200408 HUD Mortar Analysis and Permeability Study MTFA D P.pdf

PREPARED FOR:

McMullan and Associates 11800 Sunrise Valley Drive #430 Reston, VA 20191

Robert C. Weaver Federal Building:

Housing and Urban Developement

Mortar Analysis and Permeability Study

April 2020

Robert C. Weaver Federal Building (March 2020)

ROBERT C. WEAVER FEDERAL BUILDING

Table of Contents

1.0 Introduction 1

1.1 Purpose of Investigation 1

1.2 Methodology 2

1.3 Document Review 2

2.0 Field Testing 5

3.1 Field Observations 5

3.4 Permeability Test - RILEM Test Method 11.4 5

3.0 Laboratory Analysis 7

4.1 Petrographic Analysis: Stone 7

4.2 Petrographic Analysis: Mortar 7

4.3 Chloride Analysis 9

4.0 Joint Detail Analysis 11

5.0 Recommendations 13

Appendix A: RILEM Test Method 11.4 Permeability Charts Appendix B: Laboratory Analyses of a Stone and Masonry Mortars from Robert C Weaver Federal Building in Washington DC iii

MORTAR ANALYSIS AND PERMEABILITY STUDY

This page intentionally left blank

1.0 Introduction

1.1 Purpose of Investigation

In March 2020, MTFA Design + Preservation (MTFA D+P) analyzed existing conditions of mortar repointing at the southwest node of the Robert C. Weaver Federal Building for Housing and Urban Development (HUD) as part of the ongoing façade restoration. In particular, MTFA D+P was tasked with determining the source of suspected efflorescence of salts that had appeared across large expanses of newly repointed stone clad-ding. GSA and McMullan Associates requested MTFA D+P conduct a comprehensive analysis of the existing and replacement mortar to understand why the new mortar was staining the granite. The analysis included mortar sampling, laboratory testing, and onsite test-ing and observations. The project team extracted five samples for analysis including two original mortar, two new mortars, and a granite sample. The following tests were completed:

• ASTM C1721 Standard Test for Petrographic Ex-amination of Dimension Stone

• ASTM C1324 Standard Test Method for Examina-tion and Analysis of Hardened Masonry Mortar.

• Ion Chromatography for Analyzing Chlorides

• X-Ray Diffraction for Salt Mineral Identification

• RILEM Test method 11.4 E Water Absorption Un-der Low Pressure (Pipe Method) for Permeability

• Fourier Transform Infrared Spectroscopy (FTIR) to detect organic binders

• Acid Digestion of Masonry Mortar with Sand Ex-traction 1

Combining onsite testing and laboratory analysis, MTFA D+P identified deficiencies in the historic mor-tar and new mortars. These deficiencies combined with installation techniques altered from the original have contributed to the ongoing staining problem.

1 MTFA D+P to complete after Pandemic.

Mortar Sample Location, Southeast Node (March 2020)

With the results of this diagnostic study, MTFA D+P has included recommendations in this report for future repointing efforts.

1.2 Methodology

MTFA D+P met with representatives from GSA and McMullan Associates to understand the scope of the mortar problems and determine which diagnostic tools would be most successful. Prior to this investigation, the contractor had removed all original stone clad-ding on the SW node and reinstalled the stones using a Type N cement-lime mortar. Soon after reinstallation, white staining appeared across the entire façade. The general team consensus was that excessive water was penetrating the wall system and releasing through the face of the mortared joints. As the water escaped the joint, it likely transferred some component of the new mortar onto the stone face, creating a white haze be-low the joints and especially below vertical joints. The contractor removed the new mortar from all joints and installed mockups of different mortar types to evaluate their performance prior to any wholesale repointing.

McMullan Associates then brought MTFA D+P onto the project to investigate the source of this staining.

In their analysis, MTFA D+P considered the following questions:

• With a cavity wall construction, why is water com-ing from the back side of the cavity wall through the face of the mortar joint?

• If granite has low absorption and the mortar is new, why is water entering the wall system?

• Why is there a white haze at the drips below the joints? Is there salt in the granite, new mortar, or both?

• How does GSA resolve either the wall assembly or the mortar components to prevent future staining while stabilizing the walls?

MTFA D+P conducted onsite investigations and analy-sis of existing materials to understand both current conditions and the original design intent. Following this preliminary assessment, McMullan Associates collected samples from both existing mortar (pointing and bedding) at the SE Node of the building and new mortar (pointing only) at the mockups on the SW Node of the building, as well as an original stone sample.

MTFA D+P sent the mortar and stone samples to Con-struction Materials Consultants, Inc., for laboratory analysis. At the same time, MTFA D+P performed onsite permeability tests on the existing granite and two replacement mortar mockups that remained in the wall. MTFA D+P also reviewed the wall construction and typical repointing installation. The results of these observations, tests, and research are detailed in this re-port.

1.3 Document Review

Documents Reviewed:

• Marcel Breuer and Associates, Nolen-Swineburne and Associates, H.H F. A Office Building, 7th and D Streets, Washington, D.C. “End Wall Granite Fac-ing Alternate ‘L’”, Architectural Sheet 34-1, 1965.

• Quinn Evans Architects, Oehrlein & Associates Ar-chitects, Historic Structures Report: Housing and Urban Development (HUD) Building, 1999. pp 454, 543, 544.

MTFA D+P reviewed the original drawing details from the 1960s construction and the 1999 Historic Structure Report (HSR) prepared for the building. While original drawings provide some insight into original design in-tent, ongoing restoration work has revealed additional construction details not recorded in the 1960s draw-ings. The 1999 HSR documents repair work that oc-curred in the decades following original construction.

