Underground Coal Thermal Treatment as a Potential Low-Carbon Energy Source

We evaluate a novel energy strategy, underground coal thermal treatment (UCTT); it involves slowly pyrolyzing coal in-situ, transforming it to synthetic gas stream containing hydrogen and low molecular-weight hydrocarbons, liquid fuel and char. This evaluation assesses the life-cycle energy and greenhouse gas (GHG) impacts of UCTT for all process stages. It is based on experimental results at two scales, a simple heat-transfer model and literature results. The results show that UCTT can produce a high-quality liquid product and a gas mixture. UCTT's GHG emissions are in the range of those reported for in situ processing of oil shale. Net energy returns (NERs) of 0.48- 4.7 are in the range reported oil sands (2.8) and oil shale (0.48 - 2.6). Product yield at low temperatures, heater temperature, the number of heaters and the moisture content of the coal are key factors in determining the feasibility of the UCTT process.

1 Underground Coal Thermal Treatment as a Potential 2 Low-Carbon Energy Source 3 4 Kerry E. Kelly1,2*, Ding Wang2, Michal Hradisky1, Geoffrey D. Silcox2, Philip J. Smith1,2, Eric G. Eddings1,2, David W. Pershing1,2 5 1 Institute for Clean and Secure Energy, 155 South 1452 East, University of Utah, UT 84112, 6 USA 7 2 Department of Chemical Engineering, 3290 Merrill Engineering Building, 50 S. Central 8 Campus Dr., University of Utah, UT 84112, USA 9 * CORRESPONDING AUTHOR 10 Mailing address: 380 INSCC, 155 South 1452 East, University of Utah, UT 84112, USA, E- 11 mail: [e-mail]; Fax: [phone] KEYWORDS 13 Coal pyrolysis, carbon capture, in situ coal, greenhouse gas emissions, net energy return 14 HIGHLIGHTS 15 o UCTT can produce a high-quality liquid product and gas mixture. 16 o Product yield at low temperatures, heater configuration, the number of heaters and coal 17 18 19 moisture content are key factors in the feasibility of this process. o NERs and GHG emissions are in the range of those reported for in situ production of oil shale. 1 20 ABSTRACT 21 We evaluate a novel energy strategy, underground coal thermal treatment (UCTT); it involves 22 slowly pyrolyzing coal in-situ, transforming it to synthetic gas stream containing hydrogen and 23 low molecular-weight hydrocarbons, liquid fuel and char. This evaluation assesses the life-cycle 24 energy and greenhouse gas (GHG) impacts of UCTT for all process stages. It is based on 25 experimental results at two scales, a simple heat-transfer model and literature results. The results 26 show that UCTT can produce a high-quality liquid product and a gas mixture. UCTT's GHG 27 emissions are in the range of those reported for in situ processing of oil shale. Net energy returns 28 (NERs) of 0.48- 4.7 are in the range reported oil sands (2.8) and oil shale (0.48 - 2.6). Product 29 yield at low temperatures, heater temperature, the number of heaters and the moisture content of 30 the coal are key factors in determining the feasibility of the UCTT process. 31 2 32 1. INTRODUCTION 33 With current coal mining technologies and production rates, the US has approximately 270 34 years of coal reserves.1 In addition to recoverable coal reserves, the US has vast coal resources, 35 which are currently unrecoverable due to their depth, access, and other factors.2 This provides an 36 opportunity for an in situ technology to recover otherwise unrecoverable energy from coal. 37 Although natural gas prices are at historic lows in the United States, price increases are 38 projected, and prices will increase more rapidly with US natural gas exports.3 Furthermore, 39 methane recovery from coal seams is common, but coalbed methane (CBM) produces less than 40 1% of a coal bed's energy content.4 In situ pyrolysis offers the possibility of substantially more 41 energy recovery from the resource and the potential to convert the high-carbon content fuel into 42 a lower carbon content, higher heating value syngas or liquid fuel. 43 Figure 1 shows an example of the UCTT concept. This novel in situ pyrolysis process, UCTT 44 is proposed and evaluated. By slowly heating coal in-situ, the coal is transformed from long 45 chain geopolymers to a synthetic gas stream containing gas and liquid products, and char. The 46 coal contains native moisture as well as mineral mater, and these components are also heated but 47 do not transform into valuable products. This process has the potential to leave large portions of 48 the carbon from the coal in the ground in the form of char. Although UCTT requires additional 49 energy input compared to CBM, the added resource utilization and carbon management may 50 make this process worthwhile and motivates its evaluation. 51 UCTT differs significantly from underground coal gasification (UCG); it indirectly heats the 52 coal to pyrolysis temperatures (200 - 600°C) rather than injecting air/oxygen mixtures to directly 53 gasify the coal. UCTT offers several potential advantages over UCG including an improved 54 ability to discontinue operations if needed and a reduced risk of subsidence. Several 3 55 organizations have investigated UCG5, and pilot studies have been performed in China6, 56 Australia5, the US5,7, and South Africa5. Although UCG research appears to be active in some 57 countries, low oil prices have led to a decreased interest in UCG. 58 Although peer-reviewed studies of UCTT-type processes are limited, studies evaluating other 59 in situ fossil fuel processes report life-cycle GHG emissions for the production of transportation 60 fuels. These studies include thermal treatment of heavy oil, in situ production from oil sands and 61 the Shell in situ conversion process (ICP) for oil shale conversion. As conventional sources of 62 crude oil become scarce, transportation fuels are increasingly being produced from lower quality 63 resources, like heavy oil and oil sands, and potentially oil shale. Brandt and Unnsach8 examined 64 the energy intensity of thermally enhanced (steam injection) oil recovery of heavy oils in 65 California. They report well-to-pump (WTP) greenhouse gas (GHG) emissions of 32 - 47g CO2 66 e/MJ for gasoline produced from this resource (lower heating value, LHV). In comparison, well- 67 to-wheels GHG emissions from conventional petroleum sources in the US are 18.1 g CO2 e/MJ 68 (US EPA 2009).9 Brandt and Unnsach8 found that the GHG emissions vary with energy demand 69 of the heavy oil treatment (i.e., steam to oil ratio), choice of fuel used for steam generation, co- 70 generation of electric power, and the electricity mix. 71 Brandt10 evaluated GHG emissions from the production of gasoline using the Shell ICP. This 72 process heats an oil-shale field in situ, releasing liquid- and gas-phase fuels. Concerns over 73 potential groundwater contamination led to the installation of a freeze wall to isolate the 74 processing area from the water table. First, heater and producer wells are drilled. In the heating 75 wells, electrical heaters heat the oil shale to 340 - 370°C over a period of several years. The 76 liquid and gas fuels flow to the production wells for recovery. These products are then upgraded, 77 transported, and refined into gasoline. The study reports GHG emissions of 38 - 63 g CO2 e/MJ 4 78 gasoline. Work began on ICP in western Colorado in the 1980s, but activities in the US have 79 ceased. However, work on this process continues in Jordan11 and Israel12. 80 In addition to studies of in situ processes to provide transportation fuels, researchers have 81 proposed electricity production from oil shale with in situ carbon capture (EPICC). This method 82 employs a solid fuel cell underground to heat a shale formation.13 The produced gas from this 83 process flows back to the fuel cell to provide additional energy to generate electricity and to heat 84 the formation. They report that EPICC's GHG emissions are 51 - 99 kg CO2/MWh compared to 85 92-145 kg CO2/MWhr for pulverized coal with carbon capture. This work is in the conceptual 86 stages, and the authors cite EPICC's potential drawbacks including uncertain operation of 87 subsurface fuel cells, potential geologic impacts without pressure management, and economic 88 concerns associated with the value of stranded energy left in the formation and the long time 89 period for retorting. 90 The goal of this study is to begin to understand the feasibility of a UCTT process, specifically 91 by estimating UCTT's cradle-to-gate, life-cycle energy and greenhouse gas (GHG) emissions. 92 The analysis includes the impacts of well drilling, heating the formation, recovery, cleanup, and 93 transportation of the UCTT products. The energy required and product yields are based on 94 experimental results and simulations that rely on the properties from the experiments. 95 2. MATERIALS AND METHODS 96 This study uses a simplified process model life-cycle assessment approach to determine energy 97 and GHG emissions associated with UCTT. All results are presented on a LHV basis of the coal 98 and products. Figure 2 shows the processes considered in the UCTT analysis. It is envisioned as 99 a cradle-to-gate analysis with final products being transportation fuel (conventional gasoline) and 100 electricity generated from the gas-phase products. The evaluation does not include the energy 5 101 and GHG emissions associated with the construction of the refineries and power plant or the 102 manufacture of the drilling rig, the well casing, the well cement, or the associated fittings and 103 equipment. The UCTT process transforms coal in the ground into char and two product streams: 104 a two-phase liquid and a gas stream containing hydrogen and low molecular-weight 105 hydrocarbons (typically less than C4). While several options for heating a candidate formation 106 exist, this analysis focuses on electrical heating of the formation. The gas-phase products are 107 burned to reduce the purchased electricity needed to heat the formation. The liquid products are 108 refined into a transportation fuel, conventional gasoline. As discussed in Section 3.2, UCTT will 109 not likely produce sufficient gas-phase products to permit the sale of excess electricity. 110 The following subsections describe the resource, life-cycle stages, and other related processes. 111 2.1 Resource 112 This evaluation, including the experimental and simulation studies, is based on a Utah Sufco 113 coal, a high-volatile, low-moisture bituminous coal. Table 1 shows the coal properties. The 114 Sufco coal mine is located in Sevier County, UT in the Blackhawk Formation of the Wasatch 115 Plateau coalfield; it is one of the longest continuously running underground long-wall mines in 116 the US. It has approximately 126 million tons of recoverable resource and its annual production 117 in 2012 was 5.7 million tons.14 Its average thickness is approximately, 3.5 m, although thicker 118 portions of the seam exist, and the overburden depth ranges from 100 - 600 m.15 119 The evaluation assumes a 10 m thick seam of coal at a depth of 333 m; this hypothetical coal 120 seam is sufficiently contiguous in the radial direction for the heat to be completely absorbed by 121 the coal over the time period of interest. The seam is heated with a 10 m long, 0.25 m diameter 122 heater. Only one well is drilled for the analysis; the heating well is also assumed to be the 6 123 producing well. The analysis assumes a formation pressure of 30 bar and a formation 124 temperature of 20°C. 125 2.2 Drilling 126 For drilling the vertical well (a heating and producing well), diesel fuel consumption and well 127 depth come from the Utah Division of Oil Gas and Mining Well Reports,16 and the average diesel 128 consumption of 12.4 l/m for wells less than 3300 m deep is used. This value is in the range of 129 reported fuel consumption of 18.6 l/m for well drilling.17 130 2.3 Heating the Formation 131 A preliminary analysis revealed that the energy needed to heat the formation and product yield 132 are critical to the feasibility of UCTT, and these depend on the coal, char, and product properties. 133 Consequently, much of this study focused on gathering experimental results to understand these 134 key parameters. 135 requirements and GHG emissions for the other life cycle stages. As discussed in the following sections, literature data provided energy 136 UCTT experiments were performed at two scales: scoping studies in a high-temperature, high- 137 pressure, 1.9-cm diameter, fixed-bed reactor and larger scale studies in a 15-cm diameter, high- 138 pressure, high-temperature rubblized-bed reactor.18 The scoping studies identified the most 139 promising conditions as well as data for development of a yield model. The larger scale studies 140 allowed for more representative product generation. 141 The composition of the gas- and liquid-phase products comes from GC and GC/MS analysis, 142 respectively, of the products from the larger scale reactor. Although additional species are 143 present in limited amounts, in the baseline case the gas-phase products are represented by: H2 144 (3%, 25%), H2O (7%,7%), CO (10%, 6%), CH4, (35%, 39%), C2H4 (7%, 4%), C2H6 (10%, 6%), 145 and CO2 (28%, 12%) (by weight and mole %, respectively). 7 146 For the uncertainty evaluation, the experimental ranges for the gas-phase species vary from H2 147 (0.83 - 3.5%), CO (4.5 - 15%), CH4, (25 - 39%), C2H4 (1.1 - 7.1%), C2H6 (7.2 - 11%), and CO2 148 (29 - 49%) (by weight), and the ratio of liquid to total product ranges from 0.30 to 0.58 (by 149 weight). Consequently, the LHV of the product mixture ranges from 1.22 x 104 to 1.93 x 104 150 kJ/kg. 151 The liquid products contain approximately 85% carbon, 11% hydrogen, 0.5% nitrogen and 2% 152 sulfur (by weight, ranges not available). The elemental composition of the liquid is in the range 153 of that reported for conventional crude oil.19 154 distillation profile, crude oil's simulation profile, and the products single carbon number profile, 155 respectively, and illustrate that the UCTT product is "light-crude like". Figures A.2 - A.4 show the product's simulated 156 Additional measurements were performed to determine the heat capacity (Cp) and thermal 157 conductivity of the coal and char. These properties are used in the simulations and discussed in 158 the following subsection. Heat capacity for the coal and char were measured with a TA DSC- 159 Q20 differential scanning calorimeter. The measured Cp values were 1.25 ± 0.04 J/g°C for the 160 coal and 1.41 ± 0.36 J/g°C for the char. 161 uncertainty can be found in the Appendix. Thermal conductivity was inferred from the larger 162 scale experimental results as described in the Appendix. As a function of temperature, the 163 thermal conductivity of the aggregate char, product mixture is: Details of the heat capacity measurements and k = 3.37 x 10-4 T + 0.19 164 165 where, 166 T= temperature, K. 167 k = thermal conductivity, W/m K. 8 168 All properties of interest − density, heat capacity, thermal conductivity and thermal diffusivity 169 − change as a function of temperature and coal conversion, which is also a function of 170 temperature. The temperature simulation and yield estimates assume that properties remain 171 constant until 234°C, one degree above the temperature at which water boils at 30 bar. As coal is 172 converted to char and liquid and gas products, the properties in the temperature simulation 173 changed as a weighted average of the aggregate composition. 