Open Access
Add to BibSonomy
Abstract
Metal germanide thin films were investigated for infrared plasmonic applications. Thin films of copper and nickel were deposited onto amorphous germanium thin films and subsequently annealed at a range of temperatures. X-ray diffraction was used to identify stoichiometry, and SEM micrographs, energy dispersive spectroscopy, and atomic force microscopy were used to characterize composition and film quality. Electrical properties were analyzed via Hall measurements. Complex permittivity spectra were measured from 2 to 15 µm using IR ellipsometry. From this, surface plasmon polariton (SPP) characteristics such as propagation length and mode confinement were calculated and used to determine appropriate spectral windows for plasmonic applications with respect to film characteristics. Films were compared to similar palladium germanide and platinum germanide thin films and were evaluated for use with on-chip plasmonic components.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
A plasmonic waveguide is a sub-wavelength structure that enables the bound propagation of electromagnetic waves along the interface between dielectric and conductive media. These bound waves are typically referred to as surface plasmon polaritons (SPP). Plasmonic waveguides are a necessary link between photonic and electronic devices, facilitating nano-photonic devices and on-chip optoelectronic integrated circuits [1–3]. The performance of the waveguide, which includes the mode confinement and propagation length of the SPP, is based upon the designed structure geometry and the optical permittivity of the conductive host material. Early research in plasmonics primarily utilized noble metals such as gold and silver as the host material, as their plasma frequencies are in the short-wavelength visible, the real parts of their permittivity are strongly negative, and the imaginary parts are relatively small [4]. While noble metals have been very useful for plasmonic waveguides, propagation length has been limited by the loss associated with the negative part of the permittivity, whereas the mode confinement limits the spectral range to shorter wavelengths. New materials must therefore be investigated for both low loss and alternative spectral ranges.
For mid-wave infrared (MWIR, 3-5 µm) and long-wave infrared (LWIR, 8-12 µm) applications, materials such as doped silicon and germanium [5, 6], conducting transparent oxides [7, 8], silicides [5, 9] and germanides [10, 11] have been investigated. Comprehensive reviews of infrared plasmonic materials can be found in [2, 12].
Metal silicides and metal germanides have been investigated extensively for their use as contact materials in transistor designs, and have been shown to have similar properties [13, 14]. In particular, many of the phase transitions observed by annealing metal silicides show analogous trends for metal germanides. These materials are good candidates for integrated photonic circuits due to the ease of fabrication and CMOS compatibility. A previous study of metal silicides [9] focused on Pt-, Pd-, Ni- and Ti- silicides, all formed by annealing thin metal films on a silicon wafer at a temperature of 800 °C. These materials had suitable propagation lengths from the MWIR to well past LWIR spectral ranges, low loss and plasma frequencies between 1 and 2 µm. It is important to note that all four metal silicides showed similar optical properties.
Given their crystallographic similarities, metal germanides are expected to show similar properties to metal silicides. Germanium has an advantage over silicon for MWIR and LWIR applications in that it is has a higher transmissivity in this region. The real part of the permittivity for germanium is higher than for silicon, and germanium has higher carrier mobility [15]. This allows an integration of higher dielectric germanium oxides to be included in device design [16]. We have recently investigated platinum and palladium germanide films as potential SPP host materials. Films were characterized as a function of anneal temperature for crystallographic and optical properties [10,11]. While palladium germanide remained as PdGe for the temperature range, the platinum shifted from PtGe2 to Pt2Ge3 with increased temperature. For both material sets, the real and imaginary parts of the permittivity decreased with increasing anneal temperature; only films annealed at 500°C would be adequate SPP host materials in the LWIR.
This work further expands upon the metal germanide system for plasmonic applications by studying copper germanide films and nickel germanide films. These four metal germanides have been chosen due to their lower resistivity and low anneal temperatures. Film composition, texturing, and stoichiometry are characterized and correlated to resistivity and complex permittivity. Propagation length and mode confinement are calculated to determine the benefits of using metal germanide films for MWIR and LWIR SPP waveguides.
2. Materials characterization
Nickel germanide
Nickel germanide thin films were formed on p-type Si (100) wafers with vendor specified 0.001-0.005 ohm-cm resistivity. Substrates were cleaned in an O2 plasma prior to deposition of materials. Electron beam evaporation was used to deposit 200 nm germanium, immediately followed by 50 nm nickel. Wafers were then diced into 1 cm2 pieces and annealed in a nitrogen-purged tube furnace for 30 minutes each at temperatures ranging from 150 °C to 500 °C.
