Ling TC, Poon CS.

(2011) Properties of architectural mortar prepared with recycled glass with different particle sizes. Materials and Design; 32 (5): 2675-2684. http://www.sciencedirect.com/science/article/pii/S0261306911000148

Properties of Architectural Mortar Prepared with Recycled Glass with Different Particle Sizes
Tung-Chai Ling and Chi-Sun Poon* Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong

Abstract The recycling of glass waste as a source of aggregate for the production of concrete products has attracted increasing interest from the construction industry. However, the use of recycled glass in architectural mortar is still limited. This study attempts to develop a self-compacting based architectural mortar using white cement and 100% recycled blue glass as key ingredients. To improve the aesthetic qualities, a certain minimum quantity of glass cullets of larger particle size must be present. The influence of particle size of the recycled glass on the engineering properties of fresh and hardened architectural mortar is investigated. The experimental results demonstrate that it is feasible to utilize 100% recycled glass as the aggregate for the production of selfcompacting based architectural mortar. These products have an average compressive strength of 40 MPa and flexural strength of 6 MPa at 28 days which are appropriate for some architectural and building applications. Also, the overall performances of all the architectural mortars prepared with different particle sizes of glass aggregates are comparable to those of control mortar mix prepared with river sand. Keywords: Concrete, Glasses, Reutilization

1. Introduction A large volume of post-consumer beverage glass bottles is being disposed daily worldwide, only a small proportion is either washed for reuse or re-melted to manufacture new glass. In Hong Kong, the majority of waste glass is used once or a few times before being discarded into landfills [1]. According to government surveys, only 1-2% of the waste glass in Hong Kong is recycled. The use of recycled glass as aggregate for the production of concrete blocks has received considerable application interest [2, 3]. General results showed that the recycled glass cullet can be used as a replacement for natural aggregate in the production of concrete blocks without compromising its mechanical properties [4]. Indeed, in many cases, the performance is enhanced [5]. There are a number of new applications of recycled waste glass, including the use of glass cullet in granular base/fill and asphalt pavement (Glassphlat) [6]. They have also been widely used as aggregates in cement mortar and concrete mixtures [7, 8]. However, most of the previous studies reported that the use of glass as a coarse aggregate has negative effects on bonding, adverse *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 1

alkali-silica reaction (ASR) and reduction in concrete strength [9]. Thus, most of the recent works have concentrated on milling glass cullet into powder form (glass powder) to replace cement in concrete [10-13]. The implementation has gained wide acceptance due to the innocuous behaviour of ASR in concrete [13]. Furthermore, the use of supplementary cementing materials (SCM) such as fly ash, metakaolin and slag are recommended as partial replacements of cement for mechanical properties enhancement and mitigating ASR expansion [14, 15]. Recently, the use of recycled glass cullets (due to their aesthetic properties) to produce decorative concrete has attracted much interest. In order to be aesthetically pleasing and visible, glass particles of a certain minimum size must be present (for example, particle size ranging from 2.36-5 mm or 5-10 mm is needed). However, the use of glass aggregates of these sizes in concrete is also the most detrimental to the concrete in terms of strength reduction and ASR expansion [16]. This paper investigates the effect of different particle sizes of glass aggregates on the engineering properties of fresh and hardened self-compacting based architectural mortar (SCAM). In addition, the influence of different percentages of metakaolin as partial replacements of white cement in the mortar is also reported.

2. Experimental programme 2.1. Materials 2.1.1. Cement A white cement (WC) was chosen to be used in this study due to aesthetic considerations for architectural mortar applications. It is particularly suitable for exposed decorative glass aggregate finishes because it contains comparatively low alkali content which could reduce potential deterioration due to ASR expansion. The chemical compositions of the cement are presented in Table 1. Table 1: Chemical composition of cementitious materials White cement Metakaolin Chemical composition (%) Silicon dioxide (SiO2) 21.36 51.39 Aluminum oxide (Al2O3) 5.27 32.91 Ferric oxide (Fe2O3) 0.20 0.58 Calcium oxide (CaO) 67.49 0.01 Magnesium oxide (MgO) 1.14 0.01 Sodium oxide (Na2O) 0.048 0.39 Potassium oxide (K2O) 0.077 0.98 Sulfur trioxide (SO3) 2.60 Loss on ignition 1.58 13.57