174 properties for all product species at UCTT temperatures and the formation pressure led us to the 175 assumption of a simplified product mixture of 30% CH4, 30% CO2, 30% C2H6, and 10% water 176 (by weight). The liquid properties were based on the single-carbon number analysis of the liquid 177 product (Figure A.4). The Peng-Robinson Polar properties in the ProMaxTM process simulator 178 provided the liquid and gas properties as a function of temperature at 30 bar. The coal and char 179 properties are based on the experimental measurements for heat capacity and thermal 180 conductivity, described above. 181 respectively.20 182 temperature can be found in the Appendix (Figure A.5 - A.8). 183 The difficulty in obtaining The coal and char specific gravities are 1.31 and 0.99, Relationships for each of the properties and their uncertainties as a function of Thermal diffusivity was calculated using: 𝛼= 𝑘 𝜌𝐶! 184 where, 185 k = thermal conductivity, W/m K. 186 ρ = density, kg/m3. 187 Cp = heat capacity, J/kg K. 188 The thermal diffusivity for coal is 1.76 x 10-7 m2/s at temperatures below 234°C when it begins 189 to transform into char. The thermal diffusivity of the coal, char, and product aggregate at 600°C 9 190 is 1.14 x 10-6 m2/s. Literature values for thermal diffusivity of coal and char vary widely. 191 Clendenin et al.21 summarize the effect of coal treatment temperature on coal/char properties and 192 report the diffusivity of raw coal, approximately 1 x 10-7 m2/s, and char treated to 600°C, 3.5 x 193 10-7 m2/s. They also report that coal/char thermal diffusivity is constant until approximately 194 300°C. It then increases with temperature in the range of temperatures of interest to a UCTT 195 process (200 - 600°C). 196 diffusivity and uncertainty estimates as a function of temperature. Figure A.8 shows the coal, char, and product aggregate thermal 197 The evaluation also considers an enhanced thermal conductivity case, in which the coal 198 thermal conductivity is 0.87 W/m K22 and the char thermal conductivity is 1.7 W/m K23. 199 Wellington et al. 22 reported that coal in situ had a higher conductivity than pulverized coal. This 200 results in a coal diffusivity of 5.31 x 10-7 m2/s and a weighted average coal, char, product 201 diffusivity of 1.64 x 10-6 m2/s at 600°C. 202 203 Temperatures were predicted as a function of time and distance from the heater using a onedimensional unsteady heat conduction in cylindrical coordinates, given by: 1 𝜕 𝜕𝑇 1 𝜕𝑇 𝑟 = 𝑟 𝜕𝑟 𝜕𝑟 𝛼 𝜕𝑡 204 Where: 205 r = radius, m, with radial steps of 0.25 m. 206 t = time, s. The time step for the baseline case is 4750 s. 207 T = temperature, °C. 208 α = thermal diffusivity, m2/s. See Figure A.8. 209 The equation was solved using an explicit finite-difference method where the initial 210 temperature of the outside surface of the 0.25 radius heater was 800°C. This was consistent with 211 the larger scale experimental studies. The far-field background temperature of the formation was 10 212 20°C. The initial temperatures of the system were: the heater at 800°C and the coal formation at 213 20°C. Each cylindrical element occurred at intervals of 0.25 m away from the heater and the 214 elements extended 17.25 m from the heater, and each time step was 4750 s. The single-well 215 configuration's temperature prediction did not include the energy needed to heat and evaporate 216 the water, but the multiple-well configuration did include water heating and evaporation. For 217 both configurations, the energy needed to heat and evaporate was considered, as discussed 218 below. The energy needed to vaporize the pyrolysis products was neglected. Above 233°C (the 219 boiling temperature of water at 30 bar), the thermal diffusivity changed as a function of 220 temperature (Figure A.8). These predicted temperatures were used in the yield model and in 221 estimating the energy required to heat the coal, char and product mixture. 222 223 The energy requirement was calculated for the coal and water separately. For the coal, the energy required for each element at each time step in the simulation was calculated using: 224 Q = m Cp (T - T(t-Δt)) 225 where, 226 m= mass of material remaining within each element, kg. Above a temperature of 233 °C, coal 227 transforms into the product mixture, and the product leaves the system. 228 Cp = heat capacity of coal at the average temperature of the element. Above a temperature of 229 233°C, Cp changes, see Figure A.7. 230 T(t-Δt) = average temperature at the previous time step at the midpoint of the element, °C. 231 T = midpoint temperature at time t, °C. 232 233 234 The cumulative energy required to heat the coal was calculated by summing the required energy at each time step. The cumulative energy required to heat the water is given by: 11 235 Q = mΔHvap, H2O + mCp (T - To) 236 where, 237 Q = energy, kJ. 238 m= mass of water in each element. 239 To = 20°C, background temperature. 240 T = Final midpoint temperature at each element, °C. If the final temperature exceeded the 241 boiling point of water at the formation pressure, the temperature was 233°C (boiling point of 242 water @ 30 bar). 243 Cp = 4.18 kJ/kg. 244 ΔHvap, H2O (@30 bar) = 1,794 kJ/kg. This was included only if the final temperature of the element 245 exceeded 233 °C. At the end of the evaluation period if the element exceeded a temperature of 246 233 °C, all of the water in the element was assumed to evaporate and leave the system 247 248 The energy requirements for heating the water in each element to its final temperature were then summed. 249 During the UCTT process, an operator might choose to heat the formation for a given time 250 period and then collect the product, or he could choose to produce continually. If products are 251 collected continually, energy leaves the system continually. It is unclear how such a process 252 would actually operate; consequently this energy leaving the system is included because it is 253 considered more conservative, i.e., requiring more energy. This energy is given by: 254 Q = m Cp (T - To) 255 where, 256 m = cumulative mass of material leaving the system from each element, kg. 257 Cp = heat capacity of product mixture, kJ/kg. 12 258 To = initial temperature of the element, °C. 259 T = midpoint temperature at time t, °C. 260 261 An additional simulation was performed using STAR-CCM+ to investigate the effect of multiple 262 wells and heat loss to the over/underburden using the same physical properties for the coal seam 263 and heater temperature as the baseline case. Figure A.9 shows the STAR-CCM+ simulation 264 domain, which included wells spaced 12.5 m apart and a 10-m thick coal seam with 25 m of 265 overburden and underburden. The overburden and underburden were assumed to have a 266 moisture content of zero. 267 The STAR-CCM+ solves the conductive component of heat transfer: 𝜕 𝜕𝑡 ! 