Creating ultrashort laser pulses is required to perform any kind of ultrafast spectroscopy technique. Most systems use mode locked Ti:sapphire lasers. Figure 1 shows the absorption and fluorescence spectrum of Ti:sapphire. Lasing action is possible above 670 nm. The optimum lasing wavelength is about 800 nm and the optimum pump at about 515 nm. Modern Ti:sapphire lasers are pumped by the doubled output from a Nd doped YAG laser which is slightly off resonant at 532 nm. The power, compact size, and ease of operation compensates for the less efficient power transfer to the Ti:sapphire.
Figure 1 - Absorption and Florescence spectrum of Ti:sapphire.
Figure 2 diagrams the optical setup of the lasing cavity in a Spectra Physics Tsunami Ti:sapphire. When lasing is initiated, all of the fluorescence wavelengths are emitted into the resonator cavity. The fluorescence light then passes through the prism array which uses a slit to select a specific wavelength range. Most of the selected light reflects off the output coupler (OC) and returns to the Ti:Sapphire crystal. Some of these wavelengths are integer multiples of the cavity length and form standing waves, called modes. If the phases of these modes have a random phase relationship to each other the interference effects average to a near constant intensity continuous wave beam. If the phases of the modes can be organized with a fixed phase relationship towards each other they will periodically all constructively interfere at the same time and produce a strong pulse. Organizing the phases of the modes is called mode locking.
Figure 2 - Laser Optics of the Spectra physics Tsumani. The resonator cavity is between mirrors M1 and M10.The optics labeled Pr are prisms, AOM is the acousto-optical modulator and OC stands for output coupler.
Ti:sapphire lasers are self-mode locked by Kerr lensing in the Ti:sapphire crystal. Kerr lensing is a nonlinear optical effect where high intensity light is focused to a smaller diameter inside of the medium. For mode locking to occur the pump beam needs to have a much smaller diameter than the crystal. Initially the lasing modes of the wave are scrambled with random fluctuations in intensity. When a region of light with higher intensity passes through the crystal it is focused to a smaller diameter, overlaps more with the excited area of the crystal and increases in intensity. This region reflects off the cavity mirrors and returns to the crystal where it is focused harder by Kerr lensing because it has an increased intensity and is amplified even more by the overlap with the pump beam. After a several iterations a strong pulse builds up. When the pulse strikes the output coupler with sufficient intensity some of the light passes through the optic to the ouput window but some is retained in the cavity to help start a new pulse. The frequency of pulses leaving the laser is determined by the time required to complete a full trip of the cavity. Shorter cavities have higher repetition rates and longer cavities have lower repetition rates.
Self-mode locking lasers are not self-starting. Since Kerr lensing is a nonlinear effect, a region of light with the required intensity to initiate the process is often not present when lasing begins. A small external perturbation can cause a fluctuation in the light intensity to begin Kerr lensing. Some of the pioneering work on Ti:Sapphire lasers suggested “tapping a resonator mirror” to begin mode locking (Spence, 1990). The Tsunami use an acousto-optical modulator (AOM) to start mode locking. This device is a transparent crystal which changes its optical properties when a sound wave is passed through the solid. When AOM is active it cycles through a reflecting and absorbing state to create regions of higher and lower intensity that are further amplified by Kerr lensing. Once Kerr lensing has begun, the AOM may be powered off, and the laser will function normally. The shortest pulses that can be produced by self-mode locking Ti:Sapphire lasers are around 5.5 fs (Ell, 2001). Commercially available lasers vary with pulse durations between around 50 fs and pulse energy about 10 nJ.
Many experiments require higher pulse energies to collect significant data for publication. To increase the power and maintain the short pulse duration the output of the first Ti:Sapphire laser is used as the seed beam for a second Ti:Sapphire laser called a regenerative amplifier. In this setup the first Ti:Sapphire laser is called the oscillator. Amplifying very short pulses can cause extremely high peak powers in the GW range which could damage the optics. So the seed pulses are stretched to a longer pulse duration, amplified, and then compressed back into ultrashort pulses. This technique is called chirped pulse amplification as shown in figure 3. The chirp of a laser pulse refers to the order that different wavelengths appear in a light pulse. In this example the grating stretches the pulse in time with the red wavelengths before the blue wavelengths giving it a positive chirp. The stretched pulse then passes through a Ti:Sapphire crystal pumped by a more powerful YAG laser. In this crystal, since a seed beam is present, stimulated emission results in all of the pump energy amplifiing only the seed wavelengths. In most systems the beam makes multiple passes through the amplifying crystal before being released. After amplification the pulse is compressed with another set of diffraction gratings. The compression gratings remove the spectral chirp to shorten the pulse back down to around 50 fs. Following amplification, the pulse energy increases by a factor of a million to about 2 mJ at 1 kHz repetition rate. These pulses now have the correct temporal, spectral, and energetic properties for ultrafast experiments.
Figure 3 - Cartoon diagram of chirpped pulse amplification.