Life Cycle Assessment of Hydrogen Transportation Pathways via Pipelines and Truck Trailers: Implications as a Low Carbon Fuel

Abstract

Hydrogen fuel cells have the potential to play a significant role in the decarbonization of the transportation sector globally and especially in California, given the strong regulatory and policy focus. Nevertheless, numerous questions arise regarding the environmental impact of the hydrogen supply chain. Hydrogen is usually delivered on trucks in gaseous form but can also be transported via pipelines as gas or via trucks in liquid form. This study is a comparative attributional life cycle analysis of three hydrogen production methods alongside truck and pipeline transportation in gaseous form. Impacts assessed include global warming potential (GWP), nitrogen oxide, volatile organic compounds, and particulate matter 2.5 (PM_2.5). In terms of GWP, the truck transportation pathway is more energy and ecologically intensive than pipeline transportation, despite gaseous truck transport being more economical. A sensitivity analysis of pipeline transportation and life cycle inventories (LCI) attribution is included. Results are compared across multiple scenarios of the production and transportation pathways to discover the strongest candidates for minimizing the environmental footprint of hydrogen production and transportation. The results indicate the less ecologically intensive pathway is solar electrolysis through pipelines. For 1 percent pipeline attribution, the total CO_2eq produced per consuming 1 MJ of hydrogen in a fuel cell pickup truck along this pathway is 50.29 g.

1. Introduction

1.1. Background and Motivation

Hydrogen is gaining unprecedented momentum across both public and private sectors as a clean energy carrier around the world, driven by a growing focus on the decarbonization of different sectors of the economy. Measures to assist the adoption of Fuel-Cell Electric Vehicles (FCEV) have already been put in place in many countries around the world, including Japan, the European Union, as well as the state of California (CA). Governments, in collaboration with the private sector, are supporting the building of hydrogen infrastructure for early adopters across both the transportation and industrial sectors. In a recent study analyzing the feasibility of hydrogen in New Zealand, it was found that hydrogen could lead to a 25% reduction in the country’s energy system emissions in 2050 from the base year of 2018 [1]. With regard to transportation, FCEVs have a substantially greater energy conversion efficiency than gasoline-powered internal combustion engines (ICEVs) and have zero tailpipe emissions. Although FCEV operations do not produce any tailpipe emissions, it is important to investigate the life cycle emissions of such vehicles to better understand their impact on the environment. As we transition to Zero Emission Vehicles (ZEVs), which include Battery Electric Vehicles (BEVs) and FCEVs, the overall benefits of this transition will be best compared based on a life cycle assessment (LCA) approach. For example, the production of lithium-ion cells for BEVs could be compared to the production of fuel cells and the hydrogen tank in FCEVs. These could provide insights into maximizing the benefits of decarbonization.

In the US, about 10 million metric tons of hydrogen are produced annually, with petroleum processing (68%) and fertilizer production (21%) being the largest consumers [2]. The United States Department of Energy (US DOE) has set a goal to produce hydrogen at $2/kg by 2025 and $1/kg by 2030 to facilitate the transition toward net-zero carbon pathways [3].

Further, to maximize the decarbonization potential of hydrogen as a fuel, various efforts are ongoing in terms of exploring different production pathways, including carbon capture and storage, biomass-based production, and using renewable energy [4]. Although there are numerous studies on the LCA of hydrogen production, only a few investigate the LCA of hydrogen through various pathways. This paper aims to fill the knowledge gap by comparing the LCA of the two primary pathways for transferring hydrogen from a production plant to a refueling station and for consumption by an FCEV pickup truck.

1.2. Literature Review

Over the past few years, numerous countries have devoted considerable resources to the development and implementation of hydrogen energy technology. Different processes of hydrogen production are made up of a combination of feedstock and a primary energy source. Globally, 96% of hydrogen is produced through fossil fuel-based energy sources, whereas just 4% is produced through water electrolysis. Of all the fossil fuel pathways, steam methane reforming (SMR) has the largest share (80%). It uses natural gas as feedstock and fossil fuels as primary energy. SMR and the water-gas shift reaction (WGSR) are used to produce hydrogen from methane, ethanol, and methanol.

When steam reforming is performed, considerable carbon monoxide is also formed. Following these reactions, the WGSR is carried out to react with the carbon monoxide generated with steam, resulting in additional hydrogen gas and the conversion of carbon monoxide to carbon dioxide [5,6]. Coal gasification is the other method that uses fossil fuel to produce hydrogen. In this process, carbon monoxide, carbon dioxide, steam, and hydrogen are produced by reacting heated coal with a controlled amount of oxygen [7]. However, all fossil fuel pathways are associated with negative environmental impacts. In the SMR pathway, the rate of Greenhouse Gas (GHG) emission based on a life cycle basis is equal to 10 kg of CO₂ per 1 kg of H₂ [8]. According to the Environmental Protection Agency (EPA) of the United States, the SMR pathway produces 34 metric tons CO₂ annually, which is equivalent to the emissions of 7 million passenger vehicles [9]. Thus, there is a need for a method of producing hydrogen that is less harmful to the environment.

The earth’s surface absorbs approximately 120,000 TW of energy, which is greater than daily human consumption. Therefore, this energy could be considered a substitute for fossil fuel resources and used to produce hydrogen from water molecules [10]. The US DOE names water electrolysis as the most promising hydrogen production pathway for the mid-term horizon [3]. This production route entails the dissociation of water (H₂O) molecules into the constituent hydrogen and oxygen gases [11]. There are two primary methods for hydrogen distillation from water: photovoltaic electrolysis and concentrated solar power electrolysis [12,13].

