installation

Paved with good intentions A contribution to Discovering Stone by Dr Pingo Dripstone

The good Dr Dripstone (a building diagnostics expert of significant international renown) returns to the fold with one of his part humorous but vitally important articles on the problems that can and do beset our industry. The Doctor’s findings mirror some of the comments made by others in this edition in regard to knowing and understanding the product you are dealing with. In this instance the subject is paving. Many of Melbourne streets are paved in native bluestone as are landmarks

T

he anonymous photographs reproduced in this article illustrate examples of severe picture-frame staining of recent granite paving on sand-cement mortar beds in major public venues in Singapore and Macau. This writer would happily provide photographs of similar verifiable examples in Sydney, Melbourne, Canberra, the glorious fishing town of Port Lincoln and elsewhere in Australia, but resiles from so doing, having lately developed acute allergies to both literal and figurative tar, feathers and injunctions. In the course of nomadic travels, often one step ahead of writ-servers and the purportedly defamed, this author has lately seen and occasionally stopped briefly to investigate similar incidents in other cities of Australia and Asia. Some occur on landmark projects designed by eminent international architects and built by experienced multinational contractors. Picture-frame staining has blighted paving of recently refurbished railway station platforms in

Paving of Giallo Santa Cecelia gneiss at the porte cochere of a 5-star hotel in Macau, here recorded less than a year after installation.

Finance Centre Hong Kong. It has lately been reported in granite paving of newly refurbished suburban shopping centre north-west of Sydney. If not ubiquitous, it seems to be heading that way. The basic cause of picture-frame staining in paving has become widely known; see, for instance, Out Damn Spot – Or Not which first appeared in Discovering Stone issue 4 in 2003 and remains available in the Advice section of www. infotile.com under the ‘Sealing & Maintenance’ archive in Archived Articles. All hydrated (cured) Portland cement contains readily soluble calcium hydroxide, Ca(OH)2. Bedding mortars contain at least 15% and as much as 25% cement. That provides an ample reservoir of hydroxide that can be dissolved and mobilized by rainwater, cleaning water and irrigation water that penetrate fine gaps and cracks in grouted joints and seep under paving. When calcium hydroxide in dilute solution comes into contact with the air and starts to concentrate as the water evaporates, it can react

The start stage of picture-frame staining:Water and dissolved calcium hydroxide in saturated bedding rise through grouted joints and spread sideways through micro-porous exfoliated stone, in this instance possibly the famed Bear Bile Gold granite of Outer Fujian.

Caption

1 2 like Parliament House and Southern Cross Railway Station. Few problems if any are ever encountered because the end users understand the nuances of a stone which has been successfully used for more than a century. Problems occur when assumptions are made that a particular stone will behave precisely like another.

30

Sydney and plazas and stairs of an aquarium in Hong Kong. In one instance in Australia, pepper & salt granite paving of a public monument has stained so badly that it is likely to be replaced after four years’ service. The building was intended to have a design life of not less than 60 years. Severe picture-frame staining occurs in red granite paving of Singapore’s Esplanade Theatre and Concert Hall, in grey granite paving of the Singapore Management University, in grey granite paving of Macau’s Cultural Centre, in Giallo Santa Cecilia paving of a new 5-star hotel nearby and in paving and lobby flooring of the International

issue 20 discovering stone www.infotile.com/publications

with atmospheric carbon dioxide, CO2. The result is calcium carbonate, CaCO3, a dull-surfaced dense and seemingly amorphous whitish material that is insoluble in water. As calcium carbonate impregnates and clogs pores immediately below surfaces of partly-translucent crystalline stone such as granite, it changes the proportions of daylight that are reflected, scattered and absorbed. Depending on the wetness of the surface and immediately below, the stone may appear anomalously dark (and seemingly wet, even when demonstrably dry) or lighter than usual. This so-called wet spot optical effect is the

