Findings
Scaling resistance
Corrosion
Chloride ingress

Proposed Experiments
Proposed Experiment 6
Proposed Experiment 7

Scaling resistance
The severity of the Treat Island exposure is an ideal location to evaluate concrete's resistance to scaling in a harsh, "real-world" environment. All of Master Builders' experiments placed at Treat Island are monitored for scaling resistance.
_ One of the most significant findings so far is that the guidelines for air-void system parameters do not assure good scaling resistance in this environment, using the ASTM C672 visual scaling rating.
_ All of the specimens placed at the island had measured air content greater than 6%, but exhibited less resistance than expected.
_ The best performing plain hydraulic cement concrete contained Micro-Air_ air-entraining admixture. (3 year rating = 2)
_ The best performing specimens overall contained 8% silica fume (mass % of total cementitious material content). (3 year rating = 1).
_ Mass loss is not a good scaling measurement tool in this environment. The changes in moisture content due to the weather and the degree of cleanliness of the samples interferes with the measurements.
_ The bottom surface of the specimens seems to have more severe scaling than the rest of the specimens. This reduction in scaling resistance of primarily small specimens placed on the platform leads us to believe that the beams may be influenced by microbes, and the acids released from the wood as it decays.

Corrosion
The corrosion experiments placed at Treat Island were beneficial in terms of learning how not to perform corrosion experiments in this environment. However, the three experiments placed on the island produced no useful information concerning corrosion resistance due to specimen fabrication.
1. Placing the specimens on the platform creates a very tricky situation for taking periodic measurements because of the need for electrical contact with the reinforcing steel. Our first two corrosion experiments used laboratory style specimens with reinforcing bars that extended out from the concrete for the electrical connection. The bar ends were epoxy coated, and then sealed inside a rubber "boot". This boot did not protect the reinforcing bar ends in the marine environment. The rubber cracked, the seal was not perfect, and corrosion initiated under the epoxy on the ends. After two years exposure, corrosion under the epoxy caused half-cell potential measurements to drop below -300 mV CSE.
2. The specimens from the first two corrosion experiments (2 & 3) were removed from the island after three years exposure for an autopsy. Another experiment (5) took its place. The electrical connection for these specimens was designed as an internal connection between the black reinforcing steel and a rigid stainless steel wire. The ends were also painted with two coats of a two-part epoxy to prevent chloride entering the specimen along the wire. This approach also failed.
3. During the second and third year, half-cell potential measurements for Experiment 5 were < -350 mV CSE in all specimens. The potential was found highest near the ends of all specimens, indicating that galvanic corrosion may have initiated at the connection between the bars and the stainless steel wire. The connection used a zinc-plated screw to connect a grade 316 stainless steel wire to the black reinforcing bar, so there is potential for galvanic or crevice corrosion to initiate at this joint. Companion laboratory specimens showed corrosion initiated at the screw. The corrosion specimens for experiment 5 were removed in 2000 for an autopsy after 5 years exposure.

Back to Top

Chloride ingress
The most useful information obtained from the experiments at Treat Island so far has been related to chloride ingress.

Experiment 2 Findings:
The concrete mixture for this experiment had a 600 lb/yd3 cement content, and a 0.4 w/c ratio. The autopsy of corrosion specimens after 3 years shows that both corrosion-inhibiting admixtures reduced the apparent chloride diffusion coefficient. However, the apparent diffusion coefficient is only half of the story. The other parameter necessary to to describe chloride ingress is the surface chloride concentration. The chloride content profiles in the figure below shows that Rheocrete 222 (1 gal/yd3) reduced chloride ingress, but calcium nitrite (6 gal/yd3) actually increased chloride ingress.

Apparent chloride diffusion coefficients obtained after 3 years exposure

	Reference	Rheocrete 222	Calcium Nitrite
Cs, (ppm)	3196	3676	5093
Co, (ppm)	570	570	570
Da, (mm2/yr)	62	28	43
Da, (m2/sec)	1.96E-12	8.93E-13	1.37E-12

This apparent discrepancy comes from the way we calculate chloride diffusion coefficients from chloride profiles. The fitting routine has two degrees of freedom, meaning both the apparent diffusion coefficient, Da, and the surface concentration, Cs, are simultaneously adjusted to minimize the least square error fit to Fick's second law of diffusion. The background chloride content, Co, is the chloride initially present in the concrete materials. In this case, the background chloride is derived from the Great Lakes region limestone aggregate. This chloride is only released during acid digestion of the concrete powder.

