Using two-dimensional distributions to inform the mixing state of soot and salt particles produced in gas flares

Highlights

• Mixing state of a soot-sodium chloride aerosol from a turbulent flame is described

• TEM images indicate salt is in primary particles decorating the soot aggregates

• Mass-mobility distributions show distinct salt and soot-containing particle modes

• CPMA-SP2 measurements clearly indicate internal mixing of soot and salt

• mp-mrBc distribution appear to have shapes derived from coagulation dynamics

Abstract

Gas flaring is a common practice in the oil and gas industry, where droplets of flowback water with varying levels of dissolved salts (mainly composed of sodium and chloride) often become entrained in the flared gas. In this study, we examine the mixing state of the aerosol produced by a laboratory flare with and without entrained droplets of sodium chloride solutions. The resultant aerosol is cross-examined using several different methods, including: transmission electron microscopy (TEM), tandem measurements using a CPMA and a differential mobility analyzer (DMA), tandem measurements using a centrifugal particle mass analyzer (CPMA) and a single particle soot photometer (SP2), and Raman spectroscopy. A focus is placed on two-dimensional distributions of properties and the kind of morphological information contained therein. The TEM and CPMA-SP2 measurements both show that the majority of soot particles were internally mixed with salt, while TEM and CPMA-DMA measurements indicate that there are also a large number of isolated salt particles.

Introduction

Gas flaring is a process used to dispose of undesired or economically nonviable gas produced across the oil and gas industry, including for safety, maintenance, and testing purposes at extraction sites and in industrial processes (e.g., landfills, coal mines). Global gas flaring volumes between 2014 and 2018 consistently exceeded 140 billion cubic meters per year (Global Gas Flaring Reduction Partnership, 2019), amounting to an estimated 3.5% of global natural gas consumption (Elvidge, Zhizhin, Baugh, Hsu, & Ghosh, 2016). An estimated 90% of flared gas is expected to have occurred at upstream exploration and production facilities (Elvidge et al., 2016), where solution gas – i.e., natural gas dissolved in crude oil – is simply burned in pipe flares. Solution gas is predominantly composed of methane, which has a global warming potential of 28–34 times higher than that of CO₂ on a 100-year timescale (Elvidge et al., 2018; Myhre et al., 2013). Thus, flaring is favoured over venting, where flare gas is simply released to the atmosphere. Nevertheless, flaring is a significant source of gas and particle emissions and contributes around 350 million tons of CO₂ to the atmosphere (Elvidge et al., 2016). Black carbon produced by flaring may further speed the effects of climate change by directly interacting with light in the atmosphere (Bond et al., 2013) and by acting as nuclei for clouds (Spracklen, Carslaw, Pöschl, Rap, & Forster, 2011) (effects that are potentially large but uncertain which hampers climate change modeling and contributes to policy inaction). Black carbon can also reduce the albedo of earth's surface in arctic and subarctic regions (Flanner et al., 2007, 2009; Quinn et al., 2008; Sand et al., 2016). An estimated 42% of deposition of soot on sea ice may result from gas flaring (Stohl et al., 2013), an issue that may become more pressing as oil is increasingly extracted in these regions (Gautier et al., 2009).

Common practices at upstream exploration and production sites can change the particle emissions from flares. At these sites, a mixture of crude oil, solution gas, formation brine, and/or fracturing fluid is brought to the surface. Fracturing fluids are injected into shale formations at high pressure to release oil or gas reserves (Gallegos & Varela, 2015). Subsequent to fracturing, the flowback water – a mixture of formation brine and fracturing fluid – must be cleared from the well (Barbot, Vidic, Gregory, & Vidic, 2013). The remaining liquid flowing from the well after production is known as produced water. The mixture of crude oil, solution gas, and flowback/produced water are segregated in mechanical separators, where the droplets of flowback water are likely to become entrained in the flared solution gas (Milani, Jefferson, & Johnson, 2019), adding new chemical species to the resulting emissions. The two most abundant components in flowback/produced water are sodium and chloride (Barbot et al., 2013; Chapman et al., 2012; Haluszczak, Rose, & Kump, 2013; Ziemkiewicz & He, 2015), which have the potential to affect soot formation, aggregation, nanostructure, hygroscopicity (e.g., researchers have shown that small amounts of sodium chloride mixed with soot particles can significantly improve the ability of soot particles to act as nuclei for liquid clouds (Dusek, Reischl, & Hitzenberger, 2006)), and overall emission rates from flares.

