Organic Electrochemical Transistor Immuno-Sensors for Spike Protein Early Detection

The global COVID-19 pandemic has had severe consequences from the social and economic perspectives, compelling the scientific community to focus on the development of effective diagnostics that can combine a fast response and accurate sensitivity/specificity performance. Presently available commercial antigen-detecting rapid diagnostic tests (Ag-RDTs) are very fast, but still face significant criticisms, mainly related to their inability to amplify the protein signal. This translates to a limited sensitive outcome and, hence, a reduced ability to hamper the spread of SARS-CoV-2 infection. To answer the urgent need for novel platforms for the early, specific and highly sensitive detection of the virus, this paper deals with the use of organic electrochemical transistors (OECTs) as very efficient ion–electron converters and amplifiers for the detection of spike proteins and their femtomolar concentration. The electrical response of the investigated OECTs was carefully analyzed, and the changes in the parameters associated with the transconductance (i.e., the slope of the transfer curves) in the gate voltage range between 0 and 0.3 V were found to be more clearly correlated with the spike protein concentration. Moreover, the functionalization of OECT-based biosensors with anti-spike and anti-nucleocapside proteins, the major proteins involved in the disease, demonstrated the specificity of these devices, whose potentialities should also be considered in light of the recent upsurge of the so-called “long COVID” syndrome.


Introduction
The uncontrolled and devastating spread of coronavirus (COVID-19 disease) that struck the world in the last two years has led to the need to find new, rapid and effective solutions to manage and reduce its diffusion. The optimization of oro-nasopharyngeal tests for the early detection of proteins involved in SARS-CoV-2 infection has been a focus of the global scientific community [1,2]. One of the main target proteins used to identify the presence of SARS-CoV-2 is represented by the spike protein, which, located at the virus surface, plays a fundamental role in penetrating host cells and initiating infection. At this stage, however, it has been recognized that the commercially available infection Acetone, isopropanol, 3-mercaptopropionic acid (3-MPA), 11-mercaptoundecanoic acid (11-MUA), 1-ethyl-3-(3 dimethylaminopropyl) carbodiimide (EDC), Nhydroxysulfosuccinimide sodium salt (sulfo-NHS), ethanolamine, Tween 20 and phosphate-buffered saline (PBS) solution (C = 10 mM) were all purchased from Sigma-Aldrich (Missouri, MO, USA). Bovine serum albumin (BSA) was purchased by Microgem (Pozzuoli, Italy).

OECT Fabrication
OECTs were fabricated using a well-established protocol already described in detail elsewhere [24,[29][30][31], which is briefly reviewed here. A solution of PEDOT:PSS layer (Clevios PH 1000, doped with 5 volume % ethylene glycol, 0.1 volume % dodecyl benzene sulfonic acid, and 1 wt% of GOPS (3-glycidyloxypropyl)trimethoxysilane) was spincoated on the source and drain electrodes and subsequently patterned via photolithography (microchemicals AZ9260 resist) and etching with O2 plasma. A polydimethylsiloxane (PDMS) chamber, with an internal volume of 150 µL, was aligned with the PEDOT:PSS channel and irreversibly bonded on the surface (Figure 1a  Electrodes were fabricated by e-beam evaporation (ULVAC EBX-14D) through the deposition of Ti/Au films (thickness 10 nm/100 nm) on Si wafer (100) finished with 1 µm of thermal oxide. The final device aspect ratio is W/L = 6 mm/200 µm = 30 (where W is the width and L is the length of the channel). The planar gate electrodes were made by a Electrodes were fabricated by e-beam evaporation (ULVAC EBX-14D) through the deposition of Ti/Au films (thickness 10 nm/100 nm) on Si wafer (100) finished with 1 µm of thermal oxide. The final device aspect ratio is W/L = 6 mm/200 µm = 30 (where W is the width and L is the length of the channel). The planar gate electrodes were made by a Ti/Au-10 nm/100 nm film. The latter was deposited by the evaporation systems already used for the source and drain electrodes, on a Si wafer (100) finished with 1 µm of thermal oxide (SiO 2 ). Finally, they were diced in 5 × 5 or 5 × 10 mm 2 and functionalized using the protocol described in the following.

