Technical and environmental assessment of renewable dimethyl ether production from CO 2 in different european contexts

This work highlights the differences between two dimethyl ether (DME) production pathways from CO 2: (S1) intensified direct conversion of CO 2 and H 2 via sorption-enhanced DME synthesis (SEDMES) in a single reactor, (S2) indirect conversion in two steps via methanol as an intermediate product followed by dehydration to DME; For this purpose, flue gas from a cement production plant was considered as the CO 2 source and membrane separation (MS) was used for carbon capture, while hydrogen was generated via water electrolysis. ProSimPlus® software, was used for process simulation. Starting from 5,864 kg CO 2, scenario S2 produces 2,433 kg/h of DME, which is lower than 2,929.5 kg/h in S1. In addition, S1 shows significantly lower electrical and heating energy consumption per kgDME produced: 17.80 kWh and 3.07 MJ, respectively, compared to 22.59 kWh and 10.14 MJ in S2. The environmental analysis was carried out using a comprehensive Life Cycle Assessment (LCA) approach, with the production of 1 kg of high-purity DME defined as the functional unit and compared with the traditional production of DME (S3) from fossil methanol (MeOH). The results show that the electricity source plays an important role in determining the environmental impact in the context of North-West Europe with a focus on the Netherlands. If the SSP2-NDC trajectory is considered, for the Dutch electricity mix, the GWP100 impacts of both scenarios decrease significantly over year (2020–2050). For instance, in S2, GWP100 could decrease from 3.46 kg CO 2 -eq/kgDME in 2020 to −1.41 kg CO 2 -eq/kgDME by 2050 due to planned electricity defossilization in the Netherlands. Same for S1, GWP100 value dropped from 2.08 to −1.63 kg CO 2 -eq/kgDME, indicating climate positive outcomes. However, S3 maintains a stable GWP100 impact of around 1.2 kg CO 2 -eq/kg DME over years, due to its fossil fuel-based nature. This study demonstrated that S1 and S2 offer promising alternatives to the conventional fossil DME production pathway highlighting the importance of electricity sources.

As the world faces increasing pressure to reduce its dependence on fossil fuels and to decrease the carbon emissions, the transition towards renewable and low-carbon fuels and energy carriers has become a central priority. In this context, alternative fuels that can be produced from captured CO 2 have gained increasing attention (Singh et al., 2022).

Among the emerging fuels, dimethyl ether (DME) is a particularly promising candidate. DME is a non-toxic, non-carcinogenic compound with combustion characteristics similar to liquefied petroleum gas (Chai et al., 2024). DME burns cleanly, producing virtually no particulate matter or NO x emissions, and can be used in diesel engines, gas turbines, or as a feedstock in chemical synthesis. Its compatibility with existing liquefied petroleum gas infrastructure further enhances its attractiveness for large-scale deployment. Furthermore, it can be used as refrigerant in various cooling systems, as propellant in aerosol products, as specialty solvent in several chemical syntheses, and also serve as feedstock for the production of olefins via DME-to-olefins (DTO) processes (Bauer and Kruse, 2019; Cordero-Lanzac et al., 2024). Traditionally, DME is produced from fossil-derived methanol via synthesis gas (syngas), but recent advances have enabled its synthesis from renewable sources, and particularly, through carbon capture and utilization (CCU) pathways (Saravanan et al., 2017).

DME can be produced in different ways. One of the explored routes is the syngas conversion, with syngas from natural gas reforming or biomass gasification. For instance, Vakili et al. (2011) proposed a heat exchanger reactor for the single-step synthesis of DME from syngas and the co-production of benzene, via cyclohexane dehydrogenation, which supplies the necessary hydrogen. Furthermore, Farniaei et al. (2013) introduced a thermally double-coupling concept for the simultaneous production of methanol, DME, and hydrogen in a membrane reactor. However, the conversion of methanol to DME by dehydration reaction seems a promising way to produce dimethyl ether (Bakhtyari et al., 2016). Therefore, an alternative production of DME could be envisaged by first converting CO 2 and H 2 into methanol (Meunier et al., 2020), followed by a second reactor where methanol dehydration occurs. This could replace the use of fossil syngas from reforming. Another innovative approach to intensify the direct synthesis of DME from CO 2 and hydrogen (H 2) in a single reactor is the sorption enhanced DME Synthesis (SEDMES) process (Peinado et al., 2024). This method combines physically mixed methanol synthesis catalysts (copper-zinc-alumina) with methanol dehydration catalysts (alumina) in one step using an in-situ H 2 O adsorbent (zeolite) to shift reaction equilibria and enhance conversion. Both routes offer significant potential for reducing the process carbon footprint, especially when powered by renewable electricity and green hydrogen.

Several studies have paid attention to the environmental impact of DME production. For instance, Lerner et al. (2018) investigated the production of DME from methanol derived from natural gas. In their study, natural gas is first converted into syngas through reforming to be then transformed into methanol and subsequently into DME. Two methods for syngas production were considered: steam methane reforming (SMR) and autothermal reforming (ATR). The study provided results for total GHG emissions from the overall process, which includes multiple steps: syngas production, electricity, oxygen production and DME synthesis. When using the SMR route, the process emits 51.5 kg CO 2 -eq/GJ of fuel however, by ATR route this impact decreases to 35.2 kgCO 2 eq/GJ. This reduction is mainly due to the lower GHG emissions associated with oxygen production used in ATR, comparing to the high emissions from natural gas burning, used in SMR case.

