Del Mar Photonics

Femtosecond pulse measurement instruments:

Femtosecond Autocorrelators Reef-RT (scanning)
and Reef-SS (single-shot)  - Reef-RT - buy online - specs
Autocorrelation technique - Reef-RT AA-10D manual
Avoca SPIDER Spectral Phase Measurement System - manual
Rincon Cross-Correlator Pulse Profiling Systems - TOCC
Rincon SS Single shot third-order cross-correlator
Optimization of Rincon TOAC for customer laser system

Self-Referencing Spectral Interferometry for Measuring Ultrashort Optical Pulses
Chris Iaconis and Ian A. Walmsley

Abstract—This paper describes a novel self-referencing interferometric method for measuring the time-dependent intensity and phase of ultrashort optical pulses. The technique, spectral phase interferometry for direct electric-field reconstruction (SPIDER), measures the interference between a pair of spectrally sheared replicas of the input pulse. Direct (noniterative) inversion of the interferogram yields the electric field of the input pulse without ambiguity. The interferogram, which is solely a function of frequency, is resolved with a spectrometer and recorded with a slow detector. Moreover, the geometry is entirely collinear and requires no moving components. This paper describes in detail the principle of operation, apparatus, and calibration of SPIDER and gives experimental examples of reconstructed pulses. Index

Terms— Pulse characterization, pulse measurements, pulse reconstruction, ultrafast optics.

THERE ARE three known strategies for measuring the electric field of optical pulses; spectrography, tomography, and interferometry [1], [2]. The most common of these, spectrographic techniques, such as frequency-resolved optical gating (FROG) [3], temporal analysis of spectral components (TASC) [4], frequency-domain phase measurements (FDPM) [5], and spectrally and temporally resolved upconversion technique (STRUT) [6], measure a two-dimensional (2-D) representation of the one-dimensional (1-D) field and consequently require the collection of a relatively large amount of data. Moreover, unless the pulse has a simple structure, spectrography requires sophisticated iterative data inversion algorithms to reconstruct the field. Tomography also requires a large 2-D data set, but data inversion is direct (noniterative). In practice, however, it is difficult to fabricate a tomographic apparatus with a bandwidth of several terahertz and, consequently, tomography has not been applied to ultrafast pulses, although tomographic-like methods that require iterative inversion have been demonstrated [7]. Interferometry, in contrast, requires only a 1-D data set to reconstruct a 1-D field and uses direct data inversion to do so. Diels has demonstrated an interferometric method that uses an ultrafast photodiode and Schottky diode nonlinear mixer to record the beat notes more

Femtosecond Autocorrelators Reef-RT (scanning)
and Reef-SS (single-shot)  - Reef-RT - buy online - specs
Autocorrelation technique - Reef-RT AA-10D manual
Avoca SPIDER Spectral Phase Measurement System - manual
Rincon Cross-Correlator Pulse Profiling Systems - TOCC
Rincon SS Single shot third-order cross-correlator
Optimization of Rincon TOAC for customer laser system


Avoca 7 SPIDER


Wavelength  range                                600 ÷ 1060 nm

Pulse  duration                                           6 ÷ 25  fs

Minimum required input power      300 mW average for 100 MHz repetition rate;

                                                              5 mW  for  1kHz repetition rate.

Input polarization                                        horizontal

Input beam diameter                                2 ÷ 4 mm

Input beam height                                110 ÷ 140 mm



I.                    The principle of operation of  AVOCA SPIDER 7 . The principle of operation of AVOCA SPIDER 7  is a version  of spectral shearing interferometry1,2 (SSI).  This version  is known as  SPIDER – Spectral Phase Interferometry for Direct Electric field Reconstruction 3, 4. In  SSI  the pair of replicas of  the tested  ultrashort  light pulse

are shifted in frequency with  respect  to each other and then are interfered   in spectrometer.

A very simple noniterative algorithm of recorded  spectral  interferogram gives the spectral

phase of  ultrashort light pulse to be characterized . In addition with independently measured pulse spectrum  after Fourier transform the time-dependent phase and intensity   is obtained .

