Improved Test Methods for Specific Gravity and Absorption of Coarse and Fine Aggregate (2015)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

64 Practical Significance of Aggregate Specific Gravity Test Results This chapter discusses the practical significance of the relative accuracy of the specific gravity determinations. The underlying question is, how close should the result from a revised test method be to that of the current method? To answer this question, two analyses were conducted. The first analysis was to determine if differences in the results of revised and current methods were within the acceptable range of the cur- rent methods. The second analysis assessed the sensitivity of the portland cement concrete and asphalt mixture volumet- ric properties to aggregate specific gravity values. Repeatability Analysis The first analysis was based on the repeatability of the cur- rent specific gravity tests. From a practical point of view, it is acceptable if differences in results from different procedures (i.e., alternative methods) are less than differences in results of duplicate specimens tested according to the current methods. Precision estimates of the current methods are available in AASHTO T 84 and T 85 and also are determined annually by the Proficiency Sample Programs and reported on the AMRL website. The within-lab standard deviations (Sr) and accept- able range of two results (D2S) from AASHTO T 84 and T 85, respectively, are as follows: AASHTO T 84 Gsb: Sr = 0.011; D2Sr = 0.032 AASHTO T 85 Gsb: Sr = 0.009; D2Sr = 0.025 The precision indices for fine and coarse aggregate Gsbs determined annually by the Proficiency Sample Programs are shown in Tables 5-1 and 5-2, respectively. As can be seen, the precision estimates vary significantly from year to year due partially to the use of different aggregate sources in the program. For the fine aggregate precision estimates, data for the last 4 years suggest an improvement in the within-lab and between-lab statistics. However, it should be noted that the Proficiency Sample Program has used a different method of screening data since 2006. The new method detects more outliers, resulting in precision estimates that are smaller than those reported in previous years and those cited in the current standards. The precision estimates for coarse aggregate Gsb shown in Table 5-2 also show lower within-lab and between- lab statistics from 2006 to 2009, but the 2010 results increase to levels more consistent with the earlier 8 years. For this analysis, the precision statistics from the current standards as shown in Tables 5-1 and 5-2 were used to assess how well the results from alternative methods compared with the results from the standard methods evaluated in Chap- ter 3. Table 5-3 shows the Gsb results from the experiments to compare testing of the aggregates in the in-situ moisture condition and drying with the CoreDry apparatus to aggre- gates prepared by the standard oven drying method. Results that differed from the oven drying method by more than the within-lab acceptable range of two results were highlighted in shaded cells. From this table, it was evident that only testing of coarse aggregate in the in-situ moisture condition provides results that are reasonably consistent (based on practical ranges) with the oven drying method. Table 5-4 shows the Gsb results from the experiment to compare vacuum saturation of different time periods to hydrostatic soaking for 15 hours. From this table, it was evi- dent that for the coarse aggregates, results using vacuum satu- ration were reasonably consistent (based on practical ranges) with hydrostatic soaking, for each of the three time periods used in the experiment. For the fine aggregates, results using the 15-minute vacuum time were reasonably consistent with the static soaking except for the Preston sandstone. How- ever, this result was close to the within-lab range of 0.032. Sensitivity Analysis Another way to view the issue is to ask, how sensitive are concrete and asphalt mix designs to the values of the aggre- gate specific gravity? For this analysis, an asphalt mix design C H A P T E R 5

65 having a gradation of 50 percent passing the No. 4 sieve was considered. The standard deviation for the combined Gsb of the blend would be the average of the standard deviations from T 84 and T 85. Likewise, the acceptable range of two Gsb measurements of the combined aggregate should be within the average of the coarse and fine aggregate D2S ranges. Therefore, a reasonable precision estimate for Gsb of the blend is (0.032 + 0.025) = 0.0285 = 0.029, which is the practically attainable limit with the current methods. The following calculation illustrates how this acceptable range would affect the calculated VMA for an example asphalt mix. Year Sample No. of Labs Single Operator Multilaboratory No. Participated* Data Used** 1s d2s 1s d2s 2010 167/168 1293 1081 0.006 0.017 0.012 0.034 2009 163/164 1282 1091 0.007 0.020 0.017 0.048 2008 159/160 1189 1001 0.006 0.018 0.015 0.043 2007 155/156 1025 946 0.006 0.018 0.014 0.040 2006 151/152 1044 1016 0.017 0.048 0.029 0.081 2005 147/148 965 939 0.016 0.045 0.033 0.093 2004 143/144 951 936 0.019 0.054 0.041 0.115 2003 139/140 864 850 0.017 0.048 0.037 0.105 2002 135/136 753 739 0.014 0.040 0.034 0.095 2001 131/132 656 642 0.015 0.044 0.033 0.093 2000 127/128 586 579 0.021 0.060 0.041 0.115 1999 123/124 551 540 0.013 0.038 0.028 0.079 1998 119/120 483 475 0.035 0.098 0.045 0.127 Notes: *Total number of laboratories that participated in the program each year **Number of laboratories whose data were used to determine precision estimates Table 5-1. AASHTO T 84/ASTM C128 precision estimates published by AMRL. Year Sample No. of Labs Single Operator Multilaboratory No. Participated* Data Used** 1s d2s 1s d2s 2010 169/170 1384 1246 0.015 0.043 0.034 0.096 2009 165/166 1332 1093 0.003 0.010 0.006 0.017 2008 161/162 1298 1133 0.009 0.025 0.016 0.046 2007 157/158 1198 995 0.005 0.015 0.009 0.025 2006 153/154 1175 956 0.005 0.014 0.009 0.025 2005 149/150 1072 1046 0.012 0.034 0.024 0.067 2004 145/146 1031 991 0.031 0.086 0.019 0.054 2003 141/142 939 919 0.018 0.051 0.044 0.124 2002 137/138 847 838 0.016 0.044 0.026 0.074 2001 133/134 789 766 0.010 0.027 0.019 0.052 2000 129/130 696 693 0.015 0.043 0.027 0.075 1999 125/126 590 579 0.045 0.128 0.029 0.081 1998 121/122 545 542 0.019 0.053 0.031 0.088 Notes: *Total number of laboratories that participated in the program each year **Number of laboratories whose data were used to determine precision estimates Table 5-2. AASHTO T 85/ASTM C127 precision estimates reported by AMRL.

