Hydration and Early Age Properties of Cement Pastes Modified with Cellulose Nanofibrils

Abstract

In this work, the effects of cellulose nanofibrils (CNFs) on workability, hydration reaction, microstructure, early age shrinkage, fracture properties, flexural strength, and compressive strength of cement paste were investigated. Six batches with variable CNF concentrations with the same water-to-cement (w/c) ratio (0.35) were tested. Flow table test showed a reduction in the workability as CNF dosage increased. Isothermal calorimetry (IC) tests showed that after 3 days, degree of hydration (DOH) improved up to 8% because of the addition of CNFs. Thermogravimetric analysis (TGA) tests at 7 and 28 days showed no significant changes in DOH for all pastes. After 7 days, mixture with 0.15% CNF resulted in up to 31% improvement in compressive strength. For 0.09% CNF addition, cement paste showed 26% increase in compressive strength after 28 days. Tests revealed that adding a small quantity of CNF (0.06%) along with entraining 0.05 extra water reduces autogenous shrinkage by 49% at a cement paste with w/c = 0.30. For interpreting the results, a tunnels, reservoirs, and bridges (TR&B) model is proposed. This model suggests that, as proposed by others, CNFs can modify microstructure by providing tunnels for transporting water to unhydrated cement grain. Because of their hydrophilicity, CNFs retain water and work as reservoirs (internal curing), which explains the improvement in properties at low w/c ratios. Significant increases in fracture energy (up to 60%) and flexural strength (up to 116%) suggest that CNFs are an effective toughening mechanism, acting as bridges that increase the energy required for crack propagation.

Nanotechnology has shown, and continues to show, tremendous promise for enhancing the short-term and long-term properties of cement-based composites. While investigations of nanotechnology applications to cement and concrete have been proceeding for some time, recent successes illustrate some profound effects that nanomaterials can have on concrete properties ( 1 – 4 ).

A class of nanomaterial that holds promise in concrete technology is that based on cellulose. Plant-based fibers have long played a role in concrete for their role in reinforcing as well as their role in early age properties ( 5 – 8 ). Cellulosic materials have the advantages of being abundant, low toxicity, and relatively low cost.

Cellulosic nanomaterials have gradually been making their way into concrete technology ( 9 ). The effect of cellulose microcrystalline particles (MCC) on the properties of cement-based composites was studied by Gomez Hoyos et al., who found adding MCC increases the degree of hydration (DOH) ( 10 ). Onuaguluchi et al. worked on the effects of cellulose nanofiber gel suspension on the properties of cement pastes, showing that DOH, flexural strength, and energy absorption improved in pastes with cellulose nanofiber ( 11 ). Cao et al. found that the addition of cellulose nanocrystals (CNCs) enhances the performance of cement paste ( 12 ). They showed that CNCs can improve DOH and the flexural properties of cement pastes. They further showed that adding CNC can increase the flexural strength of cement paste by up to 50% ( 13 ). Other researchers worked on the effects of CNCs on the microstructure of cement paste ( 14 , 15 ). Fu et al. used nine various CNCs for improving the hydration and flexural properties of Portland cement pastes ( 16 ). Hisseine et al. performed some investigations on cellulose filaments in cement-based composites. They found that adding cellulose filaments enhances the mechanical performance of cement systems and improves DOH ( 17 – 19 ).

In the work described here, attention is focused on the potential role of cellulose nanofibrils (CNFs). CNFs are hydrophilic nanoparticles (typically, 20–50 nm in width and less than 0.2 mm in length). They can be extracted from various plants, trees, and recycled paper stock. CNFs are usually branched or forked. They are new nanoscale material given their high specific surface (31–33 m²/g), high aspect ratio, low density (1.0 g/cc slurry), low toxicity, and low cost ($1.25/lb. = $2.67/kg slurry) that enable functionalization ( 20 ). Previous work showed that the addition of small quantities of CNFs and micro cellulose fibers improved fracture toughness of high-performance concrete by up to 50% ( 21 ). Haddad Kolour et al. showed that adding a small quantity of CNF can reduce free shrinkage by 13% and improve compressive strength by up to 28% ( 22 , 23 ).

Despite these preliminary successes, important questions remain about the specific role that CNFs play in cement hydration and the development of microstructure. Therefore, the objective of this work was to examine the changes in hydration brought on by small dosages of CNF to batches of cement paste, and to tie those changes to observed early age shrinkage and strength properties. To realize this objective, an array of laboratory techniques for measuring both reaction kinetics and resulting properties were employed. In this work, the focus was on low water-to-cement (w/c) ratio systems, as preliminary work indicated that these systems realize the greatest benefit from CNF additions ( 22 ).