Gate Electrode Biofunctionalization
Gate surface was firstly cleaned in an ultrasonic bath with acetone and isopropanol for 10 min each, then the cleaned electrode was immersed in a 10 mM 3-MPA and 11-MUA (10:1 molar ratio) ethanol solution for 24 h at room temperature. This step allowed for the creation of the chemical SAM layer, exposing carboxylic groups. Then, such groups were activated by immersing the electrode in a water solution of EDC (200 mM) and NHS (50 mM) for 2 h at 25 • C. After each step, the surface was rinsed with distilled water to remove any residues and dried with air flux.
To bond the antibodies on the surface, the gate electrode was immersed in anti-spike protein solution (C = 10 nM in PBS) for 2 h at 25 • C, and then washed with 0.1% Tween PBS solution. After binding with antibodies, the gate electrode was treated with ethanolamine (1 M in PBS) and with BSA (1 µM in PBS) to saturate the remaining nonspecific sites (Figure 1c). To investigate OECT selectivity, gate electrodes were also functionalized with anti-NP protein solution at the same concentration (C = 10 nM in PBS).
The quality of the biofunctionalization protocol was tested (Section 3.1 with Figure 1 andFigure S1) by immersing gate electrodes in a solution with secondary antibodies conjugated to Alexa Fluor ® 488 dye (donkey anti-rabbit-Alexa Fluor 488 at dilution of 1:200, purchased by Thermofisher, Waltham, MA, USA) for 1 h at room temperature. Then, the samples were observed with confocal laser scanning microscopy (CLSM-Zeiss LSM5 Pascal) technique to assess the binding between the primary and secondary antibodies. Moreover, gate electrodes functionalized with anti-spike antibodies were analyzed by Fourier Transform Infrared Spectroscopy (FTIR) techniques. The latter was performed using Agilent Cary 630 FTIR (Agilent Technologies, Inc., Santa Clara, CA, USA), working with an ATR module in the range between 4000 and 650 nm. Before any FTIR spectrum acquisition, the samples were carefully dried on a very thin N 2 flow for 5 to reduce the moisture. During the analysis, more than 10 spectra were collected to find the best working conditions, i.e., environmental parameters, number of scans, resolution, overall instrument working energy and to perform several comparisons with different backgrounds.

OECT Characterization
A Keithley two-channel multimeter (Keithley 2602B) operating by Labview code was used to characterize the electrical response of the OECT devices. Output curves were obtained by applying a voltage between gate and source, V GS , ranging between −0.6 and 0.6 V with a step of 0.1 V, and measuring drain-source (I DS ) and gate-source (I GS ) current, as a function of drain-source voltage, V DS (in the range between 0 and −0.6 V). Then, curves were transferred, reporting I DS by varying V GS between −0.6 and 0.6 V with a step of 0.025 V at V DS = −0.3 V, were recorded for all the investigated functionalized gate electrodes (i.e., with anti-spike or anti-NP proteins). In the transfer curves, each point was acquired every 2 s (∆t = 2 s). For comparison, curves at different times (∆t = 0.5 and 5 s) are reported in the Supplementary Information (Figure S3). This protocol eas applied by using the gate electrodes functionalized with antibodies to drive the OECT, while PBS 10 mM was employed as electrolyte. Then, the functionalized gates were incubated with a spike-RBD protein from femto to nanomolar concentrations for about 20 min at room temperature and, after careful rinsing with bi-distilled water, the same electrodes were used to repeat the OECT measurement again in PBS 10 mM.
Before starting any biosensing experiment, the OECT channels were immersed in bidistilled water for at least 2 h. Moreover, a set of preliminary transfer curves were recorded to stabilize the OECT response in PBS 10 mM. A data analysis of transfer curves could describe the OECT electrical response in terms of the current modulation values expressed as (I DS − I 0 )/I 0 , where I 0 is the baseline current, and transconductance g m = δI DS /δV GS , which represents the transduction efficiency that is related to the channel current slope as a function of the voltage V GS . For all the experiments with bio-functionalized gate electrodes, particular care was paid to keep the immersed gating area fixed in place (A v = 8 mm 2 ).
The mean and standard deviation values for all transfer curves were achieved by acquiring five consecutive transfer curves and discarding the first one (generally affected by a major level of variability) to extract the various statistical parameters.

