Mapping the global flows of steel scraps: an alloy elements recovery perspective

Recycling steel scraps by the use of electric arc furnace is one of the most promising approaches for the steel industry to achieve net-zero emissions. Due to the uneven distribution of global steel scraps, many countries are actively involving in the global steel scraps trade. Steel scraps contain a range of critical elements, which may be transferred across borders through international trade of steel scraps. However, existing studies have paid little attention to the global flows of steel scraps and its embodied alloy elements (AEs). This study maps the journey of global steel scraps and the embodied AEs for the period of 2000–2021 for the first time by employing trade-linked material flow analysis and social network analysis. The results indicate that the global steel scraps trade had increased during the study period, with a few core countries (such as USA, Germany, and Turkey) leading the global steel scraps network. Also, critical metals had been transferred across borders in the form of AEs through the trade of steel scraps, especially from global north countries to global south countries. The largest AE flows include Chromium (Cr), nickel (Ni), manganese (Mn) and molybdenum (Mo) flows. Other AE flows, such as cobalt (Co), vanadium (V), and niobium (Nb) flows, were less, but with high values or being regarded scarce. From a global perspective, steel scraps trade and recycling can contribute to the decarbonization efforts of the global steel industry and address resource shortages in some countries. Therefore, it is urgent to promote the overall resource efficiency of steel scraps and the embodied AEs by various efforts.


Introduction
The Paris Climate Agreement established an ambitious climate target of achieving net-zero by 2050 (van Vuuren et al 2018, van Soest et al 2021).As the largest industrial emission source (Li and Hanaoka 2022), steel and iron industry alone accounted for 25% of the global direct industrial carbon emissions (Ren et al 2021), implying an urgent need to decarbonize this sector (Li et al 2019).As the most promising material to replace virgin iron ore production, steel scraps can be recycled repeatedly without losing its properties so that energy consumption and carbon emissions can be reduced (International Energy Agency 2020).Recycling steel scraps in electric arc furnaces requires an energy input of only 8.7%-25.0% of that in the primary steel production route (Harvey 2021).As such, the carbon footprint per ton of recycled steel is only 10%-20% of that of the primary steel (Fan and Friedmann 2021).However, only 40.3% of the global steel production comes from steel scraps (Wang et al 2021a).The availability of steel scraps is substantially limited by the historical production and the life spans of various steel-containing products (Pauliuk et al 2013b), especially in countries with low industrialization levels (Pauliuk et al 2013a).Several countries rely heavily on importing steel scraps to meet their domestic demand (Nes ¸er et al 2008).But these countries are facing a risk of supply chain disruptions, which may be induced by geopolitical conflicts, price volatility, and a lack of collection system (Liu and Müller 2013).
Practically, steel is not always used in its pure form, but is frequently alloyed with other elements in order to enhance its performance (Ohno et al 2015, Tan et al 2021), such as increased strength, high corrosion resistance, light weight, longer life spans, and better recyclability.For instance, zinc (Zn), aluminum (Al), and tin (Sn) are added as coating elements to enhance the corrosion resistance of steel (Panasiuk et al 2022).Other alloy elements (AEs) include manganese (Mn), chromium (Cr), nickel (Ni), molybdenum (Mo), vanadium(V), etc (Nakajima et al 2013).Functional recovery of steel scraps only occurs when AEs end up in the right place (Graedel et al 2011).However, it is difficult to remove such AEs from steel scraps, which may result in the functional loss of AEs or steel contamination (Nakamura et al 2012, Diener andTillman 2015).A typical example is that high-quality steel sheets from vehicles are downgraded into ordinary steel for construction purposes (Ohno et al 2014, Hertwich et al 2019).Although several scholars have investigated the criticality of embodied AE flows, their efforts mainly focus on several AEs with higher concentrations in the steel scraps, such as Ni, Cr, Mn, and Mo (Daigo et al 2010, Nakajima et al 2013, Ohno et al 2016, 2017, Nakamura et al 2017).A comprehensive evaluation covering all the critical elements embodied in steel scraps is still lacking, especially at the global level.Most of these AEs are critical metals for achieving the net-zero future.For example, Co and Ni play important roles in the transformation of energy systems.By 2040, the global demand for these metals may be 19 and 22 times higher than that in 2020 (International Energy Agency 2021).Some of them have higher economic value, such as V and Nb (Tan et al 2021), which deserve more attention.
Material flow analysis (MFA) is a mature method to systematically measure the flows and stocks of one material within predefined spatial and temporal boundaries (Brunner and Rechberger 2016).To date, MFA has widely been used to track the metabolic evolution of critical metals during their entire life cycles (Graedel et al 2004, Wang et al 2007, Cullen et al 2012, Glöser et al 2013, Su et al 2023).Since metal elements may cross national boundaries multiple times during production, fabrication and manufacturing stages (Liu and Müller 2013), several studies investigate the material flows embodied in international trade based on a trade-linked MFA framework (Sun et al 2017, 2019, Wang et al 2022, Gao et al 2022a).However, these studies typically investigate one single metal element from a life cycle perspective and rarely focus on all the AEs contained in a specific steel product.
Social network analysis refers to the process of investigating trade networks and relations between countries based on network science and the graph theory (Wasserman and Faust 1994).It is an effective method to identify patterns of relations among components within a system (An et al 2014), and can objectively assess a country's market dominance through metrics such as trade volume, number of trade relations, and trade intensity (Liu et al 2022, Zhou et al 2022).This method has been widely used to study the characteristics of international trade networks for various commodities, such as metal resources (Ge et al 2016, Tokito et  In summary, in order to address the above concerns, this study investigates the global flows of steel scraps and its embodied AEs by integrating tradelinked MFA with social network analysis so that the potential risks in the global steel scraps trade can be identified.We expect to recognize the dilemma of embodied AEs in global steel scraps recycling so that valuable insights can be obtained for preparing appropriate resource management policies.

