August, 19661 BERSIS AND VASSILIOU 499 A Chemiluminescence Method for Determining Ozone BY D. BERSIS AND €3. VASSILIOU (Nuclear Research Center “Democritus,” Aghia Paraskevi A ttikis, Athens, Greece) A method for determining ozone is described which is characterised by the direct recording and automatic determination of ozone within a wide range of concentrations. The development of this method is based on the use of a chemiluminescent solution that is stable, and shows a linear relation- ship between the light emitted and the ozone concentration. A combination of rhodamine I3 with gallic acid in ethanol is satisfactory in operation and does not itself emit light. The electronic instrumentation used is relatively simple. Other methods of ozone analysis based on this principle meet with much difficulty, owing t o the direct oxidation of the chemiluminescent com- pound.The present method, by contrast, involves the use of gallic acid as an ozone acceptor, and rhodamine B, which remains unchanged during the measurement, as a photon emitter. Observations made with an oscillograph of the light emitted by single bubbles of ozonised air passing through the chemiluminescent solution give valuable information about the response- time of the system. THE increasing interest in the applications of ozone, and their importance in the fields of radiation chemistry, upper atmosphere technology, industrial organic chemistry, etc., gives rise to a constantly growing number of projects dealing with basic research in ozone chemistry. Therefore, the development of good and rapid methods of ozone analysis is required.The methods for the determination of ozone that have been developed so far are mainly chemica1,l $2 e l e c t r ~ c h e m i c a l ~ ~ ~ ~ ~ - ~ and optical.’ ,8 Each of these methods has advantages and disadvantages, so that the method that is to be used must be selected according to the individual requirements and conditions. In general, however, the direct and continuous indication or, better, automatic recording of the results is desirable. Methods with procedures of this kind can be found in the l i t e r a t ~ r e , ~ ~ ~ ~ ~ but most of them are relatively slow in response, or require complicated instrumentation, Some methods are also hindered by the presence of gases, such as nitrogen dioxide and sulphur dioxide, which interfere more or less strongly.Within the range of optical methods, a field at present being developed, is one in which the light emitted by chemiluminescence reactions is used. The use of modem techniques of photon-counting combined with chemilurninescent systems of high efficiency can give rise, mainly from the point of view of sensitivity and response-time, to an ideal method of analysis. Nevertheless, a method based on these principles has not been developed to the extent expected, at least for ozone, owing to the fact that the chemilurninescent compound formed is constantly being destroyed during analysis, thus complicating the results. Re-cycling of the solution containing the chemiluminescent compound causes further complications.AIR SUPPLY- a manometer and a flow-meter were attached to control the pressure and the flow-rate. taining granulated potassium hydroxide and silica gel, respectively. OZONE PRODUCTION- EXPERIMENTAL A small rotary compressor, fluid metering type 8, Weldon Tool Co., was used, to which The air stream was freed from carbon dioxide and dried by means of two columns con- Ozonised air was produced by the following systems according to requirement- (a) A conventional Siemens ozoniser (Pyrex glass) with a wall thickness of 1 mm, gap distance of 3 mm and total volume of 11 ml. The two electrodes were filled with a sodium chloride solution (10 per cent. w/v). (b) A 5-fold ozoniser, i.e., five ozonisers, each like the one described above, connected in such a manner as to split the air stream into five equal streams as it enters.The five streams meet again at the common outlet of the ozonisers. By means of a suitable H.T. commutator switch, it was possible to energise any number of the individual ozonisers, as required by the experiment.500 BERSIS AND VASSILIOU: A CHEMILUMINESCENCE [Analyst, VOl. 91 The high voltage was supplied by a H.T. transformer (50 c/s) and controlled by means of a Variac connected to a suitable stabiliser, and was continuously monitored. To exercise additional control over the ozone concentration in the ozonised air stream, a suitable trap, K, (Fig. 1 ( b ) ) , containing granules of dry potassium hydroxide was used. The volume of ozone decomposed could be regulated by two taps, TI and T,, connected to suitable micrometric screws for fine adjustment.I Air