11.0

500

Figure 11.

1000

1500

2000

2500

Ne* PRIMARY ENERGY (@V)

ISS

Si responses showing lattice

damage from Ne bombardment.

100

200 ANGSTROMS

300

400

Figure 12. ISS responses from a 904 Å oxide layer on Si used as a standard for Si02.

2

Figure 13. ISS depth profile of Si02/Si etched down to 498 Å.

INTENSITY (ARB. UNITS)

DISCUSSION OF THE PAPER

Participant: Do you have any measure of the excess silicon versus charge density via depth?

Harrington: No, I do not know what the charge density was on these films. This is very recent work and I am continuing on it. I am sure we will do measurements like that.

Johanessen: What is the penetration depth of the helium ions?

Harrington: The penetration depth is on the order of 10 to 60 Angstroms. I think it is probably about 20 Angstroms at 1.5 keV, however the scattering occurs only from the first surface layer.

Johanessen: Is that region distorted by the measurement?

Harrington: It is distorted somewhat, but with helium scattering the damage done by helium sputtering is very very slight and at this point we are basically observing the scattering before the damage has been done. In addition, I think the surfaces do reconstruct and certainly this is borne out by many of the LEED studies. I cannot give you a good explanation as to why we see so little damage in comparison to say reconstruction or this type of thing. The scattering seems to occur before the damage exists. Even if you are able to penetrate and sputter several atoms the signal remains constant.

Johanessen: Would you repeat the value you cited for the extent of the interface?

Harrington: The silicon signal seems to rise over roughly 10 to 20 Angstroms, which looks like about 3 molecular diameters for silicon dioxide. This is about as much as I can tell on such a thick film. I am also going to pursue this more.

DiStefano: Could you estimate the contribution of multiple scattering events to the penetration depth? It is possible that this could lead to a misinterpretation of the extent of the silicon dioxide/silicon interface.

Harrington: Because you are following both the oxygen signal and the silicon signal in a situation like this and you see one rise before the other falls I do not think that is really a possibility. The probability for the scattered particles to penetrate even several stomic thicknesses and emerge as ions from a single scattering event, and we can tell the difference between a single scattering event and a multiple scattering event, is very very small. I do not know whether to say it is a percent or less, I really do not know. The fact is that you can see density differences on (111) and (110) silicon surfaces. If the beam penetrated two layers for instance the density would be the same on the two surfaces; one is basically behind the other.

Participant: How were your samples prepared?

Harrington: The silicon surfaces as prepared were as finely polished and as free of damage as we could get them. They were finely polished, a thick oxide was grown on them at, I believe, 600° or 900°C and this thick oxide was removed by etching in order to get rid of all the damage basically.

Participant: What is the background vacuum level you are operating in?

Harrington: Backfilling is done to a static pressure of about 5 x 10-5 of argon or neon or helium, whichever you are using. The vacuum before backfilling is 10-9 and titanium sublimation pumping is in effect during the whole process.

The secondary ion mass spectrographic technique implies a source of primary ions, a source of secondary ions and a spectroscopic method of analyzing the latter. Figure 1 is an artist's conception illustrating the primary and secondary ions referred to in this presentation. The primary ions in the collimated column are normally generated from a gaseous ambient either by dc or rf fields or in separate sophisticated ion guns such as duoplasmatrons. These primary ions then collide with the solid surface which results in the sputter process which in turn gives rise to a plasma containing positive and negative ions, neutral atoms and molecules having the composition of the solid surface. These sputtered ions are the secondary ions which are analyzed by mass spectroscopy. Figure 1 also illustrates the 20 angstrom possible escape depth of a secondary ion which defines the minimum surface depth.

The remainder of this presentation will be restricted to the three types of commercially available instruments. Figure 2 is a schematic of the ARL QMAS system. Primary ions are generated in the duoplasmatron, accelerated to 10 kV in the gap between the anode and pickup electrode, and focused onto the sample by the primary lens. The sputtered secondary ions are accelerated by the sample voltage back through the extraction lens into a spherical electric sector where they are energy separated. The ions are then mass analyzed by a quadrupole mass spectrometer and detected by ion to electron conversion-scintillator-photomultiplier technique. The quadrupole mass analysis scheme is illustrated in figure 3. A dc field applied to opposite poles is superimposed with a radio frequency field. This oscillating field will then start the ions oscillating between the poles. At any specified frequency only ions of a given mass will undergo stable oscillation which is necessary for them to pass through the length of the field without being collected by the electrodes. The mass is selected by choice of both the dc field and the radio frequency field. Mass spectra are recorded by continuously varying both fields keeping their ratios constant. Figure 4 shows a portion of a 1 ppm boron doped silicon spectrum taken with this instrument. The boron 10 peak shows

2

that the sensitivity is well into the sub-ppm region. This spectrum also shows the mass resolution is sufficient for most inorganic analysis.

Figure 5 is a schematic of the Cameca ion microanalyzer. This instrument is unique in that it uses a magnetic prism for mass analysis which acts as an ion emission microscope as well as a microanalyzer. The primary ions are generated in the duoplasmatron and focused onto the sample. The secondary ions are accelerated and picked up by an electrostatic immersion lens. These ions are then mass or momentum analyzed in the first leg of the magnetic prism which has both radial and tranverse focusing properties. The ions are directed towards the electrostatic mirror where only ions with an energy below a preselected threshold are reflected back through the aperture into the second leg of the magnetic prism. This leg of the magnetic prism reproduces the single mass ion image of the specimen surface on the ion to electron image converter where the electrons can be used for fluorescing a screen, exposing a film or activating a scintillator for photomultiplier detection. A given mass can be preselected by using the appropriate magnetic field, or a complete mass spectra can be recorded by sweeping the magnetic field. Figure 6 is a photograph of this instrument. The ion-gun and sample are to the left of the magnetic prism, and the ion to electron converter, fluorescent screen and photomultiplier to the right.

Figure 7 shows the scheme of the third type of instrument. This is the block diagram of the ARL ion microprobe mass analyzer instrument. The primary ions are generated in the duoplasmatron and accelerated to the primary magnet where they are mass separated. The selected ions are then focused onto the sample with a condenser-objective electrostatic lens combination. The secondary ions are accelerated to the pickup electrode, focused by a retrofocal lens, energy analyzed with a spherical electric sector, and mass separated by the secondary magnet system. The mass of interest is allowed to pass through the slit where the ions are detected with the ionelectron conversion-scintillator-photomultiplier combination. Mass spectra are obtained