Fractionation of 14N15N16O and 15N14N16O During Photolysis at 213 nm Hui Zhang, 1 Paul O. Wennberg, 1,2 Vincent H. Wu 3 and Geoffrey A. Blake 1 1 Division of Geological & Planetary Sciences, California Institute of Technology, Pasadena, CA Environmental Engineering Program, California Institute of Technology, Pasadena, CA 3 St. Catharine’s College, Cambridge University, Cambridge, UK 2 Abstract. Motivated by Yung and Miller’s suggestion [1997] that N2O is isotopically fractionated during UV photolysis in the stratosphere, we have studied the photolysis rates of the 14N15N16O and 15N14N16O structural isotopomers. In this study, we follow the concentrations of these compounds with FTIR spectroscopy during photolysis at 213 nm. Our results show that, as Yung and Miller predicted, the photolysis rate of 15N14NO is faster than 14N15NO at this wavelength. When fitted to a Rayleigh fractionation model, the observations yield single-stage enrichment factors of ε(14N15N16O / 14N14N16O) = -73 ± 5 per mil and ε(15N14N16O / 14N14N16O) = -41 ± 10 per mil. Comparisons are provided with theoretical calculation and previous measurements. 2 Introduction Nitrous oxide, N2O, is an important trace gas in Earth’s atmosphere. It is an efficient greenhouse gas and the major source of the nitrogen oxide radicals that destroy stratospheric ozone [Houghton, et al., 1995; WMO, 1995]. N2O is produced primarily by biological nitrification and denitrification processes occurring in soils and the oceans, and lost through UV photolysis and reaction with O(1D) in the stratosphere. It has been established that the present concentration of N2O in the atmosphere is 8% higher than the pre-industrial value and that it is increasing at a yearly rate of about 0.25%. N2O is targeted by the 1997 Kyoto Protocol on Climate Change for regulation. This is a difficult task, however, because the global budget for N2O is poorly quantified. The strength of the sources identified to date can only account for approximately two-thirds of the sum of the well-established sinks and the accumulation in the atmosphere. Efforts to determine the global budget of N2O and to identify the cause of its continuing increase in the atmosphere are therefore in order. Stable isotope analyses can provide useful constraints on the strength of the sources and sinks for atmospheric species. Efforts have been made to investigate the isotopic fractionation of N2O in various production and loss processes [Wahlen and Yoshinari, 1985; Yoshinari and Wahlen, 1985; Yoshida, 1988; Kim and Craig, 1990; Kim and Craig, 1993; Yoshinari, et al., 1997; Cliff and Thiemens, 1997; Rahn and Wahlen, 1997; Naqvi, et al., 1998]. These measurements reveal that, relative to tropospheric N2O, the major biological sources of N2O are light in both 15N and 18O, while stratospheric N2O is found to be isotopically heavy. To utilize these isotopic data to characterize the N2O global budget, one must understand the fractionation induced by biological processing and photolysis in the atmosphere. In an effort to explain the heavy stratospheric N2O, Yung and Miller [1997] (YM97) proposed a wavelength-dependent enrichment mechanism during UV photolysis. They suggested that the difference in the zero point vibrational energy (ZPE) for the heavier N2O isotopomers causes a blue-shift in the UV cross section, that, when convolved with the spectral characteristics of the actinic flux, results in fractionation. Experiments using laser photolysis and mass spectrometry have been conducted to test one aspect of this theory [Rahn, et al., 1998]. It has been shown that UV photolysis of N2O at 207 nm and 193 nm results in significant enrichment in both 15N and 18O with larger fractionations at longer wavelengths. This is consistent with YM97, though the observed enrichment factor is more than that predicted by the simple ZPE calculations [Rahn, et al., 