THORLABS FSAC User manual

FSAC
Interferometric
Autocorrelator
User Guide

Interferometric Autocorrelator
Table of Contents
Chapter 1Warning Symbol Definitions..............................................................................................1
Chapter 2Safety....................................................................................................................................2
Chapter 3Description...........................................................................................................................3
3.1.Introduction .....................................................................................................................3
3.2.Shipping List....................................................................................................................3
Chapter 4Setup and Operation...........................................................................................................4
4.1.Optical Requirements.....................................................................................................4
4.2.Setup and Coarse Optical Alignment............................................................................4
4.3.Operation and Fine Optical Alignment..........................................................................7
4.3.1.Power .................................................................................................................................... 7
4.3.2.Auto Shut Off ......................................................................................................................... 7
4.3.3.Output Signals ....................................................................................................................... 7
4.3.4.Controls ................................................................................................................................. 7
4.3.5.Fine Optical Alignment .......................................................................................................... 8
4.3.6.Optimizing the Autocorrelation Trace .................................................................................... 9
4.4.Common Problems .........................................................................................................9
4.4.1.Sampling Rate ....................................................................................................................... 9
4.4.2.Photodiode and Amplifier Bandwidth .................................................................................. 10
4.4.3.Saturation ............................................................................................................................ 11
4.4.4.Excessive Chirp .................................................................................................................. 11
4.4.5.Peak to Background Ratio Too High/Low ........................................................................... 12
4.5.Measuring Short (<40 fs) Pulses..................................................................................13
Chapter 5Theory and Interpretation.................................................................................................14
5.1.Working Principle..........................................................................................................14
5.2.Interferometric Autocorrelation Trace.........................................................................15
5.3.Intensity Autocorrelation Trace...................................................................................15
5.4.Excessive Chirp.............................................................................................................16
Chapter 6Maintenance........................................................................................................................18
Chapter 7Warranty..............................................................................................................................19
Chapter 8Specifications.....................................................................................................................20
8.1.General Specifications..................................................................................................20
8.2.Electrical Requirements ...............................................................................................20
8.3.Environmental Requirements ......................................................................................20
8.4.Mechanical Drawings....................................................................................................21
Chapter 9Declaration of Conformity.................................................................................................22
Chapter 10Regulatory...........................................................................................................................23
Chapter 11Thorlabs Worldwide Contacts..........................................................................................24

Interferometric Autocorrelator Chapter 1: Warning Symbol Definitions
Rev B, May 1, 2017 Page 1
Chapter 1 Warning Symbol Definitions
Below is a list of warning symbols you may encounter in this manual or on your device.
Symbol Description
Direct Current
Alternating Current
Both Direct and Alternating Current
Earth Ground Terminal
Protective Conductor Terminal
Frame or Chassis Terminal
Equipotentiality
On (Supply)
Off (Supply)
In Position of a Bi-Stable Push Control
Out Position of a Bi-Stable Push Control
Caution: Risk of Electric Shock
Caution: Hot Surface
Caution: Risk of Danger
Warning: Laser Radiation
Caution: Spinning Blades May Cause Harm

Interferometric Autocorrelator Chapter 2: Safety
Page 2 TT121442-D02
Chapter 2 Safety
All statements regarding operational safety and technical data in this manual will only apply when the unit is
operated correctly.
WARNING
Always wear appropriate laser safety eyewear during laser setup and operation.
WARNING
Use Appropriate Laser Safety Eyewear when Working with Lasers
WARNING
Visible and invisible radiation may be present.
Please note that while this is a passive device, lasers that emit high-power radiation are used in conjunction with
this device. Safe practices and proper usage of safety equipment should be taken into consideration when operating
lasers. The eye is susceptible to injury, even from very low levels of laser light. Laser emission in the visible and
near infrared spectral ranges has the greatest potential for retinal injury, as the cornea and lens are transparent to
those wavelengths, and the lens can focus the laser energy onto the retina. Follow all safety precautions in the
operator’s manual.
1. Never aim any laser at a person’s eye, skin, or clothes.
2. Always use proper laser safety eyewear.
3. Avoid wearing watches, jewelry, or other objects that may reflect or scatter the laser beam.
4. Keep the laser beam paths above or below eye level for both sitting and standing positions.
5. Ensure that individuals do not look directly into a laser beam.
6. Eliminate all unnecessary reflective surfaces from the vicinity of the laser beam path.
7. Ensure that all individuals who operate lasers are trained in laser safety and authorized to
operate a laser. Do not leave a running laser unattended if there is a chance that an unauthorized user
may attempt to operate the laser. A key switch should be used if untrained persons may gain access to
the laser. A warning light or buzzer should be used to indicate when the laser is operating.
8. Use low power settings, beam shutters, and laser output filters to reduce the beam power to less
hazardous levels when the full output power is not required.
9. Make sure that spectators are not exposed to hazardous conditions.

