Life Cycle Assessment of Greenhouse Gas Emissions in Hydrogen Production via Water Electrolysis in South Korea

Abstract

This study evaluated the greenhouse gas (GHG) emissions associated with hydrogen production in South Korea (hereafter referred to as Korea) using water electrolysis. Korea aims to advance hydrogen as a clean fuel for transportation and power generation. To support this goal, we employed a life cycle assessment (LCA) approach to evaluate the emissions across the hydrogen supply chain in a well-to-pump framework, using the Korean clean hydrogen certification tiers. Our assessment covered seven stages, from raw material extraction for power plant construction to hydrogen production, liquefaction, storage, and distribution to refueling stations. Our findings revealed that, among the sixteen power sources evaluated, hydroelectric and onshore wind power exhibited the lowest emissions, qualifying as the Tier 2 category of emissions between 0.11 and 1.00 kgCO₂e/kg H₂ under a well-to-pump framework and Tier 1 category of emissions below 0.10 kgCO₂e/kg H₂ under a well-to-gate framework. They were followed by photovoltaics, nuclear energy, and offshore wind, all of which are highly dependent on electrolysis efficiency and construction inputs. Additionally, the study uncovered a significant impact of electrolyzer type on GHG emissions, demonstrating that improvements in electrolyzer efficiency could substantially lower GHG outputs. We further explored the potential of future energy mixes for 2036, 2040, and 2050, as projected by Korea’s energy and environmental authorities, in supporting clean hydrogen production. The results suggested that with progressive decarbonization of the power sector, grid electricity could meet Tier 2 certification for hydrogen production through electrolysis, and potentially reach Tier 1 when considering well-to-gate GHG emissions.

1. Introduction

Water electrolysis, which splits water molecules into hydrogen and oxygen using electricity, is the cleanest method of hydrogen production, assuming the use of carbon-free electrical energy. This method also fulfills the demand for grid flexibility and long-term energy storage, enabling easy separation of electrolyzers to reduce grid load during supply shortages [1]. According to the roadmap for activating the hydrogen economy announced by the Ministry of Trade, Industry, and Energy in South Korea, the country aims to lead in this field through two primary pillars: hydrogen vehicles and fuel cells [2]. Regarding hydrogen production, the strategy focuses on transitioning from grey hydrogen (currently obtained from by-product hydrogen and large-scale extraction) to green hydrogen, which is produced through mass electrolysis, achieving economic feasibility, and introducing overseas carbon-free hydrogen.

To promote the use of hydrogen for mobility and energy, Korea plans to expand the deployment of various hydrogen vehicles and establish mass production systems. By 2040, the goal is to deploy 40,000 hydrogen buses, 80,000 hydrogen taxis, and 30,000 hydrogen trucks, thereby enhancing the clean public transportation system and reducing particulate matter [3]. In the energy sector, Korea aims to increase the installed capacity of fuel cells for power generation from the current 879 MW to over 15 GW by 2040, encompassing both domestic and export volumes. The capacity of residential and building-scale fuel cells will also expand from the current 13 MW to over 2.1 GW. Furthermore, there are plans to develop and commercialize hydrogen gas turbine technology [3]. As hydrogen becomes more integrated into the Korean economy, it is expected to replace approximately 5% of final energy consumption by 2040, which is equivalent to 10.4 million tons of oil equivalent [2].

Given the essential role of clean hydrogen in achieving carbon neutrality, it has become necessary to establish clean hydrogen certification standards and support measures tailored to each country’s circumstances. Korea has introduced a clean hydrogen certification system aimed at reducing investment uncertainties in hydrogen production projects and enhancing national energy security. The certification levels are categorized based on well-to-gate GHG emissions, assessing GHG emissions from electricity generation, transmission, and hydrogen production: Tier 1 with 0 to 0.10 kg CO₂e/kg H₂, Tier 2 with 0.11 to 1.00 kg CO₂e/kg H₂, Tier 3 with 1.01 to 2.00 kg CO₂e/kg H₂, and finally Tier 4 with 2.01 to 4.00 kg CO₂e/kg H₂ [4]. The United States also categorizes green hydrogen certification into four levels, but with slightly different thresholds: Tier 1 with 0 to 0.45 kg CO₂e/kg H₂, Tier 2 with 0.45–1.5 kg CO₂e/kg H₂, Tier 3 with 1.5–2.5 kg CO₂e/kg H₂, and Tier 4 2.5–4 kg CO₂e/kg H₂ [5]. This difference arises because the U.S. considers carbon capture and storage (CCS)-linked reforming using low-cost gas a key method for lowering hydrogen production costs, facilitating the large-scale deployment of clean hydrogen that is commercially viable [6]. In contrast, Europe sets a higher threshold for clean hydrogen, with 10.92 kg CO₂e/kg H₂ as the EU’s average carbon intensity for hydrogen produced through natural gas extraction without CCS [7]. Hydrogen that reduces CO₂ emissions by lowering it to around 40% (i.e., 3.38 kg CO₂e/kg H₂), is classified as premium hydrogen. This premium hydrogen can be further divided and certified as low-carbon or green hydrogen [5,7]. The scope of carbon accounting in Korea and the U.S. is currently limited to the well-to-gate stage. Well-to-gate accounting covers all processes from the extraction of raw materials (e.g., crude oil, natural gas, or biomass) to the point where the product is ready to leave the production facility (the gate). However, Korea has a plan to extend this to well-to-port or well-to-pump accounting that includes the transportation and delivery of the product to the end user [8]. Europe adopts a well-to-wheel approach, covering maritime transport, receiving terminals, land transport, and hydrogen utilization [5].

In this study, we utilized life cycle assessment (LCA) to evaluate which power sources in Korea are best suited for producing clean hydrogen at various tiers. Previous studies on the LCA of hydrogen production via water electrolysis in Korea did not address the range of well-to-pump accounting approaches [9,10], while the study in [11] only excluded the compressing stage but focused on one limited type of water electrolysis. In this study, we examined the expected changes in greenhouse gas (GHG) emissions when applying a well-to-pump evaluation with three different water electrolysis methods. Additionally, if electrolyzers use grid electricity in Korea, we assessed which tier of clean hydrogen can be achieved based on the future energy mixes projected by various studies for the Korean power sector. The key contributions of this study are summarized as follows:

• To evaluate GHG emissions from well to pump, the hydrogen supply chain in Korea was divided into seven stages. These stages encompass everything from extracting raw materials for power plant construction to electricity generation, hydrogen production, liquefaction, storage, and distribution to hydrogen refueling stations. These stages are labeled J1 through J7. The total GHG emissions from J1 to J7 provide a comprehensive well-to-pump assessment. Additionally, if only the emissions from J1 and J2 are taken into account, a well-to-gate assessment can also be conducted.
• To conduct an LCA, we collected publicly available data regarding the services, materials, and energy consumed during the evaluation stage. Our primary focus was on diverse datasets that highlight how GHG emissions significantly depend on power sources, electrolyzer efficiency, and energy mixes, especially when using grid electricity. We gathered and compared data from sixteen power sources currently in operation or planned for operation in Korea, as well as three types of water electrolyzers from various sources.
• We created a classification table (with the results of the LCA) that categorizes sixteen power sources according to Korea’s clean hydrogen certification standards. Hydroelectric dams and onshore wind turbines were considered the cleanest options, placed in Tier 2 when using the well-to-pump approach. However, they can be classified as Tier 1 if well-to-gate emissions are taken into account. Following hydroelectric and onshore wind sources, photovoltaics, nuclear energy, and offshore wind are next in line. Additionally, this study highlights that the results were highly sensitive to the methods of electrolysis used, specifically concerning their efficiency and the amount of non-energy inputs required during construction.
• Based on the future energy mixes projected by Korean energy and environmental authorities for the years 2036, 2040, and 2050, we predicted the position of hydrogen production using grid-connected water electrolysis. The results indicated that, with the expected energy mixes stemming from successful decarbonization in the Korean power sector, future grid electricity will qualify as a Tier 2 energy source for clean hydrogen production. Furthermore, when well-to-gate accounting is applied, this could reach Tier 1 status.

The remainder of this paper is organized as follows: Section 2 discusses related studies, comparing international results of LCA analysis for hydrogen production through water electrolysis. It also summarizes studies on hydrogen production in Korea, using a framework similar to that for foreign research. Based on the previous overseas and domestic studies, the types of water electrolysis to be used in our study were determined. Section 3 describes the methods employed in the assessment, outlining the system boundaries, detailing power plant specifications, and gathering data on input materials and their carbon intensities. Section 4 presents the results of the LCA analysis, detailing the GHG emissions associated with hydrogen production from sixteen power sources utilizing three electrolysis methods, each with eleven different performance specifications. Finally, Section 5 summarizes the main findings of the study.

2. Related Studies

2.1. Related Works from Overseas Studies

Table 1. Previous overseas studies and projected GHG emissions (kg ${\mathrm{CO}}_2$e/kWh).

No.	Author	Country	Scope 1	Electrolyzer	Efficiency (%)	Electricity Sources	Emission	Datas
1	Patel (2024, [12])	Finland	J2	PEMEC	71.6	PV	0.075	[13]
						Wind	0.018
						PV 50, Wind 50	0.048
2	Jolaoso (2024, [14])	U.S.	J1–J2	SOEC	71.7	PV 50, Coal 50	0.207	[15,16,17]
3	Yang (2024, [18])	China	J2	AEL	80.5	Grid 2021	0.922	[15,19]
						Hydro	0.019
4	Pawlowski (2023, [20])	Poland	J1–J2	AEL	70.0	PV standing	0.090–0.130	[15,16]
						PV tracker	0.082–0.116
5	Hamed (2023, [21])	-	J2	AEL	-	PV	0.069–0.129	-
				-		PV	0.060–0.210
				SOEC		PV	0.006
						On-wind	0.008
6	Schropp (2022, [22])	Germany	J2	PEMEC	71.0	Grid 2022	0.006–0.007	[15,23]
						Grid 2030	0.002–0.003
						Grid 2050	0.0005–0.0006
7	Elgowainy (2022, [24])	U.S.	J2	PEMEC SOEC	70.6 93.5	LWR	0.009	[25,26]
						Renewables	0
						LWR	0.012
8	Ozawa and Kudoh (2021, [27])	Japan	J1–J2, J5, J7	AEL/PEMEC	-	PV in AUS	0.089–0.240	[28]
						Wind in AUS	0.036–0.082
						Wind in NOR	0.049–0.095
						PV in UAE	0.091–0.244
9	Gerloff (2021, [29])	Germany	J2	AEL	76.2	Grid	0.523–0.806	[15]
						Renewables	0.087
				PEMEC	73.0	Grid	0.543–0.838
						Renewables	0.088
				SOEC	93.2	Grid	0.472–0.705
						Renewables	0.115
10	Sadeghi (2020, [30])	U.S.	J1–J2	AEL	68.5	CSP	0.062	[16,31,32,33]
						PV	0.092
11	Bareiß (2019, [34])	Germany	J2	PEMEC 2	71.0	Grid 2017	0.885	[13,15,16,35]
						Grid 2050	0.348
						Renewables	0.009
12	Bhandari (2014, [36])	-	-	AEL/PEMEC /SOEC	62–86	PV	0.147	[31]
						Nuclear	0.069
						Wind	0.024
						Biomass	0.105
						Coal Gasification	0.360
						Hydro	0.060
						UCTE Grid 2010	0.975
13	Cetinkaya (2012, [37])	Canada	J1–J3	PEMEC	85.0	Wind	0.029	[15,38]
			J1–J2			PV	0.072
14	Maack (2007, [39])	Iceland	J1–J3, J7	PEMEC	70.0	UCTE Grid 2020	0.852	[13,40,41,42]
15	Spath and Mann (2004, [43])	U.S.	J1–J3, J5 3	PEMEC	85	Wind	0.029	[15,38]

¹ The scopes are defined in the following Figure 1. ² Three types of PEMEC (proton exchange membrane electrolyzer cell) were considered: one manufactured in 2017 with lifetime of 8760 h and the others projected for 2050 with lifetimes of 8760 and 3000 h, respectively. ³ Only the storage was included in J5.

provides a summary of previous studies that conducted LCA analyses on hydrogen production through water electrolysis. These studies spanned multiple regions, including North America (United States, Canada), Europe (Germany, Iceland, Finland, Poland), and Asia (China, Japan). The studies were categorized based on the region of analysis, system boundary, type and efficiency of electrolyzers assumed, power sources, GHG emissions (kg CO₂e/kWh), and data sources used for analysis.

The studies showed a wide range of GHG emissions depending on the electrolyzer type, energy source, and system boundary. Elgowainy (2022, [24]) demonstrated the lowest emissions with renewable-energy-powered proton exchange membrane electrolyzer cell (PEMEC) and solid oxide electrolyzer cell (SOEC) systems (near zero). Similarly, Maack (2007, [39]), using hydropower, reported emissions of 0.019 kg CO₂e/kg H₂. The highest emissions were reported by Yang (2024, [18]), with 87.78 kg CO₂e/kg H₂ for an alkaline electrolyzer (AEL) in China’s fossil-fuel-heavy grid mix. Similarly, Sadeghi (2020, [30]) reported high emissions for coal gasification hydrogen production. The differences in GHG emissions across the electrolyzer types were consistent with the efficiency levels. SOEC consistently reported lower emissions due to higher efficiency, while AEL tended to have higher emissions, especially when fossil fuels were used in the energy mix.

Electrolyzer efficiency significantly influences GHG emissions, as a higher efficiency means less electricity consumption per kilogram of hydrogen produced. Studies such as those by Elgowainy (2022, [24]) and Gerloff (2021, [29]) reported the highest electrolysis efficiency for SOEC systems, with values up to 93.5%. These electrolyzers operate at high temperatures, making them more efficient than AEL and PEMEC systems, resulting in lower overall GHG emissions. PEMEC systems, though slightly less efficient than SOEC, still demonstrate strong performance, with efficiencies ranging from 70.6% to 85%, as reported by Spath and Mann (2004, [43]) and Elgowainy (2022, [24]). These electrolyzers are often favored in scenarios where renewable energy sources like wind and solar are employed. AEL systems generally show lower efficiency than SOEC and PEMEC systems. Pawłowski (2023, [20]) reported an AEL efficiency of 70% with fixed solar panels, while Yang (2024, [18]) noted a maximum of 80.5%. While AEL is cheaper and better established, its lower efficiency leads to higher energy consumption and emissions.

