E ﬀ ect of Aggregate Type and Specimen Conﬁguration on Concrete Compressive Strength

: In this paper, concrete mixes utilizing two sizes of natural aggregate and two sources of lightweight and recycled aggregates were used to investigate the e ﬀ ect of aggregate type and specimen size and shape on the compressive strength of concrete. In addition, samples from ready-mix concrete producers with di ﬀ erent strengths were evaluated using standard size cylinders and cubes. Results were obtained on the 7th, 28th, and 90th day. In addition, ﬂexural strength, split tension, and modulus of elasticity were evaluated on the 28th and 90th day. Statistical analyses were conducted to examine the signiﬁcance of the di ﬀ erence between the compressive strength values for each two mixes using tests of hypotheses. Moreover, other mechanical properties as a function of compressive strength were discussed and compared to those predicated by the American Concrete Institute (ACI) speciﬁcations. Results indicate specimen shape has a noticeable e ﬀ ect on the compressive strength as the Cylinder / Cube ratio on the 90 th day was ranging between 0.781 and 0.929. The concrete compressive strength and modulus of elasticity were signiﬁcantly a ﬀ ected by the aggregate type. The ﬂexural strength and split tensile strength were less a ﬀ ected by the aggregate type, which was also conﬁrmed by the values predicted with the ACI equations.


Introduction
Concrete compressive strength ( f c ) provides an indication about the ability of a specific mix to resist axial compression loads. Traditionally, axial compression test utilizes standard cubes and cylinders to determine f c and became the most commonly used test in the construction industry. In addition, other mechanical properties could be predicted as a function of f c and design codes and standards for concrete structures consider the concrete compressive strength as the main indicator of the concrete resistance to loading. Concrete compressive strength is greatly affected by cement content, water-to-cement ratio (w/c), and aggregate size and type. Other factors, summarized in Table 1, influence the test results or indirectly affect the concrete compressive strength [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16].
previous loading, and exposure conditions, as illustrated in Figure 1. Compressive strength of concrete prepared with lightweight aggregate is affected by aggregate strength, w/c ratio, surface texture, mechanical interlock, and bond with the cement paste. In addition, variation in absorption capacities leads to formation of interfacial transition zones (ITZ), which have different microstructure. However, mixing procedure and the use of supplementary materials improve the ITZ and the concrete compressive strength. Similarly, concrete strength produced utilizing recycled aggregate is affected by the previous factors, in addition to the percentage of recycled aggregate used in the mix.
The effect of petrographic characteristics of aggregates on the quality of concrete has been considered by References [29,30]. It was evident that concrete strength is significantly affected by the mineralogy and microstructure of the coarse aggregates. In addition, the aggregate strength and mechanical properties are influenced by the secondary products of serpentinites and andesites.

Effect of Aggregate Size and Shape
Aggregate gradation (size distribution), shape (flat, elongated aggregate, round), and texture (rough and smooth) have different effects on the mechanical properties according to concrete type being prepared Normal Strength Concrete (NSC), High Strength Concrete (HSC), Lightweight Concrete (LWC), and Recycled Aggregate Concrete (RAC). Several research efforts evaluated the effect of aggregate size and shape on the mechanical properties of concrete are summarized in Table  3 [6,7,[31][32][33][34]. Aggregate size and shape can affect the cement-aggregate bond strength; for example, large aggregate size creates high stress concentration in the cement paste, leading to increased cracking; therefore, small aggregate size is recommended for HSC. Similar recommendation is suggested for LWC to avoid crushing of big size aggregates while loading.

Effect of Specimen Configuration
Cylinders (100 × 200 mm, 150 × 300 mm) and cubes (100 × 100 × 100 mm, 150 × 150 × 150 mm) are commonly used to monitor compressive strength of concrete. The variation in concrete compressive strength due to shape and size might be attributed to variability in aggregate physical properties, different friction between concrete surfaces and loading platen, and variation of crack propagation and localized failure zone. Localization of the damage in a certain zone is affected by the slenderness of the sample and the boundary restraint between the loading platens and the specimen [12,35,36]. For cylinder specimens, it is expected that compressive strength should not be affected by the sample size as long as the ratio of height to maximum lateral dimension (h/d) is maintained 2:1 [37]. However, several studies [9,10,38,39] showed that compressive strength of small-size cylinders is slightly higher than that obtained using the 150/300 mm cylinders. This is due to the smaller contact area between the specimen surface and steel platen of the testing machine, which results in lower friction-based shear forces. Smaller specimens are also denser as they have less number of micro-cracks and defects, which strengthen their compressive strength. The effect of specimen shape on concrete compressive Mortar adhered during processing Potential cracks due to loading, exposure to different environmental conditions and/or crushing The effect of petrographic characteristics of aggregates on the quality of concrete has been considered by References [29,30]. It was evident that concrete strength is significantly affected by the mineralogy and microstructure of the coarse aggregates. In addition, the aggregate strength and mechanical properties are influenced by the secondary products of serpentinites and andesites.