Methods and data
We measure the global trade flows of steel scraps and its embodied AEs between different countries by applying a trade-linked MFA and then investigate the trade patterns of steel scraps and individual AEs by conducting social network analysis.The global steel scraps trade network (GSSTN) includes more than 200 countries and regions, and the temporal boundary of this study is from 2000 to 2021.The investigated AEs include manganese (Mn), nickel (Ni), zinc (Zn), stannum (Sn), plumbum (Pb), copper (Cu), aluminum (Al), chromium (Cr), cobalt (Co), vanadium (V), niobium (Nb), and molybdenum (Mo).All the detailed explanations of data processing are presented in the Supporting Information.

Trade-linked MFA
The global anthropogenic iron & steel cycle usually includes primary production, fabrication and manufacturing, final use, and waste management stages (Zhong et al 2018).In this study, we focus on the waste management stage.
We trace all AE flows in AE metallic equivalents by compiling extensive data from various sources, including reported statistics, published literature and expert interviews.Bilateral trade data of steel scraps were obtained from the UN Comtrade database (UN Comtrade 2023).Table S1 lists different steel scraps products covered in this study, along with their

Degree
In-degree The in-degree of a node refers to the number of countries importing steel scraps products from a specified country.

Out-degree
The out-degree of a node refers to the number of countries exporting steel scraps products to a specified country.

Strength
In-strength In-strength refers to the imported trade volume of the edges connected to a node.

Out-strength
Out-strength refers to the exported trade volume of the edges connected to a node.

Betweenness centrality
The betweenness centrality of one node refers to the number of these shortest paths that pass through this node.

Modularity index
The modularity index refers to the density of links inside communities.
harmonized system (HS) codes, including seven types of steel scraps products.We chose net weights (physical quantity) of steel scraps and AE contents (%) to explore the associated AE flows.The contents of AEs are listed in table S2 of the supporting information.The amount of AE x embodied in the steel scraps trade is calculated by using equation (1), where M x represents the total mass of the AE x in the steel scraps; p represents each traded steel scraps commodity and is presented by one HS code; n represents the number of involved steel scraps types, with a maximum number of 7; A p represents the mass of each steel scraps commodity under p HS code, R px represents the content of each AE embodied in each involved steel scraps commodity under p HS code.

Social network analysis
Social network analysis is used to analyze the trade network structure of global steel scraps and its embodied AEs.Gephi software is used to visualize this network so that bilateral trade relations for steel scraps can be presented, in which each node represents a country and each edge represents a bilateral relationship between two involved countries.This network model is composed of a node-set (P) and an edge-set (Q), where P= {p i : i = 1,2,…,n} (Gao et al 2022a), n is the number of nodes, and Q= {q i : i = 1,2,…,m}, m is the number of edges.The adjacency matrix of this network is expressed by equation ( 2), where ω ij represents the weight of the edge from node m to node n.
Several indicators are used to analyze the features of the GSSTN and the embodied AEs networks, including in-degree and out-degree, in-strength and  2023).Therefore, we assume the uncertainty ranges of AEs contents directly collected from databases and literature are at a low level (2%), while the uncertainty ranges derived from estimated parameters are assumed to be high (10%).The calculated results are illustrated in figures S54 and S55 of the Supporting Information.

Features of the GSSTN
By considering the top 80% of the global steel scraps trade volume during 2000-2021, we identified a total of 44 key nodes within GSSTN (figures S1, S2 and table S3). Figure 1 shows the trade patterns of key nodes within GSSTN from 2000 to 2021, in which the top three exporters were the United Kingdom (153.89Mt), Japan (143.31 Mt), the Russian Federation (131.46Mt), while the top three importers were Turkey (344.75Mt), Republic of Korea (139.26 Mt),Spain (113.73 Mt).During the study period, several countries emerged as key players in the global steel scraps trade, including the USA, Canada, and Denmark, which are key nodes in the export trade network, while Pakistan and Viet Nam emerged as key nodes in the import trade network.In contrast, The global steel scraps flows are visualized in figure 2, indicating that the total trade volume has significantly increased with more intensive trade relations.New participants in GSSTN were mostly those low and middle-income countries in Africa, South America, and Asia.However, being influenced by COVID-19, the number of participating countries in GSSTN decreased significantly after 2020 (figure S50).Also, different regions have different trade patterns.The main export destinations of those European countries include the Mediterranean and Arabian countries, while the main export destinations of Asian countries include those East and Southeast Asian countries.Moreover, the destinations for those American countries are relatively diversified, with South American and Asian countries as the main destinations.
The number of communities in GSSTN fluctuated considerably between 2000 and 2021, ranging from 5 to 10. Overall, the modularity of GSSTN has remained high, with values between 0.32 and 0.52 (figure S51), indicating a good quality of community division.However, a fluctuating decline in the modularity index can be observed during this period although it has become stabilized since 2018.This phenomenon shows that the steel scraps trade has become increasingly globalized in recent years, and the division of each community has become less significant.Also, GSSTN remained a hierarchical structure during the study period, which has been dominated by a few core countries.The community structure of GSSTN experienced a dynamic evolution trend (figure S52).Specifically, the main communities in this GSSTN have shifted from the European community to the American-Asian communities after 2007, in which Southeast Asian countries played a more important role.The community division of this GSSTN was geographically clustered in its early stage, indicating that countries with shorter distances had closer trade relations.For example, the community with the most trade relations was mainly composed of several European countries, such as Germany, the United Kingdom, and the Netherlands.
Originally, these countries had more intensive steel scraps trade relations.But later on, these countries began to have more transactions with non-European countries.In addition, the globalization further made several African countries, such as Mozambique and Zambia, involved in the core communities.Several traditional export countries, such as the United Kingdom, Germany, and the USA, have a high level of connectivity (high betweenness centrality) in this GSSTN, meaning that these countries have stronger resource control capabilities and are critical hub countries in GSSTN.Similarly, major steel scraps import countries, such as India and Turkey, gradually increased their connectivity in GSSTN, demonstrating that they had diversified trade partners for steel scraps.

Global flow patterns of AEs embodied in steel scraps trade
With the increasing steel scraps trade, the total amount of each AE embodied in global steel scraps trade increased from 1076.12 kt in 2000 to 2413.25 kt in 2021 (figure 3(a)).In particular, a surge of embodied AEs occurred in 2004.Typical AEs, such as V, Nb, and Ni, had grown faster than other AEs, increasing by more than 1.5 times.Sn had the slowest growth rate of 60.67%.The most embodied AEs include Cr (21 304.79 kt), Ni (8997.67 kt), and Mn (6759.82kt) due to their universal and vital functions in various types of steel products.These elements together accounted for more than 80% of all the AEs (figure 3(b)).Other key embodied AEs include Cu, Mo, Zn, Sn, and Pb, with a total amount ranging from 300 kt to 2500 kt.The amount of each of other remaining AEs is below 210 kt.However, although the contents of these AEs in the steel products are relatively less, their market values are high due to their scarcity and irreplaceability, such as Nb, Co, and V.We compared the structure of embodied AEs within GSSTN (figure 4 and supporting information figure S53).International steel scraps flows may not necessarily result in significant losses/acquisitions of embodied AEs.For example, although Ukraine was the third largest steel scraps exporter in 2000, its total  AEs only ranked the 12th in the world.Similarly, Egypt was the 10th largest steel scraps importer in 2014, but only ranked the 22nd in terms of its embodied AEs.The trade of steel scraps resulted in a considerable amount of embodied AEs (mainly Cr, Ni, Mn, and Mo) transferring from developed economies to developing countries (figures 3(c) and (d)).
In addition, regional disparities exist in terms of the trade structure of AEs.Countries mainly involving in stainless steel scraps trade, such as Germany, Belgium, and India, had much higher Cr, Ni, and Mo contents embodied in their traded steel scraps.In contrast, countries mainly involved in cast iron scraps and steel ingots for remelting, such as Malaysia, Belarus, and Lebanon, contributed much less to the international flows of AEs.
The global flows of different AEs embodied in steel scraps trade are illustrated in figure 5. India, Netherlands, Germany, and the United Kingdom have higher betweenness centrality values for all the AEs, indicating that these countries played a vital role in the flows of AEs embodied in global steel scraps trade.The value of modularity indicates the extent to which communities are divided in different embodied AE trade patterns.Except for Sn (0.15) and V (0.16), other AEs have clear community divisions, with modularity values ranging from 0.46 to 0.58, indicating distinct trade centers within their respective network communities.In the embodied V, Ni, and Nb trade networks, although the flows from Canada to the USA, from Mexico to the USA, and from the USA to China were always the top three trade flows, several characteristics are observed.For example, India and the United Kingdom dominated the first community (the community with the most trade relations) in both embodied Ni and Nb networks.In contrast, American and Asian countries dominated the first community in the embodied V network, such as the Republic of Korea and Canada.Similarly, India, Spain, and the United Kingdom dominated the first community in the embodied Mo network.And, several European and Asian countries dominated the first community in the embodied Sn trade network.Finally, Spain, India, United Kingdom, and Thailand dominated the first community in the embodied Cr trade network, while Germany, Finland, and Italy dominated the second community (the community with the second largest trade relations).

Environmental challenges and potential risks in the global steel scraps trade
The global distribution of steel scraps supply is currently uneven, with highly industrialized countries contributing significantly to the global market.These countries have more in-use steel stocks and can collect more steel scraps from their end-of-life products (Pauliuk et al 2013a).Also, they sell their steel scraps due to their stricter environmental regulations.Although secondary steel production is less energy and emission intensive, these developed countries still prefer to purchase steel products from other countries so that they can avoid such energy consumption and corresponding emissions (De Sa and Korinek 2021).In contrast, developing economies, such as Turkey, China, and India, are still experiencing rapid industrialization and urbanization and have higher demands for steel products.Compared with developed economies, these countries typically have lax environmental regulations and ineffective enforcement.Therefore, international steel scraps trade may further result in cross-border transfer of associated environmental burdens (Schütz et al 2004, Wang et al 2021b).For example, India imported a large amount of steel scraps during 2000-2021 (figures 1 and 2).However, the waste recycling industry in India is poorly managed (Schoot Uiterkamp et al 2011, Rathore 2020).The recycled steel from steel scraps by its formal sector accounted for only about 5% of the total recycled steel (Awasthi and Li 2017).This country is currently the second largest steel producer in the world (Dakua 2019), in which recycled steel accounted for more than 55% of the total steel production (The World Steel Association 2020).It is expected that India will account for nearly a fifth of the global steel production by 2050 (International Energy Agency 2020), which means that ineffective governance on those unauthorized recycling firms will undermine the global net zero efforts.
During the early stage of the study period, more trade activities occurred within the same continent due to their geographical proximity, which can reduce the corresponding transportation costs.Another reason is regional trade restriction policies.For example, the Basel Convention restricted EU countries to export their non-hazardous and recyclable steel scraps to non-member countries (Wang et al 2020a).However, since these countries locate in the same continent and normally have similar economic development levels and environmental governance, it is difficult to have more internal steel scraps trade.Then these countries began to seek new trade partners in other continents, especially those developing countries with higher demands for steel products.Additionally, since several core countries dominated this GSSTN, their policies on steel scraps trade can greatly influence the global steel scraps trade structure.For instance, with the implementation of Solid Waste Import Ban policy, China gradually reduced its core role in the global steel scraps trade (figure 2).Consequently, as China's original trade partner, Japan began to sell steel scraps to other Asian countries, such as Viet Nam and Bangladesh.
Moreover, our social network analysis results reflect that only a few countries have the most trade relations within GSSTN.Most countries tend to keep their original trade partners.However, due to the clear emission reduction feature of recycled steel, now more countries have determined to support steel recycling within their territories and begun to restrict the export of their steel scraps, such as Argentina (Lee 2023), the United Arab Emirates (Gerber Group 2022), and Kenya (The Scrap Metal Council 2015).The EU has also restricted the export of such steel scraps to non-OECD countries (European Commission 2021).Similarly, several countries have restricted the export of their steel scraps by charging higher export taxes (Price and Nance 2009).For instance, the export tariff rate on steel scraps is 40% in China (Recycling Today 2017), while the export tariff on steel scraps is €290/ton in Russia if the exported amount exceeds the official quota (International Information Group 2022).These policy changes may disrupt current steel scraps supply chain.In particular, trade restrictions on steel scraps by those more developed countries may lower production efficiency in emerging economics, especially advanced developing countries, since it is difficult for them to access cheap steel scraps (Yamaguchi 2018).For example, several countries (such as Turkey) heavily rely on importing steel scraps to support their steel industry (Özdemir et al 2018).These countries may suffer from getting steel scraps with unreasonable prices (Wübbeke and Heroth 2014).In addition, such export restriction policy may increase the domestic steel scraps supply in developed economies, which will lower the local steel scraps prices and reduce the incentives of local recyclers to collect more steel scraps (Ohno et al 2015).

