Supporting Information for:
Systematic characterization of generation and management of e-waste in China
Huabo Duan*,aJiukun Hub, Quanyin Tan c, LiliLiu,cYanjie Wang, band Jinhui Li,c
aCollege of Civil Engineering, Shenzhen University, 518060 Shenzhen, China.
bDongjiang Environmental Co., Ltd., 518057 Shenzhen, China.
b State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, TsinghuaUniversity, 100084Beijing, China. E-mail: ; Tel: +8610-62794143
1. Generation and collection
TheSales ObsolescenceModel (SOM) approachdeveloped byMiller (2012) and Duan et al. (2013) is used to calculate generation quantities of e-waste. A flowchart of the life cycle of electric and electronics is shown in Figure S1 as a guide for key definitions in this study. The term “generation” refers to electronics coming directly out of use (retired) or post-use storage destined for collection or disposal. Thus, “generation” is consistent with the term “ready for end-of-life [EOL] management”. One generation pathway for items is disposal (F), including landfills and incinerators. Another generation pathway already mentioned is collection for processing in a working (H) or an obsolete (G) state. An assumption is made that after two terms of use, items are obsolete. The used electronics processor, having collected the used electronic whole unit, opts either to prepare it for reuse by a new user in China (C), recover parts and materials from the item (I) and transfers them to downstream vendors (some of which may be in foreign countries), or export the used electronic product as a whole unit (J). The focus of this study is on used electronic products that are whole units. “Whole Units” refers to intact monitors, computers, mobile phones, LHAs etc. that may or may not have been refurbished. Thus, this excludes disassembled products that may be exported as several different commodity material or product streams.
Fig. S1 Life Cycle Flow Chart of Electronic Products (Miller, 2012 and Duan et al., 2013)
Unlike previous studies, this study includes uncertainty in input quantities and then propagates that uncertainty into outputs using Monte Carlo simulations. Generation quantities are modeled and then combined for the variouselectric and electronic types. This is done because these types have different consumption, use, and end of use disposition habits. The basic approach for quantifying generation and collection includes the following steps:
(1)Determine the sales of a product in a China over a time period.
(2)Determine the typical distribution of lifespan for the product over a time period using survey-based data (but from literature).
(3)Determine the annual markets shares for each type of electronic in terms of ‘sizes’ distribution. the unit weights (divided by ‘sizes’), such as screen inches of monitor and TVs, volume (capacity) of refrigerator and washing machine, and output power of air conditioner.
(4)Determine median values of the material composition of products and the content of selected common metals, precious metals, and less common metals in printed circuit boards, CRT glass and Li-ion battery for 11 types of electrics and electronics.
(5)Calculate how many products are predicted to be generated in a given year using the sales and lifespan information; calculate the weight of generated waste by multiplying unit weights and size by the quantities; and calculate the weight of generated waste by multiplying mass fractions of materials and metals by the quantities.
These generation calculation steps roughly comprise a SOM (alternatively known as market supply method). Studies cover different products, time periods, geographical regions, and vary in complexity
1.1 Sales data
Several sources offerannual production volume, export and import data as shown in TableS1, including ‘Yearbookof China Information Industry’ (Ministry of Industry and Information Technology), ‘Yearbookof China Information Industry’ (National Bureau of Statistics of China). Sales here refer to manufacturer shipments into the domestic channel (Equation 1: Sales (S)= Production Volume (P) – Export (E)+ Import (I) ). The sales data (Table S2) in surrounding years (from 2013 to 2015) with projection were allowed to vary uniformly one standard deviation from the mean (uniform distribution), by given an approximate 10% of Correlation of Variances (COV).
Equation 1: Sales data calculation
Prediction model: an association between the sales of electronics and the corresponding variables (only time series parameter is considered in this study) is to be expected.Based on this assumption, the Pearson product moment correlation coefficient was initially applied to find the coefficient of determination () between the independent variables. Subsequently, to test the hypothesis of independence between the selected explanatory variables, the Student t test was applied on the coefficient of the variable. Because the analysis included one variable, and the variable displayed a linear distribution, a multiple linear regression analysis was applied to determine the probable shape of the relation between variable and to estimate the sales quantity of electronics, which corresponds to the values of the analyzed variable. From this, it may be ascertained that the generation of sales may be explained by a multiple linear equation having the form of Equation 2. The results are shown in Table S3.
