Introduction
Plasmas of volatile organics can be used to form highly adherent, conformal coatings with novel chemistries and physical properties1. Deposits formed from plasmas of hexamethyldisiloxane (HMDSO) can be used to physically and chemically modify surfaces1. Surface analytical techniques are applied to the chemical characterisation of such films because they are usually thin and generally insoluble, thus complicating, or preventing their characterisation using conventional analytical techniques. A wide variety of reactions can occur in plasmas2 and consequently deposits generally contain a range of chemical environments3. We have previously used the position of the Si2p peak to provide an indication of the chemical state of the silicon atoms in deposits of HMDSO4. However, this work has been limited by the modest resolution of the spectrometer employed. In this study we use a high resolution XP spectrometer to obtain better energy resolution which enables curve fitting of the Si2p envelope. This provides quantitative information on the different chemical states of the silicon within the deposits. The addition of oxygen to the HMDSO plasma facilitates the production of a range of chemistries, from chemistry similar to that of the monomer, to chemistry approaching that of silica4. It is the nature of this change that we concentrate on herein.
Experimental
All samples (approx.1 micron thick) were produced on aluminium plate (approx. 1 mm thick). In the plasma apparatus the total pressure was fixed at 50 Pa, the power at 200 W, and the HMDSO flow at 20 sccm. The O2 flow rate was varied from 0 -200 sccm (standard cubic centimetres per minute). The plasma were sustained using 2.45 GHz microwaves. A Kratos Analytical AXIS 165 was used with monochromatic Al K
radiation. A take-off-angle of 90° was used throughout the study: assuming a value for 
(C1s) of 1.4 nm, the sampling depth (3sin 90) is calculated to be approx. 4.2 nm (ca.30 Si-O bonds, estimated from its length in [SiO]+ which is 0.15 nm).
Results and Discussion
Charge correction
The structure of the HMDSO molecule is illustrated. The plasma deposits contain carbon, oxygen and silicon. 
From previous work2-4 we believe that the majority of the carbon in the deposit to be present as either Si-CH3 and/or Si-CH2-CH2. We assume these to have similar binding energies which we have measured from the analysis of PDMS to be 284.4 eV (using the same method, and obtaining the same values, as in ref. 5). The O1s and C1s peaks are illustrated in Figure 3. It is apparent that the O1s peak is symmetrical and contains little structure. Examination of the C1s peak indicates that there are two components. We use the value of 284.4 eV to charge correct the spectra using the major component. The minor component appears approximately 2 eV to higher binding energy and is assumed to be ether or hydroxyl carbon.
Si2p curve fitting
We have assumed that there are only four component peaks within the Si2p envelope. This assumption was made to allow us to obtain an estimate of the proportion of the different silicon chemical environments. The major effect upon the Si2p binding energy is the addition of an oxygen bond to the silicon and the abbreviation system (Si(-O)1, Si(-O)2, Si(-O)3 and Si(-O)4) is therefore used. The position of the Si2p peak in PDMS and silica were measured, and the other two environments estimated from these. This was done by halving the difference between the silica and PDMS Si2p peak positions and assuming the influence of introducing an oxygen bond to be approximately equal to this amount. Binding energies corresponding to these four silicon environments are summarised in Table 1.
| Source | Estimate | PDMS | Estimate | Silica |
| Si 2p binding energy (eV) | 101.5 | 102.1 | 102.8 | 103.4 |
| Structures |  |  |  |  |
| Abbreviation | Si(-O)1 | Si(-O)2 | Si(-O)3 | Si(-O)4 |
The simplification to only four peaks allows us to obtain an estimate of the relative proportion of similar chemical environments and may be justified because the difference in binding energy between for Si-CH3 and Si-H and Si-CH2-CH2 is substantially less than that resulting from the addition of an oxygen to the silicon atom. It is also necessary to fit each component using a symmetrical peak, although in actuality the spin-splitting of the 2p core level results in asymmetry. While ideally both of these factors should be accounted for in the curve fit, the improvement in results does not justify the additional complexity. For each sample all peaks were linked to keep their widths equal, but variable. Allowing the component widths to change between samples takes into account the variable roughness of the deposits. The Gaussian to Lorenzian mix was kept constant at 40% for all fits. The curve fits of deposits formed from HMDSO and HMDSO/O2 (200 sccm) are illustrated in Figure 4a and b respectively.
