Dynamic Substance Flow Analysis of Aluminum and Its Alloying Elements* 1

Transcription

1 Materials Transactions, Vol 48, No 9 (27) pp 2518 to 2524 #27 The Japan Institute of Metals Dynamic Substance Flow Analysis of Aluminum and Its Alloying Elements* 1 Hiroki Hatayama* 2, Hiroyuki Yamada* 3, Ichiro Daigo, Yasunari Matsuno and Yoshihiro Adachi Department of Materials Engineering, Graduate school of Engineering, The University of Tokyo, Tokyo , Japan Aluminum demand in Japan has grown significantly during the last few decades For most uses, small amounts of other metals are added to the primary aluminum to make harder alloys, which are classified by the nature and concentrations of their alloying elements Aluminum scraps from end-of-life products, which are used as raw materials for secondary aluminum, are often mixtures of several alloys Therefore, not only the amount of scrap but also the concentrations of their alloying elements must be taken into account when assessing the maximum recycle rate of aluminum scraps This paper reports a dynamic substance flow analysis of aluminum and its alloying elements in Japan, focusing on the alloying elements Si, Fe, Cu and Mn We devised eight categories of aluminum end uses and 16 types of aluminum alloys The amount of each alloy in each end-use category was estimated from statistical data We then estimated future quantities of discarded aluminum in each of the eight categories using the population balance model At the same time, we calculated the concentrations of the alloying elements in each of the end uses It was estimated that the amount of aluminum recovered in Japan would be about 18 kt in 25, which is 212 times that recovered in 199 Calculated concentrations of alloying elements in aluminum scraps showed good correlation with those of the measured data [doi:1232/matertransmra2712] (Received May 1, 27; Accepted July 3, 27; Published August 22, 27) Keywords: material flow analysis, substance flow analysis, aluminum alloy, alloy content, aluminum scraps, population balance model 1 Introduction Increasing public concern for environmental protection and resource conservation has generated interest in the recycling of materials In Japan, technological development and legislation that protects the environment have increased the recycling of various materials such as aluminum, steel, concrete and wood from end-of-life products Therefore, materials in the products accumulated in a society will become raw materials for recycling in the future Kleijn et al 1) and Van der Voet et al 2) developed a dynamic approach for estimating the material outflow (as end-of-life products) from in-use stock Using similar approaches, previous research has estimated the amount of discarded material likely to be recovered in the future, in particular focusing on steel (Kakudate et al, Daigo et al), 3,4) construction minerals (Hashimoto et al), 5) consumer durables (Tasaki et al) 6) and TVs (Yamada et al) 7) This paper focuses on the recyclability of aluminum, which is one of the most widely used metals in Japan Aluminum is a versatile metal, owing to various properties such as its light weight, high corrosion resistance and good formability These properties are fostered by adding alloying elements; Si, Fe, Cu, Mn, Mg, Zn, and so on R ecycling of aluminum is important as the energy consumed in producing secondary aluminum is much less than that used in producing primary aluminum In producing secondary aluminum from scraps, the alloying elements in the scraps are usually reused into the recycled alloys and at the same time some alloying elements are newly added in order to satisfy standards on chemical composition On the other hand, excess presence of the alloying elements would require additional primary aluminum for dilution because some alloying elements * 1 This Paper was Originally Published in Japanese in J Japan Inst Metals 7 (26) * 2 Graduate Student, The University of Tokyo * 3 Graduate Student, The University of Tokyo, Present address: New Energy and Industrial Technology Development Organization cannot be removed at secondary smelters Therefore the chemical composition of scraps should be taken into account in assessing the recyclability of aluminum These features of aluminum recycling require MFA/SFA (material flow analysis/substance flow analysis) to be conducted for each of the respective alloys A material flow analysis studying the limitations caused by the presence of contaminants in recycling has been conducted for Cu in steel (Kakudate et al, Daigo et al) 3,4) and for Ni and Cr in stainless steel (Igarashi et al) 8) A static material flow analysis of aluminum, which shows annual aluminum flow in Japan, has already been conducted by Umezawa et al 9) However, there has been no dynamic analysis conducted because of the lack of data on such factors as product lifetime and demand for each end use The objective of this paper was to analyze dynamically the substance flow of aluminum and its alloying elements In order to do this, statistical data were analyzed in advance, as there were insufficient data for the analysis 2 Data Preparation 21 Consumption by end use Published statistics 1) provide information about aluminum demand in Japan classified into about 3 end uses In this paper, end uses were aggregated into eight categories, and demand data on these eight end uses from 196 to 23 were prepared (Table 1) These data took into account the indirect trade of aluminum and industrial scraps, after allocation (described in 23) The indirect trade is the difference between statistical domestic demand and actual consumption in Japan, as goods are manufactured in Japan and consumed in other countries and vice versa In this study, automobile export was taken into account because it is supposed to have much larger impact than other indirect trades We calculated the amount of aluminum included in exported automobiles as follows First, the average consumption of aluminum in an automobile

