Microfabricated polymer electrolyte membrane fuel cells with low catalyst loadings

Miniaturized fuel cells as compact power sources fabricated in Pyrex glass using standard polymer electrolyte membrane (PEM) and electrode materials are presented. Photolithographic patterned and wet chemically etched serpentine flow channels of 1 mm in width and 250  m in depth transport the fuels to the cell of 1.44 cm² active electrode area. Feeding H₂/O₂ a maximum power density of 149 mW cm⁻² is attained at a very low Pt loading of 0.054 mg cm⁻², ambient pressure, and room temperature. Operated with methanol and oxygen about 9 mW cm⁻² are achieved at ambient pressure, 60 °C, and 1 mg cm⁻² PtRu/Pt (anode/cathode) loading. A planar two-cell stack to demonstrate and investigate the assembly of a fuel cell system on Pyrex wafers has successfully been fabricated.     

Direct oxidation of sodium borohydride on Pt, Ag and alloyed Pt–Ag electrodes in basic media: Part II. Carbon-supported nanoparticles

We studied the borohydride oxidation reaction (BOR) by voltammetry in 0.1 M NaOH/10⁻³ M BH₄⁻ on carbon-supported Pt, Ag and alloyed PtAg nanoparticles (here-after denoted as Pt/C, Ag/C and Pt–Ag/C). In order to compare the different electrocatalysts, we measured the BOR kinetic parameters and the number of electrons exchanged per BH₄⁻ anion (faradaic efficiency). The BOR kinetics is much faster for Pt/C than for Ag/C (i_Pt=0.15, i_Ag=3.1×10⁻⁴ A cm⁻² at E=−0.65 V vs. NHE at 25 °C), but both materials present similar Tafel slope values. The n value involved in the BOR depends on the thickness of the active layer of electrocatalysts. For a “thick layer” (approximately 3 m), n is nearly 8 on Pt/C and 4 on Ag/C, whereas n decreases for thinner Pt/C active layers (n2 for thickness <1 m). These results are in favour of the sequential BH₄⁻ hydrolysis (yielding H₂) followed by hydrogen oxidation reaction (HOR), or direct sequential BOR on Pt/C, whereas Ag/C promotes direct but incomplete BOR (Ag has no activity regarding hydrogen evolution reaction, HER). The n value close to 8 for the thick Pt/C layer displays the sufficient residence time of the molecules formed (H₂ by heterogeneous hydrolysis or BOR intermediates) within the active layer, which favours the complete HOR and/or BOR. Two PtAg/C nanoparticles alloys have been tested (noted APVES-4C and APVES-E1). They show different behavior; the borohydride oxidation reaction kinetics is faster on APVES-E1 than on APVES-4C (b=0.15, and b=0.31 V dec⁻¹,  A cm⁻², respectively, at 25 °C), but the n values are higher on APVES-4C than APVES-E1 (nearly 8 vs. 3, respectively, at 25 °C). These discrepancies probably originate from the heterogeneity of such bimetallic materials, as observed from physicochemical characterizations.     

Li-ion diffusion kinetics in LiMn2O4 thin films prepared by pulsed laser deposition

LiMn₂O₄ thin films were deposited on Au substrates by pulsed laser deposition (PLD). The Li-ion chemical diffusion coefficients of the films, , were measured by cyclic voltammetry (CV), galvanostatic intermittent titration technique (GITT), potentiostatic intermittent titration technique (PITT), and electrochemical impedance spectroscopy (EIS). It was found that the values by CV and PITT were in the order of 10⁻¹³ cm² s⁻¹, and those by EIS and GITT were in the range of 10⁻¹³ to 10⁻¹¹ and 10⁻¹⁴ to 10⁻¹¹ cm² s⁻¹, respectively. These data were compared with the previously reported values.     

