INTRODUCTION

The Dual-Method Calorimeter was specifically designed to measure the heat evolution from electrolysis cells with flowing electrolyte. The system achieves a high degree of validity by performing two completely independent simultaneous measurements of the heat output of the cell. First, the electrolysis cell is equipped with inlet and outlet temperature sensors which, combined with a known and constant flow rate, provide a mass-flow calorimetry (MFC) measure of the heat generated in the cell. Second, the entire electrolysis experiment (cell, pump, and electrolyte reservoir) is housed in an insulated, instrumented enclosure which is immersed in a constant temperature air bath. The total heat power liberated in the experiment enclosure is measured as a linear function of the delta-T across the enclosure walls according to Newton's Law of Cooling (NLC).

ELECTROLYSIS CELL

The electrolysis cell is depicted in Figure A. It consists of a heavy-walled glass tube with PVDF end plugs. The end plugs are fitted with glass-jacketed temperature probes which contain precision thermistors. The probes are positioned so that the sensitive end is fully immersed in the flowing electrolyte stream. The end plugs also have elastomer seals that pass a .020" dia Pt wire into the cell at each end for the electrolysis. Each Pt wire runs through the electrolyte passage and out the open end of the plug (which faces the contents of the cell and the opposite end plug). There the wire is bent sharply outwards radially and then bent again in a circumferential direction and arranged in a single circular turn that lies against the face of the end plug.

For a typical experiment, the cell is assembled as follows. First one end plug is inserted halfway into the glass tube. This assembly is then held upright so the open end of the glass tube is facing upwards. Next a tight-fitting circular piece of fine-mesh Pt screen is inserted into the glass tube and pushed down into contact with the Pt lead wire described above. Then the charge of metal-coated beads is poured into the cell (usually 1-2 cm3) and settled as desired (e.g. by filling the cell with water and tapping). Then several circular plastic woven-mesh screens with 200-500 micron openings are inserted into the glass tube and pressed gently into contact with the bead bed. These screens serve to separate the beads from the anode to prevent short-circuiting the cell and the stack is typically 1mm thick. Next a second circular piece of the same Pt screen is inserted into the glass tube and finally the other end plug is inserted until its Pt lead wire contacts the upper Pt screen. A clamp consisting of square plates and 4 threaded rods with nuts is used to apply axial pressure to the end plugs to put the entire assembly in compression. This ensures good electrical contact throughout the cell.

INNER ENCLOSURE

Figure B shows the interior of the inner enclosure. The cell is mounted vertically on the left in a small insulated box. The purpose of this box is to minimize heat exchanges between the walls of the cell and the environment so that nearly all of the heat that leaves the cell does so through the flowing electrolyte. The electrolyte pump is a Cole-Parmer L/S peristaltic pump fitted with size 16 Norprene® tubing which provides a nominal displacement of 0.8 ml per revolution. The pump is driven at a constant 30 rpm by a synchronous AC gearmotor which is located outside the inner enclosure. A fiberglass drive shaft leads from the motor through a snug hole in the inner enclosure wall to the pump. This arrangement prevents any significant amount of the heat dissipated by the motor from entering the inner enclosure. However, the frictional losses in the pump head are dissipated in the inner enclosure so these must be accounted for in the calibration.

Also located in the inner enclosure is the electrolyte reservoir. This is a 100ml beaker fitted with an O-ring sealed top which has three ports. Electrolyte is drawn from the bottom of the reservoir by a short suction tube inside the beaker. This electrolyte then goes through the pump, through the cell, and returns to the reservoir. Returning electrolyte is delivered into a simple baffle that forces the flow up to the surface of the reservoir and prevents the very fine gas bubbles produced by the electrolysis from directly entering the suction tube and being recirculated. The gas bubbles rise to the surface of the electrolyte and the gas enters the head space above the electrolyte. The third port in the top opens directly into the head space and is connected to a small Tygon tube that runs all the way to the outside of the calorimeter system. This conveys the electrolysis gases safely out of the calorimeter enclosure and allows measurement of the gas flow rate when desired. It is important to measure the gas evolution from open electrolysis cells undergoing calorimetric measurement to determine the amount of recombination (of the hydrogen and oxygen gas) that occurs inside the cell. Typically, the amount of recombination in these experiments is negligible and, in the calorimeter calibration, it is therefore assumed that all of the caloric value of the hydrogen and oxygen gas escapes from the cell.

Air in the inner chamber is circulated by two low-power fans located in the small space below the experiment platform. This space is connected to the experiment chamber by two rectangular openings, one at each end of the platform. The fans force air to flow in a counterclockwise direction entering the experiment chamber on the right and leaving it on the left. This flow arrangement was designed to minimize dead spots and to ensure that virtually all the air in the inner chamber is circulated. See the CALIBRATION section for additional information on this subject.