basis of the phenomenon of the picture-frame staining shown in the photographs. Popular counter-measures include dip-sealing (six-sided sealing) stone pavers with various penetrating, breathable, non-film-forming water repellants and sealers; use of pore-blocking admixtures to densify bedding mortars and immobilize soluble salts in Portland cement, provision of thin plastic drainage cells between membranes and bedding layers and casting of networks of shallow channels (in one instance termed drainage superhighways) in upper surfaces of substrate slabs. These and other measures were employed in 2010 and 2011 for granite paving of a metropolitan council’s concert hall and theatre complex in Australia. The experienced and cautious main contractor engaged the services of a reputable tiling consultant specifically to propose and refine measures to prevent pictureframe and other staining associated with calcium carbonate. The consultant specified and detailed a package of conventional and comprehensive measures, all of which were duly installed. Despite these efforts, conspicuous picture-frame staining occurred before the building was completed. The contractor’s position is that it rigorously followed and even exceeded its consultant’s advice and corresponding best industry practice, for instance as explained and illustrated from time to time in articles in Discovering Stone and in design handbooks. It observes that its stonework sub-contractors complied with the architects’ specification and the published recommendations of manufacturers of component materials. The result, however, satisfies no-one. Part of the same area of paving when dry.The only barrier to flow of water between pebblecrete topping in the distance and granite paving of a colonnade in the foreground is a metal insert.

Thin stone and cement Once upon a time, much stone paving consisted of slabs so massive and thick that they did not need to be glued down. They stayed in place between (1) joints of washed sand and quarry dust packed tightly enough to resist displacement but able to accommodate some compression, and (2) deeper and more solidly fixed kerb stones. If the joint filler washed away, it could be easily replaced. If individual pavers needed to be removed and/or replaced, they could be lifted without much risk of further damage. NOTE NOTE: If this all seems strangely familiar to habitual boulevardiers among our readers, it may be because city councils in Melbourne, Sydney, Perth and beyond have returned to these quaint and antediluvian techniques in re-paving CDB footpaths, malls and pedestrianized laneways such at the shadowy but exotic thoroughfare leading to Atelier Dripstone Bar & Grill® in the fabled Paris end of Melbourne. The countervailing trend since the 1980s has been to use stone pavers of decreasing thickness and plan dimensions adhered to cementitious substrates with cementitious tile and other adhesives and separated by cementitious grouts; indeed, the methods used for natural stone paving are now often indistinguishable from those employed to fix ceramic tiles of similar dimensions. The stone pavers in the accompanying photographs range from 200 x 200 mm to 600 x 600 mm in face dimensions and 20 mm to 30 mm in thickness. The proliferation of cement in paving systems may have been a boon for the cement industry (which, in passing, I greatly admire and in which the Dripstone family-controlled Big Slippery Holding Company of Vanuatu and Cattle Fodder Agistment Trust #16 of Scratchback Downs, Queensland, have moderate ethical

“If following best industry practice – in this instance and elsewhere – conspicuously fails to forestall picture-frame staining, could there be some fundamental shortcoming in the specified stone…”

3 If following best industry practice – in this instance and elsewhere – conspicuously fails to forestall picture-frame staining, could there be some fundamental shortcoming in the specified stone (in this case one of the plethora of paving granites exported from China) or in the paving system generally? It is slowly becoming evident that in many hapless projects, the likely answer to both propositions is yes (or yes-lah!). Could it be that thin stone paving on sand-cement bedding has developed down an evolutionary path to a dead end at which satisfactory results are increasingly the exception rather than the norm? I, Dr Pingo Dripstone of no fixed abode but at home almost everywhere, believe this to be the case and so offer below some diverting thoughts on the matter.

shareholdings) but it has not necessarily improved the appearance of stone paving, particularly where neat cement is broadcast or applied as a slurry onto bedding mortar to adhere slabs...but I digress. There is, one suspects, some correlation between the thickness of stone paving and its resistance to staining by calcium carbonate originating in cement of bedding mortar. Moreover, there are rumours afoot, among wizened paviors and tilers of several Old Schools, that the proportion of soluble salts in ordinary builder’s cement is gradually increasing. This may be an indirect result of efforts to reduce energy expenditure in production of cement. One way or another, cement in paving systems should be minimized and, if possible, immobilized. Or even eliminated.

discovering stone www.infotile.com/publications issue 20

31

installation

Drainage falls To prevent dissolved calcium hydroxide rising through grouted joints and spreading onto the surface of stone pavers, it is essential to provide drainage falls at bases of porous bedding layers as well as over finished surfaces. Perhaps 99 per cent of rainwater flows to drainage across paving surfaces. Some part of the rest infiltrates shrinkage cracks and gaps in grouted joints and trickles down into the bedding. Water within bedding mortar is intended to be intercepted and to drain across waterproofing membranes. However, if water ponds on membranes before they are covered, it is at least as likely later to saturate bedding to the same depth. Water will also spread and rise through capillary action. For water in mortar bedding to rise to the surface, it is not necessary for the whole bedding layer to be saturated. If grout in joints and underlying bedding are locally saturated when the temperature of paving rises, vapour Figure 4. Calcium carbonate present in the joint between marble tiles appeared when the pressure can force that water back to the surface through room was not air-conditioned. External paving joints warmed by sunlight can exhibit the fine fissures in the grout. Figure 4 shows increasingly same phenomenon. 4 close macrophotographs of a joint between marble floor tiles in a bathroom of a five-star hotel in Macau. Small to one or all six sides of stone pavers before they are laid and hollow cones of calcium carbonate have been deposited by calcium sometimes also to exposed surfaces thereafter. Reliance on sealers hydroxide rising to the surface whenever the guest suite is empty and temporarily not air-conditioned. The same phenomenon occurs is becoming more critical for materials of imprecise provenance when sunlight warms external paving. (perhaps somewhere in Fujian), untraceable designation (Bear Bile Gold) and without genuinely comparable precedents of long-term In contemporary construction, drainage slopes under paving on performance, all laid to sub-minimal falls using excessive cement. podiums and plazas are rarely achieved by tilting suspended concrete In effect, to achieve fitness for purpose, it is increasingly demanded slabs. Tilting slabs is seen as an unnecessary and expensive complication. that natural stone be turned into a polymer-stone composite. Perhaps Slabs are cast flat and membrane waterproofing is applied to graded the industry has forgotten that paving was once, not so long ago, screeds on top of structural slabs or, as in one recent instance of severe successfully built and maintained without the myriad chemical picture-frame staining of paving at a casino in Singapore, a weak and treatments on which it increasingly relies? porous 60 mm thick screed applied directly to flat slabs without falls, all on the assumption that water within porous 60 mm thick slabs will eventually find its own way to distant outlets and drains without rising from the undersides of paving on the way. In investigating membrane waterproofing failures in paving and flooring, it is common to find membranes turned up and fixed on tops of puddle flanges of two-stage metal and plastic outlets embedded above rather than recessed below finished levels of concrete slabs and screeds. The combination of a 3 mm thick flange and a 2 mm liquidapplied membrane provides a 5 mm high dam to water flowing into an outlet. On a slab or screed with a slope of 1:100, that slight step can generate a disproportionately damp area about a metre in diameter. It is also common to find essential steps and hobs, at transitions between paving and walls, value-engineered into oblivion by architects, structural engineers and design & build contractors. They presumably do not recognize or under-estimate the possible consequences of inadequate and unreliable drainage falls and insufficient vertical freeboard surfaces against which to turn-up and terminate membranes, particularly below sills of glass walls and doors. As noted in a previous article in Discovering Stone, detailing level transitions between external paving and internal flooring at bases of glass walls is one of the most failure-prone tasks in design of contemporary commercial and institutional buildings. When the need for a step of 150 mm or more to achieve adequate drainage falls away from sills and to tailor and terminate membrane waterproofing is identified, it is often too late; concrete has been cast.

Sealers, magic bullets and confounded optimism To overcome the recognized risks in paving with thin stone on sand-cement bedding, it has become conventional to specify that penetrating sealers, impregnants and water repellants be applied

32

issue 20 discovering stone www.infotile.com/publications

In a report recently submitted in a looming dispute over pictureframe staining of the plaza of a public building, a stone consultant has speculated that the affected granite would have been fit for purpose if the paving had been made completely water-resistant by substituting polyurethane sealant for kilometres of grouted joints. He criticizes the builder for installing drainage channels under the paving on the grounds that to do so was, from the outset, a tacit admission that water would eventually reach the bedding mortar and mobilize soluble salts. He argues, in effect, that hundreds of square metres of paving of 600 x 300 x 30 mm thick stone, incorporating movement joints, grated drains and penetrations for stainless steel handrail posts, flagpoles and rainwater outlets, should have been designed and expected to perform as a fail-safe impermeable membrane to prevent water reaching under the stone. Is this actually possible? In another long-running dispute over staining of stone in wet-area floor and wall tiling in a high-rise apartment building, another consultant has declared that a particular limestone could have been made fit for purpose if 300 x 300 x 10 mm tiles had been sealed on the underside with an epoxy coating before laying and sealed on the top, from the outset and thereafter at regular intervals of 12 to 18 months, with a clear penetrating water-repellant sealer. Is that not tacit acknowledgement that the stone is not fit for purpose without these heroic modifications and life-long interventions? Surely if minimally tolerable performance of a particular variety of paving stone relies on finding and implementing just the right mix of pre-installation barrier coatings and dip-sealers and postinstallation top-sealers, the use of pore-blocking admixtures and “nano” dispersants in mortar, further magic bullet admixtures in grout, maintenance by specialist and approved stone technicians and other compensatory measures, then it is reasonable to conclude

installation

from the outset that the respective stone is not fit for purpose! Perhaps it is time to redefine fitness for purpose in this context as having the properties necessary to allow a stone to perform satisfactorily on its own merits, without successive and cumulative modification, protection and preventive measures.

Geoff Quick of the CSIRO cautioned that “...stone units produced by thermal or flame processes (exfoliated surfaces) tend to have extended

It is perhaps not widely appreciated that some of the coarse-grained granites flooding the market in Australia and Singapore have few if any successful long-term precedents in use as 20 to 30 mm thick paving slabs. The predominant use in the regions of their origin has been in rough-faced cobblestone paving and cubic stone for masonry of retaining walls in landscaping. In both applications, roughness and irregularity disguise discolouration, from cement and otherwise, that is conspicuous and objectionable in relatively flat slabs. Reassurances that the Quin Emperor’s legions tramped along Bear Bile Yellow paving en route to put down rebellions of disrespectful Mongols and recruit seafaring eunuchs do not necessarily offer hope that the same stone is suitable for 30 mm exfoliated paving on cement mortar bedding in a suburban shopping mall in 2011.

absorption coefficient (or in Europe, an imbibition factor) is a measure of the water absorbed by a small cubic sample of stone immersed in unheated water for a specified period of 24 or 48 hours. Architectural specifications and stone suppliers’ test data sheets commonly cite measurement by ASTM C97, the Standard Test Method for Absorption and Bulk Specific Gravity of Dimension Stone.

microcracking and thus

ASTM C97 and similar test methods do not identify the effects of different surface treatments on water absorption. They do not measure surface porosity, a very different property measured by different methods that require use of sophisticated and expensive instruments. Porosity of a shallow layer at one surface, however, is critical to the explanation of picture-frame staining in stone paving.

have increased absorption properties.”

The effects of flame finishing granite Selection of stone for external paving and wet-area flooring is based on many criteria, one of which is water absorption. A water

To comply with criteria for paving slip resistance, natural stone is often specified to be exfoliated, that is, flame finished. All of the examples of picture-frame staining illustrating this article occur in flame-finished or bush-hammered granite and gneiss. In flame-finishing, stone slabs pass under closely-spaced gas jets that rapidly heat the stone and cause a shallow layer of crystals to expand, shatter and fall away. The depth of the disturbed layer is a function of the intensity of heating and the crystalline structure of the stone. Some architects and engineers have appreciated since the 1980s that the bending strengths of sawn, honed or polished slabs are higher

B AM S T ON E Manufacturers and suppliers of bluestone, granite & sandstone Hamilton Rd, Port Fairy VIC 3284 Australia Phone: (03) 5568 2655 Mobile: 0419 522 709 Fax: (03) 5568 2454 Email: [email protected]

it stands the test of time Bamstone sawn bluestone tested R13 - Slip Resistance AS/NZS 4586:2004 (tests conducted by CSIRO)

00

www.bamstone.com.au

issue 20 discovering stone www.infotile.com/publications

than the bending strengths of exfoliated slabs of the same stone and thickness. Best practice in design of mechanically-fixed stone is to base engineering calculations on direct measurement of the loss of strength caused by exfoliation.

pavers. Picture-frame staining has started to recur after two months. The reputed record is seven distinct bands in 450 x 450 mm granite pavers within 18 months.

A few designers have more recently realized, some by trial and error, that exfoliated stone surfaces are also much more porous – or microporous – than those of simple sawn stone. Intense heating exposes and opens microscopic crevices and grain boundaries. In the first edition of Discovering Stone in March 2002, Geoff Quick of the CSIRO cautioned that ...stone units produced by thermal or flame processes (exfoliated surfaces) tend to have extended microcracking and thus have increased absorption properties. Recent dye stain experiments and microscopy at Atelier Dripstone Bay & Grill® have confirmed the extent to which bush-hammered, sand-blasted and other mechanically-roughened stone surfaces are also more microporousNOTE – and thereby more likely to absorb water - than sawn faces of unfinished slabs. In 48 hours, water has carried fluorescein dye upwards, through thin layers of roughened stone, three times further than through intact stone at sawn edges.

Where to from here?

5 Figure 5: Contrasting light and dark stains where water and dissolved salts rise through joints and leap-frog over successive shallow bands of exfoliated stone in which pores are clogged with calcium carbonate. NOTE: It is not unusual to find porosity confused with water absorption, even in reports of stone consultants. Pumice, a glassy volcanic rock of rapidly solidified frothy lava, is among the most porous of stones, with a typical porosity of 90%. However it does not readily absorb water; for 20 years after the explosion of the Krakatoa volcano between Java and Sumatra in 1883, rafts of pumice floated around the Pacific and occasionally incommoded passers-by. As water carries dissolved calcium hydroxide upwards, through grouted joints, it may penetrate sideways only a fraction of a millimetre into sawn and sealed edges of pavers. However, when water reaches the uppermost few millimetres of micro-porous stone, that is, the shallow layer disrupted by exfoliation or mechanical shattering, it can quickly spread much further. The sideways spread progressively deposits solid insoluble calcium carbonate in sub-surface pores, crevices and interstices. At some point, stone in the perimeter band becomes so clogged that it resists sideways flow of water from joints. Lime-tainted water may then flow over the surface of that band and then into unclogged stone more distant from joints. This process of irregular leap-frogging of solutions of different dilution and precipitation rates accounts for the occasional occurrence of multiple, roughly concentric and contrasting bands of staining. Figure 5 illustrates an example recorded in Singapore; five bands after 14 months’ exposure in granite paving of a colonnade sheltered from direct rain. Figures 6&7 show nearby areas of the same paving after efforts to remove stains and replacement of some

Staining of stone paving by calcium hydroxide in Portland cement can be prevented by encapsulating cementitious materials (structural concrete, levelling mortars, screeds cast to drainage falls, etc.) under waterproofing membranes. The technology of direct adhesion of stone to membranes is under development and is already common in swimming pool and wet-area tiling. Another approach is that described above, namely paving of massive slabs of stone varieties known to resist absorption of water and dissolved salts, restrained at perimeters by walls and kerb-stones and grouted with compacted quarry dust. A fail-safe way to prevent picture-frame staining from migration of soluble cement salts from bedding mortars, adhesives and grouts is

6 7 Figure 6&7: Examples of picture-frame staining in granite paving sheltered under a colonnade but contiguous with wholly external paving up to 10 metres distant.

to eliminate the bedding, adhesive and grout altogether. That has long been practiced in Europe and North America by mounting pavers on adjustable synthetic rubber and plastic pedestals that automatically set joint widths. Individual pavers can be lifted without damage. All rainwater flows through the joints and can be directed to retention tanks and ponds, another growing trend. This technique has been used occasionally in Australia for at least 30 years. Of course nothing is perfect; pedestal-supported white marble paving around a prominent US Government building in Washington, D.C., once displayed grey and brown discolouration wherever unstable fluids in synthetic rubber of quarter-circle shims, installed loosely on top of the pedestals, — migrated through the stone to cause tacky surfaces and indelible stains.

Dr Pingo Dripstone SJ (ret.) is an occasional but erratic contributor to Discovering Stone and to in-flight magazines of obscure regional airlines in remote parts of Africa and South America. He has recently occupied the Collapsible Chair of Experimental Bluefin Tuna and Yabby Aquaculture at the Port Lincoln TAFE, South Australia. He has not appeared on Master Chef.

discovering stone www.infotile.com/publications issue 20

35