Johnson Matthey Technol. Rev., 2018, 62, (2), 211
Inter-Diffusion of Iridium, Platinum, Palladium and Rhodium with Germanium
Improved materials for the next generation of electronic devices
- Adrian Habanyama*
- Department of Physics, Copperbelt University, PO Box 21692, Jambo Drive, Riverside, Kitwe 10101, Zambia; Department of Physics, University of Cape Town, Rondebosch 7700, South Africa
- Craig M. Comrie#
- iThemba LABS, National Research Foundation, PO Box 722, Somerset West 7129, South Africa; Department of Physics, University of Cape Town, Rondebosch 7700, South Africa
The down-scaling of nanoelectronic devices to ever smaller dimensions and greater performance has pushed silicon-based devices to their physical limits. Much effort is currently being invested in research to examine the feasibility of replacing Si by a higher mobility semiconductor, such as germanium, in niche high-performance metal oxide semiconductor (MOS) devices. Before Ge can be adopted in industry, a suitable contact material for the active areas of a transistor must be identified. It is proposed that platinum group metal (pgm) germanides be used for this purpose, in a similar manner as metal silicides are used in Si technology. Implementation of Ge-based technology requires a thorough understanding of the solid-state interactions in metal/Ge systems in order to foresee and avoid problems that may be encountered during integration. We present a systematic study of the solid-state interactions in germanide systems of four of the pgms: iridium, platinum, palladium and rhodium. Our approach was essentially twofold. Firstly, conventional thin film couples were used to study the sequence of phase formation in the germanide systems. Conventional thin film couples were also used to identify and monitor the dominant diffusing species during the formation of some of the germanides as these can influence the thermal stability of a device. Secondly, we observed and analysed several aspects of the lateral diffusion reactions in these four systems, including activation energies and diffusion mechanisms. Lateral diffusion couples were prepared by the deposition of thick rectangular islands of one material on to thin films of another material. Rutherford backscattering spectrometry (RBS) and microprobe-Rutherford backscattering spectrometry (μRBS) were used to analyse several aspects of the thin film and lateral diffusion interactions respectively. X-ray diffraction (XRD) and scanning electron microscopy (SEM) were also employed.
Ge has several attractive properties such as high mobility of charge carriers and very low carrier freeze-out temperatures (1–4). There is currently much research on high mobility semiconductors, such as Ge, with the view of using them to replace Si in niche high-performance MOS devices (5–7). Before Ge can be adopted by industry a suitable contact material to the active areas (source, drain and gate) of a transistor must be identified. It is proposed that pgm germanides be used for this purpose, similar to the manner in which metal silicides are used in present Si technology (6, 8). Our work investigates the solid-state interactions in germanide systems of four of the pgms: Ir, Pt, Pd and Rh, in thin film and lateral diffusion couples.
In the design of transistors the contact material should be stable over a wide temperature range. Conventional thin film couples are well suited for investigating the phase formation sequence and temperature stability of the phases of a system. We have also used thin film couples to identify some of the dominant diffusing species during the phase formation. For device integrity it is important to identify the dominant diffusing species during the formation of the respective germanides as this can influence their thermal stability.
The samples used for studying lateral diffusion reactions were composed of a thick island of one material on top of a much thinner film of another material. Upon annealing the island material would react with the underlying film through vertical diffusion, going through a sequence of phases until the most island-material rich phase is formed. Since no further vertical reaction with the underlying film is possible, the most island-material rich phase may then grow laterally until it attains a critical width, after which other phases appear and grow simultaneously. This is a case of multiple phase formation as would be found in bulk diffusion couples. Lateral diffusion couples thus provide the transition between thin film and bulk behaviour.
Since the island material is abundant for diffusion in the lateral diffusion couples, phase formation and reaction kinetics can be studied to a greater extent than in thin film planar structures. Lateral diffusion structures can be used to simulate bulk diffusion couples because phase formation could extend to lengths of around 100 μm (9). In kinetic studies of thin film planar structures the diffusion lengths are typically less than 0.5 μm. One can therefore study the transition from thin film to bulk diffusion couple behaviour. The study of lateral diffusion couples is particularly well suited for dealing with the challenges of achieving the required lateral abruptness of semiconductor junctions. Excessive diffusion of the substrate element, in this case Ge, during germanide formation could result in overgrowth and bridging in devices (10).
Various early techniques were developed to study lateral diffusion couples (11–20). In later studies, μRBS was used (9, 21, 22). The major advantage of this technique is its ability to give depth information.
Some previous work has been carried out in the research field of pgm/Ge junctions. Saedi et al. (23) reported a scanning tunnelling microscopy (STM) and spectroscopy study of the formation of Pt/germanide phases on Ge (111). This study gave a demonstration of the structural dependence of electronic properties in the Pt-Ge system. Schottky barrier diodes have been used in many applications such as gates for metal semiconductor field-effect transistors (MESFET), solar cells and detectors (24–27). A reduction of the PtGe/Ge electron Schottky barrier height by rapid thermal diffusion of phosphorus was reported by Henkel et al. (28). The results showed that rapid thermal diffusion from a solid diffusion doping source was effective in reducing Schottky barrier heights of Pt germanide Schottky barrier diodes on Ge. Chawanda et al. (29) investigated the change in the current-voltage (I–V) electrical properties of Pt Schottky contacts on Ge (100) at different annealing temperatures. Their results showed that the as-deposited barrier heights had values that were near the band gap of Ge for Pt/Ge (100) Schottky diodes resulting in good Schottky source/drain contact materials in p-channel Ge-MESFETs for the hole injection from source into inverted p-channel (30). Chawanda et al. (31) also studied the electrical properties of Pd Schottky contacts with Ge (100). I–V and capacitance-voltage (C–V) measurements were performed under various annealing conditions. Only one Pd germanide phase, PdGe, was formed. A hole trap at 0.33 eV above the valence band was observed after annealing at 300°C. In another study Chawanda et al. (32) investigated the change in the electrical properties of Ir Schottky barrier diodes on Ge (100). Electrical characterisation of these contacts using I–V and C–V measurements was performed under various annealing conditions. Thermal stability of the Ir/Ge (100) sample was observed up to an annealing temperature of 500°C. The results also showed that the onset temperature for agglomeration (binding of primary particles leading to phase formation) in 20 nm Ir/Ge (100) samples occurs between 600–700°C.
