Osmosis and the Blistering of Liquid
Applied Polyurethane Roof Membranes
Brian Hubbs, P.Eng
Graham Finch, MASc EIT
Rob Bombino, PE
RDH Building Engineering Ltd.
Vancouver, BC
BEST2 Portland April 2010
Background
Cold-applied asphalt-modified polyurethane
waterproofing membranes
Commonly used in inverted roofing and waterproofing
assemblies
Very common in the Pacific Northwest and Western Canada
Historical performance problems often attributed to poor
details and/or inadequate thickness.
- Concrete Pavers or Ballast
- Pedestals (Optional)
- Filter Fabric
- Extruded Polystyrene Insulation (2-4")
- Molded Polyethylene Drainage Matt (Optional)
- Cold-Applied Polyurethane Roof Membrane
- Concrete Slab (6-8")
Page 2
Background
Maintenance Manuals and Condition Assessments
Water Filled Blisters:
•
•
•
•
•
•
Low thickness
Filled with water under pressure
Blisters vary in size from small (12mm) to entire roof or deck areas
No obvious detail or discontinuity
Pavers and ballast floating in some locations
Blistering observed over both conditioned (interior) and
unconditioned space (Parkade)
• Blisters also observed on vertical planter walls and in water features
• Top of membrane almost always wet when insulation removed
Two Different membrane manufacturers
Building age varied between 3 to 12 years
Page 3
2004 - Review of Roofs For Maintenance Manuals
Moderate Blistering
Page 4
Typical Blister
Page 5
Summary of Cut Tests
2004
2008
10
12
11
12
10
Minimum Thickness - CCMC Listing (1.5mm)
Page 6
2008 - Large Blister Lifting Pavers
Page 7
2008 - Membrane Blistering – Waterbed Effect
Page 8
Membrane Blistering In Water Feature
Page 9
Membrane Pore Structure – Membrane #1, 30 mil aged
Membrane #1, 30 mil - AGED
Membrane #1, 120 mil - NEW
Page 10
Possible Causes?
Details?
No penetrations or adjacent details
Membranes fully adhered except at blisters
Vapour Diffusion or Capillary Flow?
large quantity and pressure of water cannot be explained by these
mechanisms
Interior Sources?
Vapour flow from membrane is always towards the interior
Pinholes and membrane?
Not present at blisters.
Water can not be contained under pressure
Hypothesis: Osmosis
Page 11
What is Osmosis ?
The flow of water across a semi-permeable membrane from low to
high salt concentration
Requires 2 things:
Difference in salt concentration
membrane permeable to water molecules
Page 12
Other Osmosis Research
Not well Documented by Building Industry
Either rare or unreported
Other Industries:
Fiberglass boat hulls
• Uncured resins create a chemical osmotic cell leading to destructive
fiberglass hull-blistering
Epoxy Floor Coatings
• Moisture from slabs on grade creates blisters beneath certain
membranes
Bridge decks
• De-icing salts cause blistering of coatings
Page 13
Could it be Osmosis ? - Is the Blister Water Salty ?
Blister water extracted and tested
Contains high concentrations of
dissolved metals:
Sodium: naturally occurring within cement and
aggregates
Potassium: Potash commonly used concrete additive
Silicon: naturally occurring within cement and
aggregates
Calculated osmotic suction pressures
300-400 kPa
Water extracted from membrane blisters
was under pressure
As blisters grow the membrane
delaminates
Rainwater from the top surface of the
membrane contained no relevant
concentration of minerals
Page 14
Vapour Permeance Testing
Aged 30-60 mil reinforced membranes removed from
buildings
4 to 7 US perms (230 to 400 ng/Pa s m2)
50% RH
New membranes 90-150 mil membranes
0.3 to 2.0 US perms
Importance of Inverted Wet Cup ASTM E96 Method
450
400
350
300
250
200
150
100
50
0
Wet
Water
IWC
Membrane #1 - 30 mils (Aged)
Membrane #2 - 60 mils (Aged)
Membrane #1
#3 - 150 mils (New)
0.60
Vapor Permeability
ng/Pa s m
ng/Pa s m2
Vapour Permeance
Dry
0.50
0.40
0.30
0.20
Impermeable Roof Membranes
0.10
Looks
like
(SBS, TPO, EPDM)
an 0.00
Dry Cup
Wet Cup Inverted Wet
Cup
6” Concrete
inverted
roof ?
