Thermally Tunable Glass Foams with Controllable Pore Size viaNetwork Manipulation: A Melt-Casting and Float-ManufacturableGlass Foaming MethodHongwei Guo,† Congcong Zhao,† Jian Xu,*,‡ Yuxuan Gong,*,§,∥ Mariano Velez,§

and Relva C. Buchanan∥

†School of Materials Science and Engineering, Shaanxi University of Science and Technology, 111 County Road, Xi’an 710021,China‡School of Physics and Electronic Information, Henan Polytechnic University, 1 Heping Street, Jiaozuo 454000, China§Department of Mechanical and Materials Engineering, College of Engineering and Applied Science, University of Cincinnati, 2600Clifton Avenue, Cincinnati, Ohio 45221, United States∥Glass Coatings and Concepts LLC, a subsidiary of the Shepherd Materials Science Companies, 300 Lawton Avenue, Monroe, Ohio45050, United States

ABSTRACT: Glass foams are being widely used as constructionalmaterials due to their unique properties in thermal insulation, fireretardation, and shockwave absorption. However, the cost of energyconsumption and processes in a conventional glass foam productionlimited the use of glass foams as sustainable materials. In this study,for the very first time, thermally tunable CaO−SnO2−P2O5−SiO2glass foams with controllable pore size were presented as a novelcategory of melt-casting and float-manufacturable glasses. It wasfound that the pore size and thermal properties become tunable bymanipulating the glass network, i.e., connecting linear chained Sn−Pnetwork with [SiO4] units. In addition, the unique combination ofthermal properties and porous structure of CaO−SnO2−P2O5−SiO2glasses shows potential in float glass foam production, which canproduce glass foams sheet-by-sheet with less complexity inmanufacturing processes.

KEYWORDS: Glass foams, Network manipulation, Thermal properties, Pore size, Float-manufacturable

■ INTRODUCTIONGlass foam is a category of lightweight constructional materialsthat has been widely used in the exterior and interior walls ofbuildings due to its thermal and acoustic insulating proper-ties.1−3 The porous structure, either being closed or openpores, of glass foam provides its unique properties in thermalinsulation, fire retardation, and shockwave absorption.2,4−6

However, the unit price of glass foam limited its use inconstructional needs, and tends to further restrict its use whenthe needs for energy-efficient buildings becomes moredemanding.6

The production cost of glass foams is associated with theinvestment of production equipment, large energy consump-tion, and sometimes cost of raw materials.2,6,7 During a typicalproduction process, glass foams are being made by batchingand melting of glasses with desired stoichiometry (batchingand melting are not necessary when waste glasses are beingused), grinding the as-melted glass or waste glasses (e.g.,cathode ray tube, window or bottle glass), mixing the glassparticulates with foaming agents, foaming the mixture (in amold) at elevated temperature, and annealing the as-prepared

glass foam through a lehr.5 Finally, glass foams are cut ormachined into desired size prior to being used in constructionsites. Couple of approaches has been applied to lower the costof glass foam production, among those include the use of wasteglass cullets,5,7 novel foaming agent from municipal wastes(e.g., slag, fly ash, porcine bone, egg shell, and etc.),2,3,8,9 andfast sintering schedule.2 The use of waste glass culletssomehow allows a melting-free processing of producing glassfoam, but also introduces accumulation of undesirable element(e.g., iron, and sulfides) and fluctuations in batch-to-batchcomposition (i.e., inconsistency in batch-to-batch compositionof glass cullets).6,10,11 Similarly, the use of novel foamingagents from wastes only helps with the cost of raw materials.Methods of producing glass foams at lower energyconsumption and higher production efficiency (or a muchmore automated production) have yet to be reported.

Received: March 18, 2018Revised: June 2, 2018Published: June 6, 2018

Research Article

Cite This: ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

© 2018 American Chemical Society 8875 DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

Dow

nloa

ded

via

Yux

uan

Gon

g on

Oct

ober

24,

201

8 at

14:

54:0

6 (U

TC

).

See

http

s://p

ubs.

acs.

org/

shar

ingg

uide

lines

for

opt

ions

on

how

to le

gitim

atel

y sh

are

publ

ishe

d ar

ticle

s.