Gaudet et al. (7) carried out a systematic study of thermally induced reactions of 20 transition metals with Ge substrates. They monitored metal/Ge reactions in situ during ramp anneals at 3°C s–1 using time-resolved XRD and diffusion light scattering. They also carried out resistance measurements. Their results showed that the pgms Pd and Pt were among the six most promising candidates for microelectronic applications, the other candidates being nickel, cobalt, copper and iron. Ni is the most used metal for reducing contact resistance. An example of previous research in the area of Ni/Ge junctions is the work reported by Peng et al. (33) on the I–V characteristics of Ni/Ge (100) Schottky diodes and the Ni germanide induced strain after subjecting the Schottky contacts to rapid thermal annealing in the temperature range of 300–600°C. Their results showed that the orthorhombic structure of NiGe induces epitaxial tensile strain on Ge substrates due to the difference in lattice constants. They also suggested that the increase in barrier height with increasing annealing temperature may have been due to the conduction band edge shift by the strain after the germanidation process.
Hallstedt et al. (34) studied the phase transformation and sheet resistance of Ni on single crystalline SiGe(C) layers after annealing treatments at 360–900°C. The role of strain relaxation or compensation in the reaction of Ni on Si1–x –y Gex Cy layers due to Ge or carbon out-diffusion to the underlying layer during the phase transformation was investigated. Formation of crystalline Ni(SiGe) was complete at 400–450°C but the thermal stability decreased rapidly with increased Ge amount due to agglomeration. This thermal behaviour was shifted to higher annealing temperatures when C was incorporated. Ni(SiGeC) layers formed at 500–550°C after which there was Ge segregation to the underlying layer and C accumulation at the interface. Thanailakis et al. (35) established a relationship between as-deposited Pd/Ge (111) and Ni/Ge (111) Schottky barrier height values, the metal work functions and the density of surface states of the Ge substrate.
A study of the phase formation sequence, using conventional thin film couples, was carried out prior to the lateral diffusion study. This study was carried out using RBS and XRD for phase identification. Thermally oxidised single crystal Si wafers with a (100) crystal orientation were used as substrates in all studies. In the studies of the Ir-Ge system, a thin layer of titanium (2 nm) had to be deposited onto the SiO2 prior to the deposition of the coupling materials; this Ti layer reacted with SiO2 forming a ‘glue’ without which the structure could not adhere. Electron beam vacuum deposition of coupling layers of metal and Ge was carried out at pressures in the low 10–5 Pa range.
In a further preliminary investigation a marker technique was used to monitor atomic mobility during phase formation. The term marker refers to a material deposited in the sample as a reference plane in monitoring the direction of flow of atoms. A thin layer of Ti acted as an inert marker interposed between coupling layers of metal and Ge; the Ge layer being at the surface of the sample. Upon annealing of this structure both Ge and metal atoms could have diffused across the marker at different rates causing it to shift in position towards the dominant diffusing species. The marker Ti signal was monitored by RBS for different annealing times. The dominant diffusing species (DDS) during phase formation was determined by observing the relative shift of the marker.
The lateral diffusion couples were also prepared by electron beam evaporation at a base pressure in the low 10–5 Pa range. A thin film of one material was deposited first. An ordinary Si wafer with an array of 390 × 780 μm2 rectangular windows (referred to as a Si mask), made by photolithographic techniques and selective etching, was then brought into contact with the substrate without breaking vacuum. Island material was deposited through the mask to form structures consisting of metal islands on Ge films and Ge islands on metal films.
Figure 1 shows a schematic illustration of the sample preparation setup. The source metal or Ge sample pieces were each placed in one of the three crucibles on a water-cooled copper hearth. The positions of the crucibles could be adjusted from the outside of the electron-beam evaporation chamber so that each material to be evaporated could be placed in the path of the electron beam in turn. Each crucible was shielded from the adjacent one by a 2 cm high partition to prevent cross-contamination during deposition. Above the crucibles was a shutter that could be opened and closed from outside the chamber using a bar magnet. The sample changer, which accommodated a maximum of three sample holders, hovered above the shutter. The sample changer could be rotated using an external handle in such a way as to place one sample holder at a time in the line of sight of the target vapour. In this way, it was possible to prepare up to three sets of different samples (each on one sample holder) in one experimental run. Between the shutter and the sample changer hung the Si mask holder which had a provision where a Si wafer with several rectangular 390 × 780 mm2 openings could be placed. The mask holder could be swung from side to side without breaking the vacuum. Likewise, its height could be adjusted externally. The main features of the bottom compartment of the chamber were a system of vacuum pumps and a cryopanel. A baffle valve is closed between the upper and lower compartments to ensure that the lower section of the system is maintained under vacuum when the system is not in use or during sample changing.
After removing the samples with islands from the evaporation chamber they were cleaved into twelve identical samples, each with two or three islands. Furnace-annealing was used to activate solid-state interaction after which the samples were analysed. The samples were examined using SEM to distinguish the various reaction regions and measure their diffusion lengths. Representative samples were selected for further analysis by μRBS. The distribution of elements as a function of lateral position was obtained using μRBS. This technique also provided information regarding the elemental distribution as a function of depth and the thickness of the films. A 2 MeV α-beam focused down to a pixel size of about 1 × 1 μm was scanned across a well-defined area of the samples. This area, typically of 400 × 400 μm2, was chosen to include all reaction regions observed in the SEM micrographs. Sample orientation was adjusted in such a way that the interfaces of the regions of interest lay horizontally in line with the original island edge so that the microprobe beam scanned parallel to the original island interface. RBS spectra were recorded along with position information. Typically, 128 × 128 spectra were generated in each run and once it had been established that no variation in composition was observed parallel to the interface, spectra were summed along this direction. This reduced the analysis to a one-dimensional traverse of 128 spectra perpendicular to the interface, thereby optimally exploiting beam position while improving on statistics. The RBS data was analysed using RUMP (36) simulations.
