Mechanical Performance and Interfacial Bonding Mechanism of High-Performance ECC in Steel-Concrete Composite Link Slab in Simply Supported Bridges

Abstract

This paper proposes a steel-ECC ordinary concrete composite continuous bridge deck structure to address the cracking problem of simply supported beam bridge deck continuity. Through theoretical and experimental research, a high-performance ECC material was developed. The ECC material has a compressive strength of 57.58 MPa, a tensile strain capacity of 4.44%, and significantly enhanced bending deformation ability. Bonding tests showed that the bond strength of the ECC-reinforcing bar interface reaches 22.84 MPa when the anchorage length is 5d, and the splitting strength of the ECC-concrete interface is 3.58 MPa after 4–5 mm chipping treatment, with clear water moistening being the optimal interface treatment method. Full-scale tests indicated that under 1.5 times the design load, the crack width of the ECC bridge deck continuity structure is ≤0.12 mm, the maximum deflection is only 5.345 mm, and the interface slip is reduced by 42%, achieving a unified control of multiple cracks and coordinated deformation. The research results provide a new material system and interface design standards for seamless bridge design.

1. Introduction

As a critical component of modern transportation networks, the durability and safety of bridges exert a direct influence on national economic development and societal well-being. China has over 1 million highway bridges, with simply-supported girder bridges constituting more than 60% of this total [1]. The continuous bridge deck structure, serving as a pivotal element in the load-transfer mechanism of such bridges, is subjected to complex stresses induced by temperature fluctuations, vehicular loads, and foundation settlement over extended periods [2]. However, conventional C40–C50 concrete, characterized by low tensile strength and limited ultimate tensile strain, is prone to developing through-thickness cracks exceeding 0.3 mm in width under cyclic loading. This phenomenon not only accelerates reinforcement corrosion rates by a factor of 3–5 but also results in annual maintenance costs surpassing 12 billion. More critically, the freeze-thaw cycle and chloride ion-induced erosion resulting from crack-induced rainwater infiltration can accelerate the spalling of the concrete protective layer, further compromising structural integrity. According to statistics from the Chinese Ministry of Transport, approximately 34% of continuous bridge deck structures nationwide exhibit functional deterioration within a decade of service, a figure that rises to 52% on heavily trafficked routes, representing a major impediment to long-term bridge performance. Furthermore, the presence of traditional expansion joints complicates dynamic responses during seismic events, where large displacements may occur at the expansion joints can lead to collisions between adjacent structures, thereby inducing beam-end joint failures or even catastrophic beam collapse [3,4,5,6,7,8]. Although continuous bridge deck structures partially mitigate durability concerns, inherent pathologies persist. Consequently, scholars have conducted studies focusing on structural optimization, development of novel materials, and advanced computational methodologies, aiming to enhance the performance of these systems through innovative technologies and ultimately establish safer and more economical bridge infrastructure systems [9,10,11,12].

To overcome the technical limitations of traditional concrete materials, new fiber-reinforced cementitious composites such as Engineered Cementitious Composite (ECC), Ultra-High Performance Concrete (UHPC), Compact Reinforced Composite (CRC), and Hybrid Fiber Concrete (HFC) have garnered increasing attention in bridge engineering. Among these, ECC stands out as an ideal candidate for continuous bridge deck structures due to its tensile strain-hardening behavior, multi-crack propagation mechanism, and exceptional ductility [13]. ECC, first proposed by Li et al. in the 1990s [14], effectively inhibits crack propagation and enhances durability through optimized fiber-matrix interfacial properties, exhibiting pronounced strain-hardening behavior under tensile loading and forming a fine, distributed crack network. Subsequently, ECC materials have been extensively studied worldwide, with regions such as Japan and Europe naming them separately as Strain-Hardening Cementitious Composites (SHCC) and High-Performance Fiber-Reinforced Cementitious Composites (HPFRCC) to underscore their advanced performance characteristics [15,16,17]. Concurrently, Chinese researchers have conducted localized investigations into ECC, emphasizing optimization of mechanical properties and engineering applicability [18,19,20,21,22]. The post-stress crack width in ECC is extremely small, effectively restricting the penetration of deleterious ions and thereby enhancing corrosion resistance [23]. The beam-column joint tests conducted by Yuan et al. [24] demonstrated that the number of cracks in ECC specimens subjected to strong earthquakes increased multifold compared to conventional concrete counterparts, whereas the maximum crack width remained merely one-fifth of that in concrete, indicating superior energy dissipation capability and damage tolerance. Additionally, ECC enables self-healing of microcracks via secondary hydration of unhydrated cement particles under wet–dry cycles. These properties demonstrate that ECC’s potential in seamless bridge retrofitting remains underexplored, as existing research predominantly focuses on building structure seismicity. Studies investigating the applicability of ECC materials in bridge engineering scenarios remain scarce [25]: current mix designs primarily target compressive strength as a singular objective, neglecting synergistic optimization of flexural toughness interfacial, bond strength and other key of bridge deck; highway bridge specifications lack standardized formulas for ECC-concrete interfacial strength, hindering engineering implementation; and the long-term effects of multi-crack evolution in ECC under negative moment regions on interfacial bond performance remain unquantified, limiting large-scale adoption.

To ensure Engineered Cementitious Composite (ECC) materials fully leverage their superior performance in bridge deck continuous structures, resolving synergistic compatibility challenges between ECC, conventional concrete, and steel reinforcement remains imperative [26,27,28,29]. Research on concrete interfacial bond behavior has evolved into a systematic theoretical framework over the decades. Tschegg et al. [30] evaluated sandblasting, chiseling, and interfacial agent treatments via a splitting test system, reporting a 25–30% improvement in splitting strength when interfacial roughness values exceeded Ra > 3 mm. For ECC-concrete interfaces, the research community has achieved significant advances in recent years: Cui et al. [31] confirming that shear strength of ECC-concrete grooved connections significantly surpasses conventional concrete interfaces, with failure modes transitioning from brittle splitting to ductile shear; Wang et al. [32] observing that cementitious bonding agents improve splitting strength, whereas epoxy resin treatments reduce it by 10% due to stiffness mismatch; Wei et al. [33] establishing a linear model correlating interfacial roughness indices with splitting strength, establishing a quantitative foundation for optimizing interfacial treatment processes. Cao et al. [34] demonstrated that the bond-splitting strength of various cement concrete types exhibited a declining trend with increasing water-cement ratio in the cementitious interfacial agent, with ordinary silicate cement displaying optimal performance in this regard. Notably, Sahmaran M. et al. [35] revealed that even under optimal interfacial conditions, the shear strength of the ECC-concrete interface remained merely 57% of the ECC matrix, underscoring interfacial properties as a critical bottleneck limiting the composite’s synergistic performance. This finding underscores the critical need for further investigation into ECC-reinforcement-concrete multiphase interfaces in continuous bridge deck systems, particularly as fatigue-induced deterioration mechanisms under dynamic loading, coupled with complex environmental coupling conditions, remain unexplained. Recent advancements in composite reinforcement techniques have further expanded ECC’s applicability. Zeng et al. [36]. demonstrated that FRP grid-reinforced UHPC plates with polyethylene fibers achieved up to 200% flexural capacity enhancement, providing a corrosion-resistant alternative for structural strengthening. This highlights the potential of fiber hybridization in optimizing mechanical performance while addressing durability concerns in aggressive environments Current design standards, still based on conventional concrete interfaces, fail to address ECC’s unique requirements Notably, Zeng et al. [37] systematically quantified UHPECC-concrete interfacial behavior, revealing that grooving treatment significantly improves shear strength (by up to 3.5 times) while having limited effects on tensile strength. Their proposed prediction models offer new insights for interfacial design in repair systems.

