The epoxy surface has a choice: contact with air; contact with glass; contact with water. The glass has a choice: contact with air; contact with epoxy; contact with water. When you calculate the competing choices, although glass and epoxy prefer being with each other compared to being with air (so dry adhesion is good), when water is around, the balance of choices shifts to them both preferring to be with the water: the epoxy floats off (Figure 3.6).
Figure 3.6 How water can destroy strong adhesion – by creating water-epoxy and water-glass interactions that in total are stronger than the original epoxy-glass interaction.
Although the choice is clear, why do the glass and epoxy get the chance to go with the water? How can the water get between them? Water is a very small molecule and it is the humbling experience of anyone who wants to keep water away from an interface that it can usually find a way to get there. It is a very destructive molecule for those who work with adhesives because not only does it change the surface energy balance it can also attack many of the chemical bonds and weaker “hydrogen bonding” interactions that help hold surfaces together. However, such processes take time. I assume that the handle came off my brush at that specific moment because water had slowly been getting to the interface over the years of use since I repaired it and a critical amount had, at last, arrived.
In the Kendall example, the epoxy floats off rather easily because water has an especial attraction to the silica at the surface of glass, converting it to a silicate, and the process can zip along the interface rather quickly. This effect of water at the surface of glass means, incidentally, that you cannot reliably leave two sheets of glass in contact when there is a moist atmosphere. Over time the glass at the contact point is “dissolved” by the humidity, leaving an annoying semi-opaque layer at the surface which cannot easily be removed.
3.4 USE SPARINGLY
It always used to surprise me that tubes of household glue instructed us to use a thin layer. If the glue is doing the sticking, surely more is better. Logic then stepped in. Any manufacturer wants the user to use as much as possible. If they are saying “use sparingly” they must have a very good reason.
In turns out that there are two reasons. One reason must await another chapter. The second is simple: most of these adhesives are rather useless polymers. This is not an attempt to insult the adhesive makers. Adhesion is a property of the system and the overriding priority is to find an adhesive formulation that is sufficiently liquid in the tube, with a sufficiently long shelf life before going solid in the tube, and with a sufficiently fast drying or solidifying time – along with desirable properties in terms of health, safety, price, odour etc. when in use. The end result is that (excluding crosslinked systems such as epoxies, urethanes, urea/formaldehydes etc.) the polymers are nowhere near as strong as apparently simple polymers such as PE or PET, and even less strong than the ceramic of a cup. Yet they are able to do their job if applied properly, because adhesion isn't just about strength.
This weakness of the adhesive means that if you have a thick layer in the joint, failure might be within the adhesive itself – this is called “cohesive failure” (compared to “adhesive failure” when the adhesion fails (Figure 3.7)) and is deeply unpopular with users; they don't like it when it is obvious that the adhesive itself wasn't up to the job. When the layer is as thin as possible, then the strength of the polymer itself is less relevant and/or less obvious when a failure is observed.
Figure 3.7 The difference between adhesive failure (at the interface) and cohesive failure (within the adhesive).
In Chapter 5 I will describe a special adhesive polymer which is one of the most useless polymers on the planet with very few desirable properties. If you include 100 nm of this polymer in a special type of joint, the adhesion is very poor – the joint fails by cohesive failure within this useless polymer. If, instead, you include a 20–50 nm layer then, as I once spent a few challenging months proving, the joint is almost indestructible.
3.5 THE PAIN OF CONTACT IN CONTACT LENSES
A downside of “thin” and a painful example of pure surface energy adhesion between nicely smooth surfaces is known to those who wear a pair of contact lenses rather longer than they should.
Under normal use, the lenses are held in place by surface tension forces from the liquid on the surface of the eye. You can readily get an idea of the scale of surface tension forces with a well-known hoop-like breakfast cereal. Float one hoop on the surface of your milk and with a pair of tweezers gently pull it upwards. You see the milk clinging to the hoop and you find it pulling a significant weight of liquid up from the surface till gravity exceeds surface tension and the hoop is freed. Those surface tension forces are not large, but for contact lenses they are more than enough to hold them in place. However, such forces alone can easily be overcome with a sudden shock, like blinking, so it requires an extra effect to keep lenses from popping out.
Remember Stefan's squeeze law? If you have a thin layer of liquid, it gets increasingly hard to squeeze it out as it gets thinner. Stefan's law works in reverse, too: if, instead of pushing, you try to pull two surfaces apart, it is hard for the liquid to flow inwards, at least for a while; you have to provide a significant force to allow that flow to happen at a significant speed. Fast, pull forces to the lenses are resisted by the difficulty of liquid flowing through a narrow gap. Together, this means that a normal moist eye has a contact lens floating on a thin and stable cushion of tears. Removing them is all about letting the tears flow easily under the lens as it is pulled away. Squeezing off a soft lens is easy. Blinking off a hard lens probably requires the blink to push liquid under the lens, to fight against the Stefan's law reluctance to flow.
The painful part is when the eye dries out after a long day of wearing lenses. Now you have surface energy contact with the eye. As we know from the gecko, these forces strongly resist a vertical pull and give a modest resistance to shear. If you could peel the lens then it would come off easily, but that isn't viable for most contact lens users: it is exceedingly tricky to catch the edge of a lens. So removing them is a difficult process of applying as much peel and shear (including the trick of holding the lens in place with a finger and rolling the eye away underneath it) and as little butt force as possible, all the while trying to induce tears to re-float the lens.
3.6 (DON'T) ROUGHEN BEFORE USE
One of the most enduring of adhesion myths is that you should roughen your surface before use to promote adhesion. We have already learned that this is wrong because surface energy adhesion is best when two smooth surfaces come together as intimately as possible. The more you roughen the interface, the harder it gets to ensure intimate contact. Yet the defence of roughening relies on surface energy; it says that if you roughen, you get a larger surface and therefore more surface energy adhesion. You also, they say, get mechanical interlocking between the surfaces, which sounds awesomely helpful.
Both the extra surface area and mechanical interlocking ideas are wrong in an interesting manner. Let's look at a typical rough surface measured by sliding a diamond stylus across the surface and measuring how its tip goes up and down (Figure 3.8). Although the trace in the figure is a simulation from one of my apps, it is realistic and instantly recognized by anyone who has measured a rough surface using this stylus technique.
Figure 3.8 A typical output from a surface roughness measurement device. It