Sadeghi et al. [14] compare the cost and emissions of photovoltaic solar-based hydrogen production with the SMR production pathway. Their analysis indicates that solar production is more expensive than SMR. However, this pathway has a fivefold lower global warming potential than the SMR. Zhang et al. [15] compared three different water electrolysis methods, including photovoltaic coupled with a polymer electrolyte membrane, photothermal coupled with a polymer electrolyte membrane, and thermochemical. They found that the photothermal pathway is less environmentally intensive.

Biomass-based technologies are an emerging new method for hydrogen production. Where there is a lack of water, biomass energy resources could be used instead of fossil fuels and solar electrolysis. The authors in [16,17] compared the biomass pathway to the natural gas hydrogen production pathway and found that the biomass pathway is less environmentally intensive and emits 75% less GHG per kilogram of hydrogen produced than natural gas.

This specific pathway includes two methods for producing hydrogen: biological and thermochemical. According to Balat et al. [18], the thermochemical pathway is more energy efficient than the biological pathway.

In the case of biomass gasification, the first stage of the life cycle is the collecting of Corn Stover. Corn stover is made up of the stalk, leaves, husks, and tassels left in the field after harvesting the grain with a combine. This stover can be used to make advanced biofuels or be used as a low-quality, emergency livestock feed [19]. Following the collection and selection of Corn Stover, it passes through the treatment, processing, and drying stages of the life cycle. The prepared biomass is then used to generate hydrogen gas in a biomass gasification plant. In addition, the hydrogen produced is compressed to high pressures appropriate for transit and storage.

Liquefaction or compression at a distribution terminal, transportation, and distribution are all part of the delivery pathway, as are compression, storage, precooling, and dispensing at a refueling station. Hydrogen is produced at low pressure (20 bar) via SMR and electrolysis processes, and it can be compressed before being transported via pipelines from the production plant to the distribution terminal [20].

Hydrogen is compressed or liquefied at the production facility so that it can be placed into compressed gaseous tube trailers or cryogenic-liquid tankers for delivery to refueling stations. Hydrogen can be compressed to a pressure between 200 and 500 bar and placed on gaseous tube trailers for transportation and distribution to filling stations [21]. The payload of a tube trailer is determined by the tube volume, number of tubes, and loading pressure and is limited by the 80,000-pound gross vehicle weight limit set by the US Department of Transportation [22,23].

Tube trailers are currently equipped to transport 300 to 1100 kg of compressed gaseous hydrogen, which can be discharged or switched with an “empty” tube trailer at the demand site [24]. At a refueling station, a tube trailer delivers hydrogen to a gaseous compressor, which compresses it to 1000 bar before storing it in a high-pressure buffer storage system for dispensing into vehicle tanks [25]. The dispenser uses a refrigeration mechanism to control the flow of hydrogen from the high-pressure buffer storage into the vehicle tank, preventing the FCEV tank from overheating [26].

Hydrogen can also be liquified. Given the higher density of liquid hydrogen, it allows for a greater payload when transported as a liquid compared to gas (4 metric tons vs. 1 metric ton) [27]. Liquid hydrogen delivery is generally more cost-effective than compressed hydrogen-gas delivery in tube trailers, especially over long distances [28]. However, the high energy intensity of the liquefaction process poses a barrier when viewed from a decarbonization perspective and would require a highly low carbon electricity grid [29].

Among the lowest cost options is the transportation of gaseous hydrogen through pipelines. This is done very similar to conventional natural gas pipelines as known today. Approximately 1600 miles of hydrogen pipelines are currently operating in the United States [19]. Typically, pipelines require a relatively high density of users in a region to be cost-effective, as they can supply large volumes of hydrogen fuel efficiently. In the US, these pipelines are located where large hydrogen users, such as petroleum refineries and chemical plants, are concentrated, such as in the Gulf Coast region [30].

Similar to the natural gas pipeline, hydrogen pipeline networks will have multiple distribution lines and network connections across regions to optimize the flow of hydrogen fuel to various end-use sectors, including heavy transport, export production, and industrial use, with the ability to feed the surplus hydrogen produced into gas networks [1].

As the upfront costs of new pipeline construction are a key barrier, there has been discussion around using existing natural gas pipelines to transport hydrogen via blending. It is suggested that up to 20% hydrogen concentrations by volume may be the maximum blend before significant pipeline upgrades are required [31]. As we will find in this paper, even though pipeline-based pathways for hydrogen are the least carbon-intensive, they will require certainty in future demand for commercial viability [32].

Ahmadi and Kjeang [33] consider three alternative methods of hydrogen production: electrolysis, thermochemical water splitting, and steam methane reforming of natural gas for use in an FCEV against the reference case of conventional gasoline vehicles. They further compared the results for four different provinces in Canada with different energy resource mixes. The study showed a consistent GHG emission reduction of up to 90% compared to gasoline for all three hydrogen production methods in all four Canadian provinces. The only exception was the electrolysis in Alberta, which had more emissions than the reference case scenario because of the coal and natural gas-dominated thermal power generation. In Ontario, thermochemical hydrogen production was the best scenario as a large amount of high-temperature waste heat from nuclear power plants could be used as an energy input for hydrogen production. Additionally, in provinces with major hydroelectric capacity, like Quebec and British Columbia, electrolysis provides a low-emission pathway for hydrogen production.

In the literature reviewed, the chain of operations covering packaging, transmission, distribution (T&D), and fueling are referred to as delivery. Various methods of increasing the energy density of hydrogen to enable distribution, such as liquefaction or compression of gaseous hydrogen, are referred to as packaging. Chemical transformations, such as those to and from hydrogen carriers like ammonia or methanol, can also be employed for packaging. However, current research is mostly focused on compression to gas or liquefaction as packaging methods. For example, Ahmadi and Kjeang [33] also showed that T&D emissions are modest compared to NG-SMR emissions. However, they suggested that as the carbon intensity (CI) of upcoming hydrogen production technologies drops, T&D emissions become more relevant, and may even account for most life-cycle carbon emissions.

