Select Engineering Characteristics of Crushed Glass Engineering... · Select Engineering...
Embed Size (px)
Transcript of Select Engineering Characteristics of Crushed Glass Engineering... · Select Engineering...
hed glasse low6.8 and
measuredn the ordered
Select Engineering Characteristics of Crushed GlassJoseph Wartman, M.ASCE1; Dennis G. Grubb, M.ASCE2; and A. S. M. Nasim3
Abstract: Select engineering characteristics of crushed glass produced using two processing techniques(crushing versus screening) to anAmerican Society of State Highway and Transportation Officials No. 10 gradation were experimentally evaluated. The crussamples were classified as well graded sands with gravel(SW) and exhibited excellent strength and workability characteristics. Thspecific gravity (2.49) contributed to crushed glass having compacted maximum dry densities on the order of 16.6–117.5–18.3 kN/m3 by the standard and modified Proctor compaction tests, respectively. Direct shear friction angles werebetween 47 and 62° at normal stresses ranging from 0 to 200 kPa. Friction angles obtained by drained triaxial shear were oof 48° for similar stress ranges. Measured hydraulic conductivities were on the order of 1–6310−4 cm/s. The results indicate that crushglass is a readily available, freely draining, environmentally clean, relatively low cost material whose engineering performancegenerally equal or exceed those of most natural aggregates. Despite these favorable characteristics, there are many real abarriers to increasing the beneficial use of crushed glass, and key examples are provided in an effort to illustrate these ubarriers.
CE Database subject headings: Glass; Recycling; Material properties; Aggregates.
s. Ino varyssing. This
y 1,rs. Tod withittedon
In 1998, an estimated 11 million t of glass entered the wsteam but only 29% of this material was recycled(USEPA 2000).In other words, 7.8 million t went into landfills, seemingly defeing the purpose of curbside collection and the investmenenvironmental agencies to improve recycling on the collecend. While regulations, laws, and public pressure drive thelection of recyclables, the identification and development ofondary markets for glass, however, is often beyond the influof environmental agencies.
This paper focuses on the geotechnical characteristics antors limiting the beneficial use of two general classes of glassenter the waste stream and landfills: glass cullet and waste itrial glass. Glass cullet, at least in the State of Pennsylvandefined as “postconsumer material” comprised of the mixedored glass fragments resulting from the breakage of glass coers (predominantly food, juice, beer, and liquor bottles) that cannot be reused by bottle manufacturers. Occasionally, glasswill contain fragments of broken ceramics(coffee mugs, chinplates, pottery) though these are not viewed to compromiseoverall geotechnical performance of crushed glass. Waste i
1Assistant Professor, Dept. of Civil, Architectural and EnvironmeEngineering, Drexel Univ., 3141 Chestnut St., Philadelphia, PA 191
2Program Manager, Apex Environmental Inc., 269 Great VaPkwy., Malvern, PA 19355.
3Graduate Student Researcher, Dept. of Civil, Architectural,Environmental Engineering, Drexel Univ., 3141 ChestnutPhiladelphia, PA 19104.
Note. Associate Editor: Tinh Nguyen. Discussion open until Ma2005. Separate discussions must be submitted for individual papeextend the closing date by one month, a written request must be filethe ASCE Managing Editor. The manuscript for this paper was submfor review and possible publication on April 15, 2003; approvedJanuary 13, 2004. This paper is part of theJournal of Materials in CivilEngineering, Vol. 16, No. 6, December 1, 2004. ©ASCE, ISSN 081561/2004/6-526–539/$18.00.
526 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
trial glass includes such materials as broken, obsolete, and/ospecification glass from the manufacturing of plate, window,analytical glassware, etc., but is generally considered to exglass derived from automobiles, lead crystal, TV monitors, liing fixtures, and electronics applications due to their composand coatings. Waste industrial glass is often regulated as rewaste, whereas there are frequently no regulatory constplaced on glass cullet. Crushed glass comprised of glassand/or waste industrial glass is virtually identical in every respand both must be crushed to provide consistent products ormonly accepted gradations such as those recognized bAmerican Association for State and Highway Transportationficials (AASHTO) or other organizations.
Part of the motivation for undertaking this study is tcrushed glass could be quite valuable on the local scale beccontrolling cost of many civil engineering applications and cstruction materials is the transportation. Crushed glass passi9.5 mms3/8 in.d sieve closely resembles natural or quarriedgregates and does not retain the remnant shape of the ocontainer or application shape, unless one or more of the dsions is smaller than the sieve opening. Therefore, while cruglass is perceived to have many strength and drainage aptions(Table 1), its reuse potential is hindered by the fact that this limited knowledge concerning its engineering parameteraddition, the engineering parameters have been perceived ton a supplier by supplier basis depending on the proceequipment used to control the gradation of the crushed glasswork is an effort to bridge these information gaps.
Most of the previous research on construction applicationcrushed glass has pertained to its use in concrete or asphaltmaterials. Many of these studies have investigated the reabetween silica contained in the glass and alkalis in cement p[i.e., alkali–silica reaction or(ASR)], which has a deleterious ef-
rietys havelectglassa spepac-
m twowereesen-liers I
e ofy in-
fect on concrete quality(e.g., Schmidt and Saia 1963; Johns1974; Figg 1981; Polley et al. 1998; and Dyer and Dhir 20).Studies on glass-amended asphalt mixtures, sometimes t“glasphalt,” have typically focused on asphalt debonding“stripping” from the glass aggregate(e.g., Malisch et al. 1975Day and Schaffer 1989; Chesner 1992; FHWA 1998).
