The long and ultrashort of laser pulses

Every year, manufactured devices get smaller while boasting more capabilities. Traditional manufacturing techniques cannot address many of the process requirements related to producing these devices, such as denser and smaller features, and “clean” processing without subsequent cleaning steps. The pace also is accelerating in the laser-micromachining world, a result of the growing use of picosecond (ps) and femtosecond (fs) lasers. Known as ultrafast-pulse lasers (UFPLs), their pulses are so short that the energy they deposit in a material is almost completely directed at removing material. And there are few, if any, thermal side effects with UFPLs. With longer-pulse lasers, the photon energy deposited in a workpiece does not effectively vaporize material. The result is residual heat, which produces a heat-affected zone. It usually is desirable to eliminate, or at least minimize, any HAZ.

Polyimide cut with a 50-ns, 355nm laser (top) and a 12-ps, 355nm laser (bottom). The material cut with the ps laser is cleaner and required no post-processing. Photos courtesy PhotoMachining.

It’s generally best to apply the shortest pulse possible in material-removal operations, because doing so increases peak power on-target. Typically, though, the shorter the laser pulse, the more complicated and expensive the laser. Financial considerations may dictate use of a lower-cost laser—with an accompanying sacrifice in quality.

Most lasers used to process materials today have nanosecond (10-9 second) or longer pulse lengths. A few years back, femtosecond (10-15 second) lasers appeared on the market with the promise of nonthermal processing. Within the last couple years, picosecond (10-12 second) lasers became available, giving users a wide range of pulse lengths from which to choose.

In the ultrashort category, lasers are commercially available that will output 1 millijoule of pulse energy in a 10-fs pulse, leading to peak powers exceeding 100 gigawatts. (For perspective, this is the instantaneous, combined output of about 100 large nuclear reactors!) When combined with a small spot size—say, 4µm—the result is peak-power intensities in the terawatt (1012 w) range.

Among the benefits of UFPL processing, in addition to the aforementioned nonthermal material removal, are submicron control of ablation, material independence of wavelength and below-the-surface processing.

When it comes to pulse length, you might wonder, “How short is short enough?”

Let’s look at two of the most important factors controlling heat dissipation to see how they correlate with laser-pulse length. The first is absorption, which occurs on the timescale of approximately 1 fs. The second is heat transfer outside the skin depth. This occurs on the order of 1 to 10 ps. A simple analysis suggests that in order to get true athermal ablation, the pulse lengths should be in the 1- to 100-fs range, and at least less than 1 ps. The real requirement is that the pulse thermal diffusion depth (LD) must be less than the optical skin depth (δ) for the material being processed. For any given material LD = (D tp), where D is the thermal diffusivity and tp is the pulse duration. So, the shorter the pulse, the smaller the thermal diffusion depth.

In conclusion:

For athermal ablation, LD <<< δ

For thermal ablation, LD ≥ δ

The HAZ, both radially along the surface and into the bulk, is defined by diffusion, so LHAZ ~ LD

The peak irradiance in a short pulse is sufficient for multiphoton ionization of the dielectric matrix, and it’s this ionization that drives the apparent wavelength independence in an ultrashort pulse.

Now let’s consider three examples where short-pulse-laser processing may be a good fit technically and financially.

Bioabsorbable stents. Current metal-stent technology has limitations. Primarily, if a problem arises after implanting a metal stent, it must be surgically removed. A bioabsorbable stent’s substrate, on the other hand, dissolves in the body over time; problems that occur often can be resolved with an outpatient procedure. UFPLs are used to produce biostents because their substrates are very heat-sensitive, and UFPL-processed stents require no post-process cleaning. Cleaning adds cost and could compromise the part material. Also, given the cost of stents, ultrashort-pulse-laser machining them can be cost-justified.

Aerospace/defense applications. Part volumes for the aerospace and defense sectors are usually low, but parts must meet high tolerance specifications for dimensions, cleanliness and reproducibility. Many jobs are for one-off, one-time parts, but they must be “perfect”—whatever the cost. Ultrashort-pulse lasers are the best tool for many of these oddball jobs.

Specialty applications. Lasers sometimes are used to repair large and expensive devices that incorporate circuitry, like digital X-ray detectors. The devices cost tens of thousands of dollars to produce, and if shorts are found in circuit traces at the end of the manufacturing process, they may have to be scrapped. UFPLs are used to precisely cut shorted traces without melting the metal, which can cause other shorts. The repair cost is easily supported if only a fraction of the bad devices are saved.

I have observed that an fs-pulse laser does provide heat-free processing when the laser is set up and used correctly. When done incorrectly, though, the results can be really bad. I have also observed that a ps laser does not offer true heat-free processing; the wavelength effect clearly can be seen on many different types of materials.

What about a ps-pulse laser vs. an fs unit? Generally, shorter pulses are better—all things being equal. In practice, though, all things aren’t equal. Consider operating cost. The shorter the pulse, the less energy per pulse, and the more expensive the operation on a dollars-per-photon basis. So, a ps laser’s lower dollars-per-photon operating cost sometimes makes it more attractive than the faster-pulse fs laser.

As when choosing any laser-processing option, all factors—especially time, quality, money and equipment accessibility—must be considered. µ

About the author: Ronald D. Schaeffer, Ph.D., is CEO of PhotoMachining Inc., a high-precision laser job shop and systems integrator in Pelham, N.H. E-mail: rschaeffer@photomachining.com.