An anode catalyst support for polymer electrolyte membrane fuel cells: application of organically modified titanium and silicon dioxide

This work describes an attempt to improve the physical and electrochemical parameters of PEM fuel cells that have electrodes modified by titanium and silicon dioxides. A customized design of membrane electrode assemblies was proposed which is characterized to have an around 6 times higher concentration of catalyst at the cathode side (2.0 mgPt cm−2) in order to investigate the influence of anode catalyst support treatment. Anode catalyst support materials were modified using pristine TiO2 and TiO2–SiO2–VTMS – the composite was crosslinked with the aid of vinyltrimethoxysilane. Surface area and porosity analysis was carried out with the aid of BET, BJH, t-plot and Horvath–Kawazoe (H–K) theories for particular components of the support materials and their catalyst mixtures. The experiment revealed a positive influence of TiO2–SiO2–VTMS (BET 321.9 m2 g−1, BJH 3.7 nm) on the anode catalyst layer in terms of surface area (3-times increase, 75 m2 g−1) and average pore size (decrease from 25.3 to 15.7 nm). Additionally, favourable microporosity (pores less than 2 nm) was introduced to the material according to the H–K analysis results (10.3 m2 g−1, 0.65 nm). Electrochemical experiments, which include polarization curves, electrochemical impedance spectroscopy and cyclic voltammetry, have demonstrated the change of behaviour for the fabricated fuel cells with modified anodes against the reference sample. The mitigation of charge and mass transfer resistance (by 15–20%, 50 mV at 200 mA cm−2), the improvement of power density (up to 72%, 217 mW cm−2) and a better exposure of the catalyst to the reactants of an electrochemical reaction were observed for fuel cells modified by both pristine TiO2 and the hybrid TiO2–SiO2–VTMS-based compound.


Introduction
The performance and durability of polymer electrolyte membrane fuel cells (PEMFCs) are mainly affected by two factors: the physicochemical properties of proton exchange membranes (PEM) and the rate of the ORR reaction (Oxygen Reduction Reaction) at the cathode side. These aspects attract many scientic studies including materials research and investigation of the various phenomena inuencing fuel cell operation such as activation losses, ionic conductivity and decay mechanisms. 1 Despite the sluggish kinetics of the cathode reaction and less signicant problems with the ageing processes happening at the hydrogen side, the modication of the anode is also a matter of great concern to accelerate the commercialization of fuel cells in diverse applications and various conditions of operation.
Due to the fast rate of HOR reaction (Hydrogen Oxidation Reaction), around six orders of magnitude faster than ORR, 2 negligibly-small voltage losses are present at the anode side for Pt catalyst loading down to 0.05 mg cm À2 . 3,4 However, at a very low loading, there is a high risk of severe contamination which might come from the chemical compounds carried by the fuel. The contaminants such as carbon oxide, hydrogen sulde or ammonia can interact with platinum particles creating strong, practically irreversible chemical bonds, consequently decreasing the electrochemical surface area. 1 Due to this reason, a very low anode catalyst loading may not be preferred. On the other hand, high catalyst loading at anode side accelerates the phenomenon called reverse-current decay mechanism. 5 This mechanism produces cathode layer degradation via carbon corrosion of the catalyst support happening if oxygen is present at anode side in the fuel cell stack (in most cases during start-stop cyclings). The reduction of reverse current is one of the methods to suppress degradation processes. The adjustment of catalyst amount and modication of catalyst support is therefore appropriate approach towards preparation of durable and cost-effective anode materials. One of the approaches is an attempt to modify catalyst support applying titanium dioxide as a replacement of carbon-based support materials. For instance, in ref. 6 authors proposed a new strategy for alleviating the reverse current phenomenon using a method named atmospheric resistive switching mechanism (ARSM). In this method, the resistance of anode catalyst layer composed of Ta-doped TiO 2 -supported Pt varies depending on the gas atmosphere: in the air present at the anode side, the value of cell resistance is approximately one order of magnitude larger than in hydrogen (1.2 vs. 0.13 U cm 2 ) which helps minimizing destructive corrosion of the cathode.
