Figure 9. Laser damage probabilities of a ZnS/BaF2 partial reflector at 1.06 m Figure 10. Laser damage probabilities of thin films of the component of the design used for the tests in figure (9). materials Manuscript Received 2-8-89 Laser-Induced Damage of Dielectric Systems with Gradual Interfaces at 1.064 μm D. Ristau, H. Schink, F. Mittendorf, J. Akhtar, J. Ebert and H. Welling Institut für Quantenoptik, Universität Hannover ABSTRACT of Previous work has shown that laser induced damage thresholds A the microprocessor controlled coevaporation technique is used for production of high reflective and antireflective coatings. Damage thresholds and absorption data of these systems are compared to the performance of conventional systems. An increase of damage thresholds of up to 20% is observed for some materials. This improvement is discussed by comparing the influence of intense laser radiation on gradual and abrupt interfaces. key words: damage thresholds, gradual interfaces, codeposition, absorption, antireflective coatings, highreflecting mirrors, oxide materials. 1 Introduction From the standpoint of laser induced damage the e-beam process is still of current interest for the production of dielectric layer systems. At present only sol gel processes [1] are proven to produce coatings for special applications, which can withstand higher laser power levels than e-beam deposited coatings. In future, ion processes [2] like IAD (ion assisted deposition), IBS beam sputter deposition) or IPD (ion-plating deposition) are expected to surpass the potentiality of the e-beam process because these techniques yield coatings with an improved microstructure. But, although extensive research has been done on ion processes, the aspect of laser induced damage thresholds (LIDT) has not totally been clarified [3]. Compared to the conventional process ion processes are more complicated, less economic, and up to now, they found only limited application in production. Therefore, e-beam deposition is still a major process in the field of high power coatings and its potentiality should be totally explored. °) Optics Laboratory, Pinstech, Islamabad, Pakistan as In the past several years the technique of coevaporation is discussed a means to improve the properties of e-beam deposited coatings. In preceeding studies [4] the technique of coevaporated interfaces has been demonstrated to decrease the total losses in e-beam deposited stacks with alternating materials. In some cases [5] also damage thresholds were higher in gradual 2-QWOTstacks than in conventional systems with abrupt interfaces. systems like high In this paper an extension of the technique to practical high reflecting mirrors and antireflective coatings is described. For reflecting mirrors we tested TiO2, Taz Os, HfO2 and ZrO2 in conjunction with SiO2. Damage thresholds of gradual quarterwave stacks and gradual designs with adapted standing wave field distribution were measured and compared to data of the corresponding abrupt systems. Antireflective coatings are double layer systems of Ta2 05, HfO2 and ZrO2 in combination with SiO2 and a system of CeO2 /MgF2. A comparison of damage thresholds and absorption data for gradual and abrupt systems with the same designs are presented. 2 Experimental In the past, several techniques have been tested for the production of inhomogeneous layers. One of the eldest methods is the evaporation of materials with different evaporation temperatures from a single source [6,7]. During the production of the inhomogeneous region the mixing ratio is varied by adjusting the evaporation temperature. This technique does not need any mechanical alterations in the coating plant, but it suffers from the disadvantages that the properties of the layers are very sensitive to production parameters and the technique is restricted to soft coatings. A more sophisticated technique is based on an e-beam which is alternately switched between the crucibles containing the different materials [8]. The mixing ratio can be adjusted by the exposure times of the materials to the e-beam. For the production of coatings with gradual interfaces it is sufficient to regulate the deposition rates of the two materials forming the adjacent layers. If coating designs are restricted to types involving two materials, only two evaporation sources with seperate rate regulation circuits are necessary. In such an arrangement each source is working with constant rate during most of the deposition time. Solely for depositing the interface regions both sources have to be operated simultaneously with variing deposition rates. 2.1 Experimental setup The experimental setup for simultaneous deposition of two materials with arbitrary rates is shown in figure 1. A quartz crystal monitor head is attached to each e-beam source. Each monitor head is shielded against the evaporation flux of the adjacent source with the aid of an aperture. Thermal radiation from the sources is also kept off the monitor heads by these apertures. The rate regulation is carried out by a microprocessor system which registrates the depostion rates and controls the emission currents of both sources. In order to achieve a sufficient accuracy for the rate measurements the fundamental frequency of the quartz crystals is multiplied by a factor of 16 with the aid of a PLL-circuit. During rate regulation the frequency is counted digitally and latched into the microprocessor after every 200 ms. The processor calculates appropriate emission current values which are converted by a D/A-converter into steering voltages for the e-beam power supplies. In conjunction with a control panel the microprocessor is a complete rate control system. the and For the benefit of more flexibility a desk computer is linked to microprocessor system. The computer is executing the regulation and calculation tasks meanwhile the sub-system is processing the measurement control signals. Thus the programs for the regulation, steering, controlling and presenting of the process could be developped on the level of a higher programming language (BASIC). The process software also contains routines for recording the momentary rates and process data on floppy disc for further evaluation after completion of the process. For the production of defined index profiles a calibration and the determination of regulation parameters are necessary. Single layer samples samples with defined film thicknesses were used in order to calibrate the crystal monitor . The thickness measurement was performed by using an optical thickness monitor (Leybold OMS 2000) during the process and by evaluation of spectrophotometric data. Regulation parameters were optimized with the aid of a trial and error method utilizing an optional third crystal monitor head. Thus, compared to regulation by emission current the regulation errors of the deposition rate could be decreased by a factor of three. JACOBSSON [9] showed by calculating the refractive index as a function of deposition rates that the index profile is minor influenced by rate variations. For example, a rate error of 20 % results in aberration of approximately 8% for the refractive index. 2.2 Interfaces In order to keep the expense low for the production, design and examination of the systems with gradual interfaces we restricted ourselves to interfaces with temporal linear variations; i.e. during the deposition of the interface the rate of one material is decreased linearly down to zero meanwhile, in the same time-period, the rate of the other material is increased up to the optimum value. This results in refractive index profiles as shown in figure 2 for a gradual interface between HfO2 and SiO2. Depending of the theoretical model used for the calculation the index of refraction as a function of geometrical position shows a slight curvature. In case of figure 2 calculations are based on the Lorentz-Lorenz-theory [9]. The actual profiles realized by the codeposition process can be approximately calculated with the aid of the momentary rates recorded on floppy disc during the process. As an example, an index profile for a sample of TiO2 and SiO2 is depicted in figure 3. Circles in figure 3 indicate refractive index values determined by an AES depth profile of the same sample. The deviations of the index data range from 10 % to 20 %. Calculations of the spectral behavior of systems with gradual interfaces yield only small variations for different interface Therefore, characteristics are depths. spectral not usable for the determination of an index profile. Hence, summarizing the experience from our setup, the index profiles of gradual interfaces are merely reproducable with errors in the range of 20 %. 2.3 Design and production The v-coating design is best suited for antireflective coatings in high power applications. Two layer coatings of this type yield zero reflectance for any substrate material and they have minimum optical losses. They commonly consist of a thin high index layer next to the substrate which is followed by a thicker low index layer. This design guarantees low optical losses, because the high index material usually has higher absorption and scatter losses than the low index material. According to this, we choose for antireflective systems two-layer-designs of four material combinations (table 1). The material combinations were determined by preceeding studies on conventional v-coatings [10] and on 2-QWOT gradual systems [4]. For each combination three sets of gradual and conventional systems were deposited on fused silica substrates. The interface depth was kept constant for samples of one combination by choosing constant time intervals for the codeposition periods. Values for the interface depths are calculated by using the momentary rate data of the crystal monitor heads. The high average depth for the CeO2 /MgF2-systems results from a high optimum deposition rate for MgF2. Three types of designs for high reflecting mirror were choosen (table 2) for a total of four material combinations. The first design type (type N) is a standard quarterwave stack (qw stack) with one gradual interface between the first layer pair. The other two design types (type E and D) are non-qw systems optimized [11] for low standing wave field values at the interfaces. In type E the thickness of the upper high index layer is decreased meanwhile the thickness of the adjacent low index layer is increased in a way that the total optical thickness of the layer pair remains constant. Thus, the electric field at the interface between the layer pair next to the air can be adjusted to a value equal to that at the second HL-interface. The maximum intensity at the interfaces is accordingly reduced in the system. The ratio of the maximum intensity in a standard qw stack and the corresponding intensity in the adapted stack is given by the last column (gain) in table 2. In type D the first two layer pairs are adapted leading to increased gain values. Adjusting more than two layer pairs results in a less drastic increase of the gain values. Therefore, taking also into account the additional difficulties in production, a study of these systems should not reveal any new fundamental aspects. Design type E contains gradual interfaces in the first and second |