Halogen Radical Chemistry at Aqueous Interfaces - ACS Publications

Jul 14, 2016 - Shinichi Enami,*,†. Michael R. Hoffmann,. ‡ and A. J. Colussi. ‡. †. National Institute for Environmental Studies, 16-2 Onogawa...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/JPCA

Halogen Radical Chemistry at Aqueous Interfaces Shinichi Enami,*,† Michael R. Hoffmann,‡ and A. J. Colussi‡ †

National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki 305-8506, Japan Linde Center for Global Environmental Science, California Institute of Technology, Pasadena, California 91125, United States



S Supporting Information *

ABSTRACT: Halogens play key roles in the chemical composition of marine boundary layers, the free troposphere and the stratosphere. Atmospheric halogen chemistry is dominated by reactions between gas-phase and aqueous species on the surfaces of the ocean and marine aerosol. The mechanisms of interfacial halogen radical/halide reactions, however, are not fully understood, partly due to the dearth of techniques for in situ monitoring of the products and intermediates of fast interfacial halogen radical reactions. Here, we report the online electrospray mass spectrometric identification of the species produced on the surface of aqueous Br− and I− microjets collided by I•(g) pulses generated from the 266 nm laser photolysis of CH3I/O2/N2 gas mixtures. Mass-specific identification of intermediates and products in D2O and H218O solutions and their dependences on I•(g) fluxes let us outline mechanisms of formation. We found that the uptake of I•(g) on the surface of Br− and I− microjets (effective uptake coefficient γeff ≥ 2 × 10−4) yields IBr•−/I2•− radical intermediates, which rapidly react with additional I• to produce trihalides I2Br−/IBr2−/I3− plus I3On− (n = 1, 2) species within ∼10 μs. Our findings point to a new halogen activation pathway initiated by photogenerated I•.



INTRODUCTION Halogens play unique roles in atmospheric chemistry.1−8 They are emitted by both anthropogenic and, far more importantly on the global scale, natural sources. The most significant natural source of halogens is the ocean, which continuously emits massive amounts of marine aerosols and sea-salt particles into the lower troposphere. The interaction of gas phase ozone with the surface of seawater and marine aerosols is known to release gaseous halogens X2 (X = I, Br, Cl),9−13 which, by being rapidly photolyzed by sunlight into reactive X atoms, trigger catalytic ozone depletion cycles. Among the halogens, iodine species have been found to effectively deplete ozone and form nanoparticles,3,14−17 revealing that iodine chemistry plays key roles on the global scale.6,18−20 A recent chemical model simulation has quantified the percentage of iodine-induced ozone losses in the marine boundary layer (34%), the transition layer (40%), and the free troposphere (FT) (26%).21 Model predictions, however, were based on the regeneration of IO• in the FT by unspecified reaction mechanisms.21 In the atmosphere, gas phase I atoms, I•(g), are generated by photolysis of inorganic (e.g., I2(g)) and organic (e.g., CH2I2(g), CH3I(g)) iodine compounds or via gas-phase radical−radical reactions involving IO• + IO•, IO• + RO2•, etc.3,22 The daytime concentrations of I•(g) are controlled by its gas-phase reaction with O3 (to form IO• radical, k = 1.3 × 10−12 cm3 molecule−1 s−1, τ < 1 s)23 and the rapid sunlight photolysis of IO•,14 which maintain a photodynamic equilibrium between I•(g) and IO•(g) where [I•(g)] reaches 10−22 pptv at some coastal sites.3,24−28 In such conditions, heterogeneous reactions of I•(g) with the surface of seawater and marine aerosols © XXXX American Chemical Society