66 Given: Gmb = 2.400, Pb = 5.0, Gsb = 2.650 VMA G P G mb b sb 100 1 100 2.400 100 5.0 2.650 14.0% ( )( ) ( )= − × −  = − × − = If the Gsb were changed by 0.029, the new VMA would be as follows: Given: Gmb = 2.400, Pb = 5.0, Gsb = 2.679 VMA = 14.9% This clearly shows that VMA is very sensitive to Gsb. This also illustrates one of the serious flaws in using VMA as a mix design or acceptance criterion for asphalt paving mix- tures. The test methods for the materials are so variable that it would be easy to design or accept a mixture that appears to be satisfactory but could have too low or too high a VMA. In the absolute volume method of concrete mix propor- tioning, the sand (fine aggregate) content is determined last and is based on the volume of the other ingredients subtracted from the total unit volume (e.g., cubic yard). Therefore, when the specific gravity of the coarse aggregate is changed, the net effect is a change in the volume or proportion of sand used in the concrete mixture. Note that ACI 211.1, Standard Practice for Selecting Proportions for Normal, Heavyweight, and Mass Concrete uses “oven-dry relative density (specific gravity)” (i.e., Gsb) of coarse and fine aggregates in the method for determining concrete mix designs. Also note that all spe- cific gravity values used in ACI standards are reported to the hundredths place (e.g., 2.68). The following mixture constituents are from the example mix design in Design and Control of Concrete Mixtures, (31, pp 163–166). Water = 270 lb/(1 × 62.4 pcf) = 4.33 ft3 Cement = 643 lb/(3.15 × 62.4 pcf) = 3.27 ft3 Material 15-Hr Soak 5-min. Vacuum Diff. 10-min. Vacuum Diff. 15-min. Vacuum Diff. Coarse Aggregate (D2Sr = 0.025) BF Slag 2.333 2.325 0.008 2.334 -0.001 2.328 0.005 Elmore Gravel 2.563 2.561 0.002 2.559 0.004 2.560 0.003 PS Coarse 2.499 2.496 0.003 2.496 0.003 2.492 0.007 RC Limestone 2.523 2.541 -0.018 2.540 -0.017 2.544 -0.021 RE Concrete 2.347 2.365 -0.018 2.365 -0.018 2.351 -0.004 Fine Aggregate (D2Sr = 0.032) Natural Sand 2.619 2.627 0.008 2.630 -0.011 2.624 -0.005 BF Slag Fine 2.715 2.734 -0.019 2.723 -0.008 2.720 -0.005 PS Fine 2.531 2.591 -0.060 2.578 -0.047 2.567 -0.036 RC Limestone 2.511 2.541 -0.030 2.518 -0.007 2.533 -0.022 Texas Sand 2.552 2.598 -0.046 2.594 -0.042 2.555 -0.003 Table 5-4. Comparison of Gsb results from 15-hour soak (current standard) to results from vacuum saturation. Material Oven Drying vs. In-situ Moisture Condition Oven Drying vs. CoreDry Method Oven Drying In-situ Difference Oven Drying CoreDry Difference Coarse Aggregate (D2Sr = 0.025) BF Slag 2.333 2.337 0.004 2.333 2.360 0.027 Elmore Gravel 2.563 2.565 0.002 2.563 2.566 0.003 PS Coarse 2.489 2.490 0.001 2.489 2.496 0.007 RC Limestone 2.523 2.545 0.022 2.523 2.552 0.029 RE Concrete 2.347 2.333 0.014 2.347 2.396 0.049 Fine Aggregate (D2Sr = 0.032) Natural Sand 2.619 2.631 0.012 2.619 2.633 0.014 BF Slag Fine 2.715 2.692 0.023 2.715 2.752 0.037 PS Fine 2.531 2.594 0.063 2.531 2.582 0.051 RC Limestone 2.511 2.553 0.042 2.511 2.508 0.003 Texas Sand 2.552 2.599 0.047 2.552 2.595 0.043 Table 5-3. Comparison of Gsb results from oven drying (current standard) to results from in-situ moisture condition and CoreDry methods.