The type of energy source used to power electrolyzers is the primary factor determining the environmental impact of hydrogen production, with renewable energy consistently showing the lowest emissions. Studies such as those by Elgowainy (2022, [24]) and Cetinkaya (2012, [37]) demonstrated that renewable energy sources (solar, wind, and hydropower) lead to near-zero GHG emissions during hydrogen production. For instance, the renewable-powered PEMEC system in Elgowainy (2022, [24]) showed zero GHG emissions, while the wind-powered electrolysis in Cetinkaya (2012) resulted in emissions of 0.029 kg CO₂e/kg H₂. The reliance on fossil-fuel-heavy grid mixes significantly increases emissions. Yang (2024, [18]), focusing on China’s grid mix (60.6% coal), reported substantially higher GHG emissions (87.78 kg CO₂e/kg H₂) compared to regions with cleaner energy mixes. Maack (2007, [39]), using Iceland’s primarily renewable energy, showed much lower emissions, emphasizing the impact of grid composition on hydrogen’s environmental footprint. Studies like Elgowainy (2022, [24]), which used nuclear power (Light Water Reactor, LWR) as an energy source, reported minimal emissions, with PEMEC showing emissions as low as 0.4 kg CO₂e/kg H₂. Nuclear energy presents a low-carbon option but involves considerations around waste and plant construction emissions.

The scope of the system boundary (i.e., which stages of hydrogen production and transportation were included in the LCA) also played a critical role in the GHG results. Some studies, such as Maack (2007, [39]) adopted a well-to-pump approach, evaluating emissions from raw material extraction to hydrogen ready to be used for customers. This approach provides a comprehensive assessment but tends to report higher overall emissions as more processes are included. Other studies, like Spath and Mann (2004, [43]) and Bhandari (2014, [36]), focused on well-to-gate boundaries, considering only the production stages, without downstream emissions. These studies reported lower GHG emissions but may have underestimated the full environmental impact of hydrogen use, especially if transportation and distribution stages involve fossil fuels.

Different regions showed distinct trends in hydrogen production emissions due to their unique energy mixes and policy landscapes. Elgowainy (2022, [24]) showed the U.S.’s capability to achieve low GHG emissions with renewable energy integration, particularly in regions with strong nuclear and renewable energy sources. Gerloff (2021, [29]) and Bareiß (2019, [34]) illustrated how Germany’s ambitious renewable energy targets could drastically reduce emissions from hydrogen production by 2050. Germany’s future energy mix, favoring solar and wind, highlights the critical role of government policies in shaping hydrogen’s environmental footprint. Yang (2024, [18]) underscored the challenges faced by regions heavily dependent on fossil fuels, such as China. Without significant grid decarbonization, hydrogen production in such regions will continue to have high GHG emissions, limiting its potential as a clean energy solution. Cetinkaya (2012, [37]) and Maack (2007, [39]) showed how countries (Canada and Iceland, respectively) with abundant renewable resources (wind, hydro) can achieve much lower emissions in hydrogen production, showcasing the potential of green hydrogen in these regions.

From a methodological perspective, Ozawa and Kudoh (2021, [27]) and Patel (2024, [12]) employed Monte Carlo simulations to address uncertainties in energy consumption and carbon intensity. This statistical approach helps provide a range of possible outcomes, offering more robust predictions for policy decisions.

2.2. Related Works for Korea

In

Table 2. Previous studies for Korea and projected GHG emissions (kg ${\mathrm{CO}}_2$e/kWh).

No.	Author	Country	Scope	Electrolyzer	Efficiency (%)	Electricity Sources	Emission	Data
1	Malik (2023, [9])	Korea	J1, J2, J5–J7	AEL	77.9	Wind	0.008	[15]
2	Boreum (2021, [11])	Korea	J1–J7 1	AEL	82	Grid 2021	0.360–0.411	[15,16]
						Grid 2030	0.203–0.253
						Grid 2050	0.082–0.130
						Renewables	0.040–0.088
3	Eunji (2018, [10])	Korea	J1–J2	-	65–80 56–71.5	Grid	0.578	[26,44,45]

¹ With excluding J3, only the liquefaction is included in J5.

, three studies are summarized, which assessed hydrogen production, transportation, and the associated GHG emissions in Korea. These studies provided insights into different hydrogen-related technologies and processes in Korea.

Eunji (2018, [10]) reported total GHG emissions of 19.265 kg CO₂e/kg H₂ for upstream processes and hydrogen production, with emissions adjusted to the assumed Korean grid mix. This resulted in emissions between 10.496 and 13.505 kg CO₂e/kg H₂. Malik (2023, [9]) analyzed three cases of hydrogen production and transportation, the emissions of which ranged from 4.57 to 5.35 kg CO₂e/kg H₂. This study emphasized the environmental impact of hydrogen transportation methods, concluding that compressed gas through pipelines emits the least GHG at 0.108 kg CO₂e/kg H₂. Boreum (2021, [11]) focused on GHG emissions under various energy mix scenarios for 2021, 2030, and 2050. The current mix resulted in 30.7 kg CO₂e/kg H₂, while a 2050 solar-heavy mix could reduce emissions to as low as 3.12 kg CO₂e/kg H₂.

The efficiency of hydrogen production in [10] varied depending on the location of the electrolyzers. Efficiency ranged from 65 to 80% for production outside refueling stations and from 56 to 71.5% at refueling stations. Malik (2023, [9]) focused on centralized green hydrogen production using solar power but did not report specific efficiency ranges for electrolyzers. Instead, the study evaluated the overall sustainability of the system. Boreum (2021, [11]) reported electrolyzer efficiency at 82%, but the study also compared different energy mixes and their impact on overall emissions. It assumed a 10-year lifetime for electrolyzers, affecting long-term efficiency considerations.

Eunji (2018, [10]) did not focus on transportation but provided insights into hydrogen production efficiency and GHG emissions based on electricity sources. Malik (2023, [9]) evaluated several hydrogen transportation methods, including compressed gas via pipelines, trailers, LOHC, and liquefied hydrogen. Their study concluded that compressed gas through pipelines is the most environmentally sustainable option, with emissions of 0.108 kg CO₂e/kg H₂. Boreum (2021, [11]) primarily focused on the costs of hydrogen production and different energy mixes but did not extensively cover transportation methods. The energy mix, however, directly impacted hydrogen-related emissions.

3. LCA Scope and Data

3.1. LCA System Boundary

The system boundaries assumed in this study are illustrated in Figure 1, based on the LCA methodology [46], which considers the environmental impact from raw material extraction to final disposal. The hydrogen supply chain consists of seven stages, from J1 to J7. The bottom right corner of each box representing the stages indicates the lifetime of the equipment used. In this context, J2 has a variable lifetime depending on the electrolyzer assumed for the hydrogen supply chain, and thus the lifetime indicated is based on the life cycle inventory (LCI) data considered in

Table 3. Electrolysis types, parameters, and data used in this study.

Electrolyzer	Reference	Lifetime	H2 Purity (%)	Efficiency (%)	H2 Production (kg H2/h)	Electricity (kWh/kg H2)
AEL	[47]	20 y	-	80.5	26	49
	[48]	80,000 h	<99.5	72	130	50
	[49]	30 y	99.5–99.9	70	90	49.5
	[18]	10 y	-	85.6	26	49
PEMEC	[47]	20 y	-	78.9	48	50
	[50]	40,000 h	99.99	73.5	22.1	57.5
	[49]	30 y	99.99	80	90	49
	[34] 1	20 y	-	71	18.75	55
	[39]	15 y	<99.99	87	5.38	53
SOEC	[50]	10,000 h	99.9	99	0.0297 2	65

¹ It provides two specifications for years 2017 and 2050. They are almost the same but differ in the non-energy input materials. ² This value depends on the assumed water electrolysis capacities in the references. In [50], 1 kW small experimental SOEC was assumed. In contrast, the AEL capacity assumed in [48] was 6 MW.

. Additionally, in J6, the transportation vehicles are classified as trucks and trains according to the hydrogen demand sites, and the assumptions regarding transportation are stated in

Table 4. Specifications of transport vehicles. Data used in J6.

Transportation	Pressure (bar)	Temperature (K)	Storage Capacity (kg H2)	Lifetime (year)	Fuel Efficiency (km/L)
${\mathrm{LH}}_2$ Truck	6	20	4000	14	2.55
${\mathrm{LH}}_2$ Train (Rail)	6	20	10,000	30 (60)	4.25

.

The materials input at each stage are classified into non-energy, energy, and fuel categories, represented by arrows, with the blue elements indicating the stacks. Non-energy inputs are listed in order of their significant contributions to GHG emissions, with a maximum of six inputs noted. The dotted line represents the electrical energy supplied from each power source in J1, which is utilized in the operation and maintenance of each process.

The overall hydrogen supply chain is as follows:

• J1: This stage involves the construction and operation of power sources used to generate electricity for the subsequent stages (J2 to J7).
• J2: Hydrogen is produced through electrolysis, using three types of electrolyzers: alkaline (AEL), proton exchange membrane (PEMEC), and solid oxide (SOEC). This stage includes the manufacturing of equipment and the stack inputs for the electrolysis process [18,34,39,47,48,49,50].
• J3: At this stage, the produced hydrogen is compressed from 30 bar to 70 bar in preparation for transportation via a pipeline to a liquefaction plant [51,52].
• J4: Hydrogen is transported through a 60.9 km pipeline from a hypothetical production site in Incheon to a liquefaction plant in Suwon, Korea. This stage considers both pipeline manufacturing and transportation [51,53].
• J5: Hydrogen is liquefied and stored at this stage. This involves the construction and operation of the liquefaction plant and storage tanks [9,51,54].
• J6: Liquefied hydrogen is transported to regional refueling stations. This includes the electrical energy used by transfer pumps, the manufacturing of transportation vehicles, and the fuel consumed during transport to demand sites [51,54,55,56].
• J7: The final stage involves the supply of liquefied hydrogen to refueling stations. This includes the construction and operation of the refueling stations, as well as the balance of plant (BoP) equipment [54].

3.2. Power Plant Assumptions

This study evaluates the GHG emissions associated with hydrogen production and its supply chain. It uses detailed information on the inputs and quantities of materials used across stages J1 to J7. In

Table 5. Specifications of power plants.

Power Plant	CF	Carbon Capture Rate (%)	Lifetime (y)	References
Oil	0.90	-	30	[57]
Coal	0.59	-	30	[57]
Coal-CCS	0.59	90.7	30	[57,61]
IGCC	0.86	-	30	[62]
NGCC	0.50	-	30	[57]
NGCC-CCS	0.50	90	30	[63]
WtE	0.73	-	25	[64,65]
Biomass	0.88	-	25	[66,67]
PWR	0.85	-	60	[57,59,68]
PV	0.17	-	25	[57,69]
CIGS	0.17	-	5	[57,70]
On-wind	0.24	-	30	[43,49]
Off-wind	0.36	-	30	[43,49]
Hydro	-	-	30	[57]

, the power plants used in stage J1 are specified: oil, coal, coal-50 (co-fired coal plants with 50% ammonia combustion), coal-100 (retrofitted coal plants fueled by 100% ammonia), coal-CCS (carbon capture and storage), IGCC (integrated gasification combined cycle), NGCC (natural gas combined cycle), NGCC-CCS, WtE (waste to energy), biomass, PWR (pressurized water reactor), PV (photovoltaic) and CIGS (copper indium allium selenide) solar cells, onshore and offshore wind turbines (referred to as on-wind and off-wind, respectively), and hydro dams. The plants have a diverse lifetime of 5, 25, 30, and 60 years, with infrastructure having a lifetime of 100 years in the case of hydro plants. The equipment includes turbines, generators, transformers, control systems, electrical systems, and cables. The infrastructure includes dams, pipelines, bridges, roads, and other civil structures [57]. The capacity factors (CF) for each power source were projected for 2036 [58], except for PWR (based on the 2030 energy mix given in [59]) and hydro (based on a 2037 energy mix analysis in [60]).

The analysis in this study considers different scenarios for coal and NGCC power plants with CCS. For the coal-fired power plant with CCS, additional inputs of 0.103 kg/kWh of limestone and $2.42 \times {10}^{-7}$ kg/kWh of monoethanolamine were taken into account [57,61]. This study also examined the construction of new power plants equipped with CCS, referred to as NGCC-CCS, and collected relevant data for GHG calculations [63]. In the case of IGCC plants, the study included the mining and transporting of coal, with 155.3 tons of coal required per hour [62]. Additionally, the study looked at diesel consumption during decommissioning. Due to insufficient data on coal imports for IGCC, the study only examined Korea’s main coal import countries, which include Russia (12.4%), Australia (55.1%), Indonesia (22.5%), and the United States (10.1%) [62]. Sea distances were calculated using a distance calculator from the Korea Hydrographic and Oceanographic Agency and the Sea Distances website of the United States [71,72]. The imported coal quantity (ton km/kWh) was calculated based on sea distances and proportional import amounts.

The WtE process, specifically gasification, was found to reduce GHG emissions compared to traditional incineration because it avoids the release of oxidized pollutants like NOx and SOx [73]. Data on solid refuse fuel production and electricity generation at WtE plants were collected separately [64,65]. The annual production of solid refuse fuel, which amounted to 35,220.7 tons, was converted to a per kWh value for emissions calculations based on a plant lifetime of 25 years, during which it will generate 858,870,423 kWh of electricity. Regarding biomass, the system encompassed biomass tanks, reactors, scrubbers, filters, gas engines, and the entire fuel production process, from raw material procurement to the production of plant oils [66,67].

The analysis in this study covered the life cycle of solar PV systems using monocrystalline silicon photovoltaics, from raw material extraction to polysilicon production, wafers, modules, and installation, including the balance of system (BOS). Due to data gaps, only 62.4% of the system’s material mass was accounted for [69]. Data for CIGS cover 97% of GHG emissions from cell manufacturing [70]. For wind power (onshore and offshore), the materials used in the foundations, blades, and turbines were considered. In the case of PWR, the data on uranium reduction were usually not open and excluded, as in similar studies [68], but transport data were included [57]. For hydro plants, inputs like industrial water, electricity, lubricants, and oils were considered, although transportation data were omitted [57]. All inputs for various power plants were converted to kg/kWh, and the electricity needed for stages J2 to J7 was calculated by multiplying the electricity generated by the equipment’s operational and maintenance requirements.

3.3. Input Materials and Data

Let $c_{i,j}$ represent the consumption of input material i in the process j, while let $e_i$ denote the carbon intensity when consuming material i. The total carbon emission in the hydrogen production in this study was determined by ${\Sigma }_{j=1}^7{\Sigma }_i{e}_i{c}_{i,j}$, resulting in an output in kg CO₂e/kg H₂ [27,74].

The carbon intensity for raw materials and energies for the power plant in J1 were obtained from the Korea Environmental Industry and Technology Institute [57], and for materials not listed in [57], such as various chemicals and metals, additional data were collected from external sources [75,76]. Information on elemental and cleaning agents was further obtained from relevant emissions factor databases [77,78,79,80,81,82,83,84].