This information, combined with original design intent and onsite observations, provides a more detailed un-derstanding of existing conditions.

Original drawing details indicate the facing stones were 3” thick with ¼” joints. Stainless steel anchors connected stone facing to the concrete back-up wall across a 1 ½” cavity. A relieving angle supported every third course of stone. The joints on the drawings do not show mortar details other than to say the mortar is gray. At the relieving angles, the joints are treated with backer rod and sealant. The stone itself is identified as Cold Spring – Charcoal Black, French Creek – Fox Hill Black, or Georgia Granite – Jet Mist, all with a flamed finish. During the removal of the stone facing with the current restoration project, it was confirmed that the stone was set in bedding mortar and pointed so that the entire joint was filled with mortar up to the cavity wall.

According to the 1999 HSR, original construction doc-uments specified that the joints between stone panels be pointed with a gray mortar and raked back ¼” to provide a shadow line. The architect later requested that the joints be filled flush, but this never took place.

The specification allowed for masonry cement as a mortar component but otherwise did not provide an approved mortar mix or desired flexural/compressive strength. Original specifications did identify the gran-ite as Cherry Hill granite, a dark gray stone with white, gray, and gold flecks, that was quarried in St. Peters, Pennsylvania by the French Creek Granite Company.

The HSR did not go as far as to test the stone or mor-tar materials for confirmation of information found in the specifications. The granite was repointed in 1972, following repairs to soffit stones. Documents indicate the repointing mortar was gray and struck flush with the stone face, but these documents also do not include information on the mortar mix.

2.0 Field Testing

2.1 Field Observaitons

MTFA D+P climbed scaffolding at the SE Node of the building to observe various mortar mockups and the associated staining at each instance. To address con-cerns that excessive moisture was also driving the white staining, MTFA D+P employed the RILEM II.E test method to measure permeability of both the stone and the mortar.

The location of the white residue suggests the white res-idue was the result of rain water redepositing calcium-rich creams from the cement-lime pointing mortar onto the surface of the stone below rather than a salt bloom from soluble salts in the stone or the mortar. The team observed the white residue occurred on the stone face immediately below repointed joints and extended far-thest down the stone face below vertical joints. This distribution indicates the staining was related to mois-ture passing from within the wall assembly, through the pointing mortar, and down the face of the masonry cladding. Soluble salt blooms, or efflorescence, typi-cally have a more diffuse, irregular pattern related to wet spots on the stone or on the mortar, rather than below mortar joints.2

2.1 RILEM Test Method No. 11.4

RILEM Test Method No. 11.4 is an industry standard for quantifying the rate at which water moves through a porous material. In the conservation of historic ma-sonry, this test is useful in determining the relative permeability of the various porous materials in a wall assembly – namely the masonry units and the pointing mortar.

RILEM tests require a standardized test apparatus

2 See section “Chloride Analysis.”

A

B

C

D

RILEM Test Locations, Southwest Node (March 2020): (A) Test 1, Granite (B) Test 2, Granite (C) Test 3, Mortar - Type S (D) Test 4, Mortar - Type N consisting of a graduated cylinder that terminates in a large cylindrical bulb with an opening in one side. A ring of clay temporarily attaches the bulb to the wall, so that the bulb’s side opening is flush with the test surface and the graduated cylinder is upright. Once filled with distilled water, the apparatus simulates the pressure of a wind driven rain at the test location. The water level in the graduated cylinder is noted at regu-lar time intervals until all water is absorbed or until a predetermined amount of time has passed. Analysis is simple. The faster the water absorbs, the more porous (and permeable) the substrate is; the slower, the less porous. The permeability of any given material can then be compared to established industry standards for that material as well as to the actual permeability of any other material in the wall assembly.

The RILEM test approximates the pressure of a wind driven rain between 98.1 mph (when the tube is filled to the 0 mL mark) and 44.8 mph (when the water lev-el has fallen to the 5 mL mark). Typically, we expect masonry or mortar that is effectively preventing water infiltration into the wall system to take at least 15-25 minutes to absorb 5 mL of water. If a component of the wall system absorbs 5 mL in 5 minutes or less, the wall system is not effectively preventing water infiltra-tion. At HUD, MTFA D+P performed RILEM tests on two distinct stone units and on joints pointed with two different mortar mixes. There were no original joints intact in the scaffolded area, so the original pointing mortar was not tested.

RILEM testing of the granite cladding confirmed that the existing stone is dense with low porosity and there-fore does not provide a significant pathway for water en-tering the wall assembly. According to the petrograph-ic analysis,3 the stone used to clad the HUD building is a type of diabase – a mafic igneous rock comparable to basalt or gabbro. Stones of this type are composed pri-marily of feldspar, augite, and other magnesium- and iron-rich minerals. Diabase has a density on par with many granites and commercially is often called black granite. During RILEM testing, one stone absorbed 5 mL of water in 45 minutes while the other absorbed the same amount of water in 60 minutes. The difference in absorption rate is likely related to slight variations in the surface porosity at the two test locations. Both test

3 See section “Petrographic Analysis: Stone.”

locations well out-performed the minimum effective absorption rate of 5 mL in 15-25 minutes.

In contrast, RILEM testing at pointing mortars indicat-ed existing mortar joints contribute significantly to wa-ter infiltration. Three test locations at the new pointing mortar mockups absorbed 5 mL in less than 5 minutes.

