CHAPTER 3  

 

EXCIMER LASER OVERVIEW

 

3.1 THE EXCIMER LASER

 

Excimer lasers offer the most efficient access to the ultraviolet spectral region of any laser source, leading to high pulse energies and high average and peak powers, attributes that make them a powerhouse for ultraviolet laser micromachining applications. The term “excimer” is an abbreviation for “excited dimer”, which is a reminder of the excited diatomic molecules (e.g., Xe₂) that were originally used as laser gas in the first systems. In most modern excimer lasers, excited rare gas halogen molecules (e.g., ArF, KrF) are preferred as active laser media due to their higher efficiency (Figure 3.1.1). Recently the molecular fluorine F₂ excimer laser has developed considerably due to its short wavelength, which is attractive from a resolution and materials absorption standpoint. However, it is a comparatively inefficient laser medium, and radiation is difficult to transmit to the workpiece due to significant absorption in the optical elements [1].

 

Figure 3.1.1: Emission bandwidths of various excimer lasers (adapted from [2]).

 

The origin of excimer molecules can be explained as follows: if one considers a diatomic molecule A₂ with potential energy curves as in Figure 3.1.2 for the ground and excited states, since the ground state is repulsive, the molecule cannot exist in this state (i.e., the species only exists in the monomer form A in the ground state). However, the potential energy curve for the excited state has a minimum, and the molecule A₂ can exist in the excited state (i.e., the species exists in the dimer form A₂ during the excited state lifetime). Such a molecule is called an excimer. If a large number of excimers are created in a given volume, laser action can then be produced on the transition between the upper (bound) state and the lower (free) state (bound-free transition). This transition is used in excimer lasers. This type of laser has two peculiar but important properties, both due to the fact that the ground state is repulsive [3]: 1) Once the molecule reaches the ground state after undergoing the laser transition, it rapidly dissociates. This means that the lower laser level will always be empty. 2) No well-defined rotational-vibrational transitions exist, and the transition is relatively broadband (20-100 cm^-1). However, it should be noted that in some excimer lasers the ground state energy curve does not correspond to a pure repulsive state but features a shallow minimum. In this case the transition occurs between an upper bound state and a lower (weakly) bound state (bound-bound transition). Since the ground state is only weakly bound, a molecule in this state will undergo rapid dissociation either spontaneously (predissociation) or as the result of the first collision with another molecule of the gas mixture.

 

Figure 3.1.2: Energy states of an excimer laser [3].

 

3.1.1             Rare gas-halide excimer laser

 

A particularly important class of excimer lasers is the rare gas-halide excimer laser, in which a rare gas atom (e.g., Ar, Kr, Xe) is combined, in the excited state, with a halogen atom (e.g., F, Cl). Strictly speaking these molecules should not be called excimers since they involve unlike atoms. However, the word “excimer” is now widely used in this context, and this usage will be followed. Specific examples of rare gas halide lasers are ArF (l=193 nm), KrF (l=248 nm), XeCl (l=308 nm) and XeF (l=351 nm), all emitting in the UV range. The reason why rare gas halides are readily formed in the excited state is apparent when one realises that excited rare gas atoms become chemically similar to alkali atoms, which are known to react readily with halogens. This analogy also indicates that bonding in this excited state is of ionic character: the excited electron is transferred from the rare gas atom to the halogen atom. This bound state is therefore also referred to as charge-transfer state.

The KrF laser will be considered in some detail, as this represents one of the most important lasers in this category, and is the most used throughout this work. Figure 3.1.3 shows the potential energy diagram of the KrF molecule [3]. The upper laser level is an ionically bound charge transfer state which correlates, at R = ¥, to the ²P state of the Kr positive ion and to the ¹S state of the negative F ion. The energy at R = ¥ is thus equal to the ionisation potential of the krypton atom minus the electron affinity of the fluorine atom. For large values of the internuclear distance the energy curve obeys the Coulomb law. The interaction potential between the two ions therefore extends to a greater distance (5-10 Å) than when a covalent interaction predominates. The lower state is covalently bonded and correlates, at R = ¥, to the ¹S state of the krypton atom and to the ²P state of the fluorine atom. Thus, the atomic states of the rare gas and halogen are reversed in the ground state. As a result of the interaction between the corresponding orbitals, both upper and lower states are then split, for short internuclear distance, into ²S and ²P states. Laser action occurs on the ²S ® ²S, since it has the largest cross section. During the transition, the radiating electron transfer from the F^- to the Kr⁺ ion.