1998]. The ZPE-induced fractionation theory also predicts that there should be a differential fractionation between the two structural isotopomers 14N15N16O and 15N14N16O during UV photolysis. In this paper, we use the Fourier Transform Infrared (FTIR) spectroscopy in conjunction with laser photolysis to study the fractionation between 14N15N16O and 15 14 16 N N O at 213 nm. 3 Experimental Figure 1 presents a schematic of our experimental set up. UV photons are generated by a Nd:YAG laser, while IR spectra are collected with an FTIR spectrometer. The general experimental procedure is to photolyze the N2O sample continuously while taking IR spectra to monitor the concentration of the isotopomers. The photolysis is conducted at 212.8 nm (abbreviated as 213 nm hereafter) using the 5th harmonic of a Coherent Infinity pulsed Nd:YAG laser. The YAG laser’s 532 nm output is doubled using one 2-mm thick β-BaB2O4 (BBO) crystal. In a second thin BBO crystal, 266 nm light is then mixed with the residual 1064 nm radiation to generate 213 nm pulses. The two crystals are cut for type I phase matched 4th and 5th harmonic generation of the 1064 nm fundamental. A single quartz Pellin Broca prism is used to separate 213 nm from other wavelengths, which are intercepted by bean stops. After a 45° mirror (M1), the photolysis beam enters the FTIR spectrometer (MagnaIR 560 from Nicolet) through a quartz window (W1). It passes into the sample compartment aperture at a steep angle and is redirected by a 0° mirror (M2) towards a 45° mirror (M3). Mirror M3 and another 0° mirror (M4) together pass the UV beam through the sample cell twice and then dump it on the edge of the M3 mount. All four mirrors (M1, M2, M3 and M4) are 213 nm high reflectors from Coherent. The spectrometer is oriented so that reflective losses of the photolysis beam on the windows are minimized. During the experiment, the stability of the UV power was monitored with a photodetector (UDT-555UV) through the reflection off window W1. The sample cell is made of thick-wall Pyrex glass. It has an inner diameter of 4 cm and a length of 15 cm. CaF2 was chosen as the window material because it is transparent in both the IR and UV. The two windows are glued on to the cell with Torr Seal (Varian) at slightly different wedge angles to minimize etaloning. Three isotopically labeled N2O samples were used. A pure N2O sample with natural isotopic abundance was used for 14N14N16O(99%+). Separate samples of 14N15N16O (98%+) and 15N14N16O (98%+) mixed with N2 (99.999%) at an N2:N2O ratio of 40 were purchased from Cambridge Isotopes. The different sample mixtures used in the three photolysis experiments are listed in Table 1. All the experiments are conducted at total pressure of approximately 760 torr and at room temperature. Photolysis (R1) of N2O produces O(1D), which reacts with N2O (R2): N2O + hν → N2 + O(1D) (R1) 1 O( D) + N2O → NO + NO (60%) (R2) → N2 + O2 (40%) . Because the occurrence of R2 confounds the analysis, O(1D) quenching by N2 or CO2 is necessary: O(1D) → O(3P) . (R3) In Exp. II and III, only one rare N2O isotopomer sample is involved and N2 serves the quenching gas. For an N2:N2O ratio of 75, less than 3% of O(1D) atoms are expected to undergo reaction R2 due to quenching by N2 (R3) (DeMore, et al., 1997). In Exp. I, N2 quenching is limited by the total cell pressure and the N2:N2O ratio of 40 in both the 14 15 16 N N O and 15N14N16O samples. In this case, CO2, which is four times more efficient than N2 at quenching O(1D), is used in addition. 