containing halide ions may provide unprecedented halogen activation pathways. Furthermore, photodissociation of inorganic (I2/ICl/IBr and I2Ox)29−31 and organic iodides (RI)32−35 at marine aerosols/sea surface microlayers6 are expected to emit I atoms in situ and trigger subsequent reactions with halide ions therein. In addition to the lower troposphere, an exceedingly high ratio of I•(g)/IO•(g) (greater than unity) was recently predicted in the sunlit tropical tropopause layers, where cold temperatures and low ozone abundances suppress the gas phase reaction I• + O3.36 Much is known about gas-phase halogen chemistry.3,37 In contrast, heterogeneous halogen reactions, in particular those involving radicals, are not well understood.38−41 An outstanding example is the assumed radical-driven mechanism of formation of iodine-containing ultrafine particles.17 It is believed that these particles are generated from gas-phase homogeneous nucleation initiated by radical−radical reactions (e.g., IO• + IO• ).17,42 However, a recent field measurement study performed in coastal Antarctica failed to reveal short-term correlations between IO• radical and particles.5 Furthermore, a recent experimental study found that submicron particles produced by ultraviolet photoirradiation of a gaseous mixture of CH2I2 + O3 contain multiple iodine oxides components, IOn (n = 0−4), I2On (n = 0−6), and I3On (n = 0−1), where the formation mechanism (especially I3O) cannot be explained by the current chemical model.43 Since it is known that the larger Received: April 26, 2016 Revised: June 11, 2016

A

DOI: 10.1021/acs.jpca.6b04219 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A halide ions I− and Br− are enriched in the topmost layers of water surface relative to bulk water,44−47 they are expected to be the first species to trap reactive gases.6,9,10,12,13,48−51 Indeed, it has been shown that the reactive uptake of Cl2(g)/Br2(g) on I−/Br− involves an interface-specific mechanism proceeding via the formation of trihalides such as IBr2−.52,53 In a related study, the heterogeneous reactions of I•(g) with submicron squalane and squalene droplets were recently reported.54 Given that the mechanisms of the atmospherically relevant heterogeneous halogen radical reactions are not well understood, we decided to perform experiments specifically designed to investigate such processes at the air−water interface in situ. Recently, we developed a novel mass spectrometric setup to identify and quantify in situ the intermediates and products formed on the surfaces of continuously flowing liquid microjets exposed to beams of gas-phase radicals generated by laser flash photolysis of appropriate precursor mixtures under ambient conditions (i.e., at 1 atm total pressure, 298 K).55 By using this methodology, we were able to detect for the first time peroxyl radicals and peroxides, and sulfenic and sulfinic acid intermediates on the surface of aqueous carboxylic acids and glutathione microjets exposed to •OH-radicals, respectively. 55−58 Here we apply our method to investigate heterogeneous reactions of I•(g) with aqueous halide ions. Our study simulates the chemistry taking place on the atmospheric marine aerosols or seawater surfaces as genuinely heterogeneous reaction processes where I•(g) first sticks to the surface of water to subsequently react therein with surfaceenriched halide ions.

The observed process begins when I•(g) collides with the surface of the water microjets and sticks as I•(int) (R1),69 which may react with halide ions in the topmost layers (R2), self-react to form I2 (R3), or rebound to the gas phase (R-1). I•(g) → I•(interface, int)

(R1)

I•(int) + Br − → product

(R2)

2I•(int) → I 2

(R3)

I•(int) → I•(g)

(R-1)

Since the microjets break up within ∼10 μs after being ejected from the nozzle whereas the 8 ns laser pulses inject I• every 100 ms, the phenomena we observe take place in pristine solutions.



RESULTS AND DISCUSSION Figure 1 shows a typical negative ion electrospray mass spectrum of 0.1 mM NaBr microjets exposed to CH3I(g)/ N2(g)/O2(g) mixtures with the 266 nm pulse laser beam on and off.



EXPERIMENTAL SECTION The experimental setup has been described in previous publications, which provide additional details.55,58−60 The fast (within the ∼10 μs lifetime of the intact microjets)61 formation of anionic products at the air−water interface of microjets from the reaction of aqueous halide ions with gaseous iodine atoms at 1 atm at 298 K are monitored in situ by electrospray mass spectrometry (ES-MS, Agilent 6130 Quadrupole LC/MS Electrospray System). Aqueous NaBr or NaI solutions are pumped (100 μL min−1) into the spraying chamber of the mass spectrometer through a grounded stainless steel needle (100 μm bore) coaxial with a sheath issuing nebulizer N2(g) at a high gas velocity (vg ≈ 160 m/s).59 The surface specificity of our experiments had been previously demonstrated (e.g., the signal intensity obtained by the mass spectrometer from a sample corresponds to interfacial affinity of the ion).44,45,59−67 For example, the population of I− in the interfacial layers detected as m/z = 127 signal intensity is found to be 3.04 ± 0.24 times larger than that of Br− under a typical experimental condition.62 Few other techniques can instantly monitor the formation of primary products without further processing, under essentially wall-less conditions, on the surface of fresh microjets exposed to ppmv gaseous radical levels at atmospheric pressure for less than ∼1 ms. CH3I(g) concentration is quantified by a UV−vis absorption spectrophotometer (Agilent 8453) at 266 nm (absorption cross sections σ(266 nm) = 1.0 × 10−18 at 298 K)68 immediately before entering the reaction chamber (Figures S1 and S2). Throughout, reported [CH3I(g)] values, which correspond to the concentrations actually sensed by the microjets in the reaction chamber, are ∼13 times smaller than the values determined from UV absorbance due to further dilution by the drying gas. See Supporting Information (SI) for further details.

