This is the accepted version of the journal article: Aller Pellitero, Miguel; Fremeau, Joséphine; Villa, Rosa; [et al.]. «Elec- trochromic biosensors based on screen-printed Prussian Blue electrodes». Sensors and actuators B: Chemical, Vol. 290 (July 2019), p. 591-597. DOI 10.1016/j.snb.2019.03.100 This version is available at https://ddd.uab.cat/record/274618 under the terms of the license Electrochromic biosensors based on screen-printed Prussian Blue electrodes Miguel Aller Pellitero1, Joséphine Fremeau2, Rosa Villa1,3, Boris Lakard2, Jean-Yves ihn2, and F. Javier del Campo*1 1 Instituto de Microelectrónica de Barcelona, IMB-CNM (CSIC); Esfer UAB. Campus Universidad Autónoma de Barcelona; 08193 – Bellaterra. Barcelona, in 2 Institut UTINAM, UMR 6214 CNRS/ Univ. Bourgogne Franche-Comté, 16 route de Gray, 25030 Besançon, France. 3 CIBER-BBN, Networking Centre on Bioenginee ng, Biomaterials and Nanomedicine, Barcelona, Spain. Proofs to: F. Javier del Campo. Phone: SP – 93.594.77.00 xt. 2451 Email: javier.delca o@csic.es; Date ubmitted to B&B: 13-11-2018 Date submitted to SNB: 20-01-2019 Highlights x We present the production of an electrochromic screen-printable paste based on Prussian Blue. 1 x These blue electrodes display excellent spectroelectrochemical reversibility and high catalytic activity towards the reduction of H2O2. x These blue electrodes enable both the electrochemical and optical quantitative determination of hydrogen peroxide. x The reported blue electrodes are highly suitable transducers for the construction of oxidase-based biosensors. x A glucose biosensor is presented to demonstrate the spectroelectrochemical capabi ties of these new electrodes. Abstract Prussian Blue (PB)-modified graphite screen-printed electrodes are increasingly bei g used in electrochemical biosensors. However, they do not allow the observation of the l ctrochromism of PB. This work presents the construction of PB-based, lectrochromic screen-printed biosensors. Although electrically more resistive than their graphi cou terparts, these new PB- based electrodes enable both the amperometric and colorimet detection of hydrogen peroxide. This is the first time that this has been achieved ing creen-printed electrodes, and we demonstrate it spectroelectrochemically on a glucose biosensor. The biosensor electrochemical performance equals that of previously reported PB/g aphi e electrodes, being able to detect down to 4 μM H2O2 and 54 μM glucose. At the me time, and in contrast to PB/graphite electrodes, the new PB-based electrodes afford the optical detection of these two analytes down to 1.2 μM and 15 μM, respectively. The dyn mic ranges of the glucose biosensors obtained at the PB- based electrodes are 0.1-1 mM (amperometric) and 0.025-2.5 mM (Colorimetric), matching the physiological glucose co c ntration range in body fluids other than blood or serum. Keywords: Pruss an B ue electrodes; Screen-printed electrodes; Spectroelectrochemistry; Biosensors; Hydrogen peroxide detection; Glucose sensing. 1. I rodu tion P u sian Blue consists of iron (II) and (III) atoms bridged by cyanide groups in a cubic crystal structure of formula Fe43+[FeII(CN)6]3. This composition and structure bestow Prussian Blue its characteristic colour and electrochemical properties. Prussian Blue undergoes two reversible redox 2 reactions, each associated with an ion exchange step and its respective colour change. 3 Thus, Everitt’s salt, the reduced form, is colourless K4Fe42+[FeII(CN)6]3, while the oxidized form is yellowish and known as Berlin Green, Fe43+[FeIII(CN)6A], where A represents an anion inserted in the crystal lattice to balance charge in the oxidation process. In addition to its electrochromic behavior, that makes it suitable for the fabrication of displays and smart gl ss, [1] Prussian Blue also displays excellent catalytic properties for the reduction of oxygen nd hydrogen peroxide. Specifically, this is enabled by the high-spin Fe2+ atoms in the du ed form, [2] which makes Prussian Blue films more active towards H2O2 reduction than t electrodes, and only slightly less active than peroxidase enzymes. [3] This ability to selectively catalyze the reduction of H2O2 has een successfully exploited in first generation oxidase-based electrochemical [4] and optical [5] biosensors. Karyakin et al. were first to demonstrate the construction of such biosensors. [3, 4] Since then, the number of reports featuring Prussian Blue as artificial peroxidase has increased continuously. [6-9] Surprisingly, despite its electrochromism and sen ing capabilities, very few works report the exploitation of its electrochromism for ( io)sensing purposes, and those that do usually rely on PB electrodeposited on expe transparent electrodes [10-14]. Table 1 provides a summary of key works featuring screen-printed electrodes (SPEs) modified with PB which are us for the amperometric detection of hydrogen peroxide for biosensing purposes, including h work. 4 Table 1. Comparison of different Prussian-blue screen-printed sensors and biosensors used in the amperometric determination of H2O2. DS stands for commercial Dropsens SPEs and GNT for electrodes fabricated with a screen-printing paste from Gwent Ltd, UK. Electrode Preparation of Analyte LOD Range Sensitivity Reference PB PB on GC electrodeposition Glucose 1 μM 1μM-5 0.18 A cm- [3] 2 mM M-1 PB/PPy on electrodeposition Glucose 0.1mM 0.1- 1.0-1.9 μA [15] C 20mM cm-2 mM-1 PB on ITO electrodeposition Lactic 1mM 1-20mM Unreporte [ 6] acid PB electrodeposition H2O2 0.1 μM 0.1 μM- 0.6 A cm-2 [17 (unspecified 0.1 mM M-1 electrode) Glutamate 0.1 μM 0.1-100 0 21 A cm- 2 μM M- Modified growth on H2O2 0.5 μM 0 5 μM- 0.11 A cm- [18] 2 carbon graphite 5 M M-1 paste Glucose 100 μM 0.1 10 5.6x10-4 A m cm-2 M-1 Screen- undisclosed Uric acid 3 μM 10-200 0.89 A cm- [19] 2 printed μM M-1 electrode (DS) Screen- undisclosed C techol 0 4 μM 1-90 μM Unreported [20] printed electrode (DS) Screen- undisclosed P raoxon 12 μM 5-150 Unreported [21] printed μM electrode (DS) Screen- Screen- printed und sclosed printed electrode (DS) Screen- pr nt electro e (DS) 5 0.015 A . [22] cm-2 M-1 0 5 - 5.8x10-4 A 0 [23] cm-2 M-1 . 