In addition, the foaming process and the final product werereportedly to be sensitive over glass composition as well as thechemistry of foaming agents.5,10,11 For example, a decom-position type foaming agent (e.g., carbonates, such as Na2CO3)usually forms open porosity in the glass foam; also, the ionexchange between sodium in the foaming agent and alkalis inthe glass network has been reported to change the viscosity,surface tension, and final pore structure of glass foams.5,11

Similarly, a redox-type foaming agent like SiC will change theglass chemistry by introducing extra Si and reducing metalliccations in glass network (e.g., Zn, Pb, or Cu).4,5 Particularly, astudy on the foaming agents shows that superior properties ofglass foams can be achieved with the combination of MnxOyand carbon as foaming agents.12 Therefore, the interactionbetween foaming agents and matrix glasses will greatly affectthe properties of glass foams such as morphology of porosity,distribution of porosity, thermal and mechanical properties,and more importantly create problems in product consistency.With all these being said, is there a one-step and mold-freeprocess of producing glass foam with controlled stoichiometry?Can the glass foam being produced just the way as float glasses,in other words a sheet-by-sheet production? Can such aprocess lower the cost of energy consumption, reduce the costof labor, and increase the efficiency and consistency ofproduction?In this paper, we report, for the very first time, a melt-casting

method of obtaining glass foam that satisfies the requirementsfor a float production process of glasses. Glass foams, at least inthe Si−P glass family per our study, show potential in beingproduced sheet-by-sheet in a state-of-art one step process (i.e.,float glass foam manufacturing). Also, the as-prepared glassfoams showed controlled chemistry, distribution of porosity,pore size, and thermal property with good mechanical strength(>10 MPa). Moreover, the glass foams were found to betunable in their thermal properties and pore size via themanipulation of glass network. It was found that the measuredproperties show a linear regression as a function of Si4+/Ca2+.More importantly, the cost-efficiency and potential impact ofthe shown method are discussed in detail.

■ MATERIALS AND METHODSProcessing. Chemically pure ammonium diphosphate (Fisher

Scientific, USA), tin(IV) oxide (Fisher Scientific, USA), calciumcarbonate (Fisher Scientific, USA), silica (Fisher Scientific, USA)powders were weighed (see Table 1), mixed in a rotating roller for 12

h, placed into an alumina crucible, and melted at 1100 °C for 1 h. Thecomposition of glasses was designed to vary the ratio of Si/Ca whilenot changing the amount of P2O5 and SnO2 (Sn/P less than 0.25),assuming that the pore size and thermal properties of foam glasschanges as a function of Si-network connectivity (Si4+/Ca2+). Theglass forming liquid was poured into a preheated (500 °C) andrectangular-shaped (∼2 cm × 1 cm × 1 cm) graphite mold, annealedin a resistively heated oven for 1 h, and subsequently cooledovernight. The as-melted glasses show porous structure with

entrapped gases from the decomposition of ammonium diphosphateand calcium carbonate. Therefore, the foaming process was achievedduring the melting processes without the use of foaming agent or afurther sintering process. The edges of the as-prepared glass foamswere polished prior to the measurement of CTE. Triplicates wereperformed on the CTE measurement in order to generate standarddeviations.

Phase Identification and Analysis of Bulk Composition. X-ray powder diffraction was performed on all pulverized glass sampleswith a Rigaku Dmax X-ray diffractometer. A Cu kα source, anemission current of 10 mA, and a voltage of 30 kV were used in allmeasurements. Data was collected from 10° to 70° 2θ with a step sizeof 0.01°. Bulk compositions as well as the chemical environments ofthe as-prepared glass foams were measured using a PHI VersaprobeX-ray photoelectron spectrometer. In order to obtain the bulkcomposition, the fracture surfaces (as fractured in the introductionchamber, pressure is ∼10−6 Torr) were analyzed. The quantitativeanalyses were performed using high resolution scans of Si 2p, P 2p, Sn3d5/2, and Ca 2p peaks. Relative sensitivity factors of all elements(RSFSi 2p = 1, RSFP 2p = 0.60, RSFSn 3d5/2 = 4.32, RSFCa = 9.32, andRSFO = 4.77) were derived from high resolution XPS scans of avacuum-fractured Sn−P−Ca−Si oxide glass, assuming that thecomposition of glass fracture surface was identical to the bulk glasscomposition (obtained via ICP-OES and lithium metaboratedigestion). Qualitative and quantitative analyses were performedwith PHI MultiPak V8.0 software; peak positions of all scans wereshifted/calibrated using adventitious alkyl C 1s peak located at 284.60eV (the standard deviation of peak positions is less than 0.1 eV). Thepeak area value of each high resolution scan XPS peak was fitted bythe Shirley-model after automated background subtraction. Detailsregarding XPS analyses on glasses are being reported elsewhere.13 Theobtained elemental composition was then converted into oxidecomposition, as listed in Table 2. The measured standard deviation ofSi is 0.2 mol %, P is 0.1 mol %, Sn is 0.4 mol %, Ca is 0.2 mol %, andO is 0.9 mol %.

