This is the accepted version of the following article: Quesada-González D., Stefani C., González I., de la Escosura-Muñiz A., Domingo N., Mutjé P., Merkoçi A.. Signal enhancement on gold nanoparticle-based lateral flow tests using cellulose nanofibers. Biosensors and Bioelectronics, (2019). 141. 111407: - . 10.1016/j.bios.2019.111407, which has been published in final form at https://dx.doi.org/10.1016/j.bios.2019.111407 © https://dx.doi.org/10.1016/j.bios.2019.111407. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ Signal enhancement on gold nanoparticle-based lateral flow tests using cellulose nanofibers Daniel Quesada-Gonzáleza, Christina Stefanib, Israel Gonzálezc, Alfredo de la Escosura-Muñiza, Neus Domingob, Pere Mutjéc, Arben Merkoçia,d* a Nanobioelectronics & Biosensors Group, Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193 Barcelona, Spain. b Advanced Atomic Force Microscopy Laboratory, Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193 Barcelona, Spain. c LEPAMAP Research group, University of Girona, C/Maria Aurèlia Capmany, 61, 17071 Girona, Spain. d ICREA, Institució Catalana de Recerca i Estudis Avançats, Pg. Lluís Companys 23, 08010 Barcelona, Spain. *arben.merkoci@icn2.cat Abstract Lateral flow paper-based biosensors merge as powerful tools in point-of-care diagnostics since they are cheap, portable, robust, selective, fast and easy to use. However, the sensitivity of this type of biosensors is not always as high as required, often not permitting a clear quantification. To improve the colorimetric response of standard lateral flow strips (LFs), we have applied a new enhancement strategy that increases the sensitivity of LFs based on the use of cellulose nanofibers (CNF). CNF penetrate inside the pores of LFs nitrocellulose paper, compacting the pore size only in the test line, particularly near the surface of the strip. This modification retains the bioreceptors (antibodies) close to the surface of the strips, and thus further increasing the density of selectively attached gold nanoparticles (AuNPs) in the top part of the membrane, in the test line area, only when the sample is positive. This effect boosts in average a 36.6% the sensitivity of the LFs. The optical measurements of the LFs were carried out with a mobile phone camera whose imaging resolution was improved by attaching microscopic lens on the camera objective. The characterization of CNF into paper and their effect was analyzed using atomic force microscope (AFM) and scanning electron microscope (SEM) imaging techniques. the membrane only on the test area, where the recognition 1. Introduction antibodies are dispensed. To accomplish this milestone, in this The World Health Organization (WHO) urges the work we have included for the first time cellulose nanofibers development of affordable, sensitive, selective, user-friendly, (CNF) in the test area. CNF are a nanomaterial often used as rapid and robust, equipment-free and derivable to the end-user reinforcement additive on papermaker pulps (González et al., devices (ASSURED criteria) (LaBarre et al., 2011). Lateral flow 2012, 2013, 2014, 2017; Alcalá et al. 2013; Delgado-Aguilar et strips (LFs) are paper-based biosensors that emerge as powerful al., 2015, 2016; Tarrés et al., 2016). This nanomaterial can be devices since they fulfill all the demands (Wong and Tse, 2009; produced as a gel that is easily dispensed and dried inside the Parolo and Merkoçi, 2013; Quesada-González and Merkoçi, paper pores. In addition to the structural modification, CNF are 2015; López-Marzo and Merkoçi, 2016; Zhan et al., 2017; biocompatible with antibodies, thus increasing the areas where Quesada-González, 2018). Paper substrate is inexpensive, they can be attached, so that more of them are retained near the recyclable, tunable and allows the storage of bioreceptors (e.g. paper