Alexander
Payne
,
Guillermo
Garcia-Garcia
and
Peter
Styring
*
UK Centre for CO2 Utilization, Department of Chemical and Biological Engineering, The University of Sheffield, Sir Robert Hadfield Building, Sheffield, S1 3JD UK. E-mail: p.styring@sheffield.ac.uk
First published on 3rd May 2023
With an unabated global petrochemical growth, more sustainable production methods for production of materials and fuels are essential in a decarbonised future. Although Carbon Capture Utilisation (CCU) is generally considered a sustainable production route, it is imperative to compare its environmental and economic performance with that of current methods. This article reviews the environmental impact and economics surrounding conventional production of propane and propene via natural gas liquid fractionation and crude oil refining for propane. In addition, fluid catalytic cracking and steam cracking were explored for propene production. A CCU process has been modelled using Aspen Plus and analysed through Life-Cycle Assessment and Techno-Economic Analysis. Processes simulated include carbon capture using piperazine, dry methane reforming, direct syngas to propane and methanol to propene. The results obtained show a significant reduction in environmental impacts across multiple impact categories for both products when compared to conventional production. In addition, the price of propene from CCU was competitive with conventional. However, the price of propane was significantly higher. Sensitivity analysis of hydrogen production technology and electricity grid emission intensity identified them both as key determinants of economic and environmental performance.
Decarbonisation pathways are ways in which emissions reductions can be achieved through phasing in/out of technologies, introducing new laws around emission criteria or implementing carbon taxes that incentivise sectors to reduce emissions from their direct operations and entire supply chain. Collectively, these mitigation pathways aim to limit warming to below 2 °C relative to pre-industrial levels.4 However, without additional mitigation steps beyond those present today, there is a high risk of severe, widespread, and irreversible impacts globally.5
Carbon Capture Utilisation and Storage (CCUS) consists of capturing CO2 and either storing it or using it as a raw material. As such, CCUS forms part of several decarbonisation strategies worldwide. For example, in the UK, the “Zero Carbon Humber” project aims to capture CO2 from the UK's most carbon intensive industrial cluster.6 However, Carbon Capture Storage (CCS) has several challenges such as the slow pace of assessing and exploiting storage resources, large economic costs, lack of consistent legislation, and low public awareness.7 Alternatively, the captured CO2, a waste product, can be converted into several value-added products via Carbon Capture Utilisation (CCU). Examples include propane and propene. Conventionally, propane is produced from petroleum refining or natural gas processing8 and propene from steam cracking (SC) and refinery operations. A detailed analysis of the conventional methods to produce propane and propene can be found in sections 1–2 of the ESI.† CCU can provide an alternative production pathway that both reduces greenhouse gas emissions and fossil resource depletion by producing chemicals not from fossil fuels, but from captured CO2. The state of the art of CCU, particularly regarding its use to produce propane and propene, can be found in section 3 of the ESI.†
The aim of this article is to look at the viability of CCU production methods of propane and propene by assessing them economically and environmentally. Both products are firstly explored in terms of their environmental and economic impact of their conventional production pathways. Next, a CCU alternative process is designed and modelled using Aspen Plus. Life-Cycle Assessment (LCA) is then applied to evaluate the environmental performance of such CCU production method. Techno-Economic Analysis (TEA) is finally used to combine process modelling with economic evaluation to provide a thorough understanding of the economic cost of the CCU process proposed.
In 2018, the United States was the largest natural gas producer in the world, followed by Russia.11 Conventional natural gas is stored in a naturally porous reservoir with impermeable rock strata.12 However, shale gas is unconventional as the shale rock is not naturally porous, so requires hydraulic fracking to allow the gas to migrate from pockets within the rock formation. Fracking uses a mixture of water, sand and proprietary chemicals that is pumped underground at high pressures to create a fracture network. The supply of natural gas is predominantly via fracking and in 2019 it accounted for 87% of total U.S. production.13
To calculate the environmental impacts reported in this section, emission intensities were found for each process and aggregated into each of the impact categories. Calculation and sources for this section are in the sections 4–16 of the ESI.† Results are shown in Fig. 1 and 2.
The natural gas route has a significantly higher global warming potential (GWP) for extraction compared to crude oil, as seen in Fig. 1 and 2. The completion of a well by fracking requires the “flowback” of drilling and reservoir fluids to open pits, which results in significant venting of natural gas, where the length of the period depends on the permeability of the reservoir.14 Furthermore, the figure for natural gas extraction is likely to be underestimated as a further review of three sources15–17 (section 12 of the ESI†) found that for shale gas the extraction emission intensity was 14.2 g CO2e per MJ natural gas. The extraction emission intensity used is 5.6 kg CO2e per kg propane (Fig. 2), but using the new value it would be 12.8 kg CO2e per kg propane. Nevertheless, it must be noted that diverse sources estimate different GWP for such processes, mostly due to different feedstock, technological and geographical considerations. For example, the US GREET model reported a carbon intensity 11075 g per mmbtu (0.5 kg CO2e per kg propane)18 for propane production, while the Canadian Propane Association19 estimated a carbon intensity of 74 g CO2e per MJ (3.6 kg CO2e per kg propane), including combustion.
Particulate matter formation for the crude route is significantly worse due to flaring of natural gas. In oil exploration, natural gas is less valuable, and offtake requires transportation infrastructure to deliver it to consumers which is both challenging logistically and costlier than the value of the gas.20 Therefore, fracking for natural gas exploration contributes more to environmental impacts from venting during flowback and fugitive emissions than flaring of the gas. Hence, their reduced particulate contribution.
The flowback of fracking fluids contains excess salts, high levels of trace elements and radioactive materials that can pollute groundwater.13 However, the flowback of fracking fluids is mostly recycled to frack additional wells and the remainder trucked to wastewater treatment facilities and deep injection wells. Therefore, while surface water pollution is a serious problem, most U.S. regions have significant available capacity of deep injection wells for liquid waste disposal.21
In addition, one impact associated with natural gas extraction not quantitatively considered here are the small-to-moderate magnitude seismic activity linked to hydraulic fracturing of wells and in some cases microearthquakes.13
Specific processing steps for raw natural gas include amine gas treating and dehydration. Both these processes combined totalled 0.27 g CO2 per MJ natural gas and 0.028 g CH4 per MJ natural gas. However, as data could not be found for all processes, the average for processing overall from six sources in the U.S. was used. Further environmental impacts of processing include the evaporative losses and venting of degraded products of the amine solution such as nitrosamines and nitramines which are possible carcinogens and would contribute to human toxicity impacts.22 Surveys of gas processing plants using amine solvent report average losses of 0.2 g amine per Nm3 natural gas processed.23 Similarly, in dehydration, hazardous pollutants such as benzene, toluene, ethylbenzene and xylenes (BTEX) that have an affinity for the glycol solution are vented in the regeneration step in the stripper.24 Methane also has an affinity for the amine solution and approximately 0.971 g of methane per kg of natural gas treated is vented to atmosphere from the stripper.25
Natural gas distribution has a higher GWP (2.12 kg CO2e per kg propane) than for crude oil (0.34 kg CO2e per kg propane). This is due to fugitive emissions, particularly methane from sources such as compressor stations and valves.
