6.3 Results
6.3.1 SXPS Analysis of Alkylated Surfaces
6.3.1.1 Cl-Terminated Si(111)
over an area approximately 1 µm square divided into 128 x 128 bins for a pixel width of 7.8 nm. The scan area was periodically calibrated with a scanning electron microscope (SEM) image of a nearby piece of debris in order to account for spatial drift in the image over the scan time. Between 1–5 such scans were added together for the final image. Care was taken to collect data quickly in order to minimize damage to the alkyl overlayer by the impinging electron beam. Because the amount of SiO2on the surface of alkylated samples was very small, it was not possible to collect spectra of the SiO2 signal (72–76 eV) in a short enough amount of time to prevent beam damage. Instead, signal from the Si LVV (90–92 eV)6–8 and O KLL (510 eV)9 regions were combined to determine areas of SiO2
in the SAM image. Spectra were collected with the electron analyzer held in the Si LVV region (92 eV) and the O KLL region (510 eV), which were then superimposed to create a composite image of Si and SiO2.
-2 -1 0 1 2 3 4 5
Photoelectron energy (BeV) Freshly prepared
12 hr in air 24 hr in air 36 hr in air 48 hr in air
BB Counts (Arbitrary units)
Figure 6.1: Chlorine-terminated Si(111) surface freshly prepared then allowed to sit in air for 12, 24, 36, and 48 hr. The scale is binding energy (BeV) relative to the center of the bulk Si 2p3/2 peak of the freshly prepared surface. All spectra are shown after background removal, spin-orbit stripping, and peak fitting. In all cases, cross marks are the raw data, dashed lines are the fitted curves, and the solid line is the calculated curve fit. The anomalous break in the pattern of growth of Si suboxides on the sample kept in air for 24 hr was most likely due to some contribution to the signal from the sample holder because of an ill-focused beam.
0 200
400 600
800 1000
Counts (Arbitrary units)
Photoelectron energy (BeV) a)
b) c)
O 1s
C 1sCl 2s Cl 2p
Si 2s Si 2p
Figure 6.2: XPS of Cl-Si(111) surfaces freshly prepared, then exposed to ambient air for a period of up to 54 hr. a) Freshly prepared surface; b) in air 20 hr; c) in air 54 hr.
be from a number of sources, including absorbed solvent from the wet chemical prepa- ration techniques, absorbed pump oil vapor introduced in the quick-entry load lock, or contaminating dust particles covered with oxygen-containing organic molecules that were not possible to avoid when working in standard laboratory conditions. The principal con- cern of this study was the growth and characterization of silicon oxides, and because of the uncertainty of the origin of the O 1s signal, this peak was not used to identify the presence of silicon oxides on the surfaces described here. The Si 2p region, which has specific and extensively studied spectroscopic shifts introduced by Si+–Si4+ oxides4, 11 was examined instead to address this critical question. The fact that the freshly prepared Cl-terminated surface was free of silicon oxides in the region of higher binding energy above the bulk Si 2p3/2peak is clearly demonstrated in Figure 6.1.
When exposed to ambient air, the chlorinated Si surface reacted rapidly, losing the Si–
Cl feature and growing silicon oxides at higher binding energy (Table 6.1). After 12 hr in air, the monolayer coverage of the peak approximately 0.8 eV higher in binding energy than the bulk peak, assigned to surface Si bonded to Cl atoms, dropped from 0.99 ML10to 0.87 ML, while a large, broad signal of oxide centered at +3.13 eV above the bulk Si 2p3/2 peak appeared. This feature was accompanied by smaller suboxides at 1.85 eV and 3.74 eV above the bulk Si 2p3/2 peak, representing a total equivalent monolayer coverage of 1 ML of silicon oxides.
As air exposure of the Cl-Si(111) surface continued, the growth of the surface oxide features correlated qualitatively to loss of the signal from surface Cl-bound Si atoms,≈0.9 eV above the bulk Si 2p3/2 peak. As can be seen in Table 6.1 and Figure 6.1, after the surface had been exposed to air for 12 hr, the monolayer coverage of the highest order Si oxide (Si4+, +3.74 eV) grew from below the detection limit of the instrument to 0.21 ML.
