Preparation of N-doped carbon materials from cellulose:chitosan blends and their potential application in electrocatalytic oxygen reduction

Carbon-based electrocatalysts for oxygen reduction reaction (ORR) are prepared by a direct pathway including a two-step thermal treatment process applied to porous spheres of natural biopolymer blends. Cellulose blends with chitosan are first thermally treated at moderate temperatures (e.g., 200 °C), then pyrolyzed at elevated temperature (800–1000 °C), both steps under a constant nitrogen flow. By blending of cellulose with chitosan, the nitrogen content in the final carbon-based catalyst can be considerably increased. The influence of the composition of the precursor biopolymer blend on the ORR electrocatalytic activity is analyzed in correlation with the elemental composition and other structural features of the catalyst. The polymer blend containing cellulose:chitosan = 75:25, thermally treated 1 h at 200 °C and pyrolyzed 1 h at 800 °C under nitrogen atmosphere, shows the highest electrocatalytic ORR activity. This is attributed to an increased surface area combined with relatively high nitrogen content and a higher pyridinic/pyrrolic species ratio.


Introduction
The current objective to maintain global warming below 2 °C compared to preindustrial levels is strongly connected with decreasing greenhouse gas (GHG) emissions and, especially, with the reduction of CO 2 emissions. Some of the challenges related to the zero-emissions goals were described in 2019 by Kaya et al. [1]. Reduction of the emissions can be achieved when greener and more sustainable technologies to store and convert the available green energy, by water splitting, fuel cells or in metal air batteries are further developed and improved [2]. The oxygen reduction reaction (ORR) has high technological importance in fuel cells, where it is used together with hydrogen oxidation reaction to provide energy [3]. Up to now, platinum is still considered to be the best single metal electrocatalyst for ORR [4], but its high cost, certain drawbacks and, more important, its limited availability hinder the commercial development of such technologies in very large scale [4][5][6].
In order to overcome these problems, a very high number of research studies were dedicated to the non-metal heteroatom-doped carbon materials as environmentally friendly, cheaper alternatives to noble-metal catalysts. These efforts were summarized in some recent reviews [7][8][9][10]. The reaction mechanisms involved during oxygen reduction on metal-free carbon-based catalysts were described by Ma et al. [11].
A very straightforward method for the preparation of efficient nitrogen-doped carbon-based electrocatalysts for ORR was developed by Wang and Jiang: carbon nanosheets were obtained by pyrolysis of folic acid at 900 °C under nitrogen atmosphere [12]. Zhang et al. proposed a relatively facile method for the preparation of a metal-free nitrogen-doped carbon catalyst by the thermal treatment at high temperatures under argon of zeolitic imidazolate frameworks impregnated or mixed with glucose [13]. Nitrogen-doped carbon catalysts with good activity were also obtained by the direct pyrolysis in argon atmosphere of gelatin mixtures with highly electroconductive carbon black-Ketjenblack [14]. Another direct method that allows the preparation of N-doped carbons with high surface area and good ORR activity was reported by Liu and co-workers and consisted of high-temperature pyrolysis in argon atmosphere of chitosan/urea 1/12 (w/w) composite; this method involved the use of urea in a high excess [15]. Nitrogendoped carbon nanofiber aerogels (N-CNF) were obtained from bacterial cellulose which was frozen in liquid nitrogen, freeze-dried, and then pyrolyzed under nitrogen at 800 °C. The final N-CNF material was obtained after a second pyrolysis step in NH 3 atmosphere at temperatures between 700 and 900 °C. The catalytic properties of these materials were very close to the properties of the Pt/C catalyst [16]. Very promising N-and P-doped carbon-based catalysts were also obtained by Gong et al., by pyrolysis at 800 °C under nitrogen of yeast pre-mineralized with Co(NO 3 ) 2 . 