Many other scientic papers describe the application of titanium dioxide in fuel cells 7 including its application to membranes and cathodes. TiO 2 is investigated despite the fact, the pristine TiO 2 is a semiconductor, i.e. it has the band gap of 4.85 eV and electrical conductivity as low as 10 À13 U À1 cm À1 at 298 K. 8 High conductivity of TiO 2 -supported catalyst layer can be created by introduction of crystal lattice defects to the TiO 2 , [9][10][11] via doping using various metals such as niobium, chromium or vanadium 6,[12][13][14] or preparing the mixture of TiO 2 and various allotropic forms of carbon. 9,[15][16][17] For instance, in, 11 fuel cell with titanium dioxide acting as a Pt support, achieved outstanding performance in comparison to Pt/carbon-based catalyst in terms of durability (electrochemical surface area of 11.0 vs. 3.8 m 2 g À1 aer 80 h of accelerated stress test). At the same time the sample obtained comparable initial power density 0.84 vs. 0.94 W cm À2 . Similar results were described in ref. 15, however the authors applied also carbon nanotubes (CNT) in order to increase the electrical conductivity. Experimental results revealed that CNT-Pt/TiO 2 nanober-based electrode shows enhanced performance with the maximum power density of 567 mW cm À2 compared to commercial Pt/ carbon (461 mW cm À2 ) with a signicant durability at harsh conditions of 120 C and RH 40%. The effect named as Strong Metal Support Interaction (SMSI) was demonstrated as an additional advantage of using TiO 2 , based on which the platinum dissolution and agglomeration of its particles are significantly limited leading to a great durability of fuel cell electrodes. The SMSI effect was also presented in ref. 10 for Nb-TiO 2 utilized as Pt support. In this work a signicant increase in the activity of Pt/TiO 2 for oxygen reduction reaction was shown along with the improvement of catalyst stability.
Fuel cell electrode materials are also modied using silicon. Carbides of this element were considered mainly due to good electrochemical and physical stability acting as a support for platinum particles. [18][19][20] Various applications require a special design of fuel cell stack that allows on a self-humidication process. It is relevant to applications where space and weight must by maximally limited (humidiers are not needed) or for open-cathode fuel cells which are susceptible to sudden changes of current density and temperature. Silicon dioxide, due to its hydrophilic properties, seems to be a good candidate for the electrode modication despite its poor electrical conductivity. For instance, in ref. 21 J. I. Eastcott and E. B. Easton examined the physical and electrochemical characteristics of sulfonated silica ceramic carbon electrodes for PEMFC working at low humidity. The silica-based cathodes were able to maintain stable fuel cell performance down to 20% RH, and had better performance than the Naon-based cathode. In this work, authors concluded that the presence of sulfonated silane in the cathode catalyst layer may aid in water retention and promote back-diffusion of water from the cathode to the membrane to improve MEA hydration. C.-L. Lin et al. 22 investigated silicon dioxide used as an anode catalyst layer to improve the performance of a self-humidifying proton exchange membrane fuel cells. Using just 0.01 mg cm À2 of sulfonated SiO 2 , the electrical performance of fuel cells was increased 29% and 59% when the fuel cell reaction gases were humidied at 70 C and 50 C, respectively. Similar conclusions were found in ref. 23, where various concentrations of SiO 2 were investigated. Silica nanopowders added to catalyst layers contributed to self-humidication effect improving the performance of fuel cell samples.
This work aims at the investigation of fuel cells containing both compounds i.e. TiO 2 and SiO 2 , and additionally vinyltrimethoxysilane (VTMS) acting as a binder. VTMS was used in organic modication process of titanium and silicon dioxide. Synthesized modiers were applied to anode catalyst supports. Fuel cells (FC) containing various concentrations of TiO 2 or TiO 2 -SiO 2 -VTMS in catalyst layer (CL) were prepared in order to perform electrochemical characterisation. Additionally, fuel cell with reference catalyst layer used as an anode was also fabricated. Unique properties of TiO 2 and SiO 2 change the characteristics of investigated samples in terms of performance, durability and also production costs due to lower surface density of platinum catalyst which is required for efficient operation.
The goal of this work was to show that it is possible to create efficient and not expensive PEMFCs with reduced amount of Pt by incorporation of cost-effective hybrid organic-inorganic or inorganic compounds obtained by sol-gel method.