Thermal Analysis, Characterization of Pores, and Mechan-ical Strength Test. The DSC analyses were carried out using aNETZSCH Jupiter STA 409 thermal analyzer with a heating rate of15 K/min. The CTE (coefficient of thermal expansion) values weremeasured from the casted glass foams using an Orton 1412dilatometer (equipped with quartz sample tube and push rod), anda heating rate of 10 K/min was applied to all measurements with ahigh purity alumina as the standard. The microstructural image of thefractured glass foams was captured using a Zygo NewView opticalprofiler (a white laser interferometer), and the analysis of images wasperformed using ImageJ. For statistical significance, ten images fromeach sample were used for the microstructural analysis (pore size andrather than the requirements of ASTM standard.2 In addition, theflexural strength (four point bending test) was measured at a cross-head rate speed of 2 mm/min using an 8874 axial-torsion fatiguetesting system (Instron, USA). The samples used for four pointbending test were prepared using the method discussed elsewhere.2

■ RESULTS AND DISCUSSIONPhase Identification. Figure 1 shows the powder

diffraction of all measured foam glass samples. All measuredpatterns did not show any diffraction peaks, which often can befound due to the crystallization of bulk glass as well as partial

Table 1. Batched Composition of Glasses with VaryingAmounts of Silica (mol %)

Composition (mol %) SiO2 P2O5 SnO2 CaO

Sn1 5.5 47 10 37.5Sn2 6.5 47 10 36.5Sn3 7.5 47 10 35.5Sn4 8.5 47 10 34.5

Table 2. X-ray Photoelectron Spectrometer MeasuredComposition of As-Prepared Glass Foams (mol %)

Composition (mol %) SiO2 P2O5 SnO2 CaO Si4+/Ca2+

Sn1 5.7 47.2 9.4 37.7 0.15Sn2 6.5 46.7 9.3 37.5 0.17Sn3 7.4 46.3 9.2 37.1 0.2Sn4 8.3 45.9 9.1 36.7 0.23

ACS Sustainable Chemistry & Engineering Research Article

DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

8876

decomposition of foaming agent. Therefore, the as-preparedglass foams were X-ray amorphous. It has been reported that acrystalline phase containing glass foam (so-called glass-ceramicfoam) shows better mechanical performance comparing to itsnoncrystalline counterpart.2,4,5 The crystalline phase containedin a glass-ceramic foam hinders the propagation of cracks, andtherefore reinforces the glass-ceramic foam.14 However,uncontrolled crystallization of glass foams can be challengingto product consistency, management of product performance,and engineering control of final product; also, furthercrystallization of these glass may introduce additionalmanufacturing processes as well as energy consumption.Hence, it is believed that amorphous glass foams are desirablefor the development of this new melt-casting and hence float-manufacturable (can be manufactured as float glass) method.The given CaO−SnO2−P2O5−SiO2 glasses show minimumdifferences in stoichiometries between batched (see Table 1)and as-melted compositions (see Table 2). It can bechallenging to study the chemistries of as-prepared glassfoams if crystalline phases were present in the samples, i.e., theXPS will be detecting chemical information from theamorphous phase and crystalline phase simultaneously.Therefore, the absence of crystalline phases allows the study

of correlations between glass network and physical propertiesof resultant glass foams.

