Microstructure of plasma nitrided layers on aluminium

Microstructure of plasma nitrided layers on aluminium

Surface and Coatings Technology 156 (2002) 149–154 Microstructure of plasma nitrided layers on aluminium Roman Sonnleitnera,*, Krystyna Spiradek-Hahn...

1MB Sizes 0 Downloads 1 Views

Surface and Coatings Technology 156 (2002) 149–154

Microstructure of plasma nitrided layers on aluminium Roman Sonnleitnera,*, Krystyna Spiradek-Hahna, Francois Rossib a

Austrian Research Centers Seibersdorf, A-2444 Seibersdorf, Austria b Joint Research Centre, I-21020 Ispra, Italy

Abstract Plasma immersion ion implantation (PI3 ) treatment was performed on aluminium for surface nitriding. Glow discharge optical spectroscopy (GDOS) was used to estimate the penetration depth of nitrogen. Transmission electron microscopy (TEM) was used to reveal the structure and composition of nitrided layers and that of the interface. The microstructural analysis was performed using thin foils from the cross-sections of nitrided samples. The crystallographic phases were identified by electron diffraction. It is shown that the formation of AlN is strongly correlated to the plasma treatment parameters. The structure of the surface layers and the substrate material beneath the interface depends on the plasma treatment conditions: treatment time, energy and accelerating voltage. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: PI3 treatment; TEM investigations; Nitrided layers; Aluminium

1. Introduction Aluminium is one of the favourite materials for the automotive industry because the low specific weight of aluminium and its alloys allows a reduction in the weight of cars. When replacing steel engine components, such as cylinder liners or pistons by aluminium, two disadvantages have to be taken into consideration: low load bearing capacities and bad friction behaviour. In order to overcome these properties, hard coatings have been put directly on the relatively weak aluminium surface. The large difference in the hardness between aluminium and the coating leads to adhesion problems for these hard coatings on the weak aluminium surface. Hence, it is necessary to strengthen the aluminium surface, which can be done by plasma modification of the surface, for instance, by plasma nitriding w1–3x. To strengthen the surface of aluminium, plasma immersion ion implantation experiments (PI3 ) have been carried out. The advantage of this technique is that it allows surface treatments at low temperatures, which is essential to retain the mechanical properties of most aluminium alloys. At process temperatures of more than 200 8C, most available Al-alloys show a significant decrease in their mechanical properties because of changes in their microstructure. To determine the best *Corresponding author. Tel.: q43-50-550-3330; fax: q43-50-5503366. E-mail address: [email protected] (R. Sonnleitner).

process parameters for nitriding the surface of the aluminium to the largest possible depth, experiments were performed, which varied the nitriding parameters (pressure, accelerating voltage, RF power, nitriding time and the pressure in the chamber). After investigation by glow discharge optical spectroscopy (GDOS) to gain information about the penetration of nitrogen into the Al-alloy, further microstructural investigations were performed by transmission electron microscopy (TEM). This method allows a detailed analysis of the nitrided layers with respect to the crystal structure of the AlN layer formed during the plasma immersion ion implantation and the interface between the aluminium substrate and the nitrided layer. 2. Experimental 2.1. PI3 treatment PI3 experiments were carried out in a stainless steel chamber of 1 m3 evacuated by turbomolecular pump at 2000 lys. The base pressure obtained was better than 10y6 mbar. The plasma source was placed on a sidewall of the chamber and consists of a flat ‘pancake’ antenna connected to a 3 kW RF generator (13.56 MHz). Inductive coupling was obtained through a glass window of 300 mm in diameter. The high voltage generator (GBS Electronik in Rossendorf) provided up to 30 kV, 8 amp pulses at 1 kHz maximum frequency.

