Dr. Marvin Groeb – Abrasive Solutions

Category: Allgemein

  • Musings about material removal – Part 2 – Dressing sharpening stones and inevitable contamination

    TL;DR:

    Lemma: When flattening / dressing a sharpening stone, what’s the best approach to avoid contamination?

    Methodology:

    • Made some contamination free 10 µm resin diamond sharpening stones with a low to medium grit concentration (20% by weight).
    • Flattened 1 sample each on a glass plate with a water slurry and a typical abrasive: 3 different grain sizes of SiC, on diamond powder of identical size
    • Flattened 1 sample on an electroplated stone under running water
    • Flattened 1 sample on a piece of SiC sandpaper (F1500).
    • Looked under the SEM for contamination – using a backscatter detector which shows elemental contrast and EDS, which identifies elements and thus nails contamination down.

    Results:

    • Flattening a resin stone on SiC will embedded the SiC particles in the surface, no matter how small the particles are (larger / smaller than the grit of the stone).
    • Flattening a resin stone on EP stones works very well, but you will inevitably catch a couple of those larger diamonds. It also consumes the EP stone very quickly.
    • Flattening a resin stone on diamond powder will embed the diamond in the surface layer, but also create quite a bit of glass-particles that will get stuck in the surface layer. Those particles are sub micrometer sized and I would consider them a health hazard if airborn.
    • Flattening a resin stone on SiC paper tears out the diamonds, surprisingly enough. There is very little contamination.
    • A better flattening / dressing approach is needed to keep stones pure. I employ one on the DrMarv stones that does not contaminate. Watch out for a future blog post on how to do this.

    Actual Science and long version:

    This is part of a series of blog posts, where I try to apply my professional knowledge on how chip formation and material removal happen to knife sharpening. I think this could also be called: debunking myths. Because this probably will ruffle some feathers, and is likely to be denied by some people, let me state firmly here: everything you will see in this post is real, and repeatable.

    Sharpening stones experience uneven wear. This is because they are inherently anisotropic in their composition, but also because we as humans use them inequally. Often, the end parts of the stone do not get used, as you do not want to fall “off the edge”. Moreover, different movements, pressures and just general wear sometimes require you to flatten, dress or renew a stone. For simplicity’s sake, I will from this point onwards call the process “dressing”, as it is the technical term. What you apply this to (renewing the surface, flattening the stone or actual dressing, e.g. creating a surface morphology suited to the application) is irrelevant, as the general mechanic is a 3 body abrasion on a flat surface.

    I’ve done a ton of reviews on sharpening stones so far, and the largest majority of these had particles embedded that shouldn’t actually be in them. Some of these contamination stem from non-sufficient abrasive hygiene in the factories (you’d be surprised how easily micron sized powder becomes airborn and lingers for hours!), but some also from dressing these. At higher concentrations, SiC is used as a filler material, decreasing cost and increasing the bond hardness.

    Three different approaches have been analysed for this:

    I.) Abrasive slurry on a glass plate

    A often suggested dressing method is a piece of glass (which is surprisingly flat!) or granite, and to mix an abrasive slurry on top of this glass surface. Often, SiC is used, as it’s a hard abrasive, easily and cheaply available. If you want to take a look at what the SiC I used looks like, head to this abrasive snippet.

    For the actual dressing, I took a flat piece of silica-glass. It was cleaned by rubbing it vigorously with a soap-water mixture and then rinsing under running water. A small amount (about 3 grams) of SiC powder was applied to the surface and water added until a slurry that is soft but does not run away much was created. The sample specimen was then moved in a figure 8 movement across the surface until the top surface flattened out nicely. The sample specimen was then rinsed under running water while being rubbed intensively. Afterwards, it was dried and put into a sample holder where it is unable to touch other samples. Before SEM analysis, the sample was cleaned with a de-ionized water steam cleaner (110°C, 3 bar) and rinsed with 99 % pure ethanol. It was dried with a blast of compressed and filtered air.

    Optical micrograph of the dressed surface. Abrasive used: SiC FEPA 300, 400 and 1000. Note the very even and homogenous surface and continuously smoother surface. Instrument: Leica Emspira.

    3 different grain sizes of SiC were employed to dress 3 identical samples. FEPA grades 300, 400 and 1000, corresponding to micrometre sizes of (roughly) 34, 17 and 4.5 µm. This means there is a much larger, a larger and much smaller grain size used. The microscope images clearly show that it’s very hard to pick out the SiC particles from the actual abrasive, as they are very small and every particle is reflective. Playing with the polarisator on our microscope didn’t really change anything on this.

