Characterizing emissions from open burning of military food waste and ration packaging compositions

Abstract

Emissions from open burning of military food waste and ration packaging compositions were characterized in response to health concerns from open burning disposal of waste, such as at military forward operating bases. Emissions from current and prototype Meals, Ready-to-Eat (MREs), and material options for their associated fiberboard packaging were quantified to assess contributions of the individual components. MREs account for 67–100% of the particulate matter (PM), volatile organic compounds (VOCs), polycyclic aromatic hydrocarbons (PAHs), and polychlorinated dibenzo-p-dioxins and -furans (PCDDs/PCDFs) emissions when burned in unison with the current fiberboard container and liner. The majority of the particles emitted from these burns are of median diameter 2.5 μm (PM_2.5). Metal emission factors were similar regardless of waste composition. Measurements of VOCs and PAHs indicate that targeted replacement of MRE components may be more effective in reducing emissions than variation of fiberboard-packaging types. Despite MRE composition variation, equivalent emission factors for PM, PAH, VOC, and PCDD/PCDF were seen. Similarly, for fiberboard packaging, composition variations exhibited essentially equivalent PM, PAH, VOC, and PCDD/PCDF emission factors amongst themselves. This study demonstrated a composition-specific analysis of waste burn emissions, assessing the impact of waste component substitution using military rations.

Introduction

The United States (US) Department of Defense (DOD) has traditionally depended upon field-expedient methods such as open burning, air curtain burners, and incinerators as practical means to reduce volume and generally dispose of military waste in contingency environments. Open burning, commonly referred to as “burn pits”, provides relatively small forward operating bases (FOBs) the flexibility to dispose of waste without placing additional service members in a direct line-of-fire from enemy forces while transporting the FOB waste (i.e., backhauling) to a larger base with incineration capabilities [1]. By destroying the waste, it mitigates vermin problems and vector-borne disease while also preventing enemy combatants from gaining access to potential resources and intelligence. However, service members may be exposed to pollutants originating from open burning [2, 3]. Conclusions from investigations into the long-term health effects associated with open-pit burning in a deployed military setting have generally been inconclusive [4, 5].

Military rations have been a staple of the US military, since the American Revolution [6] and the current version, the Meal, Ready-to-Eat (MRE), is often relied upon for short-term land operations. A single fiberboard package and liner, as depicted in Table 1, contains 12 MREs that provide a convenient manner for the war fighter to retain adequate caloric intake to continue necessary operations [7, 8]. The utility of the MRE can be demonstrated in overall consumption: in 2004 alone, more than 144 million MREs were purchased for United States Army (USA) camp feeding activities, leading to approximately 67,000 tons of solid waste being generated [6]. With the current MREs and their fiberboard packaging considered non-recyclable, there is increased emphasis to investigate alternative materials and designs.

Table 1.

Variations of test material

From: Characterizing emissions from open burning of military food waste and ration packaging compositions

Recent research by the US Army Natick Soldier Research, Development and Engineering Center (NSRDEC) into alternative fiberboard packaging for MREs has focused on recyclability, compostability, and replacing potentially harmful chemicals used for wet strength with environmentally friendly polymeric coatings [9]. Additional research by NSRDEC is ongoing for the prospective replacement of aluminum-based MRE pouches with multi-layer, polymer-based materials that potentially provide greater barrier and mechanical properties [10,11,12]. Current aluminum-based rations undergo retort sterilization, but recent research by NSRDEC into a microwave-assisted sterilization process, which offers faster thermal penetration and better uniformity, requires an alternative to aluminum such as a polymer [6]. By improving the designs for both the fiberboard packaging and the aluminum-based pouches for MREs, a reduction in solid waste associated with the consumption of military rations is anticipated.

Few studies have been conducted, however, to specifically characterize emissions from open burning of military waste. Other open burning situations encountered within the literature, such as biomass and household waste burns, have previously characterized pollutants such as metals, volatile organic compounds (VOCs), polycyclic aromatic hydrocarbons (PAHs), polychlorinated dibenzo-p-dioxins and -furans (PCDDs/PCDFs), and particles with an aerodynamic median diameter less than or equal to 10 μm (PM₁₀) and 2.5 μm (PM_2.5) [2, 3]. Many of these VOCs are on the US Environmental Protection Agency’s (EPA) list of hazardous air pollutants (HAP) [13] due to human toxicity concerns of varying degrees and are classified as a possible, probable, or confirmed carcinogen to humans (Group “2B” or greater) by the World Health Organization’s (WHO) International Agency for Research on Cancer (IARC) [14]. The EPA’s 16 PAH priority pollutants have been investigated for more than three decades in various capacities, with some of these pollutants having mutagenic and carcinogenic properties [15, 16]. PCDDs/PCDFs are of concern due to their toxicity and carcinogenicity, and are recognized as bioaccumulative and persistent within the environment [17]. PM_2.5 has been known to cause adverse health effects [18] and is a criteria pollutant regulated by the US EPA.

The objectives of this study were to: (1) characterize the emissions from current and prototype MRE pouches and MRE food residuals along with variations of their fiberboard-packaging material and (2) compare derived emission factors for the various packaging materials against published values. The two variables studied included: (1) the variation by mass of MRE aluminum foil-based pouches with multi-layer polymer-based pouches and (2) the presence or absence of a polymer that provides strength in high moisture conditions (i.e., wet strength) on both the current solid and prototype-corrugated fiberboard-packaging designs for MREs. To our knowledge, this is the first study in the literature to characterize emissions and emission factors from MREs and their packaging separately from a comprehensive military waste stream.

