Prioritizing Key Factors in Refrigerant Substitution for GHG Emission Reduction: An Integrated DEMATEL-ISM-MICMAC Approach

Abstract

To implement the Kigali Amendment to the Montreal Protocol, the global academic community has intensified its research on environmentally friendly refrigerant substitutes. This effort aims to effectively reduce greenhouse gas emissions and facilitate the achievement of carbon neutrality goals. In this study, 14 key influencing factors were identified through the Delphi method, and the Decision-making Trial and Evaluation Laboratory (DEMATEL) approach was innovatively applied to systematically analyze the interrelationships among these factors. The results indicate that technological innovation related to refrigerant substitution ranks first with a centrality score of 5.429, confirming it as the core driving factor for refrigerant substitution. Subsequently, through the integration of Interpretive Structural Modeling (ISM) and Cross-impact Matrix Multiplication Applied to Classification (MICMAC), a hierarchical structure of influencing factors was further developed. This clarified high-driving factors such as government policies and life-cycle costs, as well as highly interrelated factors including climate conditions, greenhouse gas emissions, and performance coefficients. The key contribution of this paper is its success in overcoming the limitations of single-factor analysis by integrating multiple dimensions of influencing factors to construct a hierarchical classification. This innovative and systematic theoretical framework not only offers a scientific basis and decision-making support for refrigerant substitution but also possesses substantial theoretical value and practical guidance. Furthermore, it serves as an essential reference for advancing the development of low-carbon refrigeration technologies.

1. Introduction

As global climate change continues to intensify, the challenge of reducing greenhouse gas emissions in the refrigeration industry has increasingly drawn significant attention from various sectors of society [1]. Traditional refrigerants, such as hydrofluorocarbons (HFCs), although not ozone-depleting in their chemical structure, possess a significant global warming potential (GWP) that is critical to address in the context of climate change. The continued use of these refrigerants contributes to the accumulation of greenhouse gases and has profound implications for global climate systems [2]. In response to the increasingly severe environmental challenges, the international community has enacted pivotal agreements, such as the Montreal Protocol and its subsequent Kigali Amendment, which actively promote the adoption of green alternatives to refrigerants [3,4]. The European Union introduced the F-gas Regulation, which aims to reduce the use of fluorinated refrigerants, minimize the leakage of fluorinated gases, and impose stringent restrictions on the application of fluorinated greenhouse gases [5,6]. The U.S. Environmental Protection Agency (EPA), under the authority of the Clean Air Act (CAA), is mandated to phase out high-GWP refrigerants and promote the adoption of environmentally friendly alternative technologies [7]. As the world’s largest producer and consumer of HFCs, China released the China Action Plan for the Phase-Out of HFCs in 2023, clearly outlining the industry-specific roadmap for the phased reduction in HFCs [8]. As a critical factor influencing greenhouse gas emissions, the research on refrigerant substitution has become increasingly urgent, and it is imperative to identify sustainable alternatives for refrigerants [9].

From the perspective of current research, many scholars have focused on the toxicity, flammability, and other physicochemical properties of refrigerants [10], there are also scholars who study the performance of various environmentally friendly refrigerants and alternative technologies [11]. However, refrigerant substitution is a complex process that involves the interplay of economy, policy, technology, market, environment, and other dimensions. Few studies comprehensively analyze the influencing factors of refrigerant substitution from multiple perspectives, and there is a lack of in-depth discussion on the interaction among multiple factors as well as the identification of key factors driving refrigerant substitution.

2. Literature Review

With the intensification of global climate change and the environmental issues arising from the refrigeration industry, refrigerant replacement has emerged as a critical research area of widespread concern. Currently, the quest for environmentally friendly alternative refrigerants has become a focal point of international attention [12]. The international community, by means of agreements such as the Montreal Protocol and the Kigali Amendment, has established specific restrictions on the use of refrigerants while promoting the research, development, and widespread adoption of low-GWP alternatives [13]. A refrigerant with a GWP of less than 150 can be classified as a low-GWP refrigerant [14]. Low-GWP HFCs, hydrofluoroolefins (HFOs), hydrochlorofluoroolefins (HCFOs), and natural working fluids (such as hydrocarbons, ammonia, CO₂, and water) are recommended as environmentally friendly refrigerants [15]. At the national level, governments have implemented a range of policies to facilitate the adoption of green refrigerants, such as providing subsidies, offering tax incentives, and allocating funding for research and development, thereby accelerating the industry’s transition [16]. New low GWP refrigerants, including natural refrigerants (e.g., carbon dioxide and ammonia) and novel synthetic refrigerants (e.g., HFOs), demonstrate superior performance and enhanced safety compared to traditional refrigerants [17,18,19]. Shen B et al. assessed the performance of R290 (propane) and R600a (isobutane) as potential replacements for R134a (an HFC) in heating systems [20]. Li Z et al. investigated the optimal combination of natural refrigerants for the two-stage Rankine cycle LNG cold energy power generation system. At a conveying pressure of 0.6 MPa, the “ethylene + propane” refrigerant combination exhibits the best power generation performance. With a net generating capacity of 9020.8 kW and an energy efficiency of 35.54%, respectively, the “ethane + propane” refrigerant combination is identified as the most suitable option to achieve both optimal power generation performance and economic performance under high-pressure transmission conditions [21]. By optimizing the heat exchanger design, Kim B et al. analyzed the superior performance of low-GWP refrigerants in comparison to traditional R-410A heat pump systems [22]. Research conducted by Nawaz K et al. demonstrates that heat pump systems utilizing R1234yf and R1234ze(E) can achieve performance levels comparable to those of conventional refrigerants, exhibiting strong capabilities in both cooling and heating capacities. Additionally, these refrigerants offer enhanced safety during charging, with reduced risks of fire and toxicity, making them highly suitable for residential applications [23]. Chae J H et al. investigated the impact of varying the charge amounts of different refrigerants on the performance of refrigeration or heat pump systems under two specific operating modes: steady-state and heating mode [24]. However, the challenge of refrigerant-related technology transformation lies in the assessment of the adaptability, cost, and safety of the new technology, which all affect market acceptance [25]. Market factors play a crucial role in refrigerant replacement. Specifically, fluctuations in market demand, consumer awareness of environmentally friendly products, and competitive dynamics within the industry can significantly influence the adoption and promotion of new refrigerants [26]. According to the research by Konrad, M. E and Macdonald, B. D enterprises exhibit caution in adopting low GWP refrigerants primarily due to the substantial costs associated with the development and transformation of new technologies [25]. Additionally, there is a need to establish an effective market incentive mechanism to enhance market acceptance and facilitate the promotion of environmentally friendly refrigerants [27]. Fu, X. and X. Yan, et al. screened the refrigerants R32, R1270, R290, R161, R600a, and R600 mixed with CO₂ based on their environmental impact, critical point, and safety data. The mass fractions of these refrigerants were set at 0.14, 0.12, 0.1, 0.06, 0.04, and 0.02, respectively. Among them, R32 was found to be the most recommended option in most cases due to its superior efficiency and lower levelized power costs [28]. The environmental impact assessment of refrigerant substitutes plays a crucial role in evaluating their sustainability. The environmental benefits of alternative refrigerants are determined not only by their GWP but also by the technological processes involved in their production and consumption patterns during usage, as demonstrated through a life-cycle assessment (LCA) [29], encompassing the environmental impacts during manufacturing, use, and disposal stages [30]. Effective environmental assessment methods can assist decision-makers in selecting appropriate alternatives to minimize the impacts of climate change. For instance, the environmental analysis of steam compression systems considers both indirect emissions resulting from the carbon emission factor of the power mix and direct emissions caused by refrigerant leaks [31]. Zhou, M. et al. reported that studies have demonstrated the environmental issues caused by refrigerants can adversely impact human health, emphasizing the need for efforts to achieve a balance between reducing greenhouse gas emissions and maintaining public health [32]. In addition, economic factors constitute another critical dimension influencing refrigerant replacement [33]. Specifically, high initial investment costs, a shortage of qualified maintenance technicians, concerns over return on investment, and the overall market economic situation can all impact enterprise decision-making [34,35]. Saif, A. and S. Elhedhli investigated the capacity, transportation, and costs of cold supply chains while considering the impacts of global warming caused by greenhouse gas emissions, utilizing low-GWP refrigerants as a potential solution [36]. In terms of economic viability, despite the high initial investment required for many low-GWP refrigerants, these costs are typically offset by savings in long-term operating expenses and the substantial environmental benefits they provide [37,38]. It is evident that refrigerant replacement constitutes a complex and systematic endeavor. While relevant scholars have investigated numerous single influencing factors in the context of refrigerant replacement, this process inherently entails the interplay of multiple factors, including policy, technology, market dynamics, environmental considerations, and economic implications [39]. The existing literature exhibits a notable gap in systematic research regarding the relationship between factors influencing refrigerant substitution. Based on this, this paper employs the DEMATEL-ISM-MICMAC method to systematically analyze the importance of each influencing factor in refrigerant substitution and the interrelationships among these factors. An explanatory structural model is constructed to clarify the hierarchical relationships among the influencing factors, as well as the influence paths and action mechanisms among them. Subsequently, the MICMAC method is utilized to conduct a driving force-dependency analysis of the factors, enabling further stratification and classification of the influencing factors. This provides a robust theoretical foundation and practical guidance for related research while exploring actionable policy recommendations to facilitate the green transformation of the refrigeration industry.

