M D F Harvey (TWI, UK), O Lunder (Sintef Materials Technology, Norway) and R Henriksen (Bandak AS, Norway).
Paper presented at International Thermal Spraying Conference, Montreal, Canada, 8-11 May 2000
Coatings have been prepared using the Diamond Jet hybrid and JP5000 high velocity oxyfuel (HVOF) systems with the objectives of improving corrosion resistance and reducing costs through increasing deposition efficiency. Models relating deposition efficiency, coating oxygen content and corrosion resistance to process parameters including fuel flow rate, oxygen flow rate and stand-off distance have been developed. A corrosion test cell has been designed and a procedure determined for studying the corrosion behaviour of large numbers of thermally sprayed coatings in an efficient manner. A significant improvement to the corrosion resistance of HVOF sprayed coatings has been achieved by spraying parameter optimisation and investigation of powder size and distribution. The project has also investigated the influence of spray angle on coating performance with a view to future on-site application. Coating materials tested and compared include nickel alloys Hastelloy C276 and 59, cobalt alloy Ultimet, duplex stainless steel S32750 and an experimental iron-based spray-fuse composition.
Corrosion resistant coatings for engineering components have traditionally been deposited by welding or electroplating, but applications have been restricted by metallurgical incompatibility between the overlay and the substrate and concerns over the environmental impact of toxic solvents and reactants respectively. To date, the porous nature of thermally sprayed coatings has led to skepticism amongst end-users regarding their suitability for corrosive environments. The main objective of the project was to establish HVOF spraying as a technology for deposition of corrosion resistant coatings.
The main problems associated with the use of corrosion resistant alloys are localized corrosion including pitting, crevice corrosion, stress corrosion cracking or galvanic coupling to less corrosion resistant alloys. Galvanic coupling is of particular importance when a corrosion resistant coating is sprayed onto a low alloy carbon steel, the substrate material used in this project. Penetration by the corrosive medium to the substrate leads to rapid corrosion along the coating to substrate interface and subsequent detachment of the coating. Hence, low porosity and high pitting resistance of the coating itself are critical factors in preventing this from occurring.
Relationship between spraying parameters and deposition efficiency, coating oxygen content and microstructure
The aim of this part of the project was to develop an understanding of the relationship between the spraying parameters, coating microstructure and corrosion behaviour. At the same time it was anticipated that these might be related to two quantifiable responses, the deposit efficiency and coating oxygen content. Two HVOF systems, the propylene-fuelled Diamond Jet hybrid (DJ2700) and the kerosene-fuelled JP5000, were chosen. Two corrosion resistant alloys, nickel alloy Hastelloy C276 and duplex steel S32750, were used, Table 1.
An initial study of nine parameters indicated that, for each HVOF system, the deposit efficiency and coating oxygen content were dependent on four main parameters:
- - JP5000: fuel flow rate, oxygen flow rate, stand off distance, carrier gas flow rate
- - Diamond Jet hybrid: fuel flow rate, oxygen flow rate, air cooling (gun) flow rate, stand off distance.
For both HVOF systems, the four most influential parameters were investigated further using a half-factorial experimental design, based around the respective manufacturer's recommended parameters. Following analysis of the data relationships were estabished between both the deposit efficiency and coating oxygen content the oxygen to fuel stoichiometric ratio (SR) for any given fuel flow rate, Figs 1-7. The labels on the curves indicate the fuel flow rates in l.min -1 (Diamond Jet) or l.hour -1 (JP5000).
For the DJ2700 (propylene-fuelled) HVOF system, increasing the stoichiometric ratio (SR) from 45 to 65%, increases the deposit efficiency significantly for both Duplex S32750 and Hastelloy C276 powders, Figs 1-2.
Fig. 1. Relationship between deposit efficiency and oxyfuel stoichiometric ratio for fixed fuel flow rates
Fig. 2. Relationship between deposit efficiency and oxyfuel stoichiometric ratio for fixed fuel flow rates
For the Diamond Jet system, it should be noted that the SR is the combined total of the oxygen supplied for combustion and the oxygen fraction from the air cooling. The maximum deposit efficiency occurs for both powders at an SR of around 65%, and increasing the SR above 65% leads to a reduction in the deposit efficiency. The reason for this is not clear, since for a fixed fuel flow rate, increasing the SR from 65% to 100% increases the total thermal energy available, increases the total gas mass flow (and presumably the particle velocity), and passes through the maximum flame temperature, reported to be at an SR of 80-85%. 
