2.1.

Chemicals and Reagents

The polymers PCDTBT and ${\mathrm{PC}}_{70}\mathrm{BM}$ were obtained from Luminescence Technology Corp. (Taipei, Taiwan). Poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS, Baytron®P) was purchased as an aqueous dispersion from H.C. Starck Trading Company Ltd. (Shanghai, China). Indium tin oxide (ITO) deposited glass slides (thickness of ITO, 100 nm; surface sheet resistance, $\sim 10\mathrm{ohms}/\mathrm{sq}$) were purchased from CSG Holding Co., Ltd. (Shenzhen, China). For the microfluidic chip, PDMS precursor and curing agent (Sylgard 184) were obtained from Dow Corning Holding Co., Ltd (Shanghai, China). $N$-[(3-trimethoxysilyl)propyl]ethylenediaminetriacetic acid, Na salt (TMS-EDTA; 50% in water) was supplied by IS Chemical Technology Ltd. (China); while (3-aminopropyl)-trimethoxysilane (APTMS; 97%) was from Sigma-Aldrich (St-Louis). SuperBlock™ phosphate-buffered saline (PBS, pH 7.2), $N\text{-}\mathrm{ethyl}\text{-}{N}^{\prime }$-(3-dimethylaminopropyl)-carbodiimide hydrochloride (EDC), $N$-hydroxysuccinimide (NHS), StartingBlock™ Blocking Buffer in PBS with 0.05% Tween-20, streptavidin horseradish peroxidase (HRP) conjugate, SuperSignal® ELISA femto luminol/enhancer solution and SuperSignal® ELISA femto stable peroxide solution were from Thermo Fisher Scientific (Shanghai, China). KeXing Biological technology Co. Ltd. (China) provided human TSH (recombinant) and its monoclonal antibody (mouse). Human tumor necrosis factor-alpha ($\mathrm{TNF}\text{-}\alpha$) and human interleukin-6 (IL-6) were supplied by Yapei Biotech. Ltd. (Shanghai, China) and Wuhan AmyJet Scientific Inc. (Wuhan, China), respectively. Polyclonal antiserum of TSH was obtained from an immunized mouse via traditional methods. After harvest, this antiserum was tested by standard dot-blot tests, which confirmed that no serious cross-reaction to the aforementioned monoclonal antibody has occurred. Biotinylation of the antiserum was conducted by Biotin Labeling Kit-${\mathrm{NH}}_2$ (Dojindo Laboratories, Kumamoto, Japan). Ultrapure water ($18\mathrm{M\Omega }/\mathrm{cm}$) was produced by a Milli-Q® system (Millipore, Milford) and used for all aqueous solutions and assay buffers.

2.2.

Device Description and Sensing Principle

A scheme of the OPD-integrated microfluidic device is shown in Fig. 1(a). The device was formed by incorporating the polycarbazole photodetector comprising the $\mathrm{ITO}/\mathrm{PEDOT}\text{:}\mathrm{PSS}/\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}/\mathrm{LiF}/\mathrm{Al}$ diode architecture [Fig. 1(b)] to a hybrid microfluidic chip of PDMS/Au-coated glass. An active area of $0.16{\mathrm{cm}}^2$ in the detector was aligned with a 30-μl volume reaction chamber on the hybrid microchip. This defined the detection zone where chemiluminescent sandwich immunoassays were developed on the Au surface. The Au-coated glass served as both a surface to immobilize antibodies and as a reflective backside substrate. The chemiluminescent oxidation of luminol by HRP labels in the presence of peroxide was used to generate $\sim 425\text{-}\mathrm{nm}$ light whose intensity was proportional to the amount of analyte targeted by the immunoassay. This light was transduced as a photocurrent signal by the OPD. Photons emitted in the CL reaction were absorbed by the $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ blend heterojunction material, resulting in photogenerated electrons and holes collected at the corresponding opaque LiF/Al cathode and transparent ITO anode.

