Metallic thin film on sige si substrates 5

Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature 5.1 Introduction In Chapter 4, we have demonstrated that at low Ni coverages, amorphous NiSi- like, NiGe-like and NiSi0.8Ge0.2-like layers were formed at RT As Ni coverages increased, eventually metallic Ni overlayers were formed on top The aim of Chapter is to use the oxidation process in air to probe the nature of bonding in these amorphous NiSi-like, NiGe-like and NiSi0.8Ge0.2-like layers It may be anticipated that at high Ni coverage, oxidation of Ni may occur, but it is unclear whether Ni oxidation will take place at low Ni coverages where NiSi-like, NiGe-like and NiSi0.8Ge0.2-like layers are proposed to be formed as described in Chapter We will use XPS to monitor the oxidation process of Ni and the substrates (Si/Ge) as a function of oxidation time in air In addition, we will also study the time-dependent oxidation process of these films 5.2 5.2.1 Oxidation of Ni thin films on H-Si(001) substrates Effect of Ni coverage on oxidation of H-Si(001) After a series of Ni thin films with different Ni atomic ratios (Ni%) were deposited onto hydrogen-terminated Si(001) surfaces at RT inside the XPS chamber, they were taken out and exposed to the atmosphere for a set timing before loading back into XPS chamber for rescanning This process has been repeated about one year for a series of sample with different Ni% so that the oxidation behavior of these Ni/H164 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Si(001) samples at RT can be monitored as a function of Ni atomic ratio and oxidation time Two regimes with different behaviors have been identified The first regime is when Ni at% is not more than 33%, while the second regime is when Ni at% is higher than 33% 33%Ni is equivalent to ~13.5 Å of Ni layer according to the growth rate measurement derived from Chapter In the following section, 10%Ni/H-Si(001) and 50%Ni/H-Si(001) will be used as typical examples to illustrate the two regimes’ oxidation behavior, while the oxidation of pure H-Si(001) will also be presented as a reference (10%Ni and 50% Ni are equivalent to 3.6 Å and 20.0 Å of Ni layers, respectively) We will firstly discuss the Si 2p spectra evolution with oxidation time and Ni coverages before we present the change of Ni 2p3/2 spectra with time Fig 5.1 shows the scaled Si 2p spectra’s evolution as a function of exposure time for these two samples as well as for the pure H-Si It can be seen that although not much changes were observed in the Si 2p spectra of 10%Ni/H-Si(001) and pure HSi(001) after one minute exposure to air (Fig 5.1(a,c)), a new peak at 103.0±0.1eV appeared immediately in the spectrum of 50%Ni/H-Si(001) after exposing to air for minute (Fig 5.1(b)), which is attributed to a mixed of Si oxides (Si3+ & Si4+)131 This new peak also appeared in 10%Ni/H-Si and pure H-Si but much later As oxidation time increased, this peak on all three samples slowly shifted towards higher binding energy and eventually it stayed around 103.5±0.1eV, a signature value of SiO2131 It indicates that the suboxides presented on the surface were slowly oxidized into SiO2 over prolonged time This is in good agreement with other researchers’ observation and coincides with the argument that SiO2 is the only stable oxide state for Si94 165 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Si SiO2 106 105 104 103 Si 2p 378 days 166 days 56 days days hours minute As-etched 102 101 100 99 98 97 106 316 days 176 days 56 days days hours minute As-deposited SiO2 Intensity (a.u.) Intensity (a.u.) Si 2p 105 104 103 102 101 100 Si 99 98 97 Binding Energy (eV) Binding Energy (eV) (b) (a) Si Intensity (a.u.) Si 2p SiO2 106 105 104 103 378 days 166 days 56 days days hours minute As-etched 102 101 100 99 98 97 Binding Energy (eV) (c) Fig 5.1 Si 2p spectra evolution as a function of oxidation time at RT for (a) 10%Ni/HSi(001) and (b) 50%Ni/H-Si(001) (c) Si 2p spectra change on hydrogen-terminated Si(001) surface is also presented for comparison The effect of Ni coverage on Si’s oxidation at the same exposure time is illustrated in more details in Fig 5.2, which shows the Si 2p spectra from samples with different Ni coverages after exposure to air for around hours It can be seen clearly from Fig 5.2 (a) that the height of Si oxide peak increased with increase of Ni% till 40% with the same oxidation time (~4hrs) Meanwhile, the B.E of the oxide shifted slowly towards higher value with higher Ni coverage, approaching that of the SiO2 However, when Ni coverage further increased to 50%, there was only a slight