1 Introduction

Sustainability has been proposed as a way to address the challenges (e.g. climate change, poverty and literacy) posed within the economic, environmental, social, and time dimensions (Brundtland, 1987), as well as their complex dynamic interrelations (Lozano, 2008). For sustainability to address such challenges, it is necessary to address them through a holistic perspective (see Elkington, 1998), i.e. their integration and the interrelations (Dalal-Clayton et al., 2002; Langer & Schön, 2003). Such complexities and broadness of the sustainability concept raise a number of challenges for its implementation (Dyllick & Hockerts, 2002; Hussey et al., 2001; Lozano, 2008), i.e. the translation of a concept, such as sustainability, from ‘theory’ (i.e. definitions) into ‘action’ (i.e. providing results and solutions) (Chofreh & Goni, 2017; Hugé et al., 2013).

A number of efforts have been developed to help implement sustainability, e.g. in organisations (Corsi & Arru, 2020; Hörisch et al., 2015; Lozano, 2020); policy design (Nadin, 2001); and, academic research (Hallstedt & Nylander, 2019; Hugé et al., 2015). Two positions can be discerned in sustainability implementation: (1) ‘implementation frameworks’; and, (2) tools, initiatives and approaches (TIAs).

Sustainability implementation frameworks are aimed at managing a complex topic in conceptual structure by providing a way to understand the active (and iterative) process through which desired objectives are achieved (Saluja et al., 2017) by promoting a multi-tiered implementation (e.g. model or a system) of the whole process (Ahmed & Sundaram, 2012). Two implementation frameworks can be found to provide general guidance; a framework divided into indicators, product-related assessment, and integrated assessment tools (Ness et al., 2007); and, a hierarchical classification of sustainability terms and their relationships by using a systematic approach (Glavič & Lukman, 2007).

The TIAs focus on activities and address the ‘Approaches’ and ‘Sub-systems’ categories proposed by Glavič and Lukman (2007). The TIAs have been used mainly in the corporate context (see Dalal-Clayton & Bass, 2012; Glavič & Lukman, 2007; Lozano, 2020; Ness et al., 2007; Robèrt et al., 2002). There have been many peer-reviewed publications on the use of the TIAs during the last five decades; however, there is limited research on how to translate ‘theory’ into ‘implementation’ for tackling sustainability in academic researchFootnote 1 (Chofreh & Goni, 2017; Moullin et al., 2020).

This paper is aimed at analysing the implementation of TIAs in academic research and compare this against the TIAs ‘definitions’. The paper is structured in the following way: Sect. 2 discusses the implementation of these tools in academic research, Sect. 3 presents the methods, Sect. 4 provides results, Sect. 5, the discussion, and Sect. 6, the final remarks.

2 A review of the TIAs implementation in academic research

A large number of TIAs have been developed, mainly by and for corporations, to better implement sustainability within their systems (Lozano, 2012a, 2012b, 2020), with comprehensive lists (see Dalal-Clayton & Bass, 2012; Glavič & Lukman, 2007; Hoogmartens et al., 2014; Ness et al., 2007; Robèrt, 2000; Robèrt et al., 2002). Some studies have proposed classifications; into sustainable systems, sub-systems, approaches, and principles (Glavič & Lukman, 2007) and, indicators or indices, product-related assessment, and integrated assessment tools (Ness et al., 2007). Other studies have focused on one or two TIAs (Ahi & Searcy, 2015; Gunarathne et al., 2021; Rex & Baumann, 2007), and while others have considered multiple TIAs (Glavič & Lukman, 2007; Lozano, 2020; Robèrt, 2000; Robèrt et al., 2002). Limited, yet increasing, research has provided empirical results on the TIAs’ use, ranging from more general (approaches) to more particular (tools), in the corporate context (Lozano, 2020).

