Correlation of chloride diffusion coefficient and microstructure parameters in concrete: A comparative analysis using NMR, MIP, and X-CT

Abstract

Permeability is a major indicator of concrete durability, and depends primarily on the microstructure characteristics of concrete, including its porosity and pore size distribution. In this study, a variety of concrete samples were prepared to investigate their microstructure characteristics via nuclear magnetic resonance (NMR), mercury intrusion porosimetry (MIP), and X-ray computed tomography (X-CT). Furthermore, the chloride diffusion coefficient of concrete was measured to explore its correlation with the microstructure of the concrete samples. Results show that the proportion of pores with diameters<1000 nm obtained by NMR exceeds that obtained by MIP, although the difference in the total porosity determined by both methods is minimal. X-CT measurements obtained a relatively small porosity; however, this likely reflects the distribution of large pores more accurately. A strong correlation is observed between the chloride diffusion coefficient and the porosity or contributive porosity of pores with sizes<1000 nm. Moreover, microstructure parameters measured via NMR reveal a lower correlation coefficient R² versus the chloride diffusion coefficient relative to the parameters determined via MIP, as NMR can measure non-connected as well as connected pores. In addition, when analyzing pores with sizes>50 µm, X-CT obtains the maximal contributive porosity, followed by MIP and NMR.

Keywords

permeability / microstructure / NMR / MIP / X-CT

Introduction

Permeability is a major factor influence on the durability of reinforced concrete (RC) structures [¹]. Many studies have shown that in chloride-rich environments, concrete structures suffer damage in the form of steel corrosion, which was caused by chloride ions penetrating into the surface of RC structures and reaching a certain critical concentration [²–⁴]. The chloride transport performance of concrete is one of the important indicators used to evaluate concrete durability [⁵,⁶]. Therefore, the chloride transport performance of concrete represents a hotspot of durability analysis for concrete structures in chloride environments and studying the microstructure of concrete is the foundation of the chloride transport process [⁵,⁷].

As a heterogeneous porous material, concrete has many pores with diverse shapes and sizes [⁸]. The slow ingress of fluid media into the pores of concrete indicates a lower permeability. Thus, the microstructure of concrete, including its porosity, pore connectivity, pore size distribution (PSD), and path tortuosity has a critical influence on its permeability [⁹]. However, these microstructural parameters affect the permeability of concrete to differing extents [¹⁰,¹¹]. Experimental studies have concluded that the permeability of concrete depends closely on the porosity. However, porosity is not the most crucial parameter that affects the permeability of concrete; moreover, the relationship between the permeability and porosity of concrete cannot be described simply. The pore connectivity and path tortuosity also have a crucial influence on the permeability of concrete [⁹,¹²]. Mehta and Paulo [¹³] suggested that permeability not only depends on the porosity of concrete, but also relies on pore connectivity and PSD. Pores and micro cracks are inevitable defects of concrete, which can affect the strength and permeability of concrete directly or indirectly [⁸,¹⁴]. Pipilikaki and Beazi-Katsioti [¹⁵] insisted that pore structure is an important microstructural characteristic of concrete, with its distribution influencing the physical and mechanical properties of concrete. Owing to their impact on the macro-properties of concrete, elucidating the microstructure parameters of concrete is vital [¹⁶]. Therefore, in this paper, the influence of microstructure parameters on the penetration of concrete by chlorides is explored.

