Design for Recycling: A Systematic Review of Approaches for Enhancing Product Recyclability

Abstract

The growing accumulation of solid waste from consumerism and the traditional linear economy poses significant risks to environmental sustainability and human health, increasing pressure on manufacturers to design more recyclable products. Design for Recycling (DfR), a key subset of Design for X (DfX), offers a promising strategy to enhance product recyclability. However, its implementation is hindered by ambiguous definitions, insufficient data, and a lack of robust methodologies and tools. This review clarifies the concept of DfR in engineering, identifies three critical research gaps, and proposes future research directions based on 89 papers: (1) the development of a comprehensive DfR knowledge base, (2) the creation of standardized recyclability evaluation indicators, and (3) the establishment of an automated, data-driven design generation system. These advancements aim to support the automated creation of highly recyclable products by leveraging accurate data, precise evaluations and minimized human bias. This review not only highlights existing research gaps but also provides valuable insights to guide future DfR studies and the development of more effective tools and methods. Furthermore, it emphasizes that the successful implementation of DfR requires active participation and commitment from the entire industrial chain and society.

1. Introduction

The increasing generation of solid waste and the looming threat of resource depletion, driven by ubiquitous consumerism, have raised the demand for more circular product designs. According to the United Nations [1], in 2019, only 17.4% of used electronics were formally collected and recycled, leaving 82.6% inadequately managed. This under-recycling poses significant risks to human health, environmental integrity, and resource efficiency. Although various measures have been implemented globally, most focus on standardizing the end-of-life (EOL) processes while overlooking upstream solutions, which leads to less desired outcomes [2,3,4].

Product design, however, offers a more effective solution at the source, as it is widely believed that 80–90% of recycling costs and revenues are determined during the design stage [5]. Moreover, recycling is considered a key characteristic of circular product design [6], and Design for Recycling (DfR) is an effective technique to enable this circularity by creating products that are easier to disassemble and recycle at the EOL through the incorporation of recycling considerations directly into the design process.

DfR, as a type of Design for X (DfX) methodology, is not a new concept, and it has consistently attracted attention both in industry and academia, ranging from plastic packaging to fashion. In this review, we mainly focus on the engineering design area, particularly mechanical and electrical products such as motors and refrigerators.

A simple search on Google Scholar using the keyword “design for recycling” shows that recent research in mechanical and electrical engineering is limited, with most papers focusing on architecture. While the importance of recycling, the development gap is to be expected. One obvious reason for this is the vague definition of DfR due to the overlaps among many DfX concepts, such as Design for Environment (DfE), Design for End-of-Life (DfEOL), and Design for Sustainability (DfS) [7]. Additionally, other factors contributing to this gap will be explored in this review. Furthermore, with the advancement of artificial intelligence (AI) and other emerging technologies, new opportunities are arising to address these challenges and develop innovative methods and tools within the DfR field. We also hope to explore the potential to develop DfR approaches in the context of these new technologies.

Overall, this review aims to clarify the definition of Design for Recycling in engineering design and propose promising future research directions by identifying current research gaps. The structure of this paper is as follows: Section 2 briefly introduces the research methods; Section 3 presents the review results (Section 3.1 defines DfR through step-by-step analysis; Section 3.2 reviews the literature on optimal EOL options; Section 3.3 examines recyclability evaluation studies; and Section 3.4 discusses improvements in product recyclability). Section 4 proposes future research directions (in Section 4.1) through the identification of current limitations (in Section 4.2). Section 5 concludes the review.

2. Methods

To minimize bias in the selection and inclusion of literature and enhance the quality and objectivity of the review, we followed the PRISMA (The Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [8].

To establish clear inclusion criteria and ensure that reliable literature was not omitted, we used the term “design for recycling” and its synonyms, such as “design for recyclability”, “recyclable design”, along with keywords like “methods”, “tools” to search in Scopus, Web of Science and Engineering Village.

Table 1. Search by keywords and document selection.

Search by Keywords	Scopus	Web of Science	Engineering Village
(‘design for recycling OR ‘design for recyclability’ OR ‘recyclable design’ OR ‘recycling-oriented product design’) AND (‘product design’ OR ‘product development’)	443	50	72
(‘design for recycling’ OR ‘design for recyclability’ OR ‘recyclable design’ OR ‘recycling-oriented product design’) AND (‘circular economy’ OR ‘end-of-life’)	478	84	60
(‘design for recycling’ OR ‘design for recyclability’ OR ‘recyclable design’ OR ‘recycling-oriented product design’) AND (‘eco-design’ OR ‘sustainable design’ OR ‘design for environment’ OR ‘green design’ OR ‘environmentally conscious design’ OR ‘life cycle design’)	343	52	20
(‘design for recycling’ OR ‘design for recyclability’ OR ‘recyclable design’ OR ‘recycling-oriented product design’) AND (‘guideline’ OR ‘method’ OR ‘tool’ OR ‘application’ OR ‘evaluation’)	710	75	47
Total	1974	261	199

outlines the detailed keywords, search process, and results.