6H 2 O, followed by Co removal by washing with 1 M HCl and a second pyrolysis at 800 °C under nitrogen [17]. Very active N-doped ORR carbon-based catalysts were also obtained from cotton treated with fuming sulfuric acid, washed and dried, which was further pyrolyzed together with NH 4 Cl in high excess at different temperatures under nitrogen atmosphere. The samples pyrolyzed at 800 °C exhibited the best activity due to the high specific surface area and high nitrogen content [18]. Effective nitrogen-doped carbon catalyst with an onset potential of 0.89 V vs. RHE was obtained by a molten salt-assisted pyrolysis of chitosan. The used salts were ZnCl 2 aided by LiCl. The final porous carbonbased material with a very high specific surface area of ~ 1317 m 2 /g was obtained after the salt removal with hydrochloric acid and washing up to neutral pH [19]. Okuda et al. showed for a similar preparation process that nitrogen-doped carbons with increased specific surface area and high total pore volume can be realized by increasing the activation temperature. At raised temperatures, the selective formation of micropores by ZnCl 2 activation occurs [20]. Nitrogen-doped carbons with superior electrocatalytic activity were prepared from chitin/chitosan by a hard template method, with Na 2 CO 3 and CaCO 3 . For such a process, an optional urea treatment increased the specific surface area and raised the nitrogen doping, leading to an improvement of the catalytic activity [21]. Very good N-doped carbonbased catalyst for ORR reaction was also obtained by a two-step pyrolysis process in argon, starting with cotton impregnated with urea and a magnesium salt. After the pyrolysis the carbons were washed with 0.5 M HCl. The response of the electrocatalyst was observed to strongly depend on the temperature of the second pyrolysis step [22]. Another method for the preparation of an efficient metal-free carbon-based catalyst for the ORR was reported by Wütscher et al. [23]. This method involved a two-step process. First, cellulose was hydrothermally treated at 200 °C. After washing and drying, the material was pyrolyzed at high temperatures in NH 3 or 10% NH 3 /He atmosphere in the second step. The most efficient catalyst was obtained after pyrolysis under NH 3 at 800 °C [23]. The preparation of very promising electrocatalysts synthetized from sustainable precursors (chitosan) was also reported by Durante et al. [24]. The catalysts were obtained by two-step pyrolysis under nitrogen of chitosan or chitosan impregnated with 1,10-phenanthroline. The addition of a supplementary nitrogen source was found to improve the textural and the electrocatalytic properties of the materials [24]. The electrocatalytic activity of such materials can be further improved by the couse of precursors for the Fe-N x active sites formation [25].
Many of the above-mentioned methods produced nitrogen-doped carbon-based metal-free catalysts with reasonable efficiency toward ORR reaction; however, they involve complex and difficult to upscale procedures [16,18], use sacrificial additives [13,17,19,21,22], involve the use of a corrosive atmosphere during pyrolysis [23] or use co-reactants which emit corrosive gases at high temperatures [15,18].
In the present work, we focused on the development of a procedure that allows the conversion of the environmentally friendly and renewable biopolymers cellulose and chitosan to nitrogen-doped carbon materials, which are active in the electrocatalytic reduction of oxygen. In previous works, we had established conditions to prepare porous spheres from cellulose:chitosan blends via dissolution in ionic liquid-based solvent and shaping via dropping cum phase inversion in water [26,27]. Such spheres with different compositions have been used as precursors for the ORR electrocatalysts developed in the present work. The employed pyrolysis process of the biopolymer spheres minimizes the use and the emissions to the atmosphere of corrosive toxic gases, and by this it has a minimum impact on the environment. No additional activation step was involved. The chemical structure of the polymer blends as well was analyzed with help of attenuated total reflection infrared (ATR-IR) spectroscopy. The composition and structure of the obtained carbon materials were investigated by means of elemental analysis, X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, scanning electron microscopy (SEM) and specific surface area determinations, while their electrocatalytic activity and selectivity was assessed from rotating disk electrode (RDE) and rotating ring disk electrode (RRDE) experiments, respectively.

Materials
For the preparation of the biopolymer solutions, cellulose Avicel PH 101 with molecular weight (Mw) between 10,000 and 300,000 g/mol from Sigma-Aldrich and chitosan, 85% deacetylated, with "medium" molecular weight, i.e., Mw between 100,000 and 300,000 g/mol, also from Sigma-Aldrich, were used as received. The ionic liquid (IL) 1-butyl-3-methylimidazolium acetate ([Bmim][OAc]) in BASF quality (≥ 95%) was also purchased from Sigma-Aldrich and used without further purification. The ionic liquid is hygroscopic, having a water content below 1%, according to supplier information. Dimethylsulfoxide (DMSO; analytical reagent; ≥ 99.5%) from VWR International was used as co-solvent for the polymer dissolution. The solvent used for the biopolymer solutions preparation consisted in a mixture [Bmim][OAc]:DMSO = 2:1 (wt:wt).

