Pulse Width Measurement
B-5
GVD Compensation in Measurement
of Ultrashort Pulses
Because the pulses produced by the Tsunami
®
laser are extremely short
(< 80 fs), the pulse broadening in optical materials from GVD makes mea-
surement of its true pulse width difficult. Also, because the GVD of glass
causes the pulse width to broaden, the pulse that reaches an experimental
sample after traveling through beam splitters, lenses, etc., may not be the
same pulse that is measured in the autocorrelator. It is thus important to
ensure that the measurement technique and experimental set up incorporate
the same amount of glass and some GVD compensation if the shortest
pulses are to be measured and delivered to a sample.
Even before the pulse leaves the laser, it travels through extra glass. For
example, if we assume the pulse in a Tsunami laser is at its shortest as it
passes through the coating of the output coupler, it then travels through the
output coupler substrate, the photodiode beam splitter and the output win-
dow. For the Tsunami laser, the total thickness of these optics is about
1.9 cm (0.75 in.). Thus, a pulse that is 60 fs at the output coupler coating
becomes 66 fs by the time it exits the laser. Include the glass of the autocor-
relator and that in any experimental setup and the pulse can be broadened
substantially.
Since most autocorrelators use beam splitters, a lens, and often a spinning
block (as in the Model 409-08), the pulse is also broadened before it is
measured. This means the pulse out of the Tsunami may be actually shorter
than that indicated by direct measurement. Consequently, GVD must also
be compensated when using an autocorrelator.
Since the sign of GVD in material is generally positive for the wavelengths
produced by the Tsunami laser, introducing negative GVD into the beam
path compensates for the broadening effect of the material. Negative GVD
can be introduced into a system with prism pairs, grating pairs, or a Gires-
Tournois Interferometer (GTI). The prism pair provides the easiest, lowest
loss means for compensating for the positive GVD of materials.
To compensate for pulse broadening from materials, a simple setup using
two high index prisms (
SF-10) is all that is necessary. Figure B-4 shows the
layout (top and side views) for an easily built pre-compensation unit. The
laser pulse travels through the first prism where different frequency compo-
nents are spread in space. Then the broadened pulse travels through the
second prism, strikes a high reflector, and reflects back along its original
path–with one exception. The high reflector is slightly tilted in the plane
perpendicular to the spectral spreading and causes the pulse to travel back
through the prisms at a slightly different vertical height. After the beam
returns through the first prism it is reflected by another mirror to the auto-
correlator and/or the experiment.