Several researchers have studied the engineering propersoil-crushed glass blends and/or unblended crushed glass. D& Moore (1993) performed a series of physical property, comption, and strength tests to study the effect of crushed glass coon the properties of soil-crushed glass blends. The crushedused in the study was field processed using conventional cruequipment to 6.3 and 19.5 mm minus gradations. The glassblended with two processed soils, consisting of either a 25.4minus gravely sand, or a 31.8 mm minus crushed rock. A vaof geotechnical tests were conducted on blended specimening crushed glass contents of 15 and 50%. Additionally, stests were performed on the unblended soil and crushedspecimens. The results indicated that the crushed glass hadcific gravity of about 2.5, standard and modified Proctor com
Table 1. Anticipated Uses of Crushed Glass
Structural applications Drainage applications
Base course Foundation drainage
Subbase Drainage blankets
Embankments French/interceptor drain
Structural fill Well packing media
Nonstructural fill Septic field media
Utility bedding and backfill Leachate collection med
Retaining wall backfill Sand filters(wastewater)Antiskid Material Soil vapor extraction med
Alternative daily landfill cover
Fig. 1. Crushed glass from Supplier I. Inset
JOURNAL OF MATERIALS IN
tion maximum densities of about 16.2 and 18.1 kN/m3, respectively, Los Angeles abrasion wear values in the range of 30–and direct shear internal friction values of about 51°. DameMoore (1993) concluded that in general, the addition of crusglass had no negative impact on the properties of the twoconsidered in the study. Shin and Sonntag(1994) summarized thDames & Moore study indicating that crushed glass is an elent supplement or replacement for natural aggregates inconstruction applications. Henry and Morin(1997) evaluated thfrost susceptibility of crushed glasss,,14 mmd and two soil-crushed glass blends usingASTM D5918(ASTM 1996a). Theyconcluded that crushed glass had a negligible to very lowsusceptibility, and that the addition of crushed glass to the sonot increase its frost susceptibility. Hagerty et al.(1993) studiedthe behavior of clean, laboratory processed crusheds,0.95 mmd, determining that highly angular materials suchcrushed glass experience more particle crushing underdimensional high-stress loads than similar, but less angularrials (e.g., glass beads).
Representative samples of crushed glass were collected frosuppliers. Samples were placed in plastic 210 L drums andsealed and shipped to Drexel Univ. Figs. 1 and 2 show reprtative “as-received” samples of the crushed glass from Suppand II, respectively.
Supplier I is a commercial quarry and aggregate suppliecated in southeastern Pennsylvania. A secondary businessinvolves the collection, stockpiling, processing, and resalglass materials from local recycling organizations and nearb
graph shows a closeup of the material(scale in cm)photo
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 527
lva-te; anwabled thecolor
es ofs areWAtural
dustrial, pharmaceutical, glass manufacturing companies.supplier’s main motivation for this secondary business is tha(1)it is increasingly difficult to obtain quarry expansion permits(2) the use of glass extends current aggregate reserves fhigher unit price opportunities. Conventional quarry operatare used to crush glass to an AASHTO No. 8 gradasø1.3 cmd or finer.
Supplier II is a recycling facility located in eastern Pennsynia that accepts, processes, and separates commingled wasresells the glass, plastics, paper, aluminum, and other renematerials. The commingled waste is manually presorted anlarger containers and fragments of glass are separated byRecycled, clear, green, and amber glass can be resold to bmanufacturers at $25–$50/ t where viable glass bottling maexist (Elliott 2001). Smaller, mixed color broken glass shardsfragments and broken ceramics(less than about 2 cm in lengt)cannot be efficiently color and/or type separated, precludingreuse in container manufacturing. These materials accumullarge stockpiles with a landfill disposal liability of up to $5unless markets can be identified. Because the processingcontained only coarse sorting screens(no actual crushing), Sup-plier II produced a 1.0 cm minus material that was nominlarger than that provided by Supplier I.
The samples from Supplier II(recycling operation) were noticeably coarser and contained more(nonglass) debris than thsamples from Supplier I(quarry). The Supplier II samples emittea noticeably musty odor, most likely from the residual liqu(soda, beer, wine, etc.) and label glue on the glass fragments.debris s,3.4 wt. %d consisted primarily of bottle labels, ametal and plastic caps. Supplier I’s samples, which incluwaste industrial glass, contained virtually no de
Fig. 2. Crushed glass from Supplier II. Inset
s,0.8 wt. %d, and did not emit any detectable odor. The crushed
528 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
glass from both suppliers was highly angular, abrasive and eited a relatively high degree of hardness. Although angularabrasive, the crushed glass particles were sufficiently smathat the rubbing of glass between bare hands posed nohazard.
A series of laboratory tests were performed to evaluatebasic physical properties of the crushed glass, including thereceived” water and debris content, specific gravity, graindistribution, and minimum and maximum density. The physproperties of the crushed glass samples and relevant test proare summarized in Table 2. The reported water contents areage values based on six tests performed on specimens obfrom different locations in the bulk shipping containers. Thwas relatively little variation in water content between specimfrom each supplier.
Gravimetric debris content tests were performed oncrushed glass. For each supplier, three 2 kg bulk specimenscollected from different locations within the bulk shipping ctainer, oven dried at 46°C for 24 h, weighed, and then evspread out in a flat aluminum pan. Debris(nonglass materials, ebottle caps, labels, plastic tops, etc.) was visually identified anmanually removed. The debris content was computed as thweight of debris divided by the dry weight of crushed glass.debris contents reported in Table 2 are averaged over tripspecimens.
Specific gravity tests on the crushed glass yielded valu2.48 and 2.49 for Suppliers I and II, respectively. These valueconsistent with the values of 2.49–2.52 reported by FH(1998), and are about 10–15% less than those of most naaggregates(Lambe and Whitman 1969). Minimum and maximumdensity tests were performed, and the reported average v(Table 2) are based on a minimum of four specimens from esupplier.
graph shows a closeup of the material(scale in cm)photo
A series of sieve analyses were conducted on specimens from
s for, thessingsn Fig.r andn the
hee asesd re-pliermost
each supplier to assess the variability in material gradation aclassify the crushed glass according to the United Soil Classtion System (USCS) and the AASHTO (1996) systems. Thspecimens were collected from four different locations wieach of the bulk shipping containers. Additional sieve anawere later repeated on exhumed samples of crushed glassthe modified Proctor tests to assess shifts in the grain-sizebution resulting from compaction-induced particle breakage.drometer tests were also performed on a single specimeneach supplier to determine the grain size characteristics ofiner material(particle size,75 mm).