5 m 0.1 A cm-2 M [24] M-1 undisclosed G l u c o s e U n r e p o r t e d 0 - 3 3 . 1 m M undisclosed H2O 2 3.6 μM 0- 100 μM electrode 6 (GNT) 7 Electrode Preparation of Analyte LOD Range Sensitivity Reference PB Glucose 0.7 mM 0-100 4.8x10-3 A mM cm-2 M-1 Lactate 1.19 mM 0-50 3.0x10-3 A mM cm-2 M-1 Uric acid 4.6 mM 0-35 4.5x10-4 A mM cm-2 M-1 Screen- undisclosed Lactate 0.15 μM 0.15μM- 0.4 A cm-2 [2 ] printed 1.1 mM M-1 electrode (GNT) Carbon SPE PB drop casted Cysteine 98.9 μM 300-700 0.06 A cm- [26] on C-SPE μM 2 M-1 PB/Carbon PB incorporated Cysteine 67.4 μM 100-600 0.09 A cm- [26] 2 SPE to C-paste μM M-1 SiO2- Growth of H2O2 3.8 μM 0 025-5 0.07 A This work ATO/PB particles mM cm-2 M-1 ITO/PB(2) Growth of H2O2 3.2 μM 0.01-5 0.08 A This work particles m cm-2 M-1 SiO2- Growth of Glucose 54.1 μM 0.1 – 1 8.8x10-3 A This work ATO/PB particles mM cm-2 M-1 ITO/PB Growth of Glucose 70.5 μM 0.1 – 1 7.4x10-3 A This work particles mM cm-2 M-1 (1) SiO2-ATO/PB: Prussian Blue was chemically grown on SiO2 core and ATO-shell 3μm particles. (2) ITO/PB: Prussian Blue was chemica grown directly on 100nm ITO nanoparticles. Regarding the prep ratio PB-modified electrodes, there are three main ways: electrodeposition, chemical growth on an electrode surface, and deposition of PB-modified graphite past s. these, since it was first reported by Neff, [27] electrodeposition has been the preferred me hod when it comes to developing electrochromic devices and electrochemical sensor . The main advantages of electrodeposition methods are not only the selective growth on electrode surfaces alone, but also the control of growth conditions. This ultimately means 8 controlling device-critical features such as film thickness, roughness, porosity, and stability. [28] Deposition methods based on potentiostatic, [17] galvanostatic, [27] cyclic, [3] and pulse voltammetry [29] control have been reported. PB is typically electrodeposited from solutions 9 containing Fe(III) and ferricyanide ions. However, electrodeposition from Fe(II) and ferrocyanide is also possible and may be desirable when a less acidic pH is required to preserve the activity of enzymes or other biomolecules.[15] Chemical growth, on the other hand, may also be used to produce Prussian Blue films, th on electrodes [26, 30, 31] and on inert surfaces [32] in many different ways. [33] Howe r, th is not an ideal approach because it lacks the surface selectivity and much of the co trol that electrochemical methods enjoy. Nevertheless, chemical synthesis of PB is used in the formulation of reen-printing materials, which is the third relevant approach that we have identified for e production of PB-modified electrodes. Up until the mid 1990s, most electrochemical s dies on Prussian Blue used electrodeposited films. However, the use of chemically synthes d Prussian Blue has been made popular through the introduction of carbon paste [34] and creen-printed electrodes [35] (SPEs). PB-modified pastes represent a straightforward way t overcome critical production issues affecting electrodeposition, such as th bat h nature of the electrodeposition process and its relatively high cost. Screen-printing also ex nds the nature of possible substrates beyond conducting electrode materials, to include p er, textiles, ceramics, polymers and, given the right combination of solvents and curing onditions, nearly any material imaginable. PB-mediated p stes have typically three key components: a conducting material, the PB itself, and a suitable binder system. The conducting material used in existing commercial products is gra hite, and different resins are used as binder agents. In this work, we have grown PB chemically over the surface of clear conducting particles, and then mixed the modified particles with a Viton® binder system to produce an electrochromic 10 paste. This material had been previously reported for the construction of electrochromic displays [36] but, to the best of our knowledge, this is its first demonstration in (bio)sensing. Thus, in contrast to previously reported PB/graphite-based screen-printed electrodes, which are limited to electrochemical detection only, the electrodes presented here enable both electrochemical a d optical detection of H2O2. The electrochemical performance of the materials pres nted here is comparable to graphite-based materials, but their electrochromism makes them ry tractive also for the construction of optical (bio)sensors. This concept is demonstrated in the context of a glucose oxidase-based biosensor. 2. Experimental 2.1 Reagents and materials Potassium ferrocyanide, ferrous sulfate, pota sium chlor de, acetic acid, glucose, bovine serum albumin (BSA), and 2-butoxyethyl acetate were pu chased from Sigma-Aldrich. Glucose oxidase (236 U·mg-1) was acquired from Sekisui Diagnostics (UK). Chitosan (M.W. 100,000 – 300,000) was purchased from Acros rga ics. SiO2-ATO conducting particles (nominal size of 3 μm) Zelec 1610-S were kindly provided by Milliken Chemical (BE). ITO conducting particles (nominal size of 5 n ), Nan Tek®, were purchased from Alfa Aesar. Viton® fluoroelastomer (GBL-600S, DuPont) w s acquired from Eagle Elastomer (USA). Carbon conducting paste ref. C203 519P4 and PB-modified carbon paste ref. C2070424P2 were purchased from Gwent E ectronic Materials, ltd (UK), Electrodag PF-455B photocurable dielectric and 725A silver pas es were obtained from Henkel (ES). 