 

Figure 3.1.3: Potential energy diagram showing the molecular structure of KrF [3].

 

As far as the excitation mechanisms are concerned, electrical excitation produces mainly excited Kr atoms and ions. When these species are present the excited KrF molecules are rapidly produced. The excited Kr atom can react with the F₂ molecule according to the reaction

 

(3.1.1)

 

Extending the previous analogy between excited rare gas atoms and alkali atoms, it is readily appreciated that the reaction rate of (3.1.1) is comparable to that between Rb (the alkali atom nearest to Kr in the periodic table) and the F₂ molecule. The Kr⁺ ion, on the other hand, reacts with the F^- ions that are formed by the dissociative attachment reaction

 

(3.1.2)

 

In order to conserve both total momentum and energy, the recombination between the two ions must occur via a three-body collision process

 

(3.1.3)

 

where M is an atom of the buffer gas (usually He or Ne). Because of the long-range interaction of the two ions, reaction (3.1.3) also proceeds at a very fast rate provided that the buffer gas pressure is sufficiently high. Although laser conditions are ideal, the production of operational lasers is quite demanding in technical terms. The reason for this can be derived from equation (3.1.3), since in order for a triple collision to occur with a reasonable high probability, a high gas pressure is required, at the very least in the bar range. At such pressure the free electrons needed for a collision can only be produced under very stringent conditions.

 

3.1.2             Technical realisation of an excimer laser

 

The excitation in an excimer laser may be performed using either an electron beam or a gas discharge [2, 4, 5]. Figure 3.1.4a shows a scheme of an excimer laser excited with an electron beam. The pulsed electron beam is ejected laterally through a thin foil in the gas reservoir, and laser emission arises at right angles to it in the resonator. Electron beam lasers are very awkward to handle, extremely expensive, and not suited for high repetition rates (typically limited to 10 Hz). For industrial applications, the excimer laser pumped with a gas discharge has found increasing use (Figure 3.1.4b). In the case of the KrF laser, the gas mixture consists typically of 6% Kr, 0.2% F₂ and the remainder is a buffer gas (Ne), reaching a total pressure between 2 and 3 bar. At this high pressure it is impossible to ignite a continuous discharge, since after a short period, a homogeneous discharge will reverse into an arc or spark discharge, which is not suitable for laser generation. Consequently, the excimer laser can only be operated in pulsed high-voltage discharge. The essential knowledge in the construction of an excimer laser resides in the technique of allowing a homogeneous gas discharge to continue for as long as possible and to stop the discharge before sparks are formed. In addition, these sparks cause erosion, which may ruin the electrodes quickly.

 

a)	b)
Figure 3.1.4: Schematic construction of an excimer laser pumped with a) an electron beam, and b) a gas discharge [6].

 

Two steps are particularly important to create a homogeneous discharge. Firstly, the gas mixture must be preionised as homogeneously as possible, which can be done either by means of X-rays, by electrons in a corona discharge (a spark discharge in extremely inhomogeneous electric fields) or by UV light. Secondly, a proper design of the electrical discharge circuit is required: the discharge voltage has to be applied between the two main electrodes with the shortest possible rise time. An optimum design can lead to pulse durations between 20 and 40 ns, which may under special conditions, extend up to 250 ns [2]. An excimer laser with a tube construction such as in Figure 3.1.4b would, however, show a clear drop in pulse energy after only a few pulses. This results from impurities in the gas due to electrode burn-off and inhomogeneties in the gas density, thus requiring the gas to circulate from pulse to pulse. Figure 3.1.5 shows a realistic construction of an excimer laser tube. The two long parallel electrodes have typical lengths of 80 cm, and the preionisation pins are arranged along the electrodes. The pulses generated at the preionisation pins (typically several tens of nanoseconds before the main discharge) generate UV light which ensures a homogeneous initial density of about 10⁸ electrons/cm³ between the electrodes. The entire tube functions as a reservoir for the laser gas mixture, which is circulated laterally through the resonator with the help of a circulation fan and a flow restrictor. The cooling rods are used as heat exchangers for the laser gas, as otherwise a steady rise in the temperature of the entire tube would occur, and this would have a negative effect on the output power.