Microstructural and Thermal Characterization. Figure2 shows the optical profile image (both two- and three-dimensional images) of the fracture surface of Sn1, Sn2, Sn3,and Sn4 samples. The three-dimensional images wereprocessed with third-order leveling in order to remove thetilt resulted from the fracture surface. Such leveling processdoes not change the accuracy of measurement related to poresize as well as its distribution.15 As shown in Figure 2, allimaged pores were homogeneously sized closed-pores withsmall pores occasionally being found in the walls between twoadjacent pores. Pores were formed via the chemical reaction ofammonium diphosphate: negatively charged nitrogen ionswere oxidized at elevated temperature to produce NOx gases,which then forms the porous structure found in this study.Generally, two types of porous structure were reported in glassfoam studies:5 (1) open pores formed by the decomposition ofcarbonates, nitric acids, structural water, and sulfides; (2) closepores formed during a redox (mostly oxidation reactions)process of ammoniums, nitrides, carbides. Given the poreswere closed-pores, it is postulated that the formation of poresin this study is primarily associated with the ammoniumdiphosphate (redox type foaming agent). Therefore, therelationship between close porosity and “foaming agent(ammonium diphosphate)” in our work agrees with resultsbeing reported elsewhere.5 It should be pointed out that theemission of NOX glass would be minimal since the pores wereclosed.The average pore size (obtained from 10 images of each

sample) of Sn1, Sn2, Sn3, and Sn4 is 99.9 ± 3.3, 88.9 ± 3.7,78.5 ± 2.4, and 60.9 ± 2.2 μm, respectively. It was found thatthe average pore size of CaO−SnO2−P2O5−SiO2 decreasedwith the addition of SiO2. The addition of SiO2 (a glassnetwork former) increased cross-linking of the glass network,increased the viscosity of the glass forming liquids at elevatedtemperature,16,17 and limited the motion of gas bubblesaccording to Stokes law (drag forces increase with increasingviscosity).18 Given that the motion of gas bubbles can belimited with higher viscosity, the forming of larger pores can bereduced because they are being formed by the merging of smallgas bubbles. Therefore, the pore size of these melt-casted glass

Figure 1. X-ray powder diffraction pattern of all foam glass samples.The missing of crystallization peak indicated the samples to be X-rayamorphous.

Figure 2. Two-dimensional optical profile image of (a) Sn1 glass foam, (c) Sn2 glass foam, (e) Sn3 glass foam, and (g) Sn4 glass foam; three-dimensional optical profile image of (b) Sn1 glass foam, (d) Sn2 glass foam, (f) Sn3 glass foam, and (h) Sn4 glass foam.

ACS Sustainable Chemistry & Engineering Research Article

DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

8877

foams can be tuned via the addition of SiO2. A tunable poresize of glass foam will not only help tailoring the final productto different applications but also decrease the dimension ofcomplexity during manufacturing processes.The effect of SiO2 addition on the network rigidity was

evaluated using thermal properties including glass Tg (glasstransition temperature) and CTE. As shown in Figure 3, the Tg

increased with the addition of SiO2; the Tg of Sn1, Sn2, Sn3,and Sn4 is 462.6, 469.1, 476.9, and 483.6 °C, respectively (seeTable 3). Tg and CTE can be used as measures of glass

network rigidity, and higher Tg and CTE values often representa higher network rigidity as well as more developed cross-linking between network formers (Si and P in this study).19,20

The addition of SiO2 increased the network rigidity of glassfoams, and further decreased the average pore size byincreasing the viscosity at elevated temperature. Similar resultson CTEs were found (see Table 3), and lowered CTEs arebeneficial to the thermal shock resistance of glass foams.21−24

The crystallizing events shown in Figure 3 were not studiedgiven the fact that the crystallization temperature (above 700°C) is much higher comparing to the annealing temperature(500 °C). During the processing of these glass foams, therewere no crystallizing events as suggested by XRD results (seeFigure 1).Correlation between Glass Networks and Properties

of Foam Glass. Aforementioned glass compositions weredesigned to have a Sn/P ratio less than 0.25, which can keepthe Sn−P network as a linear chain network.25,26 A 4-fold[PO4] unit has one nonbridging oxygen termination and threebridging oxygen. Therefore, a Sn/P ratio less than thestoichiometry of linear chained Sn(n+2)/4PnO3n+1 kept the Sn−P network as linear chains, and further allows the addition ofSiO2 as a network former. Therefore, Si and P were acting asglass network formers in all prepared samples while Ca and Sn

were acting as network modifiers. Because the Sn−P networkwas designed to be linear chains, the properties of as-preparedglass foams (e.g., network rigidity, network connectivity,thermal properties, and pore size) should hypothetically varyas a function of Si4+/Ca2+ ratio (oxidation state of Ca ion wasdetermined using XPS). In order to investigate the relationshipbetween the glass network and thermal properties, the ratio ofSi4+/Ca2+ (addition to the Sn−P network) was quantitativelyrelated to thermal properties.Figure 4 is the plot of Tg as a function of Si4+/Ca2+. Tg