The review focused on peer-reviewed English-language articles published up until the end of 2023, excluding grey literature but without restricting the publication period. After eliminating irrelevant fields such as chemistry and physics, 1974 articles were identified from Scopus, 261 from Web of Science, and 199 from Engineering Village. After removing duplicates, 1274 documents underwent title and abstract screening, with 125 articles selected for full-text review. Finally, 89 articles were included in the analysis, excluding those focused solely on end-of-pipe management. Figure 1 illustrates the research strategy employed in this literature review, following the PRISMA 2020 flow diagram [8].

The inclusion criteria were determined by three authors, with two reviewers independently assessing each record. Discrepancies were resolved through discussion, and no automation tools were used.

3. Results

3.1. Design for Recycling in Engineering

DfR is one of the DfX methods derived from concurrent engineering, first proposed by Sohlenius (1992), and has shown significant potential for enhancing product design across various stages of the product lifecycle [9]. In 1992, Boothroyd and Alting highlighted the importance of recycling used products for resource conservation in their keynote on Design for Disassembly (DfD) [10]. With the rise in sustainable design concepts, such as Life Cycle Assessment (LCA) and DfE, DfR has gained increasing attention.

A cursory literature review reveals that while some research has been conducted on DfR, the vague and inconsistent definition of DfR, along with its overlap and confusion with concepts like DfE and DfS, has hindered its development. For example, in the article by Ferrão and Amaral (2006), recycling in the context of DfR referred solely to the percentage of recycled materials at the EOL [11], while Zussman et al. (1994) defined recycling in DfR as a combination of various after-use processes, such as reusing, using on, and dumping [12]. Additionally, they measured a product’s recyclability by calculating cost–benefits rather than merely considering the proportion or weight of the recycled materials. Although both delimitations of recycling are reasonable, the absence of a standardized definition slows the implementation of DfR. Therefore, it is imperative to establish a unified specification of DfR in engineering design, facilitating its long-term and effective development to generate more recyclable products.

3.1.1. Definition of Recyclability

First, it is important to clarify the connotation of “recyclability”, as the goal of DfR is to improve a product’s recyclability. In this review, we considered recyclability as a broad concept, not limited to material recycling at EOL, but rather as a process that involves multiple stakeholders throughout the entire product lifecycle.

We examined several preferred current interpretations of “recyclability” to derive a clearer understanding of its connotation. Tonnelier et al. (2005) described recyclability as “recovery potential”, defining it as the ability of a part or assembly to meet criteria that enable it to be reused as material, parts, or energy [13]. Xing et al. (2003) argued that recyclability is an inherent product feature determined during the design phase, aimed at making products easier to reuse, remanufacture, or recover [14]. Van den Berg and Bakker (2015) emphasized that recyclability is primarily determined by material choices and the ease with which components can be separated [6].

Each of these perspectives highlights different aspects of recyclability, yet all underscore the impact that design decisions have on EOL behavior. To clarify and make the connotation of “recyclability” more comprehensive, we define it as “an inherent feature of a product, determined primarily by material choices and ease of disassembly, that allows it to be easily reused at the EOL”. Additionally, to avoid any misunderstanding, it should be noted that “reuse” in this context is a broad term encompassing material recycling, part reuse, energy recovery, and other forms of reuse.

3.1.2. Definition of DfR

Based on the proposed definition of “recyclability”, it follows naturally that enhancing a product’s recyclability involves improving its inherent ability to be reused at EOL, potentially by selecting more recyclable materials and designing for easier disassembly. Since DfR aims to improve a product’s recyclability, it also focuses on enhancing the product’s inherent potential to be reused at EOL, primarily through the selection of recyclable materials and facilitating disassembly.

However, for a more precise understanding of DfR, we also reviewed several existing definitions at first. Liu et al. (2002) suggested that DfR entails fully addressing recycling challenges related to products and components, such as recycling probability, value, processes, and technologies, to optimize resource and energy use while minimizing environmental impact [5]. Sakundarini et al. (2013) argued that DfR is a promising approach to prolong material utilization in the early design stages, focusing on harmonizing design with recycling practices to preserve valuable materials, reduce waste at EOL, and limit the use of natural resources [15]. Similarly, Zussman et al. (1994) noted that design for EOL aims to align product design with future recycling processes, echoing Sakundarini’s perspective [12,15]. From these definitions, we concluded that the phrase “a product’s inherent feature that allows it to be reused at the EOL” in our definition of recyclability corresponds to “harmonizing (or aligning) design with recycling practices (or processes)”. Additionally, the aspect we initially overlooked is that the goal of DfR is to prevent environmental degradation.

