MICROSCOPY RESEARCH AND TECHNIQUE 38:519–524 (1997) Evaporated Germanium Films as Supports for Microanalysis of Carbon and Silicon Containing Specimens B.V. JOHANSEN,1,2* AND HEIDI ORMSTAD1 1Department 2EMLB, of Environmental Medicine, National Institute of Public Health, Torshov, 0403 Oslo, Norway University of Oslo, Blindern, 0316 Oslo, Norway KEY WORDS specimen support; transmission electron microscopy; germanium films; microanalysis; suspended particular matter (SPM) ABSTRACT A preparation procedure is described for producing specimen supports of evaporated germanium. The germanium film is used as a replacement for films of carbon and silicon when microanalytical techniques like energy dispersive X-ray microanalysis (XRMA) or electron energy loss spectroscopy (EELS) are focusing on particulates containing these elements. The supports can be produced with high reproducibility within a thickness range of 15 to 30 nm and of a quality suitable also for high resolution transmission electron microscopy. Microsc. Res. Tech. 38:519–524, 1997. r 1997 Wiley-Liss, Inc. INTRODUCTION In transmission electron microscopy (TEM) evaporated carbon films have been used for more than four decades as reinforcement for plastic supports (Bradley, 1954), and for the last half of this period plain carbon films have been the support of choice in high resolution TEM of bio-molecules (e.g., Johansen, 1974). Carbon evaporation has also been used to produce replicas from a variety of specimens ranging from fractured steel surfaces (Johansen, 1968) to freeze fractured biological materials (Bullivant, 1969). Traditionally, carbon evaporation is carried out in high vacuum but recently carbon coating procedures have been described also for low vacuum equipment, like the sputter coater (e.g., Namork, 1985). In environmental analytical transmission electron microscopy, carbon extraction replicas are usually the method of choice when air particulates are sampled onto polycarbonate filters (e.g., Chatfield 1986). The importance of thin film replicas is imperative when analytical techniques like electron diffraction (ED), X-ray microanalysis (XRMA), and electron energy loss spectroscopy (EELS) are involved. For the majority of applications in analytical electron microscopy, the use of carbon as top-coating, replicating material, or specimen support is usually uncomplicated. However, coating with carbon is not very successful in situations where the presence of carbonaceous material is of interest. For example, with immuno-electron microscopy we have shown recently that allergens from cat (Fel d 1) bind mostly onto soot particles in air samples, collected from the indoor environment (Ormstad et al., 1995). One way to investigate this further is to look at the ratios between the amounts of soot (carbon) and other small airborne house dust particulates (pure and complex silicates; organic, inorganic fibres, etc.). The inhaleable suspended particulate matter (SPM) of interest is less than 10 and 2.5 µm in diameter (PM10 and PM2.5, respectively). For obvious reasons we had to find a substrate and a replicating (film) material other than carbon or silicon for our further experiments. r 1997 WILEY-LISS, INC. Germanium is a group-IVA element in the periodic table, which is known to produce amorphous films, when deposited under certain conditions (Slope and Tiller, 1963). The amorphism is of importance when investigating crystalline particulates by selected area electron diffraction (SAED), or when setting proper imaging conditions based on defocus dependent phase contrast granulation (e.g., Johansen, 1973). This paper deals with the methodology of preparing germanium films and it shows examples of how such films can be applied in the study of airborne particulates collected in the environment. MATERIALS AND METHODS Evaporation Procedures Germanium (grade 99.999%; granularity 0.7–3.5 mm f from Balzer Union) was evaporated from a twothread, twisted tungsten basket (Balzers Union, Balzers, FL), (A in Fig. 1). Before loading the germanium granules, the basket was heated for about 3 minutes to glowing-red in high vacuum (5 3 1023 Pa), to remove the oxide layer from the tungsten filament surface. The evaporation was carried out in a JEOL (Skandinaviska AB) high vacuum coating unit (JEE 4X/5B), using a liquid nitrogen LN2 trap in the baffle region to prevent back-streaming of oil vapour to the chamber. The specimens to be coated were placed above the tungsten basket, facedown. The distance between the basket and the specimen holder was 100 mm. To evaporate accurate amounts of germanium and to protect the specimens from photons during the filament heat-up, a retractable shield was placed above the tungsten filament (B in Fig. 1). A multi-purpose holder, loadable with six specimens (C in Fig. 1), was made to accommodate either 25 mm f polycarbonate filters or 17 3 17 mm mica chips. For *Correspondence to: Professor B.V. Johansen, Dept. of Environmental Medicine, National Institute of Public Health, P.O. Box 4404, Torshov, 0403 Oslo, Norway. E-mail: email@example.com Received 27 August 1996; accepted in revised form 30 April 1997 520 