Effect of lightweight aggregate and the interfacial transition zone on the durability of concrete based on grey correlation
Lijuan Kong* & Yuanbo Du
School of Materials Science and Engineering, Shijiazhuang Tiedao University, Shijiazhuang, 050043, China Received 26 November 2013; accepted 1 July 2014
The impermeability and frost-resistance of concrete prepared with different water-binder ratios containing different types (shale ceramsite, fly ash ceramsite and clay ceramsite), particle sizes (5-10 mm, 5-16 mm and 5-20 mm) and prewetting degrees (dry, 1 h pre-wetted and saturated) of lightweight aggregate (LWA) are evaluated. The microhardness and pore structure of the interfacial transition zone (ITZ) are also investigated, and the grey correlation theory is adopted to analyze the effect of LWA and ITZ with multi-factors on the durability of concrete. The results show that, the particle size of LWA and porosity of the ITZ have the highest correlation degree with the chloride diffusion coefficient of concrete, and their ri are higher than 0.9, which indicates that both of them have a significant effect on the impermeability of concrete.
Moreover, the prewetting degree of LWA should be decreased as far as possible, on the premise of meeting the concrete workability. Because the strong water absorbing force of LWA may help lower the porosity and increase the microhardness of the ITZ. For the frost-resistance of concrete, the strength of LWA is the dominant influencing factor, which suggests the significance of the frost-resistance of LWA itself, especially for concrete with a high water-binder ratio. In this study, the frost-resistance of concrete prepared with clay ceramsite, which has the lowest strength, is the worst. Furthermore, a dense ITZ is also good for the frost-resistance of concrete. Thus, the LWA with high strength and low water content can be used to prepare concrete with a high frost-resistance. The grey correlation model provides a new idea for the optimum selection of LWA to prepare highly durable concrete.
Keywords: Lightweight aggregate, Interfacial transition zone, Concrete durability, Grey correlation, Microhardness
It is well known that, the interfacial transition zone (ITZ) is the weakest region in normal weight concrete1-3, and often becomes the origin of cracks and pathway for permeation, which may have serious influence on the durability of concrete. However, the ITZ between lightweight aggregate (LWA) and cement paste is a reinforced layer, which may be attributed to the porous structure and water absorbing character of LWA4-6, and can partially improve the durability of concrete. For example, in studying the impermeability of concrete, it has been found that the impermeability of lightweight concrete is better than that of normal weight concrete in a high water-binder ratio condition7,8, however with the decrease of the water-binder ratio, the differences between them become less pronounced. For concrete with a water- cement ratio of 0.32, it has been observed that the capillary water absorption of lightweight concrete is higher than that of normal weight concrete9, which suggests that the negative effect of porous LWA on concrete impermeability is dominant. Similarly, the
frost-resistance of lightweight concrete can be far better than that of normal weight concrete10,11, but the pre-wetted LWA can somewhat decrease the frost-resistance of concrete. This may be caused by the reduction of air-entrainment and the water absorbing effect of LWA, however, the water releasing effect of pre-wetted LWA is advantageous to improve the frost-resistance of concrete in later age, especially for concrete with a low water-binder ratio12. It is obvious that lightweight aggregate concrete differs from normal weight concrete in a number of aspects, such as the increased porosity, improved ITZ, and improved cement hydration due to internal curing13. Both the LWA and the surrounding ITZ have certain effects on the concrete durability, and the actual properties of lightweight aggregate concrete depend on which of these factors are dominant. While many studies have been done on this aspect, further work is still necessary to find the key factors of influence.
Grey theory was put forward by Deng14 in 1985, its character is to study the small sample, poor information and uncertainty problem. In this theory,
——————
*Corresponding author (E-mail: konglijuan_888@163.com)
any random process is considered to be a grey value, which varied in the range of a certain magnitude and time zone. Although the objective system is complex and the data is in a mess, it always has some intrinsic rules. The grey theory method can process the limited and irregular data, and the major factors that influencing the target value can be found out.
