Analysis of Pore Structure and Its Relationship to Water Transport and Electrical Flux in Mortars Incorporated with Slag and Silica Fume

Abstract

This study investigates the effects of slag and silica fume on the mechanical properties, transport behavior, and pore structure of cement-based mortars. Mortars incorporating different proportions of supplementary materials were evaluated by compressive and flexural strength, saturated water absorption, chloride permeability, and mercury intrusion porosimetry (MIP). Fractal analysis was further applied to assess pore structure complexity. At 28 days, the slag–silica fume blend SG20SF10 reached 46.5 MPa in compressive strength and 5.8 MPa in flexural strength, exceeding OPC. MIP showed a decrease in total porosity from ~14.5% to ~11.3% (about 22% lower) with a marked reduction in pores larger than 100 nm. Consistently, SG20SF10 exhibited the lowest water absorption and chloride permeability at both ages. These results indicate that the slag–silica fume synergy refines capillary porosity and increases pore-network complexity, thereby reducing directional connectivity and transport.

1. Introduction

As one of the most widely used construction materials, cement-based materials are inevitably exposed to environmental factors throughout both production and service life [1,2,3]. Among these, water plays a pivotal role—not only serving as a medium for aggressive ion transport but also influencing the pore structure of cement-based materials. Since water absorption behavior is closely linked to pore connectivity, it ultimately determines the efficiency and depth of ion penetration [4,5,6,7,8].

Harmful ions such as chloride [9,10,11], sulfate [10], and magnesium [12], once dissolved in water, can penetrate deeply into the cement matrix and initiate physical or chemical degradation. For example, under wetting–drying conditions, chloride-laden water is absorbed via capillary suction in unsaturated mortar, accelerating steel corrosion [13], while sulfate attack and freeze–thaw deterioration are similarly facilitated by moisture ingress [14]. These coupled processes highlight the essential role of water-related transport behavior in controlling long-term durability, prompting growing academic and engineering interest in this topic [15,16].

To gain a deeper understanding of water-driven degradation, it is essential to characterize the pore structure of cement-based materials at the microstructural level [1,4,17,18,19,20]. To quantify these features, fractal analysis has been widely used, with parameters like the box-counting dimension (from MIP data) reflecting pore heterogeneity [21,22,23,24]. However, the correlation between fractal characteristics and transport properties has not yet been fully established and calls for further investigation.

Ground-granulated blast-furnace slag (GGBS) and silica fume (SF) are widely adopted supplementary cementitious materials (SCMs) for tailoring OPC microstructure and transport. Latent-hydraulic reactions of GGBS and the highly pozzolanic activity of SF promote additional C–S–H and aluminosilicate hydrates, densify the interfacial transition zone, and redistribute pore sizes toward finer gel/capillary ranges. Collectively, these processes reduce capillary continuity and increase tortuosity, which underpins the commonly reported improvements in later-age strength, sorptive water uptake, and chloride-ion penetrability compared with neat OPC at the same water–binder ratio. SF often imparts an earlier pore-refinement and flux-blocking effect due to its ultrafine particle size and filler–nucleation action, while GGBS contributes to sustained densification and enhanced chloride binding over longer curing times; when jointly used, GGBS + SF systems can deliver synergistic benefits provided that dosage and curing are properly controlled [25,26,27,28,29,30].

Nevertheless, several issues remain insufficiently resolved in the existing literature. First, most studies emphasize total porosity or bulk permeability, whereas transport in standardized tests is governed by the fraction of pores that are actually connected along the dominant transport direction; such directional connectivity is rarely quantified alongside conventional pore-volume metrics. Second, while fractal analysis of MIP has been used to capture pore-network complexity, the scale dependence of fractal descriptors and their quantitative linkage to one-face saturated absorption and electrical charge passage have not been consistently established. Third, systematic comparisons of single-slag, single-SF, and combined GGBS + SF binders under constant mix design (e.g., constant w/b) remain scarce—particularly those that jointly interrogate (i) effective connected porosity from saturated absorption, (ii) electrical flux (ASTM C1202-type), and (iii) segment-aware fractal indices from MIP. These gaps motivate the present work and frame its novelty, namely, a joint evaluation of pore volume and connectivity to interpret water- and ion-transport in SCM-modified mortars [24,27,28,29].

