1 Introduction

Cement hydration is an exothermic reaction. This reaction is a complex physical and chemical process that contributes to the development of a rigid load-bearing cementitious matrix. Early microstructural development is essential for predicting the properties of cement-based matrixes. Several methods have been used to monitor and characterize the hydration of cementitious materials and explain their hardening mechanisms [1]. For instance, conduction calorimetry has been utilized to follow the hydration heat. The obtained data can contribute to the explanation of the hydration mechanism as well as the prediction of setting times and mechanical properties. Unfortunately, the calorimetric curve does not correspond to practical conditions. In this case, evolution represents a dynamic balance between the heat produced by the hydrating cement paste and the variable environmental conditions. Other widely used methods, such as thermogravimetric and differential thermal analysis (TG–DTA), a scanning electronic microscope (SEM), and X-ray diffraction (XRD), have been utilized to identify the hydration products. TG–DTA, SEM, and XRD are unable to provide continuous information about hydration, and results can be affected by sample preparation processes, i.e., cutting, polishing, or drying [2]. Alternative non-destructive techniques, including the measurement of electrical properties, have been commonly used in the monitoring of early-age hydration [3]. The method based on measurements of electrical resistivity is especially suitable for monitoring early-age hydration. It also provides continuous data acquisition [4].

Monitoring concrete temperature is crucial because it provides information that can be directly correlated with the performance of hardened matrixes. Throughout the service life of concrete structures, cement-based materials have frequently been exposed to extreme environmental conditions, e.g., subzero or extremely high temperatures, variable humidity, and exposure to carbon dioxide. These could result in changes to the physical, mechanical, chemical, or electrical properties. During casting and curing processes, the maximum concrete temperature and temperature gradient cannot exceed certain values in order to prevent crack formation and durability issues caused by, for example, delayed ettringite formation. It is commonly accepted that the maximum permissible temperature is set at 160°F (70 °C), while the maximum temperature gradient is 35°F (19 °C) [5]. The temperature profile can also be used to estimate the maturity index and the early-age compressive strength of concrete [6]. The temperature measurements in fresh and hardened concrete are conventionally done by using either internally embedded or externally attached temperature sensors, i.e., resistance thermometers, thermocouples, or thermistors. These sensors are materials incompatible with concrete because they are made from metal. They are prone to mechanical damage and often have a short life span. Self-sensing cementitious matrixes and cement-based sensors are expected to eliminate or limit the applications of the currently used sensors. Carbon-based materials, particularly carbon fiber (CF), carbon nanofiber (CNF), and carbon nanotube (CNT), have been used to produce sensing cementitious matrixes with the ability to respond to changes in strain, stress, humidity, or temperature. Monitoring systems based on such sensors have been utilized in various structures [7] to measure temperature development and to follow early-age hydration processes [8].

Chen et al. [9] studied the temperature-resistivity of carbon fiber-reinforced concrete (CFRC). The test concrete specimens containing 0.8 vol.% of CFs showed both positive temperature coefficient (PTC) and negative temperature coefficient (NTC) effects during temperature increases. Teomete [10] and Demircilioğlu et al.[11] reported a rapid increase in electrical resistivity at 243 °C for concrete containing 1 vol.% of brass-coated steel fibers and at 150 °C for concrete with 0.8 vol.% of brass fibers. This concrete could be utilized as a pre-alarm fire warning system. McCarter et al. [12] studied temperature sensitivity over the temperature range between 10 and 60 °C. A comparison was made between plain mortar and mortar containing 0.5 vol.% of CFs (3 and 6 mm in length). The results showed the mortar containing CFs was less sensitive to temperature sensitivity and required lower activation energies than conventional mortar. Wen et al. [13] observed that cement paste containing CNFs and silica fume performed well as a negative thermistor. The electrical resistivity decreased as the temperature increased from 1 to 45 °C. Tabatabai and Aljuboori [14] embedded concrete-based sensors without any additives or conductive filers into the concrete deck of the bridge. The surface temperature was varied between −11.5 and 14 °C and the electrical resistance was measured. The electrical resistance changed considerably as the temperature changed from −11.5 to 0 °C but was nearly stable between 0 and 14 °C. The output from the laboratory experiments indicated that the prototype sensors could detect the surface of the ice and frozen concrete. Another example from Mo et al. [15], the cube mortar sensor containing 0.7% CNFs (CNFA) embedded in the self-compacting concrete cylinder to study hydration monitoring. The electrical resistance changes were likely followed in the same pattern as cement hydration. These results imply that embedded CNFAs could be used to monitor the early age of concrete.