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 2

2.1.2. Supplementary cementitious material Supplementary cementitious materials such as fly ash and metakaolin (MK) are well-known to be able to reduce the alkali-silica reaction (ASR) in concrete [16, 17]. In order to achieve better aesthetic value, white MK was chosen to be used in this study. The chemical composition of MK is given in Table 1. 2.1.3. Fine aggregates The natural fine aggregate used in this study was river sand with most of the particle sizes passing through a 2.36 mm sieve. The recycled waste glass used was discarded post-consumer blue glass bottles sourced from a glass recycling company. The recycled blue glass bottles were washed and crushed in the laboratory. The crushed glass was sieved and sorted into three size classes (small glass (SG), medium glass (MG) and large glass (LG)) according to their gradations (<2.36mm, 2.36-5mm and 5-10mm), respectively. The gradation curves of these glass aggregates and sand are shown in Fig. 1. Besides the gradation, the shape and texture of these crushed glass aggregates were different. As seen from Fig. 2, the SG aggregate appeared to be of irregularrounded shape. The MG aggregate also had a similar shape but it had a rougher surface texture and higher surface-contact angle. However, the LG aggregate is flat and angular in shape and it also has a smoother surface texture.

100 90

C u m u la tiv e p a ssin g (% )

80 70 60 50 40 30 20 10 0 10 5 2.36 1.18 0.6 0.3

LG:5-10mm M G:2.36-5mm S G: 0- 2.36mm S and

0.15

S ie ve s ize (mm)
Fig. 1. Particle size distribution of different particle sizes of blue glass aggregates

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 3

SG: <2.36mm

MG: 2.36-5mm

LG: 5-10mm
Irregular and rounded

flat and angular

Fig. 2. Shapes and textures of different particle sizes of blue glass aggregates (SG: <2.36mm, MG: 2.36-5mm and LG: 5-10mm) 2.1.4. Admixture A superplasticizer GLENIUM SP8S based on a modified polycarboxylate was used to give satisfactory fluidity for the different SCAM mixes. This product is free of chloride and complies with ASTM C 494 [18] for Type A and F admixtures. 2.2. Mix proportions In this study, the amount of cementitious materials was kept constant at 706 kg/m3 and the cementitious to fine aggregate ratio was fixed at 1:2. It aimed to provide a high volume fraction of fine materials as commonly used for self-compacting mortar design. All the mixtures were proportioned with a fixed water/cementitious materials (w/c) ratio of 0.40 and the superplasticizer dosage of 2.0% - 5.5% by weight of cementitious to obtain the targeted minislump flow values of 25010 mm. Two control mixtures named control-sand (CS) and controlglass (CG) were designed with cementitious content of (90%WC+10%MK) and fine aggregate of 100% sand (mix code: CS) and 100% crushed glass (60%SG+40%MG) (mix code: CG), respectively. To enhance the aesthetic value of SCAM, it is necessary to include glass of larger sizes in the mix proportions. To investigate the influence of larger glass particles on the mechanical properties, the MG aggregate (2.36 5mm) in the control-glass mix were systemically replaced by LG aggregate (5-10mm) at 25%, 50%, 75% and 100% by weight. The mix proportions of the two control mixes and the SCAM mixes are given in Table 2. For the CG-MK20 mix, the mix proportion was similar to that of the control-glass (CG) except that metakaolin was used to replace 20% of white cement.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 4

Table 2: Mix proportions of SCAM mixtures (Kg/m3)

Notation Control-Sand (CS) Control-Glass (CG) CG-LG25 CG-LG50 CG-LG75 CG-LG100 CG-MK20 w/c 0.4 0.4 0.4 0.4 0.4 0.4 0.4 282 282 282 282 282 282 282 Cementitio us WC 635 635 635 635 635 635 565 MK 71 71 71 71 71 71 142 1412 0 0 0 0 0 0 Sand Recycled blue glass cullet SG 0 847 847 847 847 847 847 MG 0 565 424 282 141 0 565 LG 0 0 141 282 424 565 0 % of MG replaced by LG 0 25 50 75 100 SP* 5.5 3.8 3.5 3.0 2.5 2.0 4.0 Mini slump (mm) 243 257 243 245 243 253 245

