Electrochemical Syntheses, XVII [1] Ozone Synthesis by Electrolysis of Water
Heinz P. Fritz*, Jordanis C. G. Thanos, and Dietrich W. Wabner Institut für Anorganische Chemie, Arbeitsgruppe Angewandte Elektrochemie, Technische Universität München, Lichtenbergstraße 4, D-8046 Garching
Z. Naturforsch. 34b, 1617-1627 (1979); received August 16, 1979
Ozone, Electrosynthesis, Lead Dioxide, Sulphate, Phosphate
Dimensionally stable lead dioxide anodes on a titanium substrate [2, 3] were tested for application in ozone electrosynthesis in a variety of aqueous electrolytes. At room tem- perature an ozone concentration of up to 13% by weight in the anodic gas composed of O3 and O2 is attained by using an aqueous phosphate electrolyte.
Introduction results and those of the Japanese workers is possibly In 1840 Schönbein on electrolyzing aqueous due to a special pretreatment of their anodes. sulfuric acid obtained, described and named ozone Efforts for ozone synthesis in alkaline and neutral [4]. Up to-day several studies on a high-yield aqueous solutions have also been made without synthesis of O3 by water electrolysis were published. much success [13]. In the course of our study we They have as special features the use of low tem- ended up using about neutral pH values with peratures (e.g. — 30 °C to — 56 °C [5, 6], also satisfactory results, since at very low pH, as well as —10 °C [7]), or high current densities [8-10], that is, at high current densities and low temperatures they consider physical parameters when treating excessive electrode corrosion was observed. Without concentrated aqueous H2SO4, H3PO4 or HCIO4. considering any detailed mechanism, the main Russian workers reported a room temperature reaction of water electrolysis process [9]; they electrolyze 70% H3PO4 with cur- 2 H20 ^ 02 + 4 H+ + 4 e- rent densities greater than 800 mA cm-2 and quote shows, that high proton concentrations arise at the 03-concentrations up to 6.2% v/v at 20 to 40 °C, anode and that the application of low pH in the using Pb02 deposited on platinized titanium as electrolyte will rather hinder the 03-formation. anode. Nishimura et al. claim patents for an Therefore, the concept of our study was to use about electrochemical ozone production at Pb02/Pb anodes neutral pH buffered electrolytes, Pb02/Ti anodes using the rather low current densities of 20 mA cm-2, [2, 3], and cathodes with low H2-overpotential, as obtaining a current efficiency of up to 30%, using stainless steel, titanium etc. A plausible explanation a saturated Se02 solution in water as anolyte why the previous efforts [13] in neutral bulk electro- [11, 12]. We tried to reproduce this process with a lytes were not very successful might be the use of c.d. of 100 mA cm-2 using Pb02/Ti anodes in such non-buffered electrolytes and the inapropriate an electrolyte for 2 h without succeeding to obtain choice of anodes (e.g. Pt), which show a considerably such amazing results. We also used lead anodes lower 02-overpotential than Pb02 [14, 10]. potentiostatically preanodized in H2SO4 (at 1,9 V vs SCE for 30 min). After 2 h a visible brown layer of Pb02 on the lead surface was formed. Current Experimental Procedure and Apparatus densities between 30 and 50 mA cm-2 were used, Electrochemical cells: Our measurements were however, with no improved results within 4 h of carried out in four different synthetic cells, described briefly below: experimentation. The discrepancy between our a) A cylindrical vessel with a cooling jacket and a capacity of 1 1 contained the electrolyte. A glass cylinder was coaxially immersed in the vessel. The inner cylinder was open at the lower end, dipped in * Reprint requests to Prof. Dr. H. P. Fritz. the electrolyte and was kept at a distance of about 0340-5087/79/1200-1617/$ 01.00/0 1 cm from the bottom of the container vessel. The 1618 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water other end was sealed with a teflon stopper which potentiostat Wenking, 68 TS 10, with a Tacussel, carried a reference electrode (dynamic hydrogen G9 TP voltage generator. electrode) [10], a thermometer, the anode, a capil- lary tube for the outflow of the evolved gas and a Temperature regulation was done by a ULT teflon tube for pipetting electrolyte solution for Kryostat UK-50L or a Lauda, Tisch-Kryomat, chemical analysis. Moreover, the possibility was Type TK 30. provided for introducing a small cylindrical cell Analysis of O3 was standardized by application of with a wire Pt cathode, a sheet Pt anode and a a Fischer Model 501 ozonizer. reference electrode in the immediate neighbourhood of the main wire cathode. In this auxilliary cell the Chemicals used for synthetic or analytic purposes (dissolved) O3 could be cathodically reduced and were mostly of analytical grade (Merck, Darmstadt) the potentiostatically obtained limiting