Exploring the Technological Advances and Opportunities of Developing Fuel Cell Electric Vehicles: Based on Patent Analysis

Abstract

In general, the fuel cell electric vehicle (FCEV) is regarded as more environmentally friendly than other vehicles. However, the commercialization of FCEV technology is hardly fulfilled due to high-cost fuel cells and an inadequate refueling infrastructure. Different technological trajectories of fuel cells are fiercely competitive, and related technologies are iterating quickly. It is an open issue in terms of what are the technological advances achieved or the opportunities for innovators. The paper proposes a novel approach to identify the key components of an FCEV by constructing the directed co-occurrence network of the International Patent Classification (IPC) and then adopts the Natural Language Processing (NLP) to construct the matrix of technology characteristics and functions. It is suitable to analyze the sentence structure of Subject–Action–Object (SAO) in patent documents by utilizing the NLP technology, which can help computers understand the text and communicate with us. The paper finds that the advances achieved in the fuel cell field are fuel cell composition, manufacturing fuel cells, and providing energy using fuel cells, and the advance in electric motors is supplying power for fuel cell vehicles, while the advances in hydrogen storage are to manage and store hydrogen. By contrast, the opportunities for innovators are to develop the control, diagnosis, and performance of the control system and hydrogen filling. This paper will be a contribution towards a better understanding of the advances and opportunities for developing FCEV technology.

1. Introduction

A fuel cell electric vehicle (FCEV) refers to a vehicle driven by an electric motor that is powered by fuel cells [1], which is more environmentally friendly than the traditional incumbent internal combustion engine vehicle (ICEV) and the other low-emission vehicles (LEVs). Hydrogen is widely regarded as the cleanest energy source because fuel cells can produce direct current power to drive vehicles and water is a product of the electrochemical reaction, which is harmless to the environment. Just as Manoharan et al. [2] explained, a fuel cell can produce electricity by allowing hydrogen and oxidants as reactants and then drive vehicles. To protect the environment and reduce car emissions to a greatest extent, increasing countries have adopted the strategy of prioritizing developing hydrogen since 2019 [3], because the transportation system is widely regarded as one of the primary causes of environmental pollution [4]. In an era of decarbonization, the development of FCEV has received great attention and made some progress.

From the perspective of longitudinal time, the development of FCEV is not smooth. The first wave of developing FCEV was in the period of the late 1990s and the mid-2000s, which was initiated by Daimler and then followed by other large car manufacturers. However, this wave suddenly receded in 2007 due to the high cost of fuel cells and lack of infrastructure [5]. Fortunately, the development of FCEV has been reinvigorated since 2014, under the incentive of a set of public policies launched by some countries [6].

It is an open issue in terms of what are the technological advances in the process of developing an FCEV or what are the potential opportunities for innovators. Different technological trajectories of fuel cells are fiercely competitive and related technologies are iterating quickly. Fang et al. [7] pointed out that there are six types of fuel cells based on the electrolytes employed and provided insights into the advantages and challenges between proton-exchange membrane fuel cells (PEMFCs) and solid oxide fuel cells (SOFCs). Moreover, Xu et al. [8] analyzed the advancements achieved by the third generation of SOFCs, metal-supported solid oxide fuel cells (MS-SOFCs). WIPO [9] found that fuel cells reached a high degree of technological maturity and growing patent applications in automated production and fuel cell recycling. Innovators pursue developing diversified strategies or technologies to enhance the performance, efficiency, and lifespan of fuel cells and improve the safety and convenience of hydrogen delivery and storage [10]. Manoharan, Hosseini, Butler, Alzhahrani, Senior, Ashuri and Krohn [2] investigated a set of technologies related to FCEVs and hybrid vehicles driven by fuel cells from a technical perspective and highlighted that applying fuel cells in hybrid vehicles is rational. However, there are some technical challenges to the commercialization of FCEVs due to the system complexity [11]. Samsun et al. [12] emphasized that the FCEV technology would lose its strategic position in the LEVs’ future if the dynamic of the FCEV deployment was delayed. In this context, it is crucial to understand the advances achieved and the potential opportunities in the process of developing FCEV technology.

