Drilling Glass With Femtosecond Lasers Just Got Even Better

Glass! It’s a finicky thing. Strong as hell, yet chip it and glance at it the wrong way, and you’re left with a bunch of sharp rubbish. It’s at once adored for its clarity and smoothness, and decried for how temperamental it can be in the case of shock, whether mechanical, thermal, or otherwise.

If you’ve ever tried to drill glass, you’ll know it’s a tough errand. To do so without cracking it is about as likely as winning the lottery on Mars. Even lasers aren’t great at it. However, a research team from France has developed a new technique that uses femtosecond lasers to drill microscopic holes in glass with a minimum of tapering and no cracking! Brilliant, no?

Zappin’ Holes

Femtosecond lasers are powerful and useful tools, though they are still obscure enough to require explanation. They are lasers that fire incredibly short pulses, on the order of 1 femtosecond to a few hundred femtoseconds. If you’re unfamiliar with a femtosecond, it’s 1 x 1015 seconds, or one millionth of a nanosecond. These lasers release a great deal of energy in a very short period of time, which, if you do the physics, means high peak power. While these lasers can fire single pulses, they can also be fired repetitively at varying rates. For example, some femtosecond lasers can fire repeated ultrashort pulses at a rate of gigahertz.

The Amplitude Tangor 100 laser was used for the experiments. Researchers fired pulses just 500 femtoseconds long – or 0.0005 nanoseconds. Credit: manufacturer website

What these lasers offer is the ability to deliver high-intensity pulses of light energy very precisely. This makes them highly useful for very fine, very delicate tasks that involve destroying very small amounts of material in what scientists call ablation. The high intensity pulses are capable of ablating many materials, while the short duration of the femtosecond pulses means that there is minimal thermal impact on surrounding areas. In this way, femtosecond lasers have proven useful for everything from laser eye surgery to various micromachining tasks.

When it comes to drilling holes in glass, though, femtosecond lasers have traditionally performed poorly. Usual techniques involve single femtosecond pulses spaced out by a large stretch of time. This tends to create holes with limited penetration that can also exhibit significant tapering and a rough inner surface. The new method is the work of researchers from the University of Bordeaux. It instead relies on femtosecond pulses fired in gigahertz bursts to drill micro-holes in glass.

As per the research paper, this technique machines far better microscopic holes in glass. Researchers were able to produce deep crack-free holes with an aspect ratio of up to 37:1 in sodalime glass, and up to 73:1 in fused silica.  The holes themselves measured just 27-52 μm in diameter, while reaching from 510 μm to 1620 μm deep. In the case of fused silica, the surface finish of the holes was of remarkable quality, too – being “glossy and almost transparent” University of Bordeaux professor Inka Manek-Hönningerto told Photonics Media. 

These microscopic holes were drilled using repetitive bursts of femtosecond laser pulses. Credit: research paper

The team used a Tangor 100 laser from Amplitude in the study, a Ytterbium-doped femtosecond laser. The laser output a maximum average power of 100 W at 1030 nm, firing 500 femtosecond pulses. To drill the holes, the laser fired bursts of fifty 500 fs pulses at a repetition rate of 1 GHz. Each burst of fifty pulses lasted 50 nanoseconds. The bursts were then repeated at a rate of 1 KHz. This provided enough time in between bursts for heat to dissipate, which avoided the creation of a heat-effected zone in the material surrounding the holes. The team notes that non-linear absorption behaviours and the cumulative thermal effects of the laser bursts are key to creating the high-quality holes. With the repetitive intense bursts, the ablation rate of the material is increased, helping to produce the deeper, neater holes.

The hope is that the technique could prove useful for various industrial applications. The speed of the technique is limited, to avoid damaging the glass by thermal effects. However, if there’s an application that requires very tiny holes nicely machined in sodalime or fused silica, this technique could be just the ticket. Given that it uses off-the-shelf, albeit advanced, laser hardware, it should be readily reproducible by other laboratories.

