Engineering properties of structural lightweight concrete containing expanded shale and clay with high volume class F and class C fly ash

Structural optimization is a broad term in the construction sector, whereby material efficiency, as well as cost effectiveness of structures, are optimized. In concrete technology, high‐performance lightweight concrete structures often represent such optimized properties, including cost‐effectiveness and ease of application, while having a lower structural dead load. Although the use of lightweight concrete is often viewed as a sustainable practice, it does not address the high use and dependance on Portland cement, which has a high ecological footprint. In this regard, this study evaluates the engineering properties of structural lightweight concrete containing expanded shale and clay, as coarse aggregate, with a high quantity of coal fly ash (class F and C). For this purpose, a total of 15 mixes have been produced and a comprehensive series of physico‐mechanical and durability tests have been conducted. Based on the results, it is found that expanded clay outperforms expanded shale in terms of physico‐mechanical and durability properties of the resulting concrete, potentially due to its lower particle size distribution (used in this study) and the resulting porosity compared to expanded shale. Nonetheless, comparable physico‐mechanical properties are achieved when expanded shale and clay are used, as a full replacement of limestone. In turn, the performance of class C fly ash is found to be better in mechanical, but lower in certain durability variables, compared to their class F companions. The result of this study is found significant and point to suitability of using expanded shale and clay in combination with high‐volume fly ash.