0257-8972/02/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 2 . 0 0 1 1 9 - 6


R. Sonnleitner et al. / Surface and Coatings Technology 156 (2002) 149–154

Pulses were generated from a HP 33120A pulse generator. In all experiments described below, the pulse frequency was 1 kHz and the pulse width was 70 ms. The substrate holder was a 200 mm diameter disk connected to the power supply and placed approximately 200 mm perpendicular to the source. Gas injection was made from the top wall of the chamber. Gases (N2, NH3, SiH4, Ar, CH4) were controlled by mass flow controllers. The pressure in the chamber was measured by a baratron in the range 0–6.5 Pa. The chamber was equipped with a Langmuir probe (Hiden Analytical), which shows a pure Ar discharge, an electron temperature of approximately 3 eV and a plasma density of the order of 1010 cmy3 at 0.13 Pa and 1200 W RF power. In a pure nitrogen discharge, at 0.2 Pa and 1500 W RF power, the plasma density obtained was 7=109 cmy3 with an electron temperature of 2.5 eV. To analyse the influence of accelerating voltage and RF power on the microstructure specimens with the largest differences in accelerating voltage (from 10 to 30 kV) and RF power (from 500 to 1000 W) were investigated. To exclude any influence of alloying elements (Si, Cu, Mg) on the surface reaction between aluminium and nitrogen w4x, disks of pure Al with a diameter of 30 mm and a thickness of 5 mm were used. The surfaces of the samples were polished before the plasma immersion ion implantation. For successful plasma nitriding of an aluminium surface, a pre-treatment by plasma etching is essential to remove oxygen from the surface w5x. Pre-treatment over a period of 30 min was performed in the plasma immersion ion implanter (PI3) using an accelerating voltage of 10 kV. The composition of the etching gas was 50 sccm Ar and 10 sccm H2 and the treatment power was identical to the power used in the subsequent nitrogen implantation. 2.2. Sample preparation for TEM investigations The PI3 nitrided aluminium samples were prepared for cross-sectional investigations. Silicon wafers were glued to the nitrided surface of the Al-disks, and quadrangular prism-shaped pieces were cut from the stack using a diamond wire saw. The etches were ground to produce cylinder-shaped pieces with a diameter of 2.3 mm. After embedding the cylinders in brass tubes, 200 mm thick slices were cut. The slices were polished to a thickness of 70 mm followed by mechanically dimpling to a thickness of 20 mm in the centre of the slices. Ion milling was performed by a GATAN PIPS娃 (Precision Ion Polishing System) employing two Arq ion guns. The incident angle between ions and sample was varied from 108 at the beginning to 48 at the final stages of thinning after a perforation had appeared. The accelerating voltage of the Arq ion was 5 kV w6x. TEM investigations were performed with a Philips

Table 1 Conditions of treatment for samples 1–6 wJRCx Sample (no)

N2 (sccm)

P (Pa)

RF (W)

HV (kv)

Time (min)

1 2 3 4 5 6

100 100 100 100 100 100

0.16 0.15 0.14 0.14 0.12 0.12

1250 530 300 1000 500 500

20 20 20 20 10 30

240 240 120 120 120 120

CM 20 TEMySTEM operating at 200 kV and an EDAX Phoenix system for element analysis. Selected area electron diffraction (SAED) was used for the determination of the crystal structure. 3. Results and discussion Firstly the influence of the RF power on the formation of a nitrogen-rich layer was investigated. The specimens (sample 1qsample 2) were treated in the PI3 chamber under the conditions indicated in Table 1. Depth profile investigations by GDOS show a similar nitrogen profile for both samples (1 and 2): the concentration of N (in wt.%) is higher than 25% up to a depth of 60 nm (from the surface of the specimen) and lower than 2 wt.% at a distance of 170–200 nm below the surface. An oxygen content of more than 5 wt.% is detectable up to a depth of 200 nm (Fig. 1). The oxygen profile can be explained by two mechanisms: first, the oxygen was not completely removed from the surface during the etching process, and secondly, oxygen is implanted into the specimen during the PI3 process. GDOS is an instrument for measuring the depth profiles of elements from the surface of the specimen to a certain depth, but for the determination of phases and structure in the surface near region, the use of transmission electron microscopy is essential. The TEM-image of a cross section of sample 1 at low magnification gives an overview of the interface between the Al-bulk material and nitrided layer (Fig. 2). The surface of the specimen after PI3 is relatively smooth and the thickness of the reaction zone between aluminium and the nitrogen ions is regular. TEM images at high magnification show that the layer on the sample surface is approximately 60 nm wide (Fig. 3). This layer was identified by electron diffraction as fine polycrystalline hexagonal AlN. The outer zone of the layer (directly at the surface), which has a thickness of 35 nm, shows a columnar structure perpendicular to the surface. Therefore, there is an excellent correspondence between GDOS and TEM results concerning the thickness of the AlN layer: up to the depth where the nitrogen content is higher than 15–25 wt.% in the depth profile, the formation of hexagonal AlN occurs. The thickness of the hexagonal AlN layer is 60 nm.

