Waste Heat Energy Harvesting by use of BaTiO₃ for Pyroelectric Hydrogen Generation

Preparation of BTO Powder Material

BTO powder was prepared from commercially available single crystals (supplied by MaTecK) to get a form of the pyroelectric material that provides higher geometrical surface area than large chunks needed for the present experimental investigations. For this purpose irregular pieces of BTO single crystals were crushed and ground manually with an agate mortar and pestle to achieve a fine powder. Finally, single crystal particles with a maximum size of d_P<100 µm were fractioned with a sieve. XRD analysis of the gained powder shows a good tetragonality, i. e. the splitting of the cubic 002 into 002 and 020 reflections (Figure 1).

Figure 1:

XRD analysis of the crystalline powder.

Design of Hydrogen Generator

Due to its high fugacity as much as possible hydrogen should be generated pyroelectrically in a short time. To maximize the provided surface charges ΔQ=ΔPS →A with ΔPS →=p⃗ ΔT the effective crystal surface area (A) and the value of temperature change (ΔT) are assessed as key process parameters in order to produce a large amount of hydrogen. Though the former was increased several orders of magnitude through the powder preparation. In contrast there is a strict limit for the upper temperature because of phase transition from the tetragonal to paraelectric cubic phase of BTO. Since, the hydrogen volume fraction of the sample gas is directly proportional to temperature change frequency a fast temperature change is much more important. Thus a fast heat transfer to and from the pyroelectric material is essential to facilitate short cycle times. This is accomplished by a small ashlar-formed container made of polystyrene (PS). The container is composed of two full-faced, square-cut sides with the dimensions 40×40×2 mm³ and a frame with equal size and a cutout of approx. 25×25 mm² and a material thickness of 2 mm. All parts were stacked on top of each other enclosing a void of about 1250 mm³. Temperature is measured on the inside with a type K thermocouple. Two capillary tubes serve as N₂ inlet and outlet to the gas measuring system. Before the stack was glued together with cyanoacrylate, the void was filled loosely with a mass of 3.1 g BTO crystalline powder (Figure 2). Destilled water was filled into the container with the capillary tube as shown in Figure 4 after set up characterization (see below) before the test experiments.

Figure 2:

Picture of sandwiched container made of PS filled with polarized crystalline BTO powder, capillary tube and thermocouple, sticking in the container (External dimension of the container 40×40mm²).

Figure 3:

Scheme of the sandwiched container (see Figure 4) positioned within Peltier elements (A) and picture of the experimental set-up (B), with enlarged detail (C) showing container between lower cooling Peltier elements with capillary tube for the gas measuring system.

Figure 4:

Temperature profile by cycling the container between lower cooling and upper heating Peltier elements.

Polarization of the Ferroelectric Material

In order to increase the net polarization of the BaTiO₃ particles, and thus maximize their pyroelectric activity, a polarization procedure, i. e. the application of an electric field above the coercive field strength of the material, is required for a parallelization of the randomized dipole moments. To minimize arcing, the material was poled in vacuum between two parallel stainless steel plates, to avoid contamination by the conventional “oil poling technique”. Assuming a coercive field strength of E_C=0.35 MV/m for single crystals (Shieh et al. 2009), a minimal voltage of 1.75 kV with a plate distance of 5 mm is necessary for polarization reversal. Thus, the powder filled container was poled at 2 kV for approx. 15 min.

Design of Experimental Set-up

To subject the pyroelectric material to a fast temperature cycling we decide to move the sample material between an hot and cold reservoir instead of switching between cooling and heating of the static device. Figure 3 shows the design of the test setup. Due to their high heat power density we use four Peltier heat pumps, two of them with the hot side opposite to each other as heat source and the other with the cold side opposite to each other for cooling the filled container. To ensure best operation mode (i. e. backside ambient temperature) the Peltier elements were equipped additionally with four heat dissipators and fans. The whole setup was fixed in a support frame assembled from aluminum profiles. During the experiment the filled container was easily moved up and down manually using a crank with connecting rods.

Characterization of Temperature Cycling

The experimental procedure started with the characterization of the temperature cycling. After steady-state was attained for the upper and lower temperature level with the help of the Peltier elements, the container was set between the upper heating elements until equalization. Then, the temperature cycling started waiting to reach the preset temperature at each reversal point. This yields to large cycle duration of 7 min between 25 °C and 66 °C, which could be reduced by changing the containers position before maximum resp. minimum temperature was reached inside as shown in Figure 4. Keeping the presetting of current resp. voltage of the Peltier elements constant with higher frequency of the cycling of the container, there is either the favored fast temperature change between 35 °C and 56 °C, but low ∆T or the favored high ∆T between 45 °C and up to 90 °C, but slow temperature change inside the container. For the experimentation with hydrogen generation and gas analyzing we choose a periodic temperature change from 40 °C to 70 °C (ΔT=30K) and duration of one cycle of 2 min (f=8,3mHz) as a good compromise out of it.