3. Research Method

Based on the integrated DEMATEL-ISM-MICMAC research methodology, this paper identifies the influencing factors of refrigerant substitution through a literature review, questionnaire surveys, and the Delphi method. It constructs a DEMATEL-ISM-MICMAC model to determine the key influencing factors of refrigerant substitution and provides strategic directions for the implementation of environmentally friendly refrigerants, thereby accelerating the green transformation of the refrigeration industry.

The Decision-making Trial and Evaluation Laboratory (DEMATEL) is a method proposed in 1973 by American scholar Gabriel Fontela and others. It employs graph theory and matrix tools for analyzing system factors. Through the logical relationships and direct influence matrix among factors within a system, DEMATEL calculates the degree to which each factor influences others and the degree to which it is influenced. This allows for determining the centrality and causality of each factor [40]. Its primary advantage lies in its ability to utilize matrices with numerical directions for assessing complex system factors, thereby determining the importance and influence relationships among these factors. Interpretative Structural Modeling (ISM) is a systematic modeling technique introduced by J. Warfield in 1973, specifically designed for analyzing problems associated with complex social and economic systems. It is characterized by the decomposition of a complex system into multiple subsystem elements. By leveraging practical human experience and knowledge, computer technology, graph theory, and matrix tools, it ultimately forms a multi-level hierarchical structure model [41]. Its primary advantage lies in its ability to leverage matrix operations for a deep deconstruction of the relationships among system elements, presenting them clearly via a hierarchical model. Matrix Impacts Cross-reference Multiplication Applied to a Classification (MICMAC) is utilized to further examine the correlation and significance of risk factors. This approach leverages the reaction pathways and hierarchical cycles of factors within the system to investigate the diffusivity of their interrelationships, subsequently classifying them based on the driving force and dependence of the influencing factors [42]. The primary advantage of this method lies in its ability to summarize factors into a quadrantal diagram based on the driving force-dependence degree, thereby clearly illustrating the relationships among factors through four distinct sets. The DEMATEL method, ISM method, and MICMAC method are frequently employed by researchers to analyze and manage complex systems.

DEMATEL is specifically designed to analyze the causality and strength of influence among factors, effectively identifying key drivers by means of visualized central-causality graphs. ISM focuses on constructing hierarchical structures by decomposing complex systems into multi-level hierarchical models, thereby elucidating the dependencies and conduction paths among factors in a systematic manner. MICMAC is grounded in drive-dependency analysis, which categorizes factors (such as autonomy, dependence, linkage, and independence) to assess the stability and sensitivity of a system. In recent years, the DEMATEL and MICMAC methods have been frequently employed in the engineering research domain and are often integrated with the ISM method. The integration of the three methods to create complementary and synergistic effects facilitates elucidating the role of each factor within a complex system, synthesizing causality and hierarchical structure, performing stability analysis, establishing a multidimensional systems perspective, and uncovering the comprehensive landscape of “cause-effect-hierarchy-stability”, thereby enhancing both the depth and breadth of holistic analysis [43]. At present, there are limited studies in the academic community focusing on the integration of these three methods. Guo X, Yang Z, and Zhang SY employed both DEMATEL and ISM approaches to analyze the intricate relationships within the supply chain. By constructing a causal relationship diagram, they were able to clearly pinpoint the key factors influencing supply chain performance as well as their interdependencies [44]. Kumar, Dinesh, Mangla, et al. propose an integrated DEMATEL and MICMAC approach designed to identify and analyze the factors influencing sustainable manufacturing practices, thereby offering decision-makers practical guidance for the implementation of sustainable manufacturing [45]. Zhu C, Zhu N, Zheng S, et al. assessed the challenges encountered by China’s manufacturing sector in green supply chain management using the DEMATEL and ISM methodologies [46]. Kumar S, et al. applied DEMATEL and ISM to analyze the factors influencing the adoption of mobile banking. This study offers a critical decision-making foundation for banks and financial institutions to enhance user adoption rates [47]. Ebadi Torkayesh, A. and S. Hendiani, et al.combined DEMATEL and MICMAC in the field of risk management for renewable energy projects [48]. This study, for the first time, applied the DEMATEL-ISM-MICMAC method in the field of refrigerant substitution. The aim was to analyze the importance and internal correlations of influencing factors, elucidate the hierarchical structure and overall influence relationships among these factors, and identify the key transmission paths that significantly impact refrigerant substitution and are broadly interconnected. Furthermore, this approach enhances the scientific rigor of research on refrigerant substitution, providing a theoretical foundation and practical guidance for the development of refrigerants in China. The research concept is illustrated in Figure 1 (The arrows depict the sequential order of research steps undertaken in this paper).

4. Identification of Influencing Factors

4.1. Comprehensive Collection and Analysis of Relevant Literature

In this study, the relevant literature was systematically retrieved from the Web of Science Core Collection (WOSCC) database to ensure comprehensive coverage of the research field. The search topics encompassed keywords such as “refrigerant substitution” and “influencing factors of refrigerant substitution”. The reference period for the literature in this study spans from 2000 to 2024, during which a total of 1159 articles related to refrigerant substitution were retrieved. After summarizing the retrieved literature (see Figure 2), it was observed that the number of publications has been increasing continuously and rapidly since 2011, indicating a rising research interest among experts and scholars in the field of refrigerant substitution.

Through the screening of the collected literature, we selected the studies with high relevance to this paper and excluded those unrelated to our research focus. Based on the citation frequency of the literature and a thorough analysis of their content, 29 studies focusing on refrigerant substitution or the influential factors related to this process were selected.

4.2. Preliminary Screening of Influencing Factors for Refrigerant Substitution

After completing the literature selection, the above 29 relevant studies were intensively read, the factors involved in each article were summarized and extracted (

Table 1. Literature sources on influencing factors of refrigerant substitution.

Serial Numbers	Influencing Factors	Literature Sources
1	The level of market demand for refrigerants	[8,39]
2	Awareness of social environmental protection	[39]
3	Awareness of consumer health	[32]
4	Consumers’ willingness to adopt environmentally friendly refrigerants	[39]
5	Implement the government’s supportive policies for environmental protection of refrigerants	[6,9,10,21,25,34]
6	Greenhouse gas emissions from equipment utilizing environmentally friendly refrigerants	[1,2,29,38]
7	Refrigerants ozone depletion potential (ODP)	[19,34]
8	Refrigerant global warming potential (GWP)	[2,19,22,34,38]
9	Specific climatic conditions	[1,2,11,12,18,19,31]
10	Thermal characteristics of refrigerants	[2,3,4,9,10,18,31]
11	The market price of environmentally friendly refrigerants	[20,25,33,34]
12	The amount of refrigerant charged	[1,5,16,17,21]
13	The amount of refrigerant leakage	[2,37]
14	The flammability of refrigerants	[3,5,23,34]
15	The toxicity of refrigerants	[3,25,28,34]
16	The cost of equipment transformation for environmental protection and refrigerant replacement	[3,21,23,25,29,34]
17	Technological innovation in refrigerant substitution	[25]
18	The Stability of the Supply Chain for Environmentally Friendly Refrigerants	[35]
19	System Performance Evaluation of Equipment Utilizing Environmentally Friendly Refrigerants (COP)	[2,4,11,12,20,23,25]
20	Full Life-Cycle Cost Analysis of Equipment Using Environmentally Friendly Refrigerants	[29]
21	Optimization of Key Equipment Components Utilizing Environmentally Friendly Refrigerants	[10,20,22,23,25]

), and 21 influencing factors for this study were preliminarily sorted out.

4.3. Identification of Key Influencing Factors for Refrigerant Substitution

In order to enhance the accuracy of this study on influencing factors, a questionnaire survey was employed to further investigate and refine the initially extracted influencing factors of refrigerant substitution. A 5-level Likert scale was utilized for this evaluation [49]. Based on their own experience, the respondents assess and evaluate the impact of each factor on refrigerant substitution.

A total of 360 questionnaires were distributed, and 354 valid responses were collected. In the basic information section of the first part of the questionnaire, respondents were surveyed regarding their understanding of refrigerant substitution. Out of all respondents, 27 indicated no knowledge of refrigerant substitution, accounting for 7.63% of the total. Due to this lack of familiarity with the topic, this portion of the questionnaire was considered invalid. Therefore, a total of 327 valid questionnaires were collected, resulting in an effective response rate of 90.83%. Reliability testing of the survey questionnaire revealed a Cronbach’s alpha value of 0.900, indicating that the questionnaire exhibits excellent internal consistency and satisfies the criteria for reliability analysis. The validity test of the questionnaire reveals that the KMO coefficient is 0.950, and the p-value associated with Bartlett’s test is less than 0.05, which indicates that the questionnaire data exhibit satisfactory construct validity.

