Volume Stability of Cement Paste Containing Limestone Fines

: The common cause of cracking in cement paste is shrinkage due to different reasons, such as loss of water and chemical reactions. Incorporating limestone ﬁnes (LF) as a cement replacement can affect the shrinkage of the paste. To examine this effect, ﬁve paste mixes were prepared with 0, 5, 10, 15 and 20% LF as a cement replacement and with a water-to-binder ratio (w/b) of 0.45. Four volume stability tests were conducted for each paste: chemical, autogenous and drying shrinkage and expansion. Chemical shrinkage was tested each hour for the ﬁrst 24 h and thereafter every 2 days for a total period of 90 days. The drying shrinkage, autogenous shrinkage and expansion were monitored every 2 days until 90 days. The results showed that replacing 15% LF enhanced the chemical shrinkage of the paste. However, autogenous shrinkage of the paste was found to increase between 0 and 10% LF and decline sharply at 15 and 20% LF. Drying shrinkage was found to increase with the increase in LF content. Expansion exhibited little variation between 0 and 10% LF and an increase for replacement above 15% LF. These results are discussed in terms of the formation of hydration products and self-desiccation due to hydration.


Introduction
Concrete's tendency to crack is historically recognized as inevitable. Therefore, most current concrete buildings tend to be affected by the formation of cracks. In order to reduce cracking and improve building materials, the concrete shrinkage during construction needs to be controlled, as this will have an impact on the concrete's durability. There are different types of shrinkage responsible for these cracks such as chemical, autogenous and drying shrinkage. Drying shrinkage is defined as the contracting of a hardened concrete caused by the removal of free water from the concrete, which is caused by the changes in relative humidity and temperature [1]. This can be controlled by using concrete with a lower w/c ratio or by adding supplementary cementitious materials (SCMs) that lead to its reduction [2]. Chemical shrinkage is well-defined as the internal volume change of cement paste resulting from the hydration of cementitious materials [3]. This is affected by the temperature and cement composition. Traditionally, chemical shrinkage was tested by three main measurement methods such as dilatometry, pycnometry and gravimetry. Many authors prefer using the dilatometry method to monitor chemical shrinkage [4][5][6][7][8][9][10][11][12][13][14]. The dilatometry method is based on measuring a drop of water in a hydrating cement paste through a graduated pipette [15]. Autogenous shrinkage is the macroscopic reduction in the external dimensions of cement particles during curing without any loss of moisture that occurs simultaneously with chemical shrinkage [16]. The cement content and temperature are the major factors affecting this type of shrinkage. These are controlled by the change in internal curing of the system due to self-dehydration at lower water-to-cement ratios (w/c < 0.4) [17]. Expansion of concrete occurs if exposed to water or an excessive amount of moisture. The hydration process is liable for water withdrawal from the capillary pores that are present within the concrete. Through the first hours of hydration, the capillary activities can be considered negligible. Therefore, the autogenous shrinkage is mainly

Materials
Portland cement PA-L 42.5 N type I was used in this study. This type of cement contained 0.5% free limestone. The incorporated raw limestone was extracted from Lebanon quarries. This limestone was first ground and then sieved until reaching a fines particle size below 300 µm. The particle size distributions of the limestone and cement are illustrated in Figure 1. The limestone density was 2.7 g/cm 3 . The chemical properties for the cement and limestone are shown in Table 1. The physical properties of the cement are presented in Table 2.

Mix Proportions
During this experiment, LF with 5 replacement levels of cement (0, 5, 10, 15 and 20%) were incorporated in paste specimens with a w/b ratio of 0.45. Five paste mixes were employed. The total number of specimens tested was 40, with 10 specimens for each type of shrinkage (chemical, autogenous and drying shrinkage and expansion). The details of the paste mix ratio are presented in Table 3.