Atmospheric distillation contributed mostly to GWP with a value of 0.27 kg CO2e per kg propane. In a petroleum refinery, propane is contained within light gases or gaseous refinery streams from atmospheric distillation. Two specific petroleum products containing propane are LPG and fuel gas. The energy use for each refinery product has been allocated based on energy content in certain sources which varies minimally from a mass-based allocation.26 Energy is consumed in the form of electricity, heat and steam where natural gas and refinery fuel gas are used to meet heating and steam demand.27 In 2012, 37% of processing energy at U.S. refineries was refinery fuel gas and 25% was natural gas.28
Life-cycle data for specific technologies, such as the cryogenic expansion process to recover NGLs from natural gas, are limited, so the fractionation figure was based on one U.S. source.25 However, the total greenhouse gases reported in the U.S. for natural gas processing in 2019 was 57.5 Mt CO2e.17 When combined with the total dry natural gas and NGL production (Table 1), the emission intensity can be approximated on an energy basis to 0.076 g CO2e per MJ. The value from this approximation of 0.068 kg CO2 per kg propane is minimally different to the source used:25 0.064 kg CO2 per kg.
Energy carrier | Volume | Energy content | Allocation (%) | ||
---|---|---|---|---|---|
Value | Unit | Value | Unit | ||
Dry natural gas | 962![]() |
Million m3 | 36.62554 | MJ m−3 | 83.58 |
Ethane | 667![]() |
Thousand barrels | 3.249572 | GJ per barrel | 5.14 |
Propane | 579![]() |
Thousand barrels | 4.051414 | GJ per barrel | 5.57 |
Normal butane | 157![]() |
Thousand barrels | 4.568392 | GJ per barrel | 1.71 |
Isobutane | 152![]() |
Thousand barrels | 4.568392 | GJ per barrel | 1.65 |
Pentanes plus | 203![]() |
Thousand barrels | 4.874358 | GJ per barrel | 2.35 |
Water usage for the natural gas route totalled 14.474 kg per kg propane, where extraction contributed to 83% of this figure. The values for extraction were based on two studies of U.S. and Canadian shale (section 11 of the ESI†), while for processing this was based on a Chinese source (section 13 of the ESI†). As processing requires the gas to meet a standardised pipeline specification this value should be independent of location unlike extraction. Clark et al.30 found that shale gas consumes 13 to 37 l per GJ over its life cycle, or between 12 and 33 kg per kg propane. For the crude oil route, the total water consumption amounted to 2.75 kg per kg propane, sourced from a global weighted average for extraction and the average across three U.S. refinery configurations (cracking, light and heavy cracking). Atmospheric distillation consumed 0.975 kg water per kg propane. The major contributors to water use are for cooling due to evaporative losses in cooling towers and boiler feed water for steam generation.31 However, a study of crude oil production from five North American locations by Ali and Kumar32 found that a barrel of conventional oil cycle consumes 1.71 to 8.25 barrels of fresh water (33.4–157.41 kg per kg propane) over its life and 2.4 to 9.51 barrels of fresh water are withdrawn (46.11–181.26 kg per kg propane). Water usage is a significant problem as in the U.S. the newest oil and gas developments are in drought-affected and arid regions such as the Colorado River Basin.33 Furthermore, 72% of all water used in the U.S. comes from fresh surface water sources such as rivers and lakes and 10% comes from ground water (aquifers).31
Both routes use water for preparation of the drilling fluid, which has the function of cooling the drill bit, removing drilled rocks, and providing hydrostatic pressure to prevent well collapse.34 Water consumption for oil production is used mainly for enhanced oil recovery. Typically, drilling waste and produced water is discharged to sea for offshore oil exploration if it meets environmental requirements.35 This is devastating in terms of marine ecotoxicity, as the average oil concentration of discharged produced water is 3.9 mg l−1,36 resulting in 67.4 mg kg−1 of propane. The presence of aromatic hydrocarbons, alkylphenols, heavy metals and naturally occurring radioactive material causes the most environmental concern.37
Human toxicity of the crude route totalled 3.77 × 10−8 kg 1,4-dichlorobenzene equivalents (1,4-DCBe) per of kg propane. This is due to mercury presence in the raw natural gas that is emitted during venting and flaring. Fossil resource depletion for the crude route was 717.4 kg antimony equivalents (Sbe) per kg of propane, and 588.189 kg Sbe per kg propane for natural gas.
Ozone formation and acidification potential was significantly higher for the crude oil route totalling 164.24 g non-methane volatile organic compounds equivalents (NMVOCe) per kg and 378.68 g SO2e per kg propane respectively. The natural gas route resulted in values of 5.75 g NMVOCe per kg propane and 3.83 g SO2e per kg propane respectively. For both routes, extraction was the highest contributor.
Overall, as seen in Fig. 3, production of propane from crude oil refining generated 1.51 kg CO2e per kg propane, whereas for natural gas this was 10.1 kg CO2e per kg propane. Therefore, based on GWP the natural gas route is the most environmentally damaging. However, for all other impact categories the crude oil was the most damaging. Therefore, propane production from CCU represents a good opportunity to prevent the significant environmental impacts associated with natural gas and crude oil extraction, processing, and distribution.
To prevent thermal cracking of the heavy fractions from atmospheric distillation,40 a further step of vacuum distillation is required to generate the feedstock for FCC. Therefore, the additional processing and associated energy use has contributed to a higher value across all impact categories.
To calculate the environmental impacts reported in this section, emission intensities were found for each process and aggregated into each of the impact categories. Calculation and sources for this section are in sections 17–21 of the ESI.† Results are shown in Fig. 4 and 5.
The FCC unit is the single biggest source of atmospheric pollution in an oil refinery due to sulphur oxides and particulates41 and of carbon dioxide emissions, accounting for approximately 30% of the total emitted from a refinery.42 Emissions from FCC include the combustion products from process heaters and the catalyst regenerator. FCC units are considered “self-contained” in terms of their energy sourcing and Jia et al.43 found that 82.9% of energy required for the process in China was from petroleum coke combustion. Contaminants present in the feedstock, which include metals such as nickel, vanadium and copper but also heteroatoms such as sulphur and nitrogen contaminants, end up in the coke, which is burned in the regenerator. The results show significantly higher contributions to acidification (1188.3 and 19.33 g SO2e per kg propene respectively) and particulate matter formation (940.2 and 26.3 g PM 10 per kg propene respectively) for the FCC process than for the SC process, where the greatest contributor was sulphur oxide.
The yield of propene from FCC is heavily dependent on the catalyst used, reactor configuration and operating conditions such as catalyst to oil ratio, residence time and reaction temperature.44 The results presented here were calculated using a base case yield across three studies that found that propene yield was 5.35 wt%. However, higher yields in excess of 20 wt% can be achieved when maximising propene production is the objective function of the refinery.44
Another contributor to particulate emissions and resource depletion often overlooked is the use of zeolite catalysts in FCC, which must be replaced constantly due to catalyst attrition and irreversible contamination. Cyclones and electrostatic precipitators are used to separate catalyst particles. However, a fraction becomes entrained in the exhaust gas.
Specific data for fractionation of the products of FCC and SC could not be found. However, the fractionation process is similar to NGL fractionation, so these data was used for analysis.
Water usage for the SC route amounted to 10 kg per kg propene and for the FCC route 4.1 kg per kg propene. Similar results were obtained by Yang and You,45 who found propene production from SC of naphtha in a mass-based allocation amounted to 11.25 kg per kg propene.
Although human toxicity for both routes were mostly caused by flaring and venting in crude oil extraction, the FCC units also emitted lead and arsenic. Overall, the value was 6.42 × 10−5 kg 1,4-DCBe per kg propene for the FCC route and 2.02 × 10−7 kg 1,4-DCBe per kg propene for SC.