Fig. 3 presents the range of grain size distribution curvespecimens from each supplier. In their as-received conditionsamples were predominately granular with less than 5% pathe No. 200 sieve(i.e., ,75 mm). The small fraction of finepassing the No. 200 sieve was non-plastic. The data shown i3 indicates that the crushed glass from Supplier II was coarsecontained a higher fraction of fine gravel and coarse sand tha
Table 2. Summary of Physical and Engineering Properties of Crus
Test/Index Test standard As-r
ASTM D2216-98 2.3[2.03–
Specific gravitysGsd ASTM C127-88 2.4
ASTM D4254-00 1.1[1.14–
ASTM D4253-00 1.7[1.77–
Median grain sizeD50 (mm)[range]
Coefficient of uniformity,Cu
Sand contents.075–4.75 mmd (%)[range]
Fines contents,0.075 mmd (%)[range]
USCS classification ASTM D2487-98 SW
AASHTO classification AASHTO M43-88 No.
LA abrasion(wt. %) ASTM C131-96 24%
Sodium sulfate soundness(wt. %) ASTM C88-99 6.38
Hydraulic conductivitya (cm/s)[range]
ASTM D2434-68 1.613f1.36310−4–
Modified Proctor ASTM D1557-00
gd,max skN/m3d 18.
wopt (%) 9.7
Standard Proctor ASTM D698-00
gd,max skN/m3d 16.
wopt (%) 12.
Direct shear ASTM D3080-98
Internal friction (deg)sn (kPa)0–60 61–
Consolidated drained triaxiala
internal friction (deg)USACOE 48
aSpecimens were compacted to a dry density equal to 90%s±1%d of the
material from Supplier I. Although there is variability in the grain
JOURNAL OF MATERIALS IN
size distribution among the tested specimens from Supplier Imaterial was sufficiently uniform that all specimens were clafied as “SW,” or well-graded sand, based on the USCS,“Number 10” based on the AASHTO systems. With the excepof one specimen which classified as poorly graded sand(SP), allof the material from Supplier I classified as “SW” based onUSCS. All of specimens from Supplier I classified asAASHTO No. 10 aggregate.
Fig. 4 compares the grain size distributions of individcrushed glass specimens in their “as-received” and “postcomtion” conditions following modified Proctor compaction. Tupper portion of the figure quantifies the change in grain sizthe measured increase(difference) in particle fineness. Samplexhumed from the compaction molds were more fine grainesulting from compaction-induced particle breakage. The SupII specimens experienced a greater reduction in particle sizelikely due to Supplier II’s initially coarser gradation and
pplier I Supplier II
d Postcompaction As-received Postcom
— 2.49 —
SW SW SW-SM
No. 10 No. 10 No. 10
— 25% —
— 7.1% —
— 17.5 —
— 11.2 —
— 16.6 —
— 13.6 —
— 59–62° —
— 55–59° —
— 47–55° —
— 47° —
brittle nature of glass. Despite particle breakage, the shifts in
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 529
fatey andn tesgi-
aggref 24%gerehichumts of
f antible ton didma-
gradation were not sufficient to change the USCS or AASHclassification of crushed glass from either of the suppliers.
Durability and Weathering Resistance Tests
Los Angeles (LA ) abrasion, freeze–thaw, and sodium sulsoundness tests were performed to assess the durabilitweathering resistance of the crushed glass. The LA abrasio(ASTM C131) is commonly used in highway and materials enneering to assess the abrasion resistance of constructiongates. The crushed glass yielded LA abrasion wear values o(Supplier I) and 25%(Supplier II) which are less than the ranof 30–42% provided by FHWA(1998). These wear values aslightly higher than those of most common aggregates, wtypically range from about 12 to 20%, are below the maximallowable value of 30% permitted by many State DepartmenTransportation(DOT) specifications.
Freeze–thaw cycle tests assess particle breakdown and dposition due to freezing-related expansion of water in the miractures of solids. Six 1,000 g specimens of crushed glasssubjected up to 120 cycles of repeated freezing and(heating-induced) thawing. The grain size distribution was msured after 0, 10, 20, 30, 60, and 120 cycles of freezingthawing. The particle size(D10, D50, D60) indices were routinelevaluated, where the numbered subscripts denote the ppassing by weight. The coefficient of uniformitysCud, an index omaterial uniformity defined asD60/D10, was also monitored. Materials that are well graded and uniformly graded are definethose having Cu.6 and Cu,4, respectively. Often smaamounts of material are lost during sieving and grain size labtory testing procedures. As such, it was not possible to accuperform repeated grain size distribution tests on a singlespecimen. Fig. 5 presents the test results forD10, D50, D60, andCu
versus number of freeze–thaw cycles, and a baseline
Fig. 3. As-received grain
received” specimen not subjected to freezing and thawing. While
530 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
Fig. 5 shows slight variations in grain size indices, the lack oapparent trend suggests that the crushed glass is not suscepfreeze–thaw related degradation. In other words, the gradationot become progressively finer as would be expected if theterial was deteriorating or degrading.
Sodium sulfate soundness testing was completed oncrushed glass sample. The results were 6.38 and 7.1 wt.Suppliers I and II, respectively, which is well below the 12% liestablished for coarse aggregates used in concrete applicat
Standard and modified Proctor compaction tests were perfoon the crushed glass samples, as summarized in Fig. 6. Zevoids curves(i.e., maximum dry density as a function of moistcontent for a fully saturated sample) incorporating specific gravties of 2.485(glass) and 2.65(typical soil or aggregate) are shownfor reference purposes. The compaction tests were typicallyformed using 5–6 moisture-density points. Table 2 summathe maximum dry densitysgd,maxd and optimum water conteswoptd for each crushed glass sample. For comparison, FH(1998) reports typical modified compaction densities17.7–18.6 kN/m3 for glass cullet with optimum water contentsthe range of 5.7–7.5%. As with natural aggregates, the moProctor data of the crushed glass exhibitgd,max values that exceeded those of the standard Proctor values by approxim7.5%. The moisture–density curves exhibit the characteristicvex shape of natural aggregates, suggesting that crushedcompacts similarly to natural aggregates. The flatness ocurves suggest stable compaction characteristics and goodability over a wide range of water contents given that cruglass is relatively insensitive to moisture content.