| INTRODUCTION
Starting from the 20th century and the rise of composite materials, optimal design of structures that could rapidly correspond to the oscillating state variables, such as stresses and strains have been advocated. 1 This trend, while being applied to concrete technology, resulted in the use of a variety of materials for specifically designated purposes. One of such optimized and material efficient class of concrete is high performance lightweight concrete that can not only address the structural-bearing requirements but also have a relatively lower dead-load. In practice, in the early 20th century, the tendency of certain clays and shales to expand when exposed to high temperatures have been discovered and was firstly patented by Stephen J. Hayde under the name "Haydite." 2 One of the initial uses for this was to produce expanded shale for the construction of a 7500-ton tanker named USS Selma. 3 After such successful uses, during the 1930s the manufacturing of lightweight expanded clay aggregates started in Europe with various commercial names. 3 However, it was not until the 1980s and 1990s, that the use of materials with enhanced strength to weight ratio started to grab further attention due to better earthquake performance, as well as cost competitiveness. 4 In that respect, lightweight concrete with an ovendry density of generally less than 2000 kg/m 3 started to be produced for both structural and non-structural elements. [5][6][7][8][9] Most commonly, this has been done through the use of lightweight, porous aggregates, such as expanded clay, shale, slate, natural pumice, and perlite aggregates, 4 or the inclusion of excessive air through the use of open graded aggregate mix or foaming agents. Yet, at the initial stages of producing and using lightweight foam concrete, it was proven not to be very dependable due to its very high and often uncontrollable porosity. 4 This resulted in many research studies and the development efforts to take place producing lightweight concrete through naturally available porous aggregates (e.g., expanded shale, clay, slates, etc.), industrial byproducts (e.g., coal bottom ash sand and foamed blast furnace slag) and even synthetic aggregates, such as polystyrene beads and recycled polymeric aggregates (e.g., low-density polyethylene). 2,10 In general, compared to foam concrete, lightweight concretes produced through lightweight aggregates have a relatively lower pore connectivity and it can be used for structural purposes. 4 Most commonly, expanded shale, clay, slate, ceramsite, and pumice are used as lightweight aggregates. 4 Although each have specific properties and often benefits, due to local availability, in this study, expanded shale and clay have been used as the main lightweight aggregates.
Expanded shale refers to a class of clay rich sedimentary rocks, laminated with minerals, such as feldspar, quartz, and chert with many uses in ceramic and construction industry. 11 Most commonly, shale has a relatively angular shape with a bulk density of ranging from 500 to 800 kg/m 3 with a considerably high water absorption of around 12%-14%. 12,13 To produce expanded shale and clay, they are commonly exposed to high temperatures of 1100-1300 C through kilns. 14 This process results in considerable expansion (up to five times 14 ) and causes gaseous fumes to be released from aggregates that results in the production of pores within the particles. 15 Unlike expanded shale's angular shape, clay particles are more spherical with a honeycomb structure. 4 With a density range of 250-710 kg/m 3 , 14 expanded clay is mostly composed of SiO 2 and Al 2 O 3 16 and has a variety of uses that includes but is not limited to soil stabilization, acoustic insulating materials, and drainage applications. 17 In general, lightweight aggregates offer a broad range of benefits including optimum load capacity, thermal and noise insulation, as well as often enhanced fire resistance. 15 Yet, despite such benefits, the use of lightweight aggregates does not address the high ecological footprint of Portland cement use in concrete production which is estimated to be as much as 3.5 billion tons of CO 2 production, annually. 18 In that respect, the use of supplementary cementitious materials (SCMs) such as various types of slags, fly ashes (e.g., coal and municipal incinerated ash), [19][20][21][22] agricultural waste (from rice husk ash to corn ash) has been more recently advocated and practiced. Although numerous materials have been used and found applicable in reducing the dependence on Portland cement, coal fly ash has constantly been found to be a dependable SCM that has proven to be beneficial in thermo-durability and even long-term mechanical properties. [23][24][25] Apart from such favorable effects, coal fly ash particles represent unique rheological properties due to their spherical shapes that can enhance the thixotropic properties of fresh concrete. 26 As a result, although the prospect of coal fly ash's availability is on a declining trend, its effect on thixotropic and thermo-durability of concrete has made the most novel practices, such as 3D printing of concrete to be heavily dependent on its use. [27][28][29] In that respect, this paper studies the engineering properties of structural lightweight concrete containing expanded shale and clay with high volume class F and class C fly ash. Despite the numerous studies already conducted in the use of high-volume coal fly ash, the combined and comparative effect of fly ash (class F and class C) with expanded shale and clay aggregates have not yet been reported. Ozbakkaloglu and Gholampour, 30 for instance, studied the effect of high volume class F fly ash and blast furnace slag inclusion in the production of sustainable concrete. Based on their results, it has been reported that the inclusion of high-volume fly ash significantly reduces mechanical properties of the produced specimens and results in higher water absorption when compared to the control mixes. Although a significant contribution is made in the mentioned study, the scope and results of the study do not include the effect of class C fly ash, as well as lightweight aggregates. Further, Reference 31 studied the mix design and properties of lightweight selfconsolidating concretes developed with furnace slag, expanded clay, and expanded shale aggregates by using a constant value of 40% fly ash. In their analysis, certain mechanical properties have been reported but the scope of their research did not include the physico-durability properties of high-volume fly ash and the outlined results were only associated with class F fly ash. In similar fashion, Reference 32 studied the influence of fly ash, bottom ash, and expanded clay, up to 35% (by weight), and mostly reported the corresponding effects of mechanical properties. In turn, Reference 33 studied the effect of high-volume fly ash concrete in marine environment using expanded shale and clay. However, in this study, mostly the corrosion tendency and rapid chloride permeability (RCP) test have been evaluated Reference 34 evaluated the physico-mechanical properties of concretes with fly ash as an SCM to replace OPC with different expanded cork granules and expanded clay. In that respect, a high quantity of fly ash (class F) inclusion was reported to reduce thermal conductivity when coupled with expanded clay. Despite the use of fly ash at 20-50% substitution rate, their scope of research was mostly the thermal properties of the produced mixes. Furthermore, Subasi 35 studied the including of 30% fly ash in high strength structural lightweight concrete manufactured with expanded clay aggregate. Similar to other studies mentioned, Subasi 35 analyzed the physico-mechanical properties of the concrete produced and noted that a 10% binder volume inclusion of fly ash is the most promising for mechanical properties. Finally, Lotfy et al. 31 studied the use of expanded shale and clay for selfcompacting concrete. In the mentioned study, fly ash (class F) had a replacement of 40% and it was reported that due to lower reactivity rate of coal fly ash, a significantly different bonding structure is observed. Also in this study, the scope of reports were mostly inclusive of a handful of mechanical tests, and no physico-durability properties were reported.
As can be seen in the above literature review, the effect of 30%, 50%, and 70% Portland cement substitution rate with fly ash (class F and class C) and its combined effect with the use of 100% limestone, expanded shale, and clay has not been studied nor reported. In this regard, a comprehensive experimental program was conducted in this research study that is further discussed in the following section.

| Portland cement
In this research, an ASTM C150 type I ordinary Portland cement with a density and specific surface area of 3.15 g/cm 3 and 3310 cm 2 /g has, respectively, been used. Table 1 provides further details about the physical and chemical properties of the used cement.