The CNF materials used in this study were the product of the University of Maine Process Development Center. CNFs were extracted mechanically (not chemically/thermally) from northern bleached softwood Kraft pulp (spruce/fir). As-received CNF materials were in a white slurry form. The concentration of solids was 3.0% (3 g of nanofibrils and 97 g of water in 100 g of suspension), and the density of aqueous gel was 1.0 g/cm³ ( 20 ). Because of their high surface area and surface hydroxyl (OH^–) groups, CNFs are hydrophilic and ready for surface reactivity and interactions ( 10 ).

Methods

Specimen Preparation

The paste specimens were used to study workability, hydration kinetics, early age shrinkage, compressive strength, flexural strength, and fracture properties ( 23 ). Ordinary Portland cement (OPC) Type I/II (commercial grade) that complies with ASTM C150/C150M-17, Standard Specification for Portland Cement, was used in this study ( 24 ). The OPC contained 20.1% SiO₂, 62.0% CaO, 3.6% Al₂O₃, 3.0% Fe₂O₃, 3.4% SO₃, 3.3% MgO, 0.4% Na₂O, and 1.0% K₂O with a Blaine fineness of 378 m²/kg. The cement pastes were mixed with a conventional 8-qt rotary kitchen mixer. The following procedure was used for mixing the pastes: (1) the CNF suspension was mixed with water in the kitchen mixer for 180 s at high speed (homogenization/dispersion); (2) the solution from step 1 was combined with cement powder and mixed at low speed for 60 s; (3) the mix was allowed to rest for 15 s; a spatula was used to scrape the wall and bottom of the mixing bowl; (4) mixing for 240 s at high speed. Sufficient periods of mixing were used to improve CNF dispersion ( 6 ).

The CNF concentrations of each batch were calculated based on weight with respect to cement. Cement pastes were prepared at six different CNF concentrations. Test matrix for CNF-reinforced cement paste is shown in Table 1. Water demand for each batch was modified by subtracting the amount of water brought by the CNF-water suspension (3% solid nanofibrils and 97% water).

Table 1.

Test Matrix for Cellulose Nanofibril (CNF)-Reinforced Cement Pastes

Mix no.	Water/cement (%)	CNF/cement (% weight)	Cement (g)	Water (g)	CNF slurry (g)
1 (reference)	35	0	1,000	350.00	0
2	35	0.015	1,000	345.15	5
3	35	0.030	1,000	340.30	10
4	35	0.060	1,000	330.60	20
5	35	0.090	1,000	320.90	30
6	35	0.150	1,000	301.50	50

An experimental program was developed to measure the effects of CNF dosages on the following properties: workability, hydration kinetics, early age shrinkage, fracture properties, flexural strength, and compressive strength. The specific laboratory tests are detailed below.

Workability

For each batch, and immediately after finishing the mixing procedure, a flow table test following ASTM C1437-15, Standard Test Method for Flow of Hydraulic Cement Mortar, was conducted ( 25 ). The result in percentage is the flow of that batch.

Zeta Potential

In colloidal chemistry, the zeta potential (ζ) is the potential between the fixed layer of liquid adjacent to the solid phase and the liquid constituting the bulk liquid phase ( 26 ). This test quantifies the strength of the electrostatic attraction or repulsion between particles. For studying the affinity between particles in the fresh cement paste, the zeta potentials of the cement and CNF particles were measured ( 12 ). The CNF and cement particles were, respectively, diluted in deionized (DI) water or synthetic pore solution to a concentration of about 100 mg/l ( 27 ). Synthetic pore solution (0.35 M KOH + 0.05 M NaOH in DI water) was the same solution suggested by Rajabipour et al. ( 28 ). Since the zeta potential is dependent on pH values, the tests were performed in two different pH environments at the age of 1 h: a neutral environment with a pH of 7, and the simulated pore solution with a pH of 12.7 ( 12 , 27 ).

Cement Hydration

Isothermal Calorimetry

Heat flow rate and cumulative heat release were monitored in this test in an attempt to study the hydration reaction and the DOH of the cement pastes. Immediately after mixing, 19–24 g samples of the paste were transferred to a 20 ml glass ampoule, 1 in. (25.4 mm) in diameter and 2¼ in. (57.15 mm) in height. Measurement has been performed in a sealed isothermal calorimetry (IC) chamber (maintained at 23 ± 0.1°C). Before starting the data collection, the isothermal condition was held for 45 min to reach equilibration and then steady heat measurement was run for almost 3 days. Two replicate tests were performed for each of the six different CNF dosages.