3.1 Iridium-Germanium System
3.1.1 Thin Film Couples
Ir-Ge conventional thin film couple samples for the study of the phase formation sequence had the structure: SiO2/Ge (550 nm)/Ir (90 nm). The samples were annealed in vacuum for various periods of time at temperatures ranging from 350°C to 800°C. The results showed the appearance of the compounds IrGe and Ir4Ge5 at annealing temperatures of around 350°C. Observation of the separate formation of either compound as a first phase could not be achieved. An XRD spectrum obtained following annealing at 400°C for 80 minutes is shown in Figure 2.
The figure shows the presence of IrGe and Ir4Ge5 together with unreacted Ir indicating co-existence. After these two phases, Ir3Ge7 formed; from our data it was not possible to tell whether all the Ir was consumed before Ir3Ge7 appeared. The phase IrGe4 is the most Ge rich in the Ir-Ge system, it was the final phase observed and only formed above 800°C. There was no evidence of the presence of Ir3Ge4 during the whole reaction.
A 1.2 nm layer of Ti acted as an inert marker interposed between coupling layers of Ir and Ge to monitor atomic mobility during phase formation. This structure was annealed at 400°C for different times. Figure 3 is a diagram showing the results obtained from a RUMP simulation (36) on an RBS spectrum of a sample annealed at 400°C for 20 minutes. The results show the Ti marker as lying between the phase Ir4Ge5 and the unreacted Ge at the surface.
Figure 4 shows the Ti ‘glue’ and Ti marker signals before and after annealing. The Ti marker signal shifted by 0.02 MeV to higher energies, i.e. towards the surface. From these results a conclusion can be drawn if certain assumptions are made. Firstly, if IrGe was the first phase to form, then Ge would be the sole moving species for both IrGe and Ir4Ge5 formation. This is so because had Ir been a moving species during either IrGe or Ir4Ge5 formation, it would have been observed to diffuse across the marker, which was not the case. Secondly, if Ir4Ge5 were the first phase to form, then all we can be sure of is that Ge was the sole moving species during Ir4Ge5 formation. During IrGe formation there would be no discernible indicator as to which species was moving, the marker being at the interface between Ir4Ge5 and Ge. The only firm conclusion we can draw from the marker results is that Ge was the sole diffusing species during Ir4Ge5 formation.
3.1.2 Lateral Diffusion Couples
The lateral diffusion couple results with Ir islands on Ge films showed little germanide growth upon annealing with a reaction region that was too narrow to properly resolve and monitor. The reverse configuration with Ge islands on Ir films showed substantial lateral diffusion. Figure 5 shows an SEM micrograph of part of a lateral diffusion couple, with a Ge island (250 nm) on an Ir film (25 nm), which had been annealed at 800°C for 30 minutes.
Four different reaction regions were observed and are labelled A to D in Figure 5. The original edge of the island is clearly visible as the right edge of the white vertical strip in region C. This sample then underwent μRBS in order to determine the composition and thickness of the various regions. An area to be scanned was chosen to include all regions observed.
A representative RBS spectrum picked from each of the four reaction regions labelled A to D is shown in Figure 6. A spectrum picked from the unreacted Ir film is also included in Figure 6. Peak heights of the spectra taken from regions D, C and B show the phases Ir4Ge5, Ir3Ge7 and IrGe4 respectively. The germanides in these three regions are seen at the surface position. On the other hand, the Ir peak from region A lies below the surface position, indicating that there was no Ir at the surface. Region A therefore consisted of a layer of unreacted Ge on top of the IrGe4 phase.
The RBS data from the scanned area were analysed to get stoichiometric information from integrated counts of the Ir and Ge peaks. Figure 7 shows the results as a function of lateral position. The figure shows a region of the phase Ir3Ge7 inside the original island interface. The width of this Ir3Ge7 region was observed to increase with annealing time. This suggests a gradual decomposition of the phase IrGe4 into Ir3Ge7 by the reaction 3IrGe4 → Ir3Ge7 + 5Ge.
As in the thin film work, the major difference between the lateral diffusion couples prepared by annealing at temperatures below 800°C and those at and above 800°C was the absence of the IrGe4 phase below 800°C; hence no region of IrGe4 decomposition was observed in the lateral diffusion couples below 800°C.
3.2 Platinum-Germanium System
3.2.1 Thin Film Couples
Pt-Ge conventional thin film couple samples for the study of the phase formation sequence had the structure: SiO2/Ge (500 nm)/Pt (120 nm) or equivalently, SiO2/Ge (2270 × 1015 cm–2)/Pt (780 × 1015 cm–2). These were annealed in vacuum for various periods of time at temperatures ranging from 150°C to 300°C. The results showed the appearance of Pt2Ge as the first phase formed at an annealing temperature of around 190°C. Figure 8 shows an RBS spectrum obtained following annealing at 190°C for 2 hours, together with its RUMP simulation. The inset in Figure 8 shows that at this stage of the reaction the sample consisted of Si covered with SiO2 as the substrate, followed by Ge, then Pt2Ge and some unreacted Pt at the surface. The results shown in Figure 8 were consistent with results obtained from the XRD analysis performed on this same sample.
Pt3Ge2 was observed as the second phase to form at 220°C. Figure 9 shows an XRD spectrum of a sample annealed at 220°C for 80 minutes indicating the presence of Pt3Ge2. The result was consistent with the result of a RUMP simulation on RBS data obtained from the same sample. From our data it was not possible to tell whether all the Pt was consumed before Pt3Ge2 appeared. The next detected phase was PtGe at 250°C. The last phase observed was PtGe2 which formed from the interaction of PtGe with unreacted Ge.