    The scheme of  SPIDER apparatus is shown in Fig.1. Two identical pulse copies are delayed with respect to one another by time t and frequency mixed with a stretched pulse in

a nonlinear  crystal. Each pulse replica is upconverted with a different frequency slice, and consequently, the upconverted pulses have a spectral  shift.  Note, that the central frequency of the upconverted pulses is near  2 wo (wis the central frequency of measured pulse). 

   The output signal S (wc)  of spectrometer is related to the input pulse spectrum by



                 S(wc)=| (wc)|² + | (wc +Ω)|² + 2| (wc)||(wc +Ω)|

                                            õ cos [Φω(ωc + Ω) - Φω(ωc) + wcτ]                                        (1)


Where (wis the complex  representation of the input pulse electric field,  Ω – is the frequency shift, τ is the time delay between  the two replicas, and wc is the variable center

passband frequency of  the spectrometer. The first two terms on the right side of (1) are the

individual spectra of the test pulse and its frequency shifted replica, respectively. The third term provides the spectral phase in the form of the phase difference between spectral components separated by the shear Ω. One obtains the spectral phase for a set of discrete frequencies separated  by Ω from the spectral phase difference by simply adding up the

appropriate phase differences. The spectral  amplitude at these frequencies is obtained from

the square  root of an independently recorded pulse spectrum.

       An example of the experimental  spectral  interferogram  of a pair spectrally shifted pulses is shown  on  Fig.2. The nominal fringe  spacing  is 1/τ (in linear frequency units). Spectral phase of the tested pulse manifests  itself  as deviations from the nominal fringe spacing. 

       To extract the spectral phase from the experimental  interferogram the robust direct (noniterative)  inversion procedure is used. This procedure is diagrammed schematically in Fig.3. The Fourier Transform of the experimental interferogram from Fig.2  is shown in Fig.4.

The component centered near   t = + τ  is filtered and inverse transformed back into frequency


        The next step  of  the phase reconstruction  procedure is to remove the linear phase term

τwc   from (1). The most reliable method to do it is a direct measurement of  τwc. The interferometer is calibrated by recording a spectral interferogram for the pair pulses without

imparting the spectral shear. In this case, the interfered pulses are identical and  therefore the only phase contribution is τwc. The linear phase obtained from calibration is simply subtracted from   Φω (ωc) - Φω(ωc - Ω) + τwc.

               On subtracting the term proportional to τ from the resulting phase distribution we are left with the relative phase, Φω(ωc + Ω) - Φω(ωc), between  successive pairs of frequency components separated by Ω.  The spectral phase for a discrete set of frequencies separated by

 Ω  is constructed by concatenation. This inversion procedure is noniterative;  the phase is directly retrieved from the data by a series of linear  transformations.  Thus, given appropriate data, the reconstruction error associated with the inversion routine is zero.  Of course the accuracy of the reconstruction  is limited by the signal-to-noise ratio of the experimental data,

as it is for all methods.

1. V. A. Zubov and T. I. Kuznetsova, Laser Phys. 2, 73 (1992).

2. V. Wong and I. A. Walmsley, Opt. Lett. 19, 287 (1994).

3. C. Iaconis and I. A. Walmsley, Opt. Lett. 23, 792 (1998).

4. C. Iaconis and I. A. Walmsley, IEEE J. Quantum Electron. 35, 501 (1999).


II.               The description of  AVOCA SPIDER 7 apparatus


                  The  optical scheme of AVOCA SPIDER 7 device is shown in  Fig.5. The direction of the input 

            horizontally  polarized  beam is fixed by the diaphragms  D1  and  D2. The reflections from

the front and back sides of an uncoated  fused silica ethalon ET are used as  the two time delayed  replicas of the input pulse .  Thickness of ethalon is equal 38 microns, and time

delay τ  between  these two reflections  is approximately 375 fs. After the reflection by mirrors CM, M1 and two-mirror Polarization  Rotator PR  the polarization of the replicas is changed from horizontal to vertical.  Then with the help of the flat mirrors  M2,M3  and the spherical (radius of curvature 200mm) Focusing Mirror FM the test pair of replicas is focused into Nonlinear Crystal  NC. The most part of the input pulse passed through  the ethalon and diaphragm D2 then stretched  in  the glass stretcher consisted of high dispersive TF5 glass slab 75mm in length.

 The  stretched pulse after leaving the stretcher passes the delay line DL and reflected by flat mirrors M7, HA (Height Adapter), M8, M9 and focusing mirror FM is focused on Nonlinear Crystal NC.

The parallelness of the stretched beam and the test pair beam falling on FM is providing

by appropriate adjustment of the M2 and M3  mirrors. These two beams should be parallel and lay in vertical plane with the most possible accuracy . The optical passes of the stretched pulse and a pair of test pulses from the ethalon to the focus of FM where NC is placed

is equalized by  Delay Line DL in the stretcher.  Note that the stretcher beam and the test pair beam are polarized in the perpendicular directions.  So to upconvert the test pair with the stretched pulse  Type II  40μ  BBO crystal  in non-collinear geometry is used. The NC  is cut for phase matching at  λo =800nm   and   λe =800nm ( φ=0o, θ=43º) and oriented so, that the test pair of pulses polarized along the ordinary axis. For this orientation  of  NC its bandwidth is considerably broader then  when rotated through   90º. The Type I (φ=90o, θ=32º) 10μ BBO crystal is used  to record the Reference spectrum.

    The upconverted  beam passes through Lens L1, Slit Diaphragm SD  and focused by Lens L2 ( both lenses with focal  length 60mm)  at the entrance slit- SL2 of “SIGNAL” SPECTROMETER. The input pulse spectrum is recorded by the “MAIN” SPECTROMETER . For this purpose the reflection of the input pulse from TF5 glass slab GDD is used. This  reflection with the help of  the wedge W and two flat  mirrors M10 and M11 comes to the entrance  slit SL1.

    These two diffraction grating spectrometers allow the simultaneous recording of the input pulse spectrum in the range 600÷1060nm and the upconverted spectral interferograms  in the range 345÷455nm. The recording of the spectra is carried out by two one- dimensional  CCD arrays  CCD1 and  CCD2. Every CCD array has 1024 pixel with pixel size  25μ (W)x 500μ(H).


III.    Software  Installation


Efrat Software Installation and Configuration Notes





o IBM PC compatible computer

o Processor > 600 MHz

o RAM > 128 MB

o HDD space for installation 10 MB

o Video 1024x768 HighColor or better

o USB port

o Microsoft Windows 98SE/ME/2000/XP




The software distribution consists of EfratSetup-

executable file and a set of files in Presets folder. These files contains the

information on spectral calibrations and sensitivity for spectra registration

system recorded during device fabrication. The folder Presets must be

located in the same folder as setup executable.


To install software:


o Run EfratSetup- by double-clicking it

in Windows Explorer. (If you have CD distribution and autoplay

for CD-ROM enabled then setup will start automatically after CD

insertion). On Windows 2000/XP you must have administrator privileges

to install software. The Setup Wizard will guide you for the software

installation process.


o Select 'OrminsCCD USB (SINGLE Head) DAQ Driver' item in the Components

page of a Setup Wizard to install spectra registration CCD driver on

your system.


o The driver requires vendor supplied operating system device driver

to access data acquisition hardware through the USB interface. You

select 'Install OS driver' in the Additional Tasks page of a Setup

Wizard to install vendor supplied device driver. This will launch

vendor supplied Setup Wizard. You use default settings proposed by

this wizard and wait until it finished its tasks. Then you select

'I want to restart my computer now' option from Setup Finished

page of a main Setup Wizard to complete software installation.




o Connect devices units to the computer using supplied USB cables.


o Check that device OrmisCCD appeared in the Device Manager under

USB Controllers branch. (In some cases Windows may try to

reinstall OS driver and launch 'New hardware found' wizard.