For two Type N mortar locations, both the vertical and horizontal joints absorbed 5 mL in under 20 seconds, indicating a concerning level of moisture infiltration into the wall assembly at the joints. The Type S mor-tar location fared slightly better, absorbing 5 mL in ap-proximately 75 seconds. Though significantly better than the Type N mortar joints, this is still well below the 5-minute mark. These test results indicate that the mortar joints are letting water into the wall system and since the stone is less permeable, the water remains in the joints until it comes back out the face or reaches the cavity wall for drainage.

Results of permeability testing indicate that further in-vestigation is require into the joint design itself. Both the Type N and Type S mortars are allowing for sig-nificant water infiltration. Mortars harder than Type S are not recommended for this application because they are brittle with a tendency to crack, which will also lead to excessive water infiltration. Since a harder (and presumably less permeable) mortar is not appropriate, joints must be designed to ensure water drains away from the stone face and into the wall cavity.

Analysis indicates that a Type S mortar with mini-mal lime content is most appropriate for repointing, as longs as joint design ensures water will drain back toward the wall cavity instead of through the mortar joint. A Type S mortar offers the best combination of durability and permeability. A minimal lime content will also decrease the likelihood of lime leach related to excessive water. With increased water comes increased leaching of lime and weakening of lime-based mortar.

3.0 Laboratory Analysis

3.1 Petrographic Analysis: Stone

The samples chosen for analysis included:

• Sample 1 – SW Node, stone fragment taken from veneer cladding

The stone sample from the SW Node was analyzed per ASTM C1721. The analysis indicates the stone used at HUD is a diabase, an intrusive basic igneous rock (also called dolerite). This medium-grained stone is comparable to basalt or gabbro. Stones of this type are composed primarily of feldspar, augite, and other mag-nesium- and iron-rich minerals. Diabase has a density on par with many granites and commercially is often called black granite. It is common to the mid-Atlantic igneous sills, such as the New York Palisades.

Diabase has an average porosity of approximately .30 to 2.7%.1 Its compressive strength can range from 18,000 to 44,000 psi. Its average density is .109 lb/ in3. A type N mortar is likely too soft and porous to hold up well against such a dense, strong stone. A type S mortar, with a minimum strength of 1800 psi is more appropriate to reduce water infiltration while balanc-ing strength and brittleness.

1 See Matweb.com

3.2 Petrographic Analysis: Mortar

The samples chosen for analysis included:

• Sample 2 – SE Node, Section A6, Vertical Joint, Pointing Mortar (existing)

• Sample 3 – SE Node, Section A6, Vertical Joint, Bedding Mortar (existing)

• Sample 4 – SW Node, Section D, Edison Type N (new)

• Sample 5 – SW Node, Section B, Edison Type N, Latex-Modified (new)

◊ This sample was used for compari-son purposes but was not intended to have full compositional analysis.

This sample was tested for salt iden-tification using XRD.

Three mortar samples (samples 2-4) underwent a full petrographic analysis per ASTM C1324 to identify original mix ratios, chemical composition, mineral-ogy, and physical characteristics of both original and modern pointing mortars. First, optical microscopy

Stone and Mortar Samples for Laboratory Analysis of intact samples determined overall color, texture, grain size distribution, and the percentage and shape of voids. More specialized analytical methods - includ-ing Scanning Electron Microscopy (SEM-EDS), X-Ray Fluorescence (XRF), X-Ray Diffraction (XRD) Fourier Transform Infrared Spectroscopy (FTIR), Ion Chro-matography (IC), Gravimetry, and Thermal Analyses – identified various chemical components of the mortar paste. XRD was used to identify relative abundance of various minerals in the aggregate. Acid digestion and sand extraction allowed for a more detailed examina-tion of the color, texture, and grain size distribute of the aggregate used in each mortar mix.

Analysis found that all four mortars were gray cement-based mortars typical of modern masonry mortars pro-duced since the early- to mid-20th century. Samples 2 and 3, contemporary with original construction, were made with masonry cement. Masonry cement consists of a mixture of Portland cement or blended hydraulic cement and plasticizing materials (such as limestone or hydrated or hydraulic lime), together with other ma-terials introduced to enhance one or more properties such as setting time, workability, water retention, and durability. The mixtures in masonry cements are pro-prietary and their compositions are not always known.

Sample 5 was also made with masonry cement whereas Sample 4 was a Portland cement-lime based mortar.

Cement-lime mortars are traditionally specified for historic preservation because their proportions can be controlled without unknown components.

Composition of Samples 2-5

• Sample 2: 1 part masonry cement to 2 parts sand Type N, not air-entrained

◊ High limestone fines content ◊ Portland cement ◊ Not air-entrained ◊ High permeability ◊ Low bond strength

• Sample 3: 1 part masonry cement to 3 parts sand – Type N, air-entrained

◊ High hydrated lime content ◊ Fly Ash ◊ Excessive air-entraining ◊ High permeability ◊ Low bond strength

• Sample 4: 1 part Portland cement, 1.1 part hydrated lime, 6.6 parts sand – Type N

• Sample 5: Masonry Cement – Type S (binder: sand ratio not defined in report)

◊ Latex additive ◊ Normal limestone fines per

ASTM C91

◊ Normal hydrated lime per

ASTM C91

◊ Normal Portland cement per

ASTM C91

All mortar samples contained fine crushed sand aggre-gate, which do not conform to the grain size distribu-tion required for an ASTM C144 masonry sand. The sand used in both mortars for Sample 2 and 3 shared notable similarities, as did the sand found in the two new mortars (Sample 4 and 5). Both of the older mor-tars (Sample 2 and 3) contained similar sands – in this case mostly crushed quartz that was too finely graded.