Thermogravimetric Analysis

The thermogravimetric analysis (TGA) was performed to find the DOH of CNF-reinforced cement pastes at two different ages: 7 and 28 days. At later ages, the heat release rate from IC is so small and the measuring error could be too large. So TGA was run to obtain the DOH at larger ages ( 12 ). The TGA tests were also done for the individual materials of CNF and cement. After mixing the paste, two 1 by 1 in. (25.4 by 25.4 mm) plastic cylinder molds were used to prepare two specimens for each batch (one specimen for each age). Molds were covered by plastic films and kept inside zipped bags (sealed condition) for 24 h. Then, specimens were demolded, covered by plastic films, and cured inside other zipped bags (sealed condition) for 7 and 28 days. Two replicate tests were performed for each of the six different CNF dosages.

At the ages of testing, the paste specimens were removed from the zipped bags. Immediately after removal, hydration of the cement paste was stopped by solvent exchange technique. Cement specimens were submerged in a relatively large amount of solvent for 30–40 min. Water present in the pores of the cement paste is diluted and removed by solvent. In this research, ethanol was used as solvent ( 29 ).

After stopping the hydration, specimens were stored in a sealed condition, ready for TGA. At the day of testing, they were ground into powders with mortar and pestle while evaporation was minimum. Approximately 35 mg of powder was transferred into the TGA chamber for running the test. First, the specimen was held for 5 min to reach equilibration temperature in the chamber. At the second step, the specimen was heated to 1,000°C at a rate of 10°C/min. This step is done to remove all chemically bound water (CBW).

In TGA tests, the total mass of the chemically bound water (CBW) in the hardened cement pastes is calculated. CBW is the weight loss between 105°C and 1,000°C ( 30 ). At temperatures below 105°C, some relatively small amounts of CBW could be missed from some hydration products (particularly C-S-H, ettringite, and monosulfate) ( 29 ). As a simplifying assumption, this part of CBW has been ignored in this study. The decrease in mass observed between 400°C and 500°C shows the decomposition of Ca(OH)₂ ( 29 ). Almost all of the weight loss above 600°C is related to CO₂ loss because of different forms of CaCO₃ decomposition (decarbonation). This part of weight loss, because of CO₂ evaporation, was not considered as CBW.

CBW divided by the final weight of the material shows the mass of CBW per unit gram of unhydrated cement. The assumption for the maximum amount of CBW per unit gram of fully hydrated cement was 0.23 g ( 12 ). Then, DOH is the mass of CBW per unit gram of unhydrated cement divided by 0.23 g.

Compressive Strength

After mixing the paste, six 2 by 4 in. (50.8 by 101.6 mm) cylinder molds were used to mold specimens for each batch to obtain the compressive strength at two different ages: 7 and 28 days (three specimens for each age). Molds were covered by plastic films and kept inside zipped bags (sealed condition) for 24 h. Then, specimens were demolded, covered by plastic films and cured inside other zipped bags (sealed condition) for 7 and 28 days. ASTM C39/C39M-18 (AASHTO T22), Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens, was used for breaking the specimens and measuring the compressive strengths ( 31 ). Results are the average value of the compressive strengths of three specimens for each batch at that particular age.

Early Age Shrinkage

The positive effects of CNF on both hydration reaction (3 days) and compressive strength showed that the CNF is acting as an internal curing agent in a way similar to a super-absorbent polymer (SAP) ( 32 – 35 ). That is, some of the mix water is being captured and held by the CNF during mixing but is released later on as hydration progresses. Such a mechanism—water entraining—would explain both the DOH and strength improvements, as it would play a role in minimizing self-desiccation and autogenous shrinkage.

Autogenous deformation was measured using a dilatometer equipped with an automatic data-logger and a digital length gauge. The test was performed using the corrugated tube based on ASTM C1698-09, Standard Test Method for Autogenous Strain of Cement Paste and Mortar ( 36 ). The length change of corrugated polyethylene molds, approximately 29 mm in diameter and 420 mm in length, were monitored in this test. Corrugated molds should be sealed with two end plugs and filled with fresh paste. A tamping rod was used for consolidating the fresh paste along the length of the tube. The corrugated shape of the tubes makes them much stiffer in the radial direction than in the longitudinal direction. This higher stiffness in radial direction translates almost all of the volumetric deformation of the paste to linear deformation of the mold ( 6 ).

For examining the hypothesis of water entraining by CNF, pastes with and without CNF were compared in a series of autogenous shrinkage tests. The basic w/c ratio is 0.30 in all mixes. Mixes with CNF, additionally contain entrained water. An amount of entrained water equal to (w/c)_e = 0.05 was added to mixes with CNF in an attempt to decrease self-desiccation and autogenous shrinkage based on a method proposed by Jensen et al. ( 32 , 33 ). The changes in lengths of the two samples were recorded for each mix. Each test was conducted for 7 days in a thermostatically controlled room set at 23°C. Every 30 min, for the whole duration of the test, the lengths of the samples were continuously monitored. The time of the final set is zero point for all of the readings. Before that, the major deformation is deformation of the polyethylene corrugated tube. Also, it is known that before the final set, the paste is still plastic and the reading is actually related to chemical shrinkage ( 6 ). The time of final set for each mix was determined by the Vicat needle test ( 37 ).

