Sei sulla pagina 1di 13
Construction and Building Materials 146 (2017) 238–250 Contents lists available at ScienceDirect Construction and

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www. elsevier.com/loc ate/conbuildmat Recycled glass fiber reinforced polymer additions to

Recycled glass fiber reinforced polymer additions to Portland cement concrete

Alireza Dehghan a , Karl Peterson a , , Asia Shvarzman b

a Univ. of Toronto, Dept. of Civil Eng., 35 St. George St., Toronto, ON M5S 1A4, Canada b Antex Western, 1340 Church Ave., Winnipeg, MB R3C 2L4, Canada

Western, 1340 Church Ave., Winnipeg, MB R3C 2L4, Canada highlights Recycled glass fiber polymers did not

highlights

Recycled glass fiber polymers did not cause expansive alkali silica reactions. Recycled glass fiber polymers exhibited pozzolanic behavior. An absorption test for recycled glass fiber polymer additions is presented.

article info

Article history:

Received 24 October 2016 Received in revised form 2 April 2017 Accepted 3 April 2017

Keywords:

Recycled glass fiber reinforced polymer Alkali silica reaction Drying shrinkage Concrete microscopy

abstract

With recent developments in grinding and sorting technology it is possible to recover glass fibers from waste glass fiber reinforced polymers (GFRPs). The recycled fibers still retain some of the polymer and filler materials, but have the potential to provide some of the same benefits achieved by conventional fiber additions. The research program explored the influence of recycled GFRP on compressive strength, splitting tensile strength, and drying shrinkage in concrete, and alkali silica reaction (ASR) expansion in accelerated mortar beam and concrete prism tests. Compressive strength and drying shrinkage were not improved by recycled GFRP additions at a substitution level of 5 wt% of the coarse aggregate, but splitting tensile strengths were improved in most cases. Negligible expansion was observed from the ASR testing. A scanning electron microscope investigation of the concrete prisms indicated a pozzolanic reaction of the glass fibers.

2017 Elsevier Ltd. All rights reserved.

1. Introduction

Fiber reinforced polymer (FRP) is a composite material which is usually made with glass (GFRP), carbon (CFRP), or aramid (AFRP) fibers dispersed in a thermoset polyester resin and has a wide range of applications in the construction industry. Among GFRPs, E-glass is the most common reinforcement, and represents approx- imately 99% of the commercial market [1] . The ‘‘E” in E-glass is a carryover from its initial application, as used in electrical standoff insulators [2] . The compositional ranges for E-glass used in general applications (such as GFRP) are outlined in ASTM D578 [3] ( Table 1 ). At the end of the life cycle of FRPs, fibers cannot be easily sep- arated from the resin and the resin itself cannot be easily decom- posed or recycled. Hence, landfill and incineration are the most common methods for FRP waste management [4,5] . Globally, GFRP

Corresponding author. E-mail address: karl.peterson@utoronto.ca (K. Peterson).

0950-0618/ 2017 Elsevier Ltd. All rights reserved.

production is estimated at 8 million metric tons annually, with GFRP waste production at 1.5 million metric tons [6] . Given the potential negative environmental impacts of waste FRP, the mate- rial has gained the attention of other industries to develop tech- niques and methods to recycle FRPs. Thankfully, recent advances in grinding and sorting technologies allow for the partial recovery of fibers from FRP and has made their utilization in Portland cement concrete an option. Considering that grinding and sorting equipment are readily available, and that the process produces negligible atmospheric pollution in terms of volatile organic com- pound emissions, size reduction by mechanical recycling is pre- ferred over other recycling processes [7] . The topic of recycled GFRP in concrete necessarily overlaps with a number of different areas of research, including: glass fiber rein- forced concrete (GFRC), waste glass powder concrete, and concrete made using waste plastic fine and coarse aggregate. As such, recy- cled GFRP combines some of the beneficial aspects of fibers and powdered waste glass, as well as some of the shortcomings of waste plastic aggregates.

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

239

Table 1 Certified chemical composition for glass fiber products used in general applications [3] .

