array}{cc}\cos {\beta }_k& -i\sin {\beta }_k/q_k\\ -i{q}_k\sin {\beta }_k& \cos {\beta }_k\end{array}\right]$$

(3)

where $k$ is an integer number. Additionally, ${\beta }_k$ is the phase thickness and $q_k$ is the refractive index of the corresponding layer:

$${\beta }_k={\displaystyle \frac{2\pi d_k}{{\lambda }_0}}\sqrt{{\epsilon }_k-n_1^2{\sin }^2\theta }$$

(4)

and

$$q_k={\displaystyle \frac{\sqrt{{\epsilon }_k-n_1^2{\sin }^2\theta }}{{\epsilon }_k}}$$

(5)

Here, $\theta$ is the angle of incidence, ${\lambda }_0$ is the incident wavelength light, $n_1$ is the refractive index of the prism, $d_k$ is the thickness layer, and the local dielectric function $\epsilon \left({\lambda }_0\right)$ can be adopted as $n\left({\lambda }_0\right)$. In this work, we use the data for the He-Ne laser with ${\lambda }_0=633$ nm. Hence, the total reflection analysis of the N-layer system is obtained as

$$R={\left|{\displaystyle \frac{\left(M_{11}+M_{12}{q}_N\right)q_1-\left(M_{21}+M_{22}{q}_N\right)}{\left(M_{11}+M_{12}{q}_N\right)q_1+\left(M_{21}+M_{22}{q}_N\right)}}\right|}^2$$

(6)

By using Equation (6), the SPR curve as a function of the angle of incidence is computed. It is noted that for each SPR curve, surface plasmon excitation is identified as a dip in the reflected intensity $R$, corresponding to the minimum in attenuated total reflection (ATR). The angle of incidence at the ATR minimum is called the SPR angle.

Moving forward, to analyze the performance of the biosensor, it is necessary to consider the following metrics. The sensitivity of the biosensors ($S^P$) is defined as the multiplication of the sensitivity to the refractive index change ($S_{RI}^N$) and the adsorption efficiency of the target analyte ($E$):

$$S^P=S_{RI}^P·E$$

(7)

To optimize the biosensor, we mainly focus on the sensitivity enhancement concerning conventional biosensors (i.e., P/Ag/M) (here, $P$ is the parameter to be optimized) expressed as

$${∆S}_{RI}^P={(S}_{RI}^P-S_{RI}^0)/S_{RI}^0$$

(8)

The sensitivity to the refractive index change can be expressed as

$$S_{RI}^P=∆\theta /∆n$$

(9)

The parameter $∆\theta$ represents the angle variation.

The detection accuracy (DA) (i.e., signal-to-noise ratio) can be expressed in terms of $∆\theta$ and the full width at half maximum (FWHM) as

$$\mathrm{D}\mathrm{A}=∆\theta /\mathrm{F}\mathrm{W}\mathrm{H}\mathrm{M}$$

(10)

Finally, quality factor (QF) (or figure of merit, FoM) can be expressed in terms of $S$ and FWHM as

$$\mathrm{Q}\mathrm{F}=S_{RI}^P/\mathrm{F}\mathrm{W}\mathrm{H}\mathrm{M}$$

(11)

3. Results and Discussions

3.1. Best Sensor Configuration

In this part, the objective is to determine the optimal SPR sensor configuration by comparing different multilayer designs without the presence of the analyte (SARS-CoV-2). The configurations are evaluated based on key metrics: percentage attenuation, full FWHM, and percentage sensitivity enhancement. These metrics, displayed in Figure 2 and Table S3, allow us to calculate the intrinsic performance of each configuration and identify the most effective design before introducing the target virus.

Figure 2a shows the SPR reflectance curves for each configuration (Sys₀ to Sys₉) as a function of the angle of incidence, ranging from 60° to 80°. The position and shape of each SPR dip provide insight into the resonance conditions and sensitivity of each configuration. In the basic configuration, Sys₀, a sharp and narrow SPR dip is observed at approximately 67°, reflecting the interaction between the prism–silver interface and the surrounding water medium. This sharp dip in Sys₀ suggests moderate sensitivity, as the simple silver layer structure provides a minimal enhancement to the evanescent field interaction with the medium. In contrast, the addition of silicon nitride and graphene layers shifts the SPR dip to slightly higher angles and broadens the dip width, as observed in Sys₆, Sys₈, and Sys₉, where SPR dips occur around 70° to 72°. The broader SPR dips in these configurations indicate a greater interaction between the multilayer structure and the evanescent field, enhancing the sensor’s sensitivity [27].

