Aggregate Grading

A considerable amount of testing over the past 2 years at the National Ready Mixed Concrete Association Research Laboratory, plus three producer laboratories

NRMCA

A considerable amount of testing over the past 2_ years at the National Ready Mixed Concrete Association Research Laboratory, plus three producer laboratories around the country, has yielded two detailed reports: Effect of Continuous (Well-Graded) Combined Aggregate Grading on Concrete Performance Phase A: Aggregate Voids Content (Packing Density) & Phase B: Concrete Performance. Compiled by Karthik Obla, NRMCA managing director of research and materials engineering; Haejin Kim, NRMCA Concrete Research Laboratory manager of engineering; and, Colin Lobo, NRMCA senior vice president of engineering, the studies provide results of significant potential value to practitioners, engineers and researchers.

According to the reports, packing tests demonstrated that using well-graded combined aggregates did not lead necessarily to maximum aggregate packing density. Further, concrete performance studies indicated that the use of well-graded combined aggregates failed to result conclusively in lower water demand, less shrinkage, reduced bleeding, or higher strength. Yet, improvements in finishability and segregation resistance were observed in certain situations.

Based on the research findings, the authors conclude that Coarseness Factor and 8-18 Individual Percent Retained charts are potential concrete mixture optimization tools available for evaluation and implementation by ready mixed producers. However, they should not be invoked as project specification requirements. The reports also provide a procedure for users to develop their own aggregate-grading study program with local materials.

A complete account of the studies can be found at www.nrmca.org/research/eng_articles.asp. More information can be obtained also by contacting NRMCA’s Karthik Obla, tel.: 888/846-7622, ext. 1163; e-mail: [email protected].

RESEARCH RATIONALE

It is generally believed that if aggregate volumes are selected to maximize packing density of the combined aggregates, then the amount of cementitious paste volume required for a given amount of workability (i.e., slump) is minimized. Supporting that contention is the following rationale: Cementitious paste should fill completely the voids between aggregate particles, providing as well a certain amount of excess to supply the lubrication necessary for a given workability. Consequently, reducing void content (i.e., increasing packing density) in the aggregate skeleton decreases total cementitious paste needed for the workability desired. Also commonly accepted is the proposition that minimizing cementitious paste volume in a concrete mix helps attain lower shrinkage, reduced heat of hydration, improved durability, and lower costs.

To attain maximum aggregate packing density, empirical approaches typically used in the U.S. aim to achieve a ÎcontinuousÌ, also known as Îwell-graded’, combined aggregate grading or particle-size distribution. ÎCombinedÌ in this context refers to a combination of both coarse and fine aggregate grading in the ratio that they are present in the concrete mixture. Among domestic producers, popular practice for achieving Îwell-graded combined aggregatesÌ (WG) in concrete involves the application of Coarseness Factor and 8-18 charts.

In the Coarseness Factor (CF) chart, the x-axis represents percent of combined aggregate retained on the No. 8 sieve that is also retained on the ?-in. sieve; and, the y-axis represents percent of combined aggregate that passes the No. 8 sieve. If plotted in Zone II (denoted by a box, see page 83), a combined aggregate grading is considered optimal. Workability factor should be corrected for different cementitious content, e.g., a 94-lb. decrease in cementitious weight (from a standard cementitious content of 564 lb/yd3) requires a 2.5 percent workability factor increase.

The 8-18 chart sets limits between 8 and 18 percent for large top-size aggregates, such as 1_-in., or 8 and 22 percent for smaller maximum-size aggregates, such as 1- or Ê-in., retained on each sieve below the top size and above the No. 100. ACI 302.1R-04 Section 5.4.3 adds several restrictions on certain other sieve sizes (top size, No. 30, No. 50, No. 100), while allowing limited lower individual percent retained on sieves. (A combined aggregate is plotted on the 8-18 chart on page 83.)