| Aggregate
Expanded shale and clay with a specific gravity of 1.78 and 1.59 along with a manufactured coarse and fine limestone aggregate has been used in this study. The coarse aggregate had nominal max size of 19 mm (3/4 in.) and met the specified requirements for a grade 67 aggregate according to ASTM C33. The expanded clay has been supplied from Arcosa Lightweight Inc. in Texas, originating from Erwinville, Louisiana. In addition, expanded shale aggregates have been supplied from Trinity Expanded Shale & Clay Inc. originating from Streetman plant located in Texas. Table 2 provides further details on the physical properties of the used aggregates. Figure 1 also presents the result of sieve analysis conducted based on ASTM C136 36 on different aggregates used in this study.

| Coal fly ash
In this study, locally available coal fly ash (class F and class C) with a density of 2.52 and 2.75 (g/cm 3 ) has, respectively, been used. The supplied fly ash was received from Boral Resources Inc. and the power station plant producing the fly ash was Martin Lake Plant located in Tatum, Texas. Table 3 provides further details on the properties of the fly ash used in this study.

| Superplasticizer
In this study, to increase the flowability of the mixes, a superplasticizer with a commercial name of Viscocrete 2100 and specific gravity of 1.08 kg/m 3 was added in sufficient amount to achieve workable mix.

| Mixture proportions
In this study, a total of 15 mixes with a constant water-tobinder (w/b) ratio of 0.4 has been produced. In the mixes, limestone aggregate content has been substituted by 100 vol% with expanded clay and shale. In the same way, Portland cement was replaced with 50 and 70 vol% coal fly ash (class F and class C). Table 4 provides further details on the mixture proportions. In this table the mixes are labeled as follows: letters P, L, S, C, FF, and FC refer to Portland cement, limestone, expanded shale, expanded clay, and fly ash class F and class C, respectively.
For instance, FA70S100 refers to a mix with 70% fly ash (type F) content and 100% expanded shale.

| Specimen preparation and test methods
In this study, the aggregates were kept in saturated surface dry condition to ensure they do not have a significant effect on drying shrinkage values. Further, the dry materials have initially been mixed for a period of 1 min and then after adding water the mixing continued until a coherently mixed materials would be prepared. Then, concrete samples were made by placing the freshly mixed concrete materials in proper molds and after the initial 24 h of curing, samples were moved to proper curing chamber with a constant temperature and humidity of 23 C ± 2 C and >95%, respectively. To assess the physico-mechanical properties of the produced specimens, compressive strength was conducted according to ASTM C39 37 after 7, 14, 28, and 56 days of curing on a 100 Â 200 mm cylinders, while split tensile, ASTM C496 38 and elastic modulus, ASTM C469 39 were tested on 28 days of curing using the same 100 Â 200 mm cylinders. Flexural strength test has also been conducted based on 152 Â 152 Â 500 mm samples based on ASTM C78. 40 Further, abrasion resistance, water absorption, drying shrinkage, freeze thaw resistance, and RCPT have been conducted based on ASTM T A B L E 2 Physical properties of aggregates used in this study.

Information
Expanded  45 respectively. Table 5 provides further details into the experiments conducted and their respective specimens size and timing.

