Reflections on Book Structures

In June of 1998 I gave a presentation on fan-glued binding and adhesive book structures to the Library Binding Discussion Group at the annual ALA conference. What follows is a partial writing of ideas that have existed only in outline and illustration since that time (a more complete writing will continue in part 3 of these "reflections"). My reflections herein are based on observations derived from my experience with fan-glued bindings. Most of my observations are derived from experience, a magnified examination where possible using a dissecting microscope, and by page pull tests using a tester of my own devising. Not being an adhesive engineer, the terminology herein is pretty much my own. Any feedback from my reading public is appreciated.

Thoughts on PVA

PVA is the core component in a family of glues all of which tend to be non-discriminately referred to as "PVA glue." Like the term "perfect binding" the unqualified use of the term PVA can easily mislead. The core ingredient of PVA, polyvinyl acetate, is simply the starting point for many different glue formulations. PVA in its purest state dries hard and brittle and is absolutely unsuitable for bookbinding purposes. PVA is blended with varying materials to add flexibility and other properties. "PVA glue" in binding circles has probably become shorthand for a flexible PVA. Flexibility, however, is a very flexible term. A sheet of paper is flexible as is a rubber band but one is elastic and the other isn't. Blended PVAs can exhibit a similar range of the characteristics from fairly hard, yet flexible like a sheet paper, to very soft and very elastic like a rubber band.

Understanding a PVA bond requires that we distinguish between its adhesive strength and its cohesive strength. Its adhesive strength is the force with which it attaches itself to a substrate, such as paper, board or cloth. In bookbinding for a glue to be successful it must adhere to both the paper substrate and any liners used to reinforce the spine. Ideally the strength of adhesion should be greater than any stress that is exerted on the joint between glue and substrate. If this is so then a glue is sufficiently adhesive. However, PVAs do more than adhere two substrates to each other. PVA forms a film between the substrates. This film has its own characteristics. If the film itself fails, the joint will fail. The internal strength of the film is its cohesive strength. The greater the cohesive strength the greater the force required to separate the film. Though I am not an adhesive scientist, my observations indicate an inverse relationship between cohesive strength of a PVA and its elasticity: the more elastic a PVA the lower its cohesive strength. The distinction between highly cohesive and highly elastic PVA can be described tactilely as PVAs that form a hard film (high cohesive strength) and PVAs that form a soft film (high elasticity).

These distinctions appear to be moot when gluing paper or cloth to board (but on close examination, they may not) but they may be quite significant when glue is used on the spine of a book or the joint area where the PVA assumes a structural role beyond simply that of sticking two materials together. I have found that most PVAs I have worked offer sufficient adhesion for binding purposes. It is their cohesive/elastic properties that tend to be most significant.

Reflections on the Glue Line

Pull and Peel

Books are comprised of glue, paper and cloth. All of these are flexible materials. Once glued together they are subject to various stresses when a book is used. Basically there two types of stress on a glue joint which I unscientifically distinguish as pull and peel. The various laminate structures formed throughout our bound book are subject to either one or a combination of both of these stresses when the book is use or even when the books sits upright on a shelf. Understanding which stress comes into play and how contributes to proper book design.

Pull is stress applied evenly across the entirety of the glue joint and parallel to the glue joint. Peel is stress applied to and concentrated on the edge of the glue line. (see illustration to the right). Peel represents the stress we would normally associate with removing a piece of pressure sensitive tape from a surface. We start at one end and we lift and pull back toward the other end exerting stress at the joint between the tape already lifted and the tape still attached. This approach to tape removal divides and conquers the tape's resistance to removal. If we pull the tape parallel to the glue line the stress distributes across the entire glued joint. Such would be the case if we apply a length of our pressure sensitive tape to a tabletop with a tab hanging off the edge. If we pulled this extended tab parallel to the desktop (as shown in the illustration) the tape will be almost impossible to remove. Chances are that the tape itself will break before the glue line fails.

