Dear Piotr,

have you thought also about separation via the time dependence of the fluorescence, which normally shows a much longer decay-time, than the excitation pulse (esp. for TPEF)?

(we are just doing a little research on this area :-)

Best regards

Wolfgang Paa

How weak the two-photon signal is with respect to single photon PL? If the ratio is not too bad (say better than 1:50) there might be a way. I suppose two-photon PL intensity scales differently with excitation laser intensity than single-photon PL (scales linearly). Than, you could possibly measure carefully intensity dependence of your signal (at least 10, better more points) and you could try non-negative matrix factorization. I have used this method to successfully decompose Raman signal of graphene and SiC. However, the noise level and number of spectra might be limiting factors.

Dear Piotr,

I think, the best way would be time resolved spectroscopy method as also Wolfgang Paa sugest. with the time resolving you can detect separately bott signals. Because of their different life times.

Best regards.

Rıza Demirbilek

Bonjour,

When amplitude modulating the excitation light, its intensity is proportionnal to the modulation while the TPEF varies as the square of the modulation. A large variety of electronique technique would easily discriminate the two signals.

The simplest way:

- 1) record a spectrum dividing the incoming intensity by a factor such that it does not excite the fluorescence. Probably a factor of 2 will do it?

- 2) record a spectrum at full intensity.

- 3) calculate the difference between record 2 and record 1 multiplied by the attenuation factor.

I once designed such a brutal discrimation technique for separating luminescence and Raman scattering.

LUMINESCENCE REJECTION IN RAMAN-SPECTROSCOPY

MORHANGE, JF; HIRLIMANN, C, APPLIED OPTICS Volume: 15 Issue: 12 Pages: 2969-2970 1976 DOI: 10.1364/AO.15.002969

There are two drawbacks:

- the TPEF intensity might be distorted if some is left in record 1. Not so important if you are not interested in the absolute value of the TPEF intensity.

- The recorded noise is the one of the strong signal that might grow faster than the TPEF if it is really weak.

One real way to do the recording once having tested with the above procedure. Use two synchronized beams to excite the TPEF; set a low frequency chopper on one of the two beams; send the two signals (full intensity, half intensity) into a dual amplifier in the substractive mode; set the amplifying factors such as to perform point 3) above.

Hope I was clear.

Piotr, building off of Charles' ideas, you could use a lock-in detection scheme and a scanning spectrometer. For each position of the scanned grating in the monochrometer, you demodulate the signal on your photo-detector at square of the modulation frequency to only pick out the two-photon fluorescence signal. The signal can then be amplified and filtered if you want to beat back some of the high frequency noise. If you chopped the beam at 100 Hz, you could easily find a 10 KHz oscillator that should allow you to demodulate the signal from two-photon absorption. This would be a time-consuming measurement since you have to scan across what I presume is a broad spectrum and averaging may be necessary at each particular point, but there would be no ambiguity in your signal.

Thank you Aaron for the very clear description of a possible set-up.

Probably new methods are not required when the old one does the job. According to my experience, Semrock multiphoton filters / beamsplitters work perfectly well.

Hi Piotr,

If you want to separate TPEF only from the incident laser light (and not from single photon fluorescence), you may consider using discrimination by polarization. Usually, the laser is linearly polarized, while the fluorescence is unpolarized. A good polarizer can reduce the laser light by 4 orders of magnitude or even more, while reducing the fluorescence intensity only by a factor of 2. If your TPEF signal is not too weak, you could use this method alone. Otherwise, it may be quite easily combined with all the other relevant methods mentioned above, and greatly increase their efficiency.

Good luck,

Eilon.

Hi,

If the just discremination of the excitation light, the combination of the above methods is recommended. But, if you need to compare the somthing different state, the time-gating method may cause the complex result because the fluorescence lifetime is also changing with the change of something. The polarization method also has a similar problem. I think, in general, that you must firstly optimize the optical setup and be using good filters like Semrock filters. Then, in the next step, you may consider the time-gating and polarization gating. In my case, we are not measurering TPE fluroescence but SPE fluorescence, but the medium is a highly strong scatter. This situation is more critical than your case because the spectral separation between the excitation and fluorescence is very small. Even though such case, the filters work very well. We are using filters by Semrock or Chroma.

Goro

Have you tried triple spectrometer. First two in subtractive mode and the third in additive mode. A suitable mechanical filter or modified slit at subtractive mode would give the required result.

Dear Piotr,

I am doing some modeling of recombination processes in nanocrystals and I am more precisely analyzing the time resolved PL of our samples. What I can tell you about what I know is that Francois Auzel published few works really investigating the difference and characteristics of each Two photon absorption and even an n-photon process. I am analyzing the Energy Transfer phenomenon occuring in our samples via ETU or APTE as Auzel called it. My analysis of our data mainly details the transitions responsible for different decay times for different emission bands. However in order to verify this dependence we use Pump-Power analysis. In fact, if you record the intensitiy of an emission band , say blue (405nm) for example, for several Laser powers so you can compare your curve with a P^n law and the n should give you the order of your absorption process n=2 ==> Two-photon process. I think that this article could be usefull :

Chem. Rev., 2004, 104 (1), 139-174• DOI: 10.1021/cr020357g

http://lib.semi.ac.cn:8080/tsh/dzzy/wsqk/selected%20papers/Chemical%20Reviews/104-139.pdf