1 Introduction

Paper-based fuel cells (PFCs) are being explored as a prospective power source for low-scale powering applications, besides being a clean-energy source, as paper is biodegradable. The PFCs have several advantages associated with them, such as easy availability, flexibility, porosity and low-cost [1, 2]. The porous nature of the paper facilitates the self-pumping of the fluids in the paper by the virtue of the capillary forces existing in the network of the cellulosic fibres of the paper [3, 4]. Thus eliminating the need for ancillary pumping mechanisms [5] and making them suitable for the development of portable power sources. Due to these advantages, PFCs can be considered as a potential alternative for coin/ button cells which are typically employed in digital diagnostic devices, such as digital pregnancy test kits, glucometers, uric acid meters. These devices consume low-scale power in the range of micro-milli watt. The primary use of the coin cells in these devices is to drive their micro-electronic circuits and power the visual display (LCD). Moreover, the disposal of the coin or button cells is an environmental concern, since it leads to the generation of electronic waste (e-waste), which can impact the human health and environment [6]. Besides, there is also limited availability of Li [7]. Alternatively, if these micro-electronic circuits are implemented on flexible substrates and can be driven by PFCs, this integration of the PFC-power source and flexible electronic circuits can lead to the development of flexible and disposable versions of the existing battery-operated diagnostic devices. This can aid in the upliftment of healthcare in low-resource settings by providing quick, easy and affordable diagnosis of common health conditions.

So far, methanol [8], hydrazine hydrate [9], hydrogen gas (generated in situ) [5], formic acid [3], formate [10] and hydrogen peroxide [11, 12] have been employed as fuels in different PFCs. The typical range of the reported peak power density is 0.5 to 100  mW cm−2 [5]. While, these PFCs have demonstrated promising cell performances, their integration with analytical systems would be more advantageous, if the fuel is bio-compatible with the system. For instance, developing PFCs as a power source for analytical devices, wherein the sample for the analysis serves as the fuel for driving the fuel cell (power source). Merino-Jimenez et. al have reported the working of a self-powered minimalist glucometer, wherein a microbial PFC drives the whole device with the sample (blood) serving as the fuel [13]. In this context, urine is an important biological fluid that contains vital bio-markers such as urea. To this end, urea-based PFCs can play a significant role in the development of self-powered electrochemical sensors for urine based analysis. Therefore, urea can be considered as a promising fuel as it can be sourced from human urine [14, 15], which is commonly used as a sample/reagent for performing diagnostic analysis in healthcare. Urea has an energy density of 16.9 MJ L−1 while being non-toxic, non-flammable, easy to store and transport (solid form) [16]. It is also commonly found in industrial effluents/waste water of fertilizer plants. The application of urea as a fuel has been investigated in direct urea-based fuel cells (DUFC), wherein power density up to 38.15 mW cm−2 has been achieved at 0.5 M urea at room temperature [17]. Generally, anion exchange membranes are employed in DUFCs to facilitate the transportation ionic (OH−) species in membrane-based systems. However, introducing an appropriate electrolyte, such as NaOH and KOH, which contain OH − ions, into the fuel (urea) aids in the ion migration and imparts ionic conductivity in membraneless systems. Recently, the development of PFCs powered by urea [18, 19] and urine [20] has been investigated elaborately by researchers. However, there are only two reports highlighting the practical demonstration of these fuel cells by powering the display of a digital pregnancy test kit [18, 20]. While the results obtained in these works are highly encouraging, the use of noble-metals, such as AgNO3 and Pt in the PFCs, limits their practical feasibility, as multiple such cells consisting of the noble metals need to be coupled to obtain the desired voltage (> 1 V). Thus, the use of these noble metal catalysts can compromise the vision of the development of cost-effective and affordable paper-based energy devices. Nevertheless, these findings clearly validate the idea of developing PFCs as a disposable, on-board power source in analytical systems. Further, studies relevant to the practical ability of the PFCs to drive an existing (micro) electronic system on a flexible substrate are quite limited [13].