A pull stress maximizes the effectiveness of a glue joint while peel stress minimizes it. Wherever possible we should design to a glue joint's pulling strength rather than its weaker peeling strength. This, however, is not always easy to do. Sometimes we simply have to deal with a peeling stress. The joint area between the board and the spine is frequently subject to a peeling stress as is the super to textblock. We will also see that the page to page joint in a fan-glued binding can be subject to either peel or pull stress.

The easy answer to strengthening these areas is that a stronger, more cohesive glue will give us the strongest joint. The greater the cohesive strength of the glue the greater the strength of the joint. However, the divide and conquer nature of a peel stress actually inverts this assumption. Assuming good adhesion, a highly cohesive glue tends to concentrate a peel stress whereas a less cohesive, more elastic glue, tends to distribute that stress. (see illustration to the right). By distributing stress the "weaker" less cohesive, more elastic, glue is actually more resistive to a peeling stress than a stronger, more cohesive glue. Furthermore, a highly cohesive glue places greater local stress on the substrates themselves frequently contributing to their de-lamination.

Within the last ten years or so an adhesive demonstrating this principle has emerged as a way to attach loose advertising items or cards into popular magazines. At one time many of these items were tipped in with a standard, non-elastic glue. The item usually could not be easily peeled out, without either delaminating part of itself or the page to which it was attached. These new adhesives are sufficiently elastic that the stress is distributed widely. Pulling these cards out is a task that sometimes endangers the attachment of the page to which it is connected. If you work at it carefully, however, you can actually remove the card and remove the glue without damaging either, indicating that not only are its cohesive properties low but so are its adhesive properties. But its elastic properties are impressive. The removed glue can be rolled, reformed, stretched to your heart's content and then pressed to another object to be used again. I must admit to having neither a name nor any further information as to what exactly this adhesive is but I do think it demonstrates well the contribution of elasticity to an adhesive.

Summarizing briefly, elasticity in a glue distributes stress in a peeling situation where stress tends to be concentrated in a small area. Should stress already be distributed, as in a pulling situation, an elastic glue would provide a bond that is weaker than a stiffer, more cohesive glue.

Page Attachment and Adhesive Elasticity

Assuming sufficient adhesion, a PVA with high cohesive strength does not necessarily make for better page attachment. In the dynamic environment where a page meets the glue line elasticity can offer advantages not suggested by simple page pull strength. The diagram to the right shows the joint where two different types of paper meet two different glues.

The papers are differentiated as low drape, papers that tend to bend easily or drape well, and high drape, papers that are stiff and do not drape well. A low drape paper tends to work well with either a stiff, cohesive glue or an elastic glue as the paper itself gives and, consequently, generates low potential energy at the glue line when flexed. Little is required of the glue itself.

Stiffer papers, however, presents another problem. If the paper is allowed to flex at the glue line, it will concentrate significant potential energy there. This will ultimately resolve as a peeling force on the glue line and the joint will fail as shown in the illustration above. An elastic glue, however, will both compress and stretch allowing the page to flex without failure.

A Note on Drape Factor: Anybody who reads books has experienced what I refer to as the drape quality of various papers. Some papers simply flop right over when a book is opened. Other papers prefer to stand straight up and many others fall some where in between. I have quantified this particular characteristic with what I call the drape factor and a simple tool I call a drape gauge. I define the drape factor as the distance in tenths of an inch a given sheet of paper can be extended beyond a horizontal edge (think table top) before the lead edge drops 1 inch. For instance, a typical page from a National Geographic Magazine extends 3.4 inches before dropping 1 inch. It has a drape factor of 34. Time magazine pages tend to have a drape factor of about 24. A ten point card stock has a drape factor of 44 with the grain and 56 perpendicular to the grain. The higher the drape factor the the stiffer the paper and the more potential energy there is when the page flexes at the glue line. Drape factor is measured with a single detached page. Some cross grain papers will cockle at the gutter edge when glued. If this cockling remains after the glue dries, it effectively increases the drape factor of the page.

Though elasticity allows movement, it does have less pull strength than a stiff, more cohesive glue. The key concept is movement. If the spine opens fully, the glue line is exposed to higher potential energy than if the spine is controlled in a way that the pages are not free to flex at the glue line. In the former case an elastic glue is called for and in the latter case a stiffer glue. This will not always be a clear choice but will be the best compromise between what the binder would like and what he can actually achieve.