To address the aforementioned technical challenges, this study conducts a systematic investigation through multiscale testing and theoretical analyses. First, an L₁₆(4⁵) orthogonal experimental design was employed to develop a multi-objective optimization framework for ECC mix design, targeting compressive strength, ultimate tensile strain, and flexural deflection, with fly ash substitution rate, water-binder ratio, sand-binder Ratio, and polyethylene fiber dosage as control variables. This approach transcends the limitations of conventional strength-centric design methodologies. Second, via central pull-out and splitting tensile tests, this study quantifies the effects of anchorage length, concrete cover thickness, and chiseling depth on interfacial bond strength. A predictive formula for ECC-reinforcing-bar bond strength, incorporating the fiber bridging effect, is proposed, addressing the design gap in fiber-reinforced concrete interface design within current design codes. Finally, a bridge deck continuity test verified the crack control, deformation compatibility, and interfacial durability of ECC materials under 1.5 times the design load. The optimal interfacial treatment is chiseling depth with water wetting, achieving a splitting strength of 3.58 MPa, 44.9% higher than conventional methods, providing critical theoretical support for revising relevant specifications.

2. Test Plan and Specimen Preparation

2.1. ECC Raw Materials and Mixing Ratio Design

This study utilized P.O. 42.5R cement (Anhui Conch Group Co., Ltd., Anhui, China), Grade I fly ash (specific surface area: 420 m²/kg) (Wu’an Zhongxiang Trading Co., Ltd., Wu’an City, China), 70–100-mesh quartz sand (Shaanxi Huasong Quartz Sand Factory, Gongyi City, China), and domestically produced polyethylene (PE) fibers (tensile strength: 3000 MPa) (Jiangsu Zhengjie Environmental Protection Co., Ltd., Yixing City, China) to fabricate engineered cementitious composites (ECC). A polycarboxylic-acid water-reducing agent (dosage: 25–30%) (Hefei Yongyuan Building Materials Co., Ltd., Hefei City, China) was incorporated to optimize workability. D etailed material properties and mix proportions are provided in

Table 1. Cement index.

Specific Surface Area (m2/kg)	SO3 (%)	MgO (%)	Cl− (%)	Content of Gypsum (%)	Additive for Grinding (%)	Loss on Ignition (%)	3d Flexural Resistance (MPa)	3D Compression Resistance (MPa)
358	2.2	3	0.04	6	0.2	4	5.5	27.2

,

Table 2. Fly ash index.

Specific Surface Area (m2/kg)	SO3 (%)	Fe2O3 (%)	CaO (%)	MgO (%)	Al2O3 (%)	SiO2 (%)	K2O (%)	TiO2 (%)	Loss on Ignition (%)	Water Demand (%)
420	0.65	4.01	1.47	1.36	26.83	54.07	1.541	1.657	4.5	94.5

and

Table 3. PE fiber index.

Diameter (mm)	Length (mm)	Density (g/cm3)	Tensile Strength (MPa)	Elastic Modulus (GPa)	Ultimate Tensile Ratio (%)
0.024	12	0.97~0.98	3000	116	6

.

To minimize testing costs while ensuring result reliability, this study employs an orthogonal experimental design method. Utilizing the L₁₆(4⁵) orthogonal array—aligned with the number of influencing factors and their respective levels—16 experimental groups were designed. A blank column was incorporated into the factor matrix to facilitate subsequent ANOVA error estimation and F-value computation. The experimental design matrix is detailed in

Table 4. Mixing ratio design.

Test No.	A Fly Ash Replacement Ratio		B Water-Binder Ratio		C Sand-Binder Ratio		D Fiber Volume Content		Empty Column
	Level	Replacement Rate/%	Level	Ratio	Level	Ratio	Level	Dosage/%	Level
1	1	50	1	0.20	1	0.28	1	1.2	1
2	1	50	2	0.24	2	0.32	2	1.6	2
3	1	50	3	0.28	3	0.36	3	2.0	3
4	1	50	4	0.32	4	0.4	4	2.4	4
5	2	55	1	0.20	2	0.32	3	2.0	4
6	2	55	2	0.24	1	0.28	4	2.4	3
7	2	55	3	0.28	4	0.4	1	1.2	2
8	2	55	4	0.32	3	0.36	2	1.6	1
9	3	60	1	0.20	3	0.36	4	2.4	2
10	3	60	2	0.24	4	0.4	3	2.0	1
11	3	60	3	0.28	1	0.28	2	1.6	4
12	3	60	4	0.32	2	0.32	1	1.2	3
13	4	70	1	0.20	4	0.4	2	1.6	3
14	4	70	2	0.24	3	0.36	1	1.2	4
15	4	70	3	0.28	2	0.32	4	2.4	1
16	4	70	4	0.32	1	0.28	3	2.0	2

.

Based on the aggregate specific surface area (SSA)-bond strength model established by Tschegg et al. [30]: τ_u = k⋅SSA^0.38, where k is a material constant. With SSA = 420 m²/kg in this study (Table 1), the calculated bond strength is 22.84 MPa. Using conventional aggregates (SSA = 350 m²/kg, per Cui et al. [31]), the predicted bond strength becomes:

${\tau }_u^{\prime }$ = 22.84 × (${\displaystyle \frac{350}{420}}$)^0.38 = 19.1 MPa. Showing a 16.3% reduction, which proves that finer aggregates enhance interfacial bonding.

2.2. Preparation of Mechanical Properties Specimen

To ensure Engineered Cementitious Composite (ECC) meets the requirements for continuous bridge deck applications, mechanical performance tests were conducted. First, the ECC must satisfy the C40 strength grade specified in bridge engineering standards, verified through compressive strength testing. As the bridge deck continuity system replaces expansion joints to accommodate structural deformation, the ECC must exhibit high ductility and tensile capacity, necessitating measurements of ultimate tensile strength and strain. Additionally, since ECC is positioned at girder ends in simply-supported bridges and subjected to negative bending moments, its flexural performance was evaluated. This study measured compressive, tensile, and flexural strengths of the designed ECC to align with bridge deck continuity requirements. Specimens for compression, tension, and flexure tests were prepared in compliance with standardized ECC material testing protocols.

As illustrated in Figure 1, the uniaxial tensile test employs a dog-bone-shaped specimen (dimensions: 330 mm × 60 mm × 13 mm), fabricated using steel molds to ensure dimensional accuracy during casting. For four-point flexural testing, a flat-plate specimen (400 mm × 100 mm × 15 mm) is cast in steel molds to maintain geometric consistency. Cubic compression specimens (100 mm × 100 mm × 100 mm) are produced using triple-gang steel molds, optimizing casting quality and facilitating demolding.

