Unique Features of Ultrafast Lasers
The ultra-short pulse durations of ultrafast lasers give these systems unique characteristics that distinguish them from long pulse or continuous wave (CW) lasers. In order to generate such short pulses, a wide spectral bandwidth is required. The pulse shape and center wavelength determine the minimum bandwidth required to produce a pulse of a specific duration. Typically, this relationship is described by the time bandwidth product (TBP), which is derived from the uncertainty principle. The TBP of a Gaussian distributed pulse is given by.
TBPGaussian=ΔτΔν≈0.441
Δτ is the pulse duration and Δv is the frequency bandwidth. Essentially, the equation shows that there is an inverse relationship between spectral bandwidth and pulse duration, which means that as the pulse duration decreases, the bandwidth required to generate that pulse increases. Figure 1 illustrates the minimum bandwidth required to support several different pulse durations.
Figure 1: Minimum spectral bandwidth required to support 10 ps (green), 500 fs (blue) and 50 fs (red) laser pulses
Figure 2: Depiction of average power Pavg and peak power Ppeak for a laser with pulse duration t
Technical challenges of ultrafast lasers
The wide spectral bandwidth, high peak power and short pulse duration of ultrafast lasers must be properly managed in your system. Often, one of the easiest of these challenges to address is the laser's broad-spectrum output. If you have primarily used longer pulsed or continuous wave lasers in the past, your existing inventory of optics may not be capable of reflecting or transmitting the full bandwidth of ultrafast pulses.
Laser Damage Threshold
Ultrafast optics also have a significantly different and more difficult-to-navigate laser damage threshold (LDT) than more traditional laser sources (Figure 3). When providing optics for nanosecond pulsed lasers, LDT values are typically on the order of 5-10 J/cm2. For ultrafast optics, values of this magnitude are practically unheard of, as LDT values are more likely to be on the order of <1 J/cm2, usually closer to 0.3 J/cm2.
The significant variation in LDT amplitude for different pulse durations is a consequence of the laser damage mechanism based on pulse duration. For nanosecond lasers or longer pulsed lasers, the primary mechanism leading to damage is thermal heating. The coating and substrate materials of the optics absorb the incident photons and heat up. This can lead to distortion of the material lattice. Effects such as thermal expansion, cracking, melting and lattice strain are common thermal damage mechanisms for these types of laser sources.
Figure 3: Laser damage to optical surfaces, as shown here, can degrade the performance of a laser system, rendering it useless or even dangerous. Due to the short pulse duration, the damage mechanisms when using ultrafast lasers are significantly different from those when using longer-pulse lasers.
However, with ultrafast lasers, the pulse duration itself is faster than the time scale of heat transfer from the laser to the material lattice, and therefore thermal effects are not the primary cause of laser-induced damage (Figure 4). Instead, the peak power of the ultrafast laser transforms the damage mechanism into nonlinear processes such as multiphoton absorption and ionization. This is why it is not possible to simply scale down the LDT rating of a nanosecond pulse to that of an ultrafast pulse, because the physical mechanisms of damage are different. Therefore, under the same conditions of use (e.g., wavelength, pulse duration, and repetition frequency), an optic with a sufficiently high LDT rating will be the best optic for your particular application. Optics tested under different conditions are not representative of the actual performance of the same optics in a system.
Figure 4: Laser-induced damage mechanisms for different pulse durations
Dispersion and Pulse Extension: Group Delay Dispersion
One of the most difficult technical challenges encountered when using ultrafast lasers is maintaining the duration of the ultrashort pulse initially emitted by the laser. Ultrafast pulses are very susceptible to time aberrations, which make the pulse longer. This effect becomes worse as the initial pulse duration is shortened. While ultrafast lasers can emit pulses with durations of 50 seconds, it is possible to expand the pulse in time by using mirrors and lenses to deliver the pulse to the target location, or even just to transmit the pulse through the air.
This time distortion is quantified using a metric called group delay dispersion (GDD), also known as second-order dispersion. In fact, there are also higher-order dispersion terms that may affect the temporal distribution of ultrafast laser pulses, but in practice it is usually sufficient to examine the effect of GDD. GDD is a frequency-dependent value that scales linearly with the thickness of a given material. Transmission optics like lens, window, and objective lens assemblies typically have positive GDD values, indicating that once compressed the pulse can give the transmission optics a longer pulse duration than the pulse emitted by the laser system. Lower frequency (i.e., longer wavelength) components propagate faster than higher frequency (i.e., shorter wavelength) components. As the pulse travels through more and more matter, the wavelengths in the pulse will continue to extend farther and farther in time. For shorter pulse durations, and therefore wider bandwidths, this effect is further exaggerated and can lead to significant pulse time distortion.
For longer pulses with nanosecond or even picosecond pulse durations, GDD is not a major issue. However, for shorter femtosecond pulses, even placing a 10 mm thick piece of N-BK7 in the beam path can broaden a 50 fs pulse centered at 800 nm by more than 12%, which is roughly equivalent to placing two windows or filters in the beam path.
The impact of GDD on an application depends on several factors, including the input pulse duration (τinput), the center frequency (or wavelength), and the material through which the pulse propagates.
