The detection principle is based on the variation of extraordinary optical transmission (EOT) intensity that a gold nanohole array (Au-NHA) presents when a gold nanoparticle (Au-NP) of 100nm diameter is located in closevicinity of a nanohole. Our The Au-NHA chip is functionalized with anti-ADA antibodies (clone 15C7) on specific spots as inFigure 1.a and illuminated with 660nm light to generate the EOT via a plasmonic effect at the gold surface. This effectis enhanced by the nanohole array found on our the chip. Without any sample on it, the chip has a uniform transmission. Asample solution containing the target analyte (adalimumab), activated Au-NPs functionalized with detection antibodies(clone 3C2) and reaction buffer is put on the chip 1.b. The adalimumab protein will eventually bind to the captureantibodies fixed on the Au-NHA surface. The bounded adalimumab alone induce only very little change in intensity ofthe extraordinary plasmonic transmission. A small frequency peak shift is known to happen, but which cannot be detectedwith a standard camera (it would be the scheme of a label-free detection). What triggers the detection is Instead, the bindingof a Au-NP on top of the adalimumab, as in Figure 1.ctriggers the detection. The vicinity of this NP to the NHA induces a local changein refractive index, which strongly affects the plasmonic resonance peak present on the surface of the Au-NHA. Theintensity of transmitted light at the NP locations will be reduced, thus making it possible to digitally measure variation ofintensity on a far-field image (black dots will appear, representing a NP).

In our the assay, the capture antibodies are deposited with a spotting machine (Scienion sci-FLEXARRAYER S3) on thegold nanohole array (Au-NHA). The spots have a diameter of 100 μm and are spaced every 300 μm. The machine isprogrammed to spot both the capture antibody used for the assay and a mouse antibody that is used as a negativecontrol for reference. On the chip, a silicon rubber is placed to form a well. The sample is added inside and then around cover slip is used preventing the evaporation of the sample. The volume of the sample is around 20 μm. At themoment, the mixes are prepared by the user through pipetting steps but we fully intend to automatize the process infuture iterations of our prototype. Furthermore, to facilitate the insertion of the chip in our the device, a single use holder as

The reader’s dimensions (Figure 2) are estimated at 42x30x24 cm. It is composed mainly of two parts. The hexagonaltower contains the optical setup presented in Figure 2 as well as a slot for inserting the chip and a z-axis translationmount accessible to the user for focus adjustment. As for the base support, it contains the switching power supply forthe LED and the raspberry pi used for user-machine interaction. 3 EPFSens Figure 2: Our biosensor (left) and its optical system (right).

In order to provide an accessible device, we implemented an intuitive A user interface for our touchscreen. The is implemented and the userchooses is able to choose to refer back to previously recorded results or make a measurement. The measurement process is simplifiedby a provided tutorial displayed with instructions on the handling of the cartridge and the result acquisition. The user isalso guided during the focus adjustment before starting the measurement in order to ensure the best image quality forthe analysis. The implemented software is continuously printing the image on the screen, as the user is changing thez-axis translation mount for focus adjustment. As part of the development of our prototype, we intend it is intended to implement anautofocus algorithm so as to further simplify the user interaction.The main job of the software consists in acquiring the images of the assay over a certain time period and then it usesimage analysis algorithms to detect capture antibody spots locations, where the detection takes place. For the moment,as our The device is still at a prototype stage, we has the added the possibility for the user to visually check the position of the spotsand, if needed, to adjust the detected circle position, but this step would ultimately only take place in the backgroundwithout any user input.After spot detection, the camera takes regular captures of the spots, showing the increased binding of adalimumab onthe nanohole gold array over time, which is translated by the increasing number of black dots on the captured image.The software measures the variation of the image intensity in time and outputs the detected adalimumab concentration,based on a well established calibration curve. The result is saved and printed on the screen.