Home Articles Forums Gallery Marketplace Videos

WoodCentral Forums

Knowledge You Can Build On — Est. 1998

VPS-Level Performance. Fully Managed. No Hassle.

More on glue creep *LINK*

Posts

More on glue creep *LINK*

#1

Ellis Walentine

More on glue creep *LINK*

Ellis Walentine

We had an interesting discussion a few days ago here regarding the phenomenon of "glue creep," a term that is confusing to many woodworkers myself included. Bill Tindall directed me to a fellow at Franklin International, the makers of Titebond products, for a more expert explanation of creep. Well, it turns out that that particular person wasn't available, but I was referred to Hugh Evans in their technical services department, who definitely had his facts and figures together. Hugh may or may not stop by here to recap or refute or otherwise enlighten us some more on the properties of PVA adhesives. Here are a couple of the things we discussed on the phone:

First, creep is a technical mechanical term referring to a property of all solid materials. In essence, it has to do with a deformation that a material undergoes when stretched beyond its elastic limit. PVA glues are relatively moisture resistant, so raised glue edges don't have as much to do with the glue absorbing moisture as they do with the effects of heat or the expansion and contraction of the wood being joined. In the case of my lamination example that I brought up in our earlier discussion, Hugh suggested that the raised glue edges may have had to do with the wood being at a higher moisture content when first assembled and then allowed to shrink back to a lower equilibrium MC dimension, as Bill suggested, though heat may have had something to do with it, too. Creep is also a greater risk where the joint is under stress, as in musical instruments, where the tension of strings, for example, exerts a constant force on the glue lines.

Regarding another unwanted heat effect, Titebond I and II lose as much as half their strength when the cured glue line reaches 150�F, and Titebond III can lose as much as two-thirds of its strength.

Creep is not a serious factor in veneer applications, as the yield strength of the glue is more than adequate to overcome any tendency of the veneer to shrink or expand.

I've invited Hugh to stop by here at some point and add his voice to the conversation. If there's enough interest, we can plan to invite Hugh and/or his colleagues to join us in a chat some evening.

Ellis

BTW, Hugh did mention that hide glue is quite resistant to creep, although it is much more vulnerable to moisture.


Application Guidelines

Re: More on glue creep *LINK*

#2

Re: More on glue creep

Bill Houghton, Sebastopol, CA

"Titebond I and II lose as much as half their strength when the cured glue line reaches 150�F, and Titebond III can lose as much as two-thirds of its strength."

Yet another reason not to move to Death Valley.

Thanks for the information, Ellis.

Re: More on glue creep *LINK*

#4

OK, I confess, I don't know what Creep is

Barry Irby

Apparently I have no idea what creep is. I thought it had something to do with bentwood laminations. Somehow related to spring back. The glue joint would slide over itself, allowing the piece to straighten out.

Can somebody 'plain to me what it really is?

This is not quite as dumb as it sounds. In the Home Inspection business there is a phenomenon also know as creep. If you deform some solids (plaster and glass come to mind) slowly enough you can bend them fairly far before they crack. If you try to jack up a house to level it it will crack and burst all over the place in spite of the fact that you are returning it to its original condition. It might have taken twenty or even fifty years to settle and you are trying to undo it in a few days or weeks.

Re: More on glue creep *LINK*

#5

Re: OK, I confess, I don't know what Creep is

Marv

Barry,

Others can expound on this, but glue creep as I understand it, is, once the glue cures or dries, the glued parts can move, especially if under stress or pressure. If too much creep occurs, the glue joint fails.

Glue creep will also occur if the process used for gluing is not done according to instructions, such as, controlled temperatures, proper surface preparation, improper clamping, like too much or too little or uneven. Too much moisture in the wood., etc, etc.

I don't know, I've done a lot of gluing and have given little thought to any of this stuff.

Someone correct me if I'm wrong, but I define "glue creep" as a failed glue joint or a joint that is subject to failure.

Just us epoxy and forget about all that stuff.

Marv

Re: More on glue creep *LINK*

#6

Creep clarified

bill tindall

It may help to know how creep is tested. Most commonly the term is in context of a plastic property that will be used in some structural application where the plastic will be subject to mechanical forces during use. To measure creep a test strip of material is prepared and a weight hung from it for some specified amount of time. The time will be days, or weeks or months. Creep is a measure of how much the piece stretches inelestically, that is to say, how much stretch stays stretched when the weight is removed after some amount of time. Creep is permanent plastic deformation of a material, typically a polymer/plastic, in response to stress(force) over time. Creep permanently changes the shape or size of something over some period of time.

Titebond is not recommended for preparing laminated structural beams because such a beam under load would slowly bend and permanently sag over time because the glue "creeped"(flowed, deformed). Creep has nothing to do with joint failure or fracture or elastic bending. It is just a permanent shape change over time.

