- Email: [email protected]
Design and testing of a calorimeter for measurements on electrochemical reactions with gas evolution
M-2567 j, Chem. Thermodynamics 1991, 23, 95-104
Design and testing of a c a l o r i m e t e r f o r m e a s u r e m e n t s on electrochemical reactions w i t h gas evolution GERD OLOFSSON, a INGEMAR WADSO, a and LENNART EBERSON b Thermochemistry ~ or Organic Chemistry 3, b Chemieal Center, University of Lund, P.O. Box 124, S-221 O0 Lund, Sweden (Received 17 September 1990)
A heat-conduction calorimeter suitable for the study of electrochemical processes has been constructed. The reaction vesselof 25 cm3 volume was fitted with electrodes, calibration heater, stirrer, thermocouple, and thin Teflon tube for addition of liquid. Gas evolved could leave the vessel through a gas-outlet tube. Good thermal conductivity was achieved by mounting nine thermocouple plates between the vessel and the surrounding heat sink and by letting the heat sink be in good thermal contact with the thermostat bath. The calorimeter was tested by measurements on the electrolysis of water and of heavy water and found to have an accuracy better than 0.2 per cent. I
1. Introduction Calorimetry is one of the most important methods for obtaining precise measurements on chemical systems and constitutes the basis of most of our knowledge of energy relations in chemistry. In one area, electrochemistry, the use of calorimetric methods ha~ been surprisingly scarce as attested to in a review from 1983, (1) although the large potential was demonstrated already in 1958 by Sherfey and Brenner. (2) So far, calorimetry appears to have been essentially limited to corrosion studies at open circuit and measurements of electrochemical Peltier effects at closed circuit. Reasons for this state of affairs are presumably perceived difficulties in correcting for non-chemical effects in the electrochemical cell, such as heat evolution caused by overpotential, resistive films on the electrode(s), and ohmic resistance in the cell. These are, however, factors that could be easily managed by using a high-precision calorimeter taking into account the special experimental features of the electrochemical method, of course in combination with the rigorous planning needed in all thermochemical experimentation. The set of phenomena that collectively appear under the concept of "electrochemically induced cold fusion ''(3'4) has recently focused interest on electrochemical calorimetry. Independently of whether one has faith in the concept or not, the hypothesis has to be thoroughly tested and accurate calorimetry is necessary. The 0021-9614/91/010095 + 10 $02.00/0
© 1991 Academic Press Limited
96
G. OLOFSSON, I. WADSO, AND L. EBERSON
experimental situation is challenging: electrolysis of heavy water must be performed at high current densities (up to 1 A. c m - 2 o r even higher), with strong gas evolution and under very long uninterrupted periods (as long as 103 h or more), and the electrical and thermal powers must be measured with a high accuracy to make possible a proper evaluation of the energy balance of the electrochemical cell. The original measurements were made using simple Dewar-vessel calorimeters where temperature differences between the interior vessel and the surrounding thermostat bath were measured33'4) Subsequent studies have indicated that this type of calorimeter may be difficult to calibrate accurately enough, primarily due to the large temperature increases caused by large power input. (5 8) Calorimeters working under essentially isothermal conditions appear to be more suitable for the present application. In an isothermal calorimeter the thermal power generated in the cell is dissipated to a surrounding heat sink and the heat flow is monitored. In the recent calorimeter design presented by Nfirger et al., ~9) the temperature increase of water flowing through a copper heat sink was measured. The primary aim was to construct an instrument for measurements of thermal powers generated during electrolysis of heavy water using a palladium cathode. The calorimeter could be used for power inputs up to 25 W and a device for the catalytic recombination of H 2 o r D 2 and 02 generated during electrolysis was built into the calorimeter vessel. We have found it of interest to develop a heat-conduction calorimeter for accurate measurements of the thermal power in electrochemical cells applicable to processes both with and without gas evolution. We report here a thermopile heat-conduction calorimeter with a 25 cm 3 reaction vessel. The calorimeter has been tested by measurements of the electrolysis of water and of heavy water and found to have an accuracy better than 0.2 per cent. The calorimeter is suitable for experimental periods of very long duration. With minor modifications, especially introduction of a membrane separating the cathode and anode compartments, the instrument should be useful for measuring enthalpies of reaction for electrode processes in general.
2. Instrument The calorimeter used was of the heat-conduction type. In such instruments, heat released in the reaction vessel is conducted to a surrounding heat sink, usually a metal block, by use of a thermopile positioned between the vessel and the heat sink. In the ideal case, the relation between the thermopile potential U and the thermal power P released in the vessel is given by the Tian equation: P = z'(U+v'dU/dt),
~
(1)
where e is a constant determined by calibration and ~ is the time constant for the instrument. If the thermal power is constant, equation (i) reduces to P = e.U.
