August 29, 2019
by Steven G. Naggatz, AIA, NCARB
Time is money. Developments in technology and production continue to push the envelope and advance the speed at which buildings are constructed. When properly designed, specified, and executed, details and materials can assist the construction team by significantly reducing the installation time. Even if it is well-intentioned, the use of incompatible materials to increase productivity when installing cladding elements can result in disastrous situations.
Dimension stone and limestone cladding
Stone has been used as a building material for thousands of years and continues to be utilized for decorative and functional purposes. Its aesthetics and sense of permanence have continued to make it a popular material choice among architects, contractors, and owners. Buildings of great significance, such as churches, civic centers, and government facilities, were constructed of stone to reflect the structure’s importance within the urban context and culture.
The term “dimension stone” refers to natural stone fabricated to specific sizes or shapes. Dimension stone cladding relies on inexpensive backup materials in combination with more expensive facing. Early 20th century stone structures relied on cladding panels that were at least 100 mm (4 in.) thick. These stone-clad buildings were commonly multi-wythe loadbearing assemblies combining high quality stone finished to very tight tolerances with a looser rubble or brick backup. Often, cladding of varying thicknesses was keyed into masonry backup. Iron, galvanized steel, copper alloys, and aluminum have also been used as dimension stone cladding attachments. Advances in fabrication technology and the introduction of new systems have changed how stone is anchored to substrates. Consequently, panels have decreased in thickness over the past 50 years. Modern dimension stone cladding panels are most commonly between 30 and 50 mm (1 ¼ and 2 in.) thick and are anchored with stainless steel—typically rods, straps, dowels, or a combination thereof. Recently, composite assemblies have been introduced in flat panel applications. These panels combine stiff backer materials, such as aluminum honeycomb and adhesives, and have reduced stone thickness to as little as 5 mm (¼ in.). The use of ornamental or decorative stone still requires the use of elements that are more than 100 mm thick.
Limestone is a sedimentary rock consisting of either calcium carbonate (calcite) or calcium and magnesium (dolomite) and is suitable for many architectural and structural applications. Considering its sedimentary formation, limestone has a natural bed, and in many cases, the bedding is sufficiently uniform, so it can be machined or cut with little risk of splitting. Limestone tends to be fairly uniform in strength (isotropic) when it is loaded parallel or perpendicular to its natural bedding plane compared to other types of dimension stone, such as marble, granite, or slate. It is preferential to install limestone cladding panels with its natural bed, horizontally, in a manner consistent with its original formation.
‘Face-bedded’ panels (i.e. bedding layers are vertical) are vulnerable to damage (delamination) through various mechanisms, such as crystallization of salts, dissolution of clay layers, and freeze-thaw cycles. These processes can result in separation of bedding layers within the panel.
Mechanical properties of limestone material vary between quarries, and to some extent, within the quarry itself. The range and variability of mechanical properties of the stone proposed for use should, therefore, be determined by testing.
Dimension stone attachments
The test standards and specifications most commonly used to evaluate dimension stone in the United States are published by ASTM International. Cladding elements should be evaluated further with statistical methods and attachments and designed using sound engineering principles. Metal anchors in direct contact with limestone should consist of American Iron and Steel Institute (AISI) Type 304 or 316 stainless steel. Other noncorrosive anchors, such as copper or bronze, are also acceptable. Due to the potential for chemical attack and degradation of the material itself, anchors in direct contact with limestone should not be fabricated with aluminum unless the latter is protected with a corrosion-inhibiting coating. Mill finish or anodized aluminum can be used with granite.
The anchor type varies, depending on the application, but generally consists of wires, dowels, or straps set into holes, kerfs, or other sinkages. The hole, kerf, or sinkage will routinely be oversized to accommodate installation and construction tolerances, and accommodate in-service behavior and/or incidental cladding movement. The ancillary space between the anchor itself and the hole or slot in which it is embedded should be filled with compatible material to prevent moisture accumulation within the void space.
ASTM C1242, Standard Guide for Slection, Design, and Installation of Dimension Stone Attachment Systems, recognizes the use of epoxy-filled holes for liner blocks, or in the case of precast concrete backup, as an adhesive bond for primary anchorage between the stone and stainless steel dowel. However, some epoxy adhesives may be unsuitable for limestone attachments or specific applications (Consult the 22nd edition of the Indiana Limestone Handbook by the Limestone Institute of America, Inc.).
Considering the rapid set and connection strength of epoxy-set anchors, there has been an increasing prevalence to anchor dimension stone panels on exterior façades where the anchor void is filled with epoxy. Based on available data, the coefficient of thermal expansion of epoxies used in construction can be between 10 and 30 times higher than the coefficient of thermal expansion for limestone. Considering the potential for development of cracks or spalls resulting from the inherent material property differences, the use of epoxy to fill anchor voids in limestone attachments should be carefully evaluated prior to specifying its use as part of an attachment system. The following case studies present circumstances in which the variables in epoxy-set attachments were likely not considered prior to installation.
