December, 2000 Special Technical Issue

Concrete Advantage

LOW SHRINKAGE CONCRETE RESEARCH UPDATE

Michael J. Ramerth, P.E.

Meyer, Borgman & Johnson

This article presents the results of our ongoing research with respect to producing a low-shrinkage concrete. The focus of our research is the effectiveness of using a well-graded 1½” minus aggregate. The following test results refer to this well-graded aggregate as an “8-18 gradation.” Our goal for the optimized gradation has been to fall within the % retained ranges shown on Figures 1 & 2. These %-retained ranges are based on previous research including a 1974 study by James Shilstone for the U.S. Army Corps. In addition, Figures 1& 2 are currently a part of Mn/DOT’s specification as an optional gradation incentive program.

The well graded aggregate (8-18 gradation) reduces the total surface area of the aggregate, thus reducing the water demand of the aggregate and the total amount of water required to produce a yard of contractor friendly concrete. The water content is well known as an important factor in setting the shrinkage potential of the concrete mixture.

This research as well as future MCC research is intended to quantify the effectiveness of various pozzolan blends with the 8-18 gradation aggregate, based on and their drying shrinkage performance.

INTRODUCTION

Cracking of concrete slabs on grade, walls and structural decks continues to be an industry nuisance. Therefore, finding a concrete recipe with a significantly lower shrinkage potential would be a boon to the industry and result in even higher value cast-in-place concrete structures. The purpose of these additional shrinkage tests was two-fold. First, to substantiate the results of the 1997 shrinkage tests focusing on the use of a well graded aggregate (8-18 gradation), water reducers to achieve a maximum 0.40 w/c, and a minimum amount of Portland cement as a strategy to significantly reduce drying shrinkage. Second, to test the shrinkage performance of a somewhat typical cast-in-place parking deck mix which contains 3 gallons of DCI, 5% microsilica, 10% fly ash, a w/c of 0.4 and the well graded aggregate (8-18 gradation). The premise of this research is to compare the shrinkage performance of this somewhat typical parking deck mix to that of our base mix (#6005). The comparison would hopefully shed some light on the shrinkage effects of a ternary pozzolan blend and/or other admixtures.

TESTING PROCEDURES AND RESULTS

In the spring of 1999, MCC sponsored two additional shrinkage evaluation programs on a specific construction project. These shrinkage tests are part of MCC’s ongoing mission to improve and fine-tune the performance of cast-in-place concrete for the betterment of the construction industry. Both tests were performed in general accordance with ASTM C157. The first of these tests was initiated on March 31, 1999 and conducted by American Engineering and Testing (AET). The concrete was taken from 3 separate placements of the Minnesota Life Phase II Parking Decks. The concrete was provided by Cemstone Concrete Products, using mix #6005 as described below:

Lafarge’s Type I Portland Cement	525 pcy
Mineral Solutions Fly Ash	125 pcy
Aggregate Industries’ Nelson Gravel – 1½”	741 pcy
Aggregate Industries’ Crushed Dolomite – ¾”	1090 pcy
Aggregate Industries’ Safety Grit	824 pcy
Aggregate Industries’ Nelson Sand	471 pcy
Grace’s WRDA 82	20 ocy
Water	260 pcy
W/Cm	0.40

Note: ocy = ounces per cubic yard and pcy = pounds per cubic yard

Non-chloride accelerator was added to the concrete for Sets #1 and #3 at 0.5% calcium chloride equivalent.

Shrinkage beams were cast from the plastic concrete as part of the normal sampling procedure and tested per ASTM:C157. The 4” x 4” x 11” samples were stripped at an age of 24 hours and the initial length measured. A second reading was taken after 27 days soaking in lime-saturated water. The beams were then placed in 50% relative humidity at 70°F and length change was measured periodically and calculated as the percentage difference from the original length. The following results were obtained:

MCC’s 1999 Mn. Life Shrinkage Test Results

Set No.	1	2	3	4	Average
Test Sample Date:	3/31/99	3/31/99	4/28/99	5/4/99
50% RH – 4 day	-.006	-.006	+.003	-.001	-.0025
50% RH – 8 day	-.016	-.015	-.011	-.022	-.016
50% RH – 1 month	-.029	-.025	-.025	-.029	-.027
50% RH – 2 months	-.042	-.039	-.035	-.038	-.038
50% RH – 4 months	-.047	-.041	-.042	-.049	-.045
50% RH – 8 months	-.048	-.043	-.046	-.050	-.047
50% RH – 16 months	-.049	-.043	-.046	-.050	-.047

The second of these tests was initiated on April 7, 1999. The concrete was taken from a single placement of the Northwest Airline Employee Parking Deck. The concrete was produced by Aggregate Industries – Model Ready Mix Division, using mix # 603AOM as described below:

Lafarge’s Type I Portland Cement, ASTM:C150	560 pcy
Portage 2 Fly Ash, ASTM:C618	67 pcy
Master Builders – MBSF, ASTM:C1240	33 pcy
Calcium Nitrite Corrosion Inhibitor DCI, ASTM:C494, Type C	3 gcy
Aggregate Industries – 1½” ASTM:C33	650 pcy
Aggregate Industries – ¾” ASTM:C33	600 pcy
Aggregate Industries – 3/8” ASTM:C33	450 pcy
Coarse Sand, ASTM:C33	900 pcy
Fine Sand, ASTM:C33	400 pcy
Eucon 37, ASTM:C494	As needed
Water	264 pcy
W/Cm	0.40 max

Note: gcy = gallons per cubic yard

The Length Change of Hardened Concrete tests were run in general accordance with ASTM C 157. Being placed into 10-inch molds, the concrete samples were cured in a controlled environment for 24 hours. The samples were then removed from the molds, and the initial length was recorded. The samples were placed in a lime-saturated bath for 27 days and air dried for comparative measurements which were taken at 5, 7, 14, 22 and 33 days. The length changes have been calculated as the percent difference from the original length.

MCC’s 1999 NWA Shrinkage Test Results

Set No.	1	2	3	Average
Test Sample Date	4/7/99
5 days	-.032	-.037	-.037	-.035
7 days	-.039	-.038	-.038	-.038
14 days	-.051	-.048	-.048	-.049
22 days	-.058	-.056	-.050	-.055
33 days	-.070	-.063	-.053	-.062
180 days	-.082	-.066	-.062	-.070

The following (Figure #3) is a composite graph of the shrinkage test results indicating the average length change as a percentage for the 1997 MCC study and both of the 1999 test results.

DISCUSSIONS/CONCLUSIONS

Once again the shrinkage test results for the basic 8-18 mix # 6005 were very good, yielding an average length change of -.047. This result is consistent with the average result of the previous test (1997) -.045. I believe that this most recent test substantiates the effectiveness of the 8-18 aggregate mix # 6005 through a lowering of the water demand resulting in only 260 lbs of water in a cubic yard of concrete.

The shrinkage results for the parking deck mix with the 8-18 aggregate were nearly 49% higher than those of the mix # 6005 at 180 days, with an average length change of -.070. An even more dramatic difference was observed in the first five days of the test. At 5 days, the NWA deck mix had experienced shrinkage levels nearly 5 times higher than that of those of the Mn. Life mix. This high early onset shrinkage did manifest itself on the NWA decks and in some cases we observed shrinkage cracks in the deck slabs prior to covering the slabs with insulated blankets the day of the concrete placement.

Given the rigid environmental standards of the ASTM C:157 testing procedures, we believe that the explanation for this significantly higher drying shrinkage must lie in an examination of the two mixes ingredients and their proportions. First we compared the total amount of water in these two mixes because water is one of the key factors in drying shrinkage. The total water content of these two mixes was very close with 264 lbs in the higher shrinkage NWA mix and 260 lbs of water in the Mn. Life mix. Given this slight difference of four pounds, we believe that only a minor portion of the increased shrinkage is due to the higher water content.

Next, we compared the total pozzolan content of the two mixes and found, as expected, that they too were very close in total weight with 660 pounds in the NWA mix and 650 pounds in the Mn. Life mix.

NWA (# 603AOM)		Mn. Life (# 6005)
Type I Portland	560 pcy	Type I Portland	525 pcy
Fly Ash	67 pcy	Fly Ash	125 pcy
Microsilica	33 pcy
Total:	660 pcy	Total:	650 pcy

However, the composition of these pozzolan blends is significantly different and could partially explain the difference in the shrinkage performance of these two mixes.

Another obvious difference between these two mix designs is that the NWA mix contains 3 gallons of DCI, a calcium nitrite based corrosion inhibitor, and the Mn. Life mix contains no corrosion inhibitor. It is widely known that calcium nitrite acts as an accelerator in the concrete, and less widely known that accelerators have been shown to increase drying shrinkage due to a higher degree of hydration. This leads us to believe that the acceleration effect of the calcium nitrite may have been the primary contributor to the dramatic difference in the early onset shrinkage performance of the NWA concrete.

The two mixes also contained different water reducing admixtures. Unfortunately, we don’t have enough information to draw any conclusions.

Both of the aggregate blends for these two mixes have essentially met the 8-18 gradation requirements. (Refer to figures 1&2). Thus, we do not believe that the aggregate played a significant role in the higher shrinkage performance of the NWA mix, although the Mn. Life aggregate is more uniformly graded than then NWA aggregate. (Refer to figures 1 & 2. Note that the gradation plot is smoother for the Mn. Life aggregate.)

SUMMARY

We believe that these 1999 shrinkage tests substantiate the effectiveness of the 8-18 gradation in reducing shrinkage. These tests also dramatically illustrate that significant differences in performance may result when one modifies the pozzolan blend or the admixtures in a mix design. Perhaps future testing sponsored by the MCC could focus on the specific effects of various pozzolan blends and admixtures.

The MCC has sponsored an additional shrinkage test, which is currently underway. This MCC 2000 shrinkage test is studying a concrete mix which includes the 8-18 gradation and a ternary pozzolan blend containing 25% Slag. ( G.G.B.F.S.), 10% Fly Ash and 65% Type I Portland. The results of this test will be published in 2001.