    Luckily enough, it is very easy to distinguish SiC particles in the SEM, as Silicon is a much heavier element than the carbon predominant in resin and diamond. Thus, the backscatter detector (BSD) shows these as very bright particles. Final identification can be undertaken via EDS analysis. If you are unfamiliar with these techniques, I’ve written a blog post explaining these.

    Let’s start with the FEPA 300 grit SiC powder:

    SEM micrographs of the FEPA 300 SiC powder dressed resin stone. The lighter coloured particles are SiC.

    A very large number of evenly distributed, very bright particles is visible. Because of the used detector type (BSD), this means the particles consist out of heavier elements than the surrounding material. As diamond and resin both mostly consist out of carbon, this is alreadys a very good hint that we are looking at SiC. EDS analysis reveals this:

    EDS analysis of the FEPA 300 SiC powder dressed resin stone. The lighter coloured particles are easily identifiable as SiC.

    The finer grit SiC (FEPA 400) also shows a similar picture – evenly distributed, top layer embedded SiC particles.

    SEM micrographs of the FEPA 400 SiC powder dressed resin stone. The lighter coloured particles are SiC.

    EDS analysis confirms these particles to be SiC again.

    EDS analysis of the FEPA 400 SiC powder dressed resin stone. The lighter coloured particles are easily identifiable as SiC.

    The finest SiC grit used, which is below the abrasive size (4.5 vs 10 µm diamond) also shows fine, embedded particles.

    SEM micrographs of the FEPA 1000 SiC powder dressed resin stone. The lighter coloured particles are SiC.

    EDS analysis also confirms that these are SiC particles.

    EDS analysis of the FEPA 1000 SiC powder dressed resin stone. The lighter coloured particles are easily identifiable as SiC.

    Next, I took the “expensive” route. The stone is made from 10 µm diamond powder, why not take that same-batch diamond powder to dress it? After all, it is an abrasive, and contamination can’t happen here…can it?

    Optical micrograph of the dressed surface. Abrasive used: 10 µm diamond powder. Because of the low material removal rate, only a smaller section was dressed. The different area in the top left colour is from before-dressing. Instrument: Leica Emspira.

    One thing of note is that the diamond slurry is pretty expensive to make, and it also is very slow in the dressing. I rubbed about twice as long as on the SiC slurry, and had a smaller section dressed. Nevertheless, it’s enough for SEM analysis!

    SEM micrographs of the 10 µm diamond powder dressed resin stone.

    We can immediately see, that some diamond was embedded into the top surface. It’s visible as flat grains, that are all equally sized – this is different to how the actual resin stone looks like, where diamonds are always at different depths and peaking in or out. Moreover, if one zooms in very far, small, bright particles become visible. This is interesting, as they are sub 1 µm sized!

    EDS analysis of the diamond powder dressed resin stone. The lighter coloured particles are identifiable as SiO2

    EDS analysis shows that these particles consist out of oxygen and silicon. Spot analysis confirmed a ratio of 1:2, so this is likely SiO2 – the glass plate we used seems to have abraded and embedded itself into the stone. I find this very fascinating, as we didn’t really see these particles on the SiC dressed stones! Sub micron particles out of glass can be considered a health hazard. I would advise to wear a mask on all heavy abrasive actions!

    II. Dressing on electroplated diamond plates

    One pricey alternative is to use an electroplated diamond plate. The advantage here is that the abrasive is pretty firmly bound to the plate, so chance of contamination should be lower, and we can aid the process by dressing under running water, which will automatically flush between the two abrasive bodies. The downside is besides the cost of the plate a very large wear on the EP plate.

    Optical micrograph of the dressed surface. Abrasive used: 600 grit EP diamond plate. Note the very even and homogenous surface with some very bright particles. Instrument: Leica Emspira.

    Under the optical microscope, a few large, bright particles are visible.

    I was hunting for these under the SEM, and while I mostly found a well dressed resin stone, a few of these larger particles were also visible!

    SEM micrographs of the EP diamond plate dressed resin stone.

    The particle size of > 25 µm is consistent with the 600 grit EP stone used.

    EDS analysis of the diamond EP 600 grit dressed resin stone. The large particles identify as carbon, so are likely diamond particles that have come loose from the EP stone.

