I am interested in mass producing a reusable coffee cup (à la KeepCup) with a scannable id on it.

Cup will be plastic or glass, ideally glass.

ID code will need to survive being washed regularly.

I was thinking it could be put on by whatever method is used to put beer names on pint glasses, but I really have no idea how any of this is done. Onte thing I was wondering about was if I need to print in two colours (black and white for example) or if frosting some lines on it would work.

The reason I am interested in any of this is I want to assess the feasibility of giving out takeaway coffee cups which can then be returned to a 'cup bin'. The barcodes are needed to identify which cups have been returned.

And the codes have to be unique for each glass.

  • $\begingroup$ Some etching process? $\endgroup$
    – Solar Mike
    Commented Apr 30, 2018 at 14:21
  • $\begingroup$ What have you thought about using and why do you think that it would or would not work? We would like to see what effort you have put in already so that any answers are not a rehash of what you already know. $\endgroup$
    – hazzey
    Commented Apr 30, 2018 at 18:39

1 Answer 1


A few different marking processes are available to accomplish your objective. Glass is a good choice, for reasons of durability.


One creates a mask or stencil with the desired graphic. In the case of a complex design, vinyl cutters perform the task quite well. The stencil is applied using transfer tape, after which etching chemicals are applied. After a specific period of time for the chemical, it is washed off, leaving behind a permanent image.

Will not work on plastic.

Con: requires creation of vinyl stencil for each unique image, labor intensive to weed image, apply chemicals, clean glass

Engraving (mechanical)

This particular method requires a computer controlled engraving device, which spins a diamond bit at high speeds. The computer directs the cutting point and removes glass from the surface. Different cutting bits provide some flexibility in design.

Will work on plastic, results are varied depending on plastic composition.

Con: I have no direct experience or observation of hobbyist level rotary devices for CNC routers capable of this work. Commercial versions with rotary capability are huge and expensive.

Engraving (laser) *note below

This method also uses a computer controlled device, typically a CO2 laser equipped with a rotary accessory. The glass is placed in the chuck or on rollers and rotates under the laser nozzle as it moves laterally. Microscopic pieces of glass are removed by the laser, similar to mechanical engraving. Power settings determine depth of engraving.

Will work with acrylic plastic, others may be possible. Vinyl is prohibited due to creation of chlorine gas.

Con: can be expensive (no budget referenced) and sometimes frustrating.

Marking (laser applied coating)

Substance exists (Laser Bond Tech) which when applied to glass (or other surfaces) and activated (engraving settings) with the laser, bonds to the glass. The substance creates dark, almost black marking on the surface, resistant to wear, very durable. Images created in the same manner as engraving.

Con: Additional steps involved in applying, drying and cleaning after marking.

Marking (silk screen)

Silk screen marking deposits a layer of ink on the surface of the glass (or plastic). In the case of a curved surface, special screens and squeegees are necessary to provide proper coverage. The screen is typically created using a photographic process allowing for quite good detail results. Ink is selected for durability for obvious reasons.

Will work on plastic.

Con: requires creation of silk screen for each image, also labor intensive process.

Marking (transfer ink)

This process uses a flexible rubber surface and a raised surface containing the image to be created (think rubber stamp but in metal or more durable material). The image has ink applied by a roller, which is then transferred to the rubber surface. It helps to consider a sports ball with a smooth surface contacting the inked surface. The ball then moves over the surface to be inked and is pressed into contact, transferring the image. Durability of the image depends on the ink selected.

Will work on plastic.

Con: very labor intensive and/or expensive to create unique image block, better suited to repeat marking of same image.

  • note here: These suggestions are provided as a reference for further research and are not to be considered exhaustive resource guides. Much of this material is presented from a hobbyist perspective.

As I own a laser cutter/engraver, I see this method as the most easily accomplished, although initial expense can be high. I also believe that the labor involved in quantity production of unique bar codes is lower with a laser engraver than any of the other methods.

The image generation for bar codes is about the same, but for all of the previously noted methods, one more step is involved to create a physical device to create the image on the glass. With a laser engraver, the image of the bar code is transferred to the software and the engraving begins, removing that intermediate step and attendant costs of time and money. The mechanical engraving (CNC router) is the closest to the laser process in that the image also is directly transferred to the machine without an intermediate device.

"Inexpensive" Chinese-made CO2 lasers abound on eBay, although some sources have a high failure rate, my purchase included!

Edit: testing completed, results below.

Not surprisingly, the bare naked engraved UPC-A bar code is unreadable. I tried many combinations of illumination, but none of them resulted in a read by my iPhone scanner app.

glass engraved upc code

This image is of the engraving directly on the glass. The irregular marking below is the adhesive remnants of the label.

The aforementioned Laser Bond Tech 100 purports to bond to glass, and my tests show that this is true. Excessive power levels destroy the material and the markings wash off with water. Not-excessive power levels do a good job of bonding the material. It is resistant to abrasion, although I did not take an abrasive pad to my test item.

The bar code generated on the glass was readable IF there was a white background. The scanner snapped up the code in less than one second during the positioning!

marked bar code with paper background

This image is approximately 25 mm x 50 mm.

Considering that the OPs reference is aimed at use by the general public, I envision someone scratching away at the marking material. I think with a metal device such as a knife or sharp edged spoon, one could damage the marking.

Moving right along, I figured that an inverse pattern on glass, engraved rather than marked, might work.

dark engraved bar code

The above image was taken with ambient light and the iPhone scanner snapped the image in about a second. I was surprised, considering the darkness of the area.

light engraved bar code

This photo was taken with the iPhone with the built-in light turned on. The scanner app also caught the code quickly with its light turned on.

I used an online free UPC generator, which created a mess of the SVG image I needed to perform this test. It was easily corrected, but time consuming.

For the OP, ten thousand unique codes would be better managed by someone fluent in SVG creation, with the ability to spit out those files (after testing) rather than use an online resource of questionable quality.

Generally, I'd say that the marking spray has promise, but an inverted engraving gives good durability with the least amount of labor.

In the case presented by the OP, however, I would suggest to contract to an outside business specializing in glass marking technology. Alternatively, he will spend many hours producing these codes on glasses, a few at a time.

I could produce perhaps fifty per day and likely lose my sanity after the first week.

  • $\begingroup$ Yeah laser engraver works quite well but have you tested scanners on the resulting glass object. $\endgroup$
    – joojaa
    Commented May 2, 2018 at 5:16
  • $\begingroup$ I have considered that may be a consideration. I am in a position to test this on a scrap glass surface with a scanner and will likely add the result to the answer in the next day or so. Also added laser related black marking method to answer. $\endgroup$
    – fred_dot_u
    Commented May 2, 2018 at 9:34
  • $\begingroup$ Yeah well i can do the laser part i just odnt have a scanner handily available. Anyway if OP would go this route sourely they would test it with their particular brand of scanner. $\endgroup$
    – joojaa
    Commented May 2, 2018 at 10:29
  • $\begingroup$ Very nice info on the barcodes. I wouldnt do the barcodes in svg but rather postscript since you csn find more mature routines in PS. Also making your own takes about an hour (i did it one in a lets implement something based on wikipedia session) $\endgroup$
    – joojaa
    Commented May 3, 2018 at 4:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.