Glue Joint Failures

Understanding how something fails is basic to devising a working solution. The success of a fan-glued binding depends largely on the success of the bond between the pages. This bond can suffer either failure of the glue or failure of the substrate. (see illustration to the right).

Glue failure assumes two forms. The glue suffers adhesive failure if the bond between the glue and the substrate fails. The glue suffers cohesive failure if the glue film separates. Substrate failure also assumes two forms. The substrate suffers surface failure if the adhesion of the glue is adequate but a part of the substrate or paper detaches under stress. The substrate suffers internal failure if the glue adheres but the substrate fractures or splits under stress. Substrate failure is more likely than glue failure and the way the substrate fails tends to be quite different coated and uncoated papers.

Coated Paper and Uncoated papers

Uncoated papers tend to exhibit internal failure, splitting internally when stressed. However, this turns out to be a distinct asset rather than a failure. When a book opens basic physics determine that something has to give way at the point of opening. My earlier supposition (which I stated in "Flexible Strength") was that the glue, if sufficiently elastic, provided the necessary give. My observations since then have shown that it is not the glue that gives but the paper itself. The splitting of the paper allows the book to open fully while accommodating itself to the inevitable expansion at the apex of the opening. Even though the paper splits internally the page pull strength remains high. Fan-glued books comprised of uncoated paper are very durable providing they are built with a proper adhesive and supporting structure.

While uncoated papers exhibit internal failure under stress, coated papers tend to fail at their surface. Though the high failure rate of adhesive, coated paper bindings is often attributed to poor adhesion, it is more often the case that glue sufficiently adheres to the substrate. It is the attachment of the coating to the paper that fails. A close examination of detached pages will usually show the coating delaminated from the edge of the page and still attached to the glue in the margin. Stiff papers that present the glue line with a large amount of potential energy can also act much like coated papers and have been grouped with them in the following examples.

How a substrate surface failure (or adhesive failure) resolves in a fan-glued book is shown in the illustration to the right. With an elastic glue the bonds between the opened pages fail, leaving the exposed sheets secured only at their edge. Furthermore, due to the elastic nature of the glue the adjoining pages slip and slide (no doubt aided by the low friction between coated pages) and the bonds between the adjoining pages tend to fail periodically, skipping a page or more, creating stepped blocks of pages. The leaves remain attached but resistance to forces pulling on the page is very low. Such a book will not do well under heavy use.

Should a stiff glue be used, failure is more immediate. As the glue has little elasticity, the pages at the opening tend to detach. The bonds between adjoining pages tend to shift and break in a fashion more consistent than with an elastic glue. These books perform terribly on a page pull test. The page will frequently pull loose in the process of trying to set up the pull test.

Searching for a glue with better adhesion will not solve the problem as most glues adhere adequately to the coating. Should adhesion to the coating be the problem its solution will simply transfer the point of failure to the bond between the coating and the paper. One approach to solving this problem is to remove the coating where the glue is to be applied. Sanding the edge of the pages prior to fan-gluing is one way to do this. My own page pull tests, however, have indicated that hand sanding prior to fan-gluing has no discernable effect and is merely a waste of time. The sanding action would need to remove the coating from the edge of the page without removing too much of the page. This is probably almost impossible to do manually. Sanding after milling the edges in a mechanized environment may be more effective if each page is momentarily separated out by a fanning action across a sanding drum. Not having access to such equipment I cannot comment one way or the other on its effectiveness. I believe the solution to the coated paper problem, at least in a hand bindery, lies elsewhere.

Spine Movement and Leaf Attachment

The solution to successfully fan-gluing coated papers lies in controlling the spine. The failures cited above are based on a book that opens flat. When a book opens fully as pictured above, the adhesive bond between the pages is subjected to peeling forces. The stiffer the paper the greater the peeling forces. Books using uncoated papers can be designed to work very well with a flat opening, but a relatively elastic glue should be used. An elastic glue will increase adhesion between the pages by better distributing the peeling force and will provide for the increased movement required by the spine at the point of opening.