𝜌 𝑐! 𝑇𝑑𝑉 = − 𝒒" ∙ 𝑑𝒂 ! 268 where, 269 t = time, seconds. 270 ρ = the solid density, 271 𝑐! = specific heat, 272 T = the temperature, 273 𝒒" = solid heat flux vector, 274 𝒂 = the normal area vector, 275 A = the area, and 276 V = the volume. 277 The solid heat flux vector is computed using Fourier's Law and is dictated by the fixed heater 278 temperature of 800°C. Second-order discretization for both temporal as well as spatial terms was 13 279 used along with an implicit iterative approach. A time step of 7,200 seconds was selected. On the 280 surface of the heater, the mesh size was set to 2 cm. Therefore, the heating wells were 281 sufficiently resolved to accurately capture the heat supplied by heaters into the coal seam. The 282 very fine heater mesh slowly transitioned into a coarser mesh, with maximum cell size of 20 cm, 283 for total of 21 million computation cells for the entire simulation domain. 284 2.4 Yield Model 285 A sigmiodal function was fit to the experimental scoping data (Figure 3), and this function was 286 used to predict yield as a function (mass fraction of total coal utilized) of temperature in the 287 UCTT simulations, discussed in the following subsection. The scoping data provided more 288 consistent temperature profiles compared to the larger scale studies where measuring the internal 289 temperature of the coal was challenging. The experimental data were fit by minimizing the least 290 squares difference between the predicted and measured yields. For the baseline case, the yield 291 was set at 0.4525 at temperatures of 950°C and above, from coal's proximate analysis and at less 292 than 0.03 at a temperature of 340°C. 293 The baseline sigmoidal fit is given by: 𝑌𝑖𝑒𝑙𝑑 =   𝑌𝑖𝑒𝑙𝑑!"# 1 +   𝑒 !.!"!#$(!!!"#$) 294 where, 295 Yield = the fraction of the coal that is volatilized on an as-received basis. 296 Yieldmax = the maximum yield, from the proximate analysis, 0.4525, with a range of 0.446 297 (Umin) - 0.462 (Umax) (as received). 298 k = 472 °C (baseline), with a range of 567 - 357 °C (Umin - Umax). 299 Temp = temperature, °C. 300 14 301 The Appendix provides additional detail on the yield function and the associated uncertainty. 302 Based on the experimental results, the model assumes a 50:50 split between liquid and gas-phase 303 products for the baseline cases. 304 2.5 Liquid Products and Refining 305 Because the hydrogen and carbon content are similar to that of conventional crude, US average 306 refining GHG emissions and energy requirement are selected. These are 3.2 MJ/kg crude and 307 8.69 g CO2e/MJ crude, respectively.24 The mass of crude oil moved into a refinery is slightly 308 less than the mass of the product leaving the refinery and the heating value of the liquid product 309 and crude oil differ. Consequently, the method outlined in Wang et al.25 was used to adjust from 310 a crude to gasoline basis. 311 2.6 Gas Products and Processing 312 The product gas contains approximately 30% CO2, and removing this prior to combustion is 313 relatively expensive. Consequently the gas product was presumed to be combusted in a 50% 314 efficient natural-gas combined cycle (NGCC) turbine. This process generates 147 g CO2 e/MJ of 315 electricity.26 Spath and Mann26 report on a 53% efficient NGCC, and the emissions were 316 adjusted to 50% efficiency. Because all of the gas is combusted in the NGCC to generate the 317 electricity needed to heat the coal, no excess electricity is available for sale (see Section 3.2). 318 2.7 Liquid Product Transport 319 Energy and GHG emissions from transport of the liquid product are based on a US average of 320 321 6720 J/MJ product and 0.493 g CO2 e/MJ product.24 2.8 Electricity Generation and Fuel Combustion 15 322 Energy consumption and GHG emissions were assumed to be from a 50% efficient NGCC 26 323 plant. GHG emissions from combustion of gasoline were 75.2 g CO2 e/MJ, and diesel 324 emissions were 77.0 g CO2 e/MJ. Natural gas combustion emits 56.6 g CO2 e/MJ. 325 2.9 Net Energy Return 326 The net energy return (NER) and the net external energy return (NEER) are useful metrics for 327 comparing fuel sources. The NER is the ratio of usable energy gained from an energy resource 328 to the energy used (directly and indirectly) to obtain that resource. The NEER is the ratio of 329 usable energy gained from an energy resource to the direct energy used (excluding e.g., any 330 produced fuel consumed while producing the resource). For this analysis, the NER value 331 assumes that the products are the refined liquid and gas product, and the NEER assumes that the 332 gas is combusted to generate electricity to heat the formation. The NER and NEER presented 333 are at the point of consumption, the fuel pump. 334 2.10 Sensitivity and Uncertainty 335 Table 2 summarizes selected emission factors, assumptions and properties investigated for the 336 sensitivity and uncertainty analyses. The preliminary analysis of UCTT identified the energy 337 required to heat the formation and product yield as being key to the feasibility of this process. 338 Recognizing this, the sensitivity analysis focused on several factors affecting the yield and 339 heating of the formation. These included coal moisture content, thermal conductivity, yield 340 model, refining energy requirements, heater temperature and formation heating period. The 341 uncertainty analysis explored how the range of experimental uncertainty affects the results, and it 342 propagated all experimental uncertainty. For a single well, the results present the least and most 343 favorable experimentally derived parameters (Umin - Umax). The analysis also included the 16 344 effect of multiple wells and heat loss to the overburden. This is denoted as "Multiple" in Table 2 345 and the subsequent results. 346 3 RESULTS AND DISCUSSION 347 Analysis of the products from the larger scale studies revealed that carbon tends to partition to 348 the char, while hydrogen tends to partition to the gas and liquid products (Figure 4). Oxygen 349 tends to be found in the water and gas products, while sulfur tends to partition to the char and 350 liquid products. The mass balance shown in Figure 4 uses average coal composition; however 351 inherent inhomogeneties in the coal exist. Consequently, the balance does not account for 352 approximately 3% of the carbon and 6% of the hydrogen, and it slightly over-accounts for 353 oxygen and sulfur. The experimental results show that UCTT can produce a high-quality liquid 354 product and a lower energy content gas. This suggests that UCTT is worth further evaluation. 