Samples were measured by X-ray diffraction (XRD) to determine crystalline phase and stoichiometry (see Appendix for experimental details). XRD data, shown in Fig. 1, suggest that the nickel germanide alloys go through two transitions as anneal temperature is increased. After a low temperature anneal (150 °C and 200 °C), XRD data shows small and broad peaks, suggesting a mostly amorphous structure with some weak crystallinity. The peaks at 44.6° and 46.6° both correspond to monoclinic Ni5Ge3 (PDF 00-024-0449), although the former could also be attributed to the orthorhombic Ni2Ge (013) phase (PDF 00-024-0452). After a 250 °C anneal the film transitions to orthorhombic NiGe (PDF 00-004-0545). The primary NiGe phases observed are the (111) and the (211) phases identified at 34.6° and 45.5°, respectively. The intensity is greater and peak widths narrower than observed for the lower temperature anneals; this suggests a much more crystalline structure. At 300 °C and above, films remain orthorhombic NiGe, but unreacted amorphous germanium has now recrystallized to form cubic Ge, where the Ge (111), Ge (022) and Ge (113) peaks are observed at 27.3, 45.4 and 53.7° respectively (PDF 00-004-0545). Patterson et al. observe these same transitions, although they attribute the transitions to higher anneal temperatures [17]. These higher temperatures may be due to the thicker films studied and a different anneal technique.
While peak positions do not change for anneals above 300 °C, the data suggest some morphological changes in this temperature range. To investigate this, the total intensities and full-width half maxima of select identified peaks are plotted against anneal temperature. Figure 2(a) features three peaks associated with cubic Ge. Figure 2(b) features five peaks associated with orthorhombic NiGe. For both crystal structures, the observations are the same: intensity increases with increased anneal temperature; peak width decreases with increased temperature. For NiGe, the steady increase in intensity indicates an increased crystallinity, particularly in the (111) orientation, which is the dominant and preferred phase from the powder diffraction reference. The narrowing of the peaks implies crystalline quality as a function of anneal temperature, although beyond 300 °C the width of the NiGe peaks remain roughly unchanged.
Fig. 2 Relative intensity and Full Width Half Maximum (FWHM) of (a) peaks identifying cubic Ge and (b) peaks identifying orthorhombic NiGe from the XRD spectra in Fig. 1.
Peak width is related to crystallite size by the Scherrer equation [18], which provides a lower limit for the average crystallite size. The equation takes the form
where τ is the volume weighted grain size, Δ(2θ) is the full width half maximum, θ is the Bragg angle, λ = 1.5406 Å is the Cu K-alpha wavelength and K~0.9 is the shape factor [19]. By this equation, NiGe crystals are at minimum 35 nm at 250 °C but grow to 50 nm for 500 °C anneals, an increase of about 40%.Figure 3 shows SEM images comparing the surface morphology of Ni-Ge thin films, including the unannealed film for comparison. The unannealed film is uniform and consists of grains less than 50 nm in diameter. The surface morphology is what is expected for an amorphous sputtered metal. An anneal at 200 °C results in an increased grain size, which grow into longer strands visible after a 250 °C anneal. Energy dispersive spectroscopy (EDS) indicates that while diffusion of germanium and nickel has begun, the nickel still dominates the top half of the film. After an anneal at 400 °C, the film has begun to segregate into light and dark regions, which is even more pronounced after a 450 °C anneal. According to EDS, the brighter regions are an even mixture of NiGe and Ge molecules, while the dark regions are much more Ge-rich. Overall, the 450 °C sample has a much higher concentration of Ge in the upper half of the film than was observed for samples of lower anneal temperatures. At 500 °C, larger agglomerations of Ge-rich particles have formed on the surface and have been strongly segregated from regions that are more evenly mixed between Ge crystals and NiGe crystals.
The agglomerations observed in the SEM images indicate a higher roughness at the surface of the films. Surface roughness was measured by atomic force microscopy (AFM) for each sample. Films annealed at temperatures up to 350 °C all had similar surface roughness of approximately 1.5 nm ± 0.3 nm. After a 400 °C anneal the roughness increased to 2 nm. The sample annealed at 450 °C had a roughness of 3.2 nm, while the sample annealed at 500 °C had the highest roughness at 4.9 nm. This is a significant increase in roughness and could be a limiting factor, although it is much smaller than MWIR wavelengths, and therefore is not a significant concern for most MWIR applications.
The segregation of two NiGe-Ge mixtures at high temperature anneals can be partially attributed to the initial film structure: a thick amorphous germanium layer and a thinner nickel layer must result in excess germanium. Gaudet et al. and Patterson et al. both note that germanium crystallization is much less prevalent at the surface if the metal is deposited on crystalline germanium, i.e. if the metal is deposited onto a germanium substrate and subsequently annealed [16,18].