In this study, mortar specimens with varying pore structures were prepared through the incorporation of different mineral admixtures. A series of experiments was conducted on hardened mortar samples to investigate the relationship between pore structure complexity, water absorption by capillary action, and directional pore connectivity. Water absorption and mercury intrusion porosimetry (MIP) data were collected and analyzed to better understand how pore characteristics influence transport behavior in mortar.

2. Materials and Experiments

2.1. Materials

The cement used in this study is an ordinary Portland cement (OPC, 32.5) conforming to relevant Chinese standards. It has a specific surface area of approximately 300 m²/kg and exhibits initial and final setting times of 160 and 220 min, respectively. The supplementary cementitious materials (SCMs, CaO + MgO ≈ 4.5%) include ground granulated blast-furnace slag (GGBFS) and silica fume supplied by Shenyang Wuzhou Mining Co., Ltd. (Shenyang, China). The detailed chemical composition and particle size distribution of the supplementary materials are presented in

Table 1. Chemical composition of the cement, slag and silica fume used in this study.

	SiO2	Al2O3	Fe2O3	CaO	MgO	SO3	Na2O	Loss on Ignition
Cement	21.72	5.81	4.33	62.41	1.73	2.56	0.5	1.47
GGBFS	55.01	28.5	8.05	2.39	2.19	0.02	0.08	2.45
Silica fume	86.20	0.70	0.50	0.35	0.15	0.01	0.2	2.1

and Figure 1, respectively. As shown in Table 1, the CaO/SiO₂ mass ratios are approximately 2.9 for cement, 0.04 for slag, and 0.004 for silica fume. Such low ratios in slag and silica fume suggest a higher availability of reactive silica, which is often associated with enhanced pozzolanic behavior when these materials are blended with cement. In terms of fineness, the silica fume is the finest among the three, followed by the slag, while the cement has the coarsest particle size distribution.

Five mortar mixes were prepared, including a reference mix (REF) with 100% OPC, and two mixes partially replacing OPC with either slag (OPC_SG15 and OPC_SG30), or a combination of both slag and silica fume (OPC_SG25SF5 and OPC_SG20SF10). The water-to-binder ratio and sand content were kept constant at 0.50 across all mixtures. A polycarboxylate superplasticizer (PCE, water-reducing rate 26.3%) supplied by Shenyang Dongling Concrete Admixture Co., Shenyang, China, was used at a fixed dosage of 0.1 wt.% by weight of binder. The PCE was added to enhance particle dispersion and ensure adequate flowability, which is particularly important for mixes containing 10 wt.% silica fume. To maintain experimental comparability, the dosage was kept constant across all mixtures. Dry materials were first mixed for 1 min, followed by the gradual addition of water and PCE. The mixture was then stirred for an additional 2 min to achieve uniform consistency. The detailed mix proportions are listed in

Table 2. Mix proportions of the mortar in this study.

Sample	Water (g)	Cement (g)	Sand (g)	Slag (g)	Silica Fume (g)	PCE (wt.%)
OPC_REF	50	100	150			0.1
OPC_SG15	50	85	150	15		0.1
OPC_SG30	50	70	150	30		0.1
OPC_SG25SF5	50	70	150	25	5	0.1
OPC_SG20SF10	50	70	150	20	10	0.1

.

2.2. Experiments

2.2.1. Strength Tests

Mortar specimens (40 mm × 40 mm × 160 mm) were prepared according to the mix proportions in Table 2 for mechanical strength testing. After 24 h of curing in molds, the specimens were demolded and transferred to a standard curing room. Flexural and compressive strengths were determined in accordance with EN 196-1:2016 [31]. Flexural strength was tested on the prismatic specimens, followed by compressive strength testing on the resulting halves, at both 7 and 28 days.