Based on the analysis above, we define Design for Recycling as “an optimized design approach that aims to maximize compatibility of product design with future recycling practice and minimize resource waste and environmental pollution, potentially through the selection of recyclable materials or design for easier disassembly”.

3.2. Identification of Optimal EOL Options

According to our definition of DfR, its goal is to enhance the alignment of product design with the EOL process. Ishii (1997) also emphasized the critical role of product retirement strategies in determining a product’s recyclability [16]. Therefore, predicting a product’s or its components’ potential EOL options (i.e., how they will be processed at the EOL stage) during the design stage is essential, as designing products based on these predicted optimal EOL options enhances their compatibility with recycling practices.

Additionally, it is important to note that, generally, there are two ways to interpret the term “EOL options”: one refers to EOL options for the entire product, and the other refers to EOL options for individual components. Since, in the context of DfR, disassembly is a prerequisite for recycling, most papers focus on the latter. Thus, unless otherwise stated, we refer to EOL options at the component level in this review. We identified a range of studies on EOL options determination at the design stage for more recyclable product designs and categorized them into three main classifications based on the methods employed.

3.2.1. Recovery AND/OR Graph-Based Methods

Some studies employ “disassembly and recovery AND/OR graphs” to simultaneously determine the optimal disassembly sequence for a product and the corresponding EOL strategy for its components or subcomponents [12,17,18,19,20,21]. These articles emphasized the impact of disassembly on recycling value, aiming to balance the effort required for disassembly with the revenue generated through recycling. However, although disassembly plays a significant role in recycling, and the results derived from detailed disassembly analysis are persuasive for guiding design decisions, these methods typically require extensive product data and detailed EOL information. This makes them less suitable for the early design stage when such data may not yet be available.

3.2.2. MCDM-Based Methods

Given the complexity of factors influencing EOL assessment and selection, some studies adopt Multi Criteria Decision Making (MCDM) techniques to identify optimal EOL options. These include fuzzy logic-based methods [22,23], fuzzy Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) based evaluation methods [24], and Grey Correlation Analysis [25]. Other studies propose simpler methods focusing solely on environmental and economic criteria, using mathematical EOL assessment indicators to evaluate different EOL options [26,27,28,29,30,31].

MCDM-based approaches offer flexibility, as the criteria and indicators can be adjusted according to the specific decision problem and context. However, the use of MCDM requires careful attention to subjectivity in weight assignment, the interactions between criteria, and the selection of appropriate MCDM methods to ensure the validity and reliability of the decision-making process.

3.2.3. Empirical Information-Based Methods

Some approaches propose more straightforward methods that do not require complex calculations. These methods rely on empirical or historical data to determine EOL options, utilizing tools such as case-based reasoning (CBR) models [32], the End-of-Life Design Advisor (ELDA) [33,34], or PESTEL analysis [35]. The main advantage of these approaches is their ability to provide rapid evaluations. However, they are heavily dependent on the availability of experienced data, and since data at the design stage is often lacking, this can make the design process more challenging.

3.2.4. Summary

In summary, identifying optimal EOL options for components is a necessary and fundamental step in DfR research. Designing products with recycling in mind requires prior knowledge of the optimal EOL options for each component. While current research has made valuable contributions, most approaches rely on large datasets, which may not be readily available during the early design phase. Additionally, predicting optimal EOL options remains a challenge for several reasons. Firstly, it depends on the development of suitable evaluation criteria and indicators, but these are difficult to identify as they may vary across companies and products. Secondly, human involvement in the evaluation process can introduce biases that affect the determination of the optimal option. Thirdly, an often-overlooked aspect is the uncertainty associated with EOL evaluation, particularly due to the time gap between the design phase and actual recycling. A critical question is whether a product or component can truly be recycled, as assumed during the design stage. The answer is often negative due to unstable factors such as changes in market demand, policy shifts, and other external variables.

3.3. Evaluation of Product’s Recyclability

Recyclability evaluation is central to DfR, as accurately measuring a product’s recyclability is essential for effective improvement. In this section, we reviewed papers on recyclability evaluation based on their different methods.