To close this gap, the paper aims to address the following issues: what have the technological advances in developing FCEVs achieved in the recent decade? And what are the opportunities for innovators to develop FCEVs in the future? In doing so, we propose a novel approach to identify the key components of FCEVs by constructing the directed co-occurrence networks of the International Patent Classification (IPC) and then utilize the Natural Language Processing (NLP) to construct the matrix of technology characteristics and functions. The NLP technology can help computers to understand text and communicate with human beings. It is suitable to adopt the NLP technology to analyze the sentence structure of Subject–Action–Object (SAO) in patent documents.

The paper is organized as follows. Following the introduction, we review the literature on the development of FCEV technology and the patent analysis on FCEVs. Section 3 reports the research framework and data, while Section 4 depicts the analysis results of IPC co-occurrence networks and the technical characteristics and function matrix. Finally, we draw a conclusion after discussing the advances and opportunities of developing FCEV technology in Section 5.

2. Literature Review

The literature survey is composed of three parts in the paper: the process of developing FCEV technology, the advantages and challenges for developing FCEV, and patent analysis of the FCEV technology.

2.1. The Process of Developing FCEV Technology

There is a consensus that the FCEV technology is promising because hydrogen is a clean energy carrier. The environmental and political considerations, which are to satisfy zero-emission transportation and guarantee energy security, have greatly driven the development of FCEVs in practice [13]. The combination of green hydrogen technology and fuel cells can effectively reduce car emissions and meet the growing energy demands. In this context, the wave of developing FCEVs was originally triggered in the late 1990s and the mid-2000s. However, a set of drawbacks impeded the commercialization of FCEVs, such as the high cost of fuel cells, durability, and the lack of infrastructure [5]. Thus, the first wave of developing FCEVs suddenly receded in 2007, which led to the innovation activities in this field sharply decreasing.

There is a controversy over when the new wave of developing FCEVs emerges in the existing literature. WIPO [9] pointed out there were three waves of developing the application of fuel cells in transportation based on patent analysis; the first wave was in the mid-1980s, the second wave was around 2005, and the third wave emerged in 2016. However, Yuan and Yuan [3] argued that the development of FCEVs has started to be revived since 2014 based on patent analysis, in which dominant incumbents, namely large automobile manufacturers and component suppliers, devoted themselves to developing the FCEV technology.

On the other hand, several commercial FCEVs using hydrogen fuel cell technology have been gradually launched to the market since 2014 [13]. For instance, the Honda FCX clarity, Audi Sportback A7h-tron Quattro, and Volkswagen Golf Hymotion were launched in 2014, the Hyundai Tucson Fuel Cell was sold in 2016, with the Honda Clarity Fuel Cell in 2017, and the Toyota Mirai 1 was commercialized in 2020, while Mirai 2 was in 2021 [14]. However, the global sales of FCEVs are not good. For instance, NEXO is the commercialized second-generation FCEV of Hyundai, which was sold in the world in 2018. About 11,179 Hyundai NEXOs were sold globally by 2022, but the sales of the NEXO were sharply reduced to 4709 in 2023. By contrast, 3694 Toyota Mirais were sold in 2022, with 3737 in 2023.

According to the deployment status investigated by Samsun, Rex, Antoni, and Stolten [12], 42,192 passenger FCEVs were on the roads in 2021, of which more than 90% of passenger FCEVs were in South Korea, the U.S., and Japan. Notably, FCEVs can scale up power delivery with minimized increased weight compared with battery electric vehicles, which leads to the number of commercial medium- and heavy-duty vehicles operated by fuel cells increasing [15]. Therefore, increasing FCEVs, which are launched to markets and operated in practice, demonstrate the technical and commercial feasibility of FCEVs. The development of FCEVs has been revived since 2014, but there is a long way to achieve commercial success, from a commercial perspective.

2.2. The Advantages and Challenges for Developing FCEVs

It is a hot topic of discussing how to develop hydrogen and fuel cell technology in the extant literature. There are two distinct ways to explore the advantages and challenges of developing FCEVs.

The first strand of literature focuses on exploring the advantages of FCEVs from a comparative perspective. LEVs consist of battery electric vehicles (BEVs), FCEVs, and hybrid electric vehicles [16]. Some researchers prefer to study the generic technologies of LEVs or special technologies in FCEVs. For instance, Wesseling, Faber, and Hekkert [5] regarded battery technology, motor technology, and powertrain system integration as generic technologies for each type of LEV, but fuel cells and equipment for storing hydrogen were the unique technologies for developing FCEVs. The technological advantages of FCVs are the long driving range and fast refueling time, compared with the other LEVs [14]. Bethoux [17] pointed out that FCEVs have several key advantages: the long driving range, achieving the rated traction power, fast and simple refueling, and safety, though the electric power provided by fuel cells was less efficient than lithium-ion batteries.

