Paper presented at International Thermal Spray Conference 2-4 May 2005, Basel, Switzerland.
The HVOF spraying process has recently been considered for the deposition of dense alumina coatings for dielectric coatings in semiconductor applications and as a turbine blade tip coating in aeroengines. However, due to the lower flame temperature of the HVOF process compared to the plasma spray process it is necessary to have good control over the key process parameters to achieve the correct coating characteristics. The work reported presents the results of a design of experiment study carried out on a TopGun HVOF system used to prepare coatings of alumina. The influence of several key parameters on coating characteristics such as porosity, alumina phase type, microhardness, surface roughness and adhesion have been determined. The parameters varied included oxygen and hydrogen fuel gas flow rates, and spray distance. Based on the results of these investigations recommendations are made on the control of key parameters and the range of coating characteristics that can be expected.
1. Introduction
A number of single and mixed oxide ceramic compounds, including alumina, titania, alumina-titania, chromia and zirconia, can be deposited by thermal spraying processes.[1] Thermal sprayed alumina offers good wear resistance and a high hardness which is retained at elevated temperatures. Alumina is used in various wear applications, particularly where low stress abrasion is encountered. It also has a high dielectric strength at room temperature and is commonly used to provide an electrical insulation barrier. Other applications include turbine blade tip coatings in aeroengine applications and as sliding and abrasive wear resistant coatings for pump and machinery components within the petrochemical, pulp and paper and textile industries.
Plasma spraying is commonly used to deposit oxide ceramic coatings. However, HVOF spraying is also used for some oxide ceramic coatings to achieve improvements in wear properties.[2-3] In this study a design of experiment approach has been used to investigate the spraying parameter settings employed to deposit alumina coatings using the Top Gun HVOF system. Due to the lower flame temperature of the HVOF process compared to the plasma spraying process it is necessary to have good control over the process parameters to achieve the correct coating characteristics.
The influence of oxygen and hydrogen fuel gas flow rates, and spray distance on coating characteristics, such as porosity, alumina phase type, microhardness, surface roughness and adhesion to a steel substrate have been determined. Based on the results of these investigations recommendations are made on the control of key parameters and the range of coating characteristics that can be expected.
2. Coating preparation
Alumina coatings were prepared on carbon manganese steel (S355 J2G3) using the TopGun HVOF system. Coatings were prepared with AI-1110-HP powder obtained from Praxair Surface Technolgies Inc. with a nominal size of 5-22µm. The surface of the substrates were prepared by grit blasting with alumina grit of mesh size 60 and then air washed and degreased immediately prior to coating.
The characterisation and adhesion test pieces were mounted on the perimeter of a turntable. The gun axis was aligned perpendicular to and directed at the turntable axis of rotation which was set to the vertical. The rotational speed of the turntable was set such that the traverse speed of the test pieces was 1ms-1. The gun was moved up and down in the vertical direction at 2.5mms-1 until 32 passes had been completed to deposit a coating nominally 300µm thick.
Table 1. Spraying parameters
Spray run ID | A Spray distance mm | B Hydrogen fuel flow rate scfh | C Oxygen flow rate scfh | O2/H2 ratio |
03-87 |
152 |
1530 |
620 |
0.41 |
03-88 |
152 |
1530 |
540 |
0.35 |
03-89 |
229 |
1766 |
540 |
0.31 |
03-90 |
229 |
1530 |
620 |
0.41 |
03-91 |
152 |
1766 |
540 |
0.31 |
03-92 |
229 |
1530 |
540 |
0.35 |
03-93 |
229 |
1766 |
620 |
0.35 |
03-94 |
191 |
1648 |
580 |
0.35 |
03-95 |
152 |
1766 |
620 |
0.35 |
03-96 |
191 |
1648 |
580 |
0.35 |
3. Experimental procedure
A two level factorial experimental design with two additional centre points was used to investigate the role of three parameters. The three parameters examined were spray distance, hydrogen flow rate and oxygen flow rate. The design of experiment trials involved the preparation and characterisation of 10 coatings. The spraying conditions given in Table 1. The lower setting for spray distance (A-) was 152mm and the upper setting (A+) was 229mm. Likewise for hydrogen flow B- was 1530scfh and B+ was 1766scfh, and for oxygen flow C- was 540scfh and C+ 620scfh. The coatings were then examined to determine the microstructure, deposition characteristcs, surface roughness, microhardness and coating adhesion. These measurements provided the response data for the design of experiment method.
Cross sections of the sprayed coatings were prepared using standard metallographic techniques. The coating thickness was then measured and scanning electron microscopy images taken. The Vickers microhardness was measured on the coating sections using a Duramin hardness tester manufactured by Struers and a load of 300g. 10 indents were made, and the hardness value expressed as a mean value with standard deviation.
