Free Electron Laser Nitriding of Metals: From. basis physics to industrial applications

Transcription

1 Free Electron Laser Nitriding of Metals: From basis physics to industrial applications D. Höche a, G. Rapin b, J. Kaspar c, M. Shinn d, and P. Schaaf a a Universität Göttingen, II. Physikalisches Institut, Friedrich-Hund Platz 1, Göttingen, Germany b Universität Göttingen, Institut für Numerische und Angewandte Mathematik, Lotzestrasse 16-18, Göttingen, Germany c Fraunhofer Institut für Werkstoff- und Strahltechnik, Winterbergstrasse 28, Dresden, Germany d Thomas Jefferson National Accelerator Facility, Free Electron Laser Group, Newport News, VA 23606, USA Abstract Titanium was laser nitrided by means of free-electron laser irradiation in pure nitrogen atmosphere. The variation of pulse frequency and macropulse duration of the free electron laser resulted in δ-tin x coatings with different thickness and different micro- and macroscopic morphologies. The coatings revealed characteristic values for hardness, roughness and crystallographic texture, which originate from the growth mechanism, the solid-liquid interface energy and the strain. Further investigations showed that the dendritic growth is beginning at the surface and that the alignment of dendrites is normal to the surface. A correlation of the texture with the time structure of the laser pulses was found. Numerical simulations were performed and compared with the experimental results. The simulations can explain Preprint submitted to Elsevier Science 1 November 2006

2 the experimental results. Key words: Laser surface treatment, nitrides, laser plasma, reactive laser treatment, laser nitriding, titanium PACS: Lp, Jm, Ba, y 1 Introduction Titanium and its alloys play an important role as construction material, resulting from their superior corrosion resistance, high hardness, low weight and low toxicity. Concerning the improvement of wear resistance, nitriding of metals is an established technique to improve the surface properties [1]. Plasma and gas nitriding are frequently used and have been investigated in great detail [2]. A recent development is the use of laser radiation to process such metals like Ti, Al or Fe [3] in reactive atmospheres. The treatment with a free electron laser (FEL) is a possibility which has been employed very rarely [4]. First results of FEL nitrided titanium were reported in Ref. [5]. Here, additional investigations and especially simulations were carried out, in order to get more detailed information about the growth process, solidification velocity and their effects on surface morphology and properties after this direct laser synthesis. The simulations were compared to the experimental results of FEL laser nitrided titanium and they can explain the dendritic growth beginning at the surface leading to dendrites aligned perpendicular to the surface. address: pschaaf@uni-goettingen.de (P. Schaaf). URL: (P. Schaaf). 2

3 2 Experimental The FEL irradiations were carried out at the infrared FEL of the Thomas Jefferson National Laboratory in Newport News (USA, Virginia). The FEL radiation had a wavelength of 3.1 microns, with micropulses of 0.5 ps duration and 37.4 MHz repetition rate at a mean micropulse energy of about 20 µj. Macropulses with durations of τ p from 50 to 1000 µs at repetition rates f ma from 10 to 60 Hz were formed out of these micropulses, i.e. the macropulses consist of a train of micropulses. The macropulse energy ranged from 0.04 to 0.75 J. The raw Gaussian FEL beam was focused by means of a CaF 2 lens to a spot size of D b = mm on the sample surface. Therefore, a macropulse fluence of several hundred J/cm 2 could be induced on the titanium samples. For the laser nitriding, the samples were placed in a chamber, first evacuated and then filled with nitrogen (purity %) to a pressure of 10 5 Pa. The beam reached the sample surface through a fused silica window. In order to treat the whole surface of the samples, the chamber was mounted onto a computer-controlled x-y table. A relative velocity of v = 0.5 mm/s in the x-direction was used. The y-direction movement was represented by a lateral shift δ of 100 or 200 µm after each x-scan. The whole surface of the sample was irradiated by this kind of movement (meandered scan) and the irradiation of the surface can be described by means of the dimensionless overlap parameter [5]: σ = D2 b f ma v δ (1) which indicates, how often every point on the surface is irradiated. It allows a comparison between different experimental setups and the achieved coating 3

4 properties. The samples have been analyzed by X-ray diffraction (XRD) and with the Scanning Electron Microscope (SEM). Due to the interactions of the laser beam with the titanium, a plasma plume is formed above the titanium surface, and nitrogen diffuses into the titanium as a function of the treatment parameters involved. This correlation was investigated in greater detail. 3 Results and Discussion 3.1 Numerical simulations A numerical simulation of the nitriding process was established, in order to understand the laser nitriding process details not accessible experimentally. Especially, the missing possibilities for time resolved measurements of plasma and melt bath dynamics were tried to be resolved numerically. The heat transfer equation was solved coupled with the diffusion equation in a 2D-model supported by a Finite Element Modeling (FEM) [6] with temperature dependent material parameters. The surface temperatures obtained for the three macropulse durations are shown in Fig. 1. It is seen that the surface temperature is exceeding the evaporation point (depends mainly on pulse duration) and the resulting plasma plume determines the coating growth process. In addition, melting depths of about 80 µm were calculated as displayed in 4

