Implementation of digital technologies for a circular economy and sustainability management in the manufacturing sector

Article history: Received 7 June 2022 Received in revised form 5 September 2022 Accepted 15 November 2022 Available online 19 November 2022 Editor: Dr. Abbas Mardani There is growing consensus in literature and practice that digital technologies (DTs) can offer a wide range of potentials for implementing a circular economy in companies. However, empirical insights on how the potentials of different DTs are already realized across various industries are lacking. This study addresses this research gap through descriptive, hierarchical cluster and non-parametric analyses (Kruskal-Wallis tests and Spearman rank correlations) of the use of DTs for circular economy and sustainability management based on data collected in a structured telephone surveywith 132 sustainability managers and CEOs of Austrianmanufacturing companies. The paper shows for the first time the degree and stage of implementation of four key enabling DTs for a sustainable circular economy and 31 specific applications of those DTs across eight different industries. Of the four DTs, Internet of Things (IoT) technology is most widely implemented, followed by big data analytics, artificial intelligence (AI), and blockchain technology. However, their use in sustainability management is still very limited and is currently mainly in pilot phases. Of the 31 applications surveyed, IoT technology is most frequently used for collecting data from production processes, AI for predictivemaintenance, big data analytics for demand forecasting, and blockchain technology for tracking product origins. Statistically significant differenceswere found in the degree of digitalization, in general, and for sustainability management between industries and company sizes. A strong positive correlation between the implementation of digital technologies in general and their use in a sustainability management context indicates synergies and spillover effects. The findings may help to tailor context-specific and purpose-driven strategies that selectively leverage the benefits of different DTs and supporting sustainability management effectively. Further research may identify scalable best practices, optimal enabling conditions, and environmental and social outcomes. © 2022 The Authors. Published by Elsevier Ltd on behalf of Institution of Chemical Engineers. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Digitalization and the circular economy (CE) are two emerging technological and organizational trends with the potential to improve company productivity and sustainability performance . Research on the CE and on the related potential of digital technologies (DTs) has thus increased substantially in recent years (Di Maria et al., 2022;Rosa et al., 2020). Except for occasional warnings, for example, "[…] the alleged sustainability benefits of digital technologies may turn out to be ill-founded […]" (Andersen et al., 2021, p. 97), there is now a broad consensus among scholars that DTs play an essential role in the transition from the current linear economy towards a CE (Kristoffersen et al., 2020;Lieder and Rashid, 2016;Pagoropoulos et al., 2017;Rusch et al., 2022). Much of the related research to date consists either of literature reviews identifying the potential benefits of DTs (Agrawal et al., 2021;Demestichas and Daskalakis, 2020;Lopes de Sousa Jabbour et al., 2018), or of reports on single or multiple case studies illustrating the actual benefits of using DTs for a CE and in sustainability management (Bressanelli et al., 2018;Ingemarsdotter et al., 2021;Nascimento et al., 2019). Many of these studies focus on one or more (forerunner) companies in a case study (Bressanelli et al., 2018) and/or on a selected DT in specific corporate application areas, e.g., the use of blockchain technology for sustainable supply chain management (Saberi et al., 2019). In contrast to this, empirical studies examining corporate digitalization strategies and the implementation of DTs in a sustainability and CE context are relatively few and far between. Examples include Neligan (2018), who explores digitalization strategies to enable a CE, Di Maria et al. (2022) and De Marchi and Di Maria (2020), who analyze the relationship of industry 4.0 technologies and CE, Chiarini et al. (2020) who assess industry 4.0 and manufacturing strategies (such as environmental sustainability or servitization), and Neligan et al. (2022) who reveal the degree of digitalization of business models and product service systems (PSS) of German manufacturing firms.
Overall, these studies provide an overview of selected DTs and their effects on certain CE outcomes or on sustainability performance. However, they do not explore in detail how different DTs are already being applied, either in general and/or in sustainability management, nor do they allow for comparison of the digitalization strategy or DT implementation level across industries and company sizes. Hence, while the potential of specific DTs may already have been discussed or demonstrated in various case studies, there still remains a lack of research concerning the systematic assessment and comparison of the degree of implementation of digitalization strategies and various DTs across potential applications, industries, and company sizes. Such detailed explorations are needed in order to describe the status quo and to contribute to a better understanding of current DT developments and the related implications for sustainability management and transition towards a CE.
The present study therefore attempts to address the following research question (RQ): RQ: To what extent have manufacturing companies implemented digitalization strategies and digital technologies in general and in sustainability management in a circular economy?
In order to answer this RQ, a structured telephone survey was carried out with 132 sustainability executives and CEOs from Austrian manufacturing companies. Manufacturing companies were selected because they play a central role in the CE transition and because they may benefit from the use of DTs in several different ways. The specific aim of the study was to investigate (1) the extent to which companies have already implemented digitalization strategies, (2) the extent to which they implemented different DTs in general, (3) the extent to which DTs were implemented specifically for purposes of sustainability management, and (4) how the degree of DT implementation differed across industries and company sizes. This paper is structured as follows: short introductions to digitalization and DTs (both in general and for CE and sustainability management) are given in Section 2. Section 3 deals with the study design. Section 4 covers the study results and is divided into four subsections. Section 4.1 describes the company digitalization strategies and company use of DTs. The differences in the use of DTs per industry are illustrated in Section 4.2 and analyzed statistically in Section 4.3. The correlations between the presence of digitalization strategies and the implementation of DTs, both in general and in a sustainability context, are then shown in Section 4.4. Section 5 offers some items of discussion, followed by final conclusions in Section 6.

Digitalization strategies and digital technologies
In general, digitalization refers to a process of change characterized by an increased use of DTs (Hess, 2019). Most generally, DTs can be defined as the combination of connectivity, innumerable and dispersed items of information, communication and computing technologies (Bharadwaj et al., 2013;Hanelt et al., 2021). And while digitalization strategies are recognized as essential preconditions for the implementation of DTs, the advantages also depend on the management of these digitalization efforts (Björkdahl, 2020). From the perspective of companies, DTs can facilitate modular, distributed, cross-functional, and global business processes that enable tasks to be carried out across the boundaries of time, function, and distance (Bharadwaj et al., 2013). DTs are also being increasingly embedded in products (e.g. IoT sensors) in order to take advantage of digital resources and to create new capabilities (Bharadwaj et al., 2013). The use of different DTs in a manufacturing context is also often referred to as 'Industry 4.0', i.e., the fourth industrial revolution, a term introduced in Germany in 2011 (Lu, 2017).
Given the organizational changes required for and induced by the implementation of DTs (Hess, 2019), companies need adequate digitalization strategies. Such strategies require comprehensive information sharing across digital platforms inside and outside companies, and need to facilitate multifunctional processes and closely integrate cross-enterprise information technology (IT) capabilities (Bharadwaj et al., 2013;Rai et al., 2012).
The presence of a corporate digitalization strategy has been found to correlate positively not only with a higher implementation rate of DTs, but also with higher material efficiency (Neligan, 2018). In a survey of 589 German manufacturing companies, Neligan (2018) showed that companies with "highly developed" digitalization strategies, utilize DTs more frequently and report higher rates of implementation for six out of seven measures of material efficiency, such as the optimisation of manufacturing processes, or the use of new materials. Sodhi et al. (2022) collected 405 answers from supply chain managers and also found the goal with the highest priority when using DTs is operational efficiency.

