Skip to main content

Coatings for Composite Aero-engine structures

Final Report Summary - COMPOCOAT (Coatings for Composite Aero-engine structures)

Executive Summary:
The COMPOCOAT project has developed a new surface protection system for composite aero-engine aerofoil structures with metallic leading edge. The surface protection system is based on TWI’s CompoSurf family of coating technologies. It provides erosion protection for the metallic leading edge and thermal protection for the composite body. In addition to the surface protection system, the Comeld® joining process has been utilised for achieving high strength bonding in the metal-composite interface at the leading edge. Laser welding has been tested as potential welding method for joining the leading edge without any thermal impact on the composite.
The operating temperature for the new aerofoil structure is in excess of 250°C.

Project Context and Objectives:
The main objectives of the project are:
• With the guidance of the Topic Manager, select a resin material system with high Tg value, suitable for rate manufacturing of composite blades
• Develop a new surface protection system that will be coated on the composite structure and provide a thermal barrier
• Develop a new surface protection system that will be coated on the metallic leading edge and provide erosion and FOD protection.
• Develop a new manufacturing procedure for the joining of the metallic leading edge to the composite body of the aerofoil

Project Results:
WP1: Selection and design of a surface protection system, matrix material and effective bond
Task 1.1 Detailed definition of the part to be manufactured and the operating environment
A matrix with all the specifications for the geometry of the part, the resin matrix system, the surface protection system and the tests that will need to be performed has been agreed between TWI and the Topic Manager (Table 1).
The ordering and delivery of the selected material has proven very challenging as it sourced from USA. The Topic Manager has mediated between TWI and the supplier company in order to accept a low quantity order. Delays in the order and delivery of the material have led to a request of extension of the project which has been accepted by the Project Officer.

Task 1.2 Matrix material characterisation
The initial plan for task 1.2 was to select a number of candidate resin matrix materials and perform some preliminary checks before finally choosing the most appropriate material for the specific application. However, the Topic Manager has decided to proceed with the MVK-14 system. Therefore, the characterisation work was focused on the particular system. Representative results from the cure of the MVK-14 system are shown in Figure 1

Task 1.3 Down - Selection of surface protection system
Thermal spraying of YSZ, Cr2O3 and WCCoCr materials onto metallic substrates (with or without a suitable bond coat, as appropriate) is a mature process and leads to high quality coatings. By comparison, coating deposition onto low thermal tolerance composite materials is challenging. There are a number of factors that can affect coating deposition and performance and spray parameters are often modified to enable deposition of coatings onto low thermal tolerance substrates. As such it is often necessary to sacrifice some coating quality, by modification of spraying parameters, to ensure that the coating can be successfully deposited onto the surface of CFRP substrates. Despite these challenges, successful deposition of the coatings onto the polyimide material has been demonstrated and represents a significant achievement and step forward in coatings development.
In summary:
Surface Preparation
• Renegade MVK-14 carbon-fibre reinforced polyimide CFRP was prepared by mechanised grit blasting using 80mesh alumina grit and 20psi run pressure to roughen the surface prior to provide a mechanical key for the coating to adhere.
• The surface roughness of the grit blasted surfaces prior to coating were 8µm Sa and 95µm S10z post blasting and featured resin rich areas and regions of exposed glass fibers (Figure 3).

Bond Coat Materials
• Aluminium, zinc and zinc-aluminium coatings were successfully deposited on to fully cured Renegade MVK-14 carbon-fibre reinforced polyimide using the wire flame spraying process.
• Adhesion of the Al, ZnAl and Zn WFS coatings on Renegade MVK-14 were 6.63MPa 5.64MPa and 5.87MPa respectively.
• The presence of a coating on the CFRP substrate did not appear to have any detrimental effect on the bending and fatigue performance of the substrate material.
• Based on the adhesion values and melting point of aluminum, compared to Zn and 85Zn-15Al, aluminum was selected as the bond coat material for high temperature applications.

Functional Top Coats
• YSZ and Cr2O3 APS coatings could be deposited on to Renegade MVK-14 carbon-fibre reinforced polyimide substrates prepared with WFS Al bond coats.
• WCCoCr HVOF coatings could be deposited on to Renegade MVK-14 carbon-fibre reinforced polyimide substrates prepared with WFS Al bond coats.
• WCCoCr HVOF coatings could be deposited directly onto titanium alloy substrate, but not if a WFS Al bond coat was present. This may present additional challenges when coating surfaces containing both CFRP and Ti alloy regions.
• YSZ coatings were characterised by relatively high porosity and roughness combined with low cohesive strength, but performed well in 4-point bending and fatigue tests.
• Cr2O3 coatings were smooth but exhibited cracking in the as-sprayed condition. This may be a result of coefficient of thermal expansion mismatch between the substrate, bond coat and chromia layer. Compared to the YSZ the heat input to the samples during spraying is likely to be greater due to the shorter stand-off distance.
• WCCoCr HVOF coatings were relatively smooth with low porosity and high cohesive strength and hardness.
• WCCoCr HVOF coatings were characterised by high stiffness and residual stresses and performed poorly under 4-point bending and fatigue as a result.