Overall, the total GWP was higher for the SC route and totalled 9.21 kg CO2e per kg propene, whereas for FCC the total was 5.95 kg CO2e per kg propene. As seen in Fig. 6, there is less disparity in values of impact categories when compared to propane production. However, SC across all categories was more environmentally detrimental. Furthermore, the feed for FCC is abundant in a refinery compared to naphtha, so FCC can be considered more cost effective and less environmentally impactful.46 However, the FCC unit still represents a significant opportunity as a point source for carbon capture in CCU routes and would vastly reduce the environmental impact of a refinery.
Xiang et al.47 carried out a TEA of a 1.5 Mt per a crude oil to olefins plant in China. The technology used was naphtha SC and the product cost found was $1480 per t (9340 RMB per t). The study found that 88% of the product cost was from raw materials. The assumption used was based on 2012 figures of $110 per bbl crude oil.
Zhao et al.50 carried out an economic analysis of twenty light olefin pathways. The benchmark was SC of naphtha which had a production cost of $949 per t olefins where the price of naphtha was $868 per t. One pathway utilised natural gas to form methanol followed by MTO, which had a production cost of $769 per t olefins, whereas the MTP pathway had a production cost of $850 per t olefins. Furthermore, the use of natural gas to form syngas and subsequent conversion into olefins by FT had a production cost of $2356 per t olefins due to the low selectivity for light olefins. The price of natural gas used was $405 per t, and for all cases, raw material cost dominated the production cost. However, in this study all processes excluding SC used air separation to produce syngas rather than CO2 utilisation.
Chen et al.51 found that when simulating a CO2-rich natural gas from China in a dry reforming process propene cost was $1029 per t and propane cost was $536 per t.
Ghorbani et al.53 modelled a natural gas to liquid plant (excluding CCU) and found that the cost of production for liquid fuels from FT synthesis was $89 per t ($71.39 per m3 approximated using 800 kg m−3 FT product49).
Park et al.56 performed a TEA on nine different configurations of offshore NGL recovery processes. The natural gas flowrate modelled for all configurations was 472.44 t h−1 at atmospheric temperature. The operating cost varied between $7.5 M and $13.5 M ($2–3.6 per t natural gas feed at 90% uptime), while the capital cost ranged from $0.8 M to $9.7 M. The study found that compressor duty dominated the operating cost. Similarly, AlNouss et al.57 modelled 6 different NGL recovery processes for an 84 t h−1 plant. The operating cost varied between $12 M and $18 M ($18–27 per t natural gas feed at 90% uptime), while the capital cost ranged from $16 M to $25 M.
Economic analysis of crude oil refining to produce propane or FCC to produce propene are limited. However, a study found that 86% of production costs in crude oil refining depend on raw material cost.58 Therefore, data would not provide extensive insight due to price volatility.
Modelling described within this section has used U.S. sources where applicable to allow comparison to conventional production methods in previous sections. In Aspen Plus, the property methods chosen were ELEC-NRTL for carbon capture and Peng-Robinson for the propane and propene production route as seen in literature.59 Within the Aspen simulation, heat and energy integration was carried out to achieve a more efficient energy network.
The regenerated solvent stream (LEANOUT) formed a closed loop as it was recycled back to the absorber. Heat integration was used to recover heat energy of the stream, aiding both process economics and environmental performance. In the absorber, some PZ was lost due to entrainment in the GASOUT stream (>0.2 wt%), thus the addition of a makeup stream (PZ-MK). Similarly, evaporative losses of water were present in both the stripper and absorber, hence an additional makeup stream (H2O-PZ).
CH4 + CO2 → 2CO + 2H2 | (1) |
For the propane production route, hydrogen is used directly (H2ELPRO). For methanol synthesis, a compressor (CMPH21) was used to pressurise the gas to the required 49.95 bar.
2H2 + CO ↔ CH3OH ΔH = −91 kJ mol−1 | (2) |
CO2 + 3H2 ↔ CH3OH + H2O ΔH = 50 kJ mol−1 | (3) |
CO + H2O ↔ CO2 + H2 ΔH = −41 kJ mol−1 | (4) |
Maximising methanol production requires reducing yields of methyl-formate and higher alcohols from side reactions. The reactions occurring are catalysed by a copper and zinc-based catalyst and are exothermic. Therefore, to maintain the reactor temperature, cooling water is used as a utility. The reactor outlet (MTH1) is 73 wt% methanol, while 26 wt% in descending order is made up of carbon monoxide, carbon dioxide and hydrogen. In Aspen, sensitivity analysis found the optimal conditions for the two-phase separation of the unreacted components to be 15 °C and 1 atm. The unreacted components were returned to the initial feedstock mixer (FEEDMIX) as a recycle stream to improve conversion. The liquid phase outlet (MTH5) of the flash separator (SEPMTH) is 99.82 wt% methanol which is first heated (HTRMTH1, HTRMTH2) and then compressed (CMPMTH4) to 300 °C and 16.5 bar.
The MTP process is generally accepted to involve the primary step of methanol to DME dehydration, followed by the conversion of DME to light olefins. Therefore, an intermediate reactor (DMEREAC) was set up to maximise production of DME before transfer into the propene reactor (PRPEREAC). The importance of using a zeolite catalyst and specific operating conditions in the propene reactor is because of increased selectivity for propene and to prevent further reaction of the light olefins to paraffins, aromatics and higher olefins by hydrogen transfer, alkylation and polycondensation.67–69 A recommendation for further research would be to study the direct conversion of syngas to DME to avoid the methanol production step in the production of propene through the direct reaction of DME to olefins (DTO process).
From literature, industrial operation of DME reactors are at 300 °C and 16.5 bar as a 90% methanol equilibrium conversion (eqn (5)) can be achieved.70 To maintain the thermodynamically favourable conditions of the reactor (DMEREAC) and due to its exothermicity, cooling water was chosen as a utility. The reactor outlet consisting of 62 wt% DME (DME1) with the remainder water and unreacted methanol is heated (HTRDME1) to 425 °C and 1.5 bar.
2CH3OH ↔ CH3OCH3 + H2O | (5) |
The production of DME in a separate reactor further diversifies the plant as DME is a viable alternative to diesel, producing less NOx, SOx and PM.71 Therefore, in the event of challenging propane market conditions this could be sold.
The reactor outlet (CRDPRPE1), due to the high content of water, is cooled (COOLPRP1) to 20 °C and ambient pressure before entering a flash separator. Water is an unwanted by-product and would impact the duty of downstream distillation. The water content in the crude propene stream post separation (CRDPRPE3) is 3.2 wt%.
The distillation column (DE-PRP) was set up using a DSTWU block, operated at 16.9 bar with light (propene) and heavy key (pentane) recoveries of 0.99 and 0.01 respectively.61 The bottom product (DEPRPBT) is dominated by butane and pentane composing 9.99 and 86.5 wt% respectively. The top product (PROPENE) is composed of 91.1 wt% propene. To meet polymer grade propene purity >99.5% further work would be required such as optimisation of reflux ratio to increase the purity.
