Effects of Ettringite Formation on the Stability of Cement-Treated Sediments

Abstract

This study explores the stabilization of dredged sediments classified as lean clay (CL) using hydrated lime, type III Portland cement, and compaction. While quicklime is commonly used in practice, this research explores alternative calcium-based binders with the aim of valorizing sediments for civil engineering applications. The mechanical behavior of the treated materials was evaluated through an Unconfined Compressive Strength (UCS) test campaign, with the results interpreted using the porosity/volumetric cement content ($\eta /C_{iv}$) index. This relationship assesses the influence of apparent dry density and cement content on the strength improvement of sediments, aiming to evaluate the suitability of the dredged sediments for engineering applications. A key feature of this study is the extended curing period of up to 90 days, which goes beyond the typical 28-day evaluations commonly found in the literature. Interestingly, strength degradation occurred at advanced curing ages compared to shorter curing times. To understand the mechanisms underlying this resistance degradation, the mixtures were subjected to X-ray fluorescence spectroscopy (XRF), X-ray diffraction (XRD), and thermogravimetric analysis (TGA). These tests identified the presence of the expansive sulfate-based compound ettringite, which is associated with swelling and failure in soils stabilized with calcium-based stabilizers. This research contributes to the field by demonstrating the limitations of calcium-based binders in stabilizing sulfate-bearing dredged materials and emphasizing the importance of long-term curing in assessing the durability of treated sediments.

1. Introduction

The Port of Rio Grande is the third largest port in Brazil in terms of cargo handling. Approximately every two years, maintenance dredging operations are conducted to maintain an average draft of 16.5 m, with millions of cubic meters of sediments dredged during each operation. The literature has highlighted the relationship between dredging operations and the offshore disposal of dredged material in terms of environmental impacts along the local coast [1,2]. Considering this, there is a need to investigate alternative destinations for these dredged sediments. Other authors have explored the chemical stabilization of Port of Rio Grande sediments for engineering applications [3,4], and the present study follows in their footsteps.

The primary purpose of dredging operations is to remove sediment from water bodies for bathymetric leveling or infrastructure construction. Additionally, sediment disposal in aquatic environments is linked to heavy metal accumulation, the physical disturbance and burial of benthic communities, shifts in species compositions, and the overall worsening of ecological indicators [5,6]. On the other hand, beneficial uses of dredged sediments, such as repurposing them as construction materials, have led to reduced environmental impacts [7]. Exploring effective stabilization methods for dredged sediments is essential, as this soil-like material typically has a high water content and contains organic matter and clay, resulting in low shear strength [8,9]. Soil cementing techniques are widely used to address these issues and have shown promising results in dredged sediments in numerous studies [10,11,12,13]. Several studies have shown that stabilized sediments can serve as alternative materials for geotechnical and road construction purposes, contributing to resource efficiency and waste reduction. For example, the feasibility of reusing dredged sediments from Brest Harbor (France) in road construction applications has been demonstrated, highlighting both their mechanical performance and environmental safety [14].

Some studies in the literature report good results from the treatment of marine sediments with both quicklime and hydrated lime [15,16,17]. This study employs pre-treatment of sediments with hydrated lime and use type III Portland cement as alternative calcium-based binders. The choice of the hydrated lime pre-treatment allows for particle aggregation without the risk of excessive drying, which could negatively affect the sediments’ workability. The intent is to explore the mechanical behavior and chemical durability of stabilized dredged sediments, particularly in sulfate-rich marine environments such as the Port of Rio Grande.

Although prior studies have demonstrated the feasibility of dredged sediment stabilization, most focus on short-term performance, typically evaluating strength gains within 7 to 28 days of curing. However, long-term chemical processes in sulfate-bearing environments may compromise the durability of stabilized soils, which remains a gap in the literature.

The primary objective of this research was to establish the relationship between the Unconfined Compressive Strength (UCS) and the ratio of porosity to volumetric content of cement ($\eta /C_{iv}$) [18]. The authors observed that even a small increase in cement content led to a significant rise in UCS, demonstrating a linear correlation between cement content and strength. A reduction in open porosity (or an increase in dry density) caused an exponential increase in strength due to the higher number of particle contacts within the soil. This methodology has been successfully applied to various types of soil and chemical stabilizers in other studies [19,20,21].