Based on the above, the performance varied depending on the parameter assessed. These performances are summarised below, with the best performing candidate or candidates (for that parameter) highlighted.

WP2: Application characteristics and performance
Task 2.1: Small scale coupons preparation and characterisation
TWI has completed the initial work described in this report to design and optimise the geometry of the Surfi-Sculpt feature for the COMPOCOAT application and to trial production of these in three different array orientations. These have been tested once they have bonded to the composite in order to identify the best orientation of features for the next phase of the project.

The objectives of the work reported here were to:
• Identify and select the geometry of feature on the metal.
• Produce feature using Surfi-Sculpt.
• Analyse feature to support optimisation.
• Develop up to three different arrays for the layout of the feature.
• Produce test specimens for joining to the composite

Table 3 details some of the key terms associated with Surfi-Sculpt. The work was performed on Ti64. The initial scope of work comprised selecting a suitable feature shape to be formed by Surfi-Sculpt on the metal. The main aim was to enhance the Comeld bond with the composite. The initial design selected for the feature was a cone with a height approximately 1mm to 1.5mm. A total of 37 Surfi-Sculpt treatments were created with varying geometries for the feature. Various approaches including changing the geometry, layout of the legs in the pattern and varying the amplitude of pattern, were taken to generate different features. It was expected that features designed with a shallow intrusion would show improved fatigue performance rather than those with a deep intrusion. Therefore these were preferred and developed further.
Of the samples, four were selected for metallurgical analysis (T31, 32, 33 & T37). The analysis showed the best specimen was the T37 sample (Figure 2). The mean micro hardness result was 313 ± 15.7HV. The values varied between 292HV and 335HV through both the feature and the parent material. Full results for T37 sample are in Appendix B. These values were acceptable
Once the T37 feature was selected, a suitable layout of the feature into an array was required. The three different arrays that were decided up for the next stage of testing were:
• Square array (Figure 3a)
• 0.50 offset array (Figure 3b)
• 0.33 offset array (Figure 3c)

A panel was designed which incorporated 5 specimens to aid the lay-up of the composite and the subsequent cutting of the Comeld joint. 30 specimens were produced in total.
Preliminary bead-on-plate (BoP) laser trials, 225mm length, were performed to develop welding conditions for butt-welded Ti6/4 flat test coupons, 2mm thick. A range of laser parameters has been investigated, summarised in Table 4. Typical results are shown in Figure 5. The quality of the produced welds was preliminarily assessed by inspection of the top-beads and roots, under an optical microscope. Focus was given on any occurrence of unacceptable surface discolouration due to oxidation, melt ejections and blow holes. The most successful parameters have been subjected to radiographic inspection to assess whether any defects occurred within the weld. No major defects were detected.

A number of characterisation tests have been performed on composite panels coated with WCCoCr. Figure 6 shows the panels physical appearance during a thermal exposure test. The time required for the coupon to reach temperature depended on contact between coupon and heat treatment basket, as well as location within the furnace. E.g. if the coupon face contacted the basket, heating time was ~15 minutes. If only an edge was in contact this rose to ~30 minutes. It was concluded that 30 minutes exposure to the hot furnace air was sufficient to ensure that a coupon had reached the specified temperature.
The maximum continuous operating temperature for the CFRP system was specified as ~206°C. Based on discussions with the Topic Manager, a temperature of 200°C was selected for long term thermal exposure. After 750 hours at 200°C, the furnace was deactivated and allowed to slowly cool to <45°C over a period of 20 hours. The coupons (Figure 1c) were extracted and characterised as described below

Thermal Cycling
Furnace temperatures of 275°C and 250°C were selected with water as the quenching medium. The samples were held at the maximum temperature for 30 minutes. They were then taken out of the oven and quench in water at room temperature. The cycle was repeated for 10 times. A batch of 17 coupons was loaded into the furnace on top of the heat sink (at 275°C). A second batch of 14 coated coupons was loaded into the furnace at 250°C and again subjected to the same 10 cycle thermal cycling regime. When cycled from 275°, the coating cracked in 16 out of 17 coupons, but when cycled from 250°C the coating cracked in only 1 out of 14 coupons.
Following extraction, coupons from the 250°C batch characterised as described below.