    While it’s a nice dressing result, and the contamination is rare, these diamond particles are 3x the size of the resin stone diamonds. These particles will continuously create very deep scratches.

    III. Dressing on SiC sanding paper

    Lastly, I dressed a sample on some SiC sanding paper. This was done with a bit of ethanol as a lubricant. The SIC paper was by far the quickest way to dress the surface, but left some deeper scratches visible on the surface:

    Optical micrograph of the dressed surface. Abrasive used: FEPA 1500 SiC sanding paper. Note the visible scratches. Instrument: Leica Emspira.

    SEM analysis reveals not only these scratches, but also some brighter particles. Moreover, instead of really dressing the surface, it also removed all surface layer touching diamonds! We are left with a very porous surface with few diamonds visible. I would guess that this surface will immediately clog with swarf.

    SEM micrographs of the FEPA 1500 SiC sanding paper dressed resin stone. The lighter coloured particles are SiC.

    EDS analysis confirms the bright particles as SiC.

    EDS analysis of the FEPA 1500 SiC sanding paper dressed resin stone. The lighter coloured particles are easily identifiable as SiC.

    Conclusion

    I am quite surprised. All analysed dressing methods leave us with contamination. I would say in aspects of having the best performance, the diamond slurry dressed stone is probably king – no larger or other particles are introduced. The minuscule amount of glass should be easily avoided by wearing personal protective equipment – which in my opinion, you should always wear during dressing.

    The SiC dressing embedded foreign particles. At sizes larger than the diamond, this will leave you with a stone that produces deeper scratches. At a size smaller than the diamond abrasive, you’ll just increase the grinding pressure, and make your diamond stone slower.

    The SiC sandpaper was surprisingly effective, considering it was a finer grit. Unfortunately, it also removed the diamonds from the matrix, a first here. I wouldn’t recommend it here!

  • A brief study on sharpening stones – Part 12 – PD Poltava Tools Premium CBN (2.5  µm)

    This is part of a series of blog posts – looking into the appearance and composition of commercially available sharpening stones. If you are interested in the previous episodes, check out the archive for them.

    If you have some suggestion on what I should look at next, or want to share your super secret DIY stones, I could be persuaded to open the bag of analytical devices… hit me up on Instagram under @marvgro for that.

    Disclaimer: I’m not for sale. Every review you see on this blog is bought with my own money. I have no affiliation to any manufacturer.

    Review

    Today’s stone is the “sister” one to the one review in part 10. It’s from the Ukrainian company „PT.tools“ (also known as PDT or Poltava), which are a manufacturer of abrasive tools. It uses a bronze bond, and the grit chosen (2.5 µm) is perfect for polishing, according to the manufacturer. This is a rare CBN superabrasive stone.

    Maybe a quick detour on the nature of CBN:

    Diamond, as a super-abrasive is used because it’s the hardest material we can manufacture in large quantities. It’s super hard, but suffers from one major issue: it’s the meta-stable configuration of carbon, and not only is it susceptible to high temperature graphitisation, it also experiences chemical wear in a number of materials, first and foremost in steel.

    CBN is the abbreviation for cubic boron nitride – funnily enough, it doesn’t look cubic 🙂 check out some more SEM pictures on this blog entry.

    It has the same crystal latice structure as diamond. But what’s that structure?The crystal structure is called fcc = face centered cubic, a cube with atoms on the corners and faces of the cube.

    A representation of the fcc crystal lattice.

    For the “diamond” or CBN structure, two of these lattices are intersected into each other, at a shift of 1/4 of the lattice cell:

    “diamond” lattice structure, which is two identical fcc cells intersected at 1/4 of the lattice.

    Moreover, CBN has the same bond between the atoms as diamond – a so called “sp3” orbital. This is a very strong chemical bond. The amazing hardness of CBN and diamond stem from this bond mechanic, the interatomic distance and crystal structure. While diamond clocks in at 10000 HV in the hard direction, CBN is more around 4500-5000 HV. But unlike diamond, CBN experiences no chemical wear when machining steel, which made it the abrasive of choice in commercial, precision CNC grinding of hard steels. CBN generally is more expensive by weight than the same amount and grade of diamond abrasive.

    Optical micrographs of the PT CBN 2.5 µm stone. Instrument: Leica Emspira.

    The stone is virtually indistinguishable to the naked eye from the previously reviewed diamond stone of the same grit. The stone is very firm, showing a dark grey colour, that slightly reflects reddish/bronze coloured when the light hits it. Under the microscope, a very even structure is visible. Individual grits are near impossible to make out, because the bronze binder is so reflective.