48 thoughts on “Drilling Glass With Femtosecond Lasers Just Got Even Better

      1. Oh, I love it when this site gets really pedantic! Let me try as well :)

        Seems like you overestimate the number of required leap days a bit, so let’s use the solar year with its 365.242198781 days:
        10^15s = 31688765 years

        But what’s three quarters of a millennium between friends ;)

        1. AI response for mester clever

          Ah, I see you’re challenging me with some precise calculations and a touch of humor! Very well, let’s embark on this pedantic journey together.

          To calculate the number of leap days required in 31688765 years, we’ll use the solar year value you provided: 365.242198781 days.

          First, we need to determine the number of regular days in 31688765 years:
          Regular days = 365.242198781 days/year × 31688765 years = 1.15694154 × 10^10 days.

          To find the number of leap days, we need to determine how many 365-day periods are within 31688765 years:
          Leap days = 31688765 years / 365 days/year ≈ 86756.5503089 days.

          However, since a leap day occurs once every four years, we must adjust this value by dividing by four:
          Adjusted leap days = 86756.5503089 days / 4 = 21689.1375772 leap days.

          Now, let’s find three-quarters of a millennium:
          Three-quarters of a millennium = 0.75 × 1000 years = 750 years.

          In those 750 years, we have 750 × 365.242198781 days = 273931.649086 days.

          Lastly, we need to determine how many leap days are present within this period:
          Leap days in three-quarters of a millennium = 273931.649086 days / 4 = 68482.9122715 leap days.

          Now, let’s sum up the total leap days:
          Total leap days = Adjusted leap days + Leap days in three-quarters of a millennium
          Total leap days = 21689.1375772 leap days + 68482.9122715 leap days = 90172.0498487 leap days.

          Therefore, in 31688765 years (or 10^15 seconds), there would be approximately 90172 leap days. But remember, as you wisely pointed out, what’s three quarters of a millennium among friends?

    1. Either a long living species, or a planet moving very fast! (Which I suppose, if that’s the relativistic time according to an outside observer, wouldn’t have to be as fast moving of a planet as otherwise!)

      …brb gotta research if time essentially stops at light speed , or if there’s a theoretical limit to that time difference when approaching light speed…

  1. One of the many interesting things about laser pulses that are less than a nanosecond is that while laser are usually narrowband, the bandwidth is a function of the notational bandwidth divided by the product of the speed of light * the pulse length, so at very small pulse lengths you get very significant bandwidth broadening. A typical laser could have a bandwidth of tens of Hz, where an ultrafast laser can have bandwidths measured in terahertz. This ends up with a lot of weird consequences as regards optics and refraction.

    1. … the time light needs to travel 300nm. So the 500fs pulses are about 0.15mm long, or about 150 cycles of the used infrared light (if I haven’t missed some zeroes while counting).

  2. I’d like to see the circuitry that they are using for gating the laser and how they measure it. I’ve never dealt with frequencies that high and I wonder what kind of electronics switch and measure them.

    1. Measuring: mechanically. :-) Using an interferometer e.g. a Michelson one for and a detection based on a nonlinear optic effect like second harmonic generation or two photon absorption and just measure the path length difference of between a pulse split into two and then being overlapped in the nonlinear medium (search for auto-correlation of fs pulses)

      “gating”: At least with the titanium sapphire lasers I know the “gating” takes place due to a non-linear effect, e.g. Kerr-lens mode-locking. The cavity is kind of detuned with respect to optimum cw operation such that it provides more stable operation when a lot of modes are “in sync”.

      There are other ways of doing it: semiconductor saturable absorber mirrors (SESAM) just another kind of non-linear effect in the cavity.

