Are you aware about after-pulsing effect of the photomultiplier (PMT)? It is signal-depending background. Look figure 4 in attached article and ref.[25]. You might acquire noise if PMT get a light during the laser pulse, the noise will be visible with a time delay after laser pulse. 

[25] W. Becker. ”The bh TCSPC Handbook”, Becker & Hickl GmbH, 3rd Edition, 2008; http://www.becker-hickl.com

Dear Saurabh,

I agree with Andrei, after-pulsing will account for a large fraction if not the majority of the background level (if you avoid stray light properly and cut out the laser scattered light with a proper emission filter).

The problem you mention about the fluorescence not decaying completely before the next pulse comes is often referred as incomplete decays.

As you mentioned, the first thing to do to avoid it is to minimize the repetition rate of your laser (down to 20 MHz) but as you noticed GQ have long lifetime and even 50 ns is not enough to record the whole decay.

So what you can do is to account for that in the fluorescence decay analysis. And one of the best package available for that (and other things) is called FLIMfit from Imperial College London

http://www.openmicroscopy.org/site/products/partner/flimfit

If you set the repetition rate properly in the settings, the analysis will take into account the fact that there is some fluorescence from the previous decays and the data can be fitted appropriately.

This package will also let you set the background level manually. Typically you can measure the background level from the decay by averaging the first few time bins before the rise. This measurement should take into account the after-pulsing and stray light.

If only afterpulsing is present in your background (and you can check that by blocking the laser), then the package can estimate the background automatically from the total number of photons (but you need an IRF for that, refer to documentation for details).

I hope this helps.

Romain

Dear Saurabh,

if you are using Hybrid detectors, your afterpulsing probabilty should be very low (this is the main advantage of these new detectors over standard APDs). So probably that doesn't matter much.

As far as I know the B&H analysis software ist not capable of analysing incomplete decays. So you probably have to use an external package, but you can also ask the B&H guys about this. In all our questions they were very helpful.

Bye Wolfgang

I would take Romain Lane's advice and use a fitting method that can accommodate incomplete decays. The fitting software from B&H can do it if you look into the model in the settings.

Thanks Tom, for this advice. I checked it out and you are right. B&H software can hande incomplete decays.

Thanks Wolfgang

Thank you everybody! 

I am using Picoquant's Fluofit software to fit all my fluorescence decays. However, I am not sure if this software deals with incomplete decays or not.

About the background noise, afterpulsing in my detector is very low as mentioned by Wolfgang. My problem is that the QDs I use, do not decay completely within 50 ns (the lowest rep rate 20 Mhz). As Romaine said that I should use few time bins before the decay rises to set the background level for the decay, I generally do that with samples which completely decay within the 50 ns window. However, with these QDs and my laser setup, if the decay doesn't decay completely, it gets wrapped around the decay profile and can be seen in the few time bins before the decay rises. So, these time bins actually correspond to the decay rather than the noise.

I read somewhere that pulse picker can be used to decrease rep rate and thus increase the TAC range. Is that a viable option?

I was thinking about using my solvent (toluene in this case) to scatter light to the detector for the same amount of time as the decay for quantum dots to account for the background. Is that plausible? Thank you!

Saurabh

Yes, if you're sure that after-pulsing is not a problem, you can just use the solvent with the same laser intensity, same ambient light (if you left the ceiling lights on during acquisition etc.), this will account for the stray light.

Regarding pulse-picker, before going down that road, I would try and use one of the packages that can perform incomplete decays and fix the background level. If the quality of the fit is still not acceptable, it will be worth considering lower repetition rate.

You might also want to buy a 10 MHz laser.

Good luck !

Fluofit should be also able to handle the incomplete decay, but they use other terminology, something like cyclic excitation (I guess - I do not have the software in front of me right now).

Concerning pulse picker, which must be an expensive option, it is generally designed for picking output from a mode-locked laser. I guess that you use a semiconductor pulsed laser operated with modulation current -- considering its capability of switching over 20, 50,and 80Mz. In this case, the laser can possibly accept external trigger, so that you can fully control the repetition rate by simply connecting it to a common electric pulse generator, equipment which you may find in a lab next door. 

Hey Takashi,

I am using a BDL440 diode laser. To operate it at 20,50,80MHz rep rate, a laser switch box is provided, which I think is an electronic frequency control. Do you think the electric pulse generator would be compatible with the B&H Tau130 system??

Dear Saurabh- sorry Becker & Hickl BDL cannot accept fast triggering. BTW I found a B&H web page that already shows the incomplete-decay signature - http://www.becker-hickl.de/pdf/dbsimpletau130-3.pdf . Why don’t you ask B&H (and Picoquant also - since you are the customer of both TCSPC companies! ) for the best solution based on your system? In fact, as suggested by many peers above, the problem occurs frequently, and has many solutions.

Thanks for your reply Takashi!

I am actually in touch with both the companies. B&H told me that, I can work with only 20,50 or 80 MHz, and if I want to change that I would need a pulse picker or something like that. Fluofit hasn't replied yet. However, there is a procedure in the software to account for the incomplete decays, but I am still skeptical to trust those values as it is always recommended with TCSPC to eliminate issues during experiment rather than accounting for them during fitting. However, I don't see a way to resolve this problem. However, can I use  transient absorbance system to measure lifetime on the order of microseconds as the laser setup there can work with slower rep rates?

Thanks again!

Saurabh

Dear Saurabh

>I am still skeptical to trust those values

I understand what you worry. Have you talked with your supervisor? If you do not meet good solution, I would propose the following steps.

You already got a set of lifetimes deduced by fitting with arbitrary baseline removal. So, try checking if the observed baseline is consistent with those lifetimes - you need to deal with simple mathematics, but it is easy. If you meet consistency, I believe you can trust them. If not, and the baseline is much higher than expected, it would imply the additional presence of longer-lifetime components than excitation cycle. This is when you decide how seriously you attack it. I would stop there for small contribution to the total (time-integrated) intensity.

Transient absorption is the best suitable for studying much faster dynamics. If you are close to such kind of system, the laser source might be a good alternative to the B&H source.