DLC Coating on Injection Molds: A Tool-Life Extension Case Study
Glass-filled nylon eats mould steel. A PA66 with 30% glass fibre flowing through a gate at 60 m/s drags abrasive fibres across the same cavity surface on every shot, and no amount of hardened tool steel stops it indefinitely — the tool is harder than the matrix but softer than the fibre. This case study covers what actually works: switching the wear mechanism from abrasive wear against steel to sliding wear against a diamond-like carbon film. It covers variant selection, why substrate polish matters more than coating thickness, the deposition temperature ceiling set by your steel’s temper, and the applications where DLC is the wrong call regardless of how good the film is.
The Wear Mechanism: Glass Fibre Against Tool Steel
E-glass fibre sits around HV 1000–1100 on the Vickers scale. A mould cavity hardened to 52 HRC is roughly HV 540. Any time the interface involves sliding contact, the harder phase wins — which is why gate inserts, core pins and the radii at the ends of flow paths lose dimension first. The visible symptom is usually not the wear itself but what follows: flash along the parting line, short shots because the gate geometry opened up, and sticking ejection as the surface roughness climbs and the part grips the cavity.
Three Things Usually Happen Together
- Abrasive wear removes material preferentially at flow direction changes, so gate land and corner radii drift out of tolerance first.
- Built-up edge forms when decomposed resin residue welds onto roughened steel, then tears away and leaves fresh aggressive peaks behind.
- Ejection force climbs as Ra increases, which is often the first measurable signal that the tool is degrading — worth trending rather than reacting to.
Rehardening the insert treats the symptom. The lever that changes the mechanism is putting a surface in contact with the resin that is both harder than the fibre and slippery enough that fibre does not bite. That is the argument for Diamond-Like Carbon PVD Advantages as a class of films.
Why DLC Rather Than TiN or CrN
TiN and CrN are competent general-purpose PVD coatings and they cost less. They lose on two specific counts here: coefficient of friction and deposition temperature. TiN runs roughly 0.4–0.6 against steel in dry sliding; hydrogenated DLC runs 0.05–0.15. That low friction is what lets the melt release cleanly without ejection assist and without release agent spray, and it is also what stops fibre from gripping the surface. The mechanics behind film formation are described in our How Does PVD Coating Work? overview.
| Coating | Hardness (HV) | Friction vs Steel | Deposition Temp | Best Fit |
|---|---|---|---|---|
| TiN | 2300–2500 | 0.40–0.60 | 400–500°C | Cutting tools, general wear |
| CrN | 1800–2200 | 0.35–0.50 | 400–500°C | Corrosion plus wear, resins with acid off-gassing |
| TiAlN | 2800–3300 | 0.50–0.70 | 450–550°C | High-speed dry cutting |
| a-C:H (hydrogenated DLC) | 1500–2500 | 0.05–0.15 | 150–250°C | Mould release, sliding wear |
| ta-C (tetrahedral) | 4000–7000 | 0.05–0.10 | <150°C | Severe abrasion, highest hardness |
| Cr-DLC / W-DLC multilayer | 1200–2200 | 0.08–0.20 | 150–250°C | Load carrying plus low friction |
The trade-off to note: hardness and friction pull in different directions. ta-C is the hardest and slipperiest but the most internally stressed, which caps practical film thickness. Where load carrying matters alongside low friction, a metal-doped multilayer normally outperforms a single-layer film of either extreme.
Choosing the DLC Variant
Four variants cover most mould work. Picking the wrong one is the usual reason a coating trial disappoints, and the failure mode almost always looks like adhesion loss rather than wear.
| Variant | Typical Thickness | Max Service Temp | Adhesion Behaviour | Use When |
|---|---|---|---|---|
| a-C:H single layer | 1.0–2.5 µm | 300°C | Needs Cr or Si interlayer | Standard injection tools, cost-sensitive jobs |
| ta-C | 0.5–1.5 µm | 400°C | High stress; arc-deposited with filtered source | Thin margins, very high abrasion |
| Cr-DLC graded | 1.5–3.0 µm | 350°C | Graded interface, best adhesion | Loaded sliding contacts, larger components |
| W-DLC multilayer | 1.5–3.0 µm | 400°C | Excellent load support | High contact pressure applications |
Thickness deserves a word. Thicker is not better in a mould cavity: above roughly 3 µm the internal stress starts driving edge chipping at sharp corners and radius break lines, and you also lose dimensional control on tight fits. Specify thickness by function, then verify it will survive the geometry rather than pushing it up for wear allowance.
Substrate and Pretreatment
Polish Before Coating, Not After
DLC is conformal. Every scratch, polish line and micro-pit in the substrate gets reproduced on the coating surface, and a 2 µm film amplifies rather than hides a rough base. Target Ra 0.05 µm or better going into the coating chamber on any cosmetic surface. Once coated, polishing is off the table — anything aggressive enough to true up the film will break through it. This is the single most common cause of a “DLC did not work” result we see, and it is a preparation failure rather than a coating failure.