 

Figure 3.1.5: Typical construction of a modern excimer laser [6].

 

In order to ensure an adequate population of the upper laser state of around 10¹⁵/cm³, electron densities of approximately the same order of magnitude are necessary. This results in current densities of about 10³ A/cm², which can only be obtained at dielectric strengths of 10-15 kV/cm. For this reason, commercial excimer lasers are operated with electrode spacing of 2-3 cm. As the entire discharge does not take longer than 20-40 ns, the discharge voltage has to be applied to the electrodes as quickly as possible, and it should be as far as possible above the static breakdown voltage. Figure a shows a simple electrical discharge circuit. The preionisation pins are either behind the storage capacitors C_s or behind the discharge capacitors C_p. After charging the capacitors C_s by the high-voltage source, the entire stored energy is transferred to the capacitors C_p with the aid of a thyratron. When the breakdown voltage is reached, the capacitors C_p discharge via the laser electrodes in a current pulse as short as possible. During the charge and discharge of the capacitors, a reverse current in the discharge circuit should be avoided. However, this is only possible using saturable inductivities, as shown in Figure b, or with magnetic switch control (MSC), as in Figure c. The magnetic switch decouples the slow thyratron circuit from the extremely fast laser discharge circuit and prevents reverse currents, thereby preventing a rapid destruction of the thyratron and the electrodes. When using MSC, the thyratron circuit can be adapted independently and the laser discharge circuit can be designed with very low impedance, leading to optimum component lifetime.

Figure ## Discharge circuits of modern excimer lasers

 

The amplification in excimer lasers is in the order of 10%/cm, leading to saturation after only a few round trips of the photons in the laser resonator, so that only little optical feedback is needed. The high gain of the excimer medium requires output-coupling reflectivities of 10-30 % for most efficient energy extraction. Most excimer lasers are used with stable resonators, consisting of a high reflectivity plane Al or dielectrically coated mirror and a plane CaF₂ or MgF₂ window as output mirror. The low number of round trips and the high amplification result in the oscillation of a large number of spatial transverse modes in the excimer laser (about 10⁶). Together with the relatively large bandwidth (a few Å), this results in low spatial and temporal coherence of these lasers, and high beam divergence (2-4 mrad). However, for many applications this is an advantage as interference effects caused by a high degree of coherence, such as speckle and standing waves, do not occur. When lower divergence is required, the lasers may be equipped with unstable resonators that reduce the beam divergence to 200-400 mrad in a beam with 60-70% of the pulse energy obtained with a conventional, stable resonator. Excimer lasers are characterised by their lack of a defined polarisation, which can be used in advantage since laser beam polarisation may affect processing negatively. The discharge volume of excimer lasers has a cross-section of about 1´3 cm² and a length of about 80 cm. Most excimer laser beams exhibit a reasonably flat-topped profile in their long axis, with a quasi gaussian distribution in the short axis. Typically, the pulse energy is in the range 10 mJ to 1 J, and the repetition rate can be raised to over 500 Hz. Due to the short pulse duration high power densities (>10⁸ W/cm²) are easily achieved even with low pulse energies. The efficiency is about 2%, while pulse-to-pulse stability is in the range 3-5 %. An increase in the average power can be obtained by increasing the discharge volume or by raising the repetition rate, however, both steps require large technical efforts. The use of pulsed high-frequency discharges instead of pulsed high-voltage discharges, has only generated pulse energies in the mJ range. Excimer lasers can function at several ultraviolet wavelengths depending on the gas mixture used, providing the opportunity to adapt the wavelength to the properties of the material by changing the active species. Changing the gas mixture while keeping the same halogen (e.g., from KrF to ArF) is a fairly easy process with most excimer lasers, and involves only a few hours of passivation of the laser cavity. This is achieved by flushing the laser several times and running at a manufacturer’s specified voltage. This will remove all remnants of the existing inert gas which would otherwise interfere with an efficient discharge.

