Supply and demand in the material recovery system for cathode ray tube glass

Transcription

1 Supply ad demad i the material recovery system for cathode ray tube glass The MIT Faculty has made this article opely available. Please share how this access beefits you. Your story matters. Citatio As Published Publisher Gregory, J., M.-C. Nadeau, ad R. Kirchai. Supply ad demad i the material recovery system for cathode ray tube glass. Sustaiable Systems ad Techology, ISSST '09. IEEE Iteratioal Symposium o Istitute of Electrical ad Electroics Egieers Istitute of Electrical ad Electroics Egieers Versio Fial published versio Accessed Sat Apr 28 15:12:35 EDT 2018 Citable Lik Terms of Use Detailed Terms Article is made available i accordace with the publisher's policy ad may be subject to US copyright law. Please refer to the publisher's site for terms of use.

2 Supply ad Demad i the Material Recovery System for Cathode Ray Tube Glass Jeremy Gregory, Marie-Claude Nadeau, ad Radolph Kirchai Abstract This paper presets a aalysis of the material recovery system for leaded glass from cathode ray tubes (CRTs). I particular, the global mass flow of primary ad secodary CRT glass ad the theoretical capacities for usig secodary CRT glass to make ew CRT glass are aalyzed. The global mass flow aalysis idicates that the amout of ew glass required is decreasig, but is much greater tha the amout of secodary glass collected, which is icreasig. The compariso of the ratio of secodary glass collected to the amout of ew glass required from the mass flow aalysis idicates that the material recovery system is sustaiable for the foreseeable future. However, a predictio of the time at which the market for secodary glass will collapse due to excess capacity is ot possible at the momet due to several sources of ucertaity. Idex Terms Cathode ray tube, e-waste, glass, recyclig. D I. INTRODUCTION EVICES cotaiig cathode ray tubes (CRTs), such as computer moitors or televisios (TVs), represet a sigificat ad challegig fractio of the electroics waste stream. I Europe, where regulated e-waste icludes early every product with a cord or battery, CRT devices represeted approximately 22% by weight of e-waste geerated i 2005 [1]. I the US, where regulated e-waste typically icludes oly computig equipmet ad TVs, this figure climbs to 58% [2]. The leaded glass i the CRT is the cause of several EoL treatmet challeges. Furthermore, both ecoomic ad techological barriers exist to reuse the glass from EoL CRTs i other applicatios (icludig ew CRTs). Fortuately, two viable uses for EoL CRT glass (also kow as secodary CRT glass or CRT glass cullet) exist: raw material for the productio of ew CRTs ad fluxig agets i smelters [2]. Of these two, the evirometally preferred ad historically predomiat sik has bee the former. There is a setimet i the developed world that the CRT is a dyig techology [3]. Ideed, CRT sales for moitors ad televisios are markedly dow i the US [3]. Although Jeremy Gregory is with the Materials Systems Laboratory ad the MIT Eergy Iitiative, Massachusetts Istitute of Techology, Cambridge, MA USA (phoe: ; jgregory@mit.edu). Marie-Claude Nadeau is with the Materials Systems Laboratory ad the Egieerig Systems Divisio, Massachusetts Istitute of Techology ( adeau@mit.edu). Radolph Kirchai is with the Materials Systems Laboratory ad the Departmet of Materials Sciece ad Egieerig ad the Egieerig Systems Divisio, Massachusetts Istitute of Techology ( kirchai@mit.edu). iexpesive CRTs remai popular i the developig world [4], it is clear that, eve for those markets, CRTs will evetually be supplated. Iheretly, a shift to ew display techologies will also drive a icrease i CRT device retiremet. These coupled treds directly call ito questio the ogoig viability of the evirometally preferred material recovery pathway: the reuse of CRT cullet ito ew CRTs. Will the supply of EOL cullet outstrip the capacity of ew productio? Iformatio about the timig ad extet of this oversupply should help stakeholders idetify the eed for developig or facilitatig other siks for CRT cullet. The issue has particular urgecy due to the rapidly chagig ature of the CRT market ad the emergece of govermet-madated e-waste recyclig systems. To explore this issue, this paper develops ad exercises a dyamic materials flow aalysis model of the global mass flows of primary ad secodary CRT glass. Specifically, by simultaeously comprehedig global treds i sales, retiremet, ad productio techologies for CRTs, the model projects the relative supply of ad demad for CRT glass cullet. This model is used to compare treds i supply ad demad