Equation (2) clearly shows that for the same value of GDD, a shorter pulse duration will broaden more significantly than a longer input pulse duration. This is why GDD is not discussed in the context of nanosecond or picosecond pulses. For example, a GDD of only 20,000 fs2 can widen a 1ps pulse by 0.2%. The examples in the following paragraphs show that this is equivalent to propagating a 1030 nm pulse into more than 1 m of fused silica.
The refractive index of a material depends on the frequency of light traveling through it, and GDD has a similar dependence on refractive index. When selecting transmission and refraction optics for ultrafast systems, fused silica is often recommended because it has one of the lowest GDD values in the visible and near-infrared wavelength ranges. For example, propagating a 1030 nm pulse through 1 mm of fused silica will produce a GDD of about 19 fs2 , but at the same wavelength, 1 mm of SF11 will result in a GDD of more than 125 fs2 Refractive index databases, such as refractivendex.info, are a useful resource for determining which material is the best optics for use in beam paths choice, and your accumulated GDD is a useful resource.
Due to this trend of positive GDD and time warping, it is highly recommended to use specialized ultrafast optics that produce little to no additional GDD, thus reducing the opportunity for extended pulse durations.
How do you know if you need pulse compression?
When do you need to (re)compress a laser pulse? In ultrafast imaging applications such as multiphoton microscopy, blurred images indicate that the pulse may be stretched in time. In ultrafast laser processing, pulse stretching can lead to reduced cutting accuracy and precision. Stretched pulse duration reduces the probability of multiphoton interactions, which reduces the efficiency of the ultrafast system. While it is not possible to provide strict and quick rules for every situation, the following example calculations help demonstrate some best practices for determining whether pulse compression is needed.
Consider a multiphoton microscope setup with a beam path as shown in Figure 5.
Figure 5: Example schematic of the beam path in a multiphoton microscopy experiment
A first-order approximation of the pulse expansion can be obtained by summing the GDD contributions of all elements in the system before the laser reaches the sample. Let us assume that the main contributors to the dispersion are the beam expander, the dichroic filters and the focusing objective. We will ignore the effect of the scanning mirrors since they are usually made of low GDD metal coatings. If the pulse is centered at a wavelength of 1030 nm, the system can easily add more than 600 fs2 of GDD.
Whether or not the pulse in the system needs to be compressed depends on the input pulse duration and the specific needs of the application. If you start with a 150fs pulse, the transmission through the optics will have a negligible effect on the pulse duration. However, if your application requires a temporal resolution that can only be achieved with a 10 fs laser pulse, then this amount of GDD will cause your initial pulse to expand to approximately 167 fs. In this case, recompression is required. These precise details are highly dependent on your particular beam path and application.
Ultrafast Laser Applications
Spectroscopy
Spectroscopy has been one of the main application areas of ultrafast laser light sources since their introduction. By reducing pulse durations to femtoseconds or even attoseconds, dynamic processes in physics, chemistry and biology that were historically impossible to observe are now possible. One of the key processes is atomic motion, the observation of which has improved the scientific understanding of fundamental processes such as molecular vibration, molecular dissociation and energy transfer in photosynthetic proteins.
Bioimaging
Ultrafast lasers with high peak power support nonlinear processes and improve resolution for bioimaging, such as multiphoton microscopy (Fig. 12). In a multiphoton system, two photons must overlap in space and time in order to generate a nonlinear signal from a biological medium or fluorescent target. This nonlinear mechanism improves imaging resolution by significantly reducing the background fluorescence signal that plagues studies of single-photon processes. Figure 13 illustrates this simplified signal background. The smaller excitation region of multiphoton microscopy also prevents phototoxicity and minimizes damage to the sample.
Figure 6: Multiphoton or nonlinear microscopy uses an ultrafast laser source to capture high-resolution three-dimensional (3D) images with reduced photobleaching and phototoxicity compared to conventional confocal microscopy techniques.
Figure 7: Signal position depiction of a two-photon two-photon (top) and single-photon (bottom) microscopy system. The overlap produced by the two photons results in a smaller excitation volume, while the single-photon signal is affected by the background signal from outside the focal plane.
Laser material processing
Ultrafast laser sources have also revolutionized laser micromachining and materials processing due to the unique way in which ultrashort pulses interact with materials. As mentioned earlier, when discussing LDT, the ultrafast pulse duration is faster than the time scale of thermal diffusion into the material lattice. Ultrafast lasers produce a much smaller heat-affected zone than nanosecond pulsed lasers, resulting in lower kerf loss and more precise processing. This principle also applies to medical applications, where the increased precision of ultrafast laser cutting helps to minimize damage to surrounding tissue and improve the patient experience during laser surgery.
Attosecond pulses: the future of ultrafast lasers
As research into advancing ultrafast lasers continues, new and improved light sources with shorter pulse durations are being developed. To gain insight into faster physical processes, many researchers are focusing on the generation of attosecond pulses - in the extreme ultraviolet (XUV) wavelength range, attosecond pulses are about 10-18 s. Attosecond pulses allow the tracking of electron motion and improve our understanding of electronic structure and quantum mechanics. While the integration of XUV attosecond lasers into industrial processes has not yet gained significant traction, ongoing research and advances in the field will almost certainly push this technology out of the lab and into manufacturing, as has been the case with femtosecond and picosecond laser sources.