Digital technologies enabling a sustainable and circular economy
The CE may be understood as one among several complementary strategies for sustainable development (Schöggl et al., 2020) as the concept envisions a regenerative system that minimizes the use of natural resources and the creation of waste to sustain natural capital (Di Maria et al., 2022;Geissdoerfer et al., 2017). To implement a CE and enable more sustainable consumption and production patterns, the involvement of all state and non-state entities is needed with a special emphasis on the private sector in general and manufacturing companies in particular. They are highly dependent on raw materials, including valuable metals, which could pose future challenges to operations managers due to material shortage risks (Bag et al., 2021b) but also is highly relevant for contributing to a country's economic development (Amjad et al., 2021). These factors, combined with the environmental and social costs of unsustainable production patterns, put manufacturing companies at the center of the CE transition (Bjørnbet et al., 2021).
DTs are increasingly considered as critical enablers for the CE transition, especially in terms of the collection, management, aggregation, and exchange of product data (Di Maria et al., 2022;European Commission, 2020;Nara et al., 2021) DTs can be used for monitoring products and parts in multiple lifecycles on an industrial level (Lieder and Rashid, 2016). Furthermore, they help improve the availability and quality of data necessary for sustainability-related decision making on both a process and product level (e.g., concerning a machine's realtime energy consumption) (Grischa et al., 2022), and they underpin new dematerialized business models at the organizational level (e.g., PSSs) (CEID et al., 2021;Neligan, 2018;Neligan et al., 2022). In addition to the various direct applications of DTs in a CE context, they can also support the CE transition indirectly. For example, an increased degree of digitalization along a value chain can foster inter-and intraorganizational collaboration and knowledge sharing (Di Maria et al., 2022).
Several deductive empirical studies provide evidence that use of DTs can indeed have a positive influence on the implementation of more sustainable and circular company practice. For example, Bag et al. (2021a), based on a PLS-SEM study of 124 companies in South Africa, find that DTs exert a positive effect on companies' "10R capabilities" (derived from the 10R strategies (Kirchherr et al., 2017;Potting et al., 2017)). In this study, the authors also find a positive impact of the 10R capabilities on the company's overall sustainability performance (Bag et al., 2021a). Another study by Di Maria et al. (2022) analyzed industry 4.0 technologies and the mediating role played by supply chain integration (the strategic collaboration of actors involved in valueadding customer activities) in CE outcomes. Such a mediating effect of supply chain integration was found to exist for smart manufacturing technologies. In a multiple regression based on a survey of 378 managers of Italian enterprises, Del Giudice et al. (2020) discovered that a big-data-driven supply chain moderates the association between CE HR management and CE supply chain-related company performance. In a survey of Italian manufacturing companies Chiarini et al. (2020) showed that the implementation of industry 4.0 technologies leverages specific manufacturing strategies such as supply chain integration, lean strategies, environmental sustainability, and servitization. However, they found no positive correlations with companies' environmental sustainability performance. Finally, based on a secondary data-based country-level model Hong Nham and Ha (2022) find indications that an initial development of digitalization (e.g. measured by e-commerce adoption or cloud usage) may foster the CE transition in European countries but that an overdevelopment may hinder it.
As these and other deductive studies on the implementation of DTs in companies show, several management theories help explain and contextualize it. Of them, the resource-based view (RBV) was found to be the most often used theory in studies on the adoption of DTs in supply chains (Yang et al., 2021). According to the RBV, DTs represent tangible resources of a company (Rodríguez-Espíndola et al., 2022). Via resource orchestration, an extension of the RBV, these digital resources can be effectively structured and bundled to create DT-related capabilities that can be leveraged for value creation (Kristoffersen et al., 2021). Resource orchestration specifically emphasizes the role of managers in capitalizing on a firm's resources (Sirmon et al., 2011). The need for DTspecific capabilities is also reflected in the dynamic capabilities, another pertinent theory, according to which companies need a DT-specific set of dynamic capabilities to leverage the potential of DTs and attain a competitive advantage (Ellström et al., 2022). The implementation of DTs is preceded and supported by the acquisition of DT-specific intangible resources (i.e., expertise and knowledge), for which the absorptive capacities provide a suited framework (Siachou et al., 2021). Given that our study focuses on the implementation of DTs from the perspective of sustainability managers, our study is underpinned by the resource-based and resource orchestration view.

Selected digital technologies for sustainability management
While a recent literature review listed a total of 111 terms used to denote and specify the potential applications of DTs (e.g., sensors, embedded systems, machine networks, smart products, smart manufacturing, …) (Klingenberg et al., 2021), these terms can be related to a few main underlying DTs. Our study draws on the recent classification used in the New Circular Economy Action Plan (European Commission, 2020). Here, the European Commission defined the following four DTs as being essential enablers in the CE transition: (1) artificial intelligence (AI), (2) big data analytics, (3) blockchain and (4) IoT technology. These four DTs can enable digital manufacturing in the sense of an Industry 4.0 and a growing body of scientific literature is detailing their potential for improving robustness, accuracy, efficiency and transparency in CE and sustainability-related operations and decision-making (Liu et al., 2022;Rodríguez-Espíndola et al., 2022;Rusch et al., 2022). More generally, they are key technologies for solving complex problems and allow real-time information sharing and a wide range of analytics for improved management and decision making (Gill et al., 2019;Sodhi et al., 2022). Except for blockchain technology, the studied DTs are also part of the European Commission's Digital Economy and Society Index (DESI) 2022, which enables comparison to other EU countries (European Commission, 2022a). Table 1 provides a description of the four DTs, as well as a few examples of their areas of application in CE and sustainability management. Table 1 Description and sample applications of four selected digital technologies. AI = artificial intelligence; IoT = Internet of Things.