Therefore, the grey theory has been widely applied to the economic, agriculture, meteorological and engineering15,16. In this paper, LWA with different types, particle sizes, and prewetting degrees are used to produce concrete, and the durability of concrete and microstructure of the ITZ are examined.
To quantitatively assess the effect of LWA and ITZ on the concrete durability, the grey theory method is adopted, and an attempt is made to find out the major influencing factors. The results obtained are expected to provide some theoretical basis for the optimum selection of LWA to prepare highly durable concrete.
Experimental procedure
Raw materials
Grade 42.5 ordinary Portland cement was used.
The fly ash met the Class Ι requirement of Chinese (GB/T 1596-2005) standards, which has a fineness of 45 µm square mesh sieve residue of 4.8%. Three types of lightweight aggregate, YT, NT and FT which are based on shale rock, expanded clay and sintered fly ash respectively, and with particle size between 5 mm and 10 mm, were chosen for the study, as shown in Fig. 1. The physical properties of the aggregates are summarized in Table 1. In addition, for the YT aggregate, the particle size was also changed. Table 2 shows the size distribution of LWA.
The normal weight aggregate used was crushed limestone rock with a specific gravity of 2.64 g/cm3. The fine aggregate used was river sand with a fineness modulus of 2.8 and an apparent density of 2.61 g/cm3. A naphthalene-based superplasticizer
(UNF-5) was used. Fig. 1 – LWA used in this study (a) YT, (b) NT and (c) FT Table 1 – Physical properties of lightweight aggregate
Water absorption, % Types Packing density, (kg/m3) Apparent density, (kg/m3) Strength,
MPa 1 h 24 h
Limestone 1520 2640 65 0.9 —
YT 839 1463 7.9 6.5 11.7
NT 440 758 3.2 10.0 14.1
FT 757 1251 5.7 15.3 17.4
Note: The strength of limestone aggregate is the compressive strength of source rock, whereas that of LWA is the cylinder crushing strength
Mixture proportions
Two groups of concrete were prepared with water-binder ratios of (w/b) 0.47 (A) and 0.31 (B) by mass. For each group of concrete, three different particle sizes of YT were used to prepare the lightweight concrete, and the LWAs were 1 h pre-wetted prior to mixing. They include 5-10 mm, 5-16 mm and 5-20 mm, which are designated as AX, AZ, AD and BX, BZ, BD (X indicating Dmax=10 mm, Z indicating Dmax=16 mm and D indicating Dmax=20 mm), respectively. In order to study the influence of prewetting degree of LWA on concrete durability, the dry and saturated YT with a particle size of 5-10 mm were also used, which are designated as A0, A1 and B0, B1 (0 indicating dry and 1 indicating saturated).
Moreover, the same volume fraction of the NT and FT, which have a particle size of 5-10 mm and 1 h pre-wetted degree, were also used to prepare concrete
and designated as AN, AF and BN, BF (N indicating NT and F indicating FT). The normal weight aggregate was also used to replace the LWA with the same volume fraction for comparison and designated as AL and BL. The slump of the fresh concrete was controlled within 70-90 mm, by adjusting the amounts of high-range water reducer.
Table 3 shows the specific data.
Specimens preparation and testing Impermeability test
Concrete specimens with dimensions of 100×100×50 mm were used for chloride diffusion study. They were immersed in the sodium chloride solution with a concentration of 4 mol/L, and saturated by vacuum pumping. The diffusion coefficient of chloride ion was obtained using a rapid chloride penetrant method17, which is based on the Nernst-Einstein equation:
i i
i
i Z F C
D = RT2
σ
2… (1) where Di is the diffusion coefficient of chloride ion, R is the gas constant (R=8.314 J/mol.K), T is the absolute temperature, σi is the partial conductance of chloride ion, Zi is the charge number of chloride ion, F is the Faraday constant (F=96500 C/mol), and Ci is the particle concentration of chloride ion.
