B08_Section_J_Attachment_4_Ft_Jeff_Counterscarp_Materials_Test.pdf

Atkinson-Noland & Associates

2619 Spruce Street

Boulder, CO 80302

303.444.3620

32 Old Slip, 10th Floor

New York, NY 10005

917.647.9530 ana-usa.com

REPORT

MATERIALS TESTING

Counterscarp Wall Ft. Jefferson National Monument Dry Tortugas National Park Florida

PREPARED FOR:

Josiah M. Berg, PE

HDR

8333 NW 53rd St., Suite 302 Doral, FL 33166 D 305.728.7442 M 786.863.0000 Josiah.Berg@hdrinc.com

PREPARED BY:

David Woodham, P.E.

Namratha Kondam, E.I.T.

Atkinson-Noland & Associates, Inc.

Ft. Jefferson Counterscarp Materials Tests

ANA Job No. 21-136

September 7, 2022

Pg. 2 9/7/2022

1. INTRODUCTION

At the request of representatives of HDR, Atkinson-Noland & Associates (ANA) was retained to conduct material tests on mortar and brick samples from the counterscarp wall at Ft. Jefferson in Dry Tortugas

National Park in Florida. The objective was to determine the water vapor transmission (WVT) rate for the brick samples, the 24, 48 and 72 hour cold water absorption, the compressive strength and the thermal and hygric coefficients of expansion to assist in evaluating potential replacement brick for localized rebuilding of the counterscarp wall. Mortar samples from the counterscarp wall were evaluated to determine the probable binder/aggregate ratio and the aggregate gradation and color. In addition, one mortar sample was submitted to DRP Inc. for petrographic analysis.

Samples were obtained during a site visit to Ft. Jefferson by conservator John Feinberg of the

Collaborative, inc. (tCi) and engineer Dave Woodham of Atkinson-Noland & Associates (ANA) on April

25-27, 2022. Brick and mortar samples were individually bubble wrapped and shipped from Key West to

ANA’s laboratory in Colorado.

2. MATERIALS TESTING

2.1. Brick Coefficient of Thermal Expansion

Each brick was instrumented to help determine the coefficient of thermal expansion. Two brass gage point were installed along the length of each brick and measurements between the gage points were obtained using the Whittemore gage shown in Figure 1 below. The measurements of the two gage point sets were then averaged. The gage is capable of measuring to the nearest ten-thousandth of an inch and is calibrated prior to each reading by taking a reference reading on an Invar alloy bar to reduce thermal drift due to length change of the Whittemore gauge frame. Length measurements were made at 0 ºF, 40

ºF , 84 ºF, 115 ºF, 156 ºF and 180 ºF and plotted in MS Excel. The slope of the linear curve, as shown in

Figures 2 through 4, produced the coefficient of thermal expansion for each brick tested as listed in Table

1.

Figure 1. Using the Whittemore gage to measure the length of the clay brick unit at various temperatures.

Pg. 3 9/7/2022

Figure 2. Length change vs temperature for both gage point sets for brick samples B1 through B4.

Figure 3. Length change vs temperature for both gage point sets for brick samples B5 through B8.

Pg. 4 9/7/2022

Figure 4. Length change vs temperature for both gage point sets for brick samples B9 through B12.

Table 1. Thermal Expansion Test Results

Sample ID Coefficient of Thermal Expansion

(in./in. • ºF)

B1 3.1 x 10-6

B2 3.7 x 10-6

B3 4.4 x 10-6

B4 4.1 x 10-6

B5 4.2 x 10-6

B6 4.2 x 10-6

B7 4.2 x 10-6

B8 2.7 x 10-6

B9 3.8 x 10-6

B10 5.5 x 10-6

B11 4.1 x 10-6

B12 4.4 x 10-6

Average 4.0 x 10-6

Pg. 5 9/7/2022

Typical calculation for determining the coefficient of thermal expansion:

Coefficient of thermal expansion ) x T/( gll =

Where:

Δ l is the length change, ΔT is the temperature range, lg is the gage length over which the length change is measured.