Chemical Shrinkage
A chemical shrinkage test was performed according to ASTM C 1608 [46]. The chemical shrinkage specimens and test set-up are shown in Figure 2a,b, respectively. The cement and LF were dry mixed for 2 min and then poured with water. After that, 30 gm of paste was placed in a 250-mL glass bottle for each mix, which corresponded to a depth of

Mix Proportions
During this experiment, LF with 5 replacement levels of cement (0, 5, 10, 15 and 20%) were incorporated in paste specimens with a w/b ratio of 0.45. Five paste mixes were employed. The total number of specimens tested was 40, with 10 specimens for each type of shrinkage (chemical, autogenous and drying shrinkage and expansion). The details of the paste mix ratio are presented in Table 3.  [46]. The chemical shrinkage specimens and test set-up are shown in Figure 2a,b, respectively. The cement and LF were dry mixed for 2 min and then poured with water. After that, 30 gm of paste was placed in a 250-mL glass bottle for each mix, which corresponded to a depth of around 1.8 cm. Subsequently, a 2-mL graduated pipette was inserted through a hole in a rubber stopper and filled with water. This level of water was considered the chemical shrinkage initial value (i.e., zero). At the end, a drop of oil was added to prevent any water evaporation. Two replicate specimens were tested for each mix. One bottle filled with water only playing the role of a control specimen was used to monitor any drop in the water level due to several environmental conditions (e.g., temperature variation, stopper relaxation or water absorption by the stopper or the bottle). The average value of the two readings represented the chemical shrinkage. During the first 24 h, the drop of water was monitored every 1 h and then once every 2 days until a total period of 90 days. The volume change (chemical shrinkage) was transformed to length change using the following equation [19]: Buildings 2021, 11, x FOR PEER REVIEW 4 of 18 around 1.8 cm. Subsequently, a 2-mL graduated pipette was inserted through a hole in a rubber stopper and filled with water. This level of water was considered the chemical shrinkage initial value (i.e., zero). At the end, a drop of oil was added to prevent any water evaporation. Two replicate specimens were tested for each mix. One bottle filled with water only playing the role of a control specimen was used to monitor any drop in the water level due to several environmental conditions (e.g., temperature variation, stopper relaxation or water absorption by the stopper or the bottle). The average value of the two readings represented the chemical shrinkage. During the first 24 h, the drop of water was monitored every 1 h and then once every 2 days until a total period of 90 days. The volume change (chemical shrinkage) was transformed to length change using the following equation [19]: The results obtained were expressed in micrometers/meter in order to be compared with autogenous and drying shrinkage and expansion.

Autogenous Shrinkage
Autogenous shrinkage specimens were prepared using a Hobart mixer and tested according to ASTM C192 [47]. The specimens were cast in steel molds with dimensions of 25 mm × 25 mm × 300 mm and kept for 24 h. After demolding, two demec points at 200 mm on two sides of the specimens were applied. The length change was recorded using a dial gauge ( Figure 3). The specimens were then placed in plastic bags ( Figure 4). The lengths of the specimens were recorded every 2 days for a total period of 90 days. The length changes reported were the average values of four readings from two specimens for each mix. The results obtained were expressed in micrometers/meter in order to be compared with autogenous and drying shrinkage and expansion.

Autogenous Shrinkage
Autogenous shrinkage specimens were prepared using a Hobart mixer and tested according to ASTM C192 [47]. The specimens were cast in steel molds with dimensions of 25 mm × 25 mm × 300 mm and kept for 24 h. After demolding, two demec points at 200 mm on two sides of the specimens were applied. The length change was recorded using a dial gauge ( Figure 3). The specimens were then placed in plastic bags ( Figure 4). The lengths of the specimens were recorded every 2 days for a total period of 90 days. The length changes reported were the average values of four readings from two specimens for each mix.

Drying Shrinkage
The drying shrinkage was tested according to ASTM C157 [48]. The specimens were cast in 25 mm × 25 mm × 300 mm steel molds. After 24 h, the specimens were placed in air at an average temperature of 25 °C as shown in Figure 4. The procedure for taking the readings and frequency was similar to that for the autogenous shrinkage specimens.

Expansion
The paste specimens were cast in 25 mm × 25 mm × 300 mm steel molds. Two specimens for each mix were prepared to measure expansion. After 24 h, the specimens were immersed in water at a constant temperature (20 ± 1 °C) as shown in Figure 5. The procedure for taking the readings and frequency was similar to that for the autogenous shrinkage specimens.

Drying Shrinkage
The drying shrinkage was tested according to ASTM C157 [48]. The specimens were cast in 25 mm × 25 mm × 300 mm steel molds. After 24 h, the specimens were placed in air at an average temperature of 25 °C as shown in Figure 4. The procedure for taking the readings and frequency was similar to that for the autogenous shrinkage specimens.

Expansion
The paste specimens were cast in 25 mm × 25 mm × 300 mm steel molds. Two specimens for each mix were prepared to measure expansion. After 24 h, the specimens were immersed in water at a constant temperature (20 ± 1 °C) as shown in Figure 5. The procedure for taking the readings and frequency was similar to that for the autogenous shrinkage specimens.