Attenuation, defined as the percent decrease in reflectance, indicates the strength of the SPR signal. A higher attenuation percentage corresponds to a stronger interaction between the evanescent field and the medium. As shown in Figure 2b and detailed in Table S3, the basic system Sys₀ (prism/silver/water) exhibits minimal attenuation (0.023%). For consistency, PBS was used as the medium for configurations Sys₁ through Sys₉, as PBS mimics physiological conditions and provides a more realistic refractive index for future experiments involving SARS-CoV-2 [19]. Despite the absence of an analyte, the introduction of silicon nitride and graphene layers significantly increases attenuation. For instance, Sys₆ (prism/silver/silicon nitride/graphene/PBS) achieves an attenuation of 7.81%, while Sys₈ and Sys₉ reach 7.93% and 6.59%, respectively. These results confirm that the multilayer design enhances signal strength through better interaction with the evanescent field, even in the absence of the target analyte.

FWHM, shown in Figure 2c and Table S3, is a crucial metric in SPR biosensors, as it indicates the sharpness of the SPR dip; lower FWHM values suggest a more precise resonance, which is typically desirable [28]. The basic configuration Sys₀ exhibits a relatively narrow FWHM (0.88°), consistent with a simple prism–silver design in water. For the configurations using PBS, the addition of Si₃N₄ and graphene layers slightly broadens the FWHM due to enhanced interaction with the evanescent field. Sys₆ shows an FWHM of 1.82°, while Sys₈ and Sys₉ exhibit the broadest values, 1.91° and 1.85°, respectively. Although increased FWHM indicates slightly reduced dip sharpness, it also suggests stronger interactions within the layered structure, which may enhance sensitivity upon analyte introduction.

Sensitivity enhancement (Figure 2d), measured as the percentage increase in SPR response relative to Sys₀ in water, provides a baseline measure for evaluating the impact of each layer. Sys₀ was selected as the baseline because it represents the simplest configuration in water, devoid of additional layers and reflective of the least complex SPR setup. The configurations using PBS, specifically Sys₂ through Sys₉, exhibit varying sensitivity enhancements depending on the number and type of layers included. Among these, Sys₃, Sys₈, and Sys₉ show the highest sensitivity enhancements, with values of 5.19%, 5.66%, and 5.69%, respectively (Table S3). While Sys₉ demonstrates the highest sensitivity enhancement, practical considerations, such as the complexity of manufacturing this configuration, led to the selection of Sys₈ as the best configuration. To remark, the observed drop in sensitivity for Sys₄ and Sys₅ likely results from the absence of a silicon nitride layer, which plays a critical role in stabilizing the graphene and enhancing the refractive index contrast needed for optimal SPR performance. When silicon nitride is introduced, as in Sys₆ through Sys₉, the system achieves more effective resonance conditions, resulting in higher sensitivity.

Then, based on percentage attenuation, FWHM, and percentage sensitivity enhancement metrics, Sys₈ was selected as the optimal configuration for future analyte testing. This configuration, comprising prism/silver/silicon nitride/graphene/ssDNA layers with PBS as the medium, offers high sensitivity and stability while avoiding the manufacturing complexities associated with Sys₉:

• Graphene directly on silver and topped with Si₃N₄ is unconventional, posing adhesion issues. This unusual stacking order requires precise control to avoid layer instability.
• Binding ssDNA to graphene is less straightforward than to silicon nitride, often needing intermediary molecules or additional treatments, which increases fabrication steps and variability.
• Differences in thermal expansion among silver, graphene, and Si₃N₄ can lead to strain or delamination, especially under handling or environmental changes.
• The precise deposition and functionalization required for each layer make it challenging to scale Sys₉ for commercial production, as it requires stringent quality control to maintain sensor performance.

These factors suggest Sys₈ is more practical for scalable, reproducible manufacturing while still offering high sensitivity.

3.2. Optimization of Silver Thickness

Figure 3a shows the SPR reflectance curves for varying silver thicknesses from 40 to 65 nm with PBS as the medium. The baseline configuration, Ag_base (Sys₈ with 55 nm of silver in water), is included for comparison. As the silver thickness increases, the SPR dip shifts slightly and narrows, reflecting improved coupling with the evanescent field. However, there is a trade-off between narrowing the SPR dip and managing the attenuation and sensitivity, as excessive thickness can hinder light penetration into the metal layer.

Figure 3b and Table S4 present the attenuation percentages for each silver thickness. A noticeable increase in attenuation occurs at higher silver thicknesses, particularly at 65 nm, where attenuation reaches 38.87%. The lowest attenuation is observed at 50 nm (0.43%), significantly reducing reflectance loss compared to thicker silver layers. This reduced attenuation at 50 nm suggests a more efficient SPR response, where the sensor maintains signal strength without excessive energy loss.

The FWHM values, shown in Figure 3c and Table S4, indicate the sharpness of the SPR dip, with lower values reflecting sharper resonance and potentially higher signal-to-noise ratio. FWHM decreases as the silver thickness increases, with the lowest FWHM observed at 65 nm (1.64°). However, the trade-off at 65 nm is higher attenuation, which can impact sensitivity. At 50 nm, the FWHM is 2.24°, balancing dip sharpness with manageable attenuation. This suggests that 50 nm provides an acceptable compromise, maintaining a defined resonance peak without the excessive broadening seen at lower thicknesses.

Figure 3d and Table S4 illustrate the sensitivity enhancement for each silver thickness relative to Ag_base. Figure S1 shows that sensitivity enhancement increases quasi-linearly with silver thickness, yet the whole improvement remains minimal, reaching a maximum of only 0.82% at 65 nm. At 50 nm, the sensitivity enhancement is 0.74%, offering a slight improvement over the baseline without the high attenuation and reduced FWHM associated with thicker layers. This minor increase in sensitivity suggests that thicker silver layers provide diminishing returns and that maximizing thickness beyond 50 nm is not cost-effective for practical applications.

Based on the analysis of these metrics, a silver thickness of 50 nm is selected as the optimal configuration. This thickness provides the lowest attenuation (0.43%), a balanced FWHM (2.24°), and a manageable sensitivity enhancement (0.74%). This combination ensures that the sensor maintains a robust SPR response with minimized energy loss, sufficient dip sharpness, and acceptable sensitivity.

3.3. Optimization of Silicon Nitride Thickness

To further optimize the sensor’s performance, the thickness of the silicon nitride (Si₃N₄) layer was varied from 5 to 20 nm, and its effects on SPR metrics were analyzed. Figure 4 and Table S5 display the results obtained.

Figure 4a shows the SPR reflectance curves as a function of the angle of incidence, with S₃N_{4_base} as the reference configuration in water with 5 nm thickness. As the Si₃N₄ thickness increases, the SPR dip shifts toward higher angles and broadens. This shift and broadening indicate enhanced interaction with the evanescent field, which can increase sensitivity. However, excessive thickness (such as 20 nm) leads to a pronounced broadening, which may negatively affect the sensor’s precision and lead to high attenuation.

Figure 4b and Table S5 present the attenuation percentages for each Si₃N₄ thickness. A noticeable increase in attenuation is observed as the thickness increases, especially at 20 nm, where attenuation reaches 90.64%. This high attenuation indicates substantial energy loss, which can reduce sensor effectiveness dramatically. Conversely, at 15 nm, attenuation remains relatively low (8.30%), offering a reasonable balance between signal strength and energy retention. This manageable attenuation value, at 15 nm, suggests an optimal balance, enabling effective SPR response without significant signal loss.

Figure 4c and Table S5 display the FWHM values for each thickness. The FWHM increases with the Si₃N₄ thickness, reaching a maximum of 14.16° at 20 nm, indicating a broad and less distinct SPR dip. For practical biosensing, a sharp resonance (lower FWHM) is preferable, as it correlates with higher signal-to-noise ratios. At 15 nm, the FWHM is 4.99°, which is broader than lower thicknesses but still within an acceptable range (<5°) for effective sensing.