A 1.2 nm layer of Ti acted as an inert marker interposed between coupling layers of Pt (300 × 1015 cm–2) and Ge (430 × 1015 cm–2) to monitor atomic mobility during phase formation. Figure 10 shows the results obtained from a RUMP simulation on an RBS spectrum of a sample annealed at 250°C for 20 minutes. The inset shows the Ti marker as lying between two layers of Pt2Ge. We observe some unreacted Ge at the surface and some unreacted Pt below the marker. The results show a significant amount (235 × 1015 cm–2) of PtGe above the marker, between Pt2Ge and unreacted Ge. There is no PtGe below the marker.
Figure 11 shows the Ti marker signals before and after annealing at 250°C for 20 minutes. The Ti marker signal shifted from A to B by 0.025 MeV to lower energies, i.e. towards the substrate. It is clear that both Pt and Ge atoms migrated across the marker to interact, forming Pt2Ge on either side of the marker. The amount of Pt that crossed the marker and was observed above it, in units of 1015 cm–2, is shown in Equation (i):
The amount of Ge that crossed the marker and was observed below it is shown in Equation (ii):
The atomic diffusion ratio of Pt to Ge is therefore about 4 to 1.
There are two possible mechanisms by which the PtGe above the marker could have been formed. These are: Pt2Ge → PtGe + Pt and Pt2Ge + Ge → 2PtGe. No PtGe was seen below the marker; the first of these two processes did not take place in that region as it would have left some PtGe there. It is therefore unlikely for the first process to have taken place above the marker. The second reaction is the one most likely to have taken place above the marker. The PtGe was therefore formed from Pt2Ge and Ge above the marker where there was no discernible indicator as to which species was moving during its growth. The phase Pt3Ge2 was not observed to form in the presence of the Ti marker; PtGe was formed after Pt2Ge skipping Pt3Ge2, which was observed in the absence of the Ti marker during the study of the phase formation sequence using the sample structure: SiO2/Ge (500 nm)/Pt (120 nm).
3.2.2 Lateral Diffusion Couples
The lateral diffusion couple results for Pt islands on Ge films showed little lateral diffusion upon annealing with a reaction region that was too narrow to resolve and monitor properly. The reverse configuration with Ge islands on Pt films showed substantial lateral germanide growth. Figure 12 shows an SEM micrograph of part of a lateral diffusion couple of a Ge island (145 nm) on a Pt film (35 nm) annealed at 450°C for 24 hours. The figure shows a part of the island that is close to one of its corners. Five different regions, labelled A to E, are observed where A starts in the middle of the island and E is the outermost region. The original edge of the island is clearly visible between regions C and D.
The sample in Figure 12 underwent μRBS analysis to obtain stoichiometric information by RUMP simulation of the Pt and Ge peaks. The scanned area was chosen to include all regions observed in the SEM micrograph. Figure 13 shows the results as a function of lateral position.
The regions B, C, D and E are found to consist of PtGe2, Pt2Ge3, PtGe and unreacted Pt respectively. Region A appeared to be composed of a mixture of PtGe2 and unreacted Ge. To show this more clearly, a representative spectrum picked from each of the five regions labelled A to E is shown in Figure 14. Downward pointing arrows are used to indicate the surface positions of Ge and Pt. The spectrum from region E shows a peak of unreacted Pt and no Ge. Peak heights of the spectra taken from regions D, C and B show the phases PtGe, Pt2Ge3 and PtGe2 respectively. The Ge in these three regions is seen at the surface position. It is seen from the solid line in the figure that the region A consisted of unreacted Ge and the phase PtGe2. The Pt peak of the solid line lies at the surface position. This shows that there was some Pt at the surface. The ‘shoulder’ marked in the figure however indicates that there was less Pt at the surface than deeper down. Region A therefore consisted of PtGe2 at the bottom while at the top there was unreacted Ge together with the phase PtGe2.
3.3 Palladium-Germanium System
3.3.1 Thin Film Couples
Of major concern while determining the best thickness of our Pd-Ge sample structure was the likelihood of overlap between Pd and Ge RBS signals. This is by virtue of their having atomic numbers that lie relatively close to each other in the periodic table. At the same time, samples needed to comprise of elemental layers thick enough to induce an appreciable X-ray yield. The structure used was SiO2/Ge (500 nm)/Pd (70 nm).
In this system, reaction was induced at relatively low temperatures. Figures 15 and 16 show RBS spectral data for samples annealed at temperatures of 100°C and 150°C for 2 hours and 80 minutes respectively. Data from XRD analysis were also obtained, these are displayed alongside the corresponding RBS data. Layer thicknesses obtained by RUMP simulations (solid lines) are shown. Rather straightforward behaviour is observed in this system with the two congruent phases Pd2Ge and PdGe being the only ones observed. Pd2Ge was the first to form at around 100°C.
3.3.2 Lateral Diffusion Couples
Samples for lateral diffusion study were prepared by deposition of thick Ge islands on thin Pd films. This configuration was chosen on the basis of the results observed in the Ir-Ge and Pt-Ge systems. Several lateral diffusion samples were annealed at various temperatures for different lengths of time. The Pd-Ge system exhibited relatively low temperature reaction. It was therefore necessary to carry out the investigation for this system at much lower temperatures than those used for the other systems.
Figure 17 shows an SEM micrograph of one representative sample with a 100 nm thick Ge island on a 20 nm thick Pd film, annealed at 325°C for 2 hours, showing three distinct regions labelled A to C. Areas which were chosen to include all the reaction regions observed were scanned on the nuclear microprobe for analysis by μRBS. A spectrum picked from each of the three regions of the Pd-Ge lateral diffusion sample is shown in Figure 18.
The spectrum from region C shows a peak of unreacted Pd and no Ge. The peak height of the spectrum taken from the region B shows the phase Pd2Ge. This phase is seen at the surface position. From the solid line in the figure, it can be seen that region A consisted of unreacted Ge and PdGe. The Pd peak of the solid line lies at the surface position, showing that there was some Pd at the surface. The ‘shoulder’ marked in the figure indicates that there was less Pd at the surface than deeper down. Region A therefore consisted of PdGe at the bottom while at the top there was unreacted Ge intermingled with PdGe.
