Bonding low energy plastics

6 mins read

Bob Goss of Loctite discuss a range of engineering adhesives and primers which are used for bonding these materials

Engineering plastics that are quite difficult to bond include PEEK (polyetheretherketone), PTFE, FEP (fluoropolymers), POM (acetal/Delrin), PE (polyethylene), PP (polypropylene), PBT (polybutylene terephthalate). PEEK is a really good engineering plastic; you can machine it, you can do all sorts of things with it, but when you want to stick it to something, that can be quite a challenge. PE and PP are good, low-cost, easy-to-mould plastics, very widely used, but can present their own challenges in bonding. PBT, if it is glass-filled, we don’t normally have to use any surface treatment. By itself, it can pose challenges.

By contrast, engineering plastics that are not too difficult to bond include ABS, LCP (liquid crystal polymer), PC, PC/ABS co-polymer, PMMA (acrylic), PPS (polyphenylene sulphide), nylon (polyamide), PS (polysulphone), PET (polyethylene terephthalate).

For elastomers or rubbers, those that are relatively easy to bond include nitrile rubber, butyl rubber, neoprene rubber, natural rubber and EPDM, although only some cyanoacrylates will bond to EPDM. Other difficult elastomers include Viton and silicone rubber – though an RTV (Room Temperature Vulcanising) silicone sealant will bond to silicone rubber quite well. The RTV’s are quite slow-curing, but some of the two-part silicones would be faster. Thermoplastic elastomers and thermoplastic polyurethanes are becoming more popular particularly in the medical industry where they are replacing PVC, and they can be a challenge to bond.

Surface energy is measured in millNewtons/metre. We talk about surface energy of a solid and surface tension of a liquid, but it’s the same units. PTFE, the non-stick frying pan material, has a surface energy of 15-18 mN/m. By comparison, the surface energy of PVC is 36-38 mN/m (see also Figure 1). Note that surface energy is a complex subject involving both dispersive and polar components. If your plastic has a lower total surface energy of about 33-34, it’s going to be difficult to adhere to. The adhesive will still cure on surfaces like PEEK or acetal, but it won’t stick very well because it doesn’t wet out.

You can measure surface energy with pens with different inks corresponding to different surface energies, and they either write or they don’t. If the ink spreads into globules, then you know the surface energy is too low. If it writes, it wets out and there is a greater chance of success with the adhesive.


There are two main options to overcome low surface tension. First is primer. Loctite SF 770 Primer has been around since the 1990’s but has recently been upgraded to be CMR (carcinogenic, mutagenic or toxic to reproduction) free. It can be applied by brush or spray and gives best results when used with cyanoacrylate-based adhesives on polyolefins (PP and PE), but it often shows good results with acetal and with PEEK and to a certain extent PTFE. It is only intended for cyanoacrylates. It is not particularly good in a high-peel situation, but that’s where you can look at the joint design to try to limit the peel load (see below). It does work really well; don’t underestimate it. Sample results are shown in Figure 2.

In Figure 2, with primer, the strength of the bond with a polypropylene substrate has massively improved. Notice that the bond strength performance on polycarbonate has decreased with the application of primer. That’s because the primer is not only an adhesive promoter, it is also an accelerator. It is possible to have a situation where the adhesive cures before it sticks, as has happened here. As a result, if you are bonding PEEK to PC, only apply the primer to the PEEK, not the PC.

The other way of treating low-energy surfaces is by plasma or corona treatment. This involves running a high-energy flame quickly over the surface of the plastic, at a rate of 200-300mm/sec across the plastic. Depending on the plasma, the effect can last a few days or a few months. On PTFE for example, it can raise the surface energy from 15-16 to well above 40, and other plastics are the same. We have worked with a number of different engineering companies that have bought a plasma system and then used our adhesives after they treated it.

A side-effect of plasma and corona treatment is driving the moisture off of the surface. As a cyanoacrylate relies on surface moisture to cure, a plasma-treated surface should be left for about 20 minutes for the natural moisture in the atmosphere to re-deposit before bonding in production.