increased linearly (R2 > 0.99) with the increase of Si4+/Ca2+,

which increases the cross-linking between network formingunits such as [PO4] and [SiO4]. Because the Sn−P network islinearly chained, the addition of [SiO4] can connect the Sn−Pchains via Ca2+. Calcium ion, as a divalent ion, can be used toconnect [PO4] chains (not fully charge compensated by Sn4+

because Sn/P ratio is less than 0.25) to [SiO4] units via the−O−Ca−O− bond. Therefore, the linear relationship betweenTg and Si

4+/Ca2+ can be attributed to the cross-linking betweenchained [PO4] and [SiO4] units via divalent modifying ionCa2+.27,28 A similar linear relationship (R2 > 0.99) betweenCTE and Si4+/Ca2+ was found as well, as shown in Figure 5.It needs to be pointed that thermal properties, CTE and Tg

in particular, are closely related to the network connectivityand rigidity of glasses. In addition to these two properties,viscosity of glass forming liquids can reflect the networkproperties as well. In this study, the viscosity of glass forming

Figure 3. DSC upscans of as-prepared glass foam samples. The scaleof the DSC plots was adjusted, and first derivatives were used toidentify glass transition as well as crystallization exothermic peaks.

Table 3. Thermal Properties of As-Prepared Glass Foams

Sample CTE (10−7/°C) Tg (°C)

Sn1 84.6 ± 1.1 462.6Sn2 81.2 ± 0.8 469.1Sn3 78.3 ± 0.6 476.9Sn4 74.6 ± 1.3 483.6

Figure 4. Glass transition temperature as a function of Si4+/Ca2+. AnR2 better than 0.99 was reported for the shown linear regression.

Figure 5. CTE as a function of Si4+/Ca2+. An R2 better than 0.99 wasreported for the shown linear regression.

ACS Sustainable Chemistry & Engineering Research Article

DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

8878

liquids was related to the average pore of as-prepared glassfoams since the motion of gas bubbles are determined by theviscosity of glass forming liquids. Figure 6 shows therelationship between average pore size and Si4+/Ca2+.

As shown in Figure 6, the pore size decreases linearly whenthe ratio of Si4+/Ca2+ increases. As discussed previously, theincrease of network connectivity or rigidity (e.g., Si4+/Ca2+)increases the viscosity of glass forming liquids. An increasedviscosity or drag force limits the motion of small gas bubble(formed via redox of nitrogen in ammonium diphosphate) inthe glass forming liquid, and therefore decreases the apparentpore size by limiting the merge of small gas bubbles. However,the area percentage porosity of the glass foams does notchange (within statistical error) as a function of Si4+/Ca2+, asshown in Figure 7.

Within statistical error (STD is around 2%), the areapercentage porosity does not change as a function of Si4+/Ca2+.The volume of gas bubbles retained in each glass melt shouldbe the same since the concentrations of ammoniumdiphosphate (foaming agent) from Sn1 to Sn4 remainsidentical. Therefore, the change of glass chemistry (Si4+/Ca2+) only alters the average pore size by changing the thermalproperties of the glass forming liquids. On the other hand, thearea percentage porosity is related to the amount of foamingagent added instead of being dependent on the glass chemistry.

In addition to the linear relationship between glass networkand thermal properties or pore size, the flexural strengths(a.k.a. modulus of rupture, MOR) were found to be linearlyrelated to the ratio of Si4+/Ca2+ (see Figure 7) because thenetwork rigidity is also related to strength of glasses. Weconclude that the ratio of Si4+/Ca2+determines the glassnetwork properties of this particular glass system and changesthe thermal properties of as-prepared glass foams. Therefore,the tunability of CaO−SnO2−P2O5−SiO2 glass foam inthermal properties, pore size, and mechanical strength wasachieved via the manipulation of glass network.

