Mechanical Performance, Durability, and Microstructure of Dune-Sand Concrete Incorporating Rice Husk Ash and Dried Oil Sludge

Abstract

This study evaluates sustainable dune-sand concrete produced using local dune sand, rice husk ash (RHA), and dried oil sludge. The experimental program included a control mixture (C0) and three modified mixtures (M1–M3). RHA was used as a partial replacement for Portland cement at contents of 20, 40, and 60 kg/m³, while dried oil sludge was introduced as a waste-derived additive at 10, 15, and 20 kg/m³, respectively. The total binder content, defined as Portland cement plus RHA, was maintained at 350 kg/m³, while the mixing-water content and water-to-binder ratio were kept constant at 185 kg/m³ and 0.53, respectively. The RHA was heat-treated at 800 °C for 60 min, immediately quenched in water, and subsequently dried at 105 ± 5 °C until a constant mass was achieved. Among the investigated compositions, M2, containing 310 kg/m³ Portland cement, 690 kg/m³ dune sand, 1110 kg/m³ crushed stone, 40 kg/m³ RHA, and 15 kg/m³ dried oil sludge, exhibited the most balanced overall performance. Its compressive strength reached 27.8 MPa at 7 days, 42.5 MPa at 28 days, and 47.7 MPa at 90 days, while its static modulus of elasticity was 32.7 GPa. M2 also showed the lowest water absorption (3.63%), the highest watertightness grade within the investigated series (W10), and the greatest measured freeze–thaw resistance, completing 223 cycles. Optical microscopy and X-ray diffraction observations indicated a visually more continuous matrix and lower relative intensities of several portlandite peaks in the modified concrete. For the replicate-based properties, mixture composition had a statistically significant effect (p < 0.001). Because RHA and dried oil sludge were varied simultaneously, the results represent their combined effect. M2 is therefore identified as the most balanced composition within the present limited test series, rather than as a universal optimum.

1. Introduction

Producing durable concrete in arid regions is often complicated by the quality and availability of conventional aggregates. Local dune sand is abundant, but its fine grading and rounded grains make it difficult to use as the sole fine aggregate. This creates a practical need for mixture designs that combine local raw materials with mineral additions capable of improving packing and reducing permeability.

Rice husk ash is one such addition. When the burning regime is controlled, the ash contains a substantial proportion of reactive silica; its behavior is then governed by amorphous silica content, residual carbon, fineness, specific surface area, and thermal history [1,2,3,4,5,6]. In cement paste, finely divided RHA can act as a microfiller and nucleation surface, while its amorphous SiO₂ reacts with Ca(OH)₂ to form additional calcium-silicate-hydrate products.

The beneficial range of RHA is limited. Moderate dosages have improved strength and reduced permeability in many studies, whereas higher replacement levels often increase water demand and dilute the clinker fraction [3,7,8,9]. RHA has also been investigated in self-consolidating concrete, including systems containing fly ash, where fresh-state behavior and workability were strongly dependent on its dosage [10,11]. Properly proportioned RHA has been associated with improved durability, reduced permeability, and enhanced resistance to aggressive exposure, including chloride ingress and freeze–thaw cycling [12,13,14,15]. Thermally treated RHA has additionally been used as a silica-rich component in reactive powder concrete [16].

Recent experimental work and database-based analyses make clear that RHA performance is highly dependent on how the ash is produced, the level at which it is introduced, the cement used, and the property selected for optimization. Alsaed and Al Mufti found that a controlled RHA replacement improved compressive strength and reduced air permeability, although workability declined as the ash content rose [17]. Öztürk et al., drawing on more than 1000 experimental records, likewise associated suitable RHA dosages with better durability, resource efficiency, and resistance to transport-related deterioration [18]. Albadrani reported that the preferred dosage may be relatively low in Portland pozzolana cement concrete because excessive replacement raises water demand and reduces the available clinker fraction [19]. Other authors have combined RHA with fly ash and supplementary mineral components to improve particle packing, pore structure, and long-term performance [10,13,14,16,18,20]. Hong et al. also observed improved freeze–thaw behavior in concrete containing both RHA and fly ash, but the most favorable proportions varied depending on whether strength, thermal response, or frost resistance was prioritized [20]. These results show why an optimum value reported for one binder system should not be transferred directly to another.

Dune sand adds another variable to this balance. Its fine particle size can improve filling at low contents, yet an excess may disrupt the aggregate skeleton and raise the paste demand. For this reason, dune-sand concrete generally requires a complementary coarse aggregate and careful adjustment of the binder and water contents.

Oil sludge presents a different challenge. It is a petroleum-sector residue that must be treated before use or disposal. Small quantities of dried or thermally processed material have been investigated in cement-based systems, but any proposed application must address residual hydrocarbons, chemical stability, leaching, and long-term behavior.

Studies on petroleum-derived sludge in cementitious materials generally follow three routes: direct addition after pretreatment, conversion of the residue to oily sludge ash for partial cement replacement, and stabilization or solidification of potentially hazardous constituents in a cement matrix. Direct incorporation is normally limited to small dosages because residual hydrocarbons and very fine particles can interfere with hydration, entrain air, and reduce strength when the addition is excessive. Even so, one experimental program produced concrete containing 3 mass% oil sludge and then applied thermal treatment, demonstrating that a limited amount of the waste can be incorporated technically into concrete products [21].

A second and more widely reported route is to thermally convert the material into oily sludge ash (OSA). Kankia et al. optimized cement mortars containing OSA and showed that the treated residue can replace part of the cement when composition, fineness, and dosage are properly controlled [22]. Thermal treatment removes much of the organic fraction and leaves a predominantly mineral residue, but this does not make the material uniform: its behavior still depends on calcination conditions, chemical composition, particle size, and replacement level. Kourdache et al. also considered oil sludge as an alternative fuel in cement production and discussed the resulting combustion ash as a secondary constituent in pozzolanic cement composites [23].

A related approach is co-processing at the clinker stage. Huang et al. used oil sludge as both fuel and raw material in Portland cement manufacture [24]. At kiln temperature, most organic constituents are destroyed and the mineral residue enters the clinker system, although emissions and volatile elements must be monitored. This route is fundamentally different from adding treated sludge directly to fresh concrete.

The same idea has recently been extended to alkali-activated binders. García-Díaz et al. introduced 5–20 wt.% ash from oil-refining sludge into binary matrices based on either rice husk ash or chamotte [25]. Their work is relevant here because it confirms that a petroleum-sludge-derived mineral residue can be used in the same binder system as a silica-rich secondary material. At the same time, immobilization studies consistently show that compressive strength alone is not sufficient to establish environmental acceptability. Leaching of heavy metals and other potentially hazardous constituents must be assessed separately before the material can be regarded as safe [26].

Earlier work by Bisenov, Uderbayev, and Saktaganova is particularly relevant to the possible water-repellent role of oil sludge. In non-autoclaved aerated concrete prepared from finely ground multicomponent dry mixtures, oil sludge used together with surfactants contributed to hydrophobic behavior and was incorporated through an optimized technological route [27]. A related physicochemical study of foamed concrete reported changes in phase formation, cementing bonds, and macro- and microstructural development [28]. These studies provide a basis for considering dried oil sludge not only as a waste-derived filler but also as a component that may influence water transport. However, both investigations concerned lightweight cellular concretes. The combined behavior of dried oil sludge, RHA, and dune sand in dense structural concrete has received much less attention and cannot be assumed to be identical.

Taken together, the published evidence indicates that petroleum sludge can be used in cement and concrete only after appropriate pretreatment and within a controlled dosage range. Small additions may support waste utilization and, depending on their composition, may alter particle packing or water transport; excessive or insufficiently treated sludge can hinder hydration and reduce performance. A clear gap therefore remains concerning the simultaneous use of dried oil sludge and RHA in dune-sand structural concrete. In particular, the links among sludge dosage, RHA pozzolanic activity, strength development, durability, and contaminant immobilization have not yet been established in a systematic way.