surface, where it is best appreciated the color of transducer antibodies [Wu et al., 2014]) and nanomaterials (Quesada- particles (Fig. 1A). Gold nanoparticles (AuNPs) with average González and Merkoçi, 2015; Zhan et al., 2017) in dry state. diameter of 18.0 ± 0.5 nm (Fig. S1) have been chosen as These capabilities make LFs suitable for point-of-care (PoC) transducer particles since they are easy to synthetize (Turkevich diagnostics (Burns, 2002; Kratz and Lewandrowski, 2003; et al., 1951), bioconjugate (Shyu et al., 2001; De La Escosura- Quesada-González and Merkoçi, 2018). However, in some cases Muñiz et al., 2010; Parolo et al., 2013c; Rivas et al., 2015; Ngu et LFs may present not enough sensitivity levels, which does not al., 2019) and have a strong red color readily detectable (Fu et al., allow or makes it difficult to perform quantitative analysis. 2016; Quesada-González et al., 2018). Furthermore, AuNPs have Therefore, many studies and signal enhancement strategies have demonstrated their applicability in the evaluation of real samples been proposed during last years, as they are: the addition of extra with LFs (Shim et al., 2009; Xie et al., 2015; Quesada-González amplification steps (Fu et al., 2010, 2011; Fridley et al., 2012; et al., 2018; Liu et al., 2018). Our LFs design is represented in Rodriguez et al., 2016), secondary chemical reactions (Anfossi et Fig. 1B. al., 2013; Parolo et al., 2013a; Chapman et al., 2015; Loynachan The functionality of our LFs devices is finally strongly et al., 2018), architecture modifications (Parolo et al., 2013b; dependent on the color contrast detection of the test stripe lines. Rivas et al., 2014), new labels (Quesada-González et al., 2019) or In this sense, mobile phones become important support tools in even exploring alternative signal translation methods beyond the nowadays’ biosensing (Mudanyali et al., 2012; Coskun et al., optical response (Tang et al., 2009; Marquina et al., 2012; Li et 2013a, 2013b; Roda et al., 2014, 2016; Berg et al., 2015; al., 2014; Zhu et al. 2014; Khlebtsov et al., 2019). Quesada-González and Merkoçi, 2016; Álvarez-Diduk et al., The porosity of paper is a key factor regarding the sensitivity 2017; Kühnemund et al., 2017; You et al., 2017; Brangel et al. of the strips (Wong and Tse, 2009). Small pores will lead to 2018), not only due the capacity to record and store the data, but higher sensitivity, as we demonstrated in a previous work (Rivas also because of the portability, cloud storage, accessibility and et al., 2014), nevertheless the sample may have difficulties to software capabilities that these devices offer. In this work we flow, increasing the assay time and the probability of having a attach microscopic lenses (Blips Microlenses [Smart Micro flow stop or membrane defects, further provoking low Lenses, online reference]) to the mobile phone camera improving reproducibility. Herein, we propose to decrease the pore size of the quality of close-up images taken of the test area in LFs, therefore increasing the number of pixels that can be read and 2 demonstrating the operability of the device in conjugation with CNF within nitrocellulose pores (from one to six times; i.e. 0.46 standard camera-based detection. to 2.76 %, respectively). After drying the membrane overnight at room temperature, goat anti-human IgG antibody (1.0 mg/mL in 2. Experimental section PB) was dispensed over the previous line, forming the test line 2.1. Materials. (TL). Chicken anti-goat antibody (1.0 mg/mL in PB) was dispensed at 5.5 mm of distance from the TL (Fig. 2), forming Tetrachloroauric acid (HAuCl4), bovine serum albumin the control line (CL). The detection pad was dried at 37°C for 2 (BSA), goat anti-human IgG antibody, human IgG, trisodium hours and stored at room temperature under dry conditions. citrate, sucrose, Tween 20, phosphate