ad, ultimately, estimate the time util the supply of CRT glass cullet exceeds the demad. The results of the work provide isight o the factors that will have the greatest impact o the ecoomic viability of the CRT glass recovery system. Furthermore, the methodology is a broader example of applyig material flow aalysis to idetify market vulerabilities i a material system. The case aalysis is relatively uusual i that it evaluates a material system with rapidly decliig cosumptio. Thus, the methodology may be of particular value i the aalysis of material systems with similar characteristics. From a case perspective, several studies have estimated the amout of ed-of-life CRT devices geerated regioally i the US [2, 5], Europe [1], ad South Africa [6]. I additio, Lito has used forecast the amout of CRT waste that will be collected i the US for the ext fifty years [7]. Fially, Weitzma correctly idetified that there would be decreasig demad for CRT cullet i the US because US maufacturers were producig a smaller share of the CRT market (assumig that CRT cullet geerated i the US would oly be used i CRT glass maufacturig i the US) [8]. Although each of these provides importat isights ito the problem, o sigle study is able to aswer the questios posed i that the examied scope was limited temporally, geographically (oly

3 a sigle regio), ad/or sectorally (i.e., demad or supply). The research preseted i this paper addresses these gaps for this case. Figure 1. CRT glass material system flow. N s= t Yt () = ( Wξ, c() s Sc() s λc(,) st) ds C() t ξ { FP, } = 1 s= c { TM, } II. MODELS OF SUPPLY AND DEMAND IN THE SECONDARY CRT GLASS MATERIAL SYSTEM The multiple potetial pathways of material flow for CRT glass throughout its life are depicted i Figure 1. The focus of this study is o CRT glass cullet that is used to produce ew CRTs, as highlighted i the lower portio of the figure. I particular, the aalysis focuses o the supply of CRT glass cullet, represeted by the arrow from Ed-of-Life to Recyclig, ad the demad for CRT glass cullet, represeted by the arrow from Recyclig to Productio of New CRTs. The use of CRT glass cullet i the productio of ew CRTs is the focus of this aalysis because it is the ecoomically preferable alterative (the oly CRT cullet destiatio that is reveue-geeratig for the CRT recycler), the more commo alterative (approximately 75% of CRT cullet i the US is estimated to be set to CRT glass maufacturers [2]), ad the oly closed-loop alterative. The iputs ad outputs for the supply ad demad models are show i Figure 2. Specifically, the supply model combies four elemets to estimate the aual amout of CRT glass cullet collected worldwide: historical sales, glass weights per product, product lifespas, ad EoL collectio fractios. Flows, Y, for each year, t, are separately tracked for two product types CRT TVs (T) ad moitors (M), idexed o c {T,M}, ad i terms of fuel cullet (F) ad pael cullet(p), idexed o ξ {F,P}. Additioally, cullet geeratio is modeled for a set of N regios, idexed o. CRT glass cullet collectio amouts are calculated usig the followig relatioship (variables are defied i Figure 2): (1) The cullet geerated i regio i year t, Yξ () t, is the sum of the weight of glass i a uit product i the year the product was sold (idexed o s), W ξ,c (s), multiplied times the umber of products sold i regio i year s, Sc () s, multiplied times the probability that a product sold i year s reached the ed of its life i year t, λ c (s,t), itegrated over all sales years prior to t i oe-year icremets. The cullet collected i a regio is the amout geerated multiplied times the fractio of EoL products collected i a regio i a give year, C () t. The total supply of cullet for a give year t, Y(t), is the sum of the pael ad fuel cullet collected i all regios. The demad model combies two elemets to estimate the aual capacity for usig CRT glass cullet i the productio of ew CRT glass: forecast sales ad the percetage of ew CRT glass that may be made from cullet. Potetial cullet cosumptio is calculated usig the followig relatioship: N D() t = Wξ, c() t Sc () t F () t ξ ξ { FP, } = 1 c { TM, } (2) Demad for pael or fuel cullet i a give year t, D ξ (t), is calculated by multiplyig the weight of glass (fuel or pael) sold worldwide i t times the fractio of ew glass that may be made from cullet, F ξ (t). The weight of glass sold worldwide is the weight of glass i a uit product sold i the year t, W ξ,c (t), multiplied times the umber of products sold i regio i t, S () t, summed over all N regios. The total demad for cullet c for a give year t, D(t), is the sum of the demad for the pael ad fuel cullet.