The dilemma of embodied AEs in the global steel scraps recycling
The amount of AEs embodied in the global steel scraps trade has increased rapidly, with the total volume in 2021 exceeding 1.5 times that of 2000.It indicates that more countries involved in the global steel scraps trade.The recovery rate for some special steel scraps, such as stainless-steel scraps, is normally high.A previous study shows that the recovery rates of Cr and Ni from stainless-steel scraps are nearly 95% (Team Stainless 2023).However, basic recovery technologies, such as shredding, crushing, magnetic W Cai et al sorting, are typically applied by considering economy of scale (Reck and Graedel 2012).The recovery rates of embodied AEs are on average still low (Björkman and Samuelsson 2014).In practice, downcycling is still widely applied.It is difficult to achieve full recycling of AEs due to the lack of recycling considerations during the design of complex products and difficult screening of various steel scraps (Ohno et al 2014).Additionally, most steel scraps are classified and traded based on their physical shapes (such as triangular, rectangle, round, or even irregular shapes) in the global market, indicating that the true values of such AEs cannot be fully reflected in the steel scraps markets (Ohno et al 2015).
Our results show that the global steel scraps trade has led a large amount of AEs to transfer from global north countries to global south countries, especially for Co, V, Ni and Nb.These elements are now classified as critical minerals by many developed economies (MNRC 2016, European Comission 2017, USGS 2017, DISER 2022).These countries have begun to restrict their export activities of such elements.Unfortunately, the cross-border flows of AEs embodied in steel scraps have not received adequate attention.The absence of standards for recovering AEs is one reason for this phenomenon.For example, Japan has implemented a comprehensive classification standard for steel scraps in order to differentiate steel scraps based on their AEs contents (Ohno et al 2014).However, this standard only covers the contents of three AEs (Cr, Ni, and, Mo) and two tramp elements (elements that may deteriorate steel properties during steel recycling, such as Cu and Sn), but does not classify high-valued AEs such as V, Nb, and Co.Also, we find that trade relations between global north and south countries are generally closer in embodied AEs trade networks.For instance, countries such as India, Spain, the United Kingdom, and South Africa belong to the same community in embodied Cr, Ni, Mo, Nb, Co, Pb, Al, Cu, and Zn networks.These countries carry out intensive international trade in alloyed steel scraps.However, for global south countries, even if they obtain alloyed steel scraps through international trade, they may not have the ability to extract and recycle AEs embodied in these steel scraps.Thus, it is urgent to promote technological cooperation between global north and global south countries.
In addition, copper and tin are tramp elements in steel scraps (Nakamura et al 2012).The recycling of steel scraps may increase the concentration of such tramp elements in the recycled steel and eventually lead to thermal embrittlement, cracks and fractures of secondary steel products (Daehn et al 2017, Yes ¸iltepe and S ¸es ¸en 2020).For instance, Turkey was the world's largest importer of embodied Cu and embodied Sn.Their recycling firms have to pay extra attention on their purchased steel scraps to avoid or at least reduce the accumulation of Cu and Sn.