(2)
For heat-conduction calorimeters the calibration constant ~ is nor affected by the heat capacity of the vessel or its content. This is in-c0ntrast to adiabatic or
C A L O R I M E T E R FOR E L E C T R O C H E M I C A L REACTIONS
I
97
Ie 1 cm
F I G U R E 1. Electrolysis calorimeter cell fitted with stirrer d, calibration heater e, cathode h, anode i, stainless-steel tube j containing a thermocouple for measurement of temperature in the cell, and a Teflon tube g for addition of liquid. Gas evolved could leave the cell through the central tube f.
quasiadiabatic calorimeters for which the constant is proportional to the heat capacity. This is of particular importance for the present type of application where the liquid electrolysed is partly decomposed into gases which will leave the calorimetric system during the measurement. The design of the instrument is shown in figures 1 to 3. The electrolysis calorimeter vessel, figure 1, consisted of a 25 cm 3 can c and a lid a fitted with several inserts: d, e, g, h, i, a n d j. The can and the lid were made from acid-proof steel with the inner surfaces coated with a thin layer of Teflon. An O-ring b together with a lock ring, provided a Seal between the lid and the can. There were also O-ring Seals between the inserts and the lid. The anode i consisted of a 1 mm platinum wire forming a spiral ,¢ around the cylindrical cathode h (d = 4 mm, l = 13 mm) made from acid-proof steel or palladium. The cylinder was fastened to a 1 mm wire, made from the same material. The 50 f~ calibration heater e made from manganin wire was inserted into a Teflon tube partly filled with paraffin oil. A copper-to-constantan thermocouple
98
G. OLOFSSON, I. WADSO, AND L. EBERSON
FIGURE 2. Simplified picture of calorimetric vessel showing capsule containing stirrer motor a, stainless-steel tube c surrounding the stirrer shaft and supporting the stirrer motor, gas outlet tubed connected to PVC-tube h, thermocouple leads e, thermal radiation shields f, and calorimeter cell g.
mounted in a Teflon-coated stainless-steel tube j was used to measure the temperature of the cell content. The thin Teflon tube g was used to inject liqu'id to the vessel in order to compensate for water decomposed during the electrolysis process. Gas evolved could leave the vessel through the central tube f (i.d. 4 mm) and a stirrer d made from "18 carat" gold was positioned off centre. Figure 2 shows a simplified picture of the calorimetric vessel. Above the electrolysis cell g is shown the stainless-steel tube c which surrounded the stirrer shaft and supported the stirring motor, encapsulated by a plastic tube a; the gas outlet tube d;
CALORIMETER FOR ELECTROCHEMICAL REACTIONS
~
C
H
m
r
99
/
/
/
/
/
A B FIGURE 3. Schematic picture of the calorimeter without its vessel. A, horizontal section; and B, vertical section at Y--Y. Lid a; steel tube b; outer steel can c; heat sink consisting of aluminium blocks d, e, and f; thermocouple plates i and vessel holder h.
and the thermocouple leads e. F o r clarity the Teflon tube used for injection of liquid (figure 1, g) is omitted. The gas-outlet tube was connected to a gas burette by means of a thick-walled P V C tube h (i.d. 3 mm, o.d. 7.7 mm). The a l u m i n i u m disks f serve as thermal-radiation shields. Figure 3 shows schematically vertical and horizontal sections t h r o u g h the calorimeter without its vessel. The vessel holder h, m a d e f r o m a squared stainlesssteel rod (60 m m x 30 m m × 30 mm), had a vertical hole (d = 27.5 mm) giving a close fit for the vessel. The vessel holder was surrounded by four aluminium rods e whose outside surfaces were r o u n d e d to fit into a circular cylinder with a radius of
100
G. OLOFSSON, I. WADSO, AND L. EBERSON
34 mm. Eight thermocouple plates i (Melchor CP-71, Trenton, New Jersey) were positioned between the fiat inner surfaces of the rods and the vessel holder. Two cylindrical aluminium blocks d and f with a diameter of 69 mm were positioned above and below the vessel-holder assembly. A ninth thermocouple plate was placed between the bottom of the vessel holder and block f and all the plates were connected in series. The three block units d, e, and f were bolted together to form a cylinder which fitted closely inside the steel can c. The can was immersed in a thermostatted water bath which thus would serve as a direct extension of the metal-block heat sink. The can c and its lid with upper parts were from the microcalorimeter TAM 2277-205 {ThermoMetric (LKB) Jgrf/illa, Sweden} and the instrument would thus fit into the thermostat bath of ThermoMetric's 4-channel microcalorimeter system. However, in the measurements reported here, we used a slightly less stable bath, approximately -t-0.3.10-3K during 20 h. Most heat-conduction calorimeters are used for microcalorimetric measurements and it is then usually desirable to have an isolating air gap between the heat sink and the thermostatic bath which will increase the stability of the instrument. But in the present case, where the experiments were carried out at high powers (up to 1 W or higher), an isolated heat sink would have had a significant temperature drift for a long time until it reached a steady-state value several kelvins above the temperature of the bath. For this reason, it was important to have good thermal conductivity also between the vessel and the heat sink. This was provided by the nine thermocouple plates. The time constant z for the present calorimeter charged with 21 cm 3 of water was about 250 s. Potential signals from the calorimeter were amplified using a Keithley 150 B ammeter. By the use of a microprocessor the signals were read once a second, integrated over 300 s periods, and printed out. The input of electric power during electrolysis was determined by measuring the current I¢~n and potential liceH using two Hewlett-Packard 3478A multimeters. Measurements were made once a second and the product/'cell" Io~1 was collected using a Hewlett-Packard 85 desk calculator. The power inputs during the electrical calibrations were measured in the same way. Corrections were applied for the heat effects in leads to heater and electrodes. The extent of reaction was determined by measuring the volume of gases evolved during electrolysis. The gas-collecting system consisted of a water-filled glass vessel with a glass inlet tube over which a gas burette was mounted. In the experiments with heavy water, deuterium oxide with isotopic purity better than 99.8 moles per cent (CibaGeigy) was used without further purification. The 0.1 mol" dm -3 L i O D solution was prepared by dissolving lithium metal in deuterium oxide in an argon atmosphere.
