PEEK performance - thermal spraying of a polymer

TWI Bulletin, July/August 2004

Melissa Riley joined TWI in February 2003 as a project leader in the surfacing team. Her background is in biomaterials and PVD coatings for orthopaedic implants. Since joining TWI she has been involved in a wide range of projects involving high velocity oxy-fuel (HVOF) spraying.

PEEK, or Polyether-ether-ketone is one of a range of engineering polymers which also includes PPS (polyphenylene-sulfide) and LCP (liquid crystal polymers). They are resistant to a wide range of chemicals and have a working temperature up to 200-250°C. They may be applied as a coating to protect the substrate or as functional coatings where PEEK may be used as an electrical insulator due to its low electrical conductivity. Here Melissa Riley discusses HVOF spraying of PEEK and methods of improving the adhesion of the coatings.

The conventional method of applying PEEK as a coating is by a two step process: first depositing the powder using electrostatic spraying onto the surface of the component which is then heated in an oven at about 400°C to fuse the coating. This method is difficult to carry out on large or thermally sensitive components. By using thermal spraying the polymer powder can be sprayed through a flame where it is heated and partially melts before impacting and adhering to the substrate to form a coating. This method avoids the need for a subsequent heat treatment to fuse the coating. The process offers two main advantages. Firstly, PEEK coatings can be prepared whilst maintaining a low component temperature (below 300°C). Secondly, this approach can be used to spray-form composite or graded coatings that incorporate other ceramic, metallic or polymer materials.

Thermal spraying processes

In the thermal spraying process, a powder consumable is heated to near or above its melting point and propelled onto a substrate. The impacting particles deform and build-up to form a coating. Conventional flame spraying of thermoplastic polymers is relatively common for those with lower melting temperatures such as polyamide, polyurethane and polyethylene.

In flame spraying the particle heating and melting occurs in the spray gun flame, which is then followed by fusion on a substrate held at a temperature above the melting point of the polymer. The deposition of PEEK by thermal spraying is more difficult than lower melting point polymers, because it is undesirable to hold the substrate temperature above the PEEK melting point of about 330°C.

When spraying PEEK, the aim is to heat the powder to allow particle deformation and flow on impact, together with bonding to the substrate and previously deposited material to form a cohesive coating, whilst keeping the particle temperature sufficiently low to prevent decomposition.

Of the commercially available thermal spraying processes, High Velocity Oxy Fuel (HVOF) spraying is considered the most suitable and involves oxygen fuel combustion in the chamber of the HVOF gun. The systems run on either liquid (kerosene) or gas (propylene, propane, hydrogen and acetylene), which affects the combustion temperature.

The fuel and oxygen combust at high pressure and exhaust gases exit the gun nozzle at high velocity (supersonic) speeds. The powder consumable is fed directly into the combustion chamber or into the gun nozzle where the particles become molten or semi-molten. They are carried to the substrate by the HVOF flame where they impact to form a coating built of splats of individual powder particles. The process is designed to minimise heating of the powder consumable whilst maximising the velocity of the powder particles prior to impact with the substrate. Compared to other thermal spraying processes, the kinetic energy of the spray particles can play a larger role in coating formation enabling denser coatings of PEEK to be prepared without significant decomposition. Figure 1 shows the Diamond Jet and how the HVOF process works.

HVOF is generally used for spraying metals, ceramics and cermets due to the higher velocities and lower temperatures of the process compared to plasma spraying. Difficulties associated with HVOF spraying of polymers are:

• Obtaining the correct powder size for the HVOF process
• Feeding this powder though the HVOF gun
• Preventing thermal degradation of the powder whilst in the combustion chamber and during flight to the substrate
• Poor adhesion to the substrate

Previous work has demonstrated the spraying of PEEK by HVOF. The major difficulty was obtaining powder suitable to be fed in to the HVOF guns, previous work had agglomerated spray dried powders. Two more suitable types of powder have become available from Victrex plc. These are high molecular weight (MW) 450PF and low MW 150PF, with both powders having a D50 of about 45µm.