Throughout the unconfined compression (UC) testing campaign, it was noticed that there was a decrease in UCS after longer curing periods compared to the specimens cured for 7 days. Thus, to identify the possible reasons for this strength degradation, the raw sediments were subjected to X-ray fluorescence to verify whether they contained possible contaminants that were not identified during the characterization described in a previous study [22]. UC test specimens were subject to X-ray diffraction and thermogravimetric analyses to identify undesirable compounds.

Studies have identified sulfate contamination as the source of swelling and strength degradation in soils stabilized with calcium-based compounds [23,24]. When a sulfate-bearing soil is treated with calcium-based stabilizers, such as lime and Portland cement, the sulfate and calcium combine with aluminates released from the dissolution of the clay or provided by the cement, resulting in the formation of the expansive crystal ettringite [25,26]. The chemical formula of ettringite is presented below:

$${Ca}_6{\left(Al{\left(OH\right)}_6\right)}_2{\left({SO}_4\right)}_{3}26H_{2}O$$

The mechanisms behind strength depletion in soil–cement are linked to the adsorption of water by ettringite [25]. On the one hand, the water itself can induce the expansion of ettringite. On the other hand, water can promote further ettringite growth by concentrating aluminum, calcium, and sulfate ions around the ettringite crystal. This expansion increases the pressure exerted by ettringite crystals, potentially breaking soil–cement bonds and disrupting particle interlocking.

The formation of ettringite is influenced by several factors, including the concentration of sulfate ions and the curing conditions (such as temperature and time) [27]. The rate of ettringite formation can vary depending on the concentration of sulfate and the availability of calcium and aluminum sources. In the early curing stages, high sulfate concentrations may accelerate the formation of ettringite, leading to rapid swelling and strength loss. However, over time, the continued growth of ettringite under certain curing conditions may exacerbate these effects, contributing to long-term deterioration. The kinetics of ettringite formation and its role in the failure mechanisms of soil–cement matrices are therefore crucial for understanding both the short-term and long-term stability of stabilized soils [28,29].

Additionally, a decrease in pH can result in Ca²⁺ leaching, which hinders the formation of calcium silicate hydrates (the resilient compounds in the soil–cement matrix), potentially causing the soil to behave like an unstabilized soil over time [30]. Ettringite is also linked to increased soil plasticity, excessive heaving, and the general failure of soil mass [24,30].

Marine clays are highly vulnerable to sulfate attacks due to the presence of sulfates that originate from natural sources and industrial effluents [31]. The stabilization of soils and sediments from marine environments has been extensively studied [32,33,34,35].

Sulfates, particularly sodium sulfate, can reduce the effectiveness of lime stabilization by forming expansive ettringite [32]. Before using lime to stabilize such soils, it is essential to check for the presence of acid-dissolved sulfates, as ettringite formation reduces the shear strength gained from the stabilization process [33]. On other hand, sulfate thresholds alone are not sufficient to predict ettringite formation, and the mineralogical and geotechnical properties of the marine soil must also be considered [34]. To avoid the formation of ettringite and its detrimental effects in marine sediments stabilized by calcium-based stabilizers, a binary binder composed of magnesium oxide and ground granulated blast furnace slag is a viable alternative, as it does not induce ettringite formation [35].

The primary goal of this study was to stabilize dredged sediments from the Port of Rio Grande using high-early-strength type III Portland cement [31], hydrated lime, and compaction and to assess their suitability for geotechnical applications. This research addresses a critical research gap in stabilizing sulfate-bearing dredged sediments, where the late formation of expansive minerals, such as ettringite, poses a significant challenge to their long-term performance. Although previous studies have explored soil stabilization using calcium-based compounds, they have often focused on short-term strength gains, neglecting the implications of chemical interactions at advanced curing ages. A key innovation of this research lies in its extended curing periods of up to 90 days, which allow for the identification of post-peak strength behaviors and durability issues not typically captured in conventional studies.

