Partitioning of Halogens in Volcanic Gas Emissions

Volcanic eruption columns are complex and dynamic chemical environments. Along with sulfur dioxide and water, volcanic eruption columns may contain large quantities of the halogens chlorine and bromine, and to a lesser extent iodine. Plume temperatures range from very hot at the crater rim to ambient within the umbrella cloud. Entrainment of environmental air during convection is likely to produce enhancements in water, rapid parcel cooling and resultant kinetic freeze-out of chemical species.  Until recently, it had been assumed that volcanic halogen gases primarily existed in their hydrogen halide forms, i.e., hydrogen chloride, hydrogen bromide, and hydrogen iodide.

This is important because hydrogen halides are extraordinarily soluble in water and water-soluble species are effectively screened from the stratosphere during eruption column ascent. Indeed, conventional wisdom held that volcanic eruptions posed an insignificant source of stratospheric halogen gases until Textor et al. (2003) demonstrated that, under some conditions, significant quantities of hydrogen chloride might survive transport in the eruption column to the stratosphere.

Bobrowski et al. (2003) first detected an oxidized halogen — the bromine monoxide radical, BrO — in a volcanic plume using DOAS. Subsequently, Bobrowski et al. (2007) explained the presence of BrO by invoking a bromine explosion mechanism similar to the one described by Barrie and Platt (1997) in regard to the very different phenomenon of polar sea-salt processing.

The mechanism goes like this:

Some quantity of volcanic HBr is thermally dissociated to produce the Br radical.

\text{HBr} \xrightarrow[]{\Delta} \text{H} + \text{Br}

This radical then reacts with entrained ozone to form BrO.

\text{Br} + \text{O}_3 \rightarrow \text{BrO} + \text{O}_2

The BrO may react with the hydroperoxyl radical, which may also form due to thermal chemistry in the volcanic plume, producing hypobromous acid.

\text{BrO} + \text{HO}_2 \rightarrow \text{HOBr} + \text{O}_2

HOBr can then react with dissolved bromine (e.g., in an acidic droplet on a particle surface) to produce molecular bromine.

\text{HOBr}_\text{(g)} + \text{Br}^-_{\text{(aq)}} + \text{H}^+_{\text{(aq)}} \rightarrow \text{Br}_{2\text{(g)}} + \text{H}_2\text{O}_\text{(l)}

Molecular bromine is rapidly photolyzed to produce two more bromine radicals, which react with ozone.

\text{Br}_2 + \text{h}\nu \rightarrow 2\text{Br}

2\text{Br} + 2\text{O}_3 \rightarrow 2\text{BrO} + 2\text{O}_2

Thus, for every initial bromine atom, two bromine monoxide molecules are produced, provided there is enough ozone in the entrained air volume.

Finally, bromine atoms regenerate themselves to perpetuate the autocatalytic cycle:

\text{BrO} + \text{h}\nu \rightarrow \text{Br} + \text{O}

This O atom may produce more ozone.

\text{Net}: \text{Br} + \text{Br}^-_\text{(aq)} + \text{O}_3 + \text{H}^+_\text{(aq)} \xrightarrow[]{surface} 2\text{BrO} + \text{products}

If significant quantities of volcanic halogen species were to partition to the stratosphere, they would contribute to ozone loss processes, as demonstrated by Klobas et al. (2017). Figure 1, below, demonstrates how the co-injection of volcanic halogens may significantly perturb column ozone following a Pinatubo-sized volcanic eruption in a RCP 2.6 climate change scenario. Panels a — b relate the effect from sulfate aerosol alone in the year 2018 (a) and 2100 (b). Notably, as the stratospheric burden of chlorine declines due to the decay of anthropogenic chlorofluorocarbons (CFCs), sulfate-only volcanic eruptions will have less and less impact on total column ozone — but, if the volcanic eruption also injects halogens, as in panels c — f, ozone depletion is predicted even when background halogen levels are low.



Figure 1: Ozone response to stratospheric halogen injection in contemporary and future RCP 2.6 atmospheres. (a and b) Pinatubolike eruptions (SO2 injection only), (c and d) coinjection of 0.014 HCl:SO2 (EESC effectively increased by ~0.3 ppbv), and (e and f) coinjection of 0.14 HCl:SO2 (EESC effectively increased by ~3 ppbv). Note that the color scale is nonlinear in order to encompass the broad range of ozone changes. Global averages (90°S–90°N) of total column ozone perturbation are traced atop each panel as a function of time. Temporal average ozone anomalies are traced right. Global-temporal averages are enumerated in the top right. Black triangles indicate injection latitude and time. (Klobas thesis 2018)

The existence of oxidized halogens in a volcanic plume is a game-changer when it comes to stratospheric injection efficiency. While hydrogen halides possess Henry’s law constants in the 1e13 — 1e15 mol/atm neighborhood, halogen monoxides are closer to 0.5 — 0.8 mol/atm. Compare this to the Henry’s law constant for sulfur dioxide, which we know partitions effectively to the stratosphere, of about 1e5 mol/atm. IF — and it’s a big if — volcanic halogen species can be found in great quantity in their non-hydride forms, the stratospheric injection efficiency for these halogen gases may be much higher than previously considered.  Indeed, in the recent past, groups have reported finding UTLS BrO, OClO, and IO in aged volcanic plumes with SCIAMCHY and GOME-2.