Fig. 1

Layout of the polycarbazole organic photodetector (OPD) integrated microfluidic device. (a) Three-dimensional scheme of the whole integrated device. The sensing area ($4\times 4{\mathrm{mm}}^2$) of the photodetector matches to the dimensions of the reaction chamber. (b) Polycarbazole diode design and sensing principle. A chemiluminescent immunoassay is employed to detect thyroid-stimulating hormone (TSH), with the generated light detected by the OPD. mAb, monoclonal antibody; B-pAb, biotinylated polyclonal antibody; SA-HRP, streptavidin-horseradish peroxidase conjugate. (c) Top view of the polydimethylsiloxane (PDMS) microchannel layer and side view of the OPD integrated microfluidic device (not to scale). The device is assembled by attaching the glass side of the OPD to the PDMS slab which is bonded to the PDMS-Au-glass hybrid microchip.

2.3.

Photodetector Preparation

The polycarbazole detectors were prepared on ITO-coated glass substrates. After the substrates were pretreated by ultraviolet ozone, PEDOT:PSS (used as received) was deposited on top of the substrates via spin coating. The spin speed was varied between 1200 and 4500 rpm, resulting in film thicknesses ranging from 80 nm down to 25 nm. PCDTBT (used as received) was dissolved in chloroform to make $4\mathrm{mg}/\mathrm{ml}$ solution, following by blending with ${\mathrm{PC}}_{70}\mathrm{BM}$ (used as received). The blends with a mass ratio of $1:4$ in chloroform were then spin coated onto the PEDOT:PSS layer. The thickness of the resulting $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ active layer ranged from 70 to 180 nm. An Ambios XP-100 profilometer was used to measure the layer thicknesses. After deposition of the active layer, the substrates were transferred into a glove box filled with high-purity ${\mathrm{N}}_2$ and dried at 60°C for 1 h. Furthermore, the LiF/Al electrode ($\sim 100\mathrm{nm}$) was deposited onto the polymer film by thermal evaporation under a pressure of $3\times {10}^{-4}$ Pa using a shadow mask. Following fabrication, the OPDs were encapsulated with a customized single-sided pressure-sensitive barrier foil. This ensures stability against oxygen and water vapor during device operation. Finished devices were diced into $25\times 52\mathrm{mm}$ slides.

2.4.

Microfluidic Setup and Chip Integration

The microfluidic chip was fabricated by standard PDMS casting with a SU-8 master template.³² A mixture of a Sylgard 184 precursor and a curing agent at a ratio of $10:1(\mathrm{w}/\mathrm{w})$ was degassed in a vacuum and poured onto the master template. After curing at 60°C for 1 h, the 800-μm thick PDMS cast was peeled off from the master and the structure of an 800-μm deep chamber was obtained using a scalpel blade. The microchannels connecting the inlets and chamber were 250-μm wide and 300-μm deep, whereas the channel connecting the outlet and chamber was 250-μm wide and 650-μm deep [Fig. 1(c)]. This microchannel layer was permanently attached to a 1-mm thick PDMS slab after exposing the contact surfaces to oxygen plasma.

The PDMS set was further attached to the Au-coated glass slide by means of carboxylamine coupling chemistry. A Pyrex 7740 glass wafer was previously sputtered with a 200-nm Au layer using a 20-nm Cr film as an adhesion layer. Prior to surface modification, the Au substrate was cleaned in $5:1$ ${\mathrm{H}}_2{\mathrm{SO}}_4/{\mathrm{H}}_2{\mathrm{O}}_2$ solution, dried in a ${\mathrm{N}}_2$ atmosphere, and exposed to UV ozone for 5 min. Clean substrate was immersed into a 10% ($\mathrm{v}/\mathrm{v}$) solution of TMS-EDTA in water for 2 h in order to develop carboxyl-terminated functional groups on the surface. After three washes with ultrapure water, the TMS-EDTA-modified substrate was subjected to 50-nM NHS and 200-mM EDC for 30 min at room temperature. The bonding surface of the PDMS chip was modified with amine-terminated groups after immersion in an ethanolic solution of 10% APTMS ($\mathrm{v}/\mathrm{v}$) for 1 h. The aminated substrate was then washed three times in ethanol. Both the modified PDMS and gold substrates were dried under ${\mathrm{N}}_2$ for immediate bonding. The PDMS and Au surfaces were finally sealed at room temperature for 1 h. Irreversible attachment is achieved via NHS-EDC coupling chemistry. Also, the NHS-EDC functionalization of the Au surface within the reaction chamber allows covalent binding of antibody.³³

Fluidic access holes were then added to the PDMS-Au hybrid microchip to create two channel inlets (750-μm wide), one air bleeder (750-μm wide) and one channel outlet (720-μm wide). For assembly of the integrated device, the glass side of the polycarbazole photodetector was permanently attached to the lid of the PDMS chip [Fig. 1(c)]. The glass substrate of the OPD also served as the world-to-chip interface; it was comprised of drilled access holes coinciding with the outlet, air bleeder, and inlets in the PDMS layer, whereby fluidic reservoirs were connected to the entrances of the microchannels. The reservoirs were polyether ether ketone (PEEK) capillary tubes (purchased from IDEX Health & Science, Shanghai, China) with 125-μm inner diameter (ID) and 510-μm outer diameter.

2.5.

CL Measurement and Immunoassay Procedure

Figure 2 illustrates the experimental setup used for the CL measurements. A micro gear pump mzr®-4605 (IDEX Trading Co. Ltd, Shanghai, China) operated by a PC was used to suck reagents into the OPD-integrated microfluidic device. The delivery of reagents was controlled by a valve system comprising three OMNIFIT© eight-way valves (Shanghai Beion Medical Technology Co., Ltd., China). Waste was collected by a vacuum flask connected to the pump through 762-μm ID PEEK tubing (IDEX). This tubing was also employed to connect the flask and reagent reservoirs to the corresponding OMNIFIT valve. On the other hand, the capillary reservoirs in the inlets, air bleeder, and an outlet of the integrated device were connected to the corresponding valve using 508-μm ID tubing. The air bleeder was used in this setup to remove small gas bubbles likely generated at the beginning stage of the reagent loading that may scatter the CL signal. During fluid flow operation, the pressure inside the flask generated by the pump did not exceed 50 kPa in all experiments. The photocurrent from the OPD was measured with a source measure unit (SMU, model 236, Keithley Instruments Inc., Cleveland, Ohio), and the recorded data were transferred to a PC via a USB-GPIB interface adapter (KUSB-488B, Keithley).

Fig. 2

Schematic of the experimental apparatus for the chemiluminescence (CL) immunoassay experiments. It shows the polycarbazole photodetector integrated microfluidic device connected to the readout equipment and fluid flow control system. The assay components are selectively delivered to the integrated device using external valves. HRP, horseradish peroxidase; PBS, phosphate-buffered saline. SuperSignal working solution represents the luminol/peroxide solution.

This setup was used for the CL immunoassay as depicted in Fig. 1(b). All assay components were prepared from 0.01-M PBS for immediate use. After the microchannels and reaction chamber were flushed with PBS, a 50-μl aliquot of $0.1\text{-}\mu \text{g}/\mathrm{ml}$ anti-TSH monoclonal antibody was loaded into the chamber through the inlet 1 of the integrated device [Fig. 1(c)]. The pump was stopped after 90 s to allow the incubation of the antibody solution for 2 h within the chamber. The antibody was covalently immobilized to the chamber surface. After unbounded antibody was washed out from the microchip, the surface was blocked with StartingBlock™ blocking buffer and rinsed with PBS for 5 min. A 100-μl aliquot of a TSH standard solution was then added to the microchip via inlet 1. The content was incubated within the blocked chamber at room temperature for 15 min. Further, 100 μl of biotinylated polyclonal antibody was driven into the chip for 2 min; it was injected through inlet 2 to avoid contamination. Prior to use, the biotinylated antibody was purified by AllPrep™ protein procedure kit (Qiagen, Valencia). After the integrated device was rinsed with PBS, the immune complex interacted with streptavidin-HRP conjugate at concentrations ranging from 0.005 to $0.025\mu \text{g}/\mathrm{ml}$. Subsequently, SuperSignal working solution, which was prepared by mixing equal parts of luminol/enhancer and stable peroxide solution, was transferred to the device via inlet 1 and was allowed to react with HRP. The light emitted in the CL reaction was thus detected by the OPD. All the assay experiments were performed in the dark. For the assay repetitions, the immune complex was dissociated using a glycine-HCl elution buffer (50 mM, pH 2.2) following the photocurrent measurement. The CL immunoassay was then restarted with the incubation of the TSH sample.