* % of SP dosage as cementitious weight

2.3. Sample preparation An appropriate mixing sequence and duration are essential for achieving good SCAM samples. The SCAM mixtures were prepared in a standard rotating drum-type mixer with a maximum capacity of 8 kg. Initially, fine aggregates (surface dried) and cementitious materials were mixed for about 90 s to obtain a uniform mix in dry conditions. Then, the superplasticizer (thoroughly mixed with water) was added to the mix, and the mechanical mixing process was resumed for another 90 s. To avoid the dry materials becoming stuck at the bottom part of the mixer, the mixture was mixed manually by turning it over twice or thrice using a steel trowel. Finally, the mixture was mechanically mixed for an additional 2 min to complete the whole mixing process. After the mixing was completed, the fluidity of freshly prepared mortar was evaluated by measuring the spread diameter of the mortar in two perpendicular directions specified by European specification EFNARC [19]. Twelve 4040160 mm prisms were prepared for flexural and compressive strength tests. Six additional prism samples were cast for the determination of water absorption and resistance to acid attack. The drying shrinkage and expansion due to the alkali-silica reaction (ASR) were determined on prisms with dimensions of 2525285 mm. A 707070 mm cube was cast to determine the abrasion resistance. After casting, all the specimens were covered with a thin plastic sheet and left in the laboratory at a temperature of 233 C and 75% relative humility. After demoulding for one day, three 2525285 mm prisms were used to determine the initial drying shrinkage values before they were transferred to a drying chamber at a temperature of 23 C and relative humidity of 50%. The other specimens were stored in a water tank at an average temperature of 233 C until further testing. 2.4. Testing details 2.4.1. Fresh properties A mini-slump flow cone with an internal diameter of 100 mm was used to evaluate the fluidity of fresh mortar mixture as described by European specification EFNARC [19]. Before the test, the truncated cone mould was placed on a clean metal plate and the freshly prepared mortar mixture *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 5

was poured into the cone without any compaction. Once the cone was fully filled with the mortar, the cone was lifted vertically and the spread diameters of the freshly prepared mortar in two perpendicular directions were measured. Occurrence of segregation and/or bleeding, if any, was visually observed and noted during the mini-slump flow test. 2.4.2. Permeable voids and water absorption The permeable voids and water absorption test were conducted to assess the water permeability characteristics of SCAM in accordance with ASTM 642 [20]. For this test, three 4040160 mm prism specimens were cured in water for 90 days. The surfaces of the saturated specimens were dried by removing surface moisture with a towel and the weight was determined (W90). Afterwards, the surface-dried specimens were further dried in an oven at a constant temperature of 1055 C until a constant weight was achieved (Wod). The permeable voids in percentage are given below: Permeable voids = [(W90 Wod)/ V] 100 (1) where V = volume of prism specimen Subsequently, the initial surface absorption (ISA) and final water absorption (FWA) tests were carried out by immersing the oven dried specimens completely in water. The specimens were removed from the water immersion and weighed at 30 min and 96 h to evaluate the mass gained for ISA and FWA, respectively. The ISA and FWA values were determined by the following formulation: ISA = [(W30min Wod)/ Wod] 100 (2) (3) FWA = [(W96h Wod)/ Wod] 100 where W30min = weight of surface dried specimen after 30 min of immersion W96h = weight of surface dried specimen after 96 h of immersion 2.4.3. Flexural strength A three-point flexural strength test in conformity with ASTM C348 [21] was performed at 1, 7, 28 and 90 days after casting. A centre line was marked at the top of the 4040160 mm prism specimens, using a black felt-tip marker perpendicular to its length. The SCAM specimens were tested under a central line load while simply supported over a span of 120 mm. For this test, a universal test machine with a load capacity of 50 kN was used with a displacement rate of 0.10 mm/min set. 2.4.4. Equivalent compressive strength The equivalent compressive strength test was carried out according to ASTM C349 [22]. The compressive strength was determined using a Denison compression machine with a load capacity of 3000 kN on the broken pieces (portions of the prisms broken in flexure). The reported test results were the average of six measurements. 2.4.5. Drying shrinkage A modified British standard (BS ISO, Part 8: 1920) method was used for the drying shrinkage test in this study [23]. After demoulding, the initial length of three 2525285 mm mortar bar specimens was measured. After the reading, the specimens were conveyed to a drying chamber at a temperature of 23 C and relative humidity of 50% until further measurements at 1st, 4th, 7th, 28th, 56th and 90th days, and the final length measurement was recorded at the 112th day. The *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 6