2.2. Brick Coefficient of Hygric Expansion

Similar to the determination of the coefficient of thermal expansion, we used the Whittemore gage on the gage points installed in each brick to measure changes in length as the brick became completely saturated. Length measurements were made at oven-dry room temperature and after 24, 48 and 72 hours of submersion in room-temperature water. Figures 5 through 7 show the measurements and Table

2 lists the calculated coefficients of hygric expansion from dry to fully saturated.

Figure 5. Length change vs time of submersion in room temperature water for brick samples B1 through

B4.

Pg. 6 9/7/2022

Figure 6. Length change vs time of submersion in room temperature water for brick samples B5 through

B8.

Figure 7. Length change vs time of submersion in room temperature water for brick samples B9 through

B12.

Pg. 7 9/7/2022

Table 2. Hygric Expansion Test Results

Sample ID Coefficient of Hygric Expansion

(in./in.)

B1 7.1 x 10-4

B2 2.9 x 10-4

B3 3.3 x 10-4

B4 2.1 x 10-4

B5 1.3 x 10-4

B6 1.4 x 10-4

B7 4.2 x 10-4

B8 3.3 x 10-4

B9 4.9 x 10-4

B10 3.4 x 10-4

B11 1.5 x 10-4

B12 1.2 x 10-4

Average 2.7 x 10-4

2.3 Brick Cold Water Absorption

The bricks were weighed after 38-hour and 48- hour submersion in water to measure the weight gain of the clay brick unit when submerged in room temperature water. See Table 3 for the test results.

Pg. 8 9/7/2022

Table 3. Cold Water Absorption Test Results

Sample

ID

Dry Weight

(g)

Cold Water

Saturated

Weight 38 hr., (g)

Cold Water

Saturated

Weight 48 hr., (g)

Cold Water

Absorption 38 hr.

Cold Water

Absorption 48 hr.

B1 2017.5 2206.0 2212.6 9.3% 9.7%

B2 2001.9 2111.7 2111.2 5.5% 5.5%

B3 1849.7 2016.8 2024.0 9.0% 9.4%

B4 1798.9 1953.0 1953.6 8.6% 8.6%

B5A 1863.1 1866.7 1867.0 0.2% 0.2%

B6 1926.1 2065.4 2066.3 7.2% 7.3%

B7A 2040.7 2203.3 2204.2 8.0% 8.0%

B8 1995.0 2166.7 2167.6 8.6% 8.7%

B9 2007.3 2165.9 2166.5 7.9% 7.9%

B10 1803.2 1924.6 1924.7 6.7% 6.7%

B11 1833.2 2001.2 2001.2 9.2% 9.2%

B12 1647.5 1785.5 1786.3 8.4% 8.4%

Average 8.0% 8.1%

Standard Deviation 1.2% 1.3%

Coefficient of Variation 15.3% 16.1%

A The values for bricks B5 and B7 are not included in the average, standard deviation and coefficient of variation.

2.4 Brick Compressive Strength

Brick for compression testing were sampled from the 12 bricks, based on those observed to be most representative of the original and replacement brick used in the counterscarp wall. Tests were conducted in accordance with ASTM C67, Standard Test Methods for Sampling and Testing Brick and Structural Clay

Tile.

Compressive strength results are reported in Table 4 to assist the design team in making decisions on compatible replacement brick for repairs and localized rebuilding. The compressive strength of replacement brick should roughly correspond to the strength of the original clay brick units.

Pg. 9 9/7/2022

Table 4. Compressive Strength Results

Sample ID

Width

(in.)

Length

Height

Area

(in2)

Maximum Load

(lb.)

Compressive

Strength

(psi)

B2 3.4 3.3 2.3 11.22 57,161 5,095

B7 A 3.6 3.5 2.3 12.60 83,795 6,650

B9 3.6 3.6 2.4 12.96 57,886 4,467

B11 3.6 3.35 2.3 12.06 29,895 2,479

B12 3.5 3.25 2.2 11.38 40,289 3,542

B5 A 3.4 3.35 2.2 11.39 136,451 11,980

Average 3,895 B

Standard Deviation 1,140 B

Coefficient of Variation 29% B

A Relatively newer bricks not historic bricks

B The calculations for average, standard deviation and coefficient of variation excluded the values of brick B7 and

B5 as they are assumed to be non-historic bricks.