355 Figure 5 shows the energy required and the energy produced, and Figure 6 shows the simulated 356 temperature profile for baseline conditions. In this figure for a 2.5-year heating period, the edge 357 of the element located 17 m away from the heater, which is next to the simulation boundary 358 (17.25 m away from the heater), is 0.04°C above the background temperature of 20°C; at 5 years 359 this difference increases to 0.5°C. The temperature and yield profiles follow similar trends for all 360 cases, with temperatures decreasing rapidly with distance from the heater. Figure 7 shows the 361 temperature profile for the multiple heater configuration. 362 For all cases, the produced gas contains less energy than that required to heat the formation. 363 For the baseline case, the produced gas contains approximately 45% of the energy needed to heat 364 the formation, and 7% of the energy in the products is needed replace the energy leaving the 365 system with the heated products (7%). For the 20% moisture case (Table 2), the produced gas 366 contains only 23% of the energy needed to heat the formation (at 2.5 years). For the Umax case, 17 367 the produced gas contains 85% of the energy required to heat the formation (at 2.5 years). 368 Considering all the life-cycle energy requirements, heating the formation requires the majority of 369 the energy, greater than 80%, for all cases. 370 3.1 Net Energy Return 371 As time progresses, NER values decline since relatively small volumes of coal are treated to 372 significant temperatures. Considering just the process of heating the formation, Figure 5 shows 373 how NER declines over time. It also shows the energy required and the energy produced. For 374 the baseline case, it shows that the NER declines from 1.4 at 2.5 years to approximately 1.2 at 5 375 years. Figure 5 also confirms that the energy leaving the system with the products is small 376 compared to the energy required to heat the coal and moisture. 377 Figure 8 shows that NERs decrease as a function of time for heating the formation for all 378 cases. The multiple-well case shows a slightly different trend than the single-well cases where 379 NERs increase until approximately 1 year and then decline. The temperature profiles of the 380 multiple-well case are retarded by the heating and evaporation of the water, which decreases the 381 initial NERs. In addition, the multiple-well case NER looks slightly wavier than the single-well 382 cases, which is due to the simulation mesh size away from the heater. The multiple well, Umax 383 and maximum-yield cases exhibit the greatest NERs. Figure 8 illustrates the importance of yield 384 model, liquid fraction of the product, heater temperature and configuration, and the initial 385 moisture content. As moisture content increases NER decreases. At a 20% moisture content, 386 heating the formation requires more energy than it produces in approximately 1 year, indicated 387 by an NER of less than one. At 5 years, both the minimum yield and the 10% moisture case have 388 process NERs of less than 1. 18 389 If the WTP lifecycle stages are considered, NER values are lower than those for just heating 390 the formation. Table 3 summarizes the WTP NER and NEER values for each case. This 391 suggests that the cases with a moisture content much greater than the baseline (3.2% moisture) or 392 a heater temperature much lower than the baseline are unlikely to be feasible. However, coals 393 that exhibit higher yields (>0.3) are more likely to be commercially viable. 394 The baseline NEER for gasoline produced from the UCTT process is 1.1 (range of 0.28 - 4.7), 395 and the baseline NER is 1.3 (range of 0.48- 4.5). NEER tend to be lower than NER values for 396 all cases except for the multiple well, Umax and maximum-yield cases, which have large gas 397 yields relative to the energy required. Figure 9 compares the range of UCTT NERs to other 398 sources of liquid transportation fuels and to electricity generated from coal. NERs for some 399 transportation fuels vary widely, and limited NERs are reported for oil shale.27-29 Recognizing 400 that oil shale has a limited commercial production history, particularly in situ, a good deal of 401 uncertainty is associated with oil shale NERs, with NERs ranging from less than 127 to 2.5 (if gas 402 is exported)28 or 2.629 for refined fuel. The baseline NER for UCTT is less than half that for oil 403 sands and in the range of that for oil shale (0.48 and 2.6).27,29 The in situ shale and coal processes 404 face similar challenges with energy losses, i.e., waste energy to the formation. However, the 405 multiple-well case (NER of 4.5 at 2.5 years) compares favorably with oil sands (NER of 2.8).42 406 A number of factors would limit UCTT's feasibility including coals with high moisture content 407 (>5%) and/or low volatile yield (<30%). 408 Selecting the formation heating time would likely need to be a balance between the economics 409 and NER. To overcome the fixed cost of well drilling and infrastructure, a minimum yield will 410 be necessary. However, longer formation heating times lead to larger volumes of coal that are 411 heated to low temperatures, resulting in wasted energy and lower NERs. 19 412 3.2 CO2 Emissions 413 Of the life-cycle stages, heating the formation clearly drives the CO2 emissions (Figure 10). 414 The results presented in this figure assume that all of the natural gas is combusted to generate the 415 electricity needed to heat the formation and that gasoline is the only remaining product. 416 Transport and refining CO2 emissions are similar to those for other sources of liquid fuels. On 417 an energy content basis, well-to-pump (WTP) life-cycle CO2 emissions are 178 g CO2e/MJ 418 baseline (50 - 558 g CO2 e/MJ range). 419 The WTP GHG emissions for the liquid fuel produced from the UCTT process with the low- 420 moisture coal (50 - 558 g CO2 e/MJ range) are dominated by the process of heating the coal 421 formation, which contributes 169 g CO2 e/MJ of the 178 g CO2 e/MJ WTP GHG emissions 422 (baseline at 2.5 years of heating time). Literature values for GHG emissions from in situ 423 recovery of coal resources are limited. However, in situ recovery of oil shale and sands can 424 serve as useful benchmarks, as does coal-to-liquid conversion such as the Fischer Tropsch 425 process. 426 Figure 11 compares the range of WTP GHG emissions for gasoline produced by UCTT with 427 GHG emissions from other transportation fuels. The emissions are several times greater than the 428 US average WTP GHG emissions of 18.1 g CO2 e/MJ.9 As with the NER values, the multiple- 429 well, Umax and the maximum-yield cases provide the most attractive (lowest GHG) emissions, 430 and Umin, high moisture and minimum-yield cases lead to the highest GHG emissions. This 431 evaluation's WTP GHG emissions overlap with the upper end of the range of GHG emissions 432 reported for the Shell in situ conversion process (ICP) of 38-63 g CO2 e/MJ transportation fuel10 433 and generally fall into the range of values reported by Kelly et al.27 of 60 - 120 g CO2 e/MJ for in 434 situ production of crude from oil shale. UCTT's GHG emissions also generally fall within the 20 435 wider range of values reported for ex situ oil shale operations (40-180 g CO2 e/MJ gasoline).30,31 436 These values are nearly double those reported for production of reformulated gasoline produced 437 from in situ recovery of Canadian oil sands (29 - 35 g CO2 e/MJ) reported by McKellar et al.32 438 The comparatively low GHG emissions from Canadian oil sands reflect the relative ease in 439 recovering bitumen from sands and the shallow depth of the sands. 