Re: More on glue creep *LINK*

#7

yes you do

bill tindall

You have it exactly. Referring to the bent lamination, the initial spring back is not creep, but any change in shape after weeks or months would be a result of creep.

Wood creeps, as you know. So in the case of a bent lamination the creep of the glue would tend to straighten the form over time as a result of the stress of the bent wood, but the stress over time would lessen because of the creep of the wood fibers. Creppy!

Re: More on glue creep *LINK*

#8

David Barnett

Re: More on glue creep

David Barnett

"Yet another reason not to move to Death Valley."

Or leave your guitar in the car trunk. Not that I'd glue a bridge with Titebond or LMI or Jowat or any other milky glue. Some do. Let's just say I'm still from MO on this one.

Re: More on glue creep *LINK*

#9

Re: Creep clarified

Marv

Bill,

If a particular glue creeps enough, when under stress or pressure and causes the item that was clued to be permanently reshaped, wouldn't the joint have failed?

If not, why do we use the term "glue creep"?

Marv

Re: More on glue creep *LINK*

#10

Re: yes you do

Barry Irby

OK. So its the plastic deformation of the material over time. You are right, its the same thing.

As a side note, a wood technologist I talked to said wood (as in floor joists) goes through and "episode" of deformation at abut 75 years. Might be ten or fifteen years on either side of that, but there is a point at which it sags again. I thought that was interesting. It waits Seventy five years and then does something.

What was the discussion of the glue forming little ridges at the joints? How is that related to Creep?

Re: More on glue creep *LINK*

#11

joint "failure"

Bill Tindall

As is commonly used, the description "glue joint failure", is a failure of the bond between the wood and the adhesive. Again, as is commonly used, a permanent sagging of a laminated beam do to glue creep would be called a "structural failure". If I was standing on a beam and told it was going to creep, vs I was told the joint was going to fail, the timing of my response to this information would be different, as well as the consequences. So, it pays to choose one's words carefully.

To get technical on the molecular level, creep doesn't break chemical bonds or break polymer chains. On the molecular level creep is a result of molecules moving in response to stress to result in a shape change.

Creep isn't typically thought of as a failure of a plastic (cured glue is a plastic), but rather as an inherent property of the plastic. All plastics creep more or less; all glue creeps. Rock creeps if given enough time and stress. That is how mountains form. If creep causes a failure it is typically considered that the application/structure failed as a result of flawed design or application. When a new plastic is invented one of the first properties to be measured is creep as a function of temperature, for these data will determine where/how the plastic can be used.

I was told that Titebond is not simply poly(vinyl acetate) plastic, PVA. There is some additional component that results in a reversable crosslinking of the cured glue. The crosslinking is permanent at room temperature but reversible with heat and moisture, allowing a joint to be disassembled with steam. Crosslinking makes a plastic more rigid, the molecules are not as free to move, and it creeps less. Other adhesive compositions are more rigid and creep less that Titebond. Epoxy is highly crosslinked and I suppose it only creeps on a geological time scale.

Re: More on glue creep *LINK*

#12

Re: joint "failure"

Marv

Thanks Bill for that thorough explanation.

Marv

Re: More on glue creep *LINK*

#13

glue ridges

Bill Tindall

They most commonly form when the wood shrinks and the glue doesn't. If the glue was creepy enough they wouldn't form, but then there would be other problems resulting from such a creepy glue.

I have been toying with the idea for some time to measure the creep of a wood joint made with PVA glue. The experiment is simple. Glue two wood strips together with a lap joint. Make a scribe line across the joint so one can tell if there is slippage between the strips. Nail one end of assembly to a floor joist and hang two 25 # bags of shot from the other end. Watch to see if the strips creep by watching to see of there becomes a separation of the scribe lines on adjacent wood strips resulting from one strip sliding past the other.

Re: More on glue creep *LINK*

#14

why do some attribute this observation to creep

Bill Tindall

I have wondered why these ridges have often been blamed on creep. Do people believe the glue creeps up out of the joint? Indeed this could happen if the glue was creepy enough and the glue line thick enough, and the glue line was under compression stress from the rest of the joint. But, a properly made joint is not under compression stress once the clamps are off, nor is it thick.

I have made thousands of butt joints and never seen a raised glue line at a joint. I would conclude, and be in agreement with the Franklin folks on this matter, that raised joints is a preparation or execution failure, not a glue property shortcoming.

Re: More on glue creep *LINK*

#15

David Barnett

Ah... that DOES help, Bill... *LINK*

David Barnett

I can see my definition of joint failure and glue creep extends to feeling even the slightest difference across a glue seam or joint, something that catapults me into vitriolic high dudgeon. I'd just about swap my soul to Pazuzu for seamless seams, so until that transaction is finalized, I stick to CA, resorcinols and hot hide.