buffer saline (PBS) tablets As represented on Fig. 2, the conjugate pad was assembled and the reagents to prepare phosphate buffer non-saline (PB; onto the adhesive card with a 1 mm overlap over the detection sodium phosphate basic and dibasic) and borate buffer (BB; pad, on the farthest side from the detection lines. The sample pad sodium tetraborate and boric acid) were purchased from Sigma was assembled overlapping the conjugate pad at the end of the Aldrich. Chicken anti-goat antibody was purchased from Abcam. card. An absorption pad (untreated cellulose membrane) was Nitrocellulose membrane (HF180), cotton membrane assembled on the other end of the card, overlapping the detection (CFSP001700), glass fiber (GFCP00080000) and supporting pad, at 4 mm away from the CL. The strips were cut to a width of adhesive cards were purchased from Millipore. 6 mm using a guillotine and stored at room temperature with drying pearls. CNF (0.92 %) were provided by LEPAMAP group, fabricated following a reported protocol (Delgado-Aguilar et al., 2015) by 2.5. Assay performance. using commercial bleached pine tree pulp as raw material, provided by Ence: Celulosas y Energía S.A. Human IgG solutions, from 0.01 to 1.00 µg/mL, were prepared in pH 7.4 PB (buffer that was also used as blank Blips Lens Basic Kit, from Smart Micro Optics (Smart Micro sample). 200 μL of each concentration were dropped on the Lenses, online reference), was purchased in a Kickstarter sample pad of the strips. After 5 minutes, TL and CL were campaign. observed. Images of the strips were taken at 6 mm of distance using a Moto Z mobile phone camera (13 megapixel) in which 2.2. Synthesis of gold nanoparticles. Micro Lens, from Blips Lens Basic Kit, were sticked on the Following Turkevich method (Turkevich et al., 1951), AuNPs objective (Fig. S2). A 3D-printed tool was made using BCN-3D of approximately 20 nm diameter were synthesized by citrate Sigma printer (melting polylactic acid filament) in order to keep reduction of HAuCl4. 50 mL of a 0.01 % (w/v) solution of the distance between the strips and the camera constant (video HAuCl4 were prepared in Milli-Q water. This solution was taken V1). to boiling point and 1.25 mL of 1 % (w/v) sodium citrate solution Image J software (Álvarez-Diduk et al., 2017; Kühnemund et was added under continuous vigorous stirring. The solution was al., 2017; Quesada-González et al., 2018) was used to quantify kept boiling for 10 more minutes and then allowed to cool down the TL intensity of color in the images. The intensity of white to room temperature. The AuNP solution was adjusted to a pH of pixels (maximum value 255, being 0 the blackest color) was 9.0 by addition of 0.01 mM BB. AuNPs were characterized by measured in 8-bit black-and-white format reconverted images transmission electron microscope (TEM, Tecnai F20) as shown (Fig. S3 illustrates the software and the area which was selected in Fig. S1. in a TL). Data was normalized into % color where 0 % and 100 2.3. Gold nanoparticles conjugation. % equal to 255 and 0 values respectively by the equation: 100 µL of goat anti-human IgG antibody (100 μg/mL in PB) % color = 100-(100 · IMGJ)/255 were added to 1.5 mL of AuNP solution and incubated for 20 Where “IMGJ” is the value measured with the software (from minutes at room temperature. BSA solution (0.1 % (w/v), 100 0 to 255). µL) was then added to the mixture, which was incubated for another 20 minutes and then centrifuged at 4°C (18000 rcf, 20 2.6. Characterization of CNF on nitrocellulose. minutes). The supernatant was discarded and the pellet was After the assay, two positive LFs were selected and the areas reconstituted in 0.5 mL BB (2 mM, pH 7.4) containing 10 % corresponding to the TL were cut for its study. The sample sucrose. concentration added in both strips was the same, being one strip 2.4. Lateral