An alternative novel solution to produce sensitive sensors is the use of cement covered with synthesized in-situ CNFs (nanomodified Portland cement). Studies [16,17,18] showed that mortar made with this innovative cement was sensitive to stress, strain, and humidity. The research described in this paper aimed to determine the impact of temperature on the electrical resistivity of such sensors. In addition, the feasibility of using the nanomodified Portland cement sensors (SmartCem sensors) to monitor the hydration processes was examined and compared with regular sensors.

2 Materials and methods

2.1 Materials and mix proportion

Ordinary Portland cement (CEM I 42.5N) and nanomodified Portland cement (SmartCem) were used as binders. The SmartCem was produced by synthesizing carbon nanofibers (CNFs) directly on the surface of the cement particles through the Chemical Vapor Deposition (CVD) process. The synthesis details and analysis of SmartCem are described elsewhere [16, 17]. The SmartCem contained approximately 2.71 wt.% of CNFs. The morphology of grown CNFs is displayed in Fig. 1. The scanning electron microscope image shows that CNFs generally have a diameter ranging from 10 to 50 nm, while their length was visually estimated to be in the range of approximately 3 to around 20 μm. The CNFs exhibited a highly curled structure. Sand having a maximum particle size of 150 μm was used to produce mortars. The polycarboxylate-based superplasticizer (SP) type Glenium produced by Grace Chemical was used. Its amount was constant for all mixes and equaled 0.8 wt.% of the total binder content. All mixes had a water-to-binder (w/b) ratio of 0.35 and a sand-to-binder ratio of 1.0. Mix proportions used for producing the SmartCem sensors are shown in Table 1.

Fig. 1
figure 1

The morphology of grown CNFs on the nanomodified Portland cement

Full size image
Table 1 Mix proportions for SmartCem sensors
Full size table

The studied sensors were produced as mortar beams with dimensions of 12 mm × 12 mm × 60 mm, Fig. 2a. The mortar sensors were mixed using a Bredent-type vacuum mixer and cast into house-made Teflon molds. Several mortar beams were produced and contained 0 wt.% (Ref), 2 wt.% (S2), 4 wt.% (S4), 6 wt.% (S6), 8 wt.% (S8), and 10 wt.% (S10) of the SmartCem. Three mortar beams were prepared for each set of tests. Four copper electrodes were made of 0.25-mm-thick plates with a width of 5 mm and a height of 15 mm. The copper plates were immersed vertically, 7.5 mm deep, into the mortar specimen. The distance between the inner electrodes was 30 mm and between the outer electrodes 50 mm. Electrodes were connected by the electrical wires and covered in the upper part by heat-shrink tubes, Fig. 2b. Heat-shrink tubing provides electrical insulation, electrical insulation, mechanical protection, environmental sealing, and strain relief, which can prevent undesirable resistance accumulation at the electrodes.

Fig. 2
figure 2

a Mortar sensor b Mortar sensors are wrapped with a thin plastic sheet and covered by heat shrink tubes, and c A schematic representation of the experimental setup for the temperature-sensitive study of SmartCem sensors

Full size image

Two types of concrete were prepared to monitor hydration by using cement-based sensors. Both normal strength concrete (NC) and self-compacting concrete (SCC) are made from different types of cement and compositions for the purpose of comparing the performance of the SmartCem sensor in monitoring the hydration process. Different compositions can have distinct hydration characteristics, so this comparison helps evaluate the sensor’s versatility and reliability across a range of cement composites. An NC, which contained Portland cement CEM I 42.5N, micro-sand B15 (150 µm), sand size 0–4 mm, coarse aggregate with a maximum size of 8 mm, and a superplasticizer type Glenium. An SCC was produced and delivered by a local ready-mix concrete producer. The used mix composition contained Portland cement (CEM II/A-V 52.5N) from Cementa-Sweden, Dolomite-KM 200 filler, sand 0–4 mm, and crushed granite aggregates with a maximum particle size of 16 mm. MasterGlenium SKY 823 superplasticizer was added. The detailed mix proportions are shown in Table 2.

Table 2 The concrete mix composition
Full size table

2.2 Test methods

2.2.1 Temperature monitoring

The electrical properties of the nanomodified Portland cement sensors were measured at various temperatures to study their temperature-resistivity. After casting, the specimens were water-cured for 25 days and desiccated at 20 ± 2 °C for 3 days. Next, all specimens were sealed with a thin plastic sheet, as shown in Fig. 2b.