Materials and methods

Material compositions

The aluminum/polymer composition of an MRE was varied by mass for this experiment and is expressed as the percentage of aluminum content replaced. For example, MRE-32% indicates that 32% of aluminum-based components of an MRE were replaced by mass with components composed of polymers [10]. Therefore, as depicted in Table 1, four distinct MRE materials were created and tested for this experiment: MRE-0%, MRE-32%, MRE-66%, and MRE-100%. Variations of the fiberboard packaging that contain the 12 MREs were also studied. Two fiberboard structures were utilized within this experiment: the current MRE’s solid design and a prototype coated, corrugated design [9]. In addition, these two fiberboard structural designs were characterized as having a wet-strength component either present or absent. Therefore, four distinct fiberboard-packaging materials were also tested and are also shown in Table 1.

In addition, the aluminum-based pouches consist of a five-layer structure containing Kurarister N™, a transparent, high-gas barrier, retortable film that utilizes organic and inorganic composites coated on oriented polyamide, while the prototype polymer-based pouches rely on only three polymeric layers.

Test setup

Fiberboard-packaging tests were conducted in November 2014, while MRE tests were conducted during July–August 2015. All tests were conducted at the US EPA’s National Risk Management Research Laboratory Open Burn Test Facility (OBTF) located at Research Triangle Park, North Carolina and consisted of simulated open burning. The OBTF has been described and utilized previously to study open burning conditions [3, 19,20,21] and was used in a similar experimental setting capacity (see Online Resource 1, Fig S1).

The sampling package, Fig. 1, consisted of PM_2.5, PM₁₀, VOC, and semi-volatile organic compound (SVOC) batch samplers (for PAHs and PCDD/Fs) as well as direct-reading instruments (not shown) to continuously monitor carbon dioxide (CO₂), carbon monoxide (CO), and temperature fluctuations. The sampling package operated on batteries providing 48 V and was remotely controlled for sampler operation, while data were continuously logged to the computers. Further details on the sampling package, frequently referred to as “The Flyer,” can be found in the literature [19, 20, 22,23,24] and are, therefore, only briefly described here. PM₁₀ and PM_2.5 impact samplers, using 47 mm Teflon® filters with a pore size of 2 μm, were connected to Leland Legacy pumps (SKC Inc., USA) with Tygon® tubing utilizing a flow rate-compensating pump [25] with a nominal flow rate of 10 L-min⁻¹. Filters for PM_2.5 were also analyzed by X-ray fluorescence (XRF) spectroscopy [26] for elements. VOCs were sampled with SUMMA® canisters, in accordance with US EPA Method TO-15 [27] with a 12-min sampling time. The SUMMA® canister was also analyzed for CO, CO₂, and CH₄, following US EPA Method TO-25C [28]. Quartz filters to collect PCDD/PCDFs were placed inline with sampling media and polyurethane foam (PUF), followed by the high volume pump with a flow rate of 850 L-min⁻¹ [23]. Quartz filters, PUF, and a sorbent (XAD-2) were used to collect PAHs, with a high volume pump of a flow rate of 160 L-min⁻¹ [24]. PM impact samplers and the high volume blower quartz filters were replaced when the pumps could no longer maintain the nominal 10 L-min⁻¹ flow rate and the high volume blower showed a reduced flow rate. PCDDs/PCDFs were analyzed utilizing High-Resolution Gas Chromatography/High-Resolution Mass Spectrometry (HRGC/HRMS) [29] and PAHs by Low-Resolution Gas Chromatography/Low-Resolution Mass Spectrometry (LRGC/LRMS) [20, 30]. CO₂ and CO were continuously measured by the use of a LICOR 820 (Biosciences, USA), a benchtop CO monitor (California Analytical Instruments Model 200, USA) as well as an e2V EC4–500-CO sensor (SGX Sensortech, United Kingdom). Pre-calibration as well as drift check of the CO₂ and CO sensors were conducted daily in accordance with US EPA Method 3A [31]; calibration was conducted for all other equipment in accordance with manufacturer recommendations. A DustTrak DRX 8533 (TSI Inc., USA) was used for continuous measurements of PM₁, PM_2.5, PM₄, PM_10, and total PM (up to 15 μm). A custom correction factor was applied, as per manufacturer’s recommendation, to the DustTrak-derived PM concentrations using simultaneously sampled PM_2.5; PM₁₀ concentrations (i.e., average continuous PM_2.5 or PM₁₀ concentrations were divided by the PM_2.5 or PM₁₀ filter impactor concentrations during the same time interval).

Fig. 1.

OBTF sampling package for PM, VOC, PAH, and PCDD/PCDF emissions

Test samples

Fiberboard test samples consisted of approximately 2.5 kg of material, which equated to two containers (plus liners) for each fiberboard test material. Each MRE test sample consisted of six MREs, with all MREs consisting of the same entrée, plus half of a current fiberboard container and liner (see Online Resource 1, Table S3 for content). All MRE components were opened and randomly dispersed upon the OBTF burn area, while 25% (by mass) of each MRE component (e.g., entrée, bakery items, snacks, etc.) and their individual packaging were included to simulate uneaten food and to be consistent with US Central Command guidance of wet waste not exceeding a quarter of the waste input [32]. All tests were conducted in a random order and a summary of the testing is shown in Table 2.

Table 2.

Test matrix

From: Characterizing emissions from open burning of military food waste and ration packaging compositions

	Test materiala	# of tests	Test order
Fiberboards	NA: OBTF blank	1	Pre-test
	SFP	4	3,5,7,8
	SFNP	4	1,2,12,13
	CFP	4	6,9,10,11
	CFNP	4	4,14,15,16
	NA: OBTF blank	1	Post-test
MREs	NA: SFP (baseline)	1	6
	MRE-0%	3	2,3,5
	MRE-32%	3	7,8,10
	MRE-66%	3	11,12,13
	MRE-100%	3	1,4,9
	NA: OBTF blank	1	Post-test

^aNA not applicable, SFP solid fiberboard container and liner with wet-strength polymer; SFNP solid fiberboard container and liner, no wet-strength polymer; CFP corrugated fiberboard container and liner with wet-strength polymer; CFNP corrugated fiberboard container and liner, no wet-strength polymer

The MRE flameless ration heaters (FRHs) were excluded from the experiment design due to their volatility and potential for injury [33]. The FRH primarily consists of a uniformly blended matrix of super-corroding magnesium-iron alloy powder (functioning as anode–cathode, respectively) and an electrolyte with flow and wetting agents. When this electrolyte comes into contact with water, a rapid corrosion of the magnesium particles within the matrix produces small amounts of magnesium hydroxide and gaseous hydrogen, as well as a temperature increase of at least 100 °F within 12 min [34]. Had the FRHs been included in the test samples, it is speculated that the corrosion process would have prematurely initiated, contributing a significant increase in heat and hydrogen available for combustion, as well as potentially increasing the magnesium and iron particulates found within the PM_2.5 samples [6, 34].

Burns

Tests were initiated by weighing each test sample, randomly dumping the test sample on the OBTF burn pan, and igniting the sample with a propane torch (see Online Resource 1, Fig S1 thru Fig S3). To simulate open burning conditions, the oxygen concentration was maintained at atmospheric levels by ventilating with approximately 40 air changes per hour through an exhaust duct. The number of tests per material was dictated by the amount of sample mass required to exceed detection limits of the most trace target compounds, typically PCDDs/PCDFs. This amount was estimated during the tests by the mass of carbon collected, as CO₂, as a surrogate for the amount of combustion products sampled. Data collection began once the test material was ignited (see Online Resource 1, Fig S3). SUMMA® canisters were started and stopped at a sufficiently high CO₂ concentration of 1200 ppm to avoid diluting samples with non-detectable levels of the target compounds. Sampling equipment was remotely stopped when the burns neared completion of smoldering, indicated by CO₂ levels approaching ambient levels of 450 ± 50 ppm [35]. Sampling durations were approximately 15 min for each burn; sampling was initiated upon ignition and terminated after flaming extinction and when smoke generation was minimal.

Calculations

Modified combustion efficiencies (MCE), an indicator of the effectiveness of a combustion process [36,37,38], were calculated for each sample collected: (1) CO₂/(CO₂ + CO + CH₄) for VOCs and (2) CO₂/(CO₂ + CO) for all other pollutants [39]. In addition, toxicity equivalent (TEQ) values were calculated using the World Health Organization 2005 toxic equivalent factors (TEFs) for PCDD/PCDFs [40] and TEFs relative to benzo-[a]-pyrene TEQ for PAHs [41].

Emission factors (EF) were calculated utilizing the carbon balance method [42], which assumes that all carbons in the fiberboard or MRE are emitted as CO₂, CO, CH₄, and total hydrocarbons and is thoroughly mixed with the pollutants in the smoke plume [39]. The emission factor for a pollutant, EF_pollutant, could, therefore, be determined by

$${\mathrm{EF}}_{\mathrm{pollutant}}=\frac{m_{\text{f}}+m_{\mathrm{pollutant}}}{m_{\mathrm{Total\; carbon}}}$$ (1)

where m_f = mass fraction of carbon in the respective fiberboard or MRE.

m_pollutant = above-ambient mass concentration of the pollutant of interest (typically mg pollutant m⁻³).

m_{Total carbon} = above-ambient mass concentration of carbon (typically mg carbon m⁻³) determined by the measurement of major carbon-containing combustion products [42].

Since CO₂, CO, and specifically CH₄ could be accounted for in the SUMMA® canister analysis for VOCs, all three were utilized in calculating m_{Total carbon} for VOC emission factors, while CO₂ and CO were used for all other pollutants [42]. The emission factors can, therefore, be expressed as mass of pollutant per mass of the fiberboard or MRE burned, as appropriate, hereby referred to collectively as “waste” (e.g., mg pollutant per kg waste). mfmf for the fiberboards were obtained from ultimate proximate analyses and were determined to be 0.46 for SFP, 0.47 for SFNP, 0.45 for CFP, and 0.46 for CFNP (see Online Resource 1, Table S2 for the ultimate proximate analysis). Mass fractions of the carbon for the various MRE test materials were derived by the use of literature values [43], the list of contents on the food pouches, and the fraction of each component within the MREs and were determined to be 0.42 for MRE-0%, 0.44 for MRE-32%, 0.43 for MRE-66%, and 0.46 for MRE-100% (see Online Resource 1, Table S3 and Table S4 for content breakdown).