Plans for reconstruction of an aging urban roadway called for the removal and replacement of a limestone balustrade that separated street-level pedestrian sidewalks from a river-level promenade below. The balustrades were built with a combination of traditional elements including bottom and top rails, intermediate piers, and balusters, all constructed of Indiana limestone (Figure 1). The balustrade was built over a period of approximately nine months, between June 2002 and February 2003. Vertical cracks were observed in the balusters within a few months after project completion.
Each baluster has an hourglass profile—plan dimensions at the top and bottom are approximately 200 mm (8 in.) square and each unit is 0.6 m (2 ft) tall. The widest point near the mid-height of the unit is 200 mm and the narrowest point is 90 mm (3 ½ in.). Each baluster was fabricated with a 40-mm (1 ½-in.) continuous vertical hole to accommodate a 20-mm (3/4-in.) diameter stainless steel threaded rod that, in turn, is anchored into the supporting concrete structure. The annular space between the threaded rod and baluster hole was reportedly filled with an epoxy grout. A washer and bolt on top of the baluster secured it in place.
Close-up examination of distress (Figure 2) revealed the cracks are wide (between ½ and 0.9 mm [20 and 35 mils]) near the base (Figure 3) and middle of each baluster and range between 0.18 and 0.9 mm (7 and 20 mils) near the top. The bolts on top of the baluster were found to be only hand tightened, and therefore, likely induced very little stress while tightening the bolts and provided minimal restraint in service.
Destructive removal of representative balusters revealed at least two different types of materials were used to fill the annular space between the inner walls of the hole and the threaded rod. Testing of an epoxy and epoxy grout that were reportedly used to set the balusters show the coefficient of thermal expansion vary between 6.67E-05 (epoxy) and 2.88E-05 (epoxy grout) (Refer to ASTM C531, Standard Test Method for Linear Shrinkage and Coefficient of Thermal Expansion of Chemical-Resistant Mortars, Grouts, Monolithic Surfacings, and Polymer Concretes).
The straight epoxy has a coefficient of thermal expansion of approximately 22 to 28 times the coefficient of thermal expansion of limestone (estimated to be 2.4E-6 to 3.0E-6 in/in/F).
Four limestone baluster specimens and attachments were constructed in a laboratory in order to replicate as-installed assemblies and subjected to thermal cycling in an attempt to simulate exposure of service conditions (Figure 4). Two balusters were fabricated using epoxy and another two with epoxy grout. During the fabrication of the mockups, it was difficult to place the epoxy grout, as it was very thick and had to be placed in layers into the grout hole. Further, the epoxy grout had to be rodded to promote consolidation. The plain epoxy could be easily poured into the holes. It is reasonable to assume the epoxy grout would be very difficult to place under field conditions, especially in cold weather.
Each baluster was instrumented with four stain gauges and two thermocouples. Two strain gauges were installed in the middle of each baluster (at the belled-out section) at opposing edges of the hole. Two additional ones were installed at the base of each baluster, also at opposing edges of the hole, and rotated 90 degrees with respect to gauges at the middle of the baluster. Thermocouples were installed at the top and bottom of each baluster. The balusters were cycled between 2 and 32 C (35 and 90 F) for seven cycles, and the hot and cold temperatures were each held for 12 hours. The balusters were then cycled between 2 and 43 C (35 and 110 F) for 12 cycles, held at the cool temperature for 12 hours, and for 24 hours at the high temperature. A data acquisition system recorded the temperature and strain readings every 15 minutes.
Readings obtained from the strain gauge at the bottom of the epoxy specimens demonstrated a ‘ratcheting’ effect, whereby the epoxy gains tensile strain without full release of the strain. The testing was performed on behalf of a private entity, and at their request, the thermal cycling was terminated after 20 rounds, at which point no visible cracks occurred in the test specimens. However, it is possible that with further cycling, cracks may have developed in the limestone due to tensile fatigue. Stresses could be aggravated by non-uniform expansion of the epoxy relative to the limestone during periods of increasing temperature. As the balusters heat up, the epoxy expands more than the limestone and the former can create a hoop stress in the balusters consistent with cracking observed during inspection of field conditions.
The confinement of epoxy in stainless steel dowel attachments has led to the development of similar types of cracking in cast stone balustrade elements on other projects including dowels installed between balusters and a top rail (Figure 5) and between adjacent top rail sections (Figure 6).