Back to Top

Experiment 5 Findings:
This experiment contained eight concrete mixtures. The total cementitious materials content of 658 lb/yd3 and w/cm of 0.38 were held constant. The treatments were broken into two blocks, an 8% silica fume block and a OPC block. Both blocks contained reference concrete and concrete with Rheocrete 222+ dosed at 1 gal/yd3, 30 % calcium nitrite dosed at 6 gal/yd3, and an experimental admixture.
_ All specimens in the OPC concrete block exhibited moderate to high scaling so more chloride than expected entered these specimens. As a result,there was little difference in chloride ingress between treatments.
_ The experimental admixture specimens in both blocks exhibited severe scaling so they were removed from the island in 2000 to make room for new specimens.
_ The scaling resistance of the silica fume block was significantly better than the plain concrete block. The chloride ingress for this block is shown below.
_ The data for the silica fume block shows similar trends as those described for Experiment 2.
_ The laboratory determined diffusion coefficients are a factor of 2 to 4 less than the apparent diffusion coefficient determined from curve fitting the chloride profiles.

This finding illustrates the need for more research aimed at improving the correlation between laboratory-based predictions and field chloride-ingress measurements. The apparent chloride diffusion coefficients and chloride profiles after 2 years exposure are provided in the table and figures below.

Apparent Chloride Diffusion Coefficient Measurements: Experiment 5

	Diffusion Coefficient (mm2/yr)				Surface	Background	Age
Treatment	Flux_	Migration_	Migration*	Profile*	Cs (ppm)	Co (ppm)	t (yr)
OPC Ref	30	45	53	97	8626	660	2.0
Calcium Nitrite	42	60	58	80	7875	660	2.0
Rheocrete 222+	18	54	26	93	8867	660	2.0
Silica Fume Ref	11	17	30	59	7652	660	2.0
SF + CN	13	30	39	47	9663	660	2.0
SF + Rheocrete 222+	6	14	15	29	7661	660	2.0

* The migration tests and chloride profiles for these specimens came from 2 year Treat Island exposure blocks

^_ Flux and migration tests were performed on specimens cut from cylinders.

Back to Top

Proposed Experiment 6 - Pozzolan Durability Experiment

Objectives:
1. Determine the relative scaling resistance of concrete made with ProAsh processed Class F fly ash alone and when combined with Rheomac SF100 silica fume in a severe marine exposure.
2. Determine the effect of the Treat Island environment on chloride ingress and diffusion coefficients of air entrained concrete.
3. Compare the performance of concrete exposed to the ASTM C666 Procedure A with Treat Island Exposure.

Concrete Mixtures

		Treatments
Mix Date:	June, 2000	OPC Reference
Initial Exposure:	August, 2000	15% ProAsh
		25% ProAsh
cm:	658 lb./yd3	15% ProAsh + 3% Rheomac SF100
w/cm:	0.40	15% ProAsh + 6% Rheomac SF100
Slump:	4 to 7 inches	8% Rheomac SF100
Air Content:	7 to 9%
Coarse Agg.	#57 crushed limestone

Specimens cast per treatment:
4 - 12 x 12 x 6 inch deep chloride ingress blocks for profiles at 2, 5,10 and 15 years.
3 - 3 x 4 x 16 inch beams for ASTM C666 Procedure A (w/ ASTM C 672 scaling rating)
4 x 8 inch cylinders for compressive strength, RCP, and Bulk diffusion at various ages
1 - 3 x 4 x 16 inch beam for air-void system parameters (@ surface vs interior)

Curing: 14 days moist, then covered by wet rags and sealed in plastic bags for shipping to Eastport, ME.

Back to Top

Proposed Experiment 7 - Air Entraining Admixture Experiment

Objectives:

1. Determine the relative scaling resistance of plain concrete made with vinsol resin and synthetic air-entraining admixtures.
2. Determine the relative scaling resistance of concrete made with polycarboxylate high-range water-reducing admixtures combined with vinsol resin and synthetic air-entraining admixtures.
3. Compare the performance of concrete exposed to the ASTM C666 Procedure A with Treat Island Exposure.

Concrete Mixtures

Specimens cast per treatment:
3 - 6 x 6x 21 inch beams for Treat Island Exposure
3 - 3 x 4 x 16 inch beams for ASTM C666 Procedure A (w/ ASTM C672 scaling rating)
6 - 4 x 8 inch cylinders for compressive strength
1 - 3 x 4 x 16 inch beam for air-void system parameters
Curing: 14 days moist, then covered by wet rags and sealed in plastic bags for shipping to Eastport, ME.

Back to Top

Home | Programs | Rack Maps | Keywords