Figure 4d and Table S5 illustrate the sensitivity enhancement percentage for each Si₃N₄ thickness, while Figure S2 shows a linear trend in sensitivity as Si₃N₄ thickness increases. Sensitivity enhancement improves with greater thickness, peaking at 19.82% for 20 nm. However, this increase in sensitivity comes at the cost of significant attenuation and an excessively broad FWHM at 20 nm, making it less practical. At 15 nm, sensitivity enhancement reaches 15.09%, providing a substantial improvement over lower thicknesses while avoiding excessive attenuation and broad FWHM seen at 20 nm.

Based on these metrics, a Si₃N₄ thickness of 15 nm is selected as the optimal configuration. This thickness provides a balanced performance with manageable attenuation (8.30%), an FWHM of 4.99° (ensuring adequate resonance sharpness), and a high sensitivity enhancement of 15.09%. The choice of 15 nm achieves a practical compromise between maximizing sensitivity and maintaining a sharp, low-loss SPR response.

3.4. Optimization of the Number of Graphene Layers

To determine the optimal configuration, we analyze the sensor’s performance with one to six graphene layers. Figure 5 and Table S6 summarize these results, while Figure S3 shows the sensitivity enhancement trend.

Figure 5a presents the SPR reflectance curves as the number of graphene layers increases. As more graphene layers are added, the SPR dip shifts to higher angles and broadens, indicating greater interaction with the evanescent field. However, excessive broadening with multiple layers reduces the dip sharpness, which critically impacts sensor precision. The single-layer graphene configuration (L1) maintains a distinct SPR dip at approximately 82°, balancing sensitivity and dip sharpness.

Figure 5b and Table S6 display the attenuation percentages for each configuration. Attenuation rises sharply as the graphene layers increase, with a notable jump to 66.15% at six layers. Single-layer graphene (L1) has the lowest attenuation at 8.30%, indicating minimal reflectance loss and efficient SPR response. This low attenuation with single-layer graphene suggests that it supports a strong SPR signal while retaining energy, crucial for maximizing signal-to-noise ratios.

The FWHM values, shown in Figure 5c and Table S6, increase as the graphene layer count rises. At one layer, the FWHM is 4.99°, which is within an acceptable range for achieving a sharp resonance. With additional layers, FWHM increases significantly, reaching 9.58° at five layers and 13.8° at six layers, indicating a substantial loss in resonance sharpness. Maintaining a sharp resonance is critical for high-precision detection, and single-layer graphene offers an optimal balance between sensitivity and FWHM.

Sensitivity enhancement, displayed in Figure 5d and Table S6, increases linearly up to four layers, with a maximum of 3.27% at five layers. However, beyond four layers, sensitivity enhancement plateaus and even begins to decrease slightly. Single-layer graphene (L1) provides a sensitivity enhancement of 1.28%, which is lower than configurations with multiple layers but is balanced by the advantages of lower attenuation and narrower FWHM. Figure S3 highlights this quasi-linear trend, indicating that while additional graphene layers initially improve sensitivity, they also introduce diminishing returns and adverse effects on attenuation and FWHM at higher counts.

To stress again, single-layer graphene (L1) is selected as the optimal configuration. This choice is justified by the following considerations:

• The low attenuation of single-layer graphene minimizes signal loss, ensuring a strong SPR response.
• With a manageable FWHM, single-layer graphene maintains a sharp resonance dip, balancing precision and sensitivity.
• Although sensitivity enhancement is higher with more layers, as reported in [25], the improvement with single-layer graphene is acceptable, especially given the diminishing returns and high attenuation observed with additional layers.

Then, single-layer graphene provides the best balance across these metrics, supporting efficient SPR interaction without the excessive losses and broadening seen in multilayer graphene configurations.