length of each specimen was measured within 15 min after it was removed from the drying chamber. 2.4.6. Expansion due to alkali-silica reaction For each mix, three 2525285 mm mortar bar specimens were used for the ASR test in accordance with ASTM C1260 [24] - the accelerated mortar bar method. A zero reading was taken after storing the prisms in distilled water at 80C for 24 h. The mortar bars were then transferred and immersed in 1 N NaOH solution at 80 C until the testing time. The expansion of the mortar bars was measured within 155 s after they were removed from the 80 C water or alkali storage condition by using a length comparator. The measurements were conducted at the 1st, 4th, 7th, 14th, 21st and 28th days. 2.4.7. Abrasion resistance The abrasion resistance test was determined by abrading the surface of the 707070 mm cube specimen after 28 days of water curing as specified by BS 6717 [25]. The test began by the test specimen being placed in contact with an abrasion wheel rotating at the rate of 75 revolutions in 603 s. The dimension of the groove resulting from the abrasive action was measured to evaluate the abrasion resistance of the mortar. A smaller groove indicated a better resistance to abrasion. 2.4.8. Chemical resistance The chemical resistance was studied by immersing the specimens in a sulphuric acid solution in accordance with ASTM C 267 [26]. After the 28 days of curing, three 4040160 mm prisms were removed from the water tank and each specimen was marked and tied with a nylon string around them. After the initial weight was recorded, the specimen was immersed in a 3% H2SO4 solution. The solution was replaced at 4-week regular intervals to ensure consistent acid concentration throughout the test period. The specimens were extracted from the solution and the surfaces were cleaned with a soft nylon brush before their weights were measured. The weight of specimens was measured weekly at the 4th week and subsequently the reading was measured at the 8th week and 12th week. The cumulative mass loss of each specimen expressed as a percentage is given below: Cumulative mass loss = [(Mt Mint)/ Mint] 100 (4) where Mt = mass at time t Mint = initial mass before immersion in sulphuric acid

3. Results and discussion 3.1. Fresh properties The mini-slump flow diameters of control and SCAM fresh mixtures were in the range of 240260 mm specified by European specification EFNARC [19] and the amount of superplasticizer (SP) required for each mixture is presented in Fig. 3. A total replacement of sand with glass aggregates increased the workability due to the fact that glass aggregates have smoother surface layer and nearly zero water absorption properties. It is important to note that for every 25% of MG aggregate replaced by LG aggregate, superplasticizer contents had to be reduced by about 0.5% in order to achieve the same targeted mini-slump flow diameter. A decrease in the superplasticizer dosage with increasing LG aggregate replacement level might be attributed to the larger particle size of LG aggregate which led to a reduction in total surface area per unit *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 7

volume. As shown in Table 2 and Fig. 4, for a given w/c and metakaolin content, an increase in LG aggregate led to a slight increase in bleeding and segregation. This effect was more pronounced when more LG aggregate was incorporated. This might be due to the fact that the LG aggregate had a flat and smooth shape which was prone to segregation. As seen in Fig. 4, the replacement of cement by metakaolin strongly influenced the fluidity. Increasing metakaolin from 10% to 20% enhanced the cohesiveness of the mortar which in turn reduced the bleeding and segregation. Even though a higher superplasticizer dosage of 4.0 was required for the CSMK20 mixture, no bleeding and segregation could be seen.

6.0

S u p erp la sticizer d o sa g e (% )

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 ControlSand ControlG las s CG -LG 25 CG -LG 50 CG -LG 75 CG -LG 100 CG-M K 20

Fig. 3. Fluidity (% of SP) of control and SCAM mixes

Fig. 4. Mini-slump flow appearance of SCAM mixes *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 8 80 mm

3.2. Surface appearance The pictures of blue glass aggregates exposed in SCAM after cutting the samples by a diamond saw are given in Fig. 5. As the replacement ratio of MG aggregate by LG aggregate increased, the aesthetics of the surface appearance improved although there is no objective method of measuring aesthetic value.