Effect of Aggregate Size and Shape
Aggregate gradation (size distribution), shape (flat, elongated aggregate, round), and texture (rough and smooth) have different effects on the mechanical properties according to concrete type being prepared Normal Strength Concrete (NSC), High Strength Concrete (HSC), Lightweight Concrete (LWC), and Recycled Aggregate Concrete (RAC). Several research efforts evaluated the effect of aggregate size and shape on the mechanical properties of concrete are summarized in Table 3 [6,7,[31][32][33][34]. Aggregate size and shape can affect the cement-aggregate bond strength; for example, large aggregate size creates high stress concentration in the cement paste, leading to increased cracking; therefore, small aggregate size is recommended for HSC. Similar recommendation is suggested for LWC to avoid crushing of big size aggregates while loading.  [13,[17][18][19][20][21][22][23][24][25][26][27][28].

References
Aggregate Used in the Study

Effect on Strength (Compressive and Modulus of Elasticity) Comments
Chi et al. [17] Three types of LWA Both compressive strength and elastic modulus were affected by the aggregate type and decreased with an increase in w/c ratio. RCA is less dense, more porous, and has higher water absorption than NCA due to the residual adhered mortar on RCA, which weakens the interfacial zone and reduces both compressive strength and modulus of elasticity. Wardeh et al. [27] Recycled gravel with 30%, 65%, and 100% replacement ratio.

Effect of Specimen Configuration
Cylinders (100 × 200 mm, 150 × 300 mm) and cubes (100 × 100 × 100 mm, 150 × 150 × 150 mm) are commonly used to monitor compressive strength of concrete. The variation in concrete compressive strength due to shape and size might be attributed to variability in aggregate physical properties, different friction between concrete surfaces and loading platen, and variation of crack propagation and localized failure zone. Localization of the damage in a certain zone is affected by the slenderness of the sample and the boundary restraint between the loading platens and the specimen [12,35,36]. For cylinder specimens, it is expected that compressive strength should not be affected by the sample size as long as the ratio of height to maximum lateral dimension (h/d) is maintained 2:1 [37]. However, several studies [9,10,38,39] showed that compressive strength of small-size cylinders is slightly higher than that obtained using the 150/300 mm cylinders. This is due to the smaller contact area between the specimen surface and steel platen of the testing machine, which results in lower friction-based shear forces. Smaller specimens are also denser as they have less number of micro-cracks and defects, which strengthen their compressive strength. The effect of specimen shape on concrete compressive strength is, however, recognized in the difference between cylinder and cube strength. The concrete nominal strength ( f c ) has a great effect on the specimen shape factor. It positively correlates with the cylinder/cube strength ratio as it is found that increasing the concrete strength decreases the specimen shape effect [40]. European standard (ENV 206:1990) and (BS 1881: Part 120) specifications [41,42] recommend a 0.8 ratio for the cylinder/cube strength and this ratio reaches 1.0 as concrete strength increases. Several studies showed that it is considerably hard to adopt a simple ratio for the cylinder/cube strength [40,[43][44][45][46][47][48][49][50][51]. Compressive strength of concrete utilizing lightweight (LWA) and recycled aggregate (RA) are significantly affected by the size and aspect ratio of specimens due to the decrease in the concrete unit weight. This also could be attributed to poor crack distribution and localized failure zone due to the deteriorated aggregate interlock, which causes further decrease in the compressive strength with the increase of the specimen size [1,35].
In this study, effect of aggregate type and specimens' configuration on compressive strength of concrete is investigated. The aggregate type includes natural aggregate, recycled aggregate, and lightweight aggregate. The specimens' configuration includes two sizes for cubes and cylinders. The lab tests were conducted at 7, 28, and 90 days. Tests were conducted according to the American Society for Testing and Materials (ASTM) specifications and British Standards (BS). In addition, statistical analyses were conducted to examine the significance of the difference between the compressive strength values for each two mixes using tests of hypotheses.