Digital technology
Description Examples of applications in sustainability management in a CE Internet of things (IoT) "[…] a network of entities that are connected through any form of sensor, enabling these entities, which we term as Internet-connected constituents, to be located, identified, and even operated upon." (Ng and Wakenshaw, 2017, p.4) Used to track the life cycle data of products to improve their maintenance (Bressanelli et al., 2018;Ingemarsdotter et al., 2019); improving recycling efficiency using additional data (Tao et al., 2016); addressing a company's energy management and sustainable production practices (Ren et al., 2019); product status monitoring (Zhang et al., 2017); fleet and product condition monitoring (Manavalan and Jayakrishna, 2019); collecting use-phase data to inform product design (Bressanelli et al., 2018;Ingemarsdotter et al., 2020); life cycle data collection for data-driven recycling (Kim et al., 2017), and reverse logistics decisions (Garrido-Hidalgo et al., 2020). Big data analytics "[…] is the process of examining large and varied data sets to uncover hidden patterns, unknown correlations, market trends, customer preferences and other useful information that can help organizations make more-informed business decisions […]." (Ren et al., 2019) Green supply chain management and environmental impact assessment (Kaleel Ahmed et al., 2018); preventive and predictive maintenance activities so as to facilitate the lifetime extension of products (Bressanelli et al., 2018;Li et al., 2015); increasing equipment energy efficiency (Li et al., 2015); analyzing consumer behavior (Ge and Jackson, 2014;Xu et al., 2015); enhancement of data quality and completion of data in sustainability assessment (Belaud et al., 2019); inventory management, optimizing transportation, demand analysis (Kaleel Ahmed et al., 2018); energy and operations management, collecting social sustainability data (Corbett, 2018), knowledge discovery in databases to enhance the efficiency of the design, production and service processes (Lv et al., 2018). Artificial intelligence (AI) "It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable." (McCarthy, 2004, p.2) Used to help manufacturers to select the best possible sustainable supplier and to enhance the manufacturers' sustainability (Fallahpour et al., 2016); for environmental impact assessment (Kaab et al., 2019); establishing traceability along a product's life cycle (Okorie et al., 2018); energy and environmental impact predictions (Nabavi-Pelesaraei et al., 2018). Blockchain technology "[…] a distributed data structurea distributed ledger -in which the data is shared on a peer-to-peer network. The network membersnodescommunicate and validate the data following a predefined protocol without a central authority." (Esmaeilian et al., 2020, p.3) Used for the verification of fair work practices in production, (Saberi et al., 2019); transparent record keeping of a product's origin/history ; carbon emission trading based on smart contracts (Zhang et al., 2020); tracing of materials (Teh et al., 2020); new business models (Narayan and Tidström, 2020).
J.-P. Schöggl, M. Rusch, L. Stumpf et al. Sustainable Production and Consumption 35 (2023) 401-420 More extensive discussion concerning the description, application potential, or functionalities of the four DTs is beyond the scope of this paper. For such information the reader is referred to dedicated journals 1 and literature reviews summarizing the potentials of one or more of these DTs for corporate sustainability and CE management (Cagno et al., 2021;Okorie et al., 2018;Rosa et al., 2020;Rusch et al., 2022). Among these reviews, Rusch et al. (2022) focused on all four DTs and compiled 146 examples of application and then classified these in terms of their useability along the product life cycle, and their role as enablers in CE strategies and in specific sustainable product management activities. The 146 examples analyzed by Rusch et al. (2022) frequently come from large companies and most are related to IoT technology, followed by big data analytics, blockchain technology and AI. Examples of application for all four DTs are most often linked to optimizationrelated CE strategies, and are characterized by increased efficiency and/or by improved circularity performance of products or production processes. CE strategies that entail more radical changes to company processes or business models are still relatively infrequent in the scientific literature. While the four DTs are already being used to address a wide variety of environmental issues, only a few studies have investigated their potential regarding social sustainability issues (Corbett, 2018;Saberi et al., 2019;Teh et al., 2020). However, a broader engagement with the use of DTs in the context of social sustainability is likely to be beneficial, as DT use (e.g., for predictive analytics) has been shown to be positively associated with the quality of company and supply chain environmental and social performance (Dubey et al., 2017).
In summary, Section 2 shows that first, several studies exist that focus on DTs and different mediators/moderating effects on e.g., supply chain management, environmental sustainability or servitization. And secondly, there are already theoretical descriptions as well as examples from practice how DTs can be used to enhance a CE in industry. The study builds on this knowledge and goes a step further and derives findings from an empirical sample to show how and to what extent different DTs are already being applied. This creates to derive a detailed picture of the status-quo of the implementation status of DTs in the Austrian manufacturing industry in general and for sustainability in particular. Additionally, the study shows the sustainabilityrelated application area per technology which was not yet done in another study on this detailed level.

Methods
In order to complement the above theoretical background, we now move on to the methods used in the present study and briefly describe data collection, data sample and data assessment.

Data collection and sample
Random sampling was used to select a total of 1549 manufacturing companies from a proprietary company register. Of these, 132, either managers (with responsibility for sustainability and CE activities) or CEOs (in companies where a sustainability manager was not available), responded to a telephone survey. This represents a response rate of almost 9 %. Although this is quite low it is still in line with the response rate found in other studies dealing with the social and environmental sustainability literature (e.g. (Demirel and Kesidou, 2019;Di Maria et al., 2022)). The purpose of the survey was first explained, and if needed, further information in the form of a cover letter was sent via email. The survey was conducted via telephone. The geographical area was Austria, and only companies based on manufacturing, as reflected in their NACE codes, were contacted. The focus on Austria allows insights into how DTs are utilized for a CE in OECD countries, based on the example of a country that (1) ranks among the top industries of the EU, according to its contribution to global GDP (i.e. rank 13 among EU member states), (2) is among the most advanced EU countries in terms of CE adoption (Mazur-Wierzbicka, 2021), (3) and demonstrates consistent support to Industry 4.0 implementation, besides Germany Belgium, Denmark, Finland, France, Italy, Netherlands, Portugal, Spain or Sweden (Teixeira and Tavares-Lehmann, 2022). The telephone survey was conducted between February and December 2021. Information was collected on company size (employees), annual turnover, and area of focus (Fig. 1).
The sample includes companies below 50 employees (small companies, 46 %), companies with 51 to 250 employees (medium-sized companies: 32 %), and companies with > 251 employees (large companies: 22 %) (Fig. 1c). Similar categories were employed with respect to company annual turnover. Small companies have a turnover below 10 million euros (48.4 %), medium sized companies between 10 and 50 million (29.4 %), and large companies have an annual turnover above 50 million euros (22.2 %) (Fig. 1b). Company distribution according to industry was also addressed (Fig. 1a). The most frequently represented industries in the sample are machinery and metal goods, chemicals, and food. The remaining five areas include a catchall category 'other', comprising those companies that did not match any of the other classes, for example, sports equipment, furniture, printing equipment or ceramics.

Survey questions
The survey made use of scales and constructs derived from literature. Most of the response options were displayed on a five-point Likert-type scale (ranging from "not considered" to "implemented company-wide"), adopted from  and those related to the four selected DTs were complemented with open questions. As the open questions allowed companies to add further DTs they thus 1 E.g., "Internet of Things; Engineering Cyber Physical Human Systems", "Big Data Research", "Artificial Intelligence" or "Blockchain: Research and Application". helped to control for potential omissions in the given definitions. These open questions were answered by only three of the 132 companies, and the technologies mentioned (EDI interfaces, CAM systems, internet security and cloud-based control systems) could be either related to our four overarching DTs or they referred to the use of IT more generally. Table 2 provides an overview of the scales used, the number of items, and their respective sources. Industry affiliation and company size (measured by turnover and number of employees) were included as control variables as it was assumed that they might influence the level of DT implementation. This also complies with recent demands for more detailed examination regarding the impact of company size on implementation differences (Nara et al., 2021).