10 mm Fig. 5. Architectural mortar featuring different particle sizes of blue glass aggregates 3.3. Water absorption and permeable voids The effects of LG aggregate and metakaolin content on the initial surface absorption (ISA), final water absorption (FWA) and permeable voids values of control and SCAM specimens are illustrated in Fig. 6. It can be seen that the ISA increased with an increase in LG aggregate content. As the LG was use to replace 100% MG, the ISA value was increased from 3.02% to 3.47%. This might be due to the large amount of LG aggregate leading to an increase in bleeding and a more porous microstructure.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 9

IS A (30min) 20 18 ISA , F W A and p erm eab le void s (% ) 16 14 12 10 8 6 4 2 0 ControlS and ControlG las s CG -LG 25

F W A (96h)

permeable voids

CG -LG 50

CG -LG 75

CG -LG 100

CG -M K 20

Fig. 6. ISA, FWA and permeable voids of control and SCAM mixes The FWA is considered to be the water that is continually absorbed at the inert zone of the mortar after the ISA test. It can be seen from Fig. 6 that all the mixtures continued to absorb water for an extended period from 30 min (average of 2.4% for ISA) up to 96 h (average of 7.5% for FWA). An opposite trend of absorption rate could be seen between the ISA (near surface) and the FWA (inert mortar). The water absorption ratio (FWA/ISA) decreased with increase of LG aggregate content. The decrease in the water absorption ratio with increasing LG aggregate replacement level might be attributed to the smooth surface of LG aggregate which causes lower air content in the produced mortar. On the contrary, the irregular shape of MG aggregate resulted in a larger relative surface area that trapped more air in the mortar matrix. The FWA was reduced by the inclusion of recycled glass as total replacement of sand in the mortar. This could be due to the fact that glass by its nature is an impermeable material, resulting in lower air void contained in the mortars. It was also noted that an increase in metakaolin content led to an increase in the FWA values. Similar observations concerning the effect of metakaolin on the water absorption characteristic of concrete was also reported by Khatib and Clay [27]. It can be concluded that the FWA (after 96 h) and permeable voids values were closely related, whereas as the FWA increased the permeable voids also increased. 3.4. Flexural strength The results of flexural strength of the control and SCAM mixes incorporating different sizes of glass aggregates are given in Fig. 7. The results revealed that the flexural strength of all the SCAM specimens increased when the curing age was extended. After 7 days of water curing, the flexural strength reached approximately 90% of the 28-day flexural strength. This indicated a rapid hydration during the early 7 days. It can be noted that at the age of 1 and 7 days, the *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 10

flexural strength of all the SCAM specimens was comparable, while at the ages of 28 and 90 days, the flexural strength seemed to be affected more by the grading of the glass aggregates than by the cement matrix characteristic. The flexural strength at 28 and 90 days decreased as the replacement ratio of MG aggregate with LG aggregate increased. A possible reason for this is that the LG aggregate had a smoother surface which could significantly weaken the bonding strength between the LG aggregate and the surrounding cement mortar, thus giving a lower strength. On the contrary, the smaller particle size of MG aggregate enhanced the aggregatecement matrix bonding strength.

12.00 10.00 8.00 6.00 4.00 2.00 0.00 0 20 40 60 80 100 C ontrol-S and C G-LG25 C G-LG75 C G-M K 20 C ontrol- Glass C G- LG50 C G- LG100

F lex u ra l stren g th (M P a )

Day
Fig. 7. Flexural strength of control and SCAM mixes Besides, when glass aggregates were used as a total replacement for sand, an average reduction of 36.7% and 33.6% in 28-day and 90-day flexural strength were observed, respectively. This is because the bonding strength of cement pastes with smooth surface glass aggregate was weaker than that with rough sand. It was also observed that with an increase in metakaolin from 10% to 20%, the flexural strength was decreased by about 7.5% at 90 days. In terms of strength properties, Li and Ding [28] suggested that the best performance of metakaolin replacement was achieved at 10%. 3.5. Equivalent compressive strength The compressive strength results of the control and SCAM mixes are presented in Fig. 8. The results show that the rate of gain in compressive strength was rapid up to 7 days and then slowed down afterward. At later ages of 28 and 90 days, it was observed that the compressive strength decreased with increasing LG aggregate content in a similar manner with that observed in flexural strength. As mentioned earlier, this might be due to the fact that the LG aggregate had relatively smoother surfaces and often contained a series of stepped fractures which tended to *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 11

decrease the bonding strength between the glass aggregate and the cement matrix.
60.00

C o m p ressiv e stren g th (M P a )

50.00 40.00 30.00 20.00 10.00 0.00 0 10 20 30 40 50 60 70 80 90 100 C ontrol- Sand C ontrol- Glass C G-LG25 C G-LG50 C G-LG75 C G-LG100 C G-M K 20

