Sample Prep

Details of the effects of sample topography are reported in Kita et al. (2009). Kita et al. nicely illustrates the effect of polishing relief on the reproducibility of oxygen isotope measurements in zircon standard grains (KIM5).

A polishing relief of <3 μm is generally necessary for high-precision analysis, and a relief of <1 μm is optimal. For epoxy mounts made for other analyses (e.g. U-Pb analysis by SIMS), polishing relief is sometimes ~10 μm. You can remove this type of relief by hand with 3 μm and 1 μm diamond lapping films in succession.

Effect of Topography

Once you have cast your sample, standards, and labels in an epoxy mount, it is time to lap and polish the surface. Again, the objective here is to obtain an analytical surface with no more than a few microns of relief – and negative relief is much better than positive relief. For samples that are much harder/softer than the epoxy, you will need to strike a balance between eliminating small scratches and creating surface relief by overpolishing (resulting in the softer material being removed faster).

There are a number of ways to polish an epoxy round and we can’t offer detailed advice on each. Below, courtesy of specialist Brian Hess, we outline the general method developed at UW, which uses a progression of diamond grits (both fixed and loose) on a spinning polishing wheel. No matter which method you choose to use, we strongly urge you to practice your approach before polishing the mount(s) you will analyze at WiscSIMS. Polishing is a destructive process and it is possible to polish right though your sample and/or standard grains!

Important Note: Our recipe does not gurantee success, and does not impart the requisite skills or decades of experience required to master this process.

There are some third-party companies that will polish your sample for a fee, but we do not have extensive experience with any of them. You will want to carefully describe and document the degree of polishing you desire, and be sure to communicate the scientific value of each mount. If samples show polishing relief, it may be possible to fix with a diamond lapping film.

Third-party polishing services; please contact us for an up-to-date recommendation:

• Spectrum Petrographics Inc.; Vancouver, WA
• Wagner Petrographic; Lindon, UT (thin-section specialist)

Materials used for lapping and polishing at UW:

• Fixed-diamond lapping pad (6-μm grit), 12″ diameter (Buehler UltraPrep, part#: 15-6206)
• Magnetic backing, 12″ diameter, for mounting fixed-diamond pad to rotating lapping wheel (Buehler MagnoMet, part #: 16-3072)
• Soft toothbrush and Liquinox soap [Critical cleaning liquid detergent, biodegradable, phosphate-free (part #: 1201)]
• Low-nap polishing pads with magnetic backs, 8″ diameter, mounted to rotating polishing wheel (Allied High Tech)
  – separate pads for 6- and 3-μm grits [Allied Gold Label polishing cloth (part #: 90-500-210)]
  – for 1-μm grit [Allied White Label polishing cloth (part #: 90-500-500)]
  – new pads are used after every ~5 mounts for SIMS-level polishing (would still be viable for polishing thin sections or for SEM/EPMA)
• Polycrystalline diamond suspensions (Buehler MetaDi Supreme Series)
  – 6-μm (part #: 40-6632)
  – 3-μm, (part #: 40-6632)
  – 1-μm, (part #: 40-6632)

And if you have access to a Buehler VibraMet system:

• Vibrating pad, slow-speed high-nap pad (Buehler Microcloth, 12″ diameter with adhesive back, part #: 40-7222)
• Colloidal alumina 0.05-μm poslishing suspension (Buehler MasterPrep, part #: 40-6377-032)

Step-by-step instructions:

• Grind the analytical surface with the 6-μm fixed-diamond lapping pad. At UW, we use a 12″ pad on a wheel spinning at 1140 RPM and lubricated with a constant flow of water. The goal here is to prepare a flat surface for polishing, and to appropriately expose the samples and standards as needed. Note that after lapping, the polishing process will still remove some material (a few μm thickness).
• After each lapping/polishing step, examine the surface under magnification with reflected light to assess progress, exposure and surface roughness. First, you’ll need to rinse the mount with warm tap water and scrub lightly with a soft toothbrush and concentrated Liquinox soap to remove any lapping/polishing residue. Do a final rinse swith distilled/R.O. water (to avoid deposits from tap water) and blow dry with clean air before examining under magnification.
• After lapping, and using a different low-nap pad for each grit, polish the analytical surface with a progression of 6,3,1-μm polycrystalline diamond suspensions. At UW, we use an 8″ polishing wheel spinning at 550 RPM, and only use the (slower-moving) inner 2.5″ radius for polishing. No more than 1 (one) minute is spent on each grit. Examining the sample on a reflected light microscope after each step will reveal:
  – After the 6-μm grit, the sample will not appear to be polished
  – After the 3-μm grit, the samples will be noticeably different and will visually “pop”
  – After the 1-μm grit, surfaces will appear almost clear of scratches/imperfections
• For some samples, a final polish is achieved with a 0.05-μm colloidal alumina solution. The alumina solution is placed on a vibrating pad and the mounts are placed (face-down) on the vibrating pad under ~900 g in custom weighted holders for 1-5 min (10 min for thin sections) depending on the relative softness of sample/epoxy. It is critical to repeat the cleaning with soap immediately after this step to remove colloidal alumina residue and give a final rinse with distilled/RO water. Check for polishing residue under reflected light.
• It is easy to fixate on removing small scratches during the polishing sequence, leading to overpolishing that can quickly cause topographic relief. You want no more than 3-μm of relief, and ideally <1-μm. If you introduce too much surface relief at any step, you’ll need to start over at the grinding step (fixed-diamond lapping pad). A few scratches in a polished sample are (usually) easy to avoid during analysis, and are certainly preferable to the results of overpolishing.
• Once polishing is complete, and flatness is confirmed (see next step), cut the back of the mount off so that it is 4-5 mm thick. Ensure that the cut is parallel to the polished face so the mount has an even thickness, and avoid scratching the polished surface! After cutting off the back, you can improve label legibility and remove saw scuffs from the cut surface by briefly lapping it with the 6-μm fixed-diamond lapping pad.

You can check the sample flatness by either optical microscope or white-light profilometer. However, it is not always possible to judge relief by optical microscope and samples may require scanning by ZYGO white-light profilometer or a similar instrument with sub-micron resolution.

By optical microscope, use the hash marks on the focus knob to determine the difference in height between the focal plane at the sample vs. epoxy surface.

White-light profilometers are commonly housed in the Materials Science department of a university and can produce sub-micron-resolution maps of surface topography.

Careful imaging of your polished mount is essential for analysis. You will have a 500 x 400 micron view of your sample while it is in the SIMS, and it can be difficult to orient yourself with such a small field of view. Thus, it is recommended that you have at least: 1) a single image of the entire “sweet spot”; 2) high magnification images of your entire analytical target (for 10-micron spot analyses, it is useful to have image maps where 1 pixel ~ 1 micron). If you have a large analytical target, you may need to create a large sample map by digitally stitching multiple high-magnification images together with Adobe Illustrator, FIJI, or Microsoft ICE.

Recall that non-conductive samples will be gold- or carbon-coated for WiscSIMS analysis. As a result, you will only be able to see cracks, scratches, and some grain boundaries with the WiscSIMS camera. For users that are targeting geochemical features that have been imaged by other means (e.g. SEM, fluorescence microscope), it will be helpful to have both the geochemical imaging as well as optical imaging of the gold-goated surface.

You can coat your sample at UW if necessary, otherwise a ~30 nm coat is appropriate for most applications. Note: If you plan to coat your samples before arriving in Madison (e.g. for SEM analysis), we recommend that you clean and degas your mounts before coating to prevent problems under the high vacuum in the instrument. To clean your epoxy mount (mostly of fingerprint oils), we sonicate each mount for 30 seconds in ethanol then again for 30 seconds in DI water. If possible we move the cleaned mount, using gloves, to a vacuum-oven to dry for 3+ hours. It’s a good idea to place new mounts under high-vacuum (10^-6 Torr) overnight to degas the epoxy. If you cannot clean or degas your sample before coating, we will remove the coating to perform these steps when you arrive for your analysis session.

Thin slices of biological samples (<few μm) placed flat on a Si-wafer can be analyzed without surface coating.

Following ion microprobe analysis, you will need to screen the analysis pits by SEM for aberrations like cracks or inclusions. The easiest way to do the screening is immediately following analysis and before you remove the conductive coating. After your analysis session, if you add a thin conductive coating (e.g. 30 second gold-coating at the WiscSIMS lab) to coat the pit-bottoms, you can take your sample directly to an SEM for imaging. It is possible to image pits once the coating has been removed by using Environmental Mode (low-vacuum) SEM settings, but this is not ideal.

Settings that have proven useful for imaging coated samples on the UW SEM: 15 kV beam, 40 mA current, secondary electron (SE) imaging, 4000x magnification.

Use the SEM to inspect the condition of each pit-bottom, both samples and standards. We classify the condition of each pit as either “regular”, “irregular”, or “borderline.” These classifications are somewhat qualitative and your threshold for “irregular” may change with experience, but this process is useful for independently identifying potential outliers that may be eliminated from a published dataset.