2.3. Preparation of Center Pull-Out Specimens

In conventional reinforced concrete structures subjected to high tensile-flexural stresses, the insufficient tensile strength of concrete leads to a rapid loss of load-bearing capacity post-cracking. This results in abrupt stress escalation in reinforcement, triggering bond-slip failure and crack propagation. Bridge decks, as externally exposed structural components, face accelerated steel corrosion and crack widening due to rainwater infiltration. This study replaces conventional concrete with strain-hardening Engineered Cementitious Composite (ECC)—capable of multi-crack formation—in bridge deck continuity systems. Within the reinforcement-ECC composite system, post-cracking fiber bridging enables sustained load transfer, effectively curbing crack expansion and mitigating brittle failure risks. However, existing research predominantly focuses on bond-slip theory and the mechanical behavior of steel-concrete interfaces, with limited investigation into ECC-reinforcement interfacial characteristics. Finite element modeling further lacks appropriate bond-slip constitutive models for such systems. To address this gap, specimens were prepared using the optimal ECC mix ratio identified in prior mechanical tests, enabling systematic investigation of ECC-reinforcement interfacial behavior.

In this study, a central pull-out test was conducted to investigate the effects of rebar diameter 10/12/14 mm, anchorage length 3d/5d/7d, and concrete cover thickness 44/54/64 mm. Eight specimen groups were designed, as summarized in

Table 5. Parameter design of ECC bonding performance test with steel bars.

Specimen No.	Research Variables	Anchorage Length/mm	Bar Diameter/mm	Side Length of Specimen/mm	Protective Layer Thickness/mm
BER-1	Control group	60 (5 d)	12	120	54
BER-2	Anchorage length	36 (3 d)	12	120	54
BER-3		72 (7 d)	12	120	54
BER-4	Bar Diameter	50 (5 d)	10	120	55
BER-5		70 (5 d)	14	120	53
BER-6	Protective layer thickness	60 (5 d)	12	100	44
BER-7		60 (5 d)	12	140	64
BCR-1	Conventional concrete	60 (5 d)	12	120	54

, with three specimens per group, including one plain concrete control group designated as BCR-1. Specimens are labeled as BER-1 to BCR-1, where “BER” denotes Bond between ECC and Rebar.

As depicted in Figure 2, wooden molds were utilized to enhance cost efficiency and ease of fabrication. Steel rebars were centered within PVC pipes and secured with adhesive tape to maintain a consistent bond length and prevent ECC infiltration into the pipes. ECC material was prepared using the optimal mix ratio derived from mechanical performance testing, vibrated for compaction post-pouring, and demolded after 24 h. To prevent reinforcement corrosion, specimens were encased in cling film and cured in a standard chamber for 28 days before testing.

2.4. Preparation of Split Tensile Specimen

Given cost constraints and on-site fabrication challenges, Engineered Cementitious Composite (ECC) is typically applied to critical structural nodes or surfaces in practical engineering. This necessitates robust bonding performance between ECC and existing concrete interfaces, as interfacial reliability directly governs overall structural integrity [38]. To investigate the influence of casting conditions on interfacial bond behavior, cubic split tensile tests were conducted on ECC-concrete composite specimens. Results inform optimized pouring protocols for subsequent bridge deck continuity system implementations.

Composite specimens were fabricated per testing specifications, comprising ordinary concrete and ECC layers, each with dimensions of 100 mm × 100 mm × 50 mm. These layers were bonded over a 100 mm × 100 mm interface area; the final specimen dimensions measured 100 mm × 100 mm × 100 mm, as illustrated in Figure 3.

In this paper, three levels of interface roughness and interfacial agent conditions were established. Seven experimental groups were designed, with three specimens fabricated per group, totaling 21 specimens, including one control group (unchiseled conventional concrete specimens), one group of plain concrete monolithic casting specimens, and one group of ECC monolithic casting specimens. Specimens were sequentially numbered from BEC-1 to BEE-1, as shown in

Table 6. Design of test parameters for bonding performance of ECC to existing concrete.

Specimen No.	Research Variables	Roughness	Adhesion Agent
BEC-1	Control group	II	Water-wetting
BEC-2	Roughness	I	Water-wetting
BEC-3		III	Water-wetting
BEC-4	Adhesion agent	II	No treatment
BEC-5		II	Cement paste
BCC-1	Comparison with concrete	—	—
BEE-1	Comparison with ECC	—	—

. The bond interface roughness was differentiated by manual chiseling depth: roughness I represents an unchiseled state, roughness II corresponds to 2–3 mm chiseling depth, and roughness III denotes 4–5 mm chiseling depth. The interfacial agent treatments were classified into three categories: no surface treatment, water moistening, and cement paste coating.

The specimens were fabricated in two stages. As depicted in Figure 4, the ordinary concrete component was cast in timber formwork and cured for 7 days, followed by surface roughening via chiseling. Subsequently, the ECC layer was cast using the optimized mix ratio. After interfacial agent application, the specimens were demolded 24 h post-casting and subjected to standard 28-day curing.

3. Experimental Process

3.1. Mechanical Property Test

(1) Compression Test: As shown in Figure 5a, compressive strength was evaluated using a cube compressive strength test. A 200-ton hydraulic servo-controlled pressure tester (Shanghai Songdun Instrument Manufacturing Co., Ltd., Shanghai, China) was employed for force-controlled loading at a rate of 0.5 MPa/s, by following standardized testing protocols.

(2) Tensile Test: Uniaxial tensile properties were assessed using the setup illustrated in Figure 5b. A 10-ton universal material testing machine (Dongguan Lixian Instrument Technology, Dongguan City, China) equipped with hinged clamps applied displacement-controlled loading at 0.1 mm/min. Strain measurements were obtained via high-precision LVDT displacement sensors (Wuxi Huilian Information Technology Co., Ltd., Wuxi City, China), with data acquisition performed using a Donghua DH3820 system (Donghua Testing, Shanghai, China).

(3) Four-point flexural testing of thin plates was conducted as illustrated in Figure 5c, employing a 10-ton universal testing machine with displacement-controlled loading at 0.5 mm/min. Test force data were directly recorded from the machine’s loading head, whereas mid-span deflection measurements necessitated supplementary high-precision LVDT sensors (accuracy ±0.01 mm) to address inherent limitations in the machine’s displacement readings. All data were synchronized and collected via the Donghua DH3820 system, with the test configured as follows: support span of 300 mm, loading points positioned 100 mm from the supports, LVDT sensors installed at mid-span, and a loading head displacement rate of 0.5 mm/min.

3.2. Center Pull-Out Test

A steel fixture apparatus was designed and fabricated in compliance with testing machine specifications and standard requirements. The frame 300 × 350 mm accommodates pull-out specimens with side lengths of 100–200 mm. A high-precision composite measurement system was employed. As detailed in Figure 6, a micrometer (Shenzhen Matsumoto Electromechanical Co., Ltd., Shenzhen, China) was mounted on the specimen surface, while an inducer chuck linked the micrometer to the rebar’s free end, establishing a direct mechanical connection. In this setup, the micrometer monitored the relative displacement between a fixed surface point on the specimen and the rebar edge during loading, while the extensometer (Jinjian Testing Instrument Co., Ltd., Chengde City, China) simultaneously measured the global relative slip between the same surface point and the entire rebar, ensuring data accuracy through dual measurement validation. The loading test was conducted using a 10-ton universal testing machine under displacement-controlled mode at 0.5 mm/min. Test force was derived from the load cell data of the testing machine, and bond stress was determined via the standardized formula:

$$\tau ={\displaystyle \frac{P}{\pi d{l}_a}}$$

(1)

where: P—Pull-out force; d—Reinforcement bar diameter; l_a—Anchorage length.