Notched Beam

Notched beam tests were done after 28 days for all cement pastes (four specimens for each batch in Table 1), using a simplified hybrid fracture test that combines the measurements required for the two-parameter fracture model of Jenq and Shah and the measurements required for concrete fracture energy (RILEM 1985) ( 38 ). Molds were covered by plastic films and kept inside zipped bags (sealed condition) for 24 h. Then, the specimens were demolded, covered by plastic films and paper towel, and cured inside other zipped bags (sealed condition) for 28 days. In this simplified method, midspan displacement was not measured, so G_f was redefined as the area under the complete load-crack mouth opening displacement (CMOD) curve. By this method, it was possible to measure the modulus of elasticity (E in GPa) and the material fracture parameters, critical crack length (a_c in mm), and fracture energy G_f (area under Load-CMOD curve in N.mm). The span, depth, and width of the beams were 180 mm, 40 mm, and 30 mm, respectively. The beams had an initial notch of 13 mm (1/3 of depth), and the notch was created by a diamond wet saw. The beams were tested in an Instron closed-loop testing machine. The three-point bending test of notched beams was done in CMOD control mode, and the rate of loading was 0.015 mm/min. A clip gage, mounted on the notch, controls the CMOD opening. The test consisted of loading the beam until a crack happens. Then the specimen was unloaded and reloaded so that the crack grew steadily ( 21 , 38 ).

Flexural Strength (Three-Point Bending)

Flexural strength tests were done after 28 days for all cement pastes (four specimens for each batch in Table 1). The beam sizes and curing procedure were the same as described above for the notched beam tests. By this test, it was possible to measure the flexural strength (modulus of rupture) of all specimens. Three-point bending tests were performed in an Instron closed-loop testing machine. The test consisted of loading the beam until breaking the specimen. The test was displacement (deflection) control and the rate of loading was 0.5 mm/min.

Results

Workability

Table 2 shows the effect of CNF on the flow table results for different cement batches with various amounts of CNF. The results show that increasing CNF content leads to reduced workability. It can be seen that every 0.05% of additional CNF incorporated in the cement paste decreases the flow value by almost 15%. Results from the flow table tests suggest that water in CNF slurry is not immediately available to mix with the cement, as it is being held back by the CNF. Indeed, the loss of workability suggests that some of the added mix water is held back by the CNF. Possible CNFs agglomeration because of poor dispersion (especially at high dosages) likely also plays an important role in decreasing workability.

Table 2.

Flow Table Results (Water/Cement = 35%)

Mix no.	Cellulose nanofibril (CNF) dosage (% weight)	Flow (%)
1 (reference)	0	121
2	0.015	118
3	0.030	113
4	0.060	91
5	0.090	86
6	0.150	74

Zeta Potential

The zeta potentials for the CNF and cement particles at these two different pH environments are shown in Table 3. Results show that the absolute value of the zeta potential for CNF fibers is much higher than that for cement particles. This means that cement particles have much more potential for agglomeration in comparison with CNF fibers. At pH = 12.7, synthetic pore solution, the affinities between the particles have following order:

f (cement - CNF) > f (cement - cement) > f (CNF - CNF)

Table 3.

The Zeta Potentials of Cellulose Nanofibril (CNF) and Cement Particles

Environment	pH	Cement	CNF
DI water	7	–8.4 mV	–33.8 mV
Synthetic pore solution	12.7	+0.4 mV	–9.5 mV

Note: DI = deionized.

The zeta potential results show that the affinity between cement and CNF is stronger than that between cement particles. It means that CNF fibers tend to adhere onto cement particles rather than to agglomerate themselves. For CNFs-cement adhesion, well-dispersion of both cement particles and CNF fibers is a necessity. Relatively low zeta potentials of cement (+0.4 mV) and CNF (–9.5 mV) at synthetic pore solution show that there is a potential of agglomeration within the fresh cement paste for both CNFs and cement particles.

Cement Hydration

Isothermal Calorimetry

Results of the cumulative heat of hydration for the first 3 days (72 h) of the six cement pastes with different CNF dosages are shown in Figure 1. It can be observed that the cumulative heat showed increase in all CNF-modified pastes. This increment continues until the end of the test (3 days). All CNF dosages showed higher DOH than the plain paste. After 3 days, the CNF dosages of 0.09% and 0.15% resulted in 8% improvement in DOH, while for the pastes with 0.015%, 0.03%, and 0.06% CNF, changes in DOH are not significant. The increased DOH could result from internal curing effects and short circuit diffusion because of the CNF presence ( 12 ). Figure 2 displays hydration reaction heat flow curves for the first 40 h of the six mixtures described in Table 1.