Synthesis of TiO 2
Anatase crystals of titanium dioxide were synthesized based on sol-gel method using titanium(IV) isopropoxide (TIPO) as a titanium precursor. The synthesis procedure is as follows. TIPO (4.5 ml) was dissolved in ethanol (21 ml) mixed with 3.5 ml of distilled water to obtain a titania solution. The solution was mixed using a magnetic stirrer in a plastic ask at room temperature for 6 h. During the process, the TiO 2 powder was formed and once ltered it was dried at room temperature. Thereaer TiO 2 powder was heated at 500 C for one hour.

Synthesis of organically modied TiO 2 -SiO 2
Vinyltrimethoxysilane (VTMS) was applied for organic modication of TiO 2 and SiO 2 . VTMS is an organosilicon compound with the formula H 2 C]CHSi(OCH 3 ) 3 . It was used in a form of colorless liquid. The powder of TiO 2 -SiO 2 -VTMS was obtained carrying out hydrolysis of titanium and silane precursors i.e. titanium(IV) isopropoxide (TIPO) and tetraethoxysilane (TEOS), respectively. According to the synthesis procedure, TIPO (2 ml), TEOS (2 ml) and VTMS (2 ml) were separately mixed with 96 vol% ethyl alcohol (7 ml) and distilled water (3.5 ml) at room temperature, for two hours using magnetic stirrer. NH 3 (4 ml) was used as a catalyst of the synthesis reaction. Aerwards, particular solutions were mixed one with each other in order to obtain TIPO : TEOS : VTMS molar ratio of 1 : 1.3 : 2. The combined solution was stirred additionally for 5 hours at room temperature. Finally, the yielded material was precipitated in the form of white powder and dried at room temperature.

Preparation of catalyst layers and MEAs
Naon N115 membranes (FuelCellsEtc) were used as a fuel cell electrolyte. Prior to the usage, 3 Â 4 cm Naon sheets were puried by boiling in deionized water for one hour and then rinsing in 3.75% hydrogen peroxide for an hour. Thereaer, the membranes were activated in 0.2 M H 2 SO 4 solution at boiling temperature also for one hour. Finally, they were rinsed several times in deionized water to remove any residual contaminants. To make Membrane Electrode Assemblies (MEAs), commercial gas diffusion electrodes (MLGDE 2.0 mg.cm À2 , FuelCellsEtc) were selected as a cathode and Gas Diffusion Electrodes (GDEs) were prepared as an anode. Both electrodes were cut to maintain the equal area of 1 cm 2 . Catalyst mixtures intended for anodes were fabricated by mixing of particular catalyst components in the way to obtain the uniform distribution of the platinum particles and two types of modiers (TiO 2 and TiO 2 -SiO 2 -VTMS). This was done according to the following procedure. Highly conductive Vulcan XC72R carbon black (0.33 g) was used as anode supporting material to prepare mixtures in individual 25 cm 3 beakers. Pt black powder (0.140 g) was rst mixed with a small amount of deionized water (to avoid catalyst burning) and then carbon black was added to obtain a 30% Pt/C mass fraction. Next, ethanol (96% vol.) was used in the amount of ca. 15% volume ratio. The mixtures were prepared by applying the ultrasonication stirrer for about 25 minutes. At this stage two modiers i.e. TiO 2 and TiO 2 -SiO 2 -VTMS were added, both in the form of powder. One catalyst mixture was also prepared as a reference unit. To investigate the contribution of TiO 2 and TiO 2 -SiO 2 -VTMS to the anode behavior, three different amounts of these modiers were investigated, i.e. 0.05 g, 0.1 g and 0.2 g. The given amounts were found considering the maximum concentration of a non-conductive additive in electrode structure to keep its good electron conductivity. Next, catalyst mixtures were again stirred and mixed for about 10 minutes. Finally, 20 wt% Naon solution (Sigma-Aldrich) was added into each beaker in the amount of 1.16 g followed by the refrigeration of mixtures to the temperature of about 0 C.
GDL papers with hydrophobic treatment (H2315 C4 Freudenberg Fuel Cell Products) were cut out (4 cm 2 ), weighed and used to made gas diffusion electrodes (1 cm 2 ) by casting of the prepared catalyst inks (tape casting). Obtained GDEs were dried at a temperature of 25 C for one hour and then for 15 minutes at 135 C and nally weighed again. The mass difference between GDL and GDE was used to calculate the amount of Pt, TiO 2 and TiO 2 -SiO 2 -VTMS incorporated in particular GDEs. Details with regard to constructed electrodes are given in Table 1.