The surface roughness of the deposited coatings was measured using a stylus profilometry technique with a Surfcom 300B equipment having a 5µm diamond stylus tip and a 4mN force. Values are quoted as a roughness average (Ra) in µm.
Coating adhesion was measured according to the ASTM standard C633-01. It consists of coating one face of a loading fixture and bonding this coating to the face of an uncoated loading fixture with a high strength structural adhesive (trade name FM1000). The assembly was placed in a tensile loading machine with self aligning devices. The tensile load was increased at 1mm/min and the load at failure recorded. For each coating type, the adhesion of five test pieces was measured and expressed as a mean value with standard deviation.
The crystalline phases present in the powders prior to spraying and in the deposited coatings were determined using x-ray diffraction. Patterns were collected using a Philips PW 1410 x-ray diffractometer employing a CuKα radiation source, between the angles of 20°-60° 2θ. Measured peak positions were matched to standard values on the JCPDS powder diffraction file to identify the crystalline phases present. The relative proportion of α-Al2O3 in the coatings was estimated using the direct comparison method, using the diffraction peak heights of the 113 plane for α-alumina and the 400 plane for γ-alumina.[4]
The deposition rate was measured in terms of thickness deposited per pass and weight deposited per unit time. The deposited coating thickness was determined from measurements made on cross sections of the coatings. To give deposition rate in terms of thickness per pass, the coating thickness was divided by the total number of passes needed to build up the coating.
The two characterisation test pieces were also weighed before and after being coated to determine the weight of coating deposited. Deposited weight was divided by the total time the spray jet was impinging on the characterisation test piece surface. This gave a deposition rate expressed in weight deposited per unit time. The amount of powder impinging on the characterisation test piece per unit time was also calculated. The actual measured weight of coating deposited onto the test piece was divided by this calculated weight of impinging powder to give a value for deposition efficiency.
4. Results and discussion
The measured coating characteristics are given in Table 2. This table shows that coating thickness varied from 258 to 342µm with a target coating thickness of 300µm. The deposition rate of alumina varied between 4.5 and 6.5g/min, with a deposition efficiency of between 47 and 68%.
The surface roughness of the as-sprayed coatings were all relatively low from 1.27 µmRa up to a high of 1.93 µmRa. The adhesion of the coatings ranged from 14MPa up to 38MPa with failure occurring at the substrate coating interface.
The microhardness measured on all the coating cross sections had values ranging from 856 to 1212HV. Back scattered SEM images of the coating cross sections for coatings with the smallest and largest microhardness values are shown in Fig.1. These cross sections show a variation in coating porosity, where the low mircohardness coatings exhibit the greatest amounts of porosity.
Table 2. Coating characteristics
Sample ID | Thickness, µm | Vickers microhardness (300g) | Adhesion, MPa | Surface roughness, µm Ra | Coating thickness per pass, µm | % α-Al2O3 | Deposit rate, g/min | Deposit efficiency, % |
03-87 |
342 |
1140 (±47) |
30 (±5) |
1.27 |
10.7 |
10 |
5.2 |
55 |
03-88 |
284 |
1108 (±56) |
36 (±4) |
1.43 |
8.9 |
13 |
4.7 |
50 |
03-89 |
271 |
950 (±48) |
17 (±10) |
1.57 |
8.5 |
29 |
5.7 |
60 |
03-90 |
323 |
856 (±68) |
19 (±3) |
1.86 |
10.1 |
14 |
6.1 |
64 |
03-91 |
258 |
1060 (±45) |
38 (±5) |
1.43 |
8.1 |
22 |
4.5 |
47 |
03-92 |
310 |
914 (±34) |
14 (±2) |
1.93 |
9.7 |
22 |
5.2 |
55 |
03-93 |
290 |
924 (±26) |
17 (±3) |
1.54 |
12.1 |
10 |
6.5 |
68 |
03-94 |
323 |
1020 (±60) |
28 (±3) |
1.39 |
10.1 |
12 |
5.2 |
58 |
03-95 |
290 |
1212 (±72) |
37 (±24) |
1.35 |
9.1 |
22 |
4.9 |
51 |
03-96 |
290 |
1078 (±35) |
20 (±3) |
1.36 |
9.7 |
14 |
5.5 |
58 |
Fig.1. Backscattered SEM images of a) low microhardness (03-92), and
b) high microhardness alumina coatings (03-95) showing coatings with low microhardness exhibit greater amounts of porosity within the coating.
The AI-1110-HP alumina powder consisted only of α-Al
2O
3. The sprayed alumina coatings were found to be a mixture of γ-alumina and a-alumina. The amount of α-alumina remaining in the coating was dependant on the spraying conditions but ranged from 10% up to 29% as shown in
Table 2. The collected XRD plots for alumina coatings containing 10% and 29% α-alumina are given in
Fig.2.