Interferometric Autocorrelator Chapter 3: Description
Rev B, May 1, 2017 Page 3
Chapter 3 Description
3.1. Introduction
Thorlabs’ FSAC is an interferometric autocorrelator capable of quantitative and qualitative monitoring of ultrafast
pulses from tens of femtoseconds to 1 picosecond. A modified Michelson interferometer splits an input pulse into
two copies, modulates the delay between them, and recombines the pulses at a photodiode. Because the
photodiode is not responsive over the fundamental wavelength band of typical Ti:sapphire lasers, two-photon
absorption is required to generate photocurrent. This results in nonlinear responsivity in the detector which enables
the interferometric autocorrelation.
3.2. Shipping List
The FSAC consists of the following components:
 FSAC Autocorrelator
 CL5 Mounting Clamps (Qty. 2)
 5/32" Stubby Hex Key
 VRC2SM1 Fluorescent Alignment Disk
 SM1CP2 End Cap
 R1DS3N Alignment Reticle
 LDS1212 Power Supply

Interferometric Autocorrelator Chapter 4: Setup and Operation
Page 4 TT121442-D02
Chapter 4 Setup and Operation
Environmental Requirements
The FSAC is intended for use in a laboratory setting, mounted on an optical bench, and should be operated under
controlled conditions to achieve stable performance. Humidity control is required to prevent condensation from
forming on optical surfaces. Keep the system away from air conditioning vents, which can cause sudden humidity
and temperature changes.
Electrical Requirements
The following table lists the electrical requirements of the FSAC.
Electrical Requirements
Input Voltage ±12 VDC
Power Consumption 6 W (Max)
The FSAC is designed to use the included Thorlabs LDS1212 power supply, which meets these requirements.
4.1. Optical Requirements
The FSAC is designed for horizontally polarized light (i.e., parallel to the plane of the table). In addition to allowing
the beam height to be adapted to the FSAC input aperture height (3"), a periscope such as Thorlabs’ RS99 (Figure
2) may be used to rotate the laser polarization by 90 degrees, if necessary. Alternatively, if additional dispersion is
acceptable, a half-wave plate may be used to rotate the polarization.
The noise equivalent sensitivity of the FSAC is approximately 10W (found by multiplying the average laser
power by the peak power) at 800 nm for a 1 mm beam diameter (1/). For the same beam parameters, the
response of the photodiode saturates at about 10W. Thus, for an 800 nm, Ø1 mm beam, the peak-average power
product should fall between 1 and 10W. If the peak-average power exceeds this level, then the average laser
power of the beam entering the FSAC should be attenuated (by using a neutral density filter or a Fresnel reflection
from an uncoated window, for example). The peak-average power product is roughly inversely proportional to the
square of the beam size, and may be scaled accordingly to roughly estimate the usable peak-average power range.
For example, with a Ø4 mm beam, the peak-average power product will decrease by a factor of roughly 16
compared to that for a Ø1 mm beam.
To avoid damaging the FSAC optics, the average power of the input laser should not exceed 150 mW.
The FSAC is designed for lasers with transform-limited pulse durations ranging from 40 fs to 1 ps, but may be used
for measuring pulses as short at 15 fs with dispersion compensation. See Section 4.5 for more details.
4.2. Setup and Coarse Optical Alignment
WARNING
Always wear appropriate laser safety eyewear during laser setup and operation.
The FSAC is roughly 7" x 6" (18 cm x 14 cm) in size and requires about 10" x 6" (25 cm x 14 cm) of optical table
space when two CL5 mounting clamps are used to secure the FSAC to the table (see Figure 1). In this configuration,
the autocorrelator beam height is 3" (76.2 mm) from the table surface. Alternatively, three 1/4"-20 and three M6 x 1.0
tapped holes on the bottom of the FSAC allow mounting with Thorlabs’ RS pedestal posts (not included) or other
mounting configurations. Ensure that that LDS1212 power supply is set to the appropriate line voltage, then connect
the LDS1212 power supply to the autocorrelator and set the power switch on the LDS1212 to ‘On.’ Next, use the
FSAC power switch to turn the autocorrelator on.

