The Rhode Island State House: A Capitol Challenge William D. Waterston, AIA, RRC Wiss, Janney, Elstner Associates, Inc. 311 Summer Street, Suite 300, Boston, MA 02210 Phone: 617-946-3400 • Fax: 617-946-0740 • E-mail: williamwaterston@wje.com and Barbara J. Thornton, AIA, NCARB, LEED AP Brewster Thornton Group Architects, LLP 150 Chestnut Street, Providence, RI 02903 Phone: 401-861-1600 • Fax: 401-861-5588 • E-mail: barbarat@brewsterthornton.com 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 W a t e rs t o n a n d T h o r n t o n • 1 7 9 Abstract The Rhode Island State House, an iconic 1904 McKim, Mead and White masterpiece, is a beautiful, albeit leaky, mass of white Georgia marble with brick backup. WJE and BTGA have just completed diagnosis and treatment of moisture migration at the base of its dome— the fourth-largest self-supporting marble dome in the world. Their work, instigated by gilt plasterwork falling onto Senate staffers, revealed several consequences of well-intentioned but ultimately misguided energy retrofits, cleaning efforts, and materials applications. The repairs are being monitored with humidity, temperature, and moisture sensors to observe the hoped-for return to equilibrium in this complicated masonry system. Speaker William D. Waterston, RRC, AIA – Wiss, Janney, Elstner Associates, Inc. William Waterston is an associate principal at his firm. His work includes the investigation, evaluation, and design of roofing and waterproofing systems. He is also experienced in construction document preparation and specification writing. Waterston served as chairman of the RCI Building Envelope Symposium and is the author of several articles on roofing material choices and roofing practices. He has presented at Build Boston, RCI, and Construction Specification Institute meetings and symposia. Barbara J. Thornton, AIA, NCARB, LEED AP – Brewster Thornton Group Architects Barbara Thornton is a principal at her firm and has been a partner with the company for the last 13 years. She has experience in masonry restoration, renovation, and weatherization projects and has come to specialize in exterior restoration, code, and variance work during her 27 years in professional practice. She currently consults with various universities, churches, and the Diocese of Providence regarding code implications for assembly spaces, schools, and places of worship. Thornton holds four degrees from MIT and is registered in Massachusetts, Rhode Island, and Vermont. 1 8 0 • Wa t e rs t o n a n d T h o r n t o n 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 ABSTRACT The Rhode Island State House, an iconic 1904 McKim, Mead and White masterpiece, is a beautiful—albeit leaky—mass of white Georgia marble with brick backup. Wiss, Janney, Elstner and Associates, Inc. (WJE) and Brewster Thornton Group Architects, LLP (BTGA) completed diagnosis and treatment of moisture migration at the base of its dome—the fourth-largest self-supporting marble dome in the world. Their work revealed several consequences of well-intentioned but ultimately misguided energy retrofits, cleaning efforts, and materials applications. The repairs are being monitored with humidity, temperature, and moisture sensors to observe the hoped-for return to equilibrium in this complicated masonry system. This paper will detail the investigation techniques, findings, intervention efforts, and follow-up on the tourelle corners of the Rhode Island State House. Lessons learned from this project will be shared for guidance on maintaining historic mass masonry and stone structures. While the configuration of the structure is unique, concepts of the State House’s performance are applicable to maintaining masonry joints, roofing, and waterproofing materials. Temperature and humidity fundamentals will be discussed as they relate to performance of the structure and its interior finishes. EXECUTIVE SUMMARY From 2011 through 2013, WJE and BTGA worked to diagnose and treat moisture migration issues within marble tourelles at the four corners of the dome of the RI State House. Research and investigations revealed Portland cement repointing, a sandblasted marble cladding, settled stones, inappropriate sealant installation locations, and open/cracked joints that increased exterior water infiltration and delayed drying to the exterior. At the same time, research and access openings determined that prior energy retrofits had added cellulose insulation above the decorative plaster ceilings and limited air circulation within the interstitial spaces, both of which reduced drying at the interior. A malfunctioning steam line was also adding condensation to the northeast tourelle interior sidewall. As a result, the mass masonry wall systems became saturated during certain periods of the year, and the water was held against the plaster by the insulation. This created rapid plaster deterioration, with spalling that caused great consternation