3.4 Rhodium-Germanium System
3.4.1 Thin Film Couples
Only four equilibrium phases exist in the Rh-Ge system: Rh2Ge, Rh5Ge3, RhGe and Rh17Ge22. The chance of getting excessive overlap of RBS peaks was greater in this system than in any other yet reported on in this study; therefore great care was taken to abate this. The sample structure used in this study was SiO2/Ge (500 nm)/Rh (60 nm).
Figures 19, 20 and 21 show XRD results alongside RBS spectra with RUMP simulations for samples with the structure SiO2/Ge (500 nm)/Rh (60 nm) annealed between 320°C and 400°C. Our RBS data strongly suggest the formation of the non-congruent phase Rh2Ge as the first phase but there is no firm evidence of this from the X-ray data. RBS data showed that the Rh2Ge started to form around 280°C and later gave way to RhGe. Figure 21 shows that Rh17Ge22 was formed after the RhGe phase. Only RhGe and Rh17Ge22 peaks were observed in the X-ray results, probably because the Rh2Ge layer was too thin to induce a good X-ray yield.
3.4.2 Lateral Diffusion Couples
Samples for lateral diffusion study were prepared by deposition of thick Ge islands on thin Rh films. This configuration, as for the Pd-Ge lateral diffusion samples, was chosen on the basis of the results observed in the Ir-Ge and Pt-Ge systems. Several samples were annealed at various temperatures while time was monitored in the usual way. Shown in Figure 22 is an SEM micrograph of one representative sample, with a 100 nm Ge island on a 20 nm Rh film, annealed at 600°C for 15 minutes. The figure shows four reaction regions labelled A to D. Areas which were chosen to include all the reaction regions observed were scanned on the nuclear microprobe for analysis by μRBS.
RBS spectra picked from each of the four regions, A, B, C and D are shown in Figure 23. The spectrum from region D shows a peak of unreacted Rh and no Ge. Peak heights of the spectra taken from regions C and B show the phases RhGe and Rh17Ge22 respectively. The germanides in these two regions are seen at the surface position. It can be seen from the solid line in the figure that the region A consisted of unreacted Ge and the phase Rh17Ge22. The Rh peak of the solid line lies below the surface position. This shows that there was no Rh at the surface. Region A consisted of unreacted Ge overlaying the phase Rh17Ge22. For very long annealing times at relatively high temperatures (around 600°C and above) the phase RhGe was observed to slowly stretch across the original interface into the island region.
In this work we used conventional thin film as well as lateral diffusion couples to study the germanide systems of four pgms: Ir, Pt, Pd and Rh.
4.1 Iridium-Germanium System
4.1.1 Thin Film Couples
Using conventional thin film couples, IrGe and Ir4Ge5 where observed to be the first phases to form in the Ir-Ge system and co-existed at annealing temperatures of around 350°C. Ir3Ge7 formed after these two phases while the IrGe4 phase only appeared above 800°C.
By interposing a thin layer of Ti (1.2 nm) to act as an inert marker between coupling layers in the Ir-Ge system, the direction of atomic mobility was successfully monitored during the initial stages of the reaction. In the marker samples, IrGe and Ir4Ge5 were again found to coexist from the first stages of reaction. The movement of the marker indicated that Ge was the sole moving species during Ir4Ge5 formation. It was not certain whether Ge was also the sole moving species during IrGe formation.
4.1.2 Lateral Diffusion Couples
The phases observed to form in lateral diffusion couples of the Ir-Ge system were the same as those observed in the thin film study on this system with the exception of IrGe, i.e. IrGe4, Ir3Ge7 and Ir4Ge5. The phase Ir3Ge7 was seen to stretch across the original island interface at all temperatures. As in the results of the thin film couples, the phase IrGe4 was only observed to nucleate at temperatures above 800°C.
The graph in Figure 24 is a plot of the growth in width with annealing time for a sample annealed at 800°C. In this figure the growth in width of the Ir3Ge7 phase region is labelled as Xβ while the width of the Ir4Ge5 region is labelled as Xγ. It should be pointed out that the origin in Figure 24 does not correspond to the original island interface but refers to the interface between the IrGe4 and Ir3Ge7 regions. The growth curve for the region labelled as Ir3Ge7 was a result of Ir3Ge7 grown from Ir4Ge5 and that from the decomposition of IrGe4. The growth characteristics observed are parabolic with time.
The growth of the phases Ir3Ge7 and Ir4Ge5 were monitored at temperatures of 700°C, 750°C and 800°C. Different annealing times were chosen to obtain a reasonable range of growth widths at each of the three temperatures. The squares of the growth widths were plotted against the times of annealing at each temperature and the diffusional growth constants, Kβ, were obtained from the slopes. The logarithms of the diffusional growth constants were then plotted against the reciprocals of the product, kbT, of the Boltzmann constant and the absolute temperatures. The resulting Arrhenius plot for the phase Ir4Ge5 is shown in Figure 25. The average activation energy, Ea, was obtained from the slope of the straight line fit of the Arrhenius plot. The value determined for the diffusion process in the Ir4Ge5 phase was Ea = 1.6 ± 0.1 eV.
Unfortunately the results from the samples annealed at the temperatures 700°C, 750°C and 800°C could not be used to obtain a value of the activation energy for the Ir3Ge7 phase. This was because the mechanisms at play during Ir3Ge7 growth were not the same at 800°C as at temperatures below. At 800°C the Ir3Ge7 observed inside the island region was from the decomposition of IrGe4 while that outside was formed by the interaction between Ir4Ge5 and the outward diffusing Ge atoms. Below 800°C the Ir3Ge7 outside was also formed by the interaction of Ge and Ir4Ge5 but that inside was observed as a result of exposure as Ge was consumed from the source region. On the other hand, Ir4Ge5 was formed by the interaction of Ge with the unreacted Ir film, both below and above 800°C and thus at all three temperatures.
4.2 Platinum-Germanium System
4.2.1 Thin Film Couples
In our work on the Pt-Ge system, Pt2Ge was the first phase formed. The second phase observed was Pt3Ge2. The next phase detected was PtGe and the last was PtGe2. The non-congruent phase Pt2Ge3 was skipped between the last two phases.