Day
Fig. 8. Compressive strength of control and SCAM mixes The reduction rate of SCAM incorporating 100% of glass aggregates as sand replacement was much lower in compressive strength than in flexural strength. It was possibly because the influence of weaker bonding by the presence of glass aggregates in cement mortar was much lower in compression failure than that on bending. As the flexural strength was found to be about one-fourth to one-sixth of the compressive strength, it is always interesting to establish a relationship between the two parameters. The relationship between the average flexural strength and the compressive strength is plotted in Fig. 9 which shows a strong linear correlation.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 12

12 10 8 6 y = 0.118x + 1.36 4 2 0 0 10 20 30 40 50 60 70 R = 0.96

2

F lex u ra l stren g th (M P a )

C o mpress ive stre ng th (M P a )

Fig. 9. Relationship between flexural and compressive strength of control and SCAM mixes 3.6. Drying shrinkage The drying shrinkage results of the control and SCAM mixes at various ages up to 112 days are shown in Fig. 10. In general, the drying shrinkage values of all samples were similar up to 4 days, while considerable differences were observed for the later period. The results show that the particle sizes of glass aggregates had a significant effect on the drying shrinkage. The reduction of drying shrinkage values by incorporating LG aggregate may be due to the lower absorption capacity of glass aggregates.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 13

0.09 0.08 0.07

L en g th ch a n g e (% )

0.06 0.05 0.04 0.03 0.02 0.01 0.00 0 20 40 60 80 100 120 C ontrol-S and C G-LG25 C G-LG75 C G-M K 20 C ontrol- Glass C G-LG50 C G-LG100

Day
Fig. 10. Drying shrinkage curve of control and SCAM mixes In comparison, a control-sand mix showed the highest value of drying shrinkage at 1 day up to 56 days. However, the drying shrinkage was stable as the curing time continued to increase up to 112 days. Also, increasing the replacement of white cement by metakaolin from 10% to 20% slightly reduced the drying shrinkage by 4%. This was because of the hydration of cement and secondary pozzolanic reaction of metakaolin which used up significant amounts of the free water in the cement mortar [29]. The test results at 56 days for control and all the SCAM specimens satisfied the drying shrinkage requirement (< 0.075%) of the Australian Standard AS 3600 [30]. 3.7. Alkali-silica reaction The ASR results of the control and SCAM mixes are shown in Fig. 11. As seen in the figure, the expansions due to ASR were less than 0.10% at 14 days. The use of metakaolin admixture (10% to 20%) was found to be able to successfully mitigate ASR expansion. It was also noticed that when the particle size of blue glass aggregates was in the range of 2.36-5mm and 5-10mm, the deleterious effect on ASR expansion was negligible.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 14

0.12 0.10 C ontrol- Sand C ontrol- Glass C G-LG25 C G-LG50 C G-LG75 C G-LG100 C G-M K 20

A S R ex p a n sio n (% )

0.08 0.06 0.04 0.02 0.00 0 -0.02 5 10 15 20

25

30

Day
Fig.11. ASR reactivity of control and SCAM mixes 3.8. Abrasion resistance The effect of LG aggregates and metakaolin replacement on the abrasion resistance of SCAM is shown in Fig. 12. As seen in the figure, the abrasion resistance was reduced by the inclusion of glass aggregates as total sand replacement in the mortar. Also, the abrasion resistance was reduced with an increase in the LG aggregate content. According to a previous study [31], the percentage wear of coarser glass aggregates was higher than that of finer glass aggregates. It was found that SCAM incorporating 20% of metakaolin exhibited a weaker abrasion resistance than all other SCAM mixes. This is consistent with a previous study [4] which found that the ability of concrete to withstand abrasion decreased with a decrease in concrete strength.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 15

23.0

A b ra sio n resista n ce (m m )

22.0 21.0 20.0 19.0 18.0 17.0 16.0 15.0 ControlS and ControlG las s CG -LG 25 CG -LG 50 CG -LG 75 CG -LG 100 CG -M K 20