Drying Shrinkage
The drying shrinkage was tested according to ASTM C157 [48]. The specimens were cast in 25 mm × 25 mm × 300 mm steel molds. After 24 h, the specimens were placed in air at an average temperature of 25 • C as shown in Figure 4. The procedure for taking the readings and frequency was similar to that for the autogenous shrinkage specimens.

Expansion
The paste specimens were cast in 25 mm × 25 mm × 300 mm steel molds. Two specimens for each mix were prepared to measure expansion. After 24 h, the specimens were immersed in water at a constant temperature (20 ± 1 • C) as shown in Figure 5. The procedure for taking the readings and frequency was similar to that for the autogenous shrinkage specimens. Buildings 2021, 11, x FOR PEER REVIEW 6 of 18

Chemical Shrinkage
The results for the chemical shrinkage (µ m/m) over 90 days are presented in Figure  6. For the purpose of comparison, the units of chemical shrinkage were converted to the same units as those for drying shrinkage, autogenous shrinkage and expansion (micrometer/meter) using Equation (1). It should be noted that there is a 1-day difference between chemical shrinkage and other shrinkages. Hence, the zero time for chemical shrinkage is the time immediately after placing, while the zero time for other shrinkage bars is the time after demolding (day 1). As shown in this figure, the inclusion of 0, 5, 10, 15 and 20% LF increased the chemical shrinkage within the first 24 h. Thus, the formation of calcium carboaluminate resulting from the reaction between LF and C3A in the cement increased the total volume of the hydrate products [28,29,49,50]. For the longer curing period, it was shown that chemical shrinkage kept going up as the curing time increased. For example, at 1 day, the chemical shrinkage of the 10% LF replacement level was about 619.5 µ m/m. This value rose to around 840.7 and 1239 µ m/m at 10 and 45 days, respectively. Therefore, the reaction between cement particles and LF perpetuated with time. For the first 3 days, the chemical shrinkage values were approximately similar for all mixes. At 90 days, the chemical shrinkage of pastes with 0% LF attained a value of 1578.2 µ m/m. This value increased to 1961.7 µ m/m at 5% LF and then faintly dropped with the addition of 10% LF (1917.4 µ m/m). After this slight decline, the chemical shrinkage went up to the optimum value of 2861.4 µ m/m at 15% LF. For an inclusion of 20%, the chemical shrinkage displayed a sharp drop and reached a value of 1770 µ m/m. This may be explained by the fact that with a high amount of LF (˃15%), as the curing time increased, the formation of calcium carboaluminate hydrate could consume CH. Thus, the chemical effect of LF could increase the solid volume of hydration products (external expansion), which reduced the total system volume (chemical shrinkage) [19]. Additionally, LF would mainly show a filler effect which refined the pore structure and reduced the porosity of the cement [23]. The incorporation of diverse percentages of LF enhanced the chemical shrinkage of pastes. This finding can be explained as follows: (a) Considering the fact that chemical shrinkage is a direct result of the hydration processes, the cement hydration acceleration due to the presence of LF increases the chemical shrinkage [27], and (b) LF with a fine particle size afford nucleation sites for hydration products to precipitate, accelerate the hydration process and increase the hydration degree of the cement [23].