increase of SiO2 component and the B.E stayed almost the same as that in 40%Ni As Ni% 166 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature further increased to 75% (Fig 5.2(b)), the Si oxide peaks decreased with Ni coverages and the B.E started to shift towards lower value 106 105 104 103 102 101 50% Ni/H-Si(001) 60% Ni/H-Si(001) 75% Ni/H-Si(001) H-Si(001) Intensity (a.u.) Intensity (a.u.) 50% Ni/H-Si(001) 41% Ni/H-Si(001) 33% Ni/H-Si(001) 20% Ni/H-Si(001) 10% Ni/H-Si(001) H-Si (001) 100 Binding Energy (eV) 99 98 97 106 105 104 103 102 101 100 Binding Energy (eV) (a) 99 98 97 (b) Fig 5.2 The Si 2p spectra of samples with (a) 0%~50% and (b) 50%~75% Ni after exposing to air for ~4 hours In order to quantify the growth of Si oxide as a function of Ni coverage and oxidation time, curve fitting has been carried out on the Si 2p spectra to extract the ratio between peak area of SiO2 and that of Si A typical Si 2p spectrum after curvefitting is shown in Fig 5.3, where the SiO2 peak was clearly separated from elemental Si 2p3/2 by an energy difference of 4.1±0.1eV Si Intensity (a.u.) 2p3/2 106 2p1/2 SiO2 105 104 103 102 101 100 99 98 97 Binding Energy (eV) Fig 5.3 A curve-fitting of Si 2p spectrum from 50%Ni/H-Si(001) after exposure to air for hour 167 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature By applying the same fitting procedures to all the Si 2p spectra from samples with different Ni coverages after a serial of exposure time, the evolution of SiO2/Si ratio as functions of Ni% and oxidation time is shown in Fig 5.4 It can be seen from a linear scale (Fig 5.4(a)) that there are generally two oxidation regimes divided by the dotted blue line for these set of samples In the first regime, the slope of the curve was very steep, indicating that there was extremely fast growth of oxide over short time Hence, this regime with oxidation time less than 15 days is termed as the fast growth regime In the second regime, the slope of the curve was rather gentle, suggesting a slower oxide growth rate Accordingly, this regime with oxidation time beyond 15 days is named as the slow growth regime In the fast growth regime, Si substrates with presence of Ni from 10% to 75% all had steeper slopes than that of pure H-Si(001), indicating a much higher oxidation rate than that on the pure H-Si(001) surface The enhancement effect persisted up to Ni% of 50%, while further increase in Ni% lead to a decrease of Si’s oxidation instead In the slow growth regime, all samples displayed similar slope of increase, regardless the presence of Ni Hence, Ni grown on the H-Si(001) surface at RT seems to accelerate the oxidation of Si in the fast growth regime An increase in Ni% from 10% to 50% results in an increase in oxide peak intensity and a more fully oxidized oxide (SiO2) When Ni% was further increased from 50% to 75%, it caused a slow-down in the oxidation of Si, although the Si oxide still grew much faster than on pure H-Si(001) surface By plotting the same data in a logarithmic scale, the SiO2 formation was found to be linearly dependent on the logarithm of the exposure time It indicates that the oxide growth follows a power law with the oxidation time The order of power is fitted to be 0.1 (see Appendices) This power-law dependence behavior has been similarly 168 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature observed in the alkali-metal covered clean Si and Ge substrates without hydrogen termination, where the oxidation of the substrates were also been promoted203 1.2 1.0 SiO2/Si 0.8 75% Ni/H-Si(001) 60% Ni/H-Si(001) 50% Ni/H-Si(001) 41% Ni/H-Si(001) 33% Ni/H-Si(001) 20% Ni/H-Si(001) 10% Ni/H-Si(001) H-Si(001) 0.6 0.4 0.2 0.0 50 100 150 200 250 300 350 400 450 500 550 Oxidation Time (days) (a) SiO2/Si 0.1 75% Ni/H-Si(001) 60% Ni/H-Si(001) 50% Ni/H-Si(001) 41% Ni/H-Si(001) 33% Ni/H-Si(001) 20% Ni/H-Si(001) 10% Ni/H-Si(001) H-Si(001) 0.01 1E-3 1E-3 0.01 0.1 10 100 1000 Oxidation Time (days) (b) Fig 5.4 Evolution of SiO2/Si ratio as functions of Ni coverage and oxidation time in (a) linear scale and (b) log scale In view of Ni’s enhancement effect on Si oxidation, it would be interesting to look at the Ni 2p3/2 spectra’s evolution as a function of exposure time for these two samples, which is scaled and shown in Fig 5.5 It can be seen from Fig 5.5 (a) & (b) that after exposure to air for more than 300 days, the B.E of Ni 2p3/2 main peak in both samples remained the same at 853.9±0.1eV and 853.5±0.1eV for 10%Ni/HSi(001) and 50%Ni/H-Si(001), respectively The FWHM also remained at 1.2 for both 169 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature samples throughout However, for 10%Ni/H-Si(001), there was no appearance of peaks belonging to Ni oxide even after 371 days’ exposure In contrast, a peak at 856.6±0.1eV appeared on the surface of 50%Ni/H-Si(001) right after minute’s exposure, which is attributed to Ni suboxide (NiOx, x 15 days) In the fast decay regime, Ni% decreased very sharply over short period of time, while in the slow decay regime, Ni% decreased insignificantly and slowly almost in the same rate for all the samples 80 75% Ni/H-Si(001) 60% Ni/H-Si(001) 50% Ni/H-Si(001) 20% Ni/H-Si(001) 10% Ni/H-Si(001) 70 60 Ni% 50 40 30 20 10 0 50 100 150 200 250 300 350 Oxidation Time (days) Fig 5.7 Ni% change as a function of exposure to air for coverages of Ni on HSi(001) It is worth to note that the two regions during decrease of Ni% coincide with the two regions observed earlier during the growth of SiO2 Since the SiO2 layers are growing, the only possible explanation for decrease in Ni% for all samples is that the SiO2 layers are growing above this Ni films upon oxidation, which leads to burying of Ni layer underneath and a subsequent decrease in Ni% In order to probe the depth distribution of Ni, we will resort to angle-resolved XPS on a serial of samples with different Ni coverages as a non-destructive analysis method 172 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature For 10%Ni/H-Si(001), the peak of SiO2 increased with a decrease photoelectron take-off angle from 90o to 20o (the take-off angle is defined as the angle between surface plane and photoelectron path), while Ni 2p3/2 spectra at all four takeoff angles showed a pure NiSi-like state without any oxide (Fig 5.8(a,b)) At smaller take-off angle, i.e., 20o, only the signal coming from the top surface were collected, compared to the signal more from deep bulk layers at bigger take-off angles Hence, the above results indicate a dominant presence of SiO2-rich layers on the topmost layers As the distance away from surface increased, the SiO2 component decreased accompanied by an increase in elemental Si, while Ni remained unoxidized over time from top surface layer to deeper layer As for 50%Ni/H-Si(001), both Si 2p and Ni 2p3/2 spectra displayed increases in Si and Ni oxide peaks when photoelectron take-off angle decreased from 90o to 20o, which implies that both Ni and Si were heavily oxidized at the top surface, but the extend of oxidation was less significant at the underneath layer away from the surface As the spectra for Ni coverages below 33% and above 33% are similar to those of 10%Ni and 50%Ni, respectively, they are not presented here To qualify the extent of oxidation in presence of different Ni coverage, Ni% and Si% as a function of take-off angle can be calculated based on the Ni 2p3/2 and Si 2p spectra at different take-off angles after taking into account both the sensitivity factors and transmission functions The corresponding results are shown in Fig 5.9 For samples with Ni% less than 50%, it can be seen that after oxidation Ni% increased when photoelectron take-off angle increased, which implies Ni% was less in top surface layer than in the deeper layer (Fig 5.9(a)) As the RT deposition lead to a Nirich layer on the top surface before oxidation, such a deficiency in Ni% but rich in 173 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Although the growth of both Si oxide and Ge oxide slowed down with time for both samples, the heights of SiO2 and GeO2 were significant lower than those of elemental Si and Ge in 10%Ni/H-Si0.8Ge0.2 sample, while they were much higher than those of elemental Si and Ge in 50%Ni/H-Si0.8Ge0.2 sample The higher the oxide peak, the more Si and Ge were oxidized Hence, it implies that Si0.8Ge0.2 was more heavily oxidized with a higher Ni coverage In order to probe the effect of Ni coverage on Si0.8Ge0.2’s oxidation, Si 2p and Ge 3d spectra as a function of Ni coverage after exposure to air for around hours were shown in Fig 5.21 It can be seen clearly from Fig 5.21 (a & c) that the height of Si and Ge oxide peaks increased with Ni% till 40% with the same oxidation time (~4hrs) Meanwhile, the B.E of the oxide shifted slowly towards higher value with higher Ni coverage, approaching that of the SiO2 and GeO2 However, when Ni coverage further increased from 42% to 50% and above (Fig 5.21(b)), the Si oxide peaks decreased with Ni coverages and the B.E started to shift back towards lower values Similarly, the B.E of Ge 3d started to shift back towards lower values when Ni% further increased to above 42% while the height of oxide showed a clear decrease at 84% Ni