Data analysis
The data was analyzed using a combination of descriptive and inferential statistical methods. A hierarchical clustering algorithm was used for ordering and grouping industries regarding their degree of DT implementation. Hierarchical clustering was chosen with the sample size in mind, because as an exploratory data science-based analysis, it allows to identify similar response behavior based on the total sample. Hierarchical clustering is an explorative method used to find groups or subgroups in data sets. In hierarchical clustering, groups are unknown, and hence observations are clustered according to their similarity (Handl and Kuhlenkasper, 2017). Euclidean distance was used to measure dissimilarity between observations since it can be used for objects described by vectors of numeric features (Provost and Fawcett, 2013). Subsequently, as it tends to produce well-balanced clusters, average linkage was used to measure intercluster distances (James et al., 2013). The clusters are illustrated using heatmaps. To test for potential statistically significant differences between industries and company sizes (based on number of employees and turnover), use was made of Kruskal-Wallis tests, due to non-normal distribution of the variables (Handl and Kuhlenkasper, 2017). To specify statistically significant differences, post-hoc Dunn-Bonferroni tests were computed. Spearman's rank sum correlation tests were used to test for correlations between the implementation of digitalization strategies and the implementation of DTs in general and for sustainability management. Finally, a network analysis was used to illustrate the co-usages of different technologies by the surveyed companies. R version 4.1.2 (R Core Team, 2019) and Gephi version 0.9.2 (Bastian and Heymann, 2009) were used for the analyses.

Digitalization strategies and IT resources
In the sample, around 48 % of the manufacturing companies have a digitalization strategy (either already in place company-wide or implementation is underway). Additionally, 34 % already have implemented a digitalization strategy in a pilot project. In total, that means that around 82 % of the companies either have a digitalization strategy already in place or are currently working on its implementation.
Regarding the digitalization strategies for the business areas, the results show that most companies already have a digitalization strategy for sales, supply chain management and for marketing (Fig. 2). Around 68 % of the companies have a digitalization strategy for product management, 66 % for product development tasks, and 56 % for sustainability management tasks (either already implemented company-wide, implementation is underway or in a pilot phase (sum of all bars on the right-hand side)). Sustainability management is the area with the highest percentage of companies (32 %) that have pilot projects in place (light blue bar). The qualitative notes and suggestions provided by respondents showed that production planning, human resources (HR), and production, are further areas where digitalization strategies are in place in individual cases.
The analysis of whether appropriate IT or other digital systems were available to support a CE business model, products, or services showed that around 37 % of companies have appropriate IT for a CE (already implemented company-wide (19 %) or implementation is underway (17.6 %)). 17.6 % have already implemented IT or digital systems for a CE in a pilot project. In total, that means that around 54 % of the companies have appropriate IT or digital systems for a CE either already in place, or are working on its implementation (see Table B 1 in Appendix B).

Implementation of digital technologies
In general, company-wide implementation of DTs is relatively low. For sustainability management in particular, it appears to be more of a phenomenon found in single cases than an established business strategy. Regarding the use of specific technologies, the results show that  47 % of the companies in the sample use IoT technologies (the sum of bars on the right-hand side of Fig. 3a). However, in 30 % of the cases it is used in a pilot project (light blue bar) and in only 7 % is its use company-wide. Big data analytics is being used in 33 % of companies. AI and blockchain technology applications are still relatively seldom (24 % and 14 %, respectively). Regarding the use of DTs for sustainability management (Fig. 3b), the corresponding numbers are lower, for example, for IoT technologies, we now have a figure of 18 % (sum of bars on the right-hand side) instead of 47 %. Fig. 3b also shows that big data analytics is used by 16 % of the companies, AI by 9 % and blockchain technology by only 6 %.

Use of digital technologies per industry
This section illustrates industry differences with respect to the degree of DT implementation. Subsection 4.2.1 covers the general differences, and Section 4.2.2 illustrates the differences concerning specific applications of the different DTs. Fig. 4 illustrates how frequently companies from the various industries reported using the four DTs, both in general and in the context of sustainability management. The figure is normalized to one (i.e., 100 %) per industry (i.e., row). Thus, for instance, all the five companies surveyed from the construction industry reported the use of IoT technology in general (illustrated by the dark shading of the first quadrant in the first row of Fig. 4). All cases were counted in which a DT is used at least in a pilot project (i.e., the three blue answer options on the right-hand side of Fig. 3). As shown in Fig. 4, IoT stands out as the most frequently used technology in all of the industries surveyed. It is most often used in construction (100 %) and electronics (80 %), and in companies from the category 'other' (64 %). The second most frequently used technology in general, big data analytics, is most often used in the electronics (67 %), chemical, and wood industries. Like IoT, AI is also used most often in the construction (60 %) and the electronics industries. Finally, blockchain technology is most often used by companies from 'other' industries (27 %), followed by the construction and wood industries.

Digital technologies in general and for sustainability management
While use of the digital technologies in a sustainability management context is relatively seldom, (as illustrated in Fig. 3) in their respective distributions across the various industries are generally quite similar.
On the one hand, industries that already exhibit frequent overall use of DTs, tend also to be among the most frequent users of the respective technology for sustainability management. The electronics, machinery and metal, chemical, and wood industries serve as examples here. They occur at least twice in the top half of the rankings in both categories (i.e., regarding the general and sustainability-related application of DTs). By the same token, low implementation rates in general, are also often mirrored in the sustainability context. For example, the food and textile industries are among the bottom three in both categories for all four technologies (i.e., as indicated by the light shading in Fig. 4).
The results of the cluster analysis further detail similarities and differences between DTs (in the columns) and the industries (in the rows). The cluster analysis results are illustrated in Fig. 4 by the gaps between rows and columns and the dendrograms on the left and top of the figure. Among the industries depicted, the construction and electronics industries appear to be forerunners, and the textiles and food industries are laggards, with implementation patterns across the remaining industries being relatively similar. The clusters among the DTs (i.e. the columns) illustrate that sustainability and blockchain applications are more seldomly used than IoT, big data and AI. The latter three technologies each have implementation patterns that differ from the rightmost cluster in Fig. 4 and are also distinct enough to form their own clusters (as indicated by the gaps between their columns).
In summary, the comparison across industries indicates that construction, machinery and metals, and electronics are the most advanced concerning DT use in general, and also partly with respect to sustainability management. As shown above (Fig. 3) use in sustainability management remains relatively seldom across all industries and is most advanced in the sectors construction, electronics, and 'other'. 2 The results remain largely unchanged when further differentiating them regarding the extent of implementation as can be seen in Appendix B. There, Fig. B2 illustrates more specifically, for which DT company-wide implementation is ongoing or finished (i.e., the two highest response options, blue and dark blue, in Fig. 3), or which DTs have already been implemented company-wide (i.e., the highest response option, dark blue, in Fig. 3). The construction, machinery and metals, and electronics industries are already among the most frequent company-wide users of DT, both in general, and in sustainability management, while the food, wood, and textile industries make the least frequent use of such technologies (Fig. B2b). In fact, no company among those surveyed in the latter two industries uses DTs for sustainability management already company-wide. However, while the textile industry also ranks last in terms of ongoing company-wide implementation, the wood industry moves up in ranking to join the top three sectors (Fig. B2a). Thus, the wood industry ranks among the forerunners in the use of DTs but appears to have started the implementation processes later than the construction, machinery and metal, and electronics industries, which already exhibit more frequent use of company-wide implementations.