3. Results The performance of the calorimeter was tested by measurements of the electrolysis of water and of heavy water. The chemical process in the case of water is H20(1) = H2(g) + ½02(g).
(3)
CALORIMETER FOR ELECTROCHEMICALREACTIONS
101
It takes place under isothermal conditions and the power balance during electrolysis is given by P~n = d(AH)/dt + dQ/dt. (4) The electrical power input is denoted Pceu, AH is the enthalpy of dissociation of water (3), and dQ/dt is the thermal power. All three properties were measured independently. The accuracy of the calorimeter could be established by comparing the calorimetrically measured values of dQ/dt with values calculated as the difference between Pcell and d(AH)/dt. The response of the thermocouple plates was determined from electrical calibrations using the internal heater. Values of the calibration constant e, calculated as the ratio between the electrical power input Pc,~ and the amplified response of the thermocouple plates, were determined for Peal varying between 20 mW and l l00 roW. Results of 10 experiments showed s to be linear within 0.2 per cent over this power range. The linearity of the amplifier may well have been the limiting factor. The amount of liquid in the calorimeter vessel did not influence the calibration constant as values of e determined with 20.0 cm a and 23.0 cm a of liquid agreed to within 0.01 per cent (the number of experiments was 10 and P~I = 0.5 W and 0.9 W). The rate of dissociation of water was calculated from the time measured to fill the gas burette. The atmospheric pressure, the temperature, and the hydrostatic pressure of water were measured. The water filling the burette before the gas was collected was assumed to be in equilibrium with air. When being displaced, the water was presupposed to have become saturated with H E (or D2) and 0 2 and the gas mixture to have been saturated with water vapour. Values of the vapour pressure of water were taken from reference 10 and for heavy' water from reference l l. Values of Bunsen coefficients for H2,(12) and 0 2 , (13) were used to calculate the amounts of dissolved gases. The solubility of D2 in water was assumed to be the same as for H2. The pressure in the system was observed to decrease slowly after the electrolysis was stopped. The pressure drop was ascribed to the diffusion of H 2 (or D2) through the plastic tube. From the amount of water sucked into the plastic tube, the gas diffusion was estimated to be 0.089 mm 3- s -x. The collected gas was treated as a mixture of perfect gases when calculating the amount of decomposed water. Figure 4 shows results from a typical series of measurements with an initial electrical calibration experiment followed by three electrolysis experiments conducted at different electrical powers. Results of measurements on the electrolysis of water using 21.0cm 3 of 0.1 m o l ' d m 3 N a O H solution are summarized in table 1. The value AH ° = (285.83+0.040)kJ.mol 1 was used when calculating d(AH)/dt. ~14) A small correction was applied for the evaporation of water to give a saturated gas mixture leaving the vessel. The time to fill the gas burette of (103.42___0.08)cm 3 volume varied between 37min for the highest power and 81.5 min for the lowest. The average rate of reaction and accordingly d(AH)/dt was calculated by dividing the amount of decomposed water b y the time. Values of Peel1 and dQ/dt shown in columns 1 and 3 of table 1 were the average values of the electrical power input and the thermal power, calculated from values integrated over the time of the gas collections. The potential over the cell was 3.65 V at Peel~= 0.886 W and decreased to
G. OLOFSSON, I. WADS(~, AND L. EBERSON
102 I
I
I
I
I
I
I
I
1
0.6
0.4
i.
3
"1
0.2
0
I
0
I
I
2
I
4
I
I
6
t/h
I
I
8
FIGURE 4. Recorder trace of the power P at time t during a series of experiments on the electrolysis of heavy water using the stainless-steel cathode. After an initial electrical calibration three gas collections were made during time periods 1, 2, and 3 indicated in the figure (see the last three experiments in table 2). The current was 0.246 A, 0.153 A, and 0.092 A, respectively.
3.08 V at the lowest power input. The temperature of the thermostatted bath was 298.15 K and the temperature increase in the cell was less than 0.5 K even at the highest power. The results of a corresponding series of measurements on heavy water using 20.0 cm 3 of 0.1 mol- dm 3 L i O D solution are summarized in table 2. The value AH m = 294.60 k J- m o l - 1 was used for the calculation of the enthalpy of dissociation of D 2 0 . (15) The stainless-steel cathode was then replaced by an electrode of palladium metal, the vessel was charged with 21.0cm 3 of freshly prepared 0.1 mol. dm -3 L i O D solution, and an electrolysis experiment was run continuously for 83 h. After about 6 h, the conditions in the cell became stable. The results shown
TABLE 1. Results of measurements of the electrical power input Pcell, the chemical power d(AH)/dt, and the thermal power dQ/dt during electrolysis of water (0.1 mol-dm -3 NaOH) using the stainless-steel cathode ,~
P,~,,
d(AH)/dt
dQ/dt
A {(dQ/dt)/W }
W
W
W
(expt-calc.)