    Let’s take a look under the scanning electron microscope!

    SEM micrographs of the PTD CBN 2.5 µm. Instrument: Zeiss GeminiSEM560.

    Even at 2kx magnification, it’s hard to make out the CBN. The stone is dominated by large SiC particles and the metallic bond. But fear not, avid reader, I’ve zoomed in far and found a CBN grain for you:

    There you have it! A sole grain in specification of the size stated by the manufacturer, and quite a bit of smaller debris, likely from the dressing of the stone. Around the grain, some cracking is visible. Even though it is nearly fully embedded, no good wetting (covering of the metal matrix) is achieved, giving it a loose appearance. EDS analysis reveals a couple more grains, but the bond is dominated by SiC embedded into it. The SiC particles are about one order of magnitude larger than the CBN abrasives.

    EDS analysis of the PTD CBN 2.5 µm. Instrument: Oxford Ultim Max  ∞ 40mm2 EDS sensor. Note that our EDS sensor doesn’t show elements lighter than boron.

    Zooming in a bit further, the CBN really becomes visible:

    EDS analysis of the PTD CBN 2.5 µm. Instrument: Oxford Ultim Max  ∞ 40mm2 EDS sensor. Note that our EDS sensor doesn’t show elements lighter than boron.

    The surface height maps are very similar to the diamond stone with the same matrix:

    Instrument: Bruker Alicona µCMM, 50X objective lens, 3×3 FOV high resolution focus variation scan. Data is leveled and outliers removed (0.25%).

    The ISO 25178 parameters show a slightly coarser and rougher surface than on the diamond stone. I would consider this to be larger than within the process variation of a commercial flat grinding process.

    ISO 25178 parameters of the PTD CBN 2.5 µm.

    In order to evaluate the sharpening performance and material removal mode of this stone, a blade was sharpened with it. I am using a standardised testing procedure, read about it here. Nevertheless, it’s 65 HRC M398, and sharpened to 17 DPS with resin bond diamond stones down to 10 µm. Afterwards, the tested stone is used, first in a back and forth movement until the surface becomes homogenous, and then alternating strokes (5-5-3-2) on each side, for a total of 20 strokes towards the apex per side. No pressure is applied but the weight of the apparatus.

    The edge is then analysed in the electron microscope for breakouts and morphological appearance.

    SEM micrographs of the blade finished with the PTD CBN 2.5µm. Deep, regular scratches are visible that were created by the stone, but also fine, smearing of the material. The apex is completely dulled. At least there is no burr 🙂

    The surface finish left by this stone was abhorrent. After just 2 passes, the whole surface turned matte and super dull, with lots of visible scratches. The SEM pictures show this very clearly. It is even worse than the similar sized diamond stone.

    The verdict for this stone is very similar to the equally sized diamond stone with copper bond. At this concentration, with these many large SiC particles, this isn’t a very good stone. I don’t believe that copper-bronze metal bonds at this particle size are very suitable for non-CNC based sharpening. At larger grain sizes, and with a large increase in grain density, I could see this as a fantastic, very tough, long lasting sharpening stone.

    Sharpening disclaimer: I use a standardised approach to sharpening, which basically follows how most manufacturer of guided systems tell you to use this system. I am very aware, that every stone could perform much better than this, in terms of sharpness, but I want a comparable approach. The sharpening segment mostly shows the material removal mechanism – is it burnishing? is it cutting? is the cutting pressure too high so that carbides crack? Is there massive burr or prow formation? The BESS value definitely doesn’t highlight the ultimate sharpening performance of the stone, but was an often requested information. Over time, this blog will show BESS values for different edge morphologies, but by the holy endmill – don’t read it as a „this is the max value this stone can achieve“. I would also suggest to familiarise yourself with the works of Immanuel Kant, it’s absurd I need to write such a disclaimer here.

  • How I prepare blades for the brief study of sharpening stone segments.

    I have an ongoing blog segment where I study sharpening stones. It’s typically commercially available stones, and the focus is to show what the stone looks like microscopically, what it consists of and how the surface generated relates to the stone. Each of those segments contains a sharpened blade, where pictures of the edge are taken in the SEM (and starting from segment 13ish onwards, also optical micrographs). I’ve had quite a few questions on how I prepare /sharpen the blade, so I’ve decided to put my method down.