Table 2 – Distribution of lightweight aggregate size Cumulative sieve residue, % Square mesh sieve size, mm Aggregate Nominal
size, mm
2.36 4.75 9.50 16.0 19.0 26.0
5~10 100 92 12 — — —
5~16 100 93 52 7 2 —
YT
5~20 100 90 60 10 5 1
NT 5~10 100 95 13 — — —
FT 5~10 100 93 12 — — —
Table 3 – Mixture proportions of concrete
AL AX AZ AD AN AF A0 A1 BL BX BZ BD BN BF B0 B1
w/b 0.47 0.47 0.47 0.47 0.47 0.47 0.47 0.47 0.31 0.31 0.31 0.31 0.31 0.31 0.31 0.31 Water, (kg/m3) 160 160 160 160 160 160 160 160 160 160 160 160 160 160 160 160 Cement, (kg/m3) 278 278 278 278 278 278 278 278 425 425 425 425 425 425 425 425 Sand, (kg/m3) 687 687 687 687 687 687 687 687 650 650 650 650 650 650 650 650 Fly ash, (kg/m3) 64 64 64 64 64 64 64 64 98 98 98 98 98 98 98 98
Limestone, (kg/m3) 1154 — — — — — — — 1061 — — — — — — —
YT — 583 556 541 — — 547 618 — 536 507 485 — — 503 568
NT — — — — 312 — — — — — — — 287 — — —
FT — — — — — 539 — — — — — — — 496 — —
5-10 mm — ∆ — — ∆ ∆ ∆ ∆ — ∆ — — ∆ ∆ ∆ ∆
5-20 mm — — ∆ — — — — — — — ∆ — — — — —
5-30 mm — — — ∆ — — — — — — — ∆ — — — —
Dry — — — — — — ∆ — — — — — — — ∆ —
1h-prewetted — ∆ ∆ ∆ ∆ ∆ — — — ∆ ∆ ∆ ∆ ∆ — —
LWA, (kg/m3)
Saturated — — — — — — — ∆ — — — — — — — ∆
Water reducer, % 0.45 0.50 0.45 0.45 0.55 0.60 0.75 0.45 1.40 1.45 1.40 1.40 1.50 1.60 1.85 1.45 Strength-to-weight ratio,
(×10-3MPa/kg·m-3)
20 19 18 16 13 17 20 18 29 26 24 23 18 25 28 25
Note: indicates the particle size and prewetting degree of LWA used
Frost-resistance test
Frost-resistance of concrete was done according to the quick-freeze method of Chinese (GB/T50082- 2009) standards. The specimen size was 100×100×400 mm, and after demolding, the specimens were placed in a standing curing room, where the temperature was kept at 20±2°C and the relative humidity was higher than 95%, until the day of testing. The specimens were immersed in water at the age of 24 days, and after 4 days immersion the freeze-thaw test was performed.
There are two criteria, the change of relative dynamic elastic modulus and the mass loss, for evaluating the frost-resistance of concrete. The specimens were considered to be failed if their relative dynamic elastic modulus dropped to 60% or their mass loss reached 5%.
Microhardness test
The LWA and cement paste with the same proportions of cement and fly ash as shown in Table 1 were mixed, and then placed in a Ф100×100 mm cylinder mold. After 24 h, the specimens were demolded and cured for 28 days, at 90% of relative humidity and approximately 20°C constant temperature. A disc of Ф100×10 mm was sliced from the intermediate section of the cylinder, and after polishing treatment, the specimens were used for the microhardness test of the ITZ, by using the microhardness instrument. The objective lens was moved in one direction, starting from the aggregate to the bulk cement paste, as shown in Fig. 2. The data were obtained at 10 µm intervals. As the area of indentation is large, the microhardness was tested in a polygonal line, in order to avoid overlapping the adjacent indentation. By measuring the diagonal length of the diamond-shaped indentation, the microhardness of each position can be calculated.