current except 0s04, As203 (p.A.), and (NH4)2Fe2(S04)3 served as a measure of the time dependence of the (p.A.) (Riedel-de Haen), titanium sheet metal concentration of the dissolved O3. The limiting (Contimet), surfactant FC 98 (3M) and oxygen current depended mainly on the hydrodynamic flow (pure) (Linde-AG.). Water was distilled. of the electrolyte past the cathode, on the time average being kept constant through moderate Ozone analyses in the anodic gas were done iodo- stirring of the electrolyte by a magnetic stirrer at metrically according to Boelter et al. [16]: the bottom of the container vessel. The method of The electrochemical ly generated anodic gas is quantitative analysis of O3 by a cathodic limiting bubbled through water for a time sufficient to sweep current is described in ref. [15]. out the gas originally present in the "dead volume" The above method allowed a continuous record- above the anolyte and to obtain a homogenous O3 ing, however, an exact analysis of the evolved gas distribution above the anolyte. Then the gas flow is was made as described later. The anode consisted of directed into a gas burrette by means of a two-way tap. This burrette contains a weakly alkaline 1MKI a round Pb02/Ti disc covered on one face by a teflon layer. The cathode was a titanium sheet. The solution and a nearly constant overpressure for the working anode and the catholyte were separated by gas to be collected. As the gas is led into the a sintered glass frit DIN 2 attached to the wall of burrette in form of small bubbles, ozone is reduced the inner cylinder. by iodide and the total volume of the collected oxygen is measured. The amount of O3 is calculated The uncovered anode side was facing the sintered after acidification of the iodide/iodine solution and glass, these and the cathode were parallel to each titration with sodium thiosulphate and starch other insuring a homogeneous current density over solution as indicator. The quality of this method the anodic electrode. Anolyte and catholyte could was checked by comparison with the calibration be stirred, preferentially only moderately in order curve of a commercial ozonizer. to avoid noticeable mixing of the gas bubbles of the anodic and the cathodic compartments. The It has to be noted that if the concentration of the advantage of this cell was the easy interchange- KI solution is less than approximately 0.3 M and ability of the anodes. the gas to be analysed is bubbled through with rates as high as 301/h, the analysis gives too high ozone b) The second cell was similar to a), however, values. inside the container vessel, instead of a glass Ozone dissolved in neutral aqueous solutions was cylinder, a special glass and teflon insert was placed analysed by the DPD method described by Palin to accomodate an anode with internal cooling and [17]. In weak acidic or alkaline solutions a previous larger dimensions. neutralisation is necessary. Alternatively the dis- c) This cell was also similar to a), however, the solved O3 was driven out of the solution by a stream anodic compartment was sealed to allow the use of of nitrogen and analyzed in the gas phase as de- an anolyte differing from the catholyte. Anodes scribed above. For strong acidic solutions this with and without internal cooling could be used. method proved to be the only appropriate one. Determination of O3 in solution by direct mixing d) The fourth cell had a capacity of 200 ml and with a weak alkaline potassium iodide solution and could accomodate large area planar anodes having subsequent acidification produced unsatisfactory sintered glasses as well as counter electrodes on results. either side of the anode faces. Because of the compact construction of the cell very low tem- The electrodes were studied after the experiments peratures were accessible by external and internal a) by cyclic voltammetry, using potentiostats and cooling. generators mentioned above and recorded on an x-y The results of measurements in all different cells recorder and b) by X-ray diffraction using a were reproducible to within less than 5%. Siemens D 500 diffractometer. Corrosion material from the electrodes was studied in KBr pellets by Power supplies were highly stabilized, Phillips IR spectroscopy, using a Perkin-Elmer 577/spectro- PE 1516, 0-30 V, 20 A, or Phillips PE 1512, or a photometer. 1619 H. P. Fritz et al. •Ozon e Synthesis by Electrolysis of Water