However, all the samples with Ni thin films show larger SiO2 and GeO2 signals than those of bare H-Si0.8Ge0.2, which implies that Ni has similarly prompted the oxidation of Si0.8Ge0.2 as it exercise on Si and Ge substrates 195 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Si 42% Ni 30% Ni 20% Ni 10% Ni H-Si0.8Ge0.2 SiO2 Intensity (a.u.) Intensity (a.u.) Si 107 106 105 104 103 102 101 100 99 98 97 84% Ni 60% Ni 50% Ni 42% Ni H-Si0.8Ge0.2 SiO2 107 106 105 104 103 102 101 100 Binding Energy (eV) (a) 34 33 31 27 26 GeO GeO2 GeO 32 28 84% Ni 60% Ni 50% Ni 42% Ni H-Si0.8Ge0.2 Intensity (a.u.) Intensity (a.u.) 35 97 Ge 42% Ni 30% Ni 20% Ni 10% Ni H-Si0.8Ge0.2 36 98 (b) Ge GeO2 99 Binding Energy (eV) 30 29 Binding Energy (eV) 28 27 26 36 35 34 33 32 31 30 29 Binding Energy (eV) (d) (c) Fig 5.21 (a-b) Si 2p spectra and (c-d) Ge 3d spectra as a function of Ni coverages after exposing to air for about hours at RT To quantify the growth of Si and Ge oxides as a function of Ni coverage and oxidation time, curve fitting has been carried out on the both Si 2p and Ge 3d spectra in order to extract the ratio between peak area of the oxide peaks and that of elemental peaks A typical Si 2p and Ge 3d spectra after curve-fitting are presented earlier in Fig 5.3 and Fig 5.13, respectively By applying the same fitting procedures to all the Si 2p and Ge 3d spectra from samples with different Ni coverages after a serial of exposure time in log scale, the evolution of SiO2/Si and (GeO+GeO2)/Ge ratio as functions of Ni% and oxidation time is shown in Fig 5.22 For oxidation of both Si and Ge, it can be seen from a linear scale (Fig 5.22(a & c)) that there are generally two oxidation 196 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature regimes divided by the dotted blue line for these set of samples In the first regime with oxidation time less than 25 days, the slope of the curve was very steep, indicating that there was fast growth of oxide over short time In the second regime with oxidation time greater than 25 days, the slope of the curve was gentle, suggesting a slower oxide growth rate Hence, the regimes with oxidation time less than 25 days and beyond 25 days are similarly termed as the fast growth regime and the slow growth regime, respectively H-SiGe 10% Ni 20% Ni 42% Ni 52% Ni 84% Ni 10 H-SiGe 10% Ni 20% Ni 42% Ni 52% Ni 84% Ni SiO2/Si SiO2/Si 20 18 16 14 12 10 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.1 0.01 50 100 150 200 250 300 350 1E-3 0.01 0.1 Oxidation time (days) H-SiGe 10% Ni 20% Ni 42% Ni 52% Ni 84% Ni 10 (GeO2+GeO)/Ge (GeO2+GeO)/Ge 10 100 1000 100 1000 (b) (a) 12 10 Oxidation time (days) 400 0.3 0.2 H-SiGe 10% Ni 20% Ni 42% Ni 52% Ni 84% Ni II I 0.1 0.1 0.0 0.01 50 100 150 200 250 Oxidation time (days) 300 350 400 1E-3 0.01 0.1 10 Oxidation time (days) (d) (c) Fig 5.22 Evolution of (a,b) SiO2/Si and (c,d) (GeO+GeO2)/Ge ratio as functions of Ni coverage and oxidation time in (a,c) linear scale and (b,d) log scale By plotting the same curves in log scale (Fig 5.22(b&d)), it can be seen that SiO2 was detected in all the samples with presence of Ni from 10% to 84% right after 197 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature exposing to air for minute, while it was only detected on pure H-Si0.8Ge0.2(001) after hours A mixture of GeO and GeO2 were detected in samples presence with Ni after exposing to air for minute, while they appear on pure H-Si0.8Ge0.2(001) after 11 minutes Therefore, both SiO2 and GeO2 were detected much earlier in all the samples with presence of Ni from 10% to 84% than on pure H-Si0.8Ge0.2(001) In addition, the samples with Ni thin films all displayed higher SiO2/Si and (GeO+GeO2)/Ge ratio than pure H-Si0.8Ge0.2(001) over the entire time scale Increasing Ni coverage leads to a higher SiO2/Si and (GeO+GeO2)/Ge ratio at the same oxidation time This enhancement effect continued up to Ni% of 50%, and further increase in Ni% lead to a decrease of Si0.8Ge0.2’s oxidation instead Hence, Ni grown on the HSi0.8Ge0.2(001) surface at RT has similarly promoted the oxidation of Si0.8Ge0.2 as what it does to Si and Ge substrates It is worth to compare the oxidation of Si and Ge in Ni/H-Si0.8Ge0.2 to the separate oxidation of Si in Ni/H-Si and Ge in Ni/H-Ge SiO2 formation in Ni/HSi0.8Ge0.2 was found to be linearly dependent on the logarithm of the exposure time, which indicates that the Si oxide growth follows a power law with the oxidation time with an average power order of 0.2 (see Appendices) Such a power law dependence of SiO2/Si is also observed in Ni/H-Si systems but with a smaller order of 0.1 