The temperature of the air in the inner enclosure is measured by three thermistors. Each thermistor is mounted on a small heat sink which couples it thermally to the air. Two of the sensors are located below the experiment platform (one in the inlet stream of each fan) and the third sensor is located near the top of the experiment chamber. These three sensors are monitored by the controlling computer and their average is taken as the air temperature of the inner enclosure for the NLC calorimetry.

Many wires run into the inner enclosure to connect to the temperature sensors, electrolysis cell, and fans. These wires pass into the inner enclosure through a long, separately insulated wireway which reduces thermal heat transfer through the wires to negligible levels.

OUTER ENCLOSURE

Figure C shows the overall arrangement of the outer enclosure which provides a constant temperature environment (air bath) for the inner enclosure. The air bath contains two powerful fans which circulate the air vigorously. Directly in the outlet stream of each fan is a 100 watt heater mounted to a large finned heat sink. There are also three air temperature sensors in the air bath, one in the inlet stream of each fan and a third sensor near the top of the outer enclosure. The computer monitors these three sensors, averages their readings, and modulates the duty cycle of the heaters with a proportional-integral-derivative (PID) algorithm to maintain a constant average air temperature. The outer enclosure is insulated to minimize heat losses to the room. The 200 watts of heater power is sufficient to maintain the interior of this enclosure at any desired temperature above ambient up to about 80°C.

COMPUTER SYSTEM

The calorimeter is monitored and controlled by a computer-based data acquisition system that consists of two 8-channel analog input boards (a Computer Boards Inc, CIO-DAS801 and a CIO-DAS08/Jr-AO) operating in a 40MHz 386 PC. In addition to the three temperature sensors in the inner enclosure, the inlet and outlet temperature sensors in the electrolysis cell, and the three temperature sensors in the outer enclosure, the system also monitors the voltage and current applied to the electrolysis cell and the ambient temperature in the room.

The PC runs a custom control program written in MicroSoft QuickBASIC. This program logs all raw data to disk and displays selected parameters such as cell voltage, input power, MFC output power, NLC output power, environment temperature, and room temperature on the screen. As mentioned above, the program also controls the temperature of the environment chamber during the run. If desired, the program can also integrate the measured input and output powers and plot the total energy balance for the run.

PERFORMANCE

In a large number of experiments and calibration runs it has been empirically observed that the system achieves a one-sigma measurement precision of about +/- 0.05 watts for the NLC calorimetry and +/- 0.02 watts for the MFC measurements. Measurement accuracy was determined from the calibration runs and for both methods is typically about +/- 0.05 watts or 10% relative, whichever is larger.

Despite the insulation in the walls of the outer enclosure and the tight temperature control provided by the computer, the NLC calorimetry exhibits a small but noticeable sensitivity to ambient temperature changes. The effect is positive and appears to be related to both the temperature of the room and the rate of change of the room temperature. The error due to this effect is typically less than 0.1 watts and can be seen in some of the examples presented below.

Figure D shows a screen display from a run in which the electrolysis cell was filled with a charge of Pd/Ni/Pd coated ersatz beads made by EarthTech for an attempted replication of the Patterson effect. In the plot portion of the screen, the purple trace represents the input power (Pin), the blue trace is the MFC output power (Pflow), and the red trace is the NLC output power (Pnlc). These quantities are plotted vs time on a 0-1 watt scale with P=0 represented by the horizontal gray line near the bottom of the plot. The clock time (24 hour format) for this 48 hour run is displayed along the bottom of the plot. The cell voltage is plotted in green on a 0-7 volt scale. The brown trace is the room temperature plotted on a 26 +/-5°C scale (26°C is the center of the plot and the full scale is 5°C in each direction) and the ragged gray trace in the center of the plot is the outer enclosure temperature plotted on a 50+/-1°C scale. The computer maintained the environment temperature at 50°C +/- 0.03°C throughout the run.

Periodically during this run, the electrolysis current was increased manually. This created the stair-step appearance of the power traces. Note the reasonably good agreement (within 0.1 watts) between the Pin, Pflow, and Pnlc traces throughout the run. This indicates that there were no significant heat anomalies produced by this experiment.

Figure E shows a screen display from a run in which the electrolysis cell was equipped with a resistive heater in the bead space so that excess heat production could be simulated. For the first three days of the run, there is zero input power. This period serves to demonstrate the zero stability of the system (the Pnlc and Pflow traces stay within 0.03 watts of zero most of the time). At about 0900 on the 4^th day, electrolysis was initiated with a current of 70mA which corresponds to a net input power of about .2 watts. The Pnlc (red) and Pflow (blue) traces respond fairly quickly and match the input power quite closely. At about 1800 hours that same day, 0.10 watts of power was delivered to the resistive heater by an external power supply that was not being monitored by the computer (in order to simulate a real excess heat event). This run demonstrates that the system is clearly capable of detecting an excess heat event in the electrolysis cell as small as 0.1 watts.