Deposition Temperature Versus Your Steel's Temper
The binding constraint on whether an existing tool can take DLC is the last tempering temperature of the steel. Deposit above it and you anneal the substrate under the film. The coating survives; the load-bearing steel underneath goes soft and the film collapses into it under load.
| Mould Steel | Working Hardness | Typical Temper | Low-Temp DLC Window | Note |
|---|---|---|---|---|
| H13 / 1.2344 | 48–52 HRC | 540–600°C | Room to 250°C | Wide margin, easy candidate |
| 718H / 1.2738 | 34–38 HRC | 550–600°C | Room to 250°C | Pre-hardened, generous margin |
| S136 / 1.2083 | 48–52 HRC | 200–300°C | Below 180°C | Needs genuinely low-temperature process |
| NAK80 | 38–42 HRC | Age hardened | Below 150°C | Check age-hardening curve first |
| Aluminium QC-10 | Not applicable | Age hardened | Below 150°C | Prototype tools; verify ageing response |
S136 is where most of the surprises happen. It tempers low enough that a standard 200°C PACVD cycle can sit right at the edge of its window, which is exactly why transparent-optics moulds need a genuinely low-temperature route rather than the default recipe. Our Advanced PVD Technology in Surface Finishing page covers process variants and their temperature envelopes.
Process Parameters for the Run
| Step | Parameter | Target | Why It Matters |
|---|---|---|---|
| Cleaning | Ultrasonic plus plasma etch | No residual oil | Bare metal contact for adhesion |
| Interlayer | Cr or Si, thickness | 0.1–0.3 µm | Bridges thermal expansion mismatch |
| Deposition | PACVD bias voltage | 400–800 V | Controls sp3 fraction and therefore hardness |
| Deposition | Chamber pressure | 1–10 Pa | Affects mean free path and film density |
| Deposition | Substrate temperature | Below steel temper | Protects substrate hardness |
| Cooling | Vacuum cool-down | Below 80°C before vent | Avoids oxidation of a fresh hot film |
Verification
Three measurements tell you whether the run succeeded, and none of them is appearance:
- Film thickness by ball crater or calotest — target ±10% of spec, and check several locations since cavity geometry causes uneven deposition.
- Adhesion by Rockwell indentation compared against the VDI 3198 scale (HF1–HF4 acceptable, HF5–HF6 a reject) or scratch testing with a critical load above roughly 30 N.
- Friction coefficient by pin-on-disc against the mating counterbody — this is what actually predicts release behaviour in the tool, and PVD Coating Wear Resistance details how it is measured.
Ask for the values on a coupon processed in the same batch. Coating certificates quoting nominal spec rather than batch results are not evidence of anything.
Where DLC Is the Wrong Answer
The cases below fail regularly enough to be worth stating plainly, because coating them wastes money and the failure shows up only after the tool is back in the press:
- Class-A mirror finishes. The film replicates every substrate defect, so a coating will generally make optical surfaces worse, not better.
- High-temperature resins above 350°C such as PEEK, PEI and LCP at their upper processing range. DLC oxidises in air well before those melt temperatures.
- PVC and fluoropolymers release species that attack the film chemically rather than mechanically. Corrosion-resistant CrN-based systems handle these better.
- Filled compounds with very coarse fibre. At high aspect ratios the contact pressure exceeds what a 2 µm film can support and the film spalls locally.
- Tools that will see re-machining or welding. Any repair removes the coating, and re-stripping is not always economical on small inserts.
Specifying It Properly
Put three things on the drawing or the coating PO, and the odds of getting what you wanted go up considerably: variant rather than the generic word “DLC”; target thickness with a tolerance; and the substrate’s last tempering temperature so the coater can pick a process that fits. Naming the steel spec but not the temper is the most expensive omission — it is the constraint that decides whether the tool survives the coating cycle at all.
Wear-resistant films reach well past injection tooling. Where a component ends up implanted or used surgically, the Medical-Grade PVD Coating Services route follows a very different validation path, and the broader PVD Coating Applications list covers medical, automotive and decorative work using the same coating families. Substrate limits, batch sizes and lead times sit under Physical Vapor Deposition Coating Services.
Frequently Asked Questions
How much longer will a DLC-coated mould last?
It depends on where you are starting. Benefiting most are tools running abrasive compounds where the failure mode is wear rather than corrosion or cracking. A clean result is measurable ejection force and gate-land inspection over the first few thousand shots — trend those two and you will know the answer for your specific tool within a month of running. Treat any headline multiple quoted without the resin grade and fibre loading attached as marketing.
Can an existing mould be coated, or does it need to be new?
Existing tools are usually fine, provided two conditions hold: the substrate has not been nitrided to the point where the case conflicts with the coating, and last tempering temperature clears the deposition temperature with margin. Fully strip, inspect and re-polish before coating rather than coating over a worn surface.
Does DLC change part dimensions?
Yes, by twice the film thickness on internal dimensions. On a 2 µm coating, a bore closes by roughly 4 µm and a shaft grows to match. On H7 tolerances that is visible, so either coat both mating parts or leave grind stock for the coating thickness in the tool design.
Can DLC be stripped and reapplied?
It can be removed chemically or by plasma etching, though the aggressiveness required limits how many times a given insert can go through the cycle before it needs re-grinding. For tools in a planned maintenance rotation, budget for one strip-and-recoat cycle and then re-grind.
Getting One Evaluated
Send the resin specification including filler type and loading, the steel grade with last tempering temperature, and which surfaces are cosmetic. Those three items determine both feasibility and which variant gets quoted — everything else is detail that can wait for the technical discussion. Reference data for alloy and hardness comparisons is available from ASM International.