The marker technique was also applied to the Pt-Ge system. Pt was found to be the dominant diffusing species during Pt2Ge formation. Some Ge diffusion was also observed to take place. The atomic diffusion ratio of Pt to Ge was measured as being about 4 to 1.
4.2.2 Lateral Diffusion Couples
Three phases were observed to form in the lateral diffusion couples of the Pt-Ge system: PtGe2 and Pt2Ge3 inside the original island region and PtGe outside. The Pt2Ge and Pt3Ge2 phases which were observed in the thin film study of the system were absent in the lateral diffusion couples. The graph in Figure 26 is a plot of the various growth widths against the time of annealing for a sample annealed at 550°C. The origin in this figure does not correspond to the original island interface but refers to the interface between the regions labelled as A and B in Figure 12. The growth widths of the phase regions, PtGe2, Pt2Ge3 and PtGe are labelled as Xα, Xβ and Xγ respectively.
It must be pointed out that whereas the lateral growth of the Pt2Ge3 and PtGe in the regions labelled as C and D in Figure 12 respectively are due to reaction mechanisms, the growth of region B is due to exposure of PtGe2 by the consumption of overlaying Ge. The growth characteristics observed for all regions are parabolic with time.
The lateral widths, Xα, Xβ and Xγ, were monitored at the temperatures 450°C, 500°C and 550°C. Different annealing times were chosen to obtain a reasonable range of growth widths at each of these three temperatures. The squares of the growth widths were plotted against the times of annealing at each temperature and the diffusional growth constants, Kβ, were obtained from the slopes. The logarithms of the diffusional growth constants were then plotted against the reciprocals of the product, kbT, of the Boltzmann constant and the absolute temperatures. The resulting Arrhenius plots for the phases PtGe2, Pt2Ge3 and PtGe are shown in Figure 27.
The activation energies, Ea, were obtained from the slopes of the straight line fits. The average activation energies determined from the lateral growth rates of Pt2Ge3 and PtGe were 0.9 ± 0.1 eV and 1.3 ± 0.4 eV, respectively. The activation energy corresponding to the apparent lateral growth rate of the PtGe2 region was 1.5 ± 0.2 eV.
4.3 Palladium-Germanium System
4.3.1 Thin Film Couples
The only phases observed to form in the thin film study of the Pd-Ge system were the two congruent phases PdGe and Pd2Ge.
4.3.2 Lateral Diffusion Couples
The two phases, PdGe and Pd2Ge, which were observed to form in the thin film study of the Pd-Ge system were also the only two observed in the lateral diffusion couples. The growth region outside the original island interface, labelled as region B in Figure 17, consisted of the phase Pd2Ge. The original island region, labelled as region A in Figure 17, consisted of PdGe at the bottom while at the top there was unreacted Ge intermingled with PdGe. There was no region of completely exposed PdGe without intermingled unreacted Ge, it was therefore not possible to obtain data for the growth or exposure rates of the phase PdGe.
The growth of the Pd2Ge region was monitored at the temperatures 275°C, 300°C and 325°C. The periods of annealing were chosen so as to obtain a reasonable range of growth widths at each of the three temperatures; results are presented in Figure 28. Parabolic growth characteristics were observed. Arrhenius plots were obtained from the data presented in Figure 28, in the same way as explained for the Ir-Ge and Pt-Ge lateral diffusion couples. Figure 29 is an Arrhenius plot showing the temperature dependence of the Ge diffusion rate through Pd2Ge. The average activation energy determined from the plot in Figure 29 was Ea = 1.0 ± 0.1 eV.
4.4 Rhodium-Germanium System
4.4.1 Thin Film Couples
Our RBS data strongly suggested the formation of Rh2Ge as the first phase in the Rh-Ge system but there was no firm evidence of this from the X-ray data because this phase was too thin to give a good X-ray yield. This was a consequence of the fact that the samples had to be kept thin enough to avoid excessive RBS peak overlap. The Rh2Ge phase appeared to give way to RhGe while Rh17Ge22 was formed as the last phase.
4.4.2 Lateral Diffusion Couples
The phase Rh17Ge22 was observed inside the original island region in lateral diffusion couples of the Rh-Ge system while RhGe grew outside in most cases. Under certain conditions, the latter phase was observed to slowly stretch across the original interface into the island region. This suggested a slow decomposition of Rh17Ge22 into RhGe. There were four distinct regions observed in the lateral diffusion couples of the Rh-Ge system; these are represented schematically in Figure 30. The relative position of the original island interface and different phases are shown. Knowledge of the position at which the original interface lay was vital in the analysis of the reactions taking place between different phase regions. This was particularly so in this system with wide reaction regions and a slight shift of the reaction interface between the Rh17Ge22 and RhGe regions with respect to the original interface for different times of annealing.
Our RBS data from the thin film study of this system strongly suggested the formation of the non-congruent phase Rh2Ge but this phase was not observed in the lateral diffusion couples. The growths of Rh17Ge22 and RhGe in the lateral diffusion couples were monitored at the temperatures 450°C, 500°C and 600°C. The temperature range and annealing times were chosen in such a way that the decomposition of Rh17Ge22 into RhGe was not significant. Results for carefully chosen annealing times at 500°C are presented in Figure 31. The growth characteristics were observed to be parabolic.
Arrhenius plots obtained from the data shown in Figure 31 are presented in Figure 32, showing the temperature dependence of the Ge diffusion rate through RhGe and Rh17Ge22. An average activation energy of 1.2 ± 0.3 eV was obtained for both Rh17Ge22 and RhGe.
5. Summary and Conclusion
The results of our thin film study are summarised in Table I. Temperatures at which the first reactions were observed to begin are indicated.
The lateral diffusion samples used in this study were prepared with the configuration of Ge islands evaporated onto metal films. In all systems, the germanide phases were seen to spread out from the source region in decreasing order of Ge content. The growth characteristics observed in all phases of the four systems studied were parabolic with time. This is indicative of a diffusion controlled process which, as modelled by Kidson (37), results in parabolic growth even in multiphase systems.