In order to develop cost-effective PFCs, it has to be ensured that the electrode is free of precious metals. Pt, Pd, Ag and their bimetallic alloys with other metals, such as Ni, have shown good electrocatalytic activity towards urea electro-oxidation [21,22,23]. Alternatively, non-noble metals, such as nickel (Ni) and cobalt (Co) have shown good catalytic activity towards the electrochemical oxidation of urea [24,25,26]. Generally, doping Ni with other metals such as Co can enhance the active sites and reduce the onset potential [27], thereby improve the electrocatalytic activity of Ni-Co composite catalysts. The incorporation of a suitable carbon-based support, such as rGo can further improve the overall electronic conductivity, surface area and dispersion of the catalysts [28, 29]. Nanostructured forms of Ni, Co, Cd and Cu have been employed as the anode catalysts in various DUFCs, wherein power density of 13–14.2 mW cm−2 with Ni [21, 30], 21 mW cm−2 with Co dendrites [31], 1.57 mW cm−2 with Ni-Co [32], 26.9 mW cm−2 [33] with NiCu/ZnO@multiwalled carbon nanotubes (MWCNTs) have been attained. These findings motivate the development of urea-powered PFCs with non-noble metal-based catalyst on carbon support to generate desirable cell performance (voltage or power density) for various portable applications requiring µW power range.

Keeping in view the advantages of PFCs as a power source, urea as a fuel, Ni-Co supported on rGo as composite catalysts and integration of PFCs to drive flexible electronic circuits; the present work reports the development and characterization of urea-powered PFCs with hydrogen peroxide (H2O2) as an oxidant. Membraneless and membrane-based configurations of the PFC have been employed with different electrodes, such as CP, Ni-mesh and Ni-Co@rGo for the comparison of the electrochemical performance of the cells. The PFCs with Ni-Co@rGo electrodes delivered the maximum power density at 3 M urea. The practical utility of the PFC was demonstrated by employing it as a power source to drive an 11-stage ring oscillator (RO) implemented using oxide TFTs on a polyimide substrate, for the first time. RO circuit is ubiquitous in all electronic systems like analog to digital converters (ADCs) [34], signal processing units [35] communication systems [36], PLLs [37] and microprocessors [38]. This demonstration and integration of the PFC with flexible RO highlights their ability to be implemented as a power source in flexible electronic systems.

2 Experimental

2.1 Chemicals

Urea (NH2CONH2), 30 wt% hydrogen peroxide (H2O2), nickel chloride hexahydrate (NiCl2·6H2O), cobalt chloride hexahydrate (CoCl2·6H2O), isopropyl alcohol (IPA), graphite powder, hydrazine hydrate, sodium hydroxide (NaOH), were procured from Merck (Sigma Aldrich. Acetone, ethylene glycol and toluene were purchased from Finnar. Whatman filter paper (ashless, grade 42) was used as the scaffold for PFCs. Toray carbon paper (CP), Nickel mesh (Ni-mesh) and anion exchange membrane were purchased from Sainergy Fuel Cell. Ultrapure water was obtained from Millipore system. Oxide-based RO fabricated on a 30 µm thick polyimide substrate using a commercial fab-in-a-box process. The thin film deposition process used in fabrication of the oxide-based RO consists of a combination of physical vapour deposition and atomic layer deposition as described in [39].

2.2 Instrumentation and characterization methods

The electrochemical measurements were carried out using an electrochemical analyzer, CHI604E, CH Instruments. The measurement of the electrochemical performance of the PFCs was done via chronoamperometry (I vs t) and chronopotentiometry (V vs t), while the catalysts rGo and Ni-Co@rGo were characterized via cyclic voltammetry (CV). The structural characterization of the as-prepared catalysts, rGo and Ni-Co@rGo were done via High resolution field emission scanning electron microscope (HR FESEM) Zeiss, ULTRA Plus. The composition of NiCo@rGo was studied in situ by Energy Dispersive X-ray spectrometer (EDAX). ImageJ software was used to estimate the particle size of the catalysts. The X-ray powder diffraction of the catalysts was performed on a PANalytical Empyrean X-ray diffractometer using Cu-Kα radiation. The frequency response and peak-to-peak voltage measurement of the RO were performed using a Mixed Signal Oscilloscope-MSO2012B, Tektronix.

