Bond Strength between Glass Fiber Fabrics and Low Water-to-Binder Ratio Mortar : Experimental Characterization

Powered by TCPDF (www.tcpdf.org) This material is protected by copyright and other intellectual property rights, and duplication or sale of all or part of any of the repository collections is not permitted, except that material may be duplicated by you for your research use or educational purposes in electronic or print form. You must obtain permission for any other use. Electronic or print copies may not be offered, whether for sale or otherwise to anyone who is not an authorised user. Böhling, Daniel; Cwirzen, Andrzej; Habermehl-Cwirzen, Karin


Introduction
One of the early attempts to produce ultrahigh strength concrete (UHPC) was made in the beginning of the 1970s. Very fine Portland cement clinker was mixed with potassium carbonate to control the hydration of calcium aluminates. Lignosulfonate-based plasticizer was used to provide a sufficient workability at the water-to-binder (W/B) ratio of 0.2 [1]. Developed in the 1980s, the socalled DSPs (densified with small particles) concrete utilized a large amount of silica fume, new generation of superplasticizers coupled with strong coarse aggregates [2]. As a result, the 28-day compressive strength exceeded 200 MPa. Another example of the ultrahigh strength concrete is called MDF (Macro Defect Free) [3]. e MDF concrete combined cement with polyvinyl alcohol (PVA) polymers, which acted as a dispersant and a crosslink between aluminate ions. e compressive strength reached 650 MPa. Unfortunately, regression of the strength was observed when the MDF concrete was exposed to water. Currently used ultrahigh strength concrete is derived extensively from the research initiated by De Larrard [4] and his linear packing density model (LPDM), solid suspension model (SSM), and compressive packing model (CPM). Later, Richard and Cheryrezy [5] developed a reactive powder concrete (RPC). Ultrahigh strength is achieved by optimization of the particlepacking density using ultrafine particles coupled with a very low W/B ratio, usage of modern superplasticizers, and application of heat treatment. e UHPC is very fluid in most cases, which makes it theoretically suitable to be combined with textile reinforcement. Textile-reinforced concrete (TRC) is commonly referred to as being a composite made from a fine-grained concrete matrix having the largest aggregate size in the range of 1 mm and continuous reinforcement made from textiles. e used textiles are based on multifilament yarns or highperformance fibers such as alkali-resistant glass, carbon, or aramid. e fabric reinforcement can be aligned in the load direction, and it is used in a wide variety of structures where corrosion resistivity and lower weight compared to a conventional steel reinforcement become particularly beneficial to decrease the overall thickness and the weight of the structure [6,7]. e biggest challenge in utilizing the full potential of TRC is a limited penetration of the cement paste between the fiber bundles [8,9]. erefore, the main objective of the research presented in this paper was to develop a testing procedure suitable for the determination of various parameters related to the bond strength between glass fiber fabrics and mortars with a very low W/B ratio. e study also aimed to correlate effects of the W/B ratio and workability on the penetration depth of mortars in between densely packed glass fiber bundles. Furthermore, selected mechanical properties of the composites including shear strength and pullout forces were investigated and correlated with other measured parameters.