Table II gives a summary of the various phases observed in the lateral diffusion couples of each system and the corresponding activation energies obtained.
The magnitudes of the activation energies, Ea, calculated for all phases, as shown in Table II, suggest that the lateral diffusion reactions in all four systems were not driven by surface diffusion but rather by diffusion through the interior of the lateral diffusion couples; typical values for surface diffusion being around 0.6 eV (38).
In the current design and processing of transistors the contact material should exhibit low sheet and contact resistances, form at a low temperature and be stable over a wide temperature range. A previous systematic study of the thermally induced reaction of a large number of transition metals with Ge substrates revealed that NiGe and PdGe are the most promising candidates when taking the above requirements into account (7, 8). From Table I we see that Pd germanides form at lower temperatures than any of the other pgm germanides. Pt germanides were also found to be promising candidates for microelectronic applications (7). There has been very little work carried out on the electrical properties of Ir/Ge junctions (32). Our survey of previous work in this research field showed no evidence of any systematic study of the electrical properties of Rh/Ge junctions. Our work has looked at several aspects of the interfacial phase growth and inter-diffusion kinetics at pgm/Ge junctions. However, more work needs to be carried out regarding the electrical properties of these junctions. Our current results could then be used in conjunction with the results based on the study of electrical properties in order to draw more insight and to make comprehensive suggestions on the ideally suited choice of pgm/Ge combinations for particular applications in semiconductor technology applications such as gates for metal semiconductor field-effect transistors, solar cells and detectors.
C. O. Chui, H. Kim, D. Chi, B. B. Triplett, P. C. McIntyre and K. C. Saraswat, ‘A Sub-400/spl deg/C Germanium MOSFET Technology with High-/spl Kappa/Dielectric and Metal Gate’, International Electron Devices Meeting, San Francisco, USA, 8th–11th December, 2002, Institute of Electrical and Electronics Engineers Inc, Piscataway, USA, pp. 437–440 LINK https://doi.org/10.1109/IEDM.2002.1175872
A. Ritenour, S. Yu, M. L. Lee, N. Lu, W. Bai, A. Pitera, E. A. Fitzgerald, D. L. Kwong and D. A. Antoniadis, ‘Epitaxial Strained Germanium p-MOSFETs with HfO/sub 2/Gate Dielectric and TaN Gate Electrode’, International Electron Devices Meeting, Washington, DC, USA, 8th–10th December, 2003, Institute of Electrical and Electronics Engineers Inc, Piscataway, USA, pp. 18.2.1–18.2.4 LINK https://doi.org/10.1109/IEDM.2003.1269315
H. Shang, H. Okorn-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’, International Electron Devices Meeting, San Francisco, USA, 8th–11th December, 2002, Institute of Electrical and Electronics Engineers Inc, Piscataway, USA, pp. 441–444 LINK https://doi.org/10.1109/IEDM.2002.1175873
C. O. Chui, S. Ramanathan, B. B. Triplett, P. C. McIntyre and K. C. Saraswat, IEEE Electron Dev. Lett., 2002, 23, (8), 473 LINK https://doi.org/10.1109/LED.2002.801319
C. Claeys and E. Simoen, “Germanium-Based Technologies: From Materials to Devices”, Elsevier BV, Oxford, UK, 2007, 449 pp
D. P. Brunco, B. De Jaeger, G. Eneman, J. Mitard, G. Hellings, A. Satta, V. Terzieva, L. Souriau, F. E. Leys, G. Pourtois, M. Houssa, G. Winderickx, E. Vrancken, S. Sioncke, K. Opsomer, G. Nicholas, M. Caymax, A. Stesmans, J. Van Steenbergen, P. W. Mertens, M. Meuris and M. M. Heyns, J. Electrochem. Soc., 2008, 155, (7), H552 LINK https://doi.org/10.1149/1.2919115
S. Gaudet, C. Detavernier, A. J. Kellock, P. Desjardins and C. Lavoie, J. Vac. Sci. Technol. A, 2006, 24, (3), 474 LINK https://doi.org/10.1116/1.2191861
J. A. Kittl, K. Opsomer, C. Torregiani, C. Demeurisse, S. Mertens, D. P. Brunco, M. J. H. Van Dal and A. Lauwers, Mater. Sci. Eng.: B, 2008, 154–155, 144 LINK https://doi.org/10.1016/j.mseb.2008.09.033
R. S. Nemutudi, C. M. Comrie and C. L. Churms, Thin Solid Films, 2000, 358, (1–2), 270 LINK https://doi.org/10.1016/S0040-6090(99)00679-3
S.-L. Zhang and M. Östling, Crit. Rev. Solid State Mater. Sci., 2003, 28, (1), 1 LINK https://doi.org/10.1080/10408430390802431
L. R. Zheng, L. S. Hung, J. W. Mayer, G. Majni and G. Ottaviani, Appl. Phys. Lett., 1982, 41, (7), 646 LINK https://doi.org/10.1063/1.93635
L. R. Zheng, L. S. Hung and J. W. Mayer, J. Vac. Sci. Technol. A, 1983, 1, (2), 758 LINK https://doi.org/10.1116/1.571994
L. R. Zheng, L. S. Hung and J. W. Mayer, Thin Solid Films, 1983, 104, (1–2), 207 LINK https://doi.org/10.1016/0040-6090(83)90563-1
S. H. Chen, L. R. Zheng, J. C. Barbour, E. C. Zingu, L. S. Hung, C. B. Carter and J. W. Mayer, Mater. Lett., 1984, 2, (6), 469 LINK https://doi.org/10.1016/0167-577X(84)90075-2