Chemical

% by Weight

B 2 O 3 CaO

0–10

16–25

Al 2 O 3 SiO 2 MgO Na 2 O+K 2 O TiO 2

12–16

52–62

0–5

0–2

0–1.5

Fe 2 O 3 Fluoride

0.05–0.8

0–1

ACI 544.R-96 [8] constitutes a state-of-the-art report on fiber reinforced concrete, with an entire chapter dedicated to GFRC. Fibers in concrete are used to control cracking and to improve ten- sile and flexural strength. Early formulations of GFRC utilized E- glass fibers, but their use declined after the development of alkali resistant (AR) glass fibers, which performed much better over the long-term in the high-pH environment of concrete pore water. More recently, basalt fibers, made from melting basaltic rock, have been used to produce basalt fiber reinforced concrete (BFRC) [9– 11] . In this research, fibers recovered from recycled GFRP are uti- lized, but the fibers tend to occur in grouped masses bound by residual resin, as opposed to individual clean fibers. Waste glass is increasingly being utilized in concrete, both as an aggregate and as a powdered pozzolanic addition to concrete. Shi and Zheng [12] provide a thorough review of the topic, and it remains a very active area of research today. However, the majority of the research to date focuses on waste soda lime glass sources from containers or float/plate glass, with relatively few studies on the pozzolanic aspects of recycled GFRP. Xu et al. [13] investi- gated incinerated waste GFRP from the automotive industry, and used the ash as a pozzolanic additive. Chen et al. [14] investigated the use of ground waste E-glass fibers left over from circuit board manufacture that had never been used in FRP. They documented improvements in compressive strength, chloride and sulfate resis- tance. Similarly Mastali et al. [15] reported improvements in com- pressive and flexural strength for recycled glass fibers recovered from woven fiber sheets that had never been used in FRP. However, ground recycled GFRP, a combination of E-glass fibers, resins, and filler materials, is more problematic, as the resins and fillers pro- vide no pozzolanic benefit. Gu and Ozbakkaloglu [16] conducted an extensive review of the utilization of recycled plastics in concrete from the standpoint of plastic fine aggregate, plastic coarse aggregate, and plastic fibers. While recycled plastic fibers tended to improve mechanical prop- erties of concrete, recycled plastic aggregates led to reductions in compressive and tensile strength, and in most cases increased dry- ing shrinkage. A common concern with the usage of recycled GFRP in cemen- titious binders is the potential for interference with the mechanical performance, particularly reductions in compressive strength. The

feasibility of recycled GFRP concrete has been explored by a num- ber of researchers worldwide, with a comprehensive review recently provided by Yazdanbakhsh and Bank [17] . Asokan et al. [18] reported reduced compressive strengths with increasing recy- cled GFRP powder substitution when cured in water at 20 C, but found increased compressive strengths compared to the control when oven cured at 50 C. Tittarelli and Moriconi [19] and Tittarelli and Shah [20] reported reductions in compressive strength with recycled GFRP powder additions, but some improvements in terms of reduced drying shrinkage, and lower values for capillary water absorption. Correia et al. [21] explored the use of fines produced during the cutting of pultruded GFRP, and found similar reductions on compressive strength. Osmani [22] also reported reductions in compressive strength for concrete produced with recycled GFRP powder. Alam et al. [23] and Yazdanbakhsh et al. [24] both explored the substitution of larger aggregate sized FRP scrap parti- cles, and reported reductions in both compressive and flexural strength. Alternatively, García et al. [25] explored the use of fibers recovered from recycled FRP, termed ‘‘glass fiber fluff,” and reported improvements in compressive and flexural strength when grinding and sieving is optimized. Expansive ASR is still cited as a concern for concrete as E-glass fibers are not stable in the high-alkali environment of the pore water [26,27] . Meanwhile the potential for powdered waste glass to mitigate ASR has been extensively documented, but not neces- sarily from the standpoint of waste E-glass [28–41] . Chen et al. [14] found negligible expansion due to alkali silica reaction (ASR) for concrete made with powdered recycled E-glass fibers that had never been used in GFRP. García et al. [25] reported expansions of <0.04% for concrete beams produced with glass fiber fluff recov- ered from recycled GFRP. Tittarelli and Moriconi [19] tested recy- cled GFRP powder using the recently withdrawn standard ASTM C289 [42] , and found it to be innocuous. However, it is widely rec- ognized that ASTM C289 is not a reliable test for predicting the reactivity of carbonate aggregates [43] , and calcium carbonate is a common filler material in FRP.

2. Materials and methods

Descriptions of the commercially produced GFRPs that were investigated in this study are presented in Table 2 , and images of the fibers recovered from recycled GFRP are provided in Fig. 1 . The recycled GFRP was produced using a model GM- 2411-50 ECO Grinder TM [44] single stage hammer mill grinding system with a 19 mm (3/4 in.) screen coupled to a pneumatically fed hopper with a dust collection bag and an enclosed auger feed to a series of 9.5 mm (3/8 in.) and 4.75 mm (3/16 in.) perforated opening trommel (rotary) screens. The GFRP feed consisted of sheets with a maximum thickness of 25.4 mm (1 in.) and nominal dimensions of 150 914 mm (6 36 in.). Table 3 provides a summary of the wt% glass fiber fluff retained by the screens. Materials separated by both screens were recombined for the purposes of this study. The bottom fines were not included. The results of sieve analyses performed on the combined materials are provided in Fig. 2 .

2.1. Water absorption, density, loss on ignition, and fiber cluster length

From each recycled GFRP source three representative samples, each with an approximate mass of 100 g, were produced by the ASTM C702 [45] quartering method for a determination of water absorption. Samples were completely immersed in tap water for 72 ± 2 h and stirred at least once every 24 h for one

Table 2 GFRP types, content, and abbreviations used in this study.

Type

Resin

Glass content (vol.%) as manufactured

Abbreviation

Structural sheet molding composite Structural sheet molding composite Structural sheet molding composite Light resin transfer mold E-glass fiber

Bisphenol-A epoxy vinyl ester Novolac-based epoxy vinyl ester Flame retardant epoxy vinyl ester Unsaturated polyester None

40

EVE1

40

EVE2

40

EVE3

25

UP

100

Virgin

240

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

/ Construction and Building Materials 146 (2017) 238–250 Fig. 1. Overall appearance of the recycled GFRP

Fig. 1. Overall appearance of the recycled GFRP materials (a–d), and virgin E-glass fibers (e).

