sPHENIX HCal discussion

[John Haggerty <haggerty AT bnl.gov>]

The question of whether we need to make the IHCAL performance prototype 
out of stainless steel came up at yesterday's meeting again.  It's 
harder to work with mechanically, and more expensive, and I don't think 
there will be any discernible difference, and here's my analysis of that 
question.  Of course, I could be wrong, so more eyes on the question 
would be good.

The non-magnetic stainless steel we used in the PHENIX absorber on the 
central magnet came with an assay that had these components:

>      Z       A     rho       X       %
> C    6  12.011   2.266  42.698   0.050
> Mn  25  54.938   7.473  14.640   0.960
> P   15  30.974   1.820  21.205   0.025
> S   16  32.060   2.086  19.493   0.000
> Si  14  28.086   2.329  21.823   0.560
> Ni  28  58.700   8.907  12.679  19.380
> Cr  24  51.996   7.194  14.944  24.450
> Mo  42  95.940  10.222   9.801   0.250
> Co  27  58.993   8.800  13.631   0.150
> Cu  29  63.546   8.933  12.862   0.280
> N    7  14.007   0.001  37.989   0.055
> Fe  26  55.847   7.873  13.839  53.840
>                  7.874  13.873

and if you cruise the web, you'll see that's typical for "310 stainless 
steel."  (The above table was from when we got it and I calculated the 
radiation length of the mixture for fun; if I did the weighting right, 
the radiation length of the SS310, the 13.873 g/cm^2, is pretty close to 
elemental Fe, which is >99.5% of a carbon steel like 1006 which we'll 
use in the flux return; the density is pretty much the same, too.)  It's 
not mysterious, the other big components, Ni and Cr, surround Fe on the 
periodic table, and have quite similar properties.

I thought I better check a little further, so I used my super-idealized 
model of the IHCAL based on a GEANT example to look at the difference; I 
made a nominal 1 interaction length calorimeter out of 4 Fe plates 
(that's the "4 crossing") 42 mm thick (the interaction length of Fe is 
167.7 mm), and shot some 10GeV pi+'s at it and used matScan to measure 
the interaction length.

> The calorimeter is 4 layers of: [ 42mm of G4_Fe + 7mm of G4_POLYSTYRENE ]
>  mean Energy in Absorber : 1.74625 GeV +- 1.86963 GeV
>  mean Energy in Gap      : 56.5529 MeV +- 61.824 MeV
>  mean trackLength in Absorber : 1.1057 m   +- 1.19615 m
>  mean trackLength in Gap      : 19.1849 cm  +- 19.3933 cm
>
> and the matScan agrees with my arithmetic:
>
>
> /control/matScan/singleMeasure 0 0
>
>          Theta(deg)    Phi(deg)  Length(mm)          x0     lambda0
>                   0           0      1017.6     9.62684     1.03024

The material G4_STAINLESS-STEEL packaged with GEANT isn't 310, it's 
about 74% Fe, so we should cook up the material correctly, but just to 
get a quick idea, here are the statistics from running 10 GeV pions 
through that, and the matScan:

> The calorimeter is 4 layers of: [ 42mm of G4_STAINLESS-STEEL + 7mm of 
> G4_POLYSTYRENE ]
>  mean Energy in Absorber : 1.7336 GeV +- 1.76738 GeV
>  mean Energy in Gap      : 59.5918 MeV +- 65.2185 MeV
>  mean trackLength in Absorber : 1.03557 m   +- 1.08142 m
>  mean trackLength in Gap      : 18.3985 cm  +- 18.4141 cm
>
>          Theta(deg)    Phi(deg)  Length(mm)          x0     lambda0
>                   0           0      1017.6     9.73369     1.04876

So I think it's ok to make the prototype out of pretty much any steel as 
far as its properties as an absorber, as far as I can see.  Probably not 
so true for the mechanical prototype, where we may have to learn to deal 
with a material that is rather difficult to work with.

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

Attachment: Screen Shot 2015-06-19 at 10.42.43 AM.png
Description: PNG image