The anodes used consisted of a Pb02 layer the electrolyte during electrosynthesis. The different deposited on a titanium substrate according to factors are interrelated; e.g. for a fresh electrode [2, 3, 10], from an acidic Pb(N03)2 solution. The the current density had no considerable influence on electrodeposition took place at different tempera- the O3 yield, while a remarkable improvement with tures and current densities. The following notation time was observed on an older, used electrode, if will be adopted to describe the anodes: e.g. Pb02 stronger anodic polarization was applied. For this (70/40) for an electrodeposition carried out in a reason some of the factors will be discussed at some bath of 70 °C with anodic polarization by a current stages simultaneously. density of 40 mA cm-2. First, it can be seen that if an already used anode is employed for ozone electrosynthesis no definite Experimental Results conclusion about the yield of the product may be made. In Tables la, lb and Ic the time-dependence In the beginning a thorough study of the ozone of the O3 yield is shown. At regular intervals the evolution on Pb02/Ti anodes in sulphuric acid was current density was changed and the response on made. The ozone yield was found to depend on: the O3 yield was observed. On the Table the times a) the history of the electrodes b) the current of the current changes and of the gas analyses are density at gas evolution c) the time of operation of denoted. In Table la the test of a Pb02 (70/40) the electrode d) the concentration of the electrolyte electrode is shown after previous operation in a e) the temperature of the bath at which the I.6NH2SO4 electrolyte (T=6±2°C,i=200mAcm-2) electrodes were made, and f) the temperature of for 9 h, washing in water and subsequent drying on air. The conditions of operation are shown at the Table I. Current density-03 yield [% v/v] relation. Table. The current density was changed at several time intervals and after a certain time elapsed the gaseous Then the electrode was removed from the elec- concentration was determined. Detailed description in trolyte, washed, sanded and then again placed in text. the electrolyte; the results are shown in Table lb. A remarkable improvement may be observed. The Time i O3 concentration electrode was further polarized for 10 h (T = 6 ± [min] [mA cm~2] [% v/v] 2 °C, i = 200 mA cm-2) and then washed and dried _ in air for 13 h. Further experiments according to the 0 100 procedure described above, produced the results of 228 100 0.75 Table I c, where a decrease of the O3 yield may be ob- - 238 200 -2 282 200 1.25 served. It should be noted that with i > 300mA cm 384 200 1.25 a considerable wear of the electrode especially at 390 300 - the edges occurs. 455 300 2.00 The influence of the c.d. on a newly deposited 470 300 1.85 electrode Pb02 (87/40) in a 1.6 N H2S04 bath 476 400 - (T = 6 ± 2 °C) is shown in Table II. In the previous 510 400 2.25 experiments the bath had been kept at 6 °C because 540 400 2.25 as shown later the time dependence of the O3 yield 0 300 - was more pronounced for electrolyses carried out 48 300 3.00 at higher temperatures. Moreover, in more con- 55 400 - centrated acid solutions, the O3 concentration 102 400 3.25 decreased remarkably with time. Characteristically 110 500 - 180 500 3.30 Table II. Current density O3 yield [% v/v] relation. 0 100 _ 72 100 0.80 Detailed description in the text. 120 100 0.80 180 100 0.90 Time i O3 concentration 265 100 0.90 [min] [mA cm-2] [% v/v/] 275 200 - 320 200 1.35 0 200 _ 350 200 1.50 90 200 4.25 400 200 1.40 97 400 - 450 200 1.55 170 400 4.55 - 460 300 - 175 600 520 300 1.75 195 600 4.15 590 300 2.05 200 600 4.45 - 600 400 - 205 800 640 400 2.35 220 800 4.55 690 400 2.25 230 800 4.50 - 697 500 - 234 200 725 500 2.80 254 200 4.15 1620 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water
for a Pb02 (70/40) electrode, electrolysis in 4 M H2SO4 at 22 °C after 4 h yielded an anodic gas with an ozone content of only 0.4% v/v and after 7 h no ozone could be detected at all. With an .internally cooled anode and an anolytic temperature of 3 °C, in a 8 M H2SO4 the results of Table III were
Table III. Current density-03 yield [% v/v] relation. Table description as for Table I.
Time i O3 concentration [min] [mA cm-2] [% v/v] 0 100 90 100 2.05 100 200 — 240 200 3.60 250 300 — Fig. 2. The ozone yield as a function of time differs for 360 300 3.25 electrolytes carried out at different temperatures 368 400 - (i = 200 mA cm-2, Pb02 (70/40), 1.6 N H2S04); at 420 400 4.20 18 °C (circles), at 6.5 °C (crosses). 