Characterisation Methods
The three categories of coupons were given the following notation (where X is a letter identifying the particular coupon within a series) and then characterised
AS – As Sprayed. All assumed to be nominally identical.
TE-X – Thermally Exposed (200°C in air for 750 hours, slow heating and cool)
TC-X – Thermally Cycled (250°C to water quench in perforated stainless steel mesh basket, 10 cycles)

Samples were assessed in terms of coating quality, surface roughness, adhesion, sliding wear and erosion performance in the as-sprayed, thermally exposed and thermally cycled conditions. Not all characterisation techniques were employed for all specimens. The coupons were visually inspected and photographs taken of the coated surface. Contrast enhancement was used as necessary to highlight features of interest (cracking, surface irregularity etc.). Reference coupons from all three conditions were subjected to brief, qualitative impact testing to determine whether the coating could be easily dislodged, cracked or damaged. Typical coated coupons from each series were cross-sectioned along the coupon centre, then vacuum cold mounted in epoxy resin to minimise any damage from the preparation process. The cross-sections were then prepared by grinding with a series of silicon carbide papers (up to 2400 grit), then polishing with increasingly finer diamond pastes to a finish of ¼µm. The cross-sections were examined by light microscopy and micrographs taken of typical regions and/or any regions of special interest. The microhardness values of the top coat and bond coat was measured using a Vickers hardness indenter with a load of 100g. At least 10 locations were measured per coating and an average value calculated. The coated surfaces of coupons were examined by stereo light microscopy with micrographs taken of typical regions and of regions of special interest. Three-dimensional maps of the coating surface were created using focus variation microscopy and non-contact profilometry (Alicona InfiniteFocus) Where no cracking was visible to the naked eye, maps were examined in detail to confirm or deny the presence of smaller cracks. Where cracking had been detected, maps were taken across the cracks to characterise their shape, branching and general morphology. The maps are presented following application of false colour to indicate height information. 2D and 3D roughness values were determined using the appropriate Standards (ISO 4287, ISO 4288 and ISO 25718). The adhesion of the coating to the substrate was determined for each coating condition using a modified ASTM C633 test. A minimum of six specimens were bonded per coating condition (using cold cure adhesive, Araldite Rapid) and subjected to tensile loading at a rate of 1mm.min-1 with the applied load at failure recorded. The load was then converted to a bond strength in MPa. The fracture surfaces were also assessed by visual and/or microscopic inspection and the location of failure recorded. Wear resistance tests were performed using a 10mm diameter alumina (Al2O3) ball with an applied normal force of 10N, a stroke length of 10mm, a frequency of 4Hz and a total sliding distance of 500m (3h28m20s test duration). No lubrication was used. Three tracks were applied per coating condition. In cases where multiple tracks were applied to a single coupon, it was ensured that a distance of at least 8mm was left between tracks. For each wear track, the coefficient of friction against the alumina ball was measured over time, and light micrographs were also taken of the alumina counterpart ball. The wear tracks were qualitatively examined by confocal light microscopy and by stylus to determine 2D and 3D maps across/along the track. A series of erosion resistance tests were carried out on coupons from each condition, generally following the GE Specification E50TF121 (Class B) which uses mass loss as a measure of erosion performance. A modified specimen size was used (40x40mm) in agreement with the Topic Manager. Relative erosivity values (dimensionless) were calculated (Table 5) using the following equation and images of the erosion scars recorded. In general, the erosion resistance of a sample is inversely correlated to the erosivity so that a higher erosivity indicates lower erosion resistance

WP3: Sub-element demonstrator
Task 3.1: Part and Tool definition, design and manufacture
The design and manufacturing steps of the sub-element demonstrator are show in Figure 7.