A four-story office building was constructed in 2003 to provide additional space needs for a facility constructed in 1936 in the Central Great Plains. The structural system for each building consists of reinforced concrete and is clad with Indiana limestone. The backup for the 2003 addition was a combination of cast-in-place reinforced concrete and concrete masonry unit (CMU). An air space of 75 mm (3 in.) was specified between the limestone and backup and an exterior limestone cornice is located near the top of the exterior wall above the fourth floor.
The limestone cornice units are nominally 1110 mm (44 in.) wide by 1000 mm (40 in.) high and the cornice projects out approximately 690 mm (27 in.) from the face of the building. The bottom half of the cornice has a concave profile and is nominally 203 mm thick at the bottom edge of the unit (Figure 7).
Each limestone cornice unit is supported by a pair of 250-mm (10-in.) long galvanized steel angles—the center of each angle aligns with the head joint between adjacent units and each cornice is rabbeted to allow the cornice to be seated on the angle. Each angle has a stainless steel bar shop welded to its toes. Lateral support along the top edge of each unit is provided with two 8-mm (3/16-in.) thick stainless steel straps, each located approximately 300 mm (12 in.) from the end of each unit. Each strap is nominally 75 mm wide and the downturned portion of the strap is embedded into a kerf in the top face of the limestone cornice, approximately 25 mm (1 in.). Each strap is anchored through the face of a concrete spandrel beam with a 13-mm (½-in.) diameter stainless steel expansion anchor. Each unit was estimated to weigh more than 900 kg (2000 lb) and was reported to have been set using a crane. A schematic cross-section of the detail is shown in Figure 8.
In response to reports of mortar bond separation and outward displacement of isolated cornice units, an investigation was performed in 2008 to identify the potential cause(s) of distress. A visual survey of all cornice units indicated cracked and/or bond separation of mortar existed at approximately 30 percent of all mortar joints between adjacent cornice units. In a few instances, mortar had even become dislodged between the top horizontal surface of the limestone cornice and the limestone ashlar directly above it (Figure 9). Inspection openings were made by removing the limestone cap at the top of the exterior wall, cutting through a membrane flashing, and examining the back face of the cornice. The inspection opening allowed examination of two cornice units and a total of four strap anchor connections. At the inspection opening, spalls were observed to the back face of the limestone cornice at each strap anchor connection (Figure 10). In some cases, epoxy remnants remained bonded to the limestone face of the kerf slot at limestone spalls (Figure 11), and in other instances, remaining epoxy remnants bonded to the strap anchor (Figure 12). Analysis of the strap anchor connection revealed the limestone at the kerf has sufficient strength to resist design loads.
It is unknown if the anchor voids were specified to be filled with mortar or sealant. However, the cornice anchors were likely set with epoxy to increase the installation rate of the limestone cornice units.
In its fully cured state, epoxy is typically a hard material with a modulus of elasticity comparable to that of limestone. Therefore, when the temperature is elevated relative to when the epoxy cured, thermal expansion of epoxy occurs that is significantly greater than of the adjacent limestone. Where the epoxy is confined by the limestone within dowel holes or kerfs, significant expansive forces can develop that, to an extent, are resisted by the adjacent limestone. With an increase in temperature and resulting epoxy expansion, depending on the geometry of assembly, most importantly edge distance and void volume, the limestone is no longer able to accommodate the developing forces. The expansive forces can be significantly high, resulting in shear and/or tensile stresses in the limestone exceeding the ultimate strength of the material. The cracking and spalls observed in the case studies presented are consistent with the development of high stresses at dowel and kerf connections the author has observed on other buildings and during controlled laboratory testing.
The confinement of epoxy within anchored limestone assemblies can lead to cracking in different types of assemblies. Additional analytical and laboratory studies are needed to further evaluate the mechanism(s) causing cracking for epoxy-set anchors in limestone and other materials.
Consistent with the Indiana Limestone Institute (ILI) recommendations, anchor voids in limestone should be filled with mortar or sealant, or other non-expansive, stable material. Filling anchor voids in dowel holes or kerfs with epoxy should only be considered after careful study of material properties and anchorage geometries proposed for use as part of a limestone attachment assembly.
Steven G. Naggatz, AIA, NCARB, investigates and designs repairs for distress conditions in existing buildings. His expertise lies in exterior wall systems with a focus on structural adequacy, corrosion, anchorage devices, water infiltration, and durability. Naggatz has authored papers on exterior dimension stone cladding and paving systems and has presented other papers on façade repair and rehabilitation design. He can be reached at email@example.com.
Source URL: https://www.constructionspecifier.com/stone-attachments-the-challenges-with-epoxy-set-anchors/
Copyright ©2021 Construction Specifier unless otherwise noted.