Fig. 12. Abrasion resistance of control and SCAM mixes 3.9. Chemical resistance The loss of mass of SCAM mixes upon immersing in a 3% solution of H2SO4 up to 12 weeks is shown in Fig. 13. The control-sand specimen showed the most significant deterioration with respect to loss of mass. The inclusion of glass aggregate in mortar significantly reduced the mass loss. Up to 12 weeks of immersion, the loss of mass of SCAM mixes incorporating 100% of glass aggregates ranged from 44.0% to 45.9%. When it was closely examined, the significant increase in the rate of mass loss was in the first 4 weeks. The loss of mass during this period was probably related to the removal of material from the surface of the specimens by degradation of the cementitious matrix. Besides, the rate of mass loss decreased with additional immersion time. This can be explained by the fact that after the deterioration of the surface layer, a higher percentage of the inert glass aggregates which had better acid resistance than the cement mortar were exposed to the sulphuric acid solution. However, due to the degradation of the cement matrix, some of the glass aggregates were dislodged from the specimens. It can be observed that the composition of cementitious materials has a pronounced effect on the mass loss results. For CG-MK20 specimens, the mass loss at 12 weeks was about 22.3% lower than that of the Control-Glass specimens. This can be explained by the increase in metakaolin content in the cement matrix which reduced the amount of calcium hydroxide content available for the acid-base reaction, thus resulting in a lower rate of acid attack [32].

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 16

60.00 50.00 40.00 30.00 20.00 10.00 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C ontrol-S and C G-LG25 C G-LG75 C G-M K 20 C ontrol-Glass C G- LG50 C G- LG100

M a ss lo ss (% )

W eek
Fig. 13. Rate of mass loss over time of control and SCAM mixes 4. Conclusion The results of this study concluded that it is feasible to utilize 100% recycled glass as decorative aggregates in the production of self-compacting based architectural mortar. Glass-modified cement mortar exhibited lower strength values and abrasion resistance as compared to sandcontrolled cement mortar. However, fully replacement of sand by recycled glass has improved workability, resistance in term of water absorption, drying shrinkage and acid attack of cement mortar. Also, it seems that SCAM products of various appearances for diverse applications can be achieved by using combinations of different particle sizes of glass aggregates. The overall performance of the architectural mortar was studied. Based on the laboratory results, the following conclusions may be drawn: Incorporating glass aggregates, particularly LG aggregate in SCAM increased the fluidity which in turn reduced the use of superplasticizer dosage. However, incorporating metakaolin as a supplementary mineral admixture in the mortar mixtures increased the superplasticizer dosage to maintain workability. The presence of LG aggregate in the SCAM mixtures gradually decreased the final water absorption; however, increasing the level of metakaolin admixture increased the water absorption values. In all cases, the final water absorption and permeable voids were directly related. The presence of glass aggregates caused a reduction in both flexural and compressive strengths. Also, increasing the LG aggregate and metakaolin content led to a decrease in the flexural and compressive strength, particularly between the curing age of 28 and 90 days. The drying shrinkage of SCAM decreased as the LG aggregate increased. The use of *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 17

metakaolin also marginally reduced the drying shrinkage. Using 10% of metakaolin successfully mitigated the ASR expansion of mortar. Generally, the size of blue glass aggregates ranged from of 2.36 to 5 mm and 5 to 10 mm was found to have negligible ASR expansion in SCAM. The SCAM specimens showed lower abrasion resistance when the LG aggregate and metakaolin content was increased from 0% to 100% and 10% to 20%, respectively. The use of glass aggregates and metakaolin in SCAM significantly enhanced the ability to resist acid attack. However, the acid resistance of SCAM slightly decreased when LG aggregates were used to replace MG aggregate.

Acknowledgements The authors would like to thank The Hong Kong Polytechnic University and SHK Properties for funding supports.