Objectives

The aim of this work was to produce PEEK coatings using 150PF powder via HVOF thermal spraying and to investigate the effects of preheating the substrate on the adhesion of the coatings. The use of graded metal layers to improve the adhesion of HVOF sprayed PEEK coatings was also investigated.

Results

Thermal Spraying of PEEK

Coatings were produced using the Diamond Jet Hybrid HVOF (DJ2700) system (Sulzer Metco) with 150PF PEEK powder. The closed loop fluidised bed powder feed system associated with this system enables the powder to be fed evenly through the gun and by using propylene fuel gas and process parameters for low melting point materials it is possible to spray PEEK coatings without significant degradation of the powder. Coatings were sprayed onto carbon steel prepared by grit blasting with alumina grit (mesh size 20) and degreased prior to spraying. These were approximately 250µm thick and prepared by moving the HVOF gun in a zig-zag pattern across the surface using robotic controls.

Preheating of substrates

PEEK coatings were sprayed onto substrates at different temperatures: ambient (20°C), 150°C and 250°C. The aim being that by heating the substrate above the glass transition temperature (Tg) of the PEEK (143°C), increased flow of the polymer particles would occur on impact, resulting in a more adherent and coherent coating. A thermocouple was attached to the back of the test pieces to record the temperature during the coating process and the samples preheated to the correct temperature prior to the spraying process. The same process parameters and number of passes of the spray gun were used to build up the coatings in each case.

The resulting coatings were all light tan in colour with darkening seen with increasing substrate preheat temperature ( Fig.2). The variation in colour is believed to represent decomposition of the powder and therefore a lighter colour is more desirable.

Cross sections of the coatings revealed subtle differences in the microstructure. The PEEK coating sprayed onto the ambient (20°C) substrate contains some porosity and lighter areas within the coating ( Fig.3). The coating deposited on a substrate preheated to 150°C appears to be less porous compared with that prepared on 20°C substrate but lighter areas still exist in the coating ( Fig.4). In contrast the coatings prepared on substrates preheated to 250°C contain more porosity than those produced on 20°C and 150°C substrates, however, there are no light areas within the coating, the particles have a lower aspect ratio and the particle boundaries are less visible ( Fig.5).

Previous work has shown a decrease in the porosity of PEEK coatings with increasing preheat temperature up to 400°C. Comparing the 20°C sample ( Fig.3) and the 150°C sample ( Fig.4) this may appear to be the case. However, the coating deposited on the 250°C sample ( Fig.5) has more porosity. It is unclear whether this may be due to particle pull-out during polishing or due to consolidation of the porosity along the particle boundaries during flow of the PEEK material.

The average coating thickness decreases with increasing substrate preheat temperature. The thickest coating is produced on the 20°C substrate (356µm) and the thinnest on the substrate preheated to 250°C (259µm). The particles in this coating are flatter and elongated, suggesting that the PEEK particles have flowed on impact, resulting in a thinner coating. As the spray conditions are consistent for each coating trial then the substrate temperature influences the structure of the coating. An increase in substrate temperature produces coatings with less light areas and a more lamella structure. The light areas correspond to areas of unmelted powder particles.

The temperature profiles of the substrates during the preheating and spraying process are given ( Fig.6). They are not over equal time scales due to variations in the delay between the spray passes. For samples not preheated (20°C), the substrate reached a maximum temperature of 155°C during the spraying process. The substrate temperature was only above the glass transition temperature (Tg) of PEEK (143°C) for approximately a third of the spraying time compared with the preheated samples, where the substrate temperature was always above the Tg. For samples preheated to 150°C the temperature rises to a maximum of 200°C. For samples preheated to 250°C the temperature drops to 220°C and stabilises at this temperature during spraying.