4 Figure 2. Supply ad demad models iputs ad outputs. Withi the fuctios, represets regio, t represets the year of the aalysis, ad s represets the year i which products are sold. The supply ad demad of CRT glass cullet ca be compared over time to determie relative treds ad, i particular, estimate whe supply will exceed demad. It is importat to ote that demad represets the amout of cullet that could be used i the productio of ew CRT glass. Actual cosumptio of CRT glass cullet may be limited by supply availability. However, the ecoomic viability of the system will be i jeopardy whe the supply exceeds the demad because of the limited-value alterative applicatios for CRT glass cullet. III. DATA SOURCES FOR MODEL INPUTS The complete details o data sources ad methods for model iputs are preseted i [9]; a overview of the methodology is preseted here. The temporal spa of the aalysis ad data was dictated by the eed to forecast treds sufficietly ito the future to capture potetial poits where supply would exceed demad. Forecastig treds up to the year 2025 was estimated to capture such crossover poits. Historical product sales data were required as far back 1990 i order to estimate the amout of e-waste that could be collected. For the two model parameters that are a fuctio of locatio sales ad EoL collectio fractio the aalysis is broke ito four world regios: North America; Europe, the Middle East, ad Africa (EMEA); Lati America; ad Asia (icludig Australia). These regios were selected based o the resolutio of available sales data. Aual sales data for CRT TVs ad moitors were estimated usig various sources [2, 5, 10-13] because o sigle source of publicly available data existed that was comprehesive. Gaps i historical data were estimated usig iterpolatio or mappig from treds i other regios where data were available. Gaps i forecasts were estimated usig media reports of market forecasts [e.g., 4] ad extrapolatio. No iformatio was available o sales umbers beyod 2011 due to the iheret challeges of forecastig. Although there is widespread agreemet that there will be a sigificat decrease i worldwide CRT sales i the ext decade, the timig ad ature of the sales peak are ukow. Give this ucertaity, three scearios for TV sales i Asia (the domiat factor i future sales) were selected to use i sesitivity aalyses: high, medium (baselie), ad low sales. The icreasig demad for larger TVs ad moitors has meat that the average weight of glass i these devices has icreased over time. A sales-weighted average fuel ad pael glass weight for a uit TV ad moitor was determied from [5] across the aalysis period. Studies o the age of ewaste retured for recyclig have idicated that there is a wide distributio i the product lifespa [2]. I this study, TV ad Moitor lifespa characteristics λ c (s,t) were derived from [2] by covolvig the iformatio from that report weighted i accordace with the sales fractios of specific product types (small ad large TVs ad residetial ad commercial sales). The result is probability distributios for TVs ad moitors that a product sold i year s

5 reached the ed of its life i year t. Oce a product is ready for EoL maagemet, it ca either be ladfilled or recycled, as depicted i Figure 1. Data o EoL collectio rates (the fractio of EoL products geerated that are collected for recyclig) are extremely limited. Recet studies from the Europea Uio [1] ad the Uited States [2], have estimated CRT collectio rates of approximately 30% ad 15-20%, respectively. To the authors kowledge there are o kow estimates of how these rates vary over time or i other regios of the world. EoL collectio fractios i this aalysis were assumed to follow a S-curve behavior over the period of the aalysis: collectio rates begi at a low plateau, the icrease rapidly util reachig a upper plateau. Several parameters defie the shape ad values of the fuctio icludig a referece collectio fractio value at a referece year, a upper limit, ad a shape parameter. Sice Europe has relatively high collectio rates ad a sizable populatio, a collectio rate S-curve was defied for Europe that could act as a referece for other regios. The referece poit was based o the aforemetioed EU study [1] (30% i 2006), but the upper limit (75%) ad the time delay shape parameter were estimated. Sesitivity aalyses will explore the impact of the value of the shape parameter. Curves for other locatios were assumed to have the same limits, referece poit, ad shape as the Europea curve, but are offset by a time delay shift parameter. Estimated shifts rage from 3 years for North America to 14 years for