Validation of the Gate Electrode Biofunctionalization Process
A validation step of the biofunctionalization process was initially carried out through immunofluorescence to demonstrate the presence of antibodies on the gate surface. In particular, gate electrodes previously incubated with primary antibodies were immersed in solutions with secondary antibodies conjugated to Alexa Fluor ® 488 dye. Since such molecules have a specific affinity to the primary antibodies, they can unequivocally demonstrate their presence on the gate surface (Figure 1d,e). Before the testing step, the electrodes were rinsed with PBS to remove any impurities and dried with cleaned air. Hence, confocal laser scanning microscopy (CLSM) images recorded before ( Figure 1d) and after ( Figure 1e) the functionalization step confirmed the effectiveness of the adopted protocol. A further analysis of the gate functionalization was conducted by the Fourier Transform Infrared Spectroscopy (FTIR) technique. In particular, FTIR spectra were obtained from surfaces before and after the antibody functionalization process. As shown in Figure S1, Sample A was taken from a bare cleaned Au layer; meanwhile, Sample B was recorded for a gold surface once the functionalization process with an anti-spike protein was completed. The spectrum obtained on Sample A is fully consistent with a bare gold surface, as no relevant peak was observed in the considered range. Conversely, Sample B exhibited many different peaks all across the spectrum, as the antibody itself is a protein rich in functional groups that are visible in the IR range. In this context, FTIR spectroscopy can provide important information regarding proteins' secondary structure [32]. According to the literature, increasing the size of the molecules induces a minor sensitivity, but important information about the presence of the antibody and the success of the process can still be achieved [33]. In Figure  S1, two main peaks in the secondary structure are present. The former, namely, Amide A, is located at 3252 nm and indicates the N-H stretching in resonance; meanwhile, the latter, namely, Amide I, is found at 1654 nm and shows the C=O stretching vibration. These peaks provide strong evidence of the correct bonding of the antibody on the underneath layer, representing significant markers of the presence of proteins [33,34].

OECT Initial Characterization
Before the biosensing experiments, the OECTs investigated in this work were carefully characterized to set the optimal operating conditions. Figure 2 reports a general view of a typical response achieved with an un-functionalized (bare) gold gate electrode. As shown from the output curves in Figure 2a (with V GS ranging between −0.6 V and 0.6 V, while V DS was varied between 0 and −0.6 V), the devices behave correctly as depletion-mode transistors with the I DS current (always in the range of mA) decreasing (in absolute value) following the application of positive V GS . In the output curves, only the linear and the triode regions can be observed, and this feature is to be ascribed to the gating condition that was adopted. Indeed, because of the size of the gate-immersed area (~8 mm 2 , see Materials and Methods section) and considering the volumetric capacitance (C V~4 0·F·cm −3 ), which can be associated with the entire electrolyte/PEDOT:PSS distributed interface, the response of the investigated OECT is mainly dictated by the electric double-layer (EDL) capacitance (C G ) at the gate-electrolyte interface (i.e., C G C V ) [35,36]. In the framework of the model of Bernards and Malliaras [37], this condition provides a very large value (higher than 1 V) of the so-called pinch-off (V p ) voltage, which determines the achievement of the current saturation phenomenon when V GS = 0 V and V DS < −|V p |. framework of the model of Bernards and Malliaras [37], this condition provides a very large value (higher than 1 V) of the so-called pinch-off (Vp) voltage, which determines the achievement of the current saturation phenomenon when VGS = 0 V and VDS < −|Vp|. Figure 2b shows a set of transfer curves recorded at different VDS, while, in Figure 2c, the same curves are represented as normalized with respect to the IDS value measured at VGS = −0.6 V. Finally, Figure 2d reports the corresponding transconductance (gm) values (gm = δIDS/δVGS). These plots clearly show that by increasing the absolute value of the applied VDS, both the overall IDS modulation and gm tend to rise. In particular, the transconductance receives values in the range of mS, confirming the excellent capabilities of OECT while operating as amplifying elements [17]. It is important to outline that, at a larger |VDS|, the gm curves exhibit the characteristic presence of a broad peak, for which the position and maximum value can be modified as a function of VDS. For all the transfer curves, moreover, a consistent rise in gm when VGS approaches 0.6 V can be