0.5mm-thick PET sheets (Autostat WP20) were purchased from MacDermid (UK). PB-modified screen-printed carbon electrodes (PB-SPCEs) ref. 710 were purchased from Dropsens (ES). 11 All chemicals were used as received without further purification. paste. This material had been previously reported for the construction of electrochromic displays 12 2.2 Instrumentation Electrochemical and spectroelectrochemical measurements were done using a SPELEC instrument (Dropsens) controlled by DropView SPELEC software (version 3.0), installed on a PC running Windows 10. Scanning electron microscopy images were obtained at an uriga microscope (Carl Zeiss). A Kulikke&Soffa four-point probe connected to a 3455 Digital Voltmeter (Hewlett Packard) was used to measure the sheet resistance of e bricated materials. 2.3 Electrochromic Prussian Blue screen-printed electrodes Prussian Blue screen-printing pastes were prepared following a previously published protocol. [36] Briefly, 20 g of conducting ITO or SiO2-ATO particles wer suspended, under vigorous stirring, in 200 mL of a 25 mM iron (II) sulfate aqueo s solution. Then, 30 mL of a 60 mM ferricyanide solution were added dropwise to the irred solution. The resulting blue particles were separated by decantation, washed with ilute Cl, and dried at 100ºC. Prussian Blue-modified condu ting p rticl s were next combined with a 15-20% Viton solution prepared in 2-butoxyethylace te to a 5:2 pigment to binder proportion, and thoroughly mixed in a ball mill, unt t e resu ting paste presented a honey-like texture. The particle size was measured using a 50 m grind gauge, obtaining a maximum particle size range between 5 and 12 μm. This is extre ely g od considering that the starting particle size was nominally 3μm, and that n addit ves w re used to promote particle suspension in the paste. 8 Due to the high electrical resistance of the printed Prussian Blue layers, it was needed to print 2.2o Instrumentation it er a conducting substrate. The simplest and most cost-effective case involved its application over a graphite screen-printed electrode. However, to enhance the colour contrast of the electrochromic change, a SiO2-ATO layer [37] was additionally printed between the graphite and 9 the Prussian Blue layers. All the electrodes presented in this work have been prepared as previously reported, using a manual, home-made screen printing setup. [38] 2.4 Glucose biosensor fabrication Prussian Blue screen-printed electrodes (PB-SPEs) were modified following a modif ed version of a previously reported protocol. [39] Briefly, 4 μL of a 0.5% wt. chitosan sol ion prepared in 0.1 M acetic acid was cast on the surface of the electrodes and all ed o dry at room temperature. The electrodes were subsequently rinsed with deioniz ater, and 6 μL of a 5 mg mL-1 GOx containing 15 mg mL-1 BSA were then cast and allowed to dry. The electrodes were stored at 4ºC. 2.5 Electrochemical measurements All the experiments were carried out in a 50 mM phosphate buffered solution and 0.1 M KCl, adjusted to a pH of 5.5. The selected pH ensures the stability of the Prussian Blue while keeping the activity of the oxidase enzymes on the lect ode near their optimum activity. For the electrochemical characterization, h e ectrodes were preconditioned by cyclic voltammetry, applying 20 cycles at a scan rate of 0 mVs-1, to equilibrate the crystalline structure of the Prussian Blue with the K+ cations in the supporting electrolyte. [17] Note, however, that no preconditioning step was required for the sensing of hydrogen peroxide or glucose. Limits of detect on, LoD have been determined based on conventional 3V/m methodology, where σ corresponds to the background signal standard deviation, and m to the slope of the linear range [40]. with the optical shutter closed. Background spectra were recorded over completely reduced 9 2 S ectroelectrochemical measurements the Prussian Blue layers. All the electrodes presented in this work have been prepared as Reflectance measurements were done using DropSens-SPELEC reflectance cell, using solutions of same composition as those described in section 2.4 above. The dark spectrum was recorded with the optical shutter closed. Background spectra were recorded over completely reduced 10 electrodes, which displayed the clearest colour. Unless otherwise stated, spectra were collected using an integration time of 300 ms. Data analysis was carried out using SPELEC DropView software. 3. Results and discussion 3.1 Characterization of Prussian Blue electrodes. The electrochemical and spectroelectrochemical performance of the PB- PEs was evaluated and compared to that of (i) electrodes fabricated with a commercial PB- raphite p ste, and (ii) to commercial PB-SPCEs. For the sake of clarity and simplicity, from here on we will refer to these as “commercial materials” and “graphite electrode ” i distinc ively, and to ours as “electrochromic pastes” or “ PB-electrodes”. Figure 1 shows cyclic voltammograms reco ded n sup orting electrolyte. The most striking difference in the voltammetry of the diffe t PB-modified pastes is the much higher currents observed for the blue pastes pres nted here, c mpared to those obtained from the commercial materials. One feature of ITO nan pa ticles compared to SiO2-ATO particles is that pastes with significantly higher particle o binder ratios can be formulated, so even higher PB loadings could, in principle be achieved, but the significantly higher cost of ITO nanoparticles compared to SiO2-ATO articl s should also be considered. The cyclic oltammograms in Figure 1 suggest that commercial graphite pastes contain a much lo er amount of Prussian Blue than the electrochromic pastes presented here. Also, the vo ammograms in Figure 1 show that our PB-based electrodes display relatively larger background