Interferometric Autocorrelator Chapter 4: Setup and Operation
Rev B, May 1, 2017 Page 5
Figure 1 FSAC Mounting Features
Figure 2 Thorlabs’ RS99 Periscope may be used to change the beam height while either
maintaining the beam polarization or rotating it by 90 degrees.
Attach the VRC2SM1 fluorescent alignment target to the alignment aperture of the FSAC (Figure 3). The input
aperture iris and the fluorescent alignment target provide two points which define the optical axis of the
autocorrelator. When the input beam is centered on both points, the system will be close to optimal alignment. To
achieve this, use two steering mirrors with tip and tilt adjustment in front of the autocorrelator as shown in Figure 4.

Interferometric Autocorrelator Chapter 4: Setup and Operation
Page 6 TT121442-D02
Figure 3 Front and Side Views Showing the Input Aperture (Left) and Alignment Aperture (Right)
Figure 4 Use two mirrors to center the beam on the input port and alignment aperture.
The easiest way to complete the coarse alignment of the FSAC is to start by using the steering mirror to set the
beam height to that of the autocorrelator input aperture, and ensure that the beam is parallel (level) to the table. M1
is adjusted to set the height of the beam, and M2 is used to level the beam, usually in alternating fashion. The FSAC
may be moved towards M2 to check the height, and away from M2 to check the level. Several iterations may be
required. Once the beam is level, place the FSAC at the desired location so that the beam is centered on the
aperture. Open the aperture so that the beam just passes, then carefully rotate the entire FSAC about the input
aperture (marked by the white X in Figure 4) until the beam spot on the alignment target is roughly centered. Make
minor adjustments until the beam is roughly centered on both the input aperture and the VRC2SM1 alignment