Table 3 Size-separated materials expressed as wt% of initial GFRP feed.

Type

wt%

9.5 mm screen

4.75 mm screen

Bottom fines

Loss to dust collection

EVE1

24

18

55

3

EVE2

27

7

63

3

EVE3

27

8

58

7

UP

26

4

71

Negligible

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

241

Construction and Building Materials 146 (2017) 238–250 241 Fig. 2. Size distribution of recycled GFRP materials.

Fig. 2. Size distribution of recycled GFRP materials.

Table 4 Density, water absorption, glass fiber content, and length of recycled FRPs.

Source

Density (g/cm 3 )

Absorption (%)

E-glass content (wt%)

Max. fiber cluster length (mm)

Avg. fiber cluster length (mm)

UP

1.63 ± 0.02 1.66 ± 0.02 1.56 ± 0.02 1.60 ± 0.02 2.60 ± 0.04

63.1 ± 3.4 30.1 ± 4.9 25.5 ± 4.5 28.8 ± 6.4 0.4 ± 0.05

54.3 ± 1.2 43.1 ± 1.0 41.1 ± 0.8 42.0 ± 0.5 98.1 ± 0.0

36.4

21.7 ± 6.3 17.2 ± 5.7 21.2 ± 6.3 19.5 ± 7.5 25.4 ± 0.4

EVE1

26.8

EVE2

33.5

EVE3

31.3

Virgin

25.6

minute to ensure complete submersion. After soaking, the excess water was dec- anted over a 75 m m sieve to avoid loss of material, and the samples placed in indi- vidual open containers. A saturated surface dry (SSD) condition was gradually achieved through exposure over a period of hours in an environmental chamber at 35 ± 2 C and 15 ± 2% relative humidity (RH) with frequent stirring to ensure homogenous drying. To determine if surface moisture was still present, the contain- ers were inverted onto an absorbent sheet of unbleached paper towel to check whether or not the recycled GFRP material would dampen the paper. If any mois- ture was observed on the paper, the specimen was returned to the environmental chamber and re-tested in the same manner until it reached a SSD condition. After achieving an SSD condition, the samples were weighed, and then dried in a vented oven at 100 ± 2 C. Earlier experiments conducted at a drying temperature 105 ± 5 C resulted in a strong odour and color change of the recycled GFRP, so the alternative drying approach at 100 ± 2 C was employed. A helium pycnometer was used to measure the densities of the recycled GFRP. Three representative dry samples with an approximate mass of 4–5 g were used and the volume of each sample was measured three times. The averages of nine readings are provided in Table 4 . UN1046 compressed helium as recommended by the manufacturer at a pressure of 1.2 kg/cm 2 was used in this study. Water absorption values for the recycled GFRPs are also listed in Table 4 . Loss on ignition was measured with three representative test specimens, each with an approximate mass of 5 g and conditioned at 23 ± 2 C and 50 ± 10% RH to meet the requirements of ASTM D2584 [46] . The samples were heated in a muffle furnace at 300 C for 2 h, then cooled to room temperature in a desiccator, and

weighed. Next, the samples were placed back into the muffle furnace, heated at 560 C for 2 h, cooled, and reweighed at room temperature. Preliminary investiga- tions found that heating at 560 C for 2 h was sufficient to remove all organic mate- rial. Finally, the samples were heated to 960 C for 2 h, cooled, and the mass of the residues measured at room temperature. Negligible mass loss was found between 560 C and 960 C, and indicated minimal inclusion of carbonate-based fillers. The difference between initial mass and the loss on ignition at 560 C was used as an approximation of glass content. Mass losses are presented in Fig. 3 . The glass con- tents reported for the crushed and sieved GFRP by ignition were slightly higher than the as-manufactured GFRP. The enrichment may partly be attributed to the loss of resin fines during the crushing and sieving operations. To characterize fiber lengths, representative 5 g samples were obtained, and fiber clusters manually plucked from the fluff, and images recorded with a flatbed scanner ( Fig. 4 ). Digital image processing was used to obtain the fiber cluster lengths reported in Table 4 .

2.2. ASR testing

2.2.1. Accelerated mortar bar test Susceptibility to ASR was tested in accordance with CSA A23.2-25A [47] on four 25 25 285 mm mortar bars, but with 5 wt% replacement of the coarse aggregate (an inert crushed dolostone) with recycled GFRP. CSA A23.2-25A was followed,

dolostone) with recycled GFRP. CSA A23.2-25A was followed, Fig. 3. Mass loss of recycled GFRPs and

Fig. 3. Mass loss of recycled GFRPs and virgin E-glass fibers at different temperatures.

242

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

/ Construction and Building Materials 146 (2017) 238–250 Fig. 4. Fiber clusters manually plucked from glass

Fig. 4. Fiber clusters manually plucked from glass fiber fluff (a–d), and virgin E-glass (e).