It implies that Si in Ni/H-Si0.8Ge0.2 is much heavily oxidized than the Si in Ni/H-Si system and has been observed experimentally For example, the maximum value of SiO2/Si in all Ni/H-Si samples in Fig 5.4 is only 1.2 after exposure to air for 470 days, while the maximum value is 17.5 in Ni/H-Si0.8Ge0.2 samples after exposure to air for 372 days (Fig.5.22) Such a significant increase in Si’s oxidation can be attributed to a weakened 198 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature lattice structure due to the doping of Ge into Si’s lattice structure and an expanded lattice, which facilitates oxidant’s diffusion and subsequently enhance Si’s oxidation GeO2 formation in both Ni/H-Si0.8Ge0.2 and Ni/H-Ge systems were found to follow a power law dependence of the exposure time on the logarithm scale with twogradients The average order of power is fitted to be 0.11 and 0.36 for regime-I and regime-II in Ni/H-Si0.8Ge0.2 system, while they are 0.13 and 1.26 for regime-I and regime-II in Ni/H-Ge system (see Appendices) Hence, it indicates Ge was less oxidized in Ni/H-Si0.8Ge0.2 system than in Ni/H-Ge system, which is indeed observed experimentally The (GeO+GeO2)/Ge ratio was much lower in Ni/H-Si0.8Ge0.2 than the ratio in Ni/H-Ge For example, the maximum ratio was 10.0 in Ni/H-Si0.8Ge0.2 system after exposure to air for 372 days (Fig.5.22), in contrast with a ratio value of 195.1 after exposure to air for 372 days (Fig.5.15) In addition, the turning points between two linear-gradient regimes are less distinct in Ni/H-Si0.8Ge0.2 than in Ni/H-Ge The difference could be attributed to by a smaller heat of formation of GeO2 than SiO2 As a result, in the Si-rich environment, oxidants react more preferably with Si than with Ge, leading to a smaller ratio of (GeO+GeO2)/Ge and hence less distinct slopes in Ni/H-Si0.8Ge0.2 than in Ni/H-Ge The preferential oxidation of Si over Ge is accessed by the change of Si:Ge ratio over oxidation time From Fig 5.23, it can see that Ge% ( Ge ×100% ) Ge + Si decreased over the entire oxidation process In addition, it can be seen that the decrease of Ge% can also be divided into two regimes separated by the dotted line In the first regime when oxidation time is less than 25 days, Ge% decreased very fast In the second regime when oxidation time is beyond 25 days, Ge% decreased gently This two-regime behavior can be correlated with the Ni-promoted SiGe oxidation In the 199 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature first regime of fast oxidation, Si was preferably oxidized over Ge, leading to an upwards growth of SiO2 layer and a expelling of Ge towards Si0.8Ge0.2 The faster the SiO2 grew, the faster the Ge% decreased In the second regime of slow oxidation, SiO2 grew slower, and hence, Ge% decreased also slower After year of oxidation, the eventual structure of Ni/Si0.8Ge0.2 system became SiO2-rich oxide layer on the top, followed by Ni-rich layer, Ge-rich Si1-xGexlayer and Si0.8Ge0.2 substrate Such a structure is also reported in the literatures179 H-Si0.8Ge0.2 40 35 30 Ge% 25 20 Normalized Ge% 10% Ni 20% Ni 30% Ni 42% Ni 50% Ni 60% Ni 84% Ni 15 10 50 100 150 200 250 300 Oxidation Time (days) 350 400 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 H-Si0.8Ge0.2 10% Ni 20% Ni 30% Ni 42% Ni 50% Ni 60% Ni 84% Ni 50 100 150 200 250 300 350 400 Oxidation Time (days) (b) (a) Fig 5.23 (a) Ge% and (b) normalized Ge% change as a function of oxidation time with different Ni coverages In view of Ni’s enhancement effect on Si oxidation, it would be interesting to look at the Ni 2p3/2 spectra’s evolution as a function of exposure time for these two samples, which is scaled and shown in Fig 5.24 It can be seen from Fig 5.24 (a) & (b) that after exposure to air for more than 300 days, the B.E of Ni 2p3/2 main peak for elemental Ni in both samples remained the same at 853.9±0.1eV and 853.5±0.1eV for 10% and 50% Ni/H-Si0.8Ge0.2(001), respectively The FWHM value for both main peaks stayed constant at 1.2 over time However, for 10%Ni/H-Si0.8Ge0.2(001), there was no appearance of peaks belonging to Ni oxide even after 340 days’ exposure In contrast, Ni oxide appeared on the surface of 50%Ni/H-Si0.8Ge0.2(001) right after 200 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature minute’s exposure, and it continued to grow with oxidation time And it is almost turned into Ni oxide after 372 days’ exposure 372 days 140 days 69 days days hour minute As-deposited 866 Intensity (a.u.) Intensity (a.u.) Ni 340 days 159 days 66 days days hours minute As-deposited 864 862 860 858 856 854 852 850 866 Binding Energy (eV) 