Figure 1. Negative ion electrospray mass spectra of 0.1 mM NaBr microjets exposed to CH3I(g)/N2(g)/O2(g) mixtures, [CH3I(g)] = 3.4 × 1015 molecules cm−3. Cyan: laser off. Red: 266 nm, 8 ns laser pulses at 10 Hz, on, which generate [I•(g)]0 ≈ 2.3 × 1014 molecules cm−3 (see SI for further details).

In the presence of 266 nm laser pulses, the reactant Br− signal (m/z = 79 and 81) decreases and several new peaks appear, which we attribute to reaction products initiated by I•(g) collisions with the surface of water (under [I•(g)]0 ≈ 2.3 × 1014 molecules cm−3, which correspond to exposures: E = 2.3 × 109 molecules cm−3 s, see SI for evaluation of [I•(g)]0). It is apparent that ∼50% of interfacial Br− is consumed upon exposure to ∼2.3 × 1014 molecules cm−3 I•(g), which is consistent with an effective I•(g) uptake coefficient of γeff ≈ 2 × 10−4 derived by the kinetic theory of gases and an assumed interfacial thickness ∼10−7 cm (see SI). Since [I•(g)] will decrease as a function of time during the laser pulse up to the microjets lifetime of ∼10 μs, the value is considered as a lower limit. Note that γeff ≥ 2 × 10−4 is comparable with γeff = 1.2 × 10−4 obtained from I•(g) uptake on squalene droplets.54 Mass spectra let us unambiguously identify the IBr•− radical anion (m/z = 206 and 208) as the primary product, the various trihalides IBr2− (m/z = 285, 287 and 289), I2Br− (m/z = 333 B

DOI: 10.1021/acs.jpca.6b04219 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A and 335), and I3− (m/z = 381), plus O-containing hydrolysis products such as HOIBr− (223 and 225), I2BrO− (m/z = 349 and 351), and I3O− (m/z = 397) (see below). No anionic species, such as I− and I3−, were detected on milli-Q water microjets exposed to CH3I(g) either with the 266 nm beam on or off (Figure S3). Neither we found evidence for the participation of methylperoxyl radical CH3O2• (which appears within ∼50 ns after •CH3 generation) under present conditions (see below). We verified that reactant depletion and product formation require the participation of CH3I(g) and 266 nm photons (Figures S4 and S5). Figure 2 shows how reactant and product mass spectral signal intensities depend on laser energy (A, B) and [CH3I(g)] (C), respectively.