It is clear from the Si2p peak shapes that there are multiple components within the Si2p peak. After binding energy referencing the Si 2p peak we are able to use the peak positions given in Table 1 to fit these envelopes. It is apparent that the components are able to account for the shape of the envelopes at these extremes in chemistry. The effect of oxygen flow on the concentrations of these four silicon environments (within the deposit) is shown in Figure 5. Clear trends are seen with the addition of oxygen. The initial percentage of silicon in Si(-O)1 is 33% compared with 22% Si(-O)2, 40% Si(-O)3 and 5% Si(-O)4. This alters, as oxygen is added to the plasma up to flows of 200 sccm O2 and 20 sccm, where the relative proportions are: Si(-O)1 - 3%, Si(-O)2 - 37%, Si(-O)3 - 5 % and Si(-O)4 - 55%, HMDSO.This trend is accompanied by a consistent decrease in the carbon content of the films and increase in the oxygen to silicon ratio, Figure 6.
Effect of oxygen flow on the deposit structure
From pure HMDSO, plasma deposits are formed with a relatively low degree of Si-O bond formation, relative to the monomer. The addition of oxygen reduces further the proportion of silicon in this environment. There is a concurrent increase in the proportion of silicon in the more highly oxidised environments Si(-O)3 and Si(-O)4. This is indicative of the role of oxygen within the plasma in forming gaseous COx species which are not incorporated into the deposit and Si(-O) moieties which are. These data are entirely consistent with composition data which indicates an increase in the O/Si ratio approaching, though not reaching, that of silica.
Conclusions
High resolution XPS has been used to quantify the silicon environments present in deposits formed from HMDSO/O2 plasmas. The trends observed upon increasing the oxygen flow into the plasma indicate a change from a deposit containing a high proportion of silicon environments similar to that in the HMDSO Si(-O)1, 33% to one where approximately half of the silicon atoms are equivalent to those found in silica Si(-O)4, and the majority of the remainder are bonded to three oxygen atoms, Si(-O)3. A low intensity ether/hydroxyl component was detected in the deposits. Previously the poor resolution and signal to noise of the data had obscured this information.
Acknowledgements
This work was partially funded as a Brite EuRam project under the title "Plasma polymerised coatings and interphases for improved performance carbon fibre composites", Contract No BRE2-0453. In addition to the named authors the involvement of J. Zabold of IKV in making the samples is also gratefully acknowledged.
References
1: “Plasma deposition, treatment and etching of polymers”, Edited by R.d'Agostino, Academic Press, 1990
2: M.R.Alexander, R.D.Short and F.R.Jones “Mass spectrometry of hexamethyldisiloxane plasmas”, submitted for publication in J.Phys.Chem. (1996)
3: M.R.Alexander, R.D.Short, F.R.Jones, W.Michaeli, M.Stollenwerk, G.Mathar and J.Zabold “The heterogeneous nature of hexamethyldisiloxane (HMDSMO) plasma deposits”, Ceramic films and coatings, refereed conference proceedings, British Ceramic Proceedings No. 54, Institute of Materials, pp 87-99, (1995)
4: M.R.Alexander, R.D.Short, F.R.Jones, M.Stollenwerk, J.Zabold and W.Michaeli, J.Mater.Sci., 31 pp 1879-1885 (1996)
5: “High resolution XPS of organic polymers - The Scienta ESCA300 Database” G.Beamson and D.Briggs (Wiley 1992) 72
![[top]](../buttons/TOP.GIF)
For comments, suggestions, contributions or an Instrument Demonstration