B.V. JOHANSEN AND HEIDI ORMSTAD Fig. 1. The setup for evaporating germanium onto either 25 mm f polycarbonate filters or 17 3 17 mm cleaved mica sheets. The evaporation source A is a two-thread, twisted tungsten basket located at a distance of 100 mm below the specimen holder, C. The specimens are protected by a retractable shield, B, during the heating-up procedure. The detector crystal of the film thickness monitor is located in position D of the specimen holder. Fig. 2. A diagram showing the relationship between evaporated germanium mass and deposited film thickness with a source-to-specimen distance of 100 mm. accurate measurement of deposited film thickness, the detector of a Polaron film thickness monitor (Model E-5550) was always kept in one of the specimen positions (D in Fig. 1). The coating unit power supply delivers 25 V AC. After melting the germanium granules in the basket filament, the current regulator was set to 30 A. With a steady deposition rate (1–2 nm sec21) the radiation shield is retracted from the evaporation path for as long as it takes to produce the required thickness on the thickness monitor. Without a monitor the film thickness can be approximated by relating the thickness to the mass of evaporated germanium by weighing the loaded tungsten filament before and after deposition. The thickness-tomass relationship for the evaporating conditions and setup used here is shown in Figure 2. A Mettler MT5 (Greifensee, Switzerland) balance was used, with a readability and reproducibility of 1 and 0.8 µg, respectively. the surface of distilled water in a small beaker, as is usually done for plain carbon films (e.g., Johansen, 1974). With a water suction pump, the floating film is lowered gently onto uncoated copper grids. The grids had previously been immersed and placed on a filter paper on top of a stainless steel mesh. After the water is drained well below the grid level, the steel mesh, with the grids in position, is lifted out of the beaker and placed in a petri dish to dry overnight. Specimen Support Films From Coated Mica Plain specimen supports are made by coating germanium onto freshly cleaved mica. The germanium film is subsequently released from the mica by floating it onto Coating of Polycarbonate Filters One of our applications involved quantitative and qualitative analysis of airborne suspended particulate matter (SPM). SPM was collected on 25 mm f polycarbonate filters pre-coated with 10 to 20 nm layer of germanium. The collected SPM mass per unit volume of air (µg m23 ) was determined by weighing the coated polycarbonate filter before and after air sampling. In instances where the SPM mass exceeded around 30 µg m23 an additional layer of approximately 20 nm germanium was evaporated onto the polycarbonate filter. Discs, 3 mm in diameter, were punched out from SPM-exposed filters and transferred to 3 mm, 200mesh copper grids. The grids were placed in a Jaffe- EVAPORATED GERMANIUM FILMS AS SPECIMEN SUPPORTS 521 washer containing chloroform to remove the polycarbonate part of the specimen (Jaffe, 1948). The specimens were kept in contact with chloroform for 24 hours. Further details about how to handle polycarbonate filter specimens for transmission electron microscopy can be found in an International Organisation for Standardisation catalogue (ISO 10312-1995). The treatment resulted in a germanium replica of the polycarbonate filter. Electron Microscopy and Microanalysis A JEOL 100CX transmission electron microscope was used and operated at 80 KeV. The high resolution work was carried out with a 200 and a 60 µm f aperture in the condenser II and objective lenses, respectively. Micrographs of focal series of plain germanium films were Fourier transformed to a power spectrum to determine objective lens focus and image quality using either an optical diffractometer (Johansen, 1977) or the crystallographic image processing programme CRISP (Calidris; Sollentuna Sweden). Identification of the element carbon (fractal aggregates of soot) was performed with a Gatan (Warrandale, PA) electron energy loss spectrometer (EELS) coupled to a JEOL 2000 FX transmission electron microscope using a lanthanum-hexaboride (LaB6) electron emitter. The EELS was run with a resolution of 1 eV per channel. RESULTS Evaporated germanium films thicker than 30–40 nm often obscured high resolution details in objects supported by the film. On the other hand, supports with a thickness less than 10 nm were often experienced as brittle and sometimes too fragile for practical work. With the setup described in this paper, the thickness range of 15–25 nm gave acceptable and reproducible support films. According to the diagram in Figure 2 this corresponds to the evaporation of approximately 20–35 mg of germanium. The morphology of a 15-nm-thick germanium film floated off mica is shown in Figure 3. The selected area electron diffraction (SAED) diagram from this film, presented as an insert, shows wide and diffuse diffraction rings, similar to what is seen in films of amorphous carbon and vitreous ice. The amorphousness of a germanium film, supporting carbon black particles, is revealed by the