Documentation: Setup a 2 Photon Microscope

C.Seebacher
last change 27. November 2006
Seebacher@biz.uni-muenchen.de
www.lmu.de/~chr

What you need

Hardware

Block diagram of a 2P setup

Microscope

Nikon,  Leica, Olympus, Till-Photonics, Zeiss, and many more (access to the intermediate image)
motorized z-control (you want to do 3D)
dichroic mirrors and filters

Laser

expensive femto second laser for 2 photon excitation: e.g. Spectra Physics, Coherent etc.
any CW-laser for confocal microscopy (you need an extra pinhole between dichroic and detector, or use a small detector, or use fiber coupling)

Laser attenuator

Most lasers have too much power. Reflecting ND filters can do if they do not build up too strong Kerr lenses.
An EOM is the recommendation. I have an ConOptics.

Laser beam expander

most lasers have a 1mm collimated beam but for the scanner/scanlens/objective you need several mm. I use an approx. 4 times expander.

Laser puls compressor (prechirp)

You can get some improvement using prechired pulses to get as short as possible pulses at the sample.

Scanner

Galvos: small and fast e.g. CTI6210,CTI6215, GT100 etc.
Galvo driver board:
Scanlens, a normal f ~ 50nm lens could do if you do not need a planar field and a large field and a confocal detection.
Or complete digital scanhead like this Yanus2 via Till-Photonics.

PMT-Detector

PMT: sensitive, large area, e.g. Hamamatsu R6357.
High Voltage supply, e.g Hamamatsu C4900
Amplifier with ~1MHz bandwidth, (e.g 1k Ohm in, low noise ~100x Amplification, max. +-10V output for NI-card)
AD-Card: National Instruments PCI6111, PCI6110 or similar (should have 4 MHz simultaneous sampling on each channel). You need a speed approx. 4 times the analog bandwidth for Nyquist sampling. You can work with slower cards but then you need an integrating circuit witch needs much more effort.

Cables for synchronizing galvos and AD-sampling:

CCD-Spectrograph-Detector

planned for near future, demo is working

APD-Photon counting-Detector

planned with the counters on the NI PCI611x card, but I don't have a detector.

Computer

tested on a 1.3GHz, 256BM RAM
I use a 2.8GHz, 1GB RAM

Here is a photograph of our first setup:

Photograph of our setup

Software

Labview 8 with IMAQ/Vision 8 and NIDAQmx
A small dll BinningDLL\Release\Binningdll.dll is compiled for Windows. Labview on Linux and Mac needs a new compilation.
The colibri software
or: Florian Engert Software
or: Mathlab6.1 (ScanImage) - (does not work with the digital galvo control).
Good to have: ImageJ, Irfanview, Autohotkey.

Installation of Labview

You have to install Labview first. At the moment I use version 8.0. Higher versions should also work. You need:
I needed 3 CDs: Labview, NI-Vision Development and DAQmx that comes with the Card.
You can check the hardware and the installation with the program "Measurement and Automatisation Explorer". There you should have the following entries:
Geräte und Schnitstellen:
  Ports
  NI-DAQmx, here you can test the card
Software:
  LabView8
  NI-Vision_ImageProcessing and Machine Vision:8.0.0.7845:NIVision.dll
  NI-Vision_Sevices:8.0.0.3393:NIVisSvc.dll
  NI-VISA:3.4.0.49152: Windows\System32\visa32.dll
  NI-DAQmx:8.0.0f0: NationalInstruments\NI-DAQ
  NI-IMAQ:3.5.0: NationalInstruments\NI-IMAQ

Installation of Colibri

So far there is no installation file.
You need lots of vi-files in the \2Photon and in the \Bibliothek folders; download them all.
Remove the -template from the \2Photon\ScanMicroscope.ini-template and  \2Photon\GUI.ini-template filenames (use copy if you are in SVN).
Try to open some VIs in the 2Photon folder and look if it finds VIs from the Bibliothek folder (e.g. try Variables\Variables.vi, Treiber\TJob.vi, ImaqQueue\ImaqStorage.vi..)
Create a c:\Temp folder. You can change the path in the SaveImaqFromQueue.vi and GetImage.vi. But it should be placed on a fast and large disk.
Copy /variables/Mic.ini to a new filename and change \2Photon\ScanMicroscope.ini accordingly. Modify the new file for different hardware. Delete all driver.list entries you do not need.
Modify the bitsize="6.14E-7" in /variables/Mic.ini for different hardware. We use 614 nanometer per bit in the intermediate image. With a 12bit DA (6115-card!) multiply by 16.
Option analog galvo control (1)(2) edit: galvo.type="analog"  adclock.source="internal"
Option SmartMove with digital control(3) edit: galvo.type="digital"  adclock.source="external"
Try to start ScanMicroscope.vi.

Features

Comes with complete source code.
Vector based scan description.
1 to 4 channels only limited by the AD-card.
support of the digital interface to the SmartMove boards (bidirectional scanning with all speeds).
4 MHz sampling speed (AD-card and computer dependent) with variable binning n=1...65635 (slower than 15ms/pixel)
flexible scan speeds from sub ms (40µs) to several s (memory/binning-limit?) per line.
Tested: 10µ x10µ, 4x4 pixel with 12µs/pixel achieves 0.78ms per frame, but I lost data after 60 frames (buffer too small)
Tested: 100µ x100µ, 4000x1200 Pixel with 1µs/pixel. 4000x2000 does somehow not work (out of memory..)?!
Tested: 10µ x10µ, 100x100 Pixel with 1µs/pixel achieves 46.5fps
Tested: 4µ x4µ, 80x80 pixel with 6µs/pixel, 40ms/frame 100 frames continuous
Tested: 4µx4µ, 80x80 pixel with 12µs/pixel, 78.7ms/frame 500 frames continous
Arbitrary scan tracks
Spot scan
Line scan (at the monent not tested)
Flexible Job control
Units are meter at the sample.
Some support of PCO CCD cameras.
Contionous running (almost no gap between scans)