2 Dynamic Substance Flow Analysis of Aluminum and Its Alloying Elements 2519 Table 1 End-use classifications for aluminum Table 2 Chemical compositions of aluminum alloys (23 fiscal year) End use Foil Fabricated metal General Electric communication Other products Examples foil household utensils, furniture beverage cans for agriculture and fishery home electronics appliances, audio automobiles (engines, heat exchangers etc) window frames, doors precision, ships, aircraft was calculated by dividing the statistical demand for aluminum in the automobile category by the annual automobile production 11) Actual domestic consumption was then calculated by multiplying the average consumption of aluminum by the annual registration of new automobiles 11) And this study did not take into account other indirect trades In other end uses (eg electric communication ), there are many kinds of products (eg TV, refrigerator, washing machine), therefore it is difficult to estimate the average aluminum consumption in each end uses Industrial scraps were also taken into account for mill products, which are manufactured from wrought alloy Fabricating mill products generates industrial scraps as either mill ends or irregular products The amount of industrial scrap was calculated by multiplying the demand for mill products by the yield rate of scraps from milling processes We assumed that the industrial scraps were generated in the processing year The yield rates for industrial scraps from various end-use manufacturing processes were obtained from the literature 12) It should be also noted that all the statistical data in this study related to the fiscal year 22 Production by alloy Aluminum alloys include many kinds of alloying elements Four of these alloying elements, Si, Fe, Cu, and Mn, were considered in this study because they each have the potential to restrict aluminum recycling The production of 42 types of aluminum alloys in mill products, castings and die-castings can be obtained from published statistics 1) We prepared data on production according to alloy type from 1974 to 23 by aggregating some alloy types that have similar chemical compositions or that have relatively small production (Table 2) It was assumed that the concentration of every alloying element was equal to the upper limit described in the industrial standards 13) The components of aggregated alloy types were calculated as the weighted average according to production volume The compositions of wrought alloys calculated from data obtained in 23 are shown in Table 2 Alloy compositions were calculated from data obtained in 1974 to 22 using the same method, and little fluctuation was found compared with those of 23 Therefore, the compositions of wrought alloys from 196 onwards were represented by those of 23 As for aluminum castings, their composition was assumed to be equal to that of AC2B alloy because it is the most widely used casting alloy Concentration (%) Si Fe Cu Mn 1 series series Other 3 series series Mill products Other 5 series Other 6 series series series Castings Die castings Table 3 Matrix of demand by end use and alloy End use 1 End use m Total Alloy 1 a 11 a 1m X 1 Alloy n a n1 a nm X n Total Y 1 Y m The composition of die-castings was assumed to be equal to that of ADC12 alloy for the same reason 14) 23 Data development demand by both end use and alloy Data on demand according to end use and on production according to alloy were obtained from the available statistics Data on demand by both end use and alloy were determined from these two sets of data using the following procedure Table 3 is a matrix representing the relationship between various end uses and the alloys used for them The X (row sum) and Y (column sum) values in the matrix can be obtained from statistics, where X is the demand by alloy and Y is the demand by end use The value of each matrix element was determined by demand by end use (column) and demand by alloy (row), estimated by the variables X and Y As the first step in this estimation, the alloy types used for each end use were determined from the literature 15) The matrix elements were then classified into two types, identifiable elements and inferable elements An identifiable element was an element whose value could be identified because there was a reliable relationship between end use and alloy For example, only 1 series alloys were used in products in the Foil end-use category In this case, the element a 1 series Foil was an identifiable element and the value was defined as being equal to Y Foil, while the other elements in that column were set to An inferable element was an element whose value could not be identified, as the alloy used in the end use was only assumed Values of the