Research Significance
Due to the depletion of and increased demand for natural resources, several natural aggregates were introduced to the construction industry. Therefore, one of the authors' goals is to evaluate the quality of the natural aggregates and compare the results to the previous well-established facts. In addition, another goal was to evaluate the suitability of the recycled and lightweight aggregates available in the market to produce concrete comparable to that produced by natural aggregate, hence, establish a base for comparison with concrete produced with natural aggregate.

Experimental Program
The main objective of the experimental program is to investigate the effect of aggregate type and specimen size and shape on the concrete compressive strength. In addition, correlation between compressive strength and aggregate strength is examined. The experimental program focuses on the evaluation of samples prepared in the laboratory utilizing six aggregate types, in addition to samples of five concrete strength (C45_1 (Ready mix 1-1), C75 (Ready mix 1-2), C45_2 (Ready mix 2-1), C60 (Ready mix 2-2), and C80 (Ready mix 2-3)) collected from ready-mix producers in Dubai and Sharjah, UAE. The use of the ready-mix concrete is considered to validate the results obtained based Crystals 2020, 10, 625 7 of 26 on the mixes prepared in the laboratory while conducting this research. Laboratory samples were prepared twice to ensure consistency and to account for variability of aggregate properties, especially recycled aggregate. The evaluation criteria included compressive strength of standard size cylinders and cubes, modulus of elasticity, flexural strength, and split tensile strength. Testing was conducted according to ASTM specifications [52][53][54][55][56][57][58].

Materials
All mixes prepared in the laboratory have Ordinary Portland cement type I (SG = 3.15), silica fume (SG = 2.2), Ground Granulated Blast-Furnace Slag "GGBS" (SG = 2.85), and tap water. In addition, normal weight dune sand (particle size 100% passing 0.6 mm, SG 2.60) and coarse sand (maximum particle size 4.75 mm, SG 2.60) were used as fine aggregates. Six types of coarse aggregate ( Figure 2) were used to produce the concrete mixes in the lab: 10 mm and 20 mm sizes limestone (10 mm Nat., 20 mm Nat.), sintered pulverized-fuel ash (LWA 1 ) and pumice (LWA 2 ) lightweight aggregates with aggregate size 4 to 8 mm, and recycled aggregate (RCA 1 , RCA 2 ) from two sources with size 4-14 mm. In addition, several mechanical and physical properties of the coarse aggregates were evaluated and summarized in Table 4. Detailed investigation of the aggregate properties has been reported in previous research [59,60]. In addition, all ready-mix producers' mixes used a mix of 10 mm and 20 mm natural aggregates, cement, GGBS, silica fume, fine sand, and w/cm = 0.39.    Los Angeles abrasion: results from the Los Angeles (LA) abrasion test could be used as an indicator of aggregate strength and a relative measure of resistance to crushing under a gradually applied compressive load. A high LA abrasion number reflects a larger portion of crushed aggregates (fine) compared to the original mass. However, in case of recycled aggregate, high LA values could be attributed to the mortar adhered to the aggregate during crushing of recycled concrete. In general, aggregate's composition, texture, and structure affect its strength [1]; therefore, it is recommended to evaluate the aggregate performance in concrete samples to have a better indication of its strength.
Absorption: lightweight and recycled aggregates from both sources have high absorption capacity than that of the NWA. Therefore, the LWA and RCA were pre-wet 15 min prior to mixing with part of the mixing water (about 15% of the LWA aggregate weight and 5% of the RCA aggregate weight) to compensate for the high absorption capacity and to avoid impact on short-term workability. In addition, about 5% of the cement and cementitious materials weight were added during the pre-wet time to enhance the bond strength between the aggregate and the cement paste [59,60].

Mix Proportioning
All mixes prepared in the lab were proportioned using the absolute volume method and were based on a normal weight self-consolidated concrete mix [61] with a target strength of 70 MPa. Volume fractions, Table 5, for all materials and w/c ratio were the same for all mixes prepared in the laboratory; however, weights of course aggregates were adjusted to account for the difference in specific gravities. The total volumetric fraction of the cement and supplementary cementitious materials was 16%. A low cement content (8%) per volume was selected as an approach to achieve eco-friendly concrete mixes. The volume fraction of the GGBS (6%) and silica fume (2%) are the commonly used ratios by the ready-mix producers. In addition, the w/c ratio was selected to achieve the target strength and durability requirements [61].