3.3.1. Clarification of the Two “Evaluation” Methods

However, before reviewing articles on recyclability evaluation, it is important to first note that the “recyclability evaluation” discussed in this section and the “EOL options evaluation” addressed in the previous section are distinct concepts. Although both are used to compare multiple design alternatives, they represent different research areas and should be treated as separate topics, despite the potential for confusion.

Firstly, their research objectives differ. Recyclability evaluation generally refers to assessing a product’s overall recyclability, whereas EOL options evaluation typically focuses on assessing the various EOL options for a component. Secondly, their research purposes are different. The goal of recyclability evaluation is to select the optimal design alternative for a product, while the purpose of EOL options evaluation is to select the best EOL option for a component, thereby guiding the design process to align the product’s design with EOL practices as much as possible.

Although many studies on EOL options evaluation (as reviewed in Section 3.2) claim that they can be used to select the best design alternative, the reality is that component-level EOL options evaluation is only a step of recyclability evaluation. This is because, on the one hand, a product consists of many components, and on the other, a product’s recyclability cannot be solely measured by the recycling value derived from the optimal EOL options of its components.

3.3.2. Design for Disassembly to Ease Recycling

It is widely acknowledged that disassembly is the first and crucial step in recycling, and the ease of disassembly significantly influences a product’s recyclability. Early research on DfR primarily focused on DfD, aiming to reduce disassembly time and difficulty to facilitate easier recycling [36,37,38,39,40,41,42].

However, disassembly does not equate to recyclability, and ease of disassembly does not necessarily imply ease of recycling. Therefore, using disassembly as a measure of recyclability is not reasonable.

3.3.3. Qualitative Recyclability Evaluation

In addition, qualitative recyclability evaluation methods have been developed for their user-friendliness and simplicity, such as the tree diagram and questionnaire method proposed by Tonnelier et al. (2005) and the artificial neural networks-based method presented by Liu et al. (2002) [5,13]. However, it must be noted that these methods are heavily reliant on human judgment and evaluator expertise.

3.3.4. Quantitative Recyclability Evaluation

In contrast to qualitative methods, quantitative indicators are often considered more reliable due to their mathematical foundations. Based on our definition of DfR and the previous discussion on the relationship between “recyclability evaluation” and “EOL option evaluation”, it follows that, when evaluating a product’s recyclability, we must (1) assess and select the optimal EOL option for each component to ensure alignment with recycling practices based on the current design, and (2) develop a comprehensive evaluation indicator to measure the product’s overall recyclability.

However, most research on recyclability evaluation overlooks the first point, which involves selecting the most appropriate EOL options for components. According to our review, material recovery is the most frequently considered EOL option in current studies [43,44,45,46,47,48,49], as it is both an economically viable aspect of recycling and aligned with most legislative requirements. Nevertheless, with increasing product complexity and the evolving demands of the circular economy, other EOL options are gaining importance. It is essential to incorporate these diverse EOL options, not just material recovery, into the recyclability evaluation process.

Some studies have already recognized this issue and incorporated multiple EOL options into their evaluation frameworks, effectively addressing it. However, they often overlook broader evaluation perspectives, leading to shortcomings in the second point. For instance, Papakostas et al. (2015) evaluated recyclability solely from an economic perspective, neglecting potential environmental benefits [50]. Actually, most research has considered the environmental and economic impacts of recycling and disassembly costs before recycling when evaluating a product’s recyclability. For example, Chen et al. (1994) proposed a cost–benefit analysis model to balance the disassembly efforts with the recycling revenue [51]. Favi et al. (2019) included two modules in their LeanDfD tool, helping designers quickly and reliably assess design alternatives using quantitative disassembly and recyclability criteria [45]. Zhang et al. (2004) assessed a product’s lifecycle environmental impact by considering EOL disposition from both economic and environmental perspectives. However, the social aspect is always neglected [52].

3.3.5. Summary

In summary, quantitative recyclability evaluation is preferable. It is essential to first select the optimal EOL options for each component, and a comprehensive set of evaluation indicators is necessary to assess the overall recyclability of the product. Firstly, as discussed in Section 3.2, evaluating and selecting the optimal EOL options for components at the design stage is a crucial part of the overall recyclability assessment, but further work is needed to ensure its accuracy. Secondly, regarding the construction of evaluation indicators, we believe that, in addition to disassembly aspects, a comprehensive set of indicators should incorporate all three dimensions of sustainability—economic, environmental, and social—aligning with the principles of the Triple Bottom Line (TBL) framework within a systematic circular economy approach. However, quantifying the social impacts in product design remains a challenge [53].