0.88633 0.88915 0.52498 0.34302
0.36970 0.36946 0.23950 0.16690
0.51717 0.52031 0.28639 0.17578
0.00054 0.00062 0.00091 -0.00034
103
CALORIMETER FOR ELECTROCHEMICAL REACTIONS
TABLE 2. Electrolysis of heavy water (0.1 mol. dm -3 LiOD) using the stainless-steel cathode (compare table l)
P¢¢,,
d(AH)/dt
dQ/dt
A {(dQ/dt)/W }
W
W
W
(expt-calc.)
1.01686 1.01904 1.02627 0.55709 0.30512
0.37907 0.37860 0.37878 0.23531 0.14191
0.63891 0.64027 0.64704 0.32194 0.16347
0.00112 -0.00016 -0.00045 0.00016 0.00026
TABLE 3, Electrolysis of heavy water (0.1 mol.dm -3 LiOD) using the palladium cathode (compare table 1)
P~l,
d(AH)/dt
dQ/dt
A ((dQ/dt)/W }
W
W
W
(expt-calc.)
0.49462 0,49399 1.07591 0,49172 0,49887 0,48797
0.20839 0.20799 0.38935 0.20580 0.20842 0.20331
0.28590 0.28578 0.68776 0.28565 0.29016* 0.28415
-0.00033 -0.00022 0.00120 --0.00027 -0.00029 -0.00051
in table 3 summarize measurements made during the following 3 d. The vessel was refilled once on each day by injection of 1.4 cm 3 of the LiOD solution. In one of the experiments, indicated by an asterisk in the table, extra electrical power was supplied to the calibration heater. This amount, 0.43446 W, has been subtracted from the total measured thermal power to give the value of dQ/dt shown in column 3. The temperature increase in the vessel was 0.6 K in this experiment. The cell potentials were 3.68 V or 4.30 V and they were stable to within a few mV. The corresponding current densities were 0.069 A. cm -2 and 0.129 A" cm -2. Using different equipment for power supply it would have been possible to extend the measurements to higher power levels. As can be seen from tables 1 to 3, the agreement between the measured and the calculated values of the thermal power generated during electrolysis is fully satisfactory. The 15 experiments give an average deviation of 0.04 per cent and a standard deviation of 0.14 per cent. The largest contribution to the uncertainty probably stems from the measurements of evolved gases. In all the experiments the current efficiency was within uncertainty limits 100 per cent indicating that recombination of H z (or D2) and 02 was insignificant. In conclusion, we find the described calorimeter suitable for accurate and convenient measurements of electrolysis processes, also for cases where gases are evolCeed. The results of our experiments with a palladium cathode and D 2 0 with LiOD as electrolyte solution do not show any abnormal evolution of thermal power. The financial support from the Swedish Natural Science Research Council is gratefully acknowledged.
104
G. OLOFSSON, I. WADS(3, AND L. EBERSON
REFERENCES 1. Kuhn, A. T.; Shams E1 Din, A. M. Surface Techn. 1983, 20, 55. 2. Sherfey, J. M.; Brenner, A. J. Electrochem. Soc. 1958, 105, 665. 3. Fleischmann, M.; Pons, S.; Hawkins, M. J. Electroanal. Chem. 1989, 261, 30t; erratum 1989, 263, 187. 4. Fleischmann, M.; Pons, S.; Andersson, M. W.; Li, L. J.; Hawkins, M. J. ElectroanaL Chem. 1990, 287, 293. 5. Lewis, N. S.; Barnes, C. A.; Heben, M. J.; Kumar, A.; Lunt, S. R.; McManis, G. E.; Miskelly, G. M.; Penner, R. M.; Sailor, M. J.; Santangelo, P. G.; Shreve, G. A.; Tufts, B. J.; Youngquist, M. G.; Kavanagh, R. W.; Kellogg, S. E., Vogelaar, R. B.; Wang, T. R.; Kondrat, R.; New, R. Nature 1989, 340, 525. 6. Williams, D. E.; Findlay, D. J. S.; Craston, D. H.; Sen6, M. R.; Bailey, M.; Croft, S.; Hooton, B. W.; Jones, C. P.; Kucernak, A. R. J.; Mason, J. A.; Taylor, R. I. Nature 1989, 342, 375. 7. Miskelly, G. M.; Heben, M. J.; Kumar, A.; Penner, R. M.; Sailor, M. J.; Lewis, N. S. Science 1989, 246, 793. 8. Cunnane, V. J.; Scannell, R. A.; Schiffrin, D. J. J. Electroanal. Chem. 1989, 269, 163. 9. N/irger, U.; Hayden, M. E.; Booth, J. L.; Hardy, W. N.; Whitehead, L. A.; Carolan, J. F.; Balzarini, D. A.; Wishnow, E. H.; Blake, C. C. Rev. Sci. Instrum. 1990, 61, 1504. 10. CRC Handbook of Chemistry and Physics, 50th Ed. CRC Press: Cleveland. 1969. 11. Tables of Physical and Chemical Constants. 14th Ed. Longman: London. 1973. 12. Linke, W. F. Solubilities. Inorganic and Metal-Organic Compounds. Vol. I, 4th Ed. American Chemical Society. 1958. 13. Linke, W. F. Solubilities. Inorganic and Metal-Organic Compounds. Vol. II, 4th Ed. American Chemical Society. 1965. 14. CODATA Key Values for Thermodynamics. Cox, J. D.; Wagman, D. D.; Medvedev, V. A.: editors. Hemisphere: New York. 1989. 15. The NBS Tables of Chemical Thermodynamic Properties. J. Phys. Chem. Ref. Data 1982, 11, supplement No. 2.
