Photon detection efficiency low

Cheng Li

Hardware · 11 months ago

One of our colleagues characterized the overall photon detection efficiency of the Timepix+intensifier system to be 8%. See

Vidyapin, V., Zhang, Y., England, D. et al. Characterisation of a single photon event camera for quantum imaging. Sci Rep 13, 1009 (2023).

I understand that even though the photocathode inside the intensifier has a much higher quantum efficiency, other factors could be at play. Can ASI comment on the low overall efficiency of the Timepix system and is there a way to improve this?

4 Comments | 790 Views

Andrei Nomerotski · 11 months ago

This is a good paper that provides a very useful methodology to measure the camera efficiency indeed. ASI may have other comments but I'd say that the measured 8% included two things: photocathode quantum efficiency (~20% at 810nm in Photonis graphs) and camera efficiency due to the Timepix3 threshold (estimating it as 0.08/0.20=0.4=40%). There is not much we can do about the photocathode but the camera efficiency due to the thresholding will depend on the intensifier gain (assuming here that Tpx3 is well equalized and threshold is minimal possible). The gain (and hence TOT) distribution is very wide due to the nature of MCP amplification so there will be always small amplitudes, which would be below the threshold. however if the gain is large enough then the fraction of those events will be small. I am pasting an example of TOT distribution for a high gain hi-QE-red intensifier where the thresholding effect are clearly visible and I don't think the thresholding inefficiency here would be more than 15% (hence total efficiency 0.2*(1-0.15)=0.17=17%). It would be interesting to measure the total efficiency for this intensifier using the same technique. It would be also interesting to see the corresponding TOT distribution of the dataset used for the paper.

so my bottom line would be that 8% looks a bit low and probably can be improved by factor of 2 with high gain intensifiers.

Yingwen Zhang · 11 months ago

We have actually gotten a high gain intensifier not long after we published the Sci Rep paper. We have found the gain is a little too high in which it generated 3-4 times more raw data in the camera and has significantly slowdown the data processing speed for clustering and centroiding. As a result we are now running it with a higher camera threshold to collect less events. The current quantum efficiency we got after increasing the camera threshold is still around 8-10%. I remember the efficiency I got when using a lower threshold is still similar but I will need to double check on this, it was a while back. I will do a proper analysis to check on this in the coming weeks so stay tuned.

Andrei Nomerotski · 11 months ago

that would be great, thanks, Yingwen. also please plot 'max TOT pixel' (TOT in pixel with max TOT in the cluster) as above for those cases, those plots are very indicative of the threshold effects.

Yingwen Zhang · 10 months ago

So I have made the efficiency measurement as done in Sci Rep 13, 1009 (2023) using the new high-gain intensifier. There is a slight increase in the quantum efficiency when setting a lower camera threshold, but not
by much. With Vthreshold_coarse=4 I measured a camera efficiency of around 7-8% (accounting for the overall transmission efficiency of all the filters and optics used is around 80%), and at Vthreshold_coarse=6 the efficiency is about 8-9%, so around 1% improvement in efficiency. Setting the threshold values any higher (lower in physical threshold) the camera will start saturating from the high intensifier gain.

Ultimately, I think at a certain point having higher gain in the intensifier would not help much in improving the overall efficiency. Given a fixed photocathode efficiency, having a gain of 200k or 1000k from the MCP should not matter much given that one only needs a few thousand photons to cross the camera threshold.

All

All questions and topics

Hardware

(5)

All topics on hardware

Firmware

Firmware-related questions

Software

(4)

Software-related questions

Application

Application-related questions

Other

(2)