By establishing the relationship between Unconfined Compressive Strength (UCS) and the porosity-to-volumetric cement content ratio ($\eta /C_{iv}$), this study quantitatively assesses the contribution of dry density and binder content to the initial strength gain of stabilized sediments. However, more importantly, it investigates the unexpected strength loss seen at later ages. This behavior underscores the need for a more comprehensive understanding of the underlying chemical mechanisms that drive strength degradation in sulfate-rich environments. By conducting a series of chemical analyses, including X-ray fluorescence (XRF), X-ray diffraction (XRD), and a thermogravimetric analysis (TGA), this study not only investigates the formation of expansive compounds but also evaluates their impact on the integrity of the soil–cement matrix over time.

This study thus contributes to the growing body of knowledge on sustainable sediment valorization by (i) highlighting the risks of long-term deterioration in sulfate-rich conditions, (ii) demonstrating the limitations of conventional calcium-based binders in such environments, and (iii) suggesting the need for alternative stabilization strategies in marine sediment applications.

2. Materials and Methods

2.1. Materials

The materials used in this study were dredged sediments, Portland cement, and hydrated lime. The sediments were collected from the access channel of the Port of Rio Grande and classified as a CL (inorganic lean clay) according to the Unified Soil Classification System—USCS [36]. This soil has similar characteristics to those of marine clays found along the Brazilian coast [37,38]. The Portland cement selected as the stabilizing agent was a high-early-strength type III cement, conforming to C150/C150M [39].

The hydrated lime used to improve the sediments’ workability was dolomitic hydrated lime. It chosen as a pre-treatment step to induce the flocculation of clay particles, improving workability and initial sediment structure without excessively drying the material, which would hinder its handling and compaction. Furthermore, as reported in the literature, while both quicklime and hydrated lime have been shown to be effective in the stabilization of fine-grained soils, the use of hydrated lime in this study allowed for a more uniform and controlled treatment due particularly to its lower heat release during hydration, which helps maintain better moisture conditions during mixing and compaction.

2.2. Characterization

The geotechnical characterization of the sediments studied in this work was detailed in a previous study [22]. A summary of the main properties of the sediments is presented in

Table 1. Physical properties of dredged sediments.

Parameters	Value
Initial water content (%)	110
Organic matter content (%)	7.71
Grain size > 63 μm (%)	48
63 μm > Grain size > 2 μm (%)	24
Grain size < 2 μm (%)	28
Classification (USCS)	CL
Liquid limit (%)	49
Plastic limit (%)	19
Plasticity index (%)	30
$G_s$	2.53
$\mathrm{Standard} \mathrm{Proctor} {\rho }_{dmax}$ (g/cm3)	1.46
$\mathrm{Standard} \mathrm{Proctor} W_{opt}$ (%)	25.20
$\mathrm{Modified} \mathrm{Proctor} {\rho }_{dmax}$ (g/cm3)	1.60
$\mathrm{Modified} \mathrm{Proctor} W_{opt}$ (%)	24.50
pH	7.89

. It is important to note that the soil has a large fine fraction (52%), a relatively high liquid limit (LL) of 49%, and a pH of 7.89, which is within a neutral range. The dry unit mass obtained from the standard Proctor test was 1.46 g/cm³, with an optimum moisture content of 25.20%. When using the modified Proctor test, the dry unit mass was 1.60 g/cm³, with an optimum moisture content of 24.50%.

X-ray diffraction measurements were carried out using a Panalycal X’Pert Pro diffractometer (Malvern Panalytical B.V., Almelo, The Netherlands). This unit generates X-rays via a copper tube, operating at 45 kV and 40 mA. Bragg–Brentano geometry was used, with beam collimation achieved by employing 0.04 rad Soller slits, 1° anti-scatter slits, a fixed divergence slit of 0.5°, and a 10 mm mask. The X’Celerator linear detector was used for diffraction beam detection, with an angular aperture of 2.122°. At least three scans were performed across the 2θ range from 7° to 97°, with step sizes smaller than 0.016° and a step time of approximately 60 s, resulting in a total data collection time of 2 h and 10 min per sample. The powdered samples were placed in stainless steel holders with a cavity that was 27 mm in diameter and 3 mm in depth.