3.1.

Optimization of Photodetector Design

Figure 3(a) shows the transient chemiluminescent signal of the polycarbazole detector for the detection of $10\mathrm{ng}/\mathrm{ml}$ TSH. The first CL immunoassays were performed in 400-μl volume reservoirs made of PDMS that were attached between the Au substrate and the glass slide of the OPD. These reservoirs were held in place on the Au substrate using a poly(methylcrylate) holder.^34,35 The transient response of the OPD measured under short-circuit conditions was recorded from 0 to 10 min after the addition of the SuperSignal working solution. The signal reached a steady state within $\sim 1\min$.

Fig. 3

Experimental optimization of the polycarbazole diode design. Polycarbazole devices with varied $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ (photoactive layer) and PEDOT:PSS (hole transport layer) thicknesses were fabricated and tested. (a) Transient photocurrent response due to CL detection measured as a function of photoactive layer thickness. (b) External quantum efficiency (EQE) spectra for different $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ thicknesses. (c) Background current measured at different bias voltages as a function of hole transport layer thickness. (d) Average photocurrent obtained from the plateau region of the transient response for different PEDOT:PSS thicknesses. The PEDOT:PSS was 50 nm in (a) and (b), whereas the $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ was 120 nm in (c) and (d). The CL detection experiments for (a) and (d) were conducted with $10\mathrm{ng}/\mathrm{ml}$ TSH.

The first set of CL measurements was conducted to optimize CL detection sensitivity of the polycarbazole detector. The sensitivity of OPD-based photodetection is enhanced by lowering dark current and increasing external quantum efficiency (EQE) for the desired wavelength range.³⁶ Optimizing the thickness of the active layer and hole transport layer represents a promising approach to the improvement of dark current and EQE characteristics of an OPD.^37,38 Thus, polycarbazole devices with varied $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ and PEDOT:PSS thicknesses were fabricated in this work. Experimental optimization of the OPD design is conducted via comparison of dark current and EQE performance between the fabricated devices.

Photodetectors with active layers ranging from 70 to 180 nm were first compared. The PEDOT:PSS layer was fixed at 50 nm. Figure 3(a) plots the chemiluminescent signals measured with varied active layer thicknesses. The average photocurrent at the plateau of the signal (time period between 3 and 7 min) was determined for the comparison. A significant increase in the plateau current ($I_{\mathrm{p}}$) was obtained for the $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ thickness of 120 nm, which was approximately twofold higher than the 70-nm-thick layer. The changes in $I_{\mathrm{p}}$ were in agreement with the results for EQE [Fig. 3(b)]. Here, EQE was determined from the ratio of the current density collected at device electrodes to the flux density of incident photons:³⁹

The EQE values were calculated for incident wavelengths between 400 and 500 nm at intervals of 5 nm. An EQE of 62% was obtained for the 120-nm-thick $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ device under monochromatic light irradiation at 425 nm. The decreased responses observed in Fig. 3(b) for the $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ thicknesses between 120 and 180 nm were mainly attributed to the reduced efficiency of charge transport in these thicker layers.⁴⁰ Figure 3(c) shows the background level ($I_0$) of the polycarbazole device as a function of the PEDOT:PSS thickness from 25 to 80 nm, where the $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ layer was maintained at 120 nm. $I_0$ was measured before the assay components were added to the PDMS reservoirs such that only the effect from the dark current will be considered. A minimum background current was observed for the 40-nm-thick PEDOT:PSS when no bias voltage was applied, which corresponded to an approximately sixfold improvement over the layer thickness of 50 nm. $I_{\mathrm{p}}$ was also enhanced by the 40-nm-thick PEDOT:PSS [Fig. 3(d)]. The decreased thickness of PEDOT:PSS from 80 to 40 nm improved the detector response, which may have resulted from an increased EQE (Ref. 37) and/or an enhanced chemiluminescent light absorption by the OPD.

3.2.