Areas of application per industry
This section illustrates the specific applications for which the different DTs are already in use per industry. The related questions were only answered by those companies reporting the use of the respective technology in general (i.e., shown Fig. 3). Overall, the results show that companies use IoT technologies most often for production and supply chain data collection (Fig. B1a), big data analytics and AI for demand forecasting, predictive maintenance, customer analysis and inventory management (Fig. B1b, c) and blockchain technology (Fig. B1d) to track products' origin or compliance tasks.

Internet of Things technology.
Regarding the use of IoT technology, companies were asked to report whether they already use it for collecting data from the following four product life cycle stages: supply chain, production, use, and end of life (EOL). As can be seen in Fig. 5, all industries, except for textile and food, report at least pilot projects in all four categories. The construction, electronics and machinery and metals industries do so most often. The construction and electronics industry even form two separate clusters, indicating both their difference to each other and to the other six industries. Overall, the use of IoT sensors for acquiring data in production is the most common application, as is also illustrated by the distinct cluster formed. This is followed by the cluster relating to use phase and supply chain data collection, and then by a cluster depicting EOL data acquisition, with the lowest level of implementation.
The use of IoT sensors for acquiring production data is also the most frequent (in four out of eight industries) company-wide application (Fig. B3b) and company-wide implementation is already ongoing in all industries (Fig. B3a). The electronics industry is most advanced regarding the use of IoT technology. It is the only industry that reported company-wide implementation of all four applications.

Big data analytics.
Regarding big data analytics, a total of eleven possible applications were studied. As shown in Fig. 6, almost all industries already report pilot projects for each of these applications. The only exceptions are environmental and social sustainability assessment and HR management in the food industry. Of the eleven applications, those of demand forecasting, customer analysis, predictive maintenance and financial controlling are the most frequently reported. They are relatively frequent (18-53 %) across all industries, except for food and textiles (6-13 %). The cluster analysis moves the focus away from the leading (electronics) and lagging (textile and food) industries and shows that the remaining five industry categories exhibit relatively similar patterns of DT implementation. A general distinction can be made between two application clusters. One cluster comprises the more frequent use of big data analytics with respect to internal processes and downstream-related tasks (i.e., customer analysis, financial controlling, predictive maintenance, and demand forecasting). The other cluster comprises the more seldom implemented applications relating to product design, supply chain, inventory, and HR management.
In general, however, not many industries have advanced beyond occasional pilot projects (Fig. B4a). While company-wide implementations are seldom, they were most frequently reported in the chemical, machinery and metals and the electronics industries, with 7-27 % of companies reporting the implementation of six to nine of the eleven big data applications surveyed (Fig. B4b). Of the other five industries, three reported no company-wide implementations, and two industries, others and wood, reported low implementation rates (7-17 %) of only two and three applications.

Artificial intelligence.
For AI, the same eleven potential applications were studied as for big data analysis. As shown in Fig. 7, the construction industry appears to be most advanced regarding the use of AI as 60 % of the companies surveyed report at least engaging in pilot projects for seven of the eleven potential areas of application, and 40 % using it for product design and supplier selection. Overall, predictive maintenance, inventory management, demand forecasting, and financial controlling are those areas in which applications have already been most frequently implemented. The distribution found here shows a certain resemblance to that found for big data analytics. For both technologies, predictive maintenance, demand forecasting and financial controlling rank within the top half of applications being implemented at least at the level of pilot projects, while transport planning, product design, supplier selection and HR management rank among the bottom half. Cluster analysis reveals that applications in the construction industry may also be considered as forming a cluster in their own right. Industry applications are consistently high in all areas except for HR management. However, applications in the latter field tend to be relatively low across all industries anyway. When differentiating the results in terms of the extent of implementation, the electronics, wood and construction industry (in particular), all stand out. Here, company-wide implementations are already underway in almost all of the eleven areas of AI applications (i.e. in 8-10 areas, Fig. B5a). The highest rate of already finished company-wide AI implementations can be found in the construction industry (20-40 % for five applications, Fig. B5b), and in the electronics industry (7-20 % for five applications). In other industries, only a few applications (1-4) have been already implemented company-wide. No company-wide implementation was reported in the textile sector, and only one in the wood industry.

Use of digital technologies per industry
4.2.2.4. Blockchain. Finally, Fig. 8 illustrates, for which of five potential applications, blockchain technology is already being used (at a minimum at pilot project level), and in which industry. As can be seen, blockchain technology has only been implemented up to a 20 % rate for two applications in the construction industry. This shows that the low implementation rates of blockchain technology found overall (Fig. 3) apply to all industries. In comparison, the highest implementation rates for the other three technologies range from 50 % to 100 %. In addition to construction, the wood and 'other' industries stand out from the rest, as also illustrated by the distinct clusters formed. Companies from wood and other industries report engaging in pilot projects for all five areas of application (8-18 %). From the remaining six, four industries report engaging in pilot projects in one to four areas of application, with the degree of implementation ranging from 3 % to 10 %. The food industry does not report any use of blockchain technology. Of the five areas of application surveyed, tracking the origin of products was that most frequently reported and it was also found to be the most widespread among the eight industries examined, with levels of implementation ranging from 7 % to 20 % across all industries (except food). The use of blockchain technology for compliance-related topics then occupies second place, followed by use in new business models, ecological footprints, and smart contracts. The cluster analysis reveals three clear clusters, as described above, and distinguishes the two most frequent blockchain applications (i.e., origin of products, and compliance) from the remaining three areas where applications are relatively less frequent.
Thus, differentiating the results regarding the extent of implementation reveals that blockchain technology is only being used companywide in two areas of application, i.e., tracking compliance and product origins, in one industry each (wood and electronics), and only by 8 % and 7 % of the companies in these industries (Fig. B6b). However, taking ongoing company-wide implementations into consideration shows that 8-18 % of the companies in the wood and 'other' industries are already scaling up the use of blockchain technology for five and three applications, respectively (Fig. B6a).

Differences between industries and company sizes
Investigations were also undertaken in order to ascertain whether statistically significant differences obtained concerning the degree of implementation of DTs in terms of industry and company size. As dependent variables, two constructs were used that were computed using the means of the responses to the four questions on the general degree of implementation of DTs (dt) and the degree of implementation of DTs for sustainability management (dsus) (i.e. the responses regarding the use of IoT technology, AI, blockchain technology, and big data analytics described in Table 2 and shown in Fig. 3). Both constructs exhibit Cronbach's alpha values of 0.7 for general implementation of DTs and 0.7 for DTs used for sustainability management activities.

