Grit, Scratches and Sub-Surface Damage – Part 1

It has become a somewhat common practice to follow a “progression” of increasingly finer grit abrasives when sharpening a blade. In principle, this process is comparable to the procedure for metallographic polishing, which typically consists of three stages, coarse grinding, fine grinding and polishing. These steps are usually characterized by “the scratch pattern” for want of a better option. While techniques for metallographic polishing are very well established, there is inadequate understanding of precisely what scratches patterns are, and what is actually happening to the structure of steel at the micron scale. Additionally, as I have shown in the Optical vs Electron Microscope article, bevel scratches are not useful in characterizing the apex, and in general, scratches are not unambiguously characterized optically. To better understand the so-called “scratch pattern”, electron microscope examination and analysis is necessary.

In the SWARF! article, I showed that size of metal chips increases with grit size, although more slowly than might be expected. In the Abrasion Rate vs “Grit” article, I showed that rate of metal removal somewhat increases with abrasive size, with some exceptions. In this article, I will discuss abrasion and the role of abrasive grit size in the surface texture and sub-surface damage to the steel.

The interaction of a grit particle with a metal surface is called “single-grain grinding” and (from R.S, Hahn, 1962) is generally characterized by three modes; rubbing, ploughing, and chip formation. During hand sharpening, all three of these can occur, depending on the shape and orientation of the grit particle and the applied pressure.

A grit particle with a sharp corner or edge can penetrate the surface, and as it moves will shear the metal and expel a “chip” of metal. If a sharp edge isn’t exposed, or the local pressure isn’t sufficient for the grit to penetrate, rubbing will occur. In some definitions, burnishing is distinct from rubbing in that rubbing does not cause plastic deformation while burnishing does. For this discussion, I will include burnishing as part of the rubbing phase, as rubbing without plastic deformation rarely occurs. (To further complicate this picture, as I have shown in the What does Steeling Do? article, rubbing can also remove metal via adhesive wear).

Ploughing occurs when the grit particle penetrates the surface, but doesn’t expose an efficient cutting face; upheaval of metal occurs at the sides of the groove (just like the furrow formed by ploughing through soil). Although we are focused on the dimensions of the abrasive for now, it must be appreciated that the shape, orientation, density, friability as well as how firmly embedded the grit is in the sharpening media will affect the outcome. It should also be expected that chip formation requires a minimum depth of cut (or penetration) of the grit. For example, coarser grits with insufficient local pressure will not penetrate the surface and primarily rub or burnish the surface.

Initially, I will show polished surfaces with only a finite number of isolated scratches, or “single-grit interactions.” However, when grinding or sharpening, hundreds of single-grit events occur simultaneously, and rapidly overlap one another.

In the first set of images, a hardened carbon steel coupon was first polished, and then abraded slightly with 600 grit silicon carbide sandpaper (Norton Black Ice) to create a few isolated scratches on the surface. In the images below, we observe that both ploughing and chip removal can occur simultaneously.

SEM tilt-view of a scratch from 600 grit silicon carbide in hardened carbon steel. Pile-up is visible on each side of the groove and a partial chip can be observed near the top of the image.
SEM tilt-view of a 600 grit scratch terminating when part of the grit particle came loose from the sandpaper and became embedded in the surface.
SEM tilt-view of a 600 grit scratch in polished, hardened carbon steel. Metal piles up along the edges as it is displacement from below. The morphology of the pileup is reminiscent of slip along grain dislocations.

While the tilt-view images of the scratches are informative, a cross-section view is key to understanding this process. In the image below, the scratch was cross-sectioned after a thin layer of platinum was deposited on the surface to aid both the cross-sectioning procedure and the imaging of the cross-section. Not only is the topography of the scratch clearly visible, but the grain-structure produces contrast in the image. Unlike conventional metallographic etching, this type of grain structure imaging is subtle, but provides much higher resolution.

Cross-section view of a single scratch from 600 grit silicon carbide paper in a polished piece of hardened carbon steel.

The cross section image is shown again below, with the line of the original surface marked in black. The region where the grain structure is disrupted/damaged is marked in red. The relatively large grains with underlying sub-structure are visible below the surface, but those grains are clearly distorted with sub-micron texture within about 1 micron of the scratch boundary.

Cross-section of a single scratch from 600 grit silicon carbide paper. The damaged steel is marked in red. The original surface is marked with the black line. The ploughed metal is observed as the red-marked area above the black line.

In the interaction of the single grit particle with the surface, there are three results;

  1. Metal is removed via the formation of a chip
  2. The surface is raised at the edge of the scratch
  3. The metal near and below the scratch is damaged.

