An evaluation of solar thermal heating to support a freeze-thaw anaerobic digestion system for human waste treatment in subarctic environments

Abstract

Remote locations, small communities, and weather prohibit the operation of piped sanitary sewers in many Alaska Native Villages (ANVs). Research was conducted to understand the technical feasibility of installing anaerobic digesters (ADs) in remote ANVs which would be heated by solar thermal collectors. Biochemical methane potential (BMP) assays were conducted to understand the effect of freezing and thawing on methanogenic activity of synthetic human feces. BMPs were frozen at −20 or −80 °C for 7 days and then incubated at psychrophilic (20 °C) or mesophilic (37 °C) conditions. Psychrophilic BMPs frozen at −20 or −80 °C yielded 453 ± 119 and 662 ± 77 mL CH₄/g VS, respectively. Mesophilic BMPs frozen at −20 or −80 °C yielded 337 ± 59 and 495 ± 63 mL CH₄/g VS, respectively. Freezing caused a lag period, but ultimately many of the assays reached yields similar to or even greater than the baseline, unfrozen assays. Monthly solar radiation and air temperature data were used to identify the number of solar thermal collectors that would be required to supplement heat energy to operate the ADs in several locations. Alaskan subarctic locations receive enough solar thermal energy in summer months to support seasonally operated, psychrophilic ADs.

1. Introduction

A consistent concern expressed to the United States (US) Environmental Protection Agency from Alaska Native Villages (ANVs) is the need for basic sanitation and treatment options for raw sewage. Currently over 3000 Alaskan households are estimated to use un-piped sanitation systems to collect and dispose of human waste. The remote locations of some ANVs make the transport, installation, and management of piped water, solid waste, or wastewater systems impractical or impossible [1,2]. Instead, defecation occurs in plastic buckets lined with plastic bags, which are then hauled to a collection hopper, sewage lagoon, or unlined Alaskan Class III landfill [3,4]. This system poses sanitation, public health, social, and environmental concerns including exposure to raw sewage and pathogens and potential for spillage and pollution.

Although landfilling might be cost-effective, many of the landfills in these ANVs are unlined and historically rely on permafrost to prevent groundwater pollution. As permafrost conditions continue to change, the potential for landfill leakage and environmental contamination increases [4]. Overall, a treatment system that reduces exposure to raw sewage, improves sanitation, and is feasible for remote ANVs is desired.

Anaerobic digestion is a potential alternative that addresses these concerns. However, freezing and fluctuating temperatures are a key barrier when it comes to maintaining a productive environment for methanogens in anaerobic digesters (ADs) [5–8]. Fluctuating temperatures prevent consistency in the anaerobic microbial population, causing destabilization and system failure [9,10]. Most applications of ADs exist in warmer, lower latitude areas, resulting in a lack of data on the feasibility of digestion in cold and extreme climates [11–15]. Although ADs are most productive in mesophilic (30–40 °C) and thermophilic (45–60 °C) temperature ranges, recent work has shown that psychrophilic (20 °C) digestion can result in methane yields similar to mesophilic digesters [8,12,15,16]. Psychrophilic microbial communities sourced from cold-weather regions that have acclimated to colder temperatures have also been shown to be effective [11,13,14,17]. Additionally, large-scale sewage treatment can rely on significant volumes of wastewater and climate-controlled buildings to maintain desirable temperatures. While community-scale or even family-scale ADs have been successful in Latin America and other equatorial regions because of the consistent warm ambient temperatures [18], similar attempts have not been performed in arctic or subarctic environments of Alaska where the energy needs of ADs would be greater.

Seasonal temperatures and remoteness make integrated electrical networks infeasible in many ANVs and many communities often rely on diesel fuel for their energy needs. While wind can be a valuable renewable resource, installations are limited to coastal areas where wind energy is most dense [19,20]. Alaska experiences abundant solar exposure in spring and summer months, creating an opportunity for a sustainable energy source in nearly all regions of the state [15,21]. As shown in Fig. 1A, the extended summer days mean solar radiation exposure in Alaska is similar to Ohio in spring and summer months [22].