Accordingly, this study compares a control concrete with three compositions containing different amounts of RHA and dried oil sludge. The work focuses on strength development, flexural response, static modulus, water absorption, watertightness, freeze–thaw resistance, and qualitative microstructural and phase observations. The purpose is to identify a rational composition for further development rather than to claim a universal optimum.

2. Materials and Methods

2.1. Raw Materials

Portland cement M400 was used as the primary binder. Dune sand of aeolian origin from the Belkul area of the Kyzylorda region, Kazakhstan, was used as the fine aggregate. The particle-size distribution was determined by sieve analysis. The predominant fraction was 0.16–0.315 mm, accounting for 65–80 wt.% of the sand. The content of particles larger than 0.63 mm did not exceed 0.5 wt.%, while particles smaller than 0.05 mm accounted for 2–5 wt.%. The fineness modulus ranged from 0.7 to 1.3, indicating that the material was a very fine sand. The detailed particle-size distribution is presented in

Table 1. Particle-size distribution of the dune sand.

Particle-Size Fraction, mm	Content, wt.%	General Characterization
>0.63	0–0.5	Practically absent
0.315–0.63	5–15	Coarser fraction
0.16–0.315	65–80	Predominant fraction
0.05–0.16	5–15	Fine and silty fraction
<0.05	2–5	Clay- and silt-sized particles

Note: The values are reported as ranges obtained for different samples of dune sand; therefore, the lower and upper limits of the individual fractions are not intended to be summed as a single sieve-analysis result.

. Crushed stone was used as the coarse aggregate. Rice husk ash was incorporated as a silica-rich mineral component. Oil sludge was dried before use. The mixing-water content was maintained constant at 185 kg/m³ for all concrete mixtures.

The dune sand was polymictic and consisted predominantly of quartz (51 wt.%) and feldspar minerals (28 wt.%). Calcite accounted for 10 wt.%, while clay minerals containing finely dispersed calcite represented 8 wt.%. Minor mineral components, including biotite, augite, and hornblende, collectively accounted for 3 wt.%.

The maximum particle size of the dune sand was 0.63 mm. Its water absorption ranged from 0.5 to 1.0 wt.%, and its specific gravity ranged from 2.60 to 2.65. Crushed dolomite aggregate from the Zhanakorgan deposit, Kyzylorda region, Kazakhstan, with a nominal particle-size fraction of 5–20 mm and a maximum particle size of 20 mm was used as the coarse aggregate. Its water absorption was 1.2 wt.%, and its specific gravity was 2.82.

2.2. Pretreatment of Rice Husk Ash and Oil Sludge

The oil sludge was dried in a drying chamber at 105 ± 5 °C until a constant mass was achieved before its incorporation into the concrete mixture.

The residual moisture content and total petroleum hydrocarbon content of the dried oil sludge were not independently quantified for the specific batch used in this study. Therefore, the term “dried oil sludge” refers to the material after thermal moisture removal and does not imply complete removal of petroleum hydrocarbons.

The heavy-metal and toxic-element contents of the dried oil sludge were not independently determined for the specific batch used in this study. Accordingly, the present investigation does not establish the environmental safety or complete immobilization of potentially hazardous constituents. Detailed elemental characterization and leaching tests are required before practical application.

The rice husk ash was heated at 800 °C for 60 min, immediately quenched in water, and subsequently dried at 105 ± 5 °C until a constant mass was achieved.

The phase composition and pozzolanic reactivity of RHA depend on the combined effects of treatment temperature, holding time, and cooling regime. Although high amorphous-silica contents are commonly associated with controlled treatment at approximately 500–700 °C, partial crystallization may begin at around 800 °C, particularly under prolonged thermal exposure or slow cooling [1,29]. Therefore, 800 °C is not presented here as a universally optimal treatment temperature. In the present procedure, the RHA was maintained at 800 °C for a limited period of 60 min and was then immediately quenched in water. The purpose of quenching was to terminate the high-temperature treatment rapidly and reduce the time available for further structural ordering and crystallization of silica during cooling. Rapid cooling and water subcooling have been reported to suppress crystallization and preserve a greater proportion of disordered silica in RHA treated at elevated temperatures [29]. After quenching, the RHA was dried at 105 ± 5 °C until constant mass to remove physically retained water and ensure reproducible dry-mass dosing without altering the prescribed mixing-water content and water-to-binder ratio. This drying stage was not regarded as a separate activation treatment. Because the amorphous-silica fraction and direct pozzolanic activity of the experimental batch were not quantitatively determined, the adopted procedure is presented as a processing rationale intended to limit additional crystallization rather than as evidence of enhanced RHA reactivity.

The present study forms a continuation of previous experimental work using RHA obtained from rice husk of the “Leader” variety. The available chemical-composition data for this source material are presented in

Table 2. Indicative oxide composition of the rice husk ash source material.

Component	Content (%)
SiO2	88.5
Al2O3	0.8
Fe2O3	0.5
CaO	1.2
MgO	0.7
K2O	3.2
Na2O	0.3
SO3	0.2
Loss on ignition (LOI)	4.6

Note: The values represent available characterization data for RHA obtained from rice husk of the “Leader” variety and are reported as indicative characteristics of the source material. The oxide composition was not independently redetermined for the specific experimental batch used in the present study. The reported SiO₂ value represents total silica and should not be interpreted as the quantitatively determined amorphous-silica content.

. The RHA contained 88.5 wt.% total SiO₂ and had a loss on ignition of 4.6 wt.%. However, these values were not independently redetermined for the specific batch used in the present experiments and are therefore reported as indicative source-material characteristics. The amorphous fraction of SiO₂ was not quantified separately.

The specific surface area of the obtained rice husk ash (RHA) samples was determined by the BET method using low-temperature nitrogen adsorption on an Autosorb iQ automated gas sorption analyzer (Anton Paar QuantaTec Inc., formerly Quantachrome Instruments, Boynton Beach, FL, USA). The particle-size distribution and median particle diameter, d₅₀, were determined by laser diffraction using a Mastersizer 3000 particle-size analyzer equipped with a Hydro MV wet-dispersion unit (Malvern Panalytical Ltd., Malvern, Worcestershire, UK), with the samples dispersed in water. The BET specific surface area of the RHA was 42 m²/g, while the measured d₅₀ values ranged from 15 to 30 μm. The particle density of the RHA was not determined for the experimental batch.

2.3. Mixture Proportions

The proportions of the three modified concrete mixtures (M1–M3) are presented in

Table 3. Composition of the investigated concrete mixtures (kg/m³).

Mixture	Portland Cement, kg/m3	Dune Sand, kg/m3	Crushed Stone, kg/m3	RHA, kg/m3	Dried Oil Sludge, kg/m3	Water, kg/m3	w/b
C0	350	675	1140	0	0	185	0.53
M1	330	680	1125	20	10	185	0.53
M2	310	690	1110	40	15	185	0.53
M3	290	700	1095	60	20	185	0.53

Note: All component contents are expressed in kg/m³. The water-to-binder ratio (w/b) was calculated as the mass of mixing water divided by the combined mass of Portland cement and RHA. The total binder content and mixing-water content were maintained constant at 350 and 185 kg/m³, respectively.

.

Because the RHA and dried oil sludge contents were varied simultaneously, the present mixture design permits evaluation of their combined effect but does not allow their independent contributions to be quantified.

2.4. Concrete Mixing Procedure

The concrete mixture was prepared using the following sequence:

Portland cement, dune sand, crushed stone, and the heat-treated RHA were introduced into a forced-action mixer.

Approximately 50% of the total mixing water was added, and the components were mixed thoroughly.

The remaining water was premixed with the dried oil sludge and introduced into the mixer.

Mixing was continued until a homogeneous concrete mixture was obtained.

The complete sequence of raw-material preparation, pretreatment, mixing, and concrete production is illustrated in Figure 1.