istributions for crushed glasssize d
n thed in
ilure.2 cmn pro-
idualerote theh asnter-r–rmalshear
Constant head hydraulic conductivity tests were performed ocrushed glass samples. Each test specimen was compacterigid wall permeameter to a dry density equal to 90%s±1%d ofits gd,max based on the modified Proctor test data. This levecompaction corresponds to relative density values of 88 andfor the Supplier I and II specimens, respectively. The testingcedure entailed initially spraying(moisture conditioning) thespecimens with tap water followed by placement in a 6.1high, 15.2 cm diameter rigid wall permeameter using thin land compaction with a rubber tipped pestle. Table 2 reportdraulic conductivity values of 1.61310−4 and 6.45310−4 cm/sfor Supplier I and II materials, respectively, averaged over trcate samples. As expected, the hydraulic conductivity valuethe coarser Supplier II material were slightly higher than theplier I material. The measured hydraulic conductivity valuesboth materials are within the typical range for compacted nasoils and aggregates having SW designations(e.g., NAVFAC1986). The FHWA(1998) reports fine glass cullet having hydralic conductivities as high as 6310−2 cm/s. The measured valusuggest that crushed glass is a relatively free draining mathat should perform well in filtration and drainage application
A series of consolidated, drained direct shear and triaxial cpression tests were performed to assess the shear strengthteristics of the crushed glass. The tests were performecrushed glass specimens compacted to 90%s±1%d of their gd,max
values based on the modified Proctor compaction results.
Fig. 4. Compaction-induced gr
Direct shear tests were conducted on 6–7 specimens from each
JOURNAL OF MATERIALS IN
supplier over a wide range of normal stressesssnd s15–190 kPadto assess the degree of nonlinearity in the Mohr–Coulomb faenvelope. The tests were performed using a 10.2 cm by 10by 5.1 cm deep direct shear device. The sample preparatiocedure included spraying(moisture conditioning) followed byplacement in the direct shear box using thin lifts that were cpacted with a rubber tipped pestle. The samples were thenrated with tap water to eliminate capillary tension and a nostress was then applied for several hours before testingspecimens were sheared at a rate of 1.2 mm/min yielding ato failure (peak shear stress) of about 4–6 min and displacemeon the order of 3–5 mm. The crushed glass exhibited dilbehavior during shearing with the specimens tested under hconfining stresses experiencing the highest levels of dilatan
To illustrate the nonlinearities in the direct shear data(Fig. 7),the failure envelope was represented using a best-fit secondpolynomial in lieu of the typical linear Mohr–Coulomb enveloThe coefficients of variationsR2d were greater than 0.95, indicing that the regressions fit the data well even for assumedsionsyd intercepts of zero. However, the crushed glass exhibiminor degree of apparent cohesion in triaxial compression,sibly due to particle adhesion resulting from label glues, ressugars or other cohesive(gummy) substances. Hence, the zcohesion assumption is a simplification made to demonstranon-linearities in the material. Highly angular materials, succrushed glass, are known to exhibit significant nonlinearity. Inal friction angles sfd based on conventional linear MohCoulomb criteria are presented for different ranges of nostresses in Table 2. These data are larger than the directfriction angles of 51–53° reported by FHWA(1998).
e changes by modified Proctor testain siz
A second series of shear strength tests were performed under
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 531
umetestind in action,
er ofecondnt forctor
drained triaxial compression. The tests were performed on ispically consolidated specimens using an automated triaxialing, control and data acquisition system. The specimen sizeied slightly between tests, but were about 13.8 cm high7.0 cm in diameter. Axial load, displacement, specimen volchange, and internal pore pressures were measured duringat a rate of 20 Hz. Crushed glass specimens were preparemanner consistent with the direct shear tests. After compa
Fig. 5. Grain size indices andCu as fu
specimens were placed under vacuum, and the split mold was
532 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
removed. Preliminary testing indicated that the crushed glasticles caused small punctures in the 0.3 mm thick latex membsurrounding the specimen. To remedy the situation, a laygrease was applied to the outside of the membrane, and a s0.3 mm membrane was placed over the specimen. To accouthe stiffening effect of the two membranes, a correction fawas applied to the measured strength[ASTM D4767(ASTM1995)]. After preparation, the specimen was mounted in the
of freeze–thaw cycles for crushed glassnction
axial cell, flooded with tap water, and then backpressure satu-
hr’spesglesd II,aboutfiningn thein-
. It issult ofilure
re thator theerva-ake
rated. The specimens were tested under three confining preranging from 26 to 140 kPa, equivalent to overburden dept1.5–10 m. The specimens were tested under a strain-contloading rate of 1.25 mm/min to an axial strain of approxima18–20%, yielding a typical time to failure(peak deviator stres)of about 10–15 min.
Fig. 8 shows deviator stress and volumetric strain as a funof axial strain for crushed glass from both suppliers. The devstressss1–s3d is defined as difference between the maximss1d and minimumss3d principle effective stresses. The SuppI specimens reached their peak deviator stresses at aboutaxial strains«ad. The coarser Supplier II specimens reachedpeak deviator stresses at slightly larger axial strains(about7–10%). Crushed glass specimens from both suppliers geneexhibited a limited degree of strain softening. Fig. 8 also shthe volumetric behavior of the specimens during triaxial tests.
Fig. 6. Standard and modified Proctor compaction test resultcrushed glass
Fig. 7. Failure envelopes from direct shear tests on compacrushed glass
JOURNAL OF MATERIALS IN
crushed glass from both suppliers was dilatant(i.e., exhibitingnegative volumetric strain) and the greater dilatantcy of tcrushed glass from Supplier I was attributed to its slightly hirelative density.
Fig. 9 presents the triaxial test data in the common Mocircle format with best fit linear Mohr–Coulomb failure envelofor crushed glass. The triaxial tests yielded internal friction ansfd of 48 and 47° for the crushed glass from Suppliers I anrespectively. As expected, these internal friction values are10% less than the direct shear test data for comparable constresses due to differences in boundary conditions betweetests. The Mohr–Coulomb failure envelopes shown in Fig. 9dicate that the materials exhibited a modest level of cohesionbelieved that these cohesion intercepts are the combined re(1) a linear representation of a nonlinear shear strength faenvelope, and(2) apparent cohesion resulting from adhesive lglues, sugars, and other substances. It was often observewet, gummy glass particles often adhered together when atone of the glass particles had a thin surface covering ofpaper(bottle label), or both.