2.5 Brick Water Vapor Transmission

Brick samples were prepared for water vapor transmission (WVT) testing after thermal and hygric expansion had been determined for the samples. Tests were conducted in accordance with ASTM E96, Standard Test Methods for Water Vapor Transmission of Materials using the wet cup method. One sample from each of five representative brick was sawn with a water-cooled masonry saw. The sawn samples were selected by avoiding areas with obvious inclusions that were not representative of the general matrix of the brick cross section. The samples obtained were approximately 2.1” by 2.1” by 0.45” thick.

Samples were inserted into square plastic containers and sealed with silicone at all perimeter edges as shown in Figure 8. Water was introduced into the wet cup prior to sealing the samples. All samples were placed in an environmental chamber with conditions maintained at 90 ºF and relative humidity of 50% and weighed daily to determine the water vapor transmission through the samples.

Pg. 10 9/7/2022

Figure 8. Typical sample for determining water vapor transmission.

Figure 9. Water weight loss (vertical axis) versus time (horizontal axis). The slope of the line is used to calculate the water vapor transmission rate.

Pg. 11 9/7/2022

Table 5. Results from the water vapor transmission testing.

ID

WVT

(g/h • m2)

Permeance

(Perm)

Thickness

Permeability

(Perm-in.)

B10 11.6 28.9 0.40 11.6

B1 8.0 20.1 0.45 9.0

B13 9.7 24.2 0.45 10.9

B6 7.9 19.7 0.45 8.9

B8 9.6 23.9 0.45 10.7

Average 10.2

Standard Deviation 1.20

Coefficient of Variation 11.8%

2.6 Mortar Analysis

Ten mortar samples were removed by Atkinson-Noland & Associates from the counterscarp wall for analysis. One sample was submitted for petrographic analysis and nine samples were tested by

ANA for mortar composition. The objective was to identify binder/aggregate ratio, aggregate color, and aggregate size gradation to provide an appropriate compatible replacement mortar formulation.

Aggregate Sieve Analysis

The mortar evaluation followed the requirements of ASTM C136, Standard Test Method for Sieve Analysis of Fine and Coarse Aggregates. This method is based on the use of acid digestion of the binder and sieve analysis of the aggregate. The analysis process used does not distinguish between cement and lime-based binders. The sample identification list is shown in Table 6.

Table 6. Sample identification and location.

Sample ID Location

M1 Loc. B Counterscarp Wall North side – 7 courses down

M2 Loc. B Counterscarp Wall – 5 courses down

M3* 57’ North of Loc. B – setting mortar under top header course

M4 Loc. B – taken from brick 1

M5 80’ South of breach – 2nd course bedding rebuild

M6 Counterscarp wall along Front 3

M7 At area with missing top slabs

M8 At section with mission top slabs

M9 Near North beach middle of North Front – 3rd course down

* M3 sample was not a historic mortar but was a modern mortar using white portland cement. It was tested only because it is prevalent in areas previously rebuilt.

Pg. 12 9/7/2022

The results of acid digestion of the sample are shown in Table 7. For all mortar samples, the aggregate composition is largely carbonate and was therefore partially digested by the acid. As such, the binder to aggregate ratios are skewed.

The aggregate gradation curve, plotted in Figure 10, shows that the remaining aggregates from the mortars do not fall within the gradation range of coarse and fine aggregates as specified by ASTM C144, Standard Specification for Aggregate for Masonry Mortars. Again, the gradation is skewed by the dissolution of some of the aggregate by acid digestion.

The aggregate size distribution and colors are shown in Figure 10. Ideally, the aggregate for the replacement mortar should match the color and gradation of the existing aggregates.