2.3 Cell construction and assembly

A schematic and photograph of the membraneless PFC and membrane-based PFC are shown in Fig. 1(a), (b) and (c), (d), respectively. The membraneless U-shaped PFCs were constructed by using Whatman filter paper as the substrate. As a first step, the paper was cut into a rectangular shape strip of dimensions 7 cm × 2 cm. A gap of dimensions 3.5 cm × 0.5 cm is introduced at the centre of the paper strip, giving it a U-shape. This strip was then placed on a glass slide. The electrodes (CP, Ni-mesh, Ni-Co@rGo@CP) used in the experiments have a geometric area of 0.25 cm2 and were placed each side of the U-shape strip to serve as the anode and cathode. Stainless steel plate (SS) served as the current collector. This cell assembly was referred as the membraneless PFC. The membrane-based PFC was constructed by simply replacing the glass slide with an anion exchange membrane as the base of the cell, while keeping the arrangement of all the other components same. 1 M NaOH with x M urea (x = 1, 2, 3, 4, 5) and 30 wt% H2O2 served as the anolyte and catholyte, respectively, in all the experiments. Each experiment consumed 30–40 µL of anolyte and catholyte, which were dropped manually near the electrodes using a hand held micropipette.

Fig. 1
figure 1

a Schematic and b Photograph of the urea-powered membraneless PFC. c Schematic and d Photograph of the urea-powered membrane-based PFC. Anode, Cathode: Ni-Co@rGo@CP. Anolyte: x M Urea (x = 1–5) + 1 M NaOH, Catholyte: 30 wt% H2O2. Current Collector: Stainless steel plate (SS) e Schematic and f Micro-photograph of the fabricated 11-stage RO

Full size image

2.4 Synthesis of rGo and Ni-Co@rGo

Graphene oxide (Go) was synthesized according to the Hummer’s method and subsequently rGo was synthesized following a procedure reported elsewhere [40]. First, a homogeneous dispersion was obtained by ultrasonication of Go in DI water (1 mg ml−1) for 1 h at a temperature less than 50 °C. The dispersion was then mixed with 2 mL of hydrazine hydrate and pH was adjusted to 10. The solution was kept in a hot air oven at 90 °C for 24 h. Later, the as-prepared flakes of rGo were washed several times with ethanol and DI water. Finally, this was left for drying in a hot air oven at 60 °C for 24 h. The final product obtained in the form of flaky black powder was referred as rGo. The Ni-Co@rGo nanocomposite was synthesized following an approach reported elsewhere [41]. First a solution was prepared by dispersing 50 mg of GO, 49.9 mg of NiCl2.6H2O and 50 mg of CoCl2.6H2O in 150 mL of ethylene glycol (EG), keeping the molar ration of Ni:Co = 1:1. This solution was then subjected to ultrasonication for 1 h to form a uniform dispersion. The resultant dispersion was subjected to a temperature of 110 °C in an inert (argon) atmosphere, resulting in an yellow-coloured solution. Further, a solution of 25 mL of hydrazine hydrate dissolved in NaOH (1 g) was added continuously, followed by 45 min of refluxing at 110 °C. The black product formed was then centrifuged, washed with copius amounts of water and ethanol to remove impurities. Finally, the product is dried in a vacuum oven at 45 °C. The final blackcoloured product thus obtained was referred as Ni-Co@rGo. To prepare the catalyst ink, 5 mg of catalyst was taken in 5 ml of DI water and sonicated for uniform dispersion. The catalyst ink was then drop-casted onto Toray Carbon paper and Glassy Carbon electrode, which are subsequently dried for 4–5 h.

2.5 Fabrication of the ring oscillator

The RO was fabricated on a 30 µm thick polyimide substrate using a fab-in-a-box process with a-IGZO as the semiconductor [42]. The schematic and the microphotograph of the fabricated RO are shown in Fig. 1(e) and (f), respectively.