Industry differences
A Kruskal-Wallis test shows a statistically significant difference between industries regarding the general degree of DT implementation (dt), (χ 2 (7) = 25.682, p < .001), estimated by using the mean construct described above in Section 4.3. Regarding the implementation degree for sustainability management (dsus), the results of the Kruskal-Wallis test are only significant at α = 0.1, (χ 2 (7) = 12.935, p = .073). The effect size for the general degree of DT implementation (dt) is determined to be medium (epsilon 2 = 0.202) (Cohen, 2013;King et al., 2018). Fig. 9 illustrates the data dispersion for both constructs. As can be seen, the categories construction, electronics, 'other', and machinery and metals exhibit the highest degrees of implementation, while those for the textile, food, and wood industries are the lowest. This also holds in general and for sustainability management. Post-hoc tests (Dunn-Bonferroni) show that only the two industries with the highest medians, electronics (Mdn = 2.5) and construction (Mdn = 2.25), and the two industries with the lowest medians, food (Mdn = 1) and textiles (Mdn = 1.5), differ significantly regarding the general  Areas of application for artificial intelligence (AI) per industry. The percentages summarize the following response options: implementation in a pilot project, company-wide implementation ongoing and company-wide implementation finished (i.e. the three blue bars in Fig. 3). The distribution is normalized to 1 by industry (i.e., by row) (n = 132). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) degree of DT implementation (dt) (adjusted ps < .05). No significant differences were found in pairwise comparisons of the degree of DT implementation in a sustainability management context (dsus).

Company size
The size of a company was determined based on turnover and the number of employees, and each company was classified as small, medium, or large (see Fig. 1). Kruskal-Wallis tests show significant differences in the general degree of DT implementation in terms of both company turnover (χ 2 (2) = 21.802, p < .001) and number of employees (χ 2 (2) = 18.211, p < .001). Fig. 10a and b illustrate the data dispersion for both proxies. As can be seen, in both cases, the median values of the general degree of DT implementation increase with company size. Large companies (Mdn turnover = 2.38, Mdn employees = 2.25) have the highest degree of DT implementation (dt), followed by medium (Mdn t = 1.75, Mdn e = 1.75) and small companies (Mdn t = 1.25, Mdn e = 1.25). While a similar trend can be observed in a sustainability management context, a significant difference can be found at α = 0.1 when differentiating companies by number of employees (χ 2 (2) = 5.08481, p = .080), but not by turnover (χ 2 (2) = 4.4497, p = .108). Effect sizes are found to be medium in both cases (epsilon 2 turnover = 0.166, epsilon 2 employees = 0.139). Post-hoc tests (Dunn-Bonferroni) show that, based on turnover, all differences in the general degree of DT implementation between pairs of groups are significant (adjusted ps < .05). In terms of number of employees all differences except that between medium and large companies (i.e., 50-250 and 251-10.000 employees) are significant. No significant pairwise differences between companies of different sizes were found in the sustainability management context.  Fig. 9. Degree of implementation per industry. The degree of implementation was measured using two mean constructs that summarized the response values to the two questions regarding the degree of implementation of the four technologies AI, big data analytics, blockchain technology and IoT technology in general, and in a sustainability management context (see Table 2 and Fig. 3a and b).

Correlation of digitalization strategy and technology implementation
Finally, an investigation was made in order to ascertain whether statistical correlations obtained between the two constructs used in Section 4.3, as well as between these and the company degree of implementation of digitalization strategies. The latter was measured with a construct based on aggregating the response values regarding the degree of implementation of the six specific digitalization strategies analyzed in Section 4.1.1 (Cronbach's alpha = 0.82). Furthermore, Spearman rank correlations were computed. As can be seen in Fig. 11, a) and b), the presence of a digitalization strategy correlates positively with both DT implementation in general (r = 0.53, p ≤ .001) and with DT implementation in the sustainability context (r = 0.46, p ≤ .001). As more companies are already using DTs in general than for sustainability management (see Fig. 3) an analysis was also undertaken to ascertain whether the general degree of DT implementation correlates positively with the use of DTs for sustainability management. Significance was found to exist, with a corresponding Spearman correlation coefficient of r = 0.64, (p < .001) (Fig. 11c). Finally, Fig. 11d) illustrates which of the DTs are used in conjunction with each other and how frequently this is the case. Overall, big data analytics and IoT are most often used together in general (31 times), followed by the use of IoT technology in general and for sustainability management (21 times), and the use of big data analytics in general and for sustainability management (18 times).

Discussion
In the following, the results are discussed and contextualized in relation to the current state of research (Section 5.1) and regarding the question of how to advance the use of DTs for sustainability and circularity management (Section 5.2).

Degree of implementation per technology
The analysis showed that of the four DTs investigated, IoT technologies are most frequently used, followed by big data analytics, blockchain technology and AI. This finding is in line with results from previous studies conducted in different contexts, such as a survey of 405 supply chain professionals by (Sodhi et al., 2022), or a survey of 200 Italian companies by De Marchi and Di Maria (2020). Also in line with our findings, Zheng et al. (2021) show that manufacturing companies are the best informed concerning industrial IoT and that they are least familiar with AI. The findings of our study also resemble a general European tendency found in the DESI 2022 (European Commission, 2022a) According to the DESI 2022, 48 % of large enterprises and 28 of small enterprises use any IoT, 14 % of enterprises in general conduct big data analysis, and 8 % use AI technology. Our study puts these findings into perspective regarding the DTs' degree of company-wide implementation, which is still low. Of the 24-47 % of companies using at least one of the DTs in our sample, the majority are currently still experimenting with them in pilot phases (17-30 %), and only a fraction of the companies surveyed have already implemented one of the four technologies company-wide (2-7 %).
The results of the DESI 2022's study on the integration of DT furthermore allow contextualizing Austrian companies' overall degree of digitalization. While Austria ranks 10th in the assessment, Finland, Denmark and Sweden are leading and Hungary, Bulgaria and Romania rank lowest (European Commission, 2022a).
The dominance of IoT implementation among the companies surveyed is probably explained by several factors. First, IoT technologies have acquired a higher level of maturity compared to other DTs (Zheng et al., 2021). Second, IoT is mainly seen as means for automated and continuous data collection and sharing (Ingemarsdotter et al., 2019;Liu et al., 2022). Hence, it can also be seen as a prerequisite for further DT implementations that require high quality data input, such as big data analytics or AI. Third, upcoming directives are increasing regulatory pressure concerning the adoption of IoT technologies, for example, those related to supply chain due diligence (European Commission, 2022b) or eco-design (European Commission, 2022c).
Additionally, the results showed that the implementation of a digitalization strategy in companies correlates positively and significantly with the implementation of the DTs (in general and for sustainability management). While a digitalization strategy is acknowledged as an important antecedent for the implementation of DTs, the benefit of such an implementation also depends on the comprehensive management of these digitalization efforts, especially in manufacturing companies (Björkdahl, 2020). It is likely that digitalization strategies for sustainability management and the availability of related IT resources will gain additional relevance due to legislative requirements, such as the planned introduction of digital product passports in the European Union (Berger et al., 2022;European Commission, 2020).
In our study, the share of companies that adopted DTs, in general, was higher than the share that adopted them for sustainability activities. De Marchi and Di Maria (2020) showed in their empirical investigation that for one out of four manufacturing companies sustainability concerns were a main driver behind investing in DTs. However, a larger share of the companies also reported sustainability benefits after the introduction of industry 4.0 technologies, as a sort of "by-product". In terms of our results this could mean that although DTs are not implicitly adopted for sustainability management, their implementation in general may still be beneficial for the company's sustainability management. While we were not able to determine a direct causal relationship between the implementation of DTs in general and their implementation in sustainability management, we did find a strong correlation between the two areas (Section 4.4), as well as qualitative indications for the co-use of specific technologies in both contexts (Fig. 11d).