Mortar used for repointing joints, crack repair, and rebuilding should meet the requirements of ASTM

C270, Standard Specification for Mortar for Unit Masonry, for Type N mortar, except that a natural cement would be used in lieu of portland cement. The volumetric proportions should be 1 part natural cement ,

0.75 part hydrated lime, and 4.5 to 5 parts washed beach sand.

The use of pigments may be required to match the hardened mortar color. Pigments conforming to

ASTM C979, Standard Specification for Pigments for Integrally Colored Concrete, are suitable for mortar, but should not exceed 5% by weight of binder content in the mortar. Trial mixtures may be required to arrive at a mix that matches the original mortar with respect to color and texture.

Table 7. Results of acid digestion of mortar samples.

ID.

Mass before acid digestion

(g)

Mass after acid digestion

(g)

Binder mass

(g)

Aggregate mass (g)

Binder volume

(cm3)

Aggregate volume

(cm3)

Total volume

(cm3)

Binder

Aggregate

M1 49.11 8.20 40.91 8.20 63.85 6.40 70.25 91 9

M2 51.76 9.66 42.10 9.66 65.71 7.54 73.24 90 10

M3 53.55 7.96 45.59 7.96 71.15 6.21 77.36 92 8

M4 47.66 3.43 44.23 3.43 69.03 2.68 71.71 96 4

M5 63.62 3.26 60.36 3.26 94.20 2.54 96.75 97 3

M6 61.12 8.11 53.01 8.11 82.73 6.33 89.06 93 7

M7 60.11 9.17 50.94 9.17 79.50 7.16 86.66 92 8

M8 60.40 10.26 50.14 10.26 78.25 8.01 86.26 91 9

M9 59.90 9.19 50.71 9.19 79.14 7.17 86.31 92 8

*Assumes binder only consists of hydrated lime with a bulk density of 0.641 g/cm3 (40lb/ft3).

Pg. 13 9/7/2022

Figure 10. Aggregate distribution by sieve size for mortar samples.

Pg. 14 9/7/2022

Figure 11. Aggregate gradation and color for mortar samples M1 through M6.

Sieve Size: 4 8 16 30 50 100 200 Pan

M1

M2

M3

M4

M5

M6

Pg. 15 9/7/2022

Figure 12. Aggregate distribution and color for mortar samples M7 through M9.

2.7 Coral Concrete Testing

Testing of the coral concrete was conducted previously by Atkinson-Noland & Associates and included in this report so that all the material tests can be found in one document. Previous tests were conducted for

Silman engineers in 2012. Three in situ flatjack tests were performed in the coral concrete at Ft.

Jefferson.

Compression behavior of materials used in the construction of existing walls at Ft. Jefferson was investigated using the flatjack method as described ASTM C1197, In Situ Measurement of Masonry

Deformability Properties Using the Flatjack Method. A typical test setup is shown in Figure 13. In coral concrete, two parallel slots were cut approx. 12 inches apart for flatjack insertion. As the flatjacks are pressurized, the corresponding deformations of the material between the jacks are measured using a set of surface-mounted linear variable differential transformers (LVDTs). As a general approach, two or more cycles of loading were conducted for each test. The initial cycle is used to seat the flatjacks and the following cycles provide a more accurate measure of compression behavior of the material. Applied pressures of approximately 100 psi were used in the first cycle for the coral concrete.

Compressive stiffness was calculated as the slope of the stress-strain curve for the linear elastic portion of the test. The initial portion of the stress-strain curve can be nonlinear due to flatjack deformation at low pressures and closing of small cracks; this behavior was neglected in calculating stiffness. In the case of coral concrete, each test produced compression failure in the material and compressive strength and stiffness were directly derived from test data.

M7

M8

M9

Pg. 16 9/7/2022

Figure 13. Test setup for the flatjack deformability test.

Figure 14. Stress-strain curve and test setup from Deformability Test 3 – Coral concrete. Post-peak behavior described by the softening branch of the curve.

0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 0.0060 0.0070 0.0080 0.0090 0.0100

St re ss (p si

Strain (in/in)

DEFORMABILITY TEST #3

Pg. 17 9/7/2022

Figure 15. Stress-strain curve and test setup from Deformability Test 4. – Coral concrete. Test stopped after graph reversed direction due to material failure.