Nuclear magnetic resonance (NMR), mercury intrusion porosimetry (MIP), and X-ray computed tomography (X-CT) are popular methods used for measuring the microstructure parameters and morphology of cement based materials, such as pastes, mortar and concrete [¹⁷–¹⁹]. MIP can determine the pore structure parameters of concrete effectively, track the influence of both relative humidity (RH) and surrounding temperature on the morphology of the pore structures [²⁰], and explore the effects of material composition and curing conditions on the reduction of water permeability and porosity [²¹]. Moreover, when combined with thermo gravimetric analysis (TGA) and scanning electron microscopy (SEM), the results of MIP can be used to understand the influence of moisture on the geometrical shapes and connectivity of pores [²²]. NMR technology can use the hydrogen proton spin and external magnetic field to obtain the transverse relaxation time (T₂) spectrum and the PSD. Indeed, a new procedure has been developed to obtain the PSD of concrete and soil based on the T₂ distribution results measured by NMR [²³–²⁵]. A comparison of the PSD characteristics of coals measured by both the NMR and MIP methods indicated that for the MIP method, mercury permeates pores owing to high-pressure and coupled with the pore shielding effect, the coal sample may either be deformed or destroyed, leading to inaccurate results. By contrast, NMR is a non-destructive and non-invasive method for measuring the microstructure parameters and can obtain test results that are much closer to real situations [²⁶]. In addition, X-CT is used to measure the inner pore structures of materials with scan resolutions from 100 µm to as little as 5 µm [²⁷]. It is worth noting that the crack initiation process in the surface and interior of concrete samples can be discerned, and therefore the interfacial transition zone (ITZ) of concrete can be identified by converting 2D images to a 3D structural model [²⁸].

As there are obvious differences between the principle and range for testing pore size using these three methods, the microstructures measured by NMR, MIP, and X-CT exhibit significant variation. The aperture of NMR and MIP is limited to a certain range of values (2–20000000 and 5–350000 nm, respectively), while X-CT can accurately test the pore structure for a large pore size (e.g., more than 50000 nm), thus representing an effective supplement to the other two methods. To investigate the validity and differences of measured microstructures by the three methods mentioned above, in this paper, 10 concrete samples with different water-binder (w/b) ratios (0.40, 0.45, 050, 0.55, and 0.60) and different admixtures were manufactured, and their pore structure and parameters such as porosity, PSD, and internal morphologies were determined by means of NMR, MIP, and X-CT. A comparative analysis of the results measured by the three methods is presented. Furthermore, the influences of w/b ratio and admixtures on the microstructure parameters were analyzed. Finally, the relationship between the chloride diffusion coefficient and microstructure parameters is established.

Preparation of the concrete samples

Constituents and mix proportions of concrete samples

The fine aggregate used in this experiment is river sand, which has an apparent density and fineness modulus of 2600 kg/m³ and 2.4, respectively. In addition, gravel is used as the coarse aggregate, which has a maximum size and apparent density of 20 mm and 2700 kg/m³, respectively. Tap water from the laboratory is used. The cement used is Portland cement (P.C 32.5). In addition, four types of admixture are used in the test, including fly ash (FA), slag (SG), silica fume (SF), and chopped basalt fiber (BF). The chemical compositions of the materials used in the experiment are shown in Table 1.

The dosage of water used in all the concrete mixtures is 190 kg/m³. Ten concrete mixtures (five different w/b ratios and five types of admixture) were prepared to investigate the influences of w/b ratio and admixture on the chloride permeability and microstructure parameters of concrete. To reduce the randomness effect of mixture proportions on the experimental results, for all the mixtures, the ratio of river sand to gravel is purposely set at 32%. Table 2 summarizes the concrete mixtures used in this experiment.

Preparation of the experimental concrete

According to the concrete mixture proportions listed in Table 2, for each concrete mixture, two prism specimens (length: 550 mm, width: 150 mm, height: 150 mm) were cast, and six cylinder specimens with a diameter of 100 mm and a height of 50 mm were prepared to measure the chloride diffusion coefficients of the concrete. The chloride diffusivity of concrete analyzed following the standard test method provided by ASTM C1202. All specimens were cured for 28 d at standard curing conditions (temperature of 20°C±5°C and RH>95%) in accordance with the Chinese standard SL352-2006 (Test Code for Hydraulic Concrete).

Test methods

X-CT

The radioactive source in the CT device can penetrate non-metallic materials. X-rays with different wavelengths have different penetration abilities, while specimens with different densities produce different degrees of X-ray attenuation [²⁹]. Using reconstruction algorithms, a 3D image can be calculated for which the voxel intensities correspond to the localized X-ray attenuations of the specimen material. Then, the corresponding microstructure parameters of the specimens can be determined.

