CHAPTER 6 

 

STUDY OF THE INFLUENCE OF THE EXCIMER LASER WAVELENGTH ON THE MICROMACHINING OF AL₂O₃ - TiC CERAMICS

 

The micromachining of Al₂O₃-TiC ceramics using nanosecond ArF (l=193 nm) and XeCl (l=308 nm) excimer lasers was studied. The effects of the laser fluence and number of pulses on the ablation rate and surface topography were analysed. A comparative study of the ablation of Al₂O₃-TiC using nanosecond laser pulses at different wavelengths (193, 248 and 308 nm) was performed.

 

6.1 INTRODUCTION

 

The radiation wavelength of excimer lasers covers the range 157-351 nm, corresponding to a photon energy between 7.9 and 3.5 eV, which is typically below the bandgap of dielectrics [1]. Under these conditions, single-photon absorption leading to interband electronic transition cannot take place. However, defects and impurities which exist in the material before irradiation may absorb single photons, and colour-centre formation may occur due to irradiation leading to subsequent absorption [2]. Non-linear absorption phenomena such as multiphoton absorption and transient radiation induced defects can also considerably enhance the coupling of radiation to the material [3]. The development of excimer lasers and their applications have promoted the investigation of the absorption behaviour of high intensity UV light by wide bandgap materials [4]. These materials are used in mirrors and lenses at excimer wavelengths, and for these applications avoiding absorption is critical since it leads to component failure [3]. On the other hand, excimer lasers have also been used for micromachining of dielectric materials, for example, to manufacture micro optics and dielectric masks [5, 6], and to create patterns with different refractive index inside optical materials [7]. The 157 nm wavelength of F₂ laser radiation strongly couples to glass via defects or near-band edge states, allowing for the micromachining of optical materials such as fused silica [8]. Due to its medium refractive index value and low absorption from the near-UV (>300 nm) to the IR (£5 mm) [9], Al₂O₃ is used in near-UV laser antireflection and dielectric mirror designs [10], and in combination with silicon dioxide coatings to form high damage threshold multilayer structures for UV laser applications (>10 J/cm² at 248nm [11]). The values found in literature for the Al₂O₃ bandgap vary between 7.3 [12] and 8.8 eV [13]. Thus, interband excitation due to single photon absorption cannot occur for radiation with wavelengths of 308, 248 and 193 nm corresponding to photon energies of 4.0, 5.0 and 6.4 eV, respectively. For these wavelengths, absorption of radiation by Al₂O₃ is thought to occur predominantly through coupling with defects and impurities in the material, which are frequently present in large number in polycrystalline alumina [1, 4]. The cross section for single photon absorption increases with decreasing wavelength, since the larger photon energy allows coupling to higher intermediate states within the bandgap, while in order to reach an intermediate state, electrons in a wider range of bands can be excited. Coupling of radiation by multiphoton absorption can also occur, and similarly becomes more likely with increasing photon energy [1, 4]. While it is not clear whether direct band to band excitations with two photons at 308 nm are possible at all, two-photon absorption (TPA) has been demonstrated for alumina at 308 nm (2.8´10^-11 cm/W) [14] and 266 nm (27´10^-11 cm/W) [12]. Significant two-photon absorption at 193 nm (2´10^-4 cm/W) has been demonstrated for Al₂O₃ thin films [15]. Nevertheless, for bulk materials typical TPA coefficients are lower, since the contribution of surface absorption become less relevant in total absorption [2]. Data for TPA at 193 nm for bulk Al₂O₃ could not be found in the literature, but should follow the variation observed for thin films and be larger than for 248 nm radiation.

Despite the current interest in optical materials, most of the early work discussing the influence of the radiation wavelength in excimer laser ablation was focused in polymeric materials [16, 17]. Quantitative modelling of the ablation rate d for polymers is often performed by considering ablation as a two-step process, light absorption followed by material ablation, using the equation

 

(6.1.1)

 

This equation is frequently used in different forms to relate the ablation rate with laser fluence, typically when the optical penetration depth (L₀=1/a_opt) is much longer than the heat penetration depth (e.g., most dielectrics and polymers in the UV range) [1]. Under those conditions, the laser beam may be considered as a volumetric heat source [18], and considering no phase change, the balance of energy can be expressed by

 

(6.1.2)

 

This equation relates the heat generated in the material to the laser energy absorbed in a volume equivalent to the absorption depth times the incident area (J/cm³). It assumes that heat conduction away from the irradiated volume does not occur. Consequently, for constant reflectivity R, the product Fa_opt should be proportional to DT. If ablation is a thermal mechanism, then the energy per unit volume required to cause significant ablation should be proportional to Fa_opt independently of the wavelength. Brannon et al. [16] investigated the ablation behaviour of polymide films for 248, 308 and 351 nm wavelengths, and verified that the experimental results obeyed equation

 

(6.1.3)

 

As mentioned previously, by fitting this equation to experimental ablation rate data, an effective absorption coefficient (a_eff) is estimated, which may differ significantly from the optical absorption coefficient (a_opt). Brannon et al. observed that the ablation threshold F_th decreases with increasing photon energy, but simultaneously a_eff increases so that that the product F_tha_eff » constant. According to the authors this indicates that ablation is predominantly thermal for fluences around threshold. Similar results were obtained for other polymers [19, 20].

