Atmos. Chem. Phys., 16, 4439–4449, 2016 www.atmos-chem-phys.net/16/4439/2016/ doi:10.5194/acp-16-4439-2016 © Author(s) 2016. CC Attribution 3.0 License. 12 13 Kinetic isotope effects of CH3D C OH and CH3D C OH from 278 to 313 K L. M. T. Joelsson1,3, J. A. Schmidt1, E. J. K. Nilsson2, T. Blunier3, D. W. T. Griffith4, S. Ono5, and M. S. Johnson1 1Department of Chemistry, University of Copenhagen, Copenhagen, Denmark 2Division of Combustion Physics, Department of Physics, Lund University, Lund, Sweden 3Centre for Ice and Climate (CIC), Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark 4University of Wollongong, Department of Chemistry, Wollongong, Australia 5Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA, USA Correspondence to: M. S. Johnson ([email protected]) Received: 16 September 2015 – Published in Atmos. Chem. Phys. Discuss.: 15 October 2015 Revised: 2 March 2016 – Accepted: 16 March 2016 – Published: 11 April 2016 Abstract. Methane is the second most important long-lived source signal with relatively small adjustment due to the sink greenhouse gas and plays a central role in the chemistry signal (i.e., mainly OH oxidation). of the Earth’s atmosphere. Nonetheless there are signifi- cant uncertainties in its source budget. Analysis of the iso- topic composition of atmospheric methane, including the 13 doubly substituted species CH3D, offers new insight into 1 Introduction the methane budget as the sources and sinks have dis- tinct isotopic signatures. The most important sink of at- Atmospheric methane is the subject of increasing interest mospheric methane is oxidation by OH in the troposphere, from both the climate research community and the pub- which accounts for around 84 % of all methane removal. lic due its impacts on climate change, as reported by the Here we present experimentally derived methane C OH ki- IPCC(2013). The direct radiative forcing of methane is −2 netic isotope effects and their temperature dependence over 0.64 Wm . Including feedback mechanisms and secondary 13 effects (e.g., increased O production, stratospheric water va- the range of 278 to 313 K for CH3D and CH3D; the lat- 3 por, and production of CO ), methane’s radiative forcing be- ter is reported here for the first time. We find kCH =kCH D D 2 4 3 comes 0.97 Wm−2, two-thirds of the forcing by CO over : ± : k =k13 D : ± : 2 1 31 0 01 and CH4 CH3D 1 34 0 03 at room temper- ature, implying that the methane C OH kinetic isotope effect the same time period (IPCC, 2013, Fig. 8.15). D Atmospheric methane has both natural and anthropogenic is multiplicative such that .kCH4 =k13CH /.kCH4 =kCH3D/ 4 sources and the two categories contribute about equally kCH4 =k13CH D, within the experimental uncertainty, given 3 (Ciais et al., 2013, and references therein). Wetlands are the k =k13 D : ± : the literature value of CH4 CH4 1 0039 0 0002. In addition, the kinetic isotope effects were characterized dominant natural source, and agriculture and waste are the using transition state theory with tunneling corrections. largest anthropogenic sources. Fossil fuels make smaller con- Good agreement between the experimental, quantum chem- tributions. The majority (84 %) of atmospheric methane is ical, and available literature values was obtained. Based removed by oxidation by OH in the troposphere: on the results we conclude that the OH reaction (the C ! C main sink of methane) at steady state can produce CH4 OH CH3 H2O; (R1) 13 an atmospheric clumped isotope signal (1. CH3D/ D 13 13 while oxidation in the troposphere by Cl contributes about ln.TCH4UT CH3DU=T CH4UTCH3DU/) of 0:02 ± 0:02. This 13 4 % of the total: implies that the bulk tropospheric 1. CH3D/ reflects the CH4 C Cl ! CH3 C HCl: (R2) Published by Copernicus Publications on behalf of the European Geosciences Union. 13 4440 L. M. T. Joelsson et al.: Kinetic isotope effect in CH3D C OH 13 About 8 % of methane is removed in the stratosphere by rad- A related measure is the 1. CH3D/ value that quantifies ical oxidation, such as Reactions (R1), (R2) and (R3): the extent to which rare isotopes clump together to form a multiply substituted species, as opposed to a stochastic dis- 1 CH4 C O. D/ ! CH3 C OH: (R3) tribution (Ono et al., 2014): T13CH DUT12CH U The rest (4 %) is removed by soil (Kirschke et al., 2013). 