Data for electrolyzers (AEL, PEMEC, and SOEC) used in J2 were collected from multiple sources, as shown in Table 3. Each configuration was used to evaluate GHG emissions and compared with each other. However, for SOEC, only one data source was used, which is still in the development phase and several studies have highlighted a lack of sufficient data [14,18]. The electricity consumed during water electrolysis was assumed to be supplied by the power sources analyzed in J1, with an energy consumption ranging between 49 and 65 kWh, depending on the references. Additionally, the stacks used in AEL and PEMEC involved sodium hydroxide and platinum/iridium, respectively, and unlike other inputs for equipment manufacturing, these are replaced every five years. However, Maack (2007) indicated that the GHG emissions from the catalyst are negligible and thus were not included [39].

In J3, the focus was on the non-energy materials used in compressor manufacturing and the electricity consumed during operation. The compressor’s efficiency was assumed to be 70%, with a lifetime of 22 years. Stainless steel, specifically 18/8 (18% chromium, 8% nickel), is the main material used, with 0.001 kg/kg H₂ being required, while polymers were assumed to be used as tube insulation, with $1.06 \times {10}^{-5}$ kg/kg H₂ assumed [52].

In J4, hydrogen compressed to 70 bar was transported to the liquefaction plant through pipelines at a pressure of 2 bar. The pipeline was made of L360/X52 material with a diameter of 900 mm, and included additional epoxy and low-density polyethylene layers of 300 mm and 5 mm, respectively, to address corrosion and protection issues [51,53].

In J5, the liquefaction plant operated for approximately 346 days per year with an efficiency of 94.7%. The power sources in J1 supplied the 6.76 kWh/kg H₂ of electricity consumed during hydrogen liquefaction [9]. The storage tanks for liquefied hydrogen consisted of a boss, liner, and vacuum layer. The boss and vacuum layer were made of aluminum, while the liner comprised 316 stainless steel [54]. Unlike the liquefaction plant, the electricity consumed during the operation of the storage tank was not considered.

The electricity consumption of the transfer pumps for delivering liquid hydrogen in J6 was calculated by dividing the total number of pumps required for each hydrogen demand site by the energy consumption of a single pump. GHG emissions from the diesel consumed during hydrogen transportation from storage tanks to demand sites were calculated based on specified transport vehicles and distances. For Busan, Daegu, Gwangju, and Daejeon, it was assumed that the hydrogen would be transported by train to the nearest station and then delivered to hydrogen refueling stations within the city by truck. For other regions, transportation was assumed to occur solely by truck. Since the exact locations of the hydrogen refueling stations are unknown, the latitude of each city center was used to calculate the shortest transport distance using the formula for the great circle distance on the Earth’s surface. The specifications for hydrogen trucks and trains are provided in Table 4, based on a design report published by the Ministry of Land, Infrastructure, and Transport [55]. Since the commercial use of hydrogen trucks and trains has not yet been established, accurate data for the LCI were unavailable. Therefore, inputs for constructing the truck frame and liquid hydrogen tank were collected and calculated separately [54]. For trains, data on rail construction were separately collected and included in the calculations, distinguishing the 3-year lifetime of the train body from the 60-year lifetime of the railway [56].

Finally, J7 took into account the materials and inputs needed for manufacturing, operating, and maintaining hydrogen refueling stations and BoP (balance of plant) infrastructure. The BoP included components and systems like the hydrogen supply system, air supply system, thermal management system, and control system, all designed to handle 700 bar of hydrogen pressure. Additionally, it was assumed that the 0.042 kWh/kg H₂ of electricity consumed during the operation of hydrogen refueling stations would be supplied by the power sources defined in J1 [54].

4. Results

4.1. GHG Emissions According to Power Generation Sources and Water Electrolyzers

Table 6. GHG emission resulting from stage J1 according to power sources (kg ${\mathrm{CO}}_2$e/kWh).

Power Source	Non-Energy Materials	Electricity (Operational)	Other Energy	Generation Fuel	Shipping (Fuel)	Emission
Oil	0.006	0.051	0.003	0.215	0.003	0.278
Coal	0.004	0.022	2.95 $\times {10}^{-6}$	0.275	0.003	0.303
Coal-50	0.004	0.022	2.95 $\times {10}^{-6}$	0.215	0.003	0.243
Coal-100	0.004	0.022	2.95 $\times {10}^{-6}$	0.142	0.003	0.171
Coal-CCS	0.046	0.022	2.95 $\times {10}^{-6}$	0.026	0.003	0.097
IGCC	0.025	0.002	4.23 $\times {10}^{-6}$	0.187	0.006	0.220
NGCC	0.050	0.017	1.03 $\times {10}^{-5}$	0.165	1.15 $\times {10}^{-5}$	0.232
NGCC-CCS	0.054	0.017	1.03 $\times {10}^{-5}$	0.007	1.15 $\times {10}^{-5}$	0.079
WtE	0.072	0.001	5.71 $\times {10}^{-6}$	0	0	0.073
Biomass	0.051	0.011	0.001	0	0	0.064
PWR	0.003	0.001	0.002	0.013	2.36 $\times {10}^{-7}$	0.019
PV	0.014	0.0009	1.14 $\times {10}^{-7}$	-	-	0.015
CIGS	0.010	0.0002	1.14 $\times {10}^{-7}$	-	-	0.010
On-wind	0.001	3.80 $\times {10}^{-8}$	0	-	-	0.001
Off-wind	0.010	5.99 $\times {10}^{-8}$	0	-	-	0.010
Hydro	5.35 $\times {10}^{-5}$	1.71 $\times {10}^{-7}$	5.52 $\times {10}^{-12}$	-	-	5.37 $\times {10}^{-5}$

provides the resulting GHG emissions per kilowatt-hour (kWh) in J1 for the different types of power plants, categorized by energy and material consumptions. The other energy category includes fossil fuels such as diesel, gas, and thermal energy that are not used as primary fuel [57]. A dash (-) in Table 6 indicates that there were no input data available for the specific category. The results show that renewable or carbon-free generation technologies such as PWR, PV, CIGS, on-wind, off-wind, and hydro generation produced significantly fewer GHG emissions, up to twenty times less, than conventional fossil fuel-based generation methods. WtE and biomass sources also showed great potential. Advanced fossil-based generations such as coal-50, coal-100, and IGCC were less effective at reducing GHG emissions compared to CCS technologies (such as coal-CCS and NGCC-CCS). The lowest emissions were achieved by CIGS solar cells and hydro dams, signaling a major trend toward a net-zero emission society in electricity generation.

Table 7. Total GHG emissions resulting from the hydrogen supply chain across stages J1 through J7 (kg ${\mathrm{CO}}_2$e/kg ${\mathrm{H}}_2$).