flow strips preparation. made with CNF gel on the TL and the other without. Both samples were observed on scanning electron microscopy (SEM; The conjugate pad was prepared by dipping the glass fiber FEI Magellan 400L XHR). into the AuNP-antibody conjugate suspension. The fiber was dried in a vacuum chamber for 3-4 hours, and then stored in The same samples were observed in atomic force microscopy barrier zip-lock pouches with drying pearls at 4 ºC. The sample (AFM). AFM imaging was performed with a MFP-3D Asylum pad was prepared by dipping the cotton membrane in PBS buffer AFM (Oxford Instruments, Scotts Valley, CA). In all the containing 0.05 % Tween 20 and 5 % BSA. The pad was dried in experiments, PPP-EFM tips (Nanosensors; Schaffhausen, the oven at 60°C for 2 hours and stored in barrier zip-lock Switzerland) with a stiffness constant k=2 N/m and coated with pouches with drying pearls at 4 ºC. The detection pad was made Ptlr5 were used. Multi-Frequency AFM is based in the use of assembling nitrocellulose membrane onto a supporting adhesive multiple excitation modes with different characteristic card. frequencies. In this case, the fundamental excitation mode was used together with the second eigenmode. CNF gel (0.92%) was diluted by half (to 0.46% CNF) with Milli Q water to make it less viscous and able to be dispensed In simple bimodal AFM operation mode, the tip is excited at using a Biofluidix Biospot Workstation. It was dispensed over both frequencies, with amplitudes A2 < A1. A remarkable benefit nitrocellulose membrane forming a line. The line was dispensed of MF bimodal imaging is the fact that the forces between the tip repeatedly on the same position to increase the concentration of and the sample are very small while still allowing a clear non- destructive differentiation of composition, which makes it a very phone camera with magnification lens attached on the objective. convenient way to investigate samples composed of different Results obtained are shown on Fig. 4. It can be observed how, the materials. The excitation at the fundamental mode was used to higher the concentration of CNF within nitrocellulose, the higher measure the topography in amplitude-modulation AFM, while is the color enhancement of TL up to a CNF concentration of changes in amplitude and phase for the second eigenmode, not 1.84%. At higher concentrations of the gel the intensity stops affected by the feedback loop restrictions, are monitored. The being enhanced. Seemingly an excess of CNF impedes AuNPs to contrast obtained for the second eigenmode arises mainly on the attach on nitrocellulose pores. Thus, 1.84% of CNF was chosen different composition of the sample and change of mechanical as optimal for the signal enhancement of LFs and applied on the properties such as stiffness of the material. Moreover, this further experiments. Also, it is observed how the sensitivity technique is depth sensitive and allows to observe the presence of seems to increase especially at lower concentrations; because of structures embedded in a matrix such as a gel, even below the that, on the following experiments more concentrations between surface. 0-0.5 µg/mL were evaluated. It must be highlighted that, in a previous work (Tarrés et al., Standard LFs versus CNF-LFs (containing 1.84% CNF in the 2017), CNF were characterized by XRD exhibiting the typical TL) were compared. Different concentrations of HIgG were Cellulose I crystalline structure, with peaks corresponding to added on the strips and the intensities were measured as (110), (110) and (200) reflection planes and an average diameter explained at methods section. As shown in Fig. 5, over a around 20 nm. concentration of 0.05 µg/mL HIgG the signal of the CNF-LFs is increased by an average of 36.6 % in comparison to those strips 3. Results and discussion without CNF gel. In