430 200 - 540 200 3.80 seen in Fig. 2 for two temperatures of the electro- lyte, viz. 6 °C and 18 °C. The general trend of the obtained. Heat evolution was so strong that the decrease of the O3 yield with time may be correlated coolant inside the electrode had to have a tempera- with the gradual lowering of the pH value next to ture of —10 °C. Then the electrode surface in the anode. The fact that the drop of the ozone yield reality had a temperature between 3 °C and —10 °C. was not so fast at lower temperature may plausibly At the low temperatures a continuous wear of the be explained either a) by a stronger stirring of the anode was observed at all current densities used in electrolyte due to a stronger natural convection 8 M H2SO4. In view of the observed changes with (because stronger temperatine gradients were pres- time and the better yields at lower H+ concentra- ent in the cell in order to achieve the wanted tem- tions, further experiments were made in moderately perature at the neighborhood of the anode), or b) by dilute acids and on freshly deposited lead dioxide. an increased corrosion rate of Pb02, observed at A series of Pb02 anodes (70/40) was studied under lower temperatures, and hence a partial regeneration anodic polarization in 1.6 N H2SO4 (i=200mAcm-2) of fresh anode surface. However, the observed at different cell temperatures. The yield of O3 after results may be as well due to other electrochemical 25 and 140 min of electrolysis were measured and dominant processes. the results are recorded in Fig. 1. The yield of ozone The preparation of the electrodes at various tem- as a function of time for two electrodes, prepared peratures effected a different time dependence of the under conditions similar to those of Fig. 1 may be
^ in in Z Z o o dPi °0.00 23.27 46.53 69.80 93.07 °0.00 6.98 13.96 20.9i 27.92 TEMPERATURE [°CJ TEMPERATURE [ ÖC] Fig. 3. Electrodes deposited at different temperatures Fig. 1. The effect of the temperature of the electrolyte in our standard Pb(NOs)2 bath exhibit different O3 on the amount of electrogenerated O3 as observed on producing abilities after a fixed period of electrolysis. Pb02 (70/40) anodes. Values after 25 (circles) and Electrolysis conditions: 1.6 N H2SO4, (i — 350mA cm-2, 140 min (crosses) (i = 200 mA cm-2). T = 6.5 °C ± 2 °C. 1621 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water O3 concentration in the gas phase for anodic polari- temperatures also was studied by X-ray diffraction. sations in 1.6 N H2SO4 (5 °C, i = 350 mA cm"*). The The peaks were identified from the A. S. T. M. cards yields were recorded after 105 min of electrolysis [20]. For both a-Pb02 and £-Pb02 reproducible and the values are plotted in Fig. 3. Two electrodes textures were observed. Although no detailed were prepared at each temperature. Practically no analyses were carried out, obviously the Pb02- effect on the observed yields seems to have the crystallites have a random orientation with respect current density at which the anodes were electrode- to planes perpendicular to the surface of the posited, Fig. 4. Here anodes deposited from a electrode, (but not with respect to planes parallel to the electrode surface), as evidenced by invariable X-ray spectra for two different positions of the electrodes in the diffractometer, interrelated to each other by a rotation of 22° about an axis perpendicu- lar to the surface of the electrode. For the electrodes deposited at different temperatures the relative •— in Z intensities of the observed peaks differ. However, o for the Pb02 (87/40) anodes all diffraction peaks due < t to a-Pb02 between d = 350 and d = 92,7 pm un- (v m exceptional ly seem to be smaller than for the I— Z electrodes Pb02 (70/40), Pb02 (52/40) or Pb02 UJ in ^ (32/40). More information will be given in a future O "" publication. o OV so Ozone Evolution °0.0 •1.9 83.8 12S.6 167.5 CUR. DENSITY -IOTA-1^7 from a Buffered Phosphate Electrolyte Fig. 4. Electrodes of Pb02 deposited at different Although older studies with salt solutions were current densities produce nearly the same amounts of not very successful [13], we observed a considerable O3, under otherwise constant physical and chemical increase of anodically produced ozone in an aqueous parameters. (1.6 N H2S04, i = 350 mA cm"2, T = 6 °C ± 2 °C). phosphate buffer compared to strongly acidic electrolytes. The ratio of the concentrations Pb(N03)2 bath at 70 °C were used in 1.6 N H2S04 of the two components will be symbolized by cooled down to 5 °C (i = 350 mA cm"2). To test the R=[KH2P04]/[K2HP04]. The total molar con- eventual effects of the electrode edges on O3 yields centration of phosphate, [phosphate], was 0.74 M titanium stretch metal was used as a Pb02 sub- strate ; no improvement of ozone yield was observed and R = 0.42. At 20 °C and after 16 h of electrolysis with such an anode. a 6.6% v/v concentration of O3 was obtained with It should be remarked that the IR spectrum of i = 200 mA cm-2. By changing to i = 100 mA cm-2 corroded material, which was simply washed in a constant concentration of 5.6% v/v was kept for distilled water, produced two moderately broad peaks at 1010 and 1130 cm-1, as well as two small 40 h of electrolysis in contrast to H2S04 solutions ones at 615 cm-1 and 630-1. After the corroded where a considerable decrease with time was material was washed for 2 min in water in an observed. The influence of variation of R with ultrasonic bath only the two peaks at 1130 cm-1 and constant phosphate concentration of 0.74 M on the 630 cm-1 