440 transformations of coal to liquid fuels via the Fischer Tropsch process (73 - 140 g CO2 e/MJ, 441 diesel product33) are also in the range of those reported for the conceptual UCTT process 442 described this paper (50-558 g CO2 e/MJ, gasoline product). WTP GHG for 443 This type of fuel resource, in situ coal, falls into the category of lower quality crudes that, if 444 developed, would result in increased GHG emissions compared to conventional crude sources. 445 Dooley et al.34 suggest that the development of unconventional fuels in the United States would 446 significantly increase GHG emissions. Brandt and Farrell30 suggest that transitioning to low- 447 quality crude sources, such as tar sands or coal-derived liquids, could raise upstream GHG 448 emissions by several gigatonnes annually. 449 3.3 Sensitivity and Uncertainty 450 Of the factors considered in the sensitivity and uncertainty analysis, the yield model, the heater 451 temperature, the number of heaters, and the moisture content of the coal had the greatest effect 452 on both the GHG emissions and the NER. Because of the large volumes of coal heated to 453 relatively low temperatures, using the maximum yield model (Umax includes maximum yield 454 model and the most favorable ranges of other parameters) resulted in more favorable NERs and 455 GHGs than any of the other single-well cases (Table 2). Existing yield models focus on rapid 456 heating rates, typical of pulverized coal combustion. Better models are needed to predict product 457 yields at the low ultimate temperatures and the very low heating rates typical of a UCTT process 21 458 (< 1°C/hr). This evaluation examined three yield models, all based on sigmoidal fits of the 459 experimental data. These models produce yields that follow trends reported in the literature for 460 coal devolatilization over temperatures ranging from approximately 200 - 600°C at higher 461 heating rates >0.5°C/s.35 462 Two subcomponents of the yield model are the ultimate yield and the product composition. 463 The ultimate yield for the baseline coal is typical of a high-volatile bituminous coal (38 - 45%).35 464 However, coal volatile content varies widely from 2 - 10% for anthracites, to 20 - 30% for low 465 and medium-volatile bituminous coals and 40 - 50% for subbituminous and lignites.35 The 466 product composition and the liquid to gas ratio are also related to the yield model. The liquid 467 products have a greater heating value than the gas products. The LHV of the liquid products is 468 43,000 kJ/kg, and the LHV for the gas product mixture is 16,500 kJ/kg (range of 12,200 to 469 19,300). The gas products contain approximately 30% by weight CO2 and 7% H2O, which 470 contributes to the lower energy value of the gas products. Consequently, increasing the ratio of 471 liquid to gas products improves the NER of the process. 472 Lower rank coals tend to exhibit high gas yields and low liquid yields, while high-volatile 473 bituminous coals tend to exhibit high liquid and moderate gas yields.35 The coal selected for this 474 analysis is a high-volatile bituminous coal with a low moisture content. With these attributes, it 475 is a more likely candidate for the UCTT process than a lower rank, higher moisture coal or a 476 low-volatile high rank coal. 477 Heater configuration, heater temperature and coal moisture content also have a significant 478 effect on the viability of the UCTT process. Multiple heating wells offer the most favorable 479 NERs and GHG emissions, and this array of closely spaced heaters would be more operationally 480 realistic than a single-well configuration. This dense well spacing improves product recovery per 22 481 unit energy input, making UCTT a potentially promising processes. The low-heater-temperature 482 case NER is approximately one half of the baseline case. A heater temperature of 400°C would 483 cause the UCTT process to become unfeasible. 484 temperature of 600°C, which corresponds to the experimental conditions, commercially available 485 electrical heaters may have difficulty maintaining a temperature of 600°C over the distances 486 required in a full-scale UCTT process. Although the baseline case has a heater 487 The moisture content of the baseline coal is low, and increasing the moisture content from 488 3.2% (baseline) to 10 and 20% increases the GHG emissions by a factor of 30 and 38%, 489 respectively. Different ranks of coal exhibit a wide variety of moisture contents, ranging from 490 less than 5% for anthracites and bituminous/subbituminous coals to greater than 30% for 491 lignites.35,36 492 production.37 Selecting a low-moisture resource would be a key criterion of a UCTT process. Bituminious and subbituminous coals account for more than 90% of US coal 493 Although thermal diffusivity affects the time required to heat the coal seam and this property 494 has a good deal of uncertainty, in the range considered in this evaluation its effect on the NER 495 and GHG emissions is limited. Consequently it is not included in the results. However, 496 diffusivity may be an important consideration for the economics of a UCTT process. 497 3.4 Other Uncertainties 498 The sensitivity and uncertainty analysis began to address potential uncertainties in the UCTT 499 process including formation and coal/product properties, experimental error, and product yield. 500 However additional uncertainties will affect the feasibility of UCTT. These include the behavior 501 of water in the formation, mineral matter, which affects pyrolysis behavior, the environmental 502 fate of the liquid- and gas-phase products, design and maintenance of well casing, and 503 operational challenges associated with recovering the liquid products from the formation. It is 23 504 not known whether water in the formation would flow into the heating zone as the product is 505 recovered. Any additional water would continue to absorb energy and make the energy balance 506 less favorable. If CBM was extracted from the formation, the coal seam would have been 507 dewatered, thus limiting water flow into the formation. 508 hydrocarbons may flow into the heating zone and be recovered, which would improve UCTT's 509 feasibility. This study did not consider the environmental fate of any unrecovered products. 510 However as shown in Figure 6, the pyrolysis zone extends approximately 3 meters from the 511 heater after 2.5 years of heating. Unrecovered product would likely adsorb to the heated coal 512 seam as it cools. In addition to water, other light 513 The yield models explored in this evaluation do not cover the full range of possible models, 514 such as Biagini and Tognotti38 or Kobayashi et al.39 Furthermore, the yield models considered in 515 this analysis did not account for any recovery losses associated with gathering the products from 516 the well and transporting these to the surface. 