Materials and Methods
e sulphate-resistant (SR) cement type CEM I 42.5 N (SR) manufactured by Finnsementti Oy was used, and its properties are given in Table 1. e used fine, high-purity silica fume was a by-product from the zirconium industry. As additional fine fillers, EHK quartz originating from quartzite deposits from Eastern Finland and fine sand (particle diameter of 125 to 600µm) were used ( Table 2). e fresh mix workability was adjusted by the polycarboxylate-based superplasticizer produced by BASF, type Glenium ACE 403. e fabric reinforcement was a unidirectional stitched glass fiber fabric produced by Ahlström Glassfibre Oy. e benefit of a unidirectional stitched fabric is the ability to orientate the reinforcement in the preferred direction [10]. Basic properties of the used glass fabric are given in Table 3. e used E-glass is unsuitable for concrete composites due to the known longterm degradation. However, the main research objective was the bond strength at the age of 28 days; therefore, the long-term degradation is insignificant. Four different mix designs were utilized (Table 4). e water-to-binder (W/B) ratio varied from 0.2 to 0.26, while the amount of other ingredients was constant. All mixes were produced using an Ecovac Bredent vacuum mixer. e vacuum was applied to remove the entrapped air from the paste. e mixing procedure consisted of premixing by hand to wet the mix followed by vacuum mixing for 2:00 min at the mixing speed of 290 rpm (rounds per minute). All samples were steam cured for 24 hours starting 48 hours after mixing. e workability was determined using a nonstandardized slump-flow test. In the test, 100 ml of the mix was poured from a height of approximately 10 cm on top of a flat surface covered with a sheet of a regular printing paper marked with circular lines placed at known distances, enabling to determine the slump flow ( Figure 1). All specimens were produced using Teflon molds to avoid contamination from a demolding oil. Compressive strength values were determined on specimens having dimensions of 12 × 12 × 60 mm 3 . Specimens used to test the bond strength and the shear strength had dimensions of 10 × 25 × 100 mm 3 ( Figure 2). Purpose-built Teflon formwork enabled to place the fabric in the middle of the formwork ( Figure 3). e compressive strength, flexural strength, and pullout strength were determined using a Zwick Roell RK 250/50 test rig with a loading speed of 0.35 mm/min. e single-lap shear strength test was performed following recommendation of the EN 14869-2 standard, Structural adhesives -Determination of shear behavior of structural bonds-Part 2: ick adherents shear test (ISO 11003-2 : 2001, modified). is test method was used to determine the shear behavior of an adhesive in a single-lap joint-bonded assembly when subjected to a tensile force. e specimen dimensions used for the shear strength test are shown in Figure 4 and the used setup in Figure 5. Notches were cut using a diamond saw. For each composite, three test specimens were tested at a testing speed of 1 mm/min. e average shear stress was calculated, following the EN14869-2; once the tensile force F has been applied, the average shear stress in the interface was calculated as where τ � average shear stress (N/mm 2 ), F � applied tensile load (N), a � width of the specimen (mm), and h � bond length (mm).

Advances in Civil Engineering
Pullout tests of the fabric from the cementitious matrix were performed as shown in Figure 6. For each cementitious matrix composite, three test specimens were tested at a testing speed of 1 mm/min, providing an average result and standard deviation. e microstructure and cement paste penetration depth was determined using a eld-emission scanning electron microscope (FESEM) type Quanta FEG 450 produced by FEI with an accelerating voltage set between 10 and 20 keV, a chamber pressure of 10-3 Pa, and a working distance of 15 mm. e back scattered electron (BSE) detector was used to obtain all images.

Results and Discussion
e workability of the produced UHPC mortars was determined only by measurements of the mini slump ow. e results showed an increased ow with an increasing W/B ratio (Figure 7). e measured compressive strength tended to decrease with a higher W/B ratio (Figure 8). e 7-day and the 28-day compressive strength values were nearly the same and within the calculated standard deviation. It can be related to the used steam curing which, as also shown earlier, resulted in a rapid early age strength development. e 7-day exural strength increased from 12 MPa at the W/B ratio of    e negative e ect of random small air voids was certainly enhanced by the small size of the used reference samples, 12 × 12 × 100 mm 3 . Consequently, it can be assumed that the exural strength was a ected by the used W/B ratios only to a very limited extent, and the average value was 12 MPa after 7 days. e pullout strength and the shear capacity of the study composites were determined on specimens made of the same mortars but incorporating additionally the glass ber fabric (Figures 2-4). e measured single-lap shear strength increased with a decreasing W/B ratio ( Figure 9). As shown and pointed by an arrow in Figure 5, mortars tended to form "mini beams" between the glass ber bundles, connecting layers of mortars on both sides of the fabric. It was assumed that most of the shear stress was transferred through them and their mechanical properties de ned the measured shear capacity of the composite. A higher measured shear capacity of composites using a lower W/B ratio could indicate a limited contribution of the bundles in this test setup. is assumption could explain a higher shear capacity of the composite despite a lower W/B ratio and thus worse penetration capacity. e results showed an opposite trend in the case of pullout strength. In this case, the measured maximum pullout forces were constant and reached around 3.3 kN/mm throughout the entire used W/B ratio range. e exception was the mix PR4 having the highest amount of water where the pullout strength increased to 4.5 N/mm ( Figure 10). Certainly, a higher amount of water led to a lower viscosity of the fresh mix and thus to its better penetrability. e viscosity of the produced mortars appeared to be too low for a su cient penetration between the single bers until the W/B ratio was increased to 0.26, see also the SEM test results in Figure 11. e displacement versus the applied load was measured only on one sample (PR1) which had the W/B ratio of 0.2 ( Figure 12). e results con rmed a typical pattern observed earlier [11]. In the rst stage, when the maximum pullout force was reached, the measured displacement was around 0.5 mm. In the second stage, a decrease of the pullout force to just under 50 N was observed, followed by its slight increase and accompanied by a large displacement of up to 6 mm. e observed behavior is commonly known as a "telescopic failure" in which a successive layer by layer breakdown from the sleeve laments to the core laments occurs. Once the sleeve laments fail in tension, the core laments are pulled out of the yarn. e  length of embedment of the yarn into the matrix has only a slight effect on the load-carrying capacity of the whole system [11,12]. e maximum recorded load of 250 N is related to the failure of the sleeve filaments, while the second stage with a lower force but a large displacement is related to the progressive pullout and eventual failure of the core filament.
e penetration depth of the cement matrix in between the glass fibers determined the volume of fibers, which are classified as a sleeve or a core in the load transfer. e higher the amount of the sleeve fibers, the higher the load bearing capacity of the entire system. e penetration depth on the binder matrix between the single fibers was determined visually by analyzing SEM-BSE images of resin-impregnated and polished samples. Two extreme examples are shown in Figures 12 and 13. e penetration was deeper in the case of mixes having a higher water amount and thus a lower fresh state viscosity. In the case of a mix with a W/B ratio of 0.26, the penetration was estimated to reach up to 100 µm which created the sleeve filament to result in the highest recorded pullout strength (Figure 9). On the contrary, in the mix with the lowest W/B ratio of 0.2, almost no penetration between the single fibers was observed which also translated into a very low pullout strength value. e main problem related to the penetration of mortars was the presence of fine particles including, in this case, quartz and sand. e particles even despite their very small diameter tended to be stopped by the closely spaced glass fibers, which formed an artificial sieve. Considering that a single glass fiber diameter is around 17 µm and the distance between fibers is much shorter, the maximum particle size should not exceed a few micrometers to enable their movement alongside the cement paste in between the fibers. e sieving effect is evidently visible in the studied samples where sand and quartz particles remained in the outer layer outside of the glass fiber bundle. e viscosity of the mortar and cohesive forces prevented extensive separation of the parts from the particles.
Several solutions could be used to improve the load capacity of UHPC glass fiber composites. e glass fiber fabric should be more loosely packed in comparison with the one used in these tests. e maximum particle size could be decreased to a few micrometers, thus enabling better penetration in between fibers.

Conclusions
e penetration depth of a low water-to-binder ratio mortar into a glass ber fabric and its e ects on selected mechanical properties were studied. e average penetration varied between 30 and 100 µm, depending on the used W/B ratio. e results showed that maximum particle size of quartz and sand should be below the average spacing of a single glass ber. In the case of this study, the maximum particle size was estimated to be <20 µm. e compressive and the exural strengths of the reference mortars reached 140 MPa and 15 MPa, respectively. e highest values were recorded for mortar with the lowest W/B ratio. e used experimental setup enabled also to determine the pullout force and the single lap shear strength of the mortar/glass fabric composites.
e measured values reached 4.5 N/mm and 0.25 MPa, respectively. However, the highest pullout strength was obtained for the mortar having the highest W/B ratio while, on the contrary, the highest shear strength was measured for the mortar having the lowest W/B ratio. e    Advances in Civil Engineering 9 main reason was assumed to be related to the load transfer mechanism. In this case, the lap shear was mostly affected by the mortar filling in the volume between the fiber bundles. While, in the case of the pullout force, the penetration depth in between the single fibers appeared to be the most influential factor.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.