observed, with this behavior becoming more and more pronounced at a larger |VDS|. As shown in Figure    These plots clearly show that by increasing the absolute value of the applied V DS , both the overall I DS modulation and g m tend to rise. In particular, the transconductance receives values in the range of mS, confirming the excellent capabilities of OECT while operating as amplifying elements [17]. It is important to outline that, at a larger |V DS |, the g m curves exhibit the characteristic presence of a broad peak, for which the position and maximum value can be modified as a function of V DS . For all the transfer curves, moreover, a consistent rise in g m when V GS approaches 0.6 V can be observed, with this behavior becoming more and more pronounced at a larger |V DS |. As shown in Figure Figure S2b), being four orders of magnitude smaller than the corresponding I DS current throughout the analyzed V GS range. Since the specific OECT response is determined by the synergistic combination of an electronic and an ionic circuit featuring typically different time dynamics as a function of the gating conditions and the active channel size [35,37], the effect of different sweep times in the transfer curve recording was also analyzed. Basically, various values of the time delay (∆t = 0.5, 2 and 5 s) between the V GS application (step ∆V GS = 0.02 V) and the measurement of the corresponding I DS value were taken, with the goal of assessing the related impact on the transfer curves and the corresponding g m values.
According to the results summarized in Figure S3, if the measurements are performed too fast (namely, with ∆t = 0.5 s), the devices are unable to reach a steady condition. Hence, the eventual modulation and g m values (Figure 3b) are consistently reduced. In consideration of these findings, for all the measurements discussed hereafter, ∆t = 2 s was selected as being able to provide a good trade-off between the speed and accuracy of the experiments. In the absence of Faradaic reactions at the gold gating surface, the gate current (I GATE ) plot in Figure S3c confirms the fundamental capacitive nature (i.e., displacement current) of I GATE which, indeed, is strongly dependent on the overall sweep time (i.e., the faster the measurement, the larger the I GATE ∝ δV GS /δt).
of the detection abilities of spike-RBD protein. Significantly, under the application of VDS = −0.3 V, IGATE values remain constantly lower than 200 nA ( Figure S2b), being four orders of magnitude smaller than the corresponding IDS current throughout the analyzed VGS range. Since the specific OECT response is determined by the synergistic combination of an electronic and an ionic circuit featuring typically different time dynamics as a function of the gating conditions and the active channel size [35,37], the effect of different sweep times in the transfer curve recording was also analyzed. Basically, various values of the time delay (Δt = 0.5, 2 and 5 s) between the VGS application (step ΔVGS = 0.02 V) and the measurement of the corresponding IDS value were taken, with the goal of assessing the related impact on the transfer curves and the corresponding gm values. According to the results summarized in Figure S3, if the measurements are performed too fast (namely, with Δt = 0.5 s), the devices are unable to reach a steady condition. Hence, the eventual modulation and gm values (Figure 3b) are consistently reduced. In consideration of these findings, for all the measurements discussed hereafter, ∆t = 2 s was selected as being able to provide a good trade-off between the speed and accuracy of the experiments. In the absence of Faradaic reactions at the gold gating surface, the gate current (IGATE) plot in Figure S3c confirms the fundamental capacitive nature (i.e., displacement current) of IGATE which, indeed, is strongly dependent on the overall sweep time (i.e., the faster the measurement, the larger the IGATE ∝ δVGS/δt).

OECT Detection of Spike-RBD Proteins
For all the experiments devoted to the detection of SARS-CoV-2 spike RBD protein, before any incubation step, the OECTs were preliminarily characterized in PBS. The so-

OECT Detection of Spike-RBD Proteins
For all the experiments devoted to the detection of SARS-CoV-2 spike RBD protein, before any incubation step, the OECTs were preliminarily characterized in PBS. The so-achieved transfer curves (as average of four curves consecutively measured with V DS = −0.3 V; see the Materials and Methods section) were then compared with analogous data acquired after the incubation (for 20 min) in solutions containing variable concentrations of the spike-RBD protein. According to this protocol, and similar to other experimental studies [25], the gate electrode represents the sensing and disposable components of the overall device, while the PEDOT:PSS channel can be used as an ion-to-electron amplifying transducer in combination with different gating surfaces.