2.2.2. Water Absorption Tests

The one-sided rate of water absorption was measured following ASTM C1585-20 (one-face exposure) [32], with sealing and immersion head as described below. Duplicate specimens (40 mm × 40 mm × 160 mm) of mortar were used to determine the water absorption values of mortar at 7 and 28 days. The specimens were oven-dried at 105 °C in the fan-assisted chamber to a constant weight, and the initial weights of all the tested specimens were noted simultaneously; then the specimen surfaces were coated with epoxy resin, except one exposed surface. The cured specimens were immersed in water at 20 ± 2 °C and kept 2.5 cm water head. The schematic graph of the water absorption test is shown in Figure 2.

Due to the strong sealing effect of the epoxy coating (yellow line), only one bottom surface of the specimen was exposed to water, ensuring a unidirectional water transport path. As is well known, water can migrate through mortar only when the internal pores are interconnected. Therefore, the volume of penetrable pores can theoretically be estimated from the cumulative water absorption. The connected porosity (θ) was calculated using Equation (1):

$${\theta }_{1D}=m_f/({\rho }_w\cdot V_s)$$

(1)

where θ_1D is the connecting porosity accessible in the unidirectional test, m_f is mass of water absorbed at saturation, ρ_w is density of water, and vs. is total volume of the tested specimen. Unlike studies focusing on absorption rate, this work emphasizes the total amount of water uptake at saturation, which reflects the volume of interconnected pores and their contribution to long-term transport behavior.

2.2.3. Rapid Chloride Permeability Test (RCPT)

The Rapid Chloride Permeability Test (RCPT), following ASTM C1202 (Rapid Chloride Permeability Test: ASTM C1202) (including vacuum saturation and test conditions [33]), was used to assess the chloride ion permeability of mortar. A vacuum-saturated cylindrical specimen (100 mm diameter, 50 mm thick) was placed between two chambers containing 3 wt.% NaCl (cathode) and 0.3 M NaOH (anode). A 60 V DC voltage was applied for 6 h, and the total charge passed (electrical flux in coulombs), calculated by integrating the current over time, was used to indicate chloride permeability. Lower charge values reflect lower permeability and improved durability.

2.2.4. Mercury Intrusion Porosimetry (MIP)

Mercury intrusion porosimetry (AutoPore IV 9500, Micromeritics, Norcross, GA, USA) was employed to characterize the pore structure and measure the pore size distribution of the cementitious matrix. The method involves forcing a non-wetting liquid (mercury) into the pore network under increasing pressure. According to the Washburn equation [34,35], the relationship between applied pressure P and pore radius r is expressed as

$$P=-{\displaystyle \frac{2\sigma \cos \theta }{r}}$$

(2)

where σ is the surface tension of mercury and θ is the contact angle between mercury and the pore wall (approximately 117° for cement-based materials). As the pressure increases, mercury progressively intrudes into smaller pores, enabling the determination of the pore size distribution. The cumulative intruded volume reflects the total pore volume, while the differential intrusion curve indicates the frequency of pores within specific size ranges.

To minimize the influence of inert aggregates on pore structure analysis, cement paste samples without sand were used in MIP testing. This approach avoids the diluting effect of dense, nonporous sand particles and improves the representativeness of the measured pore characteristics of the binder phase. Paste samples were gently crushed into ~mm-sized fragments, immersed in ethanol to stop hydration, and oven-dried at 40 °C for 24 h until constant mass was reached. The dried samples were then weighed, placed into penetrometers, and subjected to low- and high-pressure intrusion cycles (up to 400 MPa), covering pore sizes from the micrometer to nanometer scale.