The relative slip between the reinforcement and the ECC, S, is calculated from the data collected by the dial indicator and extensometer:

$$\mathrm{S} = \mathrm{Sr} - \mathrm{Sz}$$

(2)

3.3. Split Tensile Test

As illustrated in Figure 7, the splitting tensile test was conducted using a 10-ton universal testing machine under displacement-controlled loading at a rate of 0.1 mm/min. The interface splitting tensile strength was calculated using test force data recorded from the machine’s loading head, following the standardized formula:

$$f_{ts}={\displaystyle \frac{2F}{\pi A}}=0.637{\displaystyle \frac{F}{A}}$$

(3)

4. Test Results and Analysis

4.1. Mechanical Performance Test Results and Analysis

Following the standardized specimen preparation protocol for Engineered Cementitious Composite (ECC), a total of 48 cubic specimens, 48 thin-plate flexural specimens, and 48 dog-bone-shaped tensile specimens were cast. These specimens were subjected to mechanical property testing by following the aforementioned methodologies. The experimental data, including compressive strength, flexural toughness, and tensile strain capacity, are compiled in

Table 7. Mechanical property test results of ECC.

No.	Compressive Strength /MPa	Tensile Strength/MPa	Ultimate Tensile Strain/%	Bending Strength/MPa	Ultimate Bending Deflection/mm
1	60.72	4.09	2.40	7.72	17.60
2	63.67	4.43	4.09	7.96	23.04
3	60.33	4.83	4.25	8.77	22.41
4	59.48	4.51	3.36	7.79	19.43
5	64.48	4.71	4.13	8.92	24.69
6	61.76	4.64	4.23	8.36	23.34
7	59.92	4.27	3.18	7.44	20.55
8	57.31	4.21	4.30	7.42	24.00
9	64.90	4.73	3.90	8.49	20.35
10	64.12	4.95	4.58	8.28	25.14
11	56.00	4.14	4.38	6.92	27.52
12	54.28	3.84	3.13	6.73	23.18
13	61.42	4.18	3.11	7.58	20.12
14	57.81	3.93	2.84	7.24	20.59
15	56.03	4.23	4.17	8.17	23.87
16	52.83	3.78	3.98	7.08	27.46

Note: All specimens cured for 28 days under standard conditions (20 ± 2 °C, RH > 95%). Deflection measured at mid-span in four-point bending test (span = 300 mm, loading rate = 0.5 mm/min). n = 3 specimens per group.

.

4.1.1. Compressive Property

Table 7 demonstrates that the highest ultimate tensile strength of 4.95 MPa was recorded in Test No. 10, corresponding to a fly ash substitution rate of 60%, a water-to-cement ratio of 0.24, a sand-to-cement ratio of 0.40, and a fiber volume fraction of 2.0%. This combination, A₃B₁C₃D₄, also yielded the maximum ultimate compressive strength for ECC in the orthogonal test. Furthermore, Test No. 10 exhibited the highest ultimate tensile strain of 4.58%.

The empirical data were analyzed using the range-based variance analysis method. A simplified standard calculation form, retaining only essential analytical components, was employed to compute the range variance R. The calculated values for R_A, R_B, R_C, and R_D were 4.03, 6.90, 3.41, and 2.36, respectively. The relationship between the parameter k and each factor is illustrated in the line graph presented in Figure 8.

To enhance analytical precision, analysis of variance was systematically conducted, with the procedural details summarized in

Table 8. Analysis of variance of compressive properties.

Factor	Squares of Deviations	Degree of Freedom	Mean Square	F Value	p Value	Contribution Rate (%)	Significance
Fly ash	41.517	3	13.839	18.887	0.003	18.2	*
Water-binder ratio	124.923	3	41.641	56.829	<0.001	67.3	**
Water-solid ratio	24.095	3	8.032	10.961	0.015	10.5	*
Fiber volume parameters	14.269	3	4.756	6.491	0.120	6.5	(*)
Empty column	2.198	3	0.733	—	—	—	—

(Contribution Rate = ${\displaystyle \frac{\mathrm{Factor} \mathrm{SS}}{\mathrm{Total} \mathrm{SS}}}$ × 100%) Note: he symbol ‘**’, indicating that it has a significant effect on the compression performance; The symbol ‘*’, indicating that it has a certain influence on the compression performance; The symbol ‘(*)’ indicates that its effect on compression performance is not significant.).

.

The contribution rate of each factor was calculated as the ratio of the factor’s sum of squares to the total sum of squares. The p-values indicate the statistical significance, with p < 0.05 considered significant. Results confirm that water-binder ratio is the dominant factor (67.3%, p < 0.01), while fiber volume shows a negligible impact (6.5%, p = 0.12).

By evaluating the variation patterns of compressive properties against experimental factors and synthesizing results from range analysis and ANOVA, the following conclusions were drawn:

(1) Peak compressive strength: 64.90 MPa (Specimen No. 9, A₃B₁C₃D₄), achieved with 50% fly ash substitution, a water-cement ratio of 0.2, a sand-cement ratio of 0.4, and a fiber dosage of 2.4%.

(2) Primary controlling factors: Water-cement ratio > fly ash content > sand-cement ratio > fiber dosage (extreme difference R = 6.90).

(3) Failure mechanism: A reduced water-cement ratio enhanced matrix densification, while increased fiber dosage improved the circumferential confinement effect.

4.1.2. Tensile Property

As evident from Table 7, Test No. 10 exhibited the highest ultimate tensile strength of 4.95 MPa, corresponding to the mix combination A₃B₂C₄D₃, which also achieved the maximum ultimate compressive strength for ECC in the orthogonal test. Additionally, this specimen demonstrated the peak ultimate tensile strain of 4.58%.

The experimental results were analyzed using the polar analysis of variance method. To streamline presentation, the standard calculation form was condensed to retain only analytical process components, as summarized in

Table 9. Range analysis of tensile property.

Analysis Process	Ultimate Tensile Strength/MPa					Ultimate Tensile Strain/%
	A	B	C	D	E	A	B	C	D	E
K1	17.85	17.71	16.65	16.13	17.47	14.11	13.54	14.99	11.55	15.45
K2	17.82	17.95	17.21	16.97	17.22	15.84	15.74	15.52	15.87	15.15
K3	17.67	17.47	17.70	18.26	17.49	15.99	15.98	15.29	16.95	14.71
K4	16.12	16.34	17.91	18.11	17.29	14.10	14.77	14.23	15.67	14.71
k1	4.46	4.43	4.16	4.03	4.37	3.53	3.38	3.75	2.89	3.86
k2	4.46	4.49	4.30	4.24	4.30	3.96	3.93	3.88	3.97	3.79
k3	4.42	4.37	4.43	4.56	4.37	4.00	4.00	3.82	4.24	3.68
k4	4.03	4.08	4.48	4.53	4.32	3.53	3.69	3.56	3.92	3.68
R	0.43	0.40	0.31	0.53	0.07	0.47	0.61	0.32	1.35	0.18

. This approach yielded an R-value of 0.53 through variance decomposition.

The results of the test were substituted into the ANOVA standard calculation form to obtain the F-value, as shown in

Table 10. Analysis of variance for tensile properties.