67 Air = (7.0%/100) × 27 ft3 = 1.89 ft3 Coarse Aggregate = 1674 lb/(2.68 × 62.4 pcf) = 10.01 ft3 The weight of coarse aggregate is determined from Fig- ure 9.3 or Table 9.4 (Bulk Volume of Coarse Aggregate per Unit Volume of Concrete) and the dry-rodded unit weight of the coarse aggregate (31). The above volumes sum to 19.50 cubic feet, which is sub- tracted from 27.0 cubic feet to leave 7.50 cubic feet for the dry fine aggregate. The mass of dry fine aggregate is then determined by multiplying that volume by the Gsb of the fine aggregate and the unit weight of water, 62.4 pcf, as follows: Fine Aggregate = 7.50 ft3 × 2.64 × 62.4 pcf = 1236 lb Trial batch weights are then adjusted to account for the free (surface) water in the aggregates used to prepare the batch. This is where absorption values come into play. In the exam- ple, the coarse aggregate has an absorption capacity of 0.5 per- cent and a moisture content of 2 percent. The fine aggregate has an absorption capacity of 0.7 percent and a moisture con- tent of 6 percent. The batch weights of the aggregates are increased by multi- plying the oven-dried batch weights by 1 plus the moisture contents, as follows: Coarse Aggregate = 1674 lb × 1.02 = 1707 lb Fine Aggregate = 1236 lb × 1.06 = 1310 lb The batch water is then reduced by the amount of free water with each aggregate. The percentage of free water is the differ- ence between the moisture content and the absorption. For the coarse aggregate, the free water percentage is 2.0% - 0.5% = 1.5%, and for the fine aggregate, the free water percentage is 6.0% - 0.7% = 5.3%. The adjusted batch water is 270 lb - (1674 lb × 0.015) - (1236 lb × 0.053) = 179 lb If the aggregate specific gravities changed by the within- lab D2S range, the above example would be recalculated as follows: Water, cement, and air volumes and batch weights would remain the same. The volume of coarse aggregate would be Coarse Aggregate = 1674 lb/((2.68 + 0.025) × 62.4 pcf) = 9.92 ft3 The volume of fine aggregate would be 27.00 - (4.33 + 3.27 + 1.89 + 9.92) = 7.59 ft3 When the specific gravity of the fine aggregate was changed by the within-lab D2S, the batch weight of fine aggregate would be 7.59 cu. ft × (2.64 = 0.032) × 62.4 pcf = 1266 lb (an increase of 30 lb or 2.4%) The within-lab standard deviations (Sr) and acceptable range of two absorption results (D2S) from AASHTO T 84 and T 85 are AASHTO T 84 Percent Absorption: Sr = 0.11%, D2Sr = 0.31% AASHTO T 85 Percent Absorption: Sr = 0.088%, D2Sr = 0.25% If the example absorption values for the coarse and fine aggregates were varied by these ranges, the fine aggregate would have an absorption of 0.7% + 0.31% = 1.01%, and the coarse aggregate would have an absorption of 0.5% + 0.25% = 0.75%. The free water contents of the coarse aggregates would be 2.0% - 0.75% = 1.25%, and for the fine aggregate, the free water percentage would be 6.0% - 1.01% = 4.99%. Thus, the adjusted batch weight for water would be 270 lb - (1674 lb × 0.0125) - (1266 lb × 0.0499) = 186 lb This is a change of 7 pounds (less than 4 percent) in the batch weight for the water. According to the text on Batch Adjustments in Design and Control of Concrete Mixtures (31, p 165), increasing the mixing water content by 5 pounds will decrease the air content by 1 percent and increase the slump by one-half inch. Therefore, for concrete mix designs, changes in specific gravities for the aggregates that are less than the repeatabil- ity of the Gsb determinations could result in a small change in the proportions of coarse and fine aggregate but are not likely to have a substantial impact on properties of the fresh or hardened concrete. Changes in aggregate absorption values by the practical range of the within-lab D2S limits will have a small impact on the air content and workability of the concrete. Based on Fig- ure 9.4 in Design and Control of Concrete Mixtures (31), a loss of 1 percent air could have a moderate impact on the resis- tance of the concrete to freeze-thaw and deicing chemicals. In summary, the practical ranges of the within-lab preci- sion estimates for Gsb determinations are reasonable for con- crete. However, for asphalt, the Gsb test result has a very large impact on VMA, one of the more critical criteria for asphalt mix design and acceptance. To reduce the impact on VMA to a tolerable level, the acceptable range of differences between within-lab replicates alternative methods would have to be less than 0.010, which is about one-third of what is currently practically attainable.