Wednesday, August 20, 2014

OBG Raises Questions and Provides Answers

Made Possible by Old Brown Glue

I developed and used Old Brown Glue nearly 10 years before I began selling it to the "public." That was almost 15 years ago, and I was naturally concerned about how others would be able to use this glue and what questions they may raise.  Since that time this glue has been used for making American Indian rocket ships, artificial kidney stones, as a bonding agent for exterior sprayed finishes, applying veneer to 10' tall architectural columns in Florida, Alaskan native drums, "craquelure" finishes, and too many other odd projects for me to remember.

Of course, my intention in making and selling OBG from the start was so that other woodworkers would be able to use it and appreciate its unique features, as I have.  So I was pleased to see the increasing number of furniture makers and luthiers who are using this glue on a regular basis.  And, since my name and phone number is on the label, I am constantly hearing from these people and answering their questions. I am happy to be of assistance, and always ready to respond to emails or phone calls, since I feel somewhat responsible for the success of their projects.

In order to fully understand how glue works, it is instructive to discuss not only the glue but the properties of the wood itself.  How the wood is prepared, what species it is, how it is joined and other factors contribute to the success or failure of the work.  That is why studies of adhesion spend a lot of time discussing the material and how it is prepared.  See the article in Fine Woodworking, "How Strong is Your Glue?"

The American Society for Testing and Materials (ASTM) defines an adhesive as a substance capable of holding materials together by surface attachment.  There are a variety of forces at work which make this possible.  The most common is called "mechanical bonding" meaning that the surfaces are held together by an adhesive that has penetrated the porous surface while it is liquid, then anchored itself during solidification.  With some adhesives, there are also other physical forces of attraction, which are referred to as "specific adhesion".  These include intermolecular attraction forces which form bonds between the adhesive polymer and the molecular structure of the wood itself, such as van der Waal's forces and hydrogen bonding.  I believe that covalent bonding is also a factor, which is 11 times stronger than hydrogen bonding, but at this time there is no clear evidence that such bonds constitute an important mechanism in adhesive bonding to wood.

The reason I suspect that these molecular forces are a factor is that water has a strong molecular attraction to wood, primarily through hydrogen bonding with hydroxyl groups of the cellulose which is in the wood.  Since protein glues are carried by hot water, it is obvious to me that they would have excellent penetration into the wood surface and form specific adhesion with the molecules of the wood itself.

Animal glues are derived by the hydrolysis of the protein constituent collagen of animal hides and bones.  They are described as "hydrolyzed collagen', and are actually various amino acids which join in polypeptide linkages to form long chain polymers.  Studies have indicated that most glue molecules consist of single chains terminated at one end by an amino group and at the other end by a carboxyl group.  Cross linkage between protein molecules is possible through hydrogen, ionic and covalent bonds.

These glues are manufactured in a wide range of average molecular weights but are graded for commercial use by a test instrument called a "Bloom Gelometer." In general, woodworkers use glues rated at 192, 222, or 251 average gram strength.  The higher the number, the faster it sets and the more brittle the bond.  The lower the number, the slower it sets and the more flexible the bond.  I have used Milligan and Higgins hide glue, gram strength 192 for the past 45 years, both in the glue pot and it is the basis for the formulation of Old Brown Glue.

The addition of urea to protein glues acts as a gel depressant, simply lowering the gel point.  Franklin and Titebond both manufacture and sell a similar liquid protein glue, and they use other chemicals to achieve the same result, such as ammonium thiocyanate or dicyandiamide.  Because of the large number of hydrogen bonding sites on the protein molecule, an amazingly diverse number of additives can be used to modify animal glues, producing a wide range of results.  My goal was simply to lower the gel point, using the most basic organic chemical, and I chose urea.

The inspiration for this decision was my participation in a research group in France some 20 years ago.
The group was called ADEN, and was a joint collaboration between the Musee des Arts Deco, in Paris and the Ecole nationale Superieure des Technologies et Industries du Bois, in Nancy.  One of the research projects involved testing the use of protein glues, with and without modifiers, in bonding wood to wood and wood to metal.  Long term environmental stress testing was done to anticipate aging.  The results wer published in a paper in a paper by Aurelie Garcet in 1996, titled "Etude des Colles D'Origine Animale Utilisees Pour la Restauration de Marqueteries Anciennes."  The general results of that testing included adding thiourea to the glue but thiourea is a known carcinogen.  The only difference between urea and thiourea is that thiourea contains a sulfur molecule, so I decided to do my own testing with urea after I returned to my workshop.  It took me 37 different formulations to achieve the results I wanted.  Surprisingly, it did not take a lot of urea, but for obvious reasons, I cannot exactly state how I formulate the glue.