1.13 / ≡ 3 4 ; CH3D ln 12 13 (3) Carbon and hydrogen isotopic analyses are widely used T CH3DUT CH4U to distinguish microbial and thermal sources of atmospheric where T13CH DU, T12CH U, T12CH DU, and T13CH U represent methane (e.g., Lowe et al., 1997; Ferretti et al., 2005; Tyler 3 4 3 4 the concentrations of the different isotopologues. et al., 2007; Lassey et al., 2007). However, Reactions (R1), The kinetic isotope effects for the singly substituted (R2), and (R3) produce relatively large D/H isotope effects species CH D and 13CH have been studied previously both (Saueressig et al., 1995, 1996, 2001; Crowley et al., 1999; 3 4 experimentally and theoretically; see Tables3 and4, respec- Feilberg et al., 2005). Thus, the construction of an accurate 13 C;Dα top-down methane budget by isotopic analysis must take the tively. The kinetic isotope effect for the reaction with isotopic signatures of both sources and sinks into account OH is not described in the existing literature. The related ki- C (Quay et al., 1999; Bergamaschi et al., 2000; Allan et al., netic isotope effect for the CH4 Cl reaction was measured 2001a,b). An isotope budget based on methane source (and at room temperature with the present setup by Joelsson et al. : ± : sink) fractionations is an underdetermined systems (e.g., (2014) and found to be 1 60 0 04. The kinetic isotope ef- Pohlman et al., 2009). Recent advances in mass spectrom- fects in the doubly substituted methane isotopologue CH2D2 C etry (Eiler et al., 2013; Stolper et al., 2014) and laser infrared for the reaction CH4 Cl has been studied previously by spectroscopy (Ono et al., 2014; Wang et al., 2015) facili- Feilberg et al.(2005). Dα tate measurement of rare doubly substituted isotopologues. In the present study the kinetic isotope effects and 13C;D The abundance of these “clumped” isotopologues (clumped α are determined using the relative rate method. Species refers to the rare isotopes being clumped together) generally concentrations in the reaction cell are determined using D T12 UT13 UD Fourier transform infrared (FTIR) spectroscopy. Further, α, follows a stochastic distribution (i.e., CH4 CH3D 13 13 13 12 C C;D T CH4UT CH3DU). However, small deviations from stochas- α, and α are calculated using quantum chemistry and tic distribution can be induced by thermodynamic (Ma et al., transition state theory. 2008; Stolper et al., 2014; Liu and Liu, 2016), kinetic (Joels- son et al., 2014; Wang et al., 2015), and photolytic processes 2 Experimental procedures (Schmidt et al., 2013; Schmidt and Johnson, 2015). Analysis of the clumped isotope anomaly in methane will yield unique Sixteen experiments were conducted, numbered from 1 constraints for the methane budget. Optical methods, as will through 16 (see Table1): eight (Experiments 1–8) for be shown in this paper, provide high throughput and accuracy 12 13 CH3D and eight (Experiments 9–16) for CH3D. The for overcoming the problems of analysis of clumped CH4. experiments were conducted at four different temperatures The difference and advantage of this approach is the addi- (T D 278, 288, 298, 313 K D 5, 15, 25, 40 ◦C); two experi- tional information not available in single-isotope analysis, ments were conducted for each temperature. especially regarding the mechanism of formation, indepen- dent of the enrichment of D and 13C in the starting material. 2.1 Relative rate method j The kinetic isotope effect, Eα, is a characteristic property of each process: The experiments were carried out using the relative rate method on a semi-static gas mixture. The decaying concen- i j k. E C OH/ trations of reactants were measured as a function of the ex- Eα ≡ ; (1) k.j E C OH/ tent of reaction. Considering two bimolecular reactions with second-order rate coefficients kA and kB, where iE is the most abundant (here, the lighter) isotopo- C !kA logue, j E the rare (heavy) isotopologue, and k.E C OH/ the A OH products; (R4) C kB reaction rate coefficient for the reaction E OH. As a mea- B C OH ! products; (R5) 13 sure of how much of a fractionation of CH3D kinetic re- actions produce, the apparent clumpiness, γ , is used. It is a and assuming there were no other loss processes, the follow- measure of the effect of the clumped substitution on the re- ing relation holds: action rate, as opposed to the combined effect of two single TAU0 kA TBU0 substitutions. It is defined as (Wang et al., 2015) ln D ln : (4) TAUt kB TBUt 13C;D α Here TAU0, TAUt , TBU0, and TBUt represent the concentrations γ ≡ : (2) 13Cα×Dα of compounds A and B at times 0 and t, respectively. The Atmos. Chem. Phys., 16, 4439–4449, 2016 www.atmos-chem-phys.net/16/4439/2016/ 13 L. M. T. Joelsson et al.: Kinetic isotope effect in CH3D C OH 4441 Table 1. Experimental setup. The experiment numbers are listed in column Exp., the detector is listed in column Detect., the heavy CH4 x isotopologue included in the experiments is listed in column T CH3DU, the mean measured temperatures in the photoreactor are listed in column T , the H2O vapor concentrations at the start of the experiments (t D 0) as obtained from spectral fitting are listed in column TH2OUtD0, the mean O3 concentrations after refill (i.e., the “top” values) as obtained from spectral fitting are listed in column TO3Utop, the 12 12 CH4 concentrations at the start of the experiments (t D 0) as obtained from spectral fitting are listed in column T CH4UtD0, and the heavy x CH4 concentrations at the start of the experiments (t D 0) as obtained from spectral fitting are listed in column T CH3DUtD0.