Policy implications
Based on the main findings of this study, we propose two policy recommendations, including the optimization of steel scraps trade structure and technological cooperation.
Our findings suggest that it makes sense for all the involved countries to develop their steel scraps trade policies to avoid potential risks.Those highly industrialized economies, such as the USA, the EU and Japan, have significant influences on the global steel scraps supply chain.Therefore, key steel scraps import countries, such as Turkey and Republic of Korea, should seek to diversify their import partners.Another consideration for these import countries is how to avoid steel scraps price volatility or supply disruptions.Stockpiling is one solution, but its effective implementation relies on careful national plan and international cooperation.Key secondary steel production countries (such as Turkey and India) should have more trade cooperation with high-media countries (such as the United Kingdom, Germany, Sweden, and the USA) so that they can better express their trade opinions in the global steel scraps market.In addition, some countries imported cleaner steel scraps to produce high value-added steel products, but exported low grade steel scraps (Wang et al 2020a).For example, the copper contamination in the exported steel scraps from the USA is 14% higher than that in its domestic recycling source (Cooper et al 2020).To date, more countries start to restrict the import of low-grade steel scraps, which means that steel scraps export countries should further improve the quality of their steel scraps through pre-treatments.
Additionally, we recommend that all countries involved pay closer attention to those AEs embodied in steel scraps.Although the development of advanced recycling technologies for embodied AEs that are economically feasible still faces a number of technical difficulties, it is promising to seek potential solutions through joint research and development efforts on laser, near-infrared, or x-ray sorting technologies by all trade countries (Reck and Graedel 2012).The establishment of a formalized international research body, including key stakeholders such as scientists, engineers, entrepreneurs and policymakers, is encouraged so that they can work together to seek more innovative technologies on AE recovery.In order to make it effective, it is necessary to host regular meetings to help different stakeholders identify key technological problems, discuss how to collect adequate research funds, allocate research missions, and prepare a road map.For instance, laser-induced breakdown spectroscopy is an emerging technology for sorting steel scraps (Björkman and Samuelsson 2014).Similarly, x-ray fluorescence sorter can help further improve the sorting efficiency of steel scraps (Brooks et al 2019).Both technologies should be further promoted so that they can be better applied.Also, policymakers from different countries may work together to release an international standard on steel scraps trade so that the accumulation of copper and tin in steel scraps can be mitigated.As such, industrial standards on the recovery of AEs can help avoid high valued metals from downgrade recycling.In addition, policymakers in relevant countries should prepare economic incentives to protect critical AEs, such as resource tax, financial subsidies and appropriate pricing.Finally, global north countries should actively help those involved global south countries through technology transfer, technical secondment, capacity building activities and financial support since most global south countries do not have adequate financial, technological and institutional abilities to effectively recover AEs embodied in steel scraps.

Conclusions
Steel scraps is a promising source to replace virgin steel and can greatly contribute to the global net zero efforts.The embodied AEs have been transferred among different countries through the global steel scraps trade, leading to the leakages and losses of such critical metals.This study maps the global flows of AEs embodied in steel scraps for the first time, and uncovers the key evolution features of international steel scraps trade network and its embodied AEs from 2000 to 2021.We found that both the total trade amount and trade relations significantly increased, with a few countries (such as the USA, Germany, and Turkey) dominating the global steel scraps network.Also, we found that a few countries (such as the United Kingdom, Germany, Sweden, and the USA) were important hub countries and played a key role to connect this steel scraps trade network.In addition, the flows of AEs do not follow the patterns of international steel scraps trade since the amounts of AEs embodied in steel scraps are actually much lower in several countries with significant steel scraps trade volume.Also, a large amount of AEs was transferred from countries in the global north (such as the USA, Germany and the United Kingdom) to countries in the global south (such as China, India, and Turkey), especially the following elements, including chromium (Cr), nickel (Ni), manganese (Mn) and molybdenum (Mo).From a global perspective, trading and recycling steel scraps may contribute to decarbonization efforts in the global steel industry and address raw material shortages in several countries.This is of great significance for supporting the collaborative implementation of SDG 12 (Responsible Consumption and Production) and SDG 13 (Climate Action).However, current steel scraps prices may not fully reflect the true values of such embodied critical metals, leading to the losses of these valuable metals.Consequently, it is necessary to prepare appropriate trade policies and circular economy strategies to promote the overall resource efficiency.Finally, global north countries should pay more attention on the corresponding environmental impacts and technical challenges associated with steel scraps trade, especially in those global south countries.They should help those involved global south countries through technology transfer, technical secondment, capacity building activities and financial support.
al 2016, Zhong et al 2018, Wang et al 2019, Shao et al 2021, Zheng et al 2022, Gao et al 2022a), energy products (Du et al 2017, Chen et al 2018, Wang et al 2019) and solid waste (Wang et al 2020b, Ma et al 2021).However, although several scholars highlighted the importance of studying the trade patterns of steel scraps (Lee and Sohn 2015, Zhong et al 2018), little attention has been paid on the trade linkages between different countries or regions from an AEs transfer perspective.