photomultiplier ( 0 ) (b) K = Potassium hydroxide trap S = Silica gel trap C = Pyrex glass bubbler T, and T, = Taps M = Manometer Fig. 1. Apparatus for the automatic determination of ozone: (a), calibration unit; ( b ) , analysis unit CALIBRATION OF THE OZONISER- The calibration of the ozoniser, with respect to ozone production, was carried out as described by Ehmert .6 According to this method, a suitable reaction vessel containing a neutral 2 per cent. potassium iodide solution and a certain amount of dilute sodium thiosulphate solution is attached to the apparatus. The ozonised air bubbling through this vessel reacts with the potassium iodide, so liberating free iodine, which in turn reacts with the sodium thiosulphate.The residual sodium thiosulphate is then measured and compared with that of a blank. For this measurement, 4 platinum electrodes are used. An electric potential of about 0.18 volt is applied between two of them. At this voltage no electrolysis takes place as long as sodium thiosulphate is present, because of polarisation. The second pair of electrodes is connected to a suitable current source, so that iodine is liberated by electrolysis, and this reacts immediately with the sodium thiosulphate present. When the whole of the sodium thiosulphate has been consumed, the free iodine causes depolarisation of the first pair of electrodes, and a current flows which is linear with time. By using Faraday’s constant, the amount of iodine can be calculated from the values of current and time.PHOTOMETRIC ASSEMBLY- A Pyrex glass bubbler, C, (Fig. 1 (b)), of approximately 20 mm i.d., containing 10 ml of chemiluminescent solution, was used as a photometric cell. The porous diaphragm was of the G2 type. A second bubbler was connected in series with the first to ensure that no ozone escaped observation. An RCA 931 A photomultiplier, connected to a Varian G 11 A pen recorder through a pre-amplifier, or to an oscilloscope, and to a stabilised d.c. high tension supply, (Fig. 1 ( b ) ) , was also used.August, 19661 METHOD FOR DETERMINING OZONE 501 SOLUTION- of rhodamine B in 1 litre of ethanol (96 per cent. v/v). PROCEDURE- The gas stream to be analysed with respect to ozone concentration is passed through trap K, (Fig.1 (b)), if required (i.e., if the ozone concentration is too high), and is then bubbled through the reaction cell, C. The light emitted energises the photomultiplier, which in turn gives a signal to the recorder. This signal, as will be seen later, is proportional to the ozone concentration when the stream flow remains constant. The recorded area is used to determine the absolute amount of ozone passed through the reaction cell. CALIBRATION OF THE APPARATUS- in Fig. 1 (a). cent. v/v under the following conditions- The chemiluminescent solution was prepared by dissolving 2.5 g of gallic acid and 0.03 g METHOD AND RESULTS To calibrate the ozone-analysis apparatus, use was made of the calibration unit shown I t was found, by using Ehmert's method,6 that the production of ozone was 0-17 per gas, air; high tension, 7 kV; stream flow, 64 ml per minute; pressure, 17 inches of The ozonised air stream was led into the photometric cell, and the d.c.high tension supply connected to the photomultiplier was regulated so that the recorder showed an indication in support of the direct reading, e.g., of 17. LINEARITY AND LIMITS- To examine the linearity of the method, use was made of the calibration unit (Fig. 1 (a)), in which the single ozoniser was replaced by the 5-fold one. The five single ozonisers, as already indicated, were connected in such a manner that equal amounts of gas could pass through each. Equivalent amounts of ozone, therefore, were produced in each ozoniser under the same conditions. In practice, small differences occurring in the production of ozone by each ozoniser were corrected by fine adjustment of the high tension acting on each ozoniser.The concentration of ozone produced by each ozoniser was 0-07 per cent. v/v under the conditions described above. The percentage concentration of ozone was calculated with respect to the total stream flow, which was 64 ml per minute, and must not be confused with the flow through each ozoniser, which was approximately 13 ml per minute. water; temperature, 20" C. -l Number of operating ozonisers Fig. 2. Operation of the five-fold ozoniser The tap TI, (Fig. 1 ( b ) ) , was entirely closed; T, was opened and the d.c. high tension adjusted so that the recorder gave a reading of 100 when all five ozonisers were operating. By switching off one, two, three and four ozonisers the recorder gave readings of 80, 60, 40502 BERSIS AND VASSILIOU : A CHEMILUMINESCENCE [AutdySt, VOl. 