Testing Program
Cylinders 100 × 200 mm (4 × 8 in.) (Cy 1 ), 150 × 300 mm (6 × 12 in.) (Cy 2 ) and cubes 100 × 100 × 100 mm (4 × 4 × 4 in.) (Cu 1 ) and 150 × 150 × 150 mm (6 × 6 × 6 in.) (Cu 2 ) were prepared from each mix to evaluate the compressive strength development. In addition, flexural, splitting tensile strength, and modulus of elasticity were evaluated at 28-and 90-day for all mixes except mix LWA 1 . There was not enough material from LWA 1 to produce all samples, and the source was not available in the market. Therefore, only samples for compressive strength evaluation, for LWA 1 , were prepared and included in the related discussions. All samples were cured using wet burlaps for three days and were left to dry in ambient room temperature. Table 6 summarizes the tests, number of samples, sample size, age at testing, and specifications followed during testing.  Table 7 provides a summary of the compressive strength results for different concrete mixes at 7-, 28-and 90-day. In addition, Figures 3 and 4 show the average compressive strength for cube and cylinder specimens, respectively. The results show that generally the compressive strength for all the concrete mixes increased with age, as expected. The cube specimen of the natural concrete mixes gave 75% of its target strength (70 MPa) on the 7-day, 97% on the 28-day, and exceeded it on the 90-day. Both the lightweight aggregate and recycled aggregate concrete mixes resulted in 15-20% lower compressive strength than that of the natural mixes in the 7-, 28-, and 90-day due to the lower aggregate strength. Cylinder specimen exhibited lower strength than that of the cube as a result of the specimen shape effect. In addition, sample failure modes of cubes and cylinders from all mixes are shown in Figure 5. In addition, failure modes of other samples are illustrated in Annex A. These failure modes indicate the variability of the aggregate-cement bond strength, which is affected by the aggregate type. Similar results were reported by References [13,16,19,62,63]. Nonetheless, the lightweight and recycled aggregates showed improved bond strength due to the addition of cementitious materials during the pre-wet process before mixing, which was discussed elsewhere by References [59,60].  Table 7 provides a summary of the compressive strength results for different concrete mixes at 7-, 28-and 90-day. In addition, Figure 3 and Figure 4 show the average compressive strength for cube and cylinder specimens, respectively. The results show that generally the compressive strength for all the concrete mixes increased with age, as expected. The cube specimen of the natural concrete mixes gave 75% of its target strength (70 MPa) on the 7-day, 97% on the 28-day, and exceeded it on the 90-day. Both the lightweight aggregate and recycled aggregate concrete mixes resulted in 15-20% lower compressive strength than that of the natural mixes in the 7-, 28-, and 90-day due to the lower aggregate strength. Cylinder specimen exhibited lower strength than that of the cube as a result of the specimen shape effect. In addition, sample failure modes of cubes and cylinders from all mixes are shown in Figure 5. In addition, failure modes of other samples are illustrated in Annex A. These failure modes indicate the variability of the aggregate-cement bond strength, which is affected by the aggregate type. Similar results were reported by References [13,16,19,62,63]. Nonetheless, the lightweight and recycled aggregates showed improved bond strength due to the addition of cementitious materials during the pre-wet process before mixing, which was discussed elsewhere by References [59,60].