3 252 H Hatayama, H Yamada, I Daigo, Y Matsuno and Y Adachi Table 4 Identifiable and inferable elements defined by end use and alloy Foil Electric communication Fabricated General metal Heat Body End/Tab Fin Other Engine Other exchanger 1 series series Other 3 series series ++ Mill products Other 5 series Other 6 series series series Castings Die castings Other Exports products ++; Identifiable element +; Inferable element *; Calculated element Blank; Null inferable elements were defined according to the allocation method described later Table 4 shows the identification of various elements based on the literature: ++ denotes an identifiable element, + denotes an inferable element, and blank elements are For the assumption of which alloys were used in the category, alloy 34 was assumed to be used in the can body and alloy 5182 was assumed to be used in the can end The excess of demand over production for alloys 34 and 5182 was made up for by other 3 series alloys and the 552 alloy The allocated values for the various elements should show no discrepancy between row and column totals In this paper, demand by end use (Y i ) was allocated to each of its component elements according to the proportion of alloy production, X i However, there was difficulty in making a valid allocation in Exports and Other products categories because the alloy types used in these end-use categories could not be identified from any literature Therefore, the values of elements belonging to these two end uses were determined from the row sum, X i, after the allocation in the other columns was complete In the second step, every value of the identifiable element was defined At the same time, the value of alloy production for the inferable elements (represented as X i ) was calculated by subtracting the values of each identifiable element from the total alloy production, X i The value of end-use demand for the inferable elements (represented as Y i ) was calculated in the same way In the third step, Y i was allocated into respective inferable elements in the column, according to the proportion of X i The second and the third steps were applied to all end-use categories, except Exports and Other products, defining all values in these columns Then in the final step, the elements of the Exports and Other products end-use columns were defined In each row, the amount of alloy used for each of the two end uses was derived by subtracting all the fixed values from the total alloy production, X i The derived value was then divided into the two end uses according to the ratio of their end-use demand Statistics are kept of wrought alloy production, recording its shape, sheet and extrusion 1) The allocation method was applied individually to two of the shapes, and the total was calculated Table 5 shows the estimated demand by both end use and alloy from the 23 data to which the method was applied It should be noted that there was a difference of 2% between the grand total of alloy production (sum of X i ) and the grand total of end-use demand (sum of Y i ) in Table 5 This difference was adjusted in the Exports and Other products categories 3 Quantitative and Qualitative Analysis on Scraps The amounts of aluminum discarded from various end uses were estimated using dynamic analysis (Population Balance Model: PBM) 3) We defined some of the parameters of a product s lifetime, as this was necessary for the dynamic analysis For the Foil category, a product lifetime was not considered because little aluminum is recovered from foil products For Fabricated metal, General, Electric communication and Other products categories, the mean lifetimes were obtained from the literature 12) Other parameters of product lifetime were based on past research on consumer durables (Tasaki et al) 6) s were assumed to be discarded in the production year because their mean lifetime was less than one year For the category, a lifetime distribution was defined from the

4 Foil Fabricated metal Dynamic Substance Flow Analysis of Aluminum and Its Alloying Elements 2521 Table 5 Demand for mill products by end use and alloy (23 fiscal year) Electric communication General Body End/Tab Fin Other Heat Engine exchanger Other unit: (t) Other Exports products 1 series series Other 3 series series Mill products Other 5 series Other 6 series series series Total Total Table 6 Lifetime distribution, yield rate and collection rate in each end-use category End use Distribution function Lifetime distribution function Yield rate Mean of Collection lifetime Variance industrial rate (%) (years) scraps (%) Foil 1 Fabricated metal weibull General weibull Electric communication weibull log-normal Other products weibull Shape parameter (in weibull distribution function) empirical data The estimated figures were verified from new registrations and disposals 11,16) Window frames were the main product in the category, therefore the lifetime of products in this category was represented by that of dwellings 17) Considering the aluminum input into society in a given year, the amount of aluminum likely to be discarded in later years can be calculated by multiplying the input by the disposal probability, where disposal probability was defined according to the assumed lifetime distribution By calculating the total amount of discarded aluminum that originated from aluminum input in the past, the amount of aluminum discarded in a certain year in the future can be estimated This study estimated the amount of discarded aluminum likely to be discarded each year until 25 After the estimation of discarded aluminum, the amount of recovered aluminum was estimated based on the collection rate of each end-use category The numerator of the collection rate is the amount of aluminum