Zhang et al.75 reported that LPG could be produced from syngas using a hybrid, zeolite-methanol synthesis catalyst. The consecutive catalysis is efficiently carried out using a spherical zeolite shell and a metal-based catalyst core. The reaction mechanism involves four major steps: methanol synthesis within the core, dehydration of methanol to DME and olefins, selective hydrogenation to paraffins (C3–C4) and RWGS.76
The propane reactor (PROREAC) was based on the experimental results by Ge et al.77 on the use of a palladium-based methanol synthesis catalyst (Cu–ZnO/Pd-β) with a beta-zeolite shell. The process showed optimal performance at 350 °C, 21 bar and H2:
CO of 2.71 achieving a 44.4% hydrocarbon yield, 72.9% CO conversion and 51.5% selectivity to propane. The use of a palladium-supported catalyst is important for commercial application as standard Cu–Zn methanol synthesis catalyst experiences significant deactivation from water vapour produced by the RWGS.78 Water becomes strongly adsorbed to zeolite active sites and increases the selectivity of DME from CO2 hydrogenation.79
Operating pressure had little effect on hydrocarbon distribution but increased CO conversion. However, H2:
CO ratio was found to significantly increase propane selectivity as when the ratio was dropped from an optimal 2.5–2.7 to 1 CO conversion and propane selectivity dropped to 36% and 30% respectively while propene selectivity increased to 31%.74 High temperatures of 350 to 380 °C were found to be optimal for both CO conversion and propane selectivity.78 Conventionally, this reaction is carried out in a fixed-bed reactor. However, for commercial application this is limited by deactivation of the catalyst from sintering due to inefficient heat removal. Therefore, Zhang et al.80 investigated the use of a slurry reactor: the suspension of the bifunctional catalyst in an inert hydrocarbon liquid. Results found significant improvement of catalyst stability, increased propane selectivity and reduction in CO2 yield. Further benefits of such design include lower cost from simpler construction, increased mass and heat transfer and increased longevity of the catalyst. While this was not modelled, the results show significant relevance to both commercial viability and application.
Similar difficulty was experienced to model the propane reactor as for the propene reactor. As such, an Excel Solver was used to calculate the complete outlet yield distribution using the limited experimental results as constraints, an overall net equation (eqn (6)) occurring obtained from Zhang et al.75 and Aspen stream results.
2nCO + (n + 1)H2 ↔ CnH2n+2 + nCO2 | (6) |
The crude propane product (CRDPRO1) is cooled (COOLPRO2) before entry into a flash separator (SEPPRO1, SEPPRO2). The optimum temperature for the flash separator was found to be −72 °C through sensitivity analysis in Aspen, as it achieved the highest separation of hydrogen and methane to be recycled (RECYH21, RECYH22).
• Determine the overall environmental impact and resource consumption of a CCU production method for propane and propene.
• Determine which are the most relevant impact categories.
• Determine which parts of the production process contribute the most to environmental impact and explore alternative technologies to mitigate.
• Compare the production of propane and propene via CCU methods against their conventional counterpart.
The functional unit of the LCA is the production of 1 kg of propane and propene. CO2 is captured from a medium-sized FCC unit via chemical absorption using PZ. Propane and propene are produced via direct conversion of syngas and the MTP process, respectively. The system boundary includes the emission of CO2 from an FCC unit up to and including the production of both propane and propene (gate). End use of the product and distribution are excluded. Background processes considered include electricity generation and sourcing of natural gas and water. The chemical plant is modelled in the U.S., so data libraries selected were USLCI and ecoinvent 3.5. Data were obtained from simulation results in Aspen Plus and heuristic-based calculations where appropriate. Operational data to input into Aspen Plus was obtained through a literature review of both commercially available technologies, published studies and experimental data discussed in previous sections.
A key component in a refrigeration cycle is the compressor. This provides the work required to pressurise the saturated vapour from the evaporator, which is essential in facilitating heat transfer in the condenser from the higher temperature difference. As such, compressor duty was calculated from heuristics82 where the source of the mechanical work was provided from a gas turbine utilising natural gas, as this is the industry standard.
The duty required in Aspen was converted into compressor duty in W using 3.41 (BTU/HR) per W and 747.7 W per hp. To calculate the natural-gas consumption in the gas turbine to input into SimaPro, the calorific value of natural gas of 38.3 MJ m−3 was used and a mechanical efficiency of 0.4. Data from natural gas combusted in a U.S. industrial boiler was used. It was assumed that there are no fugitive emissions of refrigerant from the refrigeration cycles. Omission of such data results in an underestimation of majorly ozone depletion and GWP. Industrial estimated annual leakage rate can range between 7 to 25% of refrigeration volume, which is significant.83 However, the volume/inventory of refrigerants used was not determined and therefore the associated fugitive emission. Further research should investigate these effects in more detail.
Future work could also investigate the use of electric motors to mitigate against gas-turbine emissions. Only the environmental impact surrounding the compressor duty was included in this model and therefore the cooling duty required in the condenser (e.g. use of cooling water) was considered out of scope but should be included in further research. This omission results in water usage intensity and electrical consumption (fan utility in wet cooling tower) being underestimated. However, this impact is considered minor in this model.
First, the cooling water concentration must be determined which is central to the design as it dictates the flow, contact and quantity of water required to achieve the desired performance. To calculate the cooling water concentration, the following assumptions were made: hot-water temperature, 39 °C; cold-water temperature, 26 °C; wet-bulb temperature, 31 °C.
Water concentration was estimated with the sizing chart to calculate cooling water concentration for counterflow induced-draft cooling tower.84 Furthermore, the cooling water flowrate was calculated from the duty, heat capacity of water (4.18 J (g K)−1) and temperature change. The required area of cooling tower was calculated by dividing the cooling water flowrate by the cooling water concentration. The horsepower per area of cooling tower was calculated using the chart horsepower per tower area.84 Fan efficiency varies depending on the power consumption, so the correct efficiency was selected from size of the motor and belt required85 to calculate the total electrical power to input into SimaPro.
Another relevant environmental impact is the resource depletion from use of cooling water. Losses occur due to evaporative loss, drift loss (water entrainment in vapour) and blowdown (purge of water to maintain system solid concentration). Calculations and equations regarding these amounts are summarised in section 26 of the ESI.† The makeup water total equated to approximately 1.62% of cooling water flowrate and was inputted as an emission to air in SimaPro.
Natural gas was represented by “natural gas, high pressure (US), petroleum and gas production, on-shore, cut-off, S”. This assumption included energy use, infrastructure and associated emissions for onshore production in the Niger Delta (U.S.). This was justified as the model is based in the U.S. where gas supply is dominated by domestic sources.
Water for electrolysis was represented by “water, deionised, from tap water, at user (RoW) production, cut-off, S”. The dataset included the energy for operation, chemicals used for regeneration, emissions from regeneration chemicals, infrastructure of the plant and replacement of spent exchange resin. Electrolysis requires pure water (i.e. deionised) to prevent damage to the electrodes due to corrosion. Rest of the world (RoW) was chosen as no other option more relevant was available (e.g. U.S. or global (GLO)).
“Propylene (RoW), production, cut-off, S” was used to compare results to those generated by the model was chosen. This process used steam cracking of naphtha as the producing technology. RoW was chosen as above. As identified in the literature review, 47% of propene is sourced from steam cracking of naphtha. Therefore, this assumption would capture the relevant impacts.
Similarly, “propane (CA-AB) natural gas production, cut off, S” was used to compare results to those generated by the model. This includes exploration, drilling and ends at the gate of the processing plant. In addition, it includes all the fuels and emissions related to well testing, exploration, extraction and treatment (sweetening and drying): fugitive emissions, flaring, venting, and use of gas in turbines. As identified in the literature review, 60% of propane is produced from natural gas liquid fractionation. A limitation of such assumption is that the data are based on Alberta, Canada (CA-AB); however, as 98% of natural gas imports in the U.S. are from Canada, this was taken as a reasonable assumption.87 The most pressing limitation is that natural gas sourced from Canada, particularly Alberta, is where 85% of Canada's sour gas is produced. Therefore, in terms of environmental impacts with fugitive emissions, treatment, energy and material used to mitigate this is higher.88 Thus, overestimation is likely, compared to a U.S.-sourced scenario. The only other options in the databases were RoW and RER (Europe), which would not be appropriate to compare to conventional methods that are based on U.S. data.
Butane, pentane and ethane were produced in the model and were accounted for as avoided products. Therefore, the environmental impact of producing the same quantity via fossil-fuel derived sources was deducted from the total environmental impact of propane and propene. The assumptions for these were all “(CA-AB) natural gas production, cut off, S”, where the justification is the same as for propane above.
Water for cooling towers was represented by “process water, ion exchange, production mix, at plant, from surface water, RER S”. This assumption was not best represented, although there were no other alternatives for process water (industrial) within the databases. This is using data from Europe (RER), which is a limitation, however it considered that it is from a surface water source. Most water used by refineries comes from fresh water sources such as surface water.31
Finally, the allocation of emissions used an economic approach over the mass-based alternative. Use of a mass-based approach would have reduced the environmental burden of both products. For example, in the de-Propaniser (DE-PRO) along the propane production pathway, the top product (PROPANE) (99 wt% propane) mass flow was 5.96 kg s−1, and the bottom product (DEPROBT) was 3.76 kg s−1 (83 wt% butane). Therefore, based on a mass allocation, PROPANE would have been allocated circa 60%, compared to the economic approach of 85% of the total emissions. The advantage of economic allocation is that it allocates larger impacts to the products that the industry would favour their production (because of their higher prices). However, the drawback of economic allocation is the inherent instability as it is based on prices for products that vary based on market conditions.89 Therefore, future comparisons should consider the future prices with those used in this model (shown in Table 10).
Totals for propane and propene in Tables 2–5 have been specified (pre-avoided products) due to the production of valuable by-products such as ethane, butane and pentane and avoidance of carbon dioxide as an emission to air in carbon capture.
Impact category | Unit | Propane (pre-avoided products) | Propane (total) | Contributor 1 | % Total | Contributor 2 | % Total | Contributor 3 | % Total |
---|---|---|---|---|---|---|---|---|---|
Global warming potential | kg CO2e | 7.4086 | 7.3300 | Electricity Med Voltage US | 84.23% | Natural gas combustion in boiler | 19.30% | Natural gas extraction | 3.04% |
Terrestrial acidification | kg SO2e | 0.0298 | 0.0278 | Electricity Med Voltage US | 54.41% | Natural gas processing | 43.99% | Natural gas combustion in boiler | 1.47% |
Particulate matter formation | kg PM2.5 | 0.0257 | 0.0251 | Electricity Med Voltage US | 84.79% | Natural gas processing | 14.78% | Natural gas combustion in boiler | 0.52% |
Ozone formation terrestrial ecosystem | kg NOXe | 0.0080 | 0.0065 | Electricity Med Voltage US | 84.41% | Natural gas combustion in boiler | 14.86% | Natural gas, high pressure (US) production | 1.07% |
Impact category | Value (pre-avoided products) | Value | Unit | Contributor 1 | % Total | Contributor 2 | % Total | Contributor 3 | % Total |
---|---|---|---|---|---|---|---|---|---|
Marine ecotoxicity | 0.2942 | 0.2910 | kg 1,4-DCB | Electricity, medium voltage | 97.90% | Natural gas, at extraction | 1.99% | Natural gas, high pressure (US), production | 0.05% |
Freshwater ecotoxicity | 0.2171 | 0.2150 | kg 1,4-DCB | Electricity, medium voltage | 98.09% | Natural gas, at extraction | 2.14% | Crude oil, at production/RNA | 0.05% |
Human carcinogenic toxicity | 0.3620 | 0.3560 | kg 1,4-DCB | Electricity, medium voltage | 99.99% | Natural gas, high pressure (US), production | 0.10% | Water, deionised, from tap water, at user (RoW) | 0.06% |
Human non-carcinogenic toxicity | 5.3844 | 5.3300 | kg 1,4-DCB | Electricity, medium voltage | 97.50% | Natural gas, at extraction | 2.56% | Crude oil, at production/RNA | 0.09% |
Impact category | Unit | Propene (pre-avoided products) | Propene (total) | Contributor 1 | % Total | Contributor 2 | % Total | Contributor 3 | % Total |
---|---|---|---|---|---|---|---|---|---|
Global warming potential | kg CO2e | 3.4323 | 3.2500 | Electricity Med Voltage US | 78.37% | Natural gas combustion in boiler | 31.47% | Natural gas extraction | 4.95% |
Terrestrial acidification | kg SO2e | 0.0175 | 0.0130 | Natural gas, processed, at plant | 56.61% | Electricity Med Voltage US | 39.90% | Natural gas combustion in boiler | 1.90% |
Particulate matter formation | kg PM2.5 | 0.0125 | 0.0110 | Electricity Med Voltage US | 75.26% | Natural gas, processed, at plant | 23.05% | Natural gas combustion in industrial boiler | 0.81% |
Ozone formation terrestrial ecosystem | kg NOXe | 0.0040 | 0.0006 | Electricity Med Voltage US | 72.08% | Natural gas combustion in boiler | 22.32% | Natural gas, high pressure, production | 2.01% |
Impact category | Value (pre-avoided products) | Value | Unit | Contributor 1 | % Total | Contributor 2 | % Total | Contributor 3 | % Total |
---|---|---|---|---|---|---|---|---|---|
Marine ecotoxicity | 0.1291 | 0.1220 | kg 1,4-DCB | Electricity, medium voltage | 96.06% | Natural gas, at extraction site | 3.42% | Natural gas, high pressure, production | 0.11% |
Freshwater ecotoxicity | 0.0954 | 0.0906 | kg 1,4-DCB | Electricity, medium voltage | 96.04% | Natural gas, at extraction | 3.68% | Natural gas, high pressure, production | 0.10% |
Human carcinogenic toxicity | 0.1560 | 0.1420 | kg 1,4-DCB | Electricity, medium voltage | 99.98% | Natural gas, high pressure, production | 0.23% | Water, deionised, from tap water, at user RoW | 0.05% |
Human non-carcinogenic toxicity | 2.3721 | 2.2400 | kg 1,4-DCB | Electricity, medium voltage | 95.27% | Natural gas, at extraction | 4.38% | Crude oil, at production/RNA | 0.16% |
Normalisation of the impact factors was used to identify those that are of the most concern based on a comparison with a baseline. Table 3 shows the top four impact categories in descending order of importance. These impact categories represented 96.73% of the overall impact. The impact category of highest concern is marine ecotoxicity, where its main contributor (>97%) is the use of grid electricity.
Total water consumption is 0.0153 m3 kg−1 propene where the highest contributor (76%) was the use of grid electricity. However, the value is 48% the intensity for propane. When electrolysis is removed from the model, i.e. the electricity requirement of 66.741 kW h kg−1 hydrogen and 9 kg H2O kg−1 hydrogen, the water use intensity of propane reduces from 0.0321 to 0.00631 m3 kg−1 propane. Therefore, electrolysis contributed 80% to the total water use intensity. The propane reactor requires a H2/CO ratio of 2.7 compared to 2 for methanol synthesis, hence requiring significantly more hydrogen and explaining the difference in water use intensity.
The top four factors after normalisation represented 97.31% of the impact (Table 5). As with propane, the highest impact factor was marine ecotoxicity due to grid electricity use. As previously stated, U.S. medium voltage electricity was based on 2015 data. U.S. electricity generation data91 was found for 2015 which shows coal generation contributed 33% and natural gas 33%. The main components of coal are carbon, sulphur, oxygen and hydrogen with traces of heavy metals. During combustion, their respective oxides and particulate matter are formed, which can explain the high impact factor results.92 CO2 is formed during coal combustion; however, 60% of non-coal combustion emissions come from flue gas clean up, specifically limestone use.93 Overall, CO2 from coal combustion in conventional power stations (which are more prevalent than their modern integrated gasification combined cycle (IGCC) alternative) produce 50 to 60% more CO2 than natural gas in a new, efficient power plant.94 This explains the significant contribution of grid electricity use to global warming.