2.5. Experimental Program

The experimental program included fresh-state, physical, mechanical, durability, and microstructural characterization. Four principal mixtures were considered: a control concrete without rice husk ash and dried oil sludge (C0) and three modified mixtures (M1–M3). Unless otherwise stated, three specimens were tested for each composition and test condition, and replicate-based results are reported as the arithmetic mean ± standard deviation.

2.5.1. Workability of Fresh Concrete

The workability of the fresh concrete mixtures was determined by the slump test in accordance with GOST 10181-2014 [30]. Two determinations were performed for each mixture, and the mean slump value was used to assign the workability class according to GOST 7473-2010 [31]. The air content and workability retention over time were not measured; therefore, the possible effects of RHA and dried oil sludge on entrapped or entrained air and time-dependent slump loss could not be quantified.

2.5.2. Average Density

The average density of hardened concrete was determined in accordance with GOST 12730.1-2020 [32]. Three specimens were tested for each mixture. Before testing, the geometrical dimensions of each specimen were measured, and its volume was calculated. The specimens were then weighed, and the average density was calculated as the ratio of specimen mass to its geometrical volume. The results were expressed in kg/m³.

The mean value and standard deviation were calculated from the results obtained for three specimens of each mixture.

2.5.3. Compressive Strength Development

The compressive strength of concrete was determined in accordance with GOST 10180-2012 [33]. Cubic specimens with dimensions of 100 × 100 × 100 mm were prepared for each concrete mixture. After casting, the specimens were kept in the moulds for 24 h under laboratory conditions. They were then demoulded and cured under normal moist conditions at a temperature of 20 ± 2 °C and a relative humidity of not less than 95% until the testing ages of 3, 7, 28, and 90 days.

Three specimens were tested for each mixture at every curing age. Before testing, the bearing surfaces of the specimens were inspected and cleaned. The compressive strength tests were carried out using an IP-1A-1500 hydraulic compression testing machine with a maximum load capacity of 1500 kN.

Each specimen was centred on the lower platen of the testing machine and subjected to a continuously increasing compressive load until failure. The specimens were loaded continuously to failure at a constant stress-increase rate of 0.6 ± 0.2 MPa/s in accordance with GOST 10180-2012, corresponding to an applied load rate of approximately 4–8 kN/s for the 100 × 100 mm loaded cross-sectional area. The compressive strength was calculated as the ratio of the maximum failure load to the loaded cross-sectional area of the specimen. The results were expressed as the arithmetic mean of three measurements together with the corresponding standard deviation.

2.5.4. Flexural Strength

Flexural tensile strength was determined at 28 days in accordance with GOST 10180-2012 [33] using three prism specimens measuring 100 × 100 × 400 mm for each concrete composition. After demoulding at 24 h, the specimens were cured under the same normal moist conditions described in Section 2.5.3 until testing.

The tests were conducted using a PGM-500MG4A hydraulic testing machine (SKB Stroypribor LLC, Chelyabinsk, Russian Federation) equipped with an IBB-04M device for testing concrete prisms under flexural loading. A four-point loading scheme with a support span of 300 mm was used, and the total load was transmitted through two loading rollers positioned at the third points of the span (Figure 2). Each prism was centred on the supports and loaded continuously to failure at a stress-increase rate of 0.05 ± 0.01 MPa/s.

Flexural tensile strength was calculated using the following Equation (1):

Rtb = δFl/(ab²)

(1)

where F is the total failure load (N), l is the support span (300 mm), a and b are the measured width and height of the prism (mm), and δ is the scale factor. For 100 × 100 mm heavy-concrete prisms, δ = 0.92. A result was excluded when failure occurred outside the middle third of the span or when the fracture plane was inclined by more than 15° from the vertical.

For each mixture, the flexural tensile strength was reported as the arithmetic mean of three replicate measurements together with the corresponding sample standard deviation.

2.5.5. Water Absorption

Water absorption by mass was determined at 28 days in accordance with GOST 12730.3-2020 [34] using three 100 × 100 × 100 mm cube specimens for each mixture. The specimens were dried to constant mass at 105 ± 5 °C and subsequently immersed in water at 20 ± 2 °C until the difference between two consecutive mass measurements did not exceed 0.1%. Water absorption by mass, Wm, was calculated using Equation (2):

W_m = (m_s − m_d)/m_d × 100%

(2)

where m_s is the saturated surface-dry mass of the specimen (g), and m_d is the oven-dried mass of the specimen (g). The result for each mixture was reported as the arithmetic mean of three measurements together with the corresponding standard deviation.

The successive stages of specimen drying, dry weighing, water saturation, repeated weighing, and calculation are summarized in Figure 3.

2.5.6. Freeze–Thaw Resistance Test Method

Freeze–thaw resistance was evaluated at the design age according to the first basic method of GOST 10060-2012 [35], which is applicable to general-purpose concrete tested in a water-saturated condition. Separate specimens that had not been oven-dried during the water-absorption test were used, because drying concrete at approximately 100–110 °C may induce shrinkage-related microcracking and distort the subsequent frost-resistance assessment [36].

Six control specimens and twelve test cube specimens were used in the experiments. Cube specimens measuring 100 × 100 × 100 mm were used. Before testing, both the control and main specimens were saturated in water at 20 ± 2 °C in stages: the specimens were immersed to one-third of their height for 24 h, to two-thirds of their height for a further 24 h, and then fully immersed for 48 h. During full immersion, the water level was maintained at least 20 mm above the upper surfaces of the specimens.

Freezing was performed in an MK-110 programmable laboratory chamber (Russian Federation; supplied by Kazakhstan-Tasymal LLP, Ust-Kamenogorsk, Kazakhstan) at −18 ± 2 °C. Compressive strength before and after cycling was determined using the IP-1A-1500 hydraulic compression testing machine described in Section 2.5.3.

After saturation, the control specimens were wiped with a damp cloth and tested in compression to determine the reference strength. The main specimens were placed in the freezing chamber with a minimum clearance of 20 mm between adjacent specimens and chamber surfaces. For 100 mm cubes, each cycle consisted of freezing in air at −18 ± 2 °C for at least 2.5 h, followed by thawing in water at 20 ± 2 °C for 2.0 ± 0.5 h. Testing was conducted continuously at a frequency of at least one cycle per day.

At the specified inspection intervals, the specimens were examined for cracking, spalling, and edge scaling, weighed, and tested for residual compressive strength. The mass loss after N cycles, M_LN, and the strength-retention coefficient, K_f,N, were calculated as Equation (3):

M_LN = (m₀ − m_N)/m₀ × 100%
K_f,N = f_c,N/f_c,0

(3)

where M_LN is the mass loss after N freeze–thaw cycles (%); m0 is the specimen mass before cycling; m_N is the specimen mass after N cycles; K_f,N is the strength-retention coefficient after N cycles; f_c,0 is the reference compressive strength of the control specimens; and f_c,N is the residual compressive strength after N cycles. Formal conformity with GOST 10060-2012 was assessed using the standard statistical strength criterion, together with a mass loss not exceeding 2% and the absence of visible damage.

2.5.7. Watertightness (Water Penetration Resistance)

The watertightness of concrete was determined at 28 days by the wet-spot method in accordance with GOST 12730.5-2018 [37]. The tests were performed using a UVB-MG4.01 automatic concrete watertightness apparatus (SKB Stroypribor LLC, Chelyabinsk, Russia).

For each concrete mixture, a series of six cylindrical specimens measuring 150 mm in diameter and 150 mm in height was prepared. After demoulding, the specimens were cured under normal moist conditions at 20 ± 2 °C and a relative humidity of at least 95% until the test age. Before testing, the specimens were kept under laboratory conditions for 24 h. Their lateral surfaces were sealed with paraffin in steel sleeves so that water could enter only through the pressure-exposed end face.

Water pressure was initially set to 0.2 MPa and subsequently increased in increments of 0.2 MPa. Because the specimen height was 150 mm, each pressure stage was maintained for 16 h. The apparatus automatically controlled and maintained the applied pressure and registered the appearance of water on the opposite surface using moisture sensors. The test was terminated when seepage was detected in at least three of the six specimens, when the final pressure stage was completed, or when the test was stopped by the operator.