A quantitative analysis of the ablation rate based on equation (6.1.3) has also been applied successfully to ceramics, such as alumina [5, 21], mica [22], or fused silica [23]. Ihlemann et al. [5, 21] studied the ablation of Al₂O₃ at different wavelengths, and observed that the ablation rate above ~5 J/cm² is similar for 248 and 308 nm radiation (Figure 6.1.1). For 248 nm radiation the ablation threshold varies considerably from sample to sample between 1 and 3 J/cm², but is always lower than that for 308 nm. At 193 nm, the ablation threshold and the high fluence ablation rate were considerably lower than for 248 and 308 nm radiation. Equation (6.1.3) adjusted well to the experimental results for fluences below 10 J/cm². Using bulk absorption measurements, Ihlemann et al. [21] showed that there is an increase of absorption by the material which is not ablated, attributed to defect (colour-centre) formation. Incubation is more effective for 248 nm than for 308 nm radiation, consistent with the lower ablation threshold at the shorter wavelength. Ihlemann also investigated laser ablation of fused silica using 193, 248 and 308 nm excimer laser radiation [24]. The reported bandgap energy values E_g of fused silica vary between 8 [24] and 9.3 eV [8]), above the photon energy of the wavelengths considered. The ablation rate is lower and the processed surface smoother at 193 nm than at 248 and 308 nm, a result explained by the larger absorption by colour centres and two-photon absorption at the shorter wavelength. Recently, F₂ lasers emitting radiation at 157 nm have also been used to micromachine fused silica [23]. Despite the low absorption of fused silica for low intensity 157 nm radiation (a_opt=10 cm^–1 [25]), this radiation provides sufficiently strong interaction to precisely machine shapes with smooth surfaces. Hermann et al. [8] showed that the ablation rate follows a logarithmic dependence on fluence for 157 and 193 nm radiation (Figure 6.1.2). According to the authors, the high absorption coefficient (a_eff=1.7 ´ 10⁵ cm^-1) for 157 nm radiation demonstrates the importance of defect formation to enhance absorption.

 

Figure 6.1.1: Ablation rate of Al2O3 vs. laser fluence at different wavelengths (adapted from [5, 21]). The solid lines are fittings using (1/aeff)ln(F/Fth). alumina_ihlemann.opj graph 1	Figure 6.1.2: Comparison of fused silica ablation rate at 157 and 193 nm. The solid lines are fittings using (1/aeff)ln(F/Fth) [8].

 

When the heat penetration depth is larger than the optical penetration depth (e.g., in metals) the laser beam acts as a surface heat source [18]. Under these conditions, and for pulses with duration in the nanosecond range, equation typically does not represent well the experimental data, since heat conduction occurs away from the volume where radiation is absorbed [18]. Also, equation (6.1.2) can no longer be used to estimate the surface temperature rise, which can then be estimated as [1],

 

(6.1.4)

 

The influence of wavelength on cone formation.

The influence of the radiation wavelength on the surface texture of machined surfaces has been scarcely studied. Rubahn et al. [22] investigated UV ablation of muscovite (white mica). When processing was carried out using 193 nm radiation, cone-like structures formed after irradiation with multiple laser pulses at fluences close to the ablation threshold (Figure 4.1.2a). The cones were avoided at about 10 J/cm² (Figure 6.1.3b). When 248 nm radiation was used cones did not form independently of fluence and strong melting occurred. The authors suggested that cones are avoided at 193 nm because the cone diameter increases with increasing laser fluence, but do not discuss the influence of wavelength on surface texture formation. Thomas et al. [26] observed that laser treatment of polyethylenimine (PEI) with 193 and 248 nm radiation leads to the development of a surface texture. It was found that that for similar fluences cones are sharper for 193 nm radiation. This was explained by the general trend in polymers to present a lower ablation threshold for shorter wavelengths [1]. As a consequence, the cone angle varies more with the number of pulses until the effective fluence on the lateral surface of cones is around the ablation threshold (thus becoming stable) (see Chapter 4).

 

Figure 6.1.3: SEM micrograph of mica surface irradiated with 500 pulses at 9.67 J/cm2 using 193 nm radiation. In Chapter 4, Figure 4.1.2a shows cone formation when processing at 1 J/cm2 with identical radiation wavelength [22].

 

6.2 EXPERIMENTAL

 

The experiments were performed on samples of Al₂O₃-TiC ceramic containing between 30 and 34 wt. % TiC. For comparison purposes, titanium carbide (99.5 wt. %) and alumina (96 wt. %) targets were also used. The experiments were carried out in air at atmospheric pressure. The experiments with 193 nm radiation were carried out with the excimer laser Compex 110, Lambda Physik, described previously, using a ArF laser gas mixture. The maximum laser pulse energy for this wavelength was 200 mJ and the pulse duration 25 ns (FWHM). The experiments using 308 nm radiation were performed with a XeCl excimer laser EMG 203 MSC, Lambda Physik, delivering pulses with a maximum energy of 250 mJ and 30 ns duration (FWHM). Processing was performed by geometrically projecting an aperture inserted in the laser beam path on the sample surface using a lens. The laser fluence was varied up to 14 J/cm² and the repetition rate was kept below 5 Hz.

 

 

FOR RESULTS, DISCUSSION, CONCLUSIONS AND BIBLIOGRAPHY CHECK THE THESIS MAIN PAGE