Modulus of Elasticity, Split Strength, and Flexure Strength Results
Figures 6-8 summarize the test results of modulus of elasticity, split tensile strength, and flexure strength for all mixes except mix LWA 1 . The results were significantly affected by the aggregate type. Moreover, the modulus of elasticity and flexural strength test results for all the concrete mixes had increased from 28-day to 90-day, as shown in Figures 6 and 8. The modulus of elasticity was less than that of the natural aggregate and ready-mix concrete by about 25% (for RCA) and 40% (for the LWA). However, the results of split tensile strength test indicate a significant variation among different concrete mixes as illustrated in Figure 7. This could be attributed to variation of the aggregate-cement bond properties, which are greatly influenced by the aggregate type [19,39,64,65]. It should be noted that the results in Figures 6-8 are based on two (for modulus of elasticity) or four (for split test) specimens only. Therefore, the results are considered preliminary results.  Figures 6 and 8. The modulus of elasticity was less than that of the natural aggregate and ready-mix concrete by about 25% (for RCA) and 40% (for the LWA). However, the results of split tensile strength test indicate a significant variation among different concrete mixes as illustrated in Figure 7. This could be attributed to variation of the aggregate-cement bond properties, which are greatly influenced by the aggregate type [19,39,64,65]. It should be noted that the results in Figures 6-8 are based on two (for modulus of elasticity) or four (for split test) specimens only. Therefore, the results are considered preliminary results.      Figures 6 and 8. The modulus of elasticity was less than that of the natural aggregate and ready-mix concrete by about 25% (for RCA) and 40% (for the LWA). However, the results of split tensile strength test indicate a significant variation among different concrete mixes as illustrated in Figure 7. This could be attributed to variation of the aggregate-cement bond properties, which are greatly influenced by the aggregate type [19,39,64,65]. It should be noted that the results in Figures 6-8 are based on two (for modulus of elasticity) or four (for split test) specimens only. Therefore, the results are considered preliminary results.

Discussion
To evaluate the effect of sample configuration (size and shape) and aggregate strength on the concrete compressive strength, t-pooled statistical analysis was conducted on the test results. For a specific sample configuration and testing date, most of the results reported in Table 7 are average of four samples. The results for the same sample size, ratios between ′ at 7-day and 28-day, ′ at 90day and 28-day, ′ at 7-, 28-, and 90-day and the target compressive strength, and average shape factor (ratio between the cylinder and cube compressive strength values) at 7-, 28-, and 90-day are summarized in Table 8. The results indicated that there was a continuous hydration and formation of C-S-H, which is reflected by the strength gain, and most of the ratios for Cy1, Cy2, Cu1, and Cu2 specimens for different mixes were less than 1 for 7-to 28-day ratio. Previous studies [16,59,60] conclude that the natural concrete mixes typically achieve 0.7-0.8 of the 28-day strength in the first 7 days. On the other hand, most of the ratios are more than 1 for 90-to 28-day ratio. This highlights the importance of long-term evaluation especially for lightweight and recycled aggregate concrete, which is also in agreement with the recommendations for concrete with GGBS, lightweight, and recycled aggregate concrete [15,19,60]. Some scattered results in Table 8 show lower compressive strength for the 90-day than the 28-day, especially for Cy2 specimen; nonetheless, the majority of the results indicate higher strength, which is similar to results reported in the literature [31,32,44]. Table 9 illustrates a summary of the ratio between the cylinder and cube strength from the literature and from the current study.

Shape Factor Analysis
Aggregate type, target strength, and sample configuration (cylinder or cube) and sample size might affect the compressive strength values of the concrete mixes and the shape factor (the ratio of cylinder to cube compressive strength). Therefore, statistical analysis was conducted to examine the effect of these parameters on the shape factor.

Discussion
To evaluate the effect of sample configuration (size and shape) and aggregate strength on the concrete compressive strength, t-pooled statistical analysis was conducted on the test results. For a specific sample configuration and testing date, most of the results reported in Table 7 are average of four samples. The results for the same sample size, ratios between f c at 7-day and 28-day, f c at 90-day and 28-day, f c at 7-, 28-, and 90-day and the target compressive strength, and average shape factor (ratio between the cylinder and cube compressive strength values) at 7-, 28-, and 90-day are summarized in Table 8. The results indicated that there was a continuous hydration and formation of C-S-H, which is reflected by the strength gain, and most of the ratios for Cy 1 , Cy 2 , Cu 1 , and Cu 2 specimens for different mixes were less than 1 for 7-to 28-day ratio. Previous studies [16,59,60] conclude that the natural concrete mixes typically achieve 0.7-0.8 of the 28-day strength in the first 7 days. On the other hand, most of the ratios are more than 1 for 90-to 28-day ratio. This highlights the importance of long-term evaluation especially for lightweight and recycled aggregate concrete, which is also in agreement with the recommendations for concrete with GGBS, lightweight, and recycled aggregate concrete [15,19,60]. Some scattered results in Table 8 show lower compressive strength for the 90-day than the 28-day, especially for Cy 2 specimen; nonetheless, the majority of the results indicate higher strength, which is similar to results reported in the literature [31,32,44]. Table 9 illustrates a summary of the ratio between the cylinder and cube strength from the literature and from the current study.