Interferometric Autocorrelator Chapter 4: Setup and Operation
Rev B, May 1, 2017 Page 7
target. Note that when the FSAC is grossly misaligned, spurious reflections may appear on the alignment target, so
be sure to look for the brightest spot.
Once the beam is centered on both the input aperture and target, use the CL5 clamps (or other mechanism) to
secure the FSAC to the table. Finally, alternate between using the adjustment on M1 to align the beam to the input
aperture and using M2 to align the beam to the alignment target, stopping when it is very well centered in both
places. The coarse alignment is complete.
WARNING
Up to 100% of the input light may exit the alignment aperture if it is not blocked by the fluorescent
target or an end cap.
Alternatively, if space is constrained or the position of the FSAC is fixed, M1 and M2 may be used to perform the
full alignment. The input beam should first be aligned so that it is centered on the input aperture, and approximately
perpendicular to the input aperture of the autocorrelator. Attaching the R1DS3N negative crosshair reticle to the
input port and using the back reflection can help with this coarse alignment (you may need to take care that the
beam is not perfectly retroreflected back to the laser cavity). Alternating between the two, the first mirror (M1 in the
diagram) should be used to roughly center the beam on the reticle, and the second mirror (M2 in the diagram)
should be used to adjust the reflected angle to within a few degrees of the incident beam. Next, remove the reticle
from the input aperture. The beam should now appear on the VRC2SM1 fluorescent target. If not, try scanning M2
until it appears. If it still does not appear, confirm that the beam is very close to perpendicular to the FSAC input
surface. Once the beam is visible on the alignment target, alternate between using M1 to center the beam on the
input aperture and using M2 to center the beam on the fluorescent alignment disk. Note that when the FSAC is
grossly misaligned, spurious reflections may appear on the alignment target, so be sure to look for the brightest
spot.
4.3. Operation and Fine Optical Alignment
4.3.1. Power
Figure 3 shows the side panels for the FSAC. Ensure that the LDS1212 power supply is on and connected. Push
the power switch on the FSAC up to turn the device on. The momentary switch will return to the neutral position,
and the power LED will turn green to indicate that the device is on. Push the power switch down to turn the power
off.
4.3.2. Auto Shut Off
When not in use, the FSAC should be turned off by the user. In order to preserve the lifetime of the device, the
FSAC is equipped with an auto-shut-off feature. Once the device is turned on, it will automatically turn off after
approximately 12 hours.
4.3.3. Output Signals
Trigger, delay reference, and photodiode BNC outputs are available on the control panel as shown in Figure 5. The
trigger signal is a 5 Hz, 5 V TTL signal synchronized with the delay scan rate which may be used to trigger acquisition
by an oscilloscope or a digital acquisition (DAQ) card. This signal is particularly useful during setup and alignment.
Once the autocorrelation trace is visible, triggering off of the autocorrelation signal itself allows for a more stable
view. The delay reference output provides a voltage between -5 and +5 V that is proportional to the delay position,
which can be helpful when setting up the autocorrelation trace (discussed below). The photodiode output provides
the voltage signal from a variable-gain transimpedance amplifier connected to the photodiode. The output signal
will contain the autocorrelation trace, and is between 0 and +10 V.
4.3.4. Controls
In addition to the BNC outputs, the control panel also provides four controls as shown in Figure 5. The delay stage
is driven with a sinusoidal wave that is nominally centered about the zero-delay point (the center of the
autocorrelation trace). The amplitude knob is connected to a 3-turn potentiometer which controls the amplitude of

Interferometric Autocorrelator Chapter 4: Setup and Operation
Page 8 TT121442-D02
this drive wave, allowing the full delay scan range to be adjusted from 50 fs to 10 ps (i.e., 25 fs to 5 ps). The
offset knob is also a 3-turn potentiometer. It adds an overall offset to the sine wave, which in effect moves the
average position of the sine wave with respect to the zero-delay point. Its use is described in further detail below.
The photodiode focus knob physically moves the position of the photodiode along the optical axis so it can be
moved precisely to the focal point of the FSAC. A final knob switches the photodiode transimpedance amplifier gain
from 0 to 70 dB in 10 dB steps.
Figure 5 FSAC Control Panel
4.3.5. Fine Optical Alignment
After completing the coarse alignment procedure (Section 4.2), start by ensuring that the laser to be measured is
mode locked. Next, turn the delay amplitude to its maximum value and monitor the delay output to verify that the
stage is moving. The delay output should be roughly sinusoidal, though it may be somewhat distorted due to
nonlinear coupling between the drive mechanism and the stage. Next, set the delay offset to approximately zero by
turning the offset knob all the way in one direction, and then turning 1.5 turns in the opposite direction so that the
indicator line is pointing up. The average voltage of the Delay Output should be near zero. Next, increase the
photodiode gain as high as it can go without saturating any signal that may be present (the highest voltage should
be less than +10 V). Adjust the photodiode focus position using the photodiode focus knob to maximize the
photodiode output signal. You may need to decrease the photodiode gain as the signal increases. When the
photodiode position is far from optimal, you can use the DC background of the photodiode signal for optimizing the
position until the autocorrelation trace appears, at which point the peak signal of the autocorrelation trace may be
used to optimize the position.