Although many TIAs stress the importance of integrating the sustainability dimensions (see Ness et al., 2007), the majority of the TIAs have focused on the environmental and economic perspective (Atkinson et al., 2000; Lozano, 2012a, 2012b) based on their definitions (i.e. ‘theory’), such as eco-efficiency, aimed at assessing economic and environmental impacts for processes and products (OECD, 1998), and Circular Economy (CE), which links the economic and environmental dimensions (European Commission, 2015). Table 1 summarises the TIAs list provided by Lozano (2020) with its definition, and the sustainability dimensions each one addresses.

Table 1 List of tools, initiatives and approaches (TIAs) with acronyms, definitions, and sustainability dimensions
Full size table

TIAs can help to assess and monitor changes associated with strategies and efforts for implementing sustainability, which can guide decision-making and policy development (Lozano, 2020). Some TIAs have attracted more attention from policymakers (e.g. industrial ecology has potential for US environmental policy (Thomas et al., 2003) and Circular Economy has been used as a product policy framework in the European context (European Commission, 2020)).

The majority of TIAs have been analysed on a conceptual level and in case studies (Corsi & Arru, 2020; Windolph et al., 2014) but only a few studies analyse their implementation (for companies), e.g. TIAs have better implementation results when combined (Lozano, 2020), and their effective use can reduce environmental impacts (Hörisch et al., 2015). Some efforts have been undertaken to assess the interlinkages among some TIAs in companies, including the analysis of how companies adopt the CE principles in cleaner production processes in the regional context (Aranda-Usón et al., 2020) and the interactions between three tools, where it was found that their methodologies are similar enough to be used in a complementary manner (Hoogmartens et al., 2014).

Although there has been considerable research on each of the TIAs published in the literature (Corsi & Arru, 2020), there has been limited efforts on the implementation of TIAs in academic research. This includes descriptive approaches by using bibliometric methods (Meseguer‑Sánchez et al., 2021; Ye et al., 2020), or a limited number of tools, e.g. the evolution of the research output of green chemistry that has established it as a research discipline (Dichiarante et al., 2010), and the use of TIAs in academic research to foster their implementation (Corsi & Arru, 2020; Windolph et al., 2014).

3 Methods

A bibliometric analysis was carried out to analyse the implementation of TIAs in academic research. For the bibliometric analysis, the following steps were followed in this study: (1) Formulation of a search strategy to identify the output of each tool and data collection; and, (2) Development of bibliometric indicators.

The TIAs selected for this study are those proposed by Lozano (2020) with the difference of grouping the ‘Sustainability reporting’ (SR) and ‘Environmental Management System’ (EMS) ones,Footnote 2 complemented with the sustainability dimensions (environmental, economic, social, and time) that are defined based on their definition (‘theory’), partially based on Lozano (2012b) (complete list in Table 1).

The data required for bibliometric analyses were gathered between mid-November and mid-December 2020 from Clarivate Analytics’ Web of Science (WoS) Core Collection (SCI, SSCI, A&HCI). WoS is one of the most well-known multidisciplinary databases with a long and constant coverage of high-quality papers, which is widely used in bibliometric analysis (Mongeon & Paul-Hus, 2016). In this study, different search strategies have been defined for each one of the TIAs (more information on the strategies used in Table S.1). The search was done in the topic field (title, abstract, and keywords) in order to capture all the output of each TIA. All types of documents in WoS were considered, and no temporal limitation once included. Once the data was collected, the following indicators were analysed for the final dataset:

  1. (1)

    Analysing research patterns.

    • Yearly trend of each TIA.

  2. (2)

    Identifying ‘hot topics’ on each TIA.

    • Keywords burst citation. Burst detection is an analytic method to find articles that receive particular attention from the related scientific communities in a certain period of time. This assess the degree of citation intensity for a given reference and keyword. The sudden increases in the usage frequency of keywords (burst strength) based on the citations were identified by using Kleinberg's algorithm to determine the level of ‘hotness’ of the topics of each TIA in academic research (see Kleinberg, 2003).

  3. (3)

    Identifying interrelations between the TIAs.

    • Co-occurrence of the keywords assigned to each paper using the VOSviewer toolFootnote 3 to identify thematic clusters between the TIAs (higher level) within the scientific landscape. The nodes indicate the number of documents, whilst the co-occurrence links identify inter-keyword relationships and a sign of affinity and their thickness, shows the intensity. In addition, a normalisation method used was the Ling/Long modularity (see Chen, 2016) and different parameters of each cluster were extracted (e.g. link strengthavg, yearavg).