Equation 2: Prediction model
Where Y is the dependent variable; is the intercept; is the independent variable; ,are regression parameters; and ε is residuals.
1
Table S1 Production volume (In China)
Year / LCDmonitor / LCDTVs / PDPTVs / Laptop / Desktop / CRTMonitor / CRTColorTV / Mobile Phone / Refrigerator / Air Conditioner / Washing Machine1990 / 158 / 58 / 10,339
1991 / 231 / 28 / 12,051 / 1,885 / 630 / 2,147
1992 / 0 / 392 / 68 / 13,331 / 2,693 / 983 / 3,067
1993 / 0 / 639 / 126 / 13,073 / 3,847 / 1,965 / 4,381
1994 / 7 / 875 / 1,316 / 16,371 / 5,495 / 3,930 / 6,259
1995 / 1 / 836 / 566 / 19,121 / 1,310 / 6,869 / 6,830 / 7,823
1996 / 1 / 1,388 / 1,995 / 25,376 / 600 / 8,586 / 7,860 / 9,779
1997 / 8 / 2,066 / 5,485 / 27,113 / 3,780 / 9,540 / 9,740 / 10,866
1998 / 1 / 3,933 / 7,858 / 36,430 / 8,550 / 10,600 / 11,570 / 12,073
1999 / 3 / 1 / 4,055 / 15,046 / 42,620 / 23,010 / 12,100 / 13,380 / 13,421
2000 / 1,492 / 79 / 7,118 / 27,000 / 39,361 / 52,479 / 12,790 / 18,270 / 14,430
2001 / 1,636 / 283 / 8,776 / 33,269 / 40,937 / 80,317 / 13,513 / 23,130 / 13,416
2002 / 5,146 / 165 / 18 / 1,170 / 14,633 / 44,068 / 41,610 / 121,464 / 16,800 / 31,350 / 15,871
2003 / 21,731 / 202 / 112 / 12,870 / 18,830 / 51,526 / 65,215 / 182,314 / 22,426 / 48,209 / 19,645
2004 / 77,270 / 808 / 194 / 27,500 / 17,624 / 24,224 / 72,161 / 237,516 / 30,076 / 63,903 / 25,334
2005 / 130,760 / 4,530 / 770 / 45,650 / 35,188 / 29,811 / 76,677 / 303,542 / 29,871 / 67,646 / 30,355
2006 / 107,020 / 9,950 / 680 / 59,120 / 34,246 / 26,570 / 72,330 / 480,138 / 35,309 / 68,494 / 35,605
2007 / 134,790 / 17,579 / 1,089 / 86,710 / 34,020 / 9,570 / 66,040 / 548,580 / 43,971 / 80,143 / 40,051
2008 / 112,510 / 29,425 / 3,509 / 108,590 / 28,079 / 20,460 / 56,760 / 559,640 / 48,000 / 81,474 / 44,470
2009 / 127,840 / 67,653 / 1,911 / 150,090 / 31,939 / 2,300 / 29,160 / 681,930 / 59,305 / 80,783 / 49,736
2010 / 106,570 / 89,375 / 2,141 / 185,840 / 50,080 / 26,790 / 25,117 / 998,270 / 72,957 / 108,875 / 62,477
2011 / 119,040 / 104,010 / 3,121 / 238,974 / 79,748 / 7,220 / 15,152 / 1,132,580 / 86,992 / 139,125 / 67,159
2012 / 111,578 / 114,183 / 2,139 / 252,890 / 62,263 / 15,419 / 8,457 / 1,181,550 / 84,270 / 132,811 / 67,911
Table S2 Sales Data (In China)
Year / LCDmonitor / LCDTVs / PDPTVs / Laptop / Desktop / CRTMonitor / CRTColorTV / Mobile Phone / Refrigerator / Air Conditioner / Washing Machine1990 / 126 / 40 / 6,766
1991 / 184 / 19 / 7,886 / 1,188 / 413 / 1,475
1992 / 0 / 313 / 47 / 8,724 / 1,697 / 644 / 2,107
1993 / 0 / 510 / 87 / 8,555 / 2,424 / 1,288 / 3,011
1994 / 2 / 698 / 911 / 10,713 / 3,462 / 2,576 / 4,301
1995 / 0 / 667 / 392 / 12,513 / 579 / 4,328 / 4,477 / 5,376
1996 / 0 / 1,108 / 1,381 / 16,606 / 265 / 5,410 / 5,152 / 6,720
1997 / 2 / 1,649 / 3,797 / 17,743 / 1,672 / 6,011 / 6,384 / 7,467