The cubic samples with a length of 50 mm were cut down from above the middle position of the concrete specimens for the X-CT tests. An industrial CT scanner (GE Phoenix v/tome/x m; Yinghua Testing Co. Ltd., Shanghai, China) with a high-resolution (>50 µm) was used. The X-CT instrument system comprises a high-power X-ray source, a high-precision object stage, a high-resolution digital detector, and a computer for control and reconstruction. The voltage and electric current of the X-CT instrument are 190 kV and 140 µA, respectively. During the CT scanning process, the X-ray source and detector are fixed and the test sample is rotated step-by-step along the y-axial direction. Figure 1 shows photographs of the CT scanning instrument and the concrete samples prepared for X-CT testing.

NMR, MIP, and ASTM C1202 test

The microstructure of concrete was also measured using NMR (MesoMR23-060H-I, Niumag Electric Co. Ltd., Suzhou, China) and MIP (AUTOPORE 9500, Micromeritics Instrument Ltd., the USA). The detailed test principles and processes are described in our previous papers [³⁰,³¹].

The electric flux referring to the total charge Q passing through the specimens over a period of 6 h (Coulomb, C) is determined according to the requirements of ASTM C1202. Moreover, the corresponding total charge Q should be converted to the chloride diffusion coefficient D (m²/s) [³²,³³]:

(1)D=(\mathrm{0.00492}\times Q+\mathrm{2.582})\times {\mathrm{10}}^{-\mathrm{12}}.

Results and discussion

Porosity of the tested concrete

The test samples analyzed via each method were all cut from the same prism specimens. Furthermore, to avoid test errors caused by excessive differences in the sampling position, all the test samples were extracted as closely as possible from the prism specimens. Nevertheless, as the sample size for the MIP test was small, a single test on one sample may affect the final test results. Therefore, for each concrete mixture, the MIP test was performed for multiple samples. Outlying test results were removed from the subsequent analysis caused by the random selection of samples, and the final values of the microstructure parameters determined by the MIP method are the average value of at least three valid samples.

The total porosity data for the concrete samples obtained by the NMR, MIP, and X-CT methods, respectively are listed in Table 3.

These data indicate that the measured porosity obtained by NMR and MIP far exceeds that measured by X-CT. The total porosity of concrete increases gradually in response to increasing the w/b ratio, while adding admixtures into the concrete can decrease the total porosity of the concrete (except for BF, C3-BF), although the reduction effect of these admixtures is variable. Regardless of the test method, the addition of silica fume (C3-SF) produces concrete that is more compact compared with ordinary concrete—the porosity of C3-SF concrete is reduced by 12.44%, 18.34%, and 58.87% for NMR, MIP, and X-CT measurements, respectively. Conversely, the effect of slag (C3-SG) on the porosity of the concrete samples is less evident, with the porosity of C3-SG samples reduced by only 3.21%, 6.26%, and 47.06% for NMR, MIP and X-CT, respectively. It is obvious that the porosity reduction effect of admixture is considerably greater for the X-CT measurements. This phenomenon implies that the hydration of ordinary concrete (C3) is insufficient, resulting in the internal pore structures in the concrete retaining their connections at the early 28 d age, leading to a larger porosity being measured by MIP [³⁴]. However, the addition of admixtures causes some pores to become disconnected during the short hydration process. Therefore, the MIP-measured porosities are greatly reduced.