Table 5 Mix designs of mortar mixtures for the accelerated mortar bar test.

Source

Recycled GFRP (kg/m 3 )

Crushed agg. (kg/m 3 )

GU cement (kg/m 3 )

Water (kg/m 3 )

Control

0.0

1327.1

589.8

294.9

UP

65.7

1247.9

583.8

291.9

EVE1

65.7

1248.3

583.9

292.0

EVE2

65.6

1247.1

583.4

291.7

EVE3

65.7

1247.6

583.6

291.8

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

Table 6 Fresh properties of the mortar mixtures.

243

Source

Temp ( C)

Gravimetric air content (%)

 

Mix yield (kg/m 3 )

Flow (%)

Control

25.7

2.3

2243.9

83

UP

23.3

12.3

2032.5

36

EVE1

23.9

4.4

2192.0

67

EVE2

22.8

9.2

2108.9

51

EVE3

23.3

4.4

2191.9

68

Table 7

Mix

designs for concrete prism test.

 

Source

Recycled GFRP, dry mass (kg/m 3 )

Cement (kg/m 3 )

Water (kg/m 3 )

Coarse aggr. dry mass (kg/m 3 )

Fine aggr. dry mass (kg/m 3 )

NaOH pellets (kg/m 3 )

Moisture held by the SSD GFRP (kg/m 3 )

Control

415.0

186.7

1079.5

714.4

1.28

UP

19.5

415.0

186.7

1060.0

701.4

1.28

12.3

EVE1

19.9

415.0

186.7

1060.0

701.4

1.28

6.0

EVE2

18.7

415.0

186.7

1060.0

701.4

1.28

4.7

EVE3

19.2

415.0

186.7

1060.0

701.4

1.28

5.4

Virgin

31.2

415.0

186.7

1060.2

701.6

1.28

0.1

Table 8 Fresh properties of concrete mixtures for concrete prism test.

Source

Temp ( C)

Slump (mm)

Pressure meter air content (vol%)

Unit weight (kg/m 3 )

Control

21.3

145

2.5

2411.1

UP

21.1

75

4.0

2361.8

EVE1

21.3

140

2.9

2408.3

EVE2

21.1

170

3.1

2397.4

EVE3

20.7

170

2.9

2405.3

Virgin

27.1

30

4.6

2344.3

Table 9 Fresh properties of concrete mixtures.

 

Source

Temp ( C)

Slump (mm)

Pressure meter Air content (%)

Unit weight (kg/m 3 )

Control

21.5

125

2.6

2417.1

UP

20.7

80

3.9

2365.7

EVE1

22.0

165

2.9

2417.1

EVE2

22.0

165

2.7

2407.9

EVE3

20.3

175

2.9

2412.9

Virgin

27.7

50

4.3

2372.1

except that the exposure time was extended from 14 d to 28 d. The mix designs of

taining crushed aggregate, and a mass ratio of cement to aggregate of 1:2.25, but

the

mortars are presented in Table 5 , and fresh properties of the mortar mixtures

makes no definitive statement regarding corrections to the water content based

are

presented in Table 6 .

on aggregate absorption. As such, no corrections to the water content were made

CSA A23.2-25A specifies that the aggregate used for mixing be in an oven dry state. Since the recycled GFRP was being treated on the basis of a wt% replacement of aggregate, it was also brought to an oven dry state prior to mixing. CSA A23.2-

to account for water absorption by the aggregate or recycled GFRP. As shown in Table 5 , utilization of recycled GFRP increased the air content. Given the morphol- ogy, air is necessarily entrapped within the glass fiber fluff when added in a dry

25A

prescribes a fixed w / cm (water to cementitious ratio) of 0.50 for mixtures con-

state. From Table 5 , recycled GFRP additions also decreased the flowability of the mixtures. Absorption of mix water by the dry glass fiber fluff likely contributed to the reduction in workability. As such, a vibration table was selected as the

workability. As such, a vibration table was selected as the Fig. 5. 7 and 28 d

Fig. 5. 7 and 28 d compressive strengths for non-air entrained concrete cylinders made with recycled GFRP, without any GFRP (control), and with virgin E-glass fibers.

without any GFRP (control), and with virgin E-glass fibers. Fig. 6. 7 and 28 d tensile

Fig. 6. 7 and 28 d tensile strengths for non-air entrained concrete cylinders made with recycled GFRP, without any GFRP (control), and with virgin E-glass fibers.

244

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

/ Construction and Building Materials 146 (2017) 238–250 Fig. 7. Relationship between water retained in the

Fig. 7. Relationship between water retained in the SSD recycled GFRP materials and virgin fibers and the 7 and 28 d compressive strengths.

and virgin fibers and the 7 and 28 d compressive strengths. Fig. 8. Length changes due

Fig. 8. Length changes due to drying shrinkage of concrete prisms made with recycled GFRP, without any FRP (control), and with virgin E-glass fibers.

method of compaction. The reported flow results were obtained after 10 drops of an ASTM C230 [48] flow table, and all of the mortar mixtures were workable enough to be cast into the molds.