517 4 CONCLUSIONS 518 The experimental results suggest that UCTT may be promising because it can produce a high- 519 quality liquid product as well as a gas that could help heat the formation. UCTT tends to leave 520 the carbon in the ground and produce a lower carbon fuel than the parent coal. Heater 521 temperature and configuration, product yield at low temperatures and the moisture content of the 522 coal are the key factors in determining the feasibility of UCTT. Accurate low-temperature, low 523 heating rate yield information is critical to further evaluation of UCTT. 524 limitation, this evaluation suggests that high volatile, low moisture coal would be a good target 525 for UCTT. The most favorable UCTT results (NER and GHG emissions) occur for multiple 526 heater wells with a heater temperature of 600°C (or greater) combined with a low moisture (5% In spite of this 24 527 or less), high volatile content (30% or more) coal. The NER and GHG emissions for the process 528 are less favorable than those for oil sands and are in the general range of that for oil shale. 529 However due to the limited commercial adoption of in situ oil shale production and the low 530 natural gas and liquid fuel prices, UCTT is unlikely to become commercially viable in the near 531 term. 532 ACKNOWLEDGMENTS 533 This material is based upon work supported by the Department of Energy under Award 534 Number DE-FE0001243 and DE-NT0005015. The views and opinions of authors expressed 535 herein do not necessarily state or reflect those of the United States Government or any agency 536 thereof. Many thanks to Adel F. Sarofim, Presidential Professor, University of Utah, deceased, 537 for his numerous contributions. 538 REFERENCES 539 (1) British Petroleum. BP Statistical Review of World Energy; 2015. 540 541 (2) US Energy Information Administration. Recoverable Coal Reserves; Washington, DC, 2013. 542 543 (3) US Energy Information Administration. Effect of Increased Natural Gas Exports on Domestic Energy Markets; Washington, DC, 2014. 544 545 (4) Flores, R. M. Coalbed Methane: From Hazard to Resource. Int. J. Coal Geol. 1998, 35 (14), 3-26. 546 547 548 (5) Clean Air Task Force. Coal Without Carbon An Investment Plan for Federal Action Expert Reports on Research, Development, and Demonstration for Affordable Carbon Capture and Sequestration; 2009. 549 550 (6) Li Y; Liang X; Liang J. An Overview of the Chinese UCG Program. Data Sci. J. 2007, 6, S460-S466. 551 552 (7) Institute, G. R. Detailed Evaluation of Process and Environmental Data from the Rocky Mountain 1 Underground Coal Gasification Field Test; 1988. 25 553 554 (8) Brandt, A. R.; Unnasch, S. Energy Intensity and Greenhouse Gas Emissions from Thermal Enhanced Oil Recovery. Energy & Fuels 2010, 24 (8), 4581-4589. 555 556 557 (9) US Environmental Protection Agency. Lifecycle Greenhouse Gas (GHG) Emissions Results Spreadsheets. US Environmental Protection Agency 2009, pp Docket ID: EPA - HQ - OAR - 2005-0161 - 0938. 558 559 560 (10) Brandt, A. R. Converting Oil Shale to Liquid Fuels: Energy Inputs and Greenhouse Gas Emissions of the Shell in Situ Conversion Process. Environ. Sci. Technol. 2008, 42 (19), 7489-7495. 561 (11) Jordan Oil Shale Company. Jordan Oil Shale Company http://josco.jo/. 562 (12) Israel Energy Initiatives. Israel Energy Initiatives http://www.iei-energy.com/. 563 564 565 (13) Mulchandani, H.; Brandt, A. R. Oil Shale as an Energy Resource in a CO2 Constrained World: The Concept of Electricity Production with in Situ Carbon Capture. Energy and Fuels 2011, 25 (4), 1633-1641. 566 (14) BNSF Railway. Guide to Coal Mines; Fort Worth, TX, 2013. 567 568 (15) Keith, A. C. Coal Quality Characteristics of Utah's Coal Beds in and Near Potentially Producible Coal Tracts; Salt Lake City, 1989. 569 570 (16) Utah Division of Oil, G. & M. Oil and Gas Well File Search. Utah Division of Oil, Gas & Mining: Salt Lake City 2014. 571 572 (17) Advanced Resources, I. Greenhouse Gas Life-Cycle Emissions Study: Fuel Life-Cycle of U.S. Natural Gas Supplies and International LNG; Washington, DC, 2008. 573 574 (18) Smith, D. 2015. Underground Coal Thermal Treatment, Task 6 Topical Report, Utah Clean Coal Program. DOE Award Number: DE-NT0005015; Salt Lake City, 2015. 575 (19) Beychok, M. Petroleum Crude Oil. The Encyclopedia of Earth; Hogan, C. ., Ed.; 2011. 576 577 (20) Gloyn, R. E. Energy, Mineral, and Ground-Water Resources of Carbon and Emery Counties, Utah. ; Salt Lake City, 2003. 578 579 580 (21) Clendenin, W. C. C. Thermal and Electrical Properties of Anthracite and Bituminous Coals. In ransactions of the 7th annual anthracite conference of Lehigh University; Bethlemehm, PA, 1945. 581 (22) Wellington, Z. E. Emissionless Energy Recovery from Coal. 33,482, 2000. 26 582 583 584 (23) Chern, J. S.; Hayhurst, A. N. A Model for the Devolatilization of a Coal Particle Sufficiently Large to Be Controlled by Heat Transfer. Combust. Flame 2006, 146 (3), 553-571. 585 586 (24) Argonne National Laboratory. Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) Model. 587 588 589 (25) Wang, J. Allocation of Energy Use in Petroleum Refineries to Petroleum Products Implications for Life-Cycle Energy Use and Emission Inventory of Petroleum Transportation Fuels. Int. J. LCA 2004, 9 (1), 34-44. 590 (26) Spath P.L., M. K. Life Cycle Assessment of a NG Power Generation System; 2000. 591 592 593 (27) Kelly, K. E.; Wilkey, J. E.; Spinti, J. P.; Ring, T. a.; Pershing, D. W. Oxyfiring with CO2 Capture to Meet Low-Carbon Fuel Standards for Unconventional Fuels from Utah. Int. J. Greenh. Gas Control 2014, 22, 189-199. 594 595 (28) Brandt, A. R. Greenhouse Gas Impacts of Declining Hydrocarbon Resource Quality: Depletion, Dynamics, and Process Emissions, University of California, Berkeley, 2008. 596 597 598 (29) Bunger, J. W.; Russell, C. P. Thermodynamics of Shale Oile Production. In Oil Shale: A Solution to the Liquid Fuel Dilemma; Ogunsola, O. I., Hartstein, A. M., Ogunsola, O., Eds.; ACS, 2010; pp 89-101. 599 600 601 (30) Brandt, A. R.; Farrell, A. E. Scraping the Bottom of the Barrel: Greenhouse Gas Emission Consequences of a Transition to Low-Quality and Synthetic Petroleum Resources. Clim. Change 2007, 84 (3-4), 241-263. 602 603 604 605 (31) Brandt, A. R.; Boak, J.; Burnham, A. J. Brandt, A.R.; Boak, A.R., Burnham, A.R. Carbon Dioxide Emissions from Oil Shale Derived Liquid Fuels. In Oil Shale: A Solution to the Liquid Fuel Dilemma; Ogunsola, O. I., Hartstein, O. I., Ogunsola, A. M., Eds.; ACS Publication Series 1032, 2010. 606 607 608 (32) McKellar, J. M.; Charpentier, A. D.; Bergerson, J. a.; MacLean, H. L. A Life Cycle Greenhouse Gas Emissions Perspective on Liquid Fuels from Unconventional Canadian and US Fossil Sources. Int. J. Glob. Warm. 2009, 1 (1/2/3), 160. 609 610 611 (33) Jaramillo, P.; Griffin, W. M.; Matthews, H. S. Comparative Analysis of the Production Costs and Life-Cycle GHG Emissions of FT Liquid Fuels from Coal and Natural Gas. Environ. Sci. Technol. 2008, 42 (20), 7559-7565. 612 613 614 615 (34) Dooley, J. J.; Dahowski, R. T.; Davidson, C. L. The Potential for Increased Atmospheric CO2 Emissions and Accelerated Consumption of Deep Geologic CO2 Storage Resources Resulting from the Large-Scale Deployment of a CCS-Enabled Unconventional Fossil Fuels Industry in the U.S. Int. J. Greenh. Gas Control 2009, 3 (6), 720-730. 27 616 617 (35) Smith, R. J. The Structure and Reaction Processes of Coal; Plenum Press: New York, 1994. 618 (36) Radovic, L. R. Nergy and Fuels in Society; McGraw-Hill, 1992. 619 620 (37) US Energy Information Administration. Subbituminous and Bituminous Coal Dominate U.S. Coal Production; Washington, DC, 2011. 