Description of an Advantageous Optical Label-Free Biosensing Interferometric Read-Out Method to Measure Biological Species

1. Introduction

A biosensor is a device to detect specific biological species or chemical compounds. Among the different types of biosensors, label-free optical biosensors function without tags or chemical amplification. This optical direct In-Vitro Diagnostic (IVD), unlike labeled IVD, offers a direct detection of biomolecules accumulated or recognized on a given sensing surface. However, in the absence of this chemical amplification, it is a challenge for label-free biosensing to achieve the same degree of sensitivity as those exhibited by standard enzyme-linked immunosorbent assay (ELISA). In this sense, significant optical label-free biosensors are reported performing well. For example, several target biomolecules (e.g., DNA, Proteins, viruses, Bacteria or Cells) detected from different optical biosensors are well described and reviewed in recent articles [1–4].

Some of the most attractive examples are those based on: surface plasmon resonance [5,6] or ring disk resonators [7–9], Mach-Zehnder Interferometers [10], photonic crystals [11–14], Young interferometers [15], porous silicon [16,17], slot waveguide [18–21], BICELLs [22–24], among others [25,26].

What previous label-free optical biosensors share in common is the use of photonic architectures to produce resonant or interferometric optical modes suitable for enhancing the biosensing signal. When a biological component is immobilized (bioreceptor) or recognized (target biomolecule) on the sensing surface of a photonic transducer, optical resonances change producing variations in some interrogation magnitudes. Thus, the biosensing system depends not only on the photonic transducer proprieties, but also on the read-out optical method employed, as essential part to the optical sensing system as a whole. In most of the biosensors mentioned above, similar optical detection methods are employed for reading photonic transducer signals such as wavelength or angle variation of a resonant peak or dip, phase shift variation, or amplitude [27–29], or subtraction of wavelength-based optical signals are also well described in porous Si devices. The aim of this scientific report is not to compare different biophotonic transducers, but rather to describe in detail an advantageous optical read-out method for improving the LoD of any biophotonic transducer like those reported in the scientific literature.

In this article, we report an advantageous interferometric optical read-out method to enhance the biosensing response of a given photonic transducer measuring the accumulation (immobilization or recognition) of biological components and their corresponding biofilm thickness. We explain and demonstrate in detail how the biosensing response can be significantly enhanced, in comparison with the classical monitoring methods (e.g., the interfering peaks' or dips' wavelength displacements in a spectrometry profile), by operating with the interferometric optical sensitive signals coming from the transducer, and specifically converting them to optical power for a certain spectral range that depends of the biophotonic transducer type employed. In other words, we describe in detail how by measuring the optical power carefully selecting a specific spectral band, where the interfere intensity given by the optical transducer employed is higher, the LoD is drastically improved in comparison with those methods based on spectral change such as a shift in peak or dip location. Furthermore, this method can avoid common dispersive and complex optical elements such as gratings, high-numerical aperture objectives or costly interferometers, which reduce implementing costs, and makes the read-out system independent on the dispersive element resolution (e.g., wavelength, wavenumber or angle resolution).

Although in this article we explain in detail the theory involved and the experimental results for only a photonic transducer model based on a simple Fabry-Perot interferometer (FPI), what we find remarkable is that this optical read-out method can be employed for many other optical biosensors. In fact, we think that our findings reported in this article may have relevant implications for improving the majority of label-free biosensor reported in the literature. We also believe this method may be an effective approach for integrating the reading system and photonic transducers.

2. Experimental Section

The Interferometric Optical Detection Method (IODM) is characterized by the use of two interferometric signals, which allows for the optical reading system to convert the changes caused by the optical transduction into a unique, sensitive variable of detection. Therefore, two interferometric measurements are used: a first interferometric optical reference, (I_Ref^{Interferometric}[wn₁, wn₂]), which represents the measured intensity modulated by a reference interferometer (IF_Ref); and a second interferometric signal measurement (I_Out^{Interferometric}[wn₁, wn₂]) observed in the sensing region of the signal interferometer (IF_Out). The sensing surface region of IF_Out is where changes produced by the biomolecular interaction takes place, for example, due to its functionalization by incorporating molecular receptors or because, already having incorporated molecular receptors, it has been used recognizing target molecules. The first and the second interferometric measurements can be taken sequentially or in parallel. If taken in parallel, the sensing observation region is physically different but it is interpreted with the same given the equivalence of the interferometric properties.

A transduction function (f_TRANS) is then constructed from the interferometric measurements, and analyzed to determine the biosensing response caused by the biological accumulation in the sensing observation region. Figure 1A shows a schematic representation of the IODM, where an input light is modulated by the two above-mentioned interferometers IF_Ref and IF_Out. IF_Ref generates the interferometric reference and IF_Out produces the interferometric signal coming from the biological accumulation. f_TRANS is the operation between both interferometric signals and delivers an unique biosensing variable.