The produced SmartCem sensors were placed in the temperature-controlled chambers, which maintained constant temperatures of −20, 20, and 40 °C for 24 h. Later, the electrical resistance was measured using a four-probe method with a digital multimeter of type Keysight 34465A. The direct current was applied to the outer electrodes, and the potential difference was measured at the inner electrodes. The test setup is shown in Fig. 2c.

The temperature sensitivity of the SmartCem sensor was described by the temperature coefficient and the activation energy, which were calculated following Eqs. (1, 2 and 3). The electrical resistivity \(\rho_{T}\) of sensing cementitious materials at the desired temperature \(T\) was calculated according to Eq. 1 [19,20,21]:

$$ \rho_{0} = \frac{{\rho_{T} }}{{1 + \alpha (T - T_{0} )}} $$
(1)

where \(\alpha\) is a temperature coefficient and \(\rho_{0}\) is the electrical resistivity at the reference temperature \(T_{0}\). Another way to describe the temperature-resistivity behavior is to express the electrical resistivity through the motion of free electrons in the pore solution. The activation energy required for the ionic mobility in the liquid phase of a cement-based matrix can be calculated by applying the Arrhenius law [21,22,23], Eqs. 2, 3:

$$ \rho_{T} = Ae^{{ - \left( {\frac{{E_{{\text{a}}} }}{RT}} \right)}} $$
(2)
$$ \rho_{T} = \rho_{0} e^{{ - \left( {\frac{{E_{{\text{a}}} }}{R}\left( {\frac{1}{T} - \frac{1}{{T_{0} }}} \right)} \right)}} $$
(3)

where \(A\) is a pre-exponential constant, \(E_{{\text{a}}}\) is the activation energy for the conduction process (J/mol), and \(R\) is the gas constant (8.314 J mol−1 K−1)

2.2.2 Hydration monitoring

Changes in temperature, humidity, and moisture content were the main factors considered when monitoring the hydration process. The effects of these factors on the measured electrical properties of the SmartCem sensors were studied. The hydration monitoring of the normal-strength concrete (NC) was determined using reference sensors (Ref) and nanomodified Portland cement sensors containing 4 wt.% of the SmartCem.

The test sensors were embedded 50 mm below the surface in the 100 mm NC cube immediately after casting. A commercially available humidity sensor type SHT85 produced by Sensirion and a thermocouple were installed as references. The moisture content of the reference samples was measured every 24 h for 7 days using the oven-drying method and calculated using Eq. 4 [11]:

$$ {\text{Moisture}}\;(\% ) = \frac{{\left( {W - W_{{{\text{dry}}}} } \right)}}{{W_{{{\text{dry}}}} }} \times 100 $$
(4)

where W is the weight of the concrete sample (g), \(W_{{{\text{dry}}}}\) is the weight of the concrete sample after 24 h of oven-drying at 110 ± 5 °C (g).

The test setup and the location of the sensors are shown in Fig. 3. The resistivity of SmartCem sensors and readings from the reference humidity and temperature sensors were recorded for 7 days. All cube samples were covered by a plastic sheet on the top surface to prevent moisture from escaping into the surrounding environment and kept under laboratory conditions at 20 ± 2 °C and 60 ± 5% RH.

Fig. 3
figure 3

a Sensor placements, b arrangement of embedded sensors, and c the test setup for hydration monitoring of NC

Full size image

2.2.3 Monitoring of full-scale beam

SmartCem mortar sensors containing 4 wt.% of the nanomodified cement were embedded into a full-scale reinforced beam. The concrete beam had dimensions of 300 mm × 300 mm × 2500 mm, Fig. 4. The SmartCem sensor, Fig. 5a, was embedded 25 mm under the top surface in the mid-span, Fig. 5b, c. The thermocouple was also immersed near the SmartCem sensor. The monitoring was continued for 7 days.

Fig. 4
figure 4

Full-scale beam used for monitoring of hydration

Full size image
Fig. 5
figure 5

a SmartCem sensor b Location of the embedded SmartCem sensor and thermocouple c SCC beam with hydration monitoring using embedded SmartCem sensor

Full size image

3 Result and discussion

3.1 Temperature effects on the electrical resistivity

The temperature versus resistivity relation was determined for SmartCem sensors containing various amounts of the nanomodified Portland cement: Ref (0 wt.%), S2 (2 wt.%), S4 (4 wt.%), S6 (6 wt.%), S8 (8 wt.%), and S10 (10 wt.%). The change in electrical resistivity vs. the temperature determined with the reference thermocouple is presented in Fig. 6. The results showed a decreasing electrical resistivity with rising temperatures for all specimens.

Fig. 6
figure 6

Effects of temperature on the electrical resistivity of the SmartCem sensors

Full size image

The calculated sensitivity of all tested mortar sensors was very similar in the temperature range between −20 and 40 °C, Table 3. Increasing the concentration of nanomodified cement can improve electrical conductivity, which is beneficial for sensors meant to perceive variations in electrical attributes like temperature. The greater quantity of carbon nanofibers (CNFs) is thought to have created interconnected pathways for conducting electricity within the hardened binder matrix. This is likely due to a larger number of charge carriers that become active at higher temperatures. The sensor containing 10 wt.% of SmartCem exhibited the highest sensitivity to temperature. Its sensitivity was approximately 11.76% greater than that of the reference sample.

Table 3 Calculated sensitivity of the sensors over the temperature range −20 to 40 °C
Full size table

The activation energy quantifies the amount of thermal energy required for a reaction or transportation process to occur. Thus, it can be interpreted as the amount of thermal energy needed to facilitate the electrons jumping between CNFs distributed in the solidified binder matrix, the so-called electron hopping. The activation energy can be determined by multiplying the slope of the natural logarithm of resistivity and the reciprocal of temperature (Fig. 7) by the universal gas constant, see Eq. 3. The calculated activation energy values increased with an increased amount of the nanomodified Portland cement and, thus, with an increased amount of CNFs, Table 3. The presence of CNFs in the studied sensor could slightly increase the activation energy due to the hopping conduction between the cement matrix and CNFs [24]. However, the CNF amount below the percolation threshold leads to the insignificant effect of direct electronic conduction through the CNF network. The percolation threshold value for the studied mortars could be estimated based on Fig. 8. Independently of the temperature, the percolation threshold was estimated to equal ~ 5 wt.% of the nanomodified Portland cement, which corresponded to ~ 0.136 wt.% of CNFs.

Fig. 7
figure 7

Arrhenius plots for electrical resistivity versus the inverse of temperature for produced sensors

Full size image
Fig. 8
figure 8

Effects of temperature and amount of nanomodified Portland cement content on the measured electrical resistivity

Full size image

3.2 Effects of hydration on electrical resistivity

Monitoring of the early age of hydration for up to 7 days after casting was done using the NC cube samples, Fig. 3. One sample contained a SmartCem sensor (S4), a thermocouple, and a humidity sensor. In addition to the thermocouple and humidity sensors, the reference mortar sensor, which did not contain CNFs, was also installed for the reference NC sample. Variations in electrical resistivity recorded during the first 7 days are shown in Fig. 9. The reference mortar sensor showed a nearly linear increase in resistivity with time. On the contrary, the mortar sensor containing 4 wt.% of SmartCem showed a resistivity change that followed the kinetics of cement hydration.

Fig. 9
figure 9

The change in electrical resistivity of NC was measured by the Ref sensor and the SmartCem sensor for 7 days

Full size image

The correlations between the hydration temperature, humidity, and moisture content measured by normal sensors and the electrical resistivity measured by the mortar sensors were weak, Figs. 10, 11, and 12. An inverse dependency between the moisture content of NC and the electrical resistance change was observed in Fig. 12a. The increasing electrical resistivity of the NC measured by the reference sensor was related only to the decreasing internal moisture content caused by the cement hydration.

Fig. 10
figure 10

The change in electrical resistivity of NC was measured by a Ref sensor and b SmartCem sensor (S4) versus the hydration temperature of NC for 7 days

Full size image
Fig. 11
figure 11

The change in electrical resistivity of NC was measured by a Ref sensor b SmartCem sensor (S4) versus internal humidity of NC for 7 days

Full size image
Fig. 12
figure 12

The change in electrical resistivity of NC was measured by a Ref sensor b SmartCem sensor (S4) versus moisture content of NC for 7 days