The watertightness of an individual specimen was taken as the maximum water pressure at which no seepage was observed. The watertightness of a six-specimen series was defined as the maximum pressure at which at least four specimens remained free from seepage. The corresponding watertightness grade was assigned according to GOST 12730.5-2018, with pressures of 0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 MPa corresponding to grades W2, W4, W6, W8, W10, and W12, respectively.

2.5.8. Static Modulus of Elasticity

The static modulus of elasticity was determined at 28 days in accordance with GOST 24452-2023 [38]. Three prism specimens measuring 100 × 100 × 400 mm were prepared for each concrete mixture. After demoulding, the prisms were cured under the same normal moist conditions used for the compressive-strength specimens until the test age.

Axial compression was applied using an IP-1A-1500 hydraulic testing machine with a maximum load capacity of 1500 kN. Longitudinal deformation was measured on four faces of each prism using dial indicators with a resolution of 0.001 mm and a gauge length of 200 mm. After applying a seating load not exceeding 2% of the expected failure load, the load was increased in stages of approximately 10% of the expected failure load at a stress-increase rate of 0.6 ± 0.2 MPa/s. Each stage was maintained for 4–5 min, and the readings from the four indicators were averaged. The static modulus was determined within the approximately linear portion of the stress–strain response, up to 30% of the expected failure load, using Equation (4):

E = Δσ/Δε

(4)

where E is the static modulus of elasticity (MPa), Δσ is the increase in axial compressive stress between the selected load levels (MPa), and Δε is the corresponding increase in mean longitudinal strain. The result for each mixture was expressed in GPa as the arithmetic mean of the three tested prisms. Only the mixture-level mean values were retained in the available laboratory records. The individual specimen-level modulus values and complete load–deformation records were unavailable; consequently, standard deviations could not be calculated retrospectively.

2.5.9. Optical Microscopy and XRD Analysis

Microstructural and phase examinations were performed after 28 days of curing for representative specimens of the control mixture C0 and the modified mixtures M1 and M2. Samples were taken from the central zones of the hardened specimens, while surface layers and visibly damaged or contaminated areas were excluded.

For optical microscopy, representative sections were dried at a temperature not exceeding 40 °C, ground, and polished to obtain a sufficiently flat observation surface. The specimens were examined at ×50 magnification under identical illumination and imaging conditions. The comparison focused on matrix continuity, visible pores and microvoids, aggregate–paste contacts, and the distribution of the waste-derived components.

For X-ray diffraction analysis, representative fragments were dried at a temperature not exceeding 40 °C, ground to a fine powder, and passed through a 63 μm sieve. Measurements were performed using an X’Pert PRO diffractometer (PANalytical, Almelo, The Netherlands) with CuKα radiation (λ = 1.5406 Å). The diffraction patterns were recorded over a 2θ range of 10–70°.

The crystalline phases were identified by comparing the measured reflections with standard reference diffraction patterns. Particular attention was given to quartz (Q), calcite (C), portlandite (P), and ettringite (E). The comparative interpretation was based on the positions and relative intensities of the characteristic diffraction peaks.

2.5.10. Statistical Analysis

For each measured property, the arithmetic mean and sample standard deviation were calculated from three replicate measurements. The effects of concrete mixture composition were evaluated by one-way analysis of variance (ANOVA), implemented in Python 3.13.5. Separate ANOVA tests were performed for compressive strength at each curing age and for flexural strength, average density, and water absorption.

The level of statistical significance was set at α = 0.05. When the overall ANOVA indicated a statistically significant difference, Tukey’s honestly significant difference (HSD) test was applied for pairwise comparison of the mixtures while controlling the family-wise error rate. Statistical calculations were performed using Python 3.13.5 with SciPy 1.17.0 and statsmodels 0.14.6.

Because only three replicate specimens were available per mixture for the replicate-based properties, the ANOVA results were interpreted cautiously. With n = 3, the estimation of within-group variability and the assessment of distributional assumptions are limited. Therefore, the statistical results were treated as supportive comparative evidence and were interpreted together with the individual observations, mean values, standard deviations, magnitude of the differences, and Tukey HSD groupings.

3. Results

3.1. Workability and Average Density

The workability and average density results for the investigated concrete mixtures are presented in

Table 4. Slump and workability classes of the fresh concrete mixtures.

Mixture	Slump, mm	Workability Class
C0	60	P2
M1	40	P1
M2	40	P1
M3	50	P2

and

Table 5. Average density and standard deviation of hardened concrete mixtures.

Mixture	Specimen 1, kg/m3	Specimen 2, kg/m3	Specimen 3, kg/m3	Mean, kg/m3	SD, kg/m3
C0	2320	2340	2345	2335.0	13.2
M1	2280	2285	2294	2286.3	7.1
M2	2265	2271	2278	2271.3	6.5
M3	2275	2280	2290	2281.7	7.6

, respectively.

The slump of the control mixture C0 was 60 mm, corresponding to workability class P2. The introduction of rice husk ash and dried oil sludge reduced the slump to 40 mm for M1 and M2 and to 50 mm for M3. Accordingly, M1 and M2 were classified as P1, whereas M3 remained within workability class P2. Relative to the control concrete, the slump decreased by 33.3% for M1 and M2 and by 16.7% for M3. Despite this reduction, all mixtures retained workability within the P1–P2 range and could be placed and compacted without visible segregation.

C0 had an average density of 2335.0 kg/m³. After RHA and dried oil sludge were introduced, the mean values fell to 2286.3 kg/m³ for M1, 2271.3 kg/m³ for M2, and 2281.7 kg/m³ for M3. The minimum value was therefore recorded for M2, which was approximately 2.7% lighter than the control mixture. This modest reduction is consistent with the lower particle density and porous character of RHA and with the partial replacement of denser conventional constituents. Standard deviations of 6.5–13.2 kg/m³ show that the density measurements were repeatable across the three specimens.

One-way ANOVA confirmed that mixture composition had a statistically significant effect on average density (F = 29.51, p = 1.12 × 10⁻⁴). Tukey’s HSD test showed that the density of C0 was significantly higher than that of M1, M2, and M3, whereas no statistically significant differences were observed among the three modified mixtures.

3.2. Compressive Strength Development

The compressive strength development of the investigated concrete mixtures at 3, 7, 28, and 90 days is presented in

Table 6. Compressive strength development of the concrete mixtures.

Mixture	Age, Days	Specimen 1, MPa	Specimen 2, MPa	Specimen 3, MPa	Mean, MPa	SD, MPa
C0	3	13.5	13.8	14.1	13.8	0.3
C0	7	18.1	18.4	18.7	18.4	0.3
C0	28	32.0	32.4	32.8	32.4	0.4
C0	90	37.4	38.1	38.2	37.9	0.4
M1	3	20.4	21.2	20.8	20.8	0.4
M1	7	27.1	27.4	27.2	27.2	0.2
M1	28	41.0	41.7	41.3	41.3	0.4
M1	90	46.1	47.3	46.8	46.7	0.6
M2	3	20.8	21.9	21.2	21.3	0.6
M2	7	27.3	28.3	27.8	27.8	0.5
M2	28	42.3	42.7	42.5	42.5	0.2
M2	90	47.3	48.2	47.6	47.7	0.5
M3	3	20.5	20.6	20.4	20.5	0.1
M3	7	26.8	27.3	27.1	27.1	0.3
M3	28	39.5	40.8	39.8	40.0	0.7
M3	90	46.2	47.7	47.2	47.0	0.8

. All mixtures exhibited a continuous increase in compressive strength with curing age.

The modified concretes were already stronger than C0 at 3 days. Mean values were 20.8 MPa for M1, 21.3 MPa for M2, and 20.5 MPa for M3, compared with 13.8 MPa for the control. The corresponding gains were 50.7%, 54.3%, and 48.6%.