| RESULTS AND DISCUSSION
3.1 | Workability Figure 2 presents the result of slump test conducted on the mixed materials. Based on this figure, it can be seen that for a given aggregate type, mixes with and without fly ash have a mean slump value of $156 and 145 mm, respectively. The better flowability of mixes containing fly ash is well-documented in the literature 47,48 and is known to be due to the spherical shape of fly ash particles and their initial low reactivity, especially class F ashes. 49,50 The results, however, show an average of 156 mm slump value for mixes containing fly ash class C and 156.8 mm those having class F. This shows very similar performance of fly ashes of different classes. Nonetheless, for a given Portland cement and fly ash content, mixes with and without lightweight aggregates are found to have a mean slump value of $173 and 116 mm, respectively. This shows a 49% higher workability of mixes with lightweight aggregates which can be due to the less angularity of these aggregates, as well as lower specific gravity, especially expanded clay, when compared to their limestone companions. In addition, since the lightweight aggregates were kept in SSD condition, they appear to have not reduced paste water to affect the lubrication adversely. When either of lightweight aggregates are used, the flowability of mixes with fly ash class F are 175.8 mm, compared to companion mixes with class C fly ash and 173.8 mm slump value. Although such slight variation can be statistical error, it can also possibly present slightly higher early hydration rate of class C fly ashes, as reported by References 51 and 52.  Tables 3 and  1, respectively. In turn, for a given fly ash values, mixes containing lightweight aggregates (both expanded clay and shale) have a mean density value of $1883 kg/m 3 which is about 17% lower than their companion mixes with limestone with $2264 kg/m 3 . In the same way, it can be seen that the use of 100% expanded shale (P100F0S) and clay (P100F0C) aggregates have reduced density values by $15 and 18%, respectively. This reduction is associated with the higher porosity and absorption capacity of expanded shale and clay when compared to the limestone sand used in control specimens as also shown in Table 2 and discussed in References 14 and 31. Further from Figure 3, when class C and class F fly ashes are used the density values are found to have a mean of 1998 and 1990 kg/m 3 , respectively. Yet, when only lightweight aggregates are used, the mentioned values change to 1874 and 1859 kg/m 3 , respectively. This shows that when either of lightweight aggregates are used, the fresh density values have not been impacted by the type of aggregates or fly ashes, significantly. Further, the lowest density values are also found to be for mixes produced with expanded clay and class F fly ashes which correlates with the values reported in Tables 1, 2, and 3.  Figure 4, it can be seen that the inclusion of 70% fly ash (class F) has reduced the compressive strength values for $44%, 33%, 28%, and 19%, at 7, 14, 28, and 56th day of curing. For mixes containing class C fly ash, however, these values change to $30%, 24%, and 18%. Based on this result, the addition of either class of fly ash retards the compressive F I G U R E 2 Slump values of mixes. strength development initially but appear to react at later ages, as discussed at length by previous studies such as  In the same way, the mean compressive strength of mixes with and without fly ash at 56th day of curing is 42 and 49 MPa, respectively. This shows $13% reduction of compressive strength due to F I G U R E 3 Fresh density of the mixes. the lower reactivity of fly ash particles. Nonetheless, the mean value of 56th day strength of concretes produced with class F fly ash is about 20% lower than those companion samples produced with class C fly ash. However, while comparing class C fly ash containing samples with Portland cement ones, only 10% reduction is observed. This shows a comparable performance of samples containing fly ash class C.

| Compressive strength
Based on Figure 4, it can further be seen that the inclusion of 100% expanded shale and clay has resulted in mean compressive strength of $40.5 MPa while mixes produced with 100% limestone, at any given fly ash type or quantity have resulted in $49.1 MPa strength, at 56th day of curing. This shows that the inclusion of lightweight aggregates has reduced the strength for about 17.5%. This is comparable to the 17% reduction of fresh density values, reported in Section 3.2.
Based on the results, the combined effect of utilizing 70% fly ash (class F) with 100% expanded shale and clay has reduced the compressive strength values at 56th day of curing by $47 and 26%, respectively. Based on this, it can be seen that expanded shale has developed a significantly lower compressive strength value compared to the mixes with expanded clay mixes when both mixes have used a high quantity of fly ash. This can be due to the higher porosity and even larger size of expanded shale particles (as shown in Figure 1), 57 compared to their clay companions.
Further, it can be seen that after 28 days of curing, mixes with natural aggregate have not developed much strength further until the 56th day of curing. In turn, mixes containing lightweight aggregate (e.g., P100F0S and P100F0C) are found to have slightly higher strength gain values from 28 to 56th day of curing. This strength F I G U R E 5 Elastic modulus of mixes at the 28th day of curing.
F I G U R E 6 Flexural strength of mixes at the 28th day of curing. gain of lightweight aggregate containing mixes can be due to the internal curing effect of lightweight aggregates, as discussed by Reference 58. Figure 5 presents the result of elastic modulus tested after 28 days of curing. Based on this Figure, the highest and lowest values is 33.5 GPa for P100F0NA and 23.3 GPa for P30FF70S, respectively. Further from this figure, it can be seen that the effect of shale incorporation into the mix with 70% fly ash (class F) is most pronounced with $30% reduction, compared to the control mix. Although significant, when used with 100% Portland cement, mixes with 100% expanded shale (P100F0S) are found to perform similar to 100% expanded clay (P100F0C) with 7% and 5% reduction of elastic modulus, compared to control mix with limestone (P100F0NA). Similarly, at 50% fly ash class C incorporation, both P50FC50S and P50FC50C developed $9 and 2% lower elastic modulus compared to P100F0NA, respectively. The lower elastic modulus of mixes containing lightweight aggregates is reported to be due to the low ductility of specimens produced with lightweight aggregates that can result in brittle failure. 57,59 This tendency is better seen when the mean elastic modulus value of concrete mixes containing limestone, expanded shale, and clay is found to be 33, 30, and 30 GPa, respectively. Other reasons for higher modulus of elasticity of mixes produced with limestone can be the higher specific gravity and lower water absorption of limestone aggregates. This has further been reported by Reference 60 that the use of aggregates with higher specific F I G U R E 7 Splitting tensile strength of mixes at the 28th day of curing.

| Elastic modulus
F I G U R E 8 Abrasion mass loss of mixes at the 28th day of curing. gravity and lower water absorption can result in higher strength and elastic modulus.

| Flexural strength
The result of 28 days flexural strength of mixes is presented in Figure 6. Based on this figure, flexural strength values range from 6.98 MPa achieved for P100F0NA with 100% limestone to 4 MPa for P50F50NA mix. As can be seen in this figure, various mixes performed somewhat similarly. In that respect, the mean value of mixes with (50% and 70%) and without fly ash (class F) is $4.7 and 5.3 MPa, respectively. When class C fly ash is also included, the mean values of fly ash containing versus purely produced with Portland cement is both $5 MPa. Nonetheless, the lower flexural strength values achieved for mixes containing high volume class F fly ash is due to the lower reactivity of fly ash particles, 61 as stated previously. Further, the mean values of flexural strength concrete mixes with and without fly ash class C at 50% and 70% inclusion rate are 5.3 and 4.9 MPa, respectively. Moreover, the incorporation of 50% and 70% of class C fly ashes in P50FC50NA and P30FC70NA has reduced the flexural strength by 16% and 20% compared to P100F0NA. Similar to compressive strength, the influence of high-volume fly ash class C and F on the flexural strength of the concrete mixes is comparable until 50% inclusion.
Further from Figure 6, it can be seen that the mean flexural strength value of mixes with and without lightweight aggregate is $4.7 and 5.3 MPa which translates into a $14% difference. The lowered flexural strength of mixes with lightweight aggregate can be due to the lower surface adhesion of lightweight aggregates because of higher porosity, when compared to limestone aggregates as suggested by References 31 and 62 or it can be the increased porosity at the interfacial transition zone spaces in the lightweight aggregates compared to normal aggregates, as reported by References 57 and 63. Figure 7 presents the result of splitting tensile strength of various mixes tested after 28 days of curing. Based on this figure, it can be seen that the highest and lowest values is for P50FC50C and P30FF70C with 4.29 and 2.01 MPa, respectively. Further from this figure, it can be seen that mixes with either type of fly ash has a mean splitting tensile strength value of 3 MPa. However, when fly ashes are used with lightweight aggregates, class F develops a mean strength of $2.6 MPa while class C containing mixes develop a mean strength of $3.2 MPa. This shows that class C fly ash has performed better with lightweight aggregates. With either fly ashes, the use of expanded shale is found to develop some $7% higher strength value compared to companion samples produced with expanded clay. This can be due to higher specific gravity of shale aggregates, as shown in Table 2. The result of including expanded shale and clay aggregates is comparable to those reported in Reference 31 that used 20 vol% fly ash and reported a splitting tensile strength of ranging up to 3.08 MPa with mixes with lightweight aggregates experiencing a slightly reduced the splitting tensile strength values.

