University of Rochester, Department of Material Sciences
Demand by the microlithography industry for higher power LASERS, more specifically in the deep ultra violet (DUV) range at 193nm, has put further demand on optics manufactures to produce high quality optics. The ArF excimer LASER, operating at 193nm, is the illumination source of choice for the microlithography industry. Over time, pushed by the market needs for higher throughput, the source has increased in power from 30W to 120W. In addition to the LASER's optics increasing the energy converting efficiency, the industry is demanding the lifetime of thin film coated optics last longer than their predecessors designs.
Typical Rare earth metals used in DUV thin film designs are Gadolinium III Fluoride GdF3, Magnesium II Fluoride MgF2, Aluminum III Fluoride AlF3, Lanthanum III Fluoride LaF3. The materials discussed and analyzed in this report are AlF3 and GdF3. The materials underwent atomic-force microscopy (AFM) and scanning electron microscopy (SEM) to investigate the samples structural properties with additional photometric and ellipsometric inspection to aid in understanding the material's optical properties. All these methods will be assessing two different designs in hopes to gather information on how each will compare to actual spectral responses based on their structural properties.
A typical issue in manufacturing DUV coatings is modeling the design spectra of a coating against the actual spectra output. Typical practice is to make the design easy to manufacture and control by simplifying the optical layer inputs. Fig. 1 are typical partial reflector's designs vs. the actual partial reflector's outputs, where the peaks are about 5% different. More fundamental understanding of multi-layer coatings is increasingly necessary based the current demand for partial and higher reflectors coupled with longer LASER damage thresholds.
The difference in reflectivity, refer to Fig. 1, is attributed to a couple of things. First, is making the design inputs easy of manufacturing. Secondly, each layer of the multi-layer film's index of refraction changes slightly with each added layer. The multi-layer partial reflector's optical properties, index of refraction, is directly related to the structural properties of the rare earth metals growth during thin film deposition. More specifically, the structures packing density and crystal growth changes as layers are added.II. Fundamental Understanding of Material Properties:
The objective of the fundamental understanding will be to utilize the practical experience and knowledge gained from the OPT 307 SEM Practicum course to investigate multi-layer properties in Rare Earth Thin Film Materials. Applying different microscopy techniques, such as scanning electron microscopy and atomic force microscopy or scanning probe microscopy, learned in the course, coupled with photometric and ellipsometric inspection to investigate structural and optical rare earth material properties used in physical vapor deposition of DUV coating materials. More specifically, investigating 193nm DUV thin film materials used in high and partial reflectors and anti-reflection coatings. The inquiry is with rare earth materials used in 193nm DUV coating and why multi-layer thin film optical designs sometimes don’t match the actual spectral responses, refer to fig. 1. Hence, the overall objective is to use the newly learned techniques to gain fundamental understanding as to why there is a mismatch between design and actual spectral response and get the actual response (green line) overlapping the design (redline) or as close as possible while additionally looking at a next generation coating layers by manipulating the thin film structural properties during deposition.
First, the evaluation will start with a single layer deposition. The single layer understanding is the foundation of any coating. Imaging was done with an SEM utilizing the BSD and each of the material's indexes were measured using a Woollam ellipsometer and a spectrophotometer for spectral response. Furthermore, an AFM was used to analyze the single layer's surface roughness.
Images in figure 2 and figure 3 are a current design and a new design respectively, for multi-layer DUV rare earth metal coatings. The current, design's single layer GdF3 has a average surface roughness of 1.45nm over a 5um x 5um area and ALF3 average surface is 1.0nm over a 1um x 1um area. The new design's single layer GdF3 and ALF3 have average surface roughness of 0.468nm and 0.732nm respectively over same areas in Fig2. In comparing both SEM BSD images one can see there is more porosity in Fig. 2 BSD image than in Fig. 3 BSD image. Additionally, the AFM confirms this with a tighter packing density shown on the surface thus leading to a better average surface roughness of about 27%. It is a little more difficult to see the porosity of the GdF2 because of its thickness, but the AFM images does show the GdF3 have different surface roughness indicating the columnar growth of the GdF3 is better in the newer design by 67%. A decrease in porosity in the AlF3 will positively affect the optical properties by decreasing scatter and absorption caused by the scatter. Likewise, to the columnar growth of GdF3 will affect the layer's optical properties in the same way as well.