to building staffers. Repairs were initiated that replaced Portland mortar with lime putty mortar repointing, injected lime into cracked marble, treated upward-facing joints with lead cames, and removed sealants applied to flashings. The steam line was disconnected, and wetted blown-in cellulose insulation was removed from two of the tourelle ceiling areas. A deteriorated roof base flashing was reinstalled, and missing fasteners were replaced. Windows between the tourelles were restored with sealants, reopened weeps, and paint application. During repairs, temperature, humidity, and moisture monitors were installed in the four corners to record conditions on the exterior wall, above the ceiling, and on the plaster itself. Decorative plaster moldings, paint, gilt, and glazing were reapplied to the damaged interior tourelle ceiling areas. Monitoring shows a slow but consistent drying of the masonry mass, with no condensation at the back of the suspended plasterwork—a hoped-for result. The southwest tourelle shows spikes of high humidity that will continue to be investigated. HISTORY OF THE RHODE ISLAND STATE HOU SE Prior to the late 1800s, Rhode Island’s legislature met in rotation at Newport, Bristol, East Greenwich, Kingston, and Providence. In 1890, the governor began the process of building a single monumental government seat for the then-flourishing national center of textile and gun manufacturing. Rhode Islanders voted to issue bonds to raise “an amount not exceeding” $1,500,000. By competition, the wellconnected New York City firm of McKim, Mead & White was selected. The winning design, an example of the newly confident “American Classicism,” was priced in alternate materials—Indiana limestone, Georgia white marble, and Rhode Island granite. The legislature specifically voted to use the marble, sacrificing the local connection and durability of pink Westerly granite for the iconic purity of white stone. The “White City” concept of urban design, executed in part by McKim, Mead and White at the Chicago Exposition of 1893, was influential in that choice. After eight years of construction and additional bond issues, the new State House was dedicated in 1904 at a final cost of just over $3,000,000. The building included the innovative technologies of electrical lighting and a powered pump system that circulated air throughout the building. A central heating plant was located on the site to supply steam. The part-time volunteer legislature met only in the cooler months. Although the building structure is “transitional,” including steel bearing beams and column elements, the major structural loads are still carried by unreinforced brick mass-masonry walls. Concrete is used in floor and roof decks, formed over vaulted terra cotta tiles on steel ribs. Marble facing is toothed into the backup with no metal anchors. The building is 333 ft. long, 235 ft. high, and contains 15 million bricks, 1,300 tons of steel, and 27,000 cubic feet of Georgia marble. In 1903, its main dome was the largest self-supporting marble dome in the country.1 BUILDING DESCRIPTION The upper roof at the center of the State House contains the main dome in the center and a tourelle at each corner (Figure 1). The base of each tourelle is composed of four The Rhode Island State House: A Capitol Challenge 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 W a t e rs t o n a n d T h o r n t o n • 1 8 1 faces with two shorter walls that face the upper roof and the barrel below the main dome (Figure 2) and two longer walls that face the main roof level. The taller walls are approximately 28 ft. tall (Figure 3) and form the main plinth supporting the central dome. The main roof level is at the approximate level of the interior plaster ceiling of the smaller corner domes. Figure 4 shows the relationship of the components. THE ISSUE The third floor of the State House had been fully restored around 2000, complete with the ceiling receiving faux painting and gilt accents in the molded and pulled plaster corbelling (Figure 5). The roof was replaced in 1999 with a liquid- applied roofing membrane. Exterior repointing of the southwest corner had been performed in 2010. However, all was still not well; these efforts did not address the water-related issues. A request for proposals (RFP) was issued to design professionals to find an expert team to analyze and correct moisture issues that were causing visible paint deterioration in this restored ceiling. Brewster Thornton Group Architects of Providence, partnering with WJE’s Boston office, was selected to conduct the investigation and subsequent repairs. In June of 2012, before the investigation was complete and on the final day of the legislative session, the sky began to fall inside the State House. A member of the governor’s staff was hit on the head by a piece of ceiling plaster while exiting an office on the third floor, 1 8 2 • Wa t e rs t o n a n d T h o