    Disclaimer: I do know, that this will not operate every sharpening stone at it’s maximum performance, as it would require adapting to the stones behaviour.

    I’ve decided to go for a standardised approach, as my goal is to explore the abrasive behaviour, and not proof what a skilled sharpener (or not) I am. If I were to adapt every stone, the result and surface morphology would likely improve – but a comparison would still not be possible, as I wouldn’t be equally skilled on all types of stones. Therefore, all stones are kind of treated “the same”, and I look at quality parameters such as scratches, pressure induced cracking, blunting, burr formation and prows. I specifically do not take pictures of the apex width, even though our fantastic GeminiSEM560 could actually record those, having a verified sub nanometre resolution.

    Steel Blank

    The steel used is Boehler M398 Microclean, a high tech powder metallurgical steel. The nominal composition of this steel is:

    2.70 % Carbon 0.50 % Si 0.50 Mn 20.00% Cr 1.00 % Mo 7.20 % V 0.70 % W

    SEM backscatter pictures of the carbide distribution and morphology of a prepared steel blank in M398.

    EDS analysis, identifying the small, very dark carbides as Vanadium carbides, and the larger, mid-grey coloured ones as Chromium carbides.

    The blanks I used are heat treated by Roman Kasé in Switzerland. Roman has become a good friend and trusted mentor in all things sharpness, and he is known as the Master of heat treatments. His approach is very scientific, detail oriented. The heat treatment includes multiple cryo treatments, and the steel was measured at 65 HRC hardness.

    Edge Preparation

    The steel blank is thinned out on a high precision 5X CNC machine, a Kern Micro HD. I use japanese high quality tools to mill away about 1 mm on every side. The regime I am milling in is called high speed cutting, where the majority of the cutting heat is generated in the chip, and the part is cool to the touch immediately after milling. As long as there is a certain thickness, this is a fantastic preparation method which doesn’t affect the heat treatment. Afterwards, the blank is thinned and a cutting edge is created at 17 DPS via wet grinding on my Tormek T8. I use a 400 grit, 1000 grit CBN wheel. The water is changed between every stone, and the grinding process is kept VERY wet. Talk about flooding my kitchen! 🙂

    The prepared blade blanks after milling.

    Sharpening Procedure

    I use a TSPROF K03 Pro to sharpen. The blade pieces are fixed in the dual clamp holder. The angle is set at 16.9 DPS for roughing, and then ground with 120, 80, 60, 40, 20 and 10 µm resin bond stones. The stones I am using are my own design – DrMarv’s Scientific Sharpening Stones! They feature an unheard of concentration, a chemical treatment to the diamond to increase bond strength, no fillers or additives in the resin, and every batch is QC inside the SEM for particle distribution, size distribution and overall quality.

    SEM micrograph of a commercial diamond stone (Venev OSB) vs a DrMarv Scientific sharpening stone (right side). The bright particles are local charge ups of grains. Note the super high concentration, good particle size distribution and missing conglomeration.

    The blade is ground wet (splashed stones) with water, and meticously cleaned of swarf between every stone change. The sharpening movement is about 20° off the perpendicularity of the blade, and alternated in direction (“lower left to upper right” vs “upper left to lower right”) between stones. Each stone is used until the grinding marks from the previous stone have completely disappeared.

    At 20 and 10 µm, the angle is increased to 17 DPS (so that the edge is leading). All stones are used in a back-and-forth movement. The 10 µm stone finishes the blade, and the last 20 strokes per side are done towards the apex, alternating a 5-5-3-2 pattern.

    Edge after the preparation with the 10 µm DrMarv stone. The scratches are 1-2 µm wide and < 1 µm deep.

    Then, the edge is marked with a black marker, and the to-be-tested stone is adjusted to 17.0 DPS. The chamfer is then ground until the black marker is gone, and a homogenous surface appears. Then, the previous light pressure is relieved, and 20 strokes per side are applied at system weight pressure only. The strokes are applied towards the apex, alternating a 5-5-3-2 pattern, swivelling the mount in between.

    The blade is then cleaned by wiping it across a wet microfibre towel (about 30 mm) once per side. Before analysis, only thermo-chemical cleaning happens: steam cleaning at 6 bars pressure (110°C, de-ionised water) followed by an ultrasonic bath rinse in acetone, isopropyl alcohol, ethanol alcohol, ultra-pure water and then blow dried with compressed ultra pure (99.997%) nitrogen gas.