Each fuel cell (MEA) was constructed by pressing of activated membranes and electrodes at a temperature of 110 C for 5 minutes under 30 bar of pressure. Respectively, six modied fuel cells were obtained: FC_T50, FC_T100, FC_T200, FC_TS50, FC_TS100, FC_TS200 and one reference FC_ref. Fig. 1 presents the architectures of the constructed PEMFC as well as a photograph of a typical MEA.

Material characterization
Surface area and porosity analyzer Micromeritics ASAP2020 was used to calculate the physical parameters of prepared catalyst mixtures. The analysis was also performed for pristine components (TiO 2 , TiO 2 -SiO 2 -VTMS and Vulcan carbon black) to evaluate their impact on the anode behavior. Adsorption/ desorption isotherms were determined at temperature of liquid nitrogen, from high vacuum pressure (10 À6 Torr) to the relative pressure of 0.99 p/p 0 , where p 0 is the saturation pressure of the adsorbate. Prior to the measurements, catalyst mixtures were degassed at 150 C for 8 hours (maximum limit due to Naon used), whereas particular components at elevated temperature of 350 C for 8 hours to perform deep purication process (except organically modied TiO 2 -SiO 2 -VTMS, 150 C). To calculate surface area for particular samples, Brunauer-Emmett-Teller (BET) theory was applied while in case of average pore diameter and volume as well as pore size distribution, Barrett-Joyner-Halenda (BJH) method was used. Based on obtained results, additional analysis of micropore structure was carried out for selected materials using Horvath-Kawazoe and tplot methods. Gas diffusion electrodes were analyzed by scanning electron microscope (SEM) (Tescan VEGA/SBH) in order to investigate the effect of applied additives on surface morphology.
Polymer fuel cells (MEAs) were electrochemically investigated using 1 cm 2 testing fuel cell xture (Pragma Industries) with graphite single-serpentine monopolar plates. The research was carried out in situ, delivering to the measurement cell reactant or inert gases, depending on the method used. The samples were measured by cyclic voltammetry (CV), polarization curve (J-V) and electrochemical impedance spectroscopy (EIS). The equipment used in these experiments was: SI 1287 Electrochemical interface potentiostat/galvanostat (Solartron Analytical), ZS electronic load (H&H), Impedance/Gain-Phase analyzer SI 1260 (Solartron Analytical) and dedicated control programs.

Physical characterisation of materials
Porosity and the surface area of fuel cell electrodes have signicant inuence on the rate of both electrochemical reactions. High porosity and large surface area accelerates hydrogen oxidation and oxygen reduction reactions in terms of improved reactants delivery, products removal, heat energy and electrical charges, and transfers of ions. The improvement of reaction rate may happen due to the increase of active catalytic sites via larger exposure of catalyst to the reactants in a threephase boundary region. The pores of average diameter less than 2 nm (micropores) promote higher rate of electrochemical reaction, while pores larger than 50 (macropores) nm facilitate charges, reactants and products transfer. Therefore, it is intentional to maintain both structures of pores in the catalyst area and to limit the size of mesopores (2-50 nm) which are mainly responsible for negative effect of additional mass transfer resistance and reduction of electrode volumetric current. 24 Fuel cell performance strongly depends on the kinetics of electrochemical reactions, thus the modication of catalyst support is one of the methods to improve it.
According to the graphs depicted in Fig. 2a and 3a particular electrode components that were used for catalyst mixture preparation are characterized by high surface area as well as high porosity. Fig. 2a shows the shape of nitrogen adsorption/desorption isotherms for three various electrode components i.e. Vulcan carbon black, titanium dioxide and TiO 2 -SiO 2 -VTMS. Some characteristic properties can be retrieved by analyzing their low-pressure adsorption behaviour (p/p 0 less than 0.01 where microporosity is revealed). Carbon black Vulcan XC72R isotherm is assigned to type I and II according to IUPAC classication. 25 It suggests that this material has a fragmentary microporosity which is highly benecial for fuel cell operation. TiO 2 -SiO 2 -VTMS behaviour is relevant at initial adsorption stage (also type I of isotherm) however, its isotherm reveals even better properties because the increase of nitrogen adsorption at low pressure is prolonged in comparison to Vulcan giving higher surface area and bigger volume of micropores (Fig. 3a). Moreover, it contains characteristic hysteresis which is concurrent with type IV of IUPAC classication. In this background, the third component TiO 2 does not have unique structure of micropores, since the relative pressure immediately increases once adsorption begins which corresponds to a very low volume of micropores. Nevertheless, its isotherm is also characterized by a hysteresis.   Fig. 2b and 3b demonstrate result analysis of catalyst layers. TiO 2 and reference catalyst layer isotherms are equivalent. The addition of TiO 2 does not cause any improvement or deterioration of sorption phenomenon. Both isotherms are assigned to type II that describes materials composed of mainly macroporous structure. In the consequence of catalyst mixture preparation, micropores of Vulcan carbon black were closed and lled mostly by Naon. Therefore at a very low pressure no signicant adsorption could be observed. TiO 2 -SiO 2 -VTMS acts better than carbon black because some of its microporous structure is conserved once catalyst mixture is fabricated. This can be clearly seen from its isotherm - Fig. 2b. Table 2 describes the results of BET and BJH calculations obtained for particular catalyst layers and their selected components. BET analysis was performed taking the isotherm data in the range from 0.05 to 0.2 p/p 0 whereas BJH in the entire range of pressures excluding microporosity region (pressure less than 0.01 p/p 0 ). Carbon black is characterized by surface area of 249.2 m 2 g À1 and average pore width 10.1 nm according to the analysis. Preparation of catalyst mixture leads to a deterioration of physical parameters. Surface area of reference catalyst layer (CL_Ref), prepared on the basis of carbon black, is signicantly reduced (25.9 m 2 g À1 ) and average pore size is increased (25.3 nm). Simultaneously, cumulative volume of pores is decreased more than twice from 0.413 to 0.177 m 3 g À1 . TiO 2 -SiO 2 -VTMS with its superior physical properties (321.9 m 2 g À1 , 3.7 nm) applied with the concentration of just 12.5 wt% (Table 1) is capable to mitigate a negative effect of catalyst layer formation. Its catalyst mixture (CL_TS100) is characterized by 74.6 m 2 g À1 and 15.7 nm of surface area and average pore width, respectively. Cumulative volume of pores for CL_TS100 is relatively low (0.146 m 3 g À1 ), however the given value was calculated based on BJH theory that doesn't take into account micropores nor macropores of large size (the calculation of cumulative surface area and volume of pores was within the range of 1.7-300 nm). Upper result suggests that there is a large fraction of micropores implemented by TiO 2 -SiO 2 -VTMS to the catalyst layer. In contrast to CL_TS100, TiO 2 -modied catalyst layer (CL_T100) has low specic surface area (27.8 m 2 g À1 ) and large average pore size (24.0 nm)which are both comparable to the reference material. Such values result mainly from poor surface area of TiO 2 (59.5 m 2 g À1 ), which does not contribute to the improvement, however the inuence on catalyst layer is slightly positive as its average pore width (6.2 nm) is relatively small giving additional space for adsorbate.
Good results for CL_TS100 investigation were conrmed by the additional analysis with respect to the fraction of micropores. Horvath-Kawazoe (H-K) and t-plot methods were used to  estimate the volume, size and surface area of micropores for TiO 2 -SiO 2 -VTMS, Vulcan carbon black as well as one catalyst mixture. Micropores are visible at a very low pressure of isotherms for only three materials (see Fig. 2). For H-K theory the calculation was done assuming a slit geometry of micropores and taking so called interaction parameter that is 3.73 r 10 À43 erg cm 4 . For calculation of t-plot, in case of Vulcan carbon black, the thickness curve of Carbon Black STSA was used while for TiO 2 -SiO 2 -VTMS and its catalyst layer Harkins and Jura thickness curve was applied. Table 3 contains details of microporosity structure for given components.