At 7 days, C0 reached 18.4 MPa. M1, M2, and M3 reached 27.2, 27.8, and 27.1 MPa, respectively. M2 had the highest early-age value and exceeded the control by 51.1%.

The 28-day means were 32.4 MPa for C0, 41.3 MPa for M1, 42.5 MPa for M2, and 40.0 MPa for M3. Thus, the increase over C0 ranged from 23.6% to 31.2%, with the maximum again recorded for M2.

The same trend was maintained at 90 days. The mean strength of M2 reached 47.7 MPa, compared with 46.7 MPa for M1, 47.0 MPa for M3, and 37.9 MPa for C0. Relative to the control concrete, the 90-day strength increased by 23.3% for M1, 25.9% for M2, and 24.1% for M3.

The largest age-related rise occurred between 7 and 28 days. Over that interval, strength increased by 76.1% in C0 and by 47.9–52.9% in the modified mixtures. From 28 to 90 days, a further increase of 12.2–17.5% was recorded.

The statistical analysis supported the age-by-age comparison. Mixture composition affected compressive strength significantly at 3, 7, 28, and 90 days (p < 0.001). Tukey’s HSD test separated C0 from all modified mixtures at every age. M1, M2, and M3 were statistically comparable at 3 and 7 days; at 28 days, each pair differed significantly; and by 90 days the three modified mixtures again formed one statistical group. Thus, the additions produced a clear improvement over the control, while the differences within the modified series depended on curing age.

The age-dependent compressive-strength development of the control and modified concrete mixtures is shown in Figure 4.

3.3. Flexural Tensile Strength

The flexural tensile strength results obtained after 28 days of curing are presented in

Table 7. Flexural tensile strength of the concrete mixtures at 28 days.

Mixture	Specimen 1, MPa	Specimen 2, MPa	Specimen 3, MPa	Mean, MPa	SD, MPa
C0	4.35	4.41	4.40	4.39	0.03
M1	4.96	5.01	5.03	5.00	0.04
M2	5.05	5.10	5.05	5.07	0.03
M3	5.01	5.05	5.03	5.03	0.02

. The control concrete C0 exhibited a mean flexural tensile strength of 4.39 MPa, whereas the modified mixtures M1, M2, and M3 reached 5.00, 5.07, and 5.03 MPa, respectively. Thus, all modified mixtures showed higher flexural tensile strength than the control concrete, with the highest numerical mean obtained for M2.

The increases over the control were 14.0% for M1, 15.5% for M2, and 14.7% for M3. M2 had the largest numerical mean, although the absolute difference between the modified mixtures was only 0.07 MPa.

The standard deviation ranged from 0.02 to 0.04 MPa, demonstrating limited scatter among the three replicate measurements. The coefficients of variation were below 0.8% for all mixtures, indicating good repeatability of the experimental results.

One-way ANOVA confirmed that mixture composition had a statistically significant effect on flexural tensile strength (F = 353.03, p = 7.75 × 10⁻⁹). Tukey’s HSD test showed that C0 differed significantly from each modified mixture, whereas the pairwise differences among M1, M2, and M3 were not statistically significant.

M2 gave the largest numerical mean for both flexural tensile strength and compressive strength. The flexural data, however, require a more restrained interpretation: Tukey’s test did not distinguish M1, M2, and M3 from one another. The three modified mixtures can therefore be regarded as having broadly comparable bending performance, even though M2 occupied the highest numerical position.

A graphical comparison of the 28-day flexural tensile strength of all investigated mixtures is presented in Figure 5.

3.4. Water Absorption and Watertightness

The water-absorption results are presented in

Table 8. Water absorption of concrete mixtures.

Mixture	Specimen 1, %	Specimen 2, %	Specimen 3, %	Mean, %	SD, %
C0	4.80	4.85	4.75	4.80	0.05
M1	3.90	3.62	3.58	3.70	0.17
M2	3.80	3.65	3.45	3.63	0.18
M3	3.85	3.70	3.55	3.70	0.15

Note: SD, standard deviation.

. The control concrete C0 exhibited a mean water absorption of 4.80%, whereas the modified mixtures showed values of 3.70% for M1, 3.63% for M2, and 3.70% for M3. The lowest value was obtained for M2, corresponding to a reduction of approximately 24.3% relative to C0.

The watertightness results obtained by the wet-spot method are summarized in

Table 9. Watertightness grades of the investigated concrete mixtures.

Mixture	Watertightness Grade	Interpretation
C0	W6	Control concrete
M1	W8	Improved watertightness
M2	W10	Highest watertightness in the test series
M3	W6	Same grade as the control mixture

. The control mixture C0 was classified as W6. The watertightness grade increased to W8 for M1 and reached a maximum of W10 for mixture M2, which showed the highest watertightness within the investigated series. Mixture M3 was classified as W6.

The wet-spot and absorption results follow the same general pattern up to M2. Relative to C0, the watertightness grade rose from W6 to W8 for M1 and to W10 for M2, while M3 returned to W6. The simultaneous reduction in absorption suggests that the capillary network became less accessible or less continuous at the intermediate dosage. For M2, this response can reasonably be linked to the filler action and pozzolanic activity of RHA, which together may produce a denser paste when the dosage is properly balanced.

3.5. Freeze–Thaw Resistance Results

The freeze–thaw test results are summarized in

Table 10. Freeze–thaw resistance of the investigated concrete mixtures.

Mixture	Maximum Completed Cycles	Increase Relative to C0, %
C0	143	—
M1	216	51.0
M2	223	55.9
M3	221	54.5

. The control concrete C0 completed 143 cycles, whereas M1, M2, and M3 completed 216, 223, and 221 cycles, respectively. Relative to C0, the increases were 51.0% for M1, 55.9% for M2, and 54.5% for M3. Mixture M2 showed the highest measured value, although the difference between M2 and M3 was only two cycles.

The maximum number of completed freeze–thaw cycles for each concrete mixture is illustrated in Figure 6.

The freeze–thaw trend agrees with the water-absorption results. Mean water absorption decreased from 4.80% for C0 to 3.70%, 3.63%, and 3.70% for M1, M2, and M3, respectively. M2 combined the lowest water absorption with the largest number of completed cycles. This correspondence suggests that lower water uptake and reduced connectivity of water-accessible pores helped limit damage during repeated freezing and thawing.

The improved freeze–thaw performance of the modified concretes may reflect two concurrent contributions: pore-system refinement associated with RHA and a possible reduction in water accessibility caused by dried oil sludge. These effects could decrease the amount of freely mobile water in the matrix and thereby reduce the development of internal stresses during freezing. This interpretation remains tentative, because hydrophobicity and capillary sorption were not measured directly and the two additions were varied simultaneously. Pore-size distribution, sorptivity, and contact-angle measurements are required to verify the proposed mechanism.

3.6. Static Modulus of Elasticity

The 28-day static modulus of elasticity results are presented in

Table 11. Static modulus of elasticity of the investigated concrete mixtures at 28 days.

Mixture	Static Modulus at 28 Days, GPa	Increase Relative to C0, %
C0	26.5	—
M1	32.0	20.8
M2	32.7	23.4
M3	32.4	22.3

. The control concrete C0 exhibited a modulus of 26.5 GPa. The corresponding values for M1, M2, and M3 were 32.0, 32.7, and 32.4 GPa, respectively.

Because the individual specimen-level modulus values were not available, standard deviations and error bars are not reported. The static-modulus results were therefore excluded from the inferential statistical analysis and are interpreted descriptively.

Relative to C0, the static modulus increased by 20.8% for M1, 23.4% for M2, and 22.3% for M3. As illustrated in Figure 7, the modified mixtures formed a narrow stiffness range of 32.0–32.7 GPa. The maximum difference among M1–M3 was 0.7 GPa, corresponding to approximately 2.1% of the highest measured value.