Lee et al. provide evidence for the effect of liquefaction and compression at refueling stations on Well-to-Wheel (WTW) GHG emissions from hydrogen production [34]. They emphasized that the electricity mix utilized for hydrogen compression and/or liquefaction for LCA is critical as it can significantly change the WTW GHG emissions depending on the grid mix that is chosen for determining LCA. They concluded that in the WTW analysis of SMR-produced gaseous hydrogen delivered by truck, GHG emissions were reduced by 20% to 45% and NO_x by 19% to 43% depending on vehicle weight classes and driving patterns. However, they noted that compared to gaseous hydrogen, liquefied hydrogen fuels result in lower WTW emission reductions. Furthermore, for both gaseous and liquefied hydrogen pathways, because of electricity consumption for compression and liquefaction, spatial-temporal variations in electricity generation can affect the WTW results.

Frank et al. have built on previous work by adding pipelines to the delivery methods and making comparisons between the transmission of gaseous hydrogen via pipeline or tube trailers and liquified hydrogen via cryogenic trucks [35]. The results suggest that the pure-pipeline delivery scenario is insensitive to the transmission distance since comparatively modest power is needed to operate the pipeline compressors. Furthermore, they demonstrated that at a transmission distance of 500 km, GHG emissions from liquefaction would have to be 40% of those in their model to be competitive with the transmission of gaseous hydrogen with tube trailers. Given the same mass of hydrogen in gaseous and liquid forms has different amounts of energy, our study utilized a functional unit of 1 MJ of hydrogen to overcome this limitation. In addition, the authors in [35] do not take into account the LCA of pipeline construction and the assembly, disposal, and recycling of trucks and trailers, which may result in misleading results.

Sinha et al. [36] analyzed the lifecycle assessment of renewable hydrogen FCEVs in the northern and southern parts of California. Although they consider transportation pathways for short distances in their analysis, it is just limited to the compression factor prior to the loading on tube trailers. California has other resources for producing hydrogen, like SMR and biomass production sites, which are far away from the consumption stations.

1.3. Paper Contribution

This paper investigates the life cycle impact assessment of gaseous tube trucks and pipeline transportation pathways. This study aims to evaluate the life cycle impacts of hydrogen as an alternative fuel using two primary pathways. This article considers the life cycle assessment of pipeline production and construction. Additionally, the production of fuel, trailers, tubes, and vehicles, as well as the assembly, disposal, and recycling of vehicles and trailers are all part of the gaseous hydrogen delivery pathway by truck.

1.4. Paper Organization

The rest of the paper is structured as follows: Section 2 provides a concise description of the dataset and a detailed explanation of the method proposed in this paper. In Section 3, life cycle analysis results and sensitivity analysis are presented in detail. Section 4 is dedicated to a discussion of the outcome. Finally, Section 5 concludes this paper.

2. Data and Methods

2.1. Methodology

This section covers the goal, scope, and system boundary of the LCA analysis.

The target audience for this analysis includes policymakers such as the California Air Resources Board (CARB) and the US DOE, as well as industry and academia.

The scope of this LCA is well-to-wheel with a functional unit of 1 MJ of hydrogen consumed by an FCEV pick-up truck. Here, the “well” is defined as the hydrogen production plant. The foreground system includes the transportation of hydrogen fuel, either as a tube trailer or via pipeline from the production plant to a hydrogen fueling station dispenser, and final consumption by an FCEV pickup truck. Different production processes and transportation pathways are considered in this study, which are explained in the next section.

Whether produced by SMR, electrolysis, or biomass, hydrogen is in a gaseous form at the end of the production stage. Prior to delivering hydrogen fuel, whether by truck or via pipeline, it needs to be compressed. However, this part of the process is not included in our system boundary.

For this system, the hydrogen production plant is located in Los Angeles, CA, and three different production processes using SMR, electrolysis with solar energy, and biomass gasification form the different scenarios. The Los Angeles (LA) production plant was selected after discussions with experts in the field. The SMR, electrolysis with solar and biomass hydrogen production, is considered part of the system, and reference life cycle inventories (LCIs) represent this process. An 18-wheeler truck chassis is assumed for the truck-based transportation pathway. The fuel station is modeled after an existing hydrogen fuel station in East Sacramento, California. The pipeline infrastructure transfers the hydrogen from the production plant to the fuel stations in the alternate transportation pathway. The transportation distance for the pipeline and truck pathways will be from LA to Sacramento, which is about 400 miles. It is assumed the truck trailer is empty on the return trip. Thus, the distance covered by the truck is assumed to be 800 miles, because of the round trip.

As this system focuses specifically on the production, transportation, and consumption of hydrogen fuel, simplifications of the broader system were made. First, we assume that some processes would be identical regardless of the transportation form used. For example, we do not consider the different pressures of hydrogen fuel for pipeline or truck transport. Due to the GREET database limitation, the system boundary does not include the construction, maintenance, and operation of both the hydrogen production plant and fuel station. The final fueling from the pump to the end-user is considered within the system boundary. It is assumed that pipeline infrastructure is not constructed for the foreground system, and reference LCIs for pipeline production and construction are included in the system boundary.

It is assumed that the truck-trailer lifetime is 540,000 miles with a fuel efficiency of 10.5 mpg. It is also assumed that the California electricity mix (CAMX) is used to produce the hydrogen tubes for the truck trailer. The system boundary diagram is shown in Figure 1.

2.2. Scenario Development

A total of six different scenarios are included in the LCA analysis, which are a combination of three different hydrogen production processes and two different transportation pathways. We use the Greenhouse gases, Regulated Emissions, and Energy use in Technologies (GREET) Model [37] 2021^® LCI tool for obtaining all the reference LCIs. The GREET model is selected because a majority of policymakers in California and the United States rely on it as an LCA tool.

The hydrogen produced is kept in a steel tube and loaded onto a truck trailer in the truck pathway. In this case, the GREET database does not provide any LCI for steel tube production for hydrogen storage, hence we assume the reference LCI of hot-rolled steel pipe production (from GREET), with suitable weight assumptions of the tube. Carbon and low-alloy steels are commonly used in the production of high-pressure hydrogen gas tubes and pipelines [38]. In the case of the pipeline as well, the GREET model did not provide a separate LCI for pipeline production, although construction of the pipeline is included in the GREET model. Thus, we again assume the reference LCI from GREET for hot-rolled steel pipe production for the pipeline production. This assumption was made to include and determine the impact of tube and pipeline production on the hydrogen transportation pathway. In the base case, we attribute 1% of the pipeline LCA to the functional unit. The compressor’s LCI was not included in the scope of the analysis and is therefore outside the system boundary. Since both pathways utilize similar compression prior to loading, and this stage is similar among them, the compression is excluded from the pathways in order to simplify the calculation. For the end-use FCEV, we assume a pick-up truck with a 20 MPG_e fuel efficiency.

3. Results

All three pipeline pathways emit less CO_2e across the same hydrogen production process, according to the data. A solar plant with a pipeline produces less CO₂ equivalent than any of the six scenarios (three production and two transit). In addition, among all the scenarios, SMR with trucks produces the highest GHG emissions. The complete LCA results are presented in

Table A1. Life cycle analysis per 1 MJ of H₂ consumed from Natural Gas (SMR)-Truck.

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Truck trailer Production	Fuel Production	Fuel Consumption	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	1301.328	95.789	1.127	0.000	20.796	1419.039	kJ
	Total Renewable Energy	0.335	7.731	1.614	0.000	0.121	9.801	kJ
	Total water consumption	88.943	29.568	24.705	0.000	507.316	650.533	cm3
Outputs	CO2 Total	0.073	0.006	0.004	0.034	0.046	0.165	kg
	CO2	0.073	0.006	0.004	0.034	0.047	0.165	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	0.000	0.000	kg
	VOC	0.010	0.006	0.002	0.001	0.026	0.046	g
	CO	0.015	0.022	0.004	0.070	0.150	0.260	g
	NOx	0.019	0.007	0.010	0.043	0.048	0.127	g
	PM10	0.003	0.003	0.001	0.000	0.023	0.030	g
	PM2.5	0.003	0.002	0.001	0.000	0.011	0.016	g
	SOx	0.012	0.019	0.005	0.000	0.337	0.373	g
	CH4	0.178	0.015	0.031	0.000	0.119	0.343	g
	N2O	0.001	0.000	0.000	0.000	0.001	0.002	g
	CF4	0.000	0.017	0.000	0.000	0.000	0.017	g
	SF6	0.000	0.000	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.002	0.000	0.000	0.000	0.002	g
	CO2e	78.934	6.952	5.293	34.189	50.135	175.504	g

,

Table A2. Life cycle analysis per 1 MJ of H₂ consumed from solar plant-Truck.

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Truck trailer Production	Fuel Production	Fuel Consumption	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	0.000	95.789	1.127	0.000	20.796	117.711	kJ
	Total Renewable Energy	1351.000	7.731	1.614	0.000	0.121	1360.466	kJ
	Total water consumption	91.491	29.568	24.705	0.000	507.316	653.080	cm3
Outputs	CO2 Total	0.000	0.006	0.004	0.034	0.046	0.091	kg
	CO2	0.000	0.006	0.004	0.034	0.047	0.092	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	0.000	0.000	kg
	VOC	0.000	0.006	0.002	0.001	0.026	0.036	g
	CO	0.000	0.022	0.004	0.070	0.150	0.245	g
	NOx	0.000	0.007	0.010	0.043	0.048	0.108	g
	PM10	0.000	0.003	0.001	0.000	0.023	0.027	g
	PM2.5	0.000	0.002	0.001	0.000	0.011	0.013	g
	SOx	0.000	0.019	0.005	0.000	0.337	0.360	g
	CH4	0.000	0.015	0.031	0.000	0.119	0.166	g
	N2O	0.000	0.000	0.000	0.000	0.001	0.001	g
	CF4	0.000	0.017	0.000	0.000	0.000	0.017	g
	SF6	0.000	0.000	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.002	0.000	0.000	0.000	0.002	g
	CO2e	0.000	6.952	5.293	34.189	50.135	96.570	g

,

Table A3. Life cycle analysis per 1 MJ of H₂ consumed from biomass plant-Truck.

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Truck trailer Production	Fuel Production	Fuel Consumption	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	149.837	95.789	204.126	0.000	20.796	470.547	kJ
	Total Renewable Energy	2231.000	7.731	1.002	0.000	0.121	2239.854	kJ
	Total water consumption	134.304	29.568	15.334	0.000	507.316	686.523	cm3
Outputs	CO2 Total	0.010	0.006	0.004	0.034	0.046	0.102	kg
	CO2	0.011	0.006	0.004	0.034	0.047	0.102	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	0.000	0.000	kg
	VOC	0.004	0.006	0.002	0.001	0.026	0.040	g
	CO	0.018	0.022	0.004	0.070	0.150	0.263	g
	NOx	0.030	0.007	0.010	0.043	0.048	0.137	g
	PM10	0.002	0.003	0.001	0.000	0.023	0.029	g
	PM2.5	0.002	0.002	0.001	0.000	0.011	0.015	g
	SOx	0.020	0.019	0.005	0.000	0.337	0.380	g
	CH4	0.019	0.015	0.031	0.000	0.119	0.185	g
	N2O	0.006	0.000	0.000	0.000	0.001	0.008	g
	CF4	0.000	0.017	0.000	0.000	0.000	0.017	g
	SF6	0.000	0.000	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.002	0.000	0.000	0.000	0.002	g
	CO2e	12.740	6.952	5.293	34.189	50.135	109.310	g

,

Table A4. Life cycle analysis per 1 MJ of H₂ consumed from Natural Gas (SMR)—Pipeline (1% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	1301.328	1.922	20.796	1324.046	kJ
	Total Renewable Energy	0.335	0.094	0.121	0.550	kJ
	Total water consumption	88.943	0.305	507.316	596.564	cm3
Outputs	CO2 Total	0.073	0.000	0.046	0.120	kg
	CO2	0.073	0.000	0.047	0.120	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.010	0.000	0.026	0.036	g
	CO	0.015	0.001	0.150	0.166	g
	NOx	0.019	0.000	0.048	0.067	g
	PM10	0.003	0.000	0.023	0.026	g
	PM2.5	0.003	0.000	0.011	0.014	g
	SOx	0.012	0.001	0.337	0.350	g
	CH4	0.178	0.000	0.119	0.297	g
	N2O	0.001	0.000	0.001	0.002	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	78.934	0.163	50.135	129.233	g

,

Table A5. Life cycle analysis per 1 MJ of H₂ consumed from solar plant—Pipeline (1% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	0.000	1.922	20.796	22.717	kJ
	Total Renewable Energy	1351.000	0.094	0.121	1351.215	kJ
	Total water consumption	91.491	0.305	507.316	599.112	cm3
Outputs	CO2 Total	0.000	0.000	0.046	0.046	kg
	CO2	0.000	0.000	0.047	0.047	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.000	0.000	0.026	0.027	g
	CO	0.000	0.001	0.150	0.151	g
	NOx	0.000	0.000	0.048	0.048	g
	PM10	0.000	0.000	0.023	0.023	g
	PM2.5	0.000	0.000	0.011	0.011	g
	SOx	0.000	0.001	0.337	0.337	g
	CH4	0.000	0.000	0.119	0.120	g
	N2O	0.000	0.000	0.001	0.001	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	0.000	0.163	50.135	50.298	g

and

Table A6. Life cycle analysis per 1 MJ of H₂ consumed from biomass plant—Pipeline (1% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	149.837	1.922	20.796	172.554	kJ
	Total Renewable Energy	2231.000	0.094	0.121	2231.215	kJ
	Total water consumption	134.304	0.305	507.316	641.925	cm3
Outputs	CO2 Total	0.010	0.000	0.046	0.057	kg
	CO2	0.011	0.000	0.047	0.057	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.004	0.000	0.026	0.031	g
	CO	0.018	0.001	0.150	0.169	g
	NOx	0.030	0.000	0.048	0.078	g
	PM10	0.002	0.000	0.023	0.025	g
	PM2.5	0.002	0.000	0.011	0.013	g
	SOx	0.020	0.001	0.337	0.357	g
	CH4	0.019	0.000	0.119	0.139	g
	N2O	0.006	0.000	0.001	0.007	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	12.740	0.163	50.135	63.038	g

in Appendix A. The inputs to the GREET model are categorized into three main groups, such as renewable energy resources, non-renewable energy resources, and water consumption (Appendix A and Appendix B). The full input and output of the GREET model are presented in detail in [39].

Figure 2 below indicates the CO_2e emission intensity of the truck-trailer and pipeline pathways. This graph depicts the amount of CO_2e emitted per 1 MJ of hydrogen consumed in each truck and pipeline delivery scenario (assuming a 1% pipeline attribution).

Figure 3 shows the Volatile organic compound (VOC) and NO_x emissions for the six scenarios. Based on the results, solar via pipeline emits fewer VOC and NO_x and remains the best scenario for the production and transportation of hydrogen. However, in the case of NO_x, the biomass via truck is the worst-case scenario. However, SMR via truck is the worst-case scenario for emitting VOCs. As a result, if policymakers want to make decisions based on NO_x emissions, they should choose between solar or SMR via pipeline.

Figure 4 below illustrates Particulate Matter 2.5 (PM_2.5), an important impact factor on human health and air quality. Based on the findings, solar via pipeline emits less PM_2.5 than other scenarios because it is a less energy intensive pathway, and it is followed by biomass via pipeline and solar via truck. In terms of PM_2.5, the SMR and biomass via truck produce the highest PM_2.5.

3.1. Sensitivity Analysis

A key assumption in the analysis was the attribution of the pipeline construction to the LCA of 1 MJ of hydrogen consumed by an FCEV pickup truck. Given that we are considering the pathway of hydrogen from the production facility to one refueling station, it is important to understand pipeline networks and potential concerns with the assumption made in the base case presented above.

For the base case, we map all the hydrogen refueling stations that either currently exist, are in construction, or are permitted across California, and we find that they are located largely in the Bay Area–Sacramento region or in the Los Angeles region, as can be seen in Figure 5. We find that there are about 102 such stations, and thus, we assume that since we are considering 1 station, the attribution of the pipeline LCA to the total LCA estimate would be around 1% (i.e., 1 out of 102 stations = ~1%), as shown in the base case.

We studied the typical pipeline networks across the oil and gas sector. Conventionally, pipeline networks have three parts: primary network, transmission network, and distribution network. The primary pipeline connects the fuel production facility and carries the fuel till it reaches the periphery of a region, after which it is sent through the transmission network, which then divides into multiple distribution pipelines that deliver the fuel to the last-mile station. Thus, for the purpose of understanding the sensitivity of the final LCA to different assumptions of pipeline LCA attribution, we make two additional scenarios, where we assume the following:

30% pipeline LCA attribution: Given that the pipeline network is being constructed, it is assumed that the pipeline network will serve both the LA region and the Bay Area–Sacramento region. Thus, in such a scenario, the primary pipeline network will be relatively shorter, and the network will have greater pipeline length across transmission and distribution networks.

70% pipeline LCA attribution: Given the proximity of the LA region to the production facility considered in this analysis, we assume that the LA region will be served by truck and that the pipeline will be constructed to serve 67 out of 102 stations [40]. This means that the pipeline network will likely have a larger share of the primary pipeline, compared to the transmission and distribution pipeline network. Thus, we assume that a larger share of the pipeline LCA attribution (i.e., 70%) should be assumed for the reference refueling station considered in this study.

Additionally, implicit in this sensitivity analysis is the fact that while a pipeline will be built in a particular year with the intent of serving a certain existing density of end-users, it is assumed that the density of end-users will grow, and thus, greater demand for fuel through the pipeline will exist in the future. We do not assume this hypothetical scenario of increasing demand in the future to attribute a lower share of pipeline LCA to hydrogen produced and delivered today, i.e., we assume a relatively lower discount rate to reduce uncertainty in the future and to not spread the emissions impact over a longer period, thus reducing its potential effect on society today.

All the pipeline scenarios are compared to the best truck delivery scenario, i.e., solar via truck, to find the best combination of production and transportation pathways for a different share of the pipeline. As anticipated, the results revealed that solar via truck is better than all SMR combinations. It is as expected because solar via truck is less emission-intensive than the base case of SMR via pipeline. Nothing changes for other scenarios. Even with a 70% share of the pipeline, all the pipeline scenarios are still better than the best truck scenario. Figure 6 illustrates the CO_2eq sensitivity analysis for different pipeline share scenarios. The complete sensitivity analysis results are presented in

Table A7. Life cycle analysis per 1 MJ of H₂ consumed from Natural Gas (SMR)—Pipeline (30% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	1301.328	57.655	20.796	1379.779	kJ
	Total Renewable Energy	0.335	2.813	0.121	3.269	kJ
	Total water consumption	88.943	9.152	507.316	605.411	cm3
Outputs	CO2 Total	0.073	0.005	0.046	0.124	kg
	CO2	0.073	0.005	0.047	0.125	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.010	0.005	0.026	0.041	g
	CO	0.015	0.036	0.150	0.201	g
	NOx	0.019	0.005	0.048	0.072	g
	PM10	0.003	0.003	0.023	0.028	g
	PM2.5	0.003	0.001	0.011	0.015	g
	SOx	0.012	0.017	0.337	0.366	g
	CH4	0.178	0.009	0.119	0.306	g
	N2O	0.001	0.000	0.001	0.002	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	78.934	4.881	50.135	133.951	g

,

Table A8. Life cycle analysis per 1 MJ of H₂ consumed from Solar plant—Pipeline (30% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	0.000	57.655	20.796	78.450	kJ
	Total Renewable Energy	1351.000	2.813	0.121	1353.934	kJ
	Total water consumption	91.491	9.152	507.316	607.959	cm3
Outputs	CO2 Total	0.000	0.005	0.046	0.051	kg
	CO2	0.000	0.005	0.047	0.051	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.000	0.005	0.026	0.032	g
	CO	0.000	0.036	0.150	0.186	g
	NOx	0.000	0.005	0.048	0.053	g
	PM10	0.000	0.003	0.023	0.026	g
	PM2.5	0.000	0.001	0.011	0.012	g
	SOx	0.000	0.017	0.337	0.354	g
	CH4	0.000	0.009	0.119	0.128	g
	N2O	0.000	0.000	0.001	0.001	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	0.000	4.881	50.135	55.016	g

,

Table A9. Life cycle analysis per 1 MJ of H₂ consumed from biomass plant—Pipeline (30% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	149.837	57.655	20.796	228.287	kJ
	Total Renewable Energy	2231.000	2.813	0.121	2233.934	kJ
	Total water consumption	134.304	9.152	507.316	650.772	cm3
Outputs	CO2 Total	0.010	0.005	0.046	0.061	kg
	CO2	0.011	0.005	0.047	0.062	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.004	0.005	0.026	0.035	g
	CO	0.018	0.036	0.150	0.203	g
	NOx	0.030	0.005	0.048	0.082	g
	PM10	0.002	0.003	0.023	0.028	g
	PM2.5	0.002	0.001	0.011	0.014	g
	SOx	0.020	0.017	0.337	0.374	g
	CH4	0.019	0.009	0.119	0.148	g
	N2O	0.006	0.000	0.001	0.008	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	12.740	4.881	50.135	67.757	g

,

Table A10. Life cycle analysis per 1 MJ of H₂ consumed from Natural Gas (SMR)—Pipeline (70% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	1301.328	134.528	20.796	1456.652	kJ
	Total Renewable Energy	0.335	6.563	0.121	7.020	kJ
	Total water consumption	88.943	21.354	507.316	617.613	cm3
Outputs	CO2 Total	0.073	0.011	0.046	0.130	kg
	CO2	0.073	0.011	0.047	0.131	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.010	0.012	0.026	0.048	g
	CO	0.015	0.083	0.150	0.248	g
	NOx	0.019	0.011	0.048	0.078	g
	PM10	0.003	0.006	0.023	0.032	g
	PM2.5	0.003	0.003	0.011	0.016	g
	SOx	0.012	0.039	0.337	0.388	g
	CH4	0.178	0.020	0.119	0.317	g
	N2O	0.001	0.000	0.001	0.002	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	78.934	11.389	50.135	140.459	g

,

Table A11. Life cycle analysis per 1 MJ of H₂ consumed from solar plant—Pipeline (70% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	0.000	134.528	20.796	155.324	kJ
	Total Renewable Energy	1351.000	6.563	0.121	1357.685	kJ
	Total water consumption	91.491	21.354	507.316	620.161	cm3
Outputs	CO2 Total	0.000	0.011	0.046	0.057	kg
	CO2	0.000	0.011	0.047	0.057	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.000	0.012	0.026	0.038	g
	CO	0.000	0.083	0.150	0.233	g
	NOx	0.000	0.011	0.048	0.059	g
	PM10	0.000	0.006	0.023	0.029	g
	PM2.5	0.000	0.003	0.011	0.014	g
	SOx	0.000	0.039	0.337	0.376	g
	CH4	0.000	0.020	0.119	0.140	g
	N2O	0.000	0.000	0.001	0.001	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	0.000	11.389	50.135	61.524	g

and

Table A12. Life cycle analysis per 1 MJ of H₂ consumed from biomass plant—Pipeline (70% pipeline LCA attribution).

		Life Cycle Stage (per 1 MJ Hydrogen Consumed)
	Environmental Flow/Impact	H2 Production	Pipeline Production and construction	End user Pick-up truck	Total life cycle per MJ consume on FCV	Units
Inputs	Total Non-Renewable Energy	149.837	134.528	20.796	305.160	kJ
	Total Renewable Energy	2231.000	6.563	0.121	2237.685	kJ
	Total water consumption	134.304	21.354	507.316	662.974	cm3
Outputs	CO2 Total	0.010	0.011	0.046	0.067	kg
	CO2	0.011	0.011	0.047	0.068	kg
	CO2 Biogenic	0.000	0.000	0.000	0.000	kg
	VOC	0.004	0.012	0.026	0.042	g
	CO	0.018	0.083	0.150	0.251	g
	NOx	0.030	0.011	0.048	0.088	g
	PM10	0.002	0.006	0.023	0.031	g
	PM2.5	0.002	0.003	0.011	0.016	g
	SOx	0.020	0.039	0.337	0.396	g
	CH4	0.019	0.020	0.119	0.159	g
	N2O	0.006	0.000	0.001	0.008	g
	CF4	0.000	0.000	0.000	0.000	g
	SF6	0.000	0.000	0.000	0.000	g
	C2F6	0.000	0.000	0.000	0.000	g
	CO2e	12.740	11.389	50.135	74.265	g

in Appendix B.

Figure 7 shows the comparison of solar via truck as the best truck scenario with different pipeline scenarios in terms of NO_x intensity. The results indicate that all pipeline scenarios, regardless of the share of the pipeline for one specific refueling station, have better performance for controlling NO_x than solar via truck. Given the truck operates on low sulfur diesel, the solar via truck pathway emits much more NO_x than any pipeline scenario. Suppose policymakers want to make decisions based on NO_x emissions, they should choose pipeline scenarios over the truck regardless of the cost and time of construction and production of the pipeline.

The story of VOC sensitivity analysis is different from that of NO_x analysis. Figure 8 illustrates that for the 30% share of pipeline scenario, the solar via truck pathway is similar to the biomass via pipeline in terms of VOC emissions. Additionally, it is indicated that in the 70% share of pipeline scenario, all pipeline-based pathways are worse than the best truck scenario. Thus, decision-makers should consider these findings whenever the cost and time do not matter, and urban area air quality has the highest priority, given that NO_x and VOC are the most critical in ozone creation near the ground.

Figure 9 indicates that the results of the PM_2.5 sensitivity analysis lead the decision-makers to similar decisions as for VOC. It is mainly because, in the 30% share of the pipeline scenario, only solar via pipeline is better than the best truck scenario. In the 70% share of pipeline scenario, all pipeline combinations are worse than solar via truck. So, the share of the pipeline is a critical player in the life cycle analysis of different production and transportation pathways of hydrogen, especially when it comes to non-CO₂ emissions.

3.2. Policy

As can be seen from the LCA estimates presented in the earlier section, transporting hydrogen in the gaseous form via pipelines is the least carbon-intensive pathway, and more specifically, hydrogen produced using 100% solar energy in the gaseous form is the lowest carbon pathway. Understanding the potential for decarbonization across hydrogen production and delivery pathways will be key, both in terms of the economic feasibility of the pathway and the delivered cost to the consumer, as well as the potential for hydrogen as a low carbon fuel option.

The state of California implements the Low Carbon Fuel Standards (LCFS) program, which essentially rewards fuel suppliers and retailers for meeting CI reduction targets. As per CARB, the LCFS is designed to decrease the CI of California’s transportation fuel pool and provide an increasing range of low-carbon and renewable alternatives that reduce petroleum dependency and achieve air quality benefits [41].

The LCFS mechanism is an offset system or a credit-trading system, wherein fuels that qualify below the set CI benchmark for each year generate credits, which are then sold to those suppliers and retailers who do not meet the CI targets for that year. The standards are measured in terms of the CI of gasoline and diesel fuel and their respective substitutes. The measure of CI is based on the life cycle GHG emissions of each fuel and its potential pathways, which include the production, transportation, and consumption of the fuel (Figure 10).