The excess germanium also contributes to a non-uniform composition as a function of film depth. Cross sectional SEM imaging and EDS line scans were used to investigate this. Imaging confirmed the film thickness of 250 nm, with a variation less than 5% for all measured films. EDS line scans, although spatially imprecise due to spot size error, suggest that roughly the top ~50% of the films had an even mix of NiGe and Ge crystals, while the bottom half showed a smaller concentration of nickel. Furthermore, the surface nickel concentration was higher for films annealed at low temperatures than for films annealed at high temperatures. This is unsurprising, as higher anneal temperatures allow for greater diffusion of the nickel and germanium atoms.
Complex permittivity was determined using variable angle spectroscopic ellipsometry measured from 2 to 15 µm (J.A. Woollam, IR-VASE). For comparison, an unannealed sample consisting of 50 nm Ni and 200 nm Ge on silicon was measured and modeled as 2 layers on silicon. Annealed films were all modeled as a single, 250 nm layer on silicon. The thickness was fixed in the modeling.
Figure 4 presents the real (a) and imaginary (b) parts of the modeled complex permittivity spectra for the nickel germanide films. Films annealed at 150 °C-200 °C have similar permittivity to nickel; the films exhibit Drude characteristics and the real part of the permittivity remains negative in this spectral range. Films annealed between 250 °C-400 °C have similar permittivities; these films approach an epsilon near zero (ENZ) point at low frequencies, although they remain negative through the entire spectral range. The film annealed at 400 °C has an ENZ point at 1.1 µm, based on modeling. After a 450 °C anneal, the ENZ point red-shifts to about 1.5 µm, and is significantly smaller than the other films in the MWIR. However, this film has similar loss to the others. Only after a 500 °C anneal does the imaginary part of the permittivity drop significantly, while the ENZ through the entire spectral range. The film annealed at 400 in this spectral rangenealed Ge/Nipoint extends to 3.1 µm. An upswing in ε” at 2 µm occurs where ε’ is positive and a small absorption feature is present. This sample has the lowest loss, and corresponds to having the largest crystal grain size and Ge peak intensity from the XRD data; these two observations are likely correlated.
Fig. 4 Measured (a) real and (b) imaginary parts of the permittivity for annealed nickel germanide films. (c) Optical skin depth calculated from the complex permittivity.
The choice to fix the thickness of annealed films at 250 nm during modeling was made for consistency with previous analyses of platinum germanide and palladium germanide films [10,11]. However, analysis of the permittivity reveals the ellipsometry is incapable of determining thickness as the films are sufficiently optically thick. Figure 4(c) presents the optical skin depth, defined from Beers law as
where λ is the free space wavelength. Films annealed at 450 °C and below have a skin depth less than 100 nm in this spectral range. This implies that all permittivity data comes from the top half of these films, in which EDS indicates an equal mix of NiGe and Ge crystals. For the film annealed at 500 °C, the skin depth is less than the film thickness at wavelengths above the ENZ point of 3.1 µm. However, light would penetrate most if not all of the film, and hence optical properties should be considered an average of an effective medium consisting of a conglomerate of NiGe and Ge crystals, as indicated by the previous materials analyses.Copper germanide
Copper germanide samples were designed with an identical structure as the nickel germanide samples: 200 nm Ge followed by 50 nm Cu. Films were sputtered, as opposed to e-beam evaporation, onto O2 plasma-cleaned silicon substrates. Samples were annealed from 200 °C-500 °C. Thicknesses were confirmed via cross sectional SEM imaging, while XRD, EDS, and AFM characterized film quality and composition.
The composition of the copper germanide alloy does not appear to change over the observed anneal temperature range, although changes in the crystallinity are apparent. XRD data indicates that the film is composed between a mix of orthorhombic Cu3Ge (PDF 00-089-1146) and cubic Ge (PDF 00-004-0545). The intensities and full width half maxima of Ge identified peaks are shown in Fig. 5(a), while those identified with Cu3Ge are shown in Fig. 5(b). In a fashion similar to the nickel germanide films, the Ge (111) peak increases in intensity at higher temperature anneals. In this case, however, the effect appears to be more binary: for anneal temperatures through 350 °C, both the intensity and FWHM are relatively constant. At 400 °C the intensity jumps dramatically while also becoming much narrower, which remains fairly constant for even higher temperatures. Only the Cu3Ge (121) phase has a significant intensity which increases with anneal temperature, while the narrowing FWHM indicate an increase in crystallite size. Again using the Scherrer equation [18,19] to give the minimum average crystallite size, Ge crystals grow by a factor of 3, from about 16 nm to about 44 nm, while the Cu3Ge crystals grow from 24 nm to 42 nm, a little less than a factor of 2.
Figure 6 shows SEM images comparing surface morphology for a select few annealed copper germanide films. Films annealed between 200 °C and 400 °C all have similar surface characteristics, represented by the image on the left; films have small uniform grains, similar composition, and a surface roughness less than 3 nm, according to AFM measurements. EDS indicates that these films have a higher concentration of copper in the top half of the film. Interestingly, EDS data also shows a high concentration of oxygen in the top half of the film, approximately equal to the amount of germanium in the top half. This indicates the possibility of a surface oxide that was not seen in the nickel germanide films.