M2 exhibited both the highest static modulus (32.7 GPa) and the highest 28-day compressive strength (42.5 MPa). The corresponding compressive strengths were 32.4 MPa for C0, 41.3 MPa for M1, and 40.0 MPa for M3. The agreement between the highest stiffness and strength values supports the selection of M2 as the most balanced composition among the investigated mixtures.

The modulus increase was achieved together with a slight reduction in average density. The calculated specific modulus, defined here as the ratio of static modulus to average density, increased from approximately 11.35 GPa/(Mg/m³) for C0 to 14.00, 14.40, and 14.20 GPa/(Mg/m³) for M1, M2, and M3, respectively. These values correspond to improvements of approximately 23.3%, 26.9%, and 25.1% relative to the control concrete. Therefore, the modified mixtures provided greater elastic stiffness per unit mass.

Several microstructural changes could account for the higher static modulus. Better packing of the fine constituents, a more continuous cement paste, and improved stress transfer through the interfacial transition zones would all increase composite stiffness. The amorphous silica in RHA may also consume Ca(OH)₂ and form additional C–S–H, thereby reducing weak or porous regions in the hardened matrix. Because pore-size distribution and local interfacial stiffness were not measured quantitatively, these explanations should be treated as physically reasonable interpretations rather than as direct proof.

From an engineering perspective, the increase in static modulus indicates lower elastic strain under the same service-level compressive stress and may contribute to reduced short-term deformation of structural members. Among the investigated mixtures, M2 provided the most favorable overall combination of compressive strength, elastic stiffness, water resistance, and freeze–thaw performance and also exhibited the highest measured static modulus.

3.7. Microstructural Analysis

Figure 8 presents optical micrographs of the investigated concrete specimens after 28 days of curing: (a) C0; (b) M1; and (c) M2.

The microstructure of the control concrete (Figure 8a) appeared relatively heterogeneous, with clearly distinguishable aggregate particles, cementitious regions, and numerous visible voids within the matrix. Local discontinuities and non-uniform areas were also observed around some aggregate particles.

The microstructure of mixture M1 (Figure 8b) appeared visibly modified relative to C0. The cementitious matrix appeared more continuous in several regions, although comparatively large light-coloured particles and local porous zones remained distinguishable. These observations may be associated partly with the filler effect of the finely dispersed RHA particles and their contribution to improved packing within the cementitious matrix.

Mixture M2 (Figure 8c) exhibited a more uniform visual appearance and a more continuous matrix compared with C0. The boundaries between individual components were less pronounced, and the distribution of visible pores appeared more homogeneous. This may indicate improved dispersion of the waste-derived components and a modification of the internal structure of the concrete.

The optical images are therefore consistent with the mechanical and water-transport results, but they should not be overinterpreted. They provide a qualitative view of matrix continuity, visible voids, and particle contacts at the selected magnification. They do not provide total porosity, pore-size distribution, or quantitative phase information. Those characteristics would require image analysis at several scales, mercury intrusion or gas-adsorption measurements, and complementary microanalytical methods.

3.8. X-Ray Diffraction and Phase Analysis

Figure 9 compares the X-ray diffraction (XRD) patterns of mixtures C0, M1, and M2 after 28 days of curing. The main reflections were identified as quartz (Q), calcite (C), portlandite (P), and ettringite (E). A diffuse background region associated with poorly crystalline hydration products was also observed. The comparative analysis was based on the positions and relative intensities of the characteristic diffraction peaks.

The control concrete (Figure 9a) exhibited pronounced quartz reflections together with peaks assigned to portlandite, calcite, and ettringite. Quartz originated mainly from the dune sand and other siliceous constituents, whereas portlandite and ettringite were formed during cement hydration. This diffraction pattern was used as the reference for evaluating phase-related changes caused by the incorporation of RHA and dried oil sludge.

In mixture M1 (Figure 9b), the positions of the principal crystalline reflections remained generally similar to those of C0, while the relative intensity of several portlandite peaks decreased. The lower relative intensity of several portlandite reflections is compatible with possible calcium-hydroxide consumption associated with pozzolanic reactions; however, this interpretation cannot be quantified or confirmed from the present qualitative XRD data. The preservation of the principal quartz, calcite, and ettringite reflections indicates that the modified composition retained the main crystalline framework of the concrete.

Mixture M2 retained the same main crystalline phases. Relative intensities of portlandite and quartz changed further, but no new dominant crystalline phase appeared. Within the limits of the qualitative XRD interpretation, no additional dominant crystalline reflections attributable to the dried oil sludge were detected. The pattern also agrees with the more uniform appearance of the corresponding optical micrograph.

Viewed as a whole, the diffraction patterns indicate changes in the relative intensities of several crystalline reflections after the introduction of RHA and dried oil sludge, while quartz, calcite, portlandite, and ettringite remained detectable. The lower relative intensity of some portlandite peaks is compatible with possible pozzolanic consumption of calcium hydroxide, but the extent of portlandite consumption cannot be quantified from these data. Because no internal standard, Rietveld refinement, or quantitative amorphous-phase analysis was used, the XRD results should be interpreted only as qualitative supporting observations and not as direct evidence of quantified phase transformation or additional C–S–H formation.

3.9. Statistical Analysis

One-way analysis of variance confirmed that mixture composition had a statistically significant effect on every replicate-based property included in the analysis. Separate tests were performed for compressive strength at 3, 7, 28, and 90 days, flexural tensile strength at 28 days, average density, and water absorption. The ANOVA results are summarized in

Table 12. One-way ANOVA results for the investigated concrete properties.

Property	Between-Groups df	Within-Groups df	F-Value	p-Value	Significance
3-day compressive strength	3	8	265.12	2.41 × 10−8	Yes (p < 0.001)
7-day compressive strength	3	8	568.09	1.17 × 10−9	Yes (p < 0.001)
28-day compressive strength	3	8	316.80	1.19 × 10−8	Yes (p < 0.001)
90-day compressive strength	3	8	192.29	8.57 × 10−8	Yes (p < 0.001)
Flexural strength	3	8	353.03	7.75 × 10−9	Yes (p < 0.001)
Average density	3	8	29.51	1.12 × 10−4	Yes (p < 0.001)
Water absorption	3	8	43.95	2.57 × 10−5	Yes (p < 0.001)

.

Every test gave p < 0.001. F values for the strength properties ranged from 192.29 to 568.09. Density and water absorption were also affected by mixture composition, with F = 29.51 (p = 1.12 × 10⁻⁴) and F = 43.95 (p = 2.57 × 10⁻⁵), respectively.

Tukey’s HSD test provided a more detailed picture of the compressive-strength differences. C0 was significantly weaker than each modified mixture at all curing ages. M1, M2, and M3 did not differ significantly at 3 or 7 days; all pairwise contrasts were significant at 28 days; and at 90 days the three modified mixtures were again statistically comparable. The strongest separation among the modified compositions was therefore observed at the conventional 28-day test age.

For flexural tensile strength, all three modified mixtures differed from C0, whereas M1, M2, and M3 did not differ significantly from one another. M2 had the highest numerical mean. Density and water absorption produced a similar grouping: the control differed from every modified mixture, but the modified mixtures were statistically indistinguishable within each property. This pattern indicates a consistent modification effect within the present dataset without implying that every small numerical difference within M1–M3 is meaningful.

The statistical results indicate that the combined introduction of RHA and dried oil sludge changed the measured properties relative to C0 consistently within the present limited dataset. M2 provided the highest numerical strength values and the most favorable durability indicators, but the flexural results for M1–M3 belonged to one statistical group. Freeze–thaw resistance, watertightness grade, and static modulus were not included in ANOVA because complete specimen-level replicate data were unavailable for those properties; they should therefore be interpreted descriptively rather than as statistically ranked outcomes.

3.10. Sustainability and Practical Implications

The mixture design uses dune sand together with two secondary materials from local agricultural and petroleum activities. This may reduce the demand for conventional raw materials and create an additional use for wastes that otherwise require disposal.