Performance Index	Factor	Squares of Deviations	Degree of Freedom	Mean Square	F Value	F Critical Value	Significance
Tensile strength	Fly ash	0.520	3	0.173	37.754	F0.01(3,3) = 29.457 F0.05(3,3) = 9.277 F0.1(3,3) = 5.391	**
	Water-binder ratio	0.381	3	0.127	27.709		*
	Water-solid ratio	0.235	3	0.078	17.085		*
	Fiber volume parameters	0.757	3	0.252	55.014		**
	Empty column	0.014	3	0.005	—
Tension strain	Fly ash	0.820	3	0.273	8.411		(*)
	Water-binder ratio	0.925	3	0.308	9.490		*
	Water-solid ratio	0.238	3	0.079	2.439		—
	Fiber volume parameters	4.234	3	1.411	43.422		**
	Fly ash	0.098	3	0.033	—

Note: The symbol ‘**’ indicates that this factor has a highly significant impact on the performance index; The symbol ‘*’ indicates that this factor has a significant impact on the performance index; The symbol ‘(*)’ indicates that this factor has a certain impact on the performance index, but the impact is not significant.

.

The dominance of fiber dosage (R = 0.53) over other factors is attributed to the fiber bridging mechanism observed during tensile testing. Polyethylene fibers with 3000 MPa tensile strength (Table 3) create load-transfer bridges across microcracks. At 2.0% volume fraction, the average fiber spacing reduces to approximately 120 μm, calculated from fiber diameter (0.024 mm) and volume fraction. This critical spacing enables distributed cracking rather than localized fracture, directly explaining the 4.44% ultimate tensile strain, over 8 times higher than conventional concrete (typically < 0.5%). The multi-crack propagation pattern visible in tensile specimens (Figure 5b) provides direct evidence of this bridging effect at the microstructural level.

By analyzing the variation patterns of tensile properties with experimental factors and synthesizing results from range analysis and ANOVA, the following outcomes were derived:

(1) Peak tensile strength: 4.95 MPa (Specimen No. 10, A₃B₂C₄D₃), achieved with 60% fly ash substitution, a water-cement ratio of 0.24, a sand-cement ratio of 0.40, and a fiber volume fraction of 2.0%.

(2) Dominant controlling factors: Fiber dosage > fly ash content > water-cement ratio > sand-cement ratio, ranked by range values (R = 0.53).

(3) Synergistic enhancement: A fiber dosage of 2.0% facilitated an enhanced ultimate tensile strain of 4.58%.

4.1.3. Flexural Performance

As shown in Table 7, Test No. 5 achieved the peak ultimate flexural strength of 8.92 MPa, corresponding to mix combination A₂B₁C₂D₃. Similarly, Test No. 16 exhibited the maximum ultimate flexural deflection of 27.46 mm, associated with mix combination A₄B₄C₁D₃.

The experimental data were incorporated into a streamlined calculation framework for extreme variance analysis. Here, Ki (i = 1–4) denotes the sum of test indices for each factor level, while ki (i = 1–4) represents the average value of Ki (i.e., ki = Ki /sample count per level). For example, in

Table 11. Bending performance variance analysis.

Performance Index	Factor	Squares of Deviations	Degree of Freedom	Mean Square	F Value	F Critical Value	Significance
Bending strength	Fly ash	0.962	3	0.321	10.534	F0.01(3,3) = 29.457 F0.05(3,3) = 9.277 F0.1(3,3) = 5.391	*
	Water-binder ratio	1.868	3	0.623	20.448		*
	Water-solid ratio	0.531	3	0.177	5.815		(*)
	Fiber volume parameters	3.009	3	1.003	32.940		**
	Empty column	0.091	3	0.030	—
Bending deflection	Fly ash	25.711	3	8.570	18.745		*
	Water-binder ratio	22.423	3	7.474	16.348		*
	Water-solid ratio	21.227	3	7.076	15.475		*
	Fiber volume parameters	46.900	3	15.633	34.192		**
	Empty column	1.372	3	0.457	—

Note: Ki (i = 1–4): Sum of test indices per factor level; ki (i = 1–4): Mean value of Ki. The symbol ‘**’ indicates that this factor has a highly significant impact on the performance index; The symbol ‘*’ indicates that this factor has a significant impact on the performance index; The symbol ‘(*)’ indicates that this factor has a certain impact on the performance index, but the impact is not significant.

, K1–K4 represent the sums of flexural strength/deflection values for each level of a given factor, and k1–k4 are the mean values of K1–K4. A simplified tabular format (retaining critical analytical components) was employed to compute the range variance (R-value). Additionally, parameter k was systematically evaluated against each factor, as illustrated in the line graph in Figure 9.

The results of the test were substituted into the ANOVA standard calculation form to obtain the F-value as shown in Table 11.

The exceptional flexural deflection of 27.46 mm (Specimen No.16) at 2.0% fiber volume originates from interfacial slip-hardening mechanisms. The hydrophobic nature of polyethylene fibers (density 0.97–0.98 g/cm³, Table 3) creates a low water-cement ratio interfacial transition zone around each fiber. This microscale densification increases frictional bond strength from approximately 0.8 MPa to 1.6 MPa, enabling greater energy absorption during bending deformation. Consequently, specimens with optimal fiber dosage exhibited 49% higher deflection than those with 1.2% fibers, demonstrating how fiber-matrix interfacial interactions govern macroscopic ductility.

By comprehensively analyzing the variation patterns of flexural properties with experimental factors and synthesizing results from range analysis and ANOVA, the following conclusions were derived:

(1) Peak flexural strength: 8.92 MPa (Specimen No. 5, A₂B₁C₂D₃), achieved with 50% fly ash substitution, a water-cement ratio of 0.20, a sand-cement ratio of 0.32, and a fiber volume fraction of 2.0%.

(2) Dominant controlling factors: Fiber dosage > water-cement ratio > fly ash content > sand-cement ratio, ranked by range values (R = 0.98).

(3) Deformation capacity: 27.46 mm (Specimen No. 16, A₄B₄C₁D₃), corresponding to 65% fly ash substitution, a water-cement ratio of 0.32, a sand-cement ratio of 0.28, and a fiber volume fraction of 2.0%.

4.2. Bond Stress–Slip Relationship and Failure Mechanism

(1) Failure Modes

Post-loading visual inspection of tension specimens revealed distinct failure modes. The ECC-reinforcement interface exhibited shear-dominated failure mechanisms, as illustrated in Figure 10a, with fiber-bridging effects effectively restraining crack propagation to localized splitting cracks shown in Figure 10b. In contrast, plain concrete specimens demonstrated brittle splitting failure patterns visible in Figure 10c,d, characterized by full-section penetrating cracks.

(2) Bond-Slip Behavior

To characterize the interfacial interaction between ECC and steel reinforcement, the bond stress–slip relationship was experimentally derived, as illustrated in Figure 11.

The curve exhibits a development trend analogous to conventional reinforced concrete systems, comprising three primary phases: ascending, descending, and residual. Based on bond stress evolution relative to slip magnitude, the behavior is further classified into five distinct stages:

Stage 1 (micro-slip): slip < 0.1 mm, bond force dominant;

Stage 2 (slip growth): slip 0.1~0.5 mm, friction and occlusion force rise;

Stage 3 (peak stress): slip 0.5~1.0 mm, bond stress up to 22.84 MPa (BER-1);

Stage 4 (descending segment): slip >1.0 mm, fiber bridging retards stress decay;

Stage 5 (residual section): slip >2.0 mm, residual stresses stabilized at 30%~50% of the peak value.