91 atid 20, respectively, (Fig.2), It can, therefore, be concluded that in the region between 0.07 and 0.35 per cent. v/v, the light emitted b y the chemiluminescence reaction is linearly related to the ozone concentration. To determine whether this function is also linear in lower concentrations of ozone, the following operations were carried out- With all five ozonisers switched on, the ozone decomposition trap, K, (Fig. 1 ( b ) ) , was adjusted by means of taps T, and T,, so that the recorder gave a reading of 20. This repre- sented an ozone concentration of 0-07 per cent. v/v. On increasing the photomultiplier sensitivity by means of the d.c.high tension supply, a reading of 100 was given (the ozone concentration remaining at 0-07 per cent. v/v). Switching off afresh one, two, three and then four ozonisers, the recorder gave readings of 80, 60, 40 and 20, respectively. The function, therefore, is also linear within the region 0.07 to 0-014 per cent. v/v. Adopting the same technique, an ozone concentration of 0-0003 per cent. v/v was reached, the function remaining linear. STABILITY OF THE CHEMILUMINESCENT SOLUTION- Experiments with the continuous bubbling of ozonised air luminescent solution showed that the readings are stable for at stream flow is 64 ml per minute, and the ozone concentration 0.0 being made for the evaporation of alcohol. through 10ml of chemi- least 20 hours when the per cent.v/v, allowance INFLUENCE OF TEMPERATURE- This lack of effect with temperature change applies to the chemiluminescent solution only, and not to the ozonisers that were used in developing the method. Temperature change in the ozonisers largely affects the rate of ozone production. Change of temperature by * 10" C does not influence the results of analysis. INFLUENCE OF OTHER GASES- As may be appreciated, nitrogen, oxygen and similar gases do not interfere at all. Nitrogen dioxide and sulphur dioxide were each mixed with air and the mixtures were passed separately through the chemiluminescent solution to test whether they would (a) react with simultaneous emission of light, (b) destroy the solution. I t was found that in neither instance was light emitted; nor was the chemiluminescent solution destroyed, for, after passing ozone through it again, the reading remained un- changed, i e ., the reading was the same before and after nitrogen dioxide and sulphur dioxide had been passed through the solution. Ozonised - air S = Silica gel T = Tap Apparatus for increasing the stream flow without Fig. 3. change in the absolute volume of ozone per time-unit INFLUENCE OF THE STREAM FLOW- The apparatus outlined in Fig. 3 was used to study the effect of the gas-stream flow on the emitted light. The flow was increased at the outlet of the ozoniser in order to avoidlo change in the rate of ozone production.August, 19661 METHOD FOR DETERMINING OZONE 503 It was found that increasing the flow up to 200ml per minute does not influence the reading of the recorder when the rate of ozone production remains constant and within the limits already mentioned.RESPONSE-TIME OF THE CHEMILUMINESCENT SOLUTION- To determine the response-time of the solution, oscillographic observations of the light emitted by single bubbles were made (Fig. 4 ( a ) ) . Similar experiments were also conducted by using a second solution which contained no gallic acid (Fig. 4 ( b ) ) , and to make a better comparison of the results obtained, the heights of the pulses given by this second solution were arbitrarily equalised with those given by the first solution by adjusting the sensitivity of the oscilloscope. It can be seen that the time of fall, considerable (approximately second) for rhodamine B, (Fig.4 ( b ) ) , becomes almost zero for the mixture with gallic acid, (Fig. 4 ( a ) ) . I I tI % I , , I , , I I I I -1 -- Time, seconds - I 1 I . . - - Time, seconds Fig. 4 (u).. Light emitted by single bubbles of ozonised air through a mixture of gallic acid and rhodamine B in ethanol Fig. 4 (b).. Light emitted by single bubbles of ozonised air through rhodamine B in ethanol REPRODUCIBILITY OF RESULTS- It was found from a large number of recordings conducted at several ozone concentration levels under constant conditions, that the chemiluminescent solution (within its stability limits, as already stated) showed a fluctuation smaller than 21 per cent. DISCUSSION Although some attempts have been made to use the light emitted during chemi- luminescence reactions as a means of ozone analysis, all have encountered serious difficulties, mainly arising from ( a ) the continuous decrease of the concentration of the chemiluminescent compound that results in non-linear response of