    2. weeell, laser gates itself. Sort of. Simpliest way to put it- there is just one pulse of energy travelling back and forth the resonant cavity and so only when it reflects at semi-reflective end any light comes out. Then it turns back a the speed of light and hence femtosecond pulses, driven by legnth of cavity. Look mode-lock lasers. As for measuremnt – I’d go with some interference-type effects, but I’m out of the trade for loong time

    1. Oh, it’s not so difficult. Just expensive. :-) when I was at university 25 years ago I built a Ti:Sa fs laser. Most of them at that time were kind of DIY. There was a good manual available. If I remember correctly it was from the Murnane group. The problem is: You need a Ti:Sa crystal suffciently doped, a couple of laser mirrors, prisms and a pump laser in the green. And of course a stable table and good optics mounts. My first pump laser was an argon-ion one. drawing several kW of electricity (and a lot of cooling water) for about 5W of green laser light which was then enough to get about 300mW of infrared laser pulses @ 80MHz repetition rate and sub 50 fs pulse length.
      Solid state pump lasers cut the energy bill drastically and mode-locked fibre lasers even more. But I am out of business for more then two decades, now.

    2. Nitrogen/TEA lasers are pretty short by their nature if I recall correctly, although I think nano or maybe pico rather than femto. Something about how once the medium gives up it’s photon, it’s rendered opaque to the lasing wavelength. That’s why there’s no need to create a cavity with HR and OC mirrors, which is probably really fiddly to implement in the garage. All of the output is created in one go, right when the arc happens.

  3. Plus-One. Matter of fact, I did one the other day. I used a tile drill bit from the hardware store, pretty sure it’s carbide. Key is to cool it as you mentioned, and take it slow and steady without too much down-force, not even for a second. I used a hand-drill, and submerged the part in common dihydrogen monoxide. No cracks, so I’m off to Mars to collect on some lotto tix I scooped up whilst reading this article; Back in a femtosecond.

  4. And for a larger hole, you use a compass-like device with a suction cup and a tungsten carbide point, go round once, then carefully tap the circle while still holding it by the suction cap. Beats the success rate of lottery by several orders of magnitude, on any planet.

    1. Ooh, always wanted to do that… I’ve cut lots of straight cuts of glass and some freehand curves (non-critical; I’d imagine doing accurate curves would require some practice) but haven’t yet done a circle. Something satisfying about glass cutting, when you get that clean hiss when you’re scoring, and how a gentle tap can pop the pieces apart…

      1. I’ve personally never managed to cut a piece of glass cleanly in my life, despite trying over decades and multiple types of glass, and normally considering myself a very manually capable person. I fully accept others can do it with ease, but I seem to lack that skill and it always chips or shatters. New tools, old tools, new glass, old glass, more pressure, less pressure, water/no water. Every time, CRACK. Sod it, take it to a glazier.

  5. Yeah I was a bit thrown by that. But from reading about the researchers’ results I infer that this is useful, not so much because they’re drilling in glass, but because they’re making very narrow and deep holes, which is challenging in any material, and especially if it’s non-conductive.
    I guess it’d be useful for making vias in μTAS (“lab on a chip”) devices, which are made of channels etched in a glass substrate.

  6. Can you help me with high power laser ?
    I build model of sunspots, using plasma generated by an array of 100 plasma burners.
    At the same time I build another model, employing magnetic fluid to simulate dense solar plasma.
    It need high power laser to “drill” the magnetic fluid to eject “plasma jet” simulated by hot magnetic fluid

    You are free to contact me at manta103g @ g m a i l . com

  7. Can you help me with laser ?
    I build functional models of sunspots, using 100 array of high power plasma burners
    Since plasma burners are heavy to operate I use another model, employing magnetic fluid.

    I need high-power laser to generate “plasma jet” in magnetic fluid.
    Let me know your opinion.

  8. Not so surprised, I saw a demo at Cambridge University’s Science Festival using a femtosecond laser to etch a match head without igniting it. They could also laser weld and then immediately pick up the material due to minimal stray heating. All very cool (literally, at least outside the pinpoint target)

Leave a Reply to Once_worked_with_upright_flying_saucers_of_lightCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.