Application areas of digital technologies
The descriptive analysis above (Section 4.2.2) provided details on the specific applications of the different DTs in sustainability management practice. Among the eleven potential applications of big data and AI, the area of predictive maintenance stands out as being the most widely implemented AI application, and the second-most frequent application for big data. This finding corresponds with a literature review by Rusch et al. (2022), in which predictive maintenance was also the most frequently reported DT application in the scientific literature. Activities such as predictive maintenance may be used to prolong the lifetime of a product and to generate new insights into product development (Ingemarsdotter et al., 2021). This is likely to be especially relevant in manufacturing companies attempting to advance sustainability and circularity management. As the design phase is an essential determinant in the environmental impact of products (Winkler, 2011), exploring how DTs can be used to further support this stage in the product life cycle is essential.
Regarding the use of IoT technology, De Marchi and Di Maria (2020) showed that companies mainly use IoT for improving process efficiency and reducing process emissions. This finding corresponds with our study, in which companies most frequently reported using IoT sensors for collecting data from their own production processes, followed by gaining data from the supply chain and the products' use and EOL phases.
Blockchain technology, the least frequently implemented DT in our sample, is most often used for tracking the origin of products and for compliance-related topics, and least frequently in new business models, collecting environmental data and in smart contracts. This hierarchy reflects findings from a literature review by Rusch et al. (2022). Here, the Fig. 11. a) Spearman rank correlation between the implementation degree of digitalization strategies and digital technologies in general. b) Spearman rank correlation between the implementation degree of digitalization strategies and digital technologies for sustainability management. c) Spearman rank correlation between the implementation degree of digital technologies in general and for sustainability management. d) Co-use network of the four digital technologies in general and in a sustainability management context (_sus). Node size = degree of co-use, edge width = frequency of co-use. authors split blockchain applications into 14 specific areas, with the majority being related to supply chain (e.g., tracking transport, energy, conditions, locations, materials) and compliance topics. In relation to new business models, the review authors also found a few examples describing the use of blockchain technology for incentivizing users (e.g., for recycling activities).

Differences between industries
Our study further showed that the implementation level of DTs varies considerably depending on a company's industry affiliation. While differences between industries were not investigated in detail in previous research, a study by Grischa et al. (2022) does offer a few indications on this score. For example, the authors found that companies from the German automotive industry expected greater improvements in environmental sustainability due to the implementation of industry 4.0 technologies than did companies from other industries.
In our sample, forerunner companies with company-wide implementation of several specific DT applications come from the electronics, machinery and metals, chemical and construction industries. This holds for the general company-wide implementation of DTs and in a sustainability management context. This finding both confirms and adds detail to previous research focused on single industries. For instance, Lv et al. (2018) review the literature regarding the use of data mining with big data analytics in the electronics industry and find that it is already being frequently utilized. This finding is confirmed by our study, which furthermore shows that the electronics industry is not only most advanced regarding the implementation of big data analytics (Fig. 4), but that it also exhibits the highest degree of company-wide implementation of IoT technology (Fig. B3). In a perspective paper focused on the chemical industry, Fantke et al. (2021) point out that chemical companies already frequently measure extensive process parameters in real-time for purposes of visualization and optimization, and use DTs for predictive maintenance less frequently. This observation is reflected in our study, as companies from the chemical industry use IoT technology most frequently for collecting production data (Fig. 5), thus ranking among the top four industries in this respect. Furthermore, our results show that the chemical industry's commitment to DTs is already quite diverse and goes well beyond that of applications in predictive maintenance (Section 4.2).
Overall, the construction industry exhibited the highest degree of digitalization ( Fig. 9) One possible explanation for this leading role could be the fact that the use of building information models (BIM) is already well established in the construction industry (Tang et al., 2019;Volk et al., 2014). This could have led to a generally more advanced use of DTs, because learning from previous digitalization projects was used to implement new projects. However, it needs to be noted that our sample only included five construction companies (comprising two small, one medium, and two large companies, as measured by turnover).
Surprisingly, the companies from the food industry in our sample did not report much in the way of using DTs in practice, although there are several examples of how blockchain can be used to enable more transparent food supply chains (Rusch et al., 2022). For example, blockchain can be used to track the origin of food products to the end point of sale . However, it is possible that our limited sample, consisting of 12 small, four medium, and only one large company, was responsible for the relative lack of information.

Differences between company sizes
The review of the relevant literature showed that there are examples where DTs are adopted in practice and where, among other things, they facilitate sustainability and CE activities. Most examples in the literature, however, come from large international corporations (Rusch et al., 2022). Our study also found that large companies are significantly more likely to adopt DTs than small companies in general, and, tendentially also in a sustainability management context. This confirms and expands on EU-wide findings from the DESI 2022 (European Commission, 2022a), as well as findings from authors such as (Zheng et al., 2021), who found that the company size (measured by turnover) has an impact on the company degree of technology implementation.
Additionally, Zheng et al. (2020) also showed that knowledge of industry 4.0 technology increases significantly with company size and that, apart from a few exceptions, large companies generally have more resources available to implement DTs than smaller ones. A different analytical lens was used by Grischa et al. (2022) who showed that larger manufacturing companies in Germany and China have higher expectations concerning DT application than smaller companies, for example, with respect to reducing resource use or increasing material efficiency. The electronics industry in Germany, in particular, is very optimistic that industry 4.0 will lead to lower material consumption. However, in general, expectations are very heterogeneous and vary strongly across industry and country.
Finally, the use of DTs in enabling a CE and in facilitating sustainability management activities is not only related to size. It is also company specific and depends on value proposition, value chain position, and other (internal) capabilities. Therefore, companies need to ask themselves how data can be used to create value in their company, and then select the technologies that will make this possible (Klingenberg et al., 2021).