	AEL				PEMEC						SOEC
	[47]	[48]	[49]	[18]	[47]	[50]	[49]	[34] 1	[34] 2	[39]	[50]
Oil	16.199	16.665	17.926	16.213	16.478	18.571	16.221	17.867	17.857	17.350	20.816
Coal	17.632	18.124	19.372	17.646	17.936	20.222	17.655	19.454	19.444	18.886	22.660
Coal-50	13.933	14.359	15.639	13.946	14.171	15.960	13.954	15.357	15.347	14.921	17.900
Coal-100	10.234	10.593	11.906	10.247	10.405	11.697	10.253	11.260	11.250	10.957	13.140
Coal-CCS	6.094	6.379	7.729	6.107	6.191	6.927	6.112	6.675	6.665	6.520	7.814
IGCC	12.983	13.392	14.681	12.997	13.204	14.865	13.004	14.305	14.295	13.903	16.678
NGCC	13.622	14.042	15.325	13.635	13.854	15.601	13.643	15.012	15.002	14.587	17.500
NGCC-CCS	5.098	5.365	6.723	5.111	5.177	5.779	5.116	5.572	5.562	5.453	6.532
WtE	4.759	5.020	6.381	4.772	4.832	5.388	4.776	5.196	5.186	5.089	6.096
Biomass	4.250	4.502	5.868	4.264	4.315	4.802	4.268	4.633	4.623	4.544	5.441
PWR	1.756	1.963	3.351	1.769	1.776	1.928	1.773	1.871	1.861	1.871	2.232
PV	1.518	1.721	3.111	1.531	1.533	1.654	1.534	1.607	1.597	1.616	1.926
CIGS	1.267	1.465	2.857	1.280	1.277	1.364	1.283	1.328	1.318	1.347	1.602
On-wind	0.742	0.931	2.328	0.755	0.744	0.760	0.758	0.748	0.738	0.785	0.927
Off-wind	1.273	1.471	2.863	1.286	1.283	1.371	1.289	1.335	1.325	1.353	1.610
Hydro	0.699	0.888	2.285	0.713	0.700	0.710	0.715	0.700	0.690	0.739	0.872

¹ and ² Based on data from a PEMEC electrolyzer manufactured in 2017 and projected for 2050, respectively.

summarizes the well-to-pump GHG emissions associated with the hydrogen supply chain (from J1 to J7) analyzed in this study. The table lists three different types of electrolyzers, along with the referenced studies. The columns correspond to the power sources used to supply electricity for hydrogen production across the various stages: electrolysis (J2), compressors (J3), liquefaction (J5), and refueling station operation and maintenance (J7). The values in Table 7 are provided in kg CO₂e/kg H₂, and Figure 2 presents a bar chart of the same data to easily capture the differences in GHG emissions. For AEL and PEMEC, error bars indicate the variability in results based on differing electrolyzer inventory data from multiple studies, while SOEC lacks error bars due to limited data availability.

When we examine the total GHG emissions presented in Table 7 and Figure 2, we can see that the power sources hydro, on-wind, CIGS, off-wind, PV, PWR (in ascending order) can produce hydrogen with GHG emissions of less than 4 kg CO₂e/kg H₂ for any type of electrolyzer. This level of emissions qualifies as 4th-grade clean hydrogen, even when a well-to-pump accounting is considered [4].

Table 8. Classification of power sources by Korean clean hydrogen certification tiers. Both well-to-pump (from stages J1 to J7) and well-to-gate (from stages J1 and J2) approaches were applied.

Class	Range (kg CO2e/kg ${\mathbf{H}}_2$)	Well-to-Pump (J1–J7)	Well-to-Gate (J1–J2)
Out of tiers	>4.00	Oil, Coal, Coal-CCS, Coal-50 Coal-100, IGCC, NGCC NGCC-CCS, WtE, Biomass	Oil, Coal, Coal-CCS, Coal-50 Coal-100, IGCC, NGCC
Tier 4	2.01–4.00	-	NGCC-CCS 7, WtE 8, Biomass 9
Tier 3	1.01–2.00	PWR 1, PV 4, CIGS 5, Off-wind 3	-
Tier 2	0.11–1.00	Hydro 6, On-wind 2	PWR 10, Off-wind 11, PV 12, CIGS 13
Tier 1	0.00–0.10	-	On-wind 14, Hydro 15

¹ In two cases, AEL and SOEC, PWR fell into Tier 4. ² Using AEL specified in [49], on-wind fell into Tier 4. ³ Using AEL specified in [49], off-wind fell into Tier 4. ⁴ Using AEL specified in [49], PV fell into Tier 4. ⁵ Using AEL specified in [49], CIGS fell into Tier 4. ⁶ Using AEL specified in [49], hydro dams fell into Tier 4. ⁷ Using AEL in two cases, PEMEC in four cases, and SOEC, NGCC-CCS was outside of the tiers. ⁸ Using AEL in one case, PEMEC in three cases, and SOEC, WtE is outside of the tiers. ⁹ In two cases, AEL and SOEC, biomass was outside of the tiers. ¹⁰ Using AEL in two cases, PEMEC in four cases, and SOEC, PWR fell into Tier 3. ¹¹ Using the AEL specified in [49], off-wind fell into Tier 4. ¹² In two cases, AEL and SOEC, PV fell into Tiers 3 or 4. ¹³ Using the AEL specified in [49], CIGS fell into Tier 4. ¹⁴ For two cases of AEL and SOEC, on-wind fell into Tiers 2 or 3. ¹⁵ For two cases of AEL and SOEC, hydro fell into Tier 2 or 3.

categorizes the power sources based on the Korean clean hydrogen certification standard [85]. In terms of well-to-pump GHG emissions, the cleanest power sources were hydro dams and onshore wind turbines, which fall into the second tier. The classification depends on the type of electrolyzer used. When AEL specified in [52] were utilized, emissions increased to 2.285 and 2.328 kg CO₂e/kg H₂, respectively, placing them in the third tier. Without the application of AEL in [52], CIGS, offshore wind, PV, and PWR can also be categorized as third-tier power sources.

There was a notable difference between onshore and offshore wind systems due to the larger structures, durability requirements, and foundations needed for offshore wind turbines. Offshore wind consumed approximately 0.002 kg/kWh more steel and additional materials, such as epoxy (3.56 $\times {10}^{-5}$ kg/kWh), ferrosilicon (1.80 $\times {10}^{-5}$ kg/kWh), magnets (2.43 $\times {10}^{-6}$ kg/kWh), natural rubber (3.08 $\times {10}^{-6}$ kg/kWh), asphalt (2.49 $\times {10}^{-6}$ kg/kWh), paint (6.45 $\times {10}^{-7}$ kg/kWh), and silicates (2.16 $\times {10}^{-3}$ kg/kWh), which are not used in onshore Wind systems [49]. With offshore wind turbines, the portion of J1 and J2 accounted for 47% of the total emissions specified in Table 7.

For PWR, the well-to-pump GHG emissions were between PVs and biomass. It is important to note that the potential environmental impact does not include the treatment of high-level nuclear waste [68]. Similarly, for hydroelectric power (the lowest carbon technology with onshore wind), GHG emissions related to the transportation of auxiliary materials like water, power, lubricating oil, hydraulic oil, and insulating oil used during plant operation were not taken into account [57]. As a result, the figures for hydroelectric power are tentative and exclude GHG emissions from large-scale civil works such as dam construction and deforestation for channels and power plants. If long-term effects, such as biomass decay in dam reservoirs, are considered, the environmental impact of hydroelectric power could potentially exceed that of wind power. The study in [36] supported this argument by stating that total GHG emissions from hydrogen production via electrolysis using hydroelectric power amount to 1.85 kg CO₂e/kg H₂, nearly double that of hydrogen production using wind power (i.e., 0.97 kg CO₂e/kg H₂).