It was noted that the Diamond Jet hybrid gun design limits the total gas flow (i.e. fuel + oxygen + air cooling), and that it is not possible, for example, to combine a high fuel flow (i.e. 108 l.min -1) with a high oxygen flow. The SR is limited to 55% for a fuel flow rate of 108 l.min -1. The combination of a limit to the total gas flow rate and decline in deposit efficiency at higher SRs defines a deposit efficiency envelope for each particular powder alloy and powder size distribution, Fig 3.
Fig. 3. Diamond Jet hybrid - deposit efficiency envelopes for Hastelloy C276 and Duplex S32750
The JP5000 HVOF system manufacturer (Tafa Inc) recommends parameters that operate in an oxygen rich regime with an SR in excess of 100%, Fig 4. The deposit efficiency is fairly insensitive to a simultaneous variation in fuel flow rate (20.9 - 25.1 l.hour -1) and SR (100-130%), but clearly falls away with a simultaneous low fuel flow rate (17.8 l.hour -1) and a higher SR (130-150%).
Fig. 4. JP5000: Relationship between deposit efficiency and oxyfuel stoichiometric ratio for fixed fuel flow rates
For the DJ2700 (propylene fuelled) HVOF system, increasing the SR increases the coating oxygen content significantly for both Duplex S32750 and Hastelloy C276 powders, Figs 5-6. This might be expected, since both the flame oxygen content and flame temperature are increasing, and this will increase the rate of the chemical reaction (i.e. oxidation) that will occur between the metal alloy powders and the oxygen.
Fig. 5. Relationship between deposit efficiency and oxyfuel stoichiometric ratio for fixed fuel flow rates
Fig. 6. Relationship between deposit efficiency and oxyfuel stoichiometric ratio for fixed fuel flow rates
It is also observed that increasing the fuel flow at a fixed SR has the effect of significantly increasing oxidation, Fig 6, presumably because there is an increase in the number of collisions between gas particles and powder particles, and hence opportunities for the oxidation reaction to occur.
Fig. 7. JP5000: Relationship between coating oxygen content and oxyfuel stoichiometric ratio for fixed fuel flow rates
In contrast to the Diamond Jet hybrid coatings, JP5000 coating oxygen content falls with increasing SR, Fig 7. At first this appears counter intuitive, however, the increasing level of excess oxygen in the JP5000 flame as the SR increases above 100% primarily serves to cool the flame. It appears that the fall in flame temperature has a greater influence on the rate of reaction (oxidation) than the increase of oxygen in the flame.
Examination of the microstructures of the various coatings produced by both the Diamond Jet and JP5000 HVOF systems indicates that there is a consistent correlation between coatings characterised by denser microstructures with those of higher oxygen content, and vice versa, Figs 8-9.
Fig. 8. Diamond Jet hybrid - low oxygen content coating (left), high oxygen content coating (right)
Fig. 9. JP5000 - low oxygen content coating (left), high oxygen content coating (right)
Corrosion testing procedure
A corrosion test cell was designed where a crevice area, susceptible to initiation of localized corrosion, is formed on the coating surface while the bulk substrate is not in contact with the test solution. The cell is made from Pyrex glass and placed on a magnetic stirrer and heating unit, allowing constant temperature control of the electrolyte during the experiment. A ferric chloride solution, used in standard testing of stainless steels and nickel alloys(ASTM G48), was initially chosen as the test medium. A deposited coating thickness of approximately 400µm, machine finished to a thickness of 350µm was used throughout this project. An initial comparison was made between Hastelloy C276 and duplex stainless steel S32750 materials in rolled alloy form, weld overlay and HVOF coatings.
The test cell and crevice assembly was simple to set up and operate, and was reliable in terms of avoiding leaks and reproducibility of results. The screening test procedure provided a ranking of the corrosion resistance of different materials. Rolled Hastelloy C276 and duplex stainless steel S32750 exhibited a higher resistance against initiation of crevice corrosion than weld overlays of the same composition. The HVOF coatings exhibited significantly lower corrosion resistance than the weld overlay reference materials. However, HVOF sprayed Hastelloy C276 performed significantly better than alloy 625 and duplex stainless steel S32750.