Fig. 2 – Microhardness test pattern
Pore structure test
In order to investigate the pore structure, the cylinders were crushed, and the center paste located within 1 mm from the LWA surface was obtained and analyzed using mercury intrusion porosimetry.
The principle of the pore structure test is based on the intrusion of mercury in a porous network, by applying pressure from 0.1 to 6000 psi. The average radius r of the pore, considered cylindrical, reached by pressure Pc is determined by the Washburn equation:
Pc
r=−2
σ
cosθ
… (2)where r is the average radius of the pore, σ is the surface tension of mercury (σ=0.474 N/m), θ is the contact angle (θ=130º), and Pc is the operating pressure.
Results and Discussion
Test results
Durability of concrete
The frost-resistance of concrete prepared with different water-to-binder ratios containing different LWA was tested. The normal weight aggregate was also included for comparison. As the variation in mass of lightweight aggregate concrete is the superimposed results of the mass loss due to the spalling of surface mortar and the mass growth due to absorbing water, using the index of mass loss evaluate the frost resistance of concrete is not extremely accurate.
Therefore, only the relative dynamic modulus results of these concretes are shown in Fig. 3. It can be seen that, the frost-resistance of concretes prepared with dry LWA (A0, B0) are the best, followed by that of concretes with 1 h pre-wetted LWA (AX, BX), regardless of water-binder ratio. The specimen B0, which was prepared with a water-binder ratio of 0.31 containing dry LWA, even obtained 300 freeze-thaw cycles. Furthermore, both the frost-resistance of them is better than that of concretes with normal weight aggregate (AL, BL). Whereas, the frost-resistance of concretes with a w/b of 0.47 containing high water content (AN, AF) or saturated LWA (A1) is even worse than that of the concrete with normal weight concrete (AL). In other words, the high water content of LWA is disadvantageous to the frost-resistance of concrete. This can be explained as follows: on one hand, the pore structure of LWA is likely to serve a similar function of “air-entraning agent”, which can relieve the damage of frost heaving on concrete.
However, the increase of the water content of LWA would lead to the decrease of the number of pores, thus resulting in the weakened of the air-entraining effect. Besides, the high water content of LWA is also adverse to its own frost-resistance. On the other hand, the LWA with a lower water content also has a stronger water absorbing force, which can reduce the local w/b around it and form a dense ITZ. Moreover, the frost resistance of concrete prepared with small LWA is slightly better than that of concrete with large LWA. This is because the decrease of the aggregate size leads to the increase of the surface area of LWA and contents of the dense ITZ, which is beneficial to improve the frost-resistance of concrete.
In addition, the chloride diffusion coefficients of concrete at 28 days were evaluated, the results are shown in Fig. 4. It can be seen that, the impermeability of concrete prepared with dry LWA are still the best, and the chloride diffusion
coefficients of A0 and B0 are significant lower than those of the other specimens. For concrete with a w/b of 0.47, the chloride diffusion coefficient of specimen with normal weight aggregate is higher than that of specimens with LWA, which indicates the LWA can improve the impermeability of concrete. This is consistent with the earlier work by Chia and Zhang7. However, for concrete with a w/b of 0.31, the chloride diffusion coefficients of specimens with pre-wetted LWA are higher than that of specimen with normal weight aggregate. Liu et al.18,19 also found that the incorporation of coarse LWA in concrete increased the water sorptivity and permeability slightly compared to normal weight concrete with the w/c of 0.38, which is related to increased porosity of the concrete due to pores in coarse LWA. Moreover, for concrete prepared with saturated LWA, the impermeability of that with a high water-binder ratio is the worst, whereas it has some improvement when the w/b is low. This can be account for the
Fig. 3 – Effect of LWA on the frost-resistance of concrete (a) w/b=0.47 and (b) w/b=0.31
Fig. 4 – Effect of LWA on the anti-chloride diffusivity of concrete (a) w/b=0.47 and (b) w/b=0.31
internal curing effect of LWA in concrete20-22, which improved the cement hydration and microstructure of the cement paste.