defocus dependent granulation, at a much higher magnification, in a five picture focal series (Fig. 4a–e), taken from over to under focus. The corresponding optical Fourier transforms, (Fig. 4f–j) show the Thon-rings, which change through focus, in the same way as a conventional carbon support film. In the focal mini-series, Figure 4c and h were recorded close to optimum (Scherzer) focus. The fact that the phase contrast granularity of the germanium film responds to objective lens, defocusing in the same way as a conventional carbon specimen support, allows precise focusing and correction of astigmatism at high magnifications. A 10 1 15-nm-thick ‘‘sandwiched’’ germanium extraction replica of a polycarbonate filter, containing suspended particulate matter (PM10) from the indoor air of a children’s day-care center, is shown in Figure 5a. An area of the replica has been selected where the Fig. 3. An electron micrograph of a 15-nm-thick germanium film. Its corresponding selected area electron diffraction (SAED, inset) pattern reveals broad and continuous diffraction rings similar to what can be expected when electrons are scattered in amorphous materials (e.g., evaporated carbon and vitreous ice). Scale bar 5 0.5 µm. majority of the contaminating structures consist of soot aggregates (arrows). The electron energy loss spectrum (EELS) in Figure 5b, is recorded from such a soot aggregate (black circle in Fig. 5a) and confirms the presence of carbon. The spectrum in Figure 5c is recorded from an area (white circle in Fig. 5a) with no presence of carbon. DISCUSSION After carbon and silicon, germanium is the third element in group IVA of the periodic table with a physical density (5.32 g cm23 ) more than twice that of carbon (2.26 g cm23 ), which is the most commonly used material for specimen supports. Based on experiences with ion-beam sputtered films of chromium (7.19 g cm23 ) and niobium (8.4 g cm23 ) (Johansen, unpublished results), it was confirmed that the scattering contrast from the germanium film itself did not obliterate structural details in particulates supported by the germanium film, provided the film was not thicker than 30–40 nm. Going to higher accelerating voltages (.120 KeV), however, has shown that this thickness limit can be exceeded accordingly. 522 B.V. JOHANSEN AND HEIDI ORMSTAD Fig. 4. The focal series of carbon black particles supported on a 15-nm-thick germanium film imaged from over to under focus (a–e), reveals typical Fresnel diffraction around the particles. Fourier analysis of the defocus dependent granulation of the germanium film, using the PC assisted CRISP program, shows the characteristic Thon-rings through focus from f to j. The micrograph closest to Scherzer-focus (d) reveals a ‘‘nanometer crystallinity,’’ which is frequently seen in germanium but never in evaporated carbon films. The gaps between the ‘‘nanometric-islands’’ are detected in the Fourier transforms and can easiest be seen in h and i, indicated by arrows. Scale bar 5 50 nm. In attempts to establish if carbon is present in specimens like those described in this study, it is crucial that the germanium extraction replica itself does not contain carbon, for example, left by the chloroform digested polycarbonate filter. Electron energy loss analysis on sampled (Fig. 5c) and unsampled replicas, does not reveal detectable amounts of carbon. It is also our belief that germanium replicas of polycarbonate filters appear much cleaner, with no or very little filter residues after 12–24 hours chloroform digestion, as compared to what we experience with carbon-coated filters processed in a similar way. This effect may be due to the fact that the energy transfer between evaporation source and filter is significantly lower when depositing germanium (approximately 70°C) than coating filters with carbon (100–150°C). The amorphous structure of any deposited film is crucial if it is to be used as an extraction replica or a specimen support in transmission electron microscopy. Examining the texture of evaporated germanium films by selected area electron diffraction revealed a diffuse ring pattern close to what visually can be considered as amorphous. The deposition speed is known to affect the texture of any evaporated film (Slope and Tiller, 1963). With constant evaporating conditions (pressure, voltage, current) it was experienced that with reduced germanium mass in the basket filament, the time required to reach a preset thickness decreased, i.e., deposition speed increases. The maximum evaporation speed was mea- sured to 0.5 nm sec21, which is orders of magnitude below the rate required to produce crystalline or even polycrystalline germanium films (Slope and Tiller, 1963). The literature revealed some controversy about which temperature germanium films transform from amorphous to poly-chrystalline state (e.g., Slope and Tiller, 1963; Wolsky et al., 1965). When depositing 30-nm-thick films, the maximum substrate temperature measured never exceeded 70°C, which should be well below the transition temperature of 300–350°C cited as critical (Slope and Tiller, 1965). However, when inspecting the over- and