Another study, by Giuliano et al. (2021) compared the environmental impact of DME synthesis from syngas derived from digestate gasification with carbon capture and storage (CCS), against two alternative routes: DME synthesis without CCS, DME via indirect liquefaction of biomass. Their impacts were evaluated considering use as fuel and compared with fossil-based diesel with end-of-life emissions. The results show that the DME synthesis by digestate valorization with CCS outperforms the other routes with negative GWP of −113 kgCO 2 eq/GJ. The negative value is due to the avoided landfill emissions and the permanent storage of CO 2. The production of DME from biomass by indirect liquefaction shows positive GWP of 15 kgCO 2 eq/GJ, as it includes end-use combustion emissions. However, this route is still competitive to digestate-DME without CCS generating 24 kgCO 2 eq/GJ. The worst case was the fossil-based diesel route, with GWP of around 99 kgCO 2 eq/GJ, including the combustion emissions.

Fedeli et al. (2023) evaluated three biogas utilization processes: combined heat and power (CHP), dimethyl ether and methanol production. In this study, DME is produced from biogas through an indirect route involving methanol as an intermediate. The process begins with steam reforming of biogas to generate syngas, then syngas is compressed and sent to a methanol synthesis reactor. From LCA perspective, CHP has the lowest GWP impact by achieving negative value (−0.557 kg CO 2 -eq/m 3 of biogas). This is primarily due to the absence of energy intensive steps present in the chemical conversion routes (DME and methanol). However, DME production shows environmental benefits, with negative values in specific impact categories, such as eutrophication marine (−4.97E-04 kg CO 2 -eq/m 3 of biogas), human toxicity (cancer) (−3.74E-10 kg CO 2 -eq/m 3 of biogas) and ozone depletion (−5.98E-08 kg CO 2 -eq/m 3 of biogas). These results are attributed to its lower environmental burdens and high methanol to DME conversion efficiency (88.6%), highlighting its technical and environmental viability as an alternative biogas utilization route.

Styring et al. (2022) have investigated the carbon footprint of producing DME via the sorption-enhanced DME synthesis (SEDMES) process and using it as a transport fuel assimilated similar to diesel. A comparison was reported with the production and use of fossil-based diesel by setting the consumption of 1 MJ of fuel as functional unit. The well-to-wheel carbon footprint of using DME instead of diesel was calculated with the IPCC 2013 GWP100a method where results demonstrated to be 77% lower when using offshore wind power for H 2 synthesis and DME production. A 90% gain was observed if renewable energy was used for CO 2 capture and waste heat was used for the DME production and purification steps. Data from the CO 2 capture was taken from literature while data for the SEDMES process was taken from the model used by Skorikova et al. in a previous TEA study (Skorikova et al., 2020). Furthermore, they have concluded that carbon-footprint is linearly dependent on the source of electricity, since H 2 production by PEM accounts for more than 90% of the carbon footprint in the Netherlands.

Whereas the environmental benefit of renewable DME compared to fossil fuel has been quantified, an integral assessment of different routes to produce renewable DME from the same source seems lacking. Specifically, the environmental benefit of using intensified, sorption-enhanced DME synthesis processes has not yet been compared to alternative process routes. Therefore, the aim of this study is to investigate CCU technologies to produce DME, from a technical and environmental point of view. This study proposes a technical and environmental analysis by using process modeling and simulation for material and energy flows calculation and LCA for environmental impact calculation, comparing the production routes under the same conditions and from the same substrate. A prospective LCA was also evaluated using Premise tool (Sacchi et al., 2022) to assess the impacts in future economy trajectories.

In this context, two production pathways for DME from CO 2 are studied: Intensified direct synthesis via SEDMES – where CO 2 and H 2 are converted directly into DME in a single step while water is simultaneously adsorbed to shift the equilibrium toward DME.

Indirect synthesis – where CO 2 and H 2 are first converted into methanol, which is subsequently dehydrated into DME in a second processing step.

These routes are compared in terms of LCA with the conventional fossil-based production, where DME is produced from fossil fuels through well-established industrial processes. The following sections report in detail the methodology used, the collected data and assumptions applied, and the results obtained from the comparative study.

This section describes the two studied scenarios and the LCA methodology employed to assess the environmental impacts of the DME production pathways.

Dimethyl ether production scenarios

Sorption-enhanced direct DME synthesis (SEDMES) (S1)

The SEDMES represents an innovative process-intensification approach aimed at enhancing the direct synthesis of DME from captured CO 2 and H 2. This scenario is particularly promising as it combines, in a single-step, CO 2 hydrogenation and sorption-enhanced reaction technologies to overcome thermodynamic limitations typically associated with conventional methanol/DME synthesis processes (Peinado et al., 2024).