3.5. Optimization of Thiol-Tethered ssDNA Thickness

Now, we examine the impact of ssDNA layer thicknesses from 3.2 to 50 nm. Figure 6 and Table S7 summarize the results, while Figure S4 highlights the sensitivity enhancement trend. Particularly, Figure 6a shows the SPR reflectance curves as the ssDNA thickness increases. With thicker ssDNA layers, the SPR dip shifts toward higher angles and broadens. The baseline configuration, ssDNA_{3.2nm_base} (ssDNA in water), maintains a sharp dip, but with increasing thickness, excessive broadening is observed, especially at 30 nm and 50 nm. This broadening negatively affects the sensor’s precision, as it reduces the sharpness of the resonance dip.

Figure 6b and Table S7 evidence the effect of increasing ssDNA thickness on attenuation. Attenuation rises sharply with layer thickness, reaching 96.93% at 50 nm. At 5 nm, attenuation is relatively low (10.34%), providing a strong SPR response with manageable energy loss. This low attenuation at 5 nm suggests that it balances sensitivity with signal clarity, making it preferable over thicker layers that result in substantial reflectance loss.

The FWHM values in Figure 6c and Table S7 show an increasing trend with ssDNA thickness. Starting from 4.99° at 3.2 nm, FWHM rises to 5.20° at 5 nm, remaining within an acceptable range for maintaining a sharp resonance. Beyond 5 nm, FWHM increases considerably, reaching 9.29° at 20 nm and an excessive 53.86° at 50 nm. The 5 nm thickness offers a balance, allowing for a well-defined SPR dip essential for accurate and reliable biosensing.

Figure 6d and Table S7 demonstrate the sensitivity enhancement for each ssDNA thickness, while Figure S4 shows a linear trend in sensitivity up to 20 nm. Sensitivity peaks at 6.84% for a 20 nm thickness. However, this comes at the cost of high attenuation (80.48%) and broad FWHM (9.29°). For practical purposes, the 5 nm thickness provides a reasonable sensitivity enhancement (2.27%) without the detrimental increase in attenuation and FWHM associated with thicker ssDNA layers. After 20 nm, sensitivity enhancement declines sharply, making thicker layers less effective. Therefore, the 5 nm ssDNA layer offers an optimal trade-off, preserving sensitivity while maintaining low attenuation and a sharp resonance.

3.6. Application of the Optimized SPR Biosensor for SARS-CoV-2 Sensing

Table S8 presents the optimized parameters for Sys₈. The structure comprises a BK7 prism with a refractive index of 1.5151, forming the base substrate for efficient SPR excitation [29]. Over this, a 50 nm silver layer (refractive index 0.056253 + 4.2760i) is used to generate the SPR effect, achieving a balance between signal strength and resonance sharpness. A 15 nm layer of silicon nitride, with a refractive index of 2.0394, serves as a dielectric spacer, enhancing the sensor’s sensitivity by optimizing the evanescent field interaction. Single-layer graphene (0.34 nm, refractive index 2.7611 + 1.6987i) is then added to amplify the SPR signal without significant attenuation, while a 5 nm thiol-tethered ssDNA layer (refractive index 1.462) provides selectivity for SARS-CoV-2 detection through specific binding interactions. We point out that the sensor is calibrated with PBS (refractive index 1.334) as the sensing medium, which closely mimics physiological conditions. To detect SARS-CoV-2, the refractive index of the virus in PBS is presented for varying concentrations, ranging from 1.340 at 150 mM to 1.355 at 525 mM. These values, based on a study by Kumar et al. [19], demonstrate a linear relationship between refractive index and viral concentration, enabling quantitative analysis of SARS-CoV-2 levels.

Figure 7a shows the SPR reflectance curves for various SARS-CoV-2 concentrations (150 to 525 mM) in PBS, with the black curve representing the baseline optimized system in PBS without virus adsorption (refractive index of 1.334). As the virus concentration increases, corresponding to refractive indices of 1.340, 1.345, 1.350, and 1.355, the SPR dip shifts toward higher angles, demonstrating the sensor’s response to changes in the refractive index due to virus presence [30]. This shift indicates successful detection, as increased virus concentration enhances the binding events on the sensor surface.