Design and testing of a c a l o r i m e t e r f o r m e a s u r e m e n t s on electrochemical reactions w i t h gas evolution GERD OLOFSSON, a INGEMAR WADSO, a and LENNART EBERSON b Thermochemistry ~ or Organic Chemistry 3, b Chemieal Center, University of Lund, P.O. Box 124, S-221 O0 Lund, Sweden (Received 17 September 1990)
A heat-conduction calorimeter suitable for the study of electrochemical processes has been constructed. The reaction vesselof 25 cm3 volume was fitted with electrodes, calibration heater, stirrer, thermocouple, and thin Teflon tube for addition of liquid. Gas evolved could leave the vessel through a gas-outlet tube. Good thermal conductivity was achieved by mounting nine thermocouple plates between the vessel and the surrounding heat sink and by letting the heat sink be in good thermal contact with the thermostat bath. The calorimeter was tested by measurements on the electrolysis of water and of heavy water and found to have an accuracy better than 0.2 per cent. I
1. Introduction Calorimetry is one of the most important methods for obtaining precise measurements on chemical systems and constitutes the basis of most of our knowledge of energy relations in chemistry. In one area, electrochemistry, the use of calorimetric methods ha~ been surprisingly scarce as attested to in a review from 1983, (1) although the large potential was demonstrated already in 1958 by Sherfey and Brenner. (2) So far, calorimetry appears to have been essentially limited to corrosion studies at open circuit and measurements of electrochemical Peltier effects at closed circuit. Reasons for this state of affairs are presumably perceived difficulties in correcting for non-chemical effects in the electrochemical cell, such as heat evolution caused by overpotential, resistive films on the electrode(s), and ohmic resistance in the cell. These are, however, factors that could be easily managed by using a high-precision calorimeter taking into account the special experimental features of the electrochemical method, of course in combination with the rigorous planning needed in all thermochemical experimentation. The set of phenomena that collectively appear under the concept of "electrochemically induced cold fusion ''(3'4) has recently focused interest on electrochemical calorimetry. Independently of whether one has faith in the concept or not, the hypothesis has to be thoroughly tested and accurate calorimetry is necessary. The 0021-9614/91/010095 + 10 $02.00/0
© 1991 Academic Press Limited
96
G. OLOFSSON, I. WADSO, AND L. EBERSON
experimental situation is challenging: electrolysis of heavy water must be performed at high current densities (up to 1 A. c m - 2 o r even higher), with strong gas evolution and under very long uninterrupted periods (as long as 103 h or more), and the electrical and thermal powers must be measured with a high accuracy to make possible a proper evaluation of the energy balance of the electrochemical cell. The original measurements were made using simple Dewar-vessel calorimeters where temperature differences between the interior vessel and the surrounding thermostat bath were measured33'4) Subsequent studies have indicated that this type of calorimeter may be difficult to calibrate accurately enough, primarily due to the large temperature increases caused by large power input. (5 8) Calorimeters working under essentially isothermal conditions appear to be more suitable for the present application. In an isothermal calorimeter the thermal power generated in the cell is dissipated to a surrounding heat sink and the heat flow is monitored. In the recent calorimeter design presented by Nfirger et al., ~9) the temperature increase of water flowing through a copper heat sink was measured. The primary aim was to construct an instrument for measurements of thermal powers generated during electrolysis of heavy water using a palladium cathode. The calorimeter could be used for power inputs up to 25 W and a device for the catalytic recombination of H 2 o r D 2 and 02 generated during electrolysis was built into the calorimeter vessel. We have found it of interest to develop a heat-conduction calorimeter for accurate measurements of the thermal power in electrochemical cells applicable to processes both with and without gas evolution. We report here a thermopile heat-conduction calorimeter with a 25 cm 3 reaction vessel. The calorimeter has been tested by measurements of the electrolysis of water and of heavy water and found to have an accuracy better than 0.2 per cent. The calorimeter is suitable for experimental periods of very long duration. With minor modifications, especially introduction of a membrane separating the cathode and anode compartments, the instrument should be useful for measuring enthalpies of reaction for electrode processes in general.
2. Instrument The calorimeter used was of the heat-conduction type. In such instruments, heat released in the reaction vessel is conducted to a surrounding heat sink, usually a metal block, by use of a thermopile positioned between the vessel and the heat sink. In the ideal case, the relation between the thermopile potential U and the thermal power P released in the vessel is given by the Tian equation: P = z'(U+v'dU/dt),
~
(1)
where e is a constant determined by calibration and ~ is the time constant for the instrument. If the thermal power is constant, equation (i) reduces to P = e.U.