Reference [1] Environmental Protection Department. Future landfill development in Hong Kong. The Government Printer, Hong Kong; 2006. [2] Lam CS, Poon CS, Chan D. Enhancing the performance of pre-cast concrete blocks by incorporating waste glass ASR consideration Cem Concr Compos 2007;29(8):616-25. [3] Chen J, Poon CS. Photocatalytic activity of titanium dioxide modified concrete materials Influence of utilizing recycled glass cullets as aggregates. J Envir Manage 2009;90(11):3436-42. [4] Poon CS, Chan D. Effects of contaminants on the properties of concrete paving blocks prepared with recycled concrete aggregate. Constr Build Mater 2007;21(8):164-75. [5] Chen J, Poon CS. Photocatalytic construction and building materials: From fundamentals to applications. Build Envir 2009;44(9):1899-906. [6] Siddique R. Waste materials and by-products in concrete. Springer-Verlag Berlin Heidelberg; 2008. p. 151. [7] Corinaldesi V, Gnappi G., Moriconi G., Montenero A. Reuse of ground waste glass as aggregate for mortars. Waste Manage 2005;25(2):197-201. [8] Ismail ZZ, AL-Hashmi EA. Recycling of waste glass as a partial replacement for fine aggregate in concrete. Waste Manage 2009;29(2):655-9. [9] Kou SC, Poon CS. Properties of self-compacting concrete prepared with recycled glass aggregate. Cem Concr Compos 2009;31(2):107-13. [10] Topu B, Canbaz M. Properties of concrete containing waste glass. Cem Concr Res 2004;34(2):267-74. [11] Shao Y, Lefort T, Moras S. Damian Rodriguez Studies on concrete containing ground waste glass. Cem Concr Res 2000;30(1):91-100 [12] Shi C, Wu Y, Riefler C, Wang H. Characteristics and pozzolanic reactivity of glass powders. CemConcr Res 2005;35(5):987-93. [13] Schwarz N, Cam H, Neithalath N. Influence of a fine glass powder on the durability characteristics of concrete and its comparison to fly ash. Cem Concr Compos 2008;30(6):486-96. [14] Shayan A, Xu A. Value-added utilisation of waste glass in concrete. CemConcr Res 2004; 34(1): 81-9. [15] Topu B, Boa AR, Bilir T. Alkalisilica reactions of mortars produced by using waste *First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com Page 18

glass as fine aggregate and admixtures such as fly ash and Li2CO3. Waste Manage 2008;28(5):878-84. [16] Lee G, Ling TC, Wong YL, Poon CS. Effects of crushed glass cullet sizes, casting methods and pozzolanic materials on ASR of concrete blocks. Constr Build Mater 2011, doi:10.1016/j.conbuildmat.2010.12.008. [17] Taha B, Nounu G. Properties of concrete contains mixed colour waste recycled glass as sand and cement replacement. Constr Build Mater 2008;22(5):713-20. [18] ASTM C494. Standard specification for chemical admixtures for concrete. US: American Society of Testing Materials; 2010. [19] EFNARC. Specification and guidelines for self-compacting concrete. Norfolk, UK, English ed.: European Federation for Specialist Construction Chemicals and Concrete System; February 2002. [20] ASTM C 642. Standard test method for density, absorption, and voids in hardened concrete. US: American Society of Testing Materials; 2006. [21] ASTM C348. Standard test method for flexural strength of hydraulic-cement mortars. US: American Society of Testing Materials; 2008. [22] ASTM C349. Standard test method for compressive strength of hydraulic-cement mortars (using portions of prisms broken in flexure). US: American Society of Testing Materials; 2008. [23] BS ISO 1920-8. Determination of drying shrinkage of concrete for samples prepared in the field or in the laboratory. UK: British Standard Institution; 2007. [24] ASTM C 1260. Standard test method for potential alkali reactivity of aggregates (mortar-bar method). US: American Society of Testing Materials; 2007. [25] BS 6717. Precast, unreinforced concrete paving blocks-Method for measuring abrasion resistance. UK: British Standards Institution; 2001. [26] ASTM C 267. Standard test methods for chemical resistance of mortars, grouts, and monolithic surfacings and polymer concretes. US: American Society of Testing Materials; 2001. [27] Khatib JM, Clay RM. Absorption characteristics of metakaolin concrete. Cem Concr Res 2004;34(1):19-29. [28] Li Z, Ding Z. Property improvement of Portland cement by incorporating with metakaolin and slag. Cem Concr Res 2003;33(4):579-84. [29] Brooks JJ, Johari MMA. Effect of metakaolin on creep and shrinkage of concrete. Cem Concr Compos 2001;23(6):495-502. [30] AS 3600. Concrete structures-incorporating AMD 1: May 2002 and AMD 2. Australia: Australia Standard; October 2004. [31] HDR Engineering, Inc. Glass cullet utilization in civil engineering application. Nebraska State Recycling Association; 1997. [32] Al-Akhras NM. Durability of metakaolin concrete to sulfate attack. Cem Concr Res 2006;36(9):172734.

*First author: t.ling.1@bham.ac.uk; tcling611@yahoo.com

Page 19