621 622 (38) Biagini, E.; Tognotti, L. A Generalized Correlation for Coal Devolatilization Kinetics at High Temperature. Fuel Process. Technol. 2014, 126, 513-520. 623 624 (39) Kobayashi, A. F. Coal Devolatilization at High Temperatures. Symp. Combust. 1977, 16 (1), 411-425. 625 626 627 (40) Elgowainy, M. W. and J. H. and J. B. D. and H. C. and A. Well-to-Wheels Energy Use and Greenhouse Gas Emissions of Ethanol from Corn, Sugarcane and Cellulosic Biomass for US Use. Environ. Res. Lett. 2012, 7 (4), 45905. 628 (41) Inman, M. The True Cost of Fossil Fuels. Sci. Am. 2013, 308 (4), 58-61. 629 630 (42) Brandt, A. R.; Englander, J.; Bharadwaj, S. The Energy Efficiency of Oil Sands Extraction: Energy Return Ratios from 1970 to 2010. Energy 2013, 55, 693-702. 631 632 (43) Cleveland, C. J.; O'Connor, P. a. Energy Return on Investment (EROI) of Oil Shale. Sustainability 2011, 3 (11), 2307-2322. 633 634 635 (44) Brandt, A. R. Converting Oil Shale to Liquid Fuels with the Alberta Taciuk Processor: Energy Inputs and Greenhouse Gas Emissions. Energy and Fuels 2009, 23 (12), 6253- 6258. 636 637 638 (45) Searchinger, T.; Heimlich, R.; Houghton, R. a; Dong, F.; Elobeid, A.; Fabiosa, J.; Tokgoz, S.; Hayes, D.; Yu, T.-H. Use of U.S. Croplands for Biofuels Increases Greenhouse Gases through Emissions from Land-Use Change. Science 2008, 319 (5867), 1238-1240. 28 639 640 Figure 1. Example of a UCTT process. The process is not drawn to scale. 641 642 Figure 2. System boundaries for the analysis of UCTT. The shaded box indicates that heating 643 the formation is critical to the analysis. 29 644 Table 1. Sufco coal properties. These are presented on a dry ash free basis. This is based on the 645 average of 6 samples. The average moisture content is 3.21% and the average ash content is 646 5.04%. Fixed carbon (%) Heating (MJ/kg) value Carbon Hydrogen (%) Nitrogen (%) Oxygen (%) Sulfur (%) 5.61 1.72 13.56 0.65 (%) 54.75 32.7 78.46 5.00E'01% 4.50E'01% 4.00E'01% Mass Fraction 3.50E'01% 3.00E'01% Scoping 2.50E'01% RBR Baseline fit 2.00E'01% Umin Umax 1.50E'01% 1.00E'01% 5.00E'02% 0.00E+00% 0 100 200 300 400 Temperature (C) 500 600 647 648 Figure 3. Sigmoidal fit of the experimental results. The baseline model uses the scoping data, 649 while the Umin and Umax models fit the range of all experimental results and use the minimum 650 and maximum experimental yields, respectively. 651 30 652 Table 2. Emission factors, coal properties, and selected parameters used in the sensitivity and uncertainty analyses. Process Baseline Umax Coal moisture Background temperature (°C) Drilling diesel fuel consumption (l/m) Heating the formation Heater temperature (°C) Cp coal (J/g°C) Maximum yield (%) Liquid product (%) Yield model Refining (MJ/kg crude)8 Transport (J/MJ) 24 Efficiency of NGCC electricity generation 653 654 655 a) At 600°C. b) 𝑌𝑖𝑒𝑙𝑑 =   Umin 3.2 % 20 3.2% 20 3.2% 20 Max Yield 3.2% 20 12.416 12.416 12.416 12.416 12.416 12.416 12.416 12.416 12.416 800 1.41 45.25 50 b 3.2 6720 50% 800 1.14 46.2 58.2 c 3.2 6720 50% 800 1.47 44.6 29.5 d 3.2 6720 50% 800 1.41 46.2 58.2 c 3.2 6720 50% 800 1.41 44.6 29.5 d 3.2 6720 50% 800 1.41 50 60 b 2.88 6720 50% 800 1.41 40 40 b 3.52 6720 50% 800 1.41 45.25 50 b 3.2 6720 50% 800 1.41 45.25 50 b 3.2 6720 50% !.!"#" !!  ! !.!"!#$(!"#!!"#$) Min Yield 3.2% 20 Max NER 3.2% 20 Min NER 3.2% 20 10% water 10% 20 20% water 20% 20 (baseline yield model). c) 𝑌𝑖𝑒𝑙𝑑 =   !.!"# !!  ! !.!"!#$(!"#!!"#$) Low heat 3.2% Enhanced k 3.2% 20 Multiple well 3.2 % 20 12.416 12.416 800 1.41 45.25 50 b 3.2 6720 50% 800 1.41 45.25 50 b 3.2 6720 50% 600 1.41 45.25 50 b 3.2 6720 50% . d) 𝑌𝑖𝑒𝑙𝑑 =   !.!!" !!  ! !.!"!#$(!"#!!"#$) . !250!! !200!! Mole fraction in the original coal sample, char, and liquid products. C H N O S Coal 0.497 0.427 0.009 0.064 0.002 Char 0.749 0.214 0.015 0.020 0.002 Liquid 0.27 0.78 0.11 0.00 Gas! Water! Liq!prod! Char! Coal! Moles& !150!! !100!! !50!! !"!! Ccoal! Cprod! Hcoal! Hprod! Ncoal! Nprod! Ocoal! Oprod! Scoal! Sprod! Figure 4. Moles of carbon (C), hydrogen (H), nitrogen (N), oxygen (O), and sulfur (S) in the original coal (as received) and in the char, liquid, and gas. The coal was heated to an internal temperature of 540°C at ambient pressure and held for 3 hours. 32 Energy#needed#producing# Energy#needed# NER# 1.E+09# 3.0# 1.E+09# 2.5# 8.E+08# 2.0# 6.E+08# 1.5# 4.E+08# 1.0# 2.E+08# 0.5# 0.E+00# 0.0# 1.0# 2.0# 3.0# 4.0# 5.0# NER$ Energy$(kJ)$ Energy#produced# 0.0# Years$ Figure 5. Energy produced, energy required, energy required with simultaneous production and NER as a function of time for the process of heating the formation. 33 900" 0.25"m" 800" 0.50"m" 700" 0.75"m" Temperature)(C)) 600" 1.0"m" 500" 1.25"m" 1.50"m" 400" 0.25" 10.0"m" 2.25"m" 300" 2.50"m" 200" 5.0"m" 100" 10.0"m" 7.5"m" 0" 0" 2" 4" 6" 8" 10" 12" 14" 16" 18" 20" 22" 24" 26" 28" 30" 17.5"m" Time)(months)) Figure 6. Temperature profile at various radial distances from the heater (distance 0.25 m increments) a 2.5-year heating period for the baseline case. 34 Figure 7. The temperature profile for the multiple well configuration after 2.5 years of heating. 35 7.0# 6.0# ##Baseline# ##10%#water# ##20%#water# ##Umin# ##LowHeatTemp# ##MaxNER# ##MulCple# Umax Power#(##Umax)# Log.#(#MaxYield)# MaxYield# ########2#per.#Mov.#Avg.#(##MulCple)# 5.0# NER$ 4.0# 3.0# 2.0# 1.0# 0.0# 0.2# 1.1# 2.0# 3.0# 4.0# 5.0# Years$ Figure 8. Comparison of NERs for the process of heating the coal formation. MinNER is not shown because it overlaps with the lower NER cases. Table 3. Summary of WTP NER and NEER values at 2.5 and 5 years of UCTT production. Umax Umin Max yield Min yield Max NER Min NER 10% Water 20% Water Low heat Mult well Base 2.5 years NER 1.26 3.13 0.48 2.81 0.51 1.58 1.04 1.01 0.75 0.577 4.48 NEER 1.11 3.86 0.28 3.20 0.40 1.51 0.82 0.86 0.61 0.458 4.74 NER 1.06 2.85 0.14 2.96 0.39 1.13 0.87 0.84 0.62 0.421 4.14 NEER 1.01 3.33 0.08 3.66 0.30 1.27 0.74 0.77 0.54 0.327 4.68 5 years 36 12 Low$ High$ 10 NER/EROI 8 6 4 2 0 UCTT Figure 9. Oil shale Corn ethanol Canadian oil sands Gasoline from Electricity from crude coal NER/EROI for gasoline produced from the UCTT process, oil shale13,27, corn ethanol40,41, Canadian oil sands42, conventional crude43, and electricity generated from coal41. The highest oil shale NER29 in this figure was adjusted from a shale-oil crude basis to a refined fuel basis assuming a refining energy requirement of 507.5 MJ/ton shale processed and a ratio of HHV of refined product to upgraded shale oil of 0.9 from Brandt44. 37 700 Transport LF refining Extraction 600 CO2 e (g/MJ) 500 400 300 200 100 0 Baseline Multiple Umax Max yield Max NER 10% Water 20% Water Low heat Figure 10. GHG emissions (CO2 e) from baseline, maximum uncertainty (favorable), maximum yield, NER max, 10% water, 20% water, and lower heater temperature cases at a 2.5-year heating period. The Umin case is not shown on this scale. It is 558 g CO2 e/MJ. 38 600 500 CO2 e (g/MJ) 400 300 High Low 200 100 0 UCTT Oil shale Coal to liquids Corn ethanol Canadian oil Gasoline from sands crude Figure 11. Range of CO2 e emissions per MJ of transportation fuel for UCTT, oil shale (low, for Shell's in situ conversion process10 and, high, for ex situ shale30), diesel produced from coal via the Fischer Tropsch process33, corn ethanol45,40, Canadian oil sands32, and gasoline from crude oil32. 39