XRD data were analyzed using the Rietveld methodology [40], a refinement method used for peak analysis that correlates the intensities of material peaks with those of previously characterized crystals. The Inorganic Crystal Structure Database (ICSD) was used as a reference [41]. Fundamental parameter analysis was performed using TOPAS software 6.0 [42]. The analysis of dredged sediment (DS) identified quartz (55.4%), illite (33.9%), and anorthite (10.7%) in the samples, as shown in

Table 2. XRD analysis of sediments and cement.

Phase	Description	DS (%)	PC (%)	HL (%)	ICSD
SiO2	Quartz	55.4 (0.90)	2.51 (0.90)	0.88 (0.06)	174
K(Al4Si2O9(OH)3)	Illite	33.9 (0.80)	-	-	90,144
CaAl2(SiO4)2	Anorthite	10.7(0.90)	-	-	22,022
C3S	Tricalcium Silicate	-	73.30 (0.50)	-	94,742
C2S	Dicalcium Silicate	-	9.20 (0.50)	-	79,550
CaCO3	Calcite	-	7.70 (0.30)	24.50 (0.20)	79,673
C4AF	Tetra calcium Aluminoferrite	-	4.43 (0.19)	-	9197
C3A–Cubic	Tricalcium Aluminate	-	2.86 (0.14)	-	1841
Ca(OH)2	Portlandite	-	-	41.00 (0.30)	15,471
Mg(OH)2	Brucite	-	-	31.80 (0.30)	34,401
CaMg(CO3)2	Dolomite	-	-	1.87 (0.17)	10,404
$wp$		8.92	10.34	9.22

. The analysis of the Portland cement (PC) revealed its main components to be tricalcium silicate (73.3%), dicalcium silicate (9.2%), calcite (7.7%), tetra-calcium aluminoferrite (4.43%), and tricalcium aluminate (2.86%). Finally, the hydrated lime (HL) analysis identified its primary phases to be portlandite (41.0%), brucite (31.8%), and calcite (24.5%), indicating a significant carbonation of the lime, as its hydroxide content should have been at least 95%.

Figure 1 shows the XRD patterns of the hydrated lime (a), Portland cement (b), and dredged sediments (c). Based on their residual weighted profile ($wp$), which quantifies the adjustment in quality between the observed and calculated diffraction patterns, the values of $wp$ for the dredged material, cement, and hydrated lime were $wp$ = 8.92%, $wp$ = 10.34%, and $wp$ = 9.22%, respectively. These values are considered acceptable for analysis.

2.3. Unconfined Compression Tests

The specimens used in the UC tests were molded to an approximate height of 100 mm and a diameter of 50 mm. Their cement content was set at 3%, 5%, and 7% of the dry soil mass. These cement contents were pre-selected based on recommendations for soils with high fine contents [43,44]. Studies have used similar contents and achieved the stabilization of dredged material [12,45,46]. The dry unit mass was chosen to be 95% and 100% of the maximum dry mass, as determined from standard and modified Proctor tests. The water content was fixed at 25% for all dry unit masses to simplify the procedure. To improve the workability of the sediments, a pre-treatment consisting of 1% hydrated lime was applied to induce clay flocculation 24 h prior to molding.

The curing periods were set to 7, 28, and 90 days. For each curing period, two replicates of each treatment combination were molded.

Table 3. Summary of mixtures tested.

${\mathit{\rho }}_{\mathit{d}}$ (g/cm3)	w (%)	HL (%)	PC (%)	Curing Time (Days)
1.40 (95% Standard)	25	1	3, 5, 7	7, 28, 90
1.46 (100% Standard)	25	1	3, 5, 7	7, 28, 90
1.52 (95% Modified)	25	1	3, 5, 7	7, 28, 90
1.60 (100% Modified)	25	1	3, 5, 7	7, 28, 90

summarizes all the treatments adopted in the UC campaign.

Due to the high water content of the sediment, the sample was dried out in an oven at 60 °C for 24 h to prevent the combustion of its potential organic matter. The material was spread in thin layers to prevent it from becoming stiff and agglomerated after dehydration, making it easier to crumble and, once crumbled, pass through a #10 mesh sieve. The powdered soil was then mixed with the prescribed amount of water and hydrated lime. This premixture was stored in a sealed container for 24 h to prevent moisture loss.