Following the aforementioned procedure, Figure 3a-c shows the results of three experiments at 57 fM, 57 pM and 57 nM spike-RBD protein concentrations (i.e., 1.4 pg/mL, 1.4 ng/mL, 1.4 µg/mL, respectively). Firstly, it should be mentioned that a shift in the current baseline (i.e., the almost constant I DS current measured at very negative V GS values) was observed in most cases, but without any clear correlation with the incubation step and the corresponding spike-RBD protein concentration. This effect is likely related to underlying ion diffusion processes, which can slightly modify the overall conductivity of the PEDOT:PSS channel over time during the different measurements.
More interestingly, as suggested by the acquired transfer curves and their normalized counterparts, shown in Figure 3 (top and bottom panels, respectively), the incubation process tends to also modify the general shape of the transfer curves. The most evident feature is the reduction in the I DS slope, which is mainly observed in the intermediate V GS range between slightly negative values and 0.3/0.4 V. For very negative V GS values, the I DS slope remains rather low, and is quite unchanged following the incubation procedure. Conversely, as previously discussed (see Section 3.2), when V GS exceeds 0.3 V, the I DS growth with V GS becomes very steep and, in this case, turns out to be less reproducibly affected by the antibody-antigen binding occurrence. A further and clearer evidence of the I DS slope decrease in the intermediate V GS region, related to the incubation procedure, is given in Figure S4a For the sake of completeness, Figure 4 summarizes the results of an experiment where a gate functionalized with anti-SARS-CoV-2 NP (anti-NP) antibody was incubated in the spike-RBD protein solution with nanomolar (57 nM) concentration. At odds with the observations commented for Figure 3, here, the incubation step produced less characteristic changes and the related transconductance values were found to be slightly increased ( Figure S4d) in the intermediate V GS region following the incubation process. As a whole, these observations confirm that the OECT behaves coherently following the non-specific interaction of the anti-NP protein and the spike-RBD protein.
of the spike-RBD protein. According to this protocol, and similar to other experimental studies [25], the gate electrode represents the sensing and disposable components of the overall device, while the PEDOT:PSS channel can be used as an ion-to-electron amplifying transducer in combination with different gating surfaces.
Following the aforementioned procedure, Figure 3a-c shows the results of three experiments at 57 fM, 57 pM and 57 nM spike-RBD protein concentrations (i.e., 1.4 pg/mL, 1.4 ng/mL, 1.4 µg/mL, respectively). Firstly, it should be mentioned that a shift in the current baseline (i.e., the almost constant IDS current measured at very negative VGS values) was observed in most cases, but without any clear correlation with the incubation step and the corresponding spike-RBD protein concentration. This effect is likely related to underlying ion diffusion processes, which can slightly modify the overall conductivity of the PEDOT:PSS channel over time during the different measurements.
More interestingly, as suggested by the acquired transfer curves and their normalized counterparts, shown in Figure 3 (top and bottom panels, respectively), the incubation process tends to also modify the general shape of the transfer curves. The most evident feature is the reduction in the IDS slope, which is mainly observed in the intermediate VGS range between slightly negative values and 0.3/0.4 V. For very negative VGS values, the IDS slope remains rather low, and is quite unchanged following the incubation procedure. Conversely, as previously discussed (see Section 3.2), when VGS exceeds 0.3 V, the IDS growth with VGS becomes very steep and, in this case, turns out to be less reproducibly affected by the antibody-antigen binding occurrence. A further and clearer evidence of the IDS slope decrease in the intermediate VGS region, related to the incubation procedure, is given in Figure S4a For the sake of completeness, Figure 4 summarizes the results of an experiment where a gate functionalized with anti-SARS-CoV-2 NP (anti-NP) antibody was incubated in the spike-RBD protein solution with nanomolar (57 nM) concentration. At odds with the observations commented for Figure 3, here, the incubation step produced less characteristic changes and the related transconductance values were found to be slightly increased ( Figure S4d) in the intermediate VGS region following the incubation process. As a whole, these observations confirm that the OECT behaves coherently following the nonspecific interaction of the anti-NP protein and the spike-RBD protein.    