2.2.2. Concrete prism test The risk of expansive ASR was assessed in accordance with CSA A23.2-14A [49] on three 75 75 285 mm prisms per recycled GFRP source, but with one signif- icant modification: the recommended temperature of 38 C was elevated to 50 ± 3 C. The change in temperature was made to accelerate the rate of ASR, and thereby shorten the duration of testing. As with the accelerated mortar bar test, the inert coarse aggregate was substituted by 5 wt% with recycled GFRP. CSA A23.2-14A specifies that concrete be mixed according to CSA A23.2-2C [50] , so aggregate moisture corrections must be accounted for. Prior to mixing, the recycled GFRP materials were prepared in SSD condition using the same procedure used to measure absorption. After achieving a SSD condition, the recycled GFRP materials were sealed in air-tight containers and allowed to equilibrate to room temperature.

Coarse aggregate for mixing was immersed for 24 h prior to use, with excess water decanted prior to mixing. Water clinging to the coarse aggregate was accounted for as a contribution to the mix water. Similarly, fine aggregate of a known moisture condition was used for mixing, and water clinging to the fine aggregate was accounted for as a contribution to the mix water. Mix designs are shown in Table 7 . Results for fresh properties of the concrete mixtures are presented in Table 8 .

2.3. Mechanical properties

A second batch of concrete without the addition of sodium hydroxide was cast for each source of recycled GFRP using the same mixing procedure as described in Section 2.2 . Twelve 100 200 mm concrete cylinders and three 75 75 285 mm concrete prisms were cast to measure compressive strength, tensile strength, and length changes due to drying shrinkage at 23 ± 2 C and 50 ± 4% RH. Samples were cured in lime-saturated water for the periods specified by the corresponding stan- dards. Fresh concrete test results are provided in Table 9 .

Fresh concrete test results are provided in Table 9 . Fig. 9. Relationship between the water

Fig. 9. Relationship between the water retained in the SSD recycled GFRP materials and virgin fibers and drying shrinkage.

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

245

Construction and Building Materials 146 (2017) 238–250 245 Fig. 10. Expansions during accelerated mortar bar test

Fig. 10. Expansions during accelerated mortar bar test for samples made with recycled GFRP, and without GFRP (control).

samples made with recycled GFRP, and without GFRP (control). Fig. 11. Expansions for concrete prisms made

Fig. 11. Expansions for concrete prisms made with recycled GFRP, without any GFRP (control), and with virgin E-glass fibers.

3. Results and discussion

3.1. Concrete strength

CSA A23.2-9C [51] compressive strength results for the concrete samples made with recycled GFRP, virgin E-glass fibers, and with- out any fiber additions are presented in Fig. 5 . Fig. 6 presents the results of CSA A23.2-13C [52] tensile splitting tests, where the sample is failed through indirect tension by applying a compres-

sive load to the diametral plane of a cylinder. Fig. 7 plots the water retention of the SSD recycled GFRP materials versus compressive strength. Compressive strength decreased as recycled GFRP water retention increased, which suggests that the water contained within the SSD recycled GFRP glass fiber fluff may have been released during concrete mixing, locally increasing the water to cementitious ratio ( w / cm ). Alternatively, in the case of ultra high performance concrete (UHPC) with very low w / cm levels, such sources of internal curing water has proved beneficial in terms of

of internal curing water has proved beneficial in terms of Fig. 12. Mass changes for concrete

Fig. 12. Mass changes for concrete prisms made with recycled GFRP, without any GFRP (control), and with virgin E-glass fibers.

246

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

/ Construction and Building Materials 146 (2017) 238–250 Fig. 13. A porous cluster of glass fibers

Fig. 13. A porous cluster of glass fibers in thin section from sample UP concrete as observed in fluorescent mode (a), back scattered electron image (b), and close -up (c) with black arrows showing locations of characteristic X-ray spectra from interior of glass fiber (d) and exterior (reacted) glass fiber (e).

strength gain [53,54]. Tensile strength in most cases increased with recycled GFRP as compared to the control, with the exception of the EVE2 mixture.

3.2. Drying shrinkage

Length changes in concrete prisms due to drying at 23 ± 2 C and a relative humidity of 50 ± 4% were tested according to CSA A23.2-21C [55] with the results shown in Fig. 8 . During the first week of testing, the mixture containing virgin E- glass fibers experienced the least amount of drying shrinkage. However, in the subsequent weeks and months, the virgin mixture performed similarly to the EVE1 and UP mixtures, with shrinkage

values exceeding the control by 17%. The mixtures with the high- est values for water retention (EVE1 and UP) experienced the most drying shrinkage of all the recycled GFRP mixtures ( Fig. 9 ). This result was unexpected, as the purposeful addition of saturated light weight aggregate (LWA) to help reduce autogenous shrinkage is now a widely adopted strategy to counteract shrinkage [53,54,56–58] . However, at the w / cm level of 0.45 explored in this study, autogenous shrinkage is not an issue. Wang et al. [59] reported similar shrinkage issues for concrete made with polypropylene fiber mesh, as well as recycled fibers from tires or waste carpet. They attributed the increase in shrinkage to porosity contributed by the fiber masses. An alternative explanation for the slightly higher levels of drying shrinkage could be that the inher-

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

247

ently lower stiffness of the recycled GFRP provided less resistance to shrinkage [16] .