    • References and keywords co-citation cluster analysis is used to detect subtopic specialties (lower level) of each TIA with CiteSpace software.Footnote 4 G-index was used to detect the different specialties (see Eggue, 2006) used for node selection that accounts for the citation values of the articles. A correction factor of 5 was applied and the co-citation values were normalized using the cosine index, and the edges were pruned from the network with the pathfinder algorithm. This correction factor provided better visualization results (a number of comprehensive categories). The labels of each cluster were determined mainly by using the logarithm log-likelihood ratio (LLR), which assesses the strength of the bond between the term and the cluster, by considering the abstract text information (see Chen, 2016 for more details). Where there was overlap between the names within the same TIA, the Latent Semantic Indexing (LSI) algorithm was used. With the subtopics identified, a chord diagram was used, which represents connections between the different TIAs.

  4. (4)

    Developing a sustainability implementation framework:

    • The results were integrated to develop an implementation framework for academic research to provide an illustrative overview of these findings. The framework helps to summarize each TIA and discusses its main contributions in combination with the sustainability dimensions and their implementation. To analyse the implementation, the framework relates the TIAs in ‘theory’ (based on their conceptual definitions) and their ‘implementation’ (use in academic research). Each subtopic of the TIAs was classified in each sustainability dimension (Environmental; Social; Economic),Footnote 5 according to the main focus of sustainability. Some subtopics may be more nuanced and can be less easy to delineate, e.g. sustainability reporting (SR) can be divided into the four dimensions (environmental, social, economic and time). Finally, how the different TIAs might relate to each other is discussed and how the framework might be applied to current research.

3.1 Method limitations

Some of the limitations of this study include the use of the keywords for selecting each one of the tools, which conflicts with what is research ‘on’ and ‘related’ to each TIA and other studies that might presumably include ‘buzzwords’’. Once the documents were collected, a validation procedure was conducted to clean the data. Another limitation was the under-representation of other related published works by considering only the Web of Science (WoS) database, which may be indexed in other scientometric databases (e.g. Scopus, Google Scholar, Microsoft Academic, and Dimensions). Additionally, WoS does not cover all academic fields equally, and it is biased towards papers published in English. The methodology proposed may not necessarily capture the complete panorama of research related to each TIAs. Despite that all types of publications from WoS were included in the three databases from the Core Collection, some other typologies of interest (e.g. sustainability reports, grey literature) were not captured.

4 Results

This section presents a descriptive analysis of the research output results for each TIA, divided into scientific output and evolution; hot topics; and interrelations between the TIAs (between TIAs and within their subtopics).

4.1 Research output

A total of 73,672 records (all types of documents considered) were retrieved from WoS through different search strategies based on relevant terms identified from the literature. Figure 1 shows the total number of documents retrieved from the different TIAs (no temporal limit is included). LCA presents the highest number of documents (n = 23,139), followed by GCHEM (n = 14,561) and CSR (n = 12,066) which suggests the TIAS had research interest whereas others (e.g. FX and NAT) scarcely present scientific output. CSR and SR, which are at the top position (3rd and 4th) by output, also coincide with the most widely known TIAs by companies, whereas NAT and FX also appear as the lesser known tools and with less scientific output.

Fig. 1
figure 1

Scientific output of TIAs retrieved from Web of Science (WoS) (1961–2020)

Full size image

Figure 2 shows the evolution of the number of documents by year of publication (since 1987) of all TIAs. The TIAs were ranked by their total output, according to Fig. 2, where darker colours are associated with the highest output and lighter colours with the lowest. Some TIAs’ output were found in early literature (in the sixties) such as LCA, IMS, and CSR. Other TIAs, such as CS, SSCHAIN, and FX, were more recent and had output since the 2000s. As a general tendency, all TIAs’ output has a exponential growth tendency over time, with some exceptions (e.g. NAT, CC, and FX). These last ones might be associated with a scarce and discontinuous number of documents over time, i.e. the maximum number of papers in FX in 1 year is 7. The ones that presented a major growth during the period is CE (31.37), CS (29.53), GCHEM (27.98), SSCHAIN (27.89), and TBL (26.65), while others present a lower growth i.e., FX (0), NS (3.06), and CC (6.72). This shows that most of the TIAs have had the time to be implemented in academic research.