Optimization of Assay Conditions

Further immunoassay tests were conducted with the optimized OPD comprising the 120-nm-thick $\mathrm{PCDTBT}\text{:}{\mathrm{PC}}_{70}\mathrm{BM}$ and 40-nm-thick PEDOT:PSS. The OPD was integrated with the PDMS-Au hybrid microchip for the tests. The CL immunoassays performed within the integrated device were optimized before conducting quantitative TSH detection. Four concentrations of streptavidin-HRP conjugate were tested. The signal-to-noise ratio, defined here as the current ratio of $I_{\mathrm{p}}$ to $I_0$, was used for the analysis. Enhanced signal-to-noise ratio was encountered for a conjugate concentration of $15\mathrm{ng}/\mathrm{ml}$ [Fig. 4(a)]. The results were consistent for four TSH concentrations. The optimized streptavidin-HRP concentration may have led to reduced shot noise of the detection system, which may be related to fluctuations in the photon flux density.³⁶ This noise is important for systems designed to detect low-magnitude photocurrents.

Fig. 4

Optimization of assay conditions for the microfluidic CL detection. (a) Signal-to-noise ratio of the polycarbazole photodetector integrated microfluidic device for a range of streptavidin-horseradish peroxidase conjugate (SA-HRP) concentrations. Four concentrations of TSH were used for the analysis. (b) Detection of $50\mathrm{ng}/\mathrm{ml}$ TSH varying the volumetric flow rate used for delivery of luminol/peroxide solution into the integrated device.

The influence of flow rate on photocurrent measurement was also analyzed. Immunoassays were conducted to determine the optimal flow conditions for the delivery of the SuperSignal working solution into the chip. An analyte concentration of $50\mathrm{ng}/\mathrm{ml}$ was used for the tests. $I_{\mathrm{p}}$ was determined in independent experiments testing flow rates between 20 and $100\mu \text{l}/\min$. The data in Fig. 4(b) indicated a decrease in $I_{\mathrm{p}}$ with decreasing flow rate values. The reduced CL signals might have been caused by insufficient time to complete the filling of the reaction chamber with the working solution before the signal plateau was reached ($\sim 1\min$). Although flow rates of 80 and $100\mu \text{l}/\min$ led to enhanced photocurrent responses, the significant fluctuations found in the CL signals prevented the use of these flow rates for later experiments. Thus, a flow rate of $55\mu \text{l}/\min$ was found to be optimal as it provided an ideal compromise between high photocurrent response and signal stability. Notably, no fluid leakage was encountered even at the highest flow rate, indicating the robustness of the OPD-integrated microfluidic device.

3.3.

Quantitative Detection Tests

Quantitative CL immunoassay detection was conducted under optimal conditions as described above. Figure 5(a) shows the current–voltage curves of the various TSH concentrations prepared in PBS. The photocurrent data were determined from the plateau region of the chemiluminescent signal by applying a range of bias voltages to the OPD. Upon incubation of the integrated device with TSH, the photocurrent remarkably changed with cortisol concentration. The photocurrent from the detected $5\mathrm{ng}/\mathrm{ml}$ analyte was $\sim 100\text{-}\mathrm{fold}$ higher than that from a nonanalyte concentration. The current measurements were limited to short-circuit conditions to ensure minimum background levels [Fig. 3(c)] in further quantitative detection tests. To determine the calibration curve, the photocurrent data were normalized by plotting $[(I_{\mathrm{p}}-I_0)/I_0]$ against the TSH concentration [Fig. 5(b)]. Linearity was found in the range of 0.03 to $10\mathrm{ng}/\mathrm{ml}$ with a correlation coefficient of 0.995. The analytical sensitivity, determined from the slope of the linear region, was $\sim 68\mathrm{pg}/\mathrm{ml}$. The detection limit is estimated to be $<30\mathrm{pg}/\mathrm{ml}$ as indicated by the result of $3\times$ standard deviation of five blank measurements. This result is below the $500\mathrm{pg}/\mathrm{ml}$ lowest limit reported for a P3HT OPD-based optical biosensor.²³ The analytical performance of the integrated device is dependent upon the sensitivity of the OPD. Further decrease in dark current may be achieved by reducing the photoactive area.²² However, this may imply reduction in the dimensions of the reaction chamber and consequently the quantity of luminol/peroxide working solution for the analysis. A decrease in light intensity generated by the CL reaction will result in a reduced photocurrent response by the OPD.