Figure 5a reports the I DS modulation values given by the absolute value of (I F − I 0 /I 0 ), where I F and I 0 are the I DS currents in the transfer curves recorded at V GS = 0.6 V and V GS = −0.5 V, respectively. Here, the mean values from different experiments are indicated for any spike-RBD protein concentration, with the standard deviations being the error bars. The bottom panel of Figure 5a represents the percent variation in any I DS modulation value with respect to that measured prior to the corresponding incubation (BLANK). As shown, this analysis reveals the progressive reduction in the I DS modulation at an increasing spike-RBD protein concentration, with the percent decrease approaching 30% for the concentrations in the nanomolar range. It is worth highlighting that the I DS modulation value is largely used in the literature as a main parameter in biosensing experiments focused on the use of OECT [25,38] and that, similarly to what is observed here, percent variations ranging between a few units and tens per cent are typically reported for analyses dealing with antigen concentrations that are variable over several orders of magnitude [21,29]. In general, the I DS modulation reduction can be explained by considering that the incubation process provides a further decrease in the C G (i.e., the capacitance between gate and electrolyte) values, consequently attenuating the gate electrode's overall ability to modulate the I DS current. the slope of the transfer curves, Figure 5b reports the transconductance values estimated at VGS = +0.2 V (top panel), VGS = 0 V (middle panel) and VGS = −0.2 V (bottom panel) as function of the spike-RBD protein concentration. Here, the decreasing trend is only clea for the gm values at VGS = +0.2 V with a variation of up to 30%, as achieved for the nano molar concentration, in comparison to the initial value. This plot highlights the major sen sitivity of the OECT response versus the antigen concentration in the intermediate VG region, in comparison with what is observed for lower values of VGS (i.e., 0 and −0.2 V).
Moreover, in the graphs of Figure 5a and Figure 5b (top panel), the IDS modulation and transconductance at VGS = +0.2 V values obtained for the experiments carried out with anti-NP protein functionalized gate electrode before and after incubation in a solution with 57 nM of spike-RBD protein are also included. In this case, because of the lack of antigen-antibody complex, the IDS modulation and gm values before and after the incuba tion are completely indistinguishable. Moreover, in the graphs of Figure 5a,b (top panel), the I DS modulation and transconductance at V GS = +0.2 V values obtained for the experiments carried out with anti-NP protein functionalized gate electrode before and after incubation in a solution with 57 nM of spike-RBD protein are also included. In this case, because of the lack of a antigen-antibody complex, the I DS modulation and g m values before and after the incubation are completely indistinguishable.
With the aim of setting up an alternative and potentially more direct protocol to demonstrate the detection of the spike-RBD protein from the analysis of the OECT response, we elaborated the data through a different procedure focused on the modelling of the transfer curves region between −0.5 and 0.3 V. Hence, all the transfer curves in this V GS range were modelled by a simple second-order polynomial (I DS = α + β·V + γ·V 2 ). Figure S5 in the Supplementary File presents examples of this fitting procedure applied to the transfer curves initially introduced in Figures 3 and 4a. The excellent quality of the fitting curves demonstrates the properness of the approach, which can model both the initial linear and the following quadratic dependence of I DS on V GS .
Hence, the values of the β and γ fitting parameters averaged for all the experiments performed at different spike-RBD protein concentrations are plotted in Figure 6a,b, respectively. In particular, while resembling the behavior of g m at V GS = 0 and −0.2 V, even if the β fitting parameter (Figure 6a) is significantly reduced in the lowest concentration (i.e., fM) range in comparison with the BLANK test (i.e., prior to any incubation step [36]), it does not exhibit a regular dependence on the spike-RBD protein concentration, particularly for the values in the pico-and nano-molar ranges. Conversely, the γ fitting parameter quite satisfactorily reproduces the behavior achieved for the g m value estimated at the single V GS = 0.2 V, displaying a monotonously decaying behavior as a function of the spike-RBD protein concentration for the anti-spike protein functionalized gate electrode. In this case, no significant variations were observed before and after incubation in a solution with 57 nM of spike-RBD protein solution when a gate functionalized with anti-NP protein was employed. Hence, this approach confirms that the V GS region in which I DS shows a square dependence on V GS is the most sensitive to the changes produced by the gate incubation process. It should be outlined that the progressive occurrence of the g m reduction in the OECT response, as a consequence of a protein detection at the gate/electrolyte interface, has been reported in other works, confirming that transconductance behavior is strictly related to the specific V GS region [25]. This feature points to the need for alternative theoretical frameworks that can provide more insightful descriptions of the OECT electrical behavior [39,40]. range were modelled by a simple second-order polynomial (IDS = α + β • + • ). Fig  ure S5 in the Supplementary File presents examples of this fitting procedure applied t the transfer curves initially introduced in Figures 3 and 4a. The excellent quality of th fitting curves demonstrates the properness of the approach, which can model both th initial linear and the following quadratic dependence of IDS on VGS.