absence of HIgG (i.e. blank samples), the AuNPs-based LFs were prepared for the detection of human response of CNF-LFs and standard LFs is comparable. It immunoglobulin (HIgG) (as described in methods section), a demonstrates that there is no unspecific adsorption between protein often used as model on the optimization of lateral flow HIgG and the CNFs. So that, AuNPs only stop on TL due the assays (Parolo et al., 2013a, 2013b; Rivas et al., 2014; Quesada- interaction of antibodies and the analyte. González et al., 2019). CNF gel was dispensed on the TL of LFs, The calibration curve (Fig. S5) obtained for CNF-LFs follows before adding the antibodies, to increase the density of these the equation: retained near the surface of the strip and thus, enhance the colorimetric response observable. Atomic force microscopy % color = 5.688 ln [HIgG (µg/mL)] + 57.768 (AFM) images shown in Fig. 3 demonstrate the presence of CNF The r value for the equation was 0.995. The limit of detection on nitrocellulose samples. (LOD) was calculated as % colorLOD = blank + 3 σblank, i.e., the To further analyze the role of the CNF on the enhancement of corresponding value of blank sample plus 3 times its standard AuNPs binding, the same samples were also observed by AFM. deviation (Armbruster, 2008). The value obtained was of 0.01 Initially, AFM was tested as a useful tool to characterize the µg/mL in both cases. Even though the LOD is similar to standard presence of CNFs. A sample of CNF gel was studied by bimodal LFs (since it only depends on the antibodies affinity for the AFM and the presence of CNF was clearly observed due to HIgG, not on the CNF), the sensitivity is highly improved with mechanical stiffness contrast with respect to the medium (Fig. the CNF (the slope was almost two-times higher than the slope of S4). They show an average diameter of ~ 20 nm (it is to mention standard LFS, 5.7 ± 0.1 vs. 3.2 ± 0.1). The method shows a that due to the density of the sample, the fibers are entangled and reproducibility (relative standard deviation, RSD) of 2 % (n=6) it is not possible to have them isolated). The diameter fits with for a HIgG concentration of 0.05 µg/mL. the results previously observed (Tarrés et al., 2017), in which the These results reflect what was observed when characterizing CNF were characterized by XRD. To further prove the existence the strips by SEM and AFM (Fig 6). In SEM images it can be of CNF on nitrocellulose membranes, two different nitrocellulose observed how, for the same amount of HIgG, the population of samples with and without CNF were analyzed by bimodal AFM, AuNPs on the TL is greater on the CNF-LFs. It demonstrates that as shown in Fig. 3. An area of 1 µm of the surface of a more AuNPs are being retained, at least on the surface of the nitrocellulose membrane was scanned – maximum possible as strip. Still, SEM does not allow the imaging of CNFs due to lack limited by the degree of curvature of these surfaces due to the of SEM contrast against the CNFs. mean radius of the nitrocellulose fiber. The topography image shown in Fig 3a gives evidence of the intrinsic rugosity of the Finally, the binding of AuNP on the TL was also studied by nitrocellulose membrane. Still, these features do not induce AFM. It was demonstrated that for a NFC-LFs the quantity of significant contrast in the bimodal AFM images shown in Fig. 3b AuNPs is significantly greater than the case of the LFs without and 3c, indicating the homogeneity of the material. Instead, Fig CNF (Fig. 6). The high roughness level of the curved surfaces 3d clearly demonstrates the presence of CNF deposited on the impeded to selectively analyze the binding mechanisms between nitrocellulose membrane as evidenced by both, the features in the single AuNP and the CNF, but it was possible to determine that topography and their further contrast