1998 / 0 / 3,139 / 5,440 / 23,839 / 3,781 / 6,679 / 7,584 / 8,296
1999 / 1 / 0 / 3,236 / 10,417 / 27,890 / 10,175 / 7,624 / 8,770 / 9,223
2000 / 679 / 17 / 5,680 / 18,693 / 25,758 / 23,207 / 8,059 / 11,976 / 9,916
2001 / 745 / 62 / 7,004 / 23,033 / 26,789 / 35,517 / 8,515 / 15,162 / 9,219
2002 / 2,343 / 101 / 11 / 255 / 11,678 / 30,510 / 27,229 / 53,713 / 10,586 / 20,550 / 10,906
2003 / 8,289 / 202 / 95 / 1,649 / 15,070 / 20,848 / 45,232 / 109,037 / 14,130 / 31,600 / 13,499
2004 / 64,071 / 808 / 96 / 3,035 / 12,721 / 24,224 / 49,986 / 104,193 / 18,951 / 41,888 / 17,409
2005 / 60,686 / 2,354 / 535 / 12,636 / 28,979 / 10,160 / 51,219 / 87,993 / 18,822 / 44,341 / 20,860
2006 / 54,445 / 5,024 / 42 / 11,661 / 26,116 / 14,512 / 42,697 / 123,636 / 22,248 / 44,897 / 24,467
2007 / 34,411 / 9,948 / 533 / 14,624 / 25,299 / 3,574 / 42,233 / 82,001 / 24,687 / 48,157 / 26,643
2008 / 38,621 / 18,161 / 2,249 / 10,395 / 20,269 / 16,714 / 36,455 / 44,135 / 31,915 / 49,552 / 29,789
2009 / 49,724 / 50,072 / 1,297 / 40,169 / 25,342 / 1,191 / 15,120 / 123,607 / 43,992 / 57,033 / 36,036
2010 / 59,355 / 61,174 / 1,649 / 48,981 / 39,893 / 26,089 / 10,816 / 259,028 / 42,387 / 71,878 / 44,584
2011 / 58,942 / 71,684 / 2,634 / 70,754 / 70,400 / 7,011 / 4,251 / 266,738 / 54,365 / 95,498 / 46,063
2012 / 70,213 / 68,358 / 1,944 / 45,714 / 53,540 / 15,394 / 2,237 / 176,667 / 51,115 / 89,151 / 45,091
Table S3 Sales Data (In China) (Projection, mean values)
Year / LCDmonitor / LCDTVs / PDPTVs / Laptop / Desktop / CRTMonitor / CRTColorTV / Mobile Phone / Refrigerator / Air Conditioner / Washing Machine2013 / 53,141 / 75,246 / 1,428 / 66,773 / 57,901 / 9,447 / 3,465 / 300,501 / 58,515 / 90,605 / 51,832
2014 / 54,913 / 84,033 / 1,235 / 76,504 / 63,899 / 8,267 / 2,444 / 333,400 / 63,275 / 96,692 / 55,725
2015 / 56,394 / 92,820 / 1,035 / 86,596 / 70,187 / 7,199 / 1,520 / 375,495 / 68,035 / 102,779 / 59,617
2016 / 57,531 / 101,606 / 888 / 96,989 / 76,766 / 6,225 / 677 / 420,098 / 72,795 / 108,867 / 63,509
2017 / 58,328 / 110,393 / 795 / 107,622 / 83,636 / 5,328 / 338 / 467,210 / 77,555 / 114,954 / 67,402
2018 / 58,845 / 119,180 / 743 / 118,436 / 90,797 / 4,498 / 169 / 516,829 / 82,315 / 121,042 / 71,294
2019 / 59,158 / 127,967 / 716 / 129,372 / 98,249 / 3,726 / 85 / 568,957 / 87,075 / 127,129 / 75,186
2020 / 59,335 / 136,753 / 703 / 140,369 / 105,991 / 3,003 / 42 / 623,593 / 91,835 / 133,216 / 79,079
1
1.2 Determine the distribution of lifespan for the product over a time period
This method for determining typical distributions of lifespans for the product is a refinement of the model developed by Matthews et al. which accounts for two use stages (initial and reused), and accounts for different fates after each stage (Matthews et al., 1997).