Hence, the values of the β and γ fitting parameters averaged for all the experimen performed at different spike-RBD protein concentrations are plotted in Figure 6a,b, r spectively. In particular, while resembling the behavior of gm at VGS = 0 and −0.2 V, even the β fitting parameter (Figure 6a) is significantly reduced in the lowest concentration (i.e fM) range in comparison with the BLANK test (i.e., prior to any incubation step [36]), does not exhibit a regular dependence on the spike-RBD protein concentration, particu larly for the values in the pico-and nano-molar ranges. Conversely, the γ fitting paramete quite satisfactorily reproduces the behavior achieved for the gm value estimated at the sin gle VGS = 0.2 V, displaying a monotonously decaying behavior as a function of the spik RBD protein concentration for the anti-spike protein functionalized gate electrode. In th case, no significant variations were observed before and after incubation in a solution wit 57 nM of spike-RBD protein solution when a gate functionalized with anti-NP protein wa employed. Hence, this approach confirms that the VGS region in which IDS shows a squar dependence on VGS is the most sensitive to the changes produced by the gate incubatio process. It should be outlined that the progressive occurrence of the gm reduction in th OECT response, as a consequence of a protein detection at the gate/electrolyte interfac has been reported in other works, confirming that transconductance behavior is strict related to the specific VGS region [25]. This feature points to the need for alternative theo retical frameworks that can provide more insightful descriptions of the OECT electric behavior [39,40]. In summary, in this work, we assessed the selective detection of spike-RBD protein down to femto-molar range by using OECTs based on PEDOT:PSS channels. In additio to the overall IDS modulation values, our analysis revealed that the electrical paramete In summary, in this work, we assessed the selective detection of spike-RBD proteins down to femto-molar range by using OECTs based on PEDOT:PSS channels. In addition to the overall I DS modulation values, our analysis revealed that the electrical parameters associated to the transconductance (i.e., the slope of the I DS vs V GS transfer curves) and the V GS region (between 0 and 0.3 V) of the OECT response are more clearly affected by the gate incubation process. The overall time for the testing procedure is about 30 min, with 20 min being required for the incubation step and the remaining time needed to perform the electrical analysis consisting of the recording of a number of consecutive transfer curves to be averaged. Given the discussed findings, at this stage, the sensitivity performances is estimated to be approximately in the range of few tens of nanomolars of the antigen concentration (corresponding to few hundreds of ng/mL of the spike-RBD protein). Further improvements can be achieved by optimizing the functionalization protocol or using alternative probes for antigen recognition [25]. Actually, as reported by a large number of works discussed in the literature, the sensitivity performances of devices conceived for the spike protein detection are strongly dependent (i.e., over several orders of magnitudes) on the specific sensing schemes [1,2]. It should be also mentioned that, in the analyzed configuration, the organic active channel immersed in PBS can be used for multiple experiments, in combination with different properly functionalized gate electrodes upon their incubation in the solution of interest. This approach simplifies the use of these devices which, owing to their low applied voltages, are suitable for further miniaturization and can be integrated in a compact and portable measurement set-up.
Going beyond the precise diagnosis of coronavirus, all these features are very promising to the development of cheap, fast, and user-friendly tools for point-of-care analysis where OECT could be embedded in microfluidic-OECT-integrated platforms.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/bios13070739/s1, Figure S1: FITR spectra recorded for a gold surface before (SAMPLE A) and after (SAMPLE B), showing the complete functionalization process with anti-spike antibody; Figure S2: I GATE curves corresponding to the output (a) and transfer curves (b) reported, respectively, in Figure 2a