(2)
For heat-conduction calorimeters the calibration constant ~ is nor affected by the heat capacity of the vessel or its content. This is in-c0ntrast to adiabatic or
C A L O R I M E T E R FOR E L E C T R O C H E M I C A L REACTIONS
I
97
Ie 1 cm
F I G U R E 1. Electrolysis calorimeter cell fitted with stirrer d, calibration heater e, cathode h, anode i, stainless-steel tube j containing a thermocouple for measurement of temperature in the cell, and a Teflon tube g for addition of liquid. Gas evolved could leave the cell through the central tube f.
quasiadiabatic calorimeters for which the constant is proportional to the heat capacity. This is of particular importance for the present type of application where the liquid electrolysed is partly decomposed into gases which will leave the calorimetric system during the measurement. The design of the instrument is shown in figures 1 to 3. The electrolysis calorimeter vessel, figure 1, consisted of a 25 cm 3 can c and a lid a fitted with several inserts: d, e, g, h, i, a n d j. The can and the lid were made from acid-proof steel with the inner surfaces coated with a thin layer of Teflon. An O-ring b together with a lock ring, provided a Seal between the lid and the can. There were also O-ring Seals between the inserts and the lid. The anode i consisted of a 1 mm platinum wire forming a spiral ,¢ around the cylindrical cathode h (d = 4 mm, l = 13 mm) made from acid-proof steel or palladium. The cylinder was fastened to a 1 mm wire, made from the same material. The 50 f~ calibration heater e made from manganin wire was inserted into a Teflon tube partly filled with paraffin oil. A copper-to-constantan thermocouple
98
G. OLOFSSON, I. WADSO, AND L. EBERSON
FIGURE 2. Simplified picture of calorimetric vessel showing capsule containing stirrer motor a, stainless-steel tube c surrounding the stirrer shaft and supporting the stirrer motor, gas outlet tubed connected to PVC-tube h, thermocouple leads e, thermal radiation shields f, and calorimeter cell g.
mounted in a Teflon-coated stainless-steel tube j was used to measure the temperature of the cell content. The thin Teflon tube g was used to inject liqu'id to the vessel in order to compensate for water decomposed during the electrolysis process. Gas evolved could leave the vessel through the central tube f (i.d. 4 mm) and a stirrer d made from "18 carat" gold was positioned off centre. Figure 2 shows a simplified picture of the calorimetric vessel. Above the electrolysis cell g is shown the stainless-steel tube c which surrounded the stirrer shaft and supported the stirring motor, encapsulated by a plastic tube a; the gas outlet tube d;
CALORIMETER FOR ELECTROCHEMICAL REACTIONS
~
C
H
m
r
99
/
/
/
/
/
A B FIGURE 3. Schematic picture of the calorimeter without its vessel. A, horizontal section; and B, vertical section at Y--Y. Lid a; steel tube b; outer steel can c; heat sink consisting of aluminium blocks d, e, and f; thermocouple plates i and vessel holder h.
and the thermocouple leads e. F o r clarity the Teflon tube used for injection of liquid (figure 1, g) is omitted. The gas-outlet tube was connected to a gas burette by means of a thick-walled P V C tube h (i.d. 3 mm, o.d. 7.7 mm). The a l u m i n i u m disks f serve as thermal-radiation shields. Figure 3 shows schematically vertical and horizontal sections t h r o u g h the calorimeter without its vessel. The vessel holder h, m a d e f r o m a squared stainlesssteel rod (60 m m x 30 m m × 30 mm), had a vertical hole (d = 27.5 mm) giving a close fit for the vessel. The vessel holder was surrounded by four aluminium rods e whose outside surfaces were r o u n d e d to fit into a circular cylinder with a radius of
100
G. OLOFSSON, I. WADSO, AND L. EBERSON
34 mm. Eight thermocouple plates i (Melchor CP-71, Trenton, New Jersey) were positioned between the fiat inner surfaces of the rods and the vessel holder. Two cylindrical aluminium blocks d and f with a diameter of 69 mm were positioned above and below the vessel-holder assembly. A ninth thermocouple plate was placed between the bottom of the vessel holder and block f and all the plates were connected in series. The three block units d, e, and f were bolted together to form a cylinder which fitted closely inside the steel can c. The can was immersed in a thermostatted water bath which thus would serve as a direct extension of the metal-block heat sink. The can c and its lid with upper parts were from the microcalorimeter TAM 2277-205 {ThermoMetric (LKB) Jgrf/illa, Sweden} and the instrument would thus fit into the thermostat bath of ThermoMetric's 4-channel microcalorimeter system. However, in the measurements reported here, we used a slightly less stable bath, approximately -t-0.3.10-3K during 20 h. Most heat-conduction calorimeters are used for microcalorimetric measurements and it is then usually desirable to have an isolating air gap between the heat sink and the thermostatic bath which will increase the stability of the instrument. But in the present case, where the experiments were carried out at high powers (up to 1 W or higher), an isolated heat sink would have had a significant temperature drift for a long time until it reached a steady-state value several kelvins above the temperature of the bath. For this reason, it was important to have good thermal conductivity also between the vessel and the heat sink. This was provided by the nine thermocouple plates. The time constant z for the present calorimeter charged with 21 cm 3 of water was about 250 s. Potential signals from the calorimeter were amplified using a Keithley 150 B ammeter. By the use of a microprocessor the signals were read once a second, integrated over 300 s periods, and printed out. The input of electric power during electrolysis was determined by measuring the current I¢~n and potential liceH using two Hewlett-Packard 3478A multimeters. Measurements were made once a second and the product/'cell" Io~1 was collected using a Hewlett-Packard 85 desk calculator. The power inputs during the electrical calibrations were measured in the same way. Corrections were applied for the heat effects in leads to heater and electrodes. The extent of reaction was determined by measuring the volume of gases evolved during electrolysis. The gas-collecting system consisted of a water-filled glass vessel with a glass inlet tube over which a gas burette was mounted. In the experiments with heavy water, deuterium oxide with isotopic purity better than 99.8 moles per cent (CibaGeigy) was used without further purification. The 0.1 mol" dm -3 L i O D solution was prepared by dissolving lithium metal in deuterium oxide in an argon atmosphere.
3. Results The performance of the calorimeter was tested by measurements of the electrolysis of water and of heavy water. The chemical process in the case of water is H20(1) = H2(g) + ½02(g).
(3)
CALORIMETER FOR ELECTROCHEMICALREACTIONS
101
It takes place under isothermal conditions and the power balance during electrolysis is given by P~n = d(AH)/dt + dQ/dt. (4) The electrical power input is denoted Pceu, AH is the enthalpy of dissociation of water (3), and dQ/dt is the thermal power. All three properties were measured independently. The accuracy of the calorimeter could be established by comparing the calorimetrically measured values of dQ/dt with values calculated as the difference between Pcell and d(AH)/dt. The response of the thermocouple plates was determined from electrical calibrations using the internal heater. Values of the calibration constant e, calculated as the ratio between the electrical power input Pc,~ and the amplified response of the thermocouple plates, were determined for Peal varying between 20 mW and l l00 roW. Results of 10 experiments showed s to be linear within 0.2 per cent over this power range. The linearity of the amplifier may well have been the limiting factor. The amount of liquid in the calorimeter vessel did not influence the calibration constant as values of e determined with 20.0 cm a and 23.0 cm a of liquid agreed to within 0.01 per cent (the number of experiments was 10 and P~I = 0.5 W and 0.9 W). The rate of dissociation of water was calculated from the time measured to fill the gas burette. The atmospheric pressure, the temperature, and the hydrostatic pressure of water were measured. The water filling the burette before the gas was collected was assumed to be in equilibrium with air. When being displaced, the water was presupposed to have become saturated with H E (or D2) and 0 2 and the gas mixture to have been saturated with water vapour. Values of the vapour pressure of water were taken from reference 10 and for heavy' water from reference l l. Values of Bunsen coefficients for H2,(12) and 0 2 , (13) were used to calculate the amounts of dissolved gases. The solubility of D2 in water was assumed to be the same as for H2. The pressure in the system was observed to decrease slowly after the electrolysis was stopped. The pressure drop was ascribed to the diffusion of H 2 (or D2) through the plastic tube. From the amount of water sucked into the plastic tube, the gas diffusion was estimated to be 0.089 mm 3- s -x. The collected gas was treated as a mixture of perfect gases when calculating the amount of decomposed water. Figure 4 shows results from a typical series of measurements with an initial electrical calibration experiment followed by three electrolysis experiments conducted at different electrical powers. Results of measurements on the electrolysis of water using 21.0cm 3 of 0.1 m o l ' d m 3 N a O H solution are summarized in table 1. The value AH ° = (285.83+0.040)kJ.mol 1 was used when calculating d(AH)/dt. ~14) A small correction was applied for the evaporation of water to give a saturated gas mixture leaving the vessel. The time to fill the gas burette of (103.42___0.08)cm 3 volume varied between 37min for the highest power and 81.5 min for the lowest. The average rate of reaction and accordingly d(AH)/dt was calculated by dividing the amount of decomposed water b y the time. Values of Peel1 and dQ/dt shown in columns 1 and 3 of table 1 were the average values of the electrical power input and the thermal power, calculated from values integrated over the time of the gas collections. The potential over the cell was 3.65 V at Peel~= 0.886 W and decreased to
G. OLOFSSON, I. WADS(~, AND L. EBERSON
102 I
I
I
I
I
I
I
I
1
0.6
0.4
i.
3
"1
0.2
0
I
0
I
I
2
I
4
I
I
6
t/h
I
I
8
FIGURE 4. Recorder trace of the power P at time t during a series of experiments on the electrolysis of heavy water using the stainless-steel cathode. After an initial electrical calibration three gas collections were made during time periods 1, 2, and 3 indicated in the figure (see the last three experiments in table 2). The current was 0.246 A, 0.153 A, and 0.092 A, respectively.
3.08 V at the lowest power input. The temperature of the thermostatted bath was 298.15 K and the temperature increase in the cell was less than 0.5 K even at the highest power. The results of a corresponding series of measurements on heavy water using 20.0 cm 3 of 0.1 mol- dm 3 L i O D solution are summarized in table 2. The value AH m = 294.60 k J- m o l - 1 was used for the calculation of the enthalpy of dissociation of D 2 0 . (15) The stainless-steel cathode was then replaced by an electrode of palladium metal, the vessel was charged with 21.0cm 3 of freshly prepared 0.1 mol. dm -3 L i O D solution, and an electrolysis experiment was run continuously for 83 h. After about 6 h, the conditions in the cell became stable. The results shown
TABLE 1. Results of measurements of the electrical power input Pcell, the chemical power d(AH)/dt, and the thermal power dQ/dt during electrolysis of water (0.1 mol-dm -3 NaOH) using the stainless-steel cathode ,~
P,~,,
d(AH)/dt
dQ/dt
A {(dQ/dt)/W }
W
W
W
(expt-calc.)
0.88633 0.88915 0.52498 0.34302
0.36970 0.36946 0.23950 0.16690
0.51717 0.52031 0.28639 0.17578
0.00054 0.00062 0.00091 -0.00034
103
CALORIMETER FOR ELECTROCHEMICAL REACTIONS
TABLE 2. Electrolysis of heavy water (0.1 mol. dm -3 LiOD) using the stainless-steel cathode (compare table l)
P¢¢,,
d(AH)/dt
dQ/dt
A {(dQ/dt)/W }
W
W
W
(expt-calc.)