underfocused micrographs in Figure 4 (a and e) a coarser structure may be sensed. This ‘‘nano-crystallinity,’’ which approaches amorphism, is best detected in the Fourier transform of Figure 4h and i, where the reflections can be seen as a narrow circle. The circular reflection appears on the radius of the Fourier transform close to where the frequency spectrum of the defocus dependent granulation starts to expand outwards and should not create a serious problem at low and intermediate resolution and magnifications. The effect on high resolution structures has not yet been investigated. Assuming that the nano-crystallinity of the germanium films is not caused by deposition rate or temperature, the explanation may be sought in the environment within the bell jar during evaporation. It is generally known that backstreaming of hydrocarbon oil vapours to the chamber of an untrapped vacuum evaporator will affect the structure and quality EVAPORATED GERMANIUM FILMS AS SPECIMEN SUPPORTS 523 Fig. 5. An indoor air specimen of suspended particulate matter (SPM) collected in a children’s day-care center. An area of a ‘‘sandwiched’’ 10 1 15-nm-thick germanium extraction replica from the polycarbonate filter shows large soot aggregates (carbon fractal structures) as indicated with arrows in a. A similar specimen processed as a carbon extraction replica will not allow the soot particles to be separated from the carbon film support by either X-ray microanalysis (XRMA) or electron energy loss spectroscopy (EELS). The EELS profile (b) is processed from the area within the black circle B in a and reveals the K-shell excitation of carbon. The spectrum (c) is from the soot-free area in circle C (a) and includes no carbon signal. Scale bar 5 1 µm. of the evaporated films. In this coating unit the forevacuum and the diffusion pump were loaded with fluorine-based (‘‘contamination free’’) oils. It was found that germanium films evaporated with LN2 in the trap located between the vacuum chamber and the diffusion pump, seemed more stable and less fragile than those prepared without LN2 in the trap. One can therefore suggest that remaining gas molecules, within the chamber of the evaporator, are interstitially packed into the condensed atomic structure of the germanium film and hence produce the nano-crystalline morphology. The conditions will most certainly be improved by evaporating in an ion-getter or turbomolecular pumped vacuum system. This approach will benefit coating of polycarbonate filters but will most probably make separation of the germanium film from mica substrates very difficult. ACKNOWLEDGMENTS The authors are indebted to Mr. Jan Zahlin for skillful technical assistance during the evaporation 524 B.V. JOHANSEN AND HEIDI ORMSTAD experiments. We also gratefully acknowledge Mr. Erik Soerbroeden and Professor Johan Tafto, Material Science Center, University of Oslo, for performing the EELS analysis. REFERENCES Bradley, D.E. (1954) A high resolution evaporation-carbon replica technique for the electron microscope. J. Inst. Met., 83:35–42. Bullivant, S. (1969) Freeze fracturing of biological materials. Micron, 1:46–51. Chatfield, E.J. (1986) Asbestos measurements in workplaces and ambient atmospheres. In: Electron Microscopy in Forensic, Occupational and Environmental Health Sciences. S. Basu and J.R. Millette, eds. Plenum Press, New York, pp. 149–186. ISO (1995) Ambient air: Determination of asbestos fibres–Direct transfer to transmission electron microscopy method. ISO 10312: Ed1:TC 146/SC 3. Jaffe, M.S. (1948) Handling and washing of fragile replicas. J. Appl. Physics, 19:1187–1191. Johansen, B.V. (1968) A general replication technique for microfractography, which allows large and selected areas to be studied. 4. European Regional Conference on Electron Microscopy, Rome, pp. 511–512. Johansen, B.V. (1973) Bright field electron microscopy of biological specimens: I. Obtaining optimum contribution of phase contrast to image formation. Micron, 4:446–472. Johansen, B.V. (1974) Bright field electron microscopy of biological specimens: II. Preparation of ultra-thin carbon support films. Micron, 5:209–221. Johansen, B.V. (1977) High resolution electron microscopy of biological specimens. Ultramicroscopy, 2:229–239. Namork, E. (1985) Preparation of carbon films from evaporated fibres and rods at fore-vacuum pressure. J. Electr. Micrc. Tech., 2:45–51. Ormstad, H., Namork, E., Gaarder, P.I., and Johansen, B.V. (1995) Scanning electron microscopy of immunogold labeled cat allergens (Fel d 1) on the surface of airborne house dust particles. J. Immunol. Methods, 187:245–251. Slope, B.W., and Tiller, C.O. (1963) Formation conditions, structure and electrical properties of Ge films deposited on single crystal substrates. Trans. 10th. National Vacuum Symposium. The Macmillan Company, New York, pp. 339–347. Slope, B.W., and Tiller, C.O. (1965) Formation conditions and structure of Ge-films deposited on polished (111) Ca F2 substrates in an ultrahigh-vacuum system. J. Appl. Phys., 36:3174–3181. Wolsky, S.P., Piwkowski, T.R., and Wallis, G. (1965) Low-pressure sputtered germanium films. J. Vacuum Sci. Tech., 2:97–103.