Statistical analyses

Statistical analyses were accomplished utilizing the statistical software JMP® 12 (SAS Inc., USA). A confidence level of 95% (overall α = 0.05) was determined to be an acceptable threshold for statistical significance prior to conducting the experiments. Numerous multiway analysis of variance (ANOVA) tests were conducted to screen for significant differences in emission factors by material type and MCE. If the ANOVA test indicated a possible statistical distinction, verification of the assumptions required for an ANOVA was conducted, and if held true, then a Tukey honest significant difference test was conducted to show which, if any, of the test materials were statistically different from one another. To compare between the MREs and the fiberboards as a whole, a non-parametric Kruskal–Wallis (K–W) rank sums test was also performed on the aggregate data for each pollutant to account for unequal sample sizes.

Results and discussion

Two hundred and fourteen samples were collected in total for PM_2.5, PM₁₀, PAHs, VOCs, and PCDD/PCDF. When the current fiberboard and liner were burned in unison with the MREs, MRE emission factors could be calculated by the amount of fiberboard and liner present by mass, determined by ultimate proximate analysis and expressed in the form of a mass fraction. A mass balance was then conducted utilizing the mass fractions of the fiberboard with respect to the various MRE types to determine true MRE emission factors. Once calculated, the true MRE emission factors were compared to combined MRE and fiberboard emission factors to determine the contribution percentages of the MREs to overall emissions. MREs were found to account for 89–92% of PM, 84–99% of VOC, 67–96% of PAH, and 78–100% of PCDD/PCDF emissions when burned in unison with the current fiberboard container and liner. The results that follow compare MRE/fiberboard emission factors and modified combustion efficiencies to each other and the literature, presented by pollutant class.

PM and metals

Tukey tests amongst the fiberboards indicated a statistical difference between SFNP versus CFNP and CFP, as well as between SFP and CFP (p value = 0.0001) for PM₁₀. The same can be said for PM_2.5, as SFNP was statistically different from CFP (p value = 0.0109). Variability amongst the fiberboards, however, precludes a significant distinction from being identified, as Fig. 2 illustrates similar particulate matter emission factors regardless of fiberboard composition.

Fig. 2.

PM_2.5 and PM₁₀ emission factor outlier box plots from open burning of MRE and fiberboard test materials. All plots identify quartiles with outliers illustrated as dots

Figure 2 shows PM2.5 and PM10 emission factor outlier box plots from open burning of MRE and fiberboard test materials. All plots identify quartiles with outliers illustrated as dots.

PM_2.5 emission factors for the various MRE types were not determined to be statistically different (p value = 0.4073). A PM₁₀ statistical distinction was found for MRE-0% and MRE-100% versus MRE-66% (p value = 0.0051) and is believed to be due to the relatively low PM₁₀ emissions from MRE-66%. As Fig. 2 illustrates and not unlike the fiberboards, however, particulate matter emission factors for the MREs were essentially equivalent regardless of composition. Similar conclusions could be drawn for the various MRE types regarding PAHs, VOCs, and PCDD/PCDF emission factors, which were also effectively equivalent.

When comparing the PM emission factors for the MREs against the fiberboards, there is a significant difference: the MREs emitted up to five times greater PM emissions than the fiberboards (K–W p value = 0.0001) (see Fig. 2; Table 3). Furthermore, PM₁₀ and PM_2.5 emission factors were also closely correlated with one another across all material types, indicating that a majority of particulate emissions were 2.5 μm in size or smaller, regardless of material.

Table 3.

Select average emission factors ± 1 standard deviation for PM_2.5, PM₁₀, VOCs, and PCDD/PCDF^a

From: Characterizing emissions from open burning of military food waste and ration packaging compositions

Pollutant	Unit	SFP	SFNP	CFP	CFNP	MRE-0%	MRE-32%	MRE-66%	MRE-100%
PM
PM2.5	g-kg Waste−1	2.9 ± 0.63	3.4 ± 2.0	2.8 ± 0.86	3.4 ± 0.72	12.3 ± 3.3	10.5 ± 0.7	10.4 ± 1.8	11.9 ± 0.7
PM10	g-kg Waste−1	2.9 ± 0.70	3.4 ± 1.7	2.7 ± 0.89	3.6 ± 0.72	13.6 ± 1.8	11.0 ± 0.9	9.5 ± 1.3	12.6 ± 0.1
VOCs
Benzene(1),b	mg-kg Waste−1	79 ± 75	111 ± 16	153 ± 102	91 ± 49	211 ± 89	266 ± 10	182 ± 12c	225 ± 36
1,3-Butadiene(1),b	mg-kg Waste−1	19 ± 22	24 ± 5	44 ± 43	34 ± 23	103 ± 38	126 ± 12	82 ± 2.3c	94 ± 55
Styrene(2B),b	mg-kg Waste−1	5.6 ± 2.5	13 ± 1.6	60 ± 64	13 ± 7.5	206 ± 192	283 ± 198	210 ± 22c	120 ± 79
Ethylbenzene(2B),b	mg-kg Waste−1	2.4 ± 2.2	4.1 ± 0.60	15 ± 13	4.6 ± 2.9	16 ± 12	24 ± 14	18 ± 0.33c	15 ± 6.0
Toluene(3),b	mg-kg Waste−1	19 ± 19	26 ± 4.6	57 ± 44	28 ± 16	63 ± 36	79 ± 37	52 ± 11c	58 ± 18
Acrolein(3),b	mg-kg Waste−1	9.4 ± 4.8	36 ± 5.9	81 ± 64	76 ± 60	125 ± 40	145 ± 16	126 ± 12c	126 ± 27
PCDD/PCDF
2,3,7,8-TCDD(1),b	ng TEQ-kg Waste−1	0.003 ± 0.001c	0.006 ± 0.002c	0.028 ± 0.018	0.013 ± 0.002c	0.293 ± 0.236	0.064 ± 0.035c	0.033 ± 0.021c	0.065 ± 0.042
2,3,4,7,8-PeCDF(1),b	ng TEQ-kg Waste−1	0.009 ± 0.003	0.020 ± 0.013	0.026 ± 0.003	0.010 ± 0.001c	0.479 ± 0.293	0.265 ± 0.074	0.075 ± 0.047	0.250 ± 0.168
ΣPCDD/PCDF TEQ Total	ng TEQ-kg Waste−1	0.046 ± 0.016	0.075 ± 0.041	0.28 ± 0.14	0.10 ± 0.055	1.802 ± 1.040	0.904 ± 0.291	0.301 ± 0.145	0.877 ± 0.527