With the exception of the C3-BF sample, the porosity values obtained by the NMR and MIP methods are close, regardless of the existence of admixtures. However, there is no clear proportional relation between the total porosity and concrete mixtures. It is worth noting that the decay signal of hydrogen protons in the concrete samples is measured by NMR, which reveals the specific surface of the pore medium. However, in MIP, mercury is injected into the concrete samples through pores by high pressure, reflecting the influence of pore passages. Therefore, theoretically, NMR measures the connected pores and the non-connected pores simultaneously. MIP is intended for measuring open porosity, and therefore it is not suited to detecting the closed pores and micropores, leading to a smaller total porosity and pore volume than that measured by NMR. Despite this, our test results do not always reflect the phenomenon described above. Only the NMR-measured porosities relating to C1, C5, and C3-BF in Table 3 are larger than their corresponding MIP measurements. This can be attributed to the following reasons. First, external forces are applied when the test samples are cut out of the larger concrete specimens, resulting in an inevitable small shock within the samples. Consequently, the cutting process likely affects the pore structure within the samples; for example, some smaller pores may be enlarged and weak interfaces may produce smaller pores. In addition, it is impossible for mercury to be injected into coarse aggregates. Nevertheless, it should be noted that it is difficult to control the ratio of mortar to gravel in the concrete samples during the cutting process precisely, leading to a different aggregate ratio for the test samples in the NMR and MIP tests [³⁵]. Second, the pore diameter test range differs between NMR and MIP. The measuring range of MIP is approximately 5–350000 nm and the reliable aperture size should be smaller. By contrast, the measuring range of NMR is approximately 2–20000000 nm. Third, under high pressure, the pore wall may be broken because of the compressibility of sample microstructures, leading to an increase in the porosity measured by MIP [³⁶].

The pore structures of cement-based materials vary from the nanometer to the micron scale. Both capillary and gel pores can be observed by NMR and MIP, but air voids cannot be measured accurately [³⁷]. On the other hand, X-CT can measure pores with diameters>50 µm by identifying and obtaining parameters using grayscale or 3D simulations; therefore, X-CT is restricted to measuring the porosity of large pores. Accordingly, the porosity measured by X-CT is small compared with those measured by MIP and NMR.

Concrete pore morphology measured via X-CT

Each CT scan image consists of volume elements. For cross-sectional CT scan images, the number of images corresponds to the volume in the scan slices of the tested samples. It is worth noting that the different materials comprising the test samples have their own range of gray values in the CT scan images. Therefore, the number of voxels corresponding to a certain range of gray values can be used to calculate the volume fraction of the corresponding material [³⁸]. Owing to their resolution, it is difficult to distinguish single-pore boundaries from the CT images, which increases the difficulty of calculating the pore volume distribution and PSD. Therefore, it is necessary to identify and mark the boundary of each pore [³⁹]. The 3D reconstructed images of the internal structure of the concrete test samples (C1 to C3-BF) measured by X-CT are shown in Fig. 2 (different colors are used to denote different volumes according to the test principle for X-CT mentioned above).

Figure 2 demonstrates that partially replacing cement with admixtures (except C3-BF) can increase the concrete density, i.e., the porosity of the concrete can be decreased significantly. Moreover, the X-CT-measured porosity shows a consistent trend when compared with the results obtained via MIP and NMR. According to proportion of pores within the 0–50 µm range and>50 µm examined by NMR, MIP, and X-CT, the relevant contributive porosity (see Ref. [³⁰] for a definition of this indicator) can be calculated. Table 4 lists the corresponding calculation results.

As shown in Table 4, the contributive porosity of pores with sizes>50 µm measured via NMR or MIP is much less than that obtained from X-CT measurements. This indicates that pores with diameters>50 µm are being overlooked by the NMR and MIP methods, which is consistent with the results reported by Li et al. [³⁷]. In addition, the contributive porosity of pores with sizes>50 µm obtained by NMR is lower than that obtained by MIP. This can be explained by a variety of reasons. First, under high pressure, the pore walls may be damaged owing to the compressibility of the sample microstructures, leading to initially non-connected pores becoming connected pores, and thereby increasing the proportion of large pores in the samples [³¹,³⁶]. Second, for the MIP method, the sample needs to be dried before testing. A previous study has shown that when concrete samples are dried at 105°C, cracks are generated due to concrete shrinkage, resulting in the proportion of macropores increasing [⁴⁰]. Nonetheless, it can be observed that the segmental contributive porosity in the 0–50 µm range measured by MIP exceeds that measured by NMR. This phenomenon can be attributed to the presence of the ink-bottle effect.

T₂ spectrum of concrete test samples

The T₂ spectrum and the unit volume nuclear magnetic signal of the concrete test samples are shown in Fig. 3.