Figure 1.

Cumulative heat of cellulose nanofibril (CNF)-reinforced cement pastes for the first 3 days.

Figure 2.

Heat flow curves of the cellulose nanofibril (CNF)-reinforced cement pastes for the first 40 h.

Thermogravimetric Analysis

Following calculations described in the TGA methods section, after 7 days, the DOH for the control specimen (without CNF) was 61. For cement pastes with 0.015%, 0.03%, 0.06%, 0.09%, and 0.15% of CNF, DOHs were 60, 60, 60, 61, and 60, respectively. At 28 days, the DOH for control specimen was 64. For cement pastes with 0.015%, 0.03%, 0.06%, 0.09%, and 0.15% of CNF, DOHs were 64, 64, 64, 64, and 67, respectively. These results show that incorporating CNF does not change DOH significantly at the ages of 7 and 28 days. Figure 3 shows TGA results at the age of 28 days.

Figure 3.

28-day thermogravimetric analysis (TGA) results.

Compressive Strength

Results of compressive strength tests for cement pastes in Table 1 can be seen in Figure 4 for the age 7 and 28 days. After 7 days, the CNF dosages of 0.015%, 0.03%, 0.06%, 0.09%, and 0.15% resulted 20%, 3%, 18%, 28%, and 31% improvement in compressive strength, respectively. For the same CNF contents and after 28 days, compressive strengths are 2%, 23%, 12%, 26%, and 16% higher in comparison with the plain system.

Figure 4.

Cement pastes compressive strength results.

Early Age Shrinkage

Figure 5 shows the results of the autogenous deformation of cement paste specimens with and without CNF. It can be observed that adding 0.06% and 0.09% of CNF reduces autogenous shrinkage up to 49% and 26%, respectively. Reduction in internal curing effects of CNF at 0.09% could be because of potential CNFs agglomeration at higher dosage.

Figure 5.

Autogenous shrinkage versus time for pastes.

It is interesting to consider the expansion right after set in the mixtures with CNF. It happens because of water released by CNF, because of a drop in internal relative humidity inside the sealed (closed) cement paste. This phenomenon has already been observed in the cement system using SAP as an internal curing agent ( 33 ).

Compressive strengths of these mixtures have been depicted in Table 4. It shows that after 7 days, the CNF dosages of 0.06% and 0.09% resulted in 5% and 15% improvement in compressive strength, respectively. After 28 days, results for 0.06% addition of CNF are almost the same as for the plain system. At the same time, for the specimen with 0.09% CNF, compressive strength is 12% higher.

Table 4.

Compressive Strength Results for Autogenous Shrinkage Mixtures

Cellulose nanofibril (CNF) weight (%)	Compressive strength (MPa)	Compressive strength (MPa)
Cellulose nanofibril (CNF) weight (%)	Age = 7 days	Age = 28 days
0	44.8 ± 4.1	53.4 ± 5.3
0.06	47.2 ± 2.1	53.3 ± 3.9
0.09	51.5 ± 1.5	60.0 ± 3.8

Notched Beam

Table 5 shows the measured modulus of elasticity (E), the critical crack length (a_c), and fracture energy (G_f). Small changes in critical crack length show that CNFs are not very important in preventing crack initiation because they may not always be available at the crack tip. After 28 days, the CNF dosages of 0.015%, 0.03%, 0.06%, 0.09%, and 0.15% resulted in 0%, 17%, 11%, 24%, and 60% improvement in fracture energy, respectively. Significant changes in fracture energy suggest that CNFs are effective in preventing crack growth and the CNFs play an important role when the cracks try to propagate within the notched beam specimen ( 21 ).

Table 5.

Notched Beam Tests Results

Mix no.	Cellulose nanofibril (CNF) dosage (% weight)	Modulus of elasticity (GPa)	Critical crack length (mm)	Fracture energy (N.mm)
1 (reference)	0	16.1 ± 1.3	14.0 ± 1.4	6.46 ± 0.29
2	0.015	16.1 ± 1.9	14.0 ± 1.6	6.32 ± 0.48
3	0.030	17.2 ± 1.7	14.5 ± 1.7	7.57 ± 1.04
4	0.060	16.6 ± 1.0	13.8 ± 0.4	7.15 ± 0.65
5	0.090	17.4 ± 1.5	13.6 ± 1.4	8.03 ± 0.98
6	0.150	16.8 ± 0.9	12.5 ± 1.1	10.36 ± 1.16

Flexural Strength (Three-Point Bending)

Results for pastes in Table 1 can be seen in Table 6. After 28 days, the CNF dosages of 0.015%, 0.03%, 0.06%, 0.09%, and 0.15% showed 44%, 39%, 55%, 72%, and 116% improvement, respectively. Significant changes in flexural strength suggest that CNFs are effective in improving the modulus of rupture in cement pastes.