in bimodal AFM images the higher density of AuNP were placed on the side areas of (Garcia and Proksch, 2013; González-Domínguez et al., 2016). In nitrocellulose for which also higher density of CNFs were this case, the CNF seem to be retained mainly in one side of the observed to deposit, for the CNF functionalized strips, while for fiber, the one where TL antibodies are dispensed, that which is the non-functionalized ones, the attachment of AuNPs seemed to visible at bare eye. be randomly dispersed on the nitrocellulose surface. To demonstrate that CNF gel can increase the density of 4. Conclusions antibodies near the paper surface, thus enhancing the signal (i.e. quantity of AuNPs retained on positive samples), four We have devised a new signal enhancement strategy for LFs concentrations of this nanomaterial were dispensed on consisting in the addition of CNF on paper nitrocellulose to nitrocellulose. Different concentrations of HIgG were dropped on decrease the pore size in the TL, thus increasing the amount of the distinct LFs, without CNF and with the different four CNF bioreceptors (capture antibodies) near the surface. Thanks to this concentrations, and the intensities were recorded by a mobile modification, on positive samples we observed an average 4 increase of 36.6 % of the colorimetric signal, which means that Fridley, G. E.; Le, H. Q.; Fu, E.; Yager, P. Lab Chip 2012, 12, 4321– effectively we succeed increasing the amount of AuNPs retained. 4327. This allows to increase the sensitivity of the LF immunoassay Fu, E.; Lutz, B.; Kauffman, P.; Yager, P. Lab Chip 2010, 10, 918. (HIgG evaluated as model analyte) as well as to better Fu, E.; Ramsey, S. A.; Kauffman, P.; Lutz, B.; Yager, P. Microfluid. discriminate between different analyte concentrations, which is Nanofluidics 2011, 10, 29–35. of great relevance for both naked-eye and instrumental-based Fu, Q.; Wu, Z.; Xu, F.; Li, X.; Yao, C.; Xu, M.; Sheng, L.; Yu, S.; Tang, Y. Lab Chip 2016, 16, 1927–1933. detection. Unspecific adsorption of AuNPs on TL is not observed Garcia, R.; Proksch R. Eur. Polym. J. 2013, 49, 1897–1906. (when measuring blank HIgG samples on a CNF-LFs), which González, I.; Boufi, S.; Pèlach, M. A.; Alcalà, M.; Vilaseca, F.; avoids false positive results. Due to these conclusions, we believe Mutjé, P. BioResources 2012, 7, 5167–5180. that our strategy can be used to discern especially at low González, I.; Vilaseca, F.; Alcalá, M.; Pèlach, M. A.; Boufi, S.; concentrations, which would be useful especially on diagnostic Mutjé, P. Cellulose 2013, 20, 1425–1435. applications. The characterization of the nitrocellulose performed González, I.; Alcalà, M.; Chinga-Carrasco, G.; Vilaseca, F.; Boufi, with SEM and AFM demonstrates the presence of the CNF and S.; Mutjé, P. Cellulose 2014, 21, 2599–2609. how it affects to the interaction with AuNPs. The proposed González, I.; Oliver-Ortega, H.; Tarrés, Q.; Delgado-Aguilar, M.; modification is simple and cheap and can be easily applied to any Mutjé, P.; Andreu, D. Int. J. Biol. Macromol. 2017, 105, 741–748. type of LF strip, enabling its use in PoC applications. González-Domínguez, I.; Gutiérrez-Granados, S.; Cervera, L.; Gòdia, F.; Domingo, N. Biophysical Journal, 2016, 111, 1173-1179. Acknowledgments Khlebtsov, B. N.; Bratashov, D. N.; Byzova, N. A.; Dzantiev, B. B.; Khlebtsov, N. G. Nanoresearch 2019, 12(2), 413–420. The ICN2 is funded by the CERCA Programme / Generalitat Kratz, A.; Lewandrowski, K. B. Arch. Pathol. Lab. Med. 2003, 127, de Catalunya. The ICN2 is supported by the Severo Ochoa 511. program of the Spanish Ministry of Economy, Industry and Kühnemund, M.; Wei, Q.; Darai, E.; Wang, Y.; Iván, H. N.; Yang, Z.; Tseng, D.; Ahlford, A.; Mathot, L.; Sjöblom, T.; Aydogan, O.; Competitiveness (MINECO, grant No. SEV-2017-0706). Nilsson, M. Nat. Commun. 