The variation in GHG emissions across the columns in Table 7 was primarily driven by the efficiency of each electrolyzer (given in Table 3) and the type and quantity of raw materials used in their production. Higher efficiency means less electricity is needed to produce the same amount of hydrogen, leading to lower emissions. However, despite SOEC’s high efficiency of 99%, it emits significant carbon dioxide, since it needs the greatest amount of electrical energy at 65 kWh/kg H₂. This is because SOEC operates under high-temperature conditions, requiring both electrical and thermal energy, and additional energy for system maintenance [86]. The data in Table 3 indicate an electrical energy consumption for AEL ranging from 49 to 50 kWh/kg H₂. However, the results in Table 7 reveal that uncertainty was primarily attributed to the consumption of materials such as iron and nickel during manufacturing. For example, Sundin (2019, [48]) reported the use of pure iron (1.93 $\times {10}^{-2}$ kg/kg H₂), nickel (1.83 $\times {10}^{-3}$ 1.93 $\times {10}^{-2}$ kg/kg H₂), and potassium hydroxide (1.90 $\times {10}^{-2}$ 1.93 $\times {10}^{-2}$ kg/kg H₂), while Zhang (2023, [49]) identified significant amounts of nickel (2.02 $\times {10}^{-2}$ 1.93 $\times {10}^{-2}$ kg/kg H₂), iron (1.84 $\times {10}^{-1}$ 1.93 $\times {10}^{-2}$ kg/kg H₂), and platinum (1.61 $\times {10}^{-5}$ 1.93 $\times {10}^{-2}$ kg/kg H₂), all of which had a substantial impact on the emissions results. On the other hand, PEMEC data from multiple sources reported an electrical energy consumption between 49 and 57.5 kWh/kg H₂. The primary factors contributing to uncertain GHG emissions in PEMEC were the use of platinum as a catalyst, and non-energy materials such as titanium and platinum used in its construction. Lundberg (2019, [50]) reported the consumption of titanium (5.98 $\times {10}^{-4}$ kg/kg H₂) and platinum (8.49 $\times {10}^{-8}$ kg/kg H₂), while Zhang (2023, [49]) highlighted the use of steel (7.40 $\times {10}^{-3}$ kg/kg H₂), titanium (5.83 $\times {10}^{-4}$ kg/kg H₂), and copper (3.31 $\times {10}^{-4}$ kg/kg H₂). Additionally, Wulf (2018, [47]) noted that steel (4.28 $\times {10}^{-3}$ kg/kg H₂) and water (1.90 $\times {10}^1$ kg/kg H₂) are vital contributors to the uncertainty in PEMEC’s GHG emissions.

Table 9. GHG Emissions resulting from electrolysis through stages J1 and J2 (kg ${\mathrm{CO}}_2$e/kg ${\mathrm{H}}_2$).

Electrolyzer	Ref.	Electricity for Water Electrolysis	Electricity (Operational)	Non-Energy Materials	Catalyst	Water	Other Energy	Emission
AEL	[47]	0.003–14.871	0	0.007	0.002	0.005	0	0.017–14.885
	[48]	0.003–15.175	0.0005	0.087	0.037	0.008	0.069	0.205–15.377
	[49]	0.003–15.026	0	1.568	0.028	0	0.002	1.601–16.624
	[18]	0.003–14.871	0	0.015	0.002	0.009	0	0.029–14.897
PEMEC	[47]	0.003–15.175	0	0.008	0.0003	0.005	0	0.016–15.188
	[50]	0.003–17.451	0.003	0.011	0.001	0.008	0	0.026–17.474
	[49]	0.003–14.877	0	0.028	0.001	0	0	0.032–14.906
	[34] 1	0.003–16.692	0	0.008	0.003	0.002	0	0.016–16.705
	[34] 2	0.003–16.692	0	0.001	0.0002	0.002	0	0.006–16.695
	[39]	0.003–16.085	0	0.049	0	0.002	0	0.054–16.136
SOEC	[50]	0.003–19.727	0.083	0.095	0	0.007	0	0.188–19.912

¹ and ² Based on data from a PEMEC electrolyzer manufactured in 2017 and projected for 2050.

further breaks down the GHG emissions from the electrolysis stage (J2) from consuming raw materials, catalysts, electricity, and water. The overall results in Table 9 were significantly influenced by the amount of electricity consumed during electrolysis, which is given in Table 3. When comparing GHG emissions by catalysts in Table 9, PEMEC usually produced fewer emissions than AEL. Despite iridium (24.5 kgCO₂e/kg) and platinum (40,000 kgCO₂e/kg) having significantly higher carbon footprints as catalysts in PEMEC than the sodium hydroxide (1.94 kgCO₂e/kg) used in AEL, PEMEC requires 20,000 times less iridium and 200,000 times less platinum than AEL requires sodium hydroxide [18,34,39,47,48,49,50]. As a result, the overall GHG emissions from PEMEC were lower than those from AEL when considering the quantities used.

In Table 9, the AEL specifications from [49] are an outlier, especially in its high consumption of non-energy materials. Iron and nickel, which are inputs in the electrolytic cell manufacturing in AEL, are thus the primary contributors to GHG emissions in water electrolysis. In [49], the AEL assumed the consumption a larger amount of nickel at 0.0202 kg/kg H₂, iron at 0.184 kg/kg H₂, and platinum at 0.0000161 kg/kg H₂ compared to other literature. This high usage of nickel and iron led to considerable GHG emissions. Although the input of platinum was smaller compared to that of nickel and iron, its impact was substantial due to its high emission intensity of carbon emissions, with 12,500 kg CO₂e/kg H₂.

The total GHG emissions in J1 and J2 constituted well-to-gate accounting. As summarized in Table 8, hydro dams and onshore wind can be classified as first-tier methods, depending on the type of electrolysis used. All the PEMEC electrolyzers considered in this study were in the first tier, while the AEL in [49] and SOEC placed them in the poorer clean-hydrogen tiers. With most electrolyzer types, offshore wind turbines, CIGS, and PV solar cells fell into the second tier. PWR only became a second-tier source with one PEMEC electrolyzer among the considered types in this study. With well-to-gate accounting, NGCC-CCS, WtE, and biomass could be classified as fourth-tier clean methods. However, these methods were excluded if a well-to-pump approach was used.

Table 10. GHG emissions resulting from stages J3 to J7 (kg ${\mathrm{CO}}_2$e/kg ${\mathrm{H}}_2$).

Stage	Electricity	Non-Energy Materials	Other Energy	Transportation (Manufacturing + Fuel)	Emission
J3	3.78 $\times {10}^{-8}$–0.0002	0.005	0	0	0.005
J4	0	0.181	0.0006	0.0004	0.182
J5	0.0004–2.052	0.141	0	3.48 $\times {10}^{-5}$	0.141–2.160
J6	0.035	0.008	0	0.0005	0.044
J7	0.0009–0.014	0.274	0.017	0.020	0.312–0.325

further details the GHG emissions from stages J3 to J7, broken down by consumption of raw materials, electricity, diesel, thermal energy, and transportation. The stages contributed to GHG emissions (roughly) in the following order: J7 hydrogen refueling stations, J4 pipelines, J5 liquefaction plants and storage tanks, J6 rail and transport vehicles, and J3 compressors. The GHG emission in the other energies column in J4 was from diesel (8.82 $\times {10}^{-3}$ kg/kg H₂), and those in J7 were from diesel (7.00 $\times {10}^{-4}$ kg/kg H₂) along with thermal energy consumption (1.55 $\times {10}^{-1}$ MJ/kg H₂). The transportation column accounts for GHG emissions from the production of transport vehicles and the fuel used during the assembly of components for each facility. J6 included 0.0005 kg CO₂e/kg H₂ of emissions from fuel consumed in transporting liquefied hydrogen to various demand points.