^(#) Superscript number is the International Agency for Research on Cancer classification group [14]

^aFull lists can be found within SI

^bHazardous Air Pollutants in accordance with the Clean Air Act [13], 2,3,4,7,8-PeCDF specifically as a dibenzofuran

^cRelative difference

MRE PM_2.5 emission factors (12 ± 3 g-kg waste⁻¹) were comparable to previous studies burning simulated military waste, such as small-scale open burns (20 ± 6.4 g-kg waste⁻¹), burn piles (39 ± 24 g-kg waste⁻¹), and an air curtain burner (12 ± 12 g-kg waste⁻¹) [2, 3]. These comparable emission factors could be potentially attributed to combustion techniques or similarities between the various military waste compositions. However, a recent small-scale gasification study burning military waste for waste-to-energy purposes, including MREs [44], showed significantly lower PM emission factors (0.39 ± 0.22 g-kg waste⁻¹) than found in this study. It should be noted, however, that the waste utilized within Barnes’ gasification system was pre-staged; moreover, the gasification system included a combustor and a scrubber (e.g., gas-cleaning device) to reduce emissions exiting into the atmosphere.

Despite varying the presence of aluminum-based pouches for the MRE types, metal emission factors were found to be relatively equivalent amongst all MREs for all metals. For example, aluminum emission factors for MRE-0, −32, −66, and −100% ranged only between 2.3 and 2.8 mg-kg waste⁻¹. Additional metal data are available at Online Resource Table S13. Temperature readings suggest that the heat of combustion did not reach levels great enough to volatilize significant quantities of metal for a long enough period of time to distinguish between the MRE types.

Considerable MCE ranges for the MREs (0.912–0.963) and fiberboards (0.866–0.944) are believed to be attributed to material composition and the random orientation of those materials when burned [45]. These MCE ranges were comparable to Aurell et al. [2] for combustion via open burning and air curtain burners, but were much lower than Barnes’ [44] MCEs for gasification with an afterburner and scrubber via a prototype waste-to-energy unit (MCE = 0.999).

PAHs

No statistical differences were observed in fiberboard emissions factors for the sum 16-EPA PAHs among the various types (p value = 0.0854) despite SFNP showing a higher mean emission factor. MRE emission factors were equivalent amongst the various material types as well when highly influential points were excluded, with the exception of MRE-0% being statistically different from MRE-32% and MRE-66% (p value = 0.0258) with high variability preventing further distinction. Collectively, emission factors for the PAHs were essentially equivalent amongst the MREs, as well as amongst the fiberboards.

For both the MREs and fiberboard types, naphthalene was the greatest pollutant of the PAHs with at least twice the emission factor than the second greatest constituent, which was phenanthrene for MREs and acenaphthylene for the fiberboards (see Table 4). Both SFP and CFP naphthalene emission factors showed the greatest distinction from the MREs, with means of 7 ± 3 mg-kg waste⁻¹ for SFP and 15 ± 7 mg-kg waste⁻¹ for CFP. When naphthalene emission factors are compared for the greatest MRE emitter, MRE-100% (65 ± 39 mg-kg waste⁻¹), against the smallest fiberboard emitter, SFP (7 ± 3 mg-kg waste⁻¹), the polymer-based MRE emission factor surpasses the solid fiberboard with the wet-strength-coating emission factor by at least two and a half times.

Table 4.

SUM 16-EPA PAH emission factors for all material types in mg-kg waste⁻¹

From: Characterizing emissions from open burning of military food waste and ration packaging compositions