Follow the wizard and select the device driver found on the system).




o Launch Efrat using Windows Start menu or desktop shortcut. The main

window of the application will be opened.


o Under 'Acquisition' menu choose 'Configure drivers...' item. The

Drivers List dialog box will be opened.


o In Drivers List dialog box click the 'Add...' button. With the

Open dialog box navigate to the folder where you've installed

Efrat, and then to subfolder DRV\OrminsCCD. Select file named

OrminsCCD1.dll and click the 'Open' button. The OrminsCCD Configuration

dialog box will be opened (the device must be connected to computer

at that time).


o In the OrminsCCD Configuration dialog box you'll see device

serial number and selected check box to the left of it. Just

click 'OK' button in this dialog box. Two spectra acquisition

window will be opened.


o Now you click 'Close' button in the Drivers List dialog box (or you

will be unable to access the application main window).




o In spectra acquisition window named by default 'Device:OrminsCCD<0>'

click on the down arrow to the right of the 'Calibration...' button.

Choose 'Load...' item in the popup menu.


o Navigate the folder where you've installed Efrat and select the file

named 'mainsp_7.SCL'. Click the 'Open' button. Now you have spectral

calibration loaded to that spectrum channel.


o Repeat two previous steps with spectra acquisition window named

'Device:OrminsCCD<1>' but open the file named 'signalsp_7.SCL'.




o In main window choose 'Sensitivity database...' under "Acquisition'



o Click the 'File' button and choose 'Merge...' item in the popup menu.


o Navigate the folder where you've installed Efrat and select the file

named 'spsensmainsp7.SDB'. Click the 'Open' button. Now you have spectral

sensitivity data loaded to the program.


o In spectra acquisition window named 'Device:OrminsCCD<0>'  click on

the down arrow to the right of the 'Equalization...' button. Choose

'Setup...' item in the popup menu.


o In 'Availabe Sensitivity Curves' list select ''spsensmainsp7.SDB'' item and

press '>' button move it to 'Selected Sensitivity Curves' list. Click

'OK' button to finish this setting.




All settings you've made will be recorded in the program configuration file

and will be restored when you launch Efrat again. For more information on

the software operation see online help (available under 'Help' menu in

the main window of Efrat) and other supplied documentation.



IV.            Unpacking of AVOCA SPIDER 7 device.

       Carefully extract the device from the transportation box and very carefully mount it on your optical table. Use the height adjusting legs to mount the device horizontally and to choose the beam height you need. Reliably fix the legs to optical table by applied cramps. Lift the cover and accurately remove the transportation protector from MS (See Fig.6).      



V.     Measuring  Procedure


     Before you start Measuring Procedure, please look through Efrat-software Help.

Also you should check the input pulse power. For the first time it should be: 300-500mW average for 100MHz repetition rate and 5-10mW for 1 kHz repetition rate. The beam diameter shouldn’t exceed 4mm (at 1/e­2  intensity level) . The better value 2-3mm. AVOCA SPIDER 7 also expects a collimated beam without spatial chirp.


1.      Using the outer fold mirrors let the input beam pass through the installed flip diaphragms D1  and D2 with the most accuracy.

2.      Flip down the diaphragms  D1  and D2.

Attention! Be careful when flip down  D2 – don't touch the  ethalon  ET!

3.      Check the parallelness  of the stretched and the test pair beams. For this purpose carefully flip down the focusing mirror FM in horizontal position and let these two beams pass through the double diaphragm  DD (See Fig.5 and Fig.6 ). Remove the side cover just behind DD and check the beam spot separation at the distance 3÷4 meters far from DD. The beams should lay in vertical plane and beam separation at mentioned above distance should be the same as at DD – approximately 8 mm.

 Warning!  The stretcher beam   direction is factory maintained and need not for   correction. Use only the alignment with the help of  M2, M3  mirror.


4.      Slowly return the focusing mirror FM in vertical position.

5.      Set up the width of  SL1  of  “Main” Spectrometer in the range  30 ÷50 μm . Turn on the “Autorun” mode on CCD<0> window of Efrat – software. Set up the exposition time of CCD 10÷20ms and switch off the “Equalization” button. Changing the SL1 width and exposition time set up the maximum value of the input pulse spectrum signal in the range 3000÷12000 a.u. Shut the input beam and correct the zero-line by “Baseline” button. This correction should be carried out every time you run Efrat  and after every exposition time changing.


6.      Unfix the adapter with Nonlinear Crystal  NC (10μ Type I BBO for pulses 6- 12fs duration or 40μ Type II BBO for 12-25fs duration )and rotate it in the position for recording the Reference spectrum. In this position Red mark on NC-adapter should be matched with White mark on  NC mount. Fix the adapter again.

7.      Set up the width of SL2 in the middle position ~200μm.

8.      Translate the Slit Diaphragm SD in the position allowing to pass the test pair beam only – the lower beam spot after NC.

9.      Check the maximum value of the CCD<1> signal ( Reference spectrum). This signal shouldn’t be saturated. The saturation level is ~ 16000am. If the signal is saturated, then slowly close the SL2 width. Using M4 mirror adjustments try to get the maximal signal for the minimal SL2 width. This maximal signal value shouldn't exceed 13500-14500 a.u.

Note, that the signal level also depend on the central wavelength  of the input pulse spectrum because  NC is factory phase matched for the central wavelength  λc =800nm. If the central wavelength of the input pulse spectrum differs of 800 nm ,  then you need to adjust NC for this wavelength .  For this purpose use Nonlinear Crystal Adjustment Screw  NC – A . After optimizing the Reference signal level record it to your Efrat reconstruction  project.

10.   To record the Signal spectrum only Type II  40μ NC is used .  So if you recorded the Reference spectrum with the help of Type I 10μ NC you should  replace it by Type II 40μ  NC and match White marks on NC adapter and NC mount. If you used Type II 40μ NC for recording the Reference spectrum then you simply unfix the adapter NC and rotate it by 45˚ to the position for recording the  Signal interferogram or simply Signal as it named in Efrat – software. In this position the White mark on NC  adapter should match with the White mark  on NC mount . Fix NC adapter again.

11.  Slowly rotating Delay Line microscrew MS (See Fig.6) set its  position around 4.30mm. This position corresponds the upconversion with wavelength near 800nm of the stretched beam.

12.  Translate the SD in the position allowing to pass only noncollinear  upconverted beam and simultaneously  blocking the stretched  and test pair beams.

13.  Optimize the Signal  value slightly adjusting M4 mirror and the SL2 width as it described in the Subsection 9 of Measuring Procedure. Check the parallelness of the stretched and the test pair beams slightly adjusting only M3 mirror and trying to maximize the Signal value. Then try to maximize  the Signal by NC adjustment as it described in the subsection 9. If the adjustments of M3 or NC is changed  to compare  with the recording Reference Signal it is necessary to record the new Reference signal with these new adjustments of the M3 and NC  and then continue the Signal acquisition .

14.  To reconstruct the input pulse you should introduce the Stretcher parameters. These are: Slab material – TF5 Glass; Slab length – 75mm.

15.  If the repetition rate of the tested laser pulses is 10Hz or less you should use the outer synchronization of CCD. For this purpose connect the both CCD synchronization ports and

      the synchronization pulse source by means of supplied cables marked by the arrows directed from the synchronization source toward the CCD camera. The synchronization pulse should be positive, 5-20V in amplitude (the input impedance of the synchronization port is 11 kOhm) and precede the laser pulse 1.5ms or more at the level 1-1.5V. Another synchronization cable (marked by arrows directed from the CCD camera) should be used if for some purpose you need in CCD camera output synchronization pulse.


Del Mar Photonics

Femtosecond pulse measurement instruments:

Femtosecond Autocorrelators Reef-RT (scanning)
and Reef-SS (single-shot)  - Reef-RT - buy online - specs
Autocorrelation technique - Reef-RT AA-10D manual
Avoca SPIDER Spectral Phase Measurement System - manual
Rincon Cross-Correlator Pulse Profiling Systems - TOCC
Rincon SS Single shot third-order cross-correlator
Optimization of Rincon TOAC for customer laser system