Both components and the catalyst layer have comparable average pore width (around 6 nm), however according to the tplot method, surface area of CL_TS100 is signicantly lower than for carbon black as well as TiO 2 -SiO 2 -VTMS. Fabrication of catalyst layer leads to the decrease of specic surface area of pores which suggests that Naon electrolyte was partially adsorbed within the range of micropores. Such a partial introduction of this polymer into the micropore structure is possible as, according to the cluster-network model, Naon consists of an equal distribution of sulfonate ion clusters with a 40Å diameter held within a continuous uorocarbon lattice. 26 This effect is also conrmed by Horvath-Kawazoe method giving lower micropore volume (0.016 cm 3 g À1 ) for CL_TS100 in comparison to its pristine components (0.076 and 0.086 cm 3 g À1 , respectively). Micropore size distribution for Vulcan carbon black, TiO 2 -SiO 2 -VTMS and their catalyst layer is shown in Fig. 4. Derivative of pore volume (Akima Spline tting) versus pore width reveals predicted micropore distribution at a given geometry of pores (slit), in the range from ca. 0.4 to 1.0 nm. Such conguration is advantageous for hydrogen oxidation reaction due to better exposure of catalyst site and improved proton transfer because of hydrophilic properties of SiO 2 .   compounds occur in the form of round-shape particles that undoubtedly facilitates the fabrication process of catalyst layers.

Electrochemical characterization
3.2.1. The performance of fuel cells. Improved rate of anodic reaction has a positive inuence on the entire fuel cell performance assuming the kinetics at cathode side is elevated by much higher utilization of the catalyst. In the given experiment around 6-times higher platinum amount was used at cathode site (2.0 mg cm À2 ) than at anode (Table 1) in order to see the impact of modiers on fuel cell operation. However, according to ref. 3 and 4 cathodic losses are still dominant due to extremely fast rate of HOR reaction in comparison to ORR. Nevertheless, performing direct voltage-current characteristics, the effect of anode modication was investigated with regard to reference sample which contains platinum-supported carbon electrode with Pt surface density of 0.54 mg cm À2 . In Fig. 6 polarization curves are depicted as the electrical potential and power density functions of current density for reference fuel cell (FC_Ref) and fuel cells with TiO 2 -SiO 2 -VTMSand TiO 2 -based anodes (FC_TS100 and FC_T100, respectively). Table 4 contains values of parameters that were calculated based on three different methods of in situ evaluation for fuel cells i.e. polarization curves, electrochemical impedance spectroscopy, cyclic voltammetry. Samples with TiO 2 -SiO 2 -VTMSand TiO 2 were additionally analyzed with regard to different concentrations of modiers i.e. 50% and 200% of the amount applied for FC_TS100 and FC_T100 (marked as FC_T50 and FC_T200 for TiO 2 , and FC_TS50 and FC_TS200 for TiO 2 -SiO 2 -VTMS). The polarization curves that compare the inuence of various modiers concentrations are shown is Fig. 7. Each V-I characteristic was carried out maintaining the following conguration: hydrogen/air ow rate: 60/80 ml min À1 , fuel cell xture temperature 60 C, cathode output relative humidity $90%, no dead-end nor back pressure. Both cathode and anode were humidied.
The results of experiment suggest that the amount of catalyst 0.34-0.54 mg cm À2 at anode and 2.0 mg cm À2 at cathode is sufficient at given conditions to investigate the anode behaviour of entire fuel cell despite sluggish kinetics of oxygen reduction reaction. The improvement of properties can be observed based on polarization curves carried out for particular fuel cells. Reference fuel cell (FC_Ref) is characterized by maximum power density of 126 mW cm À2 and current density (at 0.25 V) of 325 mA cm À2 . Modication of anode catalyst support by addition of 12.4 wt% titanium dioxide (FC_T100) resulted in much higher power and current density (217 mW cm À2 and 483 mA cm À2 , respectively). In contrast, aer the addition of two times lower amount of TiO 2 (6.6 wt%, FC_T50) to the anode the changes of  reaction kinetics were not observed (115 mW cm À2 , 295 mA cm À2 ), whereas double increase of modier concentration (FC_T200) gave signicant gain of performance (198 mW cm À2 , 429 mA cm À2 ). FC_T200 behaviour is slightly worse than FC_T100 which shows that higher concentration of TiO 2 -SiO 2 -VTMS may not be advantageous perhaps due to poor electrical properties of SiO 2 . The efficiency at maximum power, calculated taking the higher heating value of hydrogen, shows that there are no signicant differences of voltage losses between samples. The obtained values vary from 34.1 to 37.6%. Therefore, the improvement of fuel cells operation cannot be explained by the phenomena of activation losses nor ohmic losses decrease (due to ionic resistance of the membrane). There are signicant differences in polarization concentrations, however these losses are mainly affecting the values of potential at high current density region which is beyond the nominal operating voltage range for fuel cells. The comparison of particular values of voltage losses were collected in Table 5. Activation polarization losses were calculated from the given equation a atransfer coefficient, F -Faraday's constant, icurrent density, i 0exchange current density. Both parameters a a , i 0 were taken based on the plot of Tafel equation. Ohmic losses are based on EIS experiment, where the values of resistance were obtained at high frequency region of Nyquist plot (Fig. 8).
Concentration losses were found according to the following  a OCVopen circuit voltage. b R electrodetotal transfer resistance of mass and charge through the electrode. c R electrolyteionic resistance of the electrolyte. d ECSAelectrochemical surface area. Fig. 7 The differences in performance for given fuel cells having various concentrations of TiO 2 -SiO 2 -VTMS and TiO 2 applied as anode modifiers. According to the data in Table 5  To investigate the mass and charge transfer resistance for particular samples the electrochemical impedance spectroscopy (EIS) was applied. Fuel cells were supplied with hydrogen and air (60/80 sccm) while the cells were at a temperature of 25 C and relative humidity of about 90%. The electrical load was not applied. All characteristics were measured in the frequency range from 40 kHz to 0.1 Hz with signal amplitude of 250 mV. Nyquist and Bode plots of EIS experiment for investigated fuel cells are shown in Fig. 8.
The resistance at high frequency region of Nyquist plot is a material property of the electrolyte (the sum of ionic resistance of membrane and Naon electrolyte layers). All samples were manufactured according to the same method thus the values of the resistances are comparable (Table 4) and vary from 0.23 to 0.36 U cm 2 . Small uctuation is mainly due to variations of electrolyte concentration at the anode side having xed value of 33 wt% Naon/Pt-carbon black regardless of the amount of modier used. The resistance of mass and charge transfer which corresponds to the width of semicircle in Nyquist plot varies depending on fuel cells. According to Table 4  respectively). This indicates that the low concentration of modier doesn't contribute to the electrode transport phenomena or this contribution is negligible. However, noticeable inuence is visible in case of the rest of the samples. The lowest values of resistance were obtained for FC_T100 (2.92 U cm 2 ) while slightly higher in case of FC_T200 (3.20 U cm 2 ). It is an evidence that titanium dioxide applied to the anode as a catalyst support facilitates the charge transport within the three phase boundary region or even increases the catalyst activity. TiO 2 is a well-known semiconductor and photo-catalyst with good optical properties, capable of transferring protons across its crystal lattice. [29][30][31] Relevant results were found in case of samples modied by TiO 2 -SiO 2 -VTMS, however the process that lead to the resistance mitigation is more complex. FC_TS100 and FC_TS200 with double concentration of modifying agent are characterized by similar transport resistance of 3.16 and 3.08 U cm 2 . In addition to the inuence of TiO 2 , the observed improvement results also from the presence of  silicon dioxide which provides the hydrophilic properties for the fuel cell electrode. The proper humidication of the electrode (especially anode) is of great importance in terms of increasing the power density. However, the usage of SiO 2 is reasonable assuming its low concentration due to poor electrical conductivity. Vinyltrimethoxysilane acts as a coupling agent in TiO 2 -SiO 2 -VTMS. The compound is bifunctional, featuring both a vinyl group and hydrolytically sensitive ethoxysilyl groups. As such it also contributes to the hydrogen oxidation reaction rate via incorporation of additional water molecules (increased water uptake) and hydrolysis of TiO 2 and SiO 2 that inuences the proton conduction mechanism within the electrolyte.
3.2.3. Active surface area of catalyst mixture. Surface area of platinum catalyst that takes part in a hydrogen oxidation reaction is limited due to the electrode composition, various impurities or even the excess of water present within a catalyst region. The performance of fuel cells depends on this physical magnitude which is reected by a parameter named as electrochemical surface area (ESCA). The higher the value of ECSA the better due to bigger utilization of a catalyst. To obtain the electrochemical surface area, cyclic voltammetry (CV) was used with the aid of a potentiostat. In this method particular fuel cells were scanned with the rate of 10 mV s À1 from 0 to 1 V, repeatedly and electric current response was recorded and plotted in the form of voltammograms ( Fig. 9 and 10). One electrode (HE) was fed with hydrogen (50 ml min À1 ) working as counter and reference electrode while the other was supplied with nitrogen (50 ml min À1 ) acting as working electrode.