The important thing in bonding to get the joint design right. Most applications use one of four joint types: flat surfaces, coaxial joints, tongue and groove and stepped joints. Flat surfaces use a standard lap shear joint, which is quoted in every data sheet. A number of modifications can reduce the stress of the joint, which in a lap shear joint is typically concentrated at the end of the joint.  Chamfering the ends or adding a little fillet can help with peel load and minimising stress distribution across the joint, for example.

However, a large excess of adhesive outside the joint is not always what you want to see when you come to your final component. In an example of glass-to-metal bonding, we put a little groove recess on the underside of the knob and this helped to allow a certain tolerance in dispensing and somewhere for excess adhesive to flow to; it also improved the strength of the joint (see Figure 3). The design is in the detail.

In Figure 4, a polycarbonate window requires to be bonded to an acetal frame. By having a recess in the frame, any excess adhesive can be accommodated, so when viewed from the underside the adhesive is not visible around the periphery.

Here’s a question. Which joint will be stronger? One joint is of two 25mm pieces with a 40mm overlap; another has two 50mm pieces with a 20mm overlap. In both cases, the bond area is 1000mm². If I double the overlap, and double the bond area, do I double the strength? If I had a tensile test machine, and we bonded samples and left them to cure, and we do the tests, the results could be plotted as dots on a graph (Figure 5). They would show that as the overlap is increased, there is a marginal increase in strength; but doubling the overlap does not double the strength. If the width instead of overlap were doubled, the bond is stronger again (but also not doubled). The reason is, with a lap shear joint, all the stress on the joint is at the ends of the joint. To increase bond strength, increase width, not overlap. Doubling the overlap length does not increase the tensile shear strength. Doubling the width of the joint will increase the strength.

The same applies for a cylindrical joint. If you have a joint of 6mm diameter and you want to bond a collar, you will increase the axial strength and the torque strength by a small extent by increasing the length of the collar, but to increase strength significantly, you need to increase the diameter.


Co-axial joints include assembling tubes together and putting acetal gears on shafts. If you apply adhesive on the male part of the joint, it’s more likely to come out (Figure 6). If you put it on the female part (Figure 7), there is a risk of it blocking the bore – particularly if it is small in diameter – so you have to think about where the adhesive will go. Referring to dimensions in the figure, typically D-d = 0.05 mm to 0.1 mm and L/D is between 0.3 and 4. Again we can help there. In a co-axial joint for a water mattress application (Figure 8), we included a little step in the joint to make sure the adhesive didn’t go inside the mattress.




If it’s a blind hole (Figure 9), apply the adhesive on the bottom surface so when the male part is inserted it comes up around the sides of the male part.

A similar scenario is posed by a tongue and groove joint (Figure 10), which offers good stress distribution. The tongue can be extended on one side so when the joint is closed, any excess goes one way and not another.

In thin-walled components, a tongue-groove may need to be replaced by a stepped joint (Figure 11). 

A rougher surface finish can also help adhesion. If possible, spark-erode the mould tool to give a slightly rougher surface in the bonded area. And thin bond lines are generally the best.

In conclusion, I’m always one to go for the right solution. I’ve been an engineer a long time, and I know that there are times when we’re never going to be able to stick that together. In such circumstances, one question I have asked customers is, can you change the plastic? Can you change from PP say to another plastic that’s easier to bond?

This article is an edited version of ‘Bonding low energy plastics’ given at Engineering Solutions Live, 24 March 2022, at the British Motor Museum in Gaydon, Warwickshire.

BOX: Goss retired in April after working for applications engineering in Loctite since 1985. In his presentation, he said: “One of the benefits of being an engineer at Loctite is the huge variety of engineering applications. One day you are working with a company like Dyson, and the next day it’s JCB and the day after that it’s a medical device company. It can be very varied, and I’ve had the opportunity and the privilege to go round lots of these engineering companies throughout my last 36 years to all sorts of different engineering accounts.”

BOX: Henkel and Loctite has a 70-page booklet that lists all of these plastics and more, and gives lap-shear data for bonding, for example, PC to PC or PEEK to PEEK, for cyanoacrylates and epoxies and UV acrylics and hot melts and other generic families of adhesives, both with and without a primer. It is available via