To some extent, the CF chart has been correlated with field performance; however, laboratory testing of the empirical approaches had not been performed to validate the assumption that maximum aggregate packing density or lower cementitious paste for a given degree of workability is thereby achieved. Accordingly, the NRMCA research project was undertaken with a twofold objective: 1) To determine if empirical approaches employing the CF and 8-18 charts to achieve WG actually result in maximum aggregate packing density; and, 2) To determine if the empirical approaches lead to improved concrete performance, as indicated by such criteria as lower paste content, decreased shrinkage, higher strength, and better workability.

The project’s considerable scope necessitated addressing the first objective in Phase A: Aggregate Voids Content (Packing Density) and the second objective in Phase B: Concrete Performance. Throughout the report, WG is defined as a combined aggregate grading plotted in Zone II of the CF chart (Shilstone, ACI 302) and compliant with 8-18 chart requirements recommended by ACI 302.1R-04. Combined aggregate gradings failing to meet the two requirements are referred to as Înot well graded combined aggregatesÌ or NWG.

PHASE A: PACKING DENSITY

To determine the effect on aggregate voids content of the range of combined aggregate grading, two conditions were tested: 1) No. 57 (larger coarse aggregate) combined with No. 8 (intermediate coarse aggregate) at different percentages by volume; and, 2) No. 57 combined with fine aggregate at different percentages by volume. Altogether, testing was completed for a total of 14 combined aggregate gradings.

The No. 8 aggregate resulted in an average voids content of 41.3 percent, while the No. 57 aggregate resulted in a voids content of 39.6 percent. Blending different proportions by volume of No. 57 and No. 8 aggregates resulted in slightly lower void contents, averaging between 37.3 and 37.9 percent Û not a significant difference. Thus, when the No. 57 and No. 8 aggregates were combined, resulting in a combined aggregate size distribution of 1 in. (25 mm) to No. 8 sieve size (2.36 mm), a slight reduction of about 2 percent in voids content was achieved, compared to the voids content of No. 57 aggregate.

Fine aggregate resulted in a voids content of 35.7 percent. Blending different proportions by volume of No. 57 and fine aggregates resulted in significantly lower voids content, i.e., No. 57 aggregate content between 50 and 61 percent resulted in average void contents ranging from 21.6 to 23.3 percent. In effect, when No. 57 and fine aggregates were combined, producing a combined aggregate size distribution of 1 in. (25 mm) to No. 200 sieve size (0.075 mm), a substantial reduction of about 17 percent in voids content was obtained, compared to the voids content when using No. 57 aggregate exclusively. Overall, a broader distribution of aggregate particle sizes, i.e., 1 in. (25 mm) to No. 200 sieve size (0.075 mm) versus 1 in. (25 mm) to No. 8 sieve size (2.36 mm), resulted in a greater reduction in voids content (about 17 percent versus 2 percent).

Additionally, the effect of WG on aggregate voids content was tested by considering two conditions: 1) Blend of three aggregates (No. 57, No. 8, and fine); and, 2) Blend of two aggregates (No. 57 and fine). As determined by CF and 8-18 charts, the blending of different proportions of No. 57, No. 8, and fine aggregates (Condition 1) by volume produced five WG gradings. The CF and 8-18 chart results of blending different proportions of No. 57 and fine aggregates (Condition 2) by volume Û typical of most ready mixed plants, which batch two aggregates in concrete Û indicated that all seven gradings were NWG.

Combined aggregate voids content tests for both conditions, totaling 12 combined aggregate gradings, clearly illustrated that WG yields slightly higher voids content as compared to NWG. An average voids ratio between 23.8 and 26.7 percent with an overall average of 25.5 percent was recorded for WG. The NWG demonstrated an average voids ratio between 21.6 and 23.3 percent with an overall average of 22.5 percent. Contrary to expectations, WG increased the voids content by an average of approximately 3 percent.