Fig. 1.

Solar radiation and air temperature for two locations in the contiguous United States and multiple locations in Alaska.

The freezing and thawing effects of sludges have been previously studied for their physical and chemical effects [23–25]. To understand the feasibility of small-scale ADs in remote ANVs with energy limitations, this research presents results of a laboratory analysis to understand the capability of anaerobic microorganisms to recover following freezing of combined inoculum and synthetic human feces substrate. Biochemical methane potential (BMP) assays were performed at psychrophilic and mesophilic conditions after being frozen at −20 or −80 °C for 7 days. Solar radiation and air temperature data were used to evaluate if solar thermal energy could be used to provide energy to community-scale AD systems in subarctic and arctic locations. The applicability of the results for other global locations and the potential for energy recovery from the predicted biogas volumes are discussed.

2. Materials and methods

2.1. Laboratory analyses

To test the effects of temperatures expected in the ANVs on ADs, methane production from synthetic human feces was measured using BMP assays that were inoculated and then immediately frozen. The synthetic feces was produced following protocols developed by Colón et al. [26] and Penn et al. [27]. As shown in Table 1, the recipe was modified slightly to account for the higher fat content consumed by ANVs from the ingestion of seafaring mammals [28]. The use of synthetic feces allows for comparison and evaluation of a homogenized, standardized sample without the inherent risk to laboratory personnel of handling human feces [26]. To establish the amount of “as-disposed” feces for a hypothetical scenario, moisture content and volatile solids content were measured. Moisture content of the synthetic feces was measured as mass lost after heating a sample at 105 °C for 24 h. Volatile solids (VS) content was measured as the lost mass after heating in the muffle furnace at 550 °C for 4 h.

Table 1.

Synthetic feces composition and characteristics.

Material	Manufacturer	Mass (g)	Composition (dry weight %)	Emulates
Yeast extract	Fisher Scientific	77.10	27	Biomass
Baker’s yeast	Fleischmann’s	8.59	3
Microcrystalline cellulose	Acros Organics	14.22	5	Carbohydrates
Psyllium husk	Organic India	35.58	12.5	Carbohydrates/Fiber
Miso paste (semisolid)	American Miso Company	49.93	17.5	Protein
Oleic acid (liquid)	Fisher Scientific	85.54	30	Fat
NaCl		5.68	2	Salts
KCl		5.69	2	Salts
CaCl2		2.90	1	Salts
DI water	Laboratory	714.77	0	Moisture
Total	–	1000.00	100	–

BMP assays were performed in general accordance with the protocol established by Owen et al. [29] and Hansen et al. [30]. A homogenized 0.2 g VS sample was added to a 175 mL serum bottle with 100 mL of inoculum. The inoculum was stabilized AD effluent from a municipal mesophilic sewage sludge digester in Fairfield, Ohio, USA. The sludge was maintained in the laboratory for 7 days prior to the experiment and allowed to de-gas before inoculation to digest any recalcitrant organic matter. The pH of the sludge was 7.04. Total suspended solids of the sludge (in triplicate) on a glass fiber filter was 1.2 ± 0.1% and volatile suspended solids was 100% of total solids. Microcrystalline cellulose was used as a positive control to ensure inoculum strength. Once BMP bottles were filled, the headspace was purged with ultra-high purity nitrogen gas for 2 min and sealed with a butyl rubber septa and aluminum crimp. Three batches of BMPs were performed: unfrozen (BAS), frozen at −20 °C (F20), and frozen at −80 °C (F80). BAS samples established the baseline substrate performance with which F20 and F80 could be compared. The BMP bottles were filled, sparged, capped, and moved to the appropriate temperature setting. Frozen samples were removed from freezers after 7 days and allowed to thaw for 24 h at room temperature before incubation so as not to crack the glass serum bottles. BMPs were incubated at mesophilic conditions 37 ± 2 °C in an Environmental Growth Chamber (EGC Inc., Ohio, USA), and psychrophilic conditions 20 ± 2 °C in a low-temperature incubator (Hach Inc., Colorado, USA). For BAS samples, day 0 was considered the day of inoculation. For F20 and F80 samples, day 0 was considered the day of thawing. Blank samples, containing only sludge and no substrate (synthetic feces), were used to nullify the endogenous methane production of the inoculum for each condition.