Lati America. I additio to sales forecasts, the demad model requires data o the capacity for usig CRT glass cullet i the productio of ew CRT glass. Ufortuately, there is limited iformatio i the literature o such capacities. Oe study from 2002 estimated this capacity to be 20% for pael glass ad 40% for fuel glass [10]. To update those figures, the authors surveyed curret CRT glass maufacturers. Collectig such data proved to be difficult for several reasos (e.g., they are several levels removed from moitor ad TV maufacturers i the product supply chais, laguage ad cultural barriers, uwilligess to share data), but two maufacturers shared data o their cullet capacities ad use. Their capacities were sigificatly higher tha the aforemetioed values, but they could ot collect eough cullet to meet their capacities. Additioal discussios with stakeholders i the primary ad secodary CRT glass idustry idicated that ot all CRT glass maufacturers are usig cullet i maufacturig. I light of these differeces, a costat value of 50% was used as a baselie fractioal capacity for the use of CRT fuel ad pael glass cullet i the model. Although the maufacturers who shared data o cullet use provided capacities higher tha 50%, this value attempts to represet a average for the etire idustry eve those ot usig cullet. Sesitivity aalyses will be coducted o these assumptios. glass cullet geerated ad the amout of CRT glass collected i each regio as it evolves over time. Figure 3 shows these amouts for the four regios i 2010 ad 2020 (weights are i metric tos, as i all plots). The projected amouts geerated i North America ad EMEA decrease over time as sales decrease, whereas the projected amout geerated i Asia icreases by almost a factor of two. Furthermore, the modeled collectio fractio is almost oexistet i Lati America ad Asia i 2010, but icreases sigificatly by Figure 3. Weight of CRT glass cullet geerated ad collected i the four regios i a) 2010 ad b) Predicted supply ad demad behavior for CRT cullet usig baselie iputs i the models are plotted util the year 2025 i Figure 4. As expected, cullet demad decreases over time as product sales decrease ad cullet supply icreases as more products reach EoL. The itersectio of the curves, whe supply begis to exceed demad, occurs i approximately 2014 for the baselie coditios. IV. ANALYSIS AND DISCUSSION It is iterestig to compare the modeled weight of CRT Figure 4. Supply ad demad plots of CRT glass cullet usig baselie iputs.

6 Give the ucertaity i several iputs, it is importat to examie the sesitivity of the models to these key assumptios. The areas of highest ucertaity iclude sales forecasts, collectio rates, ad the capacity for all CRT glass maufacturers to use CRT cullet. The impact of variatio i these iputs o the year i which supply ad demad itersect is depicted i Figure 5 to Figure 7. With regard to future sales, TV sales will clearly domiate moitor sales ad Asia will represet the largest market for TV sales. Thus, three future Asia TV sales scearios were used i the model. Figure 5 shows that the supply ad demad itersectio year calculated by the model is moderately sesitive to the future TV sales sceario, icreasig by two years for the high sceario ad decreasig by two years for the low sceario. The itersectio year is also oly moderately sesitive to variatio i the collectio time delay shape parameter (which determies the shape of the collectio fractio S-curves), as depicted i Figure 5. Figure 6. The impact of variatio i t D, the collectio delay with respect to Europe, o supply ad demad itersectio year. t D=10 is the baselie for Asia Pacific; Chia is always 1.4 times Asia Pacific s delay. Figure 5. The impact of variatio i future Asia TV sales sceario ad the collectio time delay shape parameter, o supply ad demad itersectio year. Baselie parameters result i the itersectio year of Aother importat parameter that defies collectio rates as a fuctio of time is the time delay shift of the collectio fractio S-curves with respect to Europe. Variatio i the shift parameter for Asia Pacific ad Chia, which collectively are Asia, the regio that will be geeratig the most CRT cullet, ca have a sigificat impact o itersectio year, as depicted i Figure 6. If Asia has a short lag behid Europe i its collectio efforts (five to te years), supply will itersect demad i a few years. However, if the lag is sigificat (o the order of 15 years), the the itersectio poit could exted to The capacity for all CRT glass maufacturers to use CRT cullet also has a sigificat impact o the itersectio year, which is illustrated i