currents. Note that similarly high background currents have been observed for electrodes made with unmodified ATO particles, [37] but also on sol-gel-derived Prussian Blue- 10 silicate electrodes [10]. We believe there are two likely reasons for such currents; first, the electrodes, which displayed the clearest colour. Unless otherwise stated, spectra were collected 11 porous structure and resulting large surface area of the electrodes (see Figure S1), but also, in the present case, the mild electroactivity of indium and antimony tin-oxides, [41, 42]. FIGURE 1 AROUND HERE The PB-electrodes presented here exhibit a superior electrochemical behavior sed on their peak-to-peak separation compared to commercial materials based on grap ite (Table S1). This is probably due to a better contact of the PB and the conduct g aterial. ITO-modified nanoparticles display the smallest peak-to-peak separation, ca. 30 mV at 5mVs-1, presumably due to the better conductivity afforded by ITO nanoparticles compar d to SiO2-ATO micro-particles. However, the much higher PB loading, arising fr m the mas ive available surface area of the nanoparticles combined with the thickness of th sc en-printed layer, results in much broader peaks compared to the other materials. Th s peak broadness, which may be detrimental for electroanalytical applications, is d to a imitation in the rate of potassium exchange during the redox process. In fact, this limi tion h s been observed for all 4 electrode types under study in the form of linear plots of peak current vs. square root of scan rate (Figure S2). Such a linear dependency points at a iffusion controlled process, where the only diffusing species involved here are potassium a ions. In contrast, thin film electrogenerated PB electrodes [43, 44] exhibit a sur ace-controlled regime as predicted by the Laviron expression. [45] The reason for this df ent behavior is in the thickness of the PB layer. While electrodeposited layers are typically in the range of nanometers (hundreds), screen printed layers are at least a few microns thick and can exchange larger amounts of cations. In all cases, charge densities have been calculated using the electrode geometric area, which allows for a very limited comparison as the thickness of the 12 PB coating on the screen-printed electrodes is unknown. Therefore, although the actual surface area of our screen-printed electrodes has not been determined, the charge density analysis of the PB peak at the different electrodes shows that our PB-based electrodes contain between 5 and 7 times more Prussian Blue than an electrodeposited layer (See Table S1). Another important difference between graphite-PB and the SiO2-ATO/ITO-PB-electrodes is their electrical conductivity. We found, using four-point probe measurements of st s ruc ures, that the resistivity of Prussian Blue modified particles increases by more than 5 times compared to a paste of unmodified particles at the same concentration (see Tabl S2). This suggests that commercial materials may have been produced simply by addin PB particles to an existing graphite paste formulation, which could also explain their smaller PB signals, and their somewhat broader peak-to-peak separation in the voltammetry. Next, the spectroelectrochemical response of e PB- PEs was compared. The black colour of graphite in the commercial PB-modified commercial electrodes precluded the observation of any spectroelectrochemical changes, in ontrast with the ITO and ATO based pastes presented here. Figure 2A and Figure 2B sho cycl c voltammograms of the two types of PB-SPEs together with the derivative of their respective voltabsorptograms. A clear correlation between the electrochemical nd the spe troscopic behavior of these electrodes is apparent, although a shift ca. 50 mV s observed between the two signals in ITO/PB electrodes. This shift in peak position between electrochemical and optical signals has been reported in the past and, again, is attributed have been adversely affected. 12 + to the mas transport limitations of K exchange from the Prussian Blue lattice, and which is PB coating on the screen-printed electrodes is unknown. Therefore, although the actual surface involved in the colouration process. [34] In spite of this slight delay between the current and the colour change, neither the electrocatalytic performance of the material nor its electrochromism have been adversely affected. 13 FIGURE 2 AROUND HERE The performance and stability of the electrochromic material under a continuous sw tching regime was also studied. The potential of the working electrode was stepped betw en -0.1 nd +0.4 V vs. Ag/Ag+ in 60-second intervals in a 0.25 mM H2O2 solution prepared in the s pporting electrolyte. The experiment was carried out in the presence of hydrogen peroxide to simulate the conditions of an oxidase-based biosensor where hydrogen perox e is produced. Figure 2C shows the change in absorbance as reducing and oxidizing potent als are applied. The data show that ITO/PB electrodes display a higher contrast between e oxidized and reduced states than SiO2-ATO/PB electrodes do. This is mainly due to two reasons. First, the higher PB loading of the ITO nanoparticle-based electrodes, confi ed by the charge densities in Table 1 above, makes these electrodes show a much deeper lue colour than SiO2-ATO based electrodes, and so when Prussian Blue is reduced into the colourless Everitt’s salt or Prussian White (PW), the contrast to the pale yellow of the underlying nanoparticles is enhanced. The second reason seems to be a wetting issue of SiO2-ATO-based electrodes. Although both types of electrodes are highly hydropho ic due o the Viton binder (see Table S3), SiO2-ATO-PB electrodes wet more slowly than ITO based electrodes. This is likely due to differences in surface topology and roughness. Eventually, however, the electrolyte wets the entire surface, allowing the interaction between more PB particles and the solution. Although this enables the maximum contrast of the material on electrochromic switching, this is still lower than at PB-ITO nanoparticle-based electrodes. 