Compound	SFP	SFNP	CFP	CFNP	MRE-0%	MRE-32%	MRE-66%	MRE-100%
Naphthalene(2B)*	7.1 ± 2.5	28 ± 21	15 ± 6.7	21 ± 9.9	46 ± 15	40 ± 8.7	31 ± 3.3	65 ± 39
Acenaphthylene	3.3 ± 1.4	13 ± 9.3	4.8 ± 2.2	6.9 ± 2.7	15 ± 5.7	12 ± 2.1	11 ± 1.3	19 ± 9.8
Acenaphthene(3)	0.17 ± 0.090	0.68 ± 0.45	0.35 ± 0.20	0.43 ± 0.19	0.69 ± 0.25	0.58 ± 0.14	0.57 ± 0.011	0.87 ± 0.57
Fluorene(3)	0.74 ± 0.32	3.0 ± 1.6	1.8 ± 0.79	1.4 ± 0.56	4.5 ± 1.9	3.2 ± 0.47	3.5 ± 0.68	4.8 ± 2.6
Phenanthrene(3)	3.0 ± 0.90	10 ± 4.7	6.8 ± 3.0	4.4 ± 1.3	16 ± 7.0	11 ± 2.0	14 ± 3.3	16 ± 7.3
Anthracene(3)	0.61 ± 0.21	2.4 ± 1.3	1.7 ± 0.87	0.99 ± 0.33	2.5 ± 1.0	2.0 ± 0.38	2.3 ± 0.36	2.9 ± 1.5
Fluoranthene(3)	1.6 ± 0.57	4.9 ± 2.7	2.7 ± 1.2	1.8 ± 0.29	3.2 ± 1.4	2.4 ± 0.37	2.9 ± 0.029	3.7 ± 1.4
Pyrene(3)	1.6 ± 0.58	4.5 ± 2.5	2.4 ± 1.1	1.7 ± 0.24	2.2 ± 1.0	1.9 ± 0.39	2.1 ± 0.18	2.9 ± 1.3
Benzo[a]anthracene(2B)	0.32 ± 0.13	1.2 ± 0.65	0.75 ± 0.35	0.44 ± 0.13	1.3 ± 0.65	0.82 ± 0.20	1.1 ± 0.023	1.4 ± 0.69
Chrysene(2B)	0.31 ± 0.11	1.1 ± 0.57	0.72 ± 0.34	0.41 ± 0.11	1.5 ± 0.79	0.99 ± 0.22	1.4 ± 0.021	1.5 ± 0.75
Benzo[b]fluoranthene(2B)	0.18 ± 0.068	0.62 ± 0.32	0.33 ± 0.16	0.23 ± 0.068	0.91 ± 0.45	0.57 ± 0.14	0.77 ± 0.016	0.85 ± 0.45
Benzo[k]fluoranthene(2B)	0.23 ± 0.085	0.91 ± 0.45	0.40 ± 0.17	0.27 ± 0.055	0.83 ± 0.47	0.60 ± 0.11	0.72 ± 0.10	0.89 ± 0.36
Benzo[a]pyrene(1)	0.25 ± 0.095	1.0 ± 0.55	0.46 ± 0.22	0.31 ± 0.067	0.79 ± 0.43	0.56 ± 0.14	0.65 ± 0.079	0.86 ± 0.44
Indeno[1,2,3-cd]pyrene(2B)	0.16 ± 0.051	0.66 ± 0.33	0.28 ± 0.13	0.20 ± 0.035	0.57 ± 0.30	0.36 ± 0.088	0.43 ± 0.072	0.59 ± 0.27
Dibenz[a,h]anthracene(2A)	0.029 ± 0.012	0.11 ± 0.067	0.067 ± 0.038	0.037 ± 0.011	0.21 ± 0.11	0.12 ± 0.031	0.14 ± 0.0048	0.19 ± 0.10
Benzo[ghi]perylene(3)	0.19 ± 0.067	0.73 ± 0.37	0.27 ± 0.11	0.21 ± 0.035	0.50 ± 0.25	0.37 ± 0.10	0.37 ± 0.11	0.58 ± 0.27
SUM 16-EPA PAHs	20 ± 7.2	74 ± 40	38 ± 16	41 ± 16	96 ± 37	77 ± 15	73 ± 1.8	122 ± 67
SUM 16-EPA PAHs TEQa	0.54 ± 0.31	1.6 ± 0.82	0.77 ± 0.38	0.51 ± 0.12	1.4 ± 0.76	0.98 ± 0.22	1.2 ± 0.093	1.5 ± 0.75

^*Hazardous air pollutant in accordance with the clean air act [13]

^(#) Superscript number is the International Agency for Research on Cancer classification group [14]

^amg B[a]P TEQ-kg waste⁻¹

When comparing the emission factors for each PAH pollutant between the MREs and the fiberboards as two distinct groups, all MRE group means were statistically greater than the fiberboard group means with the exception of fluoranthene (K–W p value = 0.0710), which were equivalent, and pyrene (K–W p value = 0.3291), where the fiberboard group mean was greater than the MRE group mean.

A comparison of the sum 16-EPA PAH emission factors between this study and Aurell et al.’s [2] air curtain burner (43 ± 50 mg-kg waste⁻¹) and burn pile (129 ± 50 mg-kg waste⁻¹) study reveals that the mean MRE PAH emission factors (96–122 mg-kg waste⁻¹) were similar, while the fiberboards, specifically SFP (20 ± 7 mg-kg waste⁻¹), were distinctly lower [2]. When compared to Barnes’ [44] gasification study, the fiberboards emitted comparable PAH emission factors, while the MRE emission factors were nearly three-to-six times greater. It is speculated that this was most likely due to combustion conditions from the open burns with lower MCEs ranging from 0.941 to 0.986 when compared to the more complete combustion occurring during the gasification/combustion process (MCE = 0.999). Woodall et al. [3] simulated that military waste study yielded significantly higher Sum 16-EPA PAH emission factors (376 ± 108 mg-kg waste⁻¹) when compared to Aurell et al. [2], Barnes [44], and this study. MRE-0% emission factors (96 ± 37 mg-kg waste⁻¹) were half that of Woodall et al.’s [3] study and CFP emission factors (38 ± 16 mg-kg waste⁻¹) nearly an order of magnitude smaller. It is believed that this difference can be attributed to the percent of food content remaining within their samples (50%), as well as the smoldering conditions the high food content percentage may have caused [3].