While stack emissions for natural gas may be cleaner than coal, the fugitive emissions, venting and flaring of natural gas during production pose a significant environmental burden. Thus, 33% electricity generation from natural gas significantly contributes to the environmental impact.
One limitation of the U.S. electricity generation data is that it illustrates the rapidly changing energy landscape, particularly in terms of coal contribution dropping to 19% in 2020. Therefore, as electrical duty is so influential in the model to emissions, future work should update databases in SimaPro for current grid energy mix. Based on this finding, the results of this model are likely to be overestimated across multiple impact factors compared to the present day. The effect of changing grid mix is discussed further in section 4.2.3.6.
The identification of marine and freshwater ecotoxicity as the top two factors in Table 5 and the significance of grid electricity can be explained through the presence of high quantities of arsenic, copper, selenium, lead and mercury in coal ash. These toxic components contaminate surface and groundwater, resulting in bioaccumulation.92
Impact category | Unit | Propane (total) | –Electrolysis | Propene (total) | –Electrolysis |
---|---|---|---|---|---|
Global warming potential | kg CO2e | 7.33 | 23.06% | 3.25 | 31.69% |
Terrestrial acidification | kg SO2e | 0.0278 | 47.48% | 0.013 | 55.54% |
Particulate matter formation | kg PM2.5 | 0.0251 | 21.51% | 0.011 | 29.55% |
Ozone formation terrestrial ecosystem | kg NOXe | 0.00651 | 6.05% | 0.000591 | −307.95% |
GWP and particulate matter formation reduced by 77% and 78% respectively and ozone formation by 94% for propane. The value for ozone formation terrestrial ecosystem is also reduced to reach a negative value, due to the inclusion of avoided products. As discussed with the differences in water use intensity, the same explanation applies to why the overall carbon intensity among other impact categories was higher for propane than propene. The propane pathway had a 35% higher H2:
CO ratio, in turn increasing electrical demand. Furthermore, as the propane reactor was less selective, higher quantities of by-products were produced. Therefore, resulting in higher overall duties from purification such as the front end de-methaniser fractionation (DE-MTH, DE-ETH, DE-PRO) compared to propene (DE-PRP).
The process was repeated along the propane production pathway in Table 7 for the units with the highest duties. This included CMPPRO, the second highest duty (electrical) after electrolysis which contributed 0.58 kg CO2e per kg propane. COOLPRO2 was the highest cooling duty (ethylene refrigeration cycle) and contributed 0.13 kgCO2e per kg propane.
Impact category | Unit | Propane (total) | –CMPPRO | –DRM | –COOLPRO2 |
---|---|---|---|---|---|
Global warming potential | kg CO2e | 7.33 | 92.09% | 94.13% | 98.23% |
Terrestrial acidification | kg SO2e | 0.0278 | 94.60% | 87.77% | 96.40% |
Particulate matter formation | kg PM2.5 | 0.0251 | 91.63% | 96.02% | 98.80% |
Ozone formation terrestrial ecosystem | kg NOXe | 0.00651 | 90.32% | 94.78% | 98.46% |
Table 8 shows the impact of removing certain units along the propene production pathway. Notably, removal of CMPMTH2 had the greatest impact, closely followed by the dry methane reformer (DRM), which consumed large quantities of natural gas due to endothermicity.
Impact category | Unit | Propene (total) | –CMPMTH2 | –DRM | –COOLPRP1 |
---|---|---|---|---|---|
Global warming potential | kg CO2e | 3.25 | 88.92% | 89.23% | 96.00% |
Terrestrial acidification | kg SO2e | 0.013 | 92.31% | 78.46% | 91.54% |
Particulate matter formation | kg PM2.5 | 0.011 | 88.55% | 92.73% | 97.27% |
Ozone formation terrestrial ecosystem | kg NOXe | 0.000591 | 33.67% | 52.79% | 82.06% |
Since the electrical duty influences the results for some environmental impact categories significantly, alternative electricity grid mix would also influence the results greatly. For instance, if all electricity was provided by wind turbines or photovoltaic panels, the overall environmental impact would be much lower. Section 4.2.3.6 compares the results obtained with those when using an alternative energy mix.
Fig. 8 shows that, overall, the novel CCU method for propene production offered a significant saving across all impact factors. The greatest savings were compared to SC, which represents 47% of global production. Impact on terrestrial acidification for the CCU route amounted to just 0.5% when compared to FCC. FCC is the single biggest source of atmospheric pollution in a refinery and production of sulphur oxides and particulates would have accounted for a large proportion of this difference and for particulate matter formation (0.65% of FCC). Therefore, despite the high electrical intensity of electrolysis and the use of an electricity grid mix with high environmental impact, the novel method still offered substantial emission savings.
The overall water use intensity (Fig. 9) for the CCU route required 32.1 kg per kg propane, whereas for natural gas and crude oil routes was 14.5 and 2.75 kg per kg propane respectively. Electrolysis accounted for 80% of the water consumption of the propane production process. Similarly, the use of fracking for the conventional natural gas route contributed 83% to the intensity figure. While the use of water for electrolysis will not result in the same consequences as fracking, future work should investigate sustainable sources of water, particularly with regard to plant design and location due to the vast quantities required.
Water use for SC and FCC conventional routes amounted to 10 and 4.1 kg per kg propene respectively, compared to 15.3 kg per kg propene for the CCU route. Therefore, while water intensity of the CCU route is 50% higher compared to the dominant technology (SC), if water can be sustainably sourced, the overall environmental impact would greatly improve on conventional methods.
Impact category | Unit | Propane (total) | Propane (ecoinvent 3) | % Change | Propene (total) | Propene (ecoinvent 3) | % Change |
---|---|---|---|---|---|---|---|
Global warming potential | kg CO2e | 7.33 | 0.112 | 98.47% | 3.25 | 1.56 | 52.00% |
Terrestrial acidification | kg SO2e | 0.0278 | 0.00274 | 90.14% | 0.013 | 0.00317 | 75.62% |
Particulate matter formation | kg PM2.5 | 0.0251 | 0.000868 | 96.54% | 0.011 | 0.00104 | 90.55% |
Ozone formation terrestrial ecosystem | kg NOXe | 0.00651 | 0.00212 | 67.43% | 0.000591 | 0.0028 | −373.77% |
Results revealed that despite changing the location and associated databases, grid electricity use accounted for on average 80% of propane and 70% of propene contributions to the impact factors. Fig. 10 shows that the UK produced reductions for both products in GWP (18%) and particulate matter formation (60%). The UK grid compared to the U.S. in 2015 had a lower natural gas (31%) and coal use (30%).95 Furthermore, renewable penetration in was around 12% for the U.S. and 20% for the UK. Nevertheless, impacts for terrestrial acidification and ozone formation were higher. In 2014, coal imports made up 78% total supply in the UK, where 85% of the total imports was from Russia.96 76% of Russia's coal export came from the Kuzbass region, where the average life expectancy is 3 to 4 times lower than the Russian average and 93.8% of drinking water sources fail to meet sanitary chemical and microbiologic standards.97 Therefore, poor environmental standards for the extraction of coal and transportation to the UK are attributed to the higher terrestrial acidification and ozone formation increase compared to the U.S.