864 862 860 Ni NiO 858 856 854 852 850 Binding Energy (eV) (a) (b) Fig 5.24 Ni 2p3/2 spectra evolution as a function of oxidation time at RT for (a) 10%Ni/H-Si0.8Ge0.2 (001) and (b) 50%Ni/H-Si0.8Ge0.2 (001) To study the effect of Ni coverage on the oxidation of Ni at the same exposure time, a series of Ni 2p3/2 after hours exposure are presented in Fig 5.25 When Ni% was not more than 30%, Ni was not oxidized When Ni% was above 30%, Ni started to be oxidized and the amount of oxide increased with an increase in the Ni% 84% Ni/H-Si0.8Ge0.2(001) 60% Ni/H-Si0.8Ge0.2(001) Intensity (a.u.) 50% Ni/H-Si0.8Ge0.2(001) 859 42% Ni/H-Si0.8Ge0.2(001) 30% Ni/H-Si0.8Ge0.2(001) 20% Ni/H-Si0.8Ge0.2(001) 10% Ni/H-Si0.8Ge0.2(001) 858 857 856 855 854 853 852 851 Binding Energy (eV) Fig 5.25 Ni 2p3/2 spectra evolution as a function of Ni coverage after exposure to air for hours at RT 201 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature The behavior of Ni on H-Si0.8Ge0.2(001) at low coverage (≤30%) was very similar to that on H-Si(001) surfaces but deviated from that on H-Ge(001) surfaces The implication is that Ni seems to be surrounded in a Si-rich environment but not a Ge-rich environment Subsequently, the Ni/H-Si0.8Ge0.2(001) should follow the oxidation behavior of Ni/H-Si(001), namely Si and Ge diffuse upward to react with oxygen while Ni continues to diffuse inward and bond with Si&Ge underneath to remain as germanosilicide during oxidation The overall Ni% change as a function of exposure time to air for coverages is shown in Fig 5.26 It can be seen that Ni% decreased in all the samples with an increase in oxidation time Two regimes can also be identified as separated by the dotted purple line for all samples: the fast decay regime (oxidation time < 25days) and the slow decay regime (oxidation time > 25 days) In the fast decay regime, Ni% decreased very sharply over short period of time, while in the slow decay regime, Ni% Ni% decreased insignificantly and slowly almost in the same rate for all the samples 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 84% Ni 60% Ni 30% Ni 10% Ni 50 100 150 200 250 300 350 Oxidation time (days) Fig 5.26 Ni% change as a function of exposure time to air for coverages of Ni on HSi0.8Ge0.2 (001) 202 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Similar to the case of Ni/H-Si and Ni/H-Ge, the two regions during decrease of Ni% coincide with the two regions where SiO2 and GeO2 grew Therefore, the decrease in Ni% for all samples is attributed to the growth of Si oxide and Ge oxide overlayers, which buries Ni layer underneath and result in a subsequent decrease in Ni% In order to probe the depth distribution of Ni, we will resort to angle-resolved XPS as a non-destructive method and a depth-profiling XPS as a destructive method for analysis For 10%Ni/H-Si0.8Ge0.2 (001), the peak of SiO2 and GeO2 increased with a decrease photoelectron take-off angle from 90o to 30o, while Ni 2p3/2 spectra at all four take-off angles showed a pure germanosilicide state without any oxide (Fig 5.27(a,c,e)) Hence, the results indicate a dominant presence of SiO2 & GeO2-rich layers on the topmost layers When take-off angle increased, the probing distance from surface increased, which is indicated by a relative decrease in the SiO2 & GeO2 component and a relative increase in elemental Si and Ge components Contrary to the oxidation of Si and Ge, Ni still remained unoxidized over time from surface to deeper layer As for 50%Ni/H-Si0.8Ge0.2 (001), Si 2p and Ge 3d exhibited the same behavior as they did in 10%Ni/H-Si0.8Ge0.2(001) (Fig 5.27(b,d)) However, Ni was oxidized (Fig 5.27(f)) In addition, Ni 2p3/2 spectra displayed an increases in Ni oxide peaks when photoelectron take-off angle decreased from 90o to 30o, which implies that both Ni and Si1-xGex were heavily oxidized at the top surface, while their oxidation was less significant at the underneath layer away from the surface 203 Intensity (a.u.) Intensity (a.u.) Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature o 30 o 45 o 60 o 90 107 106 105 104 103 102 101 Binding Energy (eV) 100 99 98 o 30 o 45 o 60 o 90 107 97 106 105 104 103 102 101 100 99 98 97 27 26 Binding Energy (eV) (b) (a) o 37 36 35 34 33 32 31 o Intensity (a.u.) Intensity (a.u.) 30 o 45 o 60 o 90 30 29 28 27 26 30 o 45 o 60 o 90 37 36 35 34 Binding Energy (eV) 33 32 31 30 29 28 Binding Energy (eV) (c) (d) Intensity (a.u.) Intensity (a.u.) o o 20 o 30 o 45 o 90 866 864 862 860 858 30 o 45 o 60 o 90 856 Binding Energy (eV) 854 852 850 866 864 862 860 858 856 854 852 850 Binding Energy (eV) (e) (f) Fig 5.27 Angle-resolved XPS (ARXPS) spectra of (a) Si 2p, (c) Ge 3d and (e) Ni 2p3/2 in 10%Ni/H-Si0.8Ge0.2 (001) after 883 days’ oxidation, and (b) Si 2p, (d) Ge 3d and (f) Ni 2p3/2 in 50%Ni/H-Si0.8Ge0.2 (001) after days’ oxidation with photoelectron takeoff angle of 30o, 45o, 60o and 90o Ni%, Si% and Ge% as a function of take-off angle can be calculated based on the Ni 2p3/2, Si 2p and Ge 3d spectra at various take-off angles after taking into 204 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature account of both the sensitivity factors and transmission functions The corresponding results are shown in Fig 5.28 It can be seen that Ni% increased when photoelectron take-off angle increased (Fig 5.28(a)) Ge% show the same trend while Si% displayed the reverse trend (Fig 5.28(b,c)) Hence, the results imply that Ni and Ge were less in top surface layers than in the deeper layer, while Si was richer in the top layer than in the deeper layer As the RT deposition lead to a Ni-rich layer on the top surface before oxidation, such a deficiency in Ni% & Ge% but rich in Si% at the top surface after oxidation suggested that Si was diffusing up towards surface and reacting with oxidants while Ni and Ge were diffusing downwards during oxidation In fact for sample having Ni% greater than 40%, Ni was barely detected on the surface after oxidation for around 800 days, an observation consistent with the fact that Ni promotes the oxidation of Si0.8Ge0.2 The more Ni, the greater the oxidation Therefore, a thicker SiO2 & GeO2 layer was formed on top of the Ni layer, leading to a dramatic decrease in Ni% 205 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature 3.5 90 3.0 88 2.5 1.5 Si% Ni% 86 2.0 10% Ni/H-Si0.8Ge0.2_Air 883 days 84 20% Ni/H-Si0.8Ge0.2_Air 877 days 1.0 30% Ni/H-Si0.8Ge0.2_Air 813 days 10% Ni/H-Si0.8Ge0.2_Air 883 days 82 20% Ni/H-Si0.8Ge0.2_Air 877 days 0.5 20 30 40 50 60 70 80 90 80 20 100 o 30% Ni/H-Si0.8Ge0.2_Air 813 days 30 40 50 60 70 80 o Photoelectron Take-off Angle ( ) (a) 90 100 Photoelectron Take-off Angle ( ) (b) 14.5 14.0 13.5 Ge% 13.0 12.5 12.0 10% Ni/H-Si0.8Ge0.2_Air 883 days 11.5 11.0 20 20% Ni/H-Si0.8Ge0.2_Air 877 days 30% Ni/H-Si0.8Ge0.2_Air 813 days 30 40 50 60 70 80 90 100 o Photoelectron Take-off Angle ( ) (c) Fig 5.28 (a) Ni%, (b) Si% and (c) Ge% composition change as a function of photoelectron take-off angles for samples with different initial Ni% In order to confirm the depth distribution of Ni/H-Si0.8Ge0.2(001) after oxidation, one sputtering experiment was carried out on 30%Ni/H-Si0.8Ge0.2(001) after exposure to air for 814 days, and the result is shown in Fig 5.29 From Fig 5.29(a), it can be seen that Ni% increased from surface to a maximum value in the deeper layer and then it started to decrease to zero Ge% also increased from surface to a peak value in deeper layer and then decreased slightly before it climbed up again to the bulk value The peak position of Ge% was slightly above that of Ni% However, the Si concentration decreased from a maximum value at top surface to a minimum value at deeper layer and then started to increase back to the bulk value Therefore, the Si 206 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature deficient region between the its minimum value point and its bulk value point is due to the fact that Si has segregated to the surface region driven by the greater heat of formation of SiO2 than those of GeO2 and NiO As a result, the maximum Si% was found on the top surface When Si was moving outwards, it actually left both Ni and Ge behind Hence, the maximum values of Ni and Ge were found some distance below the top surface In addition, because the heat of GeO2 formation is bigger than that of NiO but smaller than SiO2, Ge was also segregating outwards and leaving Ni behind, which is reflected by a shallower position of Ge’s maximum value than that of Ni By looking at Si-Ge composition change in Fig 5.29 (b), it can be seen more clearly that Si was preferably segregating from deeper layers to the top surface, leading to the formation of an interlayer where Si% was less while Ge% was more than those in the bulk substrate 90 90 85 88 Si% 86 80 84 80 Atomic Raito (%) 82 70 Atomic ratio (%) 75 65 60 Ge% 20 15 Si% 78 76 24 22 Ge% 20 18 10 16 14 12 Ni% 10 50 100 150 200 250 300 Sputtering time (s) 350 400 450 500 50 100 150 200 250 300 350 400 450 500 Sputtering time (s) (a) (b) Fig 5.29 (a) Ni-Si-Ge composition (Ni%+Ge%+Si%=100%) and (b) Si-Ge composition (Ge%+Si%=100%) change as a function of sputtering time on 30% Ni/HSi0.8Ge0.2 (001) after exposure to air for 814 days Therefore, as determined by the XPS results, the oxidation of Si0.8Ge0.2(001) with Ni thin films can be summarized as following After the initial deposition of Ni thin film on H-Si0.8Ge0.2 (001), Ni reacts with the top layers of Si0.8Ge0.2 substrate and form Ni germanosilicides (or a mixture of Ni silicide and Ni germanide) Upon 207 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature exposed to air for oxidation, Si and Ge in the Ni germanosilicides layer break their original bonds with Ni and change to bond with oxidants (O2 and -OH) and grow upwards instead, due to a larger heat of formation of SiO2 and GeO2 than that of Ni germanosilicides Among Si and Ge, oxidants are more preferable to bond with Si, leading to a Si-rich surface Ni instead prefers to re-bond with Si0.8Ge0.2 underneath to preserve the Ni germanosilicides phase due to higher heat of formation of Ni germanosilicide than that of Ni oxide Subsequently, oxidants in the air have to diffuse through the oxide layer to further react with the Ni germanosilicides layer underneath As the oxidation and re-bonding cycle continues, the oxide grows thicker and it will be more difficult for oxidants to diffuse through the oxide layer, therefore, the oxidation rate slows down eventually At low Ni coverage (≤30%), Ni was protected from oxidation by forming a Ni germanosilicide phase, similar to the observation of NiSi-like layer formation for low coverage Ni on H-Si surface At high Ni coverage (>30%), metallic Ni layer similarly formed on the top surface and was oxidized The presence of Ni promoted the oxidation of Si0.8Ge0.2 5.5 Summary For the growth of low coverage Ni (≤33%) on hydrogen-terminated Si(001) and Si0.8Ge0.2(001) substrates, Ni remained un-oxidized even after more than year’s exposure to air However, both the Si and Si0.8Ge0.2 substrates were oxidized more severe compared to the pure substrates without Ni presence Such behavior was due to the formation of NiSi-like and NiSi0.8Ge0.2-like layer immediately after deposition 208 Chapter Oxidation of Ni Thin Films on Si, Ge and Si0.8Ge0.2 Substrates at Room Temperature Following that, Ni silicide or Ni germanosilicide dissociated during oxidation because Si and Ge prefer to bond with O than with Ni The free Ni further diffused down towards substrate to bond with the substrate underneath in order to preserve the Ni silicide or Ni germanosilicide phase The diffusion therefore has left a weakened structure in the original NiSi-like and NiSi0.8Ge0.2-like layer, which facilitated the diffusion of oxidants through these layers and hence promoted the oxidation of Si/Ge At high coverage of Ni (>41%), metallic Ni thin films formed on the surfaces of H-Si and H-Si0.8Ge0.2 surfaces without forming a bond with the substrates Upon exposure to air for oxidation, both Ni and the substrate (Si/Ge) were oxidized over time The oxidation behavior of Ni on H-Ge(001) was rather different from that on H-Si and H-Si0.8Ge0.2 substrates Both Ni and Ge were oxidized throughout all Ni coverages, even when a NiGe-like layer was formed upon deposition of low coverage Ni on H-Ge surface (as evidenced by Ni 2p3/2 B.E shift) Ni was not protected from oxidation in the NiGe environment This is because Ni-O bond is stronger than Ni-Ge and Ge-Ge but weaker than Ge-O, which implies that Ni prefers to bond with O rather than Ge Therefore, both Ni and Ge were oxidized irregardless of Ni coverages The oxidation behavior of Ni in the NiSi-like and NiSi0.8Ge0.2-like layer clearly demonstrates a bonding between Ni and the H-Si and H-Si0.8Ge0.2 substrates at RT Termination the surface with hydrogen does not suppress the reaction between Ni and the substrates A power law can be employed to fit the evolution of both SiO2/Si and (GeO2+GeO)/Ge ratio as a function of oxidation time in thin oxide thickness regime However, more work is clearly required to fully understand the oxidation mechanism and to correlate both SiO2/Si and (GeO2+GeO)/Ge ratio to the Ni coverages 209 ... 20% Ni/H -Si( 001) 10% Ni/H -Si( 001) 858 857 Before oxidation 856 855 854 853 852 851 Binding Energy (eV) Intensity (a.u.) (a) 859 75% Ni/H -Si( 001) 60% Ni/H -Si( 001) 50 % Ni/H -Si( 001) 41% Ni/H -Si( 001)... 6.0 Ni 2p3/2 ~ Ge 2p3/2 5. 5 Ni 2p3/2 ~ Ge 3d Ni% 5. 0 4 .5 4.0 3 .5 3.0 2 .5 2.0 1 .5 40 45 50 55 60 65 70 75 o 80 85 90 95 Photoelectron take-off angle ( ) Fig 5. 19 Ni composition change for 38%Ni/H-Ge(001)... 41% Ni/H -Si( 001) 33% Ni/H -Si( 001) 20% Ni/H -Si( 001) 10% Ni/H -Si( 001) H -Si( 001) 0.6 0.4 0.2 0.0 50 100 150 200 250 300 350 400 450 50 0 55 0 Oxidation Time (days) (a) SiO2 /Si 0.1 75% Ni/H -Si( 001)