4 The concentration of N2O is monitored via the Q-branch of the ν2+ν3 combination band, which lies at 2798 cm-1 for 14N14N16O. Shown in Figure 2 are the ν2+ν3 spectra of the three N2O isotopomers. They are taken at 0.5 cm-1 resolution with an MCT-A liquid nitrogen cooled detector. This band is chosen for several reasons. First, it sits in a region free of interference from ambient H2O and CO2 absorption (the spectrometer is nitrogen purged). Second, the peaks for the isotopomers are shifted with respect to each other by 30 cm-1 and the small amount of cross interference among them can be easily accounted for. Third, at one atmosphere pressure, the Q-branch is collision-broadened and is fully resolved at 0.5 cm1 resolution. This leads to linear Beer’s Law behavior, which was demonstrated in this apparatus by preparing N2O samples of known concentration using manometry. At 0.5 cm-1 resolution, a high SNR can be achieved in approximately 30 minutes (see below for an explanation of the photolysis time scale). For example, in Exp. II, the S/N is more than 100 for the initial spectrum. The signal is picked at the center of the Q-branch for 14N14N16O, while the noise is the root-mean-square (RMS) noise in the baseline between 2850 and 2900 cm-1. Similar SNRs are achieved in the other two experiments. Approximately 60 mW (2 mj at 30 Hz pulse repetitive rate) of 213 nm are used. In 11 hours, 70% of the initial N2O is removed. FTIR spectra are co-added, producing one data point every 30 minutes. In addition, spectra are taken before the photolysis to test the spectral processing protocol and data fitting. In order to monitor any post-photolysis processes, IR spectra are recorded after the photolysis laser is turned off. For the background, spectra of the same pressure of N2 (in Exp. II and III) or the same mixture of N2 and CO2 (in Exp. I) are acquired. The concentration of the isotopomers is derived by first normalizing the co-added spectra against the background. The integrated absorbance of the Q-branches is determined and the retrieved signal is corrected for the small amount of absorption from the other isotopomers. Results and Discussion Figure 3 illustrates the observed fractionation. The slope of the linear fit to the data gives the fractionation factor in a Rayleigh model [Fritz and Fontes, 1980]. The results from the three experiments are as follows: ε(14N15N16O / 14N14N16O) = -73 ± 5 per mil and ε(15N14N16O / 14N14N16O) = -41 ± 10 per mil. The uncertainty results primarily from systematic error introduced in defining the spectral baseline during data processing. The N2O band is on a sloping region of the IR intensity as shown in Figure 2. An NO2 band, discussed below, is superimposed nearby on the short-wavelength side of the N2O ν2+ν3 band. These complicate the spectral analysis. Different techniques for inferring the baseline were performed and various reasonable assumptions produced the assigned uncertainty. The arithmetic average of ε(14N15N16O / 14N14N16O) and ε(15N14N16O / 14N14N16O) gives –57 ± 15 per mil for ε(15N / 14N) at 213 nm. This result is consistent with the results reported by Rahn, et al. [1998]. They observed -18.4 per mil at 193 nm and -48.7 per mil at 207 nm. We have assumed that all N2O loss is due to R1; processes other than R1 that destroy N2O would affect our interpretation. One possible interference is R2, the reaction of O(1D) with N2O producing NO. The NO can undergo further conversion to NO2 in the cell. We have observed the formation of small amounts of NO and NO2 in each of the three 5 experiments. NO is monitored via its absorption band near 1875 cm-1, NO2 is monitored at 2907 cm-1 and 1617 cm-1. The photolysis of NO2, whose cross section at 213 nm is more than 1000 times larger than that of N2O [DeMore, et al., 1997], keeps its concentration low during the experiment. However, once the photolysis beam is turned off, NO is converted into NO2 on the time scale of a few hours. Post-photolysis spectra reveal that the conversion of NO into NO2 is close to unity. By comparing the photolysis spectra with reference spectra of NO and NO2, we determine that less than 2% of the N2O destroyed in our cell occurs via R2. This is consistent with the expected