Chemical Shrinkage
The results for the chemical shrinkage (µm/m) over 90 days are presented in Figure 6. For the purpose of comparison, the units of chemical shrinkage were converted to the same units as those for drying shrinkage, autogenous shrinkage and expansion (micrometer/meter) using Equation (1). It should be noted that there is a 1-day difference between chemical shrinkage and other shrinkages. Hence, the zero time for chemical shrinkage is the time immediately after placing, while the zero time for other shrinkage bars is the time after demolding (day 1). As shown in this figure, the inclusion of 0, 5, 10, 15 and 20% LF increased the chemical shrinkage within the first 24 h. Thus, the formation of calcium carboaluminate resulting from the reaction between LF and C 3 A in the cement increased the total volume of the hydrate products [28,29,49,50]. For the longer curing period, it was shown that chemical shrinkage kept going up as the curing time increased. For example, at 1 day, the chemical shrinkage of the 10% LF replacement level was about 619.5 µm/m. This value rose to around 840.7 and 1239 µm/m at 10 and 45 days, respectively. Therefore, the reaction between cement particles and LF perpetuated with time. For the first 3 days, the chemical shrinkage values were approximately similar for all mixes. At 90 days, the chemical shrinkage of pastes with 0% LF attained a value of 1578.2 µm/m. This value increased to 1961.7 µm/m at 5% LF and then faintly dropped with the addition of 10% LF (1917.4 µm/m). After this slight decline, the chemical shrinkage went up to the optimum value of 2861.4 µm/m at 15% LF. For an inclusion of 20%, the chemical shrinkage displayed a sharp drop and reached a value of 1770 µm/m. This may be explained by the fact that with a high amount of LF (>15%), as the curing time increased, the formation of calcium carboaluminate hydrate could consume CH. Thus, the chemical effect of LF could increase the solid volume of hydration products (external expansion), which reduced the total system volume (chemical shrinkage) [19]. Additionally, LF would mainly show a filler effect which refined the pore structure and reduced the porosity of the cement [23]. The incorporation of diverse percentages of LF enhanced the chemical shrinkage of pastes. This finding can be explained as follows: (a) Considering the fact that chemical shrinkage is a direct result of the hydration processes, the cement hydration acceleration due to the presence of LF increases the chemical shrinkage [27], and (b) LF with a fine particle size afford nucleation sites for hydration products to precipitate, accelerate the hydration process and increase the hydration degree of the cement [23].

Drying Shrinkage
The results for the drying shrinkage for all paste specimens are presen 7. The incorporation of 0% LF displayed expansion for the first few days. Th non was not expected, but this small expansion has been observed by other As displayed, all plots show similar trends. Drying shrinkage values increas ing time went up. For example, at 1 day, the drying shrinkage of the 5% LF was 0 µ m/m. This value went up to a value of 900 and 1275 µ m/m at 15 and spectively. The mechanism of long-term drying shrinkage can be well exp disjoining pressure, which is relevant to the drying shrinkage in paste [52-joining pressure existing between capillary pores could result in the contr solid body.
It was noticed as well that the drying shrinkage rose as the LF content i example, at 90 days, the inclusion of 0% LF achieved a value of 950 µ m/m. Th sequently rose until achieving the highest value of 1900 µ m/m with the add LF. The reason for this elevation may be attributed to the drop of Young's m paste as well as the difference in capillary pore structure connectivity, which moisture escape [56][57][58][59].
It can be stated that the w/b ratio can play a significant role in drying s velopment. The paste in this case will be more porous with more connectivi increase the rate of water evaporation, thus increasing the drying shrinkage [17,56,60,61]. The addition of the LF content tended to amplify the drying sh was due to the strong impact of an internal reaction which enhanced the cohe the cement and LF [62]. Hence, this opposite phenomenon encourages disr compaction of the paste, resulting in a higher amount of entrapped voids, w greater drying shrinkage [62].

Drying Shrinkage
The results for the drying shrinkage for all paste specimens are presented in Figure 7. The incorporation of 0% LF displayed expansion for the first few days. This phenomenon was not expected, but this small expansion has been observed by other studies [51]. As displayed, all plots show similar trends. Drying shrinkage values increased as the curing time went up. For example, at 1 day, the drying shrinkage of the 5% LF replacement was 0 µm/m. This value went up to a value of 900 and 1275 µm/m at 15 and 45 days, respectively. The mechanism of long-term drying shrinkage can be well explained by the disjoining pressure, which is relevant to the drying shrinkage in paste [52][53][54][55]. This disjoining pressure existing between capillary pores could result in the contraction of the solid body.  Figure 8 displays the autogenous shrinkage for various pastes during the first 9 of curing. Generally, there was a tendency to have small expansion during the fir days of curing, and thereafter autogenous shrinkage started to occur. As shown, t togenous shrinkage was influenced by the LF content. At 90 days, the autogenous s age for the 0% addition was 1100 µ m/m. This value went up and reached the opt of the paste as well as the difference in capillary pore structure connectivity, which led to easier moisture escape [56][57][58][59].