remained. Here, characteristic vibrations yield of ozone was investigated and the results of of sulphates are assigned [18]. Furthermore, an anode electrically disconnected after use showed a analyses carried out after 90 min of electrolysis are rapid potential drop to about 2.2 V vs DHE fol- summarized in Table IV. The pH was measured lowed by a slower potential decrease towards 1.7 V. This phenomenon had been observed by other Table IV. Ozone yield in a buffer phosphate solution workers [19] on Pt electrodes and was attributed to for different values of R. Values from two different adsorption. Indeed, if the already used anode was experiments are quoted, [phosphate] = 0.74 M. washed and dried and subsequently tested by cyclic voltammetry in 1.6 N H2SO.4 on scanning in anodic R pH Conductivity O3 yield direction a peak at about 2.18 V vs DHE was [Ohm-1 cm-1 • 10-2] [% v/v] observed, (scan rate 50 mV sec-1) attributed to the oxidation of PbS04 to /?-Pb02. 0.19 7.6 7.87 6.2, 6.4 X-ray diffraction for such an electrode showed 0.42 7.15 7.29 6.6, 6.5 characteristic peaks [20] of a nearly randomly 1.05 6.72 6.68 6.0, 6.3 1.84 6.35 6.15 6.3,- oriented PbS04 crystallized on the lead dioxide. 5.85 4.68 6.7, 6.3 The surface of lead dioxide deposited at different 5.16 1622 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water with a glass electrode within ±0.15 units. The the decrease of the yield with temperature is conductivities of the electrolytes were measured at attributed to the deposited layer from the electro- 1 kHz and are tabulated in Table IV also. As is lyte on the anode. easily seen variation of R does not play any im- The influence of the c.d. on the amount of portant role in the measured yields, apart from the anodically evolved O3 was checked with a single fact that higher conductivities of the solutions are electrode in an electrolyte with [phosphate] = wanted for better heat dissipation. For a constant 2.22 M and R — 0.42, showing better conductivity value of R with variation of the total molar phos- and better heat dissipation at high currents. The phate concentration the yields observed at 20 °C results are given in Fig. 5b. The analyses were and 200 mA cm-2 are given in Table V with analyses carried out after successive increasings of c.d. at carried out after 120 min of electrolysis. It may be intervals of 1 h, the ozone concentrations are marked noticed that at low molar concentration the meas- with crosses. Then the electrolysis was stopped for ured yields were lower. This may be possibly 2 h and the process repeated. The new values are explained by the fact that a lower pH may occur indicated by circles. on the electrode surface when the buffer solution is not present in a concentration sufficient to compen- Fluoride and Sulphate Effects sate for the generated hydrogen ions. It is known from other workers that with binary electrolytes (H2SO4/HCIO4) [6] an increase in the O3 Temperature and Current Density Effects yield may be attained. Semchenko et al. [7] used a On account of the relative constancy of the ozone mixture of conc. HC104 and NaF (10 g/1) as electro- yield with time, we measured the O3 yields on one lyte for ozone electrosynthesis on lead dioxide electrode with internal cooling at different tem- deposited on platinized titanium with good results peratures of electrolyte. The temperature for each at —10 °C. The effect of fluoride ion in 4 M con- measurement was kept for two hours to equilibrate centration with platinum anodes in a perphosphate and then raised in steps from 3 to 23 °C as shown in synthesis was studied by Miller et al. [21]. The per- Fig. 5a ([phosphate] = 2.72 M, R = 0.42). At lower phosphate yields from electrooxidation of K2HPO4 temperatures the electrolyte solidified in the neigh- as a function of KF concentration as additional bourhood of the electrode. The current density was component in the electrolyte were reported by 100 mA cm-2 with an electrolyte having [phos- Tyurikova et al. [22]. However, the effect was phate] = 0.74 M and R = 0.42. Experiments at significant at concentrations greater than 1 M only. 8 and 25 °C showed no difference in yield and hence It was suggested that KF is actively involved in the perphosphate synthesis. On the other hand a fluoride influence on increasing the oxygen over- o CUR. DENSITY • 1 ö'/A M~2/ : So.o 116.3 232.7 3*9.0 465.3 potential on platinum during persulphate synthesis — at concentrations as low as 0.001 M was reported by _J O Smit and Hoogland [19]. > S x « * in 2 We added KF (2.5 mM or 1.4 mM) to a phosphate O buffer (R = 0.42, [phosphate] = 0.74 M), the c.d. «t * was 100 mA cm-2 and the temperature 19 ± 1 °C in fv after 25 or 85 min, respectively. Repeating the ex- periment after soaking the electrode in the elec- CO trolyte for 22 h before electrolysis 5.3 and 5.6% O3 —I v/v were