Interferometric Autocorrelator Chapter 4: Setup and Operation
Rev B, May 1, 2017 Page 9
The autocorrelation trace should now be visible. Decrease the delay amplitude to better resolve the autocorrelation.
You may now make small adjustments to M2 and the photodiode focus to optimize the signal. When the signal is
optimized, it may no longer be perfectly centered on the alignment target. This is normal. The spot on the alignment
target may also appear to oscillate at 5 Hz. This is inherent to the design of the device, and is also normal.
4.3.6. Optimizing the Autocorrelation Trace
The delay output (Figure 5) provides a voltage that is proportional to the position of the delay mirror with respect to
its nominal equilibrium position. The delay mirror is driven with a sinusoidal wave, but due to friction in the stage
bearings, the motion may deviate significantly from sinusoidal, especially near the turning points. Use the amplitude
and offset knobs to place the autocorrelation where the delay stage motion is linear, as shown in Figure 6. The
linearity may be verified by checking that the fringe spacing is the same at several positions in the interferometric
trace.
Sometimes triggering the oscilloscope off of the photodiode channel can create a more stable view. Be sure to set
the trigger holdoff on the oscilloscope to ~100 ms to avoid triggering off of both the forward and backward scan.
Next, adjust the photodiode amplifier gain to maximize the signal-to-noise ratio without saturating the photodiode.
Higher gains (>40 dB) reduce the detection bandwidth which may distort the signal. Also, ensure that the sampling
rate of the oscilloscope or DAQ is sufficient to resolve the fringes. In some cases of insufficient sampling, the trace
may appear normal but yield unexpectedly short pulse width measurements, since the fringe frequency is aliased
to a lower frequency. See Section Figure 6 for further details.
Finally, verify that the peak-to-background ratio (refer to Figure 12) is at least 7:1. Note that there may be a small,
non-zero offset when the laser is blocked, and that this should be subtracted from the peak and background
measurements. If this is not the case, please see Section 4.4.5.
Figure 6 Delay offset and amplitude are adjusted until the autocorrelation trace appears within the
linear region of the delay line, shaded in blue.
4.4. Common Problems
4.4.1. Sampling Rate
As discussed in Chapter 5, the interferometric autocorrelation contains a term oscillating at a rate corresponding to
the central wavelength of the pulse, and another term corresponding to the the second harmonic. The maximum
frequency component, f, in the interferometric autocorrelation trace corresponding to the second harmonic can be
derived from the fact that the delay mirror motion is roughly sinusoidal:

Interferometric Autocorrelator Chapter 4: Setup and Operation
Page 10 TT121442-D02
2
 

Here, 5Hz is the mirror scanning frequency,  is the the peak-to-peak time delay of the delay mirror (typically
greater than 5 times the expected pulse duration in order to operate within the linear range of the delay stage), 
300nm/fs is the speed of light, and  is the central wavelength of the input laser. According to the Nyquist sampling
theorem, the sampling rate must be at least twice  in order to avoid aliasing, otherwise the fringes may be detected
at a lower frequency resulting in an underestimated pulse duration. If the full envelope of the interferometric
autocorrelation trace is to be inferred without the aid of post-processing, then the sampling rate should be on the
order of 10 or more times . For example, to measure a 100 fs pulse with a central wavelength of 800 nm, a
conservative delay of 1 ps might be used.  is calculated to be about 12 kHz, meaning the sampling rate must be
24 kHz in order to detect the highest frequency in the interferometric autocorrelation without aliasing, and 120 kHz
in order to adequately resolve the envelope to estimate the pulse width.
4.4.2. Photodiode and Amplifier Bandwidth
The nominal detector bandwidth for each gain setting is shown in Table 1. For low gain settings the bandwidth is
limited by the photodiode, and for high gain it is limited by the transimpedance amplifier. Signal frequencies, as
described in Section 4.4.1, that are higher than the bandwidth for a given gain setting will be highly attenuated. This
results in distortion of the interferometric autocorrelation. In more extreme cases, like that shown in Figure 7, the
fringe contrast will be reduced—i.e., the peak will not be as high, and the signal will fail to go to zero near the center
of the trace. A quick way to check if the bandwidth is sufficient is to lower the gain setting to see if the contrast
improves. For a proper autocorrelation trace, ensure that the highest signal frequency (as described in
Section 4.4.1) is below the bandwidth for the chosen gain setting.
Figure 7 Left: An autocorrelation with an ideal 8:1 peak-to-background ratio.
Right: The autocorrelation signal fails to go to zero near the center of the trace and the fringe
contrast is low, both symptoms of insufficient bandwidth.
Gain Setting
(dB)
Bandwidth
(kHz)
0 120
10 120
20 120
30 120
40 100
50 32
60 11
70 3.3
Table 1 Detector Bandwidth Settings