During the sorption-enhanced reaction phase, a mixture of CO 2 and H 2 is introduced into the reactor, where the catalysts and the solid adsorbent act together to drive a series of reactions. Initially, CO 2 and H 2 react to form methanol, which is then immediately dehydrated to produce DME. Simultaneously, water, a byproduct of both reactions, is continuously adsorbed by the solid adsorbent (LTA zeolite). The removal of water avoids that the oxygen surplus ends up in CO 2, which is essential as it shifts the chemical equilibrium to improve the thermodynamic favorability of the reaction, and to promote the formation of more methanol and DME and thus enhance the overall conversion of CO 2. The process operates in cycles, alternating between reactive adsorption and adsorbent regeneration steps working as a Pressure Swing Adsorption (PSA) process. The synthesis of DME from CO 2 follows several reactions:

Hydrogenation of carbon dioxide:

Reverse water gas shift reaction (RWGS):

Hydrogenation of carbon monoxide:

Methanol dehydration:

For the direct synthesis processes, reactions Equations 1–4 occur in a single reactor where there is a high selective direct conversion of CO 2 to DME.

Process description

The sorption enhanced dimethyl ether synthesis, as depicted in Figure 1, consist into an electrolyzer to produce H 2, the SEDMES process and a distillation train for product purification.

Here, it is considered that CO 2 is obtained from flue gas exiting a cement production plant at 160 °C and 1.013 bar. This flue gas is first treated using membrane separation technology to purify it and produce a concentrated CO 2 stream (see Table 1) as presented in (Scholes et al., 2014). Earlier, it has been shown that membrane capture processes outperform chemical absorption, cryogenic separation, and high-pressure water scrubbing (Kanso et al., 2024). The purified CO 2 exits the separation process at 25 °C and 1.9 bar, and is compressed to 30 bar before being mixed with H 2 to enter the SEDMES process.

The necessary hydrogen is produced by water electrolysis as indicated in the electrolyzer section of Figure 1. The model implemented in a previous study (Sánchez et al., 2020) was used for the electrolyzer. The alkaline electrolysis process typically provides a H 2 stream at 10 °C and 1 bar. The H 2 stream is further compressed to 30 bar to be mixed with CO 2 before the SEDMES process. Compression energy is accounted as electricity needed in the SEDMES process.

The SEDMES process is designed around a multi-fixed-bed reactor system that enables the production of DME through a combination of reaction and in-situ adsorption steps (Reactor in Figure 1). The largest SEDMES pilot plant, operated by TNO in Netherlands, is shown in Supplementary Figure S1. This innovative plant consists of three identical 150 L reactors. The three-column design allows the continuous production of 3 kg/h of DME in the pilot plant using CO 2 and H 2 as feedstock by alternating the different steps of the cycle. Each cycle is composed by 4 steps as illustrated in Supplementary Figure S2. Step 1 – adsorption (ADS) - the feed is converted into DME, while the produced steam is removed by a zeolite adsorbent. During step 2 - the blowdown phase (BD) - countercurrent depressurization is applied to the system to prepare the system for pressure swing regeneration. Step 3 – purge (PURGE) - involves a countercurrent flow of dried hydrogen to desorb the water. Finally, in step 4 – repressurization (REP) - the column is pressurized concurrently with the feed stream. The duration of the regeneration phase significantly impacts the process performance, as a longer purge phase ensures drier operating conditions. To allow the purge phase to exceed the duration of the adsorption phase, at least three columns are necessary (Skorikova et al., 2020). The three reactors alternate continuously between these modes as described in Skorikova et al. (2020).

By integrating reaction and separation in a single step, and by using dynamic switching between operation and regeneration, the system achieves high CO 2 and H 2 conversion, excellent carbon utilization, and efficient integration with variable renewable energy inputs. Furthermore, the design of the system makes it suitable for decentralized applications, particularly in industrial settings where CO 2 is readily available from flue gases or other process streams (van Kampen et al., 2023).

The produced DME from the SEDMES process is then sent to the distillation train comprising three columns as displayed in Figure 1. The product is cooled and sent to a flash drum, where non-condensable gases, such as unconverted H 2, CO and CO 2 are separated. Depending on the temperature, portions of CO 2 and DME may leave as gas fractions. The flash drum operates at 29 bar. A compressor is required to recycle the non-condensable gases back to the SEDMES reactor. Next, the first distillation column of the distillation train separates CO 2 from DME. This is the most energy-intensive process due to the similar properties of the two components. The separated CO 2 is recycled back to the SEDMES process. Finally, the second distillation column purifies DME (99.5%) by separating it from methanol and water.

Indirect conversion of DME via methanol (S2)

The indirect conversion of DME is composed of two steps: methanol synthesis and methanol dehydration.

Methanol production

Currently, methanol is produced industrially by hydrogenation of carbon monoxide, typically using natural gas as a feedstock (Ban et al., 2020). The significant demand for methanol, particularly in industrial applications, has driven the establishment of large production facilities. Currently, there are over 90 methanol production plants worldwide, with a combined capacity of 110 million metric tons (The Methanol Industry, 2024).

Methanol can be produced through CO 2 and/or CO hydrogenation. The carbon monoxide can be obtained from CO 2 by the reverse water gas shift (RWGS) reaction. Thus, reactions Equations 1–3 will take place inside the reactor to produce methanol. Carbon monoxide and hydrogen are combined in the reaction (Equation 3) to produce methanol, while carbon dioxide is hydrogenated to yield methanol and water, reaction (Equation 1). Both reactions are exothermic and diminish the moles of gas, therefore are favored by high pressure and low temperature conditions. Additionally, there is a potential for the RWGS reaction (reaction Equation 2) to occur on the catalyst surface when carbon dioxide is used as a feedstock.

There are several types of catalysts that can be used in the RWGS reaction, however, the target is to increase the single pass conversion to methanol which can be done by using a catalyst active at low temperature with high surface area. For that, nowadays, CuO/ZnO/Al 2 O 3 is the most used catalyst and is highly adopted in methanol production industries due to its high activity and selectivity for the CO 2 hydrogenation reaction. Using the same catalyst, Fu et al. (2024) studied the kinetics of methanol synthesis at high pressures (35–55 bar) using CO, CO 2, and H 2 as raw materials. The kinetic model and the parameter values are provided in Section 2 and Supplementary Table S2 of the Supplementary Material.

In our work, the same process described in Ban et al. (2020) for methanol production (process configuration displayed in Figure 2) was used to implement the methanol production from CO 2 sourced by a cement production plant. The process model was implemented in ProSimPlus® (a process simulation software) as a continuous process with the kinetics presented in Supplementary Table S2. The thermodynamic model used in the simulation was the Soave-Redlich-Kwong (SRK) model. The design parameters are reported in Supplementary Table S1.

Process description

Similarly to S1, CO 2 was captured with membrane separation technology from flue gas exiting a cement production plant at 160 °C and 1.013 bar following the parameters of Table 1 and using the configuration of Scholes et al. (2014). The purified CO 2 exits the separation process at 25 °C and 1.9 bar, and is used for methanol synthesis after compression to 55 bars in a series of three compressors (CP-101 to CP-103) and cooling (in coolers C-101 to C-103). This multi-stage compression is more efficient than one stage and is typically employed to reduce energy consumption.

H 2 was produced by water electrolysis following the model implemented in a previous study (Sánchez et al., 2020). The H 2 stream coming from the alkaline electrolysis is further compressed in CP-106 from 1 bar to 30 bar and then from 30 bar to 55 bars in CP-105 and mixed in the mixer (M-101) with the CO 2 stream. This mixture is then heated (heater H-101) to reach the operating temperature of 240 °C, initiating the reactions.

Reactions Equations 1–3 take place in the chemical reactor. However, simulating reaction (Equation 3) requires the use of CO, which is produced by RWGS. Therefore, a CO stream was added and is initialized with a non-zero (small amount of carbon monoxide ∼1 mol%) to avoid convergence issues. This amount is added to the mixed CO 2 /H 2 stream before it is combined in mixer M-102 with the recycled stream, which is preheated in heater H-102 and contains unreacted gases (H 2, CO, and CO 2). In this process and even in the life cycle inventory, carbon monoxide was sourced from the Ecoinvent database via the market for heavy fuel oil: the average supply of CO associated with heavy fuel oil-based process. The output from M-102 feeds directly the reactor. The reactor’s wall temperature is maintained at 250 °C, but due to the exothermic nature of the reactions occurring inside, the temperature at the reactor outlet rises to 286 °C, under a pressure of 51 bar. This is due to the slight pressure drop within the reactor, caused by the porous catalyst bed that creates resistance to the fluid flow. Product stream is cooled down in the cooler (C-105) before entering the gas-liquid separator (TK-101), where the liquefied methanol and water are extracted from the bottom, while the uncondensed gases exit at the top of a splitter (S-101). Here, about 1% of the gases are purged to prevent accumulation, and the remaining gases are directed via the heater (H-102) back to the reactor as a recycle stream.

Two valves (BV-102 and BV-103) are used to reduce the pressure of the bottom (liquid) stream from the first flash separator to atmospheric pressure before it enters the second gas/liquid separator (TK-102) to remove any remaining gases. The bottom stream from this separator, a mixture of methanol and water, flows directly into a distillation column (ST-101) to separate the two components. The distillate from the column is methanol with 97.5% purity in vapor phase. This methanol stream is then cooled from 65 °C to 50 °C in cooler (C-106) to remove the remaining CO 2, resulting in a 99% pure methanol stream in liquid phase.

The simulation results were validated with literature data (Ban et al., 2020), as shown in Supplementary Table S4.

Dimethyl ether production

In this step, methanol is dehydrated to form DME through reaction (Equation 4). This reaction is activated at a temperature above 250 °C. However, this reaction is exothermic and theoretically favored at lower temperatures. Conversely, the formation of by-products (such as H 2, CO, etc.) is favored at higher temperatures (Azizi et al., 2014). Kinetics for methanol dehydration are considered from Azizi et al. (2014) and reported in Section 2 of Supplementary Material.

Figure 3 illustrates the process configuration for DME production through methanol dehydration, taken from the literature (Diemer and Luyben, 2010). Design parameters used in this simulation using ProSimPlus are reported in Supplementary Table S3 in SI. In this study, the thermodynamic model employed for the simulation was the “Predictive Soave-Redlich-Kwong (PSRK)” model in ProSimPlus®, using the Redlich-Kwong (RK) equation of state.