■ PERSPECTIVES OF COST-EFFICIENCY ANDPOTENTIAL IMPACT

This paper presents a series of novel glass foam that can beproduced in a sheet-by-sheet manner similar to float glass.Conventional methods of manufacturing glass foams ofteninvolve melting and grinding of glass frits, molding of glassfoams, and sintering of final products.1−3,29 The relativecomplicated processes of conventional glass foams as well asthe high energy consumption during those processes limited acleaner and more efficient manufacturing of glass foams. Incontrast, the method shown in this study presents melt-castedglass foams (mold-free and nonsintering) that are suitable tobe produced in a highly automated glass float line in just onestep. However, mechanical tools related to this melt-castingand float-manufacturable glass foams have yet to be studiedand the trade-offs between cost of tin oxide and reducedprocessing temperature or complexity of processes need to befurther studied.Apart from the manufacturing side of these CaO−SnO2−

P2O5−SiO2 glasses, the unique combination of thermallytunable properties with controllable pore size made theseglasses desirable for manufacturing processes. In addition, itwas reported that Sn−M−P (M = alkaline earth, Mg, Zn, Ca,Sr, and Ba) glass system has lower viscosity and higherchemical durability than other low melting glass systems suchas bismuth borosilicate, borates, tellurides, and lead-containingglasses.28,30,31 A lowered viscosity helps the reduction ofenergy required for the melting and annealing processes ofmanufacturing foam glass since the liquidus temperature (orso-called melting temperature) and Tg (glass transitiontemperature) are lowered.32,33 Thus, the advantages ofCaO−SnO2−P2O5−SiO2 glass foams presented in this studyare summarized as follows:

1. thermal properties and pore size can be tuned via themanipulation of glass network, i.e., connecting thelinearly chained Sn−P network with [SiO4] units usingCa2+;

2. the glass foams show potential to be manufactured in asheet-by-sheet float glass line because it is a melt-castedglass foam;

3. the network of these glasses make themselves suitable asconstructional materials with good thermal shockresistance (low CTE), insulation properties (closeporosity), and chemical durability.

■ CONCLUSIONSFloat-manufacturable CaO−SnO2−P2O5−SiO2 glasses wereprepared using a one-step melt-casting method. The closedporosity was created by the redox reaction of negativelycharged nitrogen in ammonium diphosphate. The linearly

Figure 6. Pore size as a function of Si4+/Ca2+. An R2 better than 0.98was reported for the shown linear regression.

Figure 7. Plot of area percentage porosity and flexural strength as afunction of Si4+/Ca2+. An R2 better than 0.98 was reported for theshown linear regression.

ACS Sustainable Chemistry & Engineering Research Article

DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

8879

chained Sn−P network was connected with [SiO4] units byCa2+, and such increase in cross-linking increased the viscosityof glass forming liquids. Therefore, the NOx gas bubbles wereretained in the melt to form porous structure. In addition, themanipulation of network via alkaline earth and silica makes theas-prepared glass foams tunable in their thermal properties,mechanical strengths, and pore sizes. Glass foams in thepresent study are promising in reducing the cost-efficiency ofglass foam production, and revolutionize the processes in glassfoam production by introducing a float-manufacturableprocess.

■ AUTHOR INFORMATION

Corresponding Authors*Dr. Yuxuan Gong. Tel.: +1-513-539-5300. Fax: +1-513-539-5177. E-mail: gongyauc@gmail.com.*Dr. Jian Xu. Tel.: +86-391-3987811. E-mail: xujian@hpu.edu.cn.

ORCIDYuxuan Gong: 0000-0001-5735-6508Author ContributionsY.G., J.X., and R.C.B. conceived the research. H.G., C.Z., M.V.,and Y.G. performed the experimental work. H.G. performedthe thermal analysis. Y.G. and J.X. performed the XPS andother characterization work. C.Z. performed the four pointbending test. Y.G. and J.X. researched into the industry impactof this work. All authors discussed and analyzed the data. Y.G.wrote the paper. All authors discussed and commented on thepaper.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSY.G. thanks Dr. Carlo Pantano of the Pennsylvania StateUniversity for useful discussions. Y.G. and J.X. thank RCBresearch group for the facilities used in this study. H.G. andC.Z. thank the financial support from Science and TechnologyR&D Program Focused on Special Project (No.2017YFB0310201-02), International Science and TechnologyCooperation Program of China (No. 51472151), NaturalScience Foundation of China, Science and TechnologyResearch Project of Xianyang (No. 2016K02-28), ShanxiProvince R&D Project (No. 2017ZDXM-NY-048). Finally, wethank three anonymous reviewers for their effort in improvingthe quality of our article and scientific acumen.

