Spontaneous formation of nanoparticles on electrospun nanofibres

We report the spontaneous formation of nanoparticles on smooth nanofibres in a single-step electrospinning process, as an inexpensive and scalable method for producing high-surface-area composites. Layers of nanofibres, containing the proton conducting electrolyte, caesium dihydrogen phosphate, are deposited uniformly over large area substrates from clear solutions of the electrolyte mixed with polymers. Under certain conditions, the normally smooth nanofibres develop caesium dihydrogen phosphate nanoparticles in large numbers on their external surface. The nanoparticles appear to originate from the electrolyte within the fibres, which is transported to the outer surface after the fibres are deposited, as evidenced by cross-sectional imaging of the electrospun fibres. The presence of nanoparticles on the fibre surface yields composites with increased surface area of exposed electrolyte, which ultimately enhances electrocatalytic performance. Indeed, solid acid fuel cells fabricated with electrodes from processed nanofibre-nanoparticle composites, produced higher cell voltage as compared to fuel cells fabricated with state-of-the-art electrodes.

b Photograph of the experimental setup with the fixed collector electrode, showing the digital multimeter that is connected in series with the negative DC bias power supply and the heat gun. c Schematic illustration of the nozzle-free electrospinning setup with the rotating collector electrode. d Photograph of the experimental setup with the rotating collector electrode.
e Picture of the collected electrospun CDP-PVP-PANI sample after removal from the rotating collector electrode. Scale bar is 1 µm for both images. Using 50 mg mL -1 CDP 30 mg mL -1 PVP and 0.1 mg mL -1 PANI concentration in the solution, a nanoparticle decorated nanofiber composite mat with a mean diameter of the nanofibers being 123.9 ± 32 nm, and the mean diameter of the nanoparticles being 104.5 ± 28 nm was achieved. The nanoparticle density was 29 particles µm -2 .
Furnace treatment was applied for the reduction of the polymer a. The electrospun nanofibrous sample mat (310 mg) was placed in a Barnstead box furnace (Supplementary Figure 7). After heating the sample to 200 °C with 5 °C min -1 , and holding the temperature for 1 hour, the temperature was ramped to 230 °C with 5 °C min -1 , where it was again held for 1 hour. Then the temperature was increased to 300 °C with 5 °C min -1 . The sample was left at 300 °C for 12 hours in air. The composite lost 38.5% of the PVP content (as the weight decreased from 310 mg to 191 mg). During the process, the sample changed colour from white to brown. Then the treated sample was placed on top of a Toray carbon paper, above 100 mg Platinum(II) acetylacetonate powder. After an MOCVD process, where the sample was heated to 210 °C in a vacuum oven (-27 inch.Hg pressure) in the presence of the platinum precursor and 2 mL deionized water, 62 mg Pt was deposited on the nanofibrous CDP mat. The platinum particles diffused into the sample and deposited homogeneously, increasing the sample mass to 253 mg. The sample was then sieved over a 53 µm sieve, and hand-spread on a Toray carbon paper. Finally, the powder on the carbon paper was pressed with the 2" diameter anode+electrolyte half-cell (prepared by SAFCell standard procedure) with 1 kton pressure for 3 seconds. Figure 7. The electrospun sample mat (310 mg) was placed in a box furnace (300 °C heat treatment for 12 hours in air). Upon removing the 38.5% of PVP, the weight decreased to 191 mg. The sample changed colour from white to brown. Then Pt was deposited on the sample by MOCVD process. The sample was then sieved over a 2 µm sieve, and hand-spread on a Toray carbon paper, which finalised the cathode assembly. The 2" diameter electrode was tested in a symmetric cell mode using H2 gas on the anode and cathode.

Supplementary
Impedance measurement of cell with standard SAFCell Pt on carbon "anode" as pseudo-hydrogen reference electrode, and electrospun electrode on "cathode" as working electrode was performed. High frequency intercept with real axis was taken to measure all electronic, plus electrolyte layer protonic losses across cell, with area specific resistance value of 0.32 Ω cm 2 . Real axis difference between high and low frequency intercepts measures the total electrode resistances, with a value certainly less than 70 mΩ cm 2 (0.07 Ω cm 2 ), for both electrodes. However, it is hard to estimate the exact value for the low frequency intercept with real axis, as the arc is so small, and the data became noisy at lower frequencies (Supplementary Figure 10a). Hence, a voltage versus current curve was taken in same symmetric hydrogen setup. From the data it is clear that the curve is linear in nature, and we can then assume that the cell is in the "linear Butler Vohlmer" region, where the slope of the data corresponds to the total ohmic (electrical + protonic) and non-ohmic (electrodes) losses across cell. Subtracting out the ohmic area specific resistance from this slope, gives a value of 50 mΩ cm 2 (0.05 Ω cm 2 ), for both electrodes (Supplementary Figure 10b). The value for the standard SAFCell Pt on carbon anode is 30 mΩ-cm2 (0.03 Ω cm 2 ). This means the value for the electrospun electrode in symmetric hydrogen is around 20 mΩ cm 2 (0.02 Ω cm 2 ).
Supplementary Figure 10. Impedance measurements on electrospun cathode assembly taken in a symmetric hydrogen setup. a electrochemical impedance of cell with standard SAFCell Pt on carbon "anode" as pseudo-hydrogen reference electrode, and the electrospun electrode on "cathode" as working electrode. b Voltage versus current curve. Subtracting out the ohmic area specific resistance from this slope, gives a value of 50 mΩ-cm 2 , for both electrodes. The value for the standard SAFCell Pt on carbon anode is 30 mΩ-cm 2 . This means the value for the electrospun electrode in symmetric hydrogen is around 20 mΩ-cm 2 (0.02 Ω-cm 2 ). a b Supplementary Figure 12. Electrochemical performance of the fuel cell made with the electrospun CPD-PVA after 16.19 hours of operation (solution parameters: 30 mg mL -1 CDP, 30 mg mL -1 PVP and 0.1 mg mL -1 PANI concentration). The blue data points show the real I-V curve, whereas the red data is iR-free. The dotted blue and red lines show the power density. Both the cell voltage and power density are well below the performance of the CDP-PVP-PANI samples, or the SAFCell Inc. standard.