Shape Factor Analysis
Aggregate type, target strength, and sample configuration (cylinder or cube) and sample size might affect the compressive strength values of the concrete mixes and the shape factor (the ratio of cylinder to cube compressive strength). Therefore, statistical analysis was conducted to examine the effect of these parameters on the shape factor.  The statistical testing was conducted to compare the compressive strength of the 6-inch and the 4-inch standard specimen sizes for the same specimen shape (cube and cylinder) and concrete mix. Testing was conducted for available "n1 = n2 = 4" samples of results and was carried out for the compressive strength at 7-, 28-and 90-day. The result of this test showed that the majority of the compared standard specimen sizes gave equal compressive strength, which indicates a negligible effect of using different standard sizes on the concrete compressive strength. Accordingly, the data for the cube and cylinder have been combined (having a size of 8 samples). The average concrete compressive strength is calculated as the trimmed average of 6 samples (after excluding the highest and the lowest value). The specimen shape factors "C y /C u " for six concrete mixes were compared using the t-pooled hypothesis testing, for a 95% confidence level, to investigate the effect of aggregate type on the shape factor. Table 10 shows the summary of the hypothesis testing results for the shape factors at 90-day, which is considered to represent the long-term performance of the concrete. It should be noted that the limit for "t" is ±2.228. The results shown in Table 9 indicate that the aggregate type has a significant effect on the shape factor and the following observations were concluded: (i) there is no significant difference in the shape factor values when comparing natural and lightweight aggregates. However, there was a significant difference between the two types of lightweight aggregates, which was confirmed by Reference [59]; (ii) recycled aggregate concrete mixes show a statistically significant difference in the shape factors when compared to other types of aggregate. This can be attributed to the fact that the recycled aggregate properties cannot be controlled due to the high variability of their sources even in the same batch, which may have different quality and strength [13,21,59,66].  A summary of calculated specimen shape factors Cy1/Cu1 and Cy2/Cu2 of the five ready-mix concrete with different target strengths at 7-, 28-, and 90-day are shown in Table 11. The results for samples collected from the same ready-mix producer show that, in general, when increasing the target strength level of the concrete mix, the specimen shape factor increased. Similar results were reported by other researchers [12,40].

Effect of Aggregate Type on Concrete Compressive Strength Analysis
Compressive strength of six concrete mixes was compared using the pooled t-test method to investigate the effect of the aggregate type on the concrete strength. Table 12 summarizes the t-pooled analysis for each specimen shape "Cy and Cu" at 90-day. For the cylinder specimen, Table 12, the results showed that compressive strength was affected by both the aggregate type and specimen shape. To investigate the aggregate type effect separately, the 90-day specimen shape factors were applied to the compressive strength of the cylinders and then re-compared using the t-pooled testing method. Results in Table 12 show that generally the concrete compressive strength was highly affected by the aggregate type. The following observations were concluded: (1) For normal strength concrete, changing the aggregate size did not affect the concrete compressive strength when using natural aggregate, similar findings were reported by References [7,34].
(2) Both lightweight and RCA1 mixes show less compressive strength than the two natural concrete mixes. This is attributed to their lower aggregate strength. This conclusion is compatible with the results reported by previous research [13,27], which conclude that compressive strength has decreased with replacing the natural aggregate with recycled aggregate.

Effect of Concrete Target Strength on Shape Factor
A summary of calculated specimen shape factors Cy 1 /Cu 1 and Cy 2 /Cu 2 of the five ready-mix concrete with different target strengths at 7-, 28-, and 90-day are shown in Table 11. The results for samples collected from the same ready-mix producer show that, in general, when increasing the target strength level of the concrete mix, the specimen shape factor increased. Similar results were reported by other researchers [12,40]. Table 11. Summary of Cy 1 /Cu 1 and Cy 2 /Cu 2 ratios at 7-, 28-, and 90-day.