13 Figure 2D shows this gradual contrast increase at SiO2-ATO-PB electrodes, from an initial 0.5 a.u. up to roughly 0.7 a.u. On the other hand, this effect is not observed when a polymeric layer, such as chitosan in this work, covers the electrochromic material. The chitosan coating seems to wet the el ctrode surface completely, reducing the hydrophobicity of the material and facilitating the lectrode interaction with the test solution (see Table S3). 3.2 PB electrodes for H2O2 sensing The various Prussian Blue electrodes were used in the electroche ica determination of hydrogen peroxide, and their behavior was compared. Electrodes were polarized at a reduction potential of -0.1 V vs Ag, where Prussian White is formed and e catalytic reduction of peroxide takes place. Figure 3 shows current density as a functi of e oxide concentration for the 4 electrodes tested. In all cases, a linear response is b rved between 25 μM and 5 mM H2O2. The amperometric sensitivities of the f ur el trode types are very similar, roughly 0.08 A·cm-2·M-1, and detection limits are in th range of 1 μM to 4 μM (see Table S4). This is in line with other reports on PB-modified gra hite electrodes (see Table 1), and indicative of the suitability of these electrodes s an tific al peroxidase for sensing applications, particularly in non-invasive detection s stems. FIGURE 3 AROUND HERE In addition to enabling the amperometric detection of H2O2, these PB-based electrodes also allow optical quantification through reflectance spectroscopy. It is surprising that, whilst the 14 electrochemical detection of H2O2 at Prussian Blue electrodes is amply demonstrated and 15 accepted, there are much fewer examples of optical quantification of H2O2 at Prussian Blue electrodes, including electrodeposited Prussian Blue layers. [5, 14, 46, 47] The electrochromic properties of the fabricated PB-SPEs were also exploited in the optical quantification of hydrogen peroxide. The measurement was carried out as follows. First, P ssi n Blue was reduced to Prussian White at potential of -0.1 V vs. Ag. Once a stable b ckgro nd colorimetric signal was observed, after approximately 60 seconds of electrode p ariz tion, the potentiostat was switched off, leaving the electrochemical cell at open circuit. hen, the peroxide in the solution chemically oxidized the Prussian White back o ussian Blue, and the corresponding colour change was monitored by UV-Vis reflectance (Fig. 4A). Maximum colour contrast depends on the colour of the underlying conducting particles, which were white or pale yellow respectively. These were found to be 675 m for SiO2-ATO/PB electrodes and 700 nm for ITO/PB. Figure 4B shows plots of absorb e v rsu time at these wavelengths, as a means to monitor the hydrogen peroxide reduction t the PW/PB surface. As the data show, the higher the H2O2 concentration, the faster e electrode recovers its blue colour fully. Figure 4C shows the relation between peroxide concentration and reflectance measured at 700 nm, 100 seconds after electrode depolar atio In a manner akin to coulometric sensors, which sensitivity and detection limit can be adj sted by choosing a suitable (current) integration time, here the sensor performance can be djusted to the sample concentration simply by adjusting the integration time of the spectrophotometer and/or the sampling time after electrode depolarization (in our case 300 ms d 100 s, respectively). This provides control over the sensitivity, linear range or detection lim t of the method (Table S5). FIGURE 4 AROUND HERE 16 Longer experimental times result in markedly increased sensitivities, and consequently lower detection limits. Moreover, the sensor dynamic ranges also vary, so the technique can be easily adjusted to match a sample expected concentration, and thus improve the reliability f the measurement. Note that, although dissolved oxygen is also able to oxidize Prussian White ba k to Pr ssian Blue, the rate of this reaction is considerably slower than the peroxide d iven oxidation, and so its effect may be ignored and the process considered selective towards p roxi e, barring cases of extremely low peroxide concentration. When comparing the two PB-based electrochromic electrodes, ITO electrodes display higher absorbances and sensitivity values than ATO-bas d elect ode . This may be due to the higher amount of electrochromic material present in the paste which enhances the colour contrast between the reduced and oxidized states of the material (see Table S1). Tables S4 and S5 show that e optical method is slightly more sensitive. Controlling integration and sampling times n the spectroscopic detection lets the user modify the sensitivity of the measurement and adju it to the specific requirements of the application. This makes this approach a pow rful an versatile basis for the design of enzymatic biosensors. 