The proportion of other minerals – i.e. feldspar, mica, and quartzite – to quartz was small.

Like the older mortars, the two new mortars (Sample 4 and 5) contain sands that have similar composi-tions and are too finely graded to comply with ASTM C144. This sand is also predominantly crushed quartz, though it does contain a more appreciable proportion of feldspar, mica, and quartzite. Sand aggregate that is too uniform in size weakens the strength of a mortar and increases water infiltration due to more uniform voids. Excessive fines increase the water requirements of the mortar and makes the mortars more permeable.

The binder content differed across the four samples.

While both binders for Sample 1 and 2 were masonry cements, they were predominantly gray Portland ce-ment. Sample 1 contained a significant amount of lime-stone fines while Sample 3 contained a high percentage of dolomitic hydrated lime and a negligible amount of fly ash, likely an accidental contaminant rather than an intended additive. The two mortar samples also differ in that Sample 2 exhibited poor air entrainment, while Sample 3 exhibited excessive air entrainment with small, spherical voids. However, both mortars also exhibited a high percentage of irregular voids between grains, which likely resulted in a weaker bond with the stone units and greater moisture permeability.

The binders for the two new mortars differ significant-ly, with Sample 4 containing a combination of Portland cement and dolomitic hydrated lime mixed on site and Sample 5 containing a commercially available masonry cement. Petrographic analysis identified Sample 4 as a Type N cement-lime mortar made up of 1 part Portland cement to 1.1 part dolomitic hydrated lime to 6.6 parts sand.

Thermal analyses determined Sample 4 did not exhibit adequate carbonation, the process through which lime (calcium hydroxide) reacts with water present in either the mix or the air to form calcium carbonate. If a mor-tar dries too quickly, carbonation is reduced, which can result in a mortar with poor adhesion and poor dura-bility.

Sample 5 was the best formulated off all mortars tested.

The masonry cement used in Sample 5 has better pro-portions of Portland cement, limestone fines, and do-lomitic hydrated lime than the two older mortars and it aligns with an ASTM C91 Type S. FTIR identified a latex polymer component of the mix, and the analysis concludes that this additive increased the overall flex-ibility of this particular pointing mortar, combating the brittleness often characteristic of Type S mortars. No other organic binders were detected in the rest of the mortar samples.

The petrographic analysis indicates that the older mortars were made with masonry cements and were not well-suited for the stone facing. The 1972 point-ing mortar aggregate was too fine and its lime content was rather high (see page 53 of the report), although the lime was likely in the form of fines as opposed to a hydrate. The 1960s bedding mortar was over air-entrained which interfered with its bond to the dense stone. The fine aggregate of the old and the new mor-tars is contributing to moisture permeability. Sample 4 did not fully carbonate and Sample 5, though well-formed, was still leaching lime. The analysis indicates a mortar that matches the original in composition would be an air-entrained Type N masonry cement mortar with a fine aggregate. However, the analysis also indicates a more appropriate mortar for the stone facing is a Type S cement-lime mortar with aggregate that meets ASTM C144.

3.3 Chloride Analysis

X-Ray Diffraction (XRD) and Ion Chromatography (IC) were used to determine whether white staining observed on site was related to efflorescence of harmful soluble salts or to lime leaching from the cement-lime pointing mortars. XRD is an analytical method that identifies crystalline compounds – specifically miner-als in the case of mortar analysis. IC separates and identifies water-soluble anions in the filtrate of the pul-verized mortar and determines their relative concen-trations. XRD was used to identify the white residue found on Sample 5 (the new latex-modified mortar).

IC was conducted on mortar samples 2-4 to determine salt contamination, if any.

Results from both analytical techniques indicate that white staining observed on site is the result of lime leach rather than efflorescence of soluble salts. XRD identified calcium carbonate as the white efflorescence powder on the outer surface of Sample 5. Calcium car-bonate is a necessary product of the curing process for lime mortars, so this efflorescence does not indicate that unwanted salts have been introduced into the ma-sonry construction. However, this does indicate that excessive moisture is passing through the mortar joints as the calcium carbonate is leaching out of the mortar.

IC did not detect any alarming amounts of chloride salts that are damaging any of the mortars. The sulfate content in the mortars was higher than the other salts but this is likely related to Portland cement in the com-position. The sulphate content was higher in the new mortar than in the old mortars.

The chloride analysis indicates the joint design for the HUD restoration must either reduce the amount of wa-ter infiltration through the face of the mortar or reduce the amount of lime in the mortar mix to prevent lime leaching.

4.0 Joint Construction

Detail Analysis

Field observations and laboratory testing yielded re-sults that guide the parameters of joint treatment for the HUD restoration project. The following section provides options based on the different characteristics of permeability, workability, and lime content as well as historic preservation considerations.

On March 16, 2020, MTFA D+P observed existing con-ditions at ongoing repair work on the southwest eleva-tion. Mortar had been raked from most joints. Backer rod had been inserted into all joints, including those that had been repointing with new pointing mortars.

The typical mortar installation observed onsite for the ¼” joints was 5/8” depth of mortar supported by a ½” backer rod for a 1” joint system. The granite veneer was 3” thick which left a 2” ledge for water to sit before it drained to the face of the stone or to the back cavity.

This ledge is possibly one source of water for the leach-ing calcium carbonate.

The original stone detail included a filled mortar joint with setting mortar and pointing mortar. Like suck-ing water through a straw, filled joints can slow down absorption as water is transported through the joint by capillary action. Carbonation and well-graded sands increase capillarity by decreasing the size of the void structure. The more mortar in the joint, the longer the path of capillaries, and the slower water absorption.