Fig. 10. Corrosion test cell design
Under the screening test conditions Hastelloy C276 coatings exhibited significantly higher corrosion resistance than the Duplex S32750 coatings. For a given powder, coatings sprayed by the Diamond Jet hybrid (DJ2700 propylene) andJP5000 systems behaved similarly. However, the HVOF sprayed coatings were still significantly less corrosion resistant than the reference weld overlays. Furthermore, the screening test did not discriminate between coatings of the same composition, sprayed by the different systems or with different parameters.
Subsequent corrosion tests were carried out using the same test cell, but the anodic current density was measured at an applied potential of 0.3 V SCE (saturated calomel electrode) as a function of time in synthetic seawater (ASTMD1141-90), maintained at a temperature of 25°C. Current densities were normalised with respect to the crevice area, although the true area affected by localised corrosion (both pitting and crevice corrosion) varied significantly. The average current density over a 20 hour period was recorded as a measure of the corrosion rate of the coated samples.
A relationship between the HVOF coating oxygen content and the rate of potential decay was observed, and the coatings with the lowest oxygen content were not necessarily the best in terms of corrosion resistance, Figs 11-12. The labels on the curves indicate oxygen level (wt%) in coating. It should be noted that the corrosion rate of Duplex S32750 is much greater than the Hastelloy C276 coatings, as indicated by the more rapid potential decay for Duplex S32750.
Fig. 11. Diamond Jet hybrid Hastelloy C276 coatings - corrosion potentials vs. time in synthetic seawater at 25°C
Fig. 12. JP5000 Duplex S32750 coatings - corrosion potentials vs. time in synthetic seawater at 25°C
Investigation of other alloys, powder size distribution, hydrogen fuel gas and spray angle
The scope of the program was widened to consider:
- - Alloy composition. Three new alloys were introduced, reflecting alternative commercial spray powder compositions, i.e. Alloy 59, Ultimet and an experimental powder AE7228, Table 1.
- - The influence of powder size distribution with the removal of powder fines.
- - The use of hydrogen fuel gas with the Diamond Jet hybrid (DJ2600 system).
- - The influence of spray angle.
For all four materials, the Diamond Jet hybrid process produced the highest deposit efficiency, although in two cases (both nickel alloys) it was the DJ2600 (hydrogen) and in two cases the DJ2700 (propylene) system which produced the highest deposit efficiency, Fig 13. The nickel alloy powders (Hastelloy C276 and alloy 59) produce coatings with a consistently higher oxygen content, although this is not HVOF system dependent, Fig 14.
Table 1: Spray powder size and nominal composition
|Powder||Powder size (µm)||Fe||Ni||Cr||Co||Mo||W||Mn||Si||N||C|
||15 - 45
||20 - 53
||22 - 53
||22 - 53
||22 - 53
Fig. 13. Comparison of deposit efficiency for powders sprayed in the as-received condition
Fig. 14. Comparison of coating oxygen for powders sprayed in the as-received condition
The removal of powder fines generally resulted in an increase in the deposit efficiency, and a reduction in the coating oxygen content, Figs 15-16.
Using a spray angle of up to 30° from the normal still produced a visually acceptable microstructure. For Hastelloy C276 and AE7228, there was a drop in deposit efficiency of only a few percent, Fig 17. A more marked drop in deposit efficiency of up to 10% was observed for Ultimet and alloy 59 coatings, Fig 18. The drop in deposit efficiency was accompanied by an increase in coating oxygen content, Fig 19. It can be inferred that this is the result of larger particles rebounding off the substrate.
Fig. 15. Comparison of deposit efficiency for powders sprayed in the as-received condition and with fines removed
Fig. 16. Comparison of coating oxygen content for powders sprayed in the as-received condition and with fines removed
Fig. 17. Deposit efficiency against spray angle for Hastelloy C276 and AE7228
Fig. 18. Deposit efficiency against spray angle for Alloy 59 and Ultimet
Fig. 19. Coating oxygen content against spray angle for Alloy 59 and Ultimet
Corrosion test results
The potentiostatic polarization tests in substitute seawater showed that the corrosion resistance of the HVOF sprayed Hastelloy C276 coatings were superior to Alloy 59, Ultimet and AE7228 (and in that order), Fig 20. The best corrosion resistance (lowest current density) was generally observed for HVOF coatings sprayed with the DJ2600 (hydrogen fuel) system. However, the removal of powder fines generally improved the corrosion resistance, and by narrowing the Hastelloy C276 powder size distribution corrosion resistance approaching that of the Hastelloy C276/DJ2600 (hydrogen fuel) coatings was observed with the JP5000, Fig 21.