Gas composition was measured with a gas chromatograph with thermal conductivity detector (GC-TCD; Agilent 6890, California, USA). Columns used were Porapak N and Mol sieve (Supelco, Pennsylvania, USA). The GC was used to quantify carbon dioxide, oxygen, nitrogen, and methane. On the same days of gas composition analysis, headspace gas pressure in the serum bottles were measured with a Dwyer 477B digital manometer (precision 1 mm Hg). A digital barometer (Fisherbrand, Pennsylvania, USA) measured ambient air pressure to account for elevation differences and standardize methane yields (precision 1 mbar or 0.75 mm Hg). Measurements were taken until excess gas pressure was found to be less than 1 mm Hg/day and less than 95% of the cumulative methane volume. Methane volumes were normalized to standard temperature and pressure as shown in (1).

$$V_{CH4,STP}=V_{CH4}\times \left(\frac{273.15}{273.15+T_{inc}}\right)\times \left(\frac{P_a-P_w}{P_{STP}}\right)$$ (1)

where V_CH4,STP is the volumetric methane generation at standard temperature and pressure (STP); V_CH4 is the volume of methane in the biogas (from GC analysis); T_inc is the incubation temperature (°C); P_a is the ambient pressure of the laboratory; P_w is the water vapor pressure at T_inc; and P_STP is the pressure at STP (760 mm Hg).

Cumulative methane volume is calculated from the weekly STP data (V_CH4,_STP) considering both the volume of gas extracted from the bottle (in the syringe) and the headspace within the bottle, as shown in (2).

$$V_C=V_S\times M_f+\sum _f^f{V}_n\times M_n$$ (2)

where V_C is the cumulative methane volume; V_S is the volume of reactor headspace; M_f is the final (or most recent) methane concentration; V_n is the weekly excess gas volume for the nth measurement; and M_n is the weekly methane concentration for the nth measurement.

BMPs for substrates are reported as a volume of methane per unit mass of waste (e.g., mL CH₄/g VS). The average cumulative methane generated from the blank BMPs was subtracted from each sample replicate. Finally, the corrected gas volume was divided by the mass of sample VS added.

2.2. Weather data and thermal modeling

Solar radiation data was collected from the Department of Energy National Renewable Energy Laboratory (NREL) National Solar Radiation Database (NSRDB) for five locations in Alaska (Anchorage, Bethel, Fairbanks, Homer, and Utqiagvik) and two locations in the contiguous US (Palm Springs, California and Cincinnati, Ohio), as shown in Table 2. The NSRDB Physical Solar Model (PSM) does not extend above latitude 60^◦ N, which excludes the majority of Alaskan territory [22]. Therefore, data from Typical Meteorological Year 3 (TMY3), presented in Fig. 1A on a monthly basis, were used for all sites. The TMY3s are data sets of hourly values of solar radiation and meteorological elements for a 1-year period. Their intended use is for computer simulations of solar energy conversion systems to facilitate performance comparisons of different system types, configurations, and locations in the United States and its territories. Air temperature data for these locations were collected from the National Oceanic and Atmospheric Administration (NOAA) National Centers for Environmental Information US Monthly Climate Normals [31]. The data presented in Fig. 1B were monthly averages from 2006 to 2020. The error bars represent standard deviation among those years.

Table 2.

Location of TMY3 data collected from NSRDB.