Electrochemical surface area of the catalyst was calculated based on the following equation: ECSA ¼ chargeðmC cm À2 Þ 210ðmC cm À2 Þcatalyst_loadingðg cm À2 Þ ; where the amount of charge was obtained integrating the area underneath desorption peaks in the range from 100 to 400 mV. As it is visible in Table 4, ECSA calculated for most of the samples is comparable (approx. 25 m 2 g À1 ) except FC_T100 and FC_TS100 fuel cells (63.3 and 46.5 m 2 g À1 , respectively). Slightly elevated value of ECSA was achieved for FC_TS200, i.e. 31.7 m 2 g À1 . In case of TiO 2 -SiO 2 -VTMS-modied fuel cells the increase of electrochemical surface area might be explained based on conclusions taken from physical surface area and porosity analysis. TiO 2 -SiO 2 -VTMS scaled up the area and increased microporous fraction within which platinum particles were deposited. Higher three-phase boundary region contributes to adsorption and desorption processes of hydrogen which is also observed analyzing the height of CV peaks (reecting the catalyst activity). Additionally, water distribution within the electrolyte of anode layer was increased as a consequence of SiO 2 addition.
In case of FC_T100 the improvement of electrochemical surface area is more signicant and cannot be explained by the change of anode physical structuresurface area and porosity were just slightly increased. It seems that TiO 2 may participate in a hydrogen oxidation reaction via the contribution to the proton conduction mechanism and the increase of catalyst activity in terms of improvement of reaction rate. However, such a statement cannot be proven based on experiments conducted in this paper.
3.2.4. Protability estimation of new anode vs. Pt anode for PEMFCs. Despite recent successive development of various platinum-group metals (PGM)-free catalysts, Pt-based materials are still the most popular and commonly-used catalysts for fuel cell applications. There is observed continuous drop of required amount of platinum used in electrode layers. For instance, U.S. Department of Energy (DoE) set the goal of 0.125 g kW À1 of PGM catalyst by the year 2020 for low duty vehicles. It corresponds just to 11.25 g of platinum per midsize fuel cell electric vehicle (90 kW gross ). 32 The given value is comparable with the amount of PGM materials used for internal combustion engine vehicles however further reduction is needed for successful competition. Catalyst amount drop down to 0.088 g kW À1 is assumed by 2025 according to DoE. There are various methods to achieve such a target. The approach presented in this paper could be one of them. At given conditions our method allowed to decrease the amount of catalyst by circa 30%. For the 2020 DoE target it is relevant to reduction of Pt by 37.5 mg per 1 kW or 3.4 g per 90 kW vehicle. Replacement of carbon-based support materials alongside the improvement of electrode ionomer materials seem to be necessary for development of competitive fuel cell electric vehicles.

Conclusions
Two chemical compounds were used for modication of anode support material for PEM fuel cells, i.e. TiO 2 in an anatase phase and the hybrid material (TiO 2 -SiO 2 -VTMS) composed of titanium dioxide and silicon dioxide with vinyltrimethoxysilane acting as a crosslinking agent. TiO 2 -SiO 2 -VTMS is a good candidate for the modication of support material rst of all due to its outstanding physical properties of high surface area and microporosity, and also because of its improved chemical durability and hydrophilicity, capable to stabilize catalyst nanoparticles and increase the rate of hydrogen oxidation reaction. TiO 2 -SiO 2 -VTMS applied in the amount of just 12.5 wt% leads to a threefold increase of catalyst layer surface area and it decreases the average pore size by 10 nm. In the consequence of modication, the electrochemical properties of fuel cell are improved, including enhancement of its electrochemical surface area, charge and mass transfer conductivity and the power density. It was conrmed that pristine titanium dioxide can also be successfully utilized for fuel cell electrodes acting as a catalyst support. According to the results of this work, TiO 2 added to anode of PEMFC has an inuence on the electrochemical reaction that results in the improvement of the overall fuel cell performance. However, its poor surface area and limited porosity justify the usage of SiO 2 in order to mitigate the typical catalyst degradation problems related for instance to platinum dissolution, migration and agglomeration phenomena. TiO 2 -SiO 2 -VTMS could be usable for PGM-free catalysts to solve their various problems.

Conflicts of interest
There are no conicts to declare.