sPHENIX HCal discussion

[John Haggerty <haggerty AT bnl.gov>]

I corresponded with Edward quite a bit yesterday about stressing the 
scintillator, and he's busy preparing a talk for the conference this 
week, so I'll try and summarize what he and I arrived at about how much 
we could safely stress the scintillator tiles in the HCAL.  (I'll write 
a second description of why he thinks they are thicker than we 
specified, but I might do that later today.)

I found a paper which studied this a little in extruded polystyrene 
scintillator, and I attached a figure showing the loss of light over a 
period where a tile was stressed at about 10% of breaking over three 
months.  The paper is here:

> http://deepblue.lib.umich.edu/handle/2027.42/31503

The PHENIX EMCAL was assembled under stress; Edward says: "PHENIX PbSc 
was assembled with stacking pressure ~5-6kg/cm2 - no signs of crazing 
after nearly 20 years (checked today).
Extruded plastic has much better load bearing capability. It has much 
lower density of frozen high elasticity deformations which are the 
crazing source under pressure. 15-20kg/cm2 should be OK."  (20 kg/cm^2 
is 280 PSI.)

He also said, "it is basically impossible to force this plastic 
(especially with 0.1mm foam like coating on both sides) to craze under 
any practical load (I kind of thought of suggesting to load the tiles 
while assembling the prototype - should be no problem at all). I would 
suggest to make a small test - take 10x10 cm2 piece of scintillator and 
place it into the press with pressure sensor. Do it in steps - bring and 
release and inspect then repeat at increased pressure. I may try to do 
it here at USM.
My experience is with injection molded scintillator used in all kinds of 
shashlik’s. We always compressed the structure  to the state when 
10x10x50cm2 module will hold its weight cantilevered without visible 
bending. I inspected number of layears old modules from GEPARD - no 
crazing but ~1% a year light loss - ask me why. I’ll check with Uniplast 
if they kept the record of pressure cycling which they did."

Also, it seems to me that crazing is usually degradation of the surface 
finish, reducing transmission from total internal reflection, and the 
surface of our tiles is etched, so it only diffusely reflects anyway, so 
probably it doesn't make it much different.  I suppose as the crazing 
extends a finite depth into the volume, we could reduce transmission. 
We could do a test ourselves if we have a tile (and a control tile) 
where we measure the light output over the next few months, where one is 
stressed and one is not.  I suspect doing a test over a short period 
will show nothing, but that's what experiments are for.

The outer HCAL prototype has 4 tiles with an area of about 6829 cm^2 
according to my arithmetic (someone could check that), so if the entire 
400 kg plate was supported by the tiles, the pressure is pretty small. 
Of course, effectively clamping them down could increase the stress to 
be arbitrarily high, and there are high spots and low points.

So I'd conclude from this that if we can get the tiles into the gaps, 
they will probably be ok.  We would not take tiles out in the field 
anyway, although it may make it necessary to either unstack it at 
Fermilab or bring it back after the beam test to install the high 
rapidity tiles.

For the inner HCAL, I think it means we can stack it up like a Dagwood 
sandwich.

--
John Haggerty
email: haggerty AT bnl.gov
cell: 631 741 3358

Attachment: Screen_Shot_2016-01-04_at_9.32.15_PM.png
Description: PNG image