XPS Peel Panel Tests

XPS Tests

(See 1 January 2018 update at bottom of this post)

I have been meaning to do some XPS testing for a while now.  A cold autumn day seemed like a good time to do it.

BSD's thread about sealing XPS for PE motivated me to start the tests today.

I only had time for the first half of the first test.  My objective has been to improve epoxy bonding with XPS.  A thin, low viscosity coating is what I was thinking about.

Simple tools and common materials.  Things you might find in the shed.

I cut a small XPS test panel 8" x 24" from a low density sheet of pink residential insulation (Owens-Corning); marked off four 8" x 6" quadrants and sanded with four different grits (150, 100, 60 & 40).  

I added the first coat of Secret Sauce a couple hours ago.  The sealer disappeared nicely into all the sanding grooves.  I will add another coat of sauce later tonight.  Looks good so far...Epoxy test glassing will have to wait until next weekend.



A strong shell with higher densitiy foam should be better for XPS rather than using low density foam.

First I want to do the peel test to see how much "low density foam" pulls loose with the glass.

Hopefully low viscosity sealant will penetrate the surface layer of small ruptured cells, and the higher surface area of finer grit sanding grooves, better than the higher viscosity epoxy resin does.

The first coat penetrated all roughed surfaces nicely.

I will add another coat then sand with 150 grit and do the test lamination with epoxy. 

I will need to get some of the new Dow blue mix -- less than 10% styrene polymer with 60-90% propenenitrile ethenyl benzene.  The O-C pink is about 80-95% styrene polymer.

After that, side-by-side comparison with sealed-sanded and unsealed-sanded XPS.

I applied the second coat of Secret Sauce last night.

Observations:
- The first coat dried nicely after a few hours.  All surfaces looked good.

- The foam applicator worked well for adding a second coat to the 100 & 150 grit surfaces.

- Adding a second coat with the foam applicator to the 40 & 60 grit surfaces felt like I was sealing coarse sandpaper.  I was worried I was going to rip out small chunks of applicator foam.  I did not notice any small pieces in the finish.

I may need to look at applying the sauce with a 0.25" nap roller for coarser surfaces.  I decided not to for this first test because it looked like the roller might lay down too much sauce.

I have seen several methods mentioned for improving epoxy-XPS bonding, including razor cuts.  This guy claims he tried several methods.  He says he got best bonding by "scoring" foam with a steel-toothed pet brush -- using 30 psi min. comp. strength O-C XPS:

http://ecomodder.com/forum/showthread.php/improve-bond-fiberglass-epoxy-home-depot-foam-extruded-24820.html

As with histology (internal surface area of the small intestine) and oxygen transfer in water (bubble size), surface area to volume ratio increases as object size decreases and number increases.  Hopefully, this will hold true for some given groove size (peaks and valleys) from sanding.

The trick is finding a bonding agent that will fully penetrate these grooves.  Maybe exterior, water-based polycrylic concrete/tile sealer will do that, creating a thin intermediate layer between epoxy and foam sanding grooves.

I just re-read this and realized what I said.  The "groove penetration" argument is bovine manure.  Barring differences in inter-atomic/inter-molecular forces, there should be no difference in epoxy penetration for grooves of the same size (polystyrene vs. polycrylic).

The big difference is porosity, "closed" cell vs. open (XPS vs. EPS), and resulting epoxy penetration.  Another possibility would be a difference in the strength of bond type (hydrogen, dipole, ionic, covalent, etc.) between XPS and polycrylic vs. epoxy.

More later...

Surface roughness (epoxy surface penetration) is an issue with XPS.  XPS is "closed cell" foam.  It absorbs very little water (virtually no water).  The gas bubbles in the foam are very small and closed (no interconnecting air passages like EPS has).  The only way gas can be released is if the cells are ruptured.  Even after sanding, the surface is not porous, just a thin surface layer of very small ruptured cells. As a result, the epoxy has very little surface area to bond with and virtually no foam penetration (no soaking in like in EPS).

First, delamination of XPS is related to foam density and epoxy penetration.

Lower density foam crushes (compression) and tears (shearing).I do not want to write a treatise on this.  

I will try to keep it brief.  Epoxy bonding is related to the material surface energy (adhesion/wetting) and epoxy penetration into the surface profile (roughness/porosity).  Surface roughness increase bonding surface area and more mechanical interlocking with the surface.

Epoxy has a surface energy around 45 dynes/cm.  Epoxy will bond better with materials have surface energy values greater than or equal to 45 dyne/cm.

The greater the viscosity of the coating the lower the surface penetration.