The plant modelled produces propane and propene from CO2 captured from a medium-sized FCC unit. The FCC unit is the most emission intensive in a refinery and being a stationary source can be retrofitted with a post-combustion carbon capture technology such as the piperazine system. Case studies have proven its technical feasibility.99 The model was based on a unit producing 0.5 million tons of CO2 based on a feed rate of 60000 barrels per day.
The economic analysis package within Aspen Plus was used to determine the capital cost, operation and maintenance cost (O&M) associated with the model at the desired flowrates and operating conditions. To improve the accuracy of the operational cost, literature was used to determine the price of certain utilities, feedstock and product pricing within the market. Table 10 details the assumptions used in the model. As Aspen calculates stream price in total $ per kg, the mass fractions of the stream were used to get an overall price using pure prices. For example, PROPENE has a total mass flow of around 3 kg s−1 where pure propene is 2.7 kg s−1. Therefore, the overall price of the stream was circa $3.3 per kg.
Component | Price | Unit | Comment | Ref. |
---|---|---|---|---|
Piperazine | 9 | $ per kg | Market price | 100 |
Natural gas | 0.1046 | $ per kg | U.S December 2020 | 101 |
Propane | 1.076 | $ per kg | U.S. residential March 2021 | 102 |
Propene | 1.157 | $ per kg | U.S. Polymer grade | 103 |
Ethane | 0.1471 | $ per kg | U.S. December 2020 | 101 |
Butane | 0.28621 | $ per kg | U.S. October 2020 | 101 |
Pentane and above (gasoline) | 0.3721 | $ per kg | U.S. October 2020 | 101 |
Electricity | 0.0635 | $ per kW h | U.S. Industrial January 2021 | 102 |
Metric | Price | Unit |
---|---|---|
Investment cost: propane/propene plant | 40.3 | Million USD |
Investment cost: carbon capture facility | 17.3 | Million USD |
Total investment cost | 57.6 | Million USD |
Total sales revenue | 328.7 | Million USD p.a. |
Operating cost | 370.2 | Million USD p.a. |
Raw material cost | 229.7 | Million USD p.a. |
Utility cost | 82.2 | Million USD p.a. |
Net present value (NPV) | −695.6 | Million USD |
Gross profit (GP) | 13.4 | Million USD |
Payback period | −0.77 | Years |
Propane cost | 1.14 | $ per kg |
Propene cost | 1.39 | $ per kg |
In comparison to conventional methods, the cost of propene via SC varied between $0.5 per kg45 and $1.48 per kg.47 However, the latter figure is from 2012 when crude oil was priced $110 per bbl. Current prices are around $60 per bbl, which if used in the model, would reduce the cost considerably. Therefore, the price of this technology is competitive ($1.39 per kg). However, as the price of crude oil varies significantly and can contribute up to 86% of production cost,58 comparisons must be made cautiously.
Alternatively, for propene production via MTP, cost varied in literature from $0.7 per kg (ref. 48) to $0.85 per kg. However, the use of natural gas to form syngas and subsequent conversion into olefins by FT had a production cost of $2.4 per kg.50 Furthermore, a dry reforming process using a CO2-rich natural gas from China resulted in a propene cost of $1 per kg.51
Propane cost in literature is considerably lower than the model output ($1.14 per kg). However, the majority of production requires only NGL recovery with no reactors or extensive operations. Therefore, the price will largely depend on the cost of extraction. Operational cost from NGL recovery excluding raw material cost varied between $0.04 per t to $0.26 per t.54 Propane production cost from FT synthesis varied in literature from $0.09 per kg (ref. 53) to $0.15 per kg.52
Another limitation of such a comparison is that a production cost for propene or propane that incorporated the cost of capture of CO2 was rare. However, carbon capture operational cost only contributed 4 and 8% of cost for propane and propene respectively.
The following subsections analyse each of the metrics in Table 11 more closely and show sensitivity analyses.
To calculate, Fortran code was used within Aspen to multiply the quantity of utilities used by the cost, calculate feedstock stream cost and product sales revenue based on mass flows (eqn (7)).
GP = Sales revenue − Costs − OPMT | (7) |
“OPMT” is the O&M cost, “Costs” is the cost of raw materials and utilities over the year where 330 operational days have been assumed (90% uptime).
As seen in Table 12, the value of GP is $13.4 M. Therefore, the total sales revenue generated exceeded the sum of the price of utilities, raw materials and the operation and maintenance cost. The values used in eqn (7) are summarised in Table 12.
Component | Value | Unit |
---|---|---|
Sales revenue | 41![]() |
$ per hour |
Raw cost | 29![]() |
$ per hour |
Utility cost | 10![]() |
$ per hour |
OPMT | 3.5 | Million USD |
GP | 13.4 | Million USD |
![]() | (8) |
The value of operating cost used in the NPV formula varies slightly as it considers items such as operating labour cost, plant overhead cost, general and admin cost which were all assumed to be a factor of raw material and utility cost.84 Such assumptions are found in section 27 of the ESI† which were included in the Fortran code.
The value of NPV is −$695.6 M (Table 11), indicating a project that will return a net loss. While the GP was positive, this metric did not consider the time value of money, nor did it include the total capital investment or other contributions to operating cost discussed above. Therefore, based on the current assumptions the chemical plant is not be a profitable venture.
![]() | (9) |
The value of PBP is −0.77 (Table 11), i.e. the project will never payback the initial capital investment. This is because total sales revenue is $328.7 M per year, but total operating cost is $370.2 M per year.
Fig. 12 shows the raw material and utility cost of different units in the plant. Hydrogen generates the highest cost due to significant electricity consumption, in turn increasing operational cost. The following section performs a sensitivity analysis to assess the difference in electricity required by different hydrogen production methods. Furthermore, the cost of natural gas (CH4FEED), while small in comparison to hydrogen feedstock cost, is high in comparison to other duties, thus sensitivity analysis is also performed for this.
One of the highest utilities is CMPPRO, with a duty of 30.14 MW. However, one limitation is that it does not consider the utility or environmental aspect of cooling for this compressor, which in this case would be interstage cooling. Such cooling would involve the use of refrigerants or cooling water, hence their associated duties and material usage. Evaporative/fugitive losses and cost are not examined within this model; hence make-up costs are omitted. In addition, there is a cost limitation as to pressurise the syngas from atmospheric to 21 bar would require a staged approach with multiple compressors which have not been costed individually but as one unit.
Similarly, CMPMTH2 is the second highest utility, which compresses syngas from 1.3 bar to 49 bar to be fed into the methanol reactor. Following the methanol reactor, the pressure of the system is brought to ambient conditions, so a turbine (CMPMTH3) is used to recover energy from the process that initially compresses the syngas in CMPMTH1. CMPMTH3 recovers 1.7 MW which is just over 7% of the duty for CMPMTH2, therefore, offering a saving of 0.46 kW h per second or $0.03 per s. Thus, further optimisation of the process can help to further reduce operational cost through energy and heat integration.
Propane was found to generate more revenue than propene. This is interesting as the price of propane and propene are similar ($1.076 per kg and $1.157 per kg, respectively). This is explained by the amount of product produced. While the MTP reactor had a higher selectivity for propene, there was a significant production of water that accounted for a significant loss of mass. On the contrary, the production of propane was slightly less selective (51.5% compared to 71.37%), but the impurities produced were CO and hydrogen, which were recycled.
Fig. 13 confirms literature surrounding carbon capture in that the greatest duty is that of the reboiler in the stripper (STRIPREB) due to overcoming the regeneration energy.