B. Blanpain, J. W. Mayer, J. C. Liu and K. N. Tu, J. Appl. Phys., 1990, 68, (7), 3259 LINK https://doi.org/10.1063/1.346377
B. Blanpain, J. W. Mayer, J. C. Liu and K. N. Tu, Phys. Rev. Lett., 1990, 64, (22–28), 2671 LINK https://doi.org/10.1103/PhysRevLett.64.2671
B. Blanpain, ‘Lateral Diffusion Couples and Their Contribution to Understanding Thin Film Reactions’, in “Crucial Issues in Semiconductor Materials and Processing Technologies”, eds. S. Coffa, F. Priolo, E. Rimini and J. M. Poate, Springer Science+Business Media, Dordrecht, The Netherlands, 1992, pp 421–425 LINK https://doi.org/10.1007/978-94-011-2714-1_42
J. C. Liu, J. W. Mayer and J. C. Barbour, J. Appl. Phys., 1988, 64, (2), 651 LINK https://doi.org/10.1063/1.341956
J. C. Liu, J. W. Mayer and J. C. Barbour, J. Appl. Phys., 1988, 64, (2), 656 LINK https://doi.org/10.1063/1.341957
J. C. Liu and J. W. Mayer, J. Mater. Res., 1990, 5, (2), 334 LINK https://doi.org/10.1557/JMR.1990.0334
P. J. Ding, R. Talevi, W. A. Lanford, S. Hymes and S. P. Murarka, Nucl. Instr. Meth. Phys. Res. Sect. B: Beam Int. Mater. Atoms, 1994, (1–4), 85, 167 LINK https://doi.org/10.1016/0168-583X(94)95807-6
H. S. de Waal, “The Effect of Diffusion Barriers, Stress and Lateral Diffusion on Thin-Film Phase Formation”, PhD thesis, University of Stellenbosch, South Africa, 1999
A. Saedi, B. Poelsema and H. J. W. Zandvliet, Surf. Sci., 2011, 605, (5–6), 507 LINK https://doi.org/10.1016/j.susc.2010.12.007
E. Hökelek and G. Y. Robinson, Solid State Electron., 1981, 24, (2), 99 LINK https://doi.org/10.1016/0038-1101(81)90001-0
E. H. Rhoderick, R. H. Williams, “Metal-Semiconductor Contacts”, 2nd Edn., Monographs in Electrical and Electronic Engineering, Vol. 19, Clarendon Press, Oxford, UK, 1988, 252 pp
G. A. Baraff and M. Schlüter, Phys. Rev. B, 1986, 33, (10–15), 7346 LINK https://doi.org/10.1103/PhysRevB.33.7346
S. Asubay, Ö. Güllü and A. Türüt, Appl. Surf. Sci., 2008, 254, (11), 3558 LINK https://doi.org/10.1016/j.apsusc.2007.11.050
C. Henkel, S. Abermann, O. Bethge, G. Pozzovivo, S. Puchner, H. Hutter and E. Bertagnolli, J. Electrochem. Soc., 2010, 157, (8), H815 LINK https://doi.org/10.1149/1.3425750
A. Chawanda, C. Nyamhere, F. D. Auret, W. Mtangi, M. Diale and J. M. Nel, J. Alloys Compd., 2010, 492, (1–2), 649 LINK https://doi.org/10.1016/j.jallcom.2009.11.202
H. B. Yao, C. C. Tan, S. L. Liew, C. T. Chua, C. K. Chua, R. Li, R. T. P. Lee, S. J. Lee and D. Z. Chi, ‘Material and Electrical Characterization of Ni- and Pt-Germanides for p-channel Germanium Schottky Source/Drain Transistors’, Sixth International Workshop on Junction Technology, Shanghai, China, 15th–16th May, 2006, Institute of Electrical and Electronics Engineers Inc, Piscataway, USA, pp. 164–169 LINK https://doi.org/10.1109/IWJT.2006.220884
A. Chawanda, C. Nyamhere, F. D. Auret, W. Mtangi, T. T. Hlatshwayo, M. Diale and J. M. Nel, Physica B, 2009, 404, (22), 4482 LINK https://doi.org/10.1016/j.physb.2009.09.043
A. Chawanda, S. M. M. Coelho, F. D. Auret, W. Mtangi, C. Nyamhere, J. M. Nel and M. Diale, J. Alloys Compd., 2012, 513, 44 LINK https://doi.org/10.1016/j.jallcom.2011.09.053
C.-Y. Peng, Y.-H. Yang, C.-M. Lin, Y.-J. Yang, C.-F. Huang and C. W. Liu, ‘Process Strain Induced by Nickel Germanide on (100) Ge Substrate’, 9th International Conference on Solid-State and Integrated-Circuit Technology, Beijing, China, 20th–23rd October, 2008, Institute of Electrical and Electronics Engineers Inc, Piscataway, USA, pp. 681–683 LINK https://doi.org/10.1109/ICSICT.2008.4734645
J. Hållstedt, M. Blomqvist, P. O. Å. Persson, L. Hultman and H. H. Radamson, J. Appl. Phys., 2004, 95, (5), 2397 LINK https://doi.org/10.1063/1.1645996
A. Thanailakis and D. C. Northrop, Solid State Electron., 1973, 16, (12), 1383 LINK https://doi.org/10.1016/0038-1101(73)90052-X
L. R. Doolittle, Nucl. Instr. Meth. Phys. Res. Sect. B: Beam Int. Mater. Atoms, 1986, 15, (1–6), 227 LINK https://doi.org/10.1016/0168-583X(86)90291-0
G. V. Kidson, J. Nucl. Mater., 1961, 3, (1), 21 LINK https://doi.org/10.1016/0022-3115(61)90175-1
L. Eckertová, “Physics of Thin Films”, 2nd Edn., Plenum Press, New York, USA, 1986
The authors wish to thank the University of Cape Town, the South African National Research Foundation and the University Science, Humanities and Engineering Partnerships in Africa (USHEPIA) for financial assistance. They also wish to thank the Materials Research Group at the iThemba Labs at Faure, South Africa, for the use of their facilities, Miranda Waldron in the Electron Microscope unit at the University of Cape Town, Ms Terry Davies of the X-ray unit in the Geological Science Department, University of Cape Town.
Adrian Habanyama is a Senior Lecturer at the Copperbelt University, Zambia. His research areas are solid-state physics and nanotechnology.
Professor Craig Comrie is an Emeritus Associate Professor at the University of Cape Town, South Africa. His research areas are solid-state physics and nanotechnology.