2017, 8, 1–8. Financial support was obtained under MAT2017-87202-P, LaBarre, P.; Boyle, D.; Hawkins, K.; Weigl, B. Int. Soc. Opt. Eng. funded by Agencia Estatal de Investigación (AEI) and Fondo 2011, 802902. Europeo de Desarrollo Regional (FEDER), and MINECO project Li, M.; Yang, H.; Li, S.; Zhao, K.; Li, J.; Jiang, D.; Sun, L.; Deng, A. FIS2015-73932-JIN. This work was partially funded by 2017- J. Agric. Food Chem. 2014, 62, 10896–10902. SGR-579 project from the Generalitat de Catalunya. D.Q.G. Liu, J.; Hu, X.; Cao, F.; Zhang, Y.; Lu, J.; Zeng, L. R. Soc. open sci. acknowledges Universitat Autònoma de Barcelona for the 2018, 5, 180288. possibility of carrying out this work inside the framework of its López-Marzo, A. M.; Merkoçi, A. Lab Chip 2016, 16, 3150–3176. PhD Programme in Chemistry. Ch.S. thanks the BIST for a Loynachan, C. N.; Thomas, M. R.; Gray, E. R.; Richards, D. A.; preBIST Grant (No. 754558). We thank Marcos Rosado and Kim, J.; Miller, B. S.; Brookes, J. C.; Agarwal, S.; Chudasama, V.; Belén Ballesteros for the support given on SEM and TEM McKendry, R. A.; Stevens, M. M. ACS Nano 2018, 12, 279–288. imaging. The music used on video V1 is “Converging Lines” by Marquina, C.; De Teresa, J. M.; Serrate, D.; Marzo, J.; Cardoso, F. David Hilowitz, used under the Creative Commons Attribution A.; Saurel, D.; Cardoso, S.; Freitas, P. P.; Ibarra, M. R. J. Magn. Non-Commercial 4.0 International License (CC BY-NC 4.0), Magn. Mater. 2012, 324, 3495–3498. https://bit.ly/2LxhpAJ. Mudanyali, O.; Dimitrov, S.; Sikora, U.; Padmanabhan, S.; Navruz, I.; Ozcan, A. Lab Chip 2012, 12, 2678–2686. Ngu, M. A. V.; Bergantin, J. H.; Ramos, J. D. A. Protein Peptide References Lett. 2019, 26, doi: 10.2174/0929866526666190212164751. Parolo, C.; Merkoçi, A. Chem. Soc. Rev. 2013, 42, 450–457. Alcalá, M.; González, I.; Boufi, S.; Vilaseca, F.; Mutjé, P. Cellulose Parolo, C.; de la Escosura-Muñiz, A.; Merkoçi, A. Biosens. 2013, 20, 2909–2921. Bioelectron. 2013a, 40, 412–416. Álvarez-Diduk, R.; Orozco, J.; Merkoçi, A. Sci. Rep. 2017, 7, 976. Parolo, C.; Medina-Sánchez, M.; de la Escosura-Muñiz, A.; Merkoçi, Anfossi, L.; Di Nardo, F.; Giovannoli, C.; Passini, C.; Baggiani, C. A. Lab Chip 2013b, 13, 386–390. Anal. Bioanal. Chem. 2013, 405, 9859–9867. Parolo, C.; De La Escosura-Muñiz, A.; Polo, E.; Grazú, V.; De La Armbruster, D. A.; Pry, T. Clin. Biochem. Rev. 2008, 29, 49-52. Fuente, J. M.; Merkoçi, A. ACS Appl. Mater. Interfaces 2013c, 5, Berg, B.; Cortazar, B.; Tseng, D.; Ozkan, H.; Feng, S.; Wei, Q.; 10753–10759. Chan, R. Y. L.; Burbano, J.; Farooqui, Q.; Lewinski, M.; Carlo, D.; Quesada-González, D.; Merkoçi, A. Biosens. Bioelectron. 2015, 73, Garner, O. B.; Ozcan, A. ACS Nano 2015, 9, 7857–7866. 47–63. Brangel, P.; Sobarzo, A.; Parolo, C.; Miller, B. S.; Howes, P. D.; Quesada-González, D.; Merkoçi, A. Biosens. Bioelectron. 2016, 92, Gelkop, S.; Lutwama, J. J.; Dye, J. M.; McKendry, R. A.; Lobel, L.; 549–562. Stevens, M. M. ACS Nano 2018, 12, 63–73. Quesada-González, D. “Design and application of nanomaterial- Burns, M. A. Science. 2002, 296, 1818–1819. based lateral flow devices”, Ph.D. Dissertation, Autonomous Chapman, R.; Lin, Y.; Burnapp, M.; Bentham, A.; Hillier, D.; University of Barcelona, Bellaterra, Spain, 2018. Zabron, A.; Khan, S.; Tyreman, M.; Stevens, M. M. ACS Nano 2015, Quesada-González, D.; Merkoçi, A. Chem. Soc. Rev. 2018, 47, 9, 2565–2573. 4697–4709. Coskun, A. F.; Wong, J.; Khodadadi, D.; Nagi, R.; Tey, A.; Ozcan, Quesada-González, D.; Jairo, G. A.; Blake II, R. C.; Blake, D. A.; A. Lab Chip 2013a, 13, 636–640. Merkoçi, A. Sci. Rep. 2018, 8, 16157. Coskun, A. F.; Nagi, R.; Sadeghi, K.; Phillips, S.; Ozcan, A. Lab Quesada-González, D.; Sena-Torralba, A.; Wicaksono, W. P.; de la Chip 2013b, 13, 4231–4238. Escosura-Muñiz, A.; Ivandini, T. A.; Merkoçi, A. Biosens. De La Escosura-Muñiz, A.; Parolo, C.; Merkoi, A. Mater. Today Bioelectron. 