At the mixture-proportion level, the replacement of Portland cement with RHA reduced cement consumption from 350 kg/m³ in C0 to 330, 310, and 290 kg/m³ in M1, M2, and M3, corresponding to reductions of 5.7%, 11.4%, and 17.1%, respectively. When RHA and dried oil sludge are considered together, the modified mixtures incorporated 30, 55, and 80 kg/m³ of secondary materials, respectively. For M2, which exhibited the most balanced overall performance, this corresponds to a reduction of 40 kg/m³ of Portland cement and the utilization of 55 kg/m³ of secondary materials. These values provide a preliminary resource-substitution indicator only and should not be interpreted as quantified CO₂ savings because the energy and emissions associated with RHA heat treatment, oil-sludge drying, transportation, and other life-cycle stages were not determined. A cradle-to-gate life-cycle assessment using verified local inventory data is required to quantify the net environmental impact.

Mechanical performance alone does not establish an environmental advantage. Leaching, residual hydrocarbons, heavy metals, energy used for drying and heat treatment, and the complete life-cycle of the concrete must be assessed before the environmental claim can be quantified.

3.11. Limitations

The test program covered workability, average density, compressive and flexural strength, water absorption, watertightness, freeze–thaw resistance, static modulus, optical microscopy, and XRD. The freeze–thaw assessment nevertheless remains incomplete without the full specimen-level record of mass change and residual strength at each inspection interval. Retaining those data would allow the formal frost-resistance grade to be verified more transparently against GOST 10060-2012.

Static-modulus values were available as mean results for three prisms, but the individual load–deformation curves and standard deviations were not retained in the available laboratory records. For that reason, Figure 7 contains no error bars and the modulus was not entered into ANOVA. The program also did not include chloride penetration, sulfate resistance, carbonation, long-term dimensional stability, or environmental leaching of potentially hazardous constituents from the dried oil sludge. Consequently, the present results cannot be used to establish the environmental safety of the concrete or the complete immobilization of potentially hazardous constituents. Standardized leaching tests are required before practical application.

The fresh-state characterization was limited to the initial slump test. The air content and workability retention over time were not measured; therefore, the possible effects of RHA and dried oil sludge on entrapped or entrained air and time-dependent slump loss could not be quantified.

Isothermal calorimetry was not performed; therefore, the effect of dried oil sludge on cement hydration kinetics and heat evolution could not be determined.

Future work should add direct measures of water repellency, including contact-angle and capillary-sorption tests, together with quantitative pore-structure analysis. Long-term exposure to chlorides, sulfates, carbonation, and repeated wetting–drying should also be examined. A future factorial experimental design should include RHA-only, oil-sludge-only, and combined mixtures to separate the individual effects of the two additions and determine whether a statistically significant interaction exists.

The oxide composition and loss on ignition reported for the RHA source material were based on available characterization data and were not independently verified for the specific experimental batch. The BET specific surface area and particle-size characteristics were determined experimentally; however, the amorphous-silica content was not quantified. This limitation should be considered when interpreting the pozzolanic activity of the RHA.

The microstructural characterization was limited to optical microscopy and qualitative XRD interpretation. SEM-EDS, mercury intrusion porosimetry, and quantitative image analysis were not performed. Therefore, the proposed explanations concerning matrix continuity, pore refinement, interfacial improvement, and phase development should be regarded as tentative. Future work should apply complementary microanalytical and pore-structure techniques at several observation scales to verify these interpretations.

The XRD analysis was qualitative because no internal standard, Rietveld refinement, or quantitative amorphous-phase analysis was used. Therefore, the observed reduction in the relative intensity of the portlandite reflections cannot be used to quantify portlandite consumption or additional C–S–H formation. Future studies should apply quantitative phase analysis to verify the proposed phase-development mechanism.

The use of three replicate specimens per mixture limits the statistical power and precision of the estimated variability, particularly for properties affected by concrete heterogeneity. A larger number of specimens should be used in future studies to confirm the present ANOVA-based comparisons and provide more reliable estimates of experimental variability.

4. Discussion

The slump results provide the first indication that the additions changed the fresh-mixture behavior. C0 measured 60 mm, whereas the modified mixtures were in the 40–50 mm range. The most likely reason is the fine, porous character of RHA, which increases the surface that must be wetted and therefore raises water demand when the mixing-water content is not increased [2,3,4,5,11]. This explanation fits the lower values obtained for M1 and M2. Even so, all mixtures remained within consistency classes P1–P2 and were placed and compacted without visible segregation, so the loss of flow did not prevent laboratory production.

Density changed much less than the mechanical properties. Introducing the two waste-derived components reduced average density by only 2.1–2.7%, a range that can be explained by the lower particle density and porous morphology of RHA and by partial replacement of denser constituents. Tukey’s test did not distinguish M1, M2, and M3 from one another. The strength increase was therefore not obtained by making the concrete heavier; rather, the modified mixtures developed a more efficient load-bearing matrix at a slightly lower unit mass.

Figure 4 shows a clear strength advantage for every modified mixture over C0 at all four ages. The particularly large early-age strength increase—approximately 49–54% at 3 days and 47–51% at 7 days relative to C0—cannot be attributed solely to the pozzolanic reaction of RHA. Although the Portland cement content decreased, the total binder content remained constant at 350 kg/m³, and the water-to-binder ratio was maintained at 0.53. At early ages, the fine RHA particles and their high specific surface area may improve particle packing, fill intergranular voids, and provide additional heterogeneous nucleation sites for cement-hydration products, thereby promoting the earlier formation of a denser and more continuous load-bearing matrix. An initial pozzolanic contribution may also occur during the first days of curing, while its relative importance may increase at later ages. However, because RHA and dried oil sludge were varied simultaneously, their individual contributions cannot be separated using the present mixture design. Furthermore, no isothermal calorimetry or quantitative early-age phase analysis was performed. The observed early-age strength gain is therefore interpreted as the combined result of physical filling and nucleation effects, hydration-related structure development, and a possible initial pozzolanic contribution, rather than as a purely pozzolanic effect. M2 gave the strongest overall response, reaching 27.8 MPa at 7 days, 42.5 MPa at 28 days, and 47.7 MPa at 90 days. Comparable gains at controlled RHA dosages have been reported in conventional and high-performance concretes [1,2,3,4,5,6,7,8,9,16].

The strength response was not proportional to the amount of added material. Moving from M2 to M3 increased both RHA and dried oil sludge, but the 28-day strength fell from 42.5 to 40.0 MPa. M2 therefore appears to represent a better balance among reactive silica, available calcium hydroxide, water demand, and paste continuity. At the higher dosage, the additional RHA surface may not have been wetted as effectively, while the larger sludge fraction may have increased heterogeneity or diluted the cementitious system. This non-monotonic behavior agrees with the broader RHA literature, in which benefits are normally confined to a rational dosage range [3,7,8,9].

Flexural tensile strength followed the same broad direction but with a narrower spread. C0 reached 4.39 MPa, and the modified mixtures ranged from 5.00 to 5.07 MPa, equivalent to increases of 14.0–15.5%. M2 again had the largest numerical mean. Statistical testing, however, separated only the control from the modified mixtures; it did not separate M1, M2, and M3 from one another. The main improvement in bending resistance was thus achieved once the combined modification was introduced, while further changes within the investigated dosage range had little measurable effect.

The modulus results strengthen the interpretation that the matrix became mechanically more effective. Static modulus increased from 26.5 GPa in C0 to 32.0–32.7 GPa in M1–M3, a gain of 20.8–23.4%. Because density decreased slightly, the change in specific modulus was even more pronounced: approximately 11.35 GPa/(Mg/m³) for C0 compared with 14.00, 14.40, and 14.20 GPa/(Mg/m³) for M1, M2, and M3. The narrow 0.7 GPa interval within the modified series again suggests that most of the stiffness benefit resulted from introducing the combined system, not from continually increasing its dosage. M2 combined the highest modulus with the highest 28-day compressive strength.