The water that remains in the joint then continues to move toward the surface or the cavity with evapora-tion. If the depth of mortar is reduced and the mortar is mixed with large voids (i.e. not carbonated, or with uniform sands), the mortar readily absorbs the mois-ture and pulls it through the joint quickly allowing water to sit on the stone ledge with nowhere to travel.

Consider filling the joints completely with mortar to manage absorption.

In addition to absorption, the permeability tests and chloride analysis indicated that the lime content of the joints had an effect of the staining on the face. If the Type N cement-lime mortar is continued to be used, the joints must be designed to allow for the water to drain away from the stone face and into the cavity wall.

If filling the joint with mortar is not possible, addition-al layers of backer rod could assist with filling the joints and preventing water from ponding. Additionally, re-moving surface creams of fresh mortar after initial cure with an acid wash could help with reducing the avail-able free lime that contributes to the surface staining.

In a future detail, when the stones are removed and reinstalled, trimming the hidden ledge so that is has a slope towards the cavity would help with drainage.

This is an original detail that was used for the Section

Current Joint Condition - backer rod as temporary waterproofing with no pointing mortar, Southwest Node (March, 2020)

C stone. If changing the mortar type is considered, the optimal mortar mix should contain minimal lime con-tent to impart workability and self-healing while reduc-ing absorption.

The petrographic analysis shows the older mortars were made with masonry cements and were not well-suited for the stone facing. The analysis indicates a mortar that matches the original in composition would be an air-entrained Type N masonry cement mortar with a fine aggregate. However, the analysis also indicates a more appropriate mortar for the stone facing is a Type S cement-lime mortar with aggregate that meets ASTM C144. MTFA D+P recommends using cement-lime mortars over masonry cements because the proportion of cementitious materials to lime is more controlled.

A Type S mortar (1 part Portland cement, ½ part lime, and 4 ½ parts sand) is appropriate for diabase stone that has low porosity, high density, and high strength.

Though stronger than a Type N with a minimum of 1800 psi, the Type S is much weaker in strength than the granite and so the mortar will remain sacrificial to the stone. One of the drawbacks of a Type S mortar is that the higher Portland cement content can lead to a more brittle mortar. The lime component, though reduced, helps to impart workability and flexibility.

Modern mortar specifications often call for modifying the Type S mortar with latex to allow the mortar to re-main breathable but imparting some additional flex-ibility into the mortar. Latex additives also decrease water absorption as they enrich the mortar pastes.

Consider installing a Type S cement-lime mortar with latex additives

Another option for addressing the water infiltration is-sue at the joints for HUD is to change the joints sys-tem from mortar to sealant. Sealant is an appropriate joint treatment for the historic stone because the cavity wall system allows for drainage of the wall assembly.

Sealant was an original joint treatment at every three courses for the relieving angles. A non-staining silicone sealant will prevent moisture infiltration and eliminate any staining from calcium carbonate. The draw back for this treatment is that sealant joints require a more frequent maintenance cycle than mortar joints.

Finally, the last joint treatment option is a hybrid with mortar and sealant. This system will introduce redun-dancy, eliminate the need for fully filled mortar joints, and maintain the historic appearance of the mortared stone joints. Install a sealant joint 1” from the face of the stone. Mortar the joint flush to a depth of 2 ½ times the width of the joint, or 5/8”, with a backer rod between the mortar and the sealant. The mortar will be the first line of defense against moisture but the sealant will prevent any moisture from sitting on the ledge. Any moisture that penetrated behind the seal-ant can then be drained or evaporated to the cavity wall. The mortar joint also serves to protect the sealant joints from weathering and UV exposure and therefore extending its service life. There is precedence for this type of joint but its success depends on the right depth of the mortar joint to achieve good bond with the stone.

5.0 Recommendations

To Replicate the Existing Mortar:

• Replicate the older 1972 pointing mortar with a Type N Masonry cement mortar, ASTM C144 graded aggregate, and fill the joints completely with mortar. Treat joints with acid wash to remove surface creams and available soluble calcium carbonate at the surface.

• Do not use a Type N cement-lime mortar.

To Improve the Mortar Performance:

• Replace the mortar with a Type S cement-lime mortar to improve bond, reduce available lime for leaching, and reduce water infiltration. Fill mortar joints completely to slow the rate of absorption and evaporation.

• Consider adding latex to pointing mortar to increase the flexibility of the mortar.

• Use backer rod at back of joint to prevent mortar from filling cavity.

To Eliminate Water Infiltration:

• Treat all the joints with sealant and rely on the cavity wall to drain any moisture that penetrates the wall system.

To Keep Historic Appearance but Eliminate Water Infiltration:

• Install a sealant joint recessed at least 5/8” from the face of the stone. Point the 5/8” depth with Type S cement-lime mortar.

To Alter Stone Detail During Reinstallation:

• Trim hidden top ledge of stone so it drains back to the cavity wall, similar to original sec-tion C details.

• Install a mortar stop at the cavity wall to prevent mortar drops and set each stone with filled bed joints.

Appendix A:

RILEM Test Method 11.4 Permeability Charts

Rilem Test #1

Location: Southwest Elevation

Substrate: Granite

Absorbed Water (mL) Time (min.)

0 0

0.75 10

1.5 15

3 20

4.75 30

5 45

Rilem Test #2

Location: Southwest Elevation

Substrate: Granite

Absorbed Water (mL) Time (min.)

0 0

0.5 10

0.75 15

1.5 20

2.5 30

4 45

5 60

0.5

1.5

2.5

3.5

4.5

0 10 15 20 30 45 60 75

A b s o rb e d

W a te r

(m L

Time (min.)