The second strand concentrates on studying fuel cells and hydrogen technology. The extant literature widely explores fuel cell and hydrogen technology applied in transportation but is not limited to the FCEV technology field. A fuel cell consists of at least three fundamental segments: the anode, cathode, and electrolyte. Moreover, fuel cells can be divided into different categories based on the electrolytes applied, such as alkaline fuel cells (AFCs), molten carbonate fuel cells (MCFCs), phosphoric acid fuel cells (PAFCs), PEMFCs, and SOFCs, but each one has its advantages and limitations. At present, PEMFCs may be more suitable for automotive commercial applications. However, Xu, Han, Zhu, Ni, and Yao [8] highlighted that MS-SOFCs made remarkable progress, which led to GE and Ceres Power commercializing MS-SOFCs successfully. Fang, Vairin, von Jouanne, Agamloh, and Yokochi [7] suggested collaborating in developing PEMFCs and SOFCs to address technological challenges.

Additionally, hydrogen technology is sophisticated and integrated, which relates to the process of hydrogen generation, hydrogen storage, hydrogen transportation, and hydrogen application [14]. Figure 1 shows the key components of FCEVs, which are illustrated by the alternative fuels data center, in the vehicle technologies office of the Department of Energy, U.S. [18].

However, many challenges should be met in the process of developing FCEVs. First, the price of fuel cells is the primary challenge for commercialization. Manoharan, Hosseini, Butler, Alzhahrani, Senior, Ashuri and Krohn [2] discussed the key components of FCEVs in detail, such as hydrogen storage in FCEVs, fuel cells, and control strategies, and regarded the price of fuel cells as the primary drawback for commercializing FCEV.

Second, durability is an essential factor in impeding the commercialization of fuel cells. Fan, Tu, and Chan [14] hold that stability and cost were the main challenges for developing FCEVs, and system integration, structural design, and material could hamper technology advancements and energy improvements. Bahrami et al. [19] pointed out that high cost, poor durability, and slow response in dynamic operational scenarios were key challenges for the commercialization of proton exchange membrane fuel cells (PEMFCs). There is tension between cost and durability, which remains an enormous challenge for developing FCEVs [20]. In this context, Tang et al. [21] held that controlling and regulating the operating temperature of fuel cells in different state-of-health levels would improve the performance and durability of FCEVs.

Finally, infrastructure, hydrogen production, and storage have limited the development of FCEVs [22]. From an infrastructure perspective, Bethoux [23] predicted that the large-scale commercialization of FCEVs would possibly start in the years 2030–2035 and held that a set of challenges should be addressed, such as cost, durability, and hydrogen production.

2.3. PatentAnalysis Based on the FCEV Technology

A patent document may contain technical, legal, and managerial information on an invention in a given technology field. In the era of rapid technological change, it is vital for organizations or innovators conducting patent analysis to fulfill diversified purposes. Abbas et al. [24] outlined the primary purposes of conducting patent analysis and illustrated the main techniques adopted in the existing literature, in which the methodology of text mining and visualization were utilized by most literature on patent analysis.

Moreover, a rich literature discusses the technological trend of green hydrogen and fuel cells. The first strand of literature focuses on predicting the trend in hydrogen technology and production based on structured patent data, such as technology fields, publication years, countries, applicants, etc. For instance, Yang et al. [25] classified green hydrogen technology into two types and seven subtypes, namely water electrolysis and system operation, and that found companies preferred developing system integration and control strategies for green hydrogen production, and the technologies of alkaline and polymer electrolyte membranes were at the stage of entering maturity based on patent analysis. Moreover, WIPO [9] found that the road sector accounted for the highest proportion of patents, followed by the aviation and shipping sector, and emphasized that the diffusion of hydrogen fuel cells strongly depended on infrastructure, renewable energy, and battery technology advances. Polymer electrolyte membrane fuel cells were the most promising and achieved a high degree of technological maturity based on patent analysis.