Preparation of low-platinum-loading electrocatalysts using electroless deposition method for proton exchange membrane fuel cell systems

One promising preparative method that offers the potential for improved platinum (Pt) dispersion of electrocatalysts is electroless deposition (ED). In this study, the effects of multiwalled carbon nanotubes (MWCNTs) pretreatment and synthesis procedure on properties of the four catalysts, synthesized by ED method, have been considered. The results of energy-dispersive X-ray spectroscopy (EDS), X-ray dot-mapping, X-ray fluorescence (XRF) and cyclic voltammetry (CV) analyses showed that using palladium (Pd) precursor during two-step sensitization-activation coating procedure gives uniform Pt particles distribution on MWCNTs with low aggregation and high specific surface area (∼80 m² g⁻¹). In addition, to investigate the performance of the synthesized catalysts in experimental fuel cell system, thin-film method was used to fabricate the membrane electrode assemblies (MEAs). Obtaining the polarization curves for the fabricated MEAs (Pt loading ∼0.4 mg cm⁻²) and a commercial MEA (ElectroChem, Pt loading ∼1 mg cm⁻²) demonstrated that the catalyst prepared by two-step sensitization-activation coating procedure possesses a good performance despite of its lower Pt content.     

Li-ion diffusion kinetics in LiFePO4 thin film prepared by radio frequency magnetron sputtering

LiFePO₄ thin films were prepared by radio frequency (RF) magnetron sputtering and were characterized by X-ray diffraction (XRD), scanning electron microscope (SEM) and atomic force microscope (AFM). Li-ion chemical diffusion coefficients, , were measured by potentiostatic intermittent titration technique (PITT), electrochemical impedance spectroscopy (EIS), and cyclic voltammetry (CV). The effects of Ag content, film thickness, and film orientation on the electrochemical performance and Li-ion chemical diffusion coefficients of the LiFePO₄ thin films were investigated. values were measured using the liquid electrolyte and the solid electrolyte, and the obtained values were discussed. The values by PITT and EIS were in the range of 10⁻¹⁴ to 10⁻¹² and 10⁻¹⁵ to 10⁻¹² cm² s⁻¹, respectively and that by CV was in the order of 10⁻¹⁴ cm² s⁻¹.     

Investigation on diffusion behavior of Li in LiFePO4 by capacity intermittent titration technique (CITT)

Capacity intermittent titration technique (CITT) was used to investigate the chemical diffusion coefficient () of lithium-ion in LiFePO₄ cathode material. The values of at the galvano-charge current of 0.2 and 0.4 mA were respectively found to range from 8.8 × 10⁻¹⁶ to 8.9 × 10⁻¹⁴ cm² s⁻¹ and from 1.2 × 10⁻¹⁶ to 8.5 × 10⁻¹⁴ cm² s⁻¹ in the voltage range from 3.2 to 4 V (vs. Li⁺/Li). The transfer coefficients of cathode (0.32-0.42) and anodic (0.26-0.3), and the standard rate constant (1.58 × 10⁻⁹ to 1.30 × 10⁻⁸ cm s⁻¹) were measured from the Tafel plots of LiFePO₄ in the equilibrium potential range from 3.06 to 3.45 V. From these kinetic parameters, the finite kinetics at interface was taken into account to revise the above values of . The revised values of at the galvano-charge current of 0.2 and 0.4 mA were respectively found to range from 2.44 × 10⁻¹⁵ to 2.21 × 10⁻¹³ cm² s⁻¹ and from 5.81 × 10⁻¹⁶ to 3.22 × 10⁻¹³ cm² s⁻¹ in the voltage range from 3.2 to 4 V. Results show that the approximation of infinite charge-transfer kinetics leads to a spurious value of which is lower than the revised value, and the spurious extent depends on the galvano-charge current of CITT experiment.     

Enhanced performance and improved interfacial properties of polymer electrolyte membrane fuel cells fabricated using sputter-deposited Pt thin layers

For this study, catalyst layers for polymer electrolyte membrane fuel cells (PEMFC) were prepared by spraying and sputtering to deposit Pt amount of 0.1 and 0.01 mg cm⁻², respectively. These Pt layers were then assembled to fabricate membrane electrode assemblies (MEA) having either single- or double-layered catalysts. The PEM fuel cell with double layers showed a current density of 777 mA cm⁻² at a cell voltage of 0.6 V, which is a higher current density than state-of-the-art fuel cells at 643 mA cm⁻². These results indicate that Pt loading in state-of-the-art PEMFCs could be reduced by approximately 50% with no performance loss by using both spraying and sputtering method in the MEA fabrication process.     