Locations	State	TMY3 Station ID	Latitude	Longitude	Elevation m	Environment/Biome
Palm Springs	CA	747187	33.633	−116.167	146	Desert
Cincinnati	OH	724297	39.100	−84.417	147	Temperate
Homer	AK	703410	59.650	−151.483	29.0	Subarctic/Maritime
Bethel	AK	702190	60.783	−161.833	0.914	Subarctic/Transitional
Anchorage	AK	702730	61.183	−150.000	31.1	Subarctic/Transitional
Fairbanks	AK	702610	64.817	−147.850	136	Subarctic/Continental
Utqiagvik (formerly Barrow)	AK	700260	71.320	−156.620	3.05	Arctic/Tundra

Heat losses were calculated monthly, based on heat transfer of fluids between one or more surfaces as shown in Fig. 2. Monthly average air temperatures were used to estimate the energy necessary to maintain the sludge at operational psychrophilic (20 °C) or mesophilic (37 °C) temperatures using (3) and (4).

$$Q_{loss}=U\times A\times \left(T_{AD}-T_{air}\right)$$ (3)

$$U=h_{c,i}+\sum \frac{k_y}{L_y}+h_{c,o}$$ (4)

where Q_loss (kWh) is the energy lost from the system due to temperature differences; U (W/m²_⋅°C) is the heat transfer coefficient; A (m²) is the surface area of each respective portion of the digester; T_AD is the operating temperature of the digester; T_air is the ambient air temperature (both °C); h_c,i is the thermal conductivity of the inside (air on top or sides, liquid on the bottom); k_y is the thermal conductivity (W/m⋅°C) of component y; L_y is the thickness of component y; and h_c,o is the thermal conductivity of the outside material (frozen soil on the bottom or air on the sides and top).

Fig. 2.

Solar thermal collector and AD (grey) diagram with conductive heat losses on six sides. Foam insulation (yellow hatch) is the primary means of reducing heat losses on all sides of the digester.

As shown in Fig. 2, although the AD would have a semi-circular or tubular shape, it would be wrapped in a water jacket or water bath and housed in an insulated container. Thus, the heat losses are assumed to be from a rectangular insulated container with the digester’s liner contacting the bottom insulated layer, in turn in contact with the ground. The top and sides of the digester are modeled with heat losses from air through a single insulative layer to the ambient air. The bottom is modeled considering liquid sludge/water is in contact with the AD material, which is in contact with foam, in contact with the underlying frozen soil/ice. The insulation was assumed to be 15 cm (6^′′) in thickness. The digester and foam would sit within a metal or rigid plastic container which is not shown for clarity and neglected from the model. The sides and top would exhibit the internal container temperature, assumed equal to T_AD. The soil/permafrost temperature was set equal to −10 °C when monthly ambient air temperatures were freezing and 0 °C when ambient temperatures were above freezing. The thermal characteristics for materials used in the calculations are presented in Table 3. Equations (5)–(8) were used to account for the energy required to thaw the sludge and bring it to operating temperature after winter freeze.

$$Q_{ice}=m\times C_{ice}\times \left(T_0-T_{ice}\right)$$ (5)

$$Q_{melt}=m\times C_{melt}$$ (6)

$$Q_{liquid}=m\times C_{water}\times \left(T_{AD}-T_0\right)$$ (7)

$$Q_{thaw}=Q_{ice}+Q_{melt}+Q_{liquid}$$ (8)

where Q_ice (J) is the heat energy required to warm ice to 0 °C; m is the mass of water/sludge (kg) in the digester; C_ice is the specific heat of ice (2.05 × 10³ J/kg⋅°C); T₀ is the temperature of water-ice’s melting phase change (0 °C); and T_ice is the temperature of the ice (assumed −20 °C). Q_melt (J) is the energy required to melt the ice to water (i.e., change phase) and C_melt (3.34 × 10³ J/kg) is the amount of energy required to convert solid ice to liquid water at 0 °C by mass. Q_liquid is the amount of heat energy needed to raise the digester to operating temperature; C_water is the specific heat of water, and T_AD and T₀ were previously defined. The sum of these three equations (Q_thaw) represents the total energy required to bring frozen sewage to AD operational temperature. This value is added to the other monthly estimated energy requirements (Q_loss) for the first month of operation when average air temperatures rise above freezing and when solar radiation is available to provide requisite heating. In some cases, thawing was spread over two months to minimize the number of solar thermal collectors required.