Figure 7b and Table S9 demonstrate the increasing trend in attenuation as the virus concentration rises. Starting from 18.11% at 150 mM, attenuation escalates to 78.47% at 525 mM. This high attenuation suggests that higher concentrations of SARS-CoV-2 significantly dampen the reflected light, indicating a strong interaction between the evanescent field and the viral particles. However, this also highlights a drawback: excessive attenuation at higher concentrations can lead to signal saturation, reducing the sensor’s efficiency in detecting further increases in virus concentration. This saturation effect is visible as a plateau in sensitivity at higher concentrations.

The FWHM values in Figure 7c and Table S9 indicate broadening of the SPR dip as virus concentration increases. Starting from 5.64° at 150 mM, the FWHM rises to 8.62° at 525 mM. This broadening suggests a decrease in dip sharpness, which may affect detection precision. While some increase in FWHM is expected with higher analyte concentrations due to enhanced binding, excessive broadening could reduce the sensor’s ability to resolve small changes in virus concentration, as the signal becomes less distinct.

Figure 7d and Table S9 show sensitivity enhancement percentages for each concentration. Sensitivity initially rises from 2.25% at 150 mM to 5.11% at 400 mM, following a linear trend up to this concentration, as observed in Figure S5. Beyond 400 mM, sensitivity enhancement plateaus, reaching 4.74% at 525 mM. This plateau suggests that the sensor’s responsiveness diminishes at higher concentrations, potentially due to near-complete saturation of the ssDNA functional layer. This behavior implies that while the sensor remains sensitive at low to moderate concentrations, its effectiveness decreases as it approaches higher viral loads.

Furthermore, the linear sensitivity trend up to 400 mM indicates that Sys₈ operates effectively within this range, providing reliable measurements and clear differentiation between concentrations. However, the observed plateau and high attenuation beyond 400 mM suggest that the sensor’s operational range may be limited, as increased virus concentrations lead to diminishing returns in sensitivity, broad FWHM, and excessive attenuation. Therefore, the optimized Sys₈ configuration is best suited for low to moderate SARS-CoV-2 concentrations, where it achieves a balance between sensitivity, sharpness, and manageable attenuation.

To further emphasize the potential of the optimized Sys₈ configuration, we report different performance metrics in Figure 8 and Table S10, including variation angle (Δθ), sensitivity (S), detection accuracy (DA), and figure of merit (FoM). Actually, the SPR angle shift is a critical indicator of responsiveness to refractive index changes upon SARS-CoV-2 adsorption [25]. As shown in Figure 8a and Table S10, Δθ increases progressively from 1.85° at 150 mM to 4.2° at 400 mM, reflecting the biosensor’s strong sensitivity to changes in viral concentration. This significant angle shift, particularly at 400 mM, implies a high level of viral particle interaction with the functionalized ssDNA layer. However, the reduction in Δθ to 3.9° at 525 mM suggests the onset of surface saturation, where the sensor’s available binding sites reach capacity, limiting additional angle shifts. This plateau effect is typical in biosensors and highlights Sys₈’s optimal concentration range for quantitative analysis, ideally up to 400 mM.

Sensitivity, displayed in Figure 8b and Table S10 as the angular shift per refractive index unit (°/RIU), is highest at 315.91°/RIU at 275 mM, demonstrating Sys₈’s high responsiveness to moderate concentrations. This sensitivity level is advantageous in diagnostic settings where precise measurement of viral load is necessary, especially in early-stage infections or low-viral-load scenarios [31,32]. Sensitivity decreases to 185.71°/RIU at 525 mM, reflecting the sensor’s diminishing capacity to detect finer concentration changes at high viral loads. This decrease suggests that Sys₈ is best suited for detecting SARS-CoV-2 in the 150–275 mM range, where it achieves sharp and clear angular shifts. This characteristic makes Sys₈ valuable for early diagnostics, allowing healthcare providers to detect lower virus concentrations with high accuracy.

Detection accuracy, which represents the sensor’s ability to differentiate between varying viral concentrations, peaks at 0.60 at 400 mM, as shown in Figure 8c and Table S10. The increased DA from 0.33 at 150 mM to 0.60 at 400 mM suggests that Sys₈ is highly effective in quantifying SARS-CoV-2 at these concentrations. However, DA declines to 0.45 at 525 mM, representing the trends in Δθ and S, and indicating potential saturation. Despite this, Sys₈ maintains a high degree of precision up to 400 mM, which is crucial for applications requiring accurate viral load quantification, such as patient monitoring or assessing infection progression. This level of accuracy reinforces Sys₈’s capability as a reliable diagnostic tool, especially within the concentration range where viral load is most clinically relevant.

The figure of merit, which balances sensitivity with signal resolution, is highest at 54.66 RIU⁻¹ at 150 mM and decreases to 21.55 RIU⁻¹ at 525 mM, as detailed in Figure 8d and Table S10. The high QF at lower concentrations reflects Sys₈’s ability to maintain a sharp, well-defined SPR dip, essential for distinguishing slight variations in virus concentration. As concentration increases, however, the broadening of the SPR dip reduces QF, compromising resolution. The gradual decay in QF from 54.66 RIU⁻¹ to 37.45 RIU⁻¹ up to 400 mM still supports Sys₈’s effectiveness for low and moderate concentrations. This makes Sys₈ especially valuable for scenarios where clarity and precision are crucial, such as detecting small shifts indicative of early infection or changes in treatment efficacy.

In this context, the analyzed performance metrics of Sys₈ position it as a versatile biosensor for SARS-CoV-2 detection across a range of clinically relevant concentrations. Its high sensitivity, especially at low to moderate viral loads, makes it a prime candidate for early diagnostic applications. Furthermore, the observed linearity in sensitivity up to 400 mM (as shown in Figure S5) suggests that Sys₈ can reliably quantify viral load without the need for complex calibration models, which simplifies real-time monitoring in clinical environments.

3.7. Comparison with Previous Reports

To contextualize the performance of Sys₈ configuration, a comparison with other reported SPR-based systems for virus detection [33,34,35,36] was conducted.

Table 1. Comparison of SPR biosensors for viral detection, showing sensitivity, stability, linearity, responsivity, and detection limits. The optimized Sys₈ configuration is highlighted alongside similar systems.

Ref.	Sensor Architecture	Sensitivity (°/RIU)	Stability	Linearity (Concentration Range, mM)	Responsivity (°/mM)	Detection Limit (mM)
GMC Santos et al., 2021 [33]	Gold-antibody-based sensor for HAV	265.00	High	100–350	3.7°@300	100
Moznuzzaman et al., 2021 [34]	Nano-layered SPR for SARS-CoV-2	298.00	Moderate	200–500	4.0°@450	200
Hu et al., 2014 [35]	Gold–graphene SPR for porcine circovirus	280.00	Moderate–High	50–300	3.6°@300	50
Uddin et al., 2021 [36]	Ag-based multilayer sensor for SARS-CoV-2	320.50	Moderate	150–400	4.1°@400	150
This work	Graphene-based multilayer sensor for SARS-CoV-2	315.91	High	150–400	4.2°@400	150

presents key metrics for sensitivity, stability, linearity, responsivity, and detection limit, highlighting Sys₈’s competitive standing within the field of viral diagnostics.

Sys₈ demonstrated a high sensitivity of 315.91°/RIU, comparable to the Ag-based multilayer system developed by Uddin et al. [36] for SARS-CoV-2 detection, which exhibited a sensitivity of 320.50°/RIU. This level of sensitivity is critical for detecting low concentrations of viral particles, emphasizing Sys₈’s utility in applications requiring precise viral load monitoring, especially at early infection stages. Notably, Sys₈ outperformed other configurations, such as the gold-antibody-based SPR biosensor for hepatitis A virus (HAV) by Santos et al. [33], which achieved a sensitivity of 265.00°/RIU.

In terms of stability, Sys₈ displayed high operational consistency, attributed to the combined use of silicon nitride and graphene layers, which provide enhanced structural and optical robustness. This stability aligns with the high stability reported by Santos et al. [33] for their gold-antibody-based system but exceeds that of the nano-layered SARS-CoV-2 sensor by Moznuzzaman et al. [34], which displayed moderate stability. Such stability is crucial for sustained diagnostic applications where sensor durability under repeated measurements is essential.