1.01686 1.01904 1.02627 0.55709 0.30512
0.37907 0.37860 0.37878 0.23531 0.14191
0.63891 0.64027 0.64704 0.32194 0.16347
0.00112 -0.00016 -0.00045 0.00016 0.00026
TABLE 3, Electrolysis of heavy water (0.1 mol.dm -3 LiOD) using the palladium cathode (compare table 1)
P~l,
d(AH)/dt
dQ/dt
A ((dQ/dt)/W }
W
W
W
(expt-calc.)
0.49462 0,49399 1.07591 0,49172 0,49887 0,48797
0.20839 0.20799 0.38935 0.20580 0.20842 0.20331
0.28590 0.28578 0.68776 0.28565 0.29016* 0.28415
-0.00033 -0.00022 0.00120 --0.00027 -0.00029 -0.00051
in table 3 summarize measurements made during the following 3 d. The vessel was refilled once on each day by injection of 1.4 cm 3 of the LiOD solution. In one of the experiments, indicated by an asterisk in the table, extra electrical power was supplied to the calibration heater. This amount, 0.43446 W, has been subtracted from the total measured thermal power to give the value of dQ/dt shown in column 3. The temperature increase in the vessel was 0.6 K in this experiment. The cell potentials were 3.68 V or 4.30 V and they were stable to within a few mV. The corresponding current densities were 0.069 A. cm -2 and 0.129 A" cm -2. Using different equipment for power supply it would have been possible to extend the measurements to higher power levels. As can be seen from tables 1 to 3, the agreement between the measured and the calculated values of the thermal power generated during electrolysis is fully satisfactory. The 15 experiments give an average deviation of 0.04 per cent and a standard deviation of 0.14 per cent. The largest contribution to the uncertainty probably stems from the measurements of evolved gases. In all the experiments the current efficiency was within uncertainty limits 100 per cent indicating that recombination of H z (or D2) and 02 was insignificant. In conclusion, we find the described calorimeter suitable for accurate and convenient measurements of electrolysis processes, also for cases where gases are evolCeed. The results of our experiments with a palladium cathode and D 2 0 with LiOD as electrolyte solution do not show any abnormal evolution of thermal power. The financial support from the Swedish Natural Science Research Council is gratefully acknowledged.
104
G. OLOFSSON, I. WADS(3, AND L. EBERSON
REFERENCES 1. Kuhn, A. T.; Shams E1 Din, A. M. Surface Techn. 1983, 20, 55. 2. Sherfey, J. M.; Brenner, A. J. Electrochem. Soc. 1958, 105, 665. 3. Fleischmann, M.; Pons, S.; Hawkins, M. J. Electroanal. Chem. 1989, 261, 30t; erratum 1989, 263, 187. 4. Fleischmann, M.; Pons, S.; Andersson, M. W.; Li, L. J.; Hawkins, M. J. ElectroanaL Chem. 1990, 287, 293. 5. Lewis, N. S.; Barnes, C. A.; Heben, M. J.; Kumar, A.; Lunt, S. R.; McManis, G. E.; Miskelly, G. M.; Penner, R. M.; Sailor, M. J.; Santangelo, P. G.; Shreve, G. A.; Tufts, B. J.; Youngquist, M. G.; Kavanagh, R. W.; Kellogg, S. E., Vogelaar, R. B.; Wang, T. R.; Kondrat, R.; New, R. Nature 1989, 340, 525. 6. Williams, D. E.; Findlay, D. J. S.; Craston, D. H.; Sen6, M. R.; Bailey, M.; Croft, S.; Hooton, B. W.; Jones, C. P.; Kucernak, A. R. J.; Mason, J. A.; Taylor, R. I. Nature 1989, 342, 375. 7. Miskelly, G. M.; Heben, M. J.; Kumar, A.; Penner, R. M.; Sailor, M. J.; Lewis, N. S. Science 1989, 246, 793. 8. Cunnane, V. J.; Scannell, R. A.; Schiffrin, D. J. J. Electroanal. Chem. 1989, 269, 163. 9. N/irger, U.; Hayden, M. E.; Booth, J. L.; Hardy, W. N.; Whitehead, L. A.; Carolan, J. F.; Balzarini, D. A.; Wishnow, E. H.; Blake, C. C. Rev. Sci. Instrum. 1990, 61, 1504. 10. CRC Handbook of Chemistry and Physics, 50th Ed. CRC Press: Cleveland. 1969. 11. Tables of Physical and Chemical Constants. 14th Ed. Longman: London. 1973. 12. Linke, W. F. Solubilities. Inorganic and Metal-Organic Compounds. Vol. I, 4th Ed. American Chemical Society. 1958. 13. Linke, W. F. Solubilities. Inorganic and Metal-Organic Compounds. Vol. II, 4th Ed. American Chemical Society. 1965. 14. CODATA Key Values for Thermodynamics. Cox, J. D.; Wagman, D. D.; Medvedev, V. A.: editors. Hemisphere: New York. 1989. 15. The NBS Tables of Chemical Thermodynamic Properties. J. Phys. Chem. Ref. Data 1982, 11, supplement No. 2.