emitted light as a function of ozone concen- tration, and (b) the low level of the intensity of light emitted during the chemiluminescence reactions used.It must, therefore, be concluded that an analytical method based on the conversion of chemical energy into light, with measurement of the latter, can be successful only if the concentration of the chemiluminescent compound remains unchanged during analysis, i.e., if the chemiluminescent compound does not take a direct part in the chemical reactions occurring in the reaction vessel. Because such a change occurs, solutions like those used by Biswas and Dahr,ll and Briner,12 although emitting light under the influence of ozone, are nevertheless not suitable for ozone analysis.To avoid this difficulty, Bernanose and R&n&13 used chromatographic paper impregnated with solutions of luminol or rhodamine B. Even so, however, although regular luminescence is observed, this method cannot be used for continuous recording as the concentration of the light-emitting compound diminishes rapidly. Bernanose’s method also encounters another difficulty. Chromatographic paper-discs were used, containing only a small amount of chemi- luminescent compound (of the order of 1 pg) so that the intensity of the light emitted should have been accordingly low. This may have been the reason why a Lallemand 18-step photo- multiplier, which has a sensitivity about 10 times higher than that of the IP 21 RCA, was used.In the present work it was realised that only the protection of the chemiluminescent substance by another compound could lead to the solution of the problem. The latter com- pound should react with ozone more easily than does the former and its concentration should They are mainly useful for continuous recording of the results.504 BERSIS AND VASSILIOU A CHEMILUMINESCENCE [Ana&St, VOl. 91 be relatively high. Further, during the reaction with ozone either the compound itself or its reaction products, should be able to transfer to the chemiluminescent compound an amount of energy such that the latter would be only temporarily excited, and then return to its ground state by emitting a photon.A combination of rhodamine B, as a chemiluminescent compound with gallic acid in ethanol, was found to possess the desirable properties. Use was made of gallic acid for the following 8 reasons- It contains three hydroxyl groups and therefore? has a good quantum yield. It has no induction time because of the presence: of a carboxyl group. It protects rhodamine B from direct oxidation. The oxidation products of gallic acid are not coloured and, therefore, no screening effect Its solution in ethanol is not affected at all by the atmospheric oxygen. I t is not characterised by self-emission of light. The excited molecules or reaction products of gallic acid seem to be at such energy levels (under the influence of ozone) that the energy transfer to rhodamine B, which is finally responsible for the emission of light, becomes possible with a satisfactory quantum yield.The reaction of gallic acid with ozone, the energy transfer to rhodamine B and the subsequent emission of light have, in total, a much faster response than does the reaction of rhodamine B alone with ozone (cf. Fig. 4 (a) with Fig. 4 ( b ) ) . takes place. Rhodamine B was used because- (i) It appears to be a good acceptor of the energy provided by the reaction of gallic acid with ozone. (ii) It is stable to oxygen. (iii) Unlike luminol, it does not self-emit (although the self-luminescence of luminol can be inhibited16 by using a-naphthol or 3-indazolinone-4-carboxylic acid, other com- plications arise).(iv) The light that it emits is suited to the S4 surface of the photomultiplier that was used. ( v ) It is not oxidised directly by ozone in the presence of gallic acid. The authors of the present work have produced unpublished evidence that the reactions taking place may be summarised as follows- gallic acid + ozone --+ A* + oxygen rhodamine B + A* --+ rhodamine B* + B rhodamine B* -+ hu + rhodamine B where A* is an excited intermediate, or excited intermediates, resulting from the reaction between gallic acid and ozone and B is the final product, or products, of the oxidation. Taking into account (a) the sequence of the reactions, (b) the fact that rhodamine B remains unchanged during the process, and (c) that the concentration of gallic acid is much higher than that of rhodamine B (50 : 1), the advantages of this system can be readily apprecia- ted when it is compared with other methods in which direct oxidation of a chemiluminescent compound is used.The following conclusions can be drawn- As regards the use of the 5-fold ozoniser, it can be