Strength and shrinkage properties of alkali-activated slag concrete placed into a large column

Abstract

The properties of alkali-activated slag concrete (AASC) are generally determined on small-scale laboratory-produced prisms or cylinders. This investigation reports the in situ strength and in situ strain results from measurements on an 800 × 800 × 1200-mm column composed of AASC produced at a concrete plant. The results are contrasted with those of columns made with ordinary portland cement (OPC) and blended cement (50% OPC + 50% slag) designated as GB50/50. For this type of AASC, the workability improved from the time of production to the time of concrete placement and minimal slump loss occurred over 2 hours (in contrast with OPC). The rate of in situ temperature development was similar to GB50/50, and slower than OPC; however, the maximum temperature difference between the interior and the exterior was considerably less than OPC and GB50/50. The in situ strength was identical to OPC and superior to GB50/50 at 28 and 91 days. Embedded strain gauges show small initial in situ expansion within the centre of the column followed by a rapid rate of tensile strain growth up to 14 days. At 91 days the differential strain between the centre of the column and the corner was sufficient to crack OPC concrete; however, the AASC column remained uncracked. Possible explanations for this behaviour could be due to a greater tensile strain capacity of AASC than OPC concrete caused by greater creep and lower elastic modulus and higher tensile strength of AASC.

Introduction

During construction, structural concrete typically receives minimal curing following removal of formwork. Although the literature 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 reports the significant influence of curing on the strength development and drying shrinkage of alkali-activated slag concrete (AASC), the investigations generally involve testing of laboratory-size specimens. Whether lack of curing translates into a problem with larger scale structural concrete members is seldom reported. This paper reports the results of strength of cores and in situ measurements of shrinkage strains within an 800 × 800 × 1200-mm column, which are contrasted with measurements conducted on laboratory-size specimens.

The strength of concrete containing AASC is sensitive to the curing environment. Talling and Brandstetr [1] report that loss of moisture during dry air storage reduces strength between 7 and 28 days. Kutti and colleagues [2] report 35% strength difference at 36 months age between prisms cured at 94% relative humidity (RH) compared with 33% RH. Following air storage at 95 and 70% RH, Talling [3] reports little difference in cylinder strength; however, when stored at 50% RH the strength development stops after 28 days and the strength at 91 days is 86% that of the same cylinders stored at 95% RH. Cylinders moist-cured for 7 days followed by 40% RH were 10.5% lower in compressive strength at 1 year compared with identical samples cured at 70% RH [4]. This is in contrast with Byfors [5], who recorded similar strength when comparing water-cured strength vs. air-cured (50% RH) strength after initial 7 days wet curing. Hakkinen [6] reports the same strength at 6 months age after 28 days of moist curing followed by exposure to 40% RH, as compared to identical samples exposed to 100% RH. After 1 day of curing at 40°C, cylinders subjected to 40% RH showed the same 1-year strength as identical samples subjected to 100% RH [7]. For alternating curing conditions (7 days 40% RH followed by 7 days >95% RHfor 10 cycles) Hakkinen [8] and Malalepszy and Deja [9] report similar and enhanced strength compared to cylinders stored at >95% RH. Sioulas [12] found that the in situ core strength of blended cement concretes (up to 70% slag) in large-scale columns is significantly less than that of ordinary portland cement (OPC) control concrete and up to 35% less than bath-cured cylinder strength. Hakkinen [8] showed that dry exposed AASC displays a size effect, whereby the strength of larger samples (up to 150-mm diameter cylinders) is less affected than smaller samples by surface microcracks developed during drying. The in situ strength variations of a large structure member following exposed curing is unreported.

Drying shrinkage strain of AASC is sensitive to the exposure environment, although there are differences in the reported magnitude. Kutti [2] found that for RH > 70% the drying shrinkage for AASC was similar to OPC. However, at RH 33% and RH 50% the drying shrinkage was 3.0 and 2.3 times the drying shrinkage of OPC. During bath curing AASC swells compared with OPC 4, 10, 11, whereas under exposure condition of 70% RH, AASC has 32% greater drying shrinkage than OPC [4]. There are conflicting views on the reported long-term drying shrinkage. Douglas, Bilodeau, and Malhotra [11] report higher drying shrinkage at 224 days (50% RH) for AASC than OPC concrete, whereas Hakkinen [6] and Kukko and Mannonen [7] found no significant difference after 500 days exposure at 40% RH between AASC and OPC concrete for which the proportioning of constituents by weight was the same. The shrinkage behaviour of AASC larger in size than laboratory prisms is unreported.

1 cubic meter of AASC mixture was made at a concrete plant and mixed in a 2-m³ agitator truck. A column 800 × 800 mm in cross section and 1200 mm high was made, together with samples for laboratory testing of strength and shrinkage under different curing conditions. In situ monitoring of the column has been conducted for temperature and ongoing strain using embedded thermocouples and vibrating wire strain gauges. Strength has been monitored using extracted cores.

The chemical composition and properties of the cementitious binders are summarised in Table 1. The binders used are ground granulated blast furnace slag (slag) and OPC. The term water/binder (w/b) ratio is used instead of water/cement ratio to include all the binders mentioned above. The slag is supplied with gypsum (2% SO₃), which is blended with the slag. The activators and adjuncts utilised were powdered sodium metasilicate and hydrated lime (L). The dry powdered silicate activator was preblended with the slag in the dry form prior to use for concrete manufacture. The hydrated lime was added to the mix (1% weight of slag) as a 1:3 (hydrated lime:water) mixture to improve workability. The optimisation of the activator is not discussed in this paper.

The coarse aggregate consisted of 14-mm maximum size basalt that has a specific gravity of 2.95 and 24-hour water absorption of 1.2%. The fine aggregate consisted of river sand that has a specific gravity of 2.65, 24-hour water absorption of 0.5%, and a fineness modulus of 2.19. Proportioning was for saturated and surface dry conditions.

The concrete mixture proportions for AASC are summarised in Table 2. Also shown are reference mix proportions for OPC and blended cement (50% OPC + 50% slag, GB50/50) mixes [12]. Allowance was made for moisture content in the aggregates, powdered sodium silicate activator, and hydrated lime slurry to ensure correct free water content in the mix. The development of activator optimisation and concrete mix design are reported in unpublished work.

The batching sequence was as follows:

• An initial 80-litre load of water was charged into the agitator truck using a metered pump.

• 50% of the sand and coarse aggregate was then added.

• The preblended slag and activator was hand-loaded.

• The remaining sand and coarse aggregate was then loaded.

• The hydrated lime slurry was then added, followed by increments of water to obtain the desired workability (100-mm design slump).

The total time taken for the above steps was approximately 20 minutes. The travel time from the concrete plant to the concrete laboratory was 30 minutes.

The configuration of the column formwork is shown in Fig. 1. The column was cast onto a 100-mm layer of polystyrene, over which was a sheet of polythene plastic. The column was cast in three layers, as shown in Fig. 2. The purpose of the three layers was to assist the handling and transport of concrete for subsequent coring of each layer for strength testing. The layers were separated by wet corrugated cardboard placed on either side of a masonite board. A 100-mm sacrificial layer, designed to negate the influences of boundary bleed and temperature transfer, was allowed at each boundary separator. The lower sacrificial layer also served to accommodate instrumentation, as discussed below. The formwork was removed 24 hours after concrete placement, and the column was stored at 22–25° C and 65–70% RH.

A set of thermocouples and vibrating wire strain gauges (VWG) were installed within the lower sacrificial layer, as shown in Fig. 3. Type T copper/nickel thermocouples were installed along radial arms from the centre of the column, oriented at 90, 60, and 45° to the direction of the side face of the column. The spacing of thermocouples to the side face of a column were 0, 25, 50, 100, 200, and 400 mm. Another thermocouple was located immediately outside the column to ascertain ambient temperature variations. Readings were discontinued at 72 hours when the temperatures had equilibrated.

In situ strain was measured by purpose-built embeddable VWG. The gauge length was 150 mm. The gauges were located at 25, 200, and 400 mm from the column face as well as at a corner, as shown in Fig. 3. The coring schedule is shown in Fig. 4. The core strengths at 28 and 91 days from the date of concrete placement are reported.

Samples were also made for subsequent laboratory testing as follows:

• Standard cylinders (100-mm diameter and 200-mm height) for compressive strength testing in triplicate at 1, 3, 7, 28, 56, 91 days, and 1 year (following demoulding, “bath” curing in Ca(OH)₂ saturated water at 23°C, “exposed” curing at 23°C and 50% RH, and “sealed” curing involving storage in two polythene bags and a sealed container at 23°C). The sealed and exposed curing environments were chosen to simulate curing extremes within a large column.

• Standard shrinkage prisms (75 × 75 × 285 mm) were made. The exposure conditions following demoulding for a triplicate set of samples were exposed, sealed, and followed AS1012, Part 13 (i.e., 7 days bath cured followed by exposed curing at 23°C and 50% RH).

• Properties of fresh concrete such as slump and air content were also determined.