3. Results

3.1. Compressive and Flexural Strength

Figure 3 illustrates the mechanical properties of mortars incorporating various supplementary cementitious materials at 7 and 28 days. As shown in Figure 3a, all modified mixtures except OPC_SG30 (which shows comparable strength) exhibit higher 28-day compressive strength compared to the reference (OPC_REF). Notably, the OPC_SG20SF10 mix achieves the highest compressive strength of 46.54 MPa at 28 days, indicating that the combined use of slag and silica fume can effectively enhance the long-term strength development of mortar.

Figure 3b presents the flexural strength results at 7 and 28 days. At 28 days, all mixtures incorporating supplementary materials outperformed the reference sample, with the OPC_SG20SF10 mix again exhibiting the highest flexural strength of 5.83 MPa. This indicates a synergistic effect between slag and silica fume in enhancing not only compressive strength but also flexural performance, thereby improving the toughness and overall durability of the mortar. Among the two supplementary materials, silica fume exhibited a more pronounced effect than slag. This can be attributed to its finer particle size and higher pozzolanic reactivity, which contribute both to a better filling effect and to the acceleration of hydration reactions. These observations are consistent with previous findings reported in the literature [25].

3.2. Water Absorption Test of the Mortar

Figure 4 displays the saturated water absorption of mortar samples incorporating different supplementary cementitious materials after 7 and 28 days of curing. Overall, all mixtures show a notable reduction in water absorption with curing age, indicating ongoing densification of the microstructure. Among the 7-day results, OPC_SG15 and OPC_SG30 exhibit the highest absorption values 5417.88 g/m² and 5401.1 g/m², respectively, while OPC_SG20SF10 shows the lowest about 3411.14 g/m², suggesting that slag replacement alone may not significantly refine the pore structure at early ages.

At 28 days, the trend becomes more evident-blended systems, particularly those with both slag and silica fume (OPC_SG25SF05 and OPC_SG20SF10), exhibit much lower water absorption than the reference (reference: 2721.57 g/m²; OPC_SG25SF05: 2249.96 g/m²; OPC_SG20SF10: 2056.06 g/m²). The lowest value is observed in OPC_SG20SF10, highlighting the synergistic effect of slag and silica fume in enhancing matrix compactness and reducing capillary porosity. This result is consistent with the strength improvements seen in Figure 3.

3.3. Electrical Flux of Mortar

Figure 5 shows the electrical flux results of mortar samples prepared with different SCM blends compositions at 7 and 28 days, indicating the chloride ion permeability based on the RCPT method. All samples exhibited a noticeable reduction in electrical charge over time, reflecting improved microstructure and reduced ion transport pathways with curing age.

Among all mixtures, OPC_SG20SF10 showed the lowest charge passed at both 7 and 28 days, 3815 C and 3084 C, respectively, suggesting the most effective chloride resistance. This improvement is attributed to the synergistic effect of slag and silica fume in refining pore structure and enhancing matrix densification. The addition of silica fume, particularly in OPC_SG25SF05 and OPC_SG20SF10, led to a greater reduction in flux compared to mixtures with slag alone (OPC_SG15 and OPC_SG30), indicating its more pronounced role in blocking ionic transport.

3.4. Pore Structure and the Fractal Dimension

The micro-pore structure of mortar was analyzed using mercury intrusion porosimetry (MIP). The pore size distribution of specimens cured for 7 days is shown in Figure 6. All samples exhibit a dominant peak in the range of 100–1000 nm, corresponding to capillary pores, particularly gel pores and transition pores commonly observed at early hydration stages. Compared with OPC_REF, slag and silica fume-modified mortars show broader and slightly shifted peaks, indicating a combination of pore refinement (due to pozzolanic reaction and filler effects) and pore coarsening (possibly from delayed hydration or heterogeneous packing [3,26]). Notably, OPC_SG30 and OPC_SG25SF05 exhibit higher total pore volumes in the mesopore region, suggesting slower early-age hydration and a less dense microstructure. In addition, OPC_SG20SF10 displays a noticeable increase in pores larger than 1000 nm, which may be attributed to macropores or entrapped air voids formed during mixing, rather than intrinsic capillary porosity.

After 28 days of curing, the pore structure becomes more refined across all mixtures, as shown in Figure 6b. The intensity of the main peak decreases and the curves shift towards smaller pore sizes, especially in OPC_SG25SF05 and OPC_SG20SF10. These two blends demonstrate significantly lower pore volumes in the >100 nm range, indicating the filling and closing of larger capillary pores over time. This densification effect is attributed to the pozzolanic reaction of silica fume [25,27,28] and the continued hydration of slag, which jointly reduces pore connectivity and improves matrix compactness.

Table 3. Results of pore characteristics as measured by the MIP.

	Sample	Percent in Each Range [%]				Total Porosity (θT) [%]	Average Pore Diameter [nm]
		<10 nm	10–102 nm	102–103 nm	>103 nm
Cured for 7 days	OPC_REF	4.21	21.19	61.28	13.32	17.11	242.55
	OPC_SG15	2.11	15.78	63.62	18.49	18.86	286.57
	OPC_SG30	2.42	17.44	59.46	20.68	19.30	322.72
	OPC_SG25SF5	2.71	19.54	57.7	20.05	17.93	310.29
	OPC_SG20SF10	3.52	20.94	55.51	20.04	16.65	277.38
Cured for 28 days	OPC_REF	5.43	25.26	59.77	9.54	14.51	226.87
	OPC_SG15	3.91	27.8	57.0	11.3	16.73	254.90
	OPC_SG30	3.39	21.97	58.43	16.21	15.54	303.49
	OPC_SG25SF5	5.84	29.19	51.48	13.49	11.48	293.37
	OPC_SG20SF10	7.52	41.22	38.8	12.46	11.35	233.54

summarizes the quantitative pore structure characteristics, including the proportion of pores in different size ranges, total porosity, and average pore diameter. At 28 days, both OPC_SG25SF05 and OPC_SG20SF10 show significantly lower total porosity (11.48% and 11.35%, respectively) and reduced harmful pore content (>100 nm), compared to OPC_REF. Moreover, OPC_SG20SF10 exhibits the smallest average pore diameter (233.54 nm), confirming the microstructural refinement observed in Figure 6. These data support the conclusion that the combined use of slag and silica fume leads to a denser and less porous mortar structure.

Based on fractal theory, microscopic pore characteristics in mortar can be further revealed through the use of representative geometrical models. The Menger sponge, considered an idealized fractal structure [29,30], has been widely used to simulate the complex and irregular distribution of pores in cement-based materials. One of the fundamental concepts of fractal geometry is that as the observation scale decreases, increasingly finer structural details emerge [36,37,38], enabling a more comprehensive understanding of pore connectivity and spatial distribution. In such a model, a scaling invariance exists between the number and size of pores, expressed by the following relationships:

$$N_k={(r_k/R)}^{-D}$$

(3)

$$V(r)\propto r^{D-3}$$

(4)

where D is the fractal dimension of the pore volume, N_k is the number of pores of size r_k, and R is the side length of the reference cube. The total pore volume V can be estimated as

$$Lg(V)=(D-3)Lg(r)+C$$

(5)

According to Equation (4) and the Washburn equation linking pore size to intrusion pressure, a log–log plot of cumulative pore volume versus pore radius can be used to estimate the fractal dimension D_f, obtained from the slope of the fitted linear region (100–1000 nm). This dimension provides insight into the complexity and space-filling capacity of the pore network.

The log–log relationship between pore radius (log r) and cumulative pore volume (log V) is presented in Figure 7. As shown, inflection points are observed around pore sizes of approximately 50 nm and 550 nm, indicating that the log–log data cannot be accurately fitted using a single straight line. Instead, two linear segments are required to reflect the variation in fractal characteristics across different pore size ranges. This observation aligns with the fundamental understanding that the fractal dimension may vary across different size scales, reflecting changes in the structural behavior of the pore network. Therefore, when correlating fractal parameters with macroscopic properties such as transport or strength, the scale-dependence of the fractal behavior must be taken into account. As such, the fractal dimension of the pore structure was determined segmentally using Equation (4), and the calculated values based on mercury intrusion data are summarized in

Table 4. Fractal dimension of pores and correlation coefficient from calculations (Equation (5)).