Microstructure of ITZ
The microhardness test results are shown in Fig. 5, and each data is the average of 10 test points. In addition, by observing the interval length that the microhardness experienced fluctuation until stable, the thickness of the ITZ can be inferred. The dotted and solid lines in the figure can represent the thickness of the ITZ and microhardness of the bulk paste, respectively. It can be seen that, a compact ITZ is formed around LWA, its microhardness is significant higher than that of the bulk cement paste, and the ITZ around dry LWA has the highest microhardness and largest thickness. Furthermore, with the increase of the water-binder ratio, the microhardness of both the ITZ and bulk paste have some decrease, whereas the thickness of the ITZ increase from about 60 µm to 80 µm.
Fig. 5 – Effect of LWA on the microhardness of the ITZ (a) w/b=0.47 and (b) w/b=0.31
In addition, the porosity of the ITZ around different LWA was tested, and the results are given in Table 5. However, as the limitation of sampling, the pore structure of the ITZ is the result of the cement paste located within 1 mm from a coarse aggregate surface. Although the test samples include some bulk paste, the results can reflect the influence of coarse aggregate on the ITZ to a certain extent24. In this study, the porosity results also demonstrate that LWA with different particle sizes, types and pre-wetting degree could affect the ITZ samples to a considerable degree. The porosity of the sample around dry LWA is the lowest, whereas that around saturated LWA is the highest. Moreover, the porosity of the ITZ sample increases with the decrease of the water absorption and increase of the particle size of LWA. This is because, for a constant volume fraction of aggregate, the decrease of the aggregate size leads to the increase of the surface area of LWA and contents of the dense ITZ. Liu and Zhang23 also compared the LWAs of different sizes for internal curing, and found that finer particles were more efficient in reducing the shrinkage.
Grey correlation analysis Determination of analysis series
The data series that reflect the behavior of the system is called reference series (or parent series), and can be given by
{
y( )
k k n}
Y = =1,2,L, … (3)
In this study, the chloride diffusion coefficients (DCl
-) and freeze-thaw cycles (NF), which can reflect the durability of concrete, are set as the reference series. Table 4 shows the specific data.
The data series that consist of factors influencing the behavior of a system is called the comparison series (or subsequence), and can be given by
{
X( )
k k n}
i mXi = i =1,2,L, , =1,2,L, … (4) In this study, the factors of the ITZ and LWA that influence the concrete durability, such as the surface area (SITZ), thickness (TITZ), microhardness (HITZ), porosity (PITZ) of the ITZ and particle size (RA), water absorbing force (WA), strength (fA) and density (ρA) of the LWA, are set as the comparison series. The specific data are given in Table 5.
Table 4 – Durability of concrete
Durability of concrete AL AX AZ AD AN AF A0 A1 BL BX BZ BD BN BF B0 B1 DCl-,(×10-8cm2/s) 5.71 4.61 4.85 4.93 4.73 4.88 3.29 5.42 2.57 2.97 3.11 3.19 3.15 3.27 1.86 2.78
NF, numbers 131 168 157 148 103 108 196 117 202 265 252 246 194 206 300 239 Table 5 – Parameters of the ITZ and LWA
Parameters AX AZ AD AN AF A0 A1 BX BZ BD BN BF B0 B1
SITZ, m2 36.4 34.2 35.8 36.3 37.7 36.4 36.4 36.4 34.2 35.8 36.3 37.7 36.4 36.4
TITZ, µm 70 75 70 65 75 110 60 55 60 60 55 70 90 50
HITZ, MPa 69.1 68.4 67.9 69.0 69.3 74.8 68.6 80.6 79.8 79.4 82.2 83.5 88.3 82.9
PITZ, % 22.7 23.1 23.5 20.7 21.6 21.2 24.1 17.8 18.1 18.6 17.6 17.2 16.7 18.8 RA, mm 6.27 6.99 6.71 6.62 6.37 6.27 6.27 6.27 6.99 6.71 6.62 6.37 6.27 6.27 WA, % 5.2 5.2 5.2 4.1 2.1 11.7 0.1 5.2 5.2 5.2 4.1 2.1 11.7 0.1
fA, MPa 7.9 7.9 7.9 2.5 5.7 8.2 7.3 7.9 7.9 7.9 2.5 5.7 8.2 7.3
ρA, kg/m3 757 748 763 440 839 757 757 757 748 763 440 839 757 757 Note: SITZ, TITZ, HITZ and PITZ represent the surface area, thickness, microhardness and porosity of the ITZ; RA, WA, fA and ρA represent the particle size, saturated water absorption, strength and density of the LWA
Non-dimensional processing of variables
In view of the fact that the data of various factors may be of different dimensions, it is hard to compare with each other, therefore in grey correlation analysis, the data should be handled dimensionlessly as:
( ) ( )
( )
l k n i mX k k X
x
i i
i = , =1,2,L, ; =0,1,2,L, … (5) Xi(l) can be the initial value or the mean value or the interval value of the comparison series, and they are three main methods that usually applied in dimensionless processing. In this paper, the data of various factors is comparable in order of magnitude, thus the initialization value processing method is adopted.
Then, take the chloride diffusivity of concrete with a water-binder ratio of 0.47 as an example to analyze the influence of the ITZ and LWA on it.
The matrix of processed chloride diffusion coefficient of AX~A1 is expressed as:
( )
TB= 1.0001.0521.0691.026 1.0590.7141.176 … (6) And the corresponding processed subsequence matrix is expressed a:
=
000 . 1 924 . 0 019 . 0 000 . 1 1.026 0.993 857 . 0 1.000
000 . 1 1.038 250 . 2 000 . 1 934 . 0 082 . 1 1.571 1.000
108 . 1 722 . 0 404 . 0 016 . 1 952 . 0 1.003 1.071 1.036
581 . 0 316 . 0 788 . 0 056 . 1 912 . 0 999 . 0 929 . 0 997 . 0
008 . 1 1.000 1.000 070 . 1 035 . 1 983 . 0 1.000 0.984
988 . 0 1.000 1.000 115 . 1 1.018 990 . 0 071 . 1 0.942
1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
A
… (7)
Calculation of correlation coefficient
The correlation coefficient of y(k) and xi(k) can be calculated as:
( ) ( ) ( ) ( ) ( )
( ) ( )
k x k y( ) ( )
k x ky
k x k y k
x k y k
k i i i
k i i i
k i
i − + −
− +
= −
max max
max max min
min
ρ
ε ρ
… (8) where ρ is the resolution ratio. The general range of ρ is (0, 1), in this study, ρ=0.5. εi(k) is the correlation coefficient, which indicates the relative difference between reference and comparison curves of the i factor in the k moment.
The matrix of the correlation coefficient is evaluated as
=
0.814 0.753 0.399 0.814 0.871 0.808 0.707 0.814
0.728 0.703 0.333 0.728 0.777 0.676 0.472 0.728
0.939 0.695 0.540 0.947 0.878 0.932 0.984 0.971
0.633 0.520 0.764 0.963 0.871 0.965 0.887 0.964
0.926 0.917 0.917 0.999 0.957 0.898 0.917 0.899
0.923 0.937 0.937 0.924 0.957 0.925 0.970 0.875
1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
εij
… (9)
Calculation of correlation degree
The correlation coefficients obtained above are the correlation values of reference and comparison series at various moments, hence there are more than one value. However, such scattered information is not good for the overall comparison. Consequently, it is necessary to concentrate the various correlation
Table 6 – Results of the grey correlation analysis
w/b=0.47 w/b=0.31
Chloride diffusivity Frost-resistance Chloride diffusivity Frost-resistance Factors
ri Sequence ri Sequence
Factors
ri Sequence ri Sequence
SITZ 0.893 3 0.771 4 SITZ 0.893 3 0.841 4
TITZ 0.849 6 0.736 6 TITZ 0.871 5 0.756 7
HITZ 0.886 4 0.778 3 HITZ 0.892 4 0.849 2
PITZ 0.902 2 0.746 5 PITZ 0.900 2 0.812 5
RA 0.911 1 0.720 7 RA 0.918 1 0.795 6
WA 0.699 8 0.706 8 WA 0.707 8 0.719 8
fA 0.789 7 0.820 1 fA 0.816 7 0.872 1
ρA 0.852 5 0.801 2 ρA 0.870 6 0.845 3
coefficients in one value, that is to average these coefficients as the correlation degree (ri) between reference and comparison series, which can be expressed as:
∑ ( )
=
=
= n
k i
i k k n
r n
1
, , 2 , 1 1 ,
ζ
L … (10)The matrix of correlation degree between concrete chloride diffusivity and its influencing factors is calculated as
( )
Tr= 0.8930.8490.8860.902 0.9110.6990.7890.852
… (11)
Sequence of correlation degree
At last, rank the correlation degrees. If r1 < r2, it indicates that the reference series y is more similar to the comparison series x2. The grey correlation analysis results of the concrete durability are given in Table 6.
It can be seen that, for chloride diffusivity of concrete with different water-binder ratios, the top four influencing factors are RA, PITZ, SITZ and HITZ, which illustrates that the particle size of LWA and microstructure of the ITZ have significant effect on it. At the same time, combining the chloride diffusivity results of concrete as shown in Fig. 2, we can know that, to improve the concrete impermeability, the small size should be the first consideration in choosing LWA to prepare concrete.
This is because the decrease of the aggregate size leads to the increase of the surface area of LWA and contents of the dense ITZ, moreover, the path of penetration becomes more tortuous, and all of them are good for concrete impermeability. Besides, the dense ITZ with high microhardness and low porosity,
which can be formed by the strong water absorbing force of LWA, is also very important to improve the concrete impermeability. Hence, the LWA with high water absorption and without prewetting can be used to prepare high impermeability concrete, on the premise of meeting the concrete workability.
For concrete with a water-binder ratio of 0.47, the strength of LWA (fA) is the major factor that influences the concrete frost-resistance; LWA density (ρA) is the second factor, and the correlation degrees of the other factors are all less than 0.8, which illustrates that they are less influential on the frost- resistance of concrete. From the frost-resistance results of concrete as shown in Fig. 3, it’s obvious that the higher the strength of LWA, the better the frost-resistance of concrete. The frost-resistance of concrete prepared with NT, which has the lowest strength and density, is the worst. This suggests that the frost-resistance of LWA itself has a more pronounced effect on the frost-resistance of concrete.
However, for concrete with a water-binder ratio of 0.31, fA is still the dominant influencing factor, and microhardness of the ITZ (HITZ) is the second factor that influences the frost-resistance of concrete, which indicates the microstructure of the ITZ also has an effect on it. Figure 3 shows that the frost-resistance of concrete prepared with dry LWA is the best, which is attributed to its porous structure and dense ITZ formed by its strong water absorbing effect.
Thus, the LWA with high strength and low water content can be used to prepare high frost-resistance concrete.
Conclusions
(i) Based on the grey correlation model, and combined with the test results of the durability of concrete and microstructure of the ITZ, the influence
of LWA and the ITZ on the concrete durability was analyzed quantitatively, and the major influencing factors were found. A new idea for the optimum selection of LWA to prepare the high durability concrete is provided.
(ii) The RA and PITZ have the highest correlation degree with the chloride diffusion coefficient of concrete, and their ri are higher than 0.9, which indicates that both of particle size of LWA and porosity of the ITZ have significant effect on the impermeability of concrete. Therefore, the small size should be the first consideration in choosing LWA to prepare concrete. Moreover, the prewetting degree of LWA should be decreased as far as possible, on the premise of meeting the concrete workability.
Because the strong water absorbing force of LWA may help lower the porosity and increase the microhardness of the ITZ.
(iii) For the frost-resistance of concrete, fA is the dominant influencing factor, which suggests the significance of the frost-resistance of LWA itself, and the higher the strength of LWA, the better the frost-resistance of concrete. In this study, the frost-resistance of concrete prepared with NT, which has the lowest strength, is the worst. In addition, the dense ITZ which can be formed by the water absorbing force of LWA, is also good for the frost-resistance of concrete. Thus, the LWA with high strength and low water content can be used to prepare high frost-resistance concrete.
Acknowledgements
This work was financially supported by the National Nature Science Foundation of China (51108282), Nature Science Foundation of Hebei Province of China (E2011210025) and Excellent Youth Scholars of University Science and Technology Research of Hebei Province (Y2011111).
References
1 Breton D, Ballivy G & Grandet J, Cem Concr Res, 23 (1993) 335-346.
2 Kjellsen K O, Wallevik O H & Fjallbery L, Adv Cem Res, 10 (1998) 33-40.
3 Elsharief A, Cohen M D & Olek J, Cem Concr Res, 33 (2003) 1837-1849.
4 Zhang M H & Gjorv O E, Cem Concr Res, 22 (1991) 47-55.
5 Elsharief A, Cohen M D& Olek J, Cem Concr Res, 35 (2005) 1368-1376.
6 Hu S G, Wang F Z & Ding Q J, J Chin Ceram Soc (in Chinese), 33 (2005) 713-717.
7 Chia K S & Zhang M H, Cem Concr Res, 32 (2002) 639-645.
8 Kockal N U & Ozturan T, Constr Build Mater, 25 (2011) 1430-1438.
9 Ge Y, Kong L J, Zhang B S & Yuan J, J Chin Ceram Soc (in Chinese), 36 (2008) 934-938.
10 Tang S H & Liu X B, Frost-resistance of fly ash lightweight aggregate concrete, paper presented at the 4th National Lightweight Aggregate and Lightweight Concrete Symposium, Shandong, China, 1994.
11 Guo Y S, Kimura K & Li M W, Concrete (in Chinese), 10 (2002) 8-14.
12 Kong L J, Gao L X & Ge Y, J Chin Ceram Soc (in Chinese), 39 (2011) 35-40.
13 De la Varga I, Spragg R P, Di Bella C, Castro J, Bentz D P
& Weiss J, Cem Concr Compos, 45(2014) 102-110.
14 Deng J L, Grey control system, (Huazhong University of Technology Press, Wuhan), 1985
15 Jin C M, Fan L & Qiu Y, Chin J Appl Mech, 17(2000) 74-79.
16 Tan K R & Xiao X, J Shanghai Jiaotong Univ, 41(2007) 186-188.
17 Li C L & Lu X Y, Ind Constr (in Chinese), 28(1998) 41-43 18 Liu X M, Chia K S & Zhang M H, Cem Concr Compos,
32(2010), 757-766
19 Liu X M, Chia K S & Zhang M H, Constr Build Mater, 25(2011), 335-343
20 Bentz D P, Cem Concr Compos, 31(2009) 285-289.
21 Akcay B & Tasdemir M A, Cem Concr Res, 32(2010) 611-616.
22 Kovler K, Souslikov A & Bentur A, Cem Concr Res, 26(2004) 131-138.
23 Liu X M & Zhang M H, Mag Concr Res, 62(2010) 79-89.
24 Belaid F, Arliguie G & Francois R, Cem Concr Res, 31 (2001) 1561-1566