Table 3.

Thermal conductivities and thermal coefficients for monthly heat loss calculations.

Material	Thermal conductivity, k W/m·°C	Thickness, L m	Thermal conductivity, hc W/m2·°C
Insulating foam	0.04	0.15	–
AD liner (polyethylene or similar)	0.3	0.0015	–
Ice/Frozen soil	2.2	1	–
Air, free flow	–	–	10
Water/AD sludge, 20 °C	–	–	100

Performance data from commercially available solar thermal collectors were collected and used to estimate the number of collectors that would be required to maintain digestion temperature as shown in (9).

$$N_{panels}=\text{RAD}/\left(\text{TO}\times {\text{A}}_{\text{panel}}\times EF\right)$$ (9)

where N_panels is the number of panels; RAD is the solar radiation exposure (kWh/m²); TO is the thermal collector output (kWh/panel); A_panel is the panel area (m²); and EF is the efficiency factor. The EF was assumed as 50% meaning half of the generated thermal energy was lost before utilization [32,33]. Panel output (TO) was based on commercial information and is a stepwise function of available solar radiation and the temperature difference of air temperature and desired temperature. Therefore, as temperature and solar radiation increases from Januar-y–June, fewer solar thermal collectors would be required in summer months to achieve the required energy input.

Separate from the thermal modeling, the energy of the predicted biogas was estimated in (10) to understand if significant quantities could be recovered.

$${\text{R}}_{\text{CH}4}=\text{BMP}\times \text{VS}\times (1-\text{MC})/{10}^6$$ (10)

where R_CH4 is the “as-disposed” or wet weight methane potential of feces (L CH4/kg feces); BMP is the methane potential of feces (mL CH₄/g VS feces); MC is the moisture content (%); and VS is the volatile solids content (%).

To understand how much energy could be recovered from a community-scale system, the energy value of the biogas was calculated with (11).

$${\text{E}}_{\text{BG}}=\text{LHV}\times \text{W}\times {\text{R}}_{\text{CH}4}/\text{F}$$ Equation(11)

where E_BG is the energy from the biogas (MJ), LHV is the lower heating value (22 MJ/m³ biogas), W is the cumulative mass (kg) of feces in the digester, and F is the fraction of CH₄ in biogas (60%). LHV was used because no upgrading or cleaning of the produced biogas is expected, resulting in a mixture of methane, carbon dioxide, and trace gases being combusted.

3. Results and discussion

3.1. BMPs of frozen and thawed inoculum with synthetic feces

Feces was selected as the sole substrate because concurrent research has begun in ANVs utilizing urine-separating toilets. Thus, urine would be collected and managed separately and the waste for digestion would be primarily feces. Results of the BMP experiments with synthetic feces are presented in Fig. 3. Triplicate assays were performed for the synthetic feces as well as positive controls (microcrystalline cellulose) and blanks (inoculum only) for each temperature condition. Assays incubated at mesophilic and psychrophilic conditions were all inoculated with mesophilic sludge from the same source. Mesophilic cellulose controls achieved a satisfactory methane yield, 323 mL CH₄/g VS, indicating sufficient methanogenic population (cellulose and blank assays not shown in Fig. 3).

Fig. 3.

Triplicate BMP assays of synthetic feces unfrozen (BAS), frozen at −20 °C (F20), and frozen at − 80 °C (F80). Incubation occurred at 37 °C (M) or 20 °C (P).

Mesophilic samples not frozen (M-BAS) achieved 93% of their final BMP value in 14 days and 98% in 21 days. The M-F20 and M-F80 samples lagged initially as microbes recovered functionality but ultimately the methanogens became active and produced substantial quantities of biogas. Freezing did have the effect of reducing the cumulative methane yields. M-F20, M-F80, and M-BAS were 337, 495, and 596 mL CH₄/g VS, respectively. Montusiewicz et al. [23] found freezing improved digestibility, whereas here the results were mixed. This discrepancy in findings is likely due to the substrate and inoculum being frozen together and the lower freezing conditions that might have more strongly degraded the substrate. M-F20 and M-F80 achieved over 90% of their final BMP value by day 30.