Analysis of the chemical diffusion coefficient of lithium ions in Li3V2(PO4)3 cathode material

The chemical diffusion coefficients of lithium ions (DLi⁺) in Li₃V₂(PO₄)₃ between 3.0 and 4.8 V are systematically determined by cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS) and galvanostatic intermittent titration technique (GITT). The DLi⁺ values are found to be dependent on the voltage state of charge and discharge. Based on the results from all the three techniques, the true diffusion coefficients measured in single-phase region are in the range of 10⁻⁹ to 10⁻¹⁰ cm² s⁻¹. Its apparent diffusion coefficients measured in two-phase regions by CV and GITT range from 10⁻¹⁰ to 10⁻¹¹ cm² s⁻¹ and 10⁻⁸ to 10⁻¹³ cm² s⁻¹, respectively, depending on the potentials. By the GITT, the DLi⁺ varies non-linearly in a “W” shape with the charge-discharge voltage, which is ascribed to the strong interactions of Li⁺ with surrounding ions. Finally, the chemical diffusion coefficients of lithium ions measured by CV, EIS and GITT are compared to each other.     

In situ FTIR spectroscopic studies of (bi)sulfate adsorption on electrodes of Pt nanoparticles supported on different substrates

Nanostructured Pt electrodes were prepared by electrodeposition of Pt nanoparticles on different substrates (GC, Pt and Au) under cyclic voltammetric conditions and with various number (n) of potential cycling, and were denoted as nm-Pt/S(n) (S = GC, Pt and Au). Adsorption of (bi)sulfate on the nm-Pt/S(n) was studied by in situ FTIR reflection spectroscopy. It has been revealed that the nanostructured Pt electrodes exhibit anomalous IR properties for (bi)sulfate adsorption regardless of the different reflectivity of substrate, i.e. the IR absorption of (bi)sulfate species adsorbed on all the nm-Pt/S(n) electrodes is significantly enhanced and the IR band direction is completely inverted in comparison with the same species adsorbed on a bulk Pt electrode. The two IR bands around 1200 and 1110 cm⁻¹ attributed to adsorbed (bi)sulfate species are shifted linearly with increasing electrode potential, yielding Stark tuning rates () of 152.1 and 21.1 cm⁻¹ V⁻¹ on nm-Pt/GC(20), respectively. Along with increasing n, the Stark tuning rate of the IR band around 1200 cm⁻¹ decreases quickly and declined to 7.6 cm⁻¹ V⁻¹ on nm-Pt/GC(80), while the Stark tuning rate of the IR band near 1100 cm⁻¹ is fluctuated between 23.0 and 16.2 cm⁻¹ V⁻¹. It has determined that the enhancement of IR absorption of (bi)sulfate adsorbed on nanostructured Pt electrode is varied with substrate material and n, and a maximal 16-fold enhancement of the IR band near 1200 cm⁻¹ has been measured on the nm-Pt/GC(30) electrode. The in situ FTIR studies illustrated that the adsorption of (bi)sulfate occurs mainly in the double layer potential region, and reaches a maximum around 0.80 V. The results demonstrated also that the competitive adsorption of CO and oxygen species can inhibit completely (bi)sulfate adsorption, which has evidenced a weak interaction of (bi)sulfate with nm-Pt/S(n) electrode surface.     

Ionophores in polymeric membranes for selective ion recognition; impedance studies

Synthetic calix[4]arene-crown ionophores for selective Na⁺ (ionophore L1) and Cs⁺-ions (ionophore L2) recognition find application in ion-selective membrane electrodes (ISE) for analytical purpose. Selectivity coefficients for the electrodes with compounds L1 and L2 are  = −2.6 and  = −2.4, respectively. Electrodes of two different construction: all-solid-state (ASS) (with conducting polymer layer on glassy carbon or platinum as ion-to-electron transducer) and conventional ion-selective electrode (ISE) (with liquid electrolyte and Ag/AgCl) are presented and their properties and lifetime are being compared. Resistance of PVC membrane with ionophores L1 and L2 were within the range 0.15-1.4 MΩ depending on the type of the outer electrolyte and its concentration. Conductivity of the membranes was in the range 0.7 × 10⁻⁸ to 6 × 10⁻⁸ Ω⁻¹ cm⁻¹. Warburg coefficients σ were within 0.16 × 10⁴ to 12.7 × 10⁴ Ω s^−1/2, dielectric constant values ? were in a range 28-60 depending mainly on the type of plasticizer.     