Autogenous Shrinkage
It can be stated that the w/b ratio can play a significant role in drying shrinkage development. The paste in this case will be more porous with more connectivity, which can increase the rate of water evaporation, thus increasing the drying shrinkage of the paste [17,56,60,61]. The addition of the LF content tended to amplify the drying shrinkage. This was due to the strong impact of an internal reaction which enhanced the cohesion between the cement and LF [62]. Hence, this opposite phenomenon encourages disruption in the compaction of the paste, resulting in a higher amount of entrapped voids, which leads to greater drying shrinkage [62]. Figure 8 displays the autogenous shrinkage for various pastes during the first 90 days of curing. Generally, there was a tendency to have small expansion during the first few days of curing, and thereafter autogenous shrinkage started to occur. As shown, the autogenous shrinkage was influenced by the LF content. At 90 days, the autogenous shrinkage for the 0% addition was 1100 µm/m. This value went up and reached the optimum value of 1650 µm/m with the incorporation of 10% LF. After this sharp increase, the autogenous shrinkage values dropped with the addition of 15 and 20% LF. This can be explained by the fact that LF fills the porous structures and produces higher capillary forces, which increases the self-desiccation during cement hydration. Moreover, for replacements above 10% LF, due to the internal curing effect, the released water from specimens with 15 and 20% LF can fill the emptied pore and leads to a reduction in the degree of self-desiccation and autogenous shrinkage [7,[63][64][65][66][67].

Expansion
The expansion results for the paste specimens with different percentages of LF are presented in Figure 9. The paste specimens displayed continuous expansion from zero time until 90 days. As shown in this plot, there was a logical variation in the expansion results when incorporating the LF content. For example, at 90 days, the expansion value for the 0% LF replacement was 2000 µ m/m. This value rapidly decreased to a minimum value of 1200 µ m/m with the addition of 10% LF. After this sharp decrease, the expansion value rose again up to 2100 µ m/m and 2150 µ m/m with the inclusion of 15 and 20% LF, respectively. Therefore, adding 15% LF and above enhanced the self-desiccation due to hydration. In particular, the curing time had an obvious effect on the expansion. The expansion increased as the curing time rose for all percentages. For example, at 1 day, the expansion of the 5% LF replacement was 0 µ m/m. This value went up to a value of 800 and 1400 µ m/m at 10 and 45 days, respectively. Moreover, as the curing time increased, swelling in the bars rose much more for the pastes with high LF contents (15 and 20%) than in It can be also detected that with the rising of curing time, the autogenous shrinkage will definitely rise. For example, at 1 day, the autogenous shrinkage value for 5% LF was 0 µm/m. This value rose to a value of 450 and 900 µm/m at 10 and 45 days, respectively. This was due to the pozzolanic reactions of the LF, which led to an increase in autogenous shrinkage [26,63,68,69].

Expansion
The expansion results for the paste specimens with different percentages of LF are presented in Figure 9. The paste specimens displayed continuous expansion from zero time until 90 days. As shown in this plot, there was a logical variation in the expansion results when incorporating the LF content. For example, at 90 days, the expansion value for the 0% LF replacement was 2000 µm/m. This value rapidly decreased to a minimum value of 1200 µm/m with the addition of 10% LF. After this sharp decrease, the expansion value rose again up to 2100 µm/m and 2150 µm/m with the inclusion of 15 and 20% LF, respectively. Therefore, adding 15% LF and above enhanced the self-desiccation due to hydration. In particular, the curing time had an obvious effect on the expansion. The expansion increased as the curing time rose for all percentages. For example, at 1 day, the expansion of the 5% LF replacement was 0 µm/m. This value went up to a value of 800 and 1400 µm/m at 10 and 45 days, respectively. Moreover, as the curing time increased, swelling in the bars rose much more for the pastes with high LF contents (15 and 20%) than in specimens with low LF contents. This was due to the hydrate products (monocarboaluminate and hemi or mono-carbonate) [49] that were formed at high LF contents being able to absorb more water and swell more than pastes with low LF contents [19,26]. Furthermore, it could be a result of the changes in phase composition within the hydrate network. As the CH content in the system reduced due to the addition of the LF content, the calcium aluminate was gradually replaced by carboaluminate, which has a lower density than calcium aluminate, providing an extra contribution to the expansion process [19,49].