The effect of each parameter on the measured responses for deposition rate, deposition efficiency, surface roughness, adhesion, microhardness and α-alumina content are summarised in Table 3. This indicates which main factors (hydrogen flow rate, oxygen flow rate and spray distance) and which two-way interactions, if any, have an influence on each response.
Fig.2. XRD traces for an alumina coatings containing
a) 10% α-Al2O3, indicating the greatest amount of particle melting during the spraying process, and
b) 29% α-Al2O3, indicating less melting of the alumina particles during spraying.
The analysis indicates that deposition rate was influenced by spray distance and oxygen flow rate. The coefficient estimate is the average response value for all 10 trials, while the parameter influence value is the effect of parameter setting on the average value. For example, when spray distance and oxygen flow are at the higher setting the deposit rate is given by:
DR (g/min) |
= 5.34 + A + C |
|
= 6.16 |
and when spray distance and oxygen flow rate are at the lower setting:
DR (g/min) |
= 5.34 - A - C |
|
= 4.52 |
Similarly the influence of each parameter can be determined for a particular response.
Deposition efficiency was influenced by spray distance, hydrogen flow rate and oxygen flow rate, together with two-way interactions between spray distance and oxygen flow rate. The magnitude and direction of influence of spray distance and oxygen flow rate on the deposit efficiency is shown in Fig.3.
Surface roughness was also influenced by spray distance, hydrogen flow rate and oxygen flow rate, together with a two-way interaction between spray distance and oxygen flow rate. Coating adhesion and coating microhardness were influenced only by spray distance.
The α-alumina content in the coating was influenced by hydrogen and oxygen flow rate rates, and the effect of oxygen flow rate on α-alumina content is shown in Fig.3.
Fig.3. Effects of spray distance and oxygen flow rate on deposit efficiency, microhardness and α-Al2O3 content of the coatings.
These results indicate that when preparing alumina coatings by HVOF thermal spraying it is important to maintain good control of the spray distance and oxygen flow rate in particular. High adhesion and high hardness coatings are best prepared using the shorter spraying distance of 152mm. These properties are strongly influenced by variation in spray distance but are fairly insensitive to variation (for the settings considered in this work) in the setting for hydrogen flow rate and oxygen flow rate. Using the information in Table 3, a change in spray distance of only 39mm from 152mm to 191mm will cause an estimated drop in hardness of about 10% and a drop in adhesion of about 35%.
The shorter spray distance was found to cause a drop off in deposition efficiency. Table 3 indicates that deposition efficiency is also influenced by oxygen flow rate, with the higher oxygen flow rate giving an increase in deposition efficiency. By using a higher oxygen flow rate with the shorter spray distance a good combination of high hardness, high coating adhesion and good deposition efficiency should be possible. The results also show that hydrogen flow rate has quite a strong influence on surface roughness, with the higher flow rate giving the least rough surface. Based on these considerations the recommended settings for spraying an alumina coating are:
Spray distance: |
152 mm |
Hydrogen flow rate |
1766 scfh |
Oxygen flow rate |
620 scfh |
These settings are expected to give a coating having a deposit efficency of 53%, hardness = 1131 HV, α-Al2O3 content = 17% and adhesion of 36MPa.
5. Conclusions
The hardness and adhesion values of the alumina coating are strongly influenced by the distance between the spray gun and substrate surface, with the shortest distance considered in this work (150mm) giving the highest values. A recommended set of conditions for the hydrogen and oxygen flow rates, and spraying distance have been proposed based on a design of experiment approach.
6. Acknowledgements
The authors would like to acknowledge TWI staff who contributed to this work and the project partners in the European CRAFT project: CRAF - 1999-70297, funded by the European Community under the Competitive and Sustainable Growth Programme (1998-2000).
Table 3. Summary of ANOVA results
Response | Coefficient estimate | Parameter influence |
Main factors | Two way interactions |
A Spray distance | B Hydrogen flow | C Oxygen flow | AxB | AxC | BxC |
Deposit Rate g/min |
5.34 |
+0.51 (0.0004) |
X |
+0.31 (0.0029) |
X |
X |
X |
Deposit Efficiency % |
56.2 |
+5.4 (<0.0001) |
X |
+3.23 (0.0001) |
X |
+0.98 (0.0019) |
X |
Roughness µm Ra |
1.55 |
+0.18 (0.0001) |
-0.075 (0.0041) |
-0.044 (0.0260) |
-0.096 (0.0016) |
X |
X |
Adhesion MPa |
26.1 |
-9.35 (0.0008) |
X |
X |
X |
X |
X |
Micro-hardness HV0.3 |
1021 |
-109.5 (0.0025) |
X |
X |
X |
X |
X |
α-Al2O3 Content % |
17.8 |
X |
+2.89 (0.1859) |
-3.77 (0.1020) |
X |
X |
X |
Values in brackets correspond to Probability Equation.1 non significant |