collected and consumed as aluminum scraps in Japan, and the denominator of the rate is the amount of discarded aluminum Aluminum which is merged into general waste or exported as used goods is not included in the collection rate For beverage cans, published data 18) on consumption and recovery were used in this study The collection rates of other end uses were obtained from the literature 12) The parameters assumed for this study are shown in Table 6 In the dynamic analysis, demand for each end use from 24 onwards was assumed to remain the same as the demand in 23 Other parameters, such as the lifetime distribution function, the yield rate, the indirect trade and the collection rate were also assumed to be constant in the future, as shown in Table 6 Aluminum is collected from end-of-life products for recycling However, collected aluminum cannot be separated into its alloy types Furthermore, once alloyed with aluminum, some of the alloying elements cannot be removed from

5 2522 H Hatayama, H Yamada, I Daigo, Y Matsuno and Y Adachi Amount of scrap recovery / kt Industrial scrap Other products Electric communication General Fabricated metal Statistics Table 7 Chemical compositions of scraps discarded from (a) Fabricated metal, (b), (c) General, (d) Electric communication, (e) (mill products), (f) (engine), (g) and (h) Other products end-use categories (a) Si Concentration Fe (%) Cu Mn Fiscal year their alloys (because of technical and cost limitations) These facts need to be considered when analyzing the substance flow of aluminum alloying elements Therefore, we assumed that aluminum from end-of-life products was collected without alloy-by-alloy separation, for all end uses except the category We assumed that aluminum contained in end-of-life automobiles was collected using a rough separation between mill products used in the heat exchanger and the body, and castings used in the engine Using this assumption, the chemical composition of scrap generated from every end use was calculated from the results of the PBM as a weighted average of the alloy compositions shown in Table 2 It was also assumed that the proportion of each wrought alloy in the total production of mill products for each end use would be constant over time This means that the weighted average of wrought alloys calculated for each end use was assumed to be the same as that in 23 4 Results Fig 1 Estimation of scrap recovery Figure 1 shows the estimated amount of aluminum predicted to be recovered in Japan until 25 According to the analysis, the amount of aluminum recovered in 2 would be 153 times as large as that in 199, and in 25, 212 times The increase in recovered aluminum was due to the increase in discarded aluminum, especially from and end uses The amounts of statistical scrap recovery from 199 to 2 are also shown in Fig 1 19) This estimation based on parameters in Table 6 fits well with the amount of statistical scrap recovery The compositions of scrap aluminum were estimated for every year until 25, as shown in Table 7 For the (Table 7(b)) and (Table 7(g)) categories, no fluctuation in composition was observed because only mill products were used for these two end uses For the remaining five end uses (also disregarding the (engine) category), concentrations of the alloying elements would decrease over time because of an increase in the proportion of mill products used in finished products Estimated compositions of scrap were compared with measured data, in order to validate the allocation method adopted in this paper The measured data were obtained from the literature and from an investigation of a secondary smelter 14,2) The results of this comparison are shown in (b) Si Concentration Fe (%) Cu Mn (c) Si Concentration Fe (%) Cu Mn (d) Si Concentration Fe (%) Cu Mn (e) Si Concentration Fe (%) Cu Mn (f) Si Concentration Fe (%) Cu Mn (g) Si Concentration Fe (%) Cu Mn (h) Si Concentration Fe (%) Cu Mn

6 Dynamic Substance Flow Analysis of Aluminum and Its Alloying Elements 2523 Concentration (%) Concentration (%) Si Fe Cu Mn Si Fe Cu Mn Si Fe Cu Mn (a) (b) (c) Si Fe Cu Mn Si Fe Cu Mn (d) (e) Measurement Estimation Measurement Estimation: General Estimation: (engine) Fig 2 Comparison between the measurements and the estimation of the chemical compositions of scraps discarded from (a) Fabricated metal, (b), (c), (d) General including automobile (engine) and (e) (engine) Fig 2 For some kinds of scraps, there were differences between the estimation and the measurement These differences were attributable to the following reasons (1) The types of products included in the scraps were different, (2) alloy composition was assumed incorrectly resulting in inaccurate estimation of scrap compositions, (3) contamination by impurities occurred in the separation and collection processes when aluminum was collected from end-of-life products and (4) the allocation method was invalid An incorrect assumption of alloy composition means that the alloying elements did not add up to their upper limit as described in industrial standards In this case, the estimated concentration of the alloying element in the scrap would be higher than the actual concentration In the case of contamination with impurities, the estimated concentration of the incorporated elements would be lower than the actual concentration With this in mind, we discuss below the causes of the differences between the estimated and measured scrap compositions Figure 2(a) shows the comparison between the measured