3.4. Improvement of Product’s Recyclability

Improving product recyclability through design is a key focus of DfR research. While several studies have contributed to enhancing recyclability, they mainly offer general sustainability or “green” design methods, with recyclability being only one aspect. Consequently, there is a lack of dedicated methods and tools specifically aimed at improving recyclability. In this section, we review the available research, which, due to its limited scope and somewhat fragmented nature, is categorized according to the stages at which the proposed methods can be applied: the early design stage, the later design stage, and beyond the design stage.

3.4.1. Approaches Applied in the Early Design Stage

(1)

Design Guidelines

Design guidelines are among the most widely used and earliest tools in design, with many studies applying them during the early design stage to improve a product’s recyclability [54,55,56,57]. These guidelines are typically summarized and presented in simple language, making them easy for designers to understand and apply. Their use does not require extensive product design data, which further facilitates their application in the early design stage.

However, despite their ease of use, design guidelines, when presented in verbal form, can sometimes lead to misunderstandings. Additionally, this verbal format may limit their long-term utility and storage.

As a form of design knowledge, design guidelines should be modeled and digitized to enable integration with other design methods and tools to better support the design process. Some research has attempted to implement this idea. For example, Houé and Grabot (2007) put this concept into practice by modeling DfR rules and standards, originally described in natural language, using an ontology-based Object-Role Modeling (ORM) language [58]. The modeled knowledge was then used to verify the consistency of product designs with the intended goals. Furthermore, Jalbout and Keivanpour (2023) developed a Body of Knowledge (BOK) for high-tech products to facilitate more efficient DfD and DfR [59].

(2)

Other Methods

In addition to design guidelines, several methods have been identified for use during the early design stage to improve a product’s green characteristics.

For example, Wu and Ho (2015) and Bereketli and Genevois (2013) proposed green design methods based on Quality Function Deployment (QFD) to enhance products’ eco-friendliness [60,61]. Gong et al. (2019) and Huang et al. (2020) employed MCDM to enhance sustainability by generating design candidates with greater remanufacturing potential and increased material recyclability, respectively [62,63]. Sakundarini et al. (2013) and Wichniarek et al. (2018) proposed computer-aided design (CAD) integrated methods to support the design of products with higher recyclability [15,64]. Additionally, Brissaud and Zwolinski (2004) developed a situation-based methodology that defines the design objectives of new products by integrating EOL strategies and constraints through negotiation with all stakeholders throughout the product’s lifecycle [65].

While these approaches are valuable, there is a lack of dedicated methods for enhancing recyclability. If methods such as QFD-based and MCDM-based approaches were adapted to develop dedicated DfR methods, they could facilitate the generation of more recyclable products.

A dedicated DfR method that can improve a product’s recyclability throughout the entire design process is the Integrated Recyclability and End-of-Life Design Algorithm (IREDA), introduced by Xing et al. (2003) [14]. IREDA encompasses a four-stage modeling framework: product EOL strategy prediction, modular structure formation, material and fastener selection, and recyclability assessment of design alternatives. This method holds significant potential for further improvement in each stage, including more accurate recyclability assessments and EOL strategy predictions. Since IREDA can be applied across the entire design process and has already been discussed here, it will not be revisited in the following sections.

3.4.2. Approaches Applied in the Later Design Stage

(1)

Modular Design

Modular design is a widely adopted approach for creating product architectures consisting of physically detachable units, which facilitate rapid product development, ease of disassembly, servicing, reuse, recycling, and other product lifecycle objectives [66].

To enhance product sustainability, many studies have incorporated “sustainable drivers” into traditional modular design, often employing specific clustering algorithms to create more environmentally friendly products. We refer to these modular design methods as “general sustainable modular design methods”, which are reviewed in detail in Table 2.

It can be observed that most studies are not specifically tailored to the design stage. Instead, they typically analyze existing products and generate redesign suggestions, with most case studies focusing on mature market products, such as coffee makers and refrigerators. Exceptions include the works of Li et al. (2008) and Yan et al. (2012), who explicitly state that their methods are intended for the design process [67,68].

Table 2. Analysis of general sustainable modular design methods.

References	Modular Drivers	Modular Methods	Case Study
Ma and Kremer (2015) [22]	Economic sustainability, environmental sustainability, social sustainability	Key components selection and three clustering algorithms	Coffee maker
Yan et al. (2012) [68]	Function, structure, material, manufacturability, component life, end-of-life options	DSM design structure matrix, a kernel-based fuzzy c-means (KFCM), GA	Reduction gear
Yang et al. (2011) [69]	Reuse, maintenance, recycling	Multi-objective optimization with risk constraint based on GGA	Refrigerator
Li et al. (2008) [67]	Disassembly, Reuse/recycle/disposal, material selection	Fuzzy connected graph, AHP, K ordered greedy clustering algorithm	Electrical alternator
Qian and Zhang (2009) [70]	Environmental aspects through the life cycle of the product	Fuzzy graph, fuzzy AHP, similarity analysis algorithm, independence analysis algorithm	Coffee maker
Ji et al. (2012) [71]	Facilitate life-cycle material efficiency	CCF graph and matrix, multi-attribute utility theory, leader-follower bilevel optimization model	Refrigerator
Yu et al. (2011) [72]	Function, structure, component lifetime, material compatibility, recyclability	Modular driving force, GGA	Refrigerator

GA: Genetic Algorithm; GGA: Grouped Genetic Algorithm.