Interestingly, psychrophilic BMPs on average yielded more methane than their mesophilic counterparts. P–F80 even surpassed the P-BAS samples (662 and 657 mL CH₄/g VS, respectively). However, P-BAS had only achieved 84% of its final yield by Day 51. The P-BAS and M-BAS samples were similar (596 and 575 mL CH₄/g VS) showing inoculum from mesophilic sources could be successfully and quickly adapted to psychrophilic conditions. Both psychrophilic and mesophilic temperatures, BMPs frozen at −20 °C achieved lower methane yields than the counterparts frozen at −80 °C.

The wide range in final methane yields indicates that freezing inoculum impacts digester efficiency and efficacy and requires further evaluation. It is unclear if repeated freeze and thaw cycles would build resiliency or weaken the microbial activity over time. The psychrophilic samples lagged BAS and M samples, most not reaching 90% of the final value until about day 60. This does indicate that the rate of methane production from these freeze-thaw conditions would be slower and would require the adaptation of longer waste treatment timelines that account for increased energy demands.

3.2. Energy demands of continuous operation AD

As shown in Fig. 1B, many locations in Alaska experience freezing conditions for long parts of the year. According to NOAA data, Homer, Anchorage, and Fairbanks air temperatures are above 0 °C from April–October. Bethel, which is located on the southern coast, experiences average above-freezing air temperatures from May–October. Utqiagvik, the only Arctic location analyzed in this study, is the most northern town in the US and is above freezing only from June–September. Importantly, given the BMP results, this means in all locations there are hypothetically enough days (~90 days) to treat the collected sewage and generate methane even accounting for a lag period. However, this does not account for the time required to thaw a frozen mass of material and the low ambient temperatures during these months mean heat losses would be a constant operational challenge.

The solar thermal collector model took into account available solar radiation and heating needs of an insulated digester as represented in Fig. 2. The proposed system assumes a batch-style AD that would treat the annual amount of waste for 80 people assuming a feces production rate of 0.25 kg/person/day and fecal density of 1.06 kg/L [27,34]. Fig. 4A and B presents the energy required to heat and maintain the proposed community scale mesophilic and psychrophilic AD year-round (without freezing in the winter), respectively. In this representation, the energy required to thaw would be zero due to the continuous operation of the AD. Although predicted energy demands for heating an AD decrease in summer months, solar radiation becomes negligible in winter months. Efficient operation of solar thermal collectors relies on available solar radiation and the temperature difference between ambient temperature and desired operational temperature [21,32,35, 36]. Thus, efficacy of a year-round solar thermal system is impeded not just by the minimal solar radiation but also extreme cold. While a fossil fuel could be used to supplement the system, that would not be sustainable or cost-effective for a community where fuel is scarce and in-demand. Therefore, a seasonal system which experiences freezing and thawing was investigated.

Fig. 4.

AD heat energy demands for two locations in the contiguous United States and five locations in Alaska for (A) mesophilic conditions (35 °C) and (B) psychrophilic conditions (20 °C). Assumes an 8 m³ system.

3.3. Energy demands of freeze-thaw AD

The freeze-thaw digester was assumed to be the same conceptual design as previously discussed. The amount of energy required to thaw approximately 7300 kg of feces was calculated to be 934 kWh based on (5) to (8). To decrease energy requirements during the first months with sufficient solar radiation to operate the digester, the month of thawing was established as the first month of the year when the monthly average ambient air temperature was above freezing. Based on commercial thermal collector information, the available solar energy was also required to be above approximately 50 Wh/m²/day. Table 4 presents the monthly profile of energy requirements for a freeze-thaw psychrophilic digester in Fairbanks, Alaska. As can be seen, the thawing period was extended across two months to reduce the number of solar panels required. The amount of thermal energy supplied varied based on the solar radiation available and the temperature difference (ΔT = T_air – T_AD). Coastal areas tend to experience increased cloud cover, reducing available solar radiation [21,37,38]. This was observable in data from Bethel and Anchorage, located near the southern coast, where the energy delivery was poorer than in Fairbanks, located in the interior of Alaska.