Page 15
Impact of Membrane Permeance
Use WUFI to simulate initial saturation or wetting of
Concrete surface
Polyurethane – top 10 mm of concrete
Polyurethane – entire slab
SBS mod-bit – top 10mm of concrete
SBS mod-bit – entire slab
Page 16
Laboratory Apparatus - Trial 5
Page 17
Proof of Concept – Commercial Reverse Osmosis Membrane
Initial flow
rates of up to
15 L/m2/day
Per manuf.
specs
Page 18
Measured Osmotic Flow – Control Samples
Membrane #1 - Controlled Salt Solutions - Osmotic Flow through Membrane
2500
0.1 Molar Salt - 460 kPa Osmotic pressure
2000
1.0 Molar Salt - 55,000 kPa Osmotic Pressure
1500
1000
500
220
200
180
160
140
120
100
80
60
40
20
0
0
Osmotic Flow through Membrane - g/ m2
Distilled Water - Control - No Osmotic Pressure
Days from start of test
Page 19
Osmotic Flow – Blistered Membranes & Blister Water
Membrane Sample
#1 - 30 mil
#1
2000
Membrane Sample
#2 - 60 mil
#2
1500
1000
500
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
0
Osmotic Flow through Membrane - g/ m2
Polyurethane Membranes #1 & #2 - Average Osmotic Flow through Membrane
# of Days
Page 20
Osmotic Flow – Blistered Water & Control Samples
2000
Membrane 1 - Distilled Water - 0 kPa
Membrane 1 - 0.1 M NaCl - 460 kPa
Membrane 1 - 1.0 M NaCl - 55,000 kPa
1500
Membrane 1 - Blister Water - 326 kPa
1000
500
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
0
Osmotic Flow through Membrane - g/ m2
Osmotic Flow through Membrane #1 with Different Salt Water Concentrations
Days from start of test
Page 21
Concrete Primer & Membrane Testing – New Membrane (150 mils)
Effect of Membrane Primer Type - Polyurethane vs Epoxy
Osmotic Flow through membrane - g/ m2
400
Membrane Cured during first 50
days of testing. Permeance initially
2x5 times higher. Osmotic Rate
affected
350
300
0.4 to 1.0 g/ m2/ day
250
200
150
100
Epoxy Primer on membrane - 0.5 Perms
50
Polyurethane Primer on membrane - 0.9 Perms
0
0
20
40
60
80
100
120
140
160
180
200
Time (days)
Page 22
Osmotic Flow Rate – All Samples
Aged Samples
500
M1-1 - 30 mil blistered
Osmotic Flow - g/ m2
400
M1-2 - 30 mil blistered
New Samples
M2-1 - 60 mil blistered
300
M2-2 - 70 mil blistered
200
M3-1 - 120 mil new primed
M2-3 - 60 mil new
100
M3-2 - 100 mil new field applied
M4 - 100 mil new
0
0
20
40
60
80
100
120
140
160
180
200
# days from start of test
Page 23
Control
Membranes
0.0
New M2, with Alum
Paint
1.0
New M2, unprimed
2.0
Blistered
New M1, Epoxy
Primer
New
New M1, Urethane
Primer
3.0
6
New M1, Unprimed
4.0
M2, B3 (Blistered)
Permeance (US Perms)
M2, B2 (Blistered)
Is This Good Enough
to Eliminate Blistering ?