By the time the sample had been in air for 24 hr, the amount of Cl-bound Si atoms detected on the surface had dropped to only ≈0.5 ML, while the highest order oxide (Si4+, +3.78 eV) grew to 1.45 ML. A XPS survey scan of the Cl-terminated surface collected after the sample had been exposed to air for 20 hr (Figure 6.2(b)) confirmed that the Cl 2s and 2p signals at 200 BeV and 270 BeV, respectively, had dropped significantly from the freshly prepared surface, indicating that most of the Cl present on the surface had disappeared. At
SXPSpeakshiftandintegratedareaa SurfaceSi-X“Si+”“Si2+”“Si3+”“Si4+”Totalsurface TimeinBulkpeakbshiftcoverageshiftcoverageshiftcoverageshiftcoverageshiftcoveragecoverage R-air(hr)location(eV)(eV)(ML)(eV)(ML)(eV)(ML)(eV)(ML)(eV)(ML)(ML) X=Clsiliconoxides Cl-1298.640.870.87–c –c 1.850.153.130.493.740.211.87 2498.520.940.47–c –c 1.860.573.061.303.781.454.06 3698.480.890.53–c –c 1.940.533.020.953.711.593.75 4898.440.920.49–c –c 1.990.573.151.243.791.624.12 X=Csiliconoxides CH3-1298.040.301.030.790.08––––––1.11 2498.040.310.860.760.061.76<0.013.110.053.880.041.02 3698.080.300.950.790.081.72<0.013.070.053.750.091.17 4898.050.310.890.750.071.75<0.013.100.133.800.071.16 C2H5-1298.190.190.770.610.231.570.04––––1.04 2498.210.170.660.740.231.780.073.120.063.870.031.05 3698.190.210.760.810.171.850.073.100.133.840.021.15 4898.190.210.710.810.241.850.143.100.153.840.091.33 C6H5CH2-1298.240.180.560.790.331.790.063.160.103.810.031.07 2498.250.171.000.780.271.780.043.190.083.750.031.43 Table6.1:Si2p3/2SXPSdataonfunctionalizedSisurfacesoxidizedinairforupto48hr.a CalculatedfromEq.6.3.b Absoluteenergies forthebulkSi2p3/2peakarereportedforthoroughness,butareoflimitedusebecausetheexcitationenergywasnotcalibrated.c No deconvolutionoftheshiftsofSi–ClandSi+ wasattempted.
this point it was difficult to determine if the SXPS peak previously assigned as Cl-bound surface Si atoms was actually cleanly functionalized Cl-Si(111), or rather represented an amorphous Si+substrate.
After the Cl-Si(111) surface had been exposed to air for 48 hr, the XPS survey scan shown in Figure 6.2(c) indicated only a small fraction of any Cl remained on the surface.
It is possible that the surface was stabilized to further oxidation with a submonolayer cov- erage of surficial Cl moieties still present, but SXPS data do not eliminate the possibility that our detection of the Si–Cl peak was hampered as it was convoluted with larger Si+or Si2+ oxides. Features representing Si2+(+1.99 eV), Si3+ (+3.15 eV), and Si4+ (+3.79 eV) oxides dominated the spectrum. The Si4+ oxide had a total surface coverage of 3.63 ML.
Because a penetration depth of 3.5 ˚A used in these studies only sampled approximately 2.2 ML perpendicular to the (111) crystal face, this high monolayer coverage is possibly a result of the limitations of the quantitative analysis of surface species described by Eq.
6.3, which might be inadequate for the amorphous surface described here. It is possible to state conclusively, however, that after 48 hr in ambient air the Cl-terminated Si(111) sur- face had thoroughly reacted to create an amorphous silicon oxide layer that, on the scale of our surface sensitive conditions, was reasonably thick.