sPHENIX HCal discussion

[Edward Kistenev <kistenev AT bnl.gov>]

Certainly.

There are 4 pictures attached. 

Fig.1. Showing the fiber exit similar to one in existing tiles (except the fact that fibers in this picture are centered with respect to the depth of 3.5mm (today they are shifted towards one side - took time to learn how to implement a more complicated groove shape - design from Uniplast, implementation in the picture local at UTFSM). Plastic is not coated. Fibers need polishing, if they are glued flat with an edge of the tile then coating around fiber exit will be removed by polishing procedure.

SiPM is 3x3 mm2 and is now facing open tile in the square region of the tile edge of a slightly bigger size around fibers. Light generated inside of plastic (420nm) which was supposed to be either reemitted in the fiber or simply lost is now exiting the tile and free to reach the SiPM. Attenuation length for the light of ~420nm if few cm so the light collection from the semicircle of ~that radius looks dramatically enhanced (see second picture copied from Ron’s presentation of yesterday).

Fig. 2 Tile response to LED point by point exposure. Dramatic hot spot around fiber exits. One reason is what explained above (can be alleviated by insertion of a bulk light blend between tile and SiPM), second - cladding light. In ECal epoxy is black so it kills cladding light. In HCal plastic is transparent, epoxy is optically transparent too. Cladding light either conserved in the cladding (not much because we use single clad fibers) or travels along the fiber. Same fate as with the bulk light. Blend will help but only partially;

Fig. 3 The solution proposed and tested at UTFSM. Bulk light stopper made of black plastic is glued into machined recess in the plastic. It has fixed positioning tightly spaced holes for fiber exiting. Fibers are glued into the spacer. Spacer needs to be designed as one part of the SiPM holder, second half must be attachable without any screws if we do not want to start building vacuum cleaners to extract these screws fro the gaps;

Fig. 4 Effect of the bulk light stopper on the hot spot. Tile response to LED (420nm) exposure in two configurations - without and with light stopper (everyone in HCal seen this picture) as function of the distance between exit point and excitation point (narrow band around fiber excluded - we are doing some work to reduce fiber related non uniformity too but it require cosmic data with good position resolution - probably at a later date). Light stopper entirely solves the hot spot problem.

Now few words about “tail generator” comment. The hot area is ~6cm in diameter or +-3cm  from fiber exit. Our tiles in HInner end at a depth in calorimeter just around hadronic shower maximum. They are only 10cm wide. X3 non uniformity introduced by selectively collecting a lot of stray light around shower max (10% of tile area are affected) can’t be ignored now when we understand why it is happening.

Edward

Edward Kistenev, PhD

PHENIX Physicist

On Jan 13, 2016, at 2:29 PM, EdwardOBrien <eobrien AT bnl.gov> wrote:

Hi Edward,
 Could you provide more detail about this comment in your
e-mail? I don't understand your point. Thanks.

Ed


On 1/12/2016 1:35 PM, Edward Kistenev wrote:

Finally - I am very worried by discussion about what is below the SiPM
coupler. I wrote so many mails about the need to have the bulk light
suppression blend between tile and SiPM but it looks as if all of it was
totally ignored. If it is not done - out Inner section may become a
tail-generator (not easy to simulate), it is less important in the Outer
HCal where fiber-SiPM coupling area is far from the shower maximum.