Diamond wire is moving from high-carbon steel cores to tungsten, because steel tops out around 35 µm while tungsten still holds at 30 µm and below. The published numbers for the switch are large: about 40 % more breaking load, 60 % fewer wire breaks, 15 % less silicon lost to kerf. What those numbers hide is arithmetic — a thinner wire carries fewer grit per metre, so every particle left on it has to do more work. The specification that decides whether it can is not the average size.
Why The Core Had To Change
A diamond wire is a tension member carrying abrasive. The core holds the load and the grit does the cutting, and the two constraints pull against each other: a thinner wire wastes less silicon, but a thinner core snaps at lower tension.
High-carbon steel runs out of strength before it runs out of demand. Industrial practice puts its floor near 35 µm of core diameter — below that, strength falls away faster than the kerf saving is worth, and the break rate eats the yield gain. Tungsten changes the trade: its melting point sits above 3,400 °C, its tensile strength in fine drawn form runs two to three times that of steel, and it stays stable below 30 µm.
The figures published with the substitution are consistent across suppliers: around 40 % higher breaking load at the same finished diameter, roughly 60 % fewer breaks in the cut, and about 15 % less silicon lost as kerf. At wafer-plant scale that converts to roughly a tenth off the silicon cost per gigawatt, and 5 % to 8 % more wafers from the same ingot volume.
The costs are real and worth naming, because they land on the same production line that applies the grit. Tungsten is denser, so the same length of wire weighs more and the tension control has to be more precise. It is also harder, which means the pre-plating preparation developed for steel does not transfer — activation, adhesion and grit retention all have to be re-qualified.
Fewer Grit Per Metre, More Load Per Grit
Take a wire from 40 µm down to 28 µm and the circumference falls by roughly 30 %. Grit sits on the surface, so particles per metre fall with it at the same grit size and plating thickness. The kerf narrows, which is the point — but the same cutting work now runs through fewer particles, so the average load per particle rises.
That changes what a defect costs. On a 40 µm wire a rounded or slightly dull crystal cuts a little less efficiently and the wire keeps going. On a 28 µm wire the same particle carries a load it cannot take, and it either fractures and leaves a socket, or skates and polishes instead of cutting. The substitution is not only a wire-metallurgy project; it moves the acceptable defect rate of the powder.
Three Specifications That Tighten
| Property | What worked on a 40 µm steel core | What a 28 µm core needs | Why it changed |
|---|---|---|---|
| Coarse tail | D50 quoted, upper tail tolerated | A single maximum particle size plus a coarse-particle count per 10 g | One oversize particle is a local high spot on a thin wire — it is a break site, not just a scratch |
| Distribution width | A wider band still cut, because surplus grit carried the load | Narrower, and tightened at the top end first | With fewer particles per metre there is no surplus. A wide band means a few large crystals do the work and the rest idle |
| Particle shape | Blocky or irregular both acceptable | Blocky preferred; flakes and intergrowths screened out | Flakes fracture under tension and leave a socket; intergrowths are a cluster pretending to be one particle |
The direction of the change is worth stating plainly: on a thin wire, average size is the least informative number in the certificate. Two powders with identical D50 can behave completely differently if one has a clean cut-off at the top and the other trails a coarse tail.
What Changes On The Plating Line
Plated diamond wire is a composite coating: nickel holds the grit, and the grit has to be held hard enough to survive the cut without being buried so deep that it never exposes an edge. Three powder properties that were secondary on steel become deciding on tungsten.
- Surface cleanliness. Residual graphite or catalyst from synthesis prevents nickel from nucleating on the crystal. The result is a patch where the grit sits loose or washes out — a bald spot on the wire. On a thin wire a bald spot is a wear accelerator, because the nickel around it takes the whole load.
- Magnetic content. Magnetic particles collect in the plating bath and co-deposit as nodules. A nodule is a diameter change, and a diameter change is where the wire breaks. The specification is a ceiling with a stated test method, not a passing mention.
- Dispersion. An agglomerate is plated as one body. It sits too high, takes load on a single point, and is the first thing to come off. Dry PSD will not reveal it, because the instrument measures the agglomerate as a particle.
None of this replaces the plating work itself. Tungsten needs its own pre-treatment sequence, and the current density that gives good coverage on a steel core will not necessarily do the same on tungsten. But powder sets the ceiling: a perfect plating process cannot compensate for grit that is dirty, magnetic or clumped.
Where It Goes Wrong
| Symptom | Cause | What to change |
|---|---|---|
| The wire breaks at the same position repeatedly | Nodule from magnetic contamination, or a trapped oversize particle at a guide roller | Check magnetic content and coarse-particle count before touching tension settings |
| Kerf width wanders and TTV drifts | Upper tail of the distribution is too wide, so a few large crystals set the cut depth | Tighten the top end: maximum particle size and D99, not D50 |
| Bald patches appear along the wire | Surface residue from synthesis, or pre-cleaning that was adequate for steel but not for tungsten | Add a surface-cleanliness criterion and re-qualify the pre-plate sequence |
| Edge chipping rises while the wire stays intact | Grit is too tough or too rounded to self-sharpen, so it polishes the kerf wall | Move to a more friable grade, or reduce coating weight so the particles sit prouder |
Supply Reality
The substitution will not happen overnight, and the reason is upstream of the wire plant. A fine tungsten core passes through a long chain — APT to tungsten oxide to powder to bar to swaging to drawing — and the number of plants that can hold 25 to 35 µm reliably is small. Capacity ramp sets the pace of adoption.
That has a practical implication for anyone buying grit. Tightening the distribution and cleaning up the coarse tail delivers its benefit on the steel wire you are running today, and the same benefit carries over when the tungsten core arrives. The two improvements are independent and additive, which means there is no reason to wait for the wire.
Next Step
The grade choice for wire is set out by bond type on the diamond wire sawing page — resin-bonded wire and plated wire want different crystal habits and different size bands. If the question is why a crystal behaves the way it does under load, grit toughness and crystal shape covers the measurement, and ash and magnetic content covers the two impurities that matter most in a plating bath.
For the powder itself, HDG micron powder is the general-purpose grade for plated wire, and HDY is the tighter-distribution option where the top end has to be controlled.
Tell us the finished wire diameter and the plating type and we will quote a size band, a maximum particle size and a coarse-particle limit against it.