3.3. Alkali silica reaction

Results of the accelerated mortar bar test are presented in Fig. 10. The expansion values for all of the mortar mixes made with recycled GFRP in this study were below the recommended 0.1% limit. Further- more, results of the concrete prism test for the five types of the recy- cled GFRP are provided in Fig. 11, and the expansion values are all below the recommended 0.4% expansion limit. Changes in mass were recorded, and are shown in Fig. 12. The concrete prisms made with recycled GFRP (except EVE1) experienced mass losses between the initial reading (measured after demolding) and up to 14 d of exposure at 50 C. This unusual reduction in mass could possibly be attributed to water from the SSD mats of the recycled glass fibers migrating to the exterior of the prisms at the elevated temperature of 50 C. Drying of the prisms due to the escape of water from the sealed pails was ruled out, since the level of water ponded at the base of the pails appeared constant during testing. Furthermore, the absorbent sheets lining the inside walls of the pails remained fully saturated throughout the duration of the test.

3.4. Microstructure of concrete prisms

Figs. 13–16 show images of recycled GFRP and virgin glass fibers in polished thin sections prepared from the concrete prisms after four months of ASR testing. Prior to sectioning, a fluorescent-dyed low viscosity epoxy resin was used to fill the capillary pores and air voids present in the concrete. In the flu- orescent images darker regions (less fluorescence) represent areas with low porosity, and brighter regions (more fluores- cence) represent areas with high porosity [60] . ASR-related cracking was not observed in any of the samples. Typically, con- crete that has undergone expansive ASR will exhibit deposits of alkali silica gel lining air-voids or within cracks. Alkali-silica gel deposits were not directly observed in the samples, although in some cases it appeared that the capillary pores of the hardened cement paste immediately surrounding some recycled GFRP fragments may have been filled/densified through the uptake of alkali-silica gel ( Fig. 13 a). Concrete containing recycled GFRP from the higher absorption UP material more frequently exhibited porous mats of glass fibers ( Fig. 13 ) as compared to concrete made with the relatively lower-absorption recycled GFRP sources ( Figs. 14 and 15 ). Fig. 16 shows virgin E-glass fibers.

Figs. 14 and 15 ). Fig. 16 shows virgin E-glass fibers. Fig. 14. A fragment of

Fig. 14. A fragment of flame retardant (brominated) epoxy vinyl ester resin with embedded glass fibers in thin section from sample EVE1 as observed in fluorescent mode (a), backscattered electron image (b), and close-up (d), showing location (black arrow) of characteristic X-ray spectrum collected from the FRP resin.

248

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

/ Construction and Building Materials 146 (2017) 238–250 Fig. 15. A fragment of bisphenol-A epoxy vinyl

Fig. 15. A fragment of bisphenol-A epoxy vinyl ester in thin section with embedded glass fibers from sample EVE3 as observed in fluorescent mode (a), back scatter ed electron image (b), and close-up (c) with black arrows showing locations of characteristic X-ray spectra from interior of glass fiber (d) and exterior (reacted) glass fiber (e).

4. Conclusions

Reductions in workability have been previously reported for concrete with fibers recovered from recycled GFRP and attributed to increased water demand due to the irregular geometry and the high specific surface of the fibers [25] . In the research pre- sented here, the reduction in workability was overcome by precon- ditioning the recycled GFRP to a SSD condition. However, the relatively large amounts of water retained in the SSD condition may have had a negative contribution in terms of compressive strength; the water retained by the SSD recycled GFRP in many cases appears to have had a local influence on w / cm , increasing

the capillary porosity of the adjacent cement paste (Figs. 14 a and 15 a). Furthermore, rather than being dispersed as individual fibers, in most cases the recovered fibers from recycled GFRP and from the virgin E-glass fibers occurred in the concrete as porous clumps, and these would also be expected to have a nega- tive impact on compressive strength. On the other hand, the fibers from recycled GFRP and the virgin E-glass fibers did contribute to improvements in splitting tensile strength. Drying shrinkage was not improved by the addition of recycled GFRP or virgin E-glass fibers as compared to the control concrete, a finding that contrasts with García et al. [25] who reported a mod- est improvement. But, the overall values for drying shrinkage

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

249

Construction and Building Materials 146 (2017) 238–250 249 Fig. 16. Virgin E-glass fibers in concrete thin

Fig. 16. Virgin E-glass fibers in concrete thin section, backscattered electron image (a), and close-up (b) with black arrows showing locations of characteri stic X-ray spectra from interior of glass fiber (c) and exterior (reacted) glass fiber (d).

reported here remain within a narrow range similar to the control. Although a fixed replacement level of 5 wt% of recycled GFRP for aggregate was used throughout this study, the replacement level could be revisited and refined in an effort to minimize shrinkage, and to maximize improvements to tensile strength. Accelerated mortar bar and concrete prism ASR testing showed expansion values well below the recommended expansion limits. Comparisons of X-ray energy dispersive spectra (EDS) collected from points at the perimeters of the glass fibers to points further within the glass fibers show an enrichment and calcium (Figs. 13 d & e, 15 d & e, 16 d & e), which suggests the pozzolanic reaction of the fibers to form hydration products. Given the slender profile of the exposed glass fibers, on the order of 10 m m, the glass fibers perform as a pozzolanic addition similar to expansive ASR- mitigating coal combustion fly ash or ground granulated blast fur- nace slag as opposed to performing as an ASR expansion-inducing aggregate particle.