I am using the lower viscosity polycrylic sealer because it can penetrate deeper into the XPS surface -- into smaller grooves. pores and cells.  The idea is to get a layer that has deeper penetration and more anchoring than the higher viscosity epoxy would have.   Then bond the epoxy to this thin, anchored intermediate layer.  I am assuming the sealer has better strength than the XPS.This explanation is a bit generalized."Stoneburner,A surface energy of 45 dynes/cm?  What does this refer too? Mike"It is an abstract concept that I have a very limited grasp of.  It is related to the energy needed to create and maintain a surface -- like the formation of a water drop.  All materials with surfaces have it.  For fluids, it indicates how well a fluid will spread/wet-out on solid surfaces. Here is a definition from Wikipedia:  

1. Surface energy, or interface energy, quantifies the disruption of intermolecular bonds that occur when a surface is created.  Surface energy is conventionally defined as the work that is required to build a unit area of a particular surface.

I was not really aware of it until I started learning about how epoxy bonds with surfaces.  All compounds with surfaces have their own unique surface energy value (polyethylene = 32; polystyrene = 34; 6,6 nylon = 42; glass = 250-500; etc.).

I have some links about epoxy surface energy and its bonding that I could give you.  

The take away is that epoxy adheres better with solid surfaces that have surface energy values equal to or greater than 45 dyne/cm.

I read somewhere that others have used low viscosity epoxy for the first coat to get better surface penetration and then use higher viscosity epoxy for the second coat.  I do not think they were board builders.It seems like the keying done (below) with that steel-tooth pet brush could add a lot of epoxy weight to the build.

http://ecomodder.com/forum/showthread.php/improve-bond-fiberglass-epoxy-home-depot-foam-extruded-24820.html



Well I finished laminating the test panel just before noon.  Used an old 30" x 48" piece of 6-oz e-glass thas been lying around in a drawer for years. Laminated with some old RR Kwik Kick I need to use up.  I still have enough of both for several more tests.I am skeptical about the sealer.  

I am not sure whether two coats of sealer and sanding all sealed surfaces with 150 grit was the right move.  But I will know when I pull the glass tomorrow pm or Monday and see how much foam comes up in each quadrant.

I got a carpet seam roller and have a Woodpecker covering/perforation (roller) tool on the way.  The Woodpecker tool is an interesting tool I stumbled across while searching the internet.  It may have potential.  I have many more combinations of variables to look at now.



I discoverd the propenenitrile ethenyl benzene in the new DOW blue mix is also called styrene-acrylonitrile (SAN).  The surface energy value I finally found for SAN is 40 dyne/cm.  Much closer to epoxy's 45 than styrene's 33.  Depending on the amount of SAN in the mix, the new DOW blue should bond better with epoxy than polystyrene does.

I ran a concurrent test with my updated "low budget" resin warmer this morning too -- worked great.  I will post pictures and a description in this thread also.I got my Woodpecker tool in the mail today.  I played with it on a piece of scrap XPS.  I rate it as having very high potential for improving epoxy bonding with XPS.  Paid $10.00 + $3.50 shipping on Etsy, a bargain -- like new.  Stoked.

I peeled the glass from the XPS last night.  I needed good natural lighting for decent pictures.  Hopefully the right lighting today...

Portable, adjustable glassing stand (aka ironing board)



All quadrants simultaneously laminated with the same batch of resin



Peels from XPS surfaces sanded with 40-150 grit:150-grit surface



100-grit surface



60-grit surface



40-grit surface



Another view of XPS surfaces sanded with 40-150 grit:150-grit surface



100-grit surface



60-grit surface



40-grit surface

_____

@ Lemat -- I do not have the time or equipment/money for 4-point bending tests.  I believe that low density/low min. compressive strength XPS foam tests should give me some usable information about adhesion between epoxy and XPS.

@ Dr. Strange -- I peeled several PU/PE longboards in the past.  I was surprised how little foam was pulled up.  I still had plenty of good foam to work with.  Funny you should mention EPS, I have been planning to play with Secret Sauce on it too -- different reasons.

@ Barry Snyder -- No surprises.  The greater irregular bonding surface created with coarse sandpaper caused more foam to pull up than with higher grit surfaces -- better anchoring.  Yeah, I really want to try the perforation rollers next.

@ John Mellor -- Agreed.  Sealed vs. unsealed was/is the central objective of the tests. The initial hypothesis was/is sealer will improve epoxy adhesion with XPS.  The null hypothesis is sealer will not improve epoxy adhesion with XPS.

For the first test, I was more interested in seeing whether the polycrylic sealer would bond with XPS and whether epoxy would bond well enough with the sealer.  The results suggest epoxy adhesion with sealer is stronger than sealer adhesion with XPS.  Side-by-side sealed vs. unsealed is definitely needed.  I decided that looking at different grit effects in this preliminary bonding trial could also get me some surface roughness information at the same time.
_____

Observations

*  The 1.5 pcf/15 psi XPS insulating foam clearly tears and crushes very easily.  I can see how adding a wood veneer (Lavarat) or cork skin (melikefish) to the foam surface would minimize surface damage by dissipating/dispersing point impacts over a broader area.