In summary, the direct shear and triaxial compression tesults indicate that the crushed glass was dilative and exhhigh internal friction. As discussed above, apparent cohesionbe related to adhesion of the glass particles from labeland/or other cohesive substances and should be neglectedirect shear tests indicate that internal friction was reduced uhigh confining stresses. Hence, care must be taken to ensustrength parameters selected for design are appropriate fanticipated stress conditions. Note, however, that even constive estimates of the internal friction angle of crushed glass mit equal to or greater than corresponding natural soil or aggreunder similar compactive efforts.
Toxicity characteristic leaching procedure(TCLP) (U.S. EPA1991b) and synthetic precipitation leaching procedure(SPLP)(U.S. EPA 1991a) tests for metals were performed on thereceived crushed glass from each supplier. The TCLP andtests are used to simulate metals leaching under the condassociated with typical landfill and soil conditions, respectivBy the TCLP and SPLP tests, a material is designated as aardous waste” if any detected metal occurs at concentratioexcess of 100 times the drinking water standard. Table 3 prethe leaching data versus the USEPA primary drinking waterdards(U.S. EPA 1999).
The more aggressive TCLP test results indicated trace cotrations of barium in the leachates from the Supplier I gsample, and chromium and lead in the Supplier II glass sampis not clear why the crushed glass from Supplier I contabarium, though this may be attributed to the coloring andcialty ingredients used in some commercial and industrial gware. While the exact source of the chromium and lead cannknown in the glass from Supplier II, it is not unreasonablattribute the trace chromium and lead to the presence ofprinting pigments, wine bottle capping materials, broken ceraand crystal(leaded glass) that may have appeared in the suppsamples. The SPLP test did not produce measurable conctions of metals above their detection limits in either glass samexcept for mercury, which could be the result of waste thermeter glass materials that may have entered the waste glass stream.
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 533
Real and Perceived Barriers to Using Crushed Glass
In reviewing the measured properties of crushed glass, it becclear that glass has excellent potential for use as an aggregstrength and drainage applications. And yet the U.S. EPA(2000)
Fig. 8. Stress and volume change behavior versus axial strain ois indicated in parentheses
data indicates that 7.8 million t of glass is landfilled annually.
534 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
This begs the question of how to increase the beneficial ucrushed glass in civil engineering and construction sinceindustries consume large quantities of materials. There aremain interrelated barriers to this issue: contracting mechanexisting nonperformance(material specific) based specificationand the unnecessary cross referencing of specifications. A c
pacted crushed glass in triaxial compression; effective confiningsscdf com
nient example illustrating the general problem regarding crushed
manycledate forecy-th oflica-dis-
glass is to consider that large metropolitan areas issue wasteagement contracts for refuse and curbside collection of recmaterials, and through another department purchase aggregroad construction. For example, one suburban Philadelphiacler accepting glass accumulated approximately 1,000 t /monglass cullet, but experienced trouble identifying reuse apptions. Consequently, the facility paid approximately $18/ t to
Fig. 9. Mohr–Coulomb diagrams f
Table 3. Toxicity Characteristic(TCLP) and Synthetic Precipitation(SP
U.S. EPA drinkingwater standarda
Arsenic 0.05 5.0
Barium 2.0 100
Cadmium 0.005 1.0
Chromium 0.1 5.0
Lead 0.015 5.0
Mercury 0.002 0.2
Selenium 0.05 1.0
Silver 0.05 5.0
Note: All data in milligrams per liter(mg/L).aU.S. EPA(1999).bSW-846, Chapter 7.4(Revision 3, December 1994).
JOURNAL OF MATERIALS IN
pose of the crushed glass at a New Jersey landfill where iused as daily cover; a price significantly less than the tylandfill disposal cost of $50/t. However, the facility could htripled its acceptance of glass based on plant capacity.
The cost of curbside collection and the lack of, or deprenature of, secondary glass markets may have therefore couted to New York City’s recent decision to suspend glass co
xial compression tests of crushed glass
aching Procedure Results
upplier I Supplier II Supplier I Supplier
,0.10 ,0.10 ,0.10 ,0.10
0.151 ,0.10 ,0.10 ,0.10
,0.01 ,0.01 ,0.01 ,0.01
,0.03 0.0772 ,0.03 ,0.03
,0.10 0.128 ,0.10 ,0.10
0.0002 ,0.0002 0.00024 ,0.0002
,0.20 ,0.20 ,0.20 ,0.20
,0.02 ,0.02 ,0.02 ,0.02
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 535
f itsrisk-r per
sg capon, andn forerviceancet toma-asizeionalere
the, fors the
t pased foa del
edcrite-owers arets, ama-
y ex-red to
regated in-r not onnt ofer ofifica-jec-
m ap-ies—mul-. Forptageandg is-tle 25is-ur-stan-
tion and recycling(“Recycling—waste of time” 2002). HoweverNew York’s decision is somewhat hard to understand considthat at the same time cities purchase embankment material aaggregates for construction at raw material costs of $2–$7/ttransportation($5/ton for 30 km radius), or up to $12/t. The quantities of both glass and earth materials handled by cities are ein the hundreds of thousands of tons, making the appropriatestitution of glass in earthwork applications a very viable approto managing city financial resources(with a net revenue swingup to $62/t). However, while the economics are compellingthe use of crushed glass could be easily justified from a pemance basis in a large number of urban embankment, fill, reing wall, wharf, dock construction, and aggregate applicatthe contracts for waste management, construction, and purchare issued by different city departments.
While often it may appear that government procurement ptices are key handicaps to increasing beneficial use, this igthe practices that hold sway in the engineering and construcommunity. In other words, even if cities adopted a holisticproach to materials and contracts management, many curregineering design and construction codes either deter orpletely prevent the use of crushed glass in beneficialapplications. The civil and construction industry, because oobvious concern with public welfare and safety, is generallyadverse and resistant to change unless a distinct cost and/oformance advantage can be realized(which would occur in citieand coastal plain areas lacking aggregate quarries but havintive glass sources). It is also a feature of the civil/constructienterprise that they make use of time-tested specificationsDOT specifications are often employed or referenced evenontransportation applications. This practice has been a dissto using recycled materials, and a move toward performspecifications could eliminate this bias. It is also importanrecognize that the civil and construction engineering fieldstured over some 200 years based on a system that emphraw, virgin materials, and consequently, the entire professframework, and its acquisition and distribution networks wdeveloped as such.