Problems and disadvantages of excimer lasers include:

·High performance discharge circuit is required to generate laser pulses;

·Laser gas mixture is toxic and corrosive;

·High beam divergence, and low spatial coherence preventing tight focussing;

·Limited gas lifetime and expensive mixture components (rare gases such as Xe, Kr and Ne);

·Limited lifetime of transmissive and reflective optical components;

·High cost of operation due to gas consumption and regular optics maintenance.

These limitations have been minimised in commercial excimer lasers, which have been engineered to yield reliable industrial products. Electrical components typically have lifetimes exceeding 10⁹ pulses, whereas optical components need service or replacement after about 10⁸ pulses. At a repetition rate of 100 Hz, 10⁸ pulses accumulate after 280 h of operation. Gas mixtures degrade over about 10⁷ pulses in fluoride systems (ArF and KrF) and 10⁸ pulses when using chlorine (XeCl). The use of gas regenerators in closed cycle operation extends these lifetimes while yielding better pulse-to-pulse reproducibility. These regenerators condense chemical contaminants cryogenically and filter out particulate matter created by corrosion of internal surfaces exposed to the laser discharge and its chemical products. Commercial lasers are designed to allow for the replacement of the optics without exposing the main discharge volume to oxygen. This maintains the passivation of internal components and reduces the need for gas replacement after changing optics.

Excimer lasers have various important characteristics from a material processing point of view. Moreover, they have a number of advantages as compared to other laser systems, which altogether make excimer lasers an excellent choice for micromachining applications:

·Operation at short wavelengths (157 - 351 nm);

·Short pulse duration (10-30 ns), leading to high peak power;

·Low density in the active medium, which can support high intensities without self-focusing effects;

·They are scalable to high active volume, so that high pulse energies are available from large volume devices;

·Large area multi-mode beam (a distinction from harmonics of solid state lasers);

·Uniform power density over relatively large area (a distinction from harmonics of solid state lasers);

·Low spatial and temporal coherence minimising laser speckle and fringe formation on imaging (a distinction from harmonics of solid state lasers).

 

 

 

REFERENCES

 

1. D. Bäuerle: Laser Processing and Chemistry, (Springer-Verlag, Berlin Heidelberg 2000)

2. D. Basting: Excimer laser technology: laser sources, optics, systems and applications, (Lambda Physik AG, Göttingen, 2001)

3. O. Svelto: Principles of lasers, (Plenum Press, New York 1989)

4. J. P. Sercel and P. J. Sercel: Putting Industrial Excimer Lasers to Work, Photonics Spectra (May 1990)

5. J. P. Sercel: Production Excimer Laser Equipment Overview, SPIE 1835, 172-183 (1992)

6. H. Gerhardt: Fundamentals of Laser Physics and Laser Technology, (Lambda Physik)

7. T. R. O'Keeffe and T. E. Lizotte: Excimer laser as a manufacturing tool, Proc. SPIE 2062, (1994)

8. K. H. Gerlach, J. Jersch, K. Dickmann, and L. J. Hildenhagen: Design and performance of an excimer-laser based optical system for high precision microstructuring, Optics and Laser Technology 29 (8), 439-447 (1997)

9. J. H. Brannon: Micropatterning of surfaces by excimer laser projection, J. Vac. Sci. Technol. B 7 (5), 1064-1071 (1989)

10. P. T. Rumsby, E. C. Harvey, and D. W. Thomas: Laser Microprojection for Micromechanical Device Fabrication, Proc. Experimental Mechanics: Advances and Applications 2921, 684-692 (1996)

11. I. Watt: The principles and practice of electron microscopy, (University Press, Cambridge 1985)

12. JEOL report: Invitation to the SEM world

13. D. Instruments: Multimode SPM Instruction Manual, Version 4.31ce, (1996-97)

14. M. Sander: A practical guide to the assessment of surface texture, (Feinprüf GmbH, Göttingen 1991)

15. J. Yvon: http://www.jobinyvon.com (2001)

16. J. A. B. Faria: Óptica - fundamentos e aplicações, (Editorial Presença, Lisboa 1995)

17. M. Born and E. Wolf: Principles of Optics, (Pergamon Press, Oxford 1970)

18. S. J. Burden: Comparison of Hot-Pressed Alumina-Titanium Carbide Cutting Tools, Am. Ceram. Soc. Bull. 67 (6), 1003-1005 (1988)