Process description

A methanol liquid stream (99% MeOH) containing a small amount of water (1%), as coming from the methanol synthesis process described in the previous section, is pumped to 13.5 atm and cooled to 25 °C (C-101) before being mixed in the mixer (M-101) with the recycle stream from the second distillation column (DC-102), which contains unreacted methanol. This combined stream is heated to 127 °C in the first heat exchanger (HX-101), vaporized to 150 °C (vaporizer), and then further heated in the second heat exchanger (HX-102) to 297 °C to activate the dehydration reaction at 13 atm.

The conversion within the reactor reaches 81.7%, and the output from the reactor is hot (411 °C). This hot effluent can be directed to a high-pressure steam generator to produce high-pressure steam, which can provide economic benefits in the process. In this work, the hot effluent from the reactor is cooled down before entering the distillation phase. Several heat exchangers were proposed to optimize thermal recovery. Initially, it is cooled to 317 °C in cooler (C-102). Then, as it passes through the heat exchanger (HX-102), the temperature drops to 176 °C. The first heat exchanger (HX-101) further reduces the temperature to 140 °C, and finally, cooler (C-103) cools the stream down to 82 °C.

The output stream from cooler (C-103) is a mixture of DME, H 2 O, and methanol. To purify this mixture, the process configuration includes two distillation columns. In the first distillation column (DC-101), DME is completely separated, achieving 99% mol purity at the top of the column. The bottom stream, containing a mixture of methanol and water (H 2 O 69%; methanol 29%), is fed into the second distillation column (DC-102). Here, the two components are separated, yielding a pure water stream (99% mol) at the bottom and a pure methanol stream at the top, which is recycled to enhance the overall process efficiency.

The obtained simulation results were validated according to the research conducted by Diemer and Luyben (2010), as shown in Supplementary Table S4.

Life cycle assessment

LCA was performed following the standards series 14040–14046 (ISO, 2024) for the two DME production pathways described above and for the fossil-based conventional production route. The conventional LCA provides a picture in the current economic context. In contrast, Prospective LCA takes a forward-looking approach, allowing to explore how a technology’s environmental footprint could change in the future. For instance, as energy systems become greener and industries move toward low-carbon practices, the environmental impact of producing and using a material or technology can be greatly reduced over time.

Goal and scope

With the objective of comparing the DME production scenarios, a cradle-to-gate system modeling was implemented. The selected functional unit for the analysis is the production of 1 kg of high-purity DME. The system in Figure 5 illustrates the foreground boundaries for the two scenarios described above considering flue gas from cement production plant as the CO 2 source. The life cycle system includes in the foreground the CO 2 capture implemented by membrane separation (MS), and hydrogen production by water electrolysis (WE). A pure liquefied oxygen stream is produced from water electrolysis in the two scenarios, which is considered as a co-product replacing conventional liquefied oxygen.

In addition, Figure 4 shows the emissions in each scenario (CO 2, CO, others). The water stream containing methanol, produced in S1, is considered here as a waste and it is directly sent to the waste water treatment plant.

For comparison purposes, the conventional process “dimethyl ether production, from methanol” available in the ecoinvent 3.9 database was used. This is shown in Figure 4 as Scenario 3 (S3). In this case, methanol is first synthesized from synthesis gas (a mixture of CO, CO 2, and H 2), derived from fossil-based feedstocks such as natural gas or coal. In a subsequent step, the methanol is dehydrated to produce DME. This pathway reflects established large-scale commercial technologies and is widely used as a baseline in life cycle assessments due to its mature infrastructure and well-documented environmental impacts. Direct emissions are not displayed in Figure 4 since these are already considered in the activity “dimethyl ether production, from methanol” of Ecoinvent.

The comparison of the three scenarios is based on economic allocation as it is the one applied in Ecoinvent for DME production. The same prices were used for all scenarios. The attributional, cut-off modeling was used to consider that impacts of the end-of-life phase are allocated to the life cycle that uses the recovered materials. However, the only co-product in S1 and S2 is the liquid oxygen from water electrolysis.

The production context is that of the Netherlands, in the time frame 2025–2050. To this aim, a first analysis was conducted for the present time with the currently available Ecoinvent 3.9 database. To estimate how environmental impacts might change in the future years, the Premise tool was used. This tool modifies the Ecoinvent database to align with future scenarios provided by integrated assessment models (IAM) such as REMIND or IMAGE (Sacchi et al., 2022).

Moreover, a sensitivity analysis was conducted to uncover the influence of the energy sources on the environmental impacts.

Life cycle inventory

Foreground data considering material and energy input/output flows were obtained from process simulations of carbon capture, water electrolysis, and DME production for S1 and S2, conducted as part of this study. For the SEDMES process alone, material and energy data were sourced from literature (Skorikova et al., 2020).

The study focuses on the context of DME production in North-West Europe. Background data were selected to align with the Netherlands geographic context. These included utilities, raw materials used in the processes, and wastewater treatment. The infrastructure associated with the foreground system was excluded from the analysis. In order to assess the sensitivity of the results with respect to the local conditions, the geographical context was also evaluated for Germany, France, and Sweden.