Preparation of the cellulose:chitosan spheres
A cellulose solution with the concentration of 5 wt% and a chitosan solution with the concentration of 2.5 wt% were used to prepare the biopolymer blends in solution. This ensured that the resulting polymer blends have a similar viscosity and allowed the shaping of the spheres by dropping cum phase separation in a similar manner. The solutions were prepared as follows: Cellulose or chitosan in the desired concentration was dispersed in a mortar in the solvent (IL:DMSO = 2:1 mixture). The formed polymer dispersions were transferred to a closed flask and heated at 70 °C until the complete dissolution of the polymer. During the dissolution process, the mixtures were stirred several times in order to prevent polymer accumulation at the bottom of the flask. When the two homogenous biopolymer solutions had been obtained, they were mixed with each other in certain ratios as to obtain cellulose:chitosan = 75:25 or 50:50 blends. This procedure allowed the mixing of the biopolymers at molecular level. From the prepared polymer solutions, cellulose:chitosan spheres were prepared according to the procedures described in our previous works [26,27]. This method of preparation was preferred due to its simplicity and because it allows a relatively fast preparation of a homogenous solution of the blended biopolymers which can be very easy shaped and reprecipitated in water. The ionic liquid and DMSO can afterwards be recovered from the coagulation bath.

Pyrolysis process
For the pyrolysis of the samples, glazed porcelain combustion boats were loaded with porous spheres. The combustion boats were inserted in the middle of the glass tube of the pyrolysis furnace Heraeus Kelvitron S. To create the nitrogen atmosphere, this tube was then connected to the nitrogen circuit and flushed with nitrogen for about 15 min. The samples were then annealed at 200 °C for one hour and thereafter pyrolyzed at 800 °C (or 900 or 1000 °C) for another hour under nitrogen flow. After a two-hour natural cooling phase under nitrogen, the porcelain boat with the carbonized sample was removed from the furnace. The pyrolyzed materials were washed first with 0.05 N HCl solution and thereafter with deionized water, then dried and finally stored in closed snap-cap vials until further use. Yield was calculated as follows: m before = weight of the sample before pyrolysis; m after = weight of the sample after pyrolysis.

Characterization methods
The obtained cellulose:chitosan spheres were investigated by ATR FT-IR spectroscopy using a Bruker ALPHA-Platinum FT-IR spectrometer with ATR module.
Scanning electron micrographs of the porous spheres were recorded with a FEI ESEM Quanta 400 FEG instrument. For cross-section measurements, the samples were cut after freezing them in the liquid nitrogen. The samples were sputtered with Au/Pd (80/20) at 0.1 mbar and 300 mA until a 2-3 nm layer was obtained.
The porosity of the carbonized samples was characterized by nitrogen adsorption using a Coulter SA 3100 surface analyzer. Prior to the measurement, the samples were degassed for 60 min at 120 °C and after weighing they were degassed again for 600 min at 50 °C. The analysis of the isotherms was done according to Brunauer, Emmet and Teller (BET) and Barlett, Joyner and Halenda (BJH) methods.
The relative amounts of C, H, N, O in the pyrolyzed samples were determined using a EURO EA elemental analyzer from EURO VECTOR. Each sample was measured two times and the presented result represents the average of the two measurements. In case of high variation between the results of two measurements, a third measurement was performed.
The XPS measurements were performed with an ULVAC-Phi 5000 Versaprobe II device with monochromatic Al K α source under 45° angle with respect to the sample surface. The binding energies were corrected after the C 1 s peak at 284.5 eV.
Raman spectroscopy measurements were performed using a Renishaw InViva Raman spectroscope with a 532 nm laser over a spectral range from 98 to 3200 Raman shift/cm −1 .

Oxygen reduction reaction experiments
The electrochemical measurements were performed in a three-electrode cell setup (RDE) or four-electrode cell setup (RRDE) using an Autolab PGSTAT bipotentiostat/galvanostat (Metrohm). Catalyst inks (5 mg/mL) were prepared by dispersing catalyst powders in a 49:49:2 volumetric mixture of ultrapure water, ethanol and Nafion® solution. After sonication for 15 min, the ink was drop cast on polished glassy carbon electrodes to obtain a mass loading of 210 µg/cm 2 . The modified glassy carbon electrodes, a platinum mesh, and an Ag/AgCl (3 M KCl) electrode were used as the working, the counter, and the reference electrodes, respectively. The counter electrode was kept in a compartment separated by a glass frit during the measurements to prevent contamination of the working electrode with potentially anodically dissolved Pt.