Fig. 2
figure 2

Scientific evolution of the TIAs during the period of study (1987–2020)

Full size image

4.2 Identifying ‘hot topics’ on each TIA

Table 2 lists the keywords with the strongest citation bursts, which represent the TIAs that have received increasing interest (based on citations) since the late 1980. In the period analysed, 280 different bursting keywords, according to CiteSpace software burst analysis, have been identified. From this, we can trace the development of research hotspots. Each keyword was associated with its strength which indicates the relevance of a topic, i.e. it usually indicates potentially interesting studies that have had significant attention in a short period of time. The blue bars (time span) indicate the periods that cover the burst analysis and the red section the years when the strongest bursts occurred. According to the data, the TIAs that had the strongest citation burst were: GCHEM (585.71); IE (174.5); CE (180.89); and, ECO (62.63). TIAs, such as EMS, IE, and LCA, have had more attention based on the number of citations since the early 1990s, whereas some other TIAs (e.g. CE and SSCHAIN) have been more predominant in the last five years. Certain terms that had the longest time span bursts (in red), denoting that their concepts have a keen interest during a longer period such as IE (1995–2012) and EMS (1996–2011). Some TIAs do not present any citation burst, such as CC, NS, and FX.

Table 2 Top keywords with the strongest citation burst
Full size table

4.3 Identifying interrelations between the TIAs

This section presents the TIAs interlinkages results in two levels: 1) at a higher level, the interrelations between TIAs were analysed based on a keywords co-occurrence map, whereas, 2) at a lower level, the subtopics of each one were identified based on co-citation analysis.

4.3.1 Keywords co-occurrence map

The topics addressed in TIAs’ research were illustrated in the keyword co-occurrence map of Fig. 3. 115,251 keywords were identified during the period. Considering a minimum number of 100 occurrences of a keyword, 660 keywords meet the threshold. The highest-ranking keywords were: LCA (frequency of 16,995); GCHEM (7221); and ‘Sustainability’ (5117).

Fig. 3
figure 3

Co-occurrence keywords for the TIAS (nodes = keywords; node size = proportional to publications on each node; edges = co-occurrence of keywords) (< 100 keywords). No links were found for IMS, CC, NS and FX

Full size image

The information on each cluster is presented in Table 3, which shows information such as the cluster number, label assigned, number of nodes, link strength, weight, year and the top-5 most frequent keywords. The largest cluster is #1 GCHEM (with 202 nodes), closely followed by #2 LCA which also presents the higher number of links per paper (link strengthavg of 2596.95), denoting a stronger connection between the articles of this cluster. Cluster 3 is the one that encompasses the great majority of TIAs (CE, CP, CS, CSR, ECOL, DESIGN, EMS, GMARK, SRI, SC, SSCHAIN, and TBL). Cluster 4 is comprised of two TIAs: ECO and IE, which shows that both are interrelated. The map shows that in academic research, TIAs are interrelated, with the exception of clusters 1 and 2. The biggest nodes of cluster 5 ‘waste-related issues’ (i.e. ‘waste’ and ‘recycling’), which does not include any TIA, were bridges between Cluster #2 and #3. These formers were the most recent (yearavg 2016), indicating a current interest in including the recovery perspective (Table 3).

Table 3 Summary table with cluster information
Full size table

This result shows that some TIAs were more integrative and interrelated in their implementation in academic research (e.g. CSR, SR). Some TIAs present a compartmentalized approach, e.g. GCHEM and LCA. They constitute a unique cluster by themselves, denoting its relevance. LCA is interconnected with other TIAs, whereas GCHEM does not present any.