Figure F shows a screen display from a special test run in which the experiment was a pair of NiCd "D" cells (connected in series) undergoing a charge/discharge cycle. The charging power supply was located outside the calorimeter and was monitored by the computer (to produce a Pin trace). The discharging was accomplished with a 3 ohm load resistor located in the inner chamber with the batteries. A pair of wires led out of the calorimeter system to permit the load resistor to be switched in or out of the circuit. In this experiment, only the Newton's Law of Cooling output power measurement was active and the program was set to integrate and display the total energy values. The purple line (Pin) shows the input electrical charging power and the red line (Pnlc) shows the measured heat output power. The run begins at 0830 hours and, at about 1430, the calorimeter had come to equilibrium and the charging was started. This point is marked by the reset of the Eout trace (green) which had been integrating the disequilibrium heat output signal. This point is also marked by the sudden appearance of the Pin trace (purple) which jumps up to about 3 watts of electrical input power (full scale is 4 watts on this plot). Note that during the first 5 hours of charging (constant 1 amp current) there was little or no heat evolved from the batteries, meaning that almost all of the charging power was being converted into stored chemical energy.

By 0400 hours the next day, the batteries were fully charged and the Pnlc trace had met the Pin trace indicating that all the electrical input power was now being converted to heat. At 0617, the charging current was turned off. Note that the Pin trace (purple) drops to zero promptly but the Pnlc trace (red) approaches zero slowly as the batteries and other apparatus in the inner chamber cool off. By 1500 hours thermal equilibrium was again reached and the difference between Ein and Eout was about 48000 joules. At that time the load resistor in the inner chamber was switched across the batteries. The Pnlc trace shows the resulting heat output and, rewardingly, the Eout trace (green) eventually approaches the Ein trace (blue) very closely. This indicates a "unity" energy balance for the charge-discharge cycle...which is to be expected.

 The 48000 joule difference that existed at the end of charging is reasonably consistent with the batteries' rated capacity of 5 amp-hours at 1.2 volts, which works out to 42000 joules for 2 cells. The rated capacity is a measure of the energy that can be delivered at a usable voltage and is therefore somewhat less than the total energy stored in the cells which was measured in this experiment.

CALIBRATION

The flow calorimetry provides a fundamental measurement of the heat leaving the cell and does not require an empirical calibration. The heat power flowing out of the cell via the electrolyte stream is given by:

P = F·DT·C

where F is the mass flow rate, DT is the temperature difference between the inlet and outlet electrolyte streams, and C is the specific heat of the electrolyte. The electrolyte used in these experiments is a Li2SO4 solution. In measurements made in our lab, we found that, as the concentration of Li2SO4 increases, the density of the solution goes up but the specific heat goes down such that the product of these two quantities remains reasonably close over the range 0-2 molar (within a few percent relative) to that of pure water (i.e. 4.19 joules/ml·K). Thus the heat power can also be expressed as:

  P = G·DT·CH2O

where G is the volume flow rate, DT is the delta-T and CH2O is 4.19 joule/(ml·K). We measured G manually with a graduated cylinder and stopwatch on numerous occasions and found it to be equal to 23.2 +/- 0.4 ml/min. This was deemed sufficiently constant to be treated as such in the MFC calculations and the system therefore does not have a flowmeter that is monitored by the computer.

The temperature sensors are highly reliable thermistors guaranteed by the manufacturer to be accurate to 0.2°C. We have used precision glass thermometers accurate to 0.1°C to confirm this claim. For the precise delta-T measurement required, it is necessary to characterize the offset between the inlet and outlet sensors. This is done by running the system with zero electrolysis power input.

The Newton's Law of Cooling calorimetry is calibrated empirically1 by delivering known power levels to a calibration resistor located in the inner chamber. The effect of all the various heat leaks (enclosure walls, pump drive shaft, wires, electrolysis gas vent, etc.) are lumped into one overall thermal resistance value which is typically about .6 °C/watt for this system. Despite the stirring of the air inside the inner enclosure, the response of the NLC system is somewhat sensitive to the location of the heat source. The maximum observed "location error" is approximately 10% relative1 . For optimum accuracy in the NLC calorimetry, a calibration can be performed using a calibration resistor that closely resembles the size, shape, power, and location of the experimental device being evaluated.

The small stirring fans and the peristaltic pump are heat sources that have been accounted for in the NLC calorimetry. The heat input from these devices was measured empirically and included in the offset expression for the NLC calorimetry. For reference, the fans are approximately 20 milliwatts each and the friction losses in the pump head amount to about 0.25 watts.

BACK TO EQUIPMENT PAGE