| Abrasion resistance
Abrasion resistance test is a means to evaluate concrete specimens' performance when exposed to repeated physical and frictional loads. 64 The result of mass loss of mixes F I G U R E 9 Water absorption of mixes at the 28th day of curing. subjected to abrasion test is presented in Figure 8. According to this figure, the addition of 70% fly ash to the mixes with 100% limestone, is found to reduce the mass loss by up to 15%. In the same way, the mean mass loss of mixes with fly ash (both 50% and 70%) is $0.47% while their companion mixes with 0% fly ash have a lower mean mass loss of 0.32%. The better abrasion resistance of mixes containing fly ash (class F) is also reported in Reference 65 that used up to 40% fly ash (class F) in their mixes and found $39% lower depth of wear for specimens. This can also be associated with the lower developed strength of the mentioned mixes containing either type of fly ashes. 65 Further from Figure 8, the combination of 70% fly ash (class F and C) with 100% expanded shale and clay, however, is found to increase the mass loss by $127% and 150%, respectively. Based on this, the higher mass loss of specimens with both fly ash and lightweight aggregate can be due to the lower surface adhesion of cementitious paste with the lightweight aggregates 66 or the increased overall porosity as discussed by Reference 66.
Notably, expanded clay is found to experience a relatively higher mass loss ($39% for P100F0C) when compared to the specimens made with expanded shale ($31% for P100F0S). This can possibly be due to lower specific gravity of expanded clay particles (1.59 vs. 1.784).

| Water absorption
One of the suitable ways for evaluating the durability of a concrete specimen is through the observation of the characteristics of its absorption. The result of water absorption test, conducted after 28 days of curing is presented in Figure 9. From this figure, it can be seen, that the increase in fly ash quantity has constantly increased the water absorption values. In that respect, mixes produced with either class of fly ashes are found to have a mean water absorption rate of 5% while for those produced without fly ash this value is $3.7%. Such significant difference can be due to the lower microstructural development of fly ash, as discussed by References 49 and 50.
Further form this figure, it can be seen that for a given fly ash content, the mean value of water absorption obtained from mixes with lightweight aggregate (both expanded shale and clay) is 5% which compared to mixes with limestone (1.88%) shows a 64% increase. The higher water absorption of mixes with lightweight aggregate is due to the lower density of expanded shale and clay, as well as increased overall porosity of the specimens made with those mixes. 4 Further, it can be seen that the mean water absorption of mixes containing clay ($4.3%) is relatively lower than those with shale (5.7%). This can be due to the improved particle packing of mixes containing expanded clay compared to their companion mixes with shale and explains the reason for better performance of mixes with clay aggregates.

| Drying shrinkage
Since lightweight concrete is commonly produced through an increase in the porosity values of specimens, evaporation of water from the pores coupled with lower internal solid content can induce internal cracking and a loss of mechano-durability properties. 67 The result of drying shrinkage test conducted in this study is presented in Figure 10. As can be seen, the highest and lowest drying shrinkage values after 56 days of curing is $3649 με for P30FF70C and 530.6 με for P50FF50NA, respectively. This figure shows that the mean values of mixes with higher fly ash content developed relatively lower drying shrinkage values. In that respect, the mean value of mixes with and without fly ash is found to be $1623.9 and 1708.9 με. According to Reference 68, this can be caused by the lower reactivity of coal fly ash particles that results in lower autogenous shrinkage, compared to mixes produced with 100% Portland cement.
In addition to the effect of fly ash, it can be seen that mixes containing lightweight aggregate appear to have developed considerably higher drying shrinkage values. In the respect, the mean value of micro strain for mixes with and without lightweight aggregates is found to be $2134.4 and 653.87 με. This translates into a 226% increase in drying shrinkage values. The higher shrinkage values of specimens containing lightweight aggregate can be due to the increased hydrostatic tension, as a result of moisture loss, which tends to be more deteriorating when lower content of solid materials are available to withstand this contracting force. 69 The mentioned value, however, is found to be more affected by the mixes containing expanded clay, with a mean micro strain value of $2761.8, compared to expanded shale with $1506.9 με. As shown in Table 2, this can be due to the lower specific gravity (also relative density) of expanded clay to withstand the internal stress applied as part of the shrinkage internal stress effect.