Table 6.

Flexural Strength Tests Results

Mix no.	Cellulose nanofibril (CNF) dosage (% weight)	Flexural strength (MPa)
1 (reference)	0	1.8 ± 0.3
2	0.015	2.6 ± 0.1
3	0.030	2.5 ± 0.5
4	0.060	2.8 ± 0.2
5	0.090	3.1 ± 0.3
6	0.150	3.9 ± 0.3

Discussion

Because of their high surface area and surface hydroxyl (OH^–) groups, CNFs are ready for surface reactivity and interactions ( 10 ). Water which contains hydrogen have the potential to make hydrogen bonds with surface hydroxyl (OH^–) groups of CNFs. Because of these potential hydrogen bonds, CNFs are hydrophilic and can capture some water (depending on CNF dispersion/agglomeration condition) at early ages and release it at later ages.

The improvement in 7- and 28-day compressive strength observed in all mixes can be explained once again by an effect of internal curing, which suggests CNFs retain water at early ages and release it later. Tests revealed that, like SAPs, incorporating CNFs helps to entrain extra water and reduces autogenous shrinkage. At the same time, compressive strength was enhanced in those specimens. However, looking at Figure 4 shows that this improvement is not systematic with respect to CNF dosage. The possible interpretation of these tests is that flaw sizes in the cement pastes could be increased by the potential CNF agglomerating. At low dosages, the influence of this potential agglomeration is small, but it likely plays a more important role at high dosages.

The CNF system appear to work with a mechanism that is here named as the “tunnels, reservoirs, and bridges” (TR&B) mechanism ( 23 ). The tunnels mechanism described here is based on Cao et al.’s hypothesis ( 12 ). Zeta potential results showed that CNFs tend to adhere onto cement particles. Porous space around these adhered CNFs will work as “tunnels.” It means that, similar to CNCs, CNFs provide tunnels (channels) for transporting water to unhydrated cement grain which in turn help to produce more hydration products and enhance total hydration reaction ( 12 ).

CNFs are hydrophilic and can retain some water (depending on the CNF dispersion/agglomeration situation) and, like SAPs, work as “reservoirs.” This stored water is released at later ages and works as a supplementary source of water. Effect of this supplementary water can be seen in reducing autogenous shrinkage in Figure 5.

Significant changes (60% increase) in fracture energy (Table 5) suggest that CNFs are effective in preventing crack growth, and the CNFs work as “bridges” and play an important role when the cracks try to propagate within the notched beam specimen. Figure 6 shows the schematic of the TR&B model based on the above discussions.

Figure 6.

Schematic of tunnels, reservoirs, and bridges (TR&B) model.

Conclusion

This project examined how the addition of CNFs modifies the performance of cement paste and mortar. Flow table tests on cement pastes showed that increasing CNF dosages decrease workability. Based on these results, it is possible to work on using CNFs as viscosity modifying agents, with or without other admixtures (e.g., in self-compacting concretes). IC tests disclosed that incorporating CNF improved DOH after 3 days. TGA tests at the ages of 7 and 28 days showed no significant changes in DOH because of adding CNF. After 7 and 28 days, CNF-reinforced specimens showed higher compressive strength. A new TR&B model has been proposed for interpreting the results.

Based on CNF dosage, and dispersion efficiency, the effects of CNF on the cement-based systems can vary. High dosages of CNF increase the likelihood of agglomeration and covering the surface of cement particles and, as a result, slowing down the hydration reaction at early ages ( 12 ). Good dispersion optimizes the CNF dosage, while poor dispersion results agglomeration and dysfunction. Agglomerated CNFs work as flaws in the system. At low dosages, this effect is negligible, but at high dosages these agglomerations (flaws) can deteriorate the microstructure and mechanical properties of cement paste. CNF source, cement type, mixing, and curing procedures are the other important factors that can affect the test results.

Understanding the status of water during mixing and hydration allows better control of both the effects on workability and the effects on microstructure. As a future research, studying the water kinetics of CNFs during the mixing procedure and cement hydration is highly recommended.

Footnotes

Acknowledgements

The authors gratefully acknowledge the Public-Private Partnership for Nanotechnology (P3Nano) and the USDA Forest Service for their support of this work. Also, the authors would like to acknowledge the contributions of both the University of Maine Process Development Center, who provided the CNF for this study, and Parivash Takasi, who assisted with laboratory work.

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: H. Haddad Kolour, W. Ashraf, E. N. Landis; data collection: H. Haddad Kolour; analysis and interpretation of results: H. Haddad Kolour, W. Ashraf, E. N. Landis; draft manuscript preparation: H. Haddad Kolour, W. Ashraf, E. N. Landis. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Zhu

Bartos

P. J. M.

Porro

Application of Nanotechnology in Construction: Summary of a State-of-the-Art Report. Materials and Structures, Vol. 37, No. 9, 2004, pp. 649–658. https://doi.org/10.1007/BF02483294.