The American Institute for Conservation of Historic and Artistic Works (AIC) has a subgroup, the Wooden Artifacts Group (WAG) which I have been a member of in the past.  This group publishes papers in their WAG Postprints, and in 1990 published a paper by Susan Buck, from Winterthur.  The title was "A Study of the Properties of Commercial Liquid Hide Glue and Traditional Hot Hide Glue in Response to Changes in Relative Humidity and Temperature."  Of course, this was before OBG was developed, so she used "a new, unopened bottle of Franklin Liquid Hide Glue with an expiration date of January 1991".  I should mention that Franklin was the first company in America to sell liquid hide glue.

Her conclusion states:

Based purely on strength characteristics this testing indicates that liquid hide glue is the glue of choice for repairing a join which will undergo significant stress, such as the structural join of a chair in regular use.  But, more importantly, that decision must also take into consideration the environmental conditions.  Under normal conditions of 50% RH and room temperature liquid hide glue provides the strongest bond.  However, hot hide glue proved to be the more stable of the two glues under extreme conditions of high heat or high humidity, and thus would be the ore desirable choice if fluctuation environmental conditions are anticipated."  The "extreme" conditions in her testing occurred at 84 % RH and 150 degrees F.  This is not surprising, as protein glues are reversible and those are the exact conditions under which the glue converts from solid to gel to liquid.

In fact, reversibility is one of the most important features to me in deciding to use protein glues in my work.  All furniture needs repair.  If you cannot reverse the finish or bonding agent easily, the furniture cannot be repaired easily.  One of the reasons antique furniture has survived for centuries is that it can be easily repaired.  As soon as modern adhesives or modern catalyzed finishes or epoxy fillers are used, the antique is damaged severely.

The fact that protein glues are "reversible" always raises questions, since modern synthetic adhesives are not, in fact, reversible.  You need to get both humidity (water molecules) and heat to the surface of the glue to liquify it.  That means that, if the wood joint is well made and tight, and there is a protective finish applied to the wood, and perhaps a wax is applied to the finish, and the object is in an environment which is not the surface of Venus, then the glue will hold.  However, if you throw the chair or guitar into the jacuzzi and leave it overnight, it will come apart.  Seriously, luthiers need to repair their instruments, and I understand that they use a range of glues for different reasons.  Liquid hide glues as well as hot hide glues in different molecular weights are commonly used and work perfectly, and there are instruments which are centuries old to serve as a testament to this fact.

All modified liquid protein glues have a shelf life.  Unmodified dry hide glue has an infinite shelf life, but when the protein is in solution the presence of water facilitates a chemical breakdown of the glue.  The pH changes over time.  Normal protein glue is in the 6 to 6.6pH range.  OBG is formulated to start at 5.5pH when fresh and decay to 6.0 after 18 months in the bottle, regardless of storage conditions.  Keeping it in the refrigerator prolongs the shelf life, and it can be frozen and thawed as many times as you wish, further extending the useful shelf life.

What happens to the glue over time in the bottle is that the viscosity changes as the pH changes.  It starts out as a gel in the bottle, which requires heating to use, and eventually becomes quite thin and liquid past the due date.  It also develops a strong ammonia smell, as a result of the chemical breakdown of the urea, and this is an indicator that the glue is no longer good.  I always recommend you buy it and use it fresh, before the due date, and if you are not sure, do a simple overnight test.

You can also heat and cool it for normal use as often as you want and as many times as necessary.  We do this all day long to adjust the viscosity for our work.  Simply take hot water from the tap and place the bottle in the water.  In a few minutes it is ready for use.  One advantage is that you can use the water to clean your hands since the glue gets very sticky and we often use our fingers.  Of course it is not toxic at all.

There are questions about how the urea affects the glue strength over time and why it is softer than hide glue on the surface.  My research seems to prove that OBG cures over time by loss of moisture, which is a factor of the wood and environment.  Thus, when the glue is applied to a joint or other surface and pressed together, the wood absorbs the moisture, allowing the glue to cure fairly rapidly.  However, the glue which squeezes out and rests on the surface retains a high degree of water for a fairly long time.  It might take days or weeks for it to completely dry out, which makes it softer than hot hide glue, which always dries brittle and hard.  I think this is an advantage in my work, since I often work with period finishes and the hot hide glue will damage the finish when I try to remove it.  I find the liquid glue cleans up easily with cold water and a sponge, conserving the original finish.

Thus, inside the joint, after the water has dispersed and the glue has cured, there is no longer any reaction between the protein glue and the urea modifier possible.  The glue is stable over time, and I have tested many projects over my career which gives me the confidence to continue using and promoting this wonderful material.

Further reading can be found in my article, "Why Not Period Glue?", published in the Society of American Period Furniture Makers Journal, November 2001.

1 comment:

Aymeric said...

Hi Patrick, thank you for your interesting post. I was wondering if you tried your urea recipe for OBG with different type of hot glues typically found in Europe, bone and tendon glues to name a few. I guess it should also work, but your input would be appreciated.
Also, is there any documentation on the comparison between hide glue found in North America and other hot glues found in Europe? Since you have been at Ecole Boule, you have worked with both, and an input from you would help us understand the differences.