| Freeze-thaw resistance
Freeze thaw resistance refers to the ability of concrete to withstand the impact of high temperature variation which can have destructive effect on the physicomechanical and durability properties of the concrete. Due to the increased porosity of lightweight concrete and F I G U R E 1 1 Mass loss of specimens after 324 freeze thaw cycles (a) and their respective relative dynamic modulus of elasticity (b). its potential for hosting free water, freeze thaw resistance can be the prime source of durability concern. In that sense, the result of freeze thaw test conducted in this study to measure such property is shown in Figure 11a,b. Based on Figure 11a that shows the mass loss of specimens exposed to 324 freeze thaw cycles, the largest mass loss value is for specimens prepared with the highest fly ash content (50% and 70%). In this regard, the mean value of mixes produced with and without either of fly ashes is 1.08 and 0.33%, respectively. Followed by this, mixes containing lightweight aggregates had a mean mass loss of 1.23% that compared to 0.34% of samples produced with limestone, is significant. Based on this, it can be concluded that the inclusion of fly ash has increased the mass loss values considerably but at a similar rate to the lightweight aggregates. This can be due to the lower reactivity of fly ash particles at the time of testing the freeze thaw resistance that performs similarly to the higher internal porosity of lightweight aggregates. With regards to the impact of fly ashes, it is reported by Reference 70 that the inclusion of fly ash tends to considerably reduce the interconnected voids and reduces the number and volume of capillary pores. Yet, despite this, due to the low mechanical strength of specimens produced with high volume fly ash, the expansive pressure of freeze thawing destroys the surface of concrete and causes a higher mass loss to take place. 71,72 Further from Figure 11a, at a given fly ash content, the specimens prepared with limestone, expanded shale, and clay are found to experience a mass loss mean value of 0.34%, 1.08% and 1.38%, respectively. The enhanced performance of expanded shale compared to clay can be associated to the higher specific gravity of shale particles, compared to its expanded shale companion, as reported in Table 2. Based on this, the inclusion of 70% fly ash and lightweight aggregates have had the highest mass loss values which is comparable with the reduction of compressive strength results presented in Section 3.4. Figure 11b presents the result of relative dynamic modulus of elasticity which is basically the proportion of stress to strain and reflects the elasticity performance of a given material. 73 A reduction in the dynamic modulus of elasticity then, translates into a loss of elasticity performance. 73 As can be seen in this figure, almost all specimens experienced a loss of dynamic modulus of elasticity as the number of exposures to freeze thaw cycles increased. On top of this loss, P30FF70C and P30FF70C with $34.6 and 32.3% reduction have the highest reduction in dynamic modulus of elasticity. In contrast, P100F0NA and P50FC50NA experienced the lowest relative dynamic modulus of elasticity loss with $5% and 8% reduction. This can be due to the higher compaction of mixes with limestone and their respective lower porosity to host free water compared to mixes with expanded clay and shale. Moreover, only few mixes showed the retention capabilities of <70% for dynamic modulus of elasticity. Nonetheless, all mixes had a dynamic modulus of elasticity of above 60% which is noted in ASTM C666 for being the acceptable threshold. 44