r n t o n 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 Figure 1 – Center main dome with tourelle at each corner. Figure 2 – Short wall at the base of the tourelle, viewed from the roof at the base of the main dome. Figure 3 – Typical tall wall at base of tourelle as seen from the main roof level. Figure 4 – Schematic 3-D model rendering showing tourelle, roof, and interior ceiling construction. Note: For clarity, main dome and interior walls not shown. southwest corner. The falling plaster increased attention on the project dramatically! Emergency sounding and removal of plaster were authorized as the larger investigation moved forward. THE INV ESTIGATION A search of existing original drawings and historical documents revealed little structural information. Repair records and correspondence maintained by the Rhode Island Department of Administration and the Rhode Island Historical Preservation and Heritage Commission gave valuable clues to past work. The air circulation pumps installed in the basement to assist airflow are now inoperable. These were decommissioned in the ’60s for the sake of energy conservation, and many of the airshafts were then sealed. In-room vents still connect to the above-ceiling interstitial spaces, and significant airflow still occurs above these areas. There is no air conditioning within the State House other than isolated window units. The marble facing was cleaned with a water-only low-pressure wash during the work in 1990 but had undergone earlier abuse. During a bicentennial cleaning in the 1970s, the contractor was authorized to increase the masonry cleaning scope to include sandblasting of the marble to improve “whiteness” and cut construction time. These operations damaged the marble surface and increased its porosity. Initial field observations found damp and bubbling plaster in the northeast and southwest corner domes of the ceiling, as well as flaking paint on barrel vaults in all four quadrants (Figure 6). Visual inspections of the exterior found deteriorated masonry conditions with potential avenues for moisture migration (Figure 7). An investigative program was designed by BTGA and WJE, accepted by the Department of Administration’s Facilities Division, and authorized for implementation in 2012. The investigative phase of work included inspection openings on the interior; determining exterior conditions of marble and mortar joints, windows and sealants; roofing; and selective stone removal. Conditions after rain events that precipitated wetting of the plaster areas were observed. INTERIOR SURVEY AND INSPECTION OPENINGS On the inside of the building, ceiling access openings were made (working from scaffolding) to allow inspection of the domes in the two problem corners and in four barrel vaults. These interstitial spaces were connected with small windows between sections but were not accessible to the investigators without hazard gear. Investigators found blown-in cellulose insulation above all the ceiling openings. The dome insulation was found to be saturated—actually dripping with water. Little wonder that the plaster beneath it continued to deteriorate. The access openings were also used to view the surface of the exterior walls during water testing. EXTERIOR SURVEY Visual inspection of marble surfaces revealed voids, cracks, and improperly installed sealant. These conditions were documented on elevations that were incorporated into the repair documents. Conduits and light bases were moved to inspect for penetrations in the roofing at the base of the tourelle, yet they were surfacemounted. Erosion of the mortar was identified on upward-facing joints. Flat capstones at tourelle corners were found to slope backwards, creating catchment areas that drained into open joints below column bases. Sealant was noted on the top edge of the wall flashing, likely trapping water in the wall. WATER TESTING Water testing with a spray rack was conducted on all four tourelle sidewalls, from below the roof level moving upward. Tourelle interior decks were flooded with 4 in. of standing water. Third-floor roofs and sidewalls of tourelles and flashing lines were independently sprayed with hose 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 W a t e rs t o n a n d T h o r n t o n • 1 8 3 Figure 5 – View of interior ceiling with faux painting and gilt accents. Figure 6 – Damaged ceiling paint and plaster in the northeast dome. Figure 7 – Open joints, most common at water table below tourelle. stream. Generally, water was applied from the base of the main dome roof level up the wall on the short walls. On the larger walls, a spray rack was utilized, and they were sprayed for approximately three hours (Figure 8). The short walls were sprayed with a nozzle for one hour (Figure 9). Only one of 14 areas tested revealed active leaking. The west face of the northeast tourelle had active leaking visible from the interstitial space (Figure 