(3) Failure Mechanism

The bond stress mechanism between ribbed steel reinforcement and ECC comprises three components: chemical adhesion, friction, and mechanical interlocking. The mechanical interlocking from reinforcement ribs generates radial stresses, which are counteracted by ECC’s superior tensile strength (design value 2.81 MPa), effectively suppressing splitting cracks observed in conventional concrete. Critically, the high tensile strength and ductility of ECC, coupled with the extensive fiber bridging across potential slip planes (Figure 10b), significantly enhance the frictional resistance and mechanical interlocking at the interface. This fiber-induced enhancement in frictional behavior is a key factor contributing to the observed shear-dominated failure mode in ECC (Figure 10a) compared to the brittle splitting failure in ordinary concrete (Figure 10c,d). Furthermore, fiber reinforcement actively constrains interfacial debonding propagation, resulting in an 18.1% enhancement in bond strength compared to ordinary concrete systems.

4.3. Influence of Parameters on Bonding Properties

(1) Effect of Reinforcement Diameter on Bond Behavior

This study investigated ribbed steel bars with diameters of 10 mm, 12 mm, and 14 mm to evaluate the influence of reinforcement diameter on ECC-steel interfacial bonding. Rebar diameter variations introduce multiple parametric dependencies influencing dimensional stability. To ensure analytical reliability, ribbed rebar geometry parameters, including diameter (d), transverse rib height (h), upper rib width (a), lower rib width (b), and rib spacing (k), require systematic characterization, as detailed in Figure 12.

The bond mechanism primarily relies on mechanical interlocking forces generated by transverse ribs, with performance governed by rib geometry (h, k) and the relative diameter factor. Under fixed relative anchorage length (n) and cover thickness (m), the bond stress distribution area was derived geometrically. The effective load-transfer area corresponds to the inclined wedge surface, where longitudinal stress limits linearly correlate with the ECC tensile strength design value. This relationship defines the theoretical ultimate tensile force as a function of the bond stress area, enabling the derivation of bond strength equations. The coefficient α, linking longitudinal stress components to design value, requires experimental calibration. Standardized transverse rib parameters for 8–20 mm bars exhibit incremental rib height growth (0.8–1.7 mm), stable rib spacing expansion (5.5–10 mm), and nonlinear fluctuations in relative diameter coefficients with increasing diameter, establishing a basis for bond strength calculations.

According to the relationship between ultimate bond strength and diameter, the relative diameter coefficient is defined as ${\displaystyle \frac{\left(h+d\right)h}{kd}}$ and substituted into

Table 12. Effect of rebar diameter on bonding properties.

$\mathbf{Relative} \mathbf{Diameter} \mathbf{Coefficient}/\frac{\left(\mathit{h}+\mathit{d}\right)\mathit{h}}{\mathit{k}\mathit{d}}$	Ultimate Bond Strength/MPa
0.157 (φ10)	19.34
0.165 (φ12)	22.84
0.171 (φ14)	24.08

Note: Calibration data from additional tests: (0.160, 20.15 MPa), (0.168, 23.42 MPa), (0.173, 24.75 MPa), (0.176, 25.10 MPa).

.

The relationship between the relative diameter coefficient of the reinforcement bar and the ultimate bond strength was evaluated graphically, as illustrated in Figure 13. Experimental results demonstrate a positive correlation between the relative diameter coefficient and bond strength, with incremental increases in the relative diameter coefficient from 0.157 to 0.165 and 0.171 corresponding to bond strength enhancements of 18.1% and 24.5%, respectively. A least squares regression analysis was subsequently conducted to quantify this relationship, yielding the following predictive equation:

$${\tau }_u=\left[122.431{\displaystyle \frac{\left(h+d\right)h}{kd}}-12.277\right]f_{ty}$$

(4)

The coefficients α₁ = 122.431 and α₂ = −12.277 in Equation (4) were derived through least-squares regression analysis. Using the three experimental data points in Table 12 (relative diameter coefficients of 0.157, 0.165, and 0.171 corresponding to ultimate bond strengths of 19.34 MPa, 22.84 MPa, and 24.08 MPa, respectively), a linear model ${\displaystyle \frac{{\tau }_u}{f_{ty}}}$/ =α₁⋅X + α₂ (where X = ${\displaystyle \frac{(h+d)h}{kd}}$) was fitted, resulting in R² = 0.986. To calibrate the model, four additional independent tests (see Note a in Table 12) were conducted, yielding adjusted coefficients α₁ = 119.87 and α₂ = −11.95 with a maximum change rate of 2.1%, confirming model robustness.

(2) Effect of Anchorage Length on Bond Performance

Reinforcement anchorage length significantly influences bond performance. For analytical consistency, the relative anchorage length (l_a/d) is defined as the ratio of anchorage length to reinforcement diameter. This study investigated three distinct relative anchorage lengths: a short anchorage length (l_a/d = 3), a critical/transitional anchorage length (l_a/d = 5), and a long anchorage length (l_a/d = 7) to evaluate their impact on ECC-reinforcement bond behavior, with results tabulated in

Table 13. Effect of anchorage length on bonding properties.

$\mathbf{Relative} \mathbf{Anchorage} \mathbf{Length}/\left(\frac{{\mathit{l}}_{\mathit{a}}}{\mathit{d}}\right)$	Ultimate Bond Strength /MPa
3	25.70
5	22.84
7	19.13

. (Note: The terms “short”, “critical”, and “long” are defined based on the observed bond performance and stress distribution characteristics presented in this study).

Figure 14 illustrates the inverse correlation between relative anchorage length and ultimate bond strength. Increasing the relative anchorage length from 3 to 5 and 7 reduced ultimate bond strength by 11.1% and 25.6%, respectively. This phenomenon arises because shorter anchorage lengths lead to high-stress zones occupying a greater proportion of the bond interface, thereby elevating average bond stress. Conversely, extended anchorage lengths reduce the relative contribution of high-stress regions due to the diminished proportion of total stress concentration zones, ultimately decreasing average bond stress, as demonstrated in Figure 15.

A least squares regression analysis quantified the relationship between relative anchorage length and ultimate bond strength, yielding the following equation:

$${\tau }_u=(-0.584{\displaystyle \frac{l_a}{d}}+10.937)f_{ty}$$

(5)

(3) Influence of Protective Layer Thickness on Bond Performance

To investigate the influence of protective layer thickness on ECC-reinforcement bond behavior, pull-out specimens with side lengths of 100 mm, 120 mm, and 140 mm were tested, corresponding to net protective layer thicknesses of 44 mm, 54 mm, and 64 mm after accounting for reinforcement diameter. The relative protective layer thickness was adopted as the primary variable, as bond stress distribution area scales proportionally with reinforcement diameter. Test data for c/d and ultimate bond strength are compiled in

Table 14. Effect of protective layer thickness on bonding performance.

$\mathbf{Relative} \mathbf{Protective} \mathbf{Layer} \mathbf{Thickness}/\left(\frac{\mathit{c}}{\mathit{d}}\right)$	Ultimate Bond Strength/MPa
3.67	22.16
4.50	22.84
5.33	24.50

.

Figure 16 illustrates the positive correlation between the relative protective layer thickness and ultimate bond strength. Increasing the relative protective layer thickness from 3.67 to 4.50 and 5.33 enhanced ultimate bond strength by 3.1% and 10.6%, respectively. This trend is mechanistically explained by: Limited ECC splitting resistance reduces confinement effects, enabling premature splitting failure and rapid attainment of ultimate bond strength. Expanded ECC tensile zones delay splitting initiation, requiring higher loads to trigger failure, thereby increasing ultimate bond strength.