RILEM TEST

Stone Cladding

Test #1

Test #2 meredithw Text Box

Absorbed Water (mL) Time (sec.)

0 0

0.5 5

1 10

1.75 15

2.5 30

3.5 45

4.5 60 5 75

Absorbed Water (mL_ Time (sec.)

0 0

1.5 5

3 10

5 15

Rilem Test #3

Location: Southwest Elevation

Substrate: Type S mortar, horizontal joint

Rilem Test #4

Location: Southwest Elevation

Substrate: Type S mortar, horizontal joint

0.5

1.5

2.5

3.5

4.5

0 5 10 15 30 45 60 75

A b s o rb e d

W a te r (m

L

Time (sec.)

RILEM TEST

Stone Cladding

Test #3

Test #4

Appendix B:

Laboratory Analyses of a Stone and Masonry Mortars from Robert C Weaver Federal Building in Washington DC

CONSTRUCTION MATERIALS CONSULTANTS, INC.

Laboratory Analyses of A Stone & Masonry Mortars From Robert C Weaver

Federal Building in Washington, D.C.

Robert C Weaver Federal Building (HUD) 451 7th Street SW Washington, D.C.

Prepared for:

John Milner Associates, Inc.

March 25, 2020

CMC 0320114

Construction Materials Consultants, Inc.

Berkshire Center, Suite 104

4727 Route 30 Greensburg, PA 15601 USA

Phone: 724-834-3551 Fax: 724-834-3556 www.cmc-concrete.com

Serving the Industry through Testing, Investigation, Evaluation, & Research

March 25, 2020

Amanda Edwards John Milner Associates, Inc.

3200 Lee Highway Alexandria, VA 22207

RE: ROBERT C WEAVER FEDERAL BUILDING (HUD), 451 7TH STREET SW, WASHINGTON, D.C.

Dear Ms. Edwards:

Construction Materials Consultants, Inc. (CMC) is pleased to provide the enclosed comprehensive report on “Laboratory Studies of A Stone and Masonry Mortars from Robert C Weaver Federal Building in Washington, D.C.”

Results, opinions, and conclusions presented herein are based on the information and samples provided at the time of this investigation. We reserve the right to modify the report as additional information becomes available.

Neither CMC nor its employees assume any obligation or liability for damages, including, but not limited to, consequential damages arising out of, or in conjunction with the use, or inability to use this resulting information.

Sample residues will be returned, as requested. All reports are the confidential property of clients, and information contained herein may not be published or reproduced pending our written approval.

Please feel free to contact us with any additional questions. We look forward to providing our service again for your future projects.

Sincerely Yours, Dipayan Jana, PG

President, Petrographer

DJ:jlh

Robert C Weaver Federal Buildiing (HUD), Washington, D.C.

TABLE OF CONTENTS

Laboratory Analyses Of A Stone And Masonry Mortars From Robert C Weaver Federal Building In Washinton, D.C Executive Summary Introduction Samples Methodologies Results Lapped Cross Sections & Micrographs Of Lapped Cross Sections Of Mortars Grain-Size Distribution Of Sands In Masonry Mortars Thin Sections Of Mortars Optical Microscopy Of Mortars From 1960s Void And Sand Contents From Image Analyses Of Thin Section Micrographs Of Original Mortars From 1960s Optical Microscopy Of Mortars From 2019 Void And Sand Contents From Image Analyses Of Thin Section Micrographs Of Recent Mortars From 2019 Optical Microscopy Of Stone SEM-EDS Compositional Variations Of Pastes In Mortars From 1960s Vintage SEM-EDS Compositional Variations Of Pastes In Mortars From 2019 SEM-EDS Studies Binder Fractions Of Mortars From 1960s And 2019 – A Comparison Mortar Types From Optical & Electron Microscopy Mineralogical Compositions Of 1960s Mortars From X-Ray Diffraction (XRD) Mineralogical Compositions Of A 2019 Mortar & Efflorescence Deposits From X-Ray Diffraction (XRD) Chemical Compositions Of Mortars From X-Ray Fluorescence (XRF) Chemical Compositions Of Mortars From Gravimetry Thermal Analyses Fourier Transform Infrared Spectroscopy (FTIR) Of Mortars From 1960s Fourier Transform Infrared Spectroscopy (FTIR) Of Mortars From 2019 Ion Chromatography Of Water-Soluble Salts In Masonry Mortars From 1960s Ion Chromatography Of Water-Soluble Salts In Masonry Mortars From 2019 Results From Chemical Analyses (Gravimetry, IC, XRF), Thermal, and XRD Studies Of Masonry Mortars Mix Proportions Of Masonry Mortars From Petrographic And Chemical Data Discussions Type Of Mortar & Its Ingredients Mix Calculations Mortar Condition Stone Type Tuckpointing Mortar References

Appendix –Laboratory Testing Of Masonry Mortars Introduction Sample Selection And Steps Of Laboratory Analyses Optical Microscopy For Mineralogy & Microstructure Of Mortar Scanning Electron Microscopy & X-Ray Microanalyses For Mineralogy, Microstructure, And Microchemical Compositions Of Mortar Acid Digestion For Siliceous Sand Content And Size Distribution Of Sand Cold-Acid & Hot-Alkali Digestion For Soluble Silica Content Losses On Ignition For Free & Combined Water Contents, And, Carbonate Content X-Ray Diffraction For Mineralogy Of Mortar X-Ray Fluorescence Spectroscopy For Chemical Composition Of Mortar Thermal Analyses For Determination Of Hydrous, Carbonate, And Sulfate Phases In Mortar Infrared Spectroscopy For Determination Of Organic Components In Mortar Ion Chromatography For Determination Of Water-Soluble Cations And Anions In Mortar Information Obtained From Various Laboratory Methods Steps Followed In Laboratory Analyses Mix Calculations From Petrography & Chemical Analyses Of Mortar Flow Chart Of Procedures Followed In Laboratory Analyses Of Masonry Mortars

Robert C Weaver Federal Buildiing (HUD), Washington, D.C. 1

LABORATORY ANALYSES OF A STONE AND MASONRY MORTARS FROM ROBERT C WEAVER

FEDERAL BUILDING IN WASHINTON, D.C.