Based on Fig. 3, it appears that there are about two or three peaks in the T₂ curve of each concrete sample, including a major peak and one or two minor peaks. The height and enclosed area of the major peaks are both much larger than those of the minor peaks. In addition, the major peak occupies the dominant position of the T₂ curves, spanning three orders of magnitude. By analyzing the influence of the w/b ratios, it can be concluded that as the w/b ratio increases, the signal peak distribution range gradually increases, while the PSD appears to be much stronger for the macropore signal. However, for admixture concrete, the distribution of the main peaks is obviously narrower than for ordinary concrete: the distribution of the major peak becomes narrower and the micropore signal increases in intensity. The NMR test principle indicates that the total area of each T₂ distribution curve represents the total water content within the pores, that is, the porosity of the concrete test samples. Therefore, by comparing the areas of the T₂ distribution curves, it is inferred that increasing the w/b ratio causes the total enclosed areas of the T₂ distribution curves to increases accordingly. It can be observed that for the C3-BF sample, the enclosed area of the T₂ spectrum is increased, implying that the porosity of the concrete has increased. This is ascribed to the addition of BF to the concrete inducing some internal micro cracks during the hardening of C3-BF, thereby causing the formation of weaker interfaces [³¹]. Moreover, the addition of FA and SF to the concrete can effectively reduce the total signal amount, as represented by the enclosed area of the T₂ spectrum, clearly indicating a reduction in the porosity of the concrete. It should be noted that concrete containing 30% FA (C3-FA30) reduces the total signal amount more than concrete containing 20% FA (C3-FA 20). In addition, compared with ordinary C3 concrete, adding SG (C3-SG) decreases the peak value of the main peak of the T₂ spectrum.

Cumulative pore size distribution for the concrete test samples

According to the results presented in the previous section, the cumulative pore size distribution (CPSD) of tested concrete can be calculated, as listed in Table 5.

The data in Table 5 shows that when the pore diameter is less than 100 nm or greater than 1000 nm, the porosity obtained using NMR is significantly lower than that obtained using MIP. By contrast, the opposite result can be observed when the pore diameter is less than 10 nm or greater than 1000 nm. Obviously, MIP overestimates the proportion of large pores in the concrete samples. The underlying reason for this is the application of high pressure to the samples when injecting mercury into smaller pores, which crushes pore walls to create an excessive proportion of macropores [⁴¹]. By contrast, as the specimens are saturated with water for NMR, this method is non-destructive and non-invasive. Moreover, NMR has a much wider pore diameter test range than MIP. Therefore, it can be argued that the porosity measured by NMR may be more realistic.

Figure 4 shows the total and contributive porosities (defined in Ref. [³⁰]) within different pore diameter ranges for samples C1, C5, and C3-BF.

Together, Table 5 and Fig. 4 show that the difference in the porosity measurements acquired via NMR and MIP is not significant. However, when the pore size ranges from 10 to 1000 nm, the contributive porosity of concrete obtained by NMR is larger than that obtained by MIP, with the reverse result observed for the other pore diameters. This is attributed to the differing degrees of difficulty associated with different media penetrating the same pore aperture, resulting in different permeabilities for different media, such as chloride ions or N₂ [⁴²,⁴³].

Figure 5 presents the cumulative contributive porosity of concrete within different pore size ranges measured by NMR.

The results demonstrate that the porosity and the proportion of large pores in ordinary concrete (C1-C5) increases with an increasing w/b ratio. Moreover, for the same pore diameter, the contributive porosity of concrete with admixtures (discounting C3-BF) is lower than that of ordinary C3 concrete owing to the improved shape of the pore structures and the decrease in the total porosity due to adding admixtures to the concrete.

Influence of pore size distribution on the chloride diffusion coefficient in concrete test samples

The chloride diffusion coefficients corresponding to the concrete test samples are shown in Table 6.

The cumulative contributive porosity can be calculated within different ranges of the pore diameter. Considering that a large proportion of pores have diameters<1000 nm, the cumulative aperture in this range was studied. Figure 6 shows the relationships between D and the contributive porosity (for pore diameters<1000 nm) for NMR and MIP.

The relationships between D and the total porosity of concrete for NMR and MIP are shown in Fig. 7.