Effect of Aggregate Type on Concrete Compressive Strength Analysis
Compressive strength of six concrete mixes was compared using the pooled t-test method to investigate the effect of the aggregate type on the concrete strength. Table 12 summarizes the t-pooled analysis for each specimen shape "Cy and Cu" at 90-day. For the cylinder specimen, Table 12, the results showed that compressive strength was affected by both the aggregate type and specimen shape. To investigate the aggregate type effect separately, the 90-day specimen shape factors were applied to the compressive strength of the cylinders and then re-compared using the t-pooled testing method. Results in Table 12 show that generally the concrete compressive strength was highly affected by the aggregate type. The following observations were concluded:  (1) For normal strength concrete, changing the aggregate size did not affect the concrete compressive strength when using natural aggregate, similar findings were reported by References [7,34].
(2) Both lightweight and RCA 1 mixes show less compressive strength than the two natural concrete mixes. This is attributed to their lower aggregate strength. This conclusion is compatible with the results reported by previous research [13,27], which conclude that compressive strength has decreased with replacing the natural aggregate with recycled aggregate.
(3) The RCA 2 mix resulted in an equivalent compressive strength to those of the natural concrete mixes for both cylinder and cube specimens, which matches their corresponding aggregate strength and similarly gives higher compressive strength than those of the lightweight aggregate mixes.
(4) In accordance with the higher strength shown by the RCA 2 than RCA 1 , its concrete mix exhibited a slightly higher compressive strength than that of RCA 1 concrete mix in the cube specimen and an equivalent strength in the cylinder case. This difference could be attributed to the variation of the recycled aggregate sources even within the same batch, which may have different properties and strength.

Correlation between Aggregate Strength and Concrete Compressive Strength
Regression analysis was conducted to investigate the relationship between the aggregate strength of the six types considered in this study represented with their weight loss percentage that resulted from the LA abrasion test, and their concrete mixes' trimmed average of the equivalent compressive strength. Analysis was done for both cylinder and cube specimens' compressive strength at 28-day. The results indicate a good correlation between the aggregate strength and their concrete mix compressive strength. The R 2 values, for a second-order equation, was around 0.60 for the cylinder and cube compressive strength, as shown in Figure 9. These results refer to the great role played by the aggregate type in affecting the concrete compressive strength. This contradicted with several previous studies [67][68][69][70][71] that estimate concrete compressive strength based on the water-cement ratio and ignoring the effect of aggregate type/strength in their models. Findings from the current study are in agreement with other studies [6,9,17,44,[62][63][64] that aggregate type should be considered while predicting the concrete compressive strength.
Crystals 2020, 10, x FOR PEER REVIEW 20 of 28 (3) The RCA2 mix resulted in an equivalent compressive strength to those of the natural concrete mixes for both cylinder and cube specimens, which matches their corresponding aggregate strength and similarly gives higher compressive strength than those of the lightweight aggregate mixes.
(4) In accordance with the higher strength shown by the RCA2 than RCA1, its concrete mix exhibited a slightly higher compressive strength than that of RCA1 concrete mix in the cube specimen and an equivalent strength in the cylinder case. This difference could be attributed to the variation of the recycled aggregate sources even within the same batch, which may have different properties and strength.

