PHTN1400: Principles Of Laser Systems
Femtosecond Lasers (2016W)


This is a one period (two-hour) lab in which students will investigate the use of a femtosecond laser including the conditions required for ultrafast pulse production.


The Experiment


A few things to note ...

An explanation of the wavelength selection system from the Tsunami User's Manual, pp. 3-9:

"The fs Tsunami laser is wavelength tuned using a prism sequence and a slit. This sequence provides a region in the cavity where the wavelengths are spatially spread. A variable slit is located in this dispersed beam. The output wavelength is tuned by changing the position of the horizontal slit in the vertical plane. The width of the slit can also be changed so that the bandwidth (and, hence, the temporal width) of the output pulse can be varied. This simple, straight-forward method covers the entire Ti:sapphire range for ultrashort pulses."

On the dispersion control:

From lecture notes, you should realize that the production of ultrafast pulses requires longitudinal modes to synchronize. Group Velocity Delay (GVD) is the effect whereby different wavelengths of light inside the cavity (i.e. different longitudinal modes) encounter different velocities. In any laser, optical elements induce positive delays (i.e. spreading of modes) which would foil modelocking. In our femtosecond laser GVD is controlled by inserting more, or less, of the prism glass into the intra-cavity beam path to compensate for this.

From the Tsunami User's Manual, pp. 3-9:

"The pulse width tuning characteristics of the Ti:sapphire laser are influenced by two factors: those inherent in the Ti:sapphire material itself and those from cavity parameters. While we cannot readily modify the Ti:sapphire material to change pulse width, we can modify the net group velocity dispersion (GVD) of the cavity. The optical components in the laser cavity introduce positive GVD and cause pulse spreading. Further pulse spreading is caused by self-phase modulation (SPM) in the Ti:sapphire rod, which results from the interaction of the short optical pulse with the nonlinear refractive index. In order to obtain stable, short output pulses, these effects must be compensated with negative GVD." "fs systems. Prism pairs are used to produce a net negative intracavity GVD in the fs system. Varying the amount and type of prism glass through which the intracavity beam travels changes the net intracavity GVD. Pulse widths from <35 to 150 fs can be obtained with various cavity configurations. Pulse width also depends on the operating wavelength."

On the slit control:

The prisms separate all intra-cavity wavelength spatially. A slit serves to select which of these dispersed wavelengths exists in the cavity and is used to tune the center wavelength of the laser. The width of the slit is adjusted only once: too wide and CW breakthrough occurs, too narrow and bandwidth is limited (and so pulses are broad). A narrow slit is desired when operating in purely CW mode (however using a femtosecond laser in this manner is akin to using a transport truck to drive to school each day: it will do it, but the overhead costs are high and there are easier ways to get CW tunable radiation).

You should READ APPENDICES A and B of the Tsunami manual which explains the modelocking technique, regenerating modelocking (i.e. the phase issue), GVD, etc.

Lab Report

Hand In a WORD PROCESSED (not handwritten) lab report with contents as outlined below.

Lab Report The FIRST PAGE must be a title page containing nothing more than the title of the lab, the course, and the student's name and ID number

Answer each question as "1", "2", etc with each new question starting on a NEW PAGE so that question 2 starts on the top of a new page and question 3 starts at the top of a different page, etc. Where a question has multiple parts (e.g. 3a, 3b, 3c ...) answer each in a separate paragraph with a title identifying the question in the form "3a., 3b., 3c. ...". Do NOT answer an entire question (e.g. question 3) as a single paragraph.

This format will assist you in ensuring EACH and EVERY question is answered since marks cannot be given for work not completed, nor would it be expected that you could complete the TEST QUESTIONS which will most certainly be similar to those you see here! (Hint !)

The lab must be submitted in a report cover (preferably either a three-hole punched cover or one with a clamp on the left side, not a binder), and NEVER as a stapled mass of loose papers!

Failure to follow this simple format, used for all condensed labs in this course, will result in deduction of marks


  1. Present key results as follows:
    1. The tuning range observed in CW mode (e.g. continuous of discrete) as a graph of output power vs. wavelength.
    2. An explanation (in words) of mode competition in CW mode and a drawing of the output showing this.
    3. A series of drawings showing the output starting with the phase control at the top (i.e. with no modelocking) and gradually morphing into a modelocked pulse.
    4. A paragraph explaining the effect of changing the phase control. What does this control really vary? (Hint: consider what happens if a longer cable was used between the 3955 and the laser, other than those shipped with the laser - use this as a specific example to explain how it works).
    5. The spectral width of the modelocked pulse (and the pump power, central wavelength, and control settings required to produce it).
    6. The wavelength range over which modelocked output was observed
    7. A drawing of the output showing CW breakthrough
    8. Give a description all observed conditions in this lab where CW breakthrough occurs
  2. Explain what the time-bandwidth product is with respect to a modelocked laser and why a femtosecond pulse will always have a large bandwidth.
  3. Explain the difference between a sech2 and a Gaussian pulses and how the time-bandwidth product changes. Give a simple formula for each allowing one to determine the time length of a pulse given the bandwidth.
  4. Calculate the length (in seconds) of the shortest pulse produced. You will be required to research in the manual the type of pulse expected. Be sure to show observed values and all calculations required.
  5. You probably noticed that the average output power is about the same whether the laser is operating CW or Modelocked. Calculate (showing all work) the individual pulse energy and individual peak pulse power of the laser in modelocked mode. Assume an operating frequency of 80MHz and use the observed average power from the 407 power meter.
  6. Once modelocking starts, it can continue without the 3955 modelocker (i.e. self-modelocking). Explain Kerr lensing as a possible mechanism to allow this to happen. A diagram is required as well as a paragraph describing the effect.
  7. An AUTOCORRELATOR is often used to determine the exact length of the pulse. Describe how this device works (cite where you researched this - complete URLs are required) and be aware that there are several possible configurations for such a device.
  8. What is GVD? Explain what "negative" GVD is and what "positive" GVD is in terms of which wavelengths are delayed (shorter or longer wavelengths). How does each occur and in what optical elements does each occur? What is the effect of inserting more of the glass prisms of the wavelength selector into the intra-cavity beam (i.e. does this cause more positive or more negative GVD)?