R. Sonnleitner et al. / Surface and Coatings Technology 156 (2002) 149–154


Fig. 1. Depth profile of N, O, Al.

TEM-investigations of a cross-section of sample 2 (Fig. 4), which was treated at much lower power (530 W compared to 1250 W) show that formation of fine polycrystalline hexagonal AlN occurred during the PI3 treatment. The surface of the Al disk is still smooth after nitriding and the AlN layer was formed homogeneously on the whole surface of the Al disk. The AlN layer is 60 nm thick, which is comparable to sample 1.

A columnar structure in the outer zone of the layer generated during plasma treatment is visible. Summing up these results, the conclusion can be drawn that RF power has little influence on the formation of a fine polycrystalline AlN layer. Both at 530 and 1250 W, an AlN layer with a thickness between 50 and 60 nm is formed.

Fig. 2. Sample 1 TEM (low magnification): AlN layer on aluminium bulk material.

Fig. 3. Sample 1 TEM (high magnification): interface AlN layer–Al bulk, stripes in the AlN layer perpendicular to the surface.


R. Sonnleitner et al. / Surface and Coatings Technology 156 (2002) 149–154

Fig. 4. Sample 2 TEM (high magnification): interface AlN layer–Al bulk, stripes in the AlN layer perpendicular to the surface.

To be certain that the influence of RF power on the microstructure and the thickness of the AlN-layer is low, a second pair of aluminium disks (sample 3 and sample 4) was nitrided in the PI3 chamber at varying RF powers. Gas composition and accelerating voltage were the same as used for the treatment of samples 1 and 2. Pressure was slightly lower and the duration of treatment was much shorter (Table 1). There is little difference in the microstructure between samples 3 and 4. The formation of fine polycrystalline hexagonal AlN occurs at 300 W as well as at 1000 W. The layer thickness is slightly different, it is between 30 and 50 nm at lower RF power and approximately 50–60 nm at higher RF power, which is comparable to the AlN layer thickness of samples 1 and 2. PI3 treatment at 300 W might be too low for producing thick AlN layers, but even at 300 W RF power, the difference in layer thickness compared to samples treated at much higher powers is low. Also, the treatment duration of 120 min, which is just half the time in which samples 1 and 2 were nitrided, and a slightly lower pressure do not decrease the thickness of the AlN layer. However, there is a difference in the microstructure of the first set of samples (1q2) and the second set of samples (3q4): although the crystallographic structure is polycrystalline hexagonal AlN for all samples, the columnar structure, which was found in the 240 min treated samples (1q2), was not detected in the 120 min treated samples (3q4) (Fig. 5). Having shown that the influence of the RF power on the thickness and structure of the AlN layer is very low,