This rating utilizes a 5-point Likert scale. If the average score of a factor is below 3, it indicates that the factor has minimal influence on refrigerant substitution. According to the average score presented in the table, all seven criteria, the refrigerant GWP, the refrigerant ODP, and thermal characteristics of refrigerants, as well as the amount of refrigerant charged, the amount of refrigerant leakage, the flammability of refrigerants, and the toxicity of refrigerants, scored less than 3. Consequently, these seven factors will be excluded from further consideration (

Table 2. Scoring table for influencing factors in refrigerant substitution.

Serial Numbers	Influencing Factors	Mean Score
1	Refrigerant global warming potential (GWP)	2.4
2	Refrigerants ozone depletion potential (ODP)	2.5
3	The amount of refrigerant leakage	2.7
4	The amount of refrigerant charged	2.7
5	The flammability of refrigerants	2.7
6	Thermal characteristics of refrigerants	2.7
7	The toxicity of refrigerants	2.8
8	Implement the government’s supportive policies for environmental protection of refrigerants	3.0
9	Awareness of consumer health	3.0
10	Awareness of social environmental protection	3.0
11	The Stability of the Supply Chain for Environmentally Friendly Refrigerants	3.1
12	Consumers’ willingness to adopt environmentally friendly refrigerants	3.1
13	System Performance Evaluation of Equipment Utilizing Environmentally Friendly Refrigerants (COP)	3.1
14	Technological innovation in refrigerant substitution	3.1
15	Optimization of Key Equipment Components Utilizing Environmentally Friendly Refrigerants	3.2
16	The cost of equipment transformation for environmental protection and refrigerant replacement	3.5
17	Greenhouse gas emissions from equipment utilizing environmentally friendly refrigerants	3.5
18	System Performance Evaluation of Equipment Utilizing Environmentally Friendly Refrigerants (COP)	3.6
19	Full Life-Cycle Cost Analysis of Equipment Using Environmentally Friendly Refrigerants	3.6
20	The market price of environmentally friendly refrigerants	3.6
21	Specific climatic conditions	3.6

).

Through a rigorous screening process, 14 key influencing factors for refrigerant replacement have been identified. The specific factors and their corresponding descriptions are detailed in

Table 3. Influencing factors and descriptions for refrigerant substitution.

Serial Numbers	Influencing Factors	Instructions
S1	The market price of environmentally friendly refrigerants	Market retail price of environmentally friendly refrigerants.
S2	The level of market demand for refrigerants	The total refrigerant demand from potential buyers in the market and its intensity level.
S3	Specific climatic conditions	General overview of the multi-year weather characteristics in the region where environmentally friendly refrigerant equipment is utilized.
S4	Awareness of consumer health	The extent to which consumers prioritize their own health, along with the attitudes, beliefs, and actions they adopt to maintain and enhance it.
S5	Awareness of social environmental protection	Individual and societal awareness and understanding of the critical importance of environmental protection.
S6	Consumers’ willingness to adopt environmentally friendly refrigerants	The tendency or willingness of consumers to choose environmentally friendly refrigerants when purchasing and utilizing refrigeration equipment.
S7	Technological innovation in refrigerant substitution	New technologies and methods have been developed, applied, and continuously improved to replace traditional refrigerants, such as HFCs, in the refrigeration and air conditioning industry.
S8	Implement the government’s supportive policies for environmental protection of refrigerants	The government has implemented specific policies aimed at facilitating the replacement of refrigerants.
S9	Greenhouse gas emissions from equipment utilizing environmentally friendly refrigerants	Greenhouse gas emissions, either directly or indirectly resulting from the use of environmentally friendly refrigerants in the operation of refrigeration equipment, encompass gases released during refrigerant leakage as well as those associated with equipment manufacturing, transportation, installation, maintenance, and disposal.
S10	Optimization of Key Equipment Components Utilizing Environmentally Friendly Refrigerants	Improve or upgrade key components that directly influence the cooling performance and energy efficiency of equipment by utilizing environmentally friendly refrigerants, thereby enhancing the overall performance and efficiency of the equipment.
S11	The Stability of the Supply Chain for Environmentally Friendly Refrigerants	All links in the supply chain can ensure stable and continuous operation when providing environmentally friendly refrigerants and are less susceptible to internal and external interference or changes.
S12	The cost of equipment transformation for environmental protection and refrigerant replacement	The total cost of replacing conventional refrigerants (e.g., R22, R134a, etc.) with environmentally friendly alternatives (e.g., R410A, R32, R1234yf, etc.) in existing refrigeration and air conditioning systems includes several factors. Retrofit costs are influenced by the type and size of the equipment, the compatibility of existing systems, and the specific choice of environmentally friendly refrigerants.
S13	Eco-friendly refrigerant system performance (COP)	The ratio of heat output by the equipment to the electricity consumed during operation when using environmentally friendly refrigerants is defined as the coefficient of performance (COP). A higher COP value indicates better energy efficiency of the refrigeration equipment.
S14	Full Life-Cycle Cost Analysis of Equipment Using Environmentally Friendly Refrigerants	From equipment planning, design, manufacturing, procurement, installation, operation, maintenance, renovation, and updating to eventual scrapping, all related costs throughout the entire life-cycle should be considered.

.

5. DEMATEL-ISM-MICMAC Model Construction

The DEMATEL-ISM-MICMAC integrated model serves as a systematic analysis tool that merges three methodologies (DEMATEL, ISM, and MICMAC). This model primarily focuses on analyzing the interdependencies among factors within complex systems, pinpointing critical driving factors, classifying management strategy domains, and constructing a hierarchical structural framework [50]. By conducting a quantitative analysis of influencing factors, structural decomposition, and strategic classification, we provide decision-makers with comprehensive support ranging from the identification of factor relationships to action planning. Additionally, we offer researchers a well-defined framework for study, thereby ensuring the scientific rigor and systematic nature of the research.

5.1. DEMATAL Model Construction

The DEMATEL model is constructed based on the initial direct influence matrix A, followed by the establishment of the normalized direct influence matrix F and the comprehensive influence matrix T. Ultimately, the model calculates both centrality and causality metrics [51].

Invite experts and scholars from relevant fields to evaluate the degree of interrelationship among various influencing factors in refrigerant substitution. To derive the initial correlation information for these influencing factors, an average calculation is conducted on the expert scoring matrix to construct the initial direct influence matrix A = [a_ij]_n×n (where n represents the number of influencing factors). Here, a_ij denotes the extent to which factor i influences factor j, while the diagonal elements a_ii, representing the self-influence of each factor, are all set to 0.

To address the issue of non-homogeneous dimensionality among various influencing factors in the initial direct impact matrix A, the matrix is standardized through a systematic process. First, the sums of each row and column in matrix A are calculated, and the maximum value, denoted as c, is determined. Subsequently, each element in matrix A is divided by this maximum value c, yielding normalized data that is used to construct the standardized direct impact matrix F, as illustrated in Formulas (1) and (2).

$$F=A/c$$

(1)

$$c=\max [\underset{1<i<m}{{\displaystyle \max }}{\displaystyle \sum _{i=1}^m{a}_{ij}},\left.\underset{1<i<nm}{{\displaystyle \max }}{\displaystyle \sum _{j=1}^n{a}_{ij}}\right]$$

(2)

Based on Formula (2), the comprehensive impact matrix T is constructed to further examine the indirect relationships among various factors, where E denotes the identity matrix.

$$T=F{(E-F)}^{-1}$$

(3)

The comprehensive impact matrix T partially reflects the inherent connections among various influencing factors related to refrigerant substitution. However, an imbalance in information remains evident. Consequently, the concepts of centrality and causality are incorporated to calculate the importance weights of each influencing factor more accurately. Calculate the row sums D and column sums C of the comprehensive influence matrix T. Here, D denotes the extent to which a factor influences other factors, with higher values indicating stronger influence. Conversely, C reflects the extent to which a factor is influenced by others, where larger values signify greater susceptibility to influence. D + C represents the centrality of factor j within the system, with higher values suggesting greater systemic importance. D-C indicates the causal degree of factor j; a positive value implies that the factor is primarily causal, exerting more influence on others, whereas a negative value suggests that the factor is predominantly resultant, being more influenced by others [52].

5.2. ISM Model Construction

The construction of the ISM model is grounded in the comprehensive influence matrix T, where the adjacency matrix B and the reachability matrix H are derived. This process culminates in the creation of an interpretive structural model that visually represents the logical structural relationships among various factors through a clearly defined structural diagram [53].