Osmotic Flow – (g/m2/day – High Low Average)
6.0
M1, B1 (Blistered)
Control
Membranes
New M2, with Alum
Paint
New M2, unprimed
Blistered
New M1, Epoxy
Primer
New M1, Urethane
Primer
5.0
New M1, Unprimed
M2, B3 (Blistered)
7.0
M2, B2 (Blistered)
M1, B1 (Blistered)
Permeance (US Perms – High low average)
Results – Vapour Permeance and Osmotic Flow
Osmotic Flow – (g/m2/day)
12
11
10
9
8
7
New
5
4
3
2
1
0
Page 24
Results – Vapour Permeance, Membrane vs Concrete
2.0
New 90mil membranes with primer
have much higher water vapour flow
rate through the membrane than
though the concrete. Is accumulation
and blistering still possible ?
1.6
1.4
1.2
Control membranes (SBS, EPDM,
PVC are all below the permeance
range of 6 inch concrete slab.
1.0
0.8
0.6
0.2
Reported Vapour Permeance of a 6
inch Reinforced Concrete Slab Varies
between: 0.05 and 0.2 US Perms.
0.0
Difficult to Measure.
6" Concrete Slab
Control Membranes
New M4, primer 2
New M4, primer 1
New M4 no primer
New M3 - Field Applied
New M2, with Alum Paint
New M2, unprimed
New M3, Epoxy Primer
New M3, Urethane Primer
0.4
New M3, Unprimed
Permeance (US Perms – High low average)
1.8
Page 25
0.0
6" Concrete Slab…
New M4, primer 2
New M4, primer 1
New M4 no primer
New M3 - Field Applied
New M2, with Alum Paint
New M2, unprimed
New M3, Epoxy Primer
New M3, Urethane Primer
New M3, Unprimed
Osmotic Flow – (g/m2/day – High Low Average)
Results – Osmotic Flow Through Membrane vs Concrete Vapour Flow
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.4
Estimated Vapour flow though a 6
inch Reinforced Concrete Slab in
g/m2/day.
0.6
0.2
Difficult to Measure.
Page 26
Summary - Osmotic Blistering Process
Water on top of membrane almost year round in inverted roof
Top surface of the membrane wet all year
assembly (poor slope = more water)
Moisture moves though the membrane
Concrete is initially at or close to saturation. Vapour diffusion
viaadditional
vapour moisture
diffusion
moves
though membrane to concrete
interface.
Concrete less permeable than the
Concrete
is less permeable
than membrane
and water begins to
membrane
= moisture
accumulation
saturate the concrete and accumulate at the membrane
Moisture dissolves minerals from
interface.
concrete
Mineral ions dissolve out of concrete increasing the salt
Osmosis forms small blisters at localized
concentration of the water beneath the membrane. Osmosis
voids
or debonded
areas
begins
and small
blisters are formed.
Osmosis continues expanding blisters
over time - 3-6 mm of water per year
Vapour diffusion to interior through concrete is relatively
slow compared to the rate transported by Osmosis.
Blisters grow and expand due to osmotic flow.
Page 27
Is Osmotic Blistering a Regional Issue ?
• RDH Observations
• West Coast of Canada
• Pacific Northwest
• Discussions at NBEC 2009
• Florida
• Hawaii
• Discussions at RCI San Diego 2009
• Florida
• Appears to be more prevalent in temperate, humid
climates
Page 28
Next Steps
Determine maximum safe permeance threshold for
inverted roofing and waterproofing membranes
Develop Osmotic flow test method and determine
acceptable maximum flow rates.
Revise Applicable Standards (ASTM C836-00 and
CAN/CGSB–37.58-M86) to specify:
Maximum allowable inverted wet cup permeance (0.1 ?)
Maximum allowable osmotic flow rate (1 g/m2/day ??)
Further testing of membranes currently on market
Need solutions or further research to confirm performance:
Page 29
Questions
Page 30