19. Y. Choi and S. Rhee: Effect of Precursors on the Combustion Synthesis of TiC-Al₂O₃ Composite, J. Mater. Res. 9 (7), 1761-66 (1994)

20. R. Cutler, A. Hurford, and A. Virkar: Presureless-sintered Al₂O₃-TiC composites, Mater. Sci. Eng. A195/106, 182-192 (1988)

21. R. P. Wahi and B. Ilschner: Fracture behaviour of composites based on Al₂O₃-TiC, J. Mater. Sci. 15, 875-885 (1980)

22. T. Xia, Z. Munir, Y. Tang, W. Zhao, and T. Wang: Structure formation in the combustion synthesis of Al₂O₃-TiC composites, J. Am. Ceram. Soc. 83 (3), 507-512 (2000)

23. L. Wang, Z. Munir, and Y. Maximov: Thermite reactions: their utilisation in the synthesis and processing of materials, J. Mater. Sci. 28, 3693-3708 (1993)

24. 3M Ceramic Materials Department Product bulletin

25. M. V. Swain: Structure and properties of ceramics, R.W. Cahn, P. Haasen, and E.J. Kramer (VCH, New York 1994)

26. H. Gatzen: Rigid disk slider micromachining challenges to meet microtribology needs, Trib. Int. 33, 337-342 (2000)

27. C. Lacey, C. Durán, K. Womack, and R. Simmons: Optical measurements of flying height, Proc. Future Dimensions in Storage Symposium 81-88 (1997)

28. P. d. Groot: Optical properties of alumina titanium carbide sliders used in rigid disk drives, Appl. Opt. 37 (28), 6654-6636 (1998)

29. K. Komeya: In Structure and properties of ceramics, ed. by M. Swain (VCH, New York 1994), Vol. 11, pp. 517

30. W. J. Tropf and M. E. Thomas: Aluminum Oxide (Al₂O₃) Revisited, (Academic Press, 1991), pp. 653-682

31. F. Gervais: In Handbook of Optical Constants of Solids II (Academic Press, 1991), pp. 761-775

32. J. D. Cawley and W. E. Lee: In Structure and properties of ceramics, ed. by M. Swain (VCH, New York 1994), Vol. 11, pp. 47

33. P. Liu, W. L. Smith, H. Lotem, J. H. Bechtel, N. Bloembergen, and R. S. Adhav: Absolute two-photon absorption coefficients at 355 and 266 nm, Phys. Rev. B 17 (12), 4620-4632 (1978)

34. M. J. Weber: Handbook of Laser Science and Technology, (CRC, Baton Rouge, 1986), Vol. IV

35. W. W. Duley: UV lasers: Effects and Applications in Materials Science, (Cambridge University Press, 1996)

36. H. O. Pierson: Handbook of Refractory Carbides and Nitrides: properties, characteristics, processing, and applications, (Noyes Publications, 1996)

37. Z. A. Munir and J. B. Holt: Combustion synthesis of titanium carbide: theory and experiment, J. Mater. Sci. 21, 251-259 (1986)

38. A. G. Akopyan, S. K. Dolukhanyan, and I. P. Borovinskaya: Com. Explos. Shock Wave 14, 327 (1978)

39. E. K. Storms: The refractory carbides, (Academic Press, New York 1967)

40. W. S. Williams: Phys. Rev. 135, 505-510 (1964)

41. W. S. Williams: Electrical properties of hard materials, Int. J. Refractory Metals & Hard Materials 17, 21-26 (1999)

42. S. Méçabih, N. Amrane, Z. Habi, B. Abbar, and H. Aourag: Description of structural and electronic properties of TiC and ZrC by generalized gradient approximation, Physica A 285, 392-396 (2000)

43. J. Pflüger and J. Fink: In Handbook of Optical Constants of Solids II (Academic Press, 1991), pp. 303-311

44. D. R. Lide: CRC Handbook of Chemistry and Physics, (CRC Press, 1996)

45. TAPP: Thermochemical and Physical Properties (ES Microware, 1991)

46. V. N. Tokarev, W. Marine, C. Prat, and M. Sentis: Clean processing of polymers and smoothing of ceramics by pulsed laser melting, J. Appl. Phys. 77 (9), 4714-4723 (1995)