PCDD/PCDFs

PCDD/PCDF TEQ total emission factors for CFP (0.283 ± 0.136 ng TEQ-kg waste⁻¹) were three times higher than CFNP (0.102 ± 0.055 ng TEQ-kg waste⁻¹) and six times higher than SFP (0.046 ± 0.016 ng TEQ-kg waste⁻¹) (see Table 5). It is possible that this difference is partly due to the composition of the polymer that imbues wet strength to CFP. In addition, as indicated by the MCEs for MREs (0.920–0.957) and fiberboards (0.969–0.989), all fiberboard types burned with greater combustion efficiency than all of the MRE types. It is believed that the moisture content of the food present within the MRE samples hindered the combustion process and allowed greater PCDD/PCDF emissions to be emitted from the MRE waste versus the fiberboard types. However, no statistical distinction was identified for PCDD/PCDF TEQ total emission factors amongst the various MRE and fiberboard types.

Table 5.

Select emission factor comparisons for simulated military waste burns^a

From: Characterizing emissions from open burning of military food waste and ration packaging compositions

Compound	Units	Gasification system* [44]	Air curtain burner [2]	Burn pile [2]	OBTF burn pilef [3]	MRE-0%	MRE-100%	SFP	CFP
∑PCDD/PCDF TEQ	ng TEQ-kg waste−1	3.6 ± 2.4	35 ± 24	1,765 ± 1,474	127 ± 136	1.802 ± 1.040	0.877 ± 0.527	0.046 ± 0.016	0.283 ± 0.136
∑PAHb	mg-kg waste−1	12 ± 10	43 ± 50	129 ± 50	376 ± 108	96 ± 37	122 ± 67	20 ± 7	38 ± 16
PM2.5	g-kg waste−1	0.39 ± 0.22c	12 ± 12	39 ± 24	20 ± 6.4	12 ± 3	12 ± 1	3 ± 1	3 ± 1
Iron	mg-kg waste−1	0.31 ± 0.18	0.50 ± 0.24	11 ± 23	NA	0.29 ± 0.05	0.30 ± 0.03	NA	NA
Copper	mg-kg waste−1	1.2 ± 0.33	0.18 ± 0.11	0.89 ± 0.92	NA	0.17 ± 0.06	0.21 ± 0.08	NA	NA
Cadmium	mg-kg waste−1	0.0058 ± 0.0031	0.063 ± 0.082	0.073 ± 0.033	NA	0.089 ± 0.038	0.11	NA	NA
Lead	mg-kg waste−1	0.66 ± 0.35	0.55 ± 0.42	0.37 ± 0.22	NA	0.07 ± 0.01	0.07 ± 0.05	NA	NA
Benzene	mg-kg waste−1	43 ± 98d	243 ± 299d/1,371 ± 185e	260 ± 288d/2,421 ± 1,265e	940 ± 220	211 ± 89d	225 ± 36d	79 ± 75d	153 ± 102d
1,3-Butadiene	mg-kg waste−1	ND	67 ± 95d/375 ± 31e	93 ± 150d/746 ± 228e	254 ± 52	103 ± 38d	94 ± 55d	19 ± 22d	44 ± 43d
Styrene	mg-kg waste−1	0.019	256 ± 382d/2203 ± 634e	284 ± 463d/5099 ± 3308e	1880 ± 230	206 ± 192d	120 ± 79d	6 ± 3d	60 ± 64d
Toluene	mg-kg waste−1	2.0 ± 2.8d	88 ± 130d/652 ± 111e	109 ± 170d/1,202 ± 727e	404 ± 56	63 ± 36d	58 ± 18d	19 ± 19d	57 ± 44d
Acrolein	mg-kg waste−1	0.58 ± 0.95d	133 ± 139d/463 ± 33e	98 ± 108d/757 ± 62e	564 ± 169	125 ± 40d	126 ± 27d	9 ± 5d	81 ± 64d
Vinyl chloride	mg-kg waste−1	0.25 ± 0.32d	3.7 ± 2.5d/13e	6.0 ± 5.5d/26 ± 3.3e	7.1 ± 3.5	0.55 ± 0.27d	0.52 ± 0.01d	0.22	ND
Vinyl acetate	mg-kg waste−1	0.57 ± 0.74d	79 ± 97d/324 ± 46e	43 ± 53d/688 ± 195e	705 ± 268	101 ± 41d	109 ± 27d	33 ± 13d	98 ± 70d

ND not detected, NA not analyzed

^*MAGS unit

^aRange of data denoted ±1 standard deviation, where no range is stated, only one sample with detectable levels

^bSUM 16 EPA PAHs

^cTotal PM

^dMCE > 0.95

^eMCE < 0.90

^fThese data were obtained using Woodall et al.’s [3] emission factors in μg-kg waste⁻¹ and multiplying by a correction factor (0.47) derived from their waste composition utilizing carbon fractions from each material obtained from Liu and Lipták [43]

When comparing the MREs and fiberboards with one another as two distinct groups, a majority of the PCDD/PCDF congener concentrations were statistically greater for the MREs versus the fiberboards. The PCDD/PCDF TEQ total emission factor for both the MREs and the fiberboards was significantly lower than values seen by Barnes’ [44] gasification system (3.6 ± 2.4 ng TEQ-kg waste⁻¹), Aurell et al.’s [2] air curtain burner (35 ± 24 ng TEQ-kg waste⁻¹) and burn pile (1765 ± 1474 ng TEQ-kg waste⁻¹), as well as Woodall et al.’s [3] OBTF burn pile (127 ± 136 ng TEQ-kg waste⁻¹). Although the waste compositions were certainly different, the large differences between the PCDD/PCDF TEQ total emission factors within this work and other simulated military waste burns would suggest PCDD/PCDF formation may not be primarily associated with MREs or their fiberboard packaging but may rather be attributed to other components of military waste not characterized within this experiment, as well as combustion conditions.