10). SELECTIVE STON E REMOV AL An inspection opening was made in the location of the reported leak during water testing. Our investigations determined that the original copper wall flashing was installed only three inches into the mortar joint. The face marble block was removed, and the block was found to be 6 in. thick, set in mortar on backup brick (Figure 11). The mortar in the collar joint and the setting bed was found to be friable, and we noted cracked brick, split faces, and brick fragments. The flashing base was rebuilt and a liquid-applied roofing membrane flashing installed the full depth of the marble, turned up 6 inches before the marble was reset with stainless steel pins. 1 8 4 • Wa t e rs t o n a n d T h o r n t o n 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 Figure 8 – Water testing at tall wall of tourelle base, using spray rack. Figure 10 – Leaks observed from attic below short wall of northeast tourelle. Figures 9A and 9B – Water testing at short wall of tourelle base, using spray nozzle. Figure 11 – Marble block removed from short wall of tourelle base. Note friable mortar in collar joint and brick masonry backup. Further testing showed that the previous leak was resolved by this method. MATERIALS TESTING Materials testing showed no asbestos in sealants or flashings, but it was present in bituminous dampproofing material applied over sidewall flashings above the original roof. The setting mortar between the marble and brick backup masonry was friable and damp, likely the result of repeated wetting of the mortar and freezing and thawing cycles. Laboratory observations of mortar between marble showed a single layer of gray-white mortar—likely an adhesive mortar used in pointing. The original mortar was brown, with high sand aggregate content and low binder volume. There was evidence of calcium carbonate, pointing to the original presence of lime binder. No Portland cement or slag particles were observed in the sample. ANALYSIS AND RECOMMENDATION S Although one active leak was discovered, it could only partially explain the moisture observed in that northeast tourelle, and no active leaks were identified on the southwest where the initial plaster fall had occurred. Repointing had been completed on the southwest prior to this investigation, but dampness remained. Efflorescence patterns reflect many years of moisture migration from the tourelle bases but were never observed in a rewetted condition during the investigation, including observed rain events. The high levels of air movement suggested that condensation might play a role in retaining interiorgenerated moisture. The reroof may have resolved these issues from exterior water. The insulation was retaining moisture from all sources and holding it against the plaster. As such, it was recommended that this material be removed completely from the domes and adjacent arches experiencing issues—the northeast and southwest tourelles. Additional access holes were recommended to facilitate vacuum removal. Repointing may have added to the moisture load at the southwest tourelle, although subsequent observations will show that the moisture load increases with rain events. No direct moisture paths could be identified. It was recommended that the repointing at that corner be extended to the rails and tourelle interior faces, as well as sealant replacement and fastener corrections recommended for all corners. A combination of settlement, open joints, reduced interior drying, and misplaced sealants blocking drainage resulted in the mass masonry wall becoming saturated. In addition, the sandblasting that was performed in the 1970s removed the polish finish on the marble, increasing the porosity of the marble and the amount of water in the wall. The tourelle bases at some locations are three feet thick with marble facing. The porous nature of the bricks results in significant water being absorbed at each tourelle base. When in equilibrium, the wall absorbs moisture and expels it at a rate that does not result in liquid expressed on the interior surface. However, with the physical properties of the wall components in combination with the reduction of the air movement system, the addition of insulation, and the repointing mortar and sealants reducing exterior evaporation, the system became overwhelmed. Our goal was to return the system to equilibrium in order to address interior water manifestations. CON STRUCTION REPAIRS A report based on the investigative findings was presented to the State of Rhode Island Department of Administration in July 2012. A repair budget was developed and approved for action to be completed by December of that year. Work commenced in early September with pointing crack repair (Figure 12). A mock-up area on the tourelle low sidewall was used to establish the methods to remove the inferior repointing mortar without damaging the marble. The existing repointing mortar proved quite hard and tenacious. A satisfactory solution was found with the use of a sealant cutter and hand chiseling against the edges. A repointing trial was implemented using two putty mixes—one with premixed lime putty and sand and one using hydraulic lime and sand. Brush-tamped finishing was used, and custom tools were fabricated for placing mortar lifts in the 1/8-in.