Based on the data in the comprehensive impact matrix T, the average value of all entries in the comprehensive impact matrix T can be calculated to obtain the threshold. According to Formula (4), the comprehensive impact matrix T is added to the identity matrix I to derive the overall impact matrix B.

$$H=(B+I{)}^{\mathrm{n}+1}={(B+1)}^n\ne B+I$$

(4)

The elements b_ij in the overall impact matrix B are defined as shown in Equation (5).

$$b_{ij}=\left\{\begin{array}{c}0,b_{ij }<\lambda \\ 1,b_{ij} \ge \lambda \end{array}\right. (\mathrm{i}=1,2,\dots ,15; \mathrm{j}=1,2,\dots ,15)$$

(5)

When the overall impact matrix B satisfies Formula (5), the reachability matrix H is obtained.

Based on the reachability matrix H, the reachability set R (Si), the antecedent set A (Sj), and the common set C (Si) are determined. Specifically, the reachability set R (Si) consists of all elements in each row of the reachability matrix H that equal 1, whereas the antecedent set A (Sj) includes all elements in each column of H that equal 1. The common set C (Si) is defined as the intersection of R (Si) and A (Sj), expressed as C (Si) = R (Si)∩A (Sj). This process ultimately facilitates the construction of the ISM structural model for analyzing the influencing factors related to refrigerant substitution.

$$\mathrm{R}\left({\mathrm{S}}_{\mathrm{i}}\right)=\left\{{\mathrm{S}}_{\mathrm{i}}\in \mathrm{S}|{\mathrm{m}}_{\mathrm{i}\mathrm{j}=1}\right\}$$

(6)

$$\mathrm{A}\left({\mathrm{S}}_{\mathrm{i}}\right)=\left\{{\mathrm{S}}_{\mathrm{i}}\in \mathrm{S}|{\mathrm{m}}_{\mathrm{i}\mathrm{j}=1}\right\}$$

(7)

$$\mathrm{C}\left({\mathrm{S}}_{\mathrm{i}}\right)=\mathrm{R}\left({\mathrm{S}}_{\mathrm{i}}\right)\cap \mathrm{A}\left({\mathrm{S}}_{\mathrm{i}}\right)$$

(8)

5.3. MICMAC Model Construction

The MICMAC model quantifies the degree of influence a factor exerts on other factors, referred to as the driving force Q_i, by calculating the sum of each row in the reachable matrix H. It measures the degree to which a factor is influenced by others, referred to as dependence Y_i, by calculating the sum of each column in the same matrix H. Through this approach, the driving force-dependence matrix for the influencing factors of refrigerant substitution is constructed. Subsequently, factors are categorized into autonomous factors, dependent factors, linkage factors, and independent factors, thereby elucidating the specific roles each factor plays within the system [54].

$$Q_{\mathrm{i}}={\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}+1}{m}_{ij}}$$

(9)

$$Y_{\mathrm{i}}={\displaystyle \sum _{j=1}^{\mathrm{n}+1}{m}_{\mathrm{ij}}}$$

(10)

6. Analysis Results

This article identifies 14 key factors influencing refrigerant replacement based on a literature review and questionnaire surveys. These factors are not independent of one another but rather exhibit a complex interrelationship. It remains unclear which factors should be prioritized as core elements, the hierarchical relationships among these factors, and which factors require close monitoring as unstable variables. The application of the DEMATEL-ISM-MICMAC model in the empirical analysis of influencing factors for refrigerant substitution enables a comprehensive exploration of the significance and interdependencies of each factor. By systematically analyzing the interactions among factors, the model elucidates their hierarchical and logical relationships, clearly illustrating the ranking of factor importance and the mechanisms through which they exert influence. Furthermore, it performs an in-depth driving force and dependence analysis of the influencing factors, evaluating the stability and vulnerability of different factors. This provides enterprises with a robust and insightful analytical foundation for making informed refrigerant substitution decisions.

6.1. Analysis Utilizing the DEMATEL Method

This article utilizes the Delphi method, engaging experts and scholars from universities, research institutions, and enterprises to form an expert panel. The panel will evaluate the degree of mutual influence among the identified factors affecting refrigerant alternatives by employing a scoring system ranging from 0 to 4 points (where 0 indicates no impact, 1 indicates low impact, 2 indicates moderate impact, 3 indicates high impact, and 4 indicates extremely high impact). To acquire preliminary correlation information on the various factors influencing refrigerant alternatives, the expert scoring matrix is averaged to construct the initial direct influence matrix A (

Table A1. Initial direct impact matrix A for factors influencing refrigerant replacement.

	S1	S2	S3	S4	S5	S6	S7	S8	S9	S10	S11	S12	S13	S14
S1	0.000	3.083	0.000	1.583	0.417	2.917	3.500	2.000	0.000	2.250	3.083	2.917	0.000	3.417
S2	2.500	0.000	0.000	1.250	0.333	1.083	2.667	2.167	0.000	1.333	2.333	2.000	0.000	2.250
S3	0.583	1.000	0.000	2.000	2.833	2.250	1.167	0.750	0.000	2.667	0.083	1.667	2.583	1.250
S4	2.167	2.917	0.000	0.000	3.500	3.000	2.000	2.000	0.000	0.917	1.250	0.917	1.000	0.917
S5	1.083	2.333	0.000	2.333	0.000	3.000	2.083	1.583	0.000	1.000	0.750	0.917	0.000	0.583
S6	1.750	3.500	0.000	1.667	1.417	0.000	2.667	2.250	0.000	0.917	1.583	1.583	0.667	0.667
S7	1.750	3.583	0.000	0.667	1.250	1.667	0.000	2.250	0.000	3.333	3.500	3.667	1.000	2.500
S8	1.500	3.667	0.000	2.500	2.833	3.333	2.833	0.000	0.000	2.250	2.583	2.917	0.500	1.583
S9	0.833	1.750	0.000	2.583	2.583	2.333	1.833	2.000	0.000	0.333	0.500	0.667	0.000	0.083
S10	1.667	1.500	0.000	0.583	0.333	1.000	2.500	1.000	0.000	0.000	0.833	2.667	2.917	2.583
S11	1.750	0.667	0.000	0.667	0.333	0.667	2.083	1.000	0.000	0.750	0.000	2.500	0.667	2.000
S12	1.333	2.250	0.000	1.167	1.250	2.333	2.333	2.250	0.000	1.333	2.667	0.000	0.750	2.833
S13	3.000	2.750	0.000	4.083	2.500	3.167	3.500	3.167	0.000	3.167	2.500	2.000	0.000	1.917
S14	3.583	2.583	0.000	1.750	1.000	3.083	3.500	2.583	0.000	1.667	2.083	3.500	1.000	0.000

). Based on the initial direct influence matrix A of various influencing factors replaced by the refrigerant, standardization is carried out using Formulas (1) and (2) to construct the standardized direct influence matrix F (

Table A2. Standardized direct impact matrix F for factors influencing refrigerant replacement.

	S1	S2	S3	S4	S5	S6	S7	S8	S9	S10	S11	S12	S13	S14
S1	0.000	0.097	0.000	0.050	0.013	0.092	0.110	0.063	0.000	0.071	0.097	0.092	0.000	0.108
S2	0.079	0.000	0.000	0.039	0.010	0.034	0.084	0.068	0.000	0.042	0.073	0.063	0.000	0.071
S3	0.018	0.031	0.000	0.063	0.089	0.071	0.037	0.024	0.000	0.084	0.003	0.052	0.081	0.039
S4	0.068	0.092	0.000	0.000	0.110	0.094	0.063	0.063	0.000	0.029	0.039	0.029	0.031	0.029
S5	0.034	0.073	0.000	0.073	0.000	0.094	0.066	0.050	0.000	0.031	0.024	0.029	0.000	0.018
S6	0.055	0.110	0.000	0.052	0.045	0.000	0.084	0.071	0.000	0.029	0.050	0.050	0.021	0.021
S7	0.055	0.113	0.000	0.021	0.039	0.052	0.000	0.071	0.000	0.105	0.110	0.115	0.031	0.079
S8	0.047	0.115	0.000	0.079	0.089	0.105	0.089	0.000	0.000	0.071	0.081	0.092	0.016	0.050
S9	0.026	0.055	0.000	0.081	0.081	0.073	0.058	0.063	0.000	0.010	0.016	0.021	0.000	0.003
S10	0.052	0.047	0.000	0.018	0.010	0.031	0.079	0.031	0.000	0.000	0.026	0.084	0.092	0.081
S11	0.055	0.021	0.000	0.021	0.010	0.021	0.066	0.031	0.000	0.024	0.000	0.079	0.021	0.063
S12	0.042	0.071	0.000	0.037	0.039	0.073	0.073	0.071	0.000	0.042	0.084	0.000	0.024	0.089
S13	0.094	0.087	0.000	0.129	0.079	0.100	0.110	0.100	0.000	0.100	0.079	0.063	0.000	0.060
S14	0.113	0.081	0.000	0.055	0.031	0.097	0.110	0.081	0.000	0.052	0.066	0.110	0.031	0.000

). Then, the comprehensive influence matrix T (

Table A3. Comprehensive impact matrix T for factors influencing refrigerant alternatives selection.