Table 2. Analysis of general sustainable modular design methods.

References	Modular Drivers	Modular Methods	Case Study
Ma and Kremer (2015) [22]	Economic sustainability, environmental sustainability, social sustainability	Key components selection and three clustering algorithms	Coffee maker
Yan et al. (2012) [68]	Function, structure, material, manufacturability, component life, end-of-life options	DSM design structure matrix, a kernel-based fuzzy c-means (KFCM), GA	Reduction gear
Yang et al. (2011) [69]	Reuse, maintenance, recycling	Multi-objective optimization with risk constraint based on GGA	Refrigerator
Li et al. (2008) [67]	Disassembly, Reuse/recycle/disposal, material selection	Fuzzy connected graph, AHP, K ordered greedy clustering algorithm	Electrical alternator
Qian and Zhang (2009) [70]	Environmental aspects through the life cycle of the product	Fuzzy graph, fuzzy AHP, similarity analysis algorithm, independence analysis algorithm	Coffee maker
Ji et al. (2012) [71]	Facilitate life-cycle material efficiency	CCF graph and matrix, multi-attribute utility theory, leader-follower bilevel optimization model	Refrigerator
Yu et al. (2011) [72]	Function, structure, component lifetime, material compatibility, recyclability	Modular driving force, GGA	Refrigerator

GA: Genetic Algorithm; GGA: Grouped Genetic Algorithm.

Several other modular approaches have also been identified to enhance product lifecycle attributes. For instance, Umeda et al. (2008) used self-organizing maps (SOM) to cluster components for sustainability improvement [73]. Sand et al. (2002) introduced the House of Modular Enhancement (HOME) to address multiple lifecycle objectives during product redesign [74]. Smith and Yen (2010) developed an atomic theory based green modularization method [75]. Kimura et al. (2001) and Fukushige et al. (2009) modularized products to improve the reuse of product parts [76,77]. While these methods are interesting and valuable, their contribution to recyclability improvement is unfortunately limited.

Kim and Moon (2016) proposed a novel eco-modularization and assessment method from the viewpoint of product recovery, which could inspire the development of dedicated DfR modular methods and is worth exploring further [78].

Overall, although many studies have been published on green modular design, few have focused specifically on recyclability improvement. While recyclability is a key component of sustainability and green attributes, current methods tend to address it indirectly and inadequately, as they are designed to balance trade-offs among multiple lifecycle objectives, which is already a challenge.

(2)

Other Methods

Apart from modular design, several other methods have been identified to improve a product’s recyclability during the later design stages by optimizing its disassembly ease.

For instance, Viswanathan and Allada (2001) proposed the Configuration–Value (CV) model to identify bottlenecks in product configuration and help designers pinpoint necessary changes to improve disassembly [79]. Kwak et al. (2008) introduced an eco-architecture analysis method to derive the most desirable eco-architecture that facilitates easier disassembly and recycling at the EOL stage [80]. While these methods are valuable for improving disassembly, they are primarily analytical tools that identify bottlenecks and provide suggestions for design improvements.

Furthermore, Ko (2020) proposed a novel green redesign method that offers designers a decomposing–recomposing approach to transform general products into green products [81]. This method combines extension theory with the concept of Green DNAs. Although not specifically focused on recyclability improvement, it provides a preferable approach.

Additionally, Umeda et al. (2013) proposed a design support method that quantifies the effects of design changes, including material alterations, EOL scenarios, and geometric adjustments, on recyclability rates [44]. This research is particularly beneficial, as it quantifies the impact of design changes on recyclability. Combining it with other design approaches could facilitate the automatic generation of products with enhanced recyclability.

3.4.3. Approaches Beyond the Design Stage

(1)

Disassembly Analysis for EOL Products

We also identified a category of articles focused on generating the most profitable disassembly plan for EOL products [11,82,83,84,85]. However, these articles claim that the methods they propose can also be applied at the design stage to assist designers in making informed decisions.

While disassembly experience can feasibly be treated as a form of design knowledge for conceptual design, disassembly analysis and plan generation require complete design information, which is not available at the early design stage.