Figure 4 presents an analysis of Cluster #3 (Fig. 3), in order to provide more insights into this pool of TIAs. Cluster #1 encompasses TIAs such as CE, DESIGN, and CP. Cluster #2 includes GMARK and ECOL; Cluster #3 groups TBL and SSCHAIN and Cluster #4 integrates tools related to management and reporting (CS, CSR, SRI, SR and EMS).

Fig. 4
figure 4

Co-occurrence keywords for the TIAS of sustainability-related cluster (< 100 keywords)

Full size image

4.3.2 Subtopic specialties

Figure 5 shows a circular bar plot with the top-5 research specialties for each TIA (see Table S.2.). The cluster labels were obtained from the abstract in the citing papers using the LLR algorithm. In case there is the same name for different clusters, other labels (by using LSI approach) are used, i.e. natural step has two clusters with the label ‘informing LEED’s’; one was changed to the label obtained by the LSI (‘new trend’). The size of the bars indicates the number of documents that integrate each of the topics. Some TIAs shared subtopics, for example, CSR has a subtopic, TBL; TBL has one subtopic related to SSCHAIN. This shows that some tools can be complementary and can be used for strategic sustainability implementation.

Fig. 5
figure 5

Circular bar plot of the top 5 sub-topics by the twenty TIAs

Full size image

The connections of each TIA with other TIAs were analysed by using a chord diagram, see Fig. 6. Each TIA constitutes a section of the circular layout. The numbers indicate the number of connections between the subtopics (i.e. a number one means this specific TIA only had one connection to another TIA). The arcs were drawn between the tools in case there are one (or more) subtopics interrelated. The size of the arc was proportional to the importance of the flow (i.e. number of subtopics that were shared). Only TIAs that present, at least, one connection were drawn. CE and LCA (7 connections each), SSCHAIN and DESIGN (6) present a higher number of links with the other TIAs. From the analysis, two profiles can be deduced: 1) ‘Provider’, including DESIGN (with 6 subtopics); SR (5 subtopics); and TBL (3 subtopics); 2) ‘Receiver’, such SSCHAIN (5), and CE (6). The latter is related to the fact that some TIAs are more ‘transversal’, or easily adaptable because of their integration with others (e.g. an initiative such as TBL could be better aligned with other tools).

Fig. 6
figure 6

Chord Diagram with the interrelations between the TIAs

Full size image

5 Discussion

The results show publications for most TIAs (fourteen) for more than twenty years, which evidences their rate of implementation (in line with GCHEM that is established as a discipline (see Dichiarante et al., 2010)). The burst analysis indicates that some TIAs (e.g. IE, GCHEM and CE) have become ‘hot topics’ and have a better implementation and policy potential (see European Commission, 2020; Thomas et al., 2003).

The keyword co-occurrence maps show that the majority of the TIAs (and their subtopics): (1) have a better balance, in regards to the sustainability dimension, in implementation than in ‘theory’ (i.e. TIAs’ definitions) (complementing Chofreh & Goni, 2017); and, (2) are interrelated in academic research (providing new insights to Corsi & Arru, 2020; Glavič & Lukman, 2007; Hoogmartens et al., 2014; Lozano, 2020; Ness et al., 2007; Robèrt, 2000; Robèrt et al., 2002). The most interconnected TIAs belong to the management and reporting cluster (concurring with Lozano, 2020), whereas the least connected one is GCHEM (in contrast to Lozano, 2020).

From the interrelation analyses two types of interlinking profiles were deduced (providing new insights to Glavič & Lukman, 2007; Lozano, 2020): (1) ‘Provider’ profile (e.g. DESIGN); (2); and, ‘Receiver’ (e.g. SSCHAIN and CE).

The results from the ‘theory’ (i.e. TIAs definition) were compared with their ‘implementation’ in academic research (Table S.3) showing that all TIAs address at least one sustainability dimension, with the environmental dimension being the most frequently addressed in academic research (45%) (concurring with Atkinson et al., 2000; Lozano, 2020; Lozano & Huisingh, 2011). The TIAs focusing on social dimension are higher in academic research (22%) than in companies (contrary to Lozano, 2020 in which this dimension was not sufficiently addressed).