	7 Days		28 Days
	Df	R2	Df	R2
OPC_REF	2.0723	0.9593	2.3283	0.9413
OPC_SG15	2.1954	0.9233	2.3736	0.9444
OPC_SG30	2.3344	0.9220	2.2496	0.9644
OPC_SG25SF05	2.3087	0.9522	2.3298	0.9768
OPC_SG20SF10	2.4387	0.9409	2.4663	0.9409

.

Table 4 summarizes the fractal dimension (D_f) and corresponding correlation coefficients (R²) of pore structures in mortars cured for 7 and 28 days. It is worth noting that other pore ranges, such as pores smaller than 100 nm or larger than 1000 nm may also exhibit fractal features. However, due to their relatively low volume fraction, their influence on key properties such as water transport, ionic conductivity, and pore connectivity is theoretically limited. Consequently, these pore ranges were not included in the fractal analysis presented in this study.

Higher fractal dimensions indicate a more complex and irregular pore network. Across both curing ages, all samples with supplementary materials exhibit higher D_f values than the reference, suggesting enhanced pore complexity and refinement. Notably, the OPC_SG20SF10 sample shows the highest D_f at both 7 days (2.4387) and 28 days (2.4663), implying a more developed and denser pore structure. The high R² values (>0.92) confirm the reliability of the fractal model fitting.

4. Discussion

4.1. Pore Structure and Its Relation to Water Absorption and Electrical Flux

Figure 8a illustrates the relationship between total porosity and two transport properties, saturated water absorption and electrical flux at both 7 and 28 days. A clear positive correlation is observed at both ages, where higher porosity is associated with increased water uptake and higher chloride ion permeability. Notably, the slopes of the fitted lines at 7 days are steeper than those at 28 days, suggesting that early-age porosity has a more pronounced effect on transport performance. This implies that at early hydration stages, the pore structure is more open and directly governs fluid and ion transport [39,40]. As hydration progresses, the pore network becomes denser and more tortuous, thereby weakening the direct influence of total porosity on transport behaviors. Consequently, reducing porosity at early ages appears to be more effective in improving durability-related properties.

Figure 8b shows the relationship between saturated water absorption, electrical flux, and the fractal dimension (D_f) of the pore structure. Compared to electrical flux, the variation in water absorption between 7-day and 28-day specimens is more pronounced, suggesting that water transport is more sensitive to microstructural evolution over time. In contrast to porosity, a negative correlation is observed: as D_f increases, both water absorption and chloride ion permeability tend to decrease. This implies that a higher D_f, indicating a more intricate and space-filling pore structure, contributes to enhanced resistance to fluid and ion transport. However, it should be noted that the degree of correlation is relatively low, particularly for electrical flux. This suggests that while the increase in fractal dimension generally has a positive effect on reducing transport properties, the relationship is not strictly linear and may be influenced by other factors such as pore connectivity and tortuosity.

These results reveal that both porosity and pore structure complexity, as quantified by the fractal dimension (D_f), play important roles in governing transport properties. While porosity determines the total volume of voids available for transport, fractal dimension reflects the spatial complexity, connectivity, and tortuosity of the pore network. Mixtures with low porosity and high fractal dimension generally exhibit better resistance to water and ion ingress, contributing to improved durability. The enhanced performance of silica fume-containing samples can be attributed to their ability to refine pore geometry and disrupt continuous capillary pathways. Although the correlation between fractal dimension and transport properties is not strictly linear, the observed trends suggest that increasing D_f tends to reduce water absorption and electrical flux, particularly at early ages when pore connectivity is more dominant. These findings underscore the need to consider both pore volume and structural characteristics in durability design.

