Anal. Chem. 2009, 81, 5226–5232
Mass-Dependent Isotopic Fractionation in Ozone Produced by Electrolysis S. K. Bhattacharya* Earth Sciences Division, Physical Research Laboratory, Ahmedabad, Gujarat 380009, India Joel Savarino Laboratoire de Glaciologie et Ge´ophysique de l’Environnement, CNRS/Grenoble Universite´, St. Martin d’He`res, France Boaz Luz The Institute of Earth Sciences, The Hebrew University of Jerusalem, Givat Ram, Jerusalem 91904, Israel During the electrolysis of water in an acidified medium, ozone is produced, in association with oxygen, at the anode. This ozone is found to be depleted in heavy isotopes (18O and 17O), with respect to the source water, following a strict mass-dependent rule. Our experiments also suggest that the isotopes are distributed at the apex and base positions of the bent ozone molecule in a random fashion, without obeying the zeropoint energy constraint. Endowed with these characteristics, the electrolytic ozone provides a source of reference that has a known internal heavy isotope distribution for spectroscopic studies. In addition, this ozone, when subjected to photolytic decomposition, can be used as a source of atomic oxygen with massdependent isotope ratios that can be varied by simply changing the water composition. Such an oxygen source is important for studying isotope effects in gasphase recombination/exchange reactions such as COO + O* f [COOO*] f COO* + O. The gas-phase production of ozone in the Earth’s atmosphere or in the laboratory (e.g., by discharge or ultraviolet (UV) photolysis of molecular oxygen) generates two unique isotopic effects.1,2 First, the heavy isotopes 17O and 18O are anomalously enriched or depleted (relative to the parent oxygen reservoir) in a manner that does not obey the normal rule of massdependent fractionation, i.e., δ17O ≈ 0.5 δ18O.3 Second, the internal partitioning of a heavy isotope in the singly substituted molecule, i.e., at the apex position of the ozone molecule (with a bent configuration) and the terminal end of it, neither obeys a random distribution nor has a pattern based on consideration * To whom correspondence should be addressed. E-mail:
[email protected]. (1) Thiemens, M. H.; Heidenreich, J. E., III. Science 1983, 219, 1073–1075. (2) Mauersberger, K.; La¨mmerzahl, P.; Krankowsky, D. Geophys. Res. Lett. 2001, 28, 3155–3158. (3) Craig, H. Geochim. Cosmochim. Acta 1957, 12, 133–149.
5226
Analytical Chemistry, Vol. 81, No. 13, July 1, 2009
of zero-point energy (ZPE).4 Even though it is rather easy to determine the bulk isotopic composition using mass spectrometry,1 the internal distribution can only be determined directly by spectroscopic methods.5 However, to avoid assumptions and reach high precision, any spectroscopic measurement of an ozone sample should be done in comparison to a standard whose internal distribution is known. Because there is considerable variation in the internal distribution in ozone prepared in the laboratory, and because of changes in pressure, temperature, and other factors, it is difficult to produce such a standard in a reproducible manner. Knowledge about the internal distribution of heavy isotopes in ozone4,6 ties in intimately with the current studies on understanding the origin of the isotope anomaly7 in ozone, from a chemical physics perspective. In addition, ozone is an important molecule for the chemical functioning of the atmosphere, because it initiates several oxidation chain reactions.8 Therefore, knowing and controlling its isotopic characteristics would help to resolve how the isotopic anomaly is propagated among other atmospheric molecules, such as nitrate, sulfate, or stratospheric CO2.7,9 It can, in effect, provide a new investigative tool to study the chemical reaction dynamics of many molecules.10 In this context, we have investigated the isotopic composition of ozone that is produced by the electrolytic decomposition of water. Because such ozone is not made in the gas phase but rather on the surface of the anode, it is expected that the anomalies associated with gas-phase reaction may be absent. We also used (4) Bhattacharya, S. K.; Pandey, A.; Savarino, J. J. Geophys. Res. 2008, 113, D03303. (DOI: 10.1029/2006JD008309.) (5) Janssen, C.; Tuzson, B. Appl. Phys. B 2006, 82, 487–494. (6) Michalski, G.; Bhattacharya, S. K. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 5493–5496. (7) Brenninkmeijer, C. A. M.; Janssen, C.; Kaiser, J.; Rockmann, T.; Rhee, T. S.; Assonov, S. S. Chem. Rev. 2003, 103, 5125–5161. (8) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments and Applications; Academic Press: San Diego, CA, 2000. (9) Thiemens, M. H. Science 1999, 283, 341–345. (10) Savarino, J.; Bhattacharya, S. K.; Morin, S.; Baroni, M.; Doussin, J.-F. J. Chem. Phys. 2008, 128, 194303 DOI: 10.1063/1.2917581. 10.1021/ac900283q CCC: $40.75 2009 American Chemical Society Published on Web 05/22/2009