Figure 16. Stress-strain curve and test setup from Deformability Test 5. – Coral concrete. Test stopped after graph reversed direction due to material failure.

0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 0.0060 0.0070 0.0080 0.0090 0.0100

St re ss (p si

Strain (in/in)

DEFORMABILITY TEST #4

0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 0.0060 0.0070 0.0080 0.0090 0.0100

St re ss (p si

Strain (in/in)

DEFORMABILITY TEST #5

Pg. 18 9/7/2022

Table 8. Results from flatjack deformability tests.

Test No.

Location

Material

Calculated

Compression

Modulus (psi)

Estimated

Compressive

Strength (psi)

Bastion 2 upper level Coral concrete 79,400 298

Front 3 upper level Coral concrete 349,000 360

Front 3 upper level Coral concrete 275,000 288

Average 253.030 320

Standard Deviation 150,650 40

Coefficient of Variation 60% 13%

Coral concrete was loaded to failure during each of the three tests, and both ultimate compressive strength and elastic modulus were directly derived from test data. Consistent results were obtained in terms of compressive strength, while a higher degree of variation was observed in terms of compressive modulus. An average compressive strength of 320 psi was found. Elastic modulus measured from test data was in the range of 79,400 psi to 349,000 psi. The high variability in stiffness under compression load may be related to different levels of material compaction and size of corals used as aggregate in the concrete matrix. Test data variability observed during testing suggests a wide variability in construction quality for the coral concrete. As a result, a lower-bound value of approximately 102,400 psi (mean value minus one standard deviation) is recommended as a conservative estimate of compression modulus.

3. SUMMARY

Tests were conducted on the fired clay brick and mortar sampled from the Ft. Jefferson counterscarp wall. This report also includes the results from previous tests done on the coral concrete at Ft. Jefferson.

The average coefficient of thermal expansion for the brick was determined to be 4.0 x 10-6 in./in. • ºF or

4.0 parts per million per degree Fahrenheit. This is a typical value for clay masonry units. The current masonry code in the US is TMS 402/602-16, Building Code Requirements for Masonry Structures and

Commentary, which lists the design value for the coefficient of thermal expansion for clay masonry as 4.0 x 10-6 in./in. • ºF in Table 4.2.3.

Hygric expansion from oven dry to saturation at 72 hours averaged 2.7 in./in. x 10-4 but ranged considerably from 1.2 x 10-4 in./in. to 7.1 x 10-4 in./in. Note that the average hygric expansion is two orders of magnitude greater than the coefficient of thermal expansion.

Pg. 19 9/7/2022

The cold water absorption of the brick averaged 8.0% and 8.1% for 38-hour and 48-hour submersion respectively. The results are similar to what we have measured with clay brick units from other historic structures.

The average compressive strength of selected brick was found to be 3,895 pounds per square inch. This is slightly higher than expected given the age of the structure and continuous moisture exposure of the scarp wall.

Permeability of coupons sawn from selected clay brick was found to be 10.2 Perm-in. on average. Again, this permeability value is consistent with tests on from clay brick unit obtained from other historic buildings in the US.

Mortar testing determined that mortar repointing joints, crack repair, and rebuilding should meet the requirements of ASTM C270, Standard Specification for Mortar for Unit Masonry, for Type N mortar, except that a natural cement should be used in lieu of portland cement. The volumetric proportions should be 1 part natural cement , 0.75 part hydrated lime, and 4.5 to 5 parts washed beach sand.

For the tests on coral concrete, we found an average compressive strength of 320 pounds per square inch

(psi). The elastic modulus measured from the test data was in the range of 79,400 psi to 349,000 psi. The high variability in stiffness under compression load is possibly related to different levels of material compaction and size of corals used as aggregate in the concrete matrix. As a result, a lower-bound value of approximately 102,400 psi (mean value minus one standard deviation) is recommended as a conservative estimate of compression modulus.