Fig. 6 SEM images of annealed copper germanide films. Note that the center two images are the same film, but at different magnifications to show segregation of film by composition.
There are two SEM images showing the film annealed at 450 °C. The first has the same magnification as the 300 °C film to compare grain size and roughness. This image shows some minor grain growth, while AFM measurements report roughness increased to about 3.5 nm. The second image is at a lower magnification and shows large scale compositional segregation across the film, which was not seen in the lower temperature annealed films. EDS measurements show the dark regions to be much more germanium rich than the light regions, suggesting the same grouping of excess cubic germanium crystals as was seen for nickel germanide films. This same segregation is seen for the sample annealed at 500 °C even at high magnification. Oxygen concentration is still high for this film, and while AFM measurements show surface roughness has doubled to about 7 nm, it is still significantly smaller than optical or plasmon wavelengths.
Complex permittivities were determined by modeling IR ellipsometry data. The raw data show additional absorption features in the annealed copper germanide films in the LWIR, a feature not present in the nickel germanide ellipsometry data. As EDS detected a high amount of oxygen in the film, a surface oxide was suspected. Such an oxide layer was sufficiently accounted for by considering a thin surface film with a Gaussian absorption feature around 14 µm. The thickness of the oxide layer was fixed at 1.5 nm which increased the quality of the fit. This oxide thickness was increased to 8 nm for the 500 °C annealed sample to obtain a high quality fit. These thicknesses are corroborated with data obtained by X-ray photoelectron spectroscopy (XPS, Fig. 7), which show a strong oxygen presence at the surface diminishing a few nanometers into the film; this surface oxide layer in the 500 °C film is nearly twice as thick as in the other films investigated. As copper has been shown to oxidize readily at room temperature and during processing [20], it is quite probable that a greater anneal temperature would facilitate greater oxygen diffusion into the film.
Fig. 7 XPS data for CuGe films annealed at 200 °C, 450 °C and 500 °C. Samples were sputter-etched in 15 seconds intervals, with XPS spectra measured before and after each etch.
The complex permittivities of annealed copper germanide films are presented in Fig. 8. These are obtained using the separate surface oxide layer in the model, hence any contribution of that surface oxide is ignored in this data. Unlike the other metal germanide films, the optical properties change only slightly over a wide range of anneal temperatures, despite observed morphological and crystallographic changes. Interestingly, the film annealed at 500 °C has the highest loss, shown in the imaginary part of the permittivity. As observed for nickel germanide, platinum germanide [10] and palladium germanide [11], the highest anneal temperature produced the smallest ε” values, possibly a result of the large grain size and strong crystallinity. It may be that the propensity of copper to oxidize readily could be countering any positive effect from crystal growth in this high temperature anneal. Overall, the permittivity of annealed films, particularly the imaginary part, changes very little, indicating that anneal temperature is having only a small effect on the optical properties of the material.
Fig. 8 Measured (a) real and (b) imaginary parts of the permittivity for annealed copper germanide films. (c) Optical skin depth calculated from the complex permittivity.
Electrical characterization
Figure 9 presents plots of resistivity, mobility and carrier concentration for the nickel germanide and copper germanide samples presented in this work, as well as palladium germanide samples previously studied [11]. Sheet resistance was measured via 4-point probe and van der Pauw technique used for Hall measurements. Resistivity was then calculated using the nominal 250 nm film thickness, which is consistent with previous works [10, 11] for comparison. Both methods are in good agreement. As a general trend, all films show little change over the middle range anneal temperatures (~250 °C-400 °C), which is in agreement with results from ellipsometry, XRD, and SEM data. In this range, palladium and copper germanide films each have a resistivity around 100µΩ-cm (copper germanide has a wider variation), while nickel germanide films have roughly half the resistivity, around 50-58 µΩ-cm. At the lowest temperature anneal, the nickel germanide film shows a spike in resistivity. This corresponds with Ni5Ge3, which was found to be a higher resistivity phase by both Gaudet et al. [16] and Patterson et al. [17].
Fig. 9 (Top) Resistivity measurements calculated from nominal 250 nm thickness and measured sheet resistance by van der Pauw configuration measurements. Calculated mobility (center) and carrier concentration (bottom) from Hall data. Pd-Ge is notated by the black squares; Ni-Ge is notated by the red circles; Cu-Ge is notated by the blue diamonds.
As noted earlier in this work, the metal germanide films are likely not uniform in concentration as a function of film thickness, and consist of a mixture of metal germanide and cubic germanium crystal structures. As a result, resistivity reported here is higher than found for pure metal germanide films, such as those grown by Gaudet et al. by depositing metal on germanium substrates [13]. We use the nominal thickness to stay consistent with previous studies [10, 11]. Thus, the reported electrical properties describe composite metal germanide films that contain excess germanium, which is consistent with the previous analyses.