| Rapid chloride permeability test
RCP test is commonly used to determine the resistance of concrete specimens to the penetration of chloride ions which is a critical parameter in determining the durability of concrete when exposed to salty environments. 74 It is commonly reported that a coulomb value of 1000-2000 and from 2000 to 4000 is low and moderate rate of chloride permeability, respectively. 75 The result of RCP test conducted in this study is presented in Figure 12. Based on this figure, the highest and F I G U R E 1 2 RCP of mixes at the 28th day of curing. the lowest coulomb values are for P30FF70NA with 2743 and P50FC50C with 773 coulombs. The lower coulomb values of expanded clay containing mix is aligned with Reference 33 that utilized high volume fly ash (class F) and reported a 14% reduction of coulombs values when expanded clay is used, instead of expanded shale.
Further, it is notable that mixes with 50% fly ash (class F and C) are found to have the lowest coulombs values. Mixes containing either type of fly ash are found to have a mean value of 1433.3 coulombs while mixes produced solely with Portland cement have a mean of 1746 coulombs. This can be due to the higher compaction of mixes with a combination coal fly ash and Portland cement that potentially results in lower pore network connectivity. According to Reference 76, the penetration of chloride in concrete samples generally depends on two main factors which includes the free hydroxyl ion in the pore solution, and the interconnecting voids of the concrete sample. On this basis, the addition of fly ash is found to decrease the interconnected voids in the concrete. Further, the inclusion of fly ash is known to reduce the amount of free alkali 76 that helps to bind the hydroxyl ion in the pore solution while also reducing the alkalinity of the pore solution. This can further explain as to why class C containing mixes developed lower coulombs values when compared to class F containing mixes (1300 vs. 1567 coulombs) Reference 77, however, noted that such lowered RCP result can result in higher degree of vulnerability to carbonation, which future studies in this area can further explore this phenomenon.
Similarly, the inclusion of lightweight aggregates is found to also reduce the coulombs values for about 31% (from 1893.4 to 1297.1 coulombs). For aggregates, this can be explained by the non-connectivity of pores within lightweight aggregates. 4,78 4 | CONCLUSIONS In this study, the physico-mechanical and durability properties concrete containing expanded shale and clay with high volume class F and class C fly ash have been evaluated. Based on the findings of this research, the following conclusions can be drawn: • The flowability of mixes containing fly ash, expanded clay and shale is found to be higher than the control mixes. This is believed to be due to the lower plastic viscosity contributed by the lower hydration rate of mixes containing high volume class F and even class C fly ash compared to Portland cement. In case of expanded clay and shale, their lower surface adhesion is found to be a major factor in increasing the flow values. Similarly, the results of fresh density tests show that the inclusion of fly ash slightly reduces the density values due to its lower specific gravity, compared to Portland cement. Mixes with class F, specifically experienced slightly lower density values compared to those produced with class C. As for mixes with expanded clay and shale, it is found that the inclusion of expanded clay results in a slightly lower density values compared to the other mixes containing expanded shale. The main reason for this can be aggregate gradation or higher porosity of the used aggregates. • The result of compressive, flexural and splitting tensile strength shows that the inclusion of lightweight aggregates generally reduces mechanical properties. Between expanded shale and clay, shale aggregates are found to have relatively lower strength values which is believed to be due to the higher porosity and particle size distribution of shale aggregates compared to limestone or expanded clay. In the same way, the elastic modulus of mixes prepared with expanded shale is also found to be lower than that of other mixes, potentially due to increased air-void content and reduced strength.
In terms of fly ash content, the results point to the better performance of having 50% fly ash in either type and not higher inclusion content. Nonetheless, it was observed that class C fly ashes can produce comparative strengths to mixes produced solely with Portland cement even at 50% inclusion. • The results of abrasion resistance tests show the superiority of mixes prepared with limestone aggregate, supplied with high volume fly ash. In turn, the combined effect of high-volume fly ash with lightweight aggregates (both expanded clay and shale) is found to experience higher mass loss than mixes only containing Portland cement with either of lightweight aggregates. This is believed to be due to the lower surface adhesion of lightweight aggregates that increases the abrasion mass loss values when combined with lower reactivity of fly ash. As the results were mostly impacted by the types of aggregates, the type of fly ash has been found not to have a considerable impact on the results. Similar observation is made for water absorption test whereby mixes with expanded clay generally exhibit lower water absorption values compared to their companion mixes with expanded shale. Notably, however, class C fly ashes were observed to perform worse than class F, when used in combination to lightweight aggregates. It is hypothesized that this can be due to different microstructural development or statistical error in the results. Further studies in this area can potentially shed further light on this behavior.
• The results of drying shrinkage are found to have been significantly impacted by the inclusion of fly ashes and lightweight aggregates. In this regard, the mixes produced with expanded clay exhibit a relatively higher shrinkage values compared to all other mixes. After mixes with expanded clay, mixes containing expanded shale and limestone have the highest shrinkage values, respectively. The combined effect of fly ash and expanded shale or clay, is found to consistently increase the shrinkage strains. This is hypostatized to be because of the lower content of solid materials, in case of the use of lightweight aggregates, within the specimens that can withstand internal stress caused by the shrinkage. Similarly, the results of freeze-thaw tests also show that the mixes prepared only by limestone and Portland cement have the lowest mass loss and the inclusion of fly ash and lightweight aggregates significantly increase the mass loss values. Likewise, the relative dynamic modulus of elasticity values also follows the same trend and is mostly impacted by the aggregate type. • RCP tests show that mixes containing 50% fly ash have the lowest coulomb values compared to all other mixes. This might suggest that the inclusion of fly ash potentially consumes hydroxyl ions in the pore solutions as well as resulting in the lowest pore network connectivity. This has been more observed for mixes containing class C fly ashes which can be caused by higher consumption of chloride ions and lowered alkalinity of pore solutions. In general, the mean coulomb values for different aggregates are found to be from the lowest to the highest in mixes containing expanded clay, shale and then limestone, respectively.
The findings of this research are found to be highly promising and point to the possibility of sustainable production of structural-grade lightweight concrete with high volume fly ash in combination with expanded shale and clay used as the main aggregate.

FUNDING INFORMATION
This research did not receive any specific grant or funding from agencies in the public, commercial, or not-forprofit sectors.

CONFLICT OF INTEREST
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.