Figure 1 .
Figure 1. Trade patterns of key nodes within GSSTN from 2000 to 2021.unit: kilotons (kt)].Note: 1. Countries are ranked by the net import amount from left to right, with positive values indicating imports and negative values indicating exports.2. The evolution of steel scraps trade volume among these key nodes is detailed in the Supporting Information figures S4-S47.3. The red line represents the net-import of each key node.

Figure 2 .
Figure 2. Global steel scraps trade network in 2000 (a), 2007 (b), 2014 (c), and 2021 (d).Note: 1.Each curve represents one steel scraps trade relation, flowing from an export country to an import country with a clockwise direction; 2. The size of each node represents the total trade volume of a country; 3. The width of each edge represents the trade flow between two trading partners; 4. Different colors represent different communities.Nodes within GSSTN are grouped by community detection, with nodes in the same community being more densely connected to each other than to the rest of the network.The community number represents the ranking of each community, with community 1 representing the community with the most trade relations, while community 10 representing the community with the least trade relations.5. Since countries in communities 6-10 have significantly smaller trade volumes than those in the top five communities, we do not include trade relations within communities 6-10 in this figure.

Figure 3 .
Figure 3. Trade amounts of AEs embodied in global steel scraps from 2000 to 2021.(a) Historical evolution of global exports of each alloy element embodied in steel scraps over 2000-2021 (unit: kt); (b) Shares of AEs embodied in steel scraps over 2000-2021.(c) Top ten countries in total exports of each alloy element embodied in steel scraps from 2000 to 2021 (unit: kt).(d) Top ten countries in total imports of each alloy element embodied in steel scraps from 2000 to 2021 (unit: kt).

Figure 4 .
Figure 4.The amounts of alloy elements in key countries accounting for 80% of the global steel scraps exports and imports in 2000 (a), 2007 (b), 2014 (c), and 2021 (d) (unit: kt).

Figure 5 .
Figure 5. Trade networks of alloy elements embodied in steel scraps from 2000 to 2021 (unit: kt).Note: 1.Each curve represents the trade flow of each AE embodied in steel scraps.2. The size of each node represents the total trade amount of a country.3. The width of each edge represents the trade flow between two trading partners.4. Different colors represent different communities.Nodes within GSSTN are grouped by community detection, with nodes in the same community being more densely connected to each other than to the rest of the network.The community number represents the ranking of each community, with Community 1 representing the community with the largest trade relations, while Community 10 representing the community with the least trade relations.

Table 1 .
Indicators used in social network analysis.
out-strength, betweenness centrality, and modularity indicators (table 1) (Jackson 2010, Liu et al 2016, Jiang et al 2019, Ma et al 2021, Pacini et al 2021, Gao et al 2022a).The associated calculation methods are detailed in the supporting information.All AEs flows are calculated based on the actual contents of such elements in different steel scraps products under different HS codes (Wang et al 2021b).The uncertainty of this study is mainly caused by the AEs contents (Gao et al 2022b).Simplified ranges are determined to estimate the uncertainties (Laner et al 2014).According to a previous study, the aggregate uncertainty for element concentration data may be lower than 20% (Liu et al 2013).Relevant studies on steel AEs present that such uncertainty ranges from 2% to 10% (Gao et al 2022b, Su et al