Figures 6 and 7 both show a strong positive relationship between D and the contributive/total porosity for pore diameters<1000 nm in the concrete samples. The relationships identified for the two techniques demonstrate strong similarity, indicating that although the test method can affect the quantitative results, it has little influence on the general rule between D and the total porosity or PSD.

Comparing Fig. 6 with Fig. 7, it can be seen that the correlation between D and the contributive porosity or total porosity of pores with diameters<1000 nm measured by MIP is larger than that measured by NMR, despite a consistent trend being observed. Therefore, it can be deduced that when the pore size is>1000 nm, its influence on the porosity and D of the concrete is minimal.

The main reasons for this are as follows. First, apart from connected pores, NMR can also measure the non-connected pores [³¹]. However, it has been recognized that the chloride diffusivity of concrete is related primarily to the connected pores [⁴⁴]. Second, NMR is particularly sensitive to water molecules because it is tuned to excite the hydrogen nucleus to obtain the transverse relaxation time (T₂), thereby eventually determining the microstructure parameters by calculating the relationship between T₂ and pore size. For cement-based materials, an unknown amount of hydration products can falsify the amount of hydrogen nuclei and lead to a small effect on the microstructure parameters [⁴⁵].

Although the positive correlations between the PSD measured by MIP and NMR and the value of D for the concrete samples are detected, the correlation difference confirms that MIP is widely used to measure connected pores or pores with damaged pore walls under high pressure. This is consistent with the conclusions presented by Korat et al. [⁴⁶]. On the other hand, compared with the results obtained by MIP, NMR can detect the connected pores and non-connected pores concurrently, leading to a lower correlation between the chloride diffusion coefficient and the total porosity or contributive porosity of pores with diameters<1000 nm.

Conclusions

The conclusions can be summarized as follows.

1) As the samples, principles and pore size ranges for the NMR, MIP, and X-CT tests are different, the porosity and PSD measured by NMR and MIP are different. However, the results measured by NMR are much closer to the real pore structures. NMR detects a larger proportion of the pores with diameters<1000 nm than MIP, although the difference in the porosity calculations between the two methods is not significant.

2) Because X-CT can measure the internal pore structures of cement-based materials with a scan resolution>50 µm, the 3D internal reconstructed images can reflect the distribution of large pores in concrete more intuitively. The porosity measured by X-CT is much smaller than that measured by MIP or NMR. Nonetheless, the contributive porosity of pore sizes>50 µm measured by NMR or MIP is much less than that by X-CT, and the contributive porosity of pore sizes for concrete samples measured by MIP is more than that measured by NMR, irrespective of whether the pore size ranges from 0 to 50 µm or is>50 µm.

3) There is a satisfactory positive correlation between D and microstructural parameters such as the total porosity or the contributive porosity for pores with sizes<1000 nm in the concrete samples. However, the pores measured by NMR include non-connected pores. Furthermore, the unknown amount of hydration products falsifies the amount hydrogen nuclei, eventually leading to a tiny effect on the microstructure parameters. As a result, the correlations between D and porosity as well as contributive porosity for pore sizes<1000 nm measured using NMR are both lower than for MIP.

References

[1] Alexander M G, Magee B J. Durability performance of concrete containing condensed silica fume. Cement and Concrete Research, 1999, 29(6): 917–922

[2] Yang L F, Cai R, Yu B. Investigation of computational model for surface chloride concentration of concrete in marine atmosphere zone. Ocean Engineering, 2017, 138: 105–111

[3] Isgor O B, Razaqpur A G. Modelling steel corrosion in concrete structures. Materials and Structures, 2007, 39(3): 291–302

[4] Wu L J, Li W, Yu X N. Time-dependent chloride penetration in concrete in marine environments. Construction & Building Materials, 2017, 152: 406–413

[5] Aldea C M, Shah S P, Karr A. Effect of cracking on water and chloride permeability of concrete. Journal of Materials in Civil Engineering, 1999, 11(3): 181–187

[6] Yu Y N, Yu J S, Ge Y. Water and chloride permeability research on ordinary cement mortar and concrete with compound admixture and fly ash. Construction & Building Materials, 2016, 127: 556–564