obtained after 35 or 90 min, respectively. O 0g0 It should be noted that within the first 30 min the values are low because a relative large cell (11) was used and a steady state was achieved only later. Pb02 (70/40) electrode gave 4.5 and 5.3% 03 v/v DIN 2 a 2.6% v/v O3/O2 mixture from a commercial 1624 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water ozonizer was bubbled through the appropriate Table VI. Ozone concentration in aqueous electrolytes. solution for 5 min in which the O3 later was to be 03/02 gas was bubbled into the electrolyte and analyses were carried out after 5 or 20 min. determined. The solutions were left to stand for 5 and 20 min, respectively, and the O3 was analyzed. [mg O3/I] [mg O3/I] The results are summarized in Table VI. In agree- Solution after 5 min ± after 20 min ± (0.5) min ment with Fabjan [25] ozone in 1 N KOH rapidly (0.5) min decomposes. Other previous investigations [26] H20 14.2 10.8 1 N KOH 0.0 0.0 reported longer life times of 03 solution in alkaline 2 N H2S04 10.6 10.0 environments. It is well known that 03 decomposes 8 N H2S04 7.1 4.2 in presence of H202 [27]. Also, oxygen is known to 36 N H2S04 18.6 17.6 0.48 M KH2P04 9.4 7.9 be in equilibrium with H02~ in alkaline H202 on a 0.26 M K2HP04 platinum electrode [28]. This means that impurities O.IIMKH2PO4 8.6 - (like Pt species) in catalytic quantities may cause 0.26 M K2HP04 0.37 M KH2P04 8.6 - decomposition. However, a decomposition mecha- 0.37 M K2HP04 nism for O3 in absence of an impurity as catalyst 0.63 M KH2P04 9.6 - may not be excluded, e.g. by decomposition due to 0.27 M K2HP04 0.61 M KH2P04 7.4 - OH- [25]. 1.02 M K2HP04 It is interesting to note, that samples from different parts of the anolyte, yielded varying Potential Observations on the Gas Evolution concentrations of dissolved O3. These results were The potential drop across the lead dioxide- obtained for our undivided cell and are easy to electrolyte interphase was difficult to measure in accept in view of the convectional flows of the our experiments, especially for extended times. The liquid and the mixing of anolyte and catholyte at reason is that the non-linear time-varying resistance the lower parts of the cell, as well as the relatively at the lead dioxide-titanium interphase interfers low rate of gas evolution, insufficient to achieve with the actual measurements [29]. Moreover, an complete saturation with O3. A mean value obtained additional potential drop within the electrolyte, during the electrolysis of a buffer phosphate with between Luggin capillary and the surface of the R = 0.42, and [phosphate] = 0.74 M and no anode is difficult to be accurately accounted for additions of KF was 15 mg O3 I-1 electrolyte, while from impedance measurements due to a strong in the gas phase the volumetric O3 concentration influence of the "titanium oxide" layer between lead had been 6.3% v/v. From the values in Table VI dioxide and titanium. For these reasons an assembly and Henry's law an ozone concentration of a was used with a Luggin capillary of about 1 mm saturated solution was calculated to be 24 mg l-1, diameter placed, side wise, in contact along the under these conditions. Apart from the above- anode and potentials being recorded 5 s after mentioned difficulties with an undivided cell, its use switching on the electrolyses. The used electrode was considered essential, since without convection had been electrodeposited at 70 °C (and at i — in the electrolyte and mixing of the anolyte and the 40 mA cm~2 in our standard bath); before each catholyte at the bottom of the cell a very rapid measurement the electrode was sanded and washed. decrease of the anolyte pH would occur. From the results of Table VII only qualitative Table VII. Current density-potential curves on Pb02/Ti in two different electrolytes. C. d. Potential in 1.6 N H2S04 Potential in a phosphate buffer Correction applied for [mA cm"2] [V vs SCE) [phosphate] == 0.74 M, R = 0.42 potential drop in [V vs SCE] Pb02/Ti interphase [V] 100 2.41 2.10 0.27 200 2.57 2.37 0.415 400 2.80 2.72 0.59 600 2.92 2.95 0.70 800 3.02 3.15 0.79 1625 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water conclusions may be drawn: in particular that in b) If the chemical nature and crystallographic buffer electrolyte the potential drop is lower than in state of the electrode directly influences the oxygen 1.6 N H2SO4 solution. The very rapid rise of the and ozone evolution, then water or in general measured potential in the buffer may be attributed hydrogen and oxygen incorporated in the crystal to the high voltage drop in the electrolyte, even may directly or indirectly influence the amount of within a 1 mm layer of the electrolyte. electrogenerated ozone. Hydrogen in significant More accurate measurements will be carried out amounts is known to exist in electrodeposited lead with electrodes of lead dioxide deposited on titanium dioxide [33] and to a certain extend to affect the which will have a very thin platinum wire embedded pseudo equilibrium potential of Pb02. in the lead dioxide layer between which and the It is also known for the oxygen evolution on reference electrode the potential will be measured. platinum [34], that hydroxyl radicals may be inter- mediate species of the process. Moreover, it has been mentioned before that preheated anodes were less Discussion suitable for ozone production. It must also be About the mechanism of O3 formation on the mentioned that electrodes deposited at different studied system, no novel conclusions may be made temperatures, which showed a different time at this stage. Only some orienting thoughts can be response as regards the ozone concentration in the put forward from our measurements and the evolved gas, exhibited a different surface texture collected information from literature. which was well reproducible. a) A very informative review [30] suggests that c) The action of the surfactant used for Pb02 lead dioxide reacts as an inert electrode under anodic deposition may also influence the O3/O2 ratio in the polarisation and oxygen evolution, despite previous anodic gas evolved at higher potentials, if these are considerations by Glasstone [31] on adsorbed not in the range of oxidation of the surfactant. oxygen intermediates. The main argument is that lead does not experience a valence state greater d) The possibility of oxygen from lead dioxide than IV. (Since the O2 overpotential is higher than participating in the ozone synthesis must not be on platinum [10, 14] this can explain why better completely excluded. An analogy may be drawn yields of O3 are observed on Pb02 anodes.) However, from the process taking place on platinum in a this should not necessarily preclude the possibility perchloric acid electrolyte, although here, the of more energetic, adsorbed oxygen species, since oxygen rich species participating in the anodic gas lead dioxide is not met in stoichiometric composition evolution are not met in the solid phase as in the and, moreover, even in a perfect crystal surface case of Pb02. It is known that at temperatures states are theoretically anticipated and are found lower than —25 °C [5] a considerable amount of O3 to exist also at highly purified surfaces [32]. On the is anodically synthesized. Using isotope tracer other hand if Pb02 behaves as an inert electrode techniques it has been also found that at these low with respect to oxygen evolution, the a-Pb02 may temperatures and higher anodic polarisations, offer the possibility of greater ozone yields than oxygen from the perchlorate ion is incorporated in /?-Pb02 because according to literature [30] for this the anodic gas, and indeed much more, than for reaction the exchange current density of a-Pb02 is experiments run at room temperature [35]. Such a ca. six orders of magnitude lower than for /?-Pb02, conclusion is also drawn from electrochemical in electrolytes where a rapid phase transition to studies [36]. Oxygen from K2HPO4 is also known to /?-Pb02 does not occur. appear in O2 anodically evolved at a platinum an- ode during peroxyphosphate synthesis from a It should be noted that although Pb02 electrodes K2HP04/KF electrolyte [37]. The elucidation of the deposited on titanium at 70 °C have a lower 02 mechanism of O3 evolution might be supported by overpotential than those made at 25 °C [10] the the observation of the 03" ion by ESR at low tem- latter have been found by us to be less effective for perature, (viz. —35 °C) in concentrated KOH at 03 production. The fact that Lash et al. [5] observed CdO anodes [38]. improved yields at lower pressures above the electrolyte in a HCIO4 aqueous solution and with a e) Finally the pH at the anode surface and the platinum anode needs interpretation also. nature of adsorbed species is expected to strongly 1626 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water effect the ozone current yield. From capacitance viewed from an additional observation also; al- measurements [39] at Pb02 anodes in a potassium though on a Pt anode better yields are observed in sulphate electrolyte the adsorption of sulphate ions H2SO4 than in H3PO4 solution [42] the opposite at more anodic potentials was established. Also for effect tends to exist on Pb02 anodes. However, our persulphate synthesis on Pt, the importance of the experiments with a mixed phosphate buffer-K2S04 texture of the anode on the adsorption of sulphates electrolyte suggest that no great direct influence of and the consequence of this on the surface potential SO42- on the O3 yield should be expected at and current yields of the products has been shown relatively low concentrations of the anion. From [40]. The fluoride effect in