VOCs

MRE VOC emission factors were generally one-to-seven times greater in magnitude than fiberboard VOC emission factors, with the greatest difference seen with propene. With the exception of 2-butanone (K–W p value = 0.0608), group mean VOC emission factors for MREs were statistically greater than group fiberboard emission factors. Emission factors for 1,3-butadiene amongst the fiberboards were not statistically different from one another (p value = 0.5752), while benzene was determined to be statistically different for the corrugated fiberboard with the wet-strength polymer (CFP) and the solid fiberboard without the wet-strength polymer (SFNP) versus the corrugated fiberboard without the wet-strength polymer (CFNP) (p value = 0.0001).

With similar emissions among all of the MRE types (see Tables 3 and 5), benzene was one of the greatest pollutants emitted. Overall though, no practical differences in VOC emission factors were found amongst the MRE types, nor amongst the fiberboard types.

When compared to other studies with good combustion efficiency (MCE > 0.95), Aurell et al.’s [2] air curtain burner (243 ± 299 mg-kg waste⁻¹) and burn pile (260 ± 288 mg-kg waste⁻¹) benzene emission factors were analogous to the MREs; however, in Barnes’ [44] gasification study, the benzene emission factors (43 ± 98 mg-kg waste⁻¹) were more similar to the fiberboard emissions. When comparing the MRE and fiberboard benzene emission factors to poor combustion (MCE < 0.90) conditions, Aurell et al.’s [2] air curtain burner (1371 ± 185 mg-kg waste⁻¹) and burn pile (2421 ± 1265 mg-kg waste⁻¹) emission factors were greater by an order of magnitude. Furthermore, the OBTF burn pile benzene emission factors (940 ± 220 mg-kg waste⁻¹) tested by Woodall et al. [3] were found to be nearly five times larger than MRE emission factors and an order of magnitude larger than fiberboard emission factors as well, most likely due to the larger amount of food waste present within their study, which was approximately 50% of total food waste available, compared to 25% in this study.

Reevaluated in 2012 and classified by the WHO IARC as a “carcinogenic to humans” substance [14], 1,3-butadiene mean emission levels for SFP (19 ± 22 mg-kg waste⁻¹) were found to be the lowest when compared to all other material types, to include Aurell et al.’s [2] air curtain burner (67 ± 95 mg-kg waste⁻¹) and burn pile (93 ± 150 mg-kg waste⁻¹) emission factors and Woodall et al.’s [3] OBTF burn pile (254 ± 52 mg-kg waste⁻¹), under good burning conditions (MCE > 0.95). However, no statistical significance was determined between the fiberboards (p value = 0.5752) or MREs (p value = 0.5881) for 1,3-butadiene.

MCEs for the fiberboard types (0.966–0.987) and MRE variations (0.974–0.983) appeared to be linearly correlated with VOC emission factors. This linear relationship was previously explored within the literature [36, 38, 39]. It should be noted that this MCE relationship was much less robust for each of the other pollutant classes. Figure 3 depicts normalized emission factors as calculated percentages of the maximum emission factor for each of five IARC-classified [14], hazardous air pollutants [13].

Fig. 3.

Percent of maximum VOC emission factors vs MCE for fiberboards and MREs with 95% confidence intervals

Coefficient of determination values for a linear correlation was generally greater for the fiberboards than for the MREs and is thought to be associated with the homogeneity of test material amongst the fiberboards versus the relatively heterogeneous nature of an MRE and its food contents.

Conclusions

Table 5 provides a synopsis of the simulated military waste studies discussed and provides perspective as to where select MRE and fiberboard emissions can be placed within the literature. MREs accounted for 89–92% of PM, 84–99% of VOC, 67–96% of PAH, and 78–100% of PCDD/PCDF emissions when burned in unison with the current fiberboard container and liner. Test results suggest that particulate matter equal to or less than 2.5 μm in size constitutes a majority of the particulates emitted from both MREs and their fiberboard packaging. The MREs emitted up to five times greater PM emissions than the fiberboards on a mass basis and average MRE emission factors were comparable to Aurell et al.’s air curtain burner [2] and Woodall et al.’s OBTF burn pile [3] emission factors. PAH emission factors were similar to Aurell et al.’s study [2], greater than Barnes’ study [44], and nearly four times less than Woodall et al.’s study [3]. Due to the higher emission factors, the targeted replacement of MRE constituents may be more effective than the variation of fiberboard-packaging designs to reduce PM, PAH, and VOC emissions. PCDD/PCDF emission factors were minimal for both the MREs and the fiberboards; were by far the smallest emission factors compared to similar studies [2, 3, 44]; believed to be due to material composition and burn conditions; and may be related to other components of military waste not explored within this study as well as combustion conditions. Overall, variations of the MRE and fiberboard types yielded minimal differences in PM, PAH, VOC, and PCDD/PCDF emission factors generated. Emission factors have now been established for MREs and their fiberboard packaging and can directly contribute to future military, food, and packaging waste studies. Characterized emission factors can also be utilized for predicting and modeling potential future health risks associated with exposure to the materials and pollutants described within this study. This study sought to characterize emissions from four MRE variations and four fiberboard-packaging types to establish emission factors, conduct a comparison of these emission factors to each other, and ascertain where these emission factors stand within the literature data.