1
The primary difference is the incorporation of a distribution of lifespan lengths and path probabilities so that both data quality uncertainty and variation are considered. The steps are as follows:
- Combine literature and industry estimates for the distribution of lengths of each lifespan stage(s) (eg., B. Initial Use, E. Reuse Storage) in Figure S2 (repeated above for convenience) to arrive at a mean estimate with uncertainty for each lifespan stage.
- Define pathways to generation (Figure S3) involving combinations of lifespan stages related to Figure S1.
This method is somewhat of an underestimate, because we do not estimate the second round of generation of products that underwent formal domestic reuse. Initial sensitivity analyses suggest that the result is not very sensitive to the exclusion of the second round of generation.
Fig.S2 Probability tree diagram of informal paths leading to generation. Letters and colors refer to lifespan stages in Figure 1. The probabilities of a path to a lifespan stage are represented by P( lifespan stage), or its complement P(lifespan stage’). Some probabilities are conditional on previous pathways, P(lifespan stage| previous lifespan stage) (Miller, 2012 and Duan et al., 2013)
Fig. S3 Probability Tree Diagram of Informal and Formal Paths Leading to Generation (Miller, 2012 and Duan et al., 2013)
- Combine the lengths of the lifespan stages to calculate the lengths of each pathway to generation and estimate the probability of each pathway to generation.
In Table S4 below, the equations for determining the mean path length and mean path probability are found for each of the six pathways to generation.
Table S4 Equations used to calculate mean path length and mean path probability
Six Paths () / Mean Path Length / Mean Path ProbabilityPath B, D, C, E / / 1*P(D)*P(C|D)*P(E)
Path B, D, C, E’ / / 1*P(D)*P(C|D)*P(E’)
Path B, D, C’ / / 1*P(D)*P(C’|D)
Path B, D’, C, E / / 1*P(D’)*P(C|D’)*P(E)
Path B, D’, C, E’ / / 1*P(D’)*P(C|D’)*P(E’)
Path B, D’, C’ / / 1*P(D’)*P(C’|D’)
- Determine the overall mean lifespan by aggregating the paths to generation probabilistically. Estimate the variance of the lifespan distribution from literature.
The generation model only incorporates a single mean path length, and so in Equation 1, the overall weighted mean of lifespan for all six paths is presented.
Equation 3: Overall weighted mean of lifespan for all six paths
1.3 MarketShares and Unit Weight
In terms of the statistics data released by ZDC(2014), the market shares (historical data, most from 2004 to 2013) divided by ‘sizes’ can be available. There is only possibility to collect the unit weight all electric and electronics in the year of 2014 based on product’s specification introduction from ZDC and PCONLINE (2014), we therefore assumed the unit weight data historically keep consistent but keeps various if divided by ‘sizes’.Sincethe unit weight of mobile phone keeps decreasing, we used the data from other literature (USEPA, 2011). In addition, the unit weight of desktop always keeps consistent (Mean: 10.694 kg, STD: 2.382kg) because we assume there is not ‘size’ difference. The market shares data are all shown in figures S4 and S5. The unit weight data are shown in tableS5-S13. The unit weight data were allowed to vary uniformly one to two times standard deviation from the mean (uniform distribution).