Interferometric Autocorrelator Chapter 4: Setup and Operation
Rev B, May 1, 2017 Page 11
For pulse width measurements, post-processing of the interferometric trace may be used to attain the intensity
autocorrelation. In this case, it is not necessary to detect the second harmonic frequency, and significant attenuation
of the fundamental frequency may be tolerated. See Chapter 5 for more details.
4.4.3. Saturation
If the 2-photon signal is too high, the top of the autocorrelation trace may appear distorted as shown in Figure 8.
This is independent of the voltage output of the photodiode, so it may be observed even for low output signal levels.
If this is found to be the case, then the average laser power of the beam entering the FSAC must be attenuated (by
using a neutral density filter or a Fresnel reflection from an uncoated window, for example). For an 800 nm, Ø1 mm
(1/) beam, saturation occurs when the product of the average and peak power of the laser is around 10W.
Figure 8 When the 2-photon signal becomes too high, the top of the waveform may be flattened,
resulting in an erroneous measurement.
4.4.4. Excessive Chirp
When a pulse is highly dispersed, the interferometric autocorrelation may appear as in Figure 13, where the tails of
the trace lie outside of the coherence length of the pulse and are equivalent to an intensity autocorrelation, and the
center of the trace coherently interferes, creating a spike. In this case, using the interferometric autocorrelation
alone will underestimate the pulse width. However, post-processing may still be used to find the intensity
autocorrelation trace from which the pulse width may be more accurately estimated. See Chapter 5 for more details.

Interferometric Autocorrelator Chapter 4: Setup and Operation
Page 12 TT121442-D02
4.4.5. Peak to Background Ratio Too High/Low
WARNING
Use Appropriate Laser Safety Eyewear when Working with Lasers
As discussed in Section 5.2, the theoretical peak to background ratio is 8:1. However, due to imperfect beamsplitting
and phase errors between the two arms of the FSAC interferometer, the peak to background ratio may be lower
(~7:1). If this ratio is very low (<6:1), check that the signal is not saturated (Section 4.4.3), and that the sampling
rate (Section 4.4.1) and detector bandwidth (Section 4.4.2) are sufficient. Also note that there is typically a small,
but non-zero offset in the signal that depends on the gain setting of the FSAC detector and the oscilloscope or DAQ
card. This level can be measured by blocking the input to the FSAC, and then the level may be subtracted from the
entire interferometric autocorrelation trace.
The alignment between the two arms of the interferometer is quite sensitive, and may drift over time, especially
when subjected to strong temperature fluctuations (like during shipping). It may be necessary to periodically readjust
the reference mirror alignment in order maximize the peak to background ratio. To do this, begin by removing the
FSAC lid using a 5/32" balldriver, or the included 5/32" stubby hex key. Next, with the device on, monitor the
autocorrelation trace (be sure to take proper laser safety precautions) and adjust the tip and tilt actuators on the
reference mirror (Figure 9) using the included 5/32" stubby hex key. The required adjustment is typically very small,
less than 1/8 of a turn. Also, be extremely careful not to damage the beamsplitter, or perturb any of the other optical
components. If the alignment is so far off that no autocorrelation signal is present, rough alignment may be achieved
by adjusting the reference mirror so that the delay and reference beams are overlapped on the alignment target
(Figure 3). Make fine adjustments until the peak to background ratio is optimized.
Figure 9 Left: Inside of the FSAC. The reference mirror is indicated by the red box.
Right: Adjusters on the reference mirror are indicated by boxes and
adjuster locking screws are indicated by circles.
If the peak to background ratio is too high, it is likely that the signal to noise ratio of the background signal is
insufficient to accurately measure the background level. Try increasing the gain setting of the FSAC, or increasing
the average power of the input laser beam.