Though, useful transduction functions results from IF_Out and IF_Ref interferometric signals, and IF_Out and IF_Ref may be any of the aforementioned photonic transducers, for simplicity in this article to demonstrate the concept we use: a FPI as photonic transducers model consisting of a thin layer of SiO₂ of 1,012 nm in thickness over a Si and a common transduction function f_TRANS resulting from a quotient (see Equation (1)), which lead to a reliable detection function for a given wavenumber range from wn₁ to wn₂:

Figure 1B shows the theoretical, modulated interferometric reflectance signals and how they operate by mean of f_TRANS to produce a detection function between 20,000 and 10,000 cm⁻¹. After having transformed the output light beams into numerical values, the transduction function can be mathematically represented (e.g., the wavenumber-dependent intensity or relative optical power). Thus, the transduction function, f_TRANS measures the degree of biomolecular accumulation (biosensing response) in the observation region of the signal interferometer IF_Out. Literature describing the reflectance output signal of an interferometer formed by a film layer of SiO₂ is easily found [24,30]. Later in this paper, we compare these theoretical predictions with experimental results by measuring different types of biomolecules. When biomolecules are not present in the sensing area of IF_Out, the output interferometric signals of IF_Ref and IF_Out are still the same, leading to an f_TRANS equal to one for the wavenumber range considered. However, if there is a small accumulation of biomolecules in the sensing area of IF_Out, f_TRANS changes significantly. For this theoretical simulation, we modeled biofilms ranging in thickness from 0.5 nm to 60 nm in order to simulate varying sizes of biomolecules (e.g., a coating is roughly 2.5 nm for BSA, 14 nm for anti-BSA, and 60 nm for a virus).

The theoretical f_TRANS behavior is shown in Figure 1B. To establish a single f_TRANS measurement parameter to determine the degree of biological accumulation change in the sensing area, several options can be used:

1.

A peak amplitude of function f_TRANS,

2.

A peak-to-peak amplitude of function f_TRANS,

3.

The change in slope in function f_TRANS by a pre-established wavenumber value,

4.

Considering I_Out and I_Ref of f_TRANS as their corresponding optical power (or irradiance) for a given wavenumber range [wn_a, wn_b].

For this paper, we analyzed Option 2—A peak-to-peak amplitude of function f_TRANS, in order to explain the IODM in a descriptive manner as proof of concept; and Option 4—Considering I_Out and I_Ref of f_TRANS as their corresponding optical power, for enhancing the biosensing response of a given photonic transducer.

Although options 1 to 3 can be used for explain the optical read-out method, due to the fact f_TRANS is constructed with interferometric signals coming by the biophotonic transducers, it must be remarked here that only option 4 enhances the LoD of the biosensing system. Thus, the novelty for enhancing the LoD is in choosing the appropriate wavenumber range for a given f_TRANS where to measure the optical power. This wavenumber band will depend mainly of the type of interferometric signals I_Out and I_Ref produced for the biophotonic transducers employed for the biosensing. Even more, if a wrong wavenumber range or large spectral band is chosen, the LoD not only does not improve, but the possibility of no sensing for the system at all also exists. However, we consider it appropriate to describe also option 2 for a better understanding of the optical sensing mechanism and then explain in detail option for subject of this article as a main novelty.

Option 2: Figure 1C illustrates the theoretical biosensing response of the method for the peak-to-peak signal amplitude of f_TRANS between 17,000 and 14,000 cm⁻¹, and the different observable, accumulated, biofilm thicknesses in the IF_Out. It can be observed a lineal response of this signal as the biofilm thickness increases in IF_Out. For this calculation, we considered a biofilm refractive index of 1.4 [31].

Option 4: we also studied the variation of f_TRANS considering the I_Out, and I_Ref (Equation (1)) as their corresponding P_Out and P_Ref (optical power or irradiance) from IF_Out and IF_Ref respectively, and for a given wavenumber range (wn_a, wn_b). The objective is to discover the wavenumber range (wn_a, wn_b) where to measure P_out and P_Ref of the interferometric signals IF_Out and IF_Ref of f_TRANS to significantly enhance the LoD.