TableS5 Unit Weight for LCD Monitor
LCD Monitor / >=27 / 24 / 23.6 / 22 / 21.5 / 20 / 19 / <=17Mean / 6.0 / 4.8 / 4.1 / 4.1 / 3.0 / 2.8 / 2.6 / 2.5
Std / 1.3 / 1.0 / 0.7 / 0.9 / 0.5 / 0.5 / 0.4 / 1.0
Table S6 Unit Weight for LCD TVs
LCD TVs / >=55 / 52 / 47 / 46 / 43 / 42 / 40 / 37 / 32 / <30Mean / 23.4 / 24.5 / 18.6 / 17.3 / 14.8 / 14.7 / 11.5 / 11.9 / 7.8 / 4.4
Std / 4.4 / 3.7 / 2.4 / 2.9 / 1.6 / 2.4 / 3.0 / 2.8 / 2.0 / 0.6
Table S7 Unit Weight for PDP TVs
PDP TVs / <=42 / 46 / 50 / 55 / 60 / >=65Mean / 21.3 / 29.9 / 41.3 / 47.6 / 76.7
Std / 3.3 / 5.9 / 15.0 / 17.8 / 20.3
Table S8 Unit Weight for Laptop
10.1 & 11.6 / 12.5 / 13.3 / 14 / 15.6 / >=17.3Mean / 1.3 / 1.4 / 1.5 / 2.0 / 2.3 / 3.8
Std / 0.4 / 0.1 / 0.1 / 0.3 / 0.3 / 1.0
Table S9 Unit Weight for CRT Monitor
CRT Monitor / <=14 / 15 / 17 / 19 / 21 / 22Mean / 12.7 / 14.0 / 15.9 / 21.1 / 25.3 / 29.6
Std / 2.3 / 3.4 / 2.7 / 2.6 / 7.4 / 5.9
Table S10 Unit Weight for CRT Color TVs
CRT Color TVs / <=12 / 14 / 17 / 21 / 25 / 29 / >=32Mean / 7.2 / 10.8 / 16.0 / 21.4 / 30.0 / 40.0 / 65.0
Std / 1.4 / 1.6 / 3.2 / 1.3 / 6.0 / 8.0 / 13.0
Table S11 Unit Weight for Refrigerator
Ref / >=280 / 230-280 / 200-230 / 180-200 / 100-180 / <=100Mean / 110.2 / 76.8 / 65.7 / 54.3 / 33.1 / 110.2
Std / 24.2 / 10.8 / 7.3 / 6.2 / 11.5 / 24.2
Table S12 Unit Weight for Washing Machine
Tumbling-box* / <5kg / 5kg / 6kg / 7kg / >7kgMean / 45.9 / 60.1 / 63.8 / 72.7 / 77.4
Std / 20.9 / 9.4 / 6.7 / 10.7 / 13.4
Impeller type / <5kg / 5kg / 6kg / 7kg / >7kg
Mean / 21.8 / 27.4 / 31.7 / 34.3 / 38.7
Std / 9.7 / 2.6 / 2.5 / 3.5 / 10.9
*, There is statistic data on the market shares based on the types of washing machine.
Table S13 Unit Weight for Air Conditioner
AC / 1 / 1.5 / 2 / 2.5 / 3Mean / 41.2 / 44.0 / 85.5 / 108.3
Std / 3.3 / 3.3 / 9.6 / 18.1
Fig.S4 Market Shares of Monitors, Laptops and TVs
Fig. S5 Market Shares of LHAs
1.4 Material Composition of Products and the Content of Selected Metals
Tables S14 shows the median values of the material composition of products and Table 15 and 16 the content of selected common metals, precious metals, and less common metals in printed circuit boards, CRT glass, Li-ion battery and flat panel screen for 11 types of electrics and electronics.
1
Table S14 Material Composition of 11 Types of End-of-life Electric and Electronics(Oguchi et al., 2011)*
Equipment type / LCD Monitor / LCD TVs / PDP TVs / Laptop / Desktop / CRT Monitor / CRT Color TVs / Mobile Phone / Refrigerator / Air Conditioner / Washing MachineNumber of data / 192 / 66 / 5 / 130 / 7 / 15 / 15 / 16 / 2 / 2 / 3
Ferrous material (%) / 43.2 / 50.9 / 28.5 / 23.2 / 42.1 / 12.7 / 12.7 / 0.8 / 47.6 / 45.9 / 51.7
Aluminum material (%) / 8.9 / 3.1 / 16.4 / 3.1 / 4.0 / 0.1 / 0.1 / – / 1.3 / 9.3 / 2.0
Copper cable and material (%) / 1.5 / 1.3 / 1.6 / 1.0 / 1.0 / 3.9 / 3.9 / 0.3 / 3.4 / 17.8 / 3.1
Plastic (%) / 29.3 / 28.3 / 7.7 / 28.9 / 20.4 / 17.9 / 17.9 / 37.6 / 43.7 / 17.7 / 35.3
Printed circuit board (%) / 6.6 / 7.1 / 7.9 / 9.4 / 7.7 / 8.7 / 8.7 / 30.3 / 0.5 / 2.7 / 1.7
CRT glass (%) / Panel glass / 22.9 / 22.9 / – / – / – / –
Funnel glass / 12.9 / 12.9 / – / – / – / –
Glass (%) / 9.0 / 28 / 13.7
Battery (%) / 15.7 / – / – / 20.4 / – / – / –
Unidentifiedmaterial (%) / 10.5 / 0.3 / 9.9 / 5.1 / 24.9 / 20.9 / 20.9 / 10.6 / 3.5 / 6.6 / 6.2
*, This table is the updated version of Oguchi’s work (not only cited theprevious work, includingtheir own), which includes the analysis from more other work: USEPA, 2008; Buzatu & Milea, 2008;Cryan, et al., 2010; Salhofer et al., 2011; Boni & Widmer, 2011;Petters et al., 2012;Fan et al., 2013.