After 24 h, the pre-treated sediments were mixed with the prescribed amount of Portland cement. The specimens were molded using dynamic compaction in a Mini-MCV device. This apparatus is used for Moisture Condition Value (MCV) tests and was chosen for specimen molding because its molds have convenient dimensions of a 10 mm height and 5 mm diameter. The compaction was executed using a small Proctor hammer. The amount of material used in each compaction layer was weighed on a precision scale, and the number of blows required to achieve the prescribed dry unit was applied (e.g., 6 blows were needed to reach 1.40 g/cm³). After each layer was compacted, the surface was systematically scarified to ensure good adhesion between layers. Once de-molded, three axial and diametral measurements were taken at different points using a precision caliper to ensure the specimens met the prescribed height and diameter. The specimens were also weighed on a precision scale to confirm they had the correct mass. They were cured for the three prescribed periods in a curing chamber, where the temperature was kept at 25 (±1) °C and the relative humidity at 100%.

After the prescribed curing period, the specimens were immersed in water to ensure that they were saturated for the UC tests. The specimens were then subjected to axial loading at approximately 1 mm/min, providing a constant strain rate until failure. Their maximum load was recorded with a precision of 10 N. Following this initial set of UC tests and preliminary data analysis, additional specimens were molded for treatments with UCS values that deviated by more than 10% from those of their replicates. These new specimens were molded, cured, and tested following the same methodology as previously employed.

2.4. Porosity/Cement Index ($\eta /C_{iv}$)

The ratio between the volume of the voids in soil and the total volume of the soil is the porosity (η), which is defined by Equation (1) [18]. Porosity is a function of the dry unit mass (${\rho }_D$) and content (%) of Portland cement (PC) and hydrated lime (HL). The unit masses of the solid sediments, Portland cement, and hydrated lime are represented by ${\rho }_{sSd}$, ${\rho }_{sCP}$, and ${\rho }_{sCH}$, respectively:

$$\eta =100-100\left[\left({\displaystyle \frac{{\rho }_D}{1+\frac{PC+HL}{100}}}\right)·\left({\displaystyle \frac{1}{{\rho }_{sSd}}}+{\displaystyle \frac{\frac{PC}{100}}{{\rho }_{sCP}}}+{\displaystyle \frac{\frac{HL}{100}}{{\rho }_{sCH}}}\right)\right]$$

(1)

$C_{iv}$ is the volumetric content of Portland cement, defined as the ratio between the volume of added cement and the total volume of the mixture.

2.5. Thermogravimetric Tests

For the thermogravimetric analysis, samples were taken from specimens treated with 7% Portland cement and cured for 7, 28, and 90 days. These were the same specimens previously tested in the UC tests. Analyzing samples subjected to different curing durations was crucial; ettringite is a typical product seen in the early stages of cement hydration, but with the reduction in sulfate saturation provided by the cement, ettringite is expected to transform into other compounds [26]. The thermogravimetric analysis involves observing the thermal decomposition of samples subjected to a controlled temperature increase. The temperature ramp was set at 10 °C/min, reaching a maximum temperature of 1000 °C, with 30 min isotherms. The analysis of this test reveals the relationship between the recorded mass variation, the time derivative, and the temperature gradient.

2.6. X-Ray Diffractometry and Fluorescence Tests

X-ray diffractometry was performed on a sample treated with 5% Portland cement cured for 7 days and another sample treated with 7% of Portland cement cured for 90 days. Both samples were taken from specimens used in the UC tests. The technique employed for this analysis was the same as previously described for the raw material. The selected soil–cement samples were ground to a powder with particles less than 200 μm in size for the analysis. The fluorescence tests were conducted using a powdered sample of the dry raw sediments.

2.7. Text Refinement Using AI Tools

During manuscript preparation, ChatGPT-3 by OpenAI (GPT-4-turbo) was employed solely to improve readability and correct minor grammatical errors in the text. All AI-edited portions were reviewed, adjusted as necessary, and verified by the authors to ensure accuracy and adherence to the original meaning. No AI tools were used for data interpretation, analysis, or content generation.