quenching rates. Other processes that form N2O are also taken into consideration. Among them, the three-body reaction of N2 + O(1D) to re-form N2O is extremely slow [DeMore, et al., 1997]. Exp. II and III provide a test to whether the photolysis of 14N15N16O produces 15N14N16O and vice versa. At over 70% photolysis yield, no formation of 15N14N16O is observed from the photolysis of 14N15N16O and vice versa. The fractionation found in this experiment (and that by [Rahn et al., 1998]) is significantly larger than that predicted by YM97. It appears that the ZPE model is too simple to fully account for the observed fractionation. There are a number of reasons why this might be the case. At 298 K, close to 90% of the N2O is in the ground vibrational state (000) mode while about 10% is in the first excited bending mode (010). Photodissociation dynamics studies [Neyer, et al., 1999 and refs therein] have shown that the photolysis of N2O (R1) occurs mainly via an orbitally forbidden but vibronically allowed transition through a bent excited state. Therefore, the vibrationally excited bending states of N2O have much larger Franck-Condon overlap with the dissociative state than does the (000) mode. This is corroborated by the large observed temperature dependence of the N2O cross section [Merienne, et al., 1990 and refs therein]. By deconvolving the cross section data at 225 and 296 K [Selwyn, et al., 1977] into contributions from the (000) and (010) modes, we found that the 213 nm cross section of the (010) mode is approximately 15 times larger than that of (000). This implies that at 213 nm and at 298 K, more than 50% of the photolysis occurs from the excited vibrational states. The differences in the potential energy surface for different N2O isotopomers are different for the two vibrational modes, which leads to different wavelength shifts of cross sections according to YM97. Therefore, including the vibrationally “hot” molecules in the fractionation calculation is a more appropriate approach. The preceding argument implicitly assumes that the photodissociation of N2O is via the repulsive B(1∆) electronic state, as YM97 did in light of the available data at that time. As Neyer, et al. [1999] demonstrate, however, the dynamics of this photodissociation is more complex than previously thought, with more than one electronic state involved in the excitation / dissociation. It is possible that the two different ground state modes may have different coupling with the upper states, which will further complicate the spectroscopy. In conclusion, this study supports YM97’s suggestion that the different photolysis rates of the various N2O isotopomers appear to be the predominant mechanism responsible for the observed fractionation in the stratosphere. A fully quantitative test of this theory, however, requires accurate wavelength and temperature dependent differential cross sections for these compounds; the use of simple theoretical models for prediction of these cross sections has been ruled out by this study and previous work of Rahn et al. [1998]. The difference in the cross sections of the isotopomers examined here is large enough that a classical Beer’s law study of the isotopomers (which are available commercially in high purity) can provide the required spectroscopic data. 6 References Cliff, S. S., and M. H. Thiemens, The 18O/16O and 17O/16O ratios in atmospheric nitrous oxide: A mass-independent anomaly, Science, 278, 1774-1776, 1997. DeMore, W. B., S. P. Sander, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina, Chemical kinetics and photochemical data for use in stratospheric modeling, JPL Publication 97-4, NASA/JPL, Pasadena, California, 1997. Fritz, P., and J. Ch. Fontes (Eds.), Handbook of environmental isotope geochemistry, volume 1, the terrestrial environment, A, 545 pp., Elsevier, New York, 1980. Houghton, J. T., L. G. Meira Filho, J. Bruce, H. Lee, B. A. Callander, E. Haites, N. Harris and K. Maskell (Eds.), Intergovernmental Panel on Climate Change, Climate change 1994: radiative forcing of climate change and an evaluation of the IPCC IS92 emission scenarios, Cambridge University Press, New York, 1995. Kim, K.