Interferometric Autocorrelator Chapter 4: Setup and Operation
Rev B, May 1, 2017 Page 13
4.5. Measuring Short (<40 fs) Pulses
It is important to understand that the measured pulse width using the FSAC is the width of the pulse just before the
photodetector. Dispersion in media causes broadening in unchirped and positively chirped pulses. For a Gaussian
pulse whose unchirped pulse duration is , the pulse duration will be increased as approximated by
ττ1󰇧4ln2

󰇨4ln
,
where τ is the new pulse duration,  is the total group delay dispersion (GDD) experienced by the pulse, and the
approximation holds when ≫
. The nominal GDD within the FSAC before reaching the photodetector is 230fs.
For measuring short pulses where broadening becomes significant, dispersion-compensating optics such as
Thorlabs’ DCMP175 and Thorlabs’ UMC10-15FS chirped mirrors may be used to compensate for internal dispersion
from within the FSAC.
Alternatively, if the pulse bandwidth and, consequently, the transform-limited pulse duration τ are already known,
then a system of equations can be derived and solved to estimate the pulse width entering the FSAC. Referring to
Figure 10, a bandwidth-limited pulse of duration  will be broadened by various optics of total dispersion  prior
to the FSAC to a duration of ′, for which we wish to solve. The pulse duration at the FSAC detector is the result
of broadening caused by optics prior to the FSAC, as well as optics internal to the FSAC, which have dispersion
D 230fs. This gives two equations:
1󰇧4ln2

󰇨
′1󰇡4ln2

󰇢.
The first equation may be solved for the total dispersion prior to the FSAC,


 ,
the results of which may be plugged into the second equation to estimate τ′.
Figure 10 Diagram of a typical pulse measurement setup. , ′, and are the FWHM bandwidth-
limited pulse duration, the pulse duration prior to the FSAC, and the pulse duration at the FSAC
detector, respectively. is the total dispersion introduced into the beam prior to the FSAC, and
 is the dispersion inherent in the FSAC design.
Various Optics
D0Dispersion Dac Dispersion Detector
FSAC Optics
τ0’
τ0τ

Interferometric Autocorrelator Chapter 5: Theory and Interpretation
Page 14 TT121442-D02
Chapter 5 Theory and Interpretation
5.1. Working Principle
Figure 11 shows the optical layout of the FSAC. The input beam pulse is split and copies are reflected from a delay
mirror and a reference mirror and then recombined at the beamsplitter. The recombined beam is focused onto the
detector. When the delay and reference path lengths are exactly equal so that the optical path difference (OPD) is
zero, the pulses overlap in time and add constructively to produce the maximum possible signal on the detector.
When the OPD exceeds the full width of the pulse so that the pulses no longer overlap in the combined beam, there
is no interference and the photodiode signal is at a minimum.
Figure 11 FSAC Optical Layout
As photocurrent in the FSAC is generated by a two-photon process, the output signal is proportional to the square
of the irradiance incident on the detector, which in turn is proportional to the square of the total electric field
amplitude. And since the speed of the photodiode is several orders of magnitude longer than pulse width, it must
be time integrated. This is written as


󰇛󰇜 |󰇟󰇛󰇜󰇛󰇜󰇠

|

d


,
where 󰇛󰇜 is the complex electric field of the pulse and  is the time delay between the two pulses arriving at the
detector
1
. The integrand may be expanded to give
1
Rulliere, Claude. Femtosecond laser pulses. Springer Science+ Business Media, Incorporated, 2005.

Interferometric Autocorrelator Chapter 5: Theory and Interpretation
Rev B, May 1, 2017 Page 15