Technology | Electricity consumption (kW h kg−1) | GP (Million USD) | NPV (Million USD) | Payback (Years) |
---|---|---|---|---|
AB | 66.7 | 13.4 | −695.6 | −0.77 |
PEM | 55 | 51.4 | −0.78 | −1.98 |
SOEC | 41.75 | 94.6 | 786 | 2.62 |
The use of SOEC resulted is the sole technology to provide a positive NPV (786 million USD) and a payback within 3 years. Furthermore, research surrounding SOEC has uncovered the potential of CCU to produce syngas. In addition, cost of goods sold for both products dropped by 22% (propene) and 28% (propane), further increasing price competitiveness. Kamlungsua et al.105 stated that with operation of SOEC at high temperatures, H2O and CO2 can undergo electrochemical conversion into syngas. As such, this would pose a significant recommendation for further research as not only could it prove more efficient than a dry methane reformer, but it would also reduce the need for a source of methane and therefore all the associated emission impacts with extraction, processing and transportation.
Price ($ per kg) | GP (Million USD) | NPV (Million USD) | Payback (Years) |
---|---|---|---|
0.025 | 104 | 961 | 1.72 |
0.05 | 101 | 906 | 1.93 |
0.075 | 98 | 851 | 2.19 |
0.1046 | 94 | 786 | 2.62 |
0.125 | 92 | 741 | 3.02 |
0.15 | 89 | 686 | 3.72 |
Excel Solver was used to determine that a breakeven price on NPV requires the price of a carbon incentive to be $99.87 per t (section 28 of the ESI†). Therefore, values must exceed $100 per t for the project to be economically feasible using an AB electrolysis technology. However, when SOEC is combined with a carbon tax of $25 per t similar to the section 45Q, GP increases by almost $11.5 M. Furthermore, NPV increases by 20% to $960 M and payback reduces from 2.62 years to 1.72 years. At a price of $25 per t, this represents less than 5% of sales revenue, however, the impact on overall plant economics is clear. Therefore, acquiring such an incentive would significantly improve investment prospects.
A further limitation of this model is that the price of the electrolyser AB or SOEC was not included in the model. However, capital cost can be expected of between $1100–1400 per kW for AB and greater than $2200 per kW for SOEC.107
The propane reactor utilises a bi-functional catalyst with Cu–ZnO/Pd core and beta-zeolite shell. While the Cu–ZnO catalyst is common for methanol synthesis applications, the modification to be palladium supported and contained within a zeolite shell presents a difficulty in commercial availability. Such a catalyst configuration is not widely used in industry as it remains a novel process/niche application. At an industrial scale, the catalyst requires frequent replacement due to irreversible deactivation such as sintering, which could result in production downtime due to lack of supply-chain security.
Furthermore, while the MTP process has been proven commercially, the syngas-to-propane reactor only exists in experimental studies. The technology is not mature and would be classified at a technology readiness level (TRL) of 3–4. Furthermore, although slurry reactors were suggested in literature, a pilot-scale plant for either configuration has not been developed. Such a setup would identify critical issues and lessons learned that could be eliminated when scaling up to an industrial plant, thus presenting a risk.
A key feedstock for the chemical plant is natural gas. The security of supply of natural gas in the U.S. is low risk due to the highly integrated and extensive pipeline network present and the abundance of domestic supply from shale resources. However, for a location such as the UK, it would present a reduced resilience in terms of energy security. As the UK moves towards a greater dependence on gas imports, such a plant would be exposed to market price volatility which results in a substantial change in plant economics.
• Further optimisation of the process through heat and energy integration, such as organic Rankine cycle.
• Sourcing of a sustainable water supply for electrolysis.
• Use of electric motors instead of mechanical work supplied by gas turbine for compressor duty.
• Investigate the amount of refrigerant leakage from the refrigeration cycles and cooling required and its associated environmental and operational cost burden.
• Optimisation of DE-PRP such as of reflux ratio or further separation to achieve polymer-grade propene purity of >99.5%. Such a task would illustrate the economic trade-off between increased separation costs against the increased revenue from polymer-grade propene.
• Modelling the propane reactor as a slurry reactor.
• SOEC for hydrogen production and a source of sustainable heat for the process.
• Carbon dioxide utilisation through production of syngas using SOEC instead of a dry methane reformer.
• Direct synthesis of propene from syngas, similar to the propane reactor by utilising an alternative catalyst such as SAPO-34.
• Direct conversion of syngas to DME to avoid the methanol production step in the production of propene.
• To avoid solvent degradation in carbon capture, research should investigate if flue gas from an FCC stack has SO2 and NOx below 100 ppmv. If not, suitable pre-treatment technology should be added to the front end of the process.
• Undertake an LCA and TEA looking at the supply of renewable electricity to the plant, e.g. wind energy.
• Update databases for a present-day energy mix.
• Assumptions, accuracy and clarification on the boundaries of the system for propane and propene production within databases in SimaPro.
Grid electricity was found to be very influential in the model as it was one of the highest duties and that the data in SimaPro are based on a U.S. grid where coal and natural gas combustion contributed two-thirds to total electricity generation. Other units that contributed majorly to the impact factors included the compression of syngas along the propene route (CMPMTH2) and the dry methane reformer (DRM) due to its endothermicity.
Production of propane via CCU with AB and U.S. electricity mix scenario resulted in a saving of 2.8 kg CO2 per kg propane compared to natural gas fractionation. With SOEC and a lower carbon intensity grid mix, the saving would be even higher. Similarly, water-use intensity compared to natural gas fractionation was 17.6 kg per kg propane higher. However, as electrolysis accounted for 80% of the water consumption, use of SOEC and cleaner electricity generation would further reduce this difference.
For propene, the novel CCU method also showed promising results with significant savings across all impact factors. The greatest were with respect to steam cracking of naphtha, which represents 47% of global production. The only drawback was a 50% higher water use intensity compared to steam cracking. However, if water were sustainably sourced, the environmental credentials would be much greater.
The choice of hydrogen technology was the real determinant of both economic and environmental performance across the whole model. The use of SOEC with a 37% lower electrical intensity greatly impacted profits and impact categories positively. Removal of electrolysis from the model reduced GWP by 77% and 68% for propane and propene respectively. Thus, further work should include SOEC technology and look to further optimise its performance. Furthermore, when the model utilised AB technology, it returned a negative NPV and therefore was incapable of paying back the capital investment. However, the use of SOEC produced a positive NPV of $786 M and a payback of 2.62 years. In addition, lower natural gas price and the incorporation of a carbon tax incentive produced significant and positive impacts on plant economics.
The cost of goods sold for propene was competitive with conventional production at $1.39 per kg. However, for propane the cost at $1.14 per kg was significantly higher, owing to the cheap cost of production of NGLs.
The use of the MTP in the model poses little deployment risk as it has been proven at scale. However, as the syngas-to-propane technology is at a low technology readiness level, further work must be done to prove that experimental results are achievable at a greater scale, e.g. in a pilot study.
As the world focuses on decarbonisation pathways to curb anthropogenic carbon emissions and halt the warming of the atmosphere, new, more sustainable production methods of fuels and materials are at centre stage. However, their successful implementation is based on two main criteria: economically feasible production and an environmentally superior performance compared to conventional production. This article has achieved this for propane and propene, two critical and demand-evolving products, by proving that CCU methods of production can be both environmentally and economically superior. Furthermore, not only has the model provided another example of feasibility with respect to carbon capture, but also emphasised the significant opportunity that syngas production offers in the utilisation of CO2 and extensive possibilities of transformation into valuable materials. Particularly, for hard-to-abate sectors or where electrification of heat for a process is not feasible.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2gc04721g |
This journal is © The Royal Society of Chemistry 2023 |