After the basic characterization of the temperature cycling the container was filled with distilled water and put into the test setup for carrying out the tests.

Gas Analyzer for Hydrogen Detection

With help of a highly selective and sensitive measuring system, hydrogen and oxygen can be detected with lower limits of about 50 Vol.-ppb. Figure 5 shows a schematic of the measuring system, consisting of a sample injection port, gas chromatographic columns and a solid-electrolyte cell acting as detector (Schelter et al. 2013). The gaseous sample is injected into a carrier gas flow of the gas chromatograph (MG1 from SRI Instruments Europe GmbH, Bad Honnef, Germany) by using a sampling loop with a volume of 1 ml. Two packed columns, one with silica gel and one with a molecular sieve, provide the separation of the components such as hydrogen, oxygen and water vapour within 5 min at 50 °C column temperature in Ar carrier-gas flow of 20 ml/min.

Figure 5:

Schematic viewgraph of the measuring system consisting of a sample injection system, gas chromatographic columns and a coulometric solid electrolyte detector.

The coulometric detector used in this work consists of a heated tube made of yttria-stabilized zirconia (8 mol-% yttria) with the outer diameter 5.7 mm and a wall thickness 1 mm. The tubular platinum electrodes on the inner and the outer surface of this tube are 60 mm in length. Between the working (inner) and the reference (outer) electrode a constant potential of −450 mV was applied by a self-made potentiostat, where the measuring gas is flowing through the tube while air serves as reference gas.

The principle of the coulometrically operated solid electrolyte detector is based on the complete oxidation or reduction of the gas component to be measured at the working electrode of a solid electrolyte cell heated to 650 °C resulting in an electronic current.

Figure 6 shows results of the determination of the detection limit of the measuring system for hydrogen. The measuring gas is a mixture of H₂ and N₂ adjusted by mass flow controllers (Brooks Instrument Company, Hatfield, USA). The determined detection limits, marked by dashed lines, show that in the sub-ppm concentration range the peak height h is a more accurate measure than the peak area A. Those two variables are determined by equation (I) and (II). The equations involve the measured current signal of the gas chromatograph It and the linear baseline function Ibt=m∗t+n. The baseline function is determined by choosing two points in the chromatogram, one of them at the time t₁ before the hydrogen peak and one at t₂ shortly after this peak.

1. Peak height: h=max It−Ibt

Figure 6:

Results for the detection limit of the measuring system for hydrogen (Schelter et al. 2013).

(II) Peak area: A=∫t1t2(I(t)−Ib(t)) dt

Proof of Hydrogen Generation

Before the start of the temperature cycling, a first measurement of the gas phase was done. Figure 7, shows 6 different measurements of the gas analyser. The curve progression for the measurements M5 – M10 shows at first the incoming baseline (Section 1). The relatively constant level of this baseline is mainly determined by the oxygen traces in the carrier gas of the chromatographic separation. In Section 2 this baseline shift to more negative currents is caused most probably by oxygen from the injected sample which is not retended in the column due to oxygen overload by the injected sample. Then a sharp shift in Section 3 occurs due to a switching between two current measuring ranges which are three orders of magnitude distant of each other (25 µA, 25 mA). The noise is more pronounced and a tiny deviation between the two ranges causes the shift of the baseline. At the end sharp bend to negative current increase (Section 4). This is caused by the beginning of the large oxygen peak in the sample. The curve labeled with measurement No. M5 and with the H₂ peak area shows with residual hydrogen from ambient air still contained in the overall system. Hence a purge with nitrogen was done for several minutes. Meanwhile further measurements were done (M6 … M8) until no more hydrogen was detected. Then temperature cycling started again and after a few cycles a volume fraction of hydrogen of 300 Vol.-ppb in the sample gas was detected (measurement No. M10 and labeled H₂ peak area), which is attributed to pyroelectric water electrolysis. The hydrogen concentration in the sample is calculated from the peak area (see picture) of the hydrogen peak and the sample loop volume using the Faraday law with a sample loop volume of 0.95 ml and a peak area of 2 µAs (Schelter et al. 2013).

Figure 7:

Proof of pyroelectric hydrogen generation with a highly selective measuring system with coulometric solid electrolyte detector.