-wide joints. The lime putty worked and finished well. The initial hydraulic lime putty mix from the Midwest interacted with marble to create a brown-green staining (Figure 13). A hydraulic lime from Reston, Virginia, did not exhibit this staining and was approved for use. The work was required to be complete by January 1, and the weather forced the use of the hydraulic lime mix. As weather deteriorated, two corners were enclosed to heat the space and allow curing of the mortar. As wet work outside ended, plaster restoration began. Removal of the cellulose insulation and drying of the base plaster were the first steps. Base plaster in most areas was over 4 in. thick. The southwest areas dried well and permitted repainting quickly. However, the northeast, where the heaviest damage had occurred, dried very slowly where the base plaster met the outside wall. Shaping was accomplished, but finish paint could not be applied, as the 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 W a t e rs t o n a n d T h o r n t o n • 1 8 5 Figure 12 – Repointing work underway. surface moisture (tested by pin-type meter) stayed above 25% at some locations. Faux panels were created of painted board and applied over the trouble areas to allow continued drying prior to final repaint after the next legislative session. (The Rhode Island legislature generally meets from January until June.) As not all four corners required heated enclosures, savings from that work were used to expand the repoint to all connecting walls of the dome base and included window restoration and sealant replacement. This work was accomplished during the session as outdoor-only work. In early November 2012, at the start of the heating season, an interesting condition was discovered. Upon visiting the site on consecutive days, BTGA found the conditions in the north roof access stairwell astonishingly different. Before, the stair was chilly but not uncomfortable, and the door at the top would be propped open for work access. On the next day, the temperature in the stairwell was over 80º, and water was condensing into rivulets on the walls and outside door. It was discovered that the building’s steam heat had been turned on and that a leaking steam line below the floor of the State Library on the second floor was emitting steam that came through the Library’s stained-glass ceiling perimeter vents into the interstitial space connecting with the stair and up the stairwell. According to the librarian, they were aware of the leak, but the temperature in the library was bearable. It is possible that this additional water load had been added for several years to the base of the northeast tourelle. The condition was reported, and the steam line was disconnected and the condition stopped. The library also became more comfortable. MON ITORING SYSTEM On December 3 and 4, 2012, prior to the construction work being fully complete, WJE installed wireless electronic sensors designed to measure air temperature, relative humidity, and the moisture content of the material to which they are attached. The sensors are battery-operated and automatically collect and transmit their data on an approximate five-minute interval. Data collection stations, called gateways, were also installed. The gateways are designed to collect data from the sensors and transmit that data via a cellular modem to a remote site. The data are then available to view or download from a website for remote monitoring. Twenty-one sensors and two gateway devices were installed. The sensors are located in the corner domes at the third floor of the Rhode Island State House. Sensors are located at the exterior brick masonry walls above the plaster ceiling, in the air space above the ceiling, and in the plaster on the attic side and on the interior side of the corner domes. WJE also placed sensors at the arches adjacent to the cor- 1 8 6 • Wa t e rs t o n a n d T h o r n t o n 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 Figure 13 – Staining from initial repair mortar interaction with marble. Figure 14 – Plan of the ceiling indicates relative position of the sensors. ner domes, in the exterior wall, and in the plaster on the attic side. WJE also installed two wireless surface condensation sensors on the interior of the brick masonry that measure surface temperature and potential for condensation—one on the north wall and one on the south wall. WJE accessed the space above the ceilings at the cutout openings in the dome and arch ceilings. In order to reach the exterior wall, we used an aluminum pole to mount