As can be seen in Figure 11 below, each fuel has a range of potential CI values based on the different certified pathways under the LCFS program, which include different combinations of production, transportation, and use of the fuel. Compared to the reference case, hydrogen fuel can have a relatively wide range of CI values. This is essentially driven by the choices of liquid vs. gaseous hydrogen, transportation via truck or pipeline, and so on.

At the same time, it should be noted that the potentially lowest CI value may not necessarily be the lowest cost option, as the LCA only provides the environmental impact but is not based on optimizing both cost and emissions. The most common method of producing hydrogen today is SMR, and is delivered by truck in gaseous form. As can be seen in the LCA estimates presented in the earlier sections, SMR hydrogen by truck has the highest emissions (gCO_2eq./MJ of hydrogen), and thus, its potential to help meet the overall CI reduction targets in the fuel basket within the LCFS regime may be limited as compared to a change either in the production process to biomass or solar, or in the transportation pathway from truck to pipeline.

While pipelines are costly and there will be a significant time lag in building out such infrastructure, delivery by truck will likely remain the preferred option for the foreseeable future. Thus, in terms of taking advantage of the LCFS mechanism, moving to a different production process, such as solar or biomass-based hydrogen, may yield greater low carbon results for fuel producers and retailers in the near to medium term. Thus, the feasibility of the hydrogen pathway will define the level of CI reduction possible. Although, we do see that between different pathways there is potential for marginal improvements, and it is important to consider them as the share of hydrogen fuel increases in the fuel basket. This will result in a meaningful weighted average reduction in CI values and take us closer to the intended CI reduction targets for each year.

4. Discussion

Given the relatively limited body of work focusing on detailed LCA analysis of hydrogen pathways, in this section, we highlight some key points regarding methodology and approach.

Firstly, the GREET model used in our analysis, which is relied on as an LCA tool by most policymakers in California and across the US, has relatively optimistic assumptions compared to other LCA tools such as GaBi or Ecoinvent. For example, we find that GREET has lower emissions from hydrogen production compared to other tools. Based on the Ecoinvent database, Simons et al. [44] estimated the GHG emissions of SMR hydrogen production at 125 gCO_2eq per 1 MJ of hydrogen. However, according to our findings 78.93 gCO_2eq are produced per 1 MJ of hydrogen. The authors in [35] estimated 111 gCO_2eq/1 MJ H₂ for SMR-truck and 89 gCO_2ep/1 MJ H₂ for SMR-pipeline using the GREET database. Even after excluding compression prior to loading, we estimated 125.36 gCO_2eq/1 MJ H₂ for SMR-truck and 90.323 gCO_2eq/1 MJ H₂ for SMR-pipeline (70% attribution) using the GREET database. These findings highlight the significance of taking the LCA of pipeline and tube production into account.

The goal of this study is to assess the life cycle impacts of various hydrogen production, delivery to a hydrogen refueling station, and consumption by an FCEV pickup truck; however, water consumption could be a significant limitation, particularly in California. According to this study, all truck pathways use more water than pipeline alternatives. Furthermore, water consumption does not differ significantly between pipeline attribution scenarios. This could be interesting because solar electrolysis pathways, which use water as their main feedstock, use roughly the same amount of water as SMR pathways but much less than biomass pathways (Figure 12 and Figure 13).

We also find that the production of the onboard hydrogen tank in the FCEV pick-up truck, which is the end-user of the hydrogen fuel in our analysis, is highly energy intensive. This has the potential to make the FCEV look not so efficient from an emissions standpoint. The inclusion of innovative technologies in the production of onboard hydrogen tanks has the potential to improve the environmental performance of all pathways. Furthermore, taking into account developing new technology in heavy duty fleets and improving fuel economy could enhance the environmental performance of each tube trailer truck-based approach.

GREET also has limited impact categories to measure, essentially restricted to global warming potential (GWP) impacts. While it is true that overall CO₂ reductions are the focus of most low-carbon focused policy mechanisms, it would be critical to understand the other potential impacts of fuel pathways in terms of land, water, and other resources. Thus, for future work, we recommend considering other reference LCIs to have more accurate assumptions for the LCA estimates. Additionally, in this analysis, we do not consider the compressor at production and refueling stations, a process that is highly energy intensive across the hydrogen pathway. We also do not consider the construction of the refueling station and fuel storage tank at the fuel station, which could change the potential LCA estimate.

5. Conclusions

In the past few decades, initial research has been conducted looking at the impacts of multiple segments of the hydrogen fuel life cycle. Currently, LCAs exist for conventional hydrogen production and the construction of refueling stations, but few studies have focused on the hydrogen fuel transportation pathways. This study looked at the LCA of gaseous hydrogen from SMR, solar electrolysis, and biomass-based production in LA to a refueling station in East Sacramento. The foreground system included the production of hydrogen, truck and tube production, and the production and construction of pipelines and an end-user pickup trick. Total energy and water consumption were considered for the inputs of each transportation pathway and four impact categories. Overall, it was found that the pipeline transportation pathway is less GHG intensive. Then, a sensitivity analysis is utilized to investigate the final LCA of different assumptions of pipeline pathway LCA attribution. Based on the LCI and impact assessment results, it is concluded that the outcome could change based on the different attribution of the pipeline LCA. However, solar electrolysis hydrogen production through pipeline transportation emerges as the best choice in terms of CO_2eq in all scenarios and produces 50.298 gCO_2eq/1MJ H₂. The results show that manufacturing the onboard hydrogen tank in the end user pickup truck uses a lot of energy, and emerging innovative technologies into the production of onboard hydrogen tanks has the potential to enhance the environmental sustainability of all pathways.