When grinding, multiple scratches will form and overlap preceding scratches with each pass. It is observed that the surface reaches an equilibrium of texture (depth and width of scratches) with a characteristic depth/thickness of damaged metal.

This brochure: SiCpaper.pdf from Pace Technologies ( shows measured surface roughness on steel for a range of silicon carbide grit papers. Their data show that ANSI 600 (15.3 micron particle) was able to produce a roughness of 110nm (0.11 microns), less than 1% of the particle size. This polishing was done in random directions, and obviously abrading in a single direction as is typically done with sharpening will produce somewhat more roughness.

In the image below, we observe a scratch that has been partially filled in (burnished) by subsequent nearby grit interactions.

Cross-section view of a 320 grit scratch that is partially filled by the metal displaced by subsequent scratches.

Rather than attempt to characterize the size and depth of “scratches” created by abrasive particles, which can be very misleading, I will examine the grain structure in the cross section to roughly measure the depth to which the grain structure is disrupted.

SEM tilt-view of a hardened carbon steel surface abraded with 320 silicon carbide paper.
Cross section view of 320 grit surface. Damage extends 1.5 to 2 microns below the surface.
Hardened carbon steel abraded with a Shapton Pro 320 grit stone.
Cross-section view of the Shapton Pro 320 surface. Damage extends 1-1.5 microns below the surface.

Observe that the Shapton Pro 320 grit surface is much smoother than that from the 320 grit sandpaper; however, the depth of damage is similar for both. This can be attributed to the fact that the stone is produces more burnishing than the sandpaper, resulting in a smoother surface.

Hardened carbon steel abraded with a Wicked Edge 600 grit diamond stone.
A low magnification image of a fresh/undisturbed scratch from the Wicked Edge 600 diamond stone.
Cross-section view of the Wicked Edge 600 grit surface. Damage extends 4 to 6 microns below the surface.
Hardened carbon steel surface abraded with 600 grit silicon carbide paper.
Cross section view of 600 grit surface. Damage extends 1.5 to 2 microns from the surface.
Hardened carbon steel surface abraded with 2000 grit silicon carbide paper.
Cross section view of 2000 grit surface. Damage extends 0.7 to 1 microns below the surface.
Hardened carbon steel abraded with a Shapton Glass 2k stone.
Cross section view of the Shapton Glass 2k abraded surface. Damage extend 0.5-1 microns below the surface.

Comparison of the Shapton Glass 2k stone with the 2k sandpaper shows similar depth of damage, but a much smoother surface from the stone, again presumably due to more burnishing from the stone.

Hardened carbon steel surface after abrasion with a Shapton Glass 8k grit stone.
Cross-section view of the Shapton Glass 8k surface. Damage extends approximately 0.3 to 0.5 microns below the surface.
Hardened carbon steel surface abraded with 20,000 grit Suehiro Gokumyo.
Cross section view of 20,000 grit surface. Damage extends approximately 0.3 microns below the surface.
Hardened carbon steel surface polished with 0.3 micron aluminum oxide lapping film.
Cross section view of the polished surface. Damage extends less than 0.1 micron from the surface.

The goal of metallographic polishing is not simply to produces a flat surface with a mirror finish, but to expose the bulk crystal structure by removing the damaged layer that is unavoidably formed by grinding. As the images above have shown, this is achieved by progressing from coarser to finer grit abrasive.

With the Shapton Glass 2k (nominally 7.35 micron), in Abrasion Rate vs Grit, I measured an average removal rate of only 0.02 microns per pass across the stone. At the same time (and under the same conditions), we observe individual scratches that are up to 0.5 microns deep. These numbers suggest that most of the grit-steel interactions are simply rubbing and ploughing (and burnishing) with minimal metal removed. The surface finish, or “scratch pattern” is determined more by the character of burnishing that occurs than by the dimensions of fresh scratches.

Generally, we have observed that damage typically extends well below the depth of the surface roughness. We must expect that when sharpening with all but the finest polishing grit, the apex will therefore consist of damaged steel. The obvious question is what are the properties of this modified steel? In particular, is it less brittle and more likely to deform than chip?

The size of the grit particle is only one characteristic of sharpening media, the shape and orientation of the grit will determine whether an efficient cutting edge is applied to the steel surface. The more firmly attached the grit particle, the greater the shear force that can be applied and the deeper the cutting depth that can be produced. In Part 2 of this article, I will discuss mud/slurry/loose grit; when grit particles are not attached to the sharpening stone, they can roll between the blade and stone surfaces, in a “3-body” interaction.