composition of kitchen utensil scraps and the estimated composition of the Fabricated metal category The estimation indicates a higher Si concentration and a lower Fe concentration than the measurement The overestimation of Si was considered to be because of differences in the kinds of products included in the scrap: the estimated composition was based on the weighted average of mill products, castings and die-castings used for the Fabricated metal end use, whereas kitchen utensils do not use castings and die-castings that include much Si The underestimation of Fe was attributable to the incorporation of impurities, such as handgrips of pans, into the scrap Figure 2(b) shows the comparison between the measured composition of used beverage cans and the estimated composition of the Beverage can category These two show good correlation Since the alloys used in beverage cans are identifiable and few impurities could have become incorporated into the beverage cans, the estimation was more reliable Figure 2(c) shows the comparison between the measured composition of aluminum window frame scrap and the estimated composition of the category Although the incorporation of iron into aluminum window frame scrap has been an issue, the measured data showed lower concentrations of all four of the alloying elements compared with the estimated data This result suggests that concentrations of alloying elements in produced alloys were assumed to be higher than the actual concentrations Figure 2(d) shows the comparison for scrap The scrap consisted of general and automobile engine scrap; therefore, composition of these two products was estimated separately and then compared with the measured data The measured composition lay midway between the two estimates, which correlated well with the scrap content Figure 2(e) shows the comparison between the measured composition of automobile engines and the estimated composition of the (engine) category These two showed good correlation, although the estimated concentrations of alloying elements were slightly lower than the measured ones The compositions of actual scrap were measured in 1995 (Fig 2(a), (b) and (d)) and 26 (Fig 2(c) and (e)) The estimated data for the corresponding year were used for the comparison; therefore, a small difference was observed between (engine) compositions in Fig 2(d) and (e) As shown in Fig 2, the allocation method leads to a valid estimation of the chemical compositions in aluminum scrap On the other hand, external factors such as incorrect assumptions about the alloy composition and the presence of impurities in the scrap can result in some differences between the estimated and measured compositions Further investigation into these factors will be required to improve the estimation of scrap composition 5 Conclusions A dynamic substance flow analysis of aluminum was conducted In the analysis, both the quantitative information and the chemical composition information were obtained in parallel The allocation method was developed to obtain data on demand by both end use and alloy Time-series data on the amount and composition of aluminum scrap discarded by society were estimated from the dynamic substance flow analysis Estimated compositions were compared with the measured data, which verified the allocation method However, further investigation will be required to study the effect of external factors such as incorrect assumptions and contamination with impurities Acknowledgements This research was supported by a Grant-in-Aid for Scientific Research (No ) from the Ministry of Education, Culture, Sports, Science and Technology of Japan REFERENCES 1) R Kleijn, R Huele and E Van der Voet: Ecological Economics 32 (2) ) E Van der Voet, R Kleijn, R Huele, M Ishikawa and E Verkuijlen:

7 2524 H Hatayama, H Yamada, I Daigo, Y Matsuno and Y Adachi Ecological Economics 41 (22) ) K Kakudate, Y Adachi and T Suzuki: Sci Technol Adv Mater 1 (2) 15 4) I Daigo, D Fujimaki, Y Matsuno and Y Adachi: Tetsu-to-Hagané 91 (25) ) S Hashimoto, H Tanigawa and Y Moriguchi: Proc 31st Conf on Environmental Systems (Japan Society of Civil Engineers, Kitakyushu, 23) pp ) T Tasaki, M Oguchi, T Kameya and K Urano: J Jpn Soc Waste Management Experts 12 (21) ) H Yamada, Y Matsuno, I Daigo and Y Adachi: J Jpn Soc Waste Management Experts 18 (27) ) Y Igarashi, I Daigo, Y Matsuno and Y Adachi: Tetsu-to-Hagané 91 (25) ) O Umezawa and M Okubo: Proc 18th Spring Meeting (The Japan Institute of Light Metals, 25) p 15 1) Japan Aluminium Association: Aluminium Statistics in Japan (196 23) 11) Japan Manufactures Association, Inc: Motor Vehicles Statistics of Japan (1973 2) 12) Clean Japan Center: Haikibutsu Genryouka no tameno Shakaishisutemu no Hyouka ni kannsuru Chousakenkyu Houkokusho (1999) pp ) Japanese Standards Association, JIS HB Non-Ferrous Metals and Metallurgy 23, pp ) DAIKI Aluminium Industry: Private communication (25) 15) Japan Aluminium Association: Aluminum Handbook 6th edition, pp ) Inspection & Registration Association: Wagakuni no Jidousha Hoyuu Doukou ( ) 17) Y Komatsu, Y Kato, T Yoshida and T Yashiro: Journal of Archit Plann Environ Engng, AIJ 439 (1992) ) Japan Aluminum Can Recycling Association: Aluminium Can Recycle News 19) Ministry of International Trade and Industry: Yearbook of Minerals and Non-ferrous Metals Statistics (199 21) 2) NEDO: Hitetsu Kinzokukei Sozai Risaikuru Sokusin Gijutu Kenkyuu Kaihatsu; Kiso Cchousa Kenkyuu, Youso Gijutu Kenkyuu Chousa Houkokusho (1996), pp 53 55