3 Results and discussions

3.1 Structural and electrochemical characterization of rGo and Ni-Co@rGo

The structural and morphological properties of rGo and the Ni-Co@rGo composite were analysed by SEM and the composition of the composite catalyst Ni-Co@rGo was analysed by performing EDAX in situ. The corresponding SEM micrographs are presented in Fig. 2(a), (b) and (d), (e) respectively. The images obtained for rGo at low and high magnification, Fig. 2(a) and (b), respectively, show agglomerated, flaky and rough surface structure, indicating the expanded graphite layers in rGo. The rough surface morphology present in rGo is due to the exfoliated sheets which occurs during the oxidation of graphite to Go. The presence of C and O is shown in the EDX analysis Fig. 2(c) of the rGo samples. The presence of oxygen could be due to the residual O-atoms which were not completely removed during the reduction of Go. The growth of the bimetallic Ni–Co nanoparticles on to the sheets of rGo is depicted in the micrographs Fig. 2(d) and (e). It can be inferred from the micrographs that the agglomerated nanoparticles have a spherical, bead-like structure and adhere to the surface of the rGo in a non-uniform pattern. The average diameter of the bimetallic nanoparticles is ≈95 nm. The average diameter was estimated using the ImageJ software, wherein nearly 30 particles were considered as shown in the dotted circular region in Fig. 2 (d) The EDX analysis Fig. 2(f) confirms the presence of C,O, Ni and Co in the Ni-Co@rGo composite.

Fig. 2
figure 2

Scanning Electron Micrographs of a Reduced Graphene Oxide (rGo) and b selected area for EDX analysis of rGo c EDX analysis of rGo,Scanning Electron Micrographs of d Nickel–Cobalt nanoparticles supported on Reduced Graphene Oxide (Ni-Co@rGo), eselected area of Ni-Co@rGo selected for EDX analysis f EDX analysis of Ni-Co@rGo

Full size image

The XRD patterns of the as-prepared rGo and Ni-Co@rGo are presented in Fig. 3(a). The spectra for rGo shows a well-resolved broad peak at 2θ = 24.8 o corresponding to the [002] planes of carbon. A small diffraction peak at 2θ = 44.8 o is due to the planar structure of rGo [43]. The XRD spectra of Ni-Co@rGo depict well-defined and intense peaks 2θ = 44.5°, 51.8 o and 76.3 o and are indexed to [111], [200] and [220] face-centred cubic (fcc) lattice planes, respectively, as per the PDF no. 01-1260 [41]. The sharp peaks indicate the highly crystalline structure of the composite due to the presence of Ni and Co nanoparticles. Further, it can be seen that peak for the rGo is absent in this spectra, which can be attributed to relatively high crystallinity of Ni-Co than rGo in the composite and therefore, the peak corresponding to the latter is being masked by the former. This is consistent with the finding reported by  Bai et. al and He et. al. [41, 44].

Fig. 3
figure 3

Cyclic Voltammetry studies in 1 M urea and 1 M NaOH with Working electrodes: CP, Ni-mesh and Ni-Co@rGo@GCE, Counter electrode: Pt Rod, Reference Electrode: Hg/HgO. Potential Scan rate: 50 mV s−1. GCE: Glassy Carbon Electrode

Full size image

The electrochemical characterization of CP, Ni-mesh and Ni-Co@rGo was studied by CV and the voltammograms are shown in Fig. 3(b). The working electrodes were CP, Ni-mesh with an electrode area of 0.25 cm2, each and Ni-Co@rGo drop casted onto GCE with an electrode area of 0.071 cm2. Pt rod and Hg/HgO served as the counter and reference electrodes, respectively. The analyte consisted of 1 M urea and 1 M NaOH. The potential was swept, forward and reverse, between 0 and 0.8 V at a scan rate of 50 mV s−1. The CP based working electrode shows minimal catalytic activity with a total peak current density of ≈5 mA cm−2.

The voltammogram of Ni-mesh shows a broad oxidation peak at 0.5 V and a reduction peak at 0.4 V, which can be attributed to the formation of the surface-adsorbed intermediates, NiOOH and Ni(OH)2. In alkaline media, OH− ions get adsorbed to the surface of the nickel-based electrode and form Ni(OH)2, which further undergoes oxidation to form NiOOH. During the reverse sweep, NiOOH gets reduced back to Ni(OH)2 [32]. The NiOOH intermediate species, wherein, Ni exists in the + 3 oxidation state, primarily facilitates the electro-oxidation of urea [29].

The voltammogram of Ni-Co@rGo shows two well-defined oxidation–reduction peaks at 0.35 V and 0.2 V, respectively, wherein, the onset potentials of the formation of NiOOH and urea oxidation overlap with each other at ≈0.3 V, which typically occurs at 0.356 V [45] in alkaline medium. This has also been observed in the CV profiles obtained in this work, as shown in Fig. 3(b). This indicates that the NiOOH species were able to successfully catalyze the electro-oxidation of urea. The increase in the current density towards the attainment of the switching potential, i.e. at 0.8 implicates the diffusion of the OH− ions from the bulk solution towards the electrode surface, which still has available area to form more surface adsorbed intermediate hydroxide species. In addition, the incorporation of Co in the composite enables reduction in the onset potential of conversion of Ni(OH)2 to NiOOH by adding surface defects on Ni and thus providing more active sites on Ni [27] to aid in the urea electro-oxidation. In addition to this Co also enhances the overall electronic conductivity of the composite, which leads to the increase in the current density, as depicted in the CV profiles obtained for Ni-Co@rGo (Fig. 3(b). Further, it has been reported that Co can aid in the suppression of the undesired oxygen evolution reaction [46]. An oxidation peak current density of ≈90 mA cm−2 obtained with Ni-Co@rGo as the working electrode explains its good catalytic activity towards the electro-oxidation of urea, which were employed as the one of the sets of electrodes in the proposed PFCs in this work. These PFCs could deliver the maximum OCV, current density and power density were obtained as compared to other electrodes (CP and Ni-mesh).

3.2 Electrochemical performance measurement of the urea-powered PFC

The cell geometry and the assembly of the membraneless and membrane-based PFCs are presented in Fig. 1(a) and (c). The anolyte containing variable concentrations of urea with 1 M NaOH, and the catholyte consisting of 30 wt% H2O2 are self-absorbed into the PFC when dropped (manually) near the anode and cathode, respectively. The electro-oxidation of urea is a 6e− release process and occurs at the anode in accordance to Eq. 1 [47], as shown below:

$${\text{CO}}\left( {{\text{NH}}_{{2}} } \right)_{{2}} + {\text{ 6OH}}^{ - } \rightleftharpoons {\text{N}}_{{2}} + {\text{ 5H}}_{{2}} {\text{O }} + {\text{ CO}}_{{2}} + {\text{ 6e}}^{ - } ,{\text{E}}_{{\text{o}}} = \, - 0.{\text{75V}}$$
(1)

The electrons produced at anode travel through the external circuit (connecting wires) and react with H2O2 (oxidizing agent, catholyte) to furnish OH− ions, as per Eq. 2 [48].

$${\text{H}}_{{2}} {\text{O}}_{{2}} + {\text{ 2e}}^{ - } \rightleftharpoons {\text{2OH}}^{ - } ,{\text{E}}_{{\text{o}}} = { 1}.0{\text{4V}}$$
(2)

The overall theoretical cell potential is 1.79 V, which is seemingly promising in order to generate electrical energy from the above-mentioned fuel-oxidant (urea-H2O2) pair. The open circuit voltage (OCV) of the membraneless and membrane-based PFCs at different urea concentration were recorded via chronopotentiometry at an input current of 1 µA for 200 s at a temperature of 25 °C controlled by an air conditioner. A stabilization time of 10 s was maintained for all experiments before recording the cell performance. This was done in order to ensure that the anolyte and catholyte were absorbed sufficiently into the paper. A new cell was used for a given urea concentration and set of electrodes. Prior to optimizing the OCV of each cell, multiple experiments were carried out and a standard deviation of ± 1.5% was observed. Since all the PFCs were constructed and assembled with manual precision, this deviation within the experiments is expected.