󰇛󰇜 󰇟󰇛󰇜

󰇛󰇜

󰇠d


4 󰇟󰇛󰇜󰇛󰇜Re󰇝󰇛󰇜
∗
󰇛󰇜󰇞󰇠d



2 󰇝󰇛󰇜


∗
󰇛󰇜

󰇞


4 I󰇛t󰇜I󰇛tτ󰇜d


,
where ∗ denotes the complex conjugate. The first term is a constant background. The second term is a modified
interferogram of 󰇛󰇜 which is responsible for fringes at the central frequency of its spectrum. The third term is the
interferogram of the second harmonic of 󰇛󰇜, and the fourth term is the intensity autocorrelation.
5.2. Interferometric Autocorrelation Trace
Figure 12 shows an ideal autocorrelation trace and envelope for a well-behaved (secant-squared) pulse. The signal
to background ratio is expected to be 8:1, but is typically lower (~7:1) due to non-ideal conditions such as beam
alignment and imperfect beamsplitting. The fringe spacing is proportional to the central wavelength of the input
beam, so this can be used to calibrate the time axis of the autocorrelation trace. Typically the input pulse is assumed
to be either a Gaussian or secant-squared shape, in which case the full width half maximum (FWHM) of the upper
envelope
2
of the autocorrelation trace will be directly proportional to the FWHM of the input pulse according to the
values listed in Figure 12. The pulse should be symmetric, though for short pulses, small dispersion and balance
errors between the two arms may introduce some asymmetry.
Figure 12 Ideal Interferometric Autocorrelation Trace
5.3. Intensity Autocorrelation Trace
If the fringes are filtered out (either by insufficient response time of the detector and electronics, or by post-
processing) so that the second and third terms in the equation above integrate to zero, an intensity autocorrelation
with a peak to background ratio of 3:1 is produced. In this case, if the background is also removed and a Gaussian
or secant-squared pulse shape is assumed, then the FWHM of the background-free intensity autocorrelation trace
2
Diels, Jean-Claude M., et al. "Control and measurement of ultrashort pulse shapes (in amplitude and phase) with femtosecond accuracy."
Applied Optics 24.9 (1985): 1270-1282.

Interferometric Autocorrelator Chapter 5: Theory and Interpretation
Page 16 TT121442-D02
is proportional to the input pulse FWHM by the factors in Table 2. Search for “FSAC” at Thorlabs’ website to find a
sample MATLAB or Python script for converting interferometric autocorrelation data to an intensity autocorrelation
trace. The script performs the following functions:
1. Load interferometric autocorrelation data.
2. Load offset data (data taken with laser blocked).
3. Subtract mean offset from interferometric autocorrelation data.
4. Shift time axis data, t, so that t = 0 corresponds to the peak autocorrelation signal.
5. Determine the fundamental fringe frequency, f, from the magnitude of the Fourier transform data.
6. Use the fundamental frequency to convert the time axis, t, to a delay axis using the central wavelength of
the input laser, λ, and the speed of light, c.
Delayf/
7. Set all frequencies outside of f/2 to zero to remove all fringes.
8. Perform an inverse Fourier transform on the filtered data to get an intensity autocorrelation with background.
9. Using points near the beginning and end of the trace, estimate the background and subtract it from the
intensity autocorrelation trace to get the background free intensity autocorrelation trace.
10. Determine the background-free intensity autocorrelation FWHM.
11. Convert to a pulse FWHM by assuming a pulse shape and using the values in Table 2.
Intensity (Non-Interferometric)
Autocorrelation FWHM Relations3
Gaussian τ  /1.414
sech2 τ  /1.543
Table 2
5.4. Excessive Chirp
If the pulse is significantly chirped so that the pulse length greatly exceeds the coherence length of the laser, then
the autocorrelation will appear as in Figure 13. The wings of the autocorrelation are identical to those of an intensity
autocorrelation, and the point at which the trace transitions to an interferometric autocorrelation can be used to
quantify the chirp4. In this case, it is not straightforward to interpret the pulse width without additional post-processing
(i.e., filtering out the fringes as described above). See the cited references for more information on interpreting
autocorrelation traces.
3Diels, Jean-Claude, and Wolfgang Rudolph. Ultrashort laser pulse phenomena. Academic press, 2006.
4Ibid.

Interferometric Autocorrelator Chapter 5: Theory and Interpretation
Rev B, May 1, 2017 Page 17
Figure 13 An interferometric autocorrelation of a chirped pulse transitions to an intensity
autocorrelation in the wings. Pulse width estimation, in this case, is not straightforward.

Interferometric Autocorrelator Chapter 6: Maintenance
Page 18 TT121442-D02
Chapter 6 Maintenance
The FSAC does not require regular maintenance. To maximize lifetime of the device, be sure it is powered off when
not in use.