AUSTRALIA ICOMOS CONFERENCE, 2015 FABRIC – THREADS OF CONSERVATION CONSERVATION OF THE STATE LIBRARY OF VICTORIA : CASE STUDY Dan Blake
Dan Blake Lovell Chen Architects & Heritage Consultants Level 5, 176 Wellington Parade East Melbourne VIC 3002 +61 (0)3 9667 0800 firstname.lastname@example.org
PRESENTER BIOGRAPHY Dan Blake joined Lovell Chen in 2007 and is a Senior Associate specialising in building conserva>on and restora>on. In 2012, Dan aCend the Diadrasis Martos Workshop— for the conserva>on of a sixteenth century stone monument in Spain, run in associa>on with ICRROM, University of York, University of Jean, the City of Martos, the Department of Culture, ADSUR, IAPH and IPCE. A number of projects on which Dan has contributed his skills have received peer recogni>on including the AIA (Victoria) Architecture Awards 2014 commenda>on and John George Knight Award for Heritage for the Melbourne GPO roof and façade conserva>on works. B Arch (RMIT) MA Cultural Materials ConservaZon (Melb) in ﬁeri AIA AICCM MICOMOS
ABSTRACT The State Library is one of Victoria’s most important public buildings. Our most recent works involved the external conserva>on of the Dome Reading Room and Western Façade. Constructed in several phases from the mid 19th century through to the early 20th century, the fabric of the façades exhibits changes in construc>on techniques and materials. The works to the Dome Reading Room, comprising the repair of cement stucco and the Kahn reinforced concrete system, have presented many challenges, not least the limita>ons on structural loading and diﬃcult access due to the rela>vely recent inﬁll of the surrounding courtyards. Trials of washcoa>ng systems were undertaken using diﬀerent forms of liquid Copperas and pigment cement. This raised ques>ons about the intended opacity, the current intent with respect to masking colour diﬀerences in patch repairing of render, and the longevity of an untried coa>ng system. Client expecta>ons with respect to costly and diﬃcult access, the public proﬁle of the building and how a signiﬁcant change in the colour and presenta>on of the Dome might be received were also important factors to consider. The works to the Western Eleva>on presented complex challenges in deﬁning the extent of conserva>on with respect to the nature of previous repairs.
SNAPSHOT Our recent works at the State Library of Victoria (SLV) have been focused on the conserva>on and restora>on of the exterior of the Dome Reading Room (c.1913) and the Western Façade (c.1854). These two dis>nct areas of the building represent major phases in the development of the Library in the nineteenth and into the twen>eth centuries and demonstrate signiﬁcant changes in technological advancement over sixty years of the construc>on industry. (Lovell Chen 2011 pp 7-‐43) The SLV is recognized as one of the earliest ‘free public libraries’ within the greater interna>onal context — being contemporary with those of London, Paris and New York — and at the >me of its construc>on it was one of the most visited place in the world. (Lovell Chen 2011 pp 7-‐13)
Figure 1 Diagram indicaZng stages of development of the State Library of Victoria site from 1850 – 1970. Source: ConservaZon Management Plan, Lovell Chen
In 1999 the Dome Reading room underwent major renova>ons internally and to the roof to reinstate the skylights, which had been covered over in the 1960s. These works did not focus on the render or Khan reinforced concrete dome drum, and with the arrival of the
centenary of the Dome Reading Room, works commenced to repair the dome drum and western façades. The key component of the works to the dome drum was in the discovery of remnants of the original Copperas (bright gold-‐yellow) ﬁnish to the dome render. The ini>al review of remnants of this yellow coa>ng system suggested a possible paint ﬁnish. Microscopic review of the coa>ng system indicated that the yellow colourant appeared to be a dispersion of iron par>cles in a lime-‐based (calcium carbonate) binder. Samples were removed from the site and x-‐ray ﬂorescence (XRF) analysis was undertaken on the front and back of the sample to measure the diﬀerences between the composi>on of the coa>ng, which had only been applied to one side of the render, with that of the substrate, par>cularly given the similari>es in the use of materials. Both the front and back of the sample contained traces or iron (Fe), calcium (Ca) and >tanium (Ti). Unusually the front of the sample indicated the presence of lead (Pb) in the L-‐lines, which peaked towards the end of the scan.
Figure 2 X-‐ray ﬂuorescence spectra data from rear of sample 1. Source: University of Melbourne CCMC
Figure 3 X-‐ray ﬂuorescence spectra data from front of sample 1, note the presence of lead (Pb). Source: University of Melbourne CCMC
Our ini>al response to the presence of lead was to ques>on whether the system may be lead-‐based paint. However, further assessment indicated that the presence of lead in the L-‐ line suggested the presence of another element when reviewed under the K-‐line. On this basis the presence of lead was ruled out and it was concluded the coa>ng system was a composi>on of iron-‐sulﬁde, commonly know as Copperas. With the assistance of Dr. Donald Ellsmore, we began trials with of reinstatement of the Copperas coa>ng system, with a view to achieving a uniform ﬁnish and satura>on of Copperas in new render repairs and exis>ng render. It is understood that the reac>on of Copperas in turning yellow requires ac>ve lime content. The older renders posed a problem as much of the ac>ve lime content had oxidized and would not provide the reac>on to an acceptable depth of satura>on to match the new repairs. A number of systems were trialled, some containing lime water, lime puCy and other such combina>ons of Copperas and lime. These trials achieved a yellow coloura>on with a level of varia>on that was consistent with early photographs of the dome drum. The satura>on increased as the reac>on con>nued over a number of days; however the end result presented substan>al diﬀerence in satura>on between new render and the exis>ng regardless of the method used to apply the Copperas.