The measured experimental OCV of membraneless and membrane-based PFCs at different urea concentration are presented in Fig. 4(a–d). The difference between the theoretical and experimental OCV can be attributed to several reasons based on the findings of the experiments performed in this work. First, the anisotropic nature of the Whatman filter paper effects the extent of fuel-oxidant flow rates and utilization at each of the electrodes. This in turn can impact the concentration of the species available at the electrodes and leads to an impeding cell potential in accordance to the Nernst Eq. (3) [49].

$${\text{E}}\, = \,{\text{E}}_{{\text{o}}} - \frac{{{\text{RT}}}}{{{\text{nF}}}}{\text{ln}}\,\left( {\text{Q}} \right)$$
(3)
Fig. 4
figure 4

Cell open-circuit voltage measurement w.r.t time at 1 µA current for Membraneless PFC with anode and cathode as a CP b Ni-mesh c Ni-Co@rGo@CP and d Membrane-based PFC with anode and cathode as NiCo@rGo@CP. Anolyte: 1 M urea + 1 M NaOH. Catholyte: 30 wt% H2O2

Full size image

R: universal gas constant, F: Faraday constant n: amount of electrons transferred in mol, T: is the absolute Kelvin scale temperature,

Q: reaction quotient.

According to Eq. (3) the cell potential (E) depends on the reaction quotient (Q), which in turn is governed by the concentration of the species (fuel/oxidant). The drop in OCV is more prevalent at lower urea concentration (1 M and 2 M) for all the PFCs. Second, the formation of products at the anode such as CO2 and H2O can also limit the OCV by reducing the Q, since it is the ratio of the activity coefficients of the reactants and products. Third, the formation of H2O at the anode leads to the dilution of the concentration of the anolyte. This is more prevalent at higher urea concentration of 5 M, wherein all the PFCs delivered an inferior OCV of < 0.5 V. This drop of OCV at 5 M urea is also due to the cross over and direct mixing of the anolyte-catholyte at the opposite electrodes in the present configuration of the PFCs, since there is no physical barrier between the electrodes. The separation between the electrodes is ensured by the slit between them, which cannot eliminate the fuel cross-over totally. Nevertheless, the OCV of all the PFCs is nearly stable with respect to time for all urea concentrations except at 5 M due to the prevalent cross over of anolyte at higher urea concentrations, which is also a time-dependent phenomenon. Therefore, the OCV of the all the PFCs was receding more rapidly with respect to time at 5 M as compared to lower concentrations. Finally, it is due to the above mentioned reasons the change in OCV of the as-proposed PFCs is observed. The membraneless PFC with Toray CP as the electrodes

(Fig. 4a) delivered an OCV ≈0.5 V which could be due to the limited kinetic activity of Toray

CP towards urea electrooxidation. The rest of the PFCs (Fig. 4(b–d)) delivered an OCV of ≈0.7 V at 3 M urea. Since the PFCs delivered a maximum OCV at 3 M urea, it was considered to be the optimized fuel concentration.

The electrochemical performances of the PFCs were investigated with the current–voltage (IV),

i.e. polarization studies. The IV studies were carried out via chronopotentiometry, by stepping up the current density from ≈2 µA cm−2 to its limiting (maximum) value and subsequently measuring the resulting voltage for 50 s, each time standard deviation of ± 1.5%. As shown in Fig. 4, the cell voltage was stable in the first 50 s for all the configurations. The membraneless PFC with CP electrodes (anode and cathode), Fig. 5(a), delivered a Pmax of 7 µW cm−2 and Jmax of 78 µA cm−2. This indicates that the carbon paper-based electrode has no to minimal activity towards urea electro-oxidation. When this was replaced with Ni-mesh as the electrodes (Fig. 5b), the Pmax and Jmax increased to 45 µW cm−2 and 275 µA cm−2, respectively. This can be attributed to the proven better electrocatalytic activity of Ni towards urea electro-oxidation [24]. Next, Ni-Co@rGo composite nanostructured catalysts were synthesized and loaded onto CP to serve as the electrodes (Fig. 5c). The Pmax and Jmax) delivered in this case were 55 µW cm−2 and 371 µA cm2, respectively. The enhancement in the cell performance is mainly due to the combined effects of Ni–Co nanoparticles and rGo, wherein doping Ni with Co, provides more active sites and rGo improves the surface area of the catalyst. Finally, the PFC was assembled onto an anion exchange membrane, with Ni-Co@rGo@CP electrodes resulting into a membrane-based PFC (Fig. 5d). The Pmax and Jmax obtained from the membrane-based PFC were 70 µW cm−2 and 500 µA cm−2, respectively, which were the highest among all the configurations. This is due to the fact that the anion exchange membrane facilitates the migration of OH− ions within the system and therefore the cell performance improves in comparison to the membraneless PFCs. In addition to this, it can be seen from the IV-curves that maximum power density is attained at 3 M urea across all PFC configurations.