3. Results

The peak UCS ($q_u$) results are summarized in

Table 4. Synthesis of average results of UC tests.

${\mathit{\rho }}_{\mathit{d}}$ (g/cm3)	PC (%)	7 Days		28 Days		90 Days
		${\mathit{q}}_{\mathit{u}}$ (kPa)	σ (kPa)	${\mathit{q}}_{\mathit{u}}$ (kPa)	σ (kPa)	${\mathit{q}}_{\mathit{u}}$ (kPa)	σ (kPa)
1.40	3	386	6	610	10	370	*
1.40	5	535	56	738	8	660	50
1.40	7	939	5	1173	38	780	40
1.46	3	603	3	705	15	390	10
1.46	5	818	29	1145	25	800	*
1.46	7	1130	27	1228	18	1185	225
1.52	3	610	30	678	43	275	15
1.52	5	1225	5	795	15	805	45
1.52	7	1210	80	913	13	960	90
1.60	3	1030	100	375	5	340	20
1.60	5	1355	135	1018	8	595	15
1.60	7	1670	70	1075	5	995	25

* Value could not be determined due to partial or complete specimen disintegration.

. These values represent the average and the standard deviation (σ) of the replicates of each treatment and curing time.

3.1. Cement Content (C) x UCS ($q_u$) Relationship

The relationship between cement content and UCS after 7 days of curing, presented in Figure 2a, shows that strength increases with the cement content for all dry unit masses tested. It is also evident that an increase in ${\rho }_D$ contributes to strength gains and higher angular coefficients. The relationship seen after 28 days, shown in Figure 2b, indicates an increase in UCS with curing time for dry unit masses of ${\rho }_D$ = 1.40 and 1.46 g/cm³. However, for ${\rho }_D$ = 1.52 and 1.60 g/cm³ there was a sharp decline in resistance. For these two dry unit masses, it appears that only the cement content is associated with the rise in UCS. Finally, the relationship seen after 90 days of curing, exhibited in Figure 2c, reveals a decline in resistance compared to the 7-day results. Compared to the 28-day results, the values for ${\rho }_D$ = 1.52 and 1.60 g/cm³ are similar. The R² of the equations increased with the curing time, with values as low as 0.65 after 7 days and values as high as 0.98 after 90 days.

3.2. Porosity/Cement Index ($\eta /C_{iv}$) x UCS ($q_u$) Relationship

The relationship between the porosity/cement index and UCS after 7 days of curing, as shown in Figure 3a, revealed that the increase in strength is associated with a reduction in porosity and an increase in cement content. The R² of the equation showed that it can reasonably predict the strength of the mixtures. However, the relationships seen after 28 and 90 days, shown in Figure 3b and Figure 3c, respectively, suggest that these key parameters do not correlate well with UCS for these curing periods. The R² of these equations shows that they are unsuitable for strength prediction after longer curing times.

3.3. Curing Time x UCS ($q_u$) Relationship

Curing time x UCS relationships were plotted to highlight the strength degradation of the samples over time. The graphs display individual specimens (markers) and the average values (lines) for each dry unit mass (${\rho }_D)$. The data series are grouped according to cement content (C).

When studying the relationship between curing time and UCS for dry unit masses of ${\rho }_D$ = 1.40 and 1.46 g/cm³, presented in Figure 4a,b, there is a noticeable increase in strength between 7 and 28 days of curing, followed by a sharp decrease in strength by 90 days. In contrast, for dry unit masses of ${\rho }_D$ = 1.52 and 1.60 g/cm³, shown in Figure 4c,d, a strength loss is already evident by 28 days of curing, and by 90 days it stabilizes or decreases even further.

3.4. X-Ray Fluorescence Analysis

The XRF quantification of the oxides in the sediments is presented in

Table 5. Oxide content in dredged sediments.