The key to understanding problems like this pertains tospecifications themselves. The trickle down effect of this isexample, that the local municipal engineer who also serves arecycling coordinator cannot approve the “common sense” ucrushed glass for septic field drainage media because the ging Department of Environment Protection(DEP) regulations require the use of DOT-approved aggregates. Ironically, wcrushed glass is chemically inert(noncalcareous) and is perfectlysuited for septage field applications, the crushed glass cannomany DOT aggregate specifications because they are gearconcrete and asphalt applications where glass is consideredeterious material.
Density, Aggregates and Embankment Materials
Many civil engineers associated increased soil density withcreased soil strength and vice versa. While this is generally tris then not surprising that many State DOT regulations estaminimum acceptable density criteria for embankment matand aggregates with the unintended consequence of eliminstrong lightweight materials from consideration. For examMaryland DOT(MDOT 2001) Section 916 establishes agd,max noless than 15.72 kN/m3 for all soil and soil aggregate borrowAASHTO T180, Method C(i.e., ASTM D1557). Delaware DOT(DEL DOT 2001) regulations(Section 209) set a lower limit on
536 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
gd,max 14.4 kN/m3 for all borrowed materials. While crushglass demonstrates excellent frictional strength, the densityria present a serious challenge for glass owing in part to its lspecific gravity. Worse, these minimum density specificationill suited for coastal communities with captive glass markelimited supply of quarried aggregates, and where lightweightterials are preferred for embankment construction over softpressible soil deposits. Thus, Ocean City, Md. has a very aglass recycling program and contractors often choose glaspipe backfill and other non-DOT applications because of thesavings(;$5/t) over aggregates transported from inland quar
Pennsylvania has a different approach to embankment maand aggregates. PennDOT(PennDOT 2000) Publication 408Section 206.2.(a) establishesgd,max of 95 lb/ ft3 s15.22 kN/m3dby ASTM D698for natural soil materials used in embankmconstruction with additional constraints placed on gradati(.35% passing No. 200 sieve) and plasticity. However, crushglass is technically not soil, and it therefore qualifies for usembankment material in all proportions under the granular,dom and suitable embankment material classifications of thesection. These embankment categories allow for broad accepof materials including slags, quarry fines, incinerator andashes, construction and demolition debris, etc. Many statePennsylvania base compaction criteria on a minimum pecompaction or percent density based on a material’s compaproperties, and not a universally mandated value(in kN/m3).
Though strong lightweight aggregates are addressePennDOT(PennDOT 2002) Publication 408, Section 703.2, glais intentionally treated as a deleterious material and is explexcluded from cement concrete and bituminous wearing couThe use of crushed glass is limited to 4 wt. % by two of thaggregate quality types(A, B), and to 10 wt. % for all other us[Section 703.2(c)10]. The NY DOT(NY DOT 1995) Section 703aggregate specifications are very similar, and glass is totallcluded from the section, as aggregates are specifically referby their geologic or industrial source.
While the materials-specific nature of DOT specificationgradually giving way to performance specifications, it is imptant nevertheless to realize that DOT embankment and aggspecifications were developed to avoid certain stability ancompatibility issues, and these concerns may have little orelevance in non-DOT applications. It is therefore incumbennontransportation professionals to ascertain the actual intethe DOT specifications before referencing them as a mattpractice, preference, or convenience. Force fitting DOT spections to other applications without realizing the technical obtives, criteria, and hidden costs(and time) implied by the DOTprocedures and specifications can bar recycled materials fropropriate end uses. Ironically, even environmental agencchartered with increasing recycling rates—occasionally progate regulations that impinge on recycled materials useexample, the key characteristics of drainage media in sefield applications are gradation, hydraulic conductivitychemical inertness, the governing regulations typically beinsued by state DEPs. Reproduced below are excerpts from TiPennsylvania Code, Chapters 73, theRegulations for Sewage Dposal Facilities(2001), where the underlined portions of the crent regulation demonstrate the ubiquitous reference to DOTdards, as is common in civil engineering practice:1. Section 73.55(c) Sand. Sand suppliers shall provide cert
cation in writing to the sewage enforcement officer andmitted, with the first delivery to the job site from every sasource listing the amount of sand delivered, and that all sand
cretent ofll be0.theed intioneet
ms.-viron-Pub-odsandhosetan-tionsls forucesshede ap-
e insDOTs toeringycling
onsag-fa-s isaler oft
sentsies.atedize isinget ofe
of re-101DOTpor-DEPss for
f theweciledev-It is
ng as, itsggre-proce-
supplied meets the requirements posted in the DepartmTransportation specifications Publication No. 408, sec703. The size and grading shall meet bituminous consand Type No. 1 or No. 3 requirements from a DepartmeTransportation certified stockpile. The sieve analysis shaconducted in accordance with PTM No. 616 and No. 10
2. Section 73.162(b)(4)(iv) … Coarse aggregate used intransition layer shall meet the Type B requirements postthe Department of Transportation specifications PublicaNo. 408, section 703, Table B. Size and grading shall mAASHTO No. 8 requirements, as described in… Table C,from a Department of Transportation certified stockpile.
These DOT specifications introduce a variety of probleFirst, calcareous aggregates(limestone, etc.), which are not appropriate for use in sewage applications and aggressive enments for obvious reasons, are not prohibited by PennDOTlication 408, Section 703. Also, Pennsylvania Test Meth(PTMs) developed by the PennDOT Bureau of ConstructionMaterials are additionally cited, an obscure reference for tnot conducting business with PennDOT. Equivalent ASTM sdards are more commonly referenced. The AASHTO gradaare often referenced when specifying aggregates and soistrength or drainage applications, but this infrequently introdproblems unless additional quality standards are applied. Cruglass is strongly disadvantaged from potential use in septagplications for the additional following reasons:1. While mineralogically identical to most sands, crushed g
does not meet the basic material(source) description forag-gregatesas defined in PennDOT Pub. 408, Section 703.simply, crushed glass is not sand.