Water-transport properties also favored the intermediate composition. Absorption fell from 4.80% for C0 to 3.63–3.70% for the modified concretes, and M2 reached the lowest value together with watertightness grade W10. Because the absorption and wet-spot tests measure different aspects of water ingress, their agreement gives stronger support to the conclusion that M2 developed a less continuous capillary network. This response is consistent with earlier work showing that properly processed RHA can reduce permeability through microfilling and secondary hydration-product formation [3,7,8,12,13].

The freeze–thaw results are compatible with this interpretation. The modified mixtures completed 216–223 cycles, compared with 143 cycles for C0, corresponding to increases of 51.0–55.9%. M2 again ranked first, although its advantage over M3 was only two cycles. During freezing, damage depends strongly on the amount and mobility of water in the pore system. RHA-induced pore refinement may have reduced the connected water-accessible volume, while dried oil sludge may have contributed a limited water-repellent effect. Because all modified mixtures contained both additions and no direct hydrophobicity test was performed, their separate contributions and any synergistic effect cannot be quantified. The agreement among water absorption, watertightness, and freeze–thaw resistance therefore supports this mechanism only indirectly [12,13,15].

Hong et al. [20] reported a related link between RHA, pore refinement, and freeze–thaw resistance in a concrete containing both RHA and fly ash. Their material also included a multi-stage phase-change component, so the numerical values are not directly comparable with the present dune-sand concrete. The study is nevertheless useful as supporting evidence that a properly proportioned silica-rich addition can improve cyclic durability when the pore system and water transport are favorably modified.

M3 deserves a separate comment because it completed nearly as many cycles as M2 despite having a lower 28-day compressive strength and only W6 watertightness. This result confirms that freeze–thaw resistance cannot be predicted from compressive strength alone. Pore geometry, degree of saturation, local water redistribution, and the ability of the matrix to resist crack growth all contribute. The relatively high flexural tensile strength of M3 may also have helped delay surface scaling and crack propagation. Ranking the mixtures therefore requires several durability indicators rather than a single strength value.

The optical micrographs in Figure 8 provide qualitative support for the macroscopic trends. C0 contained more visible voids and more clearly defined discontinuities around some particles. Mixture M1 appeared more continuous in several regions, and M2 had the most uniform overall appearance. These observations are compatible with better packing of the fine fraction and fewer large, connected defects. They also agree with the lower absorption and higher mechanical performance of the modified concretes, although the images alone cannot quantify porosity.

XRD adds complementary information about phase development. Quartz, calcite, portlandite, and ettringite remained present in all three patterns, so the additions did not replace the main crystalline framework of the hydrated material. Several portlandite reflections were relatively weaker in M1 and M2 than in C0, which is consistent with consumption of calcium hydroxide during a pozzolanic reaction. No additional dominant crystalline phase attributable to dried oil sludge was observed in either modified mixture, suggesting compatibility with the binder system at the investigated dosages.

The mechanical, transport, microscopy, and diffraction observations point in the same general direction. A more continuous matrix and reduced relative portlandite intensity are compatible with additional hydration products occupying available space and improving aggregate–paste contact. This offers a coherent explanation for the simultaneous increases in strength and stiffness and the reduction in water uptake. The evidence supports both a physical role of RHA through particle packing and a chemical role through pozzolanic reaction, but the phase and pore changes were not quantified directly.

The contribution of dried oil sludge cannot be isolated from these data because it was introduced together with RHA in every modified mixture. Previous studies have shown that properly treated petroleum sludge or its ash can be accommodated in cement-based materials at low dosages [21,22,23,24,25,26]. In the present series, the sludge did not prevent strength development, stiffness gain, watertightness, or freeze–thaw resistance. The best combined response occurred at 15 kg/m³ dried oil sludge in M2, indicating that this limited dosage can be incorporated within the RHA-modified matrix. A factorial experimental design is still required to identify its independent effect.

From an engineering standpoint, M2 offers the most balanced combination rather than the maximum of only one property. Relative to C0, it increased 28-day compressive strength by 31.2% and static modulus by 23.4%, reduced water absorption by 24.3%, reached W10, and completed the greatest number of freeze–thaw cycles. This combined performance is more relevant to structural use than an isolated strength result. The composition also makes use of local dune sand and two secondary resources from agricultural and petroleum activities, although the environmental benefit must be verified separately.

The study therefore establishes laboratory-scale feasibility and identifies a composition that warrants further development. Before practical use, M2 should be reproduced under production-scale mixing and curing conditions and tested for the exposure class in which it is intended to operate. Long-term chloride, sulfate, carbonation, and dimensional-stability tests are required, as are leaching and residual-hydrocarbon analyses for the dried oil sludge. Only after these checks can the RHA–oil-sludge system be assessed reliably as a sustainable option for dune-sand concrete.

5. Conclusions

1. The experiments showed that dune-sand structural concrete can be produced with RHA and dried oil sludge without losing practical consistency. Slump decreased from 60 mm in C0 to 40–50 mm in M1–M3, but all mixtures remained within classes P1–P2 and were cast and compacted without visible segregation. Average density was 2.1–2.7% lower than that of the control, so the performance gains were achieved without an increase in unit weight.

2. The modified concretes were stronger than C0 at every curing age. M2, containing 310 kg/m³ Portland cement, 690 kg/m³ dune sand, 1110 kg/m³ crushed stone, 40 kg/m³ RHA, and 15 kg/m³ dried oil sludge, produced the highest numerical compressive-strength means: 21.3 MPa at 3 days, 27.8 MPa at 7 days, 42.5 MPa at 28 days, and 47.7 MPa at 90 days. These values were 54.3%, 51.1%, 31.2%, and 25.9% above the corresponding control strengths.

3. Flexural tensile strength increased from 4.39 MPa in C0 to 5.00–5.07 MPa in the modified series. The largest mean, 5.07 MPa, was obtained for M2 and represented a 15.5% increase. Tukey’s HSD test did not distinguish M1, M2, and M3, which means that the bending-strength benefit was maintained across the investigated dosage range rather than being confined statistically to one modified composition.

4. Static modulus increased from 26.5 GPa for C0 to 32.0–32.7 GPa for the modified concretes, corresponding to gains of 20.8–23.4%. M2 reached the highest value of 32.7 GPa. Since this increase occurred while density fell slightly, the modified mixtures also developed greater elastic stiffness per unit mass, with M2 again giving the most favorable numerical result.

5. Water resistance and cyclic durability improved together. Absorption decreased from 4.80% in C0 to 3.63–3.70% in the modified mixtures; M2 recorded 3.63%, reached watertightness grade W10, and completed 223 freeze–thaw cycles, 55.9% more than C0. The combined trend is compatible with RHA-related pore refinement and a possible water-repellent contribution from dried oil sludge. This mechanism is plausible but not proven and should be examined using contact-angle, capillary-sorption, and quantitative pore-structure measurements.

6. Optical microscopy showed a visually more continuous and homogeneous matrix after RHA and dried oil sludge were introduced. XRD retained quartz, calcite, portlandite, and ettringite as the main crystalline phases, while the relative intensity of several portlandite reflections decreased in the modified samples. The observations are consistent with filler action and pozzolanic consumption of Ca(OH)₂ but remain qualitative because no quantitative image analysis, pore characterization, internal-standard XRD, or Rietveld refinement was carried out.

7. One-way ANOVA identified a significant mixture-composition effect for compressive strength at all ages, flexural tensile strength, density, and water absorption (p < 0.001). Tukey’s test consistently separated C0 from the modified concretes, whereas several differences within M1–M3 were not significant. The improvement over the control was consistent within the present limited dataset, but the small sample size requires the ANOVA-based comparisons to be interpreted cautiously.

8. On the basis of strength, modulus, water absorption, watertightness, and freeze–thaw resistance considered together, M2 is the rational composition for the next stage of work. It should not be treated as a universal optimum outside the present materials and test conditions. Production-scale validation, long-term exposure testing, chloride and sulfate resistance, carbonation, leaching, residual-hydrocarbon characterization, and life-cycle assessment are still required before the concrete can be recommended for practical implementation.