Concerning prospective LCA, the current version of Premise includes the energy mix evolution in different countries. The selected model was REMIND, which is an IAM based on a macro-economic growth model with an engineering-based energy system model. The chosen scenario is “Remind-SSP2-NDC”, which is based on the Nationally Determined Contributions (NDC) assuming a 2–2.5 °C increase on temperature in 2,100 (Nöhl et al., 2025). Premise tool generates inventories starting from the year 2000. The background inventory in this study was thus generated for each 5 years, from 2020 to 2050.

Life cycle impact assessment (LCIA)

LCIA was performed using the ReCiPe 2016 v1.13 midpoint (H) method. This method was chosen because it provides a clear and consistent way to assess environmental impacts by following cause-effect relationships. It includes 18 impact categories that are well aligned with the system studied, giving a well-rounded view of its environmental performance.

LCA was carried out using the open-source software Brightway2. To streamline the calculations and facilitate result generation and analysis, Brightway2 was connected to ProSimPlus® through a Python interface. This allowed the automatic transfer of material and energy flow data directly from the process simulations into Brightway2.

Process simulation results

The simulation results for scenarios S1 and S2 were generated using the same input flow rates of CO 2 (i.e. 134 kmol h −1) and H 2. The hydrogen flowrate was chosen to maintain a consistent CO 2:H 2 molar ratio of 1:3 in both scenarios, which corresponds to the stoichiometric ratio. To ensure a fair comparison between the two scenarios, the same CO 2 and H 2 flowrates are applied to the indirect route (S2). Each scenario was assessed based on DME productivity, as well as energy consumption per kilogram of DME generated. Table 1 presents the simulation results, detailed by individual process stages. The data and the energy consumption (for heating and electricity) in S1 is based on literature results available for SEDMES process as indicated in Supplementary Table S5, whereas in S2 it is based on the simulation results.

Figure 5 illustrates the comparison between the two scenarios (S1 & S2) in terms of energy consumption across the different process stages: membrane separation (MS), water electrolysis (WE), methanol (MeOH) and DME synthesis processes, as well as the DME production rate of each scenario, expressed in kg/h. Note that compression of CO 2 and H 2 was considered in the SEDMES process for S1 and in the MeOH process for S2.

In both scenarios, water electrolysis is the dominant contributor to electricity consumption. This is expected, as hydrogen production through electrolysis is known to be highly energy-intensive. Membrane separation that provides the pure CO 2 stream, contributes modestly to the total electricity consumption in both scenarios, with similar values observed in S1 and S2. However, in S2, electricity consumption per kg of DME is higher than in S1, primarily due to the lower DME productivity in S2. This highlights one of the advantages of the enhanced sorption DME synthesis process (SEDMES), where methanol production does not require an additional reactor, resulting in greater energy efficiency.

The heating requirements show a clear distinction between the two scenarios. While S1 shows modest heat, demand concentrated, S2 displays significantly higher heating energy needs, particularly coming from the independent methanol synthesis step. Although S1 also requires heating for methanol production, its recirculation allows for a better heat integration, which decreases heating demand. In addition, S2 achieves a lower DME production rate (∼2,433 kg/h) compared to S1 (∼2,930 kg/h). This indicated that the additional energy invested in S2 does not proportionally increase the process output, leading to lower overall process efficiency. For that, and basing on these findings, the direct DME synthesis (S1) offers a more energy efficient pathway, with fewer unit operations and lower utility requirements.

Life cycle inventory results

The inventory data used in the assessment are summarized in Table 2. S1 relies on inventory data derived from a simulation and literature data, while for S2, the simulations results were used. The Ecoinvent data for S3 are reported in Supplementary Table S6 where infrastructure was not considered.

The foreground inventory for S1 and S2 includes different types of energy consumption, mainly divided into heating and electricity, depending on the process. The system also includes water recirculation within the process. Any wastewater produced is sent to a treatment plant. Methanol produced in S1 was considered to be part of the wastewater since its concentration was very low. In the emissions section, nitrogen is emitted from CO 2 capture technology from flue gas. The avoided CO 2 emissions are reported for each scenario, corresponding to the input CO 2 streams, captured from the flue gas. This value reflects the amount of CO 2 that is prevented from being released into the atmosphere by the cement production plant.

For comparison, Ecoinvent 3.9 gives the following energy consumption for the two-steps of the S3 production route: 0.70 kWh of heat (2.52 MJ) and 0.224 kWh electricity for 1 kg DME. This heat consumption is comparable to that of S1 (0.85 kWh/kgDME or 3.07 MJ/kg DME) and lower than S2 (2.81 kWh/kgDME or 10.1 MJ/kg DME), which indicates a larger heat integration in the Ecoinvent database. However, the electricity consumption is significantly lower than S1 (17.8 kWh/kgDME) and S2 (22.6 kWh/kgDME). The large difference in electricity consumption is mainly due to the hydrogen produced by electrolysis in S1 and S2, which is not necessary in S3.