an adhesively applied hook to the surface of the brick masonry. The sensor was then hung from the hook with its probes in contact with the brick surface. For the plastermounted monitors, we used bright metal screws fastened directly into the plaster. The sensors measuring the temperature of the air within the attic were suspended within the space using fishing line. The gateway devices are mounted to the masonry walls within the closets off the northeast and southwest third-floor corridors. The gateways are plugged into a standard electrical outlet. Should there be a loss of power, the gateways will no longer collect or transmit data but will automatically resume data collection once power is restored. The software used to remotely monitor the sensors and gateways sends an alert e-mail should there be any inactivity (due to power loss, sensor failure, etc.). A plan of the ceiling is included (Figure 14) indicating the relative position of the sensors and their location within the ceiling assembly (i.e., exterior brick wall, attic-side plaster, etc.). The gateway device locations are also indicated on the plan. RESULTS After a complete annual cycle, no plaster deterioration has been observed. No moisture condensation has occurred, although conditions have been close to favorable several times. The exterior walls are slowly drying as hoped. The plaster is also drying slowly, and final paints were applied in the fall of 2013 on the two problem surfaces. Other than one persistent but muchreduced leak at the corner of the northeast tourelle, no other interior leakes have been observed. The roof vendor will be contacted to investigate this again, since all masonry efforts in the area have been exhausted. The southwest tourelle exterior wall masonry still spikes with rain events, and replacement of the emergency repointing, consistent with the final repair work on the other three corners, is being considered. The emergency repointing mortar mix contains Portland cement and may be slowing exterior evaporation. The speed with which it responds to rain events—even though water testing failed to replicate the condition— also implies that hairline cracks may not have been addressed, or bond separation has occurred in the newer mortar work. Although the tourelle bases have been stabilized, observed deterioration of the exterior tourelle components and roof continues. These areas are outside of the area studied and above the level of water infiltration. Parging of the vaulted brick roof structure is spalling at an accelerated rate. Spread cracks of the marble cornice ring were also noted. It is r e c o m m e n d e d that further study and stabilization of these f r e e – s t a n d i n g elements be conducted in the next phase of work. Plans are in place for a new HVAC system within the building that would use the interstitial spaces for duct distribution and seek to seal the remaining vent openings from individual spaces. We recommend that this plan be reevaluated in light of findings from this study to ensure that either sufficient air movement is allowed or that humidity levels are addressed to ensure that interior drying can continue at a sufficient rate. While we only removed insulation at the two problem corners, we do consider that material to be a future liability, which reduces temperatures above the ceilings and exterior walls and, hence, heightens relative humidity and the potential for condensation. Footnote 1. Ronald J. Onorato, RI State House Centennial Exhibition Catalog, 1996. 2 9 t h RC I I n t e r n a t i o n a l C o n v e n t i o n a n d T r a d e S h ow • Ma rc h 2 0 – 2 5 , 2 0 1 4 W a t e rs t o n a n d T h o r n t o n • 1 8 7 TEAM CREDITS Owner State of Rhode Island Department of Administration Operations Management Division O ne Capitol Hill, Providence, RI 02908 Executive Director: Ronald J. Renaud Facilities Management Asst. Director: Arthur J. Jochmann II Advisors Rhode Island Historic Preservation & Heritage Commission Old State House, 150 Benefit Street, Providence, RI 02903 401-222-2678 Executive Director: Edward Sanderson S enior Review Architect: Virginia Hesse Architect and Masonry Specialist B rewster Thornton Group Architects, LLP 150 Chestnut St., Providence, RI 02903 401-861-1600, www.brewsterthornton.com Principal-in-Charge: Barbara J. Thornton, AIA Project Manager: Schane Tallardy, Assoc. AIA Testing and Monitoring System Engineer and Materials Consultant Wiss, Janney, Elstner Associates, Inc. 311 Summer Street, Suite 300, B oston, MA 02210 617-946-3400, www.wje.com Principal Investigator: William Waterston General and Masonry Contractor East Coast Masonry & Restoration, Inc. 515 Greenville Ave., Johnston, RI 02919 401-232-0562 O wner: Michael St. Angelo Master Mason: Matt Miller Subcontractor for Interior Restoration E .F. O’Donnell & Sons Co., Inc. 75 Dike St., Providence, RI 02909 401-351-8505 O wner and Project Manager: Robert O’Donnell