5 temperature [K] T m - Ti T V Ti T m - TiN X time [ms] 250 µs 750 µs 1000 µs Fig. 1. Simulated surface temperatures for three pulse durations of 250, 750, and 1000 µm. Fig. 2. The melting depth determines also the nitrogen diffusion depth and thus the layer thickness. An important fact is the solidification direction. It is also seen in Fig. 2 that the solidification process starts at the surface with a velocity of a few cm/s. As a result, dendritic growth originates, starting at the surface. 3.2 Coating properties Figure 3 shows the surfaces of the prepared samples and the development of cracks, fractures and melting droplets could be observed as a function of induced energy. Comparison with the computations confirms the formation of melting droplets if the evaporation point is not exceeded. The cracks are an indication for strong tensile stress and strain. By means of XRD measurements, the development of the strain could be deduced as shown in Fig. 4 as a function of the texture parameter η (ratio 5

6 0 20 v S = 2-3 cm/s depth [µm] melting depth Ti N diffusion depth 80 N diffusion (2 P ) pulse duration corr. N diffusion 100 melting depth TiN time [ms] Fig. 2. Simulated melting depth for 1 ms pulse duration. Additionally, the calculated N - diffusion depth profile is shown for a single and 2 subsequent laser pulses. The computed solidification velocity is also represented. Fig. 3. Surface morphology of sample 1 (τ p = 250 µs, f ma = 60 Hz, δ = 100 µm), sample 2 (τ p = 750 µs, f ma = 30 Hz, δ = 100 µm), sample 3 (τ p = 1000 µs, f ma = 30 Hz, δ = 100 µm), and sample 4 (τ p = 1000 µs, f ma = 10 Hz, δ = 200 µm) showing melting droplets (especially for sample 1) and surface cracks. Scale is valid for all. 111/200 peak). Additionally, rocking curve scan of the (200) Peak were performed. The results are shown in Fig. 5. This proves the development of a (200) texture. Cross section SEM micrographs are shown in Fig. 6, where the melting depth can be determined to be 20 µm up to 80 µm for the different samples. All 6

7 strain * from (200) from (111) 1 polycrystal perfect (200) texture texture parameter region of (111) texture Fig. 4. Strain as obtained from XRD measurements as a function of the texture parameter η for the samples 1-4. intensity [a.u.] TiN (200) [degrees] Fig. 5. Rocking curves of the (200) peak for samples 1-4. They give a clear hint to strong textures. samples could be divided in nitrided, molten and heated affected zone, whose thicknesses are determined by the nitriding parameters. Cracks are only visible inside the TiN x phase. They result from from the strain which correlated with the dendritic growth. The surface properties supplemental influences the 7

8 macroscopic properties like hardness (Fig. 7). Only for sample 4 acceptable properties were achieved. Fig. 6. Cross section micrographs of the samples, where the lines show the simulated diffusion depths. The dendritic growth, the results of convection dynamics and the anisotropic nitrogen distribution are depicted. Scale is valid for all. 4 Summary It was shown that the nitrided samples reached properties of different quality as a result of the variation of production parameters. Cross sections showed layer thicknesses of about 5 to 80 µm with various nitrogen contents and different growth morphologies. The surface characteristics are mostly determined by the pulse duration and the overlap parameter. This behavior could be motivated by the numerical solution of the heat and diffusion problem. As a result of solidification velocity and temperature gradients, dendritic growth was observed. It follows after special solid-liquid interface properties such as enthalpy and stress, induced by the nitriding parameters. 8

9 hardness [GPa] pure Ti sample 1 sample 2 sample 3 sample pure Ti depth [µm] Fig. 7. Nanoindenter hardness curves showing the hardness depth profiles of the samples. They are a result of rifts, droplets and cracks on surface. Only sample 4 has a high hardness, because this sample does not exhibit cracks and has a smooth surface. In order to get coatings with optimized properties, these physical processes have to be quantified. Furthermore, we need more information by the grain size and the resulting strain. The latter was measured by X-ray diffraction and the dependence of the texture parameter and the dendritic growth could be shown. Whether these layer properties generate the desired macroscopic conditions still has to be investigated. Hardness and roughness were inspected to check the mechanical properties. A hardness of about 8 GPa could be reached, needing a surface without cracks and melting droplets. The latter are depending on heat transfer, solidification and strain, which can be minimized. In order to obtain surfaces with a small roughness, it is necessary to exceed the evaporation temperature. This seems to be the only possibility to obtain a smooth surface. However, melt dynamics limit this aim. Altogether, it seems to be possible to produce technical coatings by FEL nitriding, but still the 9

10 quality and especially the morphology have to be improved. Acknowledgements This work is supported by the Deutsche Forschungsgemeinschaft under grant DFG Scha 632/4. The Jefferson Lab is supported by the U.S. Dept. of Energy, the Office of Naval Research, the Commonwealth of Virginia and the Laser Processing Consortium. Kevin Jordan and Joseph F. Gubeli III are gratefully acknowledged for their assistance at the FEL. References [1] H. Xin, S. Mridha, and T. N. Baker. Journal of Materials Science, 31(1):22 30, [2] S. Grenier, K. Shanker, P. Tsantrizos, and F. Ajersch. Surface and Coatings Technology, 82(2): , [3] P. Schaaf. Progress in Materials Science, 47(1):1 161, [4] M. D. Shinn. Proceedings of SPIE, 4065: , [5] E. Carpene, M. Shinn, and P. Schaaf. Applied Physics A: Materials Science & Processing, 80(3): , May [6] FEMLAB AG, Göttingen. COMSOL 3.2 Multiphysics FEM Software package,