our measurements is the above discussion and the fact that at low total supposed to be due to adsorption on the surface of phosphate concentrations and low pH a drop in the the electrode, too. The addition of KF resulted in yield is observed, it is suggested that by interaction an increase of the size of the evolved gas bubbles of phosphate species with the Pb02 surface an possibly indicating a drop in the surface charge [41]. increased amount of O3 arises. However, it is not Furthermore, it is difficult to accept considering the possible to rule out that only the Pb02 is of primary results of Table VI that in more concentrated importance. sulphuric acid solutions the remarkable drop of O3 yield with time at room temperature occurred Support of this work by Fonds der Deutschen because of homogeneous decomposition in the Chemischen Industrie and Rheinisch-Westfälisches Elektrizitätswerk-AG is gratefully acknowledged, as electrolyte, since the O3 lifetime is relatively long. well as discussions and experimental help by Dr. F. The role of the anions on the anode surface may be Hindelang and Dr.F.Faber at certain stages. [1] XVI. Mitteilung: H. P. Fritz und W. Kornrumpf, [20] A. S. T. M. cards; of Joint Committee on Powder J. Electroanal. Chem. 100, 217 (1979). Diffraction Standards 1972. [2] D. Wabner, H. P. Fritz, D. Missol, R. Huß, and a) 25-447 Ref.: A. N. Winchell, "Elements of F. Hindelang, Z. Naturforsch. 31b, 39 (1976). Optical Mineralogy"; [3] D. Wabner, H. P. Fritz, and R. Huß, Chem. Ing. b) 11-549 Ref.: A. I. Zaslavsky et al., Doklady Tech. 49, 329 (1977). Akad. Nauk SSSR 75, 559 (1950); [4] H. Gmelin, Handbuch der anorganischen Chemie, c) 11-548 Ref.: J. Burbank, J. Electrochem. Verlag Chemie GmbH, Weinheim, Sauerstoff 3, Soc. 104, 693 (1957). p. 984 (1960). [21] N. B. Miller, O. G. Tyurikova, U. G. Grinberg, [5] E. I. Lash, R. D. Hornbeck, G. L. Putman, and and E. V. Kasatkin, Sov. Electrochem. 13, 363 E. D. Boelter, J. Electrochem. Soc. 98, 134 (1951). (1977). [6] A. A. Rakov, G. F. Potapova, and V. I. Veselov- [22] O. G. Tyurikova, N. B. Miller, A. A. Yakovleva, skii, Sov. Electrochem. 6, 1655 (1970). and E. V. Kasatkin, Sov. Electrochem. 14, 375 [7] D. P. Semchenko, E. T. Lyubushkina, and V. I. (1978). Lyubushkin, Izv. Ser. Kavk. Nauk. Tsentra, [23] L. J. Csanyi and F. Solymosi, Acta. Chim. Hung Vyssh. Shk. Ser.: Tech. Nauk. 3, 98 (1975). 29, 47 (1961). [8] As in 1. c. [4] 3, p. 1082 (1960). [24] A. Vogel, A Text-Book of Quantitative Inorganic [9] D. P. Semchenko, V. I. Lyubushkina, and E. T. Analysis, p. 325, 298, Third Edition, Longman, Lyubushkina, Sov. Electrochem. 9, 1641 (1973). London 1972. [10] D. W. Wabner, Habilitation, München 1976. [25] C. Fabjan, Monatsh. Chem. 106, 513-528 (1975). [11] Nishimura, Yoshikazu, C. A. 79, No. 26581 w [26] F. A. Cotton and G. Wilkinson, Advanced In- (1973); also Japan Kokai 73, 48-27, 991. organic Chemistry, 3rd Ed., p. 411, Interscience, [12] Nishimura, Yoshikazu, C. A. 83, No. 50048 f Third Edition, New York 1972. (1975); also Japan Kokai 75, 50-3, 995. [27] G. V. Samsonov, Oxide Handbook, p. 351, [13] As in I.e. [4] 3, p. 1087 (1960). Plenum New York and London 1973. [14] G. N. Kokhanov and N. G. Baranova, Sov. [28] J. P. Hoare, Electrochim. Acta 2, 549 (1966). Electrochem. 8, 838 (1972). [29] F. Hindelang, Dissertation, München 1978. [15] Ch. Fabjan and P. Bauer, Z. Anal. Chem. 280, 121 [30] J. P. Carr and N. A. Hampson, Chem. Rev. 72, (1976). 679 (1972). [16] E. D. Boelter, G. L. Putman, and E. I. Lash, [31] S. Glasstone, J. Chem. Soc. 121, 2091 (1922). Anal. Chem. 22, 1533 (1950). [32] A. Many, Y. Goldstein, and N. B. Grover, Semi- [17] A. T. Palin, J. Am. Water Works Assoc. 67, 32 conductor Surfaces, North Holland 1971. (1975). [33] J. P. Pohl, Lecture in Euchem Conference [18] K. Nakamoto, Infrared Spectra of Inorganic (30. 4. 79-3. 5. 79), Endorf, B.R.D. Compounds, p. 107, Wiley, N. Y. 1978. [34] J. Bockris and A. Reddy, Modern Electro- [19] W. Smit and J. G. Hoogland, Electrochim. Acta chemistry, p. 1103, Plenum, New York and 16, 981 (1971). London 1973. 1627 H. P. Fritz et al. • Ozone Synthesis by Electrolysis of Water [35] E. V. Kasatkin, K. I. Rozanthal, and V. I. [39] J. P. Carr, N. A. Hampson, and R. Taylor, Veselovskii, Sov. Electrochem. 4, 1261 (1968). J. Electroanal. Chem. 27, 201 (1970). [36] J. P. Hoare, in A. J. Bard (ed.): Encyclopedia of [40] W. Smit and J. G. Hoogland, Electrochim. Acta Electrochemistry of the Elements, Marcel Dekker, 16, 821 (1971). New York 1974. [41] G. E. Potapova, A. A. Rakov, E. V. Kasatkin, [37] E. V. Kasatkin, O. G. Tyurikova, N. B. Miller, and V. I. Veselovskii, Sov. Electrochem. 13, 1014 and L. P. Koroleva, Sov. Electrochem. 13, 366 (1977). (1977). [42] C. Fabjan, 3. Congress of Ozone, May 1977, Paris. [38] C. S. Casey and C. L. Garner, J. Electrochem. Soc. 122, 851 (1975).