Bond layers and graded coatings

The use of graded coatings as a way of improving the adhesion of the polymer to the substrate was investigated. Metal layers, such as NiCr, NiAl and molybdenum, are often used as bond coats to improve adhesion of thermal spray ceramic coatings to metal substrates. In this work a metal bond layer of 50/50 NiCr alloy was deposited initially followed by a 50%vol mixture of PEEK and NiCr, and then a final layer of PEEK to try to improve the adhesion characteristics of the PEEK coating. Two sets of samples were prepared with differing thickness for the metal and composite layers. Cross sections of each of the coatings produced are shown in Fig.7 and 8, and the adhesion test results are shown.

There is good distinction between the different layers of the NiCr/PEEK graded coatings with equal thickness layers ( Fig.7). The settings used to deposit the NiCr layers were those recommended for the PEEK and yet a good coating of the NiCr has been achieved. There is a small amount of porosity present and some spherical (unmelted) particles within the layer. A good coating has also been deposited using a 50% by volume mixture of NiCr and PEEK. The layer contains approximately the same volume per cent as the starting material suggesting there are no problems associated with feeding mixed powders into the HVOF gun. A coating of PEEK has been deposited on top of these layers. For coatings with thinner NiCr and NiCr/PEEK layers and a thicker PEEK layer, similar results are observed ( Fig.8).

Coating Adhesion

Adhesion of the coatings was measured using the tensile pull method as described in ASTM D4541. In this test a 20.4mm diameter steel stud was bonded to the coating surface using an epoxy adhesive (Araldite Rapid) cured at ambient temperature for a minimum of 12 hours prior to testing. The adhesion tester applied a tensile load to the stud and maintained the load normal to the coating surface using a hydraulically self-aligning mechanism. Load was increased until failure occurred and the location of failure recorded. A failure stress was calculated from the load at failure and used as a measure of coating adhesion. The results are shown in Tables 1 and 2.

Table 1 Adhesion test results for preheated substrates

Table 2 Characteristics of graded coatings

The adhesion test results for the PEEK coatings show an increase in coating adhesion with increasing substrate temperature from 8MPa to 13MPa ( Table 1). For the sample preheated to 250°C the PEEK coating was not completely detached from the test coupon indicating increased adhesion to the substrate. From this data it appears that the amount of time and the temperature the PEEK material is above its glass transition temperature (Tg) influences the structure of the coating. Above Tg the material can flow resulting in denser coatings and improved adhesion.

The adhesion of both the NiCr PEEK graded coatings at 15-16MPa is approximately double that of the directly deposited PEEK coating without substrate preheating (8MPa) ( Table 1 and 2). For the sample with three equal, 100µm thick layers the coating fails mainly at the NiCr/PEEK - PEEK layer interface. The PEEK coating is removed from the test coupon along with some NiCr particles from the graded layer. In the case of the graded coating with thin (50µm thick) metal bond and graded layers, and a thicker layer of PEEK (200µm), the failure of the coating is cohesive within the NiCr/PEEK graded layer.

The use of NiCr alloy and graded coatings can improve the adhesion of PEEK material on carbon steel substrate without the need for preheating. Other materials such as NiAl and molybdenum may achieve similar results. The use of a metal bond layer has significant practical advantages over preheating the substrates as, other than for small samples, preheating is very difficult or impractical. By developing graded coating systems PEEK coatings be can sprayed on to much larger components.

Future work in this area may consider the use of PEEK for corrosion protection and methods of further reducing the porosity levels within the PEEK coatings for single and graded coating systems.

Conclusions

• Coatings of 150PF PEEK powder can be deposited using the Diamond Jet 2700 HVOF system and powder with a nominal D ₅₀ of 45µm.
• The porosity of HVOF sprayed PEEK coatings can be reduced by preheating the substrates.
• The adhesion of PEEK coatings can be increased approximately 50%, by preheating the substrate to 250°C.
• The adhesion of PEEK coatings can also be increased, by approximately 100%, with the use of a NiCr alloy bond layer followed by an intermediate composite layer of NiCr/PEEK between the metal substrate and PEEK coating.