Comparison between Chemical, Autogenous and Drying Shrinkage and Expansion
Figures 10-14 present the length change (chemical, drying and autogenous shrinkage and expansion) of pastes for 0, 5, 10, 15 and 20% LF, respectively, over 90 days. By com paring the development of shrinkage over 90 days, it seems that the chemical shrinkage values were the highest among all other shrinkages. Likewise, the inclusion of LF achieved a noticeable enhancement in shrinkage. According to the results illustrated in these fig ures, it appears that the drying shrinkage was clearly greater than the autogenous shrink age. This is well observed in Figure 13. The results are to be expected given that autoge nous shrinkage is caused by self-desiccation in paste as water is consumed during the hydration process. This generally occurs within the early days of casting the paste. How ever, drying shrinkage is the reduction in paste volume caused by the loss of water during the drying mechanism, and this continues perhaps for years after paste curing [63][64][65][66] Additionally, all shrinkage values increased as the curing time went up. Adding 15% LF attained the maximum chemical shrinkage. However, the inclusion of 10% LF and 20% LF exhibited the highest autogenous and drying shrinkage, respectively. Concerning the ex pansion, there was a decrease between 0 and 10% LF. Beyond this level, there was an increase in shrinkage, with a maximum value of 2150 µ m/m occurring at the 20% replace ment level.

Comparison between Chemical, Autogenous and Drying Shrinkage and Expansion
Figures 10-14 present the length change (chemical, drying and autogenous shrinkage and expansion) of pastes for 0, 5, 10, 15 and 20% LF, respectively, over 90 days. By comparing the development of shrinkage over 90 days, it seems that the chemical shrinkage values were the highest among all other shrinkages. Likewise, the inclusion of LF achieved a noticeable enhancement in shrinkage. According to the results illustrated in these figures, it appears that the drying shrinkage was clearly greater than the autogenous shrinkage. This is well observed in Figure 13. The results are to be expected given that autogenous shrinkage is caused by self-desiccation in paste as water is consumed during the hydration process. This generally occurs within the early days of casting the paste. However, drying shrinkage is the reduction in paste volume caused by the loss of water during the drying mechanism, and this continues perhaps for years after paste curing [63][64][65][66]. Additionally, all shrinkage values increased as the curing time went up. Adding 15% LF attained the maximum chemical shrinkage. However, the inclusion of 10% LF and 20% LF exhibited the highest autogenous and drying shrinkage, respectively. Concerning the expansion, there was a decrease between 0 and 10% LF. Beyond this level, there was an increase in shrinkage, with a maximum value of 2150 µm/m occurring at the 20% replacement level.

Relationships between Different Length Change Parameters
As mentioned in previous studies, the total chemical shrinkage is the sum of the external and internal volume changes [8,70,71]. In general, at early ages, the external dimensions of a paste can either drop due to self-desiccation (i.e., autogenous shrinkage) or to external drying (i.e., drying shrinkage) or rise due to expansion. The internal volume change represents the chemical shrinkage [3]. On the other side, some authors define chemical shrinkage in terms of the total chemical shrinkage and external chemical shrinkage [5,9]. After the paste becomes stiff and can resist a contracting force, the external chemical shrinkage is equal to the total chemical shrinkage.