our recently published method of ozone-silver reaction4 to determine its internal isotope distribution. Ozone has a bent shape, with an apex angle of 116.8° and a bond length of 1.27 Å; during reaction with silver, only the terminal atoms participate. Therefore, the oxygen, locked as silver oxide, represents the terminal composition of the parent ozone molecule modified by a massdependent reaction fractionation. This method thus allows us to examine the mass-dependent and mass-independent isotope distribution within the ozone molecule differentiating between the terminal and central positions. The electrolysis of water has been used previously for the enrichment of deuterium and tritium (2H and 3H), and it was shown that the hydrogen gas that is produced contains ∼80% less 2H (denoted by the fractionation factor R) than the water and the fractionation factor for 3H (which is represented by the symbol β) is even higher. Bigeleisen11 deduced the following expression relating the two fractionation factors: log β ) η log R where η has a theoretical value of 1.4 but, in practice, is dependent on the electrode materials. Meijer and Li12 were the first to investigate its use for the accurate determination of δ17O and δ18O of water via analysis of the evolved oxygen gas. As part of the procedure, they determined the slope that relates 17O/16O and 18 O/16O in the oxygen gas and determined the value to be 0.489 ± 0.003. Even though it has been known for a long time that ozone is formed in electrolysis of water, the present report is the first attempt to study its isotopic composition. EXPERIMENTAL PROCEDURE The electrolytic decomposition of water produces H2 and O2, with traces of O3. The experiment was conducted at two locations: Laboratoire de Glaciologie et Ge´ophysique de l’Environnement at Grenoble University, France (LGGE) and Institute of Earth Sciences, Hebrew University of Jerusalem, Israel (HUIES) for logistical consideration. However, in both places, the same electrolytic cell was used and the vacuum extraction lines for trapping the ozone and making it react with silver were of similar design. The oxygen isotopic composition of the electrolyte was controlled using waters from diverse sources (such as laboratory Millipore water, deionized tap water, ice from Antarctica, distilled seawater, and combinations of these types). Each water sample was mixed with an appropriate amount of concentrated H2SO4 to get a 3 M solution.13,14 Approximately 75 mL of the solution was placed in a U-shaped glass chamber separated in the middle with an O-ring joint that held a Nafion membrane (Nafion 117). This membrane hinders bulk liquid flow but allows the cations (H+) to pass through it, and it inhibits the migration of hydrogen gas and SO2 produced at the cathode to the anode side, where they could contaminate the O2/O3 gases. Electrolysis was conducted at a voltage of ∼15 V DC and a current of ∼1.5 A, using two (11) Bigeleisen, J. Tritium in Physical and Biological Sciences; International Atomic Energy Agency (IAEA): Vienna, Austria, 1962; Vol. 1, pp 161168. (12) Meijer, H. A. J.; Li, W. J. Iso. Environ. Health Stud. 1998, 34, 349–369. (13) Tatapudi, P.; Fenton, J. J. Electrochem. Soc. 1993, 140, 3527–3530. (14) Foller,; P, C.; Tobias, W. J. Electrochem. Soc. 1982, 129, 506–515.
platinum electrodes (∼2 cm2). The anode was connected to the power supply using a feed-through in a rubber cork (covered with Teflon tape) that had two flushing tubes inserted in it for purging the headspace with helium. The chamber was kept in a chilled water bath (∼8 °C) to dissipate the heat released by electrolytic process. The condensable gases produced at the anode (ozone, water vapor, traces of SO2) were collected in a spiral trap kept at liquid nitrogen (LN2). The H2 produced at the cathode was allowed to vent into the atmosphere. The main product gas on the anode side is oxygen but ozone is also produced in small but significant amounts,15 typically ∼80 µmol of ozone per hour. The product O2 was collected occasionally in part using a LN2-cooled molecular sieve spiral trap that was kept in line after the ozone trap. Typically, the O3/O2 ratio was ∼0.003. Ozone was cryogenically purified to remove water and possibly SO2. An aliquot of the ozone was taken on molecular sieve at LN2, which breaks down O3 into O2 when warmed to room temperature. Isotopic analysis of this O2 reflects the bulk O3 isotope ratios. To use the ozone-silver reaction method (to determine the intramolecular distribution of isotopes), the ozone was transferred to a cold finger that was connected to a chamber that contained pieces of cleaned and degassed silver foil. After being thawed, a few percent of this ozone react rapidly with silver to produce black silver oxide; the remainder of the ozone also is converted to oxygen by catalytic decomposition. The total amount of oxygen produced by these two effects was collected for determination of the yield and isotope ratios. Next, the silver oxide oxygen was released as O2 by heating the foil to 400 °C and collected for analysis. In a few instances, for comparative analysis, we also used ozone that had been produced via Tesla discharge of oxygen. The isotopic ratios (17O/16O and 18O/16O) are expressed in δ-notation and in units of ‰, relative to the international oxygen isotope ratio standard VSMOW.16 At LGGE, molecular oxygen isotopic measurements were performed with a Finnigan MAT 253 system, with typical instrumental errors of 0.05‰ on both δ18O and δ17O, respectively, for large samples (>10 µmol) and 0.1 per mil for small samples (