Table 4.

Energy requirements for a conceptual freeze-thaw psychrophilic (20 °C) AD in Fairbanks, Alaska.

Month	Solar Radiation (Wh/m2/d)	Air Temperature (°C )	Soil Temperature (°C )a	Solar Collector Energy Demands (kWh)			Heat Losses (kWh/month)					Number of Collectors Required
				Output (kWh/panel/month)	Efficiency lossa	Outputb (kWh/panel)	Qloss, sides	Qloss, bottom	Qloss, total	Qthawc	Qtotal
Jan	6.42	−22.9	−10	0.0	50%	0.0	314	34	348	0	348	-
Feb	31.2	−17.5	−10	0.0		0.0	275	34	308	0	308	-
Mar	94.3	−11.7	−10	0.0		0.0	232	34	266	0	266	-
Apr	178	1.06	0	51.3		25.6	139	22	161	134	295	12
May	222	10.1	0	158.9		79.5	73	22	95	800	895	12
Jun	235	15.2	0	153.8		76.9	35	22	58	0	58	1
Jul	213	16.4	0	158.9		79.5	26	22	49	0	49	1
Aug	164	13.2	0	53.0		26.5	50	22	72	0	72	3
Sep	92.0	7.28	0	0.0		0.0	93	22	116	0	116	-
Oct	42.1	−2.78	−10	0.0		0.0	167	34	200	0	200	-
Nov	10.7	−16.1	−10	0.0		0.0	264	34	298	0	298	-
Dec	1.91	−19.9	−10	0.0		0.0	292	34	326	0	326	-

^aAssumed.

^bOutput is a function of the temperature difference and solar exposure. Limited solar radiation would give minimal energy for thermal heat transfer and are assumed zero in these months.

^cQ_thaw begins in the first month where the average air temperature was above 0 °C. To minimize the number of solar panels, thawing is spread to the following month. Digestion could only occur after thawing was complete.

Performance data from commercially available solar thermal collectors were used to determine the number of panels required to provide energy to the AD in each month. Fig. 5 presents the maximum number of solar panels (each sized 1.22 × 2.43 m = 2.97 m²; 4 ft × 8 ft = 32 ft²) required to thaw 7300 kg of feces and maintain psychrophilic conditions for at least two months of the year by location. Although mesophilic digestion would provide faster treatment times, ΔT was generally found to be too large even in summer months for efficient operation. Thus, an unsustainable number of solar collectors would be required, or the temperature gradient would cause excessive thermal losses.

Fig. 5.

The number of solar thermal collectors needed to supply energy to thaw and operate an 8 m³ psychrophilic AD in Alaskan locations. The number is shown in parentheses and represented by the diameter of the circle.

3.4. Energy generation from ANV-scale ADs

The viability of energy recovery from conventional wastewater ADs relies on large volumes of material (e.g., millions of gallons digested per day). The primary utility of the ADs for ANVs is to treat the raw sewage that would otherwise be disposed of in an unlined landfill or lagoon. The secondary benefit of biogas production and energy recovery was investigated. Moisture content and VS content of the synthetic feces were found to be 77% and 99%, respectively. Taking the average of 300–662 mL CH₄/g VS found in the BMP experiments (520 mL CH₄/g VS), the methane potential of 1 kg of “as-disposed” synthetic feces was 118 L. These are similar to other reported methane yields from human waste [26,39,40].