Acknowledgements

This research was funded by the NSERC Engage Grants Program, Canada, EGP #468708-14.

References

[2] J. MacDougald, C. McDougald, Insulators: A History and Guide to North American Glass Pintype Insulators, vol. 1, National Insulator Association Inc, St. Charles, Illinois, USA, 1990, p. 197 . [3] ASTM D578, Standard Specification for Glass Fiber Strands, American Society for Testing and Materials, West Conshohocken, Pennsylvania, USA, 2011, p. 19 . [4] S.J. Pickering, Recycling technologies for thermoset composite materials – current status, Compos. A Appl. Sci. Manuf. 37 (2006) 1206–1215 . [5] M.C.S. Ribeiro, A.C. Meira-Castro, F.G. Silva, J. Santos, J.P. Meixedo, A. Fiúza, M.L. Dinis, M.R. Alvim, Re-use assessment of thermoset composite wastes as aggregate and filler replacement for concrete-polymer composite materials: a case study regarding GFRP pultrusion wastes, Resour. Conserv. Recycl. 104 (2015) 417–426 . [6] Y.C. Feng, F.Q. Zhao, H. Xu, Recycling and utilization of waste glass fiber reinforced plastics, IN: MATEC Web of Conferences, 67(07012), International Symposium on Materials Application and Engineering, May 27–28, Hangzhou, China, 2016. p. 5. [7] C.A. Castro, M.C.S. Ribeiro, J. Santos, J.P. Meixedo, F.J.G. Silva, A. Fiuza, M.L. Dinis, M.R. Alvim, Sustainable waste recycling solution for glass fiber reinforced polymer composite materials industry, Constr. Build. Mater. 45 (2013) 87–94 . [8] ACI 544.1R-96, Report on Fiber Reinforced Concrete, American Concrete Institute, Farmington Hills, Michigan, USA, 1996, p. 66 . [9] B. Ahmet, A.B. Kizilkanat, N. Kabay, V. Akyüncü, S. Chowdhury, A.H. Akça, Mechanical properties and fracture behavior of basalt and glass fiber reinforced concrete: an experimental study, Constr. Build. Mater. 100 (2015) 218–224 . [10] Y.V. Lipatov, S.I. Gutnikov, M.S. Manylov, E.S. EZhukovskaya, B.I. Lazoryak, High alkali-resistant basalt fiber for reinforcing concrete, Mater. Des. (2015) 60–66 . [11] M.E. Arslan, Effects of basalt and glass chopped fibers addition on fracture energy and mechanical properties of ordinary concrete: CMOD measurement, Constr. Build. Mater. 114 (2016) 383–391 . [12] C. Shi, K. Zheng, A review on the use of waste glasses in the production of cement and concrete, Resour. Conserv. Recycl. 52 (2) (2007) 234–247 . [13] G.J. Xu, D.F. Watt, P.P. Hudec, K.A. MacDonald, D.O. Northwood, Recycling automotive related wastes in concrete, J. Mater. Process. Technol. 48 (1–4) (1995) 385–390 . [14] C.H. Chen, R. Huang, J.K. Wu, C.C. Yang, Waste E-glass particles used in cementitious mixtures, Cem. Concr. Res. 36 (3) (2006) 449–456 .

250

A. Dehghan et al. / Construction and Building Materials 146 (2017) 238–250

1538 . [21] J.R. Correia, N.M. Almeida, J.R. Figueira, Recycling of FRP composites: reusing fine GFRP waste in concrete mixtures, J. Clean. Prod. 19 (2011) 1745–1753 . [22] M. Osmani, Innovation in cleaner production through waste recycling in composites, Manage. Environ. Qual. 24 (1) (2013) 1–9 . [23] M.S. Alam, E. Slater, A.H.M. Muntasir Billah, Green concrete made with RCA and FRP scrap aggregate: fresh and hardened properties, J. Mater. Civ. Eng. 25 (2013) 1783–1794 . [24] A. Yazdanbakhsh, L.C. Bank, C. Chen, Use of recycled FRP reinforcing bar in concrete as coarse aggregate and its impact on the mechanical properties of concrete, Constr. Build. Mater. 121 (2016) 278–284 . [25] D. García, I. Vegas, I. Cacho, Mechanical recycling of GFRP waste as short-fiber reinforcements in microconcrete, Constr. Build. Mater. 64 (2014) 293–300 . [26] A. Bentur, S. Mindess, Fiber Reinforced Cementitious Composites, second ed., Taylor & Francis, New York, New York, USA, 2006, p. 601 . [27] H. Estrada, L.S. Lee, FRP composite constituent materials, in: M. Zoghi (Ed.), International Handbook of FRP Composites in Civil Engineering, Francis and Taylor, Boca Raton, Florida, USA, 2014, pp. 31–50 . [28] G.J.Z. Xu, D.F. Watt, P.P. Hudec, Effectiveness of mineral admixtures in reducing ASR expansion, Cem. Concr. Res. 25 (6) (1995) 1225–1236 . [29] C. Shi, Y. Wu, C. Riefler, H. Wang, Characteristics and pozzolanic reactivity of glass powders, Cem. Concr. Res. 35 (5) (2005) 987–993 . [30] A. Shayan, A. Xu, Performance of glass powder as a pozzolanic material in concrete: a field trial on concrete slabs, Cem. Concr. Res. 36 (3) (2006) 457–