While I've had successes with aliphatics, polyaliphatics, hard and soft (when I actually wanted flex), I've been too disappointed to risk anything that matters -- everything else, sure. Maybe I'm more obsessed by this because I work in mixed materials that resist easy joining and finishing.

Imagine... what would Tapio have done if a fingernail clicked when dragged across his Leaf Tray?

By the way, one's just become available, if you need a nut dish.

http://www.1stdibs.com/furniture_item_detail.php?id=437496


Tapio Wirkkala

Re: More on glue creep *LINK*

#16

David Barnett

But... but... how do you explain...

David Barnett

... non-wood, non-absorbant materials that exhibit this same phenomenon over time? I've seen glue ridges that formed on ceramics, stone, plastics and metals.

Re: More on glue creep *LINK*

#17

Ellis Walentine

Tapio's tray

Ellis Walentine

...is made of plywood. Phenolic glue. No creep.

Re: More on glue creep *LINK*

#18

David Barnett

Okay. I know. *LINK*

David Barnett

I've been doing tortuous things to BB plywood lately. I was making a joke, albeit lame, sort of along the lines of:


The Day the Ivory Sank at Proctor & Gamble

Re: More on glue creep *LINK*

#19

David Barnett

I suppose I'm the gadfly here...

David Barnett

But why does it not surprise me that a Franklin representative would define creep in such a way that it works to promote his product? And why, for that matter, do they apparently spend considerable effort, time and expense to dispel the perception that their glue products exhibit no such extraordinary tendencies compared to other resins held in high regard for creep resistance, if no such tendency appreciably exists?

In my experience, personal and anecdotal as it is, I have only encountered such problems when using this family of resins, whether from Borden, Franklin or any other manufacturer. Don't get me wrong. I keep fresh (another issue) bottles of the Titebonds for lots of uses, but I'm not about to commit a glued-up table-top panel to resins I no longer trust for that purpose.

Also, beyond creep, I don't worry about my resorcinol powders and hide glue granules expiring. Even West Systems epoxies lasts for years in their pump bottles. Beyond immediate no-mix convenience, I guess I'm not so hyped on these glues. By the way, when I was at One Cottage Street, I knew no one who used aliphatics. To a one they used urea-resorcinol-phenolic-formaldehydes in one form or another, as did I. I must say I breathe better now that I do not.

We each have our preferences, and I don't mean to belittle anyone's favorites for whatever their reasons. For example, I'm beginning to use more CA. I like it. It works for me. I absolutely hate the urethane glues, but have used them when it made sense. But I'm still dubious and wary when it comes to industry insiders, helpful and well-intentioned as they very well may be, telling me one thing, when my eyes, ears, and fingers tell more something altogether different.

Re: More on glue creep *LINK*

#20

a return question

bill tindall

Creep is a change in shape as a result of stress. So, if you put a strip of mushy plastic between two pieces of marble and applied clamps the mushy plastic would creep out of the joint and make a ridge. Where is the stress if two pieces of marble are glued together? I have not observed the phenomena so I have no idea what causes the ridges you observe. One would speculate it might be a result of differential expansion of materials at the joint.

You won't feel a properly prepared wood joint with properly dry wood. If you are getting raised wood to wood joints something is amiss.

Re: More on glue creep *LINK*

#21

Creep isn't defined by Franklin

bill tindall

I explained how creep is measured in a lab devoted to measuring polymer properties. I see nothing in the Franklin literature regarding definitions and explanations of creep that is contrary to usage in the polymer field. There is no stress on a well made joint in a table top. So what ever one observes can not be a result of creep. I don't observe it in my table tops, panels etc.

My daughter is a polymer chemist. I will quote from one of her texts on the subject (Billmeyer, Textbook of Polymer Science page 200 in the chapter on Structure and Properties of Bulk Polymers).

"Creep is studied by subjecting a sample...... to a constant stress (force) and observing the resulting time dependent strain (change in size) for relatively long periods of time. .....creep provides fundamental information of (polymer) performance in practical applications. "

Creep is relevant in bent lamination if the bend subjects the joint to sufficient stress. Creep will relax the bend. I have not seen any evidence of creep in anything I have bent but I have not made tight bends in think wood, either.

Re: More on glue creep *LINK*

#22

Re: I suppose I'm the gadfly here... *LINK*

Hugh Evans

I mentioned that I might join in the dialogue while talking to Ellis the other day, so here I am.