Fig. 5
figure 5

Polarization studies of Membraneless-PFC with a CP b N-mesh c Ni-Co@rGo@CP and d Membranebased PFC with anode and cathode as Ni-Co@rGo@CP. Anolyte: 1 M urea + 1 M NaOH. Catholyte: 30 wt% H2O2

Full size image

Further, the change in cell performance w.r.t urea concentration is consistent in all the PFCs. This implicates that the cell performances are reproducible. The inferior cell performance at 4 M and 5 M urea is possibly due to fuel cross over. This was also observed during OCV measurement studies, as shown in Fig. 4. The initial drop in the OCV is due to the sluggish kinetics which contribute to activation overpotential losses in the cell(s). This can be observed in the PFC performances, shown in Fig. 5b–d), where the activation losses are prevalent upto ≈50 µA cm−2. The electrochemical performance exhibited by the urea-powered PFCs, indicates that they can be used for harvesting energy from human urine as well, which is a good source of urea. This further suggests that urea/urine powered PFC can serve as the power source in analytical devices for performing urine-based analysis and thereby rendering a self-powered device which is capable of generating power from the sample (urine) itself. It is worth to note that this work is almost the first demonstration of an urea-powered PFC with non-noble metal-based electrodes. Since the membrane-based PFC with Ni-Co@rGo@CP electrodes exhibited the best performance in comparison to the membraneless PFC, it was employed for the demonstration of the practical utility of the PFC for driving flexible electronic circuits, thereafter. Prior to this, a stack of two membranebased PFCs was prepared by connecting them in series, in order to obtain the desired voltage (> 1 V) for driving the flexible electronic circuits. The stack delivered a nearly stable OCV of ≈1.4 V (standard deviation ± 3.5%) for a period of 400 s. The corresponding result is shown in Fig. 6(a). Similarly, the limiting current density of the stack was also measured w.r.t time as shown in Fig. 6(b). Initially, the stack delivered a Jmax of 500 µA cm−2, which is in-line with the performance of the single PFC (Fig. 5d). However, after 100 s the current density of the stack starts declining due to the consumption of the fuel/ oxidant at the anode/cathode. Nevertheless, the membrane-based PFC stack possessed the desirable range of OCV and current density for driving flexible electronic circuits (RO). A comparison of the performance of the various urea-powered PFC reported in literature has been presented in Table 1. It can be noticed that the PFC proposed in this work has the non-noble metal-based catalysts on both the anode and cathode side, besides generating the desirable OCV and power density. While the rest of the PFCs consists of at least one noble metal catalyst one or both the electrodes. The measurements and subsequent response of the circuits driven by the PFC have been discussed in the next Sect. 3.3.

Fig. 6
figure 6

a Cell open-circuit voltage measurement w.r.t time at 1 µA current and b Limiting current density w.r.t time at 0.1 V operating voltage for the stack of two Membrane-based PFC. Anode and Cathode: Ni-Co@rGo@CP. Anolyte: 3 M urea + 1 M NaOH. Catholyte: 30 wt% H2O2

Full size image
Table 1 Performance comparison of the urea-powered PFCs
Full size table