Figure 4 ApplicaZon of Copperas coaZng over secZon of both new and exisZng render. Note the saturaZon of the new render with acZve lime content compared with the older render lacking in lime. Source: Lovell Chen 2014
Figure 5 Further samples of Copperas using lime puey based wash. Right side of image contains a second coaZng of Copperas soluZon. Note far right side of image contains small remnants of original copperas coaZng which has weathered. Source: Lovell Chen 2014
Figure 6 Further samples of Copperas using lime water and copperas in a number of coaZngs over new and exisZng render. Note the varied level of saturaZon achieved between the exisZng render and new. Source: Lovell Chen 2015
Figure 7 Further samples of Copperas using lime water and Copperas in a number of coaZngs over new and exisZng render. Note the varied level of saturaZon achieved between the exisZng render and new. Source: Lovell Chen, photo Andrew Latreille, 2015
Further review of the ﬁnish posed complex issues to be resolved before proceeding with the applica>on of a reinstatement coa>ng system. The diﬃcult access to the site due to the in-‐building of the complex meant that opportuni>es to revisit the drum for future maintenance are limited and the cost prohibi>ve. On this basis, the performance expecta>on for the works needed to be between 30 and 40 years, with minimal maintenance. The high proﬁle loca>on of the SLV also meant that public interest in the external presenta>on of the building was another key considera>on, with our client keen to ensure it presented well. The SLV is a building to which all Melbournians and to a degree all Victorians feel a close nostalgic link and have a real sense of ownership.
With a degree of uncertainty as to the long-‐term performance of the Copperas and the degree of varia>on between new repairs, old repairs and original stucco in the test samples, it was determined that a further alterna>ve needed to be inves>gated — a similar coa>ng system that incorporates natural pigment into a cemen>>ous coa>ng known as a washcoat. This system had been extensively tried and was known to last well in excess of 20 years. It was also able to provide the uniform satura>on between new and old cement ﬁnishes, and to be mixed to match the colour achieved by Copperas whilst s>ll achieving the level of varia>on that could be iden>ﬁed in early photographs. While the washcoat may not be the most ideal treatment from a conserva>on perspec>ve, given it is a varia>on on the original intent, it does deliver the requirements of the public and the client, whilst achieving the overall aesthe>c intent of the Copperas ﬁnish. This system will age and wear in a similar way to Copperas, eventually returning the stucco to a grey render ﬁnish with only minor fragments of the original yellow colourant to be found below cornice ledges where protected from weathering.
Figure 8 North elevaZon of the SLV dome c1914. Note the variaZon in the coloured ﬁnish on the render. Source: State Library of Victoria
Figure 9 SLV dome with washcoat applied to render. Source: Lovell Chen 2015
Figure 10 SLV dome before commencement of works, note remnant yellow colouraZon to render on right of downpipe. Source: Lovell Chen 2015
Figure 11 ApplicaZon of the colour pigment washcoat over render, applied by brush over surface of render Source: Lovell Chen 2015
Figure 12 Trials of applicaZon of the colour pigment washcoat over render, applied by brush over surface of render and compared to colour of exisZng remnants of Copperas. Source: Lovell Chen 2014
Another complica>on in delivering a range of conserva>on works to the western façades was the degree to which works had been undertaken in the past (by others). The western façades where constructed in ﬁve stages, using a number of diﬀerent sandstones, including Kangaroo Point (Tas), Heatherlie Quarry Stawell and Grampians Sandstone (Vic). A later phase constructed in the 1960s comprises precast cement and rendered masonry. Many of the repairs of the last 30-‐40 years had been undertaken using sandstones from Sydney, such as Hawkesbury (NSW) and Pontville White (Tas) (King & Weston 1997 pp 179). The composi>on and characteris>c of these stones are not a good match to the original stone and, as such, appear quite diﬀerent within the overall presenta>on of the building. Other repairs where undertaken using cement-‐based mortars and, while these are not discernable from a distance, when viewed in close proximity they are generally unsympathe>c to the overall aesthe>c, and have resulted in decay to the highly decora>ve carved stone details. It was determined that limited repairs be undertaken to the western façades, acknowledging that whilst the previous repairs may appear unsympathe>c, they are stable. On this basis they have been leo un-‐altered un>l a >me when they require more substan>al repair or replacement.
Figure 13 New stone indent works to the western porZco, note the use of diﬀerent sandstones in the column where previously stone has been replaced. Source: Lovell Chen 2015
Figure 14 Previous repairs to stonework, note the varied colour, composiZon and texture of the stone used to repair the face of the cornice. Source: Lovell Chen 2014
Figure 15 Previous repairs to boele balusters where the rear secZon of the boele had been cut out to accommodate new concrete upstand for the altered gueer. Source: Lovell Chen 2014
Figure 16 A complex access system was required to undertake the works to the dome due to the lack of loading capacity of the inﬁll courtyard roofs. The scaﬀold is suspended from the dome walls.
Source: Lovell Chen, photo Andrew Latreille, 2014
Figure 17 Finished presentaZon of the dome drum. Source: Lovell Chen, photo Andrew Latreille, 2015
Figure 18 Diﬀerent phases of construcZon in diﬀerent sandstone types, note unsympatheZc repairs to the modillions (right) Source: Lovell Chen, photo Andrew Latreille, 2014
BIBLIOGRAPHY King R.L., Weston K.S., 1997, Dimension stone in Victoria, Geological Survey Report 112, Victorian State Government Department of Natural Resources and Environment, Lovell Chen, 2011, State Library of Victoria Conserva>on Management Plan Vol 1
AKNOWLEDGEMENTS Dr. Donald Ellsmore University of Melbourne CCMC
by Dan Blake