Advancing the use of digital technologies for a sustainable and circular economy
Given the low degree of implementation of DTs for sustainability practices, it is important to develop business cases to show/flagship what DTs can be used for and how they can enable better sustainability and circularity management. Klingenberg et al. (2021) state that "[…] as technical barriers disappear, topics involving value-creating technologies, such as business models and value chain transformation, become essential." (p. 585). For example, new business models could make use of DTs to improve the efficiency of PSS or pay-per-use systems (Ranta et al., 2021). While the relatively high implementation degree of predictive maintenance practices using big data and AI found in the present study hints at an increasing prevalence of PSS, the use of DTs for new business models is still quite seldom. This is reflected in the present study regarding the use of blockchain technology for new business models (Fig. 8), and in the literature. For example, Rusch et al. (2022) find that out of 146 applications of DTs, only a few focus on the business model, and the majority focus on eco-efficiency improvements. In line with this, Neligan et al. (2022) find in a representative survey among German manufacturing companies that classical business models still dominate (in 2/3 of the surveyed companies) while only around a fifth of companies expanded their business models to make them more computerized or data-driven. The latter-type business models were furthermore found to be precursors for the implementation of PSS for resource efficiency and they may also be considered as having "circular disruption" potential (Neligan et al., 2022).
Looking into the future, it can be argued that AI technology, while hardly used yet, still holds considerable and wide potential for fostering corporate CE and sustainability management and CE that needs to be explored. For instance, in addition to the potential applications stated in Table 2, machine learning techniques can help enhance the data confidentiality/privacy of its users and thereby help to increase supply chain actors' willingness to share data, such as via a digital product passport (Berger et al., 2022). Besides AI, also blockchain technology can be considered underdeveloped and its predicted potential for fostering the exchange of sustainability and CE-related data along value chains (Böckel et al., 2021;Kouhizadeh et al., 2019) has yet to be explored on a larger scale in practice.
Finally, while an increasing digitalization of companies might, indeed, be an important factor in facilitating a shift to more sustainable production and consumption patterns, it needs to be noted that the use of DTs may also lead to negative environmental impacts and induce rebound effects and/or trade-offs Itten et al., 2020). However, this goes beyond the scope of this study but other articles focus on the quantitative sustainability impacts of DTs, e.g., (Berkhout and Hertin, 2004;Ingemarsdotter et al., 2021;Itten et al., 2020).

Limitations
The main limits of the present study are its focus on Austrian manufacturing companies and its relatively small sample size (132 companies). While other studies from comparable European countries such as Italy or Germany indicate similar trends to those described above, generalization of the findings beyond the European context would require further empirical investigation. This particularly applies to the comparison between industries, which would mainly benefit from larger sample size. The fact that the study is rather exploratory in character, and given the depth of analysis employed (e.g., regarding specific applications of DTs), may nonetheless serve to validate the approach taken and help distinguish it from other studies with larger samples.

Conclusions
This study provided empirical insight into (i) manufacturing companies' general and sustainability/circularity-related degree of digitalization, (ii) the degree of implementation of four specific digital technologies (i.e., AI, big data, blockchain, IoT), and related applications per industry and (iii) digitalization strategies pursued.
The results show that manufacturing companies only seldomly use DTs for sustainability management to date, even though more than half of the companies reported possessing appropriate IT and other digital systems to support a CE business model, products, or services. Implementation of digitalization strategies and the implementation of DTs vary significantly depending on company size, with the larger companies exhibiting a higher implementation rate. Among the eight industries surveyed, the construction, electronics, and the metal and machinery industries were found to be the most advanced, while the textile and food industry exhibited the lowest degree of digitalization. Despite the general trend towards greater digitalization, the degree of implementation with respect to specific applications of the four technologies varied considerably across industries. Predictive maintenance and inventory management (AI), demand forecasting and customer analytics (big data analytics), manufacturing data and supply chain data acquisition (IoT), and product provenance and compliance tracking (blockchain) were found to be the most frequently implemented applications per technology out of a total of 31 applications studied.
Overall, only a fraction of the companies surveyed has already implemented one of the DTs company wide. Most companies that reported using a DT are doing so in pilot projects. Thus, the company-wide implementation of DTs still seems to be more of an ad hoc phenomenon than an established business strategy. However, the detailed nature of the data collection process showed that both the availability of appropriate IT and data systems for a CE as well as general digitalization strategies are further developed compared to the implementation of DTs. Hence, an operationalization of those strategies could happen relatively quickly and in the near future, as soon as economic benefits become more apparent due to changes in, e.g., the regulatory environment.
The theoretical and managerial implications of the present study are as follows: (1) This study confirms that DTs can support sustainable and circular management practices, as the surveyed companies report several related applications. However, their degree of implementation is still low, and only in relatively rare cases do companies have clearly defined sustainability-specific digitalization strategies. Given the large variety of potential applications and the related uncertainties, as reflected in the literature, companies still find defining the specific nature of their DT needs somewhat daunting. Thus, the identification of scalable best practices that can guide strategy formulation and implementation is likely to be an important prerequisite in accelerating any transition towards a digital CE.
(2) The degree of digitalization differs significantly across different industries and company sizes, with SMEs exhibiting the lowest degrees of digitalization. In light of actual and perceived resource constraints, the identification of affordable quick gains is likely to be another important enabler for furthering the implementation of DTs, especially among SMEs, both in general and in a sustainability and CE context. (3) As the general implementation degree of DTs is considerably higher than in a sustainability and CE context, the identification of synergetic applications and potential spillovers is likely to help companies move from general DT application to more specific use in targeting greater sustainability and circularity. The identification of synergies and spillovers could be fostered by increasing collaboration between IT and sustainability departments and/or by sustainability staff acquiring more IT-specific skills. (4) Finally, companies need to prevent potential lock-ins and economic and environmental rebound effects in their digitalization efforts. This entails more explicit recognition of the many purposes for which DTs may be applied. As stated above, such recognition (of potential synergies and spillovers) may come about through improving those competencies needed in optimizing DT applications for multiple operational and transformational goals.
Further research needs to be directed at gaining a deeper understanding of how small and medium sized companies in particular, can utilize DTs in their CE transition most efficiently and effectively. Future research could also be directed at the development and validation of further and more specific concepts for using DTs in the different corporate sustainability and circularity management practices. It would also be worthwhile to further detail the drivers of, and barriers to DT implementation in sustainability management and to investigate the economic and environmental outcomes of utilizing DTs in different contexts.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Areas of application of digital technologies
Response options: 1. not considered, 2. considering it, 3. implementation in pilot or stand-alone project, 4. company-wide implementation ongoing, 5. company-wide implementation finished.

# R-code Question
To what extent has your company implemented big data analytics in the following areas:

# R-code Question
The extent to which your organization has implemented internet of things technology in each area: 1 iot To obtain data from the supply chain 2 iot To obtain data in the production phase 3 iot To obtain data from the use phase of the product/service 4 iot To obtain data from the end-of-life phase of the product 5 iot_open Other: ....................

Implementation ongoing and finished
Production data acquisition SC data acquisition  Other (

Implementation ongoing and finished
Origin of products Smart contracts New business model