[7] Yang Y, Wang M. Pore-scale modeling of chloride ion diffusion in cement microstructures. Cement and Concrete Composites, 2018, 85: 92–104

[8] Bágel' L', Živica V. Relationship between pore structure and permeability of hardened cement mortars: On the choice of effective pore structure parameter. Cement and Concrete Research, 1997, 27(8): 1225–1235

[9] Das B B, Kondraivendhan B. Implication of pore size distribution parameters on compressive strength, permeability and hydraulic diffusivity of concrete. Construction & Building Materials, 2012, 28(1): 382–386

[10] Zhang Z Q, Zhang B, Yan P Y. Hydration and microstructures of concrete containing raw or densified silica fume at different curing temperatures. Construction & Building Materials, 2016, 121: 483–490

[11] Mun K J, So S Y, Soh Y S. The effect of slaked lime, anhydrous gypsum and limestone powder on properties of blast furnace slag cement mortar and concrete. Construction & Building Materials, 2007, 21(7): 1576–1582

[12] Powers T C, Brownyard T L. Studies of the physical properties of hardened Portland cement paste. Concrete International, 2003, 41: 1–6

[13] Mehta P K, Paulo J M. Concrete: Microstructure, Properties, Materials. 4th ed. New York: McGraw-Hill Press, 2006

[14] Zhang M H, Li H. Pore structure and chloride permeability of concrete containing nano-particles for pavement. Construction & Building Materials, 2011, 25(2): 608–616

[15] Pipilikaki P, Beazi-Katsioti M. The assessment of porosity and pore size distribution of limestone Portland cement pastes. Construction & Building Materials, 2009, 23(5): 1966–1970

[16] Zhou J, Ye G, van Breugel K. Cement hydration and microstructure in concrete repairs with cementitious repair materials. Construction & Building Materials, 2016, 112: 765–772

[17] Tanaka K, Kurumisawa K. Development of technique for observing pores in hardened cement paste. Cement and Concrete Research, 2002, 32(9): 1435–1441

[18] Liu J, Ou G F, Qiu Q W, Chen X C, Hong J, Xing F. Chloride transport and microstructure of concrete with/without fly ash under atmospheric chloride condition. Construction & Building Materials, 2017, 146: 493–501

[19] Stark J.Recent advances in the field of cement hydration and microstructure analysis. Cement and Concrete Research, 2011, 41(7): 666–678

[20] Brue F, Davy C A, Skoczylas F, Burlion N, Bourbon X. Effect of temperature on the water retention properties of two high performance concretes. Cement and Concrete Research, 2012, 42(2): 384–396

[21] Liu J, Xing F, Dong B Q, Ma H Y, Pan D. Study on water sorptivity of the surface layer of concrete. Materials and Structures, 2014, 47(11): 1941–1951

[22] Kong D L Y, Sanjayan J G, Sagoe-Crentsil K. Comparative performance of geopolymers made with metakaolin and fly ash after exposure to elevated temperatures. Cement and Concrete Research, 2007, 37(12): 1583–1589

[23] Tian H H, Wei C F, Wei H Z, Yan R T, Chen P. An NMR-based analysis of soil-water characteristics. Applied Magnetic Resonance, 2014, 45(1): 49–61

[24] Tian H H, Wei C F, Wei H Z, Zhou J Z. Freezing and thawing characteristics of frozen soils: Bound water content and hysteresis phenomenon. Cold Regions Science and Technology, 2014, 103: 74–81

[25] Wang X X, Shen X D, Wang H L, Zhao H X. Nuclear magnetic resonance analysis of air entraining natural pumice concrete freeze-thaw damage. Advanced Materials Research, 2014, 919–921: 1939–1943

[26] Yao Y B, Liu D M. Comparison of low-field NMR and mercury intrusion porosimetry in characterizing pore size distributions of coals. Fuel, 2012, 95: 152–158

[27] du Plessis A, Olawuyi B J, Boshoff W P, le Roux S G. Simple and fast porosity analysis of concrete using X-ray computed tomography. Materials and Structures, 2016, 49(1-2): 553–562