Electrochemical behaviour of americium ions in LiCl-KCl eutectic melt

This work presents a study on the electrochemical properties of AmCl₃ in a molten LiCl-KCl eutectic, at a temperature range of 733-833 K. Transient electrochemical techniques, such as cyclic voltammetry and chronopotentiometry, on inert metallic tungsten working electrode have been used to investigate the reduction mechanism of Am³⁺ ions. The results show that Am³⁺ is reduced to Am metal by a two-step mechanism corresponding to the Am³⁺/Am²⁺ and Am²⁺/Am⁰ transitions. Formal standard potentials of Am³⁺/Am²⁺ ( versus Cl₂/Cl⁻ at 733 K) and Am²⁺/Am⁰ ( versus Cl₂/Cl⁻ at 733 K) redox couples as well as diffusion coefficients of Am³⁺ and Am²⁺ (2.4 × 10⁻⁵ and 1.15 × 10⁻⁵ cm² s⁻¹ at 733 K, respectively) have been calculated at three different temperatures. In the studied range of temperature, the DAm3+/DAm2+ ratio was found to be around 2. In addition, thermodynamic properties have been calculated for Am³⁺ () and Am²⁺ () and compared to thermodynamic reference data in order to estimate activity coefficients (Am³⁺ = 4.7 × 10⁻³ and Am²⁺ = 2.7 × 10⁻² at 733 K) in the molten LiCl-KCl eutectic.     

Anomalous IR optical properties of aggregates of Pd nanoparticles induced through electrochemical cyclic voltammetry

IR optical properties of Pd nanoparticles with different size and aggregation state were studied in the current paper. The dispersed Pd nanoparticles () stabilized with poly(N-vinylpyrrolidone) (PVP) were synthesized by the seeding growth method, in which the seeds were formed step by step through reducing H₂PdCl₄ with ethanol. The dispersed Pd nanoparticles of much large size () were grown from the by keeping the colloid of undisturbed for 150 days at room temperature around 20 °C. The aggregates of () were prepared through an agglomeration process induced during a potential cyclic scanning between −0.25 V and 1.25 V for 20 min at a scan rate of 50 mV s⁻¹. Scanning electron microscope (SEM) patterns confirmed such aggregation of . Fourier transform infrared (FTIR) spectroscopy together with CO adsorption as probe reaction was employed in studies of IR optical properties of the prepared Pd nanoparticles. The results demonstrated that CO adsorbed on films substrated on CaF₂ IR window or glassy carbon (GC) electrode yielded two strong IR absorption bands around 1970 cm⁻¹ and 1910 cm⁻¹, which were assigned to IR absorption of CO bonded on asymmetric and symmetric bridge sites, respectively. Similar IR bands were observed in spectra of CO adsorbed on films, except the IR bands were much weak, whereas CO adsorbed on film produced an IR absorption band near 1906 cm⁻¹, and an anomalous IR absorption band whose direction has been completely inverted around 1956 cm⁻¹. The direction inversion of the IR band of CO bonded to asymmetric bridge sites on was ascribed to the interaction between Pd nanoparticles inside the aggregates. Based on FTIR spectroscopic and cyclic voltammetric results, the aggregation mechanism of Pd nanoparticles from to has been suggested that the agglomeration of Pd nanoparticles was driven by the alteration of electric field across electrode-electrolyte interface, when the PVP stabilizer was stripped via oxidation during cyclic voltammetry.     

Electroreduction of peroxycitric acid coexisting with hydrogen peroxide in aqueous solution