A 0.1 M KOH solution was used as electrolyte. Before each electrochemical experiment, the electrolyte was purged with argon or oxygen, respectively. The Arsaturated electrolyte was used to record for background currents by recording a linear sweep voltammogram (LSV) in the potential window 0 to − 1 V vs. Ag/AgCl (3 M KCl). Afterward, the electrode was rinsed with water and immersed in the O 2 -saturated electrolyte while continuously maintaining the O 2 steam on top of the electrolyte surface during the measurements. Electrochemical impedance spectra were collected at the open-circuit potential to determine the resistance of the solution (R S ). The frequency ranged from 50 kHz to 10 Hz with a 10 mV AC amplitude. Prior to the activity measurements, a conditioning step was conducted by ten cyclic voltammograms recorded in a potential window of 0 to − 1 V vs. Ag/AgCl (3 M KCl) with a scan rate of 100 mV/s without rotation. Activity measurements were done by LSV in the same potential window with a scan rate of 5 mV/s at a rotation speed of 1600 rpm. All measured potentials were iR-drop compensated and converted to the RHE scale according to Eq. (2). The measured pH value of the electrolyte was ~ 13.
A RRDE setup was used for the investigation of the catalyst selectivity. Similar to RDE experiments, LSVs were recorded in 0.1 M KOH electrolyte, purged with argon or oxygen, respectively. The LSV recorded in the Ar-saturated electrolyte was used to correct for background currents. LSVs were performed at the disk electrode in the potential region from 0 to − 1 V vs. Ag/AgCl (3 M KCl) at a scan rate of 5 mV/s and an electrode rotation speed of 1600 rpm, while a potential of 0.4 V vs. Ag/AgCl (3 M KCl) was constantly applied at the Pt ring electrode. The collection efficiency (N) of each electrode was determined in an Ar-saturated solution of 5 mM where i D and i R are the absolute values of the currents recorded at the disk and the ring electrodes, respectively. The selectivity for the generation of H 2 O 2 is calculated according to Eq. (4).

Results and discussion
The formation and the properties of the cellulose:chitosan blended polymer solutions in ionic liquid-based solvents has been amply discussed in one of our previous works [25]. While a 6 wt% biopolymer solution in [Bmim][OAc]:DMSO = 2:1 has an almost Newtonian behavior for pure cellulose, with the increase of the chitosan ratio in the blend the non-Newtonian character of the solution dramatically increases. This observation was explained by the stronger entanglement of the biopolymer chains in solution due to the presence of chitosan which has a higher molecular weight and longer chains when compared with cellulose [27]. Additionally, for a certain biopolymer concentration in solution it was observed that the water amounts necessary to destabilize the casting solution decrease with the increase of the chitosan content [26]. Both the viscosity of the biopolymer solution and its stability against the precipitation with water are depending on the solvent used, on the total biopolymer concentration in solution and on the cellulose:chitosan ratio used and strongly influence the course of the non-solvent induced phase separation process which is responsible for the porous structure formation. For example, for the same biopolymer concentration in solution it was observed that the tendency to form finger-like pores increases with the increase of the chitosan ratio due to the poorer stability of the biopolymer solution against precipitation (thermodynamic effect). On the other hand, the sponge-like porous structure in the bulk of the spheres is slightly denser for samples with higher chitosan content due to higher viscosity of the casting solutions leading to a delay in the mutual diffusion of the solvent from the casted biopolymer drop and of the water into the nascent bead (kinetic effect) [26,27]. Figure 1 presents the scanning electron micrographs of the porous cellulose:chitosan spheres before and after the pyrolysis at 800 °C.