Because of the likelihood of adhesion failures due to the weak bond between coating and paper, coated papers must be treated differently. Rather than design our leaf attachment to resist potentially destructive peeling stress, we design to avoid peeling stress altogether. Ideally, we want any stress on our leaf attachment to be a pulling force. This will ensure that the entire glue/paper joint comes into play and will significantly decrease the chance of a page detaching. We can do this by controlling the motion of the spine. The illustration to the right shows an example of a spine that is locked into a flat position with a stiffening element (in my actual experimental sample I glued a piece of binders board to the spine). A spine so stiffened will not move or "vee" up to the glue line. The pages remain locked into a vertical position at the glue line preventing an opening that exerts peel forces on the page to page junction. Whereas we maximize our strength against peeling with an elastic glue, we maximize our strength against pulling with a stiff, strongly cohesive glue.

When samples distinguishing between a freely moving spine and a stiff spine are subjected to a page pull test the difference is dramatic and varies with the strength/elasticity of the glue. My own page pull tests indicated the more elastic the glue the less the impact of stiffening the spine. I tested books comprised of pages from National Geographic Magazines and compared the page pull strength on samples using five different glues. For each glue I made two samples, one with a single super that allowed the sample to open flat to the glue line and one with a super, a strip of binder's board and a second super to further secure the binder's board to the spine. The reinforcement on the second sample immobilized the spine. The difference in pull strength from movable to immovable was at least double for the more elastic glue and up to seven times stronger for an inelastic glue. To push my experiment to the limits I used a yellow wood glue (probably PVA but without any plasticizers added). This glue dries stiff and brittle. The two samples glued with the wood glue illustrated the limits of pull and peel. They were both the strongest and the weakest in the page pull tests. The pages in the sample with the free moving spine hardly survived the opening of the book. They had the lowest pull strength of any glue tested. The wood glue's resistance to peel stress was almost nonexistent. However, the pages in the sample with the immobile spine had the highest pull strength of any glues tested and seven times the pull strength of its freely opening counterpart.

This is not to suggest that spine control be absolute as in my test samples. The degree of control required actually tends to correlate with the drape factor of the paper. The lower the drape factor (paper that drapes well) the less the binder needs to control the spine. A low drape factor paper, such as the typical Time magazine, only needs enough control to keep the spine from opening fully to an inverted "vee" opening. Control that allows a small radius at the point of opening is probably sufficient. As the drape factor increases the radius at the point of opening should increase. Unfortunately, this also means that stiff papers require the most control and, consequently, they produce the least openable or user-friendly books. This can be ameliorated by providing for a generous gutter margin.

I should also note that I did tests with uncoated paper with controlled and uncontrolled spines. With a properly elastic glue, the difference in pull tests was negligible. If a proper glue is used there is no advantage (at least in regard to leaf attachment) to adding anything beyond minimal control (usually provided by a single layer of super) to the spine. I would also reiterate that stiff papers, whether coated or uncoated, tend to behave similarly to flexible coated papers in the way they fail and often need to be treated is a similar way.

My conclusion: The adhesive attachment of a single leaf cannot be viewed in isolation from the structure to which it is attached. A structure can play to either the strengths or weaknesses of a given glue. The pages, the glue, the spine linings and the case all work together as a coherent structure regulating the action of the spine and creating the dynamic environment that ultimately affects leaf attachment. Many of us have experienced firsthand the role of structure in leaf attachment. Think of a paperback book with a thick, stiff to brittle , heavy layer of hotmelt adhesive on the spine. The hotmelt immobilizes the spine. The book may be difficult to read but the pages remain intact until the spine gives. Once the spine breaks, the pages will often come loose either individually or in blocks.

Designing durable books requires understanding how all the components work together. The single most important decision in designing a durable book structure begins with choosing the paper that makes up the individual pages. Most every structural decision that follows is based on this first decision. Unfortunately, the binder is seldom involved in selecting the paper and must compromise to find the best solution for the raw material with which he is presented.

To be continued in Reflections - part 3- Spine Control

Reflections on Book Structures - part 2

Reflections on the Glue Line