Characterization of Emissions and Residue from Measures to Improve Efficiency of In Situ Oil Burns

In-situ oil burns are often used to mitigate the potential environmental impact of floating oil originating from accidental/unintentional oil spills at sea. The resulting emissions have been characterized primarily for safety concerns related to worker inhalation exposure (Lane, 2011; Barnea, n.d.) and environmental pollution (Devai et al., 1998; Fingas, 2014). The black plumes resulting from combustion by-products have been characterized with respect to aerosol properties (Perring et al., 2011; Middlebrook et al., 2012), particulate matter equal to and less than 2.5 μm in mass median diameter (PM_2.5), and pollutants such as polycyclic aromatic hydrocarbons (PAHs), volatile organic compounds (VOCs) such as benzene, and trace amounts of polychlorinated dibenzodioxins and dibenzofurans (PCDDs/PCDFs) (Aurell and Gullett, 2010; Gullett et al., 2017). PM_2.5 is a criteria pollutant regulated by U.S. EPA as these small particles can be inhaled and cause adverse health effects to humans. Naphthalene, benzene, and PCDD/PCDFs can all be found on EPA's list of hazardous air pollutions (HAPs) (2008) as they have properties that are harmful to humans, marine life, and the environment.

Pollutants from in situ oil burns have been minimally quantified as emission factors (pollutant mass per mass of oil burned), with some exceptions (for example, (Ross et al., 1996)), limiting our ability to compare combustion efficiencies and technology improvements. Emission factors for PM_2.5 and VOCs from open burning sources have been found to decrease with increased modified combustion efficiencies (MCE) (Aurell et al., 2015; Aurell et al., 2017). For example, the PM_2.5 emission levels from burning of piles of timber slash were reduced five times when the MCE increased from 86.1% to 96.4% (Aurell et al., 2017).

This work aimed to characterize pollutants and unburned residues from in situ oil burns using variations in fire boom configurations, air-assist nozzle angles, and the presence or absence of waves.

A wave tank of interior size 14 m × 2.4 m × 2.4 m (47 ft × 8 ft × 8 ft) located at U.S. Army Corps of Engineers, Cold Regions Research and Engineering Laboratory (CRREL) in Hanover, New Hampshire was used for the testing. The wave tank contained approximately 63 m³ (16,700 gallons) of fresh water with a water level of 2.0 m (6.5 ft). The waves generated had a 12 cm amplitude and a 1.5 s period. A fire boom was positioned on the water surface and pre-weighed (approximately 31 kg (68 lbs) or 35 L (9.25 gallons)) Alaska North Slope crude oil was place inside the fire boom (Figure 1) to simulate in situ oil burning. The fire boom was equipped with compressed air nozzles to optionally supply additional air to the in situ oil burn. The crude oil was ignited with a propane torch and the burns lasted for approximately 5 minutes. After a cooling period the residue was collected from the wave tank and weighed to derive the oil mass loss.

Three different boom configurations varied the length/width aspect ratios to 1:1, 4:1 and 9:1 (Figure 1). The enclosed area by the fire boom was kept constant at 3.4 m² resulting in the same crude oil thickness for each of the boom configurations. The air-assist nozzles were configured in three different angles; across the oil (90°), up over the oil (+45°), and down toward the oil (−45°). Emissions were measured from the different boom configurations when tested with or without waves, with or without air supply, as well as a combination of waves and supply of air resulting in fourteen different test setups. A total of sixteen burns were conducted with one of the setups replicated three times to establish the reproducibility of the tests. Due to the number of factors tested and limited resources such as funding and time, only one test setup was replicated.

The target pollutants, shown in Table 1, were sampled using an instrument system that was mounted on an aluminum skid which was suspended from a crane to easily maneuver the system into the burn plume. The carbon dioxide (CO₂) and carbon monoxide (CO) sensors were calibrated daily in accordance to EPA Method 3A (2017). PM_2.5 was collected using SKC's single-stage IMPACT Sampler (SKC Inc., USA) which collects PM_2.5 on a 47 mm teflon filter with a pore size of 2 μm via a Leland Legacy pump (SKC Inc., USA) with a constant flow rate of 10 L/min. The PM filters were gravimetrically measured following procedures in 40 CFR Part 50, Appendix L (1987). The total carbon (TC), organic carbon (OC), and elemental carbon (EC) were collected on a 37 mm quartz filter using SKC's PM_2.5 single-stage Personal Modular Impactor (SKC Inc., USA) via a Gilian 5000 sampling pump (Sensidyne LP, USA) with a constant flow rate of 3 L/min. The TC/EC/OC filter samples were analyzed following NIOSH Method 504. (1999) and Khan et al. (2012). The PM_2.5 and TC sampling pumps were calibrated with a Go-cal Air Flow Calibrator (Sensidyne LP, USA) prior to sampling.

Post-burn residue samples were collected with pre-weighed oil absorbent pads (New PIG Corp., Tipton, PA, USA) to determine oil mass loss and total petroleum hydrocarbons (TPH) using Gas Chromatography (GC) with Flame Ionization Detection (FID) using modified US EPA Method 8015D (n.d.). The TPH concentration was calculated using a six-point calibration curve generated with the unweathered oil used for the tests.