Sanchez

Sobolev

Nanotechnology in Concrete - A Review. Construction and Building Materials, Vol. 24, No. 11, 2010, pp. 2060–2071. https://doi.org/10.1016/j.conbuildmat.2010.03.014.

ACI 241R-17. Report on Application of Nanotechnology and Nanomaterials in Concrete: ACI Committees 236 and 241. American Concrete Institute, MI, 2017.

Shah

S. P.

Hou

Konsta-Gdoutos

M. S.

Nano-Modification of Cementitious Material: Toward a Stronger and Durable Concrete. Journal of Sustainable Cement-Based Materials, Vol. 5, No. 1–2, 2016, pp. 1–22. https://doi.org/10.1080/21650373.2015.1086286.

Onuaguluchi

Banthia

Plant-Based Natural Fibre Reinforced Cement Composites: A Review. Cement and Concrete Composites, Vol. 68, 2016, pp. 96–108. https://doi.org/10.1016/j.cemconcomp.2016.02.014.

Kawashima

Shah

S. P.

Early-Age Autogenous and Drying Shrinkage Behavior of Cellulose Fiber-Reinforced Cementitious Materials. Cement and Concrete Composites, Vol. 33, No. 2, 2011, pp. 201–208. https://doi.org/10.1016/j.cemconcomp.2010.10.018.

Mezencevova

Garas

Nanko

Kurtis

K. E.

Influence of Thermomechanical Pulp Fiber Compositions on Internal Curing of Cementitious Materials. Journal of Materials in Civil Engineering, Vol. 24, No. 8, 2012, pp. 970–975.

Jongvisuttisun

Negrello

Kurtis

K. E.

Effect of Processing Variables on Efficiency of Eucalyptus Pulps for Internal Curing. Cement and Concrete Composites, Vol. 37, 2013, pp. 126–135. https://doi.org/10.1016/j.cemconcomp.2012.11.006.

Balea

Fuente

Blanco

Negro

Nanocelluloses: Natural-Based Materials for Fiber-Reinforced Cement Composites. A Critical Review. Polymers (Basel), Vol. 11, No. 3, 2019, p. 518.

10.

Hoyos

C. G.

Cristia

Vázquez

Effect of Cellulose Microcrystalline Particles on Properties of Cement Based Composites. Materials & Design, Vol. 51, 2013, pp. 810–818.

11.

Onuaguluchi

Panesar

D. K.

Sain

Properties of Nanofibre Reinforced Cement Composites. Construction and Building Materials, Vol. 63, 2014, pp. 119–124. https://doi.org/10.1016/j.conbuildmat.2014.04.072.

12.

Cao

Zavaterri

Youngblood

Moon

Weiss

The Influence of Cellulose Nanocrystal Additions on the Performance of Cement Paste. Cement and Concrete Composites, Vol. 56, 2015, pp. 73–83. https://doi.org/10.1016/j.cemconcomp.2014.11.008.

13.

Cao

Zavattieri

Youngblood

Moon

Weiss

The Relationship between Cellulose Nanocrystal Dispersion and Strength. Construction and Building Materials, Vol. 119, 2016, pp. 71–79. https://doi.org/10.1016/j.conbuildmat.2016.03.077.

14.

Cao

Tian

Bahr

Zavattieri

P. D.

Youngblood

Moon

R. J.

Weiss

The Influence of Cellulose Nanocrystals on the Microstructure of Cement Paste. Cement and Concrete Composites, Vol. 74, 2016, pp. 164–173. https://doi.org/10.1016/j.cemconcomp.2016.09.008.

15.

Flores

Kamali

Ghahremaninezhad

An Investigation into the Properties and Microstructure of Cement Mixtures Modified with Cellulose Nanocrystal. Materials (Basel), Vol. 10, No. 5, 2017, p. 498.

16.

Montes

Suraneni

Youngblood

Weiss

The Influence of Cellulose Nanocrystals on the Hydration and Flexural Strength of Portland Cement Pastes. Polymers, Vol. 9, No. 9, 2017, p. 424.

17.

Hisseine

O. A.

Basic

Omran

A. F.

Tagnit-Hamou

Feasibility of using Cellulose Filaments as a Viscosity Modifying Agent in Self-Consolidating Concrete. Cement and Concrete Composites, Vol. 94, 2018, pp. 327–340. https://doi.org/10.1016/j.cemconcomp.2018.09.009.

18.

Hisseine

O. A.

Omran

A. F.

Tagnit-Hamou

Influence of Cellulose Filaments on Cement Paste and Concrete. Journal of Materials in Civil Engineering, Vol. 30, No. 6, 2018, p. 04018109.