Using least squares regression on Table 14 data, the relationship between the relative protective layer thickness and ultimate bond strength was quantified as:

$${\tau }_u=(0.498{\displaystyle \frac{c}{d}}+5.991)f_{ty}$$

(6)

(4) The effect of fiber content on bonding properties

Comparisons with existing studies show: at 2.0% fiber dosage, our bond strength (22.84 MPa) is 25.4% higher than that of PVA fibers (18.21 MPa) reported by Li et al. [17], attributed to the lower friction coefficient of PE fibers (μ = 0.3 vs. 0.7). Furthermore, when fiber dosage increases from 2.0% to 2.4%, our strength increases by only 5.4% (to 24.08 MPa), lower than the 13% increase observed by Cui et al. [31] in similar matrices, mainly due to densification limits in our matrix.

(5) Statistical Regression Formula

A comprehensive regression model was developed to quantify the effects of ECC tensile strength, reinforcement diameter, anchorage length, and protective layer thickness on the ECC-reinforcement bond performance. Through appropriate simplification of Equations (4)–(6) and normalization of influence coefficients, the ultimate bond strength between ECC and steel reinforcement was formulated as:

$${\tau }_u=0.54\left[{\displaystyle \frac{\left(h+d\right)h}{kd}}-0.1\right](-{\displaystyle \frac{l_a}{d}}+18.73)({\displaystyle \frac{c}{d}}+12.011)f_{ty}$$

(7)

This statistical regression model, fitted to the test data using the least squares method, takes the form of a multiple linear regression. It quantitatively characterizes the coupled effects of steel bar geometric parameters (kd(h + d)/h), relative anchorage length (l_a/d), relative cover thickness (c/d), and ECC tensile strength (f_ty) on the bond strength (τ_u). The model validation indicators are as follows: the coefficient of determination R² = 0.92, indicating that the model explains 92% of the variability in the test data; the F-test result shows the model is overall significant (F = 38.6 > F_0.05(3,12) =3.49, p < 0.001). Each coefficient in the equation has a clear physical meaning: 0.54 is the comprehensive calibration coefficient; the term [kd(h + d)/h −0.1] quantifies the mechanical interlocking effect of the steel bar transverse ribs; the term (−l_a/d +18.73) reflects the law that anchorage efficiency decays with the increase of anchorage length (the average bond stress decreases when the anchorage length exceeds 5d), and its negative coefficient is consistent with the distribution characteristics of bond stress; the term (c/d +12.011) characterizes the strengthening effect of increased cover thickness on interface constraint, and its positive coefficient is consistent with the splitting failure delay mechanism. This model is applicable within the test parameter range of this study: steel bar diameter d = 10−14 mm, relative anchorage length l_a/d = 3−7, and relative cover thickness c/d = 3.7−5.3.

4.4. Study of Bonding Properties of ECC to Existing Concrete

4.4.1. Influence of Interface Roughness and Interfacial Agent

Following the aforementioned testing methodology and measurement protocol, splitting tensile tests were conducted on seven specimen groups, with recorded splitting loads and calculated corresponding splitting tensile strengths, as systematically compiled in

Table 15. Test results of bonding performance between ECC and existing concrete.

Specimen No.	Splitting Load/kN			Splitting Tensile Strength/MPa			Average Strength/MPa
	1	2	3	1	2	3
BEC-1	47.12	43.06	49.88	3.00	2.74	3.18	2.97
BEC-2	37.09	43.06	36.30	2.36	2.74	2.31	2.47
BEC-3	57.05	59.20	52.39	3.63	3.77	3.34	3.58
BEC-4	48.62	42.00	41.53	3.10	2.68	2.65	2.81
BEC-5	40.29	44.38	42.41	2.57	2.83	2.70	2.70
BCC-1	41.39	38.64	37.49	2.64	2.46	2.39	2.50
BEE-1	66.42	62.66	68.01	4.23	3.99	4.33	4.18

Note: Specimen BEE-1 represents the splitting tensile strength of monolithic Engineered Cementitious Composite (ECC). Specimen BCC-1 represents the splitting tensile strength of monolithic conventional concrete. Splitting tensile test: Displacement-controlled loading at 0.1 mm/min. Interface treatment: Chiseling depth 4–5 mm + water wetting (BEC-3 group shown). Values = mean ± standard deviation (n = 3).

.

As observed in Figure 17, the unchiseled interface exhibits splitting damage limited to the adhesion of the edge concrete compression zone to ECC, with the interface morphology retaining its original state. When the chiseling depth is increased to 2–3 mm, damage behavior transitions to the concrete substrate, resulting in extensive concrete fragments adhered to the ECC. Further increasing the chiseling depth to 4–5 mm induces a composite failure pattern at the interface corners, combining ECC and concrete damage, while the central region remains dominated by concrete fragment bonding with ECC.

As indicated in

Table 16. Material performance test results.

Material	Compressive Strength/MPa	Average Strength/MPa
Concrete	45.43	43.97
	42.16
	44.37
ECC	61.11	61.20
	63.65
	58.85

and

Table 17. Effect of interface roughness on splitting tensile strength.

Interface Roughness	Splitting Tensile Strength/MPa
I (No treatment available)	2.47
II (Chipping 2~3 mm)	2.97
III (Chipping 4~5 mm)	3.58

, the splitting tensile strength increases proportionally with relative chiseling depth. Compared to Roughness I, Roughness II, and Roughness III specimens exhibit strength enhancements of 20.2% and 44.9%, respectively. This improvement is attributed to the superior tensile strength of ECC relative to ordinary concrete. As chiseling depth increases, a greater volume of ECC infiltrates interfacial voids and participates in load-bearing mechanisms, thereby elevating the splitting tensile strength.

As shown in Figure 18, specimens without interfacial agent treatment exhibited damage patterns similar to those with water-moistened treatment but demonstrated less pronounced overall degradation. Post-splitting failure, fewer and thinner concrete fragments remained partially bonded to the ECC substrate. In contrast, specimens treated with a cement-based netting agent showed a significant reduction in concrete fragments at the fracture interface, with a residual film layer adhering to the ECC surface.

As demonstrated in

Table 18. Effect of interfacial agent type on split tensile strength.

Type of Interface Agents	Splitting Tensile Strength/MPa
No treatment available	2.81
Water	2.97
Cement paste	2.70

and Figure 19, water-moistened specimens exhibit a 5.7% strength enhancement compared to untreated specimens, whereas cement mortar-treated specimens show a 3.9% strength reduction. Mechanistic analysis reveals that water moistening preserves the water-cement ratio equilibrium in the interfacial transition zone by mitigating substrate water absorption, thereby preventing hydration-related defects. Conversely, cement mortar infill within chiseled grooves reduces the effective load-bearing area of ECC, resulting in a combined reduction of material strength and interfacial geometric efficacy.

4.4.2. Comparative Analysis with Monolithic Cast Specimens

As illustrated in Figure 20, conventional concrete exhibits a relatively planar fracture surface following splitting failure, whereas Engineered Cementitious Composite (ECC) retains superior structural integrity due to its intrinsic fiber-bridging mechanism. The composite specimen demonstrates an intermediate failure mode, with damage morphology transitional between monolithic concrete and ECC behaviors.

As illustrated in

Table 19. Influence of casting type on split tensile strength.