Mobility is shown to decrease with increased anneal temperature, while the magnitude of carrier concentration increases, which results in the fairly constant resistivity as . Ni5Ge3 (annealed at 200 °C) has a mobility roughly 40 times larger than all other nickel germanide films, while the carrier concentration is roughly 20 times smaller. Nickel and palladium both have negative carrier concentrations, which means the majority carriers are electrons. Copper germanide shows the opposite; the majority carrier type is holes. Krusin-Elbaum and Aboelfotoh also determined a positive carrier concentration for copper germanide films based on low temperature Hall measurements [21].
It is important to consider the electrical properties of NiGe and PdGe after a 500 °C anneal, as these films show the lowest optical loss from the imaginary part of the permittivity. For PdGe, the electron concentration is nearly 6 × 1022 cm−1, while the mobility shows an increase from the 450 °C anneal. This combination means PdGe has the lowest resistivity (i.e. is the most metallic) after the highest anneal. NiGe has a similar mobility to PdGe at this temperature, but the carrier concentration is 3 times less, which results in the resistivity being higher (i.e. the least metallic) after the highest anneal.
3. Practical outlook
It is important to evaluate and compare the metal germanides investigated in this work with other MWIR and LWIR materials. Figure 10 compares the complex permittivity of nickel germanide and copper germanide investigated in this study along with those for platinum germanide [10] and palladium germanide [11] from previous studies. Permittivity is shown for films annealed at low temperature (300 °C, 350 °C for Pt-Ge) and high temperature (500 °C). All metal germanide films have similar permittivity after the lower temperature anneal. Although morphological and crystalline changes have been observed in these films, optically speaking they are not that dissimilar from unannealed metallic components. These films have similar optical constants to those found in metal silicides, such as PtSi, PdSi and NiSi [9]. Both the real and imaginary parts of the permittivity decrease significantly after the 500 °C anneal for all but the copper germanide films. The optical constants of these films approaches those found in transparent conducting oxides, such as ITO [2], GZO [22], FTO [22] and F-CdO [23]. These materials have ENZ points close to 3 µm, although many have even still smaller ε” values than those found in the metal germanides.
Fig. 10 Complex permittivity of various annealed metal germanide films: Copper germanide and nickel germanide from this work; platinum germanide from [10]; and palladium germanide from [11]. Solid lines represent 500 °C anneals, dashed lines show 300 °C anneals (350 °C) anneal for Pt-Ge.
It is necessary to utilize standard quality factors or criteria to assess the usefulness of a host material for SPP applications, and to effectively compare films of various compositions. West et al. [7] suggest , which considers the field propagation length. Films with low loss will have a long propagation length and are therefore superior by this metric. However, this does not consider limitations based upon mode confinement, indicated by the field penetration into surrounding media. On the other hand, Soref et al. [5] suggest basing the figures of merit on the plasmon propagation length (Lx), and the field penetration depths from the interface into the into the dielectric (Ld) and the conductor (Lc). A SPP host material is useful only if the field penetration depth in the dielectric is sufficiently small (strong mode confinement) and the propagation length is sufficiently large. They argue that the lower spectral limit is given by Lx > 2λ while the upper spectral limit is found from Ld <3λ.
The plasmon propagation length is given by [24]
where the SPP wavevector isThe 1/e SPP field penetration depth from the interface into the dielectric (d) or conductor (c) is [24]It is typical to calculate the field penetration into air for an SPP host material, therefore in Eq. (5).A plot comparing the QSPP of each film is shown in Fig. 11. The sharp dips for NiGe annealed at 450 °C and 500 °C correspond to the epsilon near zero (ENZ) points of the films. The films annealed between 250 °C-350 °C have a higher QSPP by more than an order of magnitude in both the MWIR and the LWIR, although poor mode confinement would make this spectral range problematic for these materials. Copper germanide films all have similar values for QSPP, although these are almost an order of magnitude smaller than for pure copper. Again, copper would not be suitable in the MWIR due to mode confinement issues. Thus, the QSPP figure of merit is useful for comparison, but is limited in that it mostly reflects propagation lengths.
Fig. 11 Plots of QSPP for (a) nickel germanide alloys and (b) copper germanide alloys. QSPP calculation is used from ref [7].
The SPP spectral ranges of all four materials, based upon the conditions Lx >2λ and Ld < 3λ are presented below in Table 1. For films only measured in the IR, the lower limit is listed as the minimum of the measured spectral range, although the real lower limit is probably in the visible spectrum, similar to NiGe and PdGe films. For PdGe and NiGe annealed at 450°C, the lower spectral limit is in a region where ε’ > 0. Strictly speaking, this region should not be considered “plasmonic”, although it has been shown that a bound electromagnetic wave can exist when the real part of the permittivity is positive provided that ε’<<ε” [25,26].