The primary contributors to the well-to-pump GHG emissions depended on the power sources. According to Table 5, Table 8 and Table 9, as well as Figure 2, electricity consumption during electrolysis (J2) and the construction of power plants (J1) significantly contributed to emissions from most fossil fuel-based sources. On average, GHG emissions from J1 and J2 accounted for approximately 80–85% of emissions from various coal plants and 78–85% from NGCC plants. In contrast, these emissions were only 13–18% for hydro dams and onshore wind turbines. For technologies such as PV, CIGS, offshore wind, and PWR, J1 and J2 contributed roughly 47–59% of total emissions. For technologies with lower emission proportions from J1 and J2, the method of hydrogen transport becomes increasingly important when considering future accounting of well-to-pump emissions.

4.2. GHG Emissions from Grid-Connected Hydrogen Production in Korea

The energy mix scenarios presented in

Table 11. Energy mix scenarios in the Korean power sector (%).

Power Source	S1	S2	S3	S4	S5
Oil	0.3	0.1	0.1	0.1	0.1
Coal-50	15.5	9.9	7.7	0.2	0.3
Coal-100	-	1.5	1.2	5.8	5.3
IGCC	-	1.5	1.2	5.8	5.3
NGCC	5.0	-	-	-	-
NGCC-CCS	5.0	25.8	22.9	18.5	13.8
WtE	4.0	-	-	-	-
Biomass	-	0.9	0.7	0.8	0.8
PWR	37.3	18.3	14.7	14.2	13.3
PV	5.8	7.4	9.1	9.7	10.9
CIGS	13.6	17.1	21.2	22.4	25.1
On-wind	5.0	6.4	7.9	8.4	9.4
Off-wind	8.0	10.1	12.5	13.2	14.8
Hydro	0.5	1.0	0.8	0.9	0.9
Year	2036	2040 EEI	2040 EOE	2050 EEI	2050 EOE
Reference	[87]	[88]

were used in this study to observe how GHG emissions from grid-connected hydrogen production vary under different conditions. These scenarios were based on the 10th basic plan for long-term electricity supply and demand (by MOTIE [87]) and the 2022 long-term energy outlook from the Korea Enegy Economics Institute (KEEI) [88]. The energy mix scenarios from these sources were recaptured to align with the power plants investigated (at stage J1) in this study, excluding the self-generators. Moreover, KEEI’s scenarios contain two policies to achieve a low-carbon economy: energy efficiency improvement (EEI) and electrification of end-use (EOE). EEI refers to a strategy that enhances the energy efficiency of existing technologies without substituting fuels, such as advancements in energy-efficient technologies across industrial machinery, and the use of scrap steel in integrated steel-making processes. Conversely, the EOE scenario involved transitioning from technologies that rely on fossil fuels to electricity from low- or zero-emission energy sources.

For coal power, it was assumed that 50% of the fuel would be replaced with ammonia in 2026, and that no carbon capture and storage (CCS) technology would be applied. According to KEEI’s projections, it was assumed that only NGCC plants would be equipped with CCS. The installation rate for CCS was projected to reach 50% by 2036 and 100% by 2040 and 2050 [88]. In KEEI’s energy mix scenario for 2040 and 2050, 50% IGCC and 50% coal with 100% fuel replacement were also considered.

In the MOTIE’s scenario, renewable energy consisted of 59% solar, 40% wind, and 1% hydropower. In contrast, the KEEI scenario excluded hydropower from volatile renewable energy, with solar and wind power distributed at 60% and 40%, respectively. Based on the solar market research in [89], we assumed that PV was further divided into solar PV and CIGS.

Table 12. Well-to-pump (from stages J1 to J7) GHG emissions resulting from the energy mix scenarios. The numbers in parentheses represent the well-to-gate assessment, covering stages J1 and J2 (kg ${\mathrm{CO}}_2$e/kg ${\mathrm{H}}_2$).

	AEL				PEMEC						SOEC
	[47]	[48]	[49]	[18]	[47]	[50]	[49]	[34] 1	[34] 2	[39]	[50]
S1	0.733 (0.045)	0.922 (0.234)	2.319 (1.631)	0.828 (0.058)	0.734 (0.046)	0.749 (0.061)	0.753 (0.061)	0.738 (0.050)	0.728 (0.040)	0.775 (0.087)	0.916 (0.228)
S2	0.729 (0.041)	0.917 (0.230)	2.314 (1.627)	0.823 (0.054)	0.730 (0.042)	0.744 (0.056)	0.749 (0.057)	0.749 (0.045)	0.723 (0.035)	0.770 (0.083)	0.910 (0.222)
S3	0.724 (0.037)	0.912 (0.225)	2.309 (1.622)	0.818 (0.050)	0.725 (0.038)	0.738 (0.051)	0.744 (0.053)	0.727 (0.040)	0.717 (0.030)	0.765 (0.078)	0.904 (0.217)
S4	0.722 (0.035)	0.910 (0.224)	2.307 (1.621)	0.816 (0.048)	0.723 (0.036)	0.736 (0.049)	0.742 (0.051)	0.725 (0.038)	0.715 (0.028)	0.763 (0.076)	0.901 (0.214)
S5	0.719 (0.033)	0.908 (0.221)	2.305 (1.618)	0.814 (0.046)	0.720 (0.034)	0.733 (0.047)	0.739 (0.049)	0.722 (0.036)	0.712 (0.026)	0.760 (0.074)	0.898 (0.211)

¹ and ² Based on data from a PEMEC electrolyzer manufactured in 2017 and projected for 2050, respectively.

presents a projection of the well-to-pump GHG emissions from water electrolysis connected to the Korean power grid. The well-to-gate emissions are also provided in parentheses. Except for the AEL electrolyzer specified in [49], all electrolysis types enabled second-tier emissions for the projected years of 2036, 2040, and 2050. If applying well-to-gate accounting, the emissions dropped into the first-tier level for all six PEMEC types and two AEL types, regardless of the projected year. This suggests that, after 2036, when the Korean power sector will be significantly decarbonized, as projected in Table 11, grid-connected water electrolyzers will be capable of producing sufficiently clean hydrogen, even though it may not be labeled as green. The results in Table 12 also show that there was no a significant difference in GHG emissions across the various energy mix scenarios.

5. Conclusions

This study conducted an LCA of hydrogen production in Korea using water electrolysis, considering different energy sources across the hydrogen supply chain. The assessment revealed that hydroelectric and onshore wind power were the most environmentally favorable energy sources, achieving Tier 2 under a well-to-pump framework and Tier 1 under a well-to-gate framework. Photovoltaics, nuclear energy, and offshore wind also showed promise, though their classification depended on the electrolysis efficiency and energy input during construction.

The type of electrolyzer had a significant impact on GHG emissions. Higher efficiencies, particularly for PEMEC and SOEC electrolyzers, led to lower GHG outputs. As Korea advances toward a clean hydrogen economy, decarbonizing the power grid will be key to achieving low-emission hydrogen production. Future energy mixes in 2036, 2040, and 2050 project that grid electricity could reach Tier 2 certification, potentially meeting Tier 1 under well-to-gate assessments.

This study emphasizes the importance of selecting low-emission power sources [90] and enhancing electrolyzer efficiency to decrease the environmental impact of hydrogen production. This would support Korea’s goal of a sustainable hydrogen economy.