Regarding linearity, Sys₈ exhibited a linear response across a concentration range of 150–400 mM, which is on par with the Ag-based multilayer sensor by Uddin et al. [36] and slightly narrower than the nano-layered SPR sensor by Moznuzzaman et al. [34], which reported a range of 200–500 mM. However, Sys₈’s linearity range covers a clinically relevant concentration spectrum, enabling accurate quantification of SARS-CoV-2 across various viral loads encountered in diagnostic settings.

The responsivity of Sys₈ was measured at 4.2° at 400 mM, slightly higher than the 4.1° achieved by Uddin et al. [36] for a similar concentration. This level of responsivity highlights Sys₈’s capability to detect incremental changes in viral concentration, an essential feature for real-time monitoring applications. This responsivity aligns well with clinical needs, facilitating rapid response to fluctuations in viral load. Lastly, Sys₈ demonstrated a detection limit of 150 mM, comparable to the Ag-based sensor by Uddin et al. [36] and lower than the detection limit of 200 mM reported by Moznuzzaman et al. [34] for their nano-layered configuration. A lower detection limit enables Sys₈ to detect minimal viral concentrations, making it particularly suitable for early diagnostic applications where low viral loads are expected.

3.8. Experimental Feasibility of Sys₈

The proposed Sys₈ configuration, incorporating a multilayer SPR structure with silver, silicon nitride, graphene, and a thiol-tethered ssDNA layer, is designed with both theoretical optimization and practical feasibility in mind. Experimental implementation of such a configuration is feasible using well-established fabrication techniques. For instance, the deposition of silver and silicon nitride layers can be achieved using physical vapor deposition (PVD) [37] or sputtering methods [38], which provide precise control over layer thickness and uniformity. Graphene can be transferred using established chemical vapor deposition (CVD) techniques [39], which yield a monolayer or few-layer graphene with excellent consistency across the sensor surface.

The thiol-tethered ssDNA layer, serving as a binding site for SARS-CoV-2 RNA, can be introduced through a straightforward self-assembly process [40]. Thiol-modified ssDNA strands bind covalently to the silver surface, forming a stable monolayer that enhances the biosensor’s specificity. The thickness of the ssDNA layer can be controlled by adjusting the ssDNA concentration and incubation time during this process. Optimizing these parameters helps achieve a uniform layer thickness, which is critical for reproducible sensor performance. Further, rinsing with PBS solution post-immobilization removes excess ssDNA, stabilizing the layer.

4. Conclusions

In summary, the optimized graphene-based SPR biosensor configuration developed in this study, denoted as Sys₈, demonstrates significant potential for the sensitive detection of SARS-CoV-2 across various concentrations. By incorporating layers of silver, silicon nitride, single-layer graphene, and a thiol-tethered ssDNA layer, the sensor effectively utilizes surface plasmon resonance to enhance sensitivity, specificity, and stability. The combination of these materials, particularly the use of graphene and silicon nitride, enhances light-matter interaction at the metal–dielectric interface, thereby achieving a balanced response in metrics critical for biosensing applications, including sensitivity enhancement, attenuation control, and FWHM management.

Particularly, the optimized Sys₈ configuration exhibited a high sensitivity of 315.91°/RIU at moderate virus concentrations (275 mM), proving especially suitable for early-detection scenarios. Additionally, it achieved a maximum shift (Δθ) of 4.2° and a peak detection accuracy of 0.60, both of which indicate precise responsiveness to refractive index changes induced by virus binding events. The sensor maintained a high figure of merit (54.66 RIU⁻¹) at lower virus concentrations (150 mM), further supporting its application in scenarios where early detection and minimal viral loads are vital.

A linear response was observed in sensitivity enhancement up to a SARS-CoV-2 concentration of 400 mM, beyond which a plateau effect suggested sensor saturation. This highlights the sensor’s effective operational range, which is particularly relevant for detecting clinically significant viral loads. Despite challenges associated with high attenuation at elevated concentrations, the performance metrics across different virus concentrations confirm Sys₈ as a robust candidate for SARS-CoV-2 detection in point-of-care diagnostics. Future studies should explore further material refinements and real-world testing to validate the applicability of Sys₈ in clinical settings.