3.3 Electrochromic glucose biosensor Once the suita ility of the PB-SPEs as a sensing platform for hydrogen peroxide was e ablished, a glucose biosensor based on glucose oxidase was prepared. The enzymatic oxi ation of glucose yields hydrogen peroxide which, at the PB-modified electrodes, enables the quantification of glucose both electrochemically and spectroscopically. Figure 5a shows the amperometric Michaelis-Menten plots for glucose using the two 16 fabricated electrochromic PB-SPEs. The dynamic range of both biosensors is slightly shorter 17 than other reported biosensors (see Table 1), ranging from 0.1 to 1 mM glucose and reaching saturation above 2.5 mM. We believe that at this point the amount of hydrogen peroxide produced by the oxidase inhibits the enzymatic activity of GOx. This could be solved by improving the fabrication of the glucose biosensor, using different polymeric membranes nd/or stabilizing agents to control the diffusion of the analyte through the biosensor, and to increase e operational stability of the enzyme. [48] On the other hand, the analytical perfo mance of our blue biosensors in terms of detection limit and sensitivity, ca. 60 μM and ·10-3 A·cm-2·M-1 (see Table S6), is comparable to that of other reported amperometric biosens s (s e Table 1). FIGURE 5 AROUND HERE The maximum current densities registered for e glucose biosensors can be extrapolated to the hydrogen peroxide calibration plot in Figure 3 to h ve an estimation of the amount of hydrogen peroxide produced in the enzymati reacti . A value of 0.25 mM H2O2 is obtained, which in the spectroscopic experiment o gener te an absorbance of roughly 0.1 a.u. To increase this absorbance value, and also e sensitivity of the glucose sensor, the sampling time in the spectroscopic determination was modified from 100 s to 300 s. Figure 5b shows the spectroscopic calibration plots of the glucose biosensor for the two different PB-SPEs. As seen, no substantial differences are observed when either the a ometric or the spectroscopic detection method is used, as a similar dynamic range is achieved in all cases, reaching in both cases a plateau at concentrations higher than 2.5 mM. A slight gain in the detection limit is achieved as it decreases up to roughly 15 μM likely due to the 18 effect of the high background currents present in the amperometric detection that disappear when 19 a spectroscopic detection method is used but, all in all, the performance of both types of electrochromic electrodes is comparable, regardless of the detection method employed. 4. Conclusions and outlook We have presented a new kind of electrochromic, Prussian Blue based screen-print d sensors suitable for spectroelectrochemical detection. To the best of our knowledge, this is h firs report of a screen-printed electrochromic paste applied to sensing and biosensing. The Prussian Blue electrochromic pigment used is prepared by chemi ally growing a thin PB layer on the surface of two different kinds of tin oxide-based c nducting particles, namely 50- 100nm indium tin oxide (ITO) nanoparticles, and 3μm SiO2 parti les covered by a thin (ca. 100nm) antimony tin oxide (ATO) shell. These PB-modi ed p gments, combined with a Viton® binder system, yield high-quality screen-print ng pastes. However, the high electrical resistance of the resulting layers required printing the a ial on a conventional electrode surface. Also, to increase the colour contrast furt r, an undercoat of a similar SiO2-ATO paste was printed underneath the blue coating and over a graphite electrode. The electrocatalytic respon e of these blue pastes toward the reduction of hydrogen peroxide is in line with that of the commercial PB-modified screen-printed electrodes and materials. However, in contra to these, the materials presented here also enable the optical detection and quantification of hydrogen peroxide in the μM-mM range, which represents a clear advantage and opens new avenues for the design of novel multi-functional spectroelectrochemical devices. These electrochromic sensors pose important advantages compared to both electrodeposited PB and screen-printed PB-modified graphite electrodes. They overcome electrodeposition limitations, such as (i) the “wet” nature of the process, (ii) the need to control critical electrodeposition parameters such as current and potential distribution over large area substrates, 20 and (iii) it replaces an intrinsically “batch” process by a “continuous” large area process, screen- printing. On the other hand, compared to currently available PB-modified graphite pastes, the materials presented here enable the use of both electrochemical and optical detection, opening new design avenues. One possible direction would be the development of multipara etric devices based on arrays of electrochromic electrodes. A smart example of this has been recently presented by Zhang and Liu where lactic acid can be quantitatively determined by the naked eye.[16] Although in this work we have printed these electrochromic electrodes over graphite transducers, printing them over transparent electrodes would provide additional advantages, such as the possibility to observe the colour change from the back a transparent substrate. Although analysis of coloured samples, such as blood, may be m re challenging, one way to keep the colour contrast would be to protect the elect hro ic electrode with a white semipermeable coating or membrane. In any case, the possibility to screen-print electrochromic electrodes widely opens the range of possi app cations for electrochromism in analytical chemistry , including non-invasive sens rs and s a t sensing tags in food packaging, and even gas sensors. Last, this work has f us d o the use of PB as model electrochrome. However, other electrochromic aterials, including conducting polymers,[49, 50] which have also been used in the construction of ectrochromic sensors, could also be used in the mass production of new sensors using the approach presented here to formulate related screen-printing materials. Ac nowledgements MA is supported by FEDER funds managed by the Catalan Secretary of Universities and Research through project PROD-0000114 (Enterprise and Knowledge, Industry Department, Generalitat de Catalunya). JdC gratefully acknowledges financial support through a 2016 21 Leonardo grant from the BBVA Foundation. The authors are grateful to Milliken for providing the electroconductive powders (ECPs) featured in this work. Zelec materials by Milliken provide cost-effective, static-dissipative performance for coatings and polymers. They are non-volatile, non-corrosive, and resistant to heat, chemicals and humidity. Zelec enhances quality and durability by imparting resistivity throughout the lifetime of coatings, paints, security inks nd electronics products. 