The second strand of literature focuses on analyzing the development trend of fuel cell technology. For instance, Moura et al. [26] evaluated the development of hydrogen fuel cells based on the patent data between 2001 and 2020 and found that the patent applications in polymer electrolyte membranes and solid oxide fuel cells were prominent. Moreover, Tsang et al. [27] investigated 2269 patents and found that complementary resources would affect the patent-renewing decisions made by patent holders.

However, less literature explores the technological advances and opportunities of FCEVs based on patent data. Most research on FCEV technology strongly depends on structured data that are uniform in format and semantics in patent documents, such as the filing date, patent number, IPC codes, etc. By contrast, the information on technology characteristics and functions is hidden in unstructured data, such as the abstract or claims in patent documents. To close this gap, we adopt the NLP technology to construct the patent Information Relation Matrix (IRM) and explore the advances and opportunities for developing FCEV technology.

3. Research Method and Data

3.1. Methodology

The paper adopts a novel methodology of constructing the directed IPC co-occurrence networks to categorize inventions into FCEV’s key components and forms a patent information relation matrix by using the NLP technology. Due to numerous patents, we should identify the invention subjects closely related to the key components of FCEVs and exclude unrelated patent information. In doing so, we construct the directed IPC co-occurrence networks to identify the core technology components.

In general, the co-occurrence networks of IPC codes are non-directed because there is an assumption that each IPC code in a special patent is of equal importance, which can be used to conduct clustering analysis by measuring the centrality of IPC codes [28]. However, there are a primary IPC codes and or many secondary classifications if the number of IPC codes assigned to a patent application is more than two. The primary IPC code may reflect the closest technical field covered by the patent application, while the secondary IPC codes are usually considered to be necessary and complementary to the primary IPC code in a given patent application. To exclude the disturbance of irrelevant information, the paper takes the difference across the IPC codes into account and then constructs the directed co-occurrence networks of IPC codes. Figure 2 presents how to construct a directed IPC co-occurrence network in a given technology field. Suppose there are four patents, P1-4, and each patent is assigned more than two IPC codes. A directed IPC co-occurrence network is constructed by identifying the primary and secondary IPC codes in patents. The relations between these patents can be expressed as several pairs of network ties, which then form a directed IPC co-occurrence network. The clustering analysis on IPC co-occurrence networks can be used to identify the key components of FCEV technology.

Moreover, the paper adopts the NLP technology to analyze the abstracts in patent documents and extract the technology characteristics and functions of patents. It is scarcely possible to manually analyze patent documents due to numerous patents. The NLP technology can be used to analyze the structure of Subject–Action–Object (SAO) in the abstract and extract essential technology information. The form of “verb + noun” pairs may represent the technology functions, while the form of “inclusion verb + noun” in sentences refers to the technology characteristics [29]. For instance, the patent “WO2010104421-A1” filed in 2009 is about an “apparatus for generating hydrogen used for fuel cell”. The “Novelty” of this patent recorded in the Derwent Innovations Index (DII) is “A hydrogen generating apparatus has a fuel reformer (2) for reforming hydrocarbon fuel (8) to hydrogen-rich gas (10) comprising hydrogen, carbon monoxide and unconverted fuel compounds, and a molten salt reactor (12) arranged in connection with the fuel reformer for supplying hydrogen-rich gas from the reformer”. The technology characteristic of this patent is “a hydrogen generating apparatus” based on the structure of “inclusion verb + noun”. In addition, the “Use” of this patent is an “apparatus is used for generating hydrogen from a hydrocarbon fuel, used for fuel cell”. The function of this patent is “generating hydrogen” based on the form of “verb + noun” pairs. As a result, we obtain a matrix of technology characteristics and functions, namely “a hydrogen generating apparatus” and “generating hydrogen”. Thus, utilizing NLP makes it possible to construct a sophisticated IRM based on patent documents.

3.2. Research Framework

We utilize the software Python 3.9.19 to preprocess the patent records and adopt the modified method to explore the technological advances and opportunities in developing FCEVs, following the existing literature. Figure 3 presents our research framework.

First, the paper adopts priority patent applications to measure the development of FCEV technology from a panoramic perspective. A priority patent application usually refers to the first patent application in a patent family, while a patent family is a series of equivalent applications that are filed at different patent offices based on the same invention [3]. Calculating the number of priority patent applications can more accurately measure the technology innovation activities than applications in given countries.