Table S15 Metals Composition of Printed Circuit Board of 11 Types of End-of-life Electric and Electronics (mg/kg)(Oguchi et al, 2011)*
Equipment type / Number of data / Common metal / Precious metal / Less common metalAl / Cu / Fe / Pb / Sn / Zn / Ag / Au / Pd / Ba / Bi / Co / Ga / Sr / Ta
LCD Monitor / 1 / 63,000 / 180,000 / 49,000 / 17,000 / 29,000 / 20,000 / 600 / 200 / – / 3000 / – / – / – / 300 / –
LCD TVs / 1 / 63,000 / 180,000 / 49,000 / 17,000 / 29,000 / 20,000 / 600 / 200 / – / 3000 / – / – / – / 300 / –
PDP TVs / 2 / 38,000 / 210,000 / 20,000 / 7100 / 15,000 / 12,000 / 400 / 300 / – / 3900 / 100 / – / – / 650 / 100
Laptop / 2 / 18,000 / 190,000 / 37,000 / 9800 / 16,000 / 16,000 / 1100 / 630 / 200 / 5600 / 120 / 80 / 10 / 380 / 5800
Desktop / 8 / 18,000 / 200,000 / 13,000 / 23,000 / 18,000 / 2700 / 570 / 240 / 150 / 1900 / 50 / 48 / 11 / 380 / 7
CRT Monitor / 5 / 62,000 / 72,000 / 34,000 / 14,000 / 18,000 / 5300 / 120 / 5 / 20 / 2400 / 280 / 36 / – / 550 / –
CRT Color TVs / 5 / 62,000 / 72,000 / 34,000 / 14,000 / 18,000 / 5300 / 120 / 5 / 20 / 2400 / 280 / 36 / – / 550 / –
Mobile Phone / 19 / 15,000 / 330,000 / 18,000 / 13,000 / 35,000 / 5000 / 3800 / 1500 / 300 / 19,000 / 440 / 280 / 140 / 430 / 2600
Refrigerator / 1 / 16,000 / 170,000 / 21,000 / 21,000 / 83,000 / 17,000 / 42 / 44 / – / 82 / 480 / 120 / – / 51 / –
Air Conditioner / 1 / 6900 / 75,000 / 20,000 / 5800 / 19,000 / 4900 / 58 / 15 / – / 320 / – / 29 / – / 26 / –
Washing Machine / 1 / 1000 / 70,000 / 95,000 / 2200 / 9100 / 2400 / 51 / 17 / – / 65 / 51 / 16 / – / 9 / –
*, This table is the original version of Oguchi’s work (not only cited theprevious work, includingtheir own).
Table S16 Metals Composition of CRT Glass, Li-ion Battery and flat panel screen (Oguchi et al., 2011)*
Glass types / Common metal / Less common metalAl / Fe / Pb / Zn / Cu / Co / Ba / Sr / In
CRT panel glass (mg/kg) / 14,000 / 1,100 / 140 / 3,400 / 78,667 / 74,333
CRT funnel glass (mg/kg) / 19,000 / 805 / 216,250 / 1,416 / 5,333 / 5,600
Li-ion Battery (Laptop and Mobile phone) (mg/kg) / 69,750 / 222,125 / 96,500 / 167,250
In content in flat panel screen (mg/cm2)# / 356
*, This table (except Indium content value) is the simplified version of Oguchi’s work (not only cited theprevious work, includingtheir own). #, an updated version of Böni and Widmer’ s work (Boni & Widmer, 2011)(not only cited theprevious work, includingtheir own), which inlcudes the work by Gotze and Rotter,2012; Buchert et al., 2013.