(2)

Information Exchange for Design

It is widely acknowledged that effective information exchange is critical for improving design efficiency, particularly in the context of DfR. However, research on information exchange to enhance DfR remains limited.

Mangers et al. (2023) proposed a method for collecting, processing, and applying EOL process data, which provides valuable EOL knowledge at the design stage [86]. This allows designers to consider not only user requirements but also the actual EOL process chains, offering a more comprehensive understanding of recyclability early in the design process.

Additionally, Gaye et al. (2010) developed a platform for information sharing and exchange among all stakeholders involved in the recycling process to support more effective EOL recycling [87]. While their focus differs from our aim to guide design improvements, their analysis of the types of information necessary for recycling offers valuable insights for developing information systems for the design stage.

Overall, it is clear that facilitating more effective information exchange is crucial for informed DfR, and there is significant potential for further exploration in this area.

3.4.4. Summary

Many studies have already been conducted to improve product sustainability by selecting more environmentally friendly materials and designing architectures that facilitate easier disassembly. However, it is evident that the biggest issue is the lack of dedicated DfR methods and tools that can be effectively used during the design process to assist in generating products with high recyclability, rather than merely providing design suggestions and support that do not directly contribute to the design process.

4. Discussion

Design for Recycling (DfR) is not a new concept, and a substantial body of research has been conducted, as reviewed in this paper. While existing studies offer valuable insights, there is still a lot to do in achieving effective DfR for more recyclable products in the future. In this section, we first identify current research limitations (in Section 4.1) through the review of three research directions discussed in Section 3. We then propose promising research directions (in Section 4.2) to advance the development of DfR. Finally, we highlight the potential challenges in future exploration.

4.1. Limitations

A detailed review of DfR studies reveals that the most significant and pressing limitation to the effective implementation of DfR is the lack of sufficient data and information necessary for accurate and automated recyclability evaluation and design. Comprehensive data covering the entire product lifecycle, particularly product information and EOL data, is essential for precise EOL option selection, recyclability evaluation, and addressing the “uncertainty” present at the design stage. This data gap also impedes quantitative analysis and informed decision-making in most existing DfR methods and tools.

Moreover, as the core of DfR, precise recyclability evaluation is imperative. As discussed in Section 3.3, the imprecise evaluation of the designed product’s recyclability is mainly caused by insufficient data, the lack of optimal EOL option selection for components before calculating the overall product’s recyclability, and the absence of comprehensive and effective evaluation indicators.

A further limitation is the lack of dedicated DfR methods and tools for designing products with high recyclability. Our review has highlighted that, although many approaches have been proposed to improve product sustainability during the design phase, particularly through design guidelines and modular design methods, there are few strategies specifically aimed at improving recyclability. While recyclability is a key component of sustainability, its enhancement is often constrained by trade-offs among multiple sustainability objectives, limiting the potential for performance improvements.

4.2. Future Directions

To overcome current limitations and develop more effective DfR methods and tools with the aid of emerging technologies, we propose a future research framework (illustrated in Figure 2) that outlines three promising research directions (detailed in Section 4.2.1, Section 4.2.2 and Section 4.2.3) to advance DfR.

4.2.1. DfR Knowledge Base

Data serves as the foundation for informed design and automated processes with minimal human bias, particularly in the context of DfR. Product design is a complex process that involves gathering extensive information, including historical design solutions, user feedback on past products, and user needs for new products [88]. DfR introduces additional recycling considerations, making it necessary to incorporate more information from the EOL stages. Although existing information systems, such as Product Lifecycle Management (PLM), Product Data Management (PDM), and specialized EOL management systems, can support product design, a dedicated DfR knowledge and information system is needed.

Such a system should encompass DfR design guidelines, historical recycling practices, government-imposed design restrictions, as well as detailed information on materials, costs, and other relevant factors. This would enable and support the informed design of products with enhanced recyclability. This area remains underexplored and requires further investigation.

We propose that the development of a DfR knowledge base, leveraging multimodal data collection and advanced technologies, represents a significant research opportunity. Emerging technologies such as the IoT, NLP, LLM, ML, and KG can facilitate data gathering, transform natural language into computable information (e.g., processing DfR guidelines), and extract and infer specific and latent DfR needs.

4.2.2. Recyclability Evaluation Indicators

There are two main barriers in evaluating a product’s recyclability. The first is predicting the optimal EOL option, which requires minimizing human bias and reducing uncertainty at the EOL stage. Developing an automated recyclability evaluation method, potentially supported by AI and driven by sufficient data, shows promise in addressing this challenge.

