In our shop, a stainless-steel CNC machining parts factory in Ohio that does a lot of built-to-print OEM work, part marking used to be the easy part of the day. Then, in March 2023, a customer in Wisconsin wiped a serial number off a 316L stainless part with her thumb. The marking had been done by a CO2 laser a week earlier at another supplier. It came off like graphite from a broken pencil.
She asked me a question I had been asking myself for months: “Can you engrave metal with a CO2 laser?”
The short answer is no—at least not bare metal, not in any durable sense. I told her that, and then I admitted that my own shop had made the same mistake earlier that year. This is that story, plus the cost math I now use before approving any marking process.
The Short, Correct Answer
A CO2 laser emits at roughly 10.6 micrometers. Most bare metals are highly reflective at that wavelength. Stainless steel, aluminum, copper, and titanium reflect most of the energy away. If you turn the power up high enough, the small amount of absorbed energy heats the surface and creates a thin oxide layer. That layer may look dark for a few days, but it is not a real engraving. It can be rubbed off, washed off, or worn away during normal handling.
What CO2 lasers can do is mark materials that are coated, painted, anodized, or non-metallic. Wood, acrylic, leather, cardboard, plastics, and coated metals are all fair game. The laser burns the coating to expose the base material underneath. But raw stainless steel and other bare metals will not produce a permanent mark with a CO2 laser.
If the drawing says “engrave” rather than “mark,” the customer usually wants depth. Even a fiber laser has to make multiple passes to achieve a deep cut. For truly durable logos, serial numbers, and part numbers on stainless steel, the better options are:
- A fiber laser at about 1,064 nm wavelength
- CNC engraving with a carbide tool
- Electrochemical etching for thin, controlled marks
For military and aerospace work, part-marking requirements are often referenced from standards such as MIL-STD-130 or customer-specific durability tests. Medical device manufacturers may also point to UDI requirements under 21 CFR 830. If any of those apply, a CO2 laser is not going to survive the qualification process.
How We Made the Mistake Anyway
Seven weeks before that video call, a food-processing equipment OEM sent us an RFQ for a long, thin valve component made from 316L stainless. Quantity was 900 pieces. The part itself was about 6 mm in diameter and 54 mm long, with a small machined flat for a logo and a date code. The machining was not exotic for a turning center with live tooling, but the length-to-diameter ratio made it fussy.
Our process engineer quoted mechanical engraving at $2.45 per part. That included the tool cost, cycle time, and inspection time. A local laser shop quoted the marking portion at $0.85 per part. The laser shop told us they had an industrial 80 W CO2 machine and that they engraved “metal products” all the time. In fairness, they did beautiful work on anodized aluminum and stainless tumblers that were coated. They had almost no experience with raw 316L.
I approved the lower quote. The gap was $1,440 on that order, and I told myself we were being smart with the customer’s money. That decision cost us more than $4,000.
The first article looked acceptable. The engraved logo was dark enough to read. Our inspector wiped it with a dry cloth and nothing came off. But when we cleaned a production part with isopropyl alcohol the next day, the mark faded badly. It was only a superficial oxide layer, exactly what a CO2 laser leaves on bare stainless steel.
The Real Total Cost
Here is where the sticker price thinking starts to hurt. The $0.85 per part quote looked great on a spreadsheet. The actual cost of that decision looked like this:
- $765 paid to the CO2 laser shop
- 860 rejected parts that had to be reworked or scrapped
- $2,380 for emergency fiber-laser marking at a qualified shop
- Two partial air shipments to keep the OEM line running
- A permanent credibility stain with a customer we had wanted for years
The original $1,440 savings produced about $4,700 in additional costs. Every purchasing manager who compares quotes based on unit price alone should pause here. Lowest unit price is only the starting point. The total cost includes the risk of rejection, the cost of re-inspection, the cost of expedited freight, and the cost of a damaged relationship.
I now calculate total cost before comparing vendors. The formula I use in our shop is simple: total cost equals quoted price plus expected scrap rate multiplied by full process cost, plus the cost of rework, plus the cost of schedule risk. If a vendor cannot answer basic questions about laser wavelength and material compatibility, the expected scrap rate is higher than their quote suggests.
The Swiss Lathe Lesson
The same failed valve stem taught me a second lesson. The part’s slenderness made it difficult to machine on a conventional lathe without deflection. A Swiss-type lathe is built for that geometry because the guide bushing supports the bar stock right at the cutting area. I had always dismissed Swiss machines as specialized and expensive, even though many DMG MORI users run them every day for stainless medical and automotive components.
After the marking mess, we worked with an application engineer from DMG MORI Manufacturing USA in Davis, California. They showed us how a DMG MORI Swiss lathe could machine the entire part from bar stock, cut the flat, and handle secondary operations in one setup. The comparison was not just about cycle time. The real difference was stability. The guide bushing eliminated deflection, and the machine held the profile consistently across all 900 parts.
We did not buy the machine immediately. That would be a dramatic ending but not an honest one. Instead, we learned to match the process to the part geometry and to include setup risk in our quoting. For long, slender stainless parts, we now quote on Swiss-type capability, either in-house or through a trusted partner with the right machine.
A few months later, we ran a second batch of valve stems at a partner shop with a DMG MORI Swiss lathe. The parts came off with zero rejects and the marking was done mechanically while the part was still supported. It was the kind of process that looks boring in a video and saves everyone a month of headaches in real life.
What I Tell Customers Now
If you ask me today whether you can engrave metal with a CO2 laser, I will give you the same answer I gave that customer in Wisconsin: not bare metal, not durably, not for production.
If your part is stainless steel and the print calls for an engraved logo or serial number, use a fiber laser or CNC engraving. If the part is long and thin and made from 316L, ask how the supplier plans to control deflection before you worry about the marking at all. And no matter which process you choose, estimate the cost of failure before you celebrate a low quote.
Since we adopted this checklist, we have caught at least 24 potential process mistakes in the quoting stage. The most satisfying part is not the money we saved. It is the fact that I no longer have to explain to a customer why a permanent mark is rubbing off under their thumb.