Both types of spheres exhibit a predominantly macroporous structure which is conserved also after the pyrolysis process. The macroporous structure is also responsible for the relatively low values of the specific surface area which are S BET1 = 20.4 m 2 /g for the material obtained from cellulose:chitosan = 75:25 and are S BET2 = 1.5 m 2 /g for the material obtained from cellulose:chitosan = 50:50. The spheres containing cellulose:chitosan = 75:25 have a sponge-like porous structure decorated with macrovoids toward the outer surface and containing a huge macrovoid in the middle. The cellulose:chitosan = 50:50 spheres exhibit a lamellar-like macrovoid structure. Because the two used casting solutions differ both in the cellulose:chitosan ratio and in the total biopolymer concentration in solution, it is difficult to assign the differences in the formed porous structures to a certain effect (thermodynamic or kinetic). Based on the previous observations [26,27], we may however assume that both the higher solvent:biopolymer ratio and the lower stability against precipitation with water of the cellulose:chitosan = 50:50 casting solution are responsible for the formation of the lamellar-like macrovoid structure. Additionally, the spheres cellulose:chitosan = 75:25 presented in this work exhibit much larger macrovoids in  the center of the sphere accompanied by other smaller macrovoids when compared with the biopolymers composition presented in our previous work [27] because the biopolymer concentration in the casting solution was lower. Cellulose and chitosan have similar chemical structures, except for the amino groups contained by the chitosan, therefore their signals in the FT-IR spectrum are also very similar (Fig. 2). The only distinctive feature in the spectrum of chitosan is the band at ~ 1583 cm −1 which can be attributed to the NH bending vibration [28]. This band in the spectra is more pronounced for the pure chitosan and for the cellulose:chitosan = 50:50 sample. Table 1 presents the elemental analysis of the carbon-based spheres obtained after the pyrolysis at 800 °C of the biopolymer-based precursors.
As expected, the nitrogen content in the pyrolyzed material is higher for a higher chitosan fraction in the biopolymer blend. The low hydrogen content in the pyrolyzed samples confirms the successful carbonization. The still relatively high oxygen content may be due to an incomplete pyrolysis of the non-volatile residue of the primary pyrolysis, i.e., the char [29], under the conditions of a relatively fast heating rate and short pyrolysis duration. The yield was ~ 16.2% for the cellulose:chitosan = 75:25 blend and ~ 19.2% for the cellulose:chitosan = 50:50 blend.
The Raman spectra of the pyrolyzed cellulose:chitosan blends are presented in Fig. 3. Based on interpretation of the Raman spectra of disordered and amorphous carbons explained in the three-stage model proposed by Ferrari and Robertson [30], all samples show the typical structure of sp 2 carbons containing the D and G peaks. While the D band is assigned to disordered-type carbon, the G band corresponds to the sp 2 hybridized carbon in the graphite lattice. The intensity ratio I D /I G characterizes the extent of disorder present within the material [31]. All four samples exhibit similar values for the I D /I G ratio reflecting a similar level of disorder. Only a small increase of the ratio can be observed for the samples obtained from cellulose:chitosan = 50:50, most probably as a consequence of the defects due to the larger fraction of nitrogen in the lattice of the pyrolyzed material. Changing the pyrolysis temperature from 800 °C to 1000 °C does not seem to have an important impact on the extent of disorder.  Figure 4 presents the N 1 s XPS spectra of the carbon-based spheres obtained after the pyrolysis at 800 °C of the biopolymer-based precursors. As discussed by Artyuschova [32], the N 1 s XPS spectra of the nitrogen-doped carbons prepared by thermal routes are difficult to interpret because the materials contain different nitrogen bonding configurations and effects like hydrogenation, protonation or oxidation may also affect the position and intensity of the peaks. The N 1 s XPS spectrum of the pyrolyzed cellulose:chitosan = 75:25 reveals four peaks at ~ 398 eV, 399.5 eV, 400.7 eV and 402 eV, which can be ascribed to pyridinic nitrogen, pyrrolic nitrogen, hydrogenated nitrogen and protonated nitrogen, respectively [32,33]. In the XPS spectra of the corresponding material obtained from the cellulose:chitosan = 50:50 blend, however, the contents of hydrogenated nitrogen and of the pyrrolic nitrogen are increasing while the protonated nitrogen is decreasing considerably in comparison with the 75:25 blend. The ORR electrochemical activity of the two N-containing carbon materials obtained after pyrolysis at 800 °C was evaluated in O 2 saturated 0.1 M KOH, using rotating disk electrode (RDE) experiments (Fig. 5a). Despite not reaching the high activity of Pt/C (half-wave potential of 0.89 V; current density of − 1 mA/cm 2 recorded at ~ 0.98 V vs. RHE [34]), nor the expected limited current density generated by a complete reduction of O 2 to H 2 O, the results indicate that the modulation of the activity can be achieved by blending the two biopolymers in different ratios. The carbon material obtained via pyrolysis of the cellulose:chitosan = 75:25 blend shows a significantly increased ORR catalytic activity compared with the material obtained from the cellulose:chitosan = 50:50 blend. While for the 50:50 blend − 1 mA/cm 2 current density is obtained at 0.29 V vs. RHE, for the 75:25 blend, this is recorded at 0.66 V vs. RHE. Besides, the current densities recorded on the 75:25 catalyst are more than double than on the 50:50, − 1.94 mA/cm 2 compared with − 0.79 mA/cm 2 when analyzing the two catalysts at 0.5 V vs. RHE. The difference in activity observed for the two carbon materials may be caused by the different types of N and their relative ratio on the surface. It was proposed that the ratio between the species identified at binding energy (BE) lower than 400 eV (pyridinic and pyrrolic) and of those obtained at higher BE than 400 eV (protonated and hydrogenated N) can be used to compare the electrocatalysts qualitatively and a higher ratio indicates a more active ORR electrocatalyst [32]. For both carbon materials, a relatively low ratio is calculated of 0.47 and 0.64 for the 75:25 and 50:50 blends, respectively, values fitting to those reported for not-so-active ORR electrocatalysts [32]. Although the 50:50 carbon material contains a higher N amount and a relatively higher pyridinic and pyrrolic N species content than the 75:25 blend, its ORR electrocatalytic activity is inferior. Considering the low N amount in the carbon materials and the limited resolution of the XPS, further conclusions based on the XPS data cannot be made. A second factor that may explain the high difference in the electrocatalytic activity is the relatively higher surface area (~ one order of magnitude) of the 75:25 blend-based carbon material compared with the one obtained from the 50:50 blend. For the less active catalyst (50:50), we evaluated the effect of mass loading on the recorded electrochemical ORR activity. The polarization curves presented in Fig. 5b show that the poor performance of the 50:50 blend pyrolyzed at 800 °C is maintained even if we double the mass loading, the increase in the recorded current density at a potential of 0.5 V vs. RHE being only of 0.1 mA/cm 2 . Therefore, the high activity of the 75:25 carbon material can still not be achieved. Based on this, we can conclude that the 50:50 material indeed has a lower catalytic activity than the 75:25 material and the poor recorded catalytic activity is not caused by an inhomogeneous film formation.
Since the catalyst prepared from the cellulose:chitosan = 50:50 precursor possesses considerably more nitrogen, as a consequence of the increased fraction of nitrogen in the polymer blend, we further explored the possibility to optimize the pyrolysis temperature aiming to induce the formation of a higher content of pyrrolic nitrogen. The pyrolysis of the cellulose:chitosan 50:50 blend was also performed at 900 and 1000 ℃ and their structural as well as their ORR electrocatalytic activity are further discussed. With the increase of the pyrolysis temperature from 800 °C to 900 °C, no decrease in the nitrogen content of the carbonized material was observed ( Table 2).
A further increase of the pyrolysis temperature to 1000 °C led, however, to a significant decrease in the nitrogen content in the pyrolyzed material. The oxygen content in the pyrolyzed material is clearly decreasing with the increase of the pyrolysis temperature, suggesting that a more advanced carbonization was achieved at 1000 °C. Concomitantly a slight decrease of the yields from ~ 19.2% after pyrolysis at 800 °C to ~ 18.0% after pyrolysis at 900 °C and ~ 16.7% after pyrolysis at 1000 °C was observed.
The N 1 s XPS spectra of the cellulose:chitosan = 50:50 samples pyrolyzed at different temperatures reveal a dependence of the nitrogen species as a function of the pyrolysis temperature (Fig. 6). All three samples contain the four peaks at ~ 398 eV, 399.5 eV, 400.7 eV and 402.5 eV, which can be ascribed to pyridinic nitrogen,  pyrrolic nitrogen, hydrogenated nitrogen and protonated nitrogen, respectively [32,33]. The total N content, the relative amounts of the different N species present in the samples and the corresponding ratios of the pyridinic/pyrrolic species are summarized in Fig. 7 for all cellulose:chitosan samples. The highest amount of nitrogen was detected in the sample cellulose:chitosan 50:50 (900 °C). The nitrogen content in the sample cellulose:chitosan 50:50 (1000 °C) was considerably lower (< 2 at %) (Fig. 6c). A large composition change was obtained when increasing the cellulose content in the blend from 50 to 75% and conducting the pyrolysis at 800 °C. Cellulose:chitosan 50:50 samples treated at 800 °C and 900 °C exhibit identical composition and differences in the fitting result seem to be due to the noise. Then the cellulose:chitosan 50:50 (1000 °C) sample seems to undergo slightly more change in the N species content. However, as the nitrogen content in the sample surface is very low, the low signal to noise level and the uncertainties in defining the background (Fig. 6c) make the interpretation very difficult. The increase of the pyrolysis temperature from 800 °C to 1000 °C for the cellulose:chitosan = 50:50 blend leads to a slight change of the LSVs recorded during ORR (Fig. 8). However, the electrochemical activity of these carbon materials is still considerably lower than cellulose:chitosan = 75.25 pyrolyzed at 800 °C and the slight variation of the N species formed at different synthesis temperatures cannot modulate the ORR electrocatalytic activity of the carbon material obtained from the 50:50 blend. These results are in agreement with the findings of Cheng et al. [35].