Oxide	SiO2	Al2O3	Fe2O3	K2O	CaO	SO3	TiO2	Other
Content (%)	51.84	19.78	16.63	4.18	2.40	2.07	1.75	1.33

. Our analysis revealed that more than 88% of the sediment is composed of silicon, aluminum, and iron oxides, which are generally the most common components of soils. It also identified potassium oxide, which is very common for estuarine sediments and is associated with the mineral illite; calcium oxide; and sulfate, which is associated with biogenic remains. The identification of sulfates is important because of their potential to react with calcium compounds, creating the expansive mineral ettringite.

3.5. Thermogravimetric Analysis

All tested samples lost a significant part of their mass (around 20%) within the temperature range of 20 to 100 °C, which likely corresponded to moisture loss. Figure 5a presents the thermogravimetric analysis results for a specimen cured for 7 days. In the TGA of ettringite, an endothermic peak is typically expected between 80 and 130 °C [47]. However, in the present study, the intense moisture loss observed in this range may have masked the dehydration of ettringite. Prominent peaks were detected around 500 °C and 650 °C, corresponding to the thermal decomposition of calcium-based compounds, such as C-S-H gels, calcium hydroxide, and calcite. The analysis of the 90-day sample, displayed in Figure 5b, reveals a similar trend, with peak formations occurring within the same temperature range.

3.6. X-Ray Diffractometry Analysis

The X-ray diffraction analysis of a sample cured for 7 days, presented in Figure 6a, revealed the presence of sulfates and ettringite in the soil minerals. Similarly, the analysis of the 90-day-cured sample, presented in Figure 6b, also revealed the same compounds.

4. Discussion

Based solely on the UCS results, it is evident that most of the mixtures cured for 7 days and treated with 5 to 7% cement achieved strength values ranging from 700 to 1500 kPa. This range aligns with the expected strength values of soil mixtures treated with 5 to 15% cement [26]. This outcome can be attributed to the use of type III Portland cement, which promotes early strength gains. Additionally, analyzing the porosity/cement content x UCS relationship revealed that an increase in dry density significantly enhanced the strength of the samples. This finding suggests that combining chemical stabilization with compaction yields higher UCS values than chemical stabilization alone [4].

As for the UCS results after 28 days of curing, it was observed that the strength of specimens with dry unit masses of 1.40 and 1.46 g/cm³ increased, whereas for dry unit masses of 1.52 and 1.60 g/cm³ there was a decrease in strength. Given that all specimens were subjected to identical curing, water immersion, and testing conditions, the observed strength depletion appears to be related to the dry unit mass. This hypothesis is further supported by the 90-day results, as no clear relationship between strength and dry unit mass was observed. The curing time x UCS relationships presented in Figure 4 draw attention to the sudden depletion in strength and the differences between different dry unit masses.

As shown in Figure 3b,c, the R² of these curing times demonstrate weak correlations between the porosity/cement index and UCS,; however, the R² of the relationship between the cement content and UCS (Figure 2b,c) yielded values close to 1. Furthermore, the magnitude of the strength values seen across different unit masses was comparable, indicating that the cement bonds continued to contribute to the strength of the samples. These observations suggest that particle interlocking may have been compromised due to the formation of expansive compounds. The strength improvement seen for those durations of curing were only explained by cementation alone and not by the degree of compaction.

The XRF analysis of the sediments’ composition indicated that sulfates were the soil component most likely to interfere with stabilization using cement and lime, as this combination led to a series of catastrophic failures in stabilized soils because of ettringite formation. The sulfate content in our tested sediments falls within the range identified in the literature as being capable of triggering ettringite formation, although with the tests we analyzed here, it was not possible to confirm the availability of these compounds for such reactions [25]. The Portland cement supplied portlandite and tricalcium aluminate, both precursors of ettringite in the presence of sulfate ions. The pre-treatment with hydrated lime enriched the mixture with hydroxyl and calcium ions. This could have potentially contributed to the formation of ettringite since calcium is one of its precursors and the hydroxyl solubilizes alumina from the clay, especially in the later stages of curing.

XRD confirmed the presence of ettringite, even after longer curing times, suggesting that it was responsible for the strength reduction observed during UC testing. In the 28-day samples, expansive mineral formation primarily affected only denser specimens with a lower porosity, as limited void space restricted ettringite growth without disrupting particle interlocking. By 90 days, however, sufficient time had elapsed for the ettringite crystals to grow indiscriminately, utilizing the components released by soil–lime reactions.