As described in the methods, because no upgrading/cleaning of the gas is expected, the lower heating value (LHV) was used to estimate energy generation. LHV was estimated as 22 MJ/m³ CH₄ [41]. Assuming approximately 7300 kg of feces would result in 3.16 × 10⁴ MJ, enough heat to boil only 28 L of water. This also assumed a starting water temperature of 10 °C, boiling (phase change) at 100 °C, and conversion of half of the water mass to steam at 100 °C. Mudasar and Kim [40] reported 766 Wh/kg feces which is 64% of the value determined here (1.20 Wh/kg). This is likely due to system-specific power delivery in-efficiencies. Alternately, the amount of biogas necessary to heat a home, or operate an appliance would depend greatly on building materials and efficiency rating [42]. In this scenario, the low human defecation rate and the low population density means that human feces AD might produce only small quantities of biogas for recovery. Adding food scraps or other readily biodegradable wastes could be explored to boost methane yields [16,43].

3.5. Limitations of the analysis

The study objective was to understand the technical feasibility of solar-augmented AD for ANVs. These are cold climate communities that lack the physical and financial resources to install and maintain typical wastewater infrastructure. This necessitates an investigation into more innovative and sustainable options that accommodate subsistence living arrangements that are more common in developing countries. The locations examined were exclusively in the United States but have broad latitudinal coverage that reflect desert, temperate, subarctic and arctic environments. The conceptual design may be applicable to many remote, high-altitude, and cold-climate communities in North and South America, Africa, and Asia that face similar challenges with human waste management. While the results demonstrate technical feasibility, the study did not attempt to understand the social or economic complexities of implementing a new waste management system. Several studies illustrate successful implementation of small-scale AD, but the utilization of the systems were dictated by the needs of the communities [44]. Additionally, this study does not address what to do with the treated effluent. Future research should investigate the potential for beneficial reuse or disposal options, which would be community-specific.

For locations that lack monthly site-specific meteorological data, Zhang, et al. [21] identifies a detailed methodology to simulate the data necessary for such an approach. Finally, the impact or effect of solar collector designs was not investigated here. Many styles of solar thermal collectors are available, each with their strengths and weaknesses [21]. The two types most applicable for ANVs might be flat-plate or unglazed solar collectors [35]. In climates with temperatures less than 30 °C, unglazed collectors are most common. These collectors are not insulated, so much of the thermal heat can be lost at low air temperatures [35,36]. However, their simple design and material requirements could be advantageous to the ANV scenario. Although other types of collectors can reach higher temperatures, the specialized equipment and fragility under extremely cold seasons might limit their utility. Heat transmission would come from sensible heat transfer of liquid water wrapped or encasing an AD lined with polyethylene or similar geo-membrane plastic. Zhong et al. [45] reported that 70% of solar energy collected for a solar thermal-AD was utilized. To conservatively estimate energy needs and number of panels required for an 80-person system in Alaska, efficiency of the digester output was assumed to be 50% in all months (50% solar heat energy collected was lost). Future work should focus on the development of a pilot-scale system in replicated or actual field conditions to understand the actual energy losses and maintenance requirements of the entire AD and solar collector system.

4. Conclusions

This work evaluated the potential for a solar-augmented AD to treat synthetic human feces with a combination of BMP assays and thermal modeling. The substrate was inoculated with AD sludge, immediately frozen for 7 days at −20 °C or −80 °C, and then incubated in psychrophilic (20 °C) or mesophilic (37 °C) conditions. Freezing and thawing of inoculum did induce a lag period of approximately 5–10 days in the lab-scale reactors. Freezing, in some instances, actually augmented the methane production. Methane yields ranged from 330 to 670 mL CH₄/g VS synthetic feces; the large range due to various conditions tested. BMP assays reach 95% of the final methane yield within 14 days under mesophilic conditions and 60 days under psychrophilic conditions. The research indicated that psychrophilic digestion could achieve equal methane production as mesophilic systems over time. Solar radiation and air temperature data were used to model a hypothetical freeze-thaw AD augmented by solar thermal collectors in ANVs. Although the limited time available for treatment varied by location, subarctic conditions indicated enough time for seasonal waste treatment (60–90 days). Arctic locations likely cannot support such a treatment system due to limited solar radiation and low annual ambient temperatures. Thus, the research posits a scenario where single-use, community-scale ADs could be used to collect and treat human waste even in remote, subarctic conditions such as ANVs.

Data availability

All data in this manuscript are publicly available.