photon counting

[pdidier]

Hi i am setting up a fluorescence correlation spectroscopy (FCS) experiment based on a home made confocal microscope coupled to a single avalanche photodiode (SAPD Perkin Elmer) that is able to count photon by delivering TTL pulses (2.5 V, 9 ns). To perform FCS or single molecule FRET experiment analysis I need to record photon count during several minutes. We have installed a PCIe 6361 with a connector box. I am a newbie in the labview "world" and I tried to record photon flux by using the attached vi. as you will see it is not really optimized in term of speed acquisition. it works well if i record 1 000 000 points with 1µs time resolution. unfortunatly for ma purpose it is definitively not enough moreover the NI card is a able to run 100Mhz.

would it be possible to use some tricks ? like FIFO mode or a sliding buffer ?

many thanks for your help.

regards

[johnsold]

I cannot help with the DAQ and speed issues.  You may also have a memory issue.  Consider acquiring the data as integer (not scaled to DBL) and recording to a binary file.  Convert to scaled values later if needed.

A 100 MHz sampling rate is not fast enough to capture all of your 9 ns pulses, because the sampling interval is 10 ns. You must also have more than 10 ns between the trailing edge of one pulse and the leading edge of the next to be able to resolve the pulses. What kind of time resolution (between pulses) do you need to measure the correlation within your error budget?

Lynn

[pdidier]

Dear John,

thanks for your answer. I apologize; I did not give enough information in my post. If fact, in our experiment the count rate is at maximum 1 MHz, moreover the APD is saturated when the count rate reach 10 MHz. Basically we perform two kind of experiments; in the FCS scheme, the photon flux is recorded and we have an algorithm that is computing the autocorrelation of the fluorescence signal. As you can see on the picture, we typically need a microsecond resolution on a time scale that covers several orders of magnitude. In this case we need to record the flux with µs resolution for let says 1 minute. I am wondering if it is possible instead of counting the photon to record the arrival time of the photon. Indeed when the count rate is low (10-50 kHz) we will have zero values most of the time so it is maybe more clever to record arrival time …. The second kind of experiment is the energy transfer scheme in which we monitor the interaction between two biomolecules by using two fluorophores. A laser is exciting the first one (donor) and if the second one (acceptor) is at less than 10 nm you will observe an energy transfer from the donor to the acceptor. It is hugely used in biology to evidence molecular interaction. In this case you need to record with the detectors the fluorescence signal for several minutes with a time resolution of 1-100 µs. Again the size of the file can be very big.

I hope this will help you in your answer.

Regards,

pascal

[pdidier]

here is the picture of a typycal FCS curve.

[johnsold]

The APD produces 9 ns pulses.  At 10 MHz it is saturated.  Does that mean that after a 9 ns pulse it cannot generate another pulse until 100 ns have elapsed, whether another photon occurs or not?

If this is the case, then by stretching the 9 ns pulses to 50 ns and sampling at a rate greater than 20 MHz you could be assured of capturing every pulse.

Time of arrival would reduce your data if pulses are relatively scarce.

Lynn

[pdidier]

Exactlythis is the dead time of the detector in this case if a photon arrives within 100 ns it is not detected. How can I change the pulse form 9 to 50 ns ?

For the time of arrival this could be usefull in single molecule experiments where the count rate is generally 20-40 kHz which means a photon every 25 µs on average.

regards,

pascal

[phuongmai]

Hi Pascal,

Did you get the labview program for you studying ?  I am interested in that field right now. If possible pls send me email and we can discuss more in detail

My informaion : minhphuong.engi@gmail.com

Regards,
Phuong

[johnsold]

Pascal has not logged in since 2011 so it is unlikely that you will get an answer. Also posting your email address on a public forum leaves you vulnerable to spam.

Lynn