The second challenge is constructing a set of recyclability evaluation indicators by identifying the comprehensive factors that influence a product’s recyclability and determining how to measure them. This is a complex but essential task. There are three main steps to constructing a comprehensive indicator set. First, identify the factors influencing a product’s recyclability, i.e., determining which factors impact its recyclability. Second, identify key evaluation metrics for recyclability. This step aims to establish how a product’s recyclability should be evaluated. The final step is to construct indicators that define how to measure a product’s recyclability, which is crucial for assessing recyclability improvements during the DfR process.

When constructing the indicator set, it is vital to consider recyclability from all three dimensions of sustainability—economic, social, and environmental—as “recyclability” is not limited to environmental benefits alone. Additionally, incorporating product EOL strategies comprehensively is key to ensuring an accurate recyclability evaluation at the design stage.

4.2.3. Data-Driven Automated Design

The ideal dedicated DfR tool should be capable of automatically generating products with high recyclability. The three research categories reviewed in this study are closely interrelated. Identifying optimal EOL options is an integral part of evaluating a product’s recyclability, while accurate recyclability evaluation is essential for effectively enhancing recyclability during the design process. This interdependence underscores the efficiency and natural alignment of developing a data-driven, automated generative system for DfR, which can be built upon a well-established knowledge base and a robust set of recyclability evaluation indicators.

A data-driven automated generative design system enables the creation of innovative concepts and solutions with minimal human bias, learning from historical data and DfR knowledge, and aided by ML, as illustrated in Figure 2.

However, as Moghaddam et al. (2023) noted, while technology can revolutionize product design, it remains a fundamentally human endeavor, heavily reliant on the ingenuity and creativity of designers [89]. In DfR, designers play a crucial role in defining and incorporating specific design needs related to recyclability (as required by the company or specific products) into the automated generative system, ensuring the generation of the desired product concept. Therefore, a user-friendly HCI interface is essential for the automated DfR generative system.

Furthermore, new recyclability methods based on the comprehensive DfR indicator set and enabled by AI and XR are necessary for automating the evaluation and generation of design concepts. Incorporating automated design evaluation into the generative process allows for the automatic selection of optimal concepts, facilitating the creation of design solutions with the highest recyclability.

4.3. Challenges

However, we must acknowledge that automatically generating products with high recyclability, as we proposed in Figure 2, presents significant challenges. There are substantial limitations to applying AI in the design process, including high operational costs and difficulties in controlling generative algorithms. Much work remains to be done to develop practical tools and platforms that can effectively augment intelligent DfR.

Moreover, it is important to acknowledge that the successful implementation of DfR depends not only on the application of DfR methods and tools but also on the active participation and commitment of the entire industrial chain and society.

Firstly, all stakeholders within the product lifecycle must recognize the importance of enhancing product recyclability and take concrete steps to facilitate its improvement. At the societal level, raising public awareness and education is crucial in order to deepen understanding of DfR concepts, as user opinions and behaviors play a pivotal role in improving product recyclability. Technically, improving stakeholders’ access to product design and recycling knowledge is essential for informed decision-making.

Secondly, economic viability is a critical factor in encouraging manufacturers to adopt DfR. To realize a true circular economy, government incentives and subsidies are necessary to support companies that implement DfR strategies.

Furthermore, strengthening collaboration between academic researchers, industry practitioners, and government agencies is vital for the effective development and implementation of DfR tools that are accessible to both experts and industry practitioners [90].

5. Conclusions

Design for Recycling (DfR) is a promising approach to address the growing challenges of solid waste by enabling the design of more recyclable products in engineering. To advance the development of effective and practical DfR methods and tools, we first clarified its definition by reviewing and analyzing the concepts of “recyclability” and “recyclability improvement”. Through a systematic analysis of 89 papers, we identified key research gaps in the field, including insufficient data availability, imprecise recyclability evaluation methods, and the lack of dedicated and efficient DfR methodologies.

To address limitations, we proposed a future research framework comprising three key solutions: (1) the construction of a comprehensive DfR knowledge base, (2) the development of a robust recyclability evaluation indicator set, and (3) the creation of a data-driven, automated DfR system. These solutions leverage emerging technologies to advance DfR research and practice, particularly in light of growing environmental concerns and the increasing availability of innovative technologies. The proposed framework aims to provide valuable insights for advancing DfR within the engineering field.

It is also important to emphasize that while advanced technologies play a critical role in supporting DfR methods and tools, human designers remain central to the design process. The ultimate goal is to achieve human-AI collaboration, enabling more automated and efficient design processes. Furthermore, the successful and sustainable implementation of DfR requires active participation and commitment from all stakeholders in society.