Environmental impact results

Figure 6 illustrates the global warming potential (GWP100), expressed in kg CO 2 -eq per kilogram of DME produced, for the three scenarios. The bars show the contributions from various process stages such as hydrogen production, carbon dioxide capture, DME or methanol synthesis, in addition to the avoided CO 2 emissions. S1 and S2 are considered operating in the Netherlands (also mentioned on the graph by NL), and exhibit net GWP results much higher than for S3, with the maximum of 5.93 kg CO 2 -eq/kg DME for S2. Hydrogen production needed for S1 and S2 causes the largest difference with respect to S3. Furthermore, it is observed that SEDMES process is more efficient in terms of hydrogen production and utilization since a smaller impact is displayed even if the same stoichiometric ratio of 3:1 H 2:CO 2 is used for S1 and S2. The high impact of hydrogen production has also been highlighted for the production of several molecules from CO 2 such as, methanol and olefins where hydrogen electricity contributed between 20%–30% to the global GWP100 impact (Rosental et al., 2020) and up to 90% in a study referring to SEDMES process (Styring et al., 2022). A comparison of performances of S1 and S2 is reported in Supplementary Table S5 of the SI.

Finally, the conventional route (S3), based on fossil fuel data from the ecoinvent 3.9 database, shows a net GWP100 impact of 1.42 kg CO 2 -eq/kg DME. Interestingly, despite its reliance on fossil resources, it performs better than the CCU scenarios because those are powered by high-carbon-based electricity. The overall results for the 18 impact categories are presented in Figure 7. To better compare and assess these impacts, the results are normalized to one. All the impact categories are highly affected by the electricity consumption due to hydrogen production process. When using Dutch electricity (based on 59% natural gas), all net values for impact categories show positive values. Furthermore, across all 18 categories (notably, all toxicity, ecotoxicity, eutrophication categories), S3 consistently exhibits the lower impact values. Supplementary Figures S4–S6 in the SI provide a more detailed illustration of the impacts with a contribution analysis.

Overall, S2 performs worse than S1, largely due to the higher burdens associated with the methanol synthesis and electrolysis stage. These stages are energy intensive and highly sensitive to the carbon intensity of the electricity mix coming from natural gas and coal. The important difference between the CCU and S3 scenarios is definitely explained by the higher energy consumption (for hydrogen and CO 2 capture) and thus their contribution is reflected in all impact categories, as shown in Supplementary Figures S3–S5 in the SI. Hence a sensitivity analysis and a prospective LCA based on changes in electricity mix is conducted in the next sections.

In this LCA, the cradle-to-gate approach was used, considering that the use and end-of-life stages of DME are identical in the three scenarios. If combustion is the end-of-life, the amount of CO 2 resulting from 1 kg DME, i.e. 1.91 kg CO 2, should be added to the global warming impact. In this case, the GWP100 results are 5.90, 7.84 and 3.33 kg CO 2 -eq for S1, S2 and S3, respectively. This demonstrates that there is neither mitigation, nor negative emissions for the life cycle of DME, irrespective of the production route for the use case studied in this work. Only an important storage period of several years in long-lived materials could offer a climate benefit (Kanso et al., 2025).

Sensitivity analysis

A sensitivity analysis was conducted to assess the influence of the electricity mix in the global warming potential in scenario S1. Figure 8 shows the results obtained with data from Ecoinvent 3.9 for four European countries: Germany (S1_DE), Netherlands (S1_NL), France (S1_FR), Sweden (S1_SE).

The results clearly confirm that hydrogen production is the dominant contributor to the overall GWP across the studied countries. In France and Sweden, where electricity generation is based on low carbon sources (nuclear and hydrothermal), the net GWP becomes negative, reaching −0.84 kg CO 2 -eq/kg DME and −1.36 kg CO 2 -eq/kg DME, respectively. Moreover, both results outperform the conventional production S3. Conversely, in Netherlands and Germany, the net GWP is positive (3.99 kg CO 2 -eq/kg DME and 3.38 kg CO 2 -eq/kg DME, respectively) due to the electricity mix considered, despite the avoided CO 2 emissions in the process. However, if the electricity mix evolves, these results may approach the net value of the Swedish case. The results indicate that the SEDMES process can become climate-beneficial only if coupled with a low fossil electricity system. This comparison emphasizes that the environmental benefits of emerging DME production pathways are highly dependent on the defossilization of the electricity used in key stages, especially the water electrolysis. Furthermore, other alternatives to produce DME could be tested such as the substitution of CO 2 /H 2 syngas from flue gas and water electrolysis by syngas from biomass gasification (Liuzzi et al., 2020).

The LCA methodology applied, i.e., type of allocation, system expansion or subtraction for co-products, is decisive for the results and their interpretation, as highlighted by Garcia-Garcia et al. (2021). In their review, the climate change performance of DME obtained by CCU with wind energy, on a cradle-to gate base, achieves −0.48 to −1.07 kg CO 2 -eq, or 0.12–0.15 kg CO 2 -eq when end-of-life combustion is included, according to the cited sources. The GWP100 results obtained in this study when defossilized electricity is used, corroborate those found in the literature.

Prospective LCA results

The purpose of this study is to highlight how changes in the Dutch electricity mix over the years influence the overall environmental impact of the three scenarios of DME production. The results were obtained using the premise tool which transforms the Ecoinvent database according to IAMs (Sacchi et al., 2022). This means that the Ecoinvent database for prospective LCA is different from the one employed in conventional LCA in Sections 3.3, 3.4. Hence, the impact results at present time are different. The “Remind-SSP2-NDC” model calculates projections from 2000, and the results for 2020 are lower than the ones of Sections 3.3, 3.4. However, the trends are interesting to be analyzed. The results for prospective LCA are illustrated in Figure 9 for GWP100 impact (values reported in Supplementary Table S7 in SI), and Supplementary Figures S7–S10 in SI represent the other impacts.

