Establishment of performance criteria
In our case, the pavilion had to be built under an extremely limited budget, withstand arctic temperatures and be constructed by an amateur building crew. Accordingly, the ease-of-assembly of the construction and need for stable properties over a wide temperature range proved to be more critical aspects for the adhesive selection than obtaining maximized strength and acquiring a fully transparent, flawless, appearance. The focus was placed primarily in finding a structural adhesive that functions similarly to a mortar in traditional brickwork in order to facilitate assembly: the adhesive should provide sufficient strength and at the same time absorb, within its thickness, the intolerances in size of the bricks and of the entire construction and allow for a fast and simple assembly. Subsequently, the prioritized performance criteria for the adhesive selection for the Qaammat pavilion, are fundamentally different to the ones followed by the previous realized examples of adhesively-bonded glass brick envelopes, namely the Crystal Houses façade [4] and the Atocha Memorial [10],and are more similar to the bonding solution followed at the Qwalala Sculpture (see Table 1), although in our case all joints should be sealed afterwards to prevent water/frost and dirt from entering.
In terms of structural performance, due to the lack of a structural analysis model, several assumptions had to be made for selecting a suitable adhesive. Owing to the high degree of perforation of the structure that reduces wind pressure due to lateral wind gusts, tensile resistance properties were not considered crucial. A shear strength ≥ 1 MPa was established as desirable at a wide temperature range, based on the previously realized examples. However, due to the lesser overlap of the blocks (resulting to a reduced bonding cross-section) and the bending stresses occurring due to the inclined cantilevering of the two walls at the lower part of the glass structure (Fig. 6), an adhesive of a higher strength would be more favourable at this zone. A high creep resistance is not considered critical for this structure: considering the total dimensions and weight of each wall (circa 2 tn) and assuming an even load distribution, the expected pre-compression due to the own weight of the structure at a brick in the first row of the pavilion with 20% of its total surface bonded is < 0.22 MPa. The chosen adhesive should present stable properties at a wide temperature range, particularly against ambient temperatures as low as -35 ˚C, recorded in this location.Footnote 3
Ease-of-assembly was equally critical: the thickness of the adhesive should be able to accommodate the manufacturing tolerances of the glass bricks (± 1.5 mm) and further size discrepancies which may occur during construction, thus a 3 mm gap-fill capacity was considered essentialFootnote 4; this was also desired to compensate for movements due to thermal differentials of the glass bricks. Moreover, the selected adhesive should allow for a fast fixing and curing time, which were set at < 30 min and < 24 h respectively. A quick fixing time was important for preventing the overflow of the adhesive and accidental movement (sliding) of the blocks but also for enabling a relatively quick construction, essential due to the short Greenlandic summer: the pavilion should be built within a few weeks, thus, the adhesive should set quickly enough to allow for the built-up of several rows (3–4) in one day. Lastly, due to the lack of electricity and of other common commodities in the specific location, it was essential that the construction could be realized without the need of strictly controlled environmental conditions (i.e. regulated humidity and temperature levels). In terms of visual performance, although a fully-transparent adhesive was the most desirable, adhesives of a white or light grey colour were also acceptable as a solution by the architect.
In specific, the following key factors, in terms of structural performance, visual result and ease of assembly, have been established for the adhesive selection:
-
1.
Structural performance:
-
Shear strength ≥ 1 MPa and adequate tensile strength – a higher strength is more favourable at the lower levels of the construction
-
Stable mechanical properties over a wide temperature range, as low as -35o C
-
ability to equalize stresses (prevention of stress concentrations, e.g. due to locally insufficient contact with the glass substrate or due to voids within the adhesive layer)
-
ability to accommodate movements due to thermal expansion to prevent thermal breakage
-
2.
Visual performance:
-
Transparent, translucent or white/light grey in colour, in order to maintain a high level of transparency
-
Very good resistance to UV-radiation
-
can be homogeneously spread (prevention of overflow and of bubbles, gaps and dendritic patterns)
-
3.
Ease of assembly:
-
fast fixing (< 30 min) & curing time (< 48 h)
-
> 3 mm gap filling ability
-
no emissions of noxious or poisonous chemicals
-
no need for strictly regulated environmental conditions during construction
Choice of most suitable adhesive family
The arctic climatic conditions of Greenland pointed out towards two-component flexible adhesives, from the polyurethane and silicone-modified families as the most suitable adhesive family. These families are known for their excellent stability of mechanical properties over a broad temperature range (see Fig. 7). Moreover, such flexible adhesives typically present tensile and shear strength > 1 MPa and a bond thickness sufficient for accommodating dimensional tolerances and for equalizing stresses [12]. The strength of the Si–O bond provides silicones with high UV-resistance and allows them to be extruded even in temperatures < 0˚C [13]. Equally importantly, the chemical hardening process of such adhesives is less influenced by external climate conditions, allowing for construction conditions that do not require strictly regulated levels of temperature and humidity, a necessity in this case. In particular, silicone sealants have been previously successfully used in bonding and sealing applications in arctic climatic conditions, such as in the Princess Elizabeth Research Station in Antarctica [14], but also for the bonding of a similar cast glass structure, i.e. the Qwalala sculpture in Italy [5].
In specific, one-component moisture/heat activated adhesives were quickly discarded as an option due to their physical hardening process. This type of cure chemistry requires a favourable water vapour pressure in the atmosphere, which is a function of both temperature and humidity; which in our case could not be fully regulated. Moreover, the curing of such adhesives takes place from outside to inside at a relatively slow rate (of a few mm per day), rendering them unsuitable for wide joints: as the adhesive solidifies and thus, shrinks on its surface, tensile forces develop that can be sufficient to tear the still soft, uncured adhesive at the interior of the bond.
Epoxies and acrylates, despite presenting the highest strength among the adhesive families typically used in structural glass applications (see Fig. 7), including in the construction of both the Crystal Houses façade [15] and the Atocha Memorial [10], were in principle considered unsuitable for this case-study, due to their reduced application thickness/gap-filling property (typically between 0.1—0.5 mm) that does not allow them to accommodate construction tolerances[12].Footnote 5 Moreover, their application calls for thoroughly controlled environmental conditions during construction, which could not be secured at the pavilion’s location.
Furthermore, the uneven surface of the cast glass bricks hindered the application of double-sided transparent tapes, previously used in float glass layered sculptures (see Table 1), as they are in principle unable to accommodate the dimensional tolerances of the cast blocks and contraction and expansion movements expected due to the extreme climatic conditions.
Cement-based mortars used for hollow glass bricks, although initially considered, were soon eliminated as an option as well, as, besides not meeting the visual requirements due to their darker colouring, they further require a rougher surface to achieve a good bond than the smooth surface of solid glass blocks. Indeed previous shear experiments at TU Delft have pointed out that even with the application of a primer, such mortars still do not tend to properly bond to the glossy surface of solid glass blocks and can easily lead to adhesive failure at low strength values [16]. Experimental work on solid glass blocks bonded with a selection of mortars by [17] further confirmed the adhesion collapse mode and indicated a shear strength considerably less than 1 MPa on glass blocks with a smooth finishing surface.Footnote 6 Tile adhesives were also discarded as a choice, due to the fact that they are engineered primarily for indoor applications and are in principle not suitable for the low temperatures of Greenland.
Adhesive preselection
Based on all the above, on market availability and upon consultation with the Institute of Building Construction of TU Dresden, Dow Silicones Belgium and Siko B.V, a selection of suitable transparent and white two-component adhesives in the polyurethane and silicone-modified families were selected for further exploration. A two-component adhesive from the acrylate family was also selected as it fulfilled the established criteria and presents high bonding strength.
In specific, the following four adhesives were selected for further investigation: (A) 3 M™ Scotch-Weld™ Polyurethane Adhesive DP610, (B) Teroson MS 9399, (C) Experimental Fast Adhesive by Dow Inc., (D) Siko Clearbond. More specifically, adhesives A, B and D are available in the market, while adhesive C was specially formulated by DOW Silicones Belgium for the purposes of this project, as none of the commercially available bonding solutions from Dow’s High Performance Building Solutions range checked all of the project’s requirements: This adhesive has been formulated by DOW using a 4:1 Vol. mixing ratio with the aim to reach a lap shear strength of ~ 1 MPa in 1 h versus the standard mixing ratio of 100:14 weight that requires 24 h to reach the same strength. The snap time is reduced to 4—6 min and the time to handle strength to approx. 24 h. Moreover, DOW has removed the colouring pigment of the reacting component in order to achieve a final white colour instead of dark grey [13].Footnote 7 The properties of the selected adhesives, as provided by the manufacturers, can be found in Table 2 below.
Table 2 Characteristic properties of the selected adhesives as provided by the manufacturers