SELF-SEALING CEMENT-BASED MATERIALS USING SUPERABSORBENT POLYMERS

Transcription

1 SELF-SEALING CEMENT-BASED MATERIALS USING SUPERABSORBENT POLYMERS Hai Xiang Dennis Lee (1), Hong Seong Wong (1) and Nick Buenfeld (1) (1) Concrete Durability Group, Imperial College London, UK Abstract Superabsorbent polymers (SAP) have the potential to be used as an admixture to enhance the self-sealing of cracks in cement-based materials such as concrete. In this study, cement paste and mortar specimens containing different types of SAP were cast and mechanically loaded to produce a single through-crack with controllable width between 100μm to 400μm. These specimens were then exposed to 0.02M sodium chloride solution at a pressure gradient of 4 and the flow rates were measured with time to evaluate the efficiency of the SAP for crack sealing. The cumulative flow through specimens containing 5% SAP by weight of cement was up to 85% lower than that of the control specimen. 1. Introduction Well-constructed uncracked concrete with a low water/cement ratio has very low permeability, but the formation of cracks in concrete can increase its permeability by orders of magnitude [1-3]. Some of these cracks can self-heal over a long period of time due to several mechanisms including the formation of CaCO 3 and mechanical blocking by eroded cement paste particles, [4-6]. However, this self-healing is limited to cracks of up to μm [4,6] and is not always reliable. A state-of-the-art report on self-healing phenomena in cementbased materials is currently being prepared by RILEM TC 221-SHC. SAP has the potential to be used as an admixture for self-sealing of concrete [7]. A unique characteristic of SAP is that its swelling capacity changes significantly with the ionic content of the solution. When added to a concrete mix, SAP absorbs about 3-20 grams of water per gram of SAP and swells only slightly [7,8]. When the concrete sets and hardens, the SAP gradually releases its absorbed water and collapses. The SAP lies dormant in the concrete until cracking occurs that exposes the SAP. If the cracked concrete is subjected to external

2 wetting by a dilute aqueous solution, much more significant swelling of the SAP occurs, hence sealing the crack. The aim of this study is to demonstrate the effect of self-sealing using SAP in cement pastes and mortars. Specimens containing 5% Poly(acrylate) (Poly(AA)) or 4% Poly(acrylate-coacraylamide) (Poly(AA-co-AM)) SAP were prepared. A single through crack of between 100 and 400μm wide was induced in each specimen. The flow rate of dilute sodium chloride solution through the cracked specimens was measured and the results were compared with control specimens that did not contain SAP. 2. Materials and methods SAP-cement pastes and mortars were prepared using general-purpose cement, CEM II/B- V 32.5R, (Table 1), in accordance with EN [9]. The sand used in the mortars was siliceous sharp sand. The composition of the cement and the properties of the SAP are shown in Table 1 and Table 2 respectively. Three types of SAP from different sources were used. The SAP were either poly(acrylate) or poly(acrylate-co-acrylamide). The mix compositions are outlined in Table 3. The total water/cement (w/c) ratio indicates the amount of water required to produce a mix with similar consistence to the control mix CP1 or Mor. The total w/c ratios for specimens containing SAP are higher than CP1 or Mor because a significant portion of the batch water is absorbed by the SAP. SAP was added to cement and mixed thoroughly for 2 mins in a pan mixer. The water was then added and mixed for another 3 mins. The mixes were cast into specifically-designed cylindrical moulds (dia.=100mm, height=150mm) with inserts to produce grooves in the sides of the specimens for inducing cracks. The moulds were filled in three equal layers and each layer was compacted on a vibrating table. The specimens were then covered with plastic sheets and stored at 100% RH for 5 days. The hardened specimens were then demoulded, wrapped with cling film and stored for another 9 days. The top and bottom of each specimen was ground to create a flat surface. The specimen was then wrapped again in cling film and stored at ambient temperature (~21 o C) for up to 2 weeks. Table 1: Composition of the cement CEM II/B-V 32.5 R*. Composition (%) SiO 2 Al 2 O 3 Fe 2 O 3 CaO MgO SO K 2 O Na 2 O EqNa 2 O P 2 O 5 LOI %PFA

3 Table 2: Properties of the SAP used in the study*. SAP Source Diameter (μm) Bulk density (kgm -3 ) Polymer type S1 BASF, Germany < Poly(AA) S2 Evonik, Germany n/a Poly(AA) S5 ETi Poly(AA-co-AM) *Information provided by the manufacturer Table 3: Mix proportions. Specimen Designation Total w/c Batch quantities (kg/m 3 ) Cement SAP Sand Cement paste (control) CP Cement paste - 5% S1 CP1-5S Cement paste - 5% S2 CP1-5S Cement paste - 4% S5 CP1-4S Mortar (control) Mor Mortar - 5% S2 Mor-5S A single through crack was induced in each specimen with a loading frame. A metal bar was placed in the tip of each side groove of the specimen and pressure was then applied on the metal bar. The applied pressure was gently increased until cracking occurs. The cracked specimen was then briefly split apart and reassembled to ensure that a complete through crack was induced. A rubber seal attached to a thin stainless-steel plate was fitted into the side grooves of the specimen. A set of perspex strips were then inserted into the side grooves (Fig 1) and the assembled specimen was held together using three stainless-steel hose clamps. The crack width was adjusted by tightening or loosening the hose clamps and a stereomicroscope was used to measure the average crack width. An angled light source was used to enhance the contrast of the crack and to increase the accuracy of the crack width measurement. The ends of the specimen were then attached to the inlet and outlet cells using rubber sealant (Fig 2). These cells were fitted with sensors to monitor the temperature of the inlet solution, and ph and resistivity of the outlet solution. The inlet cell was connected to a tank containing 0.02M sodium chloride solution positioned to apply a constant 60cm pressure head to the specimen. The experimental set up, shown in Fig. 3, simulates a cracked concrete element subjected to groundwater pressure gradient of 4 meters of water per meter of sample thickness. The flow through the specimen was collected and was continuously monitored for at least 5 days using a data logging system.

4 Tightening nut Induced crack Perspex strip Steel plate Rubber seal Hose clamp Figure 1: Top view of the assembled cracked specimen Connections to sensors and ph probe Outlet cell Inlet cell Figure 2: Specimen attached to the inlet and outlet cell 3. Results and discussion The peak flow rates for all of the paste specimens are shown in Fig. 4. The results are compared with the theoretical flow rate calculated from the Poiseuille equation. It can be observed that for each crack width, the flow rate calculated with the Poiseuille equation is higher than that for the control CP1. This is almost certainly due to the tortuosity and roughness of the crack which is not accounted for in the Poiseuille equation. Replicate specimens CP1-5S1 and CP1-5S2 (200μm crack width) gave similar results. Most of the specimens containing SAP showed lower peak flow rates compared to the control. This is attributable to the swelling of SAP, which partially blocks the crack.

5 International RILEM Conference on Use of Superabsorbent Polymers and Other New Additives in Concrete Figure 3: Schematic showing the experimental setup for measuring the flow through cracked specimens 3 Peak flow rate (mm /s) Poi seui l l e Equat i on 1000 CP1 CP1-5S1 100 CP1-5S Surface crack width (micrometers) Figure 4: Peak flow rate for all paste specimens plotted against crack width. Fig 5 shows the change in measured flow rate and cumulative flow with time for pastes with a 200μm wide crack. The peak flow rate for CP1 is highest, followed by CP1-5S1, CP1-5S2, and CP1-4S5. The cumulative flow at the end of the experiment (after 100 hours of testing) is highest for CP1-5S1, followed by CP1, CP1-5S2 and CP1-4S5. It is interesting to note that the cumulative flow for CP1-5S1 is 80% higher than that of CP1. In contrast, CP1-5S2, which also contains a Poly(AA) SAP, but of larger particle size, shows an 80% reduction in the cumulative flow. This suggests that S1 is not effective for crack blocking because of its

6 smaller particle size. CP1-4S5 gave the lowest peak flow rate and cumulative flow, despite having the lowest SAP dosage. This suggests that Poly(AA-co-AM) is more effective than Poly(AA) for sealing cracks. The results for the mortars are shown in Fig 6. The peak flow rate and cumulative flow for Mor-5S2 is lower than that of the control specimen by about 50% and 85% respectively. It is interesting to note that the mortar achieved a similar level of reduction in cumulative flow to CP1-4S5, despite the lower amount of SAP used in Mor-5S2 (Table 3). This suggests that the cracking in the mortar occurred mainly through the paste fraction rather than the aggregate particles so that more SAP was exposed than would otherwise be Flow rate (mm^3/s) CP1 CP1-5S1 CP1-5S2 CP1-4S Cumulative flow (L) Time (mins) (a) CP1 CP1-5S1 CP1-5S2 CP1-4S Time (mins) (b) Figure 5: Flow rate (a) and cumulative flow (b) for pastes specimens with a 200μm crack

7 Flow rate (mm^3/s) Mor Mor-5S Time (mins) (a) Cumulative flow (mm^3/s) Mor Mor-5S Time (mins) (b) Figure 6: Flow rate (a) and cumulative flow (b) for mortars specimens with a 300μm crack

8 4. Conclusions The flow of dilute sodium chloride solution through cracked cement paste and mortar specimens containing SAP was measured to assess the potential of SAP as an admixture for self-sealing cracks in concrete. The cumulative flow through specimens containing SAP was significantly lower than for the control. Paste specimens containing 5% SAP by weight of cement showed a reduction in the cumulative flow of up to 80%. The mortar specimens also achieved a significant reduction in the cumulative flow, about 85%, despite having a lower SAP content compared to the pastes. The results suggest that Poly(acrylate-co-acrylamide) is more effective than Poly(acrylate) for crack sealing. The results also suggest that a larger SAP particle size is beneficial for sealing cracks. However, a more detailed study is required to confirm these findings and to gain a better understanding of the effect of SAP on the mechanical properties and durability of cracked concrete. References [1] C. Aldea, S. Shah and A. Karr, (1999), "Effect of cracking on water and chloride permeability of concrete", J. Mater. Civ. Eng., 11, 3, [2] Z. Bazant, S. Sener and J. Kim, (1987), "Effect of cracking on drying permeability and diffusivity of concrete", ACI Mater. J., 84, [3] K. J. Wang, D. C. Jansen and S. P. Shah, (1997), "Permeability study of cracked concrete", Cem. Concr. Res., 27, 3, [4] C. Edvardsen, (1999), "Water permeability and autogenous healing of cracks in concrete", ACI Mater. J., 96, 4, [5] W. Khushefati (2004), "Healing of cracks in concrete", Department of Civil and Environmental Engineering, Imperial College London, PhD Thesis. [6] C. A. Clear. (1985), "The effect of autogenous heaing upon leaking of water through cracks in concrete", Cement and concrete association, Technical Report 559. [7] H.X.D. Lee, H.S. Wong and N.R. Buenfeld, (2010), Potential of superabsorbent polymer for self-sealing cracks in concrete, Adv. Appl. Ceram., 109, 5, [8] O. Jensen and P. Hansen, (2002), "Water-entrained cement-based materials II. Experimental observations", Cem. Concr. Res., 32, 6, [9] EN 197-1, (2000), Cement: Composition, specifications and conformity criteria for common cements, European Committee for Standardisation