	S1	S2	S3	S4	S5	S6	S7	S8	S9	S10	S11	S12	S13	S14
S1	0.139	0.268	0.000	0.147	0.100	0.232	0.283	0.198	0.000	0.185	0.244	0.256	0.057	0.241
S2	0.175	0.133	0.000	0.112	0.075	0.144	0.213	0.166	0.000	0.129	0.184	0.187	0.041	0.173
S3	0.123	0.169	0.000	0.147	0.159	0.186	0.174	0.130	0.000	0.171	0.113	0.171	0.124	0.140
S4	0.174	0.235	0.000	0.088	0.177	0.212	0.208	0.173	0.000	0.124	0.160	0.161	0.071	0.138
S5	0.118	0.185	0.000	0.134	0.059	0.182	0.174	0.134	0.000	0.103	0.117	0.130	0.034	0.103
S6	0.155	0.238	0.000	0.128	0.110	0.112	0.214	0.171	0.000	0.119	0.164	0.173	0.059	0.127
S7	0.189	0.276	0.000	0.122	0.121	0.195	0.181	0.202	0.000	0.214	0.252	0.273	0.087	0.216
S8	0.184	0.289	0.000	0.179	0.174	0.248	0.267	0.141	0.000	0.185	0.229	0.252	0.071	0.188
S9	0.105	0.164	0.000	0.143	0.138	0.163	0.161	0.143	0.000	0.080	0.104	0.115	0.030	0.080
S10	0.159	0.184	0.000	0.102	0.080	0.148	0.216	0.141	0.000	0.096	0.146	0.209	0.130	0.187
S11	0.130	0.122	0.000	0.079	0.060	0.106	0.165	0.110	0.000	0.092	0.089	0.170	0.051	0.141
S12	0.156	0.215	0.000	0.122	0.111	0.193	0.220	0.181	0.000	0.138	0.205	0.140	0.068	0.198
S13	0.260	0.309	0.000	0.252	0.191	0.283	0.330	0.267	0.000	0.243	0.263	0.267	0.070	0.231
S14	0.249	0.269	0.000	0.162	0.125	0.250	0.296	0.224	0.000	0.178	0.227	0.281	0.087	0.152

) is constructed according to Formula (3), and, finally, the centrality and causality are calculated and analyzed.

Based on the summation of each row D and each column C in the comprehensive influence matrix T, the centrality (D + C) and causality (D-C) values are computed, as presented in Table 1. The scatter plot depicting the relationship between centrality and causality for various factors involved in refrigerant substitution is shown in Figure 3. The higher the centrality of a factor, the more critical its role within the system. A stronger core degree indicates a higher cause degree, which in turn signifies a greater influence of the factor on other factors. Factors with a cause degree greater than 0 are referred to as cause factors, whereas those with a cause degree less than 0 are termed result factors [55]. From the analysis of

Table 4. Centrality and causal degree of influencing factors in refrigerant substitution.

	Degree of Influence (D Value)	Degree of Affected (C Value)	Degree of Centrality (D + C Value)	Degree of Reason (D-C Value)
S1	2.350	2.315	4.665	0.035
S2	1.733	3.057	4.790	−1.324
S3	1.808	0.000	1.808	1.808
S4	1.921	1.917	3.838	0.004
S5	1.472	1.683	3.155	−0.210
S6	1.770	2.654	4.424	−0.883
S7	2.328	3.101	5.429	−0.773
S8	2.407	2.381	4.788	0.026
S9	1.425	0.000	1.425	1.425
S10	1.799	2.057	3.856	−0.259
S11	1.315	2.498	3.813	−1.182
S12	1.947	2.784	4.731	−0.837
S13	2.965	0.980	3.945	1.985
S14	2.501	2.314	4.815	0.187

and Figure 3, it is evident that the centrality score of technology innovation related to refrigerant replacement (S₇) is 5.429, ranking first among all factors. This suggests that technological innovation in refrigerant replacement serves as the most critical core factor influencing the transition to alternative refrigerants. Strengthening efforts in this area of technology innovation can effectively facilitate the process of refrigerant replacement. Secondly, the factors with a degree of centrality exceeding 4 include consumers’ willingness to adopt eco-friendly refrigerants (S₆), the market demand for eco-friendly refrigerants (S₂), government policies supporting the promotion of eco-friendly refrigerants (S₈), the market price of eco-friendly refrigerants (S₁), the total life-cycle cost of equipment utilizing eco-friendly refrigerants (S₁₄), and the retrofitting cost for substituting existing equipment with eco-friendly refrigerants (S₁₂). These six factors are identified as having a substantial influence on the transition to alternative refrigerants. Finally, the centrality scores for key components in the optimization of environmentally friendly refrigerant equipment (S₁₀), social environmental awareness (S₅), stability in the supply chain of environmentally friendly refrigerants (S₁₁), consumer health awareness (S₄), and eco-friendly refrigerant system performance (COP) (S₁₃) range between 3 and 4. In contrast, factors with lower centrality scores include climate conditions (S₃) and greenhouse gas emissions resulting from the use of environmentally friendly refrigerant equipment (S₉).

According to the degree of influence among the factors related to refrigerant substitution, there are seven causal factors and seven resultant factors. Eco-friendly refrigerant system performance (COP) (S₁₃) exerts the greatest influence on other factors. The level of market demand for refrigerants (S₂) is most susceptible to the influence of other factors. It can be observed from Table 4 and Figure 3 that the system performance evaluation of equipment utilizing environmentally friendly refrigerants (COP) (S₁₃) exerts the most significant influence on other factors and demonstrates strong restraining and driving capabilities. Specific climatic conditions (S₃), greenhouse gas emissions from equipment utilizing environmentally friendly refrigerants (S₉), the market price of environmentally friendly refrigerants (S₁), awareness of consumer health (S₄), implementation of the government’s supportive policies for environmental protection refrigerants (S₈), and full life-cycle cost analysis of equipment using environmentally friendly refrigerants (S₁₄) all have values greater than zero. These factors also play a foundational role in promoting the process of refrigerant replacement and exert a certain degree of influence on its advancement. The factors most susceptible to external influences include the level of market demand for refrigerants (S₂), awareness of social environmental protection (S₅), the stability of the supply chain for environmentally friendly refrigerants (S₁₁), and consumers’ willingness to adopt environmentally friendly refrigerants (S₆). Technological innovation in refrigerant substitution (S₇), optimization of key equipment components utilizing environmentally friendly refrigerants (S₁₀), and the cost of equipment transformation for environmental protection and refrigerant replacement (S₁₂) are all less than 0, indicating that these factors are primarily shaped by the impact of the other seven factors.

6.2. Analysis Utilizing the ISM Method

First, based on the data in the comprehensive impact matrix T, the average value of all elements in T is calculated to determine the threshold. According to Formula (4), the comprehensive impact matrix T is added to the identity matrix I to obtain the overall impact matrix B (

Table A4. Overall impact matrix B for factors influencing refrigerant alternatives selection.

	S1	S2	S3	S4	S5	S6	S7	S8	S9	S10	S11	S12	S13	S14
S1	1.139	0.268	0.000	0.147	0.100	0.232	0.283	0.198	0.000	0.185	0.244	0.256	0.057	0.241
S2	0.175	1.133	0.000	0.112	0.075	0.144	0.213	0.166	0.000	0.129	0.184	0.187	0.041	0.173
S3	0.123	0.169	1.000	0.147	0.159	0.186	0.174	0.130	0.000	0.171	0.113	0.171	0.124	0.140
S4	0.174	0.235	0.000	1.088	0.177	0.212	0.208	0.173	0.000	0.124	0.160	0.161	0.071	0.138
S5	0.118	0.185	0.000	0.134	1.059	0.182	0.174	0.134	0.000	0.103	0.117	0.130	0.034	0.103
S6	0.155	0.238	0.000	0.128	0.110	1.112	0.214	0.171	0.000	0.119	0.164	0.173	0.059	0.127
S7	0.189	0.276	0.000	0.122	0.121	0.195	1.181	0.202	0.000	0.214	0.252	0.273	0.087	0.216
S8	0.184	0.289	0.000	0.179	0.174	0.248	0.267	1.141	0.000	0.185	0.229	0.252	0.071	0.188
S9	0.105	0.164	0.000	0.143	0.138	0.163	0.161	0.143	1.000	0.080	0.104	0.115	0.030	0.080
S10	0.159	0.184	0.000	0.102	0.080	0.148	0.216	0.141	0.000	1.096	0.146	0.209	0.130	0.187
S11	0.13	0.122	0.000	0.079	0.060	0.106	0.165	0.110	0.000	0.092	1.089	0.170	0.051	0.141
S12	0.156	0.215	0.000	0.122	0.111	0.193	0.220	0.181	0.000	0.138	0.205	1.140	0.068	0.198
S13	0.26	0.309	0.000	0.252	0.191	0.283	0.330	0.267	0.000	0.243	0.263	0.267	1.070	0.231
S14	0.249	0.269	0.000	0.162	0.125	0.250	0.296	0.224	0.000	0.178	0.227	0.281	0.087	1.152

). When the overall impact matrix B satisfies Formula (5), the reachable matrix H is obtained. Using the average value of all elements in the comprehensive impact matrix T, the threshold is calculated, and through calculation, the empirical threshold is set to 0.14. Substituting this threshold into Formula (5), the reachable matrix H is derived. By using the criterion C (Si) = R (Si) to partition the system hierarchy of influencing factors, the ISM structural model for refrigerant replacement influencing factors is constructed and analyzed.