3.3 Characterization of flexible electronic circuit (11-stage RO) driven by PFC

An 11-stage RO fabricated on a flexible wafer was driven by the stack of two PFCs to demonstrate the PFC driving capability, as shown in Fig. 7. This setup shows the two-PFC stack, a multi meter measuring the DC output voltage of the PFC stack, an RO (at wafer level) placed on a probe station and an oscilloscope showing the measured response of the RO. Figure 8(a) to (e) show the RO response when it is driven by the PFC stack with x M urea concentration (where x = 1 to 5). Table 2 presents voltage generated by the PFC stack (VDD), RO output voltage peak to peak (VPP) value and frequency of oscillation (f) for different urea concentration levels. This has been also presented in Fig. 9, which can serve as the calibration curve between the measured frequency and urea concentration. A maximum f of 37.52 kHz and VPP of 1.04 V were obtained at 3 M urea concentration. As it is expected from the functionality of the RO, f and VPP are proportional to the VDD generated by the stack of two PFCs under same loading condition. It can be noticed from Table 2 and Fig. 9, the whole setup can also be used as a sensor and readout circuit as the variation in the urea concentration is converted in to frequency. This setup opens a window for self-powered flexible POC devices to sense different bio-signals. Table 3 represents the power consumption of different oscillator circuits on flexible wafer as reported in the literature [50,51,52,53,54] at different VDD. It should be noted that only this work has shown the integration of the RO on flexible wafer with the PFC powered by urea, whereas other works are employing external DC power supply.

Fig. 7
figure 7

Test set-up demonstrating the characterization of a 11-stage RO on a flexible substrate driven by a proposed stack of two-membrane-based urea-powered-PFC

Full size image
Fig. 8
figure 8

The circuit response from experimental characterization, when it is driven by the stack of two-membranebased PFCs at a 1 M b 2 M c 3 M d 4 M e 5 M urea. Oxidant: 30 wt% H2O2. Anode, Cathode: NiCo@rGo@CP

Full size image
Table 2 Performance comparison of the proposed comparator
Full size table
Fig. 9
figure 9

Frequency vs urea concentration calibration curve based on the frequency response of 11-stage RO driven by stack of PFC, anolyte: x M Urea + 1 M NaOH (x = 1 to 5), Oxidant: 30 wt% H2O2. Anode, Cathode: Ni-Co@rGo@CP

Full size image
Table 3 Power consumption of different flexible electronic circuits as reported in literature at different power supply voltages
Full size table

4 Conclusion

In this work the development, working and practical application of urea-powered PFCs have been discussed. The electrochemical performance of the membraneless PFC was studied for different configurations employing different non-precious electrodes, namely, CP, Ni-mesh, whereas, Ni-Co@rGo composite catalyst loaded on CP was employed as the electrodes in both membraneless and membrane-based PFC. The CV profiles of the electrodes depicted that Ni-Co@rGo exhibited the maximum catalytic activity towards urea electro-oxidation with an onset potential of ≈0.35 V and peak current density ≈90 mA cm−2. Out of all the cell configurations, the membrane-based

PFC with Ni-Co@rGo@CP electrodes could deliver a Pmax of 70 µW cm−2 with an OCV of ≈0.7 V. Activation overpotential losses and low fuel utilization in paper-based systems are generally responsible for low-scale (micro-watt) power densities. A stack of two membrane-based PFCs with Ni-Co@rGo@CP as electrodes could deliver a nearly stable OCV of 1.4 V and current density of 500 µA cm−2 for 400 s. This stack could successfully drive an 11-stage RO, which was designed using a-IGZO TFT technology on a flexible polymeric substrate. The measured oscillating frequency and peak-to-peak voltage were 37.52 kHz and 1.04 V respectively at 3 M urea concentration. This demonstration evidently elucidates the serviceability of the urea-powered PFC with non-noble metal-based electrodes, as a power source for driving flexible electronic circuits and portable diagnostic/sensing devices which employ these kinds of circuits, which require low-scale power. When coupled together with appropriate sensing components, this leads off towards the development of self-powered flexible-portable sensing systems for diagnostic applications. Such analytical systems with flexible electronics for performing the urine-based diagnosis would be independent of external power sources for their operation as they will be driven by PFCs, powered by urea (a commonly occurring component in urine).