4.2. Relationship Between Pore Connectivity and Fractal Dimension

In the following analysis, it is important to emphasize that during the saturated water absorption and electrical flux tests, pore connectivity along a one-dimensional direction is a prerequisite for fluid or ion transport. In other words, only those pores that are interconnected along the test direction can contribute to measured transport behavior. Mercury intrusion porosimetry was used to determine the percolation-accessible porosity and the distribution of entry-throat sizes based on the Washburn equation [17,34] with an assumed contact angle. The measured porosity also depends on connectivity and intrusion pathways. Closed pores and pores shielded by constrictions that are not penetrated at the applied maximum pressure are not captured. If we define the effective porosity in the water transport direction and express it as a fraction of the total porosity (as described in Equation (6)), this ratio can serve as an indicator of pore connectivity.

$$C_{1D}={\displaystyle \frac{{\theta }_{1D}}{{\theta }_T}}\times 100$$

(6)

Such directional connectivity is theoretically governed by the development paths of pores at the microscale, which may not always correspond directly to total pore volume or even fractal dimension. A conceptual illustration of this idea is shown in Figure 9, where (a) demonstrates a case of low fractal dimension but high pore connectivity, while (b) shows a more complex but disconnected pore network. This phenomenon can be attributed to the nature of fractal growth and the behavior of hydration products. A lower fractal dimension often implies a more anisotropic or directionally biased structure, where pore-forming mechanisms—such as the oriented growth of hydration products or localized bleeding and settlement—lead to preferential alignment and thus facilitate one-dimensional connectivity. In contrast, a higher fractal dimension reflects a more space-filling and isotropic development of pores, which increases overall complexity but may disrupt continuous paths along a single direction.

Fractal dimension describes how tightly the pore space fills at smaller scales, but by itself it does not indicate connectivity. Directional transport needs a continuous chain of pore openings large enough to carry fluid or ions along the test direction. In the 100 to 1000 nm window, for samples with similar total porosity, a higher fractal dimension often means a finer, more branched network with more dead ends, so the chance of a through path is lower. This is conditional: hierarchical networks with a few larger bridging pores can raise connectivity even at high fractal dimensions, while narrow throats can reduce it at the same fractal dimension. Figure 10 reflects this conditional inverse trend and is consistent with reduced flux and improved impermeability. This hypothesized mechanism is conceptually depicted in Figure 9. Gas transport shows the same pattern. Permeability is controlled by whether pore throats connect to form a path, and MIP-derived throat sizes often track gas permeability in cement-based materials. At sub-micron pores, Knudsen effects further reduce gas flow [41].

However, discrepancies in this correlation are observed. The reliability of pore structure data obtained from mercury intrusion porosimetry (MIP) can be affected by the so-called “inkbottle” effect [24], which may distort the estimation of total porosity and consequently affect the derived fractal dimension. To better understand the development of pore connectivity, future studies should consider incorporating in situ characterization techniques capable of capturing the real-time evolution of microstructure during hydration. Approaches such as X-ray microtomography, neutron imaging, or advanced nuclear magnetic resonance (NMR) methods could offer deeper insight into the origin and progression of connected pore networks from a microstructural and hydration kinetics perspective. Further work in this direction is needed to validate and extend the current findings.

5. Conclusions

This study investigated the influence of slag and silica fume on the mechanical properties, transport behavior, and pore structure of specimen. Their combined use improved strength, reduced water absorption and chloride permeability, and refined pore structure. Mercury intrusion porosimetry and further fractal analysis showed that lower porosity was associated with significantly improved resistance to water and ion transport. However, the relationship was not strictly proportional, as pore connectivity (inversely related to fractal dimension) critically influenced transport behavior. Further analysis indicated that higher fractal dimensions tended to reduce directional connectivity, suggesting that increased structural complexity may hinder the formation of continuous transport pathways. These findings underscore the importance of considering both pore volume and connectivity in durability evaluation. In situ techniques are recommended to better capture the evolution of pore networks during hydration.