Re: Low energy thresholds

[Matthew Shepherd <mashephe@indiana.edu>]

Hi George,

I agree with your avenues and they sound viable.  I think we first  
need to figure out what intersection we are at for sure.  I'm really  
anxious to see what you guys see in cosmic studies with the real  
segmentation, light guides, and also prototype SiPMs.  That will be a  
huge.  Then we can decide in which direction to push.

As far as physics simulation and altering the full GlueX  
reconstruction model, I think the key inputs are:

1. # PE/MeV deposited in fiber
2. dark rate
3. fraction of volume occupied by fiber
4. parametrization of energy dependent sampling fluctuations

1&2 really play together to drive the threshold, which drives  
resolution and physics, etc., and, unfortunately, I think these are  
the "softest" numbers we have.  (By the way this is assuming the the  
noise is dominated by dark rate and not by, say, pickup on the signal  
cables.)  If 1 goes down it can be compensated by pushing 2 down.   
Hopefully you will get 1 pretty solidly from your cosmic tests with  
the full module and SiPM.  We may have to do some educated guessing  
to make sure we have a reasonable estimate of 2.

3. can be obtained trivially

4. comes from standalone MC.  This may have to get a little more  
sophisticated in the case of the of a "hybrid" module. There might be  
different functions for the inside cells and outside cells.

If we know these 4 things, retuning the BCAL model in the full  
simulation is pretty straightforward.

-Matt


On Nov 1, 2007, at 1:21 PM, George Lolos wrote:

> Hi all:
>
> I believe that by now we all have been convinced that detection of  
> low energy photons is very important due to the physics interest in  
> that region of phase space for reactions of interest.   The main  
> problem here is the number of photons reaching the sensors, the  
> number of pe's we collect, and the noise we allow to leak in that  
> imposes higher thresholds.  From Elton's plots, the main culprit is  
> the limitations due to sampling fraction so whatever attempts we  
> make to improve the sensor side of the problem is guided by the  
> sampling fraction contribution to the energy resolution.  So, the  
> way I see it, we have two avenues to push the limits:
>
> 1.  Push SensL for as high a PDE and as low a DR as possible.  We  
> will do this, however, I believe that we will reach a stage where  
> we will hit the ceiling of technology and will need substantial  
> cooling of the sensors to keep high PDE and low DR.  This has  
> budgetary and technical challenges we may not wish to face.  In  
> addition, after pulsing becomes more severe with lower temperatures  
> and this imposes an additional R&D to reduce this.  What I am  
> saying is that we can only push DR down so far without cooling and  
> we should have other options.
>
> 2.  Increase the number of photons reaching the SiPM's by  
> increasing the number of SciFi's read out by them.  This, in  
> practical terms means increasing the SciFi to Pb ratio, so  
> increasing the sampling fraction.  This has two benefits by  
> improving both the pe statistics and improving the contributions  
> due to the sampling fraction fluctuations.  However, the down side  
> may be more leakage out the back and many more SciFi's than we can  
> afford or want, for that matter, if we go for, say, 3 mm instead of  
> the present 5 mm of Pb thickness.
>
> One now has to ask, what region of the BCAL is really the one that  
> needs the increased photo-statistics and higher sampling fraction?   
> This is clear, the critical region of the BCAL is that of the inner  
> layers.   Using 3 mm Pb for the inner layers will increase the  
> sampling fraction by almost a factor of 5/3 over the present one.   
> This will also increase the number of photons by almost the same  
> amount while keeping the DR constant.  The outer layers of ~ 12 cm,  
> on the other hand, have no need for such improvements and we are  
> better off to keep the Pb thickness to 5 mm to contain as much of  
> the shower energy as possible.
>
> I am proposing then that we pursue a MC study of the "hybrid" 3 mm  
> and 5 mm Pb thickness and investigate how many layers do we need  
> for the former and how many for the latter to reach an optimum  
> figure of merit.
>
> Any comments/shoot-downs/new ideas?
>
> George
>