Second, a cluster analysis on IPC co-occurrence networks is conducted to identify the core technology components. The FCEV technology is an integrated and sophisticated system that encompasses a set of technical components or elements, such as fuel cells, electric motors, the control system, hydrogen storage, etc. [2]. We disassemble the FCEV technology into different components and then identify the core components to illustrate the technological advances and opportunities.

Third, we identify the core components of FCEV technology based on clustering analysis. The paper constructs directed IPC co-occurrence networks and uses the software Gephi 0.9.2 to conduct the clustering analysis. The nodes that occupy a dominant position in the IPC co-occurrence networks are identified as the key components of FCEV technology.

Fourth, the paper adopts the method proposed by Ki and Kim [29] to construct the IRM. We leverage the spaCy library to analyze the structure of Subject–Action–Object (SAO) and extract technical information from text data. Representative information is identified by counting the frequency and technical suitability. Then, a matrix is created using the binary relation in which technology characteristics and functions are used as axes.

Finally, we reveal the advances and opportunities for developing FCEV technology. The number of tags represents the number of co-occurrences of technology characteristics and functions. Thus, the vacancy or void in the matrix means technological opportunity, while the number with a higher value implies that the technology advance is achieved.

3.3. Data Collection

The paper adopted the patent search strategy proposed by Borgstedt et al. [30]. We ran the retrieval strategy by combining keywords and patent technical classifications in the Derwent Innovations Index (DII), which were IP = (B60W-010/28 OR B60L-011/18 OR H01M-008*) AND TS = (vehicle* OR car OR automobile*) AND TS = (fuel cell*). The DII is very suitable for patent analysis in a given technology because it is a famous patent database for collecting a great volume of global patent documents since 1963.

We collected 19,626 patent families for FCEV technology between 2010 and 13 December 2023, and these patent records were downloaded on 1 June 2024. The time frame merely covers 2014–2023 in our patent analysis because the research purpose of this paper is to reveal the technological advances and opportunities for developing FCEV technology in recent years. There were only 781 priority patent applications filed in 2023 because there is a time lag between patent application and publication, namely eighteen months. However, it still is worth conducting a patent analysis on the priority applications in 2023 because it may represent the latest technology development of FCEVs.

To simplify, we merely take the first priority date in patent families to conduct the patent analysis, though some patent families may have more than two priority rights. As a result, 13,071 patent families are identified between 2014 and 2023.

4. Research Results

4.1. The Evolution of Developing FCEV Technology

Figure 4 shows the number of annual and accumulated patent applications based on the patent priority date between 2010 and 2023. The annual number of priority applications maintains a stable growth trend, except in 2022 and 2023. The priority applications in 2021 were 2,086 and reached the peak, while the priority applications in 2023 were only 781 due to the time lag between patent application and publication. In total, the accumulated applications reached 17,366 between 2010 and 2023. This suggests that innovation activities have flourished in the recent decade.

4.2. Identifying the Key Components of FCEV

Figure 5 reports the results of constructing directed IPC co-occurrence networks, in which Figure 5a depicts the IPC co-occurrence networks between 2010 and 2023, while Figure 5b illustrates the IPC co-occurrence networks from 2014 to 2023. The IPC co-occurrence network in 2014–2023 can more clearly reflect the latest status of developing FCEV technology, though the structure and scale of IPC co-occurrence networks are similar between 2010–2023 and 2014–2023.

Table 1. The results of clustering analysis based on the IPC co-occurrence networks (2014–2023).

Label	Key Components	Indegree	Outdegree	Weighted Indegree	Weighted Outdegree	Pagerank	Clustering	Eigencentrality
H01M	fuel cell	206	188	19,624	3165	0.128225	0.021278	1
B60L	electric motors	73	156	3416	11,830	0.029669	0.038828	0.54635
B60K	electric motors	39	99	577	3547	0.013384	0.073295	0.360956
B01D	fuel cell	22	83	147	1116	0.006599	0.085004	0.231471
C01B	-	28	61	268	794	0.010284	0.081511	0.257132
F17C	hydrogen storage	41	55	432	786	0.017295	0.113458	0.382421
B60H	the control system	10	54	80	1431	0.003683	0.132792	0.148564
B01J	fuel cell	12	39	194	981	0.003533	0.150407	0.136687
B60R	-	21	31	216	192	0.00791	0.203453	0.247281
C08G	-	11	31	42	315	0.004588	0.150624	0.127599
C25B	fuel cell	34	29	346	413	0.01187	0.169218	0.317172
B60S	-	11	24	63	133	0.003444	0.23399	0.152524
G01R	the control system	16	22	315	445	0.008233	0.243386	0.225025
H02J	the control system	40	14	850	121	0.016436	0.176263	0.359873

depicts the results of clustering analysis based on the four-digit IPC co-occurrence networks between 2014 and 2023. The primary IPC code is unique and closer to the technical field covered by a given invention. The directed IPC co-occurrence networks can describe how key technology fields have integrated with complement technology fields. This paper adopts the indicator out-degree to rank the distribution of four-digit IPC codes, though a set of network characteristics is reported in Table 1. As a result, we obtain the key components of FCEV technology, in which fuel cell technology includes the four-digit IPC codes H01M, B01D, C25B, and B01J, electric motors are composed of B60L and B60K, the control system includes H02J, G01R, and B60H, and hydrogen storage relates to the four-digit IPC code F17C. Some four-digit IPC codes are not identified as the key components, though listed in the top 10. For instance, the four-digit IPC code “B60R” refers to the other vehicle fittings or parts, but it is not recognized as the key component of FCEV technology.

Moreover, Figure 6 depicts the patent applications concerned with key components of FCEV technology. Each patent merely belongs to one type of key component without duplication because the primary IPC code assigned to each patent is unique. Figure 6 shows that the number of patent applications concerned with key components maintains a growth trend, but the majority of patent applications concentrate on the fuel cell and electric motors, while patents related to the control system and hydrogen storage are limited.

4.3. The Matrix of Technology Characteristics and Functions

The paper extracts technology information from the “Abstract” in DII and then constructs the matrix of technology characteristics and functions based on the information “Novelty” and “Use”. The information “Novelty” presents the technology characteristics, while “Use” describes the main technology functions of a special invention.

The first step is to extract the text of “Novelty” and “Use” in the “Abstract” by utilizing Python. Then, we adopt Part-of-speech (POS) Tagging to assign word types to tokens by using the spaCy library. The third step is to extract the form of “inclusion verb + noun” in “Novelty” and the form of “verb + dobj” pairs in “Use”, based on the structure of Subject–Action–Object (SAO) in sentences. Finally, we identify the Subject in “Novelty” as technology characteristics and adopt the Action–Object (AO) in “Use” as technology functions.

Moreover, we apply two criteria to identify technology information: frequency and technology suitability, following the method proposed by Ki and Kim [29]. In doing so, we adopt the openpyxl in Python to label and merge similar expressions into technology information and use Gephi to conduct the clustering analysis. Figure 7 presents the co-occurrence networks of technology characteristics and functions in the key components.

5. Discussion

This paper adopts the directed IPC co-occurrence networks to identify the key components of FCEV technology from 2014 to 2023 and then utilizes the NLP technology to construct the matrix of technology characteristics and functions. Our research makes some valuable findings. (1) The key components of FCEV technology, including fuel cells, electric motors, the control system, and hydrogen storage, have maintained a trend of gradual growth. (2) Both fuel cells and electric motors are the priority areas of developing FCEV technology due to a sharp increase in patent applications. (3) There are many technology opportunities for developing FCEVs in the future.

5.1. The Key Components of Developing FCEVs

Our finding depicts that patenting activities in developing FCEV technology maintain an increasing trend, which is in line with the conclusion drawn by Khan et al. [31]. However, we find out that the key components of FCEV technology still keep increasing from 2014 to 2023. Our finding provides detailed evidence that the development of FCEV technology is reviving indeed, due to the increase in patent applications on key components.

Our research finds that innovation activities are uneven across the key components of FCEV technology. This paper identifies the key components, including fuel cells, electric motors, the control system, and hydrogen storage, which have occupied an important position in the directed IPC co-occurrence networks. However, the development of key components is not in disequilibrium, in which most of the inventions focus on fuel cells and electric motors, while the innovation activities in developing control systems and hydrogen storage are not very active. Notably, developing fuel cells or stacks is the most active area. Table 1 shows that either the in-degree or out-degree of subclass “H01M” is the highest, which covers the technology of fuel cells or stacks. This implies that great progress has been achieved in the direct conversion of chemical energy into electrical energy in recent years.

5.2. The Advance of Developing FCEV Technology

Table 2. The matrix of technology characteristics and functions in FCEVs.

(a). The matrix of technology characteristics and functions in fuel cells.
	Fuel Cell	Manufacture	Power	Drive	Control	Cooling		Diagnose
anode	51	13	15	13	21	3		3
assembly	31	25	11	6	0	2		1
catalyst	103	48	59	6	6	1		0
cathode	259	60	60	40	38	8		10
device	519	212	187	60	72	55		40
electrode	66	82	42	9	3	7		1
electrolyte	120	55	107	15	4	3		1
method	41	24	9	11	42	8		13
system	138	16	53	13	4	20		2
(b). The matrix of technology characteristics and functions in electric motors.
	Charging	Controlling	Cooling	Diagnose	for Vehicle		Supplying Power
cooling system	1	19	92	3	92		44
control apparatus	26	84	11	21	389		186
control method	5	25	0	9	31		12
fuel cell stack	5	19	3	9	175		39
fuel cell system	4	13	4	8	58		30
power control	51	116	40	20	495		375
(c). The matrix of technology characteristics and functions in the control system.
		Controlling	Diagnosing	Driving	Performing		Vehicle
assembly		6	17	8	8		14
control method		0	1	1	2		0
device		1	0	1	2		1
fuel cell		6	8	8	2		11
test		0	8	4	3		2
(d). The matrix of technology characteristics and functions in hydrogen storage.
		Filling	Generating Power	Managing	Storing		Supplying
charging		2	1	3	9		3
control		3	0	3	12		1
storage		6	3	10	37		4
supply		4	1	5	15		14
tank		1	2	8	26		5

depicts the matrix of technology characteristics and functions in the key components. Table 2a shows that inventions related to fuel cell technology include a set of technology characteristics: anode, cathode, electrode, electrolyte, catalyst, and the related method or system, while these patent applications may have a series of functions, such as composing the fuel cell, manufacturing the fuel cell, and providing energy.

The given number in Table 2a is on behalf of patent applications filed in the specific technology field. For instance, the number in the first row and column is 51 in Table 2a, which means that the number of patent applications related to the anode of fuel cells is 51. The greater the number in a given technology field, the more active the innovation activities are, and vice versa. According to the matrix of technology characteristics and functions, the paper identifies the technological advances in developing FCEV technology.

Table 3. The advances and opportunities in developing FCEV technology.

Key Components	Technological Advances	Technological Opportunities
Fuel cell	fuel cell composition, manufacturing fuel cells, providing energy	controlling, cooling, and diagnosing fuel cells
Electric motors	supplying power for vehicles	cooling and diagnosing fuel cell stack
The control system	-	controlling, diagnosing, driving, performing methods, and devices
Hydrogen storage	managing and storing hydrogen	filling hydrogen

presents the technological advances based on the matrix of technology characteristics and functions.

First, there are a great number of patent applications focusing on the fuel cell composition, manufacturing fuel cells, and providing energy using fuel cells. Our finding is in line with the view of Manoharan, Hosseini, Butler, Alzhahrani, Senior, Ashuri and Krohn [2], in which the development of fuel cells has achieved great advances. Notably, Xu, Han, Zhu, Ni, and Yao [8] pointed out that the advances achieved in MS-SOFCs were thin-film deposition, active electrode catalyst preparation, protective coatings technology, etc. Fang, Vairin, von Jouanne, Agamloh, and Yokochi [7] highlighted that SOFCs could handle a variety of fuels, either hydrocarbons or hydrogen, and PEMFCs were widely applied in cars due to the quick response and low operating temperature. Moreover, hydrogen is one type of chemical energy carrier, and it has a high energy density. Fuel cells can directly convert chemical energy into electrical energy with high efficiency and continually provide operational power as long as the fuel is sufficient. Thus, great progress has been made in the technology concerned with providing energy via fuel cells.

Second, the technological advance achieved in electric motors is to supply better power for fuel cell vehicles. The electric motor is an essential component of FCEVs because vehicles are driven by electric motors. The technology of supplying power has achieved great attention and made significant progress.

Third, the advances achieved in hydrogen storage are to manage and store hydrogen. Both fuel cells and the equipment for storing hydrogen are special technology for FCEVs [5]. Particularly, the onboard hydrogen storage should be improved for developing FCEVs [2]. The technologies related to managing and storing hydrogen have made some progress.