The electrochemical reduction of peroxycitric acid (PCA) coexisting with citric acid and hydrogen peroxide (H₂O₂) in the equilibrium mixture was extensively studied at a gold electrode in acetate buffer solutions containing 0.1 M Na₂SO₄ (pH 2.0-6.0) using cyclic and hydrodynamic voltammetric, and hydrodynamic chronocoulometric measurements. The reduction of PCA was characterized to be an irreversible, diffusion-controlled process, and the cyclic voltammetric reduction peak potential () was found to be more positive by ca. 1.0 V than that of the coexisting H₂O₂, e.g., the values obtained at 0.1 V s⁻¹ for PCA and H₂O₂ were 0.35 and −0.35 V, respectively, vs. Ag|AgCl|KCl (sat.) at pH 3.3. The of PCA was found to depend on pH, i.e., at pH > 4.5, the plot of vs. pH gave the slope (−64 mV decade⁻¹) which is close to the theoretical value (−59 mV decade⁻¹) for an electrode process involving the equal number of electron and proton in the rate-determining step, while at pH < 4.5, the was almost independent of pH. The relevant electrochemical parameters, Tafel slope, number of electrons, formal potential (E⁰′), cathodic transfer coefficient and standard heterogeneous rate constant (k⁰′) for the reduction of PCA and the diffusion coefficient of PCA were determined to be ca. 100 mV decade⁻¹, 2, 1.53 V (at pH 2.6), 0.29, 1.2 × 10⁻¹² cm s⁻¹ and 0.29 × 10⁻⁵ cm² s⁻¹, respectively, and except for E⁰′, the obtained values were almost independent of the solution pH. The overall mechanism of the reduction of PCA was discussed.     

Kinetic analysis on LiFePO4 thin films by CV, GITT, and EIS

LiFePO₄ thin films were deposited on Ti substrates by pulsed laser deposition (PLD). The apparent chemical diffusion coefficients of lithium in the films, , were measured by cyclic voltammetry (CV), galvanostatic intermittent titration technique (GITT), and electrochemical impedance spectroscopy (EIS). The average values calculated from CV results were in the order of 10⁻¹⁴ cm² s^–1. The values obtained by GITT, and EIS techniques were in the range of 10^–14–10^–18 cm² s^–1, 10^–14–10^–18 cm² s^–1, respectively. The values obtained by the two methods show a minimum point at x ∼ 0.5 for Li_1−xFePO₄. However, the overpotential values of the LiFePO₄ thin film electrodes obtained from the GITT results and the diffusion impedance deduced from the impedance spectra also show the minimum values at x ∼ 0.5 for Li_1–xFePO₄. This contradict could be caused by the improper use of GITT and EIS techniques for measuring the chemical diffusion coefficient of Li in Li_1–xFePO₄ which constitutes two phase, i.e., LiFePO₄ and FePO₄ in this region.     

Scaling analysis of nanofiltration systems fed with saturated calcium sulfate solutions in the presence of carbonate ions

Nanofiltration of saturated calcium sulfate solution of 0.02 mol/L calcium content, with molar ratio ranging from 0.0500 to 2.8 × 10⁻³ was carried out in 6.5 × 10⁻⁴ m² active membrane area laboratory module at 1.2 × 10⁶ Pa transmembrane pressure using the total retentate recycle mode. Permeate mass flow and retentate calcium concentration vs. time and concentration factor (CF) curves allowed identification of calcium sulfate crystallization mechanisms. Though both bulk and surface crystallization mechanisms were identified, they were, however, strongly affected by water quality. A non-fouling CF range up to 2, which is probably due to the existence of metastable supersaturated CaSO₄ solution, was also observed in the case of molar ratio equal to 2.8 × 10^−3. Inorganic scales at the end of each experiment were removed from a NF module, dried at room temperature during 24 h and then examined using the X-ray diffraction method. Gypsum and aragonite were identified as the most common calcium sulfate and calcium carbonate precipitates, respectively. A mixed salt Ca₂(CO₃)(SO₄)·4H₂O (so-called rapidcreekite) was also identified as a result of carbonate and sulfate co-precipitation under low carbonate content conditions.     

Ionic conductivity of polymer gels deriving from alkali metal ionic liquids and negatively charged polyelectrolytes

We have prepared polymer gel electrolytes with alkali metal ionic liquids (AMILs) that inherently contain alkali metal ions. The AMIL consisted of sulfate anion, imidazolium cation, and alkali metal cation. AMILs were mixed directly with poly(3-sulfopropyl acrylate) lithium salt or poly(2-acrylamido-2-methylpropanesulfonic acid) lithium salt to form polymer gels. The ionic conductivity of these gels decreased with increasing polymer fraction, as in general ionic liquid/polymer mixed systems. At low polymer concentrations, these gels displayed excellent ionic conductivity of 10⁻⁴ to 10⁻³ S cm⁻¹ at room temperature. Gelation was found to cause little change in the diffusion coefficient of the ionic liquid, as measured by pulse-field-gradient NMR. These data strongly suggest that the lithium cation migrates in successive pathways provided by the ionic liquids.     

