Understanding heterogeneity in Genesis diamond-like carbon film using SIMS analysis of implants

An amorphous diamond-like carbon film deposited on silicon made at Sandia National Laboratory by pulsed laser deposition was one of several solar wind (SW) collectors used by the Genesis Mission (NASA Discovery Class Mission #5). The film was ~1 μm thick, amorphous, anhydrous, and had a high ratio of sp 3–sp 2 bonds (>50%). For 27 months of exposure to space at the first Lagrange point, the collectors were passively irradiated with SW (H fluence ~2 × 1016 ions cm−2; He fluence ~8 × 1014 ions cm−2). The radiation damage caused by the implanted H ions peaked at 12–14 nm below the surface of the film and that of He about 20–23 nm. To enable quantitative measurement of the SW fluences by secondary ion mass spectroscopy, minor isotopes of Mg (25Mg and 26Mg) were commercially implanted into flight-spare collectors at 75 keV and a fluence of 1 × 1014 ions cm−2. The shapes of analytical depth profiles, the rate at which the profiles were sputtered by a given beam current, and the intensity of ion yields are used to characterize the structure of the material in small areas (~200 × 200 ± 50 μm). Data were consistent with the hypothesis that minor structural changes in the film were induced by SW exposure. Electronic supplementary material The online version of this article (doi:10.1007/s10853-017-1267-3) contains supplementary material, which is available to authorized users.


A Close Look at Figure 7c
Background: the image is of the floor of a SIMS analysis pit, and only carbon was present in the EDS analysis. Accordingly, the 1/2 µm crystal was inferred to be diamond. Here, measurements of crystal faces confirm cubic lattice. In addition, light colored (high density) lineations in the matrix were interpreted as features of stress relief. Geometry highlights hexagonal geometry, consistent with the geometry of uniform 2-dimensional shrinkage; basaltic columns or mud cracks are common examples of relief of (tensile) 2D stress. Figure 7c -Why is this beautiful crystal here? A discussion of why it isn't channeling (pages 5-9) Background: multiple crystals were seen in the floors of SIMS pits, the most spectacular in Fig. 7c. Because it is possible for crystals to be oriented with respect to stress, it is plausible that stress-relief in these DLC films formed oriented crystals. If so, then perhaps oriented diamond crystals were excavated (not pulverized) by the SIMS primary beam because the incoming primary ions where channeled along openings in their lattice. Here, we discuss issues with this interpretation. Figure 3: a sample spreadsheet (page 10) Background: each line in Fig. 3 is the best fit to multiple calculations and each line uses a unique set of parameters. This sample spreadsheet shows some of the data behind Fig. 3. Note: the reason why the lines in Fig. 3 require curve fitting is that SRIM presents its results using 10 bins [16], which adds error to the calculation of the depths representing X peak and X half . Changing the model film thickness changes the size of the bins, and therefore may give a different apparent model depth distribution, but within error.

Profile 4
 Reasons why we don't think this weirdness was operator error (pages 11-13)  Could strange intensities and molecular counts be the result of a transition to Static SIMS mode? (page

14)
Background: data produced during the SIMS depth profile changed dramatically after refocusing the primary ion beam (Fig. 11). Normally, (1) this extreme behavior would be considered operator error and (2) the analytical conditions are well into the range of dynamic SIMS [10]. These pages explore why (1) this may not have been operator error and why (2) the initial sputtering (i.e., prior to refocus) may have damaged only the surface layer of the DLC such that the analytical conditions may not have been in the dynamic range needed for depth profiling. Field of Diamonds (white speckles)

jurewicz SOM
A close look at figure 7c Why are these corners and edges so sharp?
Note small coherent diamonds on surface. How did they survive? They are about the size of the scattered beam (for a incoming "point" ion track) in the dlc.

jurewicz SOM
A close look at figure 7c -why is this beautiful crystal there! Model for perfect channeling: • crystal oriented to primary beam • no diffraction or scattering at either at surfaces of crystal or boundaries of "channels" through crystal •~4keV O primary Assume • Constant sputter rate = 0.266A/S (ave) • Amount of exposed crystal =~3000A