The oil mass loss was derived by subtracting the residue weight from the initial crude oil weight in each of the tests. The modified combustion efficiency (MCE_T) in this study includes TC from particles (CO₂/(CO₂+CO+TC)) and was calculated for each of the tests to establish how well the oil burned in each of the different test conditions. The carbon mass balance approach was used to determine emission factors for each of the measured pollutants. The carbon mass balance assumes non-differential mixing of the plume pollutants. The sampled pollutant mass is divided by the co-sampled carbon mass (determined from CO and CO₂ measurements and particulate total carbon) and then multiplied by the carbon fraction of the fuel, 0.85 (Aurell and Gullett, 2010; Nelson, 1982). The derived emission factors are expressed as mass of pollutant per mass of initial oil. TPH in the residue is expressed as mass TPH per mass residue.

The results showed a linear (R² of 0.839) decrease of PM_2.5 emission levels with increased MCE_T values (Figure 2). Figure 2 also indicates that the 9:1 Boom Aspect Ratio (red markers) had lower PM_2.5 emission levels than the 1:1 Boom Aspect Ratio (yellow markers). By plotting the MCE_T and PM_2.5 emission levels as a function of Boom Aspect Ratio as shown in Figures 3a and 3b, it becomes apparent that the higher Boom Aspect Ratio results in higher MCE_T values and lower PM_2.5 emission levels regardless of adding additional injection air or in the presence of waves. The Boom Aspect Ratio effect is most probably due to air penetrating more efficiently into the thinner fuel rich zone of the 9:1 versus the 1:1 configuration (see Figure 1). The rate and extent of the combustion is limited by diffusion of oxygen into the region of volatilized fuel so the shorter penetration distance on the 9:1 configuration allows for more efficient combustion. The PM_2.5 emission factor for the replicate runs had a relative standard deviation (RSD) of 8% which indicates good reproducibility of the conducted tests. This RSD is lower than from a previous in situ oil burn study (25%) when a test scenario was replicated 14 times (Gullett et al., 2017).

Oil mass loss varied from 88.8% to 99.6% (Table 2). A larger mass loss percentage was found with higher Boom Aspect Ratio but only when waves were absent (Figure 3c). The waves are speculated to circulate water which cools the bottom layer of the oil forcing the oil to reheat for every passing wave. This reduces the heat transfer to the oil and the resulting oil vaporization, limiting combustion efficiency. The addition of injection air did not have any large effect on the mass loss as the Control tests had almost the same percentage as the Air tests. The three replicate runs had an RSD of 1.6% indicating very good reproducibility.

The TPH in the residue were found to decrease with increased percent oil mass lost (Figure 4a). No direct correlation could be found plotting the TPH as a function of Boom Aspect Ratio and MCE_T as shown in Figures 3d and 4b, respectively. This suggests that the oil vaporization (and TPH concentration) was not related to the subsequent air/fuel mixing and combustion efficiency. Although, a higher TPH concentration in the residue was found in tests when waves were added to Boom Aspect Ratios 4:1 and 9:1 indicating a decline in the percent of oil mass lost (Figures 3c and 3d). The TPH values are more related to the extent of burning (mass loss %) than the burn quality (MCE_T). When a greater fraction of original oil is burned (high mass loss), higher loss of organics occurred resulting in lower TPH concentration in the residue. For example, combustion quality may be very good (high MCE_T) but very little of the oil was burned (low mass loss). With low mass loss or extent of oil burned, a lower fraction of the oil will have been devolatilized and combusted, resulting in higher quantities of organics remaining in the residue (higher TPH concentration.. A similar trend was observed in previous studies with lower mass loss resulting in higher TPH and PAH concentrations in the post-burn residue (Wang et al., 1999; Fingas, 2017). The increase of TPH in the residue when waves were added suggests that the waves negatively affect mass loss. The waves result in the oil slick oscillating lower into the water, resulting in greater heat loss to the water and less combustion efficiency. The TPH RSD for the replicate runs was 7% which indicates good reproducibility of the conducted test condition.

The PM_2.5 emission factor was found to decrease with increased combustion efficiency. The MCE_T and PM_2.5 emission factors ranged from 85% to 93% and 57 to 154 g/kg initial oil, respectively, corresponding to about three times lower PM_2.5 emission factor from the highest MCE_T to the lowest. Higher Boom Ratio configurations resulted in higher MCE_T regardless of added air or presence of waves.

Oil mass loss ranged from 88.8% to 99.6% where a higher oil mass loss occurred with higher Boom Aspect Ratios except when waves were present which always lowered the oil weight loss. The TPH in the residue was found to decrease with increased percent oil mass lost but no such correlation was found with MCE_T values. A higher TPH concentration was found when waves were added to the larger Boom Aspect Ratios.

This work was funded through an interagency agreement between the Department of Interior, Bureau of Safety and Environmental Enforcement and the U.S. EPA's Office of Research and Development. Messrs. Dale Greenwell, Bill Mitchell, and Dennis Tabor of U.S. EPA/ORD contributed critically to sampling, instrumentation, and analyses. Messrs. Brandon Booker and William Burch ably assisted with test execution and sampling.