The results show that some TIAs in their implementation have a more integrated approach by covering more dimensions than in ‘theory’ (e.g. CE, DESIGN, ECO, ECOL, EMS, GMARK, IE and LCA) and that the TIAs related to management and reporting (Fig. 3, Cluster 3) are the most interconnected and address the most sustainability dimensions (except for SR and SSCHAIN) (providing new insights into the corporate context discussed by Glavič & Lukman, 2007; Lozano, 2020).

The results were integrated to develop the ‘Sustainability Implementation Framework’ aimed at helping to understand the implementation of TIAs in academic research (Fig. 7). The framework compares the ‘theory’ (i.e. definitions), left side of the figure, with the ‘implementation’ (i.e. use) in academic research, right side of the figure. The TIAs were divided into three different levels (a multi-layered implementation, concurring with Ahmed & Sundaram, 2012) from general to particular (i.e. Initiatives, Approaches and Tools, based on Glavič & Lukman's, 2007 framework and complementing it with the proposals by Lozano, 2012a, 2012b, 2020). The Tools (more particular focus), the Approaches, and the Initiatives (more general focus) should be aligned for a better contribution to sustainability, and the theory and implementation should be more congruent. The triangles aim to illustrate that the sustainability dimensions are more clearly separated, whereas the circles depict the more connected.

Fig. 7
figure 7

Sustainability Implementation Framework (SIF) for academic research. For the theory (left figure), triangles were used to show the dimensions of sustainability more clearly separated, while for the ‘implementation’ circles (right figure) were used to show a stronger connection between the dimensions

Full size image

6 Conclusions

In the last five decades, there has been an increasing interest in the use of tools, initiatives, and approaches (TIAs) in academic research. Sustainability plays a pivotal role in addressing their implementation and has to encompass a holistic perspective, including the four dimensions (economic, environmental, social, and time), as well as their interrelations. However, the majority of such efforts in academic research have focused on descriptive approaches (e.g. bibliometric approaches) with a limited number of tools. In addition, most of the implementation frameworks developed to date for those tools remain theoretical and none of them have been applied in the academic research.

A bibliometric analysis was conducted to analyse the implementation of the most widely used TIAs in academic research. The top twenty TIAs were analysed, covering the period 1961–2020, and bibliometric indicators such as research patterns (yearly trend), hot topics (burst analysis) as well as their interrelations (co-occurrence maps and co-citation cluster) were examined.

This study analyses the implementation of TIAs in academic research and compared against the TIAs ‘theory’. The results show TIAS implementation in academic research can foster sustainability. The interrelationships between the TIAs (and their subtopics) in their implementation highlight that their use is more holistic and can better address the complexity of sustainability.

The results were integrated into a Sustainability Implementation Framework (SIF), which is aimed at helping to understand the implementation of tools in academic research. SIF demonstrates that TIAs have a more holistic and balanced approach in their implementation rather than in ‘theory’. SIF shows that TIAs can improve their congruence by better linking the implementation of TIAs with their theory, which is paramount for sustainability change. SIF can be helpful to sustainability researchers for organising the information (e.g. by levels) of the implementation of the tools and can provide guidance on the different interactions in academic research, thus helping to advance the sustainability transformation.

For a better implementation of TIAs in academic research, it is necessary to address sustainability dimensions (economic, environmental, social, and time) in a holistic and balanced way, considering the alignment of general and specific efforts, i.e. TIAs, and congruence (linking ‘theory’ and ‘implementation’). The TIAs’ implementation should follow more strictly the definitions, or, perhaps, the TIAs’ definitions should be redefined to encompass the insights from their implementation.

Further research should be carried out on specific case studies, countries and sectors to test the framework, which can help to gain an insight into the practical implementation of the tools. The practical use of TIAs by different types of organisations should also be explored, as well as their motivation (e.g. reason why the tools are used) and limitations. In addition, the use of these tools in innovation (e.g. research and development projects), and how to incorporate the time dimension, should also be investigated.