Full size image

The temperature measured in NC cubes varied between 23 and 25 °C, and in this range, the temperature sensitivity of the SmartCem sensor was very low, Figs. 6, 10. This result indicated that the electrical resistivity during the early age of the hydration process is independent of the developed temperature. Generally, the electrical resistivity of the Portland cement-based matrixes is affected by the chemical composition of the pore solution, the connectivity of the pores, fraction of capillary porosity, and the volume fractions of the hydration products [25]. Thus, the observed initial rapid decrease of the resistivity could be related to the increasing concentration of calcium (Ca2+), sodium (Na+), potassium (K+), hydroxyl ions (OH−), and sulfate (SO42−) ions due to the ongoing dissolution of Portland cement particles [26]. Later, with progressing hydration processes, the electrical resistivity tended to increase. A continuous increase was observed for the reference mortar sensor during the entire measuring period. In the case of the SmartCem sensors, the maximum measured resistivity reached around 6000 Ω.cm, 48 h after mixing. At a later age, a nearly constant and slightly decreased resistivity was measured. The reduced number of ions present in the pore solution due to progressing hydration could be indicated as the main cause. Other factors could also include worsened pore connectivity [26,27,28]. The later, gradual decrease of the electrical resistivity observed after 120 h could be related to the densification of the binder matrix microstructure and facilitated electron jumping between CNFs.

The electrical resistivity of the hydrating cement matrix is affected by several factors, including, for example, humidity, ion concentration in the pore solution, or the curing temperature. Pure water is not a good conductor, as the present ions transport the electrical current. It can be increased with an increase in ion concentration. In the case of fresh concrete, the number of ions depends strongly on the amount and type forming the hydration phase, e.g., a rapidly decreasing amount of Ca2+ is directly linked to forming C-S–H and Portlandite. As such, this information could be used for more detailed monitoring of ongoing processes. However, based on the collected data, it is impossible to determine the exact effect of each of these factors.

The SmartCem sensors (S4) were used to monitor the hydration processes in the large-scale SCC beam, Figs. 4, 5. Generally, the electrical resistivity of the SCC beam measured by the embedded SmartCem sensor displays almost the same trend as the thermocouple. The recorded electrical resistivity of SmartCem sensors and the temperature measured by thermocouples are shown in Fig. 13.

Fig. 13
figure 13

The measured electrical resistivity of the SCC beam obtained by the SmartCem sensor and the hydration temperature evolution for 7 days

Full size image

In period I, (0–5 h), after embedding the SmartCem sensor into the SCC beam, the electrical resistivity decreased significantly. Rapidly released from Portland cement ions, i.e., K+, Na+, Ca2+, OH−, and SO42− dissolved in water and enhanced the electrical conductivity of the pore solution [1]. During the IInd period, the electrical resistivity tended to fluctuate. It could be related to the competitive balance between the ion dissolution and formation of the hydration products, especially ettringite and calcium hydroxide. The electrical resistivity increases because the hydration products are formed on the cement particles and cause a delay in the hydration process. At the same time, ions from cement continue to dissolve in the water and form an electrolytic solution, resulting in a decrease in the electrical resistivity of the SCC beam. The end of this period is typically characterized by the initial set. At a later age, during the IIIrd period known as acceleration, the formation of hydration products is enhanced, thus leading to increased electrical resistivity. This could also be related to the significantly decreased number of charged ions present in the pore solution [26]. The hydration temperature in the SCC beam increased rapidly to reach its maximum, mainly due to the hydration of the tricalcium silicate and the tricalcium aluminate. During the IVth period, the hydration enters a deceleration controlled by the diffusion of ions [26]. The hydration slows down, thus leading to the observed decrease in electrical resistivity. The hydration products tend to be thicker and larger, and the capillary porosity is significantly reduced, leading to decreased electrical resistivity. In period V, the hydration temperature of the SCC beam reached the ambient temperature and remained at that level. The hydration process gradually slowed down and resulted in a slight increase in the measured electrical resistivity.

4 Conclusions

The first part of the research focused on the effects of temperature on the electrical resistivity of SmartCem sensors. The second part, preliminary assessed the feasibility of using these sensors for monitoring the hydration processes. The main conclusions can be summed up as follows:

  • SmartCem sensors are capable of temperature sensing, similar to the negative thermistor, whose electrical resistivity decreases when the temperature is increased.

  • SmartCem sensors containing 10 wt.% of the nanomodified Portland cement (~ 0.271 wt.% of CNFs) showed the highest sensitivity, which was approximately 11.76% higher in comparison with the reference mortar sensor. The temperature coefficient of resistivity was −0.0152/ °C.

  • The percolation threshold (the amount of the nanomodified Portland cement) for temperature sensitivity was measured to be 5 wt.%, which corresponded to ~ 0.136 wt.% of CNFs.

  • The change in electrical resistivity occurring during hydration in an early period of the studied concrete could be sensed by the embedded SmartCem sensors.

  • The variations in electrical resistivity recorded by the SmartCem sensor could reflect physical and chemical changes occurring due to cement hydration.

  • The change in electrical resistivity has been less affected by the hydration temperature.