First, based on the data in the comprehensive influence matrix T, calculate the average value of all elements in the comprehensive influence matrix T to obtain the threshold. According to Formula (4), add the comprehensive influence matrix T to the identity matrix I to establish the overall influence matrix B (Table A4). By calculation, an empirical threshold of 0.14 is set, and the threshold is substituted into Formula (5). When the overall influence matrix B satisfies Formula (5), the reachable matrix H (

Table A5. Reachability matrix H for influencing factors of refrigerant substitution.

Factor	S1	S2	S3	S4	S5	S6	S7	S8	S9	S10	S11	S12	S13	S14
S1	1	1	0	1	0	1	1	1	0	1	1	1	0	1
S2	1	1	0	0	0	1	1	1	0	0	1	1	0	1
S3	0	1	1	1	1	1	1	0	0	1	0	1	0	0
S4	1	1	0	1	1	1	1	1	0	0	1	1	0	0
S5	0	1	0	0	1	1	1	0	0	0	0	0	0	0
S6	1	1	0	0	0	1	1	1	0	0	1	1	0	0
S7	1	1	0	0	0	1	1	1	0	1	1	1	0	1
S8	1	1	0	1	1	1	1	1	0	1	1	1	0	1
S9	0	1	0	1	0	1	1	1	1	0	0	0	0	0
S10	1	1	0	0	0	1	1	0	0	1	1	1	0	1
S11	0	0	0	0	0	0	1	0	0	0	1	1	0	0
S12	1	1	0	0	0	1	1	1	0	0	1	1	0	1
S13	1	1	0	1	1	1	1	1	0	1	1	1	1	1
S14	1	1	0	1	0	1	1	1	0	1	1	1	0	1

) is obtained. Using C (Si) = R (Si) as the criterion, the hierarchical structure of the influencing factors system is divided, and the ISM structural model of the influencing factors for refrigerant substitution is derived and analyzed.

The first level is determined by dividing the hierarchy according to the criterion C (Si) = R (Si). Once the first level is established, the rows and columns corresponding to the factors of the first level are excluded. Subsequently, the second level is divided following the same principle of C (Si) = R (Si), and this process continues iteratively. Ultimately, the set of influencing factors derived from the reachable matrix is obtained, as presented in

Table 5. Sets of influencing factors in the reachable matrix.

Si	Reachable Set R (Si)	Antecedent Set A (Sj)	Intersection C (Si)
S1	1,2,4,6,7,8,10,11,12,14	1,2,4,6,7,8,10,12,13,14	1,2,4,6,7,8,10,12,14
S2	1,2,6,7,8,11,12,14	1,2,3,4,5,6,7,8,9,10,12,13,14	1,2,6,7,8,12,14
S3	2,3,4,5,6,7,10,12	3	3
S4	1,2,4,5,6,7,8,11,12	1,3,4,8,9,13,14	8,1,4
S5	2,5,6,7	3,4,5,8,13	5
S6	1,2,6,7,8,11,12	1,2,3,4,5,6,7,8,9,10,12,13,14	1,2,6,7,8,12
S7	1,2,6,7,8,10,11,12,14	1,2,3,4,5,6,7,8,9,10,11,12,13,14	1,2,6,7,8,10,11,12,14
S8	1,2,4,5,6,7,8,10,11,12,14	1,2,4,6,7,8,9,12,13,14	1,2,4,6,7,8,12,14
S9	2,4,6,7,8,9	9	9
S10	1,2,6,7,10,11,12,14	1,3,7,8,10,13,14	1,10,14,7
S11	7,11,12	1,2,4,6,7,8,10,11,12,13,14	11,12,7
S12	1,2,6,7,8,11,12,14	1,2,3,4,6,7,8,10,11,12,13,14	1,2,6,7,8,11,12,14
S13	1,2,4,5,6,7,8,10,11,12,13,14	13	13
S14	1,2,4,6,7,8,10,11,12,14	1,2,7,8,10,12,13,14	1,2,7,8,10,12,14

.

According to the results presented in Table 5, the ISM structure model of influencing factors for refrigerant substitution was constructed, as illustrated in Figure 4. In Figure 4, the arrows depict the influence relationships between factors. An arrow pointing from one node to another signifies that the factor represented by the starting node directly influences or impacts the factor represented by the ending node. The factors in the diagram are divided into three levels: Level L1 represents direct factors, which are the most apparent and form the outermost layer; Levels L2 and L3 are indirect factors that influence other factors indirectly through multiple pathways; Levels L4 and L5 are deep-rooted factors, serving as the fundamental underlying causes.

It is evident from Figure 4 that the ISM structural model of refrigerant substitution factors comprises six levels. Specifically, the surface direct influencing factors correspond to the first and second levels, the middle indirect influencing factors are located at the third and fourth levels, and the deep root influencing factors are situated at the fifth and sixth levels.

Analysis of direct influencing factors on the surface: Technological innovation in refrigerant substitution (S₇) in the first layer, the stability of the supply chain for environmentally friendly refrigerants (S₁₁), the cost of equipment transformation for environmental protection and refrigerant replacement (S₁₂), the market price of environmentally friendly refrigerants (S₁) in the second layer, the level of market demand for refrigerants (S₂), and consumers’ willingness to adopt environmentally friendly refrigerants (S₆) are all direct factors influencing refrigerant substitution. Among these factors, technological innovation related to refrigerant replacement, the stability of the supply chain for environmentally friendly refrigerants, and the cost of equipment transformation for adopting environmentally friendly refrigerants serve as key facilitators of refrigerant substitution. Meanwhile, the market price and demand level of environmentally friendly refrigerants constitute the core competitiveness in this transition. Additionally, consumer willingness to adopt environmentally friendly refrigerants acts as the primary driving force. These six factors collectively play a direct and critical role in shaping the development of refrigerant substitution.

Analysis of intermediate indirect influencing factors: Awareness of social environmental protection (S₅) in the third layer, optimization of key equipment components utilizing environmentally friendly refrigerants (S₁₀), awareness of consumer health in the fourth layer (S₄), implementation of the government’s supportive policies for environmental protection refrigerants (S₈), and full life-cycle cost analysis of equipment using environmentally friendly refrigerants (S₁₄) serve as intermediate indirect influencing factors. Among these factors, the optimization of key equipment components using environmentally friendly refrigerants provides a critical technical foundation for the development of refrigerants. Meanwhile, heightened social environmental awareness and increasing consumer health consciousness ensure the sustained momentum and appeal of refrigerant replacement initiatives. Additionally, supportive government policies promoting environmentally friendly refrigerants offer an effective institutional framework.

Analysis of influencing factors of deep-rooted causes: In the fifth layer, specific climatic conditions (S₃), greenhouse gas emissions from equipment utilizing environmentally friendly refrigerants (S₉), and eco-friendly refrigerant system performance (COP) (S₁₃) further reinforce these factors. Among them, climate conditions and greenhouse gas emissions constitute the fundamental driving forces behind refrigerant replacement. Meanwhile, the superior performance and low life-cycle cost of environmentally friendly refrigerant systems provide a solid foundation for promoting refrigerant replacement.

6.3. Analysis Utilizing the MICMAC Method

Based on the reachability matrix H, the driving forces and dependencies of each influencing factor were systematically calculated. The results were derived in accordance with Formulas (8) and (9). Subsequently, each factor was categorized into four quadrants—namely, autonomous factors, dependent factors, associated factors, and independent factors—represented on a two-dimensional coordinate axis, as illustrated in Figure 5. Figure 5 classifies influencing factors according to two dimensions: “Driving force Q” and “Degree of dependence Y”. The horizontal axis (Degree of dependence Y) indicates the extent to which a factor relies on other factors; a higher value signifies greater dependence. The vertical axis (Driving force Q) denotes the influence of a factor on the overall system; a higher value implies a stronger impact of that factor on other factors. In addition, the upper left quadrant (Dependent factors) is defined by high dependency (larger Y) and low driving power (smaller Q), whereas the upper right quadrant (Related factors) is marked by high dependency (larger Y) and high driving power (larger Q). The lower left quadrant (Autonomy factors) is defined by low dependence (smaller Y) and low driving force (smaller Q), whereas the lower right quadrant (Independent factors) is marked by low dependence (smaller Y) and high driving force (larger Q). Specifically, the autonomous factors exhibit both weak driving forces and weak dependence. The dependent factors demonstrate strong dependence but a weak driving force. The associated factors show both strong driving force and strong dependence. Lastly, the independent factors exhibit a strong driving force but weak dependence [56].