5. References [1] R.J. Mortimer, D.R. Rosseinsky, P.M.S. Monk, Electrochromic Materials and Devices, Weinheim, Germany: Wiley-VCH; 2015. [2] K. Itaya, N. Shoji, I. Uchida, Catalysis of the Reduction of Molecula Oxygen to water at Prussian Blue Modified Electrodes, J Am Chem Soc, 106(1 84) 34 -9. [3] A.A. Karyakin, O.V. Gitelmacher, E.E. Karyakina, A High-Se sitive Glucose Amperometric Biosensor Based on Prussian Blue Modified Electrodes, n l ett, 27(1994) 2861-9. [4] A.A. Karyakin, O.V. Gitelmacher, E.E. Karyakina, russian Blue-Based First-Generation Biosensor. A Sensitive Amperometric Electrode for Glucose Anal Chem, 67(1995) 2419-23. [5] R. Koncki, T. Lenarczuk, A. Radomska, S. Głąb, Optical biosensors based on Prussian Blue films, The Analyst, 126(2001) 1080-5. [6] Q. Chi, S. Dong, Amperometric biosen based on the immobilization of oxidases in a Prussian blue film by electrochemical codeposition, Anal Chim Acta, 310(1995) 429-36. [7] A.A. Karyakin, E.E. Karyakin L. rton, Prussian-Blue-based amperometric biosensors in flow-injection analysis, Talant , 43(1 96) 1597-606. [8] S.A. Jaffari, J.C. Pick p, N vel hexacyanoferrate (III)-modified carbon electrodes- application in miniaturized bio nsors with potential for in vivo glucose sensing, Biosens Bioelectron, 11(1996) 1167-75. [9] J. Kim, J.R. Sempionatto, S. Imani, M.C. Hartel, A. Barfidokht, G. Tang, et al., Simultaneous Monitoring of Sweat a Interstitial Fluid Using a Single Wearable Biosensor Platform, Adv Sci, (2018) 1800880. [10] S. Bhar thi, O Lev, Sol-gel-derived Prussian blue-silicate amperometric glucose biosensor, Appl Biochem Biote hnol, 89(2000) 209-16. [11] . S. Tseng, L.-C. Chen, K.-C. Ho, Amperometric detection of hydrogen peroxide at a Pruss n Blue-modified FTO electrode, Sensors and Actuators B: Chemical, 108(2005) 738-45. [12] L. Chen, K. Ho, Multimode optoelectrochemical detection of cysteine based on an el trochromic Prussian blue electrode‫ۼ‬, Sensors and Actuators B: Chemical, 130(2008) 418- 24. 22 [13] P. Virbickas, A. Valiūnienė, A. Ramanavičius, Towards electrochromic ammonium ion sensors, Electrochem Commun, 94(2018) 41-4. Leonardo [14] grant from the A. Zloczewska, A.BBVA Foundation. Celebanska, The authors K. Szot, are grateful to D. Tomaszewska, M.Milliken Opallo,for M.providing Jonsson- Niedziolka, Self-powered biosensor for ascorbic acid with a Prussian blue electrochromic display, Biosens Bioelectron, 54(2014) 455-61. 23 [15] A. Ramanavicius, A.I. Rekertaitė, R. Valiūnas, A. Valiūnienė, Single-step procedure for the modification of graphite electrode by composite layer based on polypyrrole, Prussian blue and glucose oxidase, Sensors and Actuators B: Chemical, 240(2017) 220-3. [16] F. Zhang, T. Cai, L. Ma, L. Zhan, H. Liu, A Paper-Based Electrochromic Array for Visualized Electrochemical Sensing, Sensors (Basel), 17(2017). [17] A.A. Karyakin, E.E. Karyakina, Prussian Blue-based `artificial peroxidase' as a transducer for hydrogen peroxide detection. Application to biosensors, Sens Actuators, B, 57(1999) 2 -73. [18] D. Moscone, D. D'Ottavi, D. Compagnone, G. Palleschi, Construction and An tical Characterization of Prussian Blue-Based Carbon Paste Electrodes and Their s m ly as Oxidase Enzyme Sensors, Anal Chem, 73(2001) 2529-35. [19] F.S. da Cruz, F.d.S. Paula, D.L. Franco, W.T.P. dos Santos, L.F. Ferreira, E tro hemical detection of uric acid using graphite screen-printed electrodes mod fied with Prussian blue/poly(4-aminosalicylic acid)/Uricase, J Electroanal Chem, 806(2017) 172-9. [20] M. Buleandra, A.A. Rabinca, C. Mihailciuc, A. Balan, C. Nichita, I. Stamat n et al., Screen- printed Prussian Blue modified electrode for simultaneous detection of hydroquinone and catechol, Sens Actuators, B, 203(2014) 824-32. [21] V. Scognamiglio, I. Pezzotti, G. Pezzotti, J. Cano, I. Manfredon a K. Buonasera, et al., Towards an integrated biosensor array for simultaneous and rapid multi-analysis of endocrine disrupting chemicals, Anal Chim Acta, 751(2012) 161-70. [22] E. Costa Rama, J. Biscay, M.B. Gonzalez Garcia, A. Julio Reviejo, J.M. Pingarron Carrazon, A. Costa Garcia, Comparative study of different alcohol sensors based on Screen- Printed Carbon Electrodes, Anal Chim Acta, 728(2012) 69-76. [23] J. Noiphung, T. Songjaroen, W. ungchai, C.S. Henry, O. Chailapakul, W. Laiwattanapaisal, Electrochemical detect n of glucose from whole blood using paper-based microfluidic devices, Anal Chim Acta, 788( 13) 3 -45. [24] W. Dungchai, O. Chailapakul, S. Henry, Electrochemical Detection for Paper-Based Microfluidics, Anal Chem, 81(20 ) 5821-6. [25] T. Shimomura, T. Sumiy , M. o, T. Itoh, T.-a. Hanaoka, An Electrochemical Biosensor for the Determination of La tic Acid in Expiration, Procedia Chem, 6(2012) 46-51. [26] J.A. Bonacin, P.L. Santos, V Katic, C.W. Foster, C.E. Banks, Use of ScreenǦprinted Electrodes Modified by Pr ian Blue and Analogues in Sensing of Cysteine, Electroanal, 30(2017) 170-9. [27] V.D. Neff, El ctrochemical Oxidation and Reduction of Thin Films of Prussian Blue, J Electrochem oc, 25(1978) 886-7. [28] C.L. Lin, K.C H , A study on the deposition efficiency, porosity and redox behavior of 24 Prussian blue thin f lms using an EQCM, J Electroanal Chem, 524-525(2002) 286-93. [29] P. Najafisayar, M.E. Bahrololoom, Pulse electrodeposition of Prussian Blue thin films, Thin Solid Films, 5 2(2013) 45-51. [ A.N. Sekretaryova, V. Beni, M. Eriksson, A.A. Karyakin, A.P.F. Turner, M.Y. Vagin, Cholesterol Self-Powered Biosensor, Anal Chem, 86(2014) 9540-7. [31] E.V. Karpova, E.E. Karyakina, A.A. Karyakin, Communication—Accessing Stability of Oxidase-Based Biosensors, J Electrochem Soc, 164(2017) B3056-B8. [32] S. Cinti, R. Cusenza, D. Moscone, F. Arduini, Paper-based synthesis of Prussian Blue Nanoparticles for the development of whole blood glucose electrochemical biosensor, Talanta, 187(2018) 59-64. 