Figure 7. It is importat to remember that the capacity for cullet us is a idustry-wide parameter; it could iclude some maufacturers who are usig o cullet ad others who use cullet extesively. If most CRT glass maufacturers are usig cullet i high fractios, the the itersectio will be later tha the baselie predictio of However, low usage of cullet meas the itersectio could occur quite soo. Figure 7. The impact of variatio i the fractioal capacity for cullet use (pael ad fuel are equal). The baselie value is 50%. Predictig the exact year whe supply will exceed demad is ot possible, but these aalyses idicate that it could happe withi the ext te years. The itersectio year uder the baselie sceario was predicted to be 2014, but this itersectio could be sooer if there are low TV sales, but particularly if there are high collectio rates i Asia ad if CRT glass maufacturers maitai low usage of CRT cullet. Notably, the supply ad demad curves depicted i Figure 4 shift depedig o the modeled assumptios (vertically or horizotally), but the shapes essetially remai the same. Thus, a shift i a curve chages the positio of the kee i each curve, which ca have a sigificat impact o the supply ad demad itersectio poit. There are two major implicatios of this research. First, more data eeds to be collected o the curret ad expected capacity of the etire CRT glass maufacturig idustry to use CRT cullet because this has a sigificat impact o the demad curve. Secod, more research is eed i the area of alterative value-drive applicatios for CRT cullet. Others have called for such research [11] ad while there has bee research ito alterative applicatios for CRT cullet [12, 13], the applicatios are geerally low or egative value for cullet ad are ulikely to meet the supply of CRT cullet geerated. The breakthrough that is eeded to create a viable valuedrive market for CRT cullet is the capability to extract lead ad other udesirable elemets from the cullet such that it ca be used i the same applicatios as other glass cullet, such as

7 architectural, automotive, ad packagig applicatios. Although techologies exist to remove lead from CRT cullet [11], the cullet is still oly accepted i low-value applicatios. Thus, research efforts should focus o trasformig CRT cullet ito a secodary commodity that is valuable. ACKNOWLEDGMENT We gratefully ackowledge cotributios from the followig people who provided cotacts, iformatio, ad data: Jay Celorie, Ha Che, Joh Dickeso, Eckart Dörig, Silvio Fachim, James Garder, Jaco Huisma, Robi Igethro, Joseph Nardoe, Gregory Sampso, ad Dai Tsuda. REFERENCES [1] J. Huisma, et al., "2008 review of directive 2002/96 o waste electrical ad electroic equipmet (WEEE): Fial report," Uited Natios Uiversity, [2] Uited States Evirometal Protectio Agecy, "Maagemet of electroic waste i the Uited States: Approach 1," [3] Aoymous, "CRTs are fadig," E-Scrap News, vol. 6, o. 9, pp. 4, September [4] Aoymous. (2006). CRT still rules TV market, isuppli says. Electroic News. Available: arter=eews. [5] A. Mochamp, et al., "Cathode ray tube maufacturig ad recyclig: Aalysis of idustry survey," Electroic Idustries Alliace, Arligto, VA, [6] D. Zumbuehl, "Mass flow assessmet (MFA) ad assessmet of recyclig strategies for cathode ray tubes (CRTs) for the Cape Metropolita Area (CMA), South Africa," Thesis, EMPA, [7] J.D. Lito, J.S. Yeomas, ad R. Yoogaligam, "Eablig idustrial ecology through the forecastig of durable goods disposal: Televisios as a exemplar case study," Caadia Joural of Admiistrative Scieces-Revue Caadiee Des Scieces De L Admiistratio, vol. 21, o. 2, pp , Ju [8] D.H. Weitzma, "Is CRT glass-to-lead recyclig safe ad evirometally friedly?," i Proceedigs of the 2003 IEEE Iteratioal Symposium o Electroics ad the Eviromet, Bosto, MA, USA, 2003, pp [9] J. Gregory, M.-C. Nadeau, ad R. Kirchai, "Evaluatig the ecoomic viability of a material recovery system: The case of cathode ray tube glass," submitted to: Evirometal Sciece & Techology [10] Demaufacturig of Electroic Equipmet for Reuse ad Recyclig, "Fial iformatio report subtask gl-04: Electroics demaufacturig exploratio (CRT glass)," Washigto, DC, [11] C.S. Poo, "Maagemet of CRT glass from discarded computer moitors ad TV sets," Waste Maagemet, vol. 28, o. 9, pp , [12] F. Adreola, et al., "CRT glass state of the art: A case study: Recyclig i ceramic glazes," Joural of the Europea Ceramic Society, vol. 27, o. 2-3, pp , [13] D. Kim, M. Quila, ad T.F. Ye, "Ecapsulatio of lead from hazardous CRT glass wastes usig biopolymer cross-liked cocrete systems," Waste Maagemet, vol. 29, o. 1, pp , 2009.