Task 3.2: Sub-element erosion, impact resistance and fatigue testing, and post failure investigations
A series of 40x40mm samples were machined from selected SEDs from three regions:
• Coated CFRP area,
• Coated Ti alloy area
• Coated interface between the Ti and CFRP.
A series of erosion resistance tests were carried out on the extracted coupons, generally following the GE Specification E50TF121 (Class B) which uses mass loss as a measure of erosion performance. A modified specimen size was used (40x40mm) in agreement with the TM. Samples were degreased and weighed (to an accuracy of 0.0001g) and mounted in the rig at 20° relative to the blasting nozzle with a stand-off distance of 4 inches (101.6mm). The samples were then subjected to an erodent of 600g of white alumina blast media (240 mesh grit) with a blast pressure of 0.2MPa. The grit was consistently consumed within 135±2 seconds. Each test was terminated after 150 seconds. The specimens were extracted, cleaned with an air jet and then weighed again to an accuracy of 0.0001g. Polycarbonate (Lexan) samples were tested for calibration purposes as required in the Specification. These exhibited a relatively high mass loss, in the 0.13-0.41g range. Results are presented in Table 6
Impact Resistance
Due to the complex geometry and inhomogeneity of the SEDs, no impact testing standard was directly applicable in this instance, but a test method generally following ASTM D2794 was agreed with the TM. A series of drop-weight impacts were made on the uncoated SED TWI/20870/5 D1. These impacts, using a ¾” diameter hemispherical indenter, were located mid-width on the CFRP at impact energies varying from 4.41J to 17.64J. Based on these results, an impact energy of 17.64J was selected as it resulted in a 0.3mm indent in the CFRP surface of the uncoated SED. Several other coated SEDs were then subjected to impact testing at this energy, on both the coated CFRP and Ti alloy surfaces, again mid-width. This test regime also included multiple impacts on a single location. Focus variation microscopy was used to create 3D profiles of typical impact craters and measure the total impact depth relative to the surface, on both the Ti and CFRP substrates. Several examples of impact craters were also cross-sectioned and examined using the methods detailed above. Representative examples of impacted SEDs are shown in Figure 8.
Mechanical performance
A series of 100x15mm specimens were prepared across the Ti-CFRP interface from two SEDs for 4-point bending tests. As no test standard was directly applicable to these multi-material specimens, a test method was agreed with the Topic Manager. Varying specimen displacements were used but specimens were not taken to failure of the Ti alloy. Figure 9 shows the experimental setup and positioning of the SED in the test. The coated surface was facing downwards. The test specimens were photographed in the fully loaded/displaced condition and then unloaded and visually examined as seen in Figure 10. Cross-sections were prepared from those specimens and examined under microscope as shown in Figure 11

WP4: Manufacturing of the component demonstrator
Task 4.1: Manufacturing of the component demonstrator
The component demonstrator design is shown in Figure 12. Three component demonstrators were manufactured. One was delivered to the Topic Manager.

Task 4.2: Application of surface protection system to component demonstrator
The methodology used for the application of the surface protection in the SEDs was also followed in the case of the component demonstrator. Minor modifications that have been made in order to tailor the application to the component demonstrator are summarised below. Some suggestions for further development and optimisation of the process are also provided:
• The development work in was carried out on Renegade CFRP MVK-14 with a glass fibre layer at the surface, whereas the CFRP prepreg material consisted only of the MVK-14 prepreg. This led to a different surface finish and lower adhesion between TWAS Al bond coat and the CFRP substrate. This can potentially be mitigated by inclusion of a glass fibre layer in future components.
• The use of a slightly cooler HVOF spray plume was required to improve coating quality. This was accomplished by increasing kerosene flow through the system for smaller coupons, but this was not feasible for SEDs as it led to unstable combustion and texturing of the surface. Further refinement of spray parameters may be required to reduce spray temperature whilst maintaining coating quality (e.g. standoff distance, powder feed rate, oxygen flow rate etc.)
• Steps or gaps were present at the Ti-CFRP interface in the component demonstrator parts. As a result, the coating processes follow these surface features and they are visible post-processing. This could be mitigated by improved prepreg curing and component design to better match the ply thickness, particularly for tapered areas, and improve overall coating performance in service.
• HVOF WCCoCr would normally be applied directly to Ti alloy without requiring a bond coat; this system has extremely high adhesive strength. However, due to the dissimilar CFRP-Ti joint, the use of a TWAS Al bond coat across the entire component was necessary to mitigate the effect of the discontinuous Ti-CFRP interface. However this results in a lower bond strength to the Ti alloy compared with not using a bond coat.

Potential Impact:
A number of technologies have been developed and applied to specimens and demonstrator parts, representative to aero-engine components. These technologies are: i) the sculpting of Ti64 using electron beam in order to produce protrusions that will enhance the joint between CFRP and Ti, ii) the bonding of aerospace grade CFRP to the Ti64 using a number of surface pre-treatments, iii) the laser welding of the Ti64 in close proximity of the composite and iv) the application of a surface protection system to both the composite and the metal.

The major impact from adopting the above technologies is weight savings of up to 90% (as calculated for the demonstrator parts) for composite-Ti64 aerofoils compared to an all Ti64 aerofoil. This could result in significant gains in the aero-engine efficiency and overall weight of the aircraft.

List of Websites:
The project had no website.