19.

Hisseine

O. A.

Wilson

Sorelli

Tolnai

Tagnit-Hamou

Nanocellulose for Improved Concrete Performance: A Macro-to-Micro Investigation for Disclosing the Effects of Cellulose Filaments on Strength of Cement Systems. Construction and Building Materials, Vol. 206, 2019, pp. 84–96. https://doi.org/10.1016/j.conbuildmat.2019.02.042.

20.

UMaine Process Development Center. Cellulose Nano-Fibrils (CNF) Product Specification Sheet. University of Maine, 2019, p. 2. https://umaine.edu/pdc/wp-content/uploads/sites/398/2016/03/Specs-CNF.pdf.

21.

Peters

S. J.

Rushing

T. S.

Landis

E. N.

Cummins

T. K.

Nanocellulose and Microcellulose Fibers for Concrete. Transportation Research Record: Journal of the Transportation Research Board, 2010. 2142: 25–28.

22.

Haddad Kolour

Ahmed

Alyaseen

Landis

E. N.

An Investigation on the Effects of Cellulose Nanofibrils on the Performance of Cement Paste and Concrete. Advances in Civil Engineering Materials, Vol. 7, No. 1, 2018, pp. 463–478. https://doi.org/10.1520/ACEM20180048.

23.

Haddad Kolour

An Investigation on the Effects of Cellulose Nanofibrils on the Performance of Cement Based Composites. University of Maine, 2019. https://digitalcommons.library.umaine.edu/etd/3128.

24.

ASTM C150/C150M: Standard Specification for Portland Cement . ASTM International, American Society for Testing and Materials, West Conshohocken, PA, Vol. 552, 2019, pp. 1–9.

25.

ASTM C1437: Standard Test Method for Flow of Hydraulic Cement Mortar . ASTM International, American Society for Testing and Materials, West Conshohocken, PA, 2015, pp. 12–15.

26.

Nägele

The Zeta-Potential of Cement. Cement and Concrete Research, Vol. 15, No. 3, 1985, pp. 453–462.

27.

Nägele

The Zeta-Potential of Cement: Part II: Effect of pH-Value. Cement and Concrete Research, Vol. 16, No. 6, 1986, pp. 853–863.

28.

Rajabipour

Sant

Weiss

Interactions between Shrinkage Reducing Admixtures (SRA) and Cement Paste’s Pore Solution. Cement and Concrete Research, Vol. 38, No. 5, 2008, pp. 606–615.

29.

Scrivener

Snellings

Lothenbach

A Practical Guide to Microstructural Analysis of Cementitious Materials. CRC Press, Boca Raton, FL, 2016, p. 540.

30.

Mouret

Bascoul

Escadeillas

Study of the Degree of Hydration of Concrete by Means of Image Analysis and Chemically Bound Water. Advanced Cement Based Materials, Vol. 6, No. 3–4, 1997, pp. 109–115.

31.

ASTM C39/C39M: Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens . ASTM International, American Society for Testing and Materials, West Conshohocken, PA, 2018, pp. 1–7.

32.

Jensen

O. M.

Hansen

P. F.

Water-Entrained Cement-Based Materials: I. Principles and Theoretical Background. Cement and Concrete Research, Vol. 31, No. 4, 2001, pp. 647–654.

33.

Jensen

O. M.

Hansen

P. F.

Water-Entrained Cement-Based Materials: II. Experimental Observations. Cement and Concrete Research, Vol. 32, No. 6, 2002, pp. 973–978.

34.

Mechtcherine

Reinhardt

H. W.

Application of Superabsorbent Polymers (SAP) in Concrete Construction: State of the Art Report Prepared by RILEM Technical Committee 225-SAP. Springer, New York, 2012.

35.

Snoeck

Pel

De Belie

The Water Kinetics of Superabsorbent Polymers during Cement Hydration and Internal Curing Visualized and Studied by NMR. Scientific Reports, Vol. 7, No. 1, 2017, p. 9514. https://doi.org/10.1038/s41598-017-10306-0.

36.

ASTM C1698: Standard Test Method for Autogenous Strain of Cement Paste and Mortar . ASTM International, American Society for Testing and Materials, West Conshohocken, PA, 2009, pp. 1–8.

37.

ASTM C191: Standard Test Method for Time of Setting of Hydraulic Cement by Vicat Needle . ASTM International, American Society for Testing and Materials, West Conshohocken, PA, Vol. 7, 2018, pp. 1–8.

38.

Shah

S. P.

Swartz

S. E.

Ouyang

Fracture Mechanics of Concrete: Applications of Fracture Mechanics to Concrete, Rock and Other Quasi-Brittle Materials. John Wiley & Sons, New York, 1995, p. 588.