said that it is the most dependable source of ozone in that concentrations are kept constant and related accurately by certain ratios (e.g., 1 to 2, 1 to 3, 1 to 4 and 1 to 5 ) . If, instead of the taps TI and T, of the potassium trap (Fig. 1 ( b ) ) , a capillary tube is used in one or the other branch of the apparatus, a fixed proportion of the ozone would be destroyed. Partial decomposition of the ozone could also be achieved thermally. This method, however, requires closer control.t It has been observed that in reactions of polyphenols with ozone the light sum increases rapidly In the presence of carboxyl groups, the emission of light begins simultaneously with the reaction. with increasing numbers of hydroxyl groups. The delay otherwise observed is called the induction time.August, 19661 METHOD FOR DETERMINING OZONE 505 A G2 porous diaphragm was used in the construction of the bubbler, which acted as the reaction cell, C (Fig. 1 ( b ) ) , as this grade of diaphragm gives fine bubbles without the need to increase considerably the pressure of 17 inches of water of the ozonised air. It must be pointed out that the production of ozone in the ozoniser falls as the pressure increases and, therefore, all experiments must be carried out at the same pressure.For the single ozoniser that was used in the present work, it was found that by changing the pressure from 5 to 35 inches of water (the stream flow being constant and equal to 64 ml per minute), the ozone concentration changed from 0.10 to 0.08 per cent. v/v. I t has already been mentioned that the time of linear response of 10ml of the chemi- luminescent solution is 20 hours for a stream flow of 64ml per minute, and an ozone con- centration of 0.01 per cent. v/v. This time can be extended, either by increasing the dimensions of the reaction cell, or by suitable adjustment of the potassium hydroxide trap, K (Fig. 1 ( b ) ) . If by the latter, the time of linear response can easily become 700 hours with the conditions described above.In the instance of the solution containing gallic acid (Fig. 2), the occurrence of light fluctuation, which is almost entirely absent from similar curves taken by using rhodamine B alone in ethanol, is an additional measure of the fast response of this solution. The incidental noise originates from the statistical behaviour of the bubbles passing through the chemiluminescent solution. The intensity of light emitted is a measure of the ozone concentration only if the stream flow is constant. Generally, however, it is a measure of the absolute amount of ozone passing through the reaction cell per unit of time. The fact that by this method concentrations of ozone down to 0.0003 per cent. v/v only were recorded, does not exclude at all the possibility of measuring much lower con- centrations.For example, the modern low noise photomultipliers make it possible to count photons readily, one by one. A rough calculation of the quantum yield of the system used in the present work, shows that an emitted photon corresponds to about lo5 molecules of ozone, i.e., about 10-11pg of ozone. These extreme figures, which lie beyond the limits even of radioactivation analysis, show the future possibilities of chemiluminescence as a tool in analytical chemistry. This work was done under the auspices of the Greek Atomic Energy Commission. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. REFERENCES Morozov, N. M., Zh. Analit. Khim., 1960, 15, 367. Zehender, F., and Stumm, W., Mitt. Geb. Lebensmittelunters. u. Hyg., 1953, 44, 206. Wartburg, A. F., Brewer, A. W., and Lodge, J. P., jun., A i r & Wat. Pollut., 1964, 8 (l), 21. Mast, G. M., and Saunders, H. E., 1.S.A. Trans., 1962, 1, 325. Hersch, P., and Deuringer, R., Analyt. Chem., 1963, 35, 897. Ehmert, A., in “Ozone Chemistry and Technology,” Advances in Chemistry Series 21, American Osherovich, A. L., and Rodionov, S. F., A t m . Ozone (Moskow: Mosk. Univ.), Sb., 1961, 72. Alway, C. D., and Slomp, G., jun., in “Ozone Chemistry and Technology,” Advances in Chemistry Regener, V. H., Ibid., 1959, 21, 121. Bersis, D., and Katakis, D., J . Chem. Phys., 1964, 40 (7), 1997. Biswas, J., and Dhar, N., 2. anorg. allg. Chem., 1928, 173, 125. - _ _ , Ibid., 1931, 199, 400. Brink, E., Helv. Chim. Acta, 1940, 23, 320. Bernanose, A. J., and R h 6 , M. G., in “Ozone Chemistry and Technology,” Advances in Chemistry Bersis, D. S., 2. phys. Chem., 1969, 22, 328. Uemura, K., Science and Crime Detection, 1954, 7, 6. Chemical Society, Washington, D.C., 1959, p. 128. Series 21, American Chemical Society, Washington, D.C., 1959, p. 108. Series 21, American Chemical Society, Washington, D.C., 1959, p. 7. Received J u l y 5th, 1965