Relationships between Different Length Change Parameters
As mentioned in previous studies, the total chemical shrinkage is the sum o ternal and internal volume changes [8,70,71]. In general, at early ages, the externa sions of a paste can either drop due to self-desiccation (i.e., autogenous shrinka external drying (i.e., drying shrinkage) or rise due to expansion. The internal change represents the chemical shrinkage [3]. On the other side, some author chemical shrinkage in terms of the total chemical shrinkage and external chemica age [5,9]. After the paste becomes stiff and can resist a contracting force, the extern ical shrinkage is equal to the total chemical shrinkage.
The relationships between chemical shrinkage and the other length change ters-autogenous shrinkage, drying shrinkage and expansion at different curing 7, 28 and 90 days)-are illustrated in Figures 15-17, respectively. As shown in F a positive relation existed between chemical shrinkage and autogenous shrinka plot draws a linear relationship, with a coefficient of determination R 2 of 0.89, 0 0.96 and 0.66 for 0, 5, 10, 15 and 20% LF, respectively. This means that during the h process and after setting started, chemical shrinkage did not occur spontaneously was delayed by the rigidity of the paste itself [3]. Hence, water took its place in th pores of the cement paste. As the hydration process progressed, the volume of t in the pores dropped. Thus, a process known as self-desiccation took place an creating autogenous shrinkage. Additionally, and before the initial setting of th the capillary forces were almost negligible. Thus, chemical shrinkage was essen only motivator for autogenous shrinkage.
Similarly, the relation between chemical shrinkage and drying shrinkage portional. The coefficient of determination R 2 was 0.8, 0.67, 0.69, 0.92 and 0.89 fo 15 and 20% LF, respectively ( Figure 16). This indicates that chemical shrinkage i with the increase in drying shrinkage. Chemical shrinkage occurred before th time of the paste while drying shrinkage took place after the setting time or w paste started to harden. Thus, chemical shrinkage was a basic contributor fo shrinkage. Furthermore, the driving force of drying shrinkage was the loss of resulting from the self-desiccation in the cement paste [26].
There was a negative linear relation between chemical shrinkage and expans ure 17). This negative relation indicates that as the chemical shrinkage values in the expansion of the paste dropped. In fact, when the bars were cured in water, w self-desiccation existed, the bars exhibited continuous expansion over time. T The relationships between chemical shrinkage and the other length change parametersautogenous shrinkage, drying shrinkage and expansion at different curing ages (1, 7, 28 and 90 days)-are illustrated in Figures 15-17, respectively. As shown in Figure 15, a positive relation existed between chemical shrinkage and autogenous shrinkage. This plot draws a linear relationship, with a coefficient of determination R 2 of 0.89, 0,89, 0.73, 0.96 and 0.66 for 0, 5, 10, 15 and 20% LF, respectively. This means that during the hydration process and after setting started, chemical shrinkage did not occur spontaneously, since it was delayed by the rigidity of the paste itself [3]. Hence, water took its place in the finesse pores of the cement paste. As the hydration process progressed, the volume of the water in the pores dropped. Thus, a process known as self-desiccation took place and led to creating autogenous shrinkage. Additionally, and before the initial setting of the paste, the capillary forces were almost negligible. Thus, chemical shrinkage was essentially the only motivator for autogenous shrinkage.
Buildings 2021, 11, x FOR PEER REVIEW shrinkage did not take place. Furthermore, as the amount of the LF content i (above 15% LF), the less shrinkage and expansion would always exist.

Conclusions
This paper discusses the volume stability of paste incorporating limestone fin cement replacement. The volume stability included three types of shrinkage-che autogenous and drying-and expansion. Based on the results obtained in this e mental work, the following conclusions can be made:  Similarly, the relation between chemical shrinkage and drying shrinkage was proportional. The coefficient of determination R 2 was 0.8, 0.67, 0.69, 0.92 and 0.89 for 0, 5, 10, 15 and 20% LF, respectively ( Figure 16). This indicates that chemical shrinkage increased with the increase in drying shrinkage. Chemical shrinkage occurred before the setting time of the paste while drying shrinkage took place after the setting time or when the paste started to harden. Thus, chemical shrinkage was a basic contributor for drying shrinkage. Furthermore, the driving force of drying shrinkage was the loss of moisture resulting from the self-desiccation in the cement paste [26].
There was a negative linear relation between chemical shrinkage and expansion ( Figure 17). This negative relation indicates that as the chemical shrinkage values increased, the expansion of the paste dropped. In fact, when the bars were cured in water, where no self-desiccation existed, the bars exhibited continuous expansion over time. Therefore, shrinkage did not take place. Furthermore, as the amount of the LF content increased (above 15% LF), the less shrinkage and expansion would always exist.

Conclusions
This paper discusses the volume stability of paste incorporating limestone fines as a cement replacement. The volume stability included three types of shrinkage-chemical, autogenous and drying-and expansion. Based on the results obtained in this experimental work, the following conclusions can be made:

•
Chemical shrinkage increased for substitutions levels between 0 and 15% LF, where it achieved the highest value of 2861.4 µm/m at 90 days. Beyond 15% LF substitution, the chemical shrinkage decreased. • Drying shrinkage increased as the LF content in the paste increased. At 90 days, the maximum drying shrinkage value of 1900 µm/m was achieved for 20% LF replacement.

•
The length change (autogenous shrinkage) showed positive values (i.e., expansion) for the first few days (7 days) for all replacements except for 10% LF replacement. After this period, all replacements showed contraction. Autogenous shrinkage increased between 0 and 10% LF, where it achieved the highest value of 1650 µm/m at 90 days followed by a sharp drop for compositions above 10% LF.

•
The expansion values for the paste specimens showed a slight decrease for replacements between 0 and 10% LF. Above 10% LF, there was sharp increase in expansion, and the highest value of 2150 µm/m was achieved for the paste with 20% LF. • A positive correlation existed between chemical shrinkage and both autogenous and drying shrinkage. However, a negative correlation occurred between chemical shrinkage and expansion.