III. Measurement Methods of Matrial Properties:
a. Sample Preparation:
Sample preparation is critical when setting up for use in the SEM. The samples analyzed for this paper needed to be cleaved to enable us to evaluate the cross section multi-layers with the SEM. In doing so, the layer's were compromised as seen in some of the cross sectional views, also refer to Fig 4. Additionally, samples analyzed in this paper where all non-conductive, so the technique of applying a conductive (gold) thin film was needed to prepare the sample for the SEM's electron beam interaction and avoid charged on the surface of the sample. There were several occasions where additional conductive coating needed to be applied to stop charging build-up on the sample.
b. Scanning Electron Microscopy:
The Scanning Electron Microscope is a powerful instrument in the area microscopy. It differs from typical microscopy most people are familiar with, called light microscopy, when using a light microscope. Each of their definitions are in their names. Light microscopy uses light as a source to image a sample. While the SEM or electron microscopy, uses electrons, in the form of an electron beam, to interact with the sample to provide an image of a sample. There are several different modes that can be used depending on what the users intentions are and what type of image is needed. The SEM can image samples using Secondary Electron Detectors (SED), Backscatter Electron Detector (BSD), InLens Detector, and further be used to analyze elemental make-up of a sample using the x-ray emissions of a sample and the Edax hardware and software. As mentioned in a Lab, all detectors are not created equal Each serves a purpose or is better at evaluating different areas of a sample. The SED detector provides the best insight of the surface based on their relative low energy. The InLens is also another type of secondary electron detector, but its location is perpendicular to the sample. The BSD detects higher energy electrons that go below the surface of the sample and have an interaction volume much greater than any of the other types of electrons. Hence the BSD is a preferred detector to use for elemental analysis. The Edax works in conjunction with the SEM's BDS to determine elemental analysis from the x-ray emission of a specimen. The Edax module is used to evaluate the x-ray emissions from the area being evaluated. There are a couple ways to review elemental ID's utilizing the Edax system. The first is by mapping the backscatter image, where the image is the area under test by which the information comes. From the map area comes the individual elements and an image breakdown of type and location of the elements. Each element is color coded for easier separate identification.
c. Atomic Force Microscopy:Uses a scanning probe to evaluate a surface of a sample. The samples in this paper used the scanning probe while in tapping mode. Tapping mode is one of the safest ways to evaluate a surface of a sample with out destroying the sample, called a non-destructive test. In tapping mode, the AFM probe only touches the sample based on a frequency of the cantilever and attached tip. There are forces between the probe and surface, electrostatic in nature, that dampen the vibration and intern is viewed as peaks and valleys of the sample as it scans across a preferred area.
d. Ellipsometery:A Woollam ellipsometer are another powerful instrument to aid in the area of thin film and optical design based on optical properties of material used in such a design. The ellipsometer aided in determining the optical constants called, index of refraction, for each of the materials tested. The ellipsometer was also useful as a secondary measurement of thickness, based on the index of refraction, to compliment the spectral results of the spectrophotometer.
Optical characteristic and properties can effectively be evaluated using spectrophotometers. Spectrophotometers breakdown UV and visible light into separate wavelengths via a diffraction grating in a monochromator and either bounced off a sample (reflection) or pass through the sample (transmission). Spectrophotometers were used to analyze the spectral response to the multi-layered thin film response to the designed wavelength at which the optic under test is to perform.
As proven, the multi-layer rare earth metal properties for Deep Ultra-Violet thin film coating material can be effectively analyzed through Microscopy coupled with spectral and ellipsometric techniques for the thin film optical coating industry. This is another area where microscopy can be exploited. The material properties, both mechanical and optical can be manipulated to improve a films structure thus improving the optical properties for DUV thin film coatings. Figure 6 shows as the layers are stacked the layers grow worse. Basically, the better the packing density and smoother the first layer, then all the other sequential layers will be smoother. Indexes will still change slightly as layers are stacked but should change at a slower rate. There is a saying in thin films, “Lay junk on top of junk and you get more junk” Once Again, sample preparation is critical to not altering the sample thus altering possible conclusions. The sample preparation with respect to the cleaving of the sample needs to be improved so the material structures are not compromised in the future.
1. Improve on sample preparation so the integrity of the layers in not altered.
2. Use TEM to collect a cross section to better image what material properties can be evaluated at the interface.
3. Continue practicing all the concepts of Scanning Electron Microscopy.
I would like to thank Brian McIntyre for his patients and all the knowledge he shared about microscopy during the semester. I will continuously use the knowledge gained from this class in my future work. Additionally, I need to give a big thank you to TA Fakhraddin Akbari Dourbash, better known as "Kev", for all your support. You never turned down a request for help.