The definition of mechanical creep I provided does not apply solely to Franklin wood glues and is, in fact, what I learned while studying Materials Science as part of my Mechanical Engineering curriculum. In fact, Titebond Original, II, and III are sufficiently susceptible to creep that they are never recommended for structural/load bearing applications.

As a Technical Specialist, my job is to provide the most accurate information possible and am ethically bound to suggest the most appropriate product for a given application, even if that means recommending a competitor. For example, let's say you have a wobbly piece of furniture to repair and you don't want to attempt any woodworking (this happens from time to time), an epoxy is clearly the most appropriate choice and that's what I would advise... We don't have an epoxy in our line. My job is not sales or marketing: I'm a scientist.

Glue ridges can occur and tend to result from rather weak bonds with excessively thick gluelines. All PVAs are susceptible to this, and Titebond is no exception.

I can certainly appreciate your weariness of industry insiders and hope I'm a step closer to earning your trust in this regard.


Wikipedia Article Regarding Creep

Re: More on glue creep *LINK*

#23

David Barnett

I stand corrected

David Barnett

What I've been calling creep for two decades is apparently not the same phenomenon as defined by materials science. I have only once encountered a glue ridge, that is to say, a thin ridge of glue plastically deformed and raised between the two surfaces conjoined by that glue (well after squeeze-out at the time of assembly), on a laminated workbench top, and that was epoxy, not PVA. However, I have encountered boards in a previously flat, dead-smooth glued-up panel becoming proud of adjoining boards in that seam. I never thought it was the glue itself I was feeling but that the glue had insufficient hardness, allowing the board to flex in the joint. It's something much easier to feel by running one's hand across that surface than it is to explain as I'm trying to do right now.

What does one properly call this phenomenon, then, and why have I observed it on glued-up seams that use PVA or epoxies, and not urea-resorcinol-phenolic-formaldehyde resins, hot hide glue, or cyanoacrylates?

I am sitting here, looking at and running my fingers across two small tables with glued-up tops made from the same boards at the same time nearly twenty years ago. One has joints of rubbed hot hide glue, the other PVA. Clearly, the HHG top remains perfectly smooth to my touch while the other does not. I'm aware that many variables could enter the picture. I will grant that it could be differences in technique, the age of the PVA at time of application, ambient factors of temperature and RH, or something else. But the difference is there.

Now, I didn't glue these tables differently as an experiment. I glued the HHG panel as my first attempt at a rubbed joint. On the second table I used PVA because I didn't entirely trust the outcome of the first panel after scraping the gelled hide glue from the edge-joined boards. In a matter of one weekend, however, the difference was obvious. Everyone I showed it to, accomplished woodworkers all, ascribed it to PVA creep, and I've used these tables to demonstrate that phenomenon ever since. A chess table I once had access to was an even more severe example.

Much that I recall from those days has become oversimplified, compressed and codified into roughly-drawn cartoons of real events and persons. Discussions in shops, arguments with reps at trade shows and woodworking seminars are now jumbled in my memory. I do know that there was a strongly shared belief that while PVA glues were convenient and suitable for many woodworking purposes, they were not a viable choice for curved veneer panels, sculpted stacked lam chairs and cabinets in certain tropical woods, and marquetry. The studio furniture craftsmen I knew didn't use it where it would show or be felt, and two decades of my observing these effects have only reinforced this perception.

"So what ever one observes can not be a result of creep."

Fair enough. I shall endeavor to be more accurate. I will no longer refer to this tendency of PVAs* to more freely allow, almost visco-elastically when compared to other glues, wood movement in edge-jointed panels and other constructions as glue creep. I must say, however, that this tendency remains nonetheless demonstrably real and I should like to know what to call it.

*exploited in PVAs such as Jade No. 403 and No. 711 used multipurposely in book arts

Re: More on glue creep *LINK*

#24

David Barnett

Thank you, Hugh...

David Barnett

...for coming here and patiently explaining this. Your unbiased approach and tone is refreshing and I'm entirely pleased to learn what creep really is and to revise my nomenclature. Actually, both Ellis and Bill elucidated the distinction really well, but my eyes glazed over as soon as I heard that term. That was so knee-jerk. As much as I like to think I'm open to new learning, I know I'm a little slower on the uptake these days, with "rather weak bonds" and "excessively thick gluelines" at the synaptic level.

Re: More on glue creep *LINK*

#25

Can't use PVA glues for load bearing applications?

wilbur

"In fact, Titebond Original, II, and III are sufficiently susceptible to creep that they are never recommended for structural/load bearing applications."

I know I'm missing something here. Wouldn't gluing joints used in furniture making be a structural/load bearing application? Chairs, bookcases, and tables certainly bear loads.

Couple this with David's observation of panel glue ups with PVA glues going out of flat, and I'm not sure what woodworking application would be suitable for PVA glues.