A round robin program was organized to evaluate the above-cited NRMCA Research Laboratory findings in other regions of the country using aggregates local to each area. Three NRMCA producer members participating in the research project included: 1) Titan America Technical Services, Jacksonville, Fla.; 2) HTC/Lehigh Research Facility, Atlanta, Ga.; and, 3) Aggregate Industries, Denver, Colo. Participants conducted basic aggregate tests including relative density, absorption, sieve analysis, and dry-rodded unit weight. For combining the aggregates, proportions were suggested by NRMCA on the basis of typical slab-on-grade mixture specifications, resulting in six aggregate proportions for each participant in order to evaluate both WG and NWG gradings. Included in the aggregate mineralogy were the following classifications :

  • Florida Û Coarse = crushed limestone; Fine = natural silica sand
  • Georgia Û Coarse = crushed limestone; Intermediate Coarse = crushed granite; Fine = alluvial concrete sand
  • Colorado Û Coarse = gravel; Fine = natural sand; all from the same source

At each location, intermediate coarse aggregate was obtained from the same quarry as the larger coarse aggregates. This protocol served to maintain consistent particle shape, discounting the influence that a different particle shape can have on test results, and provided a more realistic simulation of concrete plant operations.

A summary of data from the three locations indicates negligible difference in voids content between WG and NWG. From the four locations, including Maryland, data highlighting minimum, maximum, and average voids content attained for both WG and NWG was tabulated. Results indicated that combining aggregates to a maximum density as proposed by empirical methods such as the CF or the 8-18 charts does not have a significant impact on the voids content of the combined aggregate. In fact, as cited previously, Maryland yielded a higher voids content for WG than for NWG.

Yet, the research indicated some variation in voids content by location. In Florida, Maryland, Georgia, and Colorado, average voids contents were 23.5, 23.9, 25.7, and 27.9 percent, respectively. Since combined aggregate gradings as measured by the CF and 8-18 charts were similar, the variation in voids content most likely was due to differences in aggregate shape and texture. A subsequent Phase II study found that for a target 4-in. slump, Florida, Maryland, and Georgia aggregates required 290 lb/yd 3, 290 lb/yd 3, and 320 lb/yd 3, respectively. Thus, while variation in voids content may lead to differences in water demand, voids content variability can not be attributed to a difference in combined aggregate gradings.

Overall, aggregate tests from four sources (Md., Fla., Ga., and Colo.) indicated that WG does not lead necessarily to maximum aggregate packing density and, thereby, a minimum voids content considered conducive to improved workability, durability and economy in concrete mixes. No difference in voids content between the WG and NWG gradings was evident in three of the sources; at one site, NWG yielded about a 3 percent lower voids content.

PHASE B: CONCRETE PERFORMANCE

While Phase A research concluded that using Îwell graded combined aggregatesÌ (WG) does not necessarily lead to a reduction in voids content of the aggregate as compared to Înot well graded combined aggregatesÌ (NWG), field reports suggest that WG may help improve concrete performance. Accordingly, NRMCA’s Phase B program aimed to evaluate whether the use of WG improves fresh and hardened properties of concrete.

A major portion of the Phase B research, comprising four stages, was conducted at the NRMCA Research Laboratory in College Park, Md. Additionally, a planned testing program was conducted with local materials at sites involving two of the three labs participating in Phase A Û Titan America Technical Services and Heidelberg/Lehigh Research Facility.

The following materials were used in the experimental study conducted at Maryland:

  • ASTM C 150 Type I portland cement (Stage II high slump)
  • ASTM C 618 Class F fly ash
  • ASTM C 989 GGBF slag
  • ASTM C 260 tall oil air entraining admixture
  • ASTM C 494 Type A lignin-based WRA
  • ASTM C 494 Type F naphthalene sulfonate high range water reducing admixture
  • ASTM C 33 natural sand
  • ASTM C 33 No. 57 crushed limestone
  • ASTM C 33 No. 467 crushed limestone
  • ASTM C 33 No. 8 crushed limestone

Local materials were used at the two other locations. In all cases, intermediate coarse aggregate (No. 8, No. 89) was obtained from the same quarry as the larger coarse aggregates (No. 467, No. 57) to maintain consistent particle shape. Moreover, concrete mixes tested in Florida and Georgia were nonair-entrained, had a cementitous materials content of 517 lb/yd3, and did not contain any chemical admixtures or supplementary cementitious materials.