Pouring Type	Splitting Tensile Strength/MPa
Integral concrete pouring	2.50
Composite pouring	2.97
ECC monolithic pouring	4.18

and Figure 21, the splitting tensile strength follows a descending order: ECC-monolithic specimens > composite-cast specimens > concrete-monolithic specimens. Compared to the concrete-monolithic specimens, the composite and ECC-monolithic specimens exhibit 18.8% and 67.2% higher splitting tensile strengths, respectively. Notably, the interfacial tensile strength exceeds the inherent tensile capacity of the concrete matrix itself, demonstrating effective interfacial compatibility between ECC and conventional concrete.

5. Bridge Deck Continuous Test Validation

Building upon the theoretical framework of bonded interface strength and damage mechanisms established earlier, this study evaluates the practical performance of Engineered Cementitious Composite (ECC) materials in bridge deck continuity systems through experimental testing. As illustrated in Figure 22 and Figure 23, three full-scale specimens were fabricated for comparative analysis: LS-1 (traditional link slab concrete), LS-2 (double-arch steel plate-concrete), and LS-3 (double-arch steel plate-ECC). Displacement-controlled loading was applied to simulate real-world bridge deformation patterns, ensuring alignment with operational conditions.

A comparative analysis of test results is presented in

Table 20. Comparison of key test results.

Performance Index	LS-1	LS-2	LS-3 (ECC)	Regulatory Limits
Cracking load (°)	0.183	0.194	0.267	-
Design load crack width (mm)	0.28	0.23	0.07	≤0.2
1.5 times load crack depth (mm)	110	100	70	≤120
Maximum deflection (mm)	10.217	7.406	5.345	L/600 = 8.3
Maximum strain of reinforcement (με)	2000	1200	950	2300 (Yield)

Note: L = 5000 mm is the test simulation span.

:

(1) Crack Control: LS-3 exhibited characteristic ECC multi-crack propagation, with crack density 3–5 times higher than LS-1. However, maximum crack width under 1.5 design load remained at 0.12 mm (60% of code limits), while conventional concrete specimens developed through-thickness cracks. ECC crack depths consistently stayed below 50% of the section thickness.

(2) Structural Compatibility: LS-3 demonstrated a 42% reduced interfacial slip compared to LS-2, with steel plate strains constrained below 150 με, validating the interfacial bond parameters established in prior analyses. Enhanced ECC-steel deformation synergy increased structural stiffness by 28%.

(3) Ductility: Without rebar yielding, LS-3 achieved an ultimate rotation angle of 1.376° (Figure 24), 49.6% higher than LS-1. Its load–deflection curve displayed a distinct plastic plateau (Figure 25), accompanied by a 3.2% improvement in energy dissipation capacity. These outcomes confirm the feasibility of the proposed ECC mix and interfacial treatment for steel-ECC-concrete composite bridge deck continuity systems.

6. Conclusions

(1) High-Performance ECC Mix Ratio

This study proposes an optimized mix ratio for Engineered Cementitious Composites (ECC) tailored to seamless bridge deck applications, with the following parameters: 55% fly ash content, 0.24 water-to-cement ratio, 0.36 sand-to-cement ratio, and 2.0% polyethylene fiber content. Experimental verification confirms that the mechanical properties of this ECC formulation demonstrate exceptional performance, achieving a compressive strength of 57.58 MPa, a tensile strength of 4.64 MPa, and a flexural strength of 7.93 MPa, thereby meeting the C40 concrete strength requirements for bridge engineering. Critically, the material’s inherent cohesive strength, as measured by the splitting tensile strength of monolithic ECC specimens (BEE-1, Table 15), reached 4.18 MPa. This value significantly exceeds the splitting strength of conventional concrete monolithic specimens (BCC-1, 2.50 MPa, Table 15). Meanwhile, the material exhibits superior ductility, with an ultimate tensile strain of 4.44% and a mid-span deflection of 25.58 mm, effectively accommodating deformation demands in bridge structures. Orthogonal test results reveal that among the factors affecting the mechanical properties of ECC materials, the water-cement ratio is the primary factor governing compressive strength, while fiber content predominantly controls tensile and flexural performance.

(2) Bonding mechanism of ECC–structure interface

Regarding reinforcement bonding performance, an anchorage length of 5d yielded an ultimate bond strength of 22.84 MPa, representing an 18.1% enhancement over conventional concrete, with shear-dominated failure modes. For the bonding interface between ECC and existing conventional concrete, the optimal treatment (chiseling depth of 4–5 mm with water moistening) achieved a cohesive/bond strength (measured by splitting tensile test) of 3.58 MPa. This represents a significant 43.2% increase over the bond strength (2.50 MPa) observed in conventional concrete monolithic specimens (BCC-1) and demonstrates the effectiveness of the proposed interface treatment. Furthermore, a clear water wetted interfacial agent is recommended for interface treatment to significantly enhance interfacial bonding efficacy. The superior bonding performance, particularly the prevalence of shear failure modes at ECC-rebar interfaces and the attainment of bond strengths exceeding the tensile capacity of conventional concrete at ECC-concrete interfaces (as evidenced by concrete substrate failure in deep-chiseled specimens, Figure 17e,f), highlights the enhanced frictional and interlocking characteristics imparted by ECC. This enhancement is primarily attributed to the material’s intrinsic high tensile strength and, most importantly, the potent fiber bridging mechanism that effectively mobilizes friction and resists interfacial slip.

(3) Validation Findings

Full-scale tests demonstrate that the integration of a double-arch structure with ECC enables distributed cracking at the 0.1 mm scale in crack control, ensuring compliance with waterproofing and durability specifications. In terms of deformation coordination, the maximum deflection of 5.345 mm is below the L/600 limit (L = 5000 mm), ensuring traffic smoothness. The significantly reduced interface slip (42% reduction compared to LS-2) and low reinforcement strains (950 µε) further confirm synergistic reliability. Critically, this successful performance under the complex combined stresses (primarily tension-flexure-shear induced by negative bending moments at the link slab location) at 1.5 times the design load provides strong indirect validation of the ECC material’s capability to perform effectively in the multi-axial stress states encountered in bridge deck continuity systems. The inherent high ductility, multi-crack formation, and fiber bridging mechanisms underpin this resilience. While this study focused on fundamental material properties and key interface behaviors under dominant loading modes (tension for rebar pull-out, shear-dominant for concrete splitting), the demonstrated performance suggests promising behavior under combined loads (e.g., tension-shear, compression-shear). Detailed investigation of ECC and its interfaces under explicit combined loading scenarios represents an important direction for future research.

(4) Environmental Degradation Resistance of Developed ECC Beyond Static Mechanical Performance

Although this study mainly centered on static mechanical performance, the developed ECC shows inherent resistance to environmental degradation, which is demonstrated by the following aspects. It has freeze-thaw resistance, and autonomous crack-healing occurs through secondary hydration under wet–dry cycles, reducing chloride ion penetration by more than 60% compared with conventional concrete [23]. In terms of corrosion inhibition, the crack width is limited to ≤0.12 mm under service loads, which suppresses reinforcement corrosion rates by 3–5 times [1,2]. For seismic resilience, the multi-crack dispersion mechanism improves energy dissipation, and previous studies have confirmed that the crack widths under cyclic loading are 80% smaller [24]. Quantitative validation under dynamic and extreme thermal conditions will be studied in subsequent research.