So, an interesting paper showed up in a google scholar alert in my email inbox this morning: Bobrowski and colleagues measured the evolution of halogen gases using both in situ and remote sensing techniques AND quantified the oxidized fractions. What they found confirmed my suspicions: while chlorine exists mainly in the form of hydrogen chloride, up to 35% of total bromine and 18% of total iodine are bound in oxidized species. Mind you this wasn’t an explosively erupting volcano with stratospheric significance — they measured emissions from a lava lake — so this isn’t direct evidence of oxidized halogen species in a stratospheric eruption column (and eruption columns are dark and ozone-poor, so photochemistry will be suppressed everywhere except the boundaries), but it’s another step in that direction.

The stratospheric injection efficiency of volcanic halogens is a developing story and more research is needed to really constrain the magnitude of the threat to stratospheric ozone. The fact that oxidized halogen species are significantly present in volcanic emissions is another indication that prior estimates of the very low halogen injection efficiency may have been too conservative.


Interactions between Climate Change and Volcanism

An interesting review detailing the inter-linkage between climate change and volcanic activity hit my google scholar alert inbox.  Cooper et al. (2018) discuss not only how large explosive volcanic eruptions might impact surface radiative forcing via sulfate aerosols scattering incoming sunlight, but also how symptoms of the changing climate might influence the frequency of volcanic eruptions.


The story goes like this: as the cryosphere is destabilized, landlocked ice sheets will melt. In places with very thick ice sheets, this will produce a very abrupt (on a geological timescale) shift in lithospheric pressure loading. As one might infer from figure 1, below, such a shift might significantly enhance the fraction of melt partitioning in the vicinity. More melt, of course, translates to more frequent volcanism, though there isn’t an instantaneous response — it takes centuries for the new magma to reach the surface.


Figure 1: Example of decompression melting. An abrupt change in surface loading, (e.g., isothermal decompression) provokes a phase transition.  (adapted from Arham Bahar / Universiti Malaysia Kelantan)

As I was reading this review I was reminded of a work relating another lesser-known link between climate change and volcanism. Aubry et al. (2016) explore how climate change might influence eruption column height.


Initially, a volcanic plume may be considered in loose analogy to a molecular beam — the expansion of a high pressure gas into a vacuum through an aperture. The expanding volcanic gas carries with it some momentum, thrusting itself upward. At some point, friction with (and entrainment of) the surrounding atmosphere counters this momentum. Depending on the strength of the eruption, the gas thrust region may extend from the crater rim to several km above the crater rim. Above the gas thrust region, convective thrust drives vertical transport within the eruption column. Air parcels within the eruption column are considerably buoyant and will convect to their point of neutral buoyancy. Finally, at this point, as in a cloud top, the parcels will diffuse horizontally, forming the umbrella cloud.

The maximum height of an eruption column is limited by the amount of heat transferred to the atmosphere per unit time (power, Watts), as it is convection which ultimately drives this process. Ultimate column height can be expressed according to the following equation, in which N is the the Brunt–Väisälä frequency, M is the mass ejection rate, and horizontal transport coefficients k1 + k2 = 1.


The Brunt–Väisälä frequency is dependent on the lapse rate (\Gamma = -\frac{\text{dT}}{\text{dz}} ), the specific heat capacity \text{c}_\text{p} , the gravitational constant g, and the zonal temperature T.
\text{N}^2 = \frac{\text{g}}{\text{T}}\left(\frac{\text{g}}{\text{c}_\text{p}}-\Gamma \right)
In order for a volcanic eruption to influence global climate (and stratospheric chemistry), its eruption column must penetrate the tropopause (defined as the point of inflection in the lapse rate) — and climate change is expected to produce changes in both the tropopause height and the lapse rate. In the future, in most regions, the Brunt–Väisälä frequency will decrease. In turn, the plume height for any given eruption mass ejection rate will be shorter. Aubry et al. (2016) conclude that stratospheric input of volcanic gases will be less frequent in a future climate change scenario.


Figure 2: Lapse rate (-\frac{dT}{dz} ) as a function of year within the RCP 6.0 emissions storyline. Note that lapse rate increases as the climate changes. (Klobas thesis 2018)

Finally, since we’re on the topic of unexpected climate-change/volcanism interactions, I’d like to bring attention to a work by Laakso et al. (2016) in which they simulate what might happen if a large volcanic eruption were to occur during albedo modification geoengineering by stratospheric aerosol injection –aka solar radiation management (SRM).


Figure 3 below provides the total aerosol burden of sulfate in Tg. For controls, they simulate a volcano-only scenario (black solid), a SRM-only scenario (black dashes), the sum of the SRM-only and volcano-only scenarios (magenta dashes). There are two experimental perturbations: SRM is immediately stopped when the volcano erupts (blue solid) and SRM continues as normal despite the volcanic eruption (red solid).

Surprisingly, the time required to recover to the pre-eruption sulfate burden during SRM is halved relative to the volcano-only case, even if the SRM program continues as normal after the eruption. Why? Because coagulation processes will produce larger aerosols more quickly, reducing the lifetime of sulfate aerosol via gravitational sedimentation. They go on to show that the cessation of SRM following a volcanic eruption could cause very rapid surface warming as a result of reduced aerosol lifetimes. The authors obviously explore more than this — and it’s open access — so I encourage you to read it yourself for more info.