the influence of high voltage was investigated. The treatment parameters for samples 5 and 6 are listed in Table 1. The depth profiles of both samples look rather different from each other. The nitrogen profile of sample 5, which was nitrided at 10 kV, has a maximum of 10 wt.% at a depth of 25 nm from the surface and a value less than 2 wt.% at a depth of 125 nm. In comparison, sample 6, which was treated at 30 kV, has a maximum in the nitrogen profile (15 wt.%) at 20 nm, but more than 2 wt.% nitrogen is still detectable up to a depth of 400–500 nm. TEM investigations of the cross-section show that, even with an accelerating voltage of only 10 kV, the formation of very fine polycrystalline hexagonal AlN occurs at the surface. The thickness of the layer lies between 25 and 60 nm. The surface of the polished sample is still smooth after PI3 treatment (Fig. 6), no roughening of the surface occurs through a nitriding process at this accelerating voltage. No columnar structure in the AlN layer is visible. The microstructure of the cross-section of sample 6, treated at 30 kV, was found to differ strongly from all other samples. The high accelerating voltage causes a strong and non-uniform removal of Al, or even newlyformed AlN, from the surface, which leads to a deep roughening of the treated surface (Fig. 7). Notches up to 500 mn can be seen. The formation of fine polycrystalline AlN over the whole surface was detected. The thickness of AlN varies between 100 and 300 nm. Fig. 8 shows a 250 nm thick AlN layer at a higher magni-

Fig. 5. Sample 4 TEM (high magnification): AlN layer on aluminium, no stripes perpendicular to specimen surface.

R. Sonnleitner et al. / Surface and Coatings Technology 156 (2002) 149–154

Fig. 6. Sample 5 TEM (low magnification): smooth surface after PI3 treatment.


Fig. 8. Sample 6 TEM (high magnification): 250 nm thick AlN layer.

4. Summary fication. Broad stripes are visible. The surface roughness up to 500 nm is the explanation for the detection of nitrogen at a low concentration more than 500 nm below the surface measured in the depth profile. A continuous 500 nm thick smooth AlN layer was not formed during PI3 treatment, just a much thinner zone of non-uniformly thick AlN along the rough surface.

The formation of polycrystalline hexagonal AlN after PI3 treatment was detected on all samples, independent of accelerating voltage, RF power, treatment duration and pressure. However, the appearance and the thickness of the AlN is dependent on the treatment parameters. A strong influence on the surface structure after nitriding was the accelerating voltage. At voltages of 10 or 20 kV, the AlN layer is smooth and the thickness of the layer is relatively regular. An increase of the accelerating voltage to 30 kV causes a roughening of the specimen surface and the thickness of the AlN layer varies a lot. The influence of the RF power is low. The thickness of the AlN layers is comparable for all samples treated up to 20 kV accelerating voltage, rather independent from the RF power. Only at very low energies (300 W) is the thickness of the AlN layer lower. At treatment times of 240 min, the AlN layers show a columnar structure (samples 1q2), whereas after 120 min treatment time, no columnar structure is detected (samples 3q4). The results confirm the possibility of the production of AlN layers with a polycrystalline microstructure by PI3 treatment. Acknowledgments

Fig. 7. Sample 6 TEM (low magnification): high surface roughness after PI3 treatment at 30 kV.

The authors would like to thank the support received by partners of the Brite Euram project DUALCO BE97 4615 under which these investigations were performed.


R. Sonnleitner et al. / Surface and Coatings Technology 156 (2002) 149–154

References w1x C. Jarms, H.R. Stock, P. Mayr, HTM 51 2 (1996) 113–117. w2x W. Paatsch, Galvanotechnik, MO 51 9 (1997) 678–682. w3x R.J. Rodriguez, A. Sanz, A. Medrano, J.A. Garcia-Lorente, Vacuum 52 (1999) 187–192.

w4x B. Reinhold, J. Naumann, H.-J. Spies, Plasma Nitriding of Al Alloys; 11th Conference of IFHT, 1999, pp. 341–349. w5x H.R. Stock, C. Jarms, F. Seidel, J.E. Doring, ¨ Surf. Coat. Technol. 94–95 (1997) 247–254. w6x E. Stangl, K. Spiradek, M. Brabetz, Zinc-Based Coating Systems: Production and Performance; The Minerals, Metals & Materials Society, 1998, pp. 165–171.