EXECUTIVE SUMMARY

Located at 451 7th Street, SW in Washington, D.C. the Robert C. Weaver Federal Building is the headquarters of the U.S. Department of Housing and Urban Development (HUD). Constructed from 1965 to 1968 by the renowned architect Marcel Breuer, the building is recognized as the first federal building in the country to utilize precast concrete as the primary structural and exterior finish material, as well as the first fully modular design for a federal office building.

As part of the renovation process, two sets of masonry mortar samples were provided for detailed laboratory examinations. The first set comprises two samples from the original 1960s vintage, and the second set of two samples reportedly came from 2019 construction. The purpose of this laboratory examination is to determine the compositions and mix proportions of mortars from the original 1960s and recent 2019 constructions and evaluate their compatibilities and suitability for long-term performance of masonry walls. Additionally, a dark gray, dense hard crystalline stone masonry sample from the building was provided in two small pieces to determine the stone type.

Mortar samples were analyzed by comprehensive laboratory examinations following various industry standards, e.g., ASTM C 1324 and RILEM methods (Middendorf et al. 2004, 2005) starting with visual examinations, extensive optical microscopy, scanning electron microscopy and energy-dispersive X-ray microanalyses (SEM- EDS), followed by wet chemical (gravimetric) analyses, X-ray diffraction (XRD), energy-dispersive X-ray fluorescence spectroscopy (ED-XRF), thermal analyses (TGA, DTG, DSC), and, ion chromatography (IC) of water-soluble salts in mortars.

Visual examinations of mortar fragments from the original as well as recent construction showed apparent similarities in typical gray color tones of mortars found in many modern-day cement-lime or masonry cement mortars. Pieces in each mortar sample are reasonably dense, hard, and uniform in appearance confirming representative samples show no evidence of any contamination or odd-looking peace to indicate multiple pointing events. One of the two mortars (#5) from 2019 construction showed white efflorescence powder on a bedding face, which was subsequently determined by XRD studies to be calcium carbonate (calcite).

Optical microscopical examinations of mortars showed both similarities in basic ingredients across four mortars from original to recent constructions, as well as noticeable differences in sand and binder compositions. For the sand used in two original mortars from 1960s construction, both mortars (#2 and 3) showed use of compositionally similar sands, which are crushed siliceous sands consisting of major amounts of crushed silica (quartz) and minor amounts of quartzite, feldspar, and other siliceous components. Sands extracted from original mortars showed grain-size distributions having an abundance of finer grain sizes concentrated from crushing operations, which do not conform to grain size distribution of ASTM C 144 masonry sand. By contrast, sands used in the recent 2019 construction, though also siliceous, contain a noticeable amount of feldspar (albite), quartzite, mica and other siliceous components at subordinate amounts after quartz as the dominant siliceous constituents.

Since sand constitutes the dominant volume fraction of mortars, such variabilities in and mineralogies between original and recent mortars is reflected in their bulk chemical (oxide) compositions as well. Grain size distribution of sand extracted from a recent 2019 mortar after acid digestion showed dominance of finer size fractions as also seen in the sands from the original mortars, having sand finer than the grain size distribution of sands recommended for ASTM C 144 masonry sands. Grain size distribution plots of sands from two original mortars and one recent mortar are very similar showing excessive fines and deviation of ranges of size distribution recommended for ASTM C 144 masonry sands. Excessive fines in the sands increases the water requirements of mortar, thereby increases the overall water-cementitious materials ratios, makes mortar more permeable to moisture, and reduces long-term moisture tightness of mortar joints. Sand particles are dense, hard, clear to off-

Robert C Weaver Federal Buildiing (HUD), Washington, D.C. 2 white to light brown, and present in sound conditions without any evidence of potentially deleterious alkali-aggregate reactions in the mortars.

Perhaps the most interesting similarities as well as differences between the two original (1960s) and two recent (2019) mortars are found in their binders. The two original mortars both used masonry cement as the binder, therefore, pastes in both mortars showed the characteristic carbonated nature, which is a hallmark microstructural feature of masonry cement mortars. Use of masonry cement was not uncommon in many 1960s masonry constructions, but the compositions of the masonry cements varied significantly. In mortar #2, masonry cement contained a significant amount of limestone fines, which has given an overall granular appearance of mortar as noticed in optical and electron microscopy. By contrast, masonry cement used in mortar #3 had major amount of dolomitic hydrated lime, which has given an overall very fine-grained porous and carbonated appearance of paste as opposed to coarser granular appearance from limestone fine particles seen in mortar #2. Besides this major difference in masonry cements’ components, masonry cements in both mortars contained Portland cement as the main cementitious (hydraulic) phase, often detected by its residual cement particles leaving only dark brown skeletal remains of interstitial ferrite phases. Contrary to the abundant limestone fines in mortar #2, its presence is scarce in mortar #3, which, as mentioned, is dominated by hydrated lime. A trace amount of dark brown to black spherical fly ash particles are detected in the high-lime masonry cement paste in mortar #3 whose amount is less than that anticipated from intentional incorporation as in the blended cement, but enough to be detected in many areas of thin section to indicate perhaps it was accidentally incorporated either in the masonry cement during the manufacturing phase of cement and/or during the construction phase of the mortar.