-R., and H. Craig, Two isotope characterization of N2O in the Pacific Ocean and constraints on its origin in deep water, Nature, 347, 58-61, 1990. Kim, K.-R., and H. Craig, Nitrogen-15 and oxygen-18 characteristics of nitrous oxide: A global perspective, Science, 262, 1855-1857, 1993. Merienne, M. F., B. Coquart, and A. Jenouvrier, Temperature effect on the ultraviolet absorption of CFCl3, CF2Cl2 and N2O, Planet. Space Sci., 38, 617-625, 1990. Naqvi, S. W. A., T. Yoshinari, D. A. Jayakumar, M. A. Altabet, P. V. Narvekar, A. H. Devol, J. A. Brandes, and L. A. Codispoti, Budgetary and biogeochemical implications of N2O isotope signatures in the Arabian Sea, Nature, 394, 462-464, 1998. Neyer, D. W., A. J. R. Heck, and D. W. Chandler, Photodissociation of N2O: J-dependent anisotropy revealed in N2 photofragment images, J. Chem. Phys., 110, 3411-3417, 1999. Rahn, T., and M. Wahlen, Stable isotope enrichment in stratospheric nitrous oxide, Science, 278, 1776-1778, 1997. Rahn, T., H. Zhang, M. Wahlen, and G. A. Blake, Stable isotope fractionation during ultraviolet photolysis of N2O, Geophys. Res. Lett., 25, 4489-4492, 1998. Selwyn, G., J. Podolske, and H. S. Johnston, Nitrous oxide ultraviolet absorption spectrum at stratospheric temperatures, Geophys. Res. Lett., 4, 427-430, 1977. Wahlen, M., and T. Yoshinari, Oxygen isotope ratios in N2O from different environments, Nature, 313, 780-782, 1985. World Meteorological Organization, Scientific assessment of ozone depletion: 1994, Global Ozone Research and Monitoring Project-Report 37, World Meteorological Organization, Geneva, Switzerland, 1995. Yoshida, N., N-15-depleted N2O as a product of nitrification, Nature, 335, 528-529, 1988. Yoshinari, T., and M. Wahlen, Oxygen isotope ratios in N2O from nitrification at a wastewater treatment facility, Nature, 317, 349-350, 1985. Yoshinari, T., et al., Nitrogen and oxygen isotopic composition of N2O from suboxic waters of the eastern tropical North Pacific and the Arabian Sea – measurement by continuousflow isotope-ratio monitoring, Marine Chem., 56, 253-264, 1997. Yung, Y. L., and C. E. Miller, Isotopic fractionation of stratospheric nitrous oxide, Science, 278, 1778-1780, 1997. 7 H. Zhang, P. O. Wennberg, and G. A. Blake, Division of Geological and Planetary Sciences, M/S 150-21, California Institute of Technology, Pasadena, California 91125. (e-mail: hui@gps.caltech.edu; wennberg@gps.caltech.edu; gab@gps.caltech.edu) V. H. Wu, St. Catharine’s College, University of Cambridge, Cambridge, CB2 1RL, UK. (email: vw205@hermes.cam.ac.uk) (Received X X, 1999; Revised X X, X; Accepted X X, X;) Table 1. Gas mixtures used in the photolysis experiments. 14 Exp. N14NO (torr) 14 N15NO (torr) 15 N14NO (torr) Quenching Gas (torr) 460 (N2) 280 (CO2) I 5 6 5.5 II 6 4 - 750 (N2) III 6 - 5 750 (N2) Figure 1. Light from an Nd:YAG laser is directed into the sample compartment of an FTIR spectrometer via a side port. Gas samples are located inside the cell. Figure 2. FTIR spectra of the three N2O isotopomers, taken between 2650 cm-1 and 2900 cm1 at 0.5 cm-1 resolution. Figure 3. The fractionation data from 213 nm photolysis, fitted with Rayleigh fractionation model. δ = (Ri / Rstd - 1) x 1000, where R’s are the slow-to-fast photolysis isotopic ratio. Rstd is for the pre-photolysis samples and Ri is for the photolyzed samples. f is the fraction of N2O remaining. ε(15N14N16O) = ε(15N14N16O / 14N14N16O) and ε(14N15N16O) = ε(14N15N16O / 14 14 16 N N O). ZHANG ET AL.: FRACTIONATION OF N2O DURING PHOTOLYSIS AT 213 NM * Author for correspondence. E-mail: hui@gps.caltech.edu, Fax: (626) 585-1917