1
1.5 Prediction of E-Waste and Scrap Metals Generation
The quantity of e-waste generated in year y is based on the sales in year s and the probability that a product sold in year s is generated in year y. The probability distribution is created using parameters from the lifespan estimates. Here, a lognormal distribution was assumed. Equation 4shows the how the quantity is calculated. The materials composition and metals contents are further estimated when the mass fractions and metals contents(indicated in section 1.4) are multiplied, see Equaiton 5and 6. Here, is fraction of each mateiral, is the type of materials; is conent of each mateiral, and is the type of metals
Equation 4: Quantity of e-waste generated in year y
Equation 5and 6: Quantity of materials and scrap metal generated in year y
1.6 Data and Intermediate Results
1.6.1 Life span
Ideally, lifespan stage assumptions would be disaggregated by electric and electronic type, owner type, and purchase year and distinguishing first use, reuse, and storage. However, lifespans were modeled separately for the following types: TVs and monitors; laptop, desktop, mobile phone and LHAs. The relevant estimates for each electric and electronic type from Table S17 were included in the development of lifespan stage length estimates.
TableS17 Modeled Lifespan Stage Lengths (Years)
B. Initial Use / D. Initial Storage / C. Reuse / E. Reuse StorageTVs / Huang et al., 2006 / µ / 6.94 / 3.47 / 3.47 / 1.74
σ / 2.31 / 1.16 / 1.16 / 0.58
Laptop / Zheng, 2009 / µ / 4.31 / 2.16 / 2.16 / 1.08
σ / 1.67 / 0.84 / 0.84 / 0.42
Desktop / Zheng, 2009 / µ / 5.63 / 2.82 / 2.82 / 1.41
σ / 1.55 / 0.78 / 0.78 / 0.39
Mobile phone / Huang et al, 2006
Yin et al., 2014 / µ / 2.78 / 1.39 / 1.39 / 0.70
σ / 1.45 / 0.73 / 0.73 / 0.36
Refrigerator / Huang et al., 2006 / µ / 7.42 / 3.71 / 3.71 / 1.86
σ / 4.06 / 2.03 / 2.03 / 1.02
Air Conditioner / Huang et al., 2006 / µ / 5.98 / 2.99 / 2.99 / 1.50
σ / 3.21 / 1.61 / 1.61 / 0.80
Washing Machine* / µ / 5.98 / 2.99 / 2.99 / 1.50
σ / 3.21 / 1.61 / 1.61 / 0.80
*, Assumed to the same to Air conditioner.
1.6.2 Probability of Paths Leading to Generation
With a goal of modeling generation in two decades (2005 to 2025), the analysis included lifespan stage estimates from twenty one years prior in 1989, which allows for a generous total lifespan of electics (such as TVs) purchased in 1989. Because most of the data sources only reported the total life span and do not differentiate the electric and electronic type, lifespan stage estimates for each type were only included the survey data provided by Huang, et al., 2006 and Yin, et al., 2014. The mean µ and standard deviation σ for each lifespan stage for each type are shown in Table S18.