2019, 132, 132-135. 2010, 13, 24–34. Rivas, L.; Medina-Sánchez, M.; de la Escosura-Muñiz, A.; Merkoçi, Delgado-Aguilar, M.; González, I.; Tarrés, Q.; Alcalà, M.; Pèlach, A. Lab Chip 2014, 14, 4406–4414. M. À.; Mutjé, P. BioResources 2015, 10, 5330–5344. Rivas, L.; de la Escosura-Muñiz, A.; Serrano, L.; Altet, L.; Francino, Delgado-Aguilar, M.; González, I.; Tarrés, Q.; Pèlach, M. À.; Alcalà, O.; Sánchez; Merkoçi, A. Nanoresearch 2015, 8(11), 3704–3714. M.; Mutjé, P. Ind. Crops Prod. 2016, 86, 295–300. Rodriguez, N. M.; Wong, W. S.; Liu, L.; Dewar, R.; Klapperich, C. M. Lab Chip 2016, 16, 753–763. Roda, A.; Michelini, E.; Cevenini, L.; Calabria, D.; Calabretta, M. M.; Simoni, P. Anal. Chem. 2014, 86, 7299–7304. Roda, A.; Michelini, E.; Zangheri, M.; Di Fusco, M.; Calabria, D.; Simoni, P. TrAC - Trends Anal. Chem. 2016, 79, 317–325. Shim, W. B.; Kim, K. Y.; Chung, D. H. J. Agric. Food Chem. 2009, 57(10), 4035–4041. Shyu, R. H.; Shyu, H. F.; Liu, H. W.; Tang, S. S. Toxicon 2001, 40, 255–258. Smart Micro Optics: Blips lenses [online reference] https://www.smartmicrooptics.com/blips/ (accessed Mar 16, 2019). Tang, D.; Sauceda, J. C.; Lin, Z.; Ott, S.; Basova, E.; Goryacheva, I.; Biselli, S.; Lin, J.; Niessner, R.; Knopp, D. Biosens. Bioelectron. 2009, 25, 514–518. Tarrés, Q.; Delgado-Aguilar, M.; Pèlach, M. A.; González, I.; Boufi, S.; Mutjé, P. Cellulose 2016, 23, 3939–3950. Tarrés. Q.; Boufi, S; Mutjé, P.; Delgado-Aguilar, M. Cellulose 2017, 24, 3943–3954. Turkevich, J.; Stevenson, P. C.; Hillier, J. A Faraday Soc. 1951, 11, 55–75. Wong, R.; Tse, H. “Lateral Flow Immunoassay”, Springer 2009. Wu, G.; Srivastava, J.; Zaman, M. H. Anal. Biochem. 2014, 449, 147-154. Xie, Y.; Chang, H.; Zhao, K.; Li, J.; Yang, H.; Mei, L.; Xua, S.; Deng, A. Anal. Methods, 2015, 7, 513–520. You, M.; Lin, M.; Gong, Y.; Wang, S.; Li, A.; Ji, L.; Zhao, H.; Ling, K.; Wen, T.; Huang, Y.; Gao, D.; Ma, Q.; Wang, T.; Ma, A.; Li, X.; Xu, F. ACS Nano 2017, 11, 6261–6270. Zhan, L.; Guo, S. Z.; Song, F.; Gong, Y.; Xu, F.; Boulware, D. R.; McAlpine, M. C.; Chan, W. C. W.; Bischof, J. C. Nano Lett. 2017, 17(12), 7207-7212. Zhu, X.; Shah, P.; Stoff, S.; Liu, H.; Li, C. A Analyst 2014, 139, 2850–2857. Supplementary Material Supplementary material (TEM characterization of AuNPs; Image J software screenshot; AFM imaging of CNF; calibration curves; a video of a 3D printed tool used to control the distance between mobile phone camera-BLIPS Lens and the LFs) is available in the online version of this article at http://xxxxxxxxxxxxxxxx. 6 Figures Figure 1. A) Schematic representation of a positive TL not modified (up) and modified with cellulose nanofibers dispensed (down). B) Schematic representation of a LFs with all its components. Figure 2. Representation of a LFs construction; the different pads that compose it and its position (in green, size of the adhesive card; in orange, size of the pads; in purple, some positioning parameters). Figure 3. AFM imaging of sample without and with CNF. (a) to (c) refers to the sample without NFC and (d) to (f) refers to the sample with CNF (a) and (d) AFM 3D topography imaging. (b) and (e) Second amplitude (A2) images of bimodal AFM, corresponding to the dynamics of the first excited eigenmode and (c) and (f) correspond to the second frequency (f2) accordingly. The dimension of all images is 1 µm. From phase and amplitude image is obvious the existence of an extra material (CNF). Figure 4. Comparison, at different concentrations of HIgG, of the effect of adding different concentrations of CNF in the TL of LFs. 8 Figure 5. Data obtained from LFs assay performed with both standard and CNF-LFs and the images of the strips, also taken with Blips Micro Lens (inset). Figure 6. SEM (up) and AFM 3D topography (down) images of nitrocellulose corresponding to the TL of a standard LF (left) and of a CNF-LF (right).