We observe three spectral ranges where the optical power rises as the biofilm thickness increases specifically for the photonic transducer considered in this article (see Figure 2A). The observation wavenumber ranges (wn_a, wn_b) are: (13,192.6, 11,890.6) cm⁻¹, (16,447.4, 15,243.9) cm⁻¹ and (9,685.0, 18,587.4) cm⁻¹. In order to simulate f_TRANS for this case, we have to obtain the optical power of the IF_Ref and IF_Out for one of these wavenumber ranges considered. Thus f_TRANS for Option 4, can be calculated as the integral of reflectance signal of IF_Ref (P_Ref) divided by the integral of reflectance of the IF_Out (P_Out) for the particular wavenumber range (wn_a, wn_b). For this case, f_TRANS directly represents the relative optical power, which is directly correlated to the biofilm thickness accumulated in the IF_Out sensing surface. As abovementioned, in order to produce the best biosensing response, it is critical to select the proper wavenumber range (wn_a, wn_b) where f_TRANS produces a maximum of relative optical power.

Once the suitable location of the wavenumber range is selected, the transduction function f_TRANS makes it possible to assess the value of the increased relative power as a function of the biofilm thickness. A proper wavenumber range with a good signal-to-noise ratio (SNR) in our optical set-up is (13,192.6, 11,890.6) cm⁻¹(from 758 to 841 nm), where f_TRANS amplitude rises when the biofilm thickness increases. Therefore, the increased relative optical power (IROP) is a function of the biofilm thickness accumulated in the IF_Out, and this IROP can be calculated (see Equation (2)):

Therefore, the IODM is the IROP caused by the interference of the two interferometers IF_Ref and IF_Out. It is important to remark here that the chosen wavenumber range optimizes the sensitivity of the biosensing system (defined as the variation of the IROP as a function of the biofilm thickness) for the best SNR, and therefore the manner in which we can improve significantly the LoD. Figure 2B shows a schematic representation of the IODM for option 4 for the model FPI considered. The method for Option 4 directly obtains the biosensing curve just by reading the optical power of each interferometer. Figure 2C,D shows the biosensing response IROP (Equation 2) as a function of the biofilm thickness accumulated onto the IF_Out. A linear response for biomolecules smaller than 60 nm is observed. In fact, the theoretical biosensing sensitivity (measured as the slope of the sensing curve) is 0.975% nm⁻¹, and it is lower for biomolecules higher of 40 nm (see Figure 2C).

3. Results and Discussion

For all of the experiments, the sensing areas of IF_Out and IF_Ref were both square cells of 1 cm² size and the FPI employed was formed by 1,012 nm of SiO₂ over Si as substrate (see Figure 3). All measurements were taken with a FT-VIS-IR spectrometer whose spot size was adjusted by the edge apertures to the 1 cm² of the sensing area.

4. Conclusions

In this article we assess a new read-out IODM in an effort to significantly improve the LoD of a FPI used as photonic transducer model, and tested it for two different immunoassays. It is based on operation of the optical power of interferometric signals produced by a given transducer specifically for a spectral band, and not the LoD's dependence on the wavelength or wavenumber resolution of the spectrometer employed. Thus, we have demonstrated that this improvement is at least as good as two orders of magnitude for the same transducer and experimental conditions with this novel approach based on the IODM based on IROP if we select the optimum specific wavenumber range (or spectral band) to optimize sensitivity (IROP variation as a function of the biomolecules concentration) where the SNR is lower. This IODM is a viable and promising alternative to other biosensing methods reported in the scientific literature.

The theoretical predictions and experimental results demonstrate the advantages of using this optical biosensing read-out method for measuring biological species such as proteins and viruses, and estimating their biofilm thickness. With this IODM based on IROP we were capable of enhancing the LoD of a really simple photonic transducer based on a single layer of SiO₂ from 450 ng·mL⁻¹ to 1.37 ng·mL⁻¹ recognizing anti-BSA, certifying the benefits of using this IODM. Moreover, we can calculate the anti-BSA and BSA biofilm thickness, achieving similar values to those reported in the literature, demonstrating the reliability of this method. Furthermore, we tested the IODM for a more complex experiment detecting easily SU-8 thin film deposited onto the IF_Out sensing surface, the immobilization of the Rotavirus antigen, and the recognition of its corresponding antibody, achieving competitive LoD figures.

Finally, we can conclude that the IODM based on IROP is a promising approach for several reasons. First, IODM can function for other interferometric photonic transducers such as those based on ring or disk resonators, Mach-Zehnder interferometers, Young interferometers, BICELLs, and others reported for label-free biosensing. Secondly, the fact that the LoD does not depend on the resolution of costly optical dispersive elements such as gratings (wavelength), interferometers in the reader (wavenumber) or high numerical aperture objectives (angle); make IODM as a promising alternative for developing compact lab-on-a-chip devices, with plenty of room in the clinical settings.

It is expected that the LoD figures for other photonic transducers can be significantly enhanced according the results presented in this paper by measuring the IROP within the spectral band defined by the photonic transducer employed.