Fig. 20. Average current density of coatings during 20 hour exposure in synthetic seawater at an applied potential of 0.3 V SCE. (Note logarithmic scale)
Fig. 21. Average current density of Hastelloy C276 coatings during 20 hour exposure in synthetic seawater at an applied potential of 0.3 V SCE
No significant effect of spraying angle (up to 30° from the normal) on the corrosion resistance was observed for any of the powder and HVOF system combinations, suggesting that the corrosion resistance obtained in the workshop is transferable to manual and on-site application, Fig 22-23
Fig. 22. Corrosion rate (expressed as anodic current density) of HVOF sprayed coatings at an applied potential of 0.3 V SCE in synthetic seawater at 25°C. (Note logarithmic scale)
Fig. 23. Corrosion rate (expressed as anodic current density) of HVOF sprayed coatings at an applied potential of 0.3 V SCE in synthetic seawater at 25°C
After corrosion testing, no evidence of corrosion was seen on the surface of the coatings demonstrating the lowest corrosion rate (i.e. the lowest average current density). Weld overlays of nickel alloys have a typical average current density of 2 - 5µA.cm-2, compared to the best HVOF coatings with values of around 10µA.cm-2. It should be noted that the typical HVOF coating thickness tested in this project was 350µm, compared to 3mm for the weld overlay, and it is possible that thicker HVOF coatings may have improved corrosion resistance.
- The key parameters influencing the deposit efficiency and the oxygen content of coatings produced by both systems are oxygen flow rate, fuel flow rate and the related oxygen to fuel stoichiometric ratio (SR).
- The Diamond Jet hybrid (propylene fuelled) system:
- operates with fuel rich parameters (SR < 100%)
- has a maximum deposit efficiency (that appears to be limited by total gas flow rate) at an SR of about 65%
- The JP5000 HVOF system: - operates with oxygen rich parameters (SR > 100%)
- has a maximum deposit efficiency at an SR between 100-130%
- For both HVOF systems, coatings characterised by denser microstructures and higher coating oxygen content demonstrate superior corrosion resistance.
- For both the Diamond Jet hybrid (including both fuel gas variants - propylene and hydrogen) and JP5000 systems there is a consistent ranking of the corrosion resistance of the four alloys tested:
- Hastelloy C276
- Alloy 59
- The best corrosion resistance is achieved using the Diamond Jet hybrid system with hydrogen fuel gas and as-received powder, and with the JP5000 with powder fines removed
- For both the Diamond Jet hybrid (including both fuel gas variants - propylene and hydrogen) and JP5000 systems varying the spray angle from the normal by a trailing or leading angle of 30°, has no statistically significant impact on the corrosion rate of any of the four alloys tested. For two materials, a 30° spray angle results in a lower deposit efficiency, and higher coating oxygen content, indicating that larger spray particles rebound off the substrate first with increasing spray angle.
A consortium of eight companies contributed to this project, which was funded by the European Union through the Co-operative Research (CRAFT) scheme as part of the Framework IV programme. The purpose of the CRAFT scheme is to develop technologies adapted to the needs of small to medium sized enterprises (SMEs) through a transnational network of SMEs, research organisations and large companies. The consortium comprised the following companies:
- Bandak AS (thermal spraying)
- Midsund Bruk (weld overlays)
- Brodrene Johnsen (thermal spraying)
- Sintef Materials Technology (corrosion testing development)
- Provacuum (thermal spraying)
- Lucas-Milhaupt (powder development)
- ICI Engineering (corrosion testing development)
- TWI (thermal spraying process development).
- A D Hewitt: Technology of oxy-fuel gas processes, Part 2: Comparative combustion properties of fuel gases, Welding and Metal Fabrication, September 1972.