5.3. The Opportunities for Developing FCEV Technology

The commercialization and large-scale application of FCEVs still suffer from many challenges in the reviving phase. There are numerous opportunities for developing FCEV technology in the future. Both the control system and hydrogen storage need to be further improved, from a holistic perspective. Particularly, hydrogen storage technology is unique for developing FCEVs, but onboard hydrogen storage is still one of the obstacles for fuel cell vehicles at present [2]. Moreover, the control system is very scarce, though some patent applications have been filed.

On the other hand, there are numerous opportunities for each key component of FCEVs. Table 3 also depicts the opportunities for developing FCEV technology. First, there are opportunities for developing the control, cooling, and diagnosis of fuel cells, from the perspective of technology functions. For instance, a thermal system, namely a cooling system, is a fundamental component of FCEVs, which can maintain a proper operating temperature range of FCEVs, including fuel cells, electric motors, power electronics, etc. [18]. Nowadays, inventions related to controlling, cooling, and diagnosing fuel cells are limited, which may provide great opportunities for innovators.

Second, there are many opportunities to develop the cooling and diagnosis of fuel cell stacks in the field of electric motors. The assembly of individual membrane electrodes, which apply hydrogen and oxygen to produce electricity, constitutes a fuel cell stack [18]. Either cooling or diagnosing the fuel cell stack is very important for developing electric motors. Controlling the operating temperature of fuel cells can improve the durability and dynamic performance of fuel cells [21]. However, there are different degrees of degradation during the long-term operation, which lead to the temperature-sensitivity characteristics of fuel cells gradually changing [32]. In this context, Meng et al. [33] pointed out that performance degradation during operation has greatly impeded the commercialization of FCEVs and highlighted that the prediction of fuel cells’ performance degradation could improve the diagnosis and managing fuel cells’ health.

Third, there are also rich opportunities for developing the control system and hydrogen storage. The control system manages the flow of electrical energy and controls the speed of the electric motor and the torque produced by it. Moreover, it is not easy to store hydrogen due to the low density, and hydrogen needs to be compressed and cooled. Thus, technological opportunities exist in the development of controlling, diagnosing, and performing the control system, as well as filling hydrogen.

6. Conclusions

The FCEV is a disruptive technology and promising for the future. Once this disruptive technology achieves a breakthrough, FCEV technology will immediately fulfill commercialization and have extensive applications. In the era of decarbonization and technological revolution, it is an exciting topic about what are the technological advances and opportunities for developing FCEVs. Our contribution is two-fold.

This paper proposes a novel approach to identify the key components of FCEVs by constructing directed IPC co-occurrence networks. An FCEV is a complex and integrated technology system, which needs to combine various sophisticated technologies and components. To explore the technological advances and opportunities, we should disassemble the FCEV technology into key components and then reveal the technology characteristics and functions of FCEV’s components. The traditional method of identifying key components depends on experts’ judgment. To reduce subjective judgment bias by experts to the maximum extent possible, we construct the directed IPC co-occurrence networks by distinguishing the primary and secondary IPC codes in a given patent application and identifying the critical components of a given technology field based on the results of clustering analysis. The approach of constructing directed IPC co-occurrence networks is convenient and rational for identifying key components in practice. More importantly, adopting the NLP technology makes it possible to construct sophisticated information relation matrixes based on numerous patent documents, which is almost impossible to complete through manual patent analysis. The method proposed in this paper can be used to identify advances achieved in a given technology field and detect technical opportunities for innovators.

On the other hand, this paper gains an insight into what are the technological advances in recent years and opportunities for developing FCEVs in the future. The paper highlights that the advances achieved in the fuel cell field are fuel cell composition, manufacturing fuel cells, and providing energy using fuel cells, and the advance achieved in electric motors is supplying power for fuel cell vehicles, while the advances in hydrogen storage are to manage and store hydrogen. By contrast, there are many opportunities for developing the control and diagnosis of fuel cells, performing the control system, and filling hydrogen, which may be the future research directions.

Our study also has some limitations. As innovation activities are very active in the FCEV technology field, the number of patent applications continues to increase and renew. The conclusions drawn by this paper primarily depend on patent applications between 2014 and 2023. The technological advances and opportunities are dynamic and elastic because the technological trajectories of fuel cells are fiercely competitive and technologies are iterating quickly. We continue to pay attention to the development of FCEV technology and further explore the opportunities for FCEV development.