Table 1. SPP spectral ranges for all metal germanide alloys, and for all measured anneal temperatures.
The spectral ranges provided in Table 1 have been calculated considering the metal germanide as the conductive material and air as the dielectric. This would be applicable to an insulator-metal-insulator (IMI) waveguide, such as a stripe (slab) or slotted stripe waveguide design, or even a nanowire [5,27]. For such configurations, the copper germanide films are not suitable for MWIR and LWIR applications, but the other three metal germanide systems would be suitable host materials, especially if annealed at 500 °C. At that anneal temperature, NiGe and PdGe can function as IMI waveguides well beyond LWIR applications.
The wavelength of interest in this investigation was the MWIR and LWIR, thus ellipsometry was measured for all samples in this range. As mentioned above, ENZ points were observed for NiGe and PdGe after high temperature anneals. In an attempt to observe a shift in the ENZ point as a function of anneal temperature, NiGe and PdGe films were also measured in the visible spectrum. While no ENZ point was in fact found for these films, the additional measurements provided the full spectral range for SPP applications. No copper germanide films were ever measured in the visible as the curvature in the real permittivity did not suggest an ENZ point would be observed.
The metal germanide films can also be utilized in metal-oxide-semiconductor waveguides, such as a gap plasmon mode waveguide [28,29]. In such a waveguide, the SPP travels along the interface between the metal germanide and a dielectric material. As the permittivity of the dielectric will be greater than that of air, the SPP spectral range of the metal germanide will be red-shifted. To investigate this, the spectral ranges were calculated assuming the dielectric to be germanium (ε’ = 16, ε” = 0) to keep with the same Group IV material basis. The result is a wave that is much more tightly confined at the expense of propagation length. The penetration depth into the dielectric decreases by almost an order of magnitude, which pushes the upper spectral limit well beyond the measured spectral range in this investigation. The propagation length is decreased by more than an order of magnitude; therefore the lower spectral limit based upon Lx>2λ is shifted to higher wavelengths. It should also be noted that the plasmon wavelength is significantly decreased, and therefore surface roughness may play a larger role in limiting performance. Table 2 presents only the lower spectral limit for the metal germanide films encased in a germanium layer, as the higher limit is beyond the measurement range of this paper. We note that choice of a lower index dielectric, such as silicon in keeping with Group IV materials, would result in good spectral ranges that fall in-between what is shown in Tables 1 and 2.
Table 2. SPP lower spectral limit for metal germanide alloys encased in a germanium dielectric layer.
4. Summary
Nickel and copper thin films were deposited on amorphous germanium thin films and subsequently annealed to form metal germanide alloys. These alloys were investigated for crystalline structure and optical permittivity. These are compared to palladium and platinum germanides. While copper germanide films showed little change over a wide range of temperatures, the other materials changed significantly as anneal temperatures increased. Most notably, higher anneal temperatures resulted in increased grain sizes and crystallinity, although at a cost of increased surface roughness. After a 500 °C anneal, the dominant phases are NiGe, PdGe and Pt3Ge for each alloy, respectively. Nickel and palladium germanides also have significant Ge crystallinity, as shown by XRD analysis, which is not apparent in platinum germanide films. This may contribute to the smaller values for the negative part of the permittivity in the MWIR and LWIR, which itself contributes to nickel and palladium germanide alloys annealed at 500 °C having the widest spectral ranges of all materials investigated. Copper germanide is not suggested for IR applications due to its propensity to oxidize, but the other materials are good candidates for MWIR SPP applications.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Appendix Experimental details
Surface morphology and structure were characterized by scanning electron microscopy (SEM-Hitachi SU70). Energy dispersive x-ray spectroscopy (EDS, EDAX) was used to determine relative film composition using beam energies ranging from 2-7kV. Higher beam energies would penetrate the entire film, as determined by the presence of a silicon peak in the spectrum, while lower beam energies would penetrate less and therefore be more surface sensitive. Analysis software (TEAM EDS software suite) was used to identify the characteristic energy levels and identify relative film composition.
Sheet resistance values were measured using a 4-point probe. Samples were measured multiple times in different locations using different values for the input current, and the sheet resistance values were averaged. Film resistivity was calculated as the product of the sheet resistance and the nominal film thickness (250 nm for annealed films and 50 nm for unannealed metal films). Electrical data were also determined from Hall Effect measurements using an Accent Optical Technologies system in the Van der Pauw configuration. The magnetic field strength was 0.486 T with the current fixed at 19.9 mA each film measurement. The measurements were used to determine carrier type and carrier concentration N. Resistivity ρ was calculated from measured sheet resistance in the Van der Pauw configuration, which was also used to calculate the mobility µ.
Samples were measured by X-ray diffraction (XRD, PANalytical Empyrean) using an asymmetric out-of-plane configuration to determine crystalline phase and stoichiometry. The Cu Kα1 X-ray beam at 1.5406 Å was incident at a fixed angle of ω = 10 deg. The incident beam was passed through a hybrid monochromater with a 1/16° divergence slit. The divergent beam had a 0.04° soller slit and was measured using a PIXcel detector comprised of a 255 × 255 array. The detector was used in a scanning line mode to integrate the signal across each line, thus increasing the effective dwell time. The scattered intensity was recorded as a function of scattering angle 2θ° between 20-90 deg.
IR ellipsometry (JA Woollam, IR-VASE) was used to determine the complex optical constants for each film. The spectrometer has a range from 1.25-40 µm, although depolarization noise resulting from low signal on the detector means that data at the extremes of this range is unreliable. The limitation of the spectra to 2-15 µm represents a high confidence in the accuracy of the data and the modeling for this range.
Complex permittivity was modeled using the Woollam WVASE32 software. To aid in modeling, a bare silicon substrate was measured and modeled first using a simple Drude term. This substrate was then coated with 200 nm Ge and measured again. Multiple samples with only Ge were also annealed at select temperatures and measured for comparison. Annealed films were treated as a single, 250 nm layer. A model was built using a Drude term and an offset for the real part of the permittivity. In select cases a small Gaussian oscillator term was used as well. These model parameters, which were Kramers-Kronig consistent, were fitted to match the raw ellipsometry data for Psi and Delta. All models matched the data with a mean squared error (MSE) less than 3, all terms had a correlation less than 1, and all parameter terms were physical.
In the case of copper germanide films, a separate general oscillator layer was used on top of the 250 nm layer. This layer consisted of a Gaussian oscillator and a permittivity offset, and was designed to simulate a thin oxide layer on the surface. The thickness of this layer was set at 1.5 nm, although this had to be increased for the 500 °C film.
Funding
Air Force Office of Scientific Research (FA9550-15RYCOR162).
Acknowledgments
This material is based upon work supported by the Air Force Office of Scientific Research under award number FA9550-15RYCOR162. We thank T. A. Cooper for the Hall-effect measurements.
References and links
1. R. Soref, “Mid-infrared photonics in silicon and germanium,” Nat. Photonics 4(8), 495–497 (2010). [CrossRef]
2. G. V. Naik, V. M. Shalaev, and A. Boltasseva, “Alternative Plasmonic Materials: Beyond Gold and Silver,” Adv. Mater. 25(24), 3264–3294 (2013). [CrossRef] [PubMed]
3. S. Lal, S. Link, and N. Halas, “Nano-optics from sensing to waveguiding,” Nat. Photonics 1(11), 641–648 (2007). [CrossRef]
4. S. Maier, Plasmonics: Fundamentals and Applications (Springer, 2007).
5. R. Soref, R. E. Peale, and W. Buchwald, “Longwave plasmonics on doped silicon and silicides,” Opt. Express 16(9), 6507–6514 (2008). [CrossRef] [PubMed]
6. R. Soref, J. Hendrickson, and J. W. Cleary, “Mid- to long-wavelength infrared plasmonic-photonics using heavily doped n-Ge/Ge and n-GeSn/GeSn heterostructures,” Opt. Express 20(4), 3814–3824 (2012). [CrossRef] [PubMed]
7. P. R. West, S. Ishii, G. V. Naik, N. K. Emani, V. M. Shalaev, and A. Boltasseva, “Searching for better plasmonic materials,” Laser Photonics Rev. 4(6), 795–808 (2010). [CrossRef]
8. J. W. Cleary, M. R. Snure, K. D. Leedy, D. C. Look, K. Eyink, and A. Tiwari, “Mid- to long-wavelength infrared surface plasmon polariton properties in doped zinc oxides,” Proc. SPIE 8545, 854504 (2012). [CrossRef]
9. J. W. Cleary, R. E. Peale, D. J. Shelton, G. D. Boreman, C. W. Smith, M. Ishigami, R. Soref, A. Drehman, and W. R. Buchwald, “IR permittivities for silicides and doped silicon,” J. Opt. Soc. Am. B 27(4), 730–734 (2010). [CrossRef]
10. J. W. Cleary, W. H. Streyer, N. Nader, S. Vangala, I. Avrutsky, B. Claflin, J. Hendrickson, D. Wasserman, R. E. Peale, W. Buchwald, and R. Soref, “Platinum germanides for mid- and long-wave infrared plasmonics,” Opt. Express 23(3), 3316–3326 (2015). [CrossRef] [PubMed]
11. E. M. Smith, W. H. Streyer, N. Nader, S. Vangala, R. Soref, D. Wasserman, and J. W. Cleary, “Palladium Germanides for Mid- and Long-Wave Infrared Plasmonics,” MRS Adv. 2(44), 2385–2390 (2017). [CrossRef]
12. S. Law, V. Podolskiy, and D. Wasserman, “Towards nano-scale photonics with micro-scale photons: the opportunities and challenges of mid-infrared plasmonics,” Nanophotonics 2(2), 103–130 (2013). [CrossRef]
13. S. Gaudet, C. Detavernier, A. J. Kellock, P. Desjardins, and C. Lavoie, “Thin film reaction of transition metals with germanium,” J. Vac. Sci. Technol. 24(3), 474–485 (2006). [CrossRef]
14. Y. F. Hsieh and L. J. Chen, “Interfacial reactions of palladium thin films on Ge(111) and Ge(001),” Thin Solid Films 162, 295–303 (1988). [CrossRef]
15. H. Shang, H. O. Schmidt, K. K. Chan, M. Copel, J. A. Ott, P. M. Kozlowski, S. E. Steen, S. A. Cordes, H. S. P. Wong, E. C. Jones, and W. E. Haensch, “High mobility p-channel germanium MOSFETs with a thin Ge oxynitride gate dielectric,” Dig. Int. Electron. Devices Meet. 17(4), 441–444 (2002).
16. S. Gaudet, C. Detavernier, C. Lavoie, and P. Desjardins, “Reaction of thin Ni films with Ge: Phase formation and texture,” J. Appl. Phys. 100(3), 034306 (2006). [CrossRef]
17. J. K. Patterson, B. J. Park, K. Ritley, H. Z. Xiao, L. H. Allen, and A. Rockett, “Kinetrics of Ni/a-Ge bilayer reactions,” Thin Solid Films 253(1-2), 456–461 (1994). [CrossRef]
18. A. Patterson, “The Scherrer formula for X-ray particle size determination,” Phys. Rev. 56(10), 978–982 (1939). [CrossRef]
19. J. I. Langford and A. J. C. Wilson, “Scherrer after sixty years: A survey and some new results in the determination of crystallite size,” J. Appl. Cryst. 11(2), 102–113 (1978). [CrossRef]
20. J. Li, J. W. Mayer, and E. G. Colgan, “Oxidation and protection in copper and copper alloy thin films,” J. Appl. Phys. 70(5), 2820–2827 (1991). [CrossRef]
21. L. Krusin-Elbaum and M. O. Aboelfotoh, “Unusually low resistivity of copper germanide thin films formed at low temperatures,” Appl. Phys. Lett. 58(12), 1341–1343 (1991). [CrossRef]
22. J. W. Cleary, R. Gibson, E. M. Smith, S. Vangala, I. O. Oladeji, F. Khalilzadeh-Razaie, K. Leedy, and R. E. Peale, “Infrared photonic to plasmonic couplers using spray deposited conductive metal oxides,” Proc. SPIE 10105, 101050E (2017).
23. E. L. Runnerstrom, K. P. Kelley, E. Sachet, C. T. Shelton, and J. P. Maria, “Epsilon-near-zero modes and surface plasmon resonance in fluorine-doped cadmium oxide thin films,” ACS Photonics 4(8), 1885–1892 (2017). [CrossRef]
24. H. Raether, Surface Plasmons on Smooth and Rough Surfaces and on Gratings (Springer, 1998).
25. J. W. Cleary, G. Medhi, M. Shahzad, I. Rezadad, D. Maukonen, R. E. Peale, G. D. Boreman, S. Wentzell, and W. R. Buchwald, “Infrared surface polaritons on antimony,” Opt. Express 20(3), 2693–2705 (2012). [CrossRef] [PubMed]
26. F. Yang, J. R. Sambles, and G. W. Bradberry, “Long-range surface modes supported by thin films,” Phys. Rev. B Condens. Matter 44(11), 5855–5872 (1991). [CrossRef] [PubMed]
27. N. Kinsey, M. Ferrera, V. M. Shalaev, and A. Boltasseva, “Examining nanophotonics for integrated hybrid systems: a review of plasmonic interconnects and modulators using traditional and alternative materials,” J. Opt. Soc. Am. B 32(1), 121–142 (2015). [CrossRef]
28. J. Mu, R. Soref, L. C. Kimerling, and J. Michel, “Silicon-on-nitride structures for mid-infrared gap-plasmon waveguiding,” Appl. Phys. Lett. 104(3), 031115 (2014). [CrossRef]
29. R. F. Oulton, V. J. Sorger, D. A. Genov, D. F. P. Pile, and X. Zhang, “A hybrid plasmonic waveguide for subwavelength confinement and long-range propagation,” Nat. Photonics 2(8), 496–500 (2008). [CrossRef]