Fig. 5

Quantitative CL immunoassay detection of TSH. (a) Current-voltage profiles at different TSH concentrations. (b) Calibration curve obtained from the normalized photocurrent for a range of TSH concentrations in phosphate-buffered saline. Inset represents the linear region of the calibration curve. All measurements of current in (b) were conducted with no bias applied. The error bars represent standard deviation for triplicate assays.

3.4.

Specificity and Reproducibility Tests

The specificity of the method was tested using various reagent blanks, cytokines $\mathrm{TNF}\text{-}\alpha$ and IL-6 as negative controls and $2\mathrm{ng}/\mathrm{ml}$ TSH as a positive control. The anti-TSH monoclonal antibody, blocking buffer, streptavidin-HRP conjugate and luminol/peroxide working solution were used for the reagent blanks. Figure 6 plots the results of the specificity tests. No distinguishable interference from any of the prepared blanks on the photocurrent measurements was observed. Furthermore, the variation of photocurrent for $\mathrm{TNF}\text{-}\alpha$ and IL-6, biomarkers also related with thyroid disease and stress disorders,^31,41 was negligible compared with TSH.

Fig. 6

Specific detection of TSH by the polycarbazole photodetector integrated microfluidic device. The specificity tests were performed with respect to various reagent blanks, analyzing the interference from monoclonal antibody (mAb), blocking buffer (BF), streptavidin-horseradish peroxidase conjugate (SA-HRP) and luminol/peroxide working solution. The negative controls were the cytokines human tumor necrosis factor-alpha ($\mathrm{TNF}\text{-}\alpha$) and human interleukin-6 (IL-6) while the positive control was $2\mathrm{ng}/\mathrm{ml}$ TSH. The error bars represent standard deviation for triplicate tests.

The CL assay variability was also characterized. Intra-assay variability, obtained from the relative standard deviation, was calculated from the five assay repetitions for $2\mathrm{ng}/\mathrm{ml}$ TSH concentration using two polycarbazole devices prepared in the same batch. Inter-assay variability was determined in the same way as intra-assay variability but testing three OPDs prepared in three batches. The values inferior to 3% and 7% for intra-assay and inter-assay, respectively, indicated good detection reproducibility of the developed microfluidic system.

3.5.

Analysis of Clinical Samples

The analysis of complex biological samples is commonly conducted to further evaluate the feasibility of a biosensor system.^42–44 To investigate the feasibility of the developed OPD-integrated microfluidic device for POCT applications, six serum samples of adult subjects were tested in parallel with six devices. The TSH levels in the serum samples were within the range of 0.035 to $0.12\mathrm{ng}/\mathrm{ml}$ as detected by a commercially available enzyme-linked immunosorbent assay (ELISA) method (Abcam®, Cambridge). All tests were conducted with 100 μl of diluted serum. The procedures for the commercial ELISA including assay preparation and sample dilution were performed according to the instructions of the manufacturer.

Upon incubation of the integrated devices with the corresponding diluted serum samples, the measured photocurrents were normalized via $[(I_{\mathrm{p}}-I_0)/I_0]$ and the TSH levels were thus estimated using the calibration curve shown in Fig. 5(b). The estimated concentrations were then compared with those obtained with commercial ELISA. To this end, the absolute percentage difference in TSH concentration between the two methods of detection was determined and plotted for each serum sample (Fig. 7). This difference did not exceed 15% which indicated good correlation between the measurements by the integrated device and those made by ELISA.

Fig. 7

Absolute percentage variation in six serum concentrations of TSH detected by two methods. Each variation value was obtained from the ratio of the TSH concentration estimated with the polycarbazole photodetector integrated microfluidic device to that detected using a standard enzyme-linked immunosorbent assay method.

The integrated device is originally designed and developed for single use. The low cost of both the PDMS-Au hybrid microchip and polycarbazole OPD ensures feasibility of a POCT disposable device. In addition, the PDMS microfluidic substrate guarantees sufficient assay miniaturization and reduction of reagent consumption. Most POCT applications typically demand fast sample analysis.⁴⁵ The analyte detection described in this study required 15 min for incubation of a sample and then 1 min for equilibration with luminol/peroxide working solution. A reduction in sample incubation time could be an immediate solution to further decrease the overall duration of the detection process.