In fact, laboratory studies have long established a connection between the addition of metal salts and soot formation and soot emission. In particular, the alkali and alkaline earth metal salts, including sodium chloride, have been studied for several decades, mostly using in situ (in-flame) optical measurements (Bonczyk, 1983, 1988; Di Stasio, LeGarrec, & Mitchell, 2011; Haynes, Jander, & Wagner, 1979; Mitchell & Miller, 1989; Simonsson et al., 2017, 2018; Tappe, Haynes, & Kent, 1993) ex situ electron microscopy (Bladt, Ivleva, & Niessner, 2014; Bulewicz, Evans, & Padley, 1975; Kazemimanesh, Kuang, Kostiuk, & Olfert, 2020; Moallemi, Kazemimanesh, Kostiuk, & Olfert, 2018; Simonsson et al., 2017, 2018), or ex situ (out-of-flame) particle sizing (Moallemi et al., 2018; Kazemimanesh et al., 2020). Notably, Haynes et al. (1979) performed some of the first laboratory studies of this kind, showing that the addition of NaCl reduces the coagulation rate of soot primary particles within the flame, resulting in smaller aggregate soot particles with a higher number concentration within the flame. Similar observations have been reported in more recent studies (Kazemimanesh et al., 2020; Moallemi et al., 2018; Simonsson et al., 2017, 2018). Kazemimanesh et al. (2020) subsequently reported that NaCl addition results in the inhibition of soot oxidation near the flame tip, which resulted in larger particles (and higher mass concentrations) in the post-flame region. That same study also showed that NaCl particles homogenously nucleate downstream from the flame and coagulate with soot.

Such fundamental studies are complimented by investigations of the particulate emissions and mixing state of salt-containing aerosols downstream from the flame. Laboratory-scale turbulent flames with entrained droplets of salt solutions have been used to investigate the size distribution and morphology (Kazemimanesh et al., 2021) of particle emissions expected from flares. These studies demonstrated that salt can be internally mixed with soot particles in the aerosols produced by these flames. Complimentary studies have investigated aerosols formed by the mixing of sea spray aerosol with soot (Laskin et al., 2012; Li et al., 2011) and core-shell sodium chloride-soot particles generated in the laboratory (Choi, Stipe, Koshland, & Lucas, 2006). Motivated by the latter studies, several studies have considered the effects that changes in particle morphology can have on the particle's climate warming potential. For instance, modelling indicates that salts may also have a lensing effect¹ when internally mixed with soot (Scarnato, Vahidinia, Richard, & Kirchstetter, 2013). Within this context, the distribution of salt on the particle, which may vary between the atmospheric and flare scenarios, has significant consequences for the optical and cloud condensation nuclei (CCN) (Dusek et al., 2006) properties, requiring detailed characterization of the mixing state of flare-generated soot.

In this study, we seek a detailed experimental characterization of the morphological properties and mixing state of soot and salt particle emissions from a large-scale, turbulent flame, used as an approximation to flares. Aqueous solutions of sodium chloride were atomized in a custom pipe flare and entrained in a turbulent flame. Since only sodium chloride is considered in this work, hereafter salt will refer specifically to sodium chloride. Propane at a constant flow rate was considered throughout the experiments while the concentration of salts in the aqueous solutions was varied from zero (deionized water) up to 3% (by mass). Particle morphology was characterized using a suite of instrumentation, with a particular focus on tandem measurements. This includes tandem centrifugal particle mass analyzer (CPMA)-differential mobility analyzer (DMA)-condensation particle counter (CPC) measurements, which are used to map out mass-mobility distributions (Buckley, Kimoto, Lee, Fukushima, & Hogan, 2017; Rawat et al., 2016; Sipkens, Olfert, & Rogak, 2020a); and CPMA-single particle soot photometer (SP2) measurements, which are used to map out refractory black carbon-total particle mass (Broda et al., 2018; Naseri, Sipkens, Rogak, & Olfert, 2020) distributions. These two-dimensional distributions, as opposed to single parameter size distributions and summary parameters (e.g., average effective density), are a natural interpretation of tandem measurements and reveal increasingly detailed information about the particle morphology, allowing practitioners to resolve distinct soot-containing and salt particle modes. Transmission electron microscopy (TEM), CPMA-SP2, and CPMA-DMA-CPC are used to assess the mixing state (i.e., whether the soot and salt particles exist as independent populations or the particles contain both soot and salt). Changes in soot nanostructure are assessed using Raman spectroscopy.