2. Even if the PennDOT source material definitions were inpreted liberally, Type B aggregate quality limits the glcontent to 4 or 10 wt. %. Blending and testing to guarapercentage limits are costly.
3. PennDOT certification implies an onsite quality assuraquality control(QA/QC) materials testing program that cothe typical Pennsylvania quarry on the order$25,000–$30,000/year to maintain, and furthermorequires a company listing per specific aggregate typPennDOT Publication 34, Bulletin 14,Aggregate Producer.
While quarry operations can easily adsorb the costs ofcertification, recyclers simply do not have financial resourceadsorb this cost, even though Table 2 shows that the engineproperties of the crushed glass processed at quarries and recfacilities are virtually identical.
The financial implications of the current PADEP regulatiare staggering. The typical quarry price of AASHTO No. 8gregate and bituminous sands(B1 and B3) are on the order o$7/ t (Pennsylvania prices). Bituminous sands are roughly equivlent to an AASHTO No. 10 gradation, for which crushed glasgenerally the same prices$7/ td and often less. Assuming locdelivery, the sale price of crushed glass would be on the ord$10/ t versus a landfill disposal cost of $50/ t(landfilling), or a nerevenue swing of approximately $60/ t.
The typical sand mound uses 100 t of sand and repre$700 in lost economic opportunity to glass recycling facilitWith 67 Pennsylvania counties each constructing an estimannual average of 100 sand mounds, the annual market sabout 670,000 t, or a $4.7 million dollar industry. Assum100% capture by crushed glass, modification of this single sstateregulations could increase thenational glass recycling ratby an additional 7%, based on the national estimate11 million t (U.S. EPA 2000).
JOURNAL OF MATERIALS IN
Pennsylvania has become a leader in the beneficial usecycled materials and industrial byproducts by virtue of its Act(1988). Through research and joint efforts, PADEP and Penncontinue to identify and remove barriers to recycling in transtation applications. While regulations are hard to change, PAhas issued guidance allowing the substitution of crushed glaaggregates with third party laboratory certification based onwork. In this way, a crushed glass stockpile could be certwith a 1 week turnaround time for approximately $200 andentire DOT regulatory framework is avoided without comproming the performance of the septage field application. Certificwould accompany the bill of sale, as is the usual practice inaggregate industry.
The provided example illustrates the complicated nature orecycling problem, i.e., the devil is in the details. And whilehave illustrated how Pennsylvania has identified and reconthe problem to increase recycling, examples like this exist inery state, in every branch of engineering and construction.hoped that this data set on crushed glass and the economvolved serve as motivation for policy makers, planners andneers to thoroughly examine and eliminate unnecessary bato the beneficial use of crushed glass.
Select engineering characteristics of crushed glass producedtwo processing techniques(crushing versus screening) to anAASHTO No. 10 gradation were experimentally evaluated.crushed glass samples were classified as well graded sangravel (SW) and exhibited excellent strength and workabcharacteristics. The low specific gravity(2.485) contributed tocrushed glass having compacted maximum dry densities oorder of 16.6–16.8 and 17.5–18.3 kN/m3, by the standard anmodified Procter compaction test, respectively. Direct sheartion angles were measured between 47 and 62° at normal stranging from 0 to 200 kPa. Friction angles obtained by triashear were on the order of 48° for a similar stress ranges.sured hydraulic conductivities were on the order of 1310−4 cm/s. These characteristics suggest that crushed glabe used for many civil, construction, and geotechnical engiing applications, including compacted fill, trench backfill, reting wall, or mechanically stabilized earth(MSE) wall backfill,and roadway subbase among others. A move toward performspecifications and the elimination of other unnecessary reaperceived barriers against recycling glass will enable cruglass to be increasingly used in these applications.
Both crushed glass suppliers were able to provide cruglass with consistent, reproducible properties. The enginecharacteristics of the crushed glass varied slightly betweenpliers, although it appears that these variations were more crelated to grain size distribution than the parent glass charactics or processing procedures. This suggests that as locrushed glass meets AASHTO No. 8 or 10 classificationsproperties will be comparable to or exceed those of natural agates of the same gradation, regardless of the processingdures(i.e., quarry crushing equipment versus recycling equipmscreening).
The Strategic Environmental Management Program Office oPennsylvania Department of Transportation provided financial
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 537
support for this research. D.M. Stoltzfus & Son, Inc.(TalmagePa.), and Todd Heller, Inc.(Northamption, Pa.), provided matching funds and the raw materials for the testing program. KenJ. Thornton, (PennDOT), Mark Arnold (Stoltzfus), and ToddHeller are thanked for their cooperation. James Gustine andcente Morales are thanked for their assistance in experimentLA Abrasion testing was completed by Construction TechnoLaboratories, Skokie, Il. Sodium sulfate soundness testingcompleted by Valley Forge Laboratories, Devon, Pa. The Tand SPLP tests were completed by Lancaster Laboratoriescaster, Pa. The views expressed here are solely those of theers and endorsement is not implied by the project sponsors
The following symbols are used in this paper:Cu 5 coefficient of uniformity;
D10 5 diameter in particle size distribution curvecorresponding to 10% finer by weight(cm);
D50 5 diameter in particle size distribution curvecorresponding to 50% finer by weight(cm);
D60 5 diameter in particle size distribution curvecorresponding to 60% finer by weight(cm);
wopt 5 optimum water content(%);gd,max 5 maximum dry densityskN/m3d;
«a 5 axial strain(%);sc 5 effective normal stress during triaxial shear testing
(kPa);sn 5 applied effective normal stress during direct shear
testing(kPa);s1 5 maximum principle effective stresses(kPa);s3 5 minimum principle effective stresses(kPa);
s1–s3 5 deviator stress(kPa); andf 5 friction angle(deg).
American Society of State Highway and Transportation Offic(AASHTO). (1996). “Standard specification for: sizes of aggregafor road and bridge construction.”AASTHO designation M 43-8,Washington, D.C.