*  Notice the greater amount of foam on the (numbered) edge opposite the side used to peel up the glass patch.  The peel edge was taped to avoid this effect.  The opposing/numbered edge without tape -- with resin drip over the foam edge -- appeared to have much greater mechanical bonding/leverage (bottle-opener effect?).

*  Uniformity -- while coarser grit appears to improve adhesion, it lacks uniform distribution over the entire surface.  The deeper grit cuts pulled up the most foam.  But the deeper cuts are random with gaps between them (areas of lower bonding?).

I believe perforation rollers could create more uniform mechanical bonding over the entire foam surface to be glassed.

I think I will look at the effects of perforation next.  Sealed vs. unsealed?  What final sanding grit before perforation?  Too coarse, and the deep sanding grooves will likely impact the effects of perforation -- especially uniform distribution effects.

Adding perforation rollers for testing too has significantly increased the number of potential variable combinations...
_____

As mentioned earlier, we still need unsealed vs. sealed comparisons.  There is a wide range of variables to play with.  It would be better for you to pick the combinations that interest you.  I welcome any test data you are willing to collect and bring to the table.

Here are a few potential variables to combine for testing:

* Dow (blue) vs. Owens-Corning (pink) foams: SAN + styrene vs. styrene, respectively = 2

* Foam min. compressive strengths (psi):  15, 25, 40, 60, 100 = 5

* Final sanding grits:  40, 60, 100, 150 (more?) = 4+

* Sealed foam vs. unsealed foam = 2

* Sanded sealer vs. un-sanded sealer (different grits too?) = 2

* Sanded foam vs. unsanded foam = 2

* Perforating rollers:  Woodpecker vs. Carpet Seam Sealer = 2

* XPS vs. EPS = 2

I am sure there are several more that I have omitted in the moment.  
Please use combinations that interest you.

For now, I am only testing four (4) treatments at a time with my quadrant panels.  Carefully selecting the right variable combinations can eliminate the need for testing all variable combinations.  For example, I want to look at perforation roller effects next:

* Perforating rollers:  Woodpecker vs. Carpet Seam Sealer = 2

* Sealed vs. unsealed  foam = 2

* Sanded sealer vs. un-sanded sealer = 2

* perforation before sealing vs. peforation after sealing = 2

* potential sanding grits for final foam surface prep: 150, 100, 60, 40 = 4

I can think of a few more variables for "roller testing" alone but am ignoring them for the time being.

Please correct me if my ancient recollection about this calculation is wrong.  If I remember my math correctly, the number of possible variable combinations/treatments for roller testing alone would be:

2 x 2 x 2 x 2 x 4 = 64

I only have 4 test patches in my next panel (limited by scraps) at the moment -- a big panel would be cumbersome to laminate on an ironing board.  So these are the first variables I will choose:

* I have plenty of low-density, pink O-C XPS to work with = 1 variable (all treatments)

* I will use 150 grit for final surface prep = 1 variable (all treatments)

* Sealed foam vs. unsealed foam = 2 variables

* Perforating rollers: Woodpecker vs. carpet seam roller = 2 variables

I will use the 150 grit to minimize sanding groove interference with perforation effects.I think I can get a good idea overall about sealed vs. unsealed effects on this one and will have eliminated one surface grit to test.

I still want to the test the effects of peforations made before and after sealing -- 2 more variables.  Next round...
_____ 

I was planning to color some polycrylic sealer with RIT fabric dye for use on EPS this weekend.  My original thought was using the concrete sealer on EPS as a lightweight low viscosity sealer that would penetrate deeper into EPS than epoxy without adding a lot of weight.  It would also prevent too much epoxy from soaking into the foam.  But the main idea was to make the EPS more water tight in the event of dings -- especially lower density EPS foam.

Next, I thought maybe adding dye to the sealer might be an interesting technique for coloring EPS foam, creating a unique mottling pattern with the styrene beads.  Last weekend it suddenly ocurred to me that I could determine the degree of sealer penetration into EPS foam using the dyed sealer.  Degrees of penetration could be controlled by application method (foam pad, nap roller, brush, etc.) and/or timed surface flooding of the EPS.

Ironically, it never occurred to me to use the dyed sealer on sanded/roughed XPS surfaces to observe penetration.  I have a bottle of teal and a bottle of blue RIT dye.  I can get one more round of testing from the first XPS panel now.  More basic experiments.  Love it...
_____

I suspect many of the XPS failures can be related to the use of low density/low compressive strength foam (e.g. 1.3 pcf or 20.8 kg/m3).I have posted the specifications for commonly available XPS foam, Dow Blue and Owens-Corning Pink, at this link:

http://bgboard.blogspot.com/2013/12/polystyrene-foam-types-and.html
_____

For statistically valid comparisons, replication is needed to get a large enough sample size.  My number does not include replication for statistical analyses.