Crystal dlc substrate
• Time of "excavating" crystal =~3 hours Shape of beam as penetrated DLC first 100A of surface and edge of crystal • Every excavated surface has been exposed the 3.75keV/nucleon primary beam as scattered by collisions for approximately 6.3 minutes

jurewicz SOM
A close look at figure 7c -why is this beautiful crystal is not there because of channeling.... 75keV 26Mg into graphite, 7 degree tilt Results of calculations used for lines in Figure 3 1.00E+00 All data collected for first 300Å of standard Profile 4.
Rebound at refocus: • 12 C increases 3 orders of magnitude before refocus and stays.
• 12 C 2 increases less than 2 orders of magnitude. 25Mg, 26Mg drop slightly less than 1 order of magnitude and stays. • 24Mg increases 2orders of magnitude, possibly due to surface dirt and/or deflected beam?
Full Profile of standard Profile 4 and comparison to Profiles 2, 3. • Mg species profiles collected for Profile 4 looked reasonably shaped after refocus. • Near surface contamination larger than might be hoped.
(Possible deflection / misalignment / small particle? Note that beam had been realigned at refocus by senior analyst). • Note difference in early 12C2/12C ratio in early refocused profile compared to Profile 7: in comparison, it seems anomalously high in the transient zone here, too. Modeling of Profile 4 Did the weirdness in the first 307 seconds represent going from dynamic mode to static mode on the SIMS Because the area was so diamond-like?
McPhail model of stacking spheres if the diameter = 5 n m diameter of the circle (nm)= 0.0000005 cm # of circles per cm2 = 4 E + 1 2 formula is (1cm/(diameter of circle in cm))^2 actual area of circles = 0.785398163 cm2 actual area of raster = 1 c m 2 So McPhail stacking sphere model leave about 20% of area untouched by ion.

SRIM of impact of Ar atom into silicon.
I chose Si because that is the major material these guys look at, so I thought it a good guess.
Ar @ 10keV range = 15.5 nm lateral range = 5 . 5 n m But, the reason that the diamond-like areas need a larger compound correction is because diamond has a large band gap, which non-ideally dissipates the energy, so that less goes into the breaking of bonds and sputtering. If this is happening, the range for depth should be different as well. So, we were running between ~2E14 and ~4E14. SRIM doesn't do a great job on low energy implants, especially into non-ideal solids. However, if we scale to what is observed, at a density of 3.3 gm/cm2 for the non-ideal DoS static SIMS starts at 2E14. So, we were running at the very edge of static SIMS. If so, focusing and defocusing the primary beam would have changed the current

It is plausible that with the large, round beam we could have been in Static Mode SIMS, but with the point beam, we could have been back in dynamic mode!!!!!
We note that low sputtering yields (as per diamond/large band gap) would also explain the low counts for all species in Profile 4.

jurewicz SOM
Profile 4: Could strange intensities and molecular counts be the result of a transition to Static SIMS mode? Physics problems related to sputtering and possible increased formation of molecular ions during the SIMS analysis.
• At first glance, it appears that the sphere will charge completely in a matter of seconds. • It is likely that the amount of charging will vary (at least at first) with the net distance measured from the surface of the sphere to the conductive layer. (12kev to 0 kev on primary O ion beam). • Primary ion beam may be deflected, or may simply be decelerated, creating little damage to the crystals.
2. Presence of insulating layers in analysis region. Open circles are estimated ion density assuming SRIM is correct at these levels and that the material is behaving uniformly as #906 Nuclear Grade Graphite.
Estimated charging models using SRIM if the insulating layer is still behaving as 3.2 gm/cc #906 Nuclear Grade Graphite.
jurewicz SOM Profiles 2, 3, 4: Could inhomogenieties make very small areas charge resulting in extra contributions of molecules (cf., Static SIMS mode)?
Since SRIM overestimates the radial damage at these energies, the calculated ion densities are a lower limit for the current needed to saturate the surface of the DLC.