Autonomous factor sets with lower driving force and less dependence, such as awareness of consumer health (S₄), greenhouse gas emissions from equipment utilizing environmentally friendly refrigerants (S₉), and optimization of key equipment components utilizing environmentally friendly refrigerants (S₁₀), exhibit limited influence from other factors. These factors are relatively independent and easier to control. Positioned in the middle layer of the ISM model, they serve as critical connectors bridging the preceding and succeeding layers.

The dependent factor set characterized by low driving force and high dependence comprises the level of market demand for refrigerants (S₂), the stability of the supply chain for environmentally friendly refrigerants (S₁₁), and the cost of equipment transformation for environmental protection and refrigerant replacement (S₁₂). These factors are highly susceptible to influence from other factors. With the exception of the level of market demand for refrigerants (S₂), all these factors are positioned at the upper levels of the ISM model, indicating that their resolution typically relies on the development and improvement of other interrelated factors.

The independent factor set characterized by high driving force and low dependence includes the market price of environmentally friendly refrigerants (S₁), eco-friendly refrigerant system performance (COP) (S₁₃), and full life-cycle cost analysis of equipment using environmentally friendly refrigerants (S₁₄). Apart from the market price of environmentally friendly refrigerants (S₁), the remaining two factors are positioned in the middle and lower levels of the ISM model. Although these factors have minimal influence on others, they are significantly impacted by other factors. This highlights the necessity of prioritizing the system performance of environmentally friendly refrigerants and the overall life-cycle cost of environmentally friendly refrigerant equipment during the refrigerant replacement process.

At the same time, the research results show that the non-influencing factors belong to the correlation factor set, indicating that the influencing factors selected in the study are relatively stable. In addition, consumers’ willingness to adopt environmentally friendly refrigerants (S₆) is between the autonomous factor and the dependent factor, and its dependence and driving force are low; technological innovation in refrigerant substitution (S₇) is between the dependent factor and the associated factor, and its driving force is low but its dependence is high, and it is more susceptible to other factors. Implementing the government’s supportive policies for environmental protection refrigerants (S₈) for the implementation of environmentally friendly refrigerants is between the autonomous factor and the independent factor, and its dependence is low but the driving force is slightly stronger, which can affect the upper factor to a certain extent.

7. Discussion

This paper employs the DEMATEL-ISM-MICMAC method to investigate the interrelationships and driving mechanisms among the influencing factors of refrigerant substitution, and further proposes targeted countermeasures and recommendations. It systematically facilitates the refrigerant substitution process, offering a scientific foundation for enterprises to select environmentally friendly refrigerants while promoting sustainable social development. Specifically, this paper employs the DEMATEL method to analyze the relationships among factors within the complex system of refrigerant substitution and ascertain the significance of each factor. The results indicate that three factors—government support policies for promoting environmentally friendly refrigerants (S₈), the market price of environmentally friendly refrigerants (S₁), and the total life-cycle cost of environmentally friendly refrigerant equipment (S₁₄)—exhibit relatively high centrality and serve as causal factors. These factors play a crucial role in the influencing factors of refrigerant substitution. Moreover, technological innovation related to refrigerant substitution (S₇) demonstrates the highest centrality and is identified as the most critical core influencing factor. Additionally, the environmental refrigerant system performance (COP) (S₁₃) exhibits the highest degree of causality, suggesting that this factor has the most significant influence on other factors. In contrast, the demand level of the environmental refrigerant market (S₂) demonstrates the lowest degree of causality, rendering it highly sensitive and susceptible to influence from other factors. This aspect should also be carefully considered by enterprises. The ISM multi-layer hierarchical structure model categorizes the influencing factors into five distinct levels, providing a clear and effective representation of the intrinsic relationships among refrigerant substitution factors. Additionally, it employs the MICMAC driver-dependence matrix to elucidate the critical roles of these influencing factors within the system. Technological innovations related to refrigerant substitution (S₇), the stability of the eco-friendly refrigerant supply chain (S₁₁), retrofitting costs for equipment transitioning to eco-friendly refrigerants (S₁₂), demand levels in the eco-friendly refrigerant market (S₂), and consumers’ willingness to adopt eco-friendly refrigerants (S₆) constitute surface-level direct influencing factors that impact refrigerant substitution. While these factors exhibit relatively low driving force and high dependency, they can directly accelerate the advancement of refrigerant substitution. Social environmental awareness (S₅) and the optimization of key components in equipment utilizing eco-friendly refrigerants (S₁₀) serve as mid-level indirect influencing factors, playing a crucial role in facilitating the transition to alternative refrigerants. Consumer health awareness (S₄), government support policies for promoting environmentally friendly refrigerants (S₈), and the full life-cycle cost of environmentally friendly refrigerant equipment (S₁₄) are fundamental driving factors with profound and significant influence. These factors not only provide strong leadership but also establish a robust foundation for the development of the industry. Moreover, climate conditions (S₃), greenhouse gas emissions from environmentally friendly refrigerant equipment (S₉), and the performance of environmentally friendly refrigerant systems (COP) (S₁₃) are positioned at the lowest level of the SIM structural model. These elements represent critical root causes that must not be overlooked in the context of refrigerant replacement. Based on the aforementioned research conclusions, it is recommended that government support policies be reinforced, environmental policies related to refrigerants be formulated and implemented more effectively, financial subsidies and tax incentives be provided to enterprises, and encouragement be given to adopt environmentally friendly refrigerants. Additionally, technological innovation should be promoted through increased investment in the research and development of environmentally friendly refrigerant technologies, as well as supporting breakthroughs in key core technologies and equipment innovation. At the same time, optimize the cost structure of environmentally friendly refrigerants, reduce their market prices, lower costs through scaled production, control equipment retrofitting expenses, and encourage equipment manufacturers to optimize the design of environmentally friendly refrigerant equipment to decrease the overall life-cycle costs. Establish a stable and efficient supply chain for environmentally friendly refrigerants to ensure the timely supply of raw materials and products, prevent supply chain disruptions, foster collaboration among upstream and downstream enterprises, and strengthen supply chain resilience. Focus on enhancing market demand and consumer awareness by strengthening the promotion of environmental protection and health consciousness, increasing consumers’ recognition of and willingness to use eco-friendly refrigerants, cultivating market demand, and fostering a positive cycle.

The existing literature that analyzes the influencing factors often emphasizes either a single dimension or partial correlations. This study integrates the DEMATEL, ISM, and MICMAC methods to systematically explore the hierarchical relationships and driving mechanisms among these factors. A comparative analysis of the findings from this study and the existing literature indicates that the DEMATEL method was employed to systematically identify key factors within the complex refrigerant substitution system. Particular attention was given to the pivotal roles of technological innovation related to refrigerant substitution (S₇), government support policies encouraging environmentally friendly refrigerants (S₈), the market price of environmentally friendly refrigerants (S₁), and the life-cycle cost of environmentally friendly refrigerant equipment (S₁₄). These findings are consistent with numerous existing studies that emphasize technological innovation as a critical driver for refrigerant substitution [29]. It is evident that technological innovation not only propels the research and development of alternative refrigerants but also facilitates their application and dissemination. Moreover, the significance of government support policies reinforces the crucial role of policy incentives in promoting environmentally friendly refrigerants, as emphasized by relevant studies [5,7,16]. The dual focus on market price and life-cycle cost not only complements existing research on the impact of economic factors but also provides a more comprehensive perspective. Many studies have identified economic feasibility as a critical constraint in promoting alternative refrigerants, making this dual emphasis particularly relevant [24]. This study revealed that the performance of environmentally friendly refrigerant systems, as measured by the coefficient of performance (COP) (S₁₃), exerts the most significant influence on other factors. This suggests that enhancing system performance can substantially promote improvements in other related elements. This conclusion is consistent with prior studies, which emphasize that superior performance serves as the primary motivator for users to adopt new refrigerants [9,12]. Furthermore, it is intrinsically linked to the dual objectives of improving energy efficiency and mitigating environmental impact. This paper utilizes the ISM multi-layer hierarchical structure and the MICMAC matrix to systematically classify influencing factors, explicitly distinguishing deep-rooted fundamental factors, mid-level indirect factors, and surface-level direct factors. By addressing the limitations of existing studies that focus narrowly on single dimensions, it offers a more comprehensive and actionable hierarchical governance framework. Furthermore, the analysis of driving power and dependence enhances the guidance for policy formulation and corporate strategy development, thereby aligning with the integrated systemic perspectives advocated for in some of the literature. The research methods and conclusions of this paper reinforce and extend the existing literature’s understanding of the factors influencing refrigerant substitution in several key dimensions. Specifically, it presents a more innovative and systematic theoretical framework that highlights the pivotal role of technological innovation, the catalytic impact of policies, and a hierarchical classification of factors affecting refrigerant substitution. Additionally, it underscores the significance of performance evaluation and socio-environmental considerations, offering actionable guidance and valuable references for future policy formulation and corporate strategic planning. Existing studies on refrigerant substitution tend to focus on a single factor, such as cost or environmental impact, while paying insufficient attention to the complex interdependencies among factors. To address this limitation, we innovatively applied the DEMATEL method to systematically identify and analyze the intricate relationships among 14 key factors in refrigerant substitution, thereby offering a comprehensive solution approach. Based on the DEMATEL results, we constructed an ISM structural model to visually illustrate the hierarchical relationships and internal connections among the factors, thereby facilitating the formulation of more effective alternative strategies. Moreover, through MICMAC analysis, the driving forces and dependencies of each influencing factor were systematically identified, further enhancing the understanding of the underlying mechanism of refrigerant substitution.