Correlation between Aggregate Strength and Concrete Compressive Strength
Regression analysis was conducted to investigate the relationship between the aggregate strength of the six types considered in this study represented with their weight loss percentage that resulted from the LA abrasion test, and their concrete mixes' trimmed average of the equivalent compressive strength. Analysis was done for both cylinder and cube specimens' compressive strength at 28-day. The results indicate a good correlation between the aggregate strength and their concrete mix compressive strength. The R 2 values, for a second-order equation, was around 0.60 for the cylinder and cube compressive strength, as shown in Figure 9. These results refer to the great role played by the aggregate type in affecting the concrete compressive strength. This contradicted with several previous studies [67][68][69][70][71] that estimate concrete compressive strength based on the watercement ratio and ignoring the effect of aggregate type/strength in their models. Findings from the current study are in agreement with other studies [6,9,17,44,[62][63][64] that aggregate type should be considered while predicting the concrete compressive strength.
Modulus of elasticity E = 3320 f c + 6900 for f c 21 MPa < f c < 83 MPa (1) Split tensile strength f spt = 0.59 f c for f c 21 MPa < f c < 83 MPa (3) where, f c is the cylinder Cy 2 compressive strength at 28-day in MPa. λ is a reduction factor to account for aggregate type. λ = 0.85 for sand lightweight concrete.
In the current study, modulus of elasticity, split tensile strength, and flexural strength were calculated using equations from Reference [72] for natural aggregates and equations from Reference [73] for lightweight and recycled aggregate concrete mixes. The main goal of the comparison is to check the applicability of using ACI equations to predict the mechanical properties of concrete with different aggregate types. Table 13 summarizes the comparison of the calculated and tested values for the modulus of elasticity, the flexure strength, and the split strength. It is observed that the tested modulus of elasticity for the ready mixes and the mixes with natural aggregate ranged between 0.85 and 1.06 of the calculated values according to ACI equations. However, in the case of lightweight and recycled aggregates, the ACI equations overestimated the modulus of elasticity, which indicates unsuitability of these equations with RCA and LWA samples. Values from the tested samples were much lower than their standard predicted values, which were in the range of 0.55 to 0.74 of the calculated values. Lightweight aggregate and recycled aggregate samples require more curing time, which might affect the test results at 28 days, as discussed earlier in the paper.
The tested split tensile strength for the majority of the concrete mixes is lower than the standard predicted value, especially for the 10mm and 20mm concrete mixes. This indicates weak bond strength, resulting in an early failure in the lateral direction. In addition, ready-mix concrete mixes have resulted in tested flexural strength that is comparable to the calculated values (ratio of 0.83 to 1.09). The natural aggregates resulted in a lower flexure strength than the calculated values. The pumice lightweight aggregate and recycled aggregate concrete mixes exhibited higher flexural strength than those predicted, especially for the pumice lightweight aggregate concrete mix. This could be a result of its small size and rough surface, which increased the interlocking between aggregate and the cement paste and the special presoaking procedure followed for both lightweight and recycled aggregate concrete mixes, which enhanced the bond in the ITZ. Table 14 summarizes some of the results of the modulus of elasticity, split strength, and flexural strength from the literature and the current study results, which came compatible with each other. Variation in values is attributed to the different types of aggregate used and the strength level targeted in those studies.

Conclusions
This paper presents a preliminary experimental investigation of the effect of specimen configuration and aggregate type on the concrete compressive strength. Six concrete mixes with different aggregate types were cast and tested for compressive strength, elastic modulus, splitting tensile strength, and flexural strength. In addition, samples from five ready-mix concrete were collected from two producers and were evaluated for the same mechanical properties. Two specimen shapes, cylinder and cube, were used with 100 × 200 mm (4 × 8 in.) and 150 × 300 mm (6 × 12 in.) sizes. Compressive strength was evaluated on the 7-, 28-and 90-day, while the elastic modulus, split tensile strength, and flexural strength were evaluated on the 28-and 90-day. The results from the current study are based on a small number of samples, however, the following could be concluded from the findings: For the same cementitious materials, w/c ratio and curing conditions, compressive strength of the recycled and lightweight aggregate concrete was about 15-20% lower than that of natural aggregate and ready-mix concrete mixes. The modulus of elasticity was less than that of the natural aggregate and ready-mix concrete by about 25% (for RCA) and 40% for the LWA. These lower values can be attributed to the fact that these aggregates have lower strength than natural aggregates.
For all the concrete mixes in the study, the concrete compressive strength shows significant correlation with the aggregate strength (with an R2 of 0.61). Both flexural and split tensile strengths were less affected by the aggregate type than the compressive strength and elastic modulus.
The specimen size and aggregate unit weight effect on compressive strength is negligible regardless of the specimen shape and the aggregate type. In addition, for normal strength concrete and for the aggregate gradation used in the study (10 mm and 20 mm), aggregate size has no effect on the concrete compressive strength.
The cylinder/cube ratio ranged between 0.781 and 0.929 for the 90-day, which is in agreement with previous research. Using different types of aggregate in concrete mixes resulted in a cylinder/cube ratio influenced by the aggregate strength and, in turn, was affected by the specimen shape.
Based on the preliminary results obtained in this research, it seems that ACI equations [72,73] overestimated modulus of elasticity and split tensile strength values for LWC and RAC, which indicates inapplicability of such equations for lightweight and recycled concrete.
The presoaking procedure followed during mixing of lightweight and recycled concrete enhanced the ITZ, which was reflected by the improved bond, interlocking with the cement paste and better flexural performance. In this study, pumice lightweight aggregate and recycled aggregate concrete mixes exhibited higher flexural strength than those predicted by the ACI standard [69,70].

Funding:
The APC was funded, in part, by Open Access Program from the American University of Sharjah.