In contrast, the mixtures used in a previous study contained higher water contents (above the liquid limit), and their longest curing duration was 28 days [4]. These conditions likely prevented UCS reductions, even if ettringite formed, as their mixtures had larger void spaces that could accommodate crystal growth without disrupting the soil–cement matrix.

This study’s findings suggest that the stabilization of these sediments with Portland cement and other calcium-based compounds may face significant challenges. While the early stages of curing showed promising strength gains, and even longer curing durations retained notable strength improvements compared to untreated sediments, the long-term interactions between the soil–cement matrix and sulfates may eventually degrade the material’s strength, making it behave like unstabilized soil.

Nonetheless, alternative stabilization approaches offer potential solutions to these issues, such as using low-calcium stabilizers like Class F fly or sulfate-resistant Portland cement with a low tricalcium aluminate content. However, it is important to note that the clay’s minerals can also supply aluminate ions for ettringite formation over longer curing periods [48]. Non-commercial stabilizers like barium hydroxide or alkali-activated cements could be effective and free such constraints, though their large-scale application warrants further systematic investigations [25,49]. The addition of gypsum has proven effective in minimizing the release of Ca²⁺ ions in sulfate-bearing soils, attenuating the formation of ettringite in soil mixtures containing calcium-based compounds [50]. Additionally, encapsulation techniques such as ion encapsulation could be explored as a strategy for preventing late-stage ettringite formation. These techniques may offer promising avenues by which to mitigate expansive reactions and improve the long-term stability of treated sediments. Recent studies suggest that encapsulating ions could significantly reduce the risk of ettringite formation, thereby enhancing the durability of stabilized materials [51,52].

Moreover, future studies could focus on microstructural analyses to determine whether ettringite formation occurs uniformly across specimens or is concentrated at specific interfaces (e.g., between aggregates). Understanding these patterns could provide insights into optimizing stabilization methods to mitigate strength loss. Additionally, investigating the behavior of these materials under cyclic or periodic loading could reveal whether mechanical fatigue exacerbates ettringite-related degradation, offering valuable data for real-world geotechnical applications.

5. Conclusions

Throughout this study, it has been shown that dredged sediments from the Port of Rio Grande treated with type III Portland cement and hydrated lime exhibited significant improvements in their UCS with short curing periods. However, a decline in resistance was observed with longer curing periods, specifically 90 days, a curing period not commonly evaluated in similar studies. This longer-term analysis allowed for the identification of strength depletion mechanisms that may be overlooked in standard short-term investigations (of up to 28 days).

The XRF analysis revealed the presence of sulfate in the sediment. Given the well-documented challenges in stabilizing sulfate-bearing soils with calcium-based compounds, further investigations using TGA and XRD confirmed the formation of ettringite. This expansive mineral was identified as the primary contributor to the strength loss observed over longer curing times. The following key conclusions can be drawn from this study:

• The combination of chemical stabilizers with compaction shows potential as a technique for stabilizing sediments from the Port of Rio Grande.
• The presence of sulfates in these sediments poses a challenge to their stabilization using calcium-based compounds like Portland cement and lime, despite the promising results observed with shorter curing times.
• Low-calcium stabilizers should be considered for these sediments, with particular attention paid to the risk of ettringite formation over longer curing times.
• Extended curing age evaluations are essential, as strength loss mechanisms may only become apparent beyond the typical 28-day period.
• Non-calcium stabilizers are preferable for stabilizing sulfate-bearing soils. Their efficacy in these sediments and their practical application at larger scales warrant further investigation.
• The combination of advanced chemical analysis with mechanical performance testing offers a comprehensive approach to understanding long-term stabilization behavior, which is essential for practical engineering applications.

As a suggestion for future research, a detailed microstructural analysis could be conducted to investigate whether ettringite formation occurs uniformly across specimens or is concentrated at specific interfaces (e.g., between aggregates). This would help clarify the mechanisms of strength loss at work and contribute to the development of more effective stabilization techniques. Furthermore, investigations into the behavior of materials after periodic loading can be performed to clarify whether this affects their strength.