In both S1 and S2 the global warming impact is quite high during the early years (2020–2025), especially for S2 in 2020 which peaks at 3.46 kg CO 2 -eq/kg DME. This high impact is due to the carbon-intensive electricity mix reflecting the Dutch energy grid. In both scenarios, hydrogen production is the dominated contributor. However, since S1 is the direct conversion route, this scenario benefits from a reduction in overall energy use by combining multiple process steps and enabling more efficient heat integration between the methanol and DME production stages over the years.

According to the SSP2-NDC trajectory, and over years, the Dutch electricity mix gradually defossilizes, and the GWP impacts of S1 and S2 will decrease substantially, where, around 2035, the net GWP100 becomes negative. By 2050, both production scenarios achieve climate benefits, with S1 reaching nearly −1.63 kg CO 2 -eq/kg DME and S2 just below −1.41 kg CO 2 -eq/kg DME. These negative values reflect the effect of a low-fossil energy mix over years, leading to the avoidance of more CO 2 than is emitted during the process. However, if the end-of-life stage of DME is included in the system boundaries, there is no more negative net impact, but the values become 0.28 and 0.5 kg CO 2 -eq for S1 and S2, respectively. By 2050, the CCU routes, with net GWP100 cradle-to-grave lower than S3 (3.05 kg CO 2 -eq including combustion), could mitigate the climate change impact with respect to the conventional, fossil production route.

Scenario S3, which represents the conventional production, does not benefit from the electricity defossilization trends over years, because it is based on the fossil sources. This scenario shows a minimal variation over the 30 years period, conserving a constant GWP100 impact around 1.2 kg CO 2 -eq/kg DME. This underscores the limited environmental improvement potential of S3 under existing industrial configurations.

Concerning the other impact categories, although the most impact values decrease in time, only four out of the 18 evaluated demonstrate lower values for S1 than S3 after 2030: photochemical oxidant formation (human and ecosystems), water consumption, and, obviously, fossil fuel consumption. The trends of the impacts regarding land use and ecotoxicity are different from GWP100 as is displayed in Supplementary Figures S6,S8. These trends are correlated with the prospective electricity mixes which will use double of wind power and eight times more solar power while reducing all coal dependence. These mixes are reported in Supplementary Table S8. Indeed, the different changes correspond to the increase of mineral use, such as lead and copper, and the use of more forest land to implement photovoltaic panels for solar power.

The results of this analysis show that the CCU-based DME production routes (S1 and S2) become more advantageous over the years, for the global warming potential, and for the most of impact categories. The progressive decarbonization of the energy system plays an important role in transforming these emerging technologies into climate-beneficial solutions.

Two CCU routes for DME production were investigated for technical and environmental performances: enhanced direct synthesis process from CO 2 and H 2 (S1) (the SEDMES process, existent at pilot scale), and the two-steps route with methanol as intermediary compound (S2). Both scenarios were compared with the conventional production of DME (S3), i.e., a two-steps process using syngas from natural gas reforming as raw material.

In terms of energy consumption, S2 (indirect route) consumes more energy than S1 (SEDMES route): 17.80 kWh and 3.07 MJ per kg DME produced, for electrical and heat energy respectively in S1, compared to 22.59 kWh and 10.14 MJ per kg DME in S2. This energy analysis highlights some advantages of the SEDMES process regarding a better heat integration and a more hydrogen efficient process in addition to the higher production of DME. The analysis revealed that hydrogen production is the dominant electricity consumer, which is reflected also in the environmental impacts.

Concerning the global warming potential impact of the cradle-to-gate system, S1 (3.99 kg CO 2 -eq) performs better than S2 (5.93 kg CO 2 -eq), but both CCU scenarios perform worse than the conventional S3 (1.42 kg CO 2 -eq), in the current Dutch context.

Nevertheless, S1 demonstrates promising environmental performance over time. For instance, as the Dutch electricity grid becomes less carbon-intensive in the coming years, S1 not only maintains high CO 2 conversion efficiency but also holds the potential to mitigate climate change with respect to the traditional production route, and under optimized conditions. If the SSP2-NDC trajectory is followed, this impact decreases to −1.63 kgCO 2 -eq/kgDME by 2050 in the cradle-to-gate boundaries, and 0.28 kgCO 2 -eq if the DME combustion is considered as end-of-life.

Similar to S1, the indirect route S2 has a high global warming impact at present according to prospective LCA (1.37 kgCO 2 -eq/kgDME in 2025) due to the high energy needed in the two-steps process (methanol and DME synthesis), and especially the electrical energy needed for hydrogen production. The climate impact decreases over years, and, like S1, S2 could mitigate the impact with respect to the conventional route S3. This later has a positive, relative static GWP impact over years, of around 1.2 kgCO 2 -eq/kgDME.

Only when aligned with a low-fossil electricity system, CCU production of DME offers a viable and a better environmental alternative to fossil-derived DME production processes. Improving the energy efficiency of hydrogen production and synthesis technologies and especially the use of defossilized energy sources are crucial for the environmental acceptability of the new CCU technologies.