Perhaps the most important difference between the two original mortars is in their air entrainment. Mortar #2 shows marginal air entrainment but has numerous irregular-shaped entrapped voids between sand particles as well as pore spaces between individual calcite grains of limestone fines, which have given an overall porous microstructure of this mortar susceptible to moisture penetration. By contrast, mortar #3 shows excellent air entrainment in having numerous fine, discrete, spherical and near-spherical entrained air voids of sizes 1 mm or less, as well as many coarse voids, all of which, in fact, have given an excessive air entrainment appearance in the mortar. Even though masonry cement mortars are characteristically air entrained in having 16 to 18 percent air, air content in mortar #3 is found to be at the high end of this range, which is detrimental to development of good bond to the adjacent masonry units. Abundant irregular-shaped voids and interstitial capillary pore spaces in mortar #2 and abundant spherical entrained air bubbles in mortar #3 are both detrimental to development of a good bond to adjacent masonry units as well as their resistance to moisture transmission. Therefore, both original mortars are found to be candidates to have weak bond to their adjacent masonry units and have high moisture permeabilities.

Contrary to the masonry cement binders in two original mortars, mortar #4 from the recent 2019 construction showed use of Portland cement and dolomitic hydrated lime as the two essential binder components. Therefore, paste in this mortar is non-carbonated, and contains abundant residual Portland cement particles, cement hydration products including coarser patches of calcium hydroxide crystals of cement hydration mixed with finer-grained porous hydrated lime, which is still not carbonated enough due to limited interaction with atmospheric carbon dioxide.

By contrast, mortar #5 from 2019 showed use of masonry cement as found in the original mortars, but the latest masonry cement was very different from 1960s masonry cements. In mortar #5, masonry cement does not show any noticeable abundance of either limestone fines (as found in #2) or hydrated lime (as found in #3) but more reasonable proportions of cement, lime, and limestone fines commonly found in many modern-day well-proportioned masonry cement mortars. Abundance of residual Portland cement particles in paste indicate a Portland cement-rich composition of the masonry cement, e.g., similar to an ASTM C 91 Type S masonry cement.

Mortar #5 showed another difference in paste from the rest of the mortars in having a polymer component, which has densified the paste more than other mortars. Such polymer-induced added densification of paste beyond carbonated lime and cement hydration products is not only detected in optical microscopy but more so during

Robert C Weaver Federal Buildiing (HUD), Washington, D.C. 3

SEM-EDS studies. FTIR analysis of this mortar detected a latex-based polymer addition, consistent with reported latex-modified nature of this mortar.

SEM-EDS studies of paste per se in two original and two recent mortars showed overall enrichment of silica in siliceous sand grains and interstitial lime-based paste reflected in elemental (Ca-Si-Al-Mg-etc.) maps. Use of dolomitic lime in all mortars is reflected from the high magnesia contents of pastes in SEM-EDS studies.

Cementation indices of paste (paste-CI after Eckel 1922) shows wide range within and between the mortars which is not unexpected considering proprietary mix composition of masonry mortars as well as variable degrees of mixing of cement, lime, and limestone fine components in the paste. A systematic trend in compositional plots of silica and lime contents of pastes against paste-CI from SEM-EDS studies demonstrated their lime, limestone fines, and Portland cement-based compositions of original binders.

Optical microscopy and scanning electron microscopy showed evidence of leaching of lime paste from around sand grains in mortar #2 from original mortar, which is responsible for overall porous microstructure of original mortar.

Compositional analyses of original and recent mortars by energy-dispersive X-ray fluorescence spectroscopy (ED- XRF) showed overall similar oxide compositions of two original mortars (despite their differences in compositions of masonry cements used), which are very different from the two recent mortars. Two recent mortars from 2019 construction showed overall similarities in oxide compositions. Two original mortars are found to be richer in silica mostly due to higher sand contents and quartz-rich composition of sands as opposed to many feldspar particles found in the sands in recent mortars, which are responsible for elevated alumina contents in the recent mortars compared to the original ones. High feldspar content of sand in the recent mortars also elevated the potassium contents in these mortars compared to the original ones. Since sand constitutes bulk volume of the mortar, changes in the sand mineralogies between the 1960s and 2019 mortars is reflected in their bulk chemistries. Oxide compositions of all four mortars reflect their similarities and differences in the mineralogies of sands used, as well as compositions of binder phases.

Mineralogical compositions of mortars determined from X-ray diffraction studies showed simple quartz-based mineralogy of sand in original mortars, more complex quartz-feldspar-mica based mineralogy of sand in recent mortars, dominance of quartz in the sand in all four mortars, and subordinate amount of calcite from carbonated paste and calcite fine particles of masonry cements. XRD studies of soft, white efflorescence powder on mortar #5 showed calcium carbonate (calcite) composition of the efflorescence due to moisture-induced leaching of lime and cement hydration products from mortar followed by evaporation, precipitation of dissolved salts on the mortar’s outer surface, and atmospheric carbonation of precipitated salts.

Thermal analyses of all four mortars showed simple thermograms for original mortars mostly from the loss of free water and water in hydrate salts during heating up to 200ºC, polymorphic transition of quartz at 575ºC, and major endothermic peak from decarbonation of carbonated lime and calcite fine particles.