■ REFERENCES(1) Arulrajah, A.; Disfani, M. M.; Maghoolpilehrood, F.;Horpibulsuk, S.; Udonchai, A.; Imteaz, M.; Du, Y.-J. Engineeringand Environmental Properties of Foamed Recycled Glass as aLightweight Engineering Material. J. Cleaner Prod. 2015, 94, 369−375.(2) Gong, Y.; Dongol, R.; Yatongchai, C.; Wren, A.; Sundaram, S.;Mellott, N. Recycling of Waste Amber Glass and Porcine Bone intoFast Sintered and High Strength Glass Foams. J. Cleaner Prod. 2016,112, 4534−4539.(3) Ponsot, I.; Bernardo, E. Self Glazed Glass Ceramic Foams fromMetallurgical Slag and Recycled Glass. J. Cleaner Prod. 2013, 59, 245−250.(4) Guo, H.; Gong, Y.; Gao, S. Preparation of High Strength FoamGlass−Ceramics from Waste Cathode Ray Tube. Mater. Lett. 2010,64, 997−999.

(5) Taurino, R.; Lancellotti, I.; Barbieri, L.; Leonelli, C. Glass−Ceramic Foams from Borosilicate Glass Waste. INT J. APPL GLASSSCI 2014, 5, 136−145.(6) Konig, J.; Petersen, R. R.; Yue, Y. Influence of the Glass−Calcium Carbonate Mixture’s Characteristics on the Foaming Processand the Properties of the Foam Glass. J. Eur. Ceram. Soc. 2014, 34,1591−1598.(7) D’Amore, G. K. O.; Caniato, M.; Travan, A.; Turco, G.; Marsich,L.; Ferluga, A.; Schmid, C. Innovative Thermal and AcousticInsulation Foam from Recycled Waste Glass Powder. J. CleanerProd. 2017, 165, 1306−1315.(8) Fernandes, H. R.; Ferreira, D. D.; Andreola, F.; Lancellotti, I.;Barbieri, L.; Ferreira, J. M. Environmental Friendly Management ofCrt Glass by Foaming with Waste Egg Shells, Calcite or Dolomite.Ceram. Int. 2014, 40, 13371−13379.(9) Fernandes, H.; Tulyaganov, D.; Ferreira, J. Preparation andCharacterization of Foams from Sheet Glass and Fly Ash UsingCarbonates as Foaming Agents. Ceram. Int. 2009, 35, 229−235.(10) Petersen, R. R.; Konig, J.; Yue, Y. The Mechanism of Foamingand Thermal Conductivity of Glasses Foamed with Mno2. J. Non-Cryst. Solids 2015, 425, 74−82.(11) Petersen, R. R.; Konig, J.; Smedskjaer, M. M.; Yue, Y. Effect ofNa2co3 as Foaming Agent on Dynamics and Structure of Foam GlassMelts. J. Non-Cryst. Solids 2014, 400, 1−5.(12) Konig, J.; Petersen, R. R.; Yue, Y. Fabrication of HighlyInsulating Foam Glass Made from Crt Panel Glass. Ceram. Int. 2015,41, 9793−9800.(13) Gong, Y.; Wren, A. W.; Mellott, N. P. QuantitativeMorphological and Compositional Evaluation of Laboratory PreparedAluminoborosilicate Glass Surfaces. Appl. Surf. Sci. 2015, 324, 594−604.(14) Apel, E.; Deubener, J.; Bernard, A.; Holand, M.; Muller, R.;Kappert, H.; Rheinberger, V.; Holand, W. Phenomena andMechanisms of Crack Propagation in Glass-Ceramics. J. MECHBEHAV BIOMED MATER 2008, 1, 313−325.(15) Wieland, M.; Hanggi, P.; Hotz, W.; Textor, M.; Keller, B.;Spencer, N. Wavelength-Dependent Measurement and Evaluation ofSurface Topographies: Application of a New Concept of WindowRoughness and Surface Transfer Function.Wear 2000, 237, 231−252.(16) Eichler, K.; Solow, G.; Otschik, P.; Schaffrath, W. Bas (Bao·Al2o3· Sio2)-Glasses for High Temperature Applications. J. Eur.Ceram. Soc. 1999, 19, 1101−1104.(17) Mauro, J. C.; Yue, Y.; Ellison, A. J.; Gupta, P. K.; Allan, D. C.Viscosity of Glass-Forming Liquids. Proc. Natl. Acad. Sci. U. S. A.2009, 106, 19780−19784.(18) Ramos, J. Behavior of Multicomponent Gas Bubbles in GlassMelts. J. Am. Ceram. Soc. 1986, 69, 149−154.(19) Zhang, M.; Mancini, S.; Bresser, W.; Boolchand, P. Variation ofGlass Transition Temperature, Tg, with Average CoordinationNumber,> M<, in Network Glasses: Evidence of a ThresholdBehavior in the Slope| Dtg/D> M<| at the Rigidity PercolationThreshold (> M≤ 2.4). J. Non-Cryst. Solids 1992, 151, 149−154.(20) Micoulaut, M.; Phillips, J. C. Rings and Rigidity Transitions inNetwork Glasses. Phys. Rev. B: Condens. Matter Mater. Phys. 2003, 67,104204.(21) Di Girolamo, G.; Marra, F.; Blasi, C.; Serra, E.; Valente, T.Microstructure, Mechanical Properties and Thermal Shock Resistanceof Plasma Sprayed Nanostructured Zirconia Coatings. Ceram. Int.2011, 37, 2711−2717.(22) Zimmermann, J. W.; Hilmas, G. E.; Fahrenholtz, W. G.Thermal Shock Resistance of Zrb2 and Zrb2−30% Sic. Mater. Chem.Phys. 2008, 112, 140−145.(23) Tulyaganov, D.; Fernandes, H.; Agathopoulos, S.; Ferreira, J.Preparation and Characterization of High Compressive StrengthFoams from Sheet Glass. J. Porous Mater. 2006, 13, 133−139.(24) Mear, F.; Yot, P.; Cambon, M.; Ribes, M. Elaboration andCharacterisation of Foam Glass from Cathode Ray Tubes. Adv. Appl.Ceram. 2005, 104, 123−130.