Study on physical and electrostatic interactions of counterions in poly(perfluorosulfonic) acid matrix: Characterization of diffusion properties of membrane using radiotracers

The self-diffusion coefficients of water and ions were used to study the physical (tortuosity) and electrostatic interactions of counterions in poly(perfluorosulfonic) acid membrane (Nafion-117) matrix. The self-diffusion coefficients of water were measured in the water swollen Nafion-117 membrane with Zn²⁺, Ca²⁺, Sr²⁺, and Fe²⁺ counterions by analyzing the experimental exchange rates between tritium tagged water (HTO) in membrane and equilibrating water. In order to study the effects of equilibrating solution, the HTO-desorption rate profiles between the membrane samples in H⁺ or Cs⁺ forms and equilibrating solution containing CsCl or HCl (0.25 mol/L) were measured. It was observed that the HTO-exchange rate profile was slower in case of membrane sample in Cs⁺-from equilibrated with salt/acid solution than that equilibrated with deionized water in same ionic form. However, HTO-exchange rate profile did not alter in case of H⁺-form of membrane on equilibration with salt or acid solution. The variation of ln  with polymer volume function V_p/(1 − V_p), where V_p is polymer volume fraction, indicated that: (i) in the membrane with multivalent counterions was lower than that reported for membrane with monovalent counterions at same V_p, and (ii) the linear trends observed in variation of ln  with V_p/(1 − V_p) for multivalent and monovalent counterions were significantly different. The values of in membrane normalized with at V_p = 0 were taken as an estimate of the tortuosity factor for self-diffusion of ions in the membrane matrix. The self-diffusion coefficients of ions reported in the literature along with tortuosity factor obtained from in the corresponding ionic forms of the membrane were analyzed to obtain the charge (Z_i) independent electrostatic interaction parameter g(φ) of monovalent and divalent ions in the membrane. This analysis indicated that g(φ) also vary exponentially as a function of V_p/(1 − V_p) irrespective of charge on counterions. In order to study the influence of V_p on diffusional transport rates of Na⁺ and Cs⁺ ions in membrane, a permeation experiment was carried out using H⁺-form of membrane having high water volume fraction. The diffusional transport rates of Cs⁺ and Na⁺ in H⁺-form of membrane were found to be similar indicating that the water volume fraction in membrane has strong influence on the parameters that govern the diffusion across the Nafion-117 membrane.     

Kinetics of dissociative adsorption of ethylene glycol on Pt(1 0 0) electrode surface in sulfuric acid solutions

The dissociative adsorption of ethylene glycol (EG) on Pt(1 0 0) electrode surface cooled in air after flame annealing was investigated by using programmed potential step technique and in situ FTIR spectroscopy. The stable adsorbates derived from EG dissociative adsorption on Pt(1 0 0) were determined by in situ FTIR spectroscopy as linear- and bridge-bonded CO. The quantitative results demonstrated that the average rate of dissociative adsorption of EG on Pt(1 0 0) surface varies with electrode potential, yielding a volcano-type distribution with a maximum value located near 0.10 V versus SCE. From the variation of the quantity of CO adsorbates generated in EG dissociative adsorption with the adsorption time t_ad, the initial rate (ν_i) of this surface reaction was evaluated quantitatively. The maximum value of ν_i has been determined to be 2.64 × 10⁻¹¹ mol cm⁻² s⁻¹ in a solution containing 2 × 10⁻³ mol L⁻¹ EG. The influence of the surface structure of Pt(1 0 0) electrode obtained by different pretreatment as well as of the specific adsorption of (bi)sulfate anions on the kinetics of EG dissociative adsorption has been also investigated and discussed. In comparison with a Pt(1 0 0) surface cooled in air atmosphere after flame treatment, the Pt(1 0 0) surface cooled in an Ar-H₂ stream or subjected to a treatment of fast potential cycling decreased significantly the initial rate ν_i of EG dissociative adsorption. Similar effect was also observed for the specific adsorption of (bi)sulfate anions. However, the maximum attainable coverage () of adsorbates derived from EG dissociative adsorption is not affected either by the surface structure of Pt(1 0 0) or by (bi)sulfate anions adsorption.