The reduction of oxygen can lead to H 2 O or H 2 O 2 formation and transfer of different number of protons coupled to electrons associated with the two products. While  (Fig. 9).
In RRDE, the H 2 O 2 formed at the disk electrode can be reduced on the ring electrode, thus leading to a Faradaic current and enabling the calculation of the amount  Table 3 (Fig. 9a). For the cellulose:chitosan = 50:50 blends pyrolyzed   Fig. 9b-d).
The results indicate that the ORR occurs via a mixed mechanism involving the transfer of 4 or 2 electrons, often reported for metal-free N-doped carbons. In contrast to the metal-free N-doped carbon materials, on commercial Pt/C (20 wt% Pt) the ORR typically involves the transfer of close to 4 electrons and only a small amount of H 2 O 2 is found at 0.50 V vs. RHE [36,37].
Artyushkova et al. [38] have found that the E 1/2 and the electrocatalytic activity of a metal-free nitrogen-doped carbons increase with the increase of the nitrogen content in the sample. They demonstrated that for metal-free nitrogen-doped electrocatalysts O 2 reduction occurs in a two-step process: O 2 → H 2 O 2 → H 2 O. They also have stipulated that pyrrolic N serves as an active site for the first step of the oxygen reduction reaction, while it was observed that pyridinic N does convert H 2 O 2 further to water but in a lower amount [38]. Based on this, we can assume that the sample cellulose:chitosan 50:50 (900 °C), although it has a slightly higher nitrogen content and a higher pyrrolic N content, due to a lower pyridinic/pyrrolic species ratio (Fig. 7a), exhibits a poorer overall electrocatalytic activity than the cellulose:chitosan 75:25 (800 °C) material.

Conclusions
In the present work, we propose the preparation of nitrogen-doped carbons as metalfree catalysts for the oxygen reduction reaction, starting with sustainable abundant natural precursors like cellulose and chitosan mixed at molecular level in ionic liquid-based solutions and shaped by dropping cum phase inversion in water to porous spheres. The proposed method, the direct pyrolysis of the porous biopolymer blends in N 2 atmosphere, aims to reduce toxic emission during the catalyst's synthesis and avoids the used of additional sacrificial activating agents.
The polymer blend containing cellulose:chitosan = 75:25 annealed at 200 °C and pyrolyzed at 800 °C yields a carbon material with improved catalytic activity compared with the blend containing cellulose:chitosan = 50:50. This is attributed to an increased surface area combined with relatively high nitrogen content and a higher pyridinic/pyrrolic species ratio. Increasing the pyrolysis temperature for the cellulose:chitosan 50:50 blend from 800 to 900 °C does not lead to an improvement of the pyridinic/pyrrolic species ratio. A further increase of the pyrolysis temperature to 1000 °C does increase the pyridinic/pyrrolic species ratio; however, the pyrrolic N content which promotes the first stage of the ORR process and the total N content in the pyrolyzed sample are strongly reduced.
Even though the electrocatalytic activity of the cellulose:chitosan 75:25 (800 °C) blend does not raise to the activity of similar materials reported in the specialized literature, which were prepared from similar precursors but where the pyrolysis procedures involved also activation steps, we showed that the electrocatalytic activity could be modulated by the ratio between the two biopolymers as well as by the pyrolysis conditions. In further studies, we aim at a further improvement of both the structural properties of the carbon materials (e.g., by increasing the specific surface area by using non-toxic nano NaCl templating agents) and of the N-species composition of the carbonized material by the fine tuning of the pyrolysis process.