Question on SPhenoMSSM in high-scale models with parameters more general than mSUGRA

[BjHerrmann]

Hi Florian,

I have run a few tests with SPhenoMSSM (generated from SARAH 4.12.3 including the patch that you posted in the other thread -- thanks!) and based on SPheno 4.0.3. Since my issue is not related to the other one, I open a new thread for this.

So now basically SPhenoMSSM runs fine when using the provided LesHouches.in.MSSM. I have only modified some input values to the parameter region of our interest, and taken out some calculation options such as decays, low-energy stuff, etc. for the moment. So basically I ask SPhenoMSSM to compute simply the mass spectrum from the high-scale input parameters. The model that we want to study ultimately is the MSSM with SU(5)-like boundary conditions at the GUT scale, so basically an "mSUGRA" variant from the SLHA point of view with certain relations between certain soft mass blocks. We also want to include generation mixing in the squark and slepton sectors, so we need fully general 6x6 mass matrices at the high scale.

The issue I have is that when running a given point (without caring about the SU(5) stuff at that stage), and I use the exact format from the LesHouches.in.MSSM, it works and I obtain the mass spectrum from SPhenoMSSM.

However, when I switch to giving in addition to the m0 and A0 parameters in MINPAR, one or more full mass blocks (e.g. MSQ2IN etc) with the same diagonal value as m0 (and without offdiagonal values for the moment) together with the trilinear blocks (with a value corresponding roughly to the A0 value), the code stops saying:

~/Projects/GUT_NMFV/SU5xA4> ./SPhenoMSSM Test_MSSM.in Test_MSSM.spc
NaN appearing in GToParameters213
At position 60
Note: The following floating-point exceptions are signalling: IEEE_INVALID_FLAG IEEE_UNDERFLOW_FLAG IEEE_DENORMAL
STOP Subroutine TerminateProgram

My impression is that something goes wrong with the reading of the MSQ2IN, MSU2IN, MSD2IN, TUIN, TDIN, TEIN blocks...? Do you have an idea what happens and how to solve this issue? Or am I missing something in the input?

I join the two input files of (in my opinion) the same point: Test_MSSM_1.in respecting the format of LesHouches.in.MSSM, and Test_MSSM_2.in where in addition in provide the soft mass and trilinear blocks (with the same values, but in fact the exact values do not seem to matter regarding this question).

Viele Grüsse,
Björn

[FStaub]

Hi Bjoern,

the reason for the problem is

Code: Select all

 51 1              # Write Output in CKM basis 

By doