Case 1: TPO Roofing Membrane Installation

We investigated the failure of a thermoplastic olefin (TPO) roofing membrane as­sembly at a four-building complex. The buildings consist of precast concrete exterior walls, cast-in-place concrete interior walls, and precast concrete roof planks with a concrete topping slab. Approximately one year after construction, the building owner observed inflated and billowing roofing membrane at two of the buildings, and at one of these buildings there was an irregular roofing sur­face from displaced insulation below the membrane. During our field work to investigate the failure, we observed occasional roofing membrane inflation and billowing at all structures; this was caused by internal building pressure (positive pressure) and pressure created by wind blowing across the roofs (negative pres­sure) (Fig. 1). Emergency retention bars installed over the roof surfaces reduced, but did not eliminate, the extent of the membrane inflation and billowing.

The typical roofing assembly consisted of, from exterior to interior, TPO roofing membrane, reportedly adhered with a polymer-based contact adhesive; multiple layers of paper-faced polyisocyanurate insulation that varied from 3 to 12 in. (7.62 to 30.48 cm) in total thickness, reportedly adhered with a one-com­ponent foamed adhesive; a 2 to 3 in. (5.08 to 7.62 cm) thick cast-in-place con­crete topping slab; and approximately 6 in. (15.24 cm) thick precast concrete roof deck planks. Prior to any sample roof openings, we observed notable insu­lation board deformation reflected through the TPO membrane (Fig. 2). We made 29 sample openings and observed the following:

• At 29 of the 29 sample openings (100 %), as we peeled the membrane from the insulation board, some portion of the top insulation board facer experienced a cohesive failure and delaminated within itself (Fig. 3).

FIG. 1—TPO roofing membrane inflation and billowing.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

FIG. 2—Notable insulation board deformation reflected through the TPO membrane.

• At 24 of the 29 openings (83 %), we observed moisture on the insulation board facers; we verified dampness or moisture with moisture-sensitive paper (Fig. 4). At 23 of the 29 sample openings (79 %), we observed moisture between the top layer of insulation and the TPO roofing membrane.

• At 25 of the 29 openings (86 %), we observed staining on the insulation board facers (Fig. 5).

• At 28 of the 29 openings (97 %), we observed apparent microbial growth on the insulation board facers (Fig. 6).

• At seven of the 29 openings (24 %), the insulation boards were displaced below the TPO membrane prior to our making our exploratory opening (Fig. 7).

• At eight of the 29 openings (28 %), the insulation boards were bowed or curled. At some of these openings, the edges of the tapered insulation board were curled up and no longer adhered to the insulation board below (Fig. 8). At seven of the eight openings with curled or deformed insulation (88 %), the insulation curling or deformation occurred at areas that were wet or which showed signs of previous moisture (i. e., stains or apparent biological growth).

• At five of the 29 roof openings (17 %), the concrete topping slab sub­strate was cracked. At three of these five locations, cracks occurred at openings adjacent to the parapet, and at two of these three locations we verified that the crack was above a joint in the hollow core precast con­crete roof deck planks.

FIG. 4—Top insulation facer is wet, as indicated by moisture-sensitive paper that turns pink in the presence of moisture.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

We tested an insulation board sample in accordance with ASTM C1616 [4] and calculated a sample moisture content of 66.29 % (i. e., the sample was wet).

We measured the temperature and relative humidity of the concrete top­ping slab using a modified version of ASTM F2170 [5] (we had to limit the time between readings due to facility constraints), and after 24 h the average relative humidity of the topping slab was 75 % (relative humidity ranged between 58.6% and 91.1 %).