ACS Sustainable Chemistry & Engineering Research Article

DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

8880

(25) Masai, H.; Takahashi, Y.; Fujiwara, T.; Matsumoto, S.; Yoko, T.High Photoluminescent Property of Low-Melting Sn-DopedPhosphate Glass. Appl. Phys. Express 2010, 3, 082102.(26) Courtney, I. A.; Dunlap, R.; Dahn, J. In-Situ 119sn MossbauerEffect Studies of the Reaction of Lithium with Sno and Sno: 0.25B2o3:0.25 P2o5 Glass. Electrochim. Acta 1999, 45, 51−58.(27) Krohn, M. H.; Hellmann, J. R.; Mahieu, B.; Pantano, C. G.Effect of Tin-Oxide on the Physical Properties of Soda-Lime−SilicaGlass. J. Non-Cryst. Solids 2005, 351, 455−465.(28) Bekaert, E.; Montagne, L.; Delevoye, L.; Palavit, G.; Revel, B.Structure and Properties of X Sno−(100−X) P2o5 Glasses. C. R.Chim. 2004, 7, 377−382.(29) Silva, R.; de Brito, J.; Lye, C.; Dhir, R. The Role of Glass Wastein the Production of Ceramic-Based Products and Other Applica-tions: A Review. J. Cleaner Prod. 2017, 167, 346−364.(30) Hong, J.; Zhao, D.; Gao, J.; He, M.; Li, H.; He, G. Lead-FreeLow-Melting Point Sealing Glass in Sno−Cao−P2o5 System. J. Non-Cryst. Solids 2010, 356, 1400−1403.(31) Masai, H.; Fujiwara, T.; Matsumoto, S.; Tokuda, Y.; Yoko, T.Emission Property of Sn2+-Doped Zno−P2o5 Glass. J. Non-Cryst.Solids 2014, 383, 184−187.(32) Kim, K.; Kim, K. Valuable Recycling of Waste Glass Generatedfrom the Liquid Crystal Display Panel Industry. J. Cleaner Prod. 2018,174, 191−198.(33) Singh, N.; Li, J. An Efficient Extraction of Lead Metal fromWaste Cathode Ray Tubes (Crts) through Mechano-Thermal Processby Using Carbon as a Reducing Agent. J. Cleaner Prod. 2017, 148,103−110.

ACS Sustainable Chemistry & Engineering Research Article

DOI: 10.1021/acssuschemeng.8b01225ACS Sustainable Chem. Eng. 2018, 6, 8875−8881

8881