In Figure 2A, B, laser energies (x-axis) at 1, 5, 10, 20, 30, and 40 mJ pulse−1 correspond to [I•(g)]0 ≈ 9.8 × 1012, 5.4 × 1013, 1.1 × 1014, 2.2 × 1014, 3.2 × 1014, and 4.3 × 1014 molecules cm−3, respectively. Figure 2C extends the [I•(g)]0 range, i.e., up to [I•(g)]0 ≈ 1.1 × 1015 molecules cm−3 (or E = 1.1 × 1010 molecules cm−3 s) at [CH3I(g)] = 1.6 × 1016 molecules cm−3 (x-axis). Notice that (1) all product signals are more intense than Br−, as expected from the exponential dependence of ion surface affinities on ionic radius established in previous experiments by Cheng et al.,45,70 and (2) the sum of signal intensities, i.e., total ion charge, is larger than the intensity of initial Br− signals. These observations imply that products are generated in I• reactions with Br− and intermediate species in intact aqueous microjets rather than in the daughter droplets produced from microjet fragmentation. The fact that Br− signals bottom out rather than decaying exponentially to zero implies that the outermost interfacial layers where the chemistry takes place are being competitively replenished by Br− diffusion from the bulk solution, as observed before.10,49,50,71 In this connection, it is important to note that the net charges registered by the ion detector are produced from the statistical separation of pre-existing anions and cations during the aerodynamic fragmentation of the neutral aqueous microjets into charged microdroplets.59,60,62 The resulting ensemble of negatively charged microdroplets, which contain excess anions in proportion to their populations in the interfacial layers of the microjets, subsequently undergo a series of consecutive Coulomb explosions (due to charge crowding from solvent evaporation) in which all anions are ultimately released to the gas-phase. Thus, since the net charge of the ensemble of charged droplets is conserved during Coulomb explosions, the observed partitioning of ions as a function of ion radius implies that products are generated on intact microjets rather than in secondary droplets. To reiterate, the ions detected by our mass spectrometer are ions that were already present or produced by chemical reactions in the interfacial layers of microjets (see previous publications for further details).59−61 Product signals evolve specifically as functions of [I•(g)]0. Notably, the dissimilar evolution of signal intensities at the lowest I• exposures (Figure 2A) reveals that the trihalide I2Br− is produced before IBr2− and has mechanistic implications (Scheme 1). We confirmed that HOIBr− has an exchangeable proton in experiments performed in D2O (Figure S6). At larger Scheme 1. Mechanism of Iodine Atom Initiated Reactions of Halide Ions at the Air−Water Interface

Figure 2. (A) Electrospray mass spectral signal intensities from aqueous 0.1 mM NaBr microjets exposed to gaseous CH3I/O2/N2 at [CH3I(g)] = 6.4 × 1015 molecules cm−3 irradiated by 266 nm laser pulses as a function of laser energy mJ pulse−1 up to 5 mJ pulse−1, (B) up to 40 mJ pulse−1, and (C) as a function of [CH3I(g)] under 40 mJ pulse−1 irradiation. Note the similar signal intensities in B and C experiments under identical 6.4 × 1015 molecules cm−3, 40 mJ pulse−1 conditions. C

DOI: 10.1021/acs.jpca.6b04219 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A I• exposures, I3On− (n = 1, 2) products appear. Similar experiments on I− microjets led to the generation of I3−, I2•−, IO−, HI2O2−, and I3On− (n = 1 and 2, m/z = 397 and 413, respectively) (Figures S7 and S8). The similar γeff ≈ 2 × 10−4 is obtained for I•(g) uptake on I− microjets. Thus, Br− and I− both react with I• at the air−water interface via similar mechanisms. Figure 3 shows electrospray mass spectra of 0.1 mM NaBr in 95 vol % H218O microjets exposed to gaseous CH3I/O2/N2 irradiated by 266 nm laser pulses. Clearly, all the O contained in the products derives from the solvent. In fact, the same I3On− (n = 1, 2) products were observed under anoxic conditions (i.e., in the absence of O2(g)) (Figure S9), thereby confirming that (1) O atom(s) come from water rather than from CH3O2•, (2) neither I* (2P1/2 excited

iodine atom, which is quenched by O2, see SI) nor CH3• radicals contribute to the observed products. We infer the rapid (≤10 μs) solvolysis of halogen-containing species at the air− water interfaces. On the basis of these observations, and in accordance with previous literature reports on polyhalide chemistry,4,14,72 we propose the mechanisms shown in Schemes 1 and 2: Scheme 2. Mechanism of the On-Incorporated Halogen Products Formation at the Air−Water Interface

The initial stages of the process involve the accommodation of I•(g) on the water surface (R1), followed by an extremely fast reaction of I•(int) with interfacial Br− (R2),47,63 leading to IBr•− (m/z = 206 and 208). We estimate k2 ≈ 1010 M−1 s−1 from k(I• + I−) = 9.8 × 109 M−1 s−1 and k(Br• + Br−) = 1.1 × 1010 M−1 s−1 in bulk water.4,72,73 Thus, the reaction I• + Br− at the air−water interface should take place in