Table S18 Probability of Paths Leading to Generation
Types / Source# / Storage rate / Reuse rate / Reuse Storage rate* / Collected processing rate+ / Reuse rate after processing+P(D) / P(D’) / P(C) / P(C’) / P(E) / P(E’) / P(F) / P(F’) / P(H) / P(H’)
TVs & Monitor / Mean / (1)(2) / 35% / 65% / 52% / 48% / 9% / 91% / 90% / 10% / 10% / 90%
Std / 4% / 8% / 1% / 9% / 1%
Laptop / Mean / (1)(3) / 22% / 78% / 46% / 54% / 6% / 94% / 90% / 10% / 10% / 90%
Std / 9% / 5% / 2% / 9% / 1%
Desktop / Mean / (1) / 47% / 53% / 46% / 54% / 12% / 88% / 90% / 10% / 10% / 90%
Std / 7% / 7% / 2% / 9% / 1%
Mobile phone / Mean / (1)(2)(4) / 35% / 65% / 55% / 45% / 9% / 91% / 90% / 10% / 10% / 90%
Std / 5% / 27% / 1% / 9% / 1%
Refri-
gerator / Mean / (1)(2) / 22% / 78% / 55% / 45% / 5% / 95% / 90% / 10% / 10% / 90%
Std / 6% / 3% / 0% / 9% / 1%
Air Con-
ditioner / Mean / (1)(2) / 15% / 85% / 50% / 50% / 4% / 96% / 90% / 10% / 10% / 90%
Std / 5% / 8% / 1% / 9% / 1%
Washing Machine / Mean / (1) / 27% / 73% / 56% / 44% / 7% / 93% / 90% / 10% / 10% / 90%
Std / 4% / 8% / 1% / 9% / 1%
*, Reuse Storage rate is assumed half of Initial Storage rate.#; (1) Streicher et al., 2011; (2) Huang et al., 2006; (3) Song et al., 2012; (4) Li et al., 2012.&, The paths data for desktop with projection were allowed to vary uniformly one standard deviation from the mean (uniform distribution), by given an approximate 15% of Correlation of Variances (COV). +, We assumed roughly 90% of collected after processing rate and 10% of Reuse rate after processing in this study based on interview with recyclers and experts.
1.6.3 Results
(1) Metals used contented in PCBs, CRT glass, flat panel screen and Li-ion battery (manufacturing process)
Fig.S6 Metals used contented in PCBs, CRT glass, flat panel screen and Li-ion battery (manufacturing process)
(2) E-waste generation and uncertain capture.
Fig.S7 E-waste generation and uncertain capture
Fig.S8 Scrap Metals generation and uncertain capture.
Appendix
Abbreviation / Full titleMIIT / Ministry of Industry and Information Technology
NBSC / National Bureau of Statistics of China
NGOs / Non-Governmental Organizations
BAN / Basel Action Network
UNEP / United Nations Environment Programme
ITU / International Telecommunication Union
EPA / Environmental Protection Agency
EPB / Environmental Protection Bureaus
ZDC / Zhongguancun Data Center
WHO-TEQ / World Health Organization toxic equivalent (WHO-TEQ)
SOM / Sales Obsolescence Model
WEEE / Waste Electrical and Electronic Equipment Directive
EoL / End-Of-Life
Ee-waste / Electronic Waste
BFR / Brominated Flame Retardants
FPDs / Flat Panel Displays
NMMs / Nonmetallic Materials
LCDs / Liquid Crystal Displays
PCBs / Printed Circuit Boards
CRT / Cathode Ray Tube
LHAs / Large Home Appliances
TBBPA / Tetrabromobisphenol A
PBDEs / Polybrominated Diphenyl Ethers
PCDDs, PCDFs / Polychlorinated dibenzo-p-dioxins (PCDDs), Polychlorinated dibenzofurans (PCDFs)
STD / Standard Deviation
COV / Correlation of Variances
References
Böni H, Widmer R. Disposal of Flat Panel Display Monitors in Switzerland, Final Report. EMPA, SWICO Recycling, St. Gallen, Switzerland; 2011.
Buchert M, Manhart A, Bleher D, Pingel D. Recycling critical raw materials from waste electronic equipment. Freiburg: Öko-Institut eV; 2012.
Buzatu M, Milea NB. Recycling the liquid crystal displays. UPB Sci Bull, Series B 2008, 70 (4): 93-102
CryanJ, Freegard K, Morrish L, Myles N. Final report on the demonstration trials into Flat Panel Display recycling technologies; WRAP and Axion Consulting; 2010.
Duan, H.; Miller, T.R.; Gregory, J.; Kirchain, R. Quantitative Characterization of Transboundary Flows of Used Electronics: Analysis of Generation, Collection, and Export in the United States. December, 2013. Materials Systems Laboratory, MIT.
Fan S, Fan C, Yang J, Liu K. Disassembly and recycling cost analysis of waste notebook and the efficiency improvement by re-design process. J Clean Prod 2013,39(0): 209-219.
Gotze R, Rotter VS. In Challenges for the recovery of critical metals from waste electronic equipment - A case study of indium in LCD panels, Electronics Goes Green 2012+ (EGG), 9-12 Sept. 2012; pp 1-8.