My testing here is about feasibility and at this time is not replicated.  Feasibility trials can be a good way to cull and select variables/treatments for further testing.

I am only speculating about the dried weight of a thin, single coat of concrete sealer.  I am assuming a thin coat concrete sealer becomes lighter after application because drying (evaporation?) seems to be involved.   At some point, I plan to weigh a piece of foam before and after sealing with polycrylic, but before lamination.  
_____

XPS Dye Test

XPS panel with 60-150 grit sanded then dyed polycrylic sealer surfaces + sanded, Woodpecker Roller peforated, then dyed polycrylic sealer surfaces -- 4 oz of polycrylic sealer was dyed with 15 ml of dark green, liquid Rit Dye.

Dyed 150 & 100 grit surfaces

Dyed 100 & 60 grit surfaces

Dyed 60 & 400 grit surfaces

Dyed XPS X-Sections

The natural lighting late Christmas afternoon was good enough for some cross(X)-section photos of the dye panel.  Dye penetration on the sanded surfaces is barely detectable to the un-aided eye on the 40-grit surface and "maybe" the 60-grit surface.  The much deeper Woodpecker perforations stand out noticeably in contrast.  The Woodpecker tines are approximately the thickness, maybe slightly thicker, of a single-edge razor blade and slice nicely into the foam surface.

Dyed 40-grit surfaces, most of the deep Woodpecker tine perforations noticeable on bottom surface
(Some dyed grit grooves barely visible  on top surface.)

Dyed 60-grit surfaces, deep Woodpecker tine perforations noticeable on bottom surface
(Some grit grooves maybe visible on top surface in second picture?)

Dyed 150-grit surfaces, deep most Woodpecker tine perforations noticeable on top surface
(No grit grooves visible on surfaces.)

XPS Perforation Tests

Problems that popped up were related to uniformity and depth of perforations (first photo below),  It was very difficicult to maintain even pressure across the whole carpet seam roller, getting half deep holes and half shallow holes.  I had to over-perforate with the carpet seam roller: I feel like the holes were too deep which caused too much foam "crushing."  There is an obvious learning curve for the carpet seam roller -- maybe Lavarat can offer some guidance.

I laminated the perforation test panel yesterday afternoon,  I will peel the glass late tomorrow afternoon.

The Woodpecker perforator made smaller, nice clean deep holes and was much easier to control.  Of course, I got a lot of practice with it perforating both sides of the first test panel before dying.

The experiments will continue...

XPS test panel sanded first 100 grit then perforated with Woodpecker (small holes) and Carpet Seam (larger holes) rollers -- left 2 quadrants sealed with polycrilic concrete sealer after perforation; right 2 quadrants are unsealed.

Perforated, sealed

Perforated, unsealed

_____

Here are some quick shots from my son's camera phone to ponder.

The close-up, money shots will have to wait for some good lighting -- hopefully tomorrow...

Foam side of the peeled glass patches:  left 2 patches, carpet seam roller; right 2 patches, Woodpecker roller.

XPS test panel surface after the laminated glass was peeled:  left half was sealled, right half was unsealed.

_____

XPS Woodpecker Roller (WR) and Carpet Sealer Roller (CR) Peel Test Photos:  Sealed vs. Unsealed

(Discussion & observations later...)

The money shots:

Perforated XPS, all quadrants laminated with the same batch of epoxy

Left half of panel:  sealed CR and WR quadrants, respectively

Right half of panel:  unsealed CR and WR quadrants, respectively

All quadrants peeled:  left to right -- sealed CR & WR and unsealed CR & WR (all foam side up)

XPS panel:  peforated surface after peeling

Sealed CR (left) & unsealed CR (right)

Unsealed CR (left) & unsealed WR (right)

Unsealed WR (left) & sealed WR (right)

Sealed WR

Unsealed WR

Sealed CR front, unsealed CR back

Unsealed CR front, sealed CR back

01 January 2018

It has been a few years.  In retrospect, I abandoned the sealer -- filled the roughed peaks and valleys without adding any apparent improved bonding with epoxy.

I liked the Topflite Woodpecker tool best.   Perforation is more about increasing surface area for bonding – many smaller holes/cuts would bond better than fewer big holes. The big holes add more resin weight. In my testing, the stock Woodpecker tool (WPT) seemed as effective as, or more so than, the carpet roller and appeared to use less resin (lighter?). I modified the Woodpecker tool, adding more tynes -- photos below -- to increase perforation numbers and total surface area for resin penetration/bonding. I had planned to use the modified WPT for a late summer-autumn build this year. But several issues postponed that project.  

Modified and Standard WPT.
(Standard WPT @ bottom of first two photos)