468 . [31] N. Schwarz, H. Cam, N. Neithalath, Influence of a fine glass powder on the durability characteristics of concrete and its comparison to fly ash, Cem. Concr. Compos. 30 (6) (2008) 486–496 . [32] B. Taha, G. Nounu, Using lithium nitrate and pozzolanic glass powder in concrete as ASR suppressors, Cem. Concr. Compos., 30(6) 497–505. [33] L.M. Federico, S.E. Chidiac, Waste glass as a supplementary cementitious material in concrete – critical review of treatment methods, Cem. Concr. Compos. 31 (8) (2009) 606–610 . [34] Z.Z. Ismail, E.A. AL-Hashmi, Recycling of waste glass as a partial replacement for fine aggregate in concrete, Waste Manage. 29 (2) (2009) 655–659 . [35] H. Du, K.H. Tan, Use of waste glass as sand in mortar: Part II – alkali–silica reaction and mitigation methods, Cem. Concr. Compos. 35 (1) (2013) 118–126 . [36] F. Rajabipour, H. Maraghechi, G. Fischer, Investigating the alkali-silica reaction of recycled glass aggregates in concrete materials, J. Mater. Civ. Eng. 22 (12) (2010) 1201–1208 . [37] I.R. Rachida, M. Cyr, A. Tagnit-Hamou, Use of fine glass as ASR inhibitor in glass aggregate mortars, Constr. Build. Mater. 24 (7) (2010) 1309–1312 . [38] K. Afshinnia, P. Rangaraju, Effectiveness of ground glass powder from recycled glass in mitigating alkali-silica reaction in concrete, J. Transp. Res. Board 2508 (2015) 10 . [39] K. Afshinnia, P.R. Rangaraju, Influence of fineness of ground recycled glass on mitigation of alkali–silica reaction in mortars, Constr. Build. Mater. 81 (2016) 257–267 .

[41] K. Afshinnia, P.R. Rangaraju, Impact of combined use of ground glass powder and crushed glass aggregate on selected properties of Portland cement concrete, Constr. Build. Mater. 117 (2016) 263–272 . [42] ASTM C289, Standard Test Method for Potential Alkali-Silica Reactivity of Aggregates (Chemical Method) West Conshohocken, Pennsylvania, USA, 2007, p. 7 . [43] D. Stark, Alkali-silica reactions in concrete, in: J.F. Lamond, J. Pielert (Eds.), Significance of Tests and Properties of Concrete and Concrete Making Materials, ASTM STP 1679C, ASTM International, Bridgeport, New Jersey, USA, 2006, pp. 401–409 . [44] R.J. Wolf, Method of recycling Fiberglass Reinforced Plastics, U.S. Patent 8,361,358, 2013. [45] ASTM C702, Standard Practice for Reducing Samples of Aggregate to Testing Size West Conshohocken, Pennsylvania, USA, 2011, p. 5 . [46] ASTM D2584, Standard Test Method for Ignition Loss of Cured Reinforced Resins West Conshohocken, Pennsylvania, USA, 2011, p. 6 . [47] CSA A23.2-25A, Test Method for detection of alkali-silica reactive aggregate by accelerated expansion of mortar bars, in: CSA A23.1/A23.2 Concrete Materials and Methods of Concrete Construction/Methods of Test and Standard Practices for Concrete, Canadian Standards Association, Mississauga, Ontario, Canada, 2015, pp. 425–433 . [48] ASTM C230, Standard Specification for Flow Table for Use in Tests of Hydraulic Cement West Conshohocken, Pennsylvania, USA, 2014, p. 6 . [49] CSA A23.2-14A, Potential expansivity of aggregates (procedure for length change due to alkali-aggregate reaction in concrete prisms at 38 C), in: CSA A23.1/A23.2 Concrete Materials and Methods of Concrete Construction/ Methods of Test and Standard Practices for Concrete, Canadian Standards Association, Mississauga, Ontario, Canada, 2015, pp. 350–362 . [50] CSA A23.2-2C, Making concrete mixes in the laboratory, in: CSA A23.1/A23.2 Concrete Materials and Methods of Concrete Construction/Methods of Test and Standard Practices for Concrete, Canadian Standards Association, Mississauga, Ontario, Canada, 2015, pp. 552–564 . [51] CSA A23.2-9C, Compressive strength of cylindrical concrete specimens, in: