Keywords

1 Introduction

To date, cementitious composites are still one of the most widely used construction materials in the world [1, 2]. With increasing demand for long-span structures, offshore platform, and precast prefabricated structures, the development of high-strength, light-weight cementitious composites (HSLWCCs) is attracting attention. The common methods used to produce them are the incorporation of air/bubbles or the use of light-weight aggregates. However, instability of the air bubbles leads to poor mechanical properties of the HSLWCC. Therefore, the use of light-weight aggregates is considered to be a more promising approach [3, 4]. Although some light-weight aggregates have good bonding to the cement matrix because of their porous structure and pozzolanic reactivity [5,6,7], the developed light-weight aggregate cementitious composites are still compromised by the lesser mechanical properties of the light-weight aggregates [6]. This has prompted researchers to look for more suitable alternatives.

Hollow glass microspheres (HGMs) are a high-performance, ultra-light-weight material consisting of a thin outer shell and an inert gas inside [8,9,10]. Compared with conventional light-weight aggregates, HGMs have the advantages of superior particle size distribution, lower density, higher strength, and sustainability [11], which creates a new impetus to producing HSLWCCs. HGMs may have pozzolanic reaction in the cement matrix due to their high content of SiO2 and Al2O3 [12, 13]. It has been shown that HGM can react with the surrounding cement matrix at low addition amounts to form new gel particles and increase the bonding properties of HGM with the matrix [14]. Nevertheless, the contribution of low additions of HGM to density is not obvious, and the pozzolanic reactivity of HGMs at different additions has not been well investigated.

Building on previous studies, the density and property changes of cementitious composites produced with larger additions of HGM need to be clarified. The pozzolanic reactivity of HGMs at different additions also needs to be discussed in depth. Therefore, in this study, different additions of HGMs at 30, 40, 50 and 60% by weight of cement were used to develop a HSLWCC, with a special aim of producing a high-strength floatable cementitious composite. The variation in the density and compressive strength of HSLWCC were investigated and the engineering properties of HSLWCC were evaluated by structural efficiency. Finally, the dispersion of HGM in the matrix and pozzolanic reactivity were evaluated by scanning electron microscopy (SEM).

2 Methods

2.1 Raw Materials

In this study, ordinary Portland cement (OPC, CEM I 52.5R) was used as the main binder for the mixture. The HGMs were incorporated into the cement paste to reduce the weight. The HGMs used in the mixture were ≈1–100 µm in size and had an average particle density of ≈460 kg/m3. They had a high compressive strength of ≈55 MPa. Figure 1 shows the SEM images and X-ray diffraction results of the HGMs. As can be seen, the HGMs had a good non-crystalline structure and potential pozzolanic reactivity. Highly efficient polycarboxylate superplasticizer (SP) was used to obtain suitable workability. The particle size distributions of both the OPC and HGMs were obtained by a laser diffraction particle size analyzer, as shown in Fig. 2. The chemical composition of the raw materials was determined using X-ray fluorescence spectroscopy (XRF) and listed in Table 1.

Fig. 1
A S E M image and an X-ray diffraction. a. A scanning electron microscopy image indicates a good non-crystalline structure. b. An X-ray diffraction pattern with respect to theta in degrees Celsius. It increases at first and decreases after 25 degrees Celsius.

Scanning electron microscopy image a and X-ray diffraction b of hollow glass microspheres

Full size image
Fig. 2
A graph of volume in percentage versus particle size in micrometers for cement and H G M. The volume of H G M was high when compared to cement.

Particle size distribution of ordinary Portland cement and hollow glass microspheres (HGMs)

Full size image
Table 1 Chemical composition of materials, wt%
Full size table

2.2 Mix Proportions

The mix proportions are listed in Table 2. The amounts of HGMs added were 30, 40, 50, and 60% by weight of cement. Samples without HGMs were marked as REF, while samples with HGMs were denoted by HSL and the replacement ratio of HGMs. For example, the content of HGMs in HSL-30 was 30% of cement by weight. The water-to-cement ratio was kept at 0.5. As extra HGMs were added to the mixture, the SP content was adjusted to ensure similar workability of the different samples.

Table 2 Mix proportions, wt%
Full size table

To prepare the samples with HGMs, SP was first solved in the water and the solution was then added to the cement. The mixture was stirred at high speed for 3 min. The HGMs were added to the fresh mixture and the mixture were stirred for another 6 min to achieve good workability. Subsequently, the mixture was poured in to the mold on a vibrating table. After casting, samples were covered with plastic films and stored under laboratory condition. After 24 h, the samples were demolded and placed in a standard curing chamber until age 28 days under standard curing conditions (temperature: 20 °C, humidity: >90%).

2.3 Experimental Methods

Cubes of 50 × 50 × 50 mm3 were prepared for the measurement of compressive strength and density. The volume of the sample was tested by the water displacement method, from which the density of the sample could be calculated. For the compression test, the samples were loaded at a rate of 1.5 kN/s. Three samples were used for each mixture and the average value is reported. The microstructure of the different samples was observed by SEM. The samples were soaked in the ethanol to stop the hydration process and then dried in a vacuum chamber.

3 Results and Discussion

3.1 Density, Strength and Structural Efficiency

Figure 3a shows the density and compressive strength of the different samples. As expected, the density of the HSLWCC decreased linearly with increasing HGM content. With the inclusion of 30% of HGM, the density of HSL-30 decreased by 30% compared with the REF. When the HGM content increased to 60% of cement, the density of HSL-60 reduced to 970 kg/m3 and it became floatable.

Fig. 3
Two vertical bar graphs labeled a and b depict compressive strength in megapascals, density in kilograms per meter cube, and structural efficiency of R E F, H S L-30, H S L-40, H S L-50, and H SL-60. The compressive strength of R E F is high and H S L-60 is low.

a Density, strength and b structural efficiency of high-strength light-weight cementitious composite

Full size image

Figure 3a also shows that the inclusion of HGMs reduced the compressive strength of the HSLWCC. However, it was interesting to discover that the decreasing rate of compressive strength was not linear with increasing HGM content. With 30% HGMs, the compressive strength of HSL-30 only reduced slightly by ~ 10.1% compared with the REF samples. This might be caused by good bonding between the HGMs and the cement matrix (see SEM results and the associated discussion in Sect. 3.2) and the high compressive strength of HGM as mentioned in Sect. 2.1. With further increasing of HGM content from 30 to 40%, the compressive strength of the HSLWCC reduced significantly by 42% compared with the Ref samples. This may be due to the large amount of HGMs disrupting the continuity of the cement paste [14]. However, the decreasing rate of compressive strength slowed when the content of HGMs exceeded 40%. When the HGM content increased from 50 and 60%, the compressive strength only reduced slightly. Although HSL-60 showed an ultra-low density <1000 kg/m3, its compressive strength was still maintained at >30 MPa.

Structural efficiency, which is the ratio between compressive strength and density (unit: kN·m/kg), is the main factor evaluating the lightness and strength of concrete. A higher value of structural efficiency represents a higher specific strength (i.e., high strength and light weight). Figure 3b shows the structural efficiency of the samples with different additions of HGMs. As evident, the HSL-30 specimens showed the highest structural efficiency. However, the structural efficiency decreased significantly in HSL-40, caused by the significant reduction in compressive strength as the density reduced linearly. When the HGM content was >40%, the changes in structural efficiency were only slight. However, it should be noted that with high HGM content, the structural efficiency of HSL-40, HSL-50 and HSL-60 did not reduce significantly compared with the REF. HSL-50 showed the lowest structural efficiency among the samples in this study, at 31.4, compared with the structural efficiency of the REF at 36.3.

3.2 SEM Analysis

Figure 4 shows the SEM images of the samples. The fracture surface of the HSL-30 sample shown in Fig. 4a indicates that the HGMs were uniformly dispersed in the matrix. Besides, Fig. 4b also shows that the HGMs were well bonded to the matrix and the hydration product was still partially attached to the surfaces of the HGMs after fracture. As shown in Fig. 1b, HGMs had potential pozzolanic reactivity. The reaction between HGMs and the cement matrix could further enhance the bonding [15, 16]. However, it should be noted that the bonding between the HGMs and cement matrix was still weak, as the reaction was limited, resulting in debonding between the HGMs and the cement matrix under loading, as shown in Fig. 4c. Besides, crushing of HGMs was also found in the samples after the compression test, as shown in Fig. 4d. Generally, two failure modes occurred in the HSLWCC: (i) debonding of the interface and (ii) crushing of HGMs. In the samples with lower HGM additions, failure mode (i) was the dominant damage mode, and in the samples with high HGM additions, damage mode (ii) dominated.

Fig. 4
Four SEM images labeled a to d indicate uniformly dispersed H G M in the matrix, well-bonded H G M to the matrix and the hydration product, bonding between the H G M and cement matrix, and debonding of the interface and crushing of H G M.

a Fracture image of HSL-30; b HGM nucleation reaction; c morphology of HSL-60; d failure mode of the samples

Full size image

4 Conclusions

In this study, a HSLWCC was developed using HGMs as the light-weight filler. In particular, a novel floatable cementitious composite with an apparent density of ~970 kg/m3 and compressive strength of ~31 MPa was developed.

By incorporating HGMs with a content of 30–60% by weight of cement, the density of the developed HSLWCC ranged from 970 to 1340 kg/m3 and the compressive strength ranged from 31 to 62 MPa. The compressive strength of the HSLWCC decreased significantly when the HGM content increased from 30 to 40%. With further increasing of HGM content, the compressive strength only reduced slightly. The density of the HSLWCC almost decreased linearly with increasing HGM content. The structural efficiency of the HSLWCC showed a sudden significant increase at a HGM content of 30%, while the structural efficiency of other HSLWCC samples was slightly lower than that of the reference sample.

Debonding of the interface and crushing of the HGM were both found at the fracture surface. The debonding of interface dominated in mixtures with high HGM content, whereas crushing of HGMs was usually found in the mixtures with low HGM content.