The replacement of refrigerants constitutes a multifaceted challenge influenced by various factors across different levels, necessitating collaborative efforts from governments, enterprises, and consumers alike. From a national perspective, by focusing on core driving factors—technological innovations in refrigerant substitution (S₇), government support policies for environmentally friendly refrigerants (S₈), and the life-cycle cost of environmentally friendly refrigerant equipment (S₁₄)—the state should develop long-term subsidy or tax incentive policies to promote research and development of environmentally friendly refrigerants (e.g., low-GWP refrigerants) and domestic production of key equipment, thereby reducing the application costs of such technologies. Additionally, a dedicated fund should be established to support improvements in the coefficient of performance (COP) of environmentally friendly refrigerant systems (S₁₃) and optimization of critical components of environmentally friendly refrigerant equipment (S₁₀). This will foster collaboration among industry, academia, and research institutions to overcome technical challenges. Furthermore, it is essential to expedite the establishment of mandatory regulations for phasing out high-GWP refrigerants and market access standards for environmentally friendly refrigerants, ensuring the stability and reliability of the environmentally friendly refrigerant supply chain (S₁₁). Incorporate refrigerant substitution into “low-carbon” assessments and incentivize enterprises to undergo transformation via the carbon trading mechanism. Strengthen public environmental awareness (S₅) and consumer health consciousness (S₄) through targeted public campaigns, thereby guiding consumers to prioritize environmentally friendly products. Prioritize the procurement of eco-friendly refrigeration equipment in government purchasing and public infrastructure development to both exemplify commitment and stimulate demand within the eco-friendly refrigerant market (S₂). Simultaneously, the government should proactively engage in the establishment of global refrigerant standards, facilitate the introduction of advanced technologies, and address cross-border challenges such as varying climate conditions (S₃). From an enterprise perspective, it is crucial to optimize the supply chain and full life-cycle costs, collaborate with universities or research institutions to address the retrofitting costs associated with replacing refrigerants (S₁₂), and enhance the performance of eco-friendly refrigerant systems (S₁₃). High-efficiency, low-cost replacement solutions should also be developed. Large-scale production of environmentally friendly refrigerants can reduce their prices (S₁), thereby lowering the initial investment burden for users. Furthermore, diversified supply channels for eco-friendly refrigerants should be established to mitigate supply chain risks. Additionally, refrigerant solutions with enhanced adaptability should be developed to accommodate the climate characteristics of various regions, such as high-temperature or high-humidity environments. From the consumer perspective, consumers are encouraged to proactively consider the environmental attributes of refrigerants (e.g., GWP values) (S₅) and their potential health impacts (S₄). By prioritizing low-carbon products, they can incentivize enterprises to accelerate transformation toward sustainability. Furthermore, sharing personal usage experiences via social media platforms can amplify market demand for environmentally friendly refrigerants (S₂). Consumers should rationally evaluate the trade-offs between costs and long-term benefits, refraining from overemphasizing the initial price of refrigerants (S₁) and instead conducting a comprehensive assessment of the full life-cycle costs of equipment, including energy consumption and maintenance expenses (S₁₄). Participation in trade-in programs or government subsidy schemes can help reduce the financial barriers associated with adopting alternative refrigerants. Additionally, providing constructive feedback to companies and regulatory bodies regarding performance limitations (e.g., insufficient COP) can stimulate technological advancements in refrigerant substitution (S₇). Such actions will not only facilitate refrigerant replacement but also promote technological innovation, economic growth in relevant industries, and ecological protection.

8. Conclusions

Studying the factors influencing refrigerant substitution is essential to addressing global environmental challenges, facilitating energy structure optimization, and fostering technological innovation. The primary contribution of this paper is the comprehensive consideration of various factors affecting refrigerant replacement, coupled with the innovative application of the DEMATEL method to systematically analyze the interrelationships among these factors. This study identifies four critical factors—technological innovation (S₇), government supportive policies (S₈), market price (S₁), and the full life-cycle cost of equipment (S₁₄)—that play a pivotal role in the refrigerant replacement process. Among these factors, the environmentally friendly refrigerant system performance (COP) (S₁₃) exhibits the highest degree of causality, highlighting its substantial influence on other variables. In contrast, the market demand level (S₂) demonstrates the lowest degree of causality, rendering it highly sensitive and susceptible to influence from other factors, thereby necessitating particular attention from enterprises.

This paper presents, for the first time, a multi-layer hierarchical model based on ISM, which effectively categorizes the influencing factors of refrigerant substitution into direct factors, indirect factors, and fundamental factors. Specifically, technological innovation (S₇), supply chain stability (S₁₁), equipment renovation costs (S₁₂), market demand level (S₂), and consumer willingness to use (S₆) serve as surface-level direct influencing factors. Social environmental awareness (S₅) and key component optimization (S₁₀) function as mid-level indirect influencing factors. Meanwhile, consumer health awareness (S₄), government support policies (S₈), and the total life-cycle cost of equipment (S₁₄) constitute deep-rooted fundamental influencing factors. At the same time, climate conditions (S₃), greenhouse gas emissions (S₉), and system performance (S₁₃) constitute the foundational layer of the model and represent critical factors that must be considered in refrigerant substitution. The ISM model successfully elucidates the intrinsic relationships and hierarchical structure among these influencing factors, facilitating a comprehensive analysis of the complex refrigerant substitution system and offering structured guidance for decision-making in policy formulation, technological advancement, and industrial transition.

Based on the MICMAC analysis, market price (S₁), system performance (S₁₃), and total equipment life-cycle cost (S₁₄) are identified as independent key factors with strong driving power but low dependence. This finding is highly consistent with the critical influencing factors determined by the DEMATEL method, thereby reinforcing the importance for enterprises to prioritize these three core factors when promoting refrigerant replacement.

Research has demonstrated that government support policies and life-cycle costs serve as critical “driving forces”. Future refrigerant selection should prioritize enhancing relevant policies and subsidy incentives, decreasing the life-cycle costs of equipment utilizing environmentally friendly refrigerants, and invigorating both upstream and downstream markets alongside refrigeration technology research and development, ultimately fostering a virtuous cycle. The three “active factors”—climate conditions, greenhouse gas emissions, and system performance—collectively drive the operation of the entire system. When developing refrigeration equipment with environmentally friendly refrigerants, it is essential to evaluate the performance of these refrigerants (e.g., energy efficiency as measured by COP), their practical environmental adaptability (e.g., suitability for varying climates), and their associated greenhouse gas emission levels. This ensures that improvements in overall system performance are achieved concurrently with advancements toward environmental objectives.

Existing research on refrigerant alternatives predominantly emphasizes individual factors, such as cost or environmental impact, often neglecting the intricate interdependencies among them. Consequently, this gap complicates the development of effective strategies for expediting the transition to environmentally friendly refrigerants. In response, our study innovatively employs the DEMATEL method to systematically identify and analyze the complex interrelationships among 14 critical factors influencing refrigerant replacement, thereby providing a comprehensive solution to this challenge. Our research demonstrates that technological innovation (S₇) serves as a pivotal factor and exerts the most significant influence on other factors within the system. This underscores the necessity of prioritizing investment in the development of new technologies while fostering an enabling environment for their adoption. Moreover, our findings reveal that the level of market demand (S₂) is the most responsive to changes in other factors, thereby emphasizing the critical role of implementing strategies to stimulate demand for environmentally friendly refrigerants. Finally, we leveraged the DEMATEL results to construct an ISM structural model that not only visualizes the hierarchical relationships and interconnections among the factors but also aids in the formulation of effective strategies for refrigerant replacement. Although the DEMATEL method is highly authoritative, its results may still contain a degree of subjectivity. Moreover, the selection of influencing factors relies predominantly on existing research, which limits the comprehensiveness of the analysis and may result in discrepancies when compared to real-world conditions. In future studies, refining expert review criteria and increasing the size of the expert sample could further enhance the representativeness and scientific rigor of the findings.