[28] Huang Y J, Yang Z J, Ren W Y, Liu G H, Zhang C. 3D meso-scale fracture modelling and validation of concrete based on in-situ X-ray computed tomography images using damage plasticity model. International Journal of Solids and Structures, 2015, 67–68: 340–352

[29] Qin X, Xu Q. Statistical analysis of initial defects between concrete layers of dam using X-ray computed tomography. Construction & Building Materials, 2016, 125: 1101–1113

[30] Zhang J Z, Guo J, Li D H, Zhang Y R, Bian F, Fang Z F. The influence of admixture on chloride time-varying diffusivity and microstructure of concrete by low-field NMR. Ocean Engineering, 2017, 142: 94–101

[31] Zhang J Z, Bian F, Zhang Y R, Fang Z F, Fu C Q, Guo J. Effect of pore structures on gas permeability and chloride diffusivity of concrete. Construction & Building Materials, 2018, 163: 402–413

[32] Feng N Q, Xing F. High performance concrete technology. Beijing: Atomic Energy Press, 2000

[33] Feng N Q, Feng X X, Hao T Y, Xing F. Effect of ultrafine mineral powder on the charge passed of the concrete. Cement and Concrete Research, 2002, 32(4): 623–627

[34] Ye G. Percolation of capillary pores in hardening cement pastes. Cement and Concrete Research, 2005, 35(1): 167–176

[35] Kumar R, Bhattacharjee B. Study on some factors affecting the results in the use of MIP method in concrete research. Cement and Concrete Research, 2003, 33(3): 417–424

[36] Diamond S. Mercury porosimetry: An inappropriate method for the measurement of pore size distributions in cement-based materials. Cement and Concrete Research, 2000, 30(10): 1517–1525

[37] Li B, Mao J Z, Nawa T, Han T Y. Mesoscopic damage model of concrete subjected to freeze-thaw cycles using mercury intrusion porosimetry and differential scanning calorimetry (MIP-DSC). Construction & Building Materials, 2017, 147: 79–90

[38] Zhang Y, Verwaal W, van de Ven M F C, Molenaar A A A, Wu S P. Using high-resolution industrial CT scan to detect the distribution of rejuvenation products in porous asphalt concrete. Construction & Building Materials, 2015, 100: 1–10

[39] Huang Y J, Yan D M, Yang Z J, Liu G H. 2D and 3D homogenization and fracture analysis of concrete based on in-situ X-ray computed tomography images and monte carlo simulations. Engineering Fracture Mechanics, 2016, 163: 37–54

[40] Zhou J, Ye G, van Breugel K. Characterization of pore structure in cement-based materials using pressurization-depressurization cycling mercury intrusion porosimetry (PDC-MIP). Cement and Concrete Research, 2010, 40(7): 1120–1128

[41] Wang Y, Yuan Q, Deng D H, Ye T, Fang L. Measuring the pore structure of cement asphalt mortar by nuclear magnetic resonance. Construction & Building Materials, 2017, 137: 450–458

[42] Zhou Y, Hou D S, Jiang J Y, Wang P G. Chloride ions transport and adsorption in the nano-pores of silicate calcium hydrate: Experimental and molecular dynamics studies. Construction & Building Materials, 2016, 126: 991–1001

[43] Sun C, Bai B. Diffusion of gas molecules on multilayer graphene surfaces: Dependence on the number of graphene layers. Applied Thermal Engineering, 2017, 116: 724–730

[44] Sun W G, Sun W, Wang C H. Relationship between the transport behavior of modern concrete and its microstructures: Research methods and progress. Materials Review, 2018, 32: 3010–3022

[45] Zhang P, Wittmann F H, Zhao T J, Lehmann E H, Vontobel P. Neutron radiography, a powerful method to determine time-dependent moisture distributions in concrete. Nuclear Engineering and Design, 2011, 241(12): 4758–4766

[46] Korat L, Ducman V, Legat A, Mirtiĉ B. Characterisation of the pore-forming process in lightweight aggregate based on silica sludge by means of X-ray micro-tomography (micro-CT) and mercury intrusion porosimetry (MIP). Ceramics International, 2013, 39(6): 6997–7005

RIGHTS & PERMISSIONS

Higher Education Press