25 [33] B. Kong, C. Selomulya, G. Zheng, D. Zhao, New faces of porous Prussian blue: interfacial assembly of integrated hetero-structures for sensing applications, Chem Soc Rev, 44(2015) 7997-8018. [34] N.F. Zakharchuk, B. Meyer, H. Hennig, F. Scholz, A. Jaworksi, Z. Stojek, A comparative study of Prussian-Blue-modified graphite paste electrodes and solid graphite electrodes with mechanically immobilized Prussian Blue, J Electroanal Chem, 398(1995) 23-5. [35] M.P. O'Halloran, M. Pravda, G.G. Guibault, Prussian Blue bulk modified screen-print d electrodes for H2O2 detection and for biosensors, Talanta, 55(2001) 605-11. [36] J.P. Coleman, A.T. Lynch, P. Madhukar, J.H. Wagenknecht, Printed, flex le electrochromic displays using interdigitated electrodes Sol Energy Mater Sol C lls, 56(19 9) 395-418. [37] M.A. Pellitero, Á. Colina, R. Villa, F.J. del Campo, Antimony tin oxi e (ATO) screen- printed electrodes and their application to spectroelectrochemistry, Elec ochem Commun, 93(2018) 123-7. [38] G. Ruiz-Vega, M. Kitsara, M.A. Pellitero, E. Baldrich, F.J. el am o Electrochemical Lateral Flow Devices: Towards Rapid Immunomagnetic Assays, ChemElectroChem, 4(2017) 880-9. [39] A. Abellan-Llobregat, I. Jeerapan, A. Bandodkar, L. Vi al, . Canals, J. Wang, et al., A stretchable and screen-printed electrochemical sensor for gl cose determination in human perspiration, Biosens Bioelectron, 91(2017) 885-91. [40] G.L. Long, J.D. Winefordner, Limit of Dete ion. A closer look at the IUPAC definition, Anal Chem, 55(1983) 712A-24A. [41] J.P. Coleman, A.T. Lynch, P. Madhuka , J.H. Wagenknecht, Antimony-doped tin oxide powders: Electrochromic materials for pr nted displays, Sol Energy Mater Sol Cells, 56(1999) 375-94. [42] R. Schiller, G. Battistig, J. Rab i, Rev rsible electrochemical coloration of indium tin oxide (ITO) in aqueous solutions, diat hys Chem, 72(2005) 217-23. [43] F. Ricci, A. Amine, G. P lleschi, D. M scone, Pussian blue based screen printed biosensors with improved characteristics of long-term lifetime and pH stability, Biosens Bioelectron, 18(2003) 165-74. [44] N.B. Li, J.H. P rk, . Park, S.J. Kwon, H. Shin, J. Kwak, Characterization and electrocatalytic prop ties of Prussian blue electrochemically deposited on nano-Au/PAMAM dendrimer-modified gol electrode, Biosens Bioelectron, 23(2008) 1519-26. [45] E. Laviron, G ne al expression of the linear potential sweep voltammogram in the case of diffusionless electroch mical systems, J Electroanal Chem, 101(1979) 19-28. [46] H. Liu, R.M. C ooks, Paper-based Electrochemical Sensing Platform with Integral Battery and Electrochrom c Read-out, Anal Chem, 84(2012) 2528-32. [47] A. Khorami, J.F. Botero-Cadavid, P. Wild, N. Djilali, Spectroscopic detection of 22 Hydrogen peroxide with an optical fiber probe using chemically deposited Prussian blue, Electrochim Acta, 115(2014) 416-24. [48] A.P.F. Turner, I. Karube, G.S. Wilson, Biosensors. Fundamentals and Applications, 1st ed., New York: Oxford University Press; 1987. [49] M.A. Deshmukh, M. Gicevicius, A. Ramanaviciene, M.D. Shirsat, R. Viter, A. Ramanavicius, Hybrid electrochemical/electrochromic Cu(II) ion sensor prototype based on PANI/ITO- electrode, Sensors and Actuators B: Chemical, 248(2017) 527-35. 23 [50] R. Celiesiute, A. Rmanaviciene, M. Gicevicius, A. Ramanavicius, Electrochromic sensors based on conducting polymers, metal oxides, and coordination complexes, Crit Rev Anal Chem, 4(2018) 1-14. 24 Captions to Figures Figure 1. Cyclic voltammograms obtained in supporting electrolyte for the different types of PB- SPEs. Scan rate of 5 mVs-1. SiO2-ATO/PB (Black), ITO/PB (Blue), DropSens C/PB (Orange), Gwent C/PB (Green). Figure 2. Cyclic voltammogram in supporting electrolyte and derivative of the voltabsortogram for a iO2-ATO/PB (a) and ITO/PB (b) electrode. Scan rate 5 mVs-1. (c) Plot of the evolution of ab rbance when potential was stepped from -0.1 to +0.4 V vs Ag/Ag+ at intervals of 60s in a 0.25 mM H2O2 solution in supporting electrolyte. (d) Representation of the contrast variation of the experiment in (c) for a longer time scale. 25 Figure 3. Plot of the current d sity against the hydrogen peroxide concentration. Applied potential of -0.1 V vs. Ag/Ag+ pseudo reference electrode. DropSens C/PB (Orange triangles), ITO/PB (Blue diamonds), Gwent C/PB (Green squares), and SiO2-ATO/PB (Black circles). 26 Figure 4. (a) Progress with time of the a sorbance spectra for an ITO/PB electrode in a 0.25 mM H2O2 solution in supporting electrolyte. (b) Evolution of the absorbance measured at 700 nm with time for different hydrogen p roxide concentrations in an ITO/PB electrode (see Figure S3 for the plots in ( ) and (b obtained with a SiO2-ATO/PB electrode). (c) Plots of the absorbance measured at 700 n f r a fixed time of 100 s in ITO/PB (Blue triangles) and SiO 2-ATO/PB (Black circles) electrodes. Error bars represent the standard deviation for three different e ectrodes. 27 Figure 5. (a) Amperometric and (b) spectroscopic calibration curves or the glucose biosensor obtained with ITO/PB (Blue triangles) and SiO2-ATO/PB (Black circles) ele trodes. Error bars represent the standard deviation of three different electrodes. Captions t Tables Table 1 Comparison of different Prussian-blue screen-printed sensors and biosensors used in the amperometric determination of H2O2. 28