American Society of Testing and Materials(ASTM). (1968). “Standardtest method for permeability of granular soils(constant head).” ASTMD2434-68, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (1988). “Standardtest method for specific gravity and absorption of coarse aggregASTM C127-88, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (1995). “Standardtest method for consolidated undrained triaxial compression tecohesive soils.”ASTM D4767-95, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (1996a). “Standardtest methods for frost heave and thaw weakening susceptibilisoils.” ASTM D 5918-96, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (1996b). “Standardtest method for resistance to degradation of small-size coarsegate by abrasion and impact in the Los Angeles machine.”ASTMC131-96, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (1998a). “Standardpractice for classification of soils for engineering purposes(UnifiedSoil Classification System).” ASTM D2487-98, West ConshohockePa.
American Society for Testing and Materials(ASTM). (1998b). “Standardtest method for direct shear test of soils under consolidated drained
538 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEM
conditions.”ASTM D3080-98, West Conshohocken, Pa.American Society for Testing and Materials(ASTM). (1998c). “Standard
test method for laboratory determination of water(moisture) contenof soil and rock by mass.”ASTM D2216-98, West Conshohocken, P
American Society for Testing and Materials(ASTM). (1999b). “Standardtest method for soundness of aggregates by use of sodium sulmagnesium sulfate.”ASTM C88-99, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (2000a). “Standardtest method for maximum index density and unit weight of soils ua vibratory table.”ASTM D4253-00, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (2000b). “Standardtest method for minimum index density and unit weight of soilscalculation of relative density.”ASTM D4254-00, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (2000c). “Standardtest methods for laboratory compaction characteristics of soilstandard effort.”ASTM D698-00, West Conshohocken, Pa.
American Society for Testing and Materials(ASTM). (2000d). “Standardtest methods for laboratory compaction characteristics of soilmodified effort.”ASTM D1557-00, West Conshohocken, Pa.
Chesner, W. H.(1992). “Waste Glass and Sewage Sludge Frit UsAsphalt Pavements.”Proc., Utilization of Waste Material in Civil Engineering Construction, ASCE, New York, 296–307.
Dames & Moore.(1993). “Glass feedstock evaluation project: engineing suitability evaluation.”Rep. Prepared for the Clean WashingCenter, Seattle.
Day, D. E., and Schaffer, R.(1989). Glassphalt paving handbook, Univ.of Missouri-Rolla, Rolla, Mo.
Delaware Department of Transportation(DELDOT). (2001). “Specifica-tions for bridge and road construction.” Dover, Del.
Dyer, T. D., and Dhir, R. K.(2001). “Chemical reactions of glass culused as cement component.”J. Mater. Civ. Eng., 13(6), 412–417.
Elliott, R. W. (2001). “Evaluation of the use of crushed recycled glasa filter medium: Part 1.”Water Eng. Manage., 13–18.
Federal Highway Administration.(FHWA). (1998). “User guidelines fowaste and by-product materials in pavement construction.”^http://www.tfhrc.gov///hnr20/recycle/waste/& (March 10, 2003)
Figg, J. W.(1981). “Reaction between cement and artificial glass in ccrete.” Proc., Conf. on Alkali-Aggregate Reaction in Concrete, CapeTown, South Africa.
Hagerty, M. M., Hite, D. R., Ullrich, C. R., and Hagerty, D. J.(1993).“One-dimensional high-pressure compression of granular mediJ.Geotech. Eng., 119(1), 1–18.
Henry, K. S., and Morin, S. H.(1997). “Frost susceptibility of crusheglass used as construction aggregate.”J. Cold Reg. Eng., 11(4), 326–333.
Johnston, C. D.(1974). “Waste glass as coarse aggregate for concretJ.Test. Eval., 2(5), 344–350.
Lambe, T. W., and Whitman, R. V.(1969). Soil mechanics, Wiley, NewYork.
Malisch, W. R., Day, D. E., and Wixon, B. G.(1975). “Use of domestiwaste glass for urban paving.”Rep. No. EPA-670/2-75-053, U.S. En-vironmental Protection Agency, Washington, D.C.
Maryland Department of Transportation,(MDOT). (2001). “Standardspecifications, construction and materials.” State Highway Admtration, Baltimore.
Naval Facilities Engineering Command,(NAVFAC). (1986). Designmanual 7.02: Foundations & earth structures, Dept. of the NavyAlexandria, Va.
New York State Department of Transportation.(NYDOT). (1995). “Stan-dard Specifications, Construction and Materials.” Metric Units, Oof Engineering, Albany, N.Y.
Pennsylvania Department of Transportation.(PennDOT). (2000). “Ap-proved aggregate producers.”Publication No. 34, Bulletin 14, Bureauof Office Services—Forms and Publications, Harrisburg, Pa.
Pennsylvania Department of Transportation,(PennDOT). (2002). “Speci-fications.”Publication 408/2000 (Change 9), Effective 7/2/02, Harris-burg, Pa.
Polley, C., Cramer, S. M., and de la Cruz, R. V.(1998). “Potential challenges for using waste glass in Portland cement concrete.”J. Mater.Civ. Eng., 10(4), 210–219.
“Recycling-waste of time.”(2002). The Economist, July 3, 32.
Schmidt, A., and Saia, W. H. F.(1963). “Alkali-aggregate reaction teson glass used for exposed aggregate wall panel work.”J. Am. ConcrInst., 60, 1235–1236.
Shin, C. J., and Sonntag, V.(1994). “Using recovered glass as constrtion aggregate feedstock.”Transportation Research Record 14,Transportation Research Board, Washington, D.C., 8–18.
U.S. Environmental Protection Agency,(U.S. EPA). (1991a). “Synthetic
JOURNAL OF MATERIALS IN
precipitation leaching procedure(SPLP).” Federal Register, 55(March 29), Washington, D.C.
U.S. Environmental Protection Agency,(U.S. EPA). (1991b). “Toxicitycharacteristic leaching procedure(TCLP).” Method 1311, FederalRegister, 55 (March 29), Washington, D.C.
U.S. Environmental Protection Agency,(U.S. EPA). (1999). “Nationalprimary drinking water standards.”EPA-F-94-001, Washington, D.C
United States Environmental Protection Agency(USEPA). (2000). “Mu-nicipal solid waste generation-recycling and disposal in the UStates: Facts and figures for 1998.”Rep. No. EPA530-F-00-02,Washington, D.C. ^http://www.epa.gov/epaoswer/non-hw/munpubs/99charac.pdf& (March 10, 2003).
CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2004 / 539