A 2.5-ft3 revolving drum mixer was used to blend the concrete. Except for 1.6 ft3 at Stage I, batch size was 1.0 ft3. Concrete was mixed in accordance with ASTM C 192, although two minutes beyond the recommended standard mixing time were added (during the first mixing cycle) when a HRWR admixture was introduced (Stage IV).

Following ASTM standardized procedures to the extent possible, basic concrete tests were conducted, such as slump (ASTM C 143), density (ASTM C 138), air content (ASTM C 231), 28-day compressive strength (ASTM C 39), shrinkage (ASTM C 157), and bleeding (ASTM 232). NRMCA engineering staff selected five aggregate proportions for each study participant to allow evaluation of both WG and NWG combinations.

In assessing the results to draw conclusions regarding the concrete performance of WG and NWG mixtures, data was culled from multiple scenarios: NRMCA/MD (Stages I-IV, including HS stages, denoting a higher slump mixture), plus FL and GA representing the experiments conducted at those locations. Each scenario contained several WG and NWG mixtures that were compared to the control ACI (NWG) mixture. Since every scenario did not include all the tests (e.g., FL, GA did not conduct finishability or segregation tests), comparisons were made only where applicable.

Water demand

Compared to the control (NWG) ACI mixture, test batches using WG resulted in similar slumps in 67 percent of the cases, while 33 percent of the test samples exhibited lower slumps. NWG mixtures resulted in higher slumps than the control when coarse aggregate content was higher than the control mixture. Two of three NWG mixtures demonstrating lower slumps had lower coarse aggregate content.

Entrapped air content and density

A slight increase in entrapped air content was observed when a greater quantity of fine aggregate was used. Otherwise, no noticeable change was observed among the mixtures in entrapped air content and density.

Bleed water

Compared to the control NWG mix, concrete mixtures using WG demonstrated similar bleeding in 75 percent of the cases and increased bleeding in 25 percent. NWG mixtures showed higher bleeding characteristics than the control when a higher coarse aggregate content was used. At low w/cm (0.40), the mixtures had less bleeding despite their high slump ratings (Stage IV).

Finishability

Reviewing the distribution of finishability ratings (FR) among the various scenarios, improved finishability was not evident for the WG mixtures compared to the NWG samples. For both the WG and NWG mixes, FR was either 2T-1 or 2T-2 in 80 percent of the batches evaluated. All WG mixtures assigned a higher FR value than the control had combined aggregate grading resulting in higher coarseness factor and lower workability (about 68/33). Most of the WG mixtures rating a similar FR had intermediate coarseness and workability factors (about 60/35). A low w/cm (Stage IV) led to deterioration in finishability despite exhibiting adequate slump, regardless of the aggregate grading. Finishability was marginally improved in such cases when the mixture’s coarse aggregate content was increased.

Compressive strength

Measured against the control, WG mixtures demonstrated similar strengths in 67 percent of the batches, lower strength in 22 percent of the cases, and higher strength in 11 percent. Overall, the differences were not highly significant.

Drying shrinkage

Compared to the control, WG mixtures exhibited similar shrinkage (±0.005 percent) in 92 percent of the batches, while 8 percent of the WG cases demonstrated marginally higher shrinkage. Given the precision of ASTM C 157, a difference in length change of about 0.01 percent is not considered statistically significant.

Segregation

All WG mixtures showing lower segregation than the control included a combined aggregate grading with intermediate coarseness and workability factors (about 60/35). For WG mixtures with a higher coarseness factor and lower workability factor (about 68/33), similar or higher segregation was evident. An extremely low w/cm led to lower segregation (Stage IV) in spite of high slump. NWG mixtures showed higher segregation than the control when they contained a higher coarse aggregate content.

For each scenario, with the exception of Florida, quantities of cementitious materials and mixing water were kept constant for all batches, and the resulting slump was measured as a relative indicator of the mixture’s water demand. Accordingly, a higher measured slump would suggest that water reductions are feasible for that mix. If WG batches had been designed at lower paste content (lower cementitious and lower water content) as compared to the control NWG-ACI mixture, they possibly could have exhibited lower shrinkage. However, in such a situation, the slump of the WG mixture would have been lower than that of the NWG-ACI control. Since concrete slump has to meet specifications, the only way to compensate for a lower slump is to increase water or water-reducing admixture dosage. While increasing water content increases the paste content and may lead to higher shrinkage and lower strength, reducing paste contents through the use of a water-reducing admixture is feasible for both WG and NWG mixtures.

In sum, test results from three locations (Md., Fla., Ga.) indicated that a concrete mixture containing Îwell-gradedÌ combined aggregates would not exhibit lower water demand, decreased bleeding, less shrinkage, or higher strength as compared to a mixture designed according to ACI 211 procedure that incorporates Înot well-gradedÌ aggregates. When NWG mixtures included a coarse aggregate content exceeding that required by ACI 211 (by 5 to 10 percent), the concrete tended to demonstrate a lower water demand, but increased bleeding and higher segregation. Thus, the researchers concluded that a concrete specification requiring WG through compliance with CF and/or 8-18 charts will not ensure the goals typical of such controls on aggregate grading, i.e., reduced mixing water content or lower shrinkage.

The report’s authors emphasize that their conclusion does not imply that aggregate grading is unimportant for concrete performance. If adequate fine material is not present, for example, concrete can be prone to segregation and high bleeding. On the other hand, excessive fines may create a sticky mix that is difficult to finish. While ACI 211.3R addresses such conditions Û Table 6.3.6 recommends compensating for a high proportion of fine aggregate material (resulting in lower fineness modulus) by increasing the amount of coarse aggregate Û the use of WG by application of CF and 8-18 charts is generally proposed as an improvement over the ACI 211 procedure. Although the CF and 8-18 charts did not achieve reductions in water demand or shrinkage, they did help attain better finishability and lower segregation relative to the control mixture, depending on where the WG mixture was located inside Zone II of the CF chart. At CF=60/WF=35, lower segregation was attained with similar finishability, whereas CF=68/WF=33 yielded better finishability with similar or higher segregation. The CF and 8-18 charts, therefore, can assist ready mixed concrete producers in evaluating whether such factors benefit concrete performance, given local materials and production constraints. Not to be invoked as project specification requirements, CF and 8-18 charts nonetheless are potential concrete mixture optimization tools.

ÎWGÌ VS. ÎNWGÌ CONTROL CONCRETE PERFORMANCE

Testing at three locations demonstrated the following results when concrete batches containing well-graded combined aggregates (WG) were compared against the control mixture for which aggregate contents were apportioned according to ACI 211 (NWG).

  1. Water demand: Similar in 67% and higher in 33% of test cases
  2. Bleeding water amount: Similar in 75% and higher in 25% of cases
  3. Compressive strength: Similar in 67% and lower in 22% of cases
  4. Shrinkage: Similar in 92% and higher in 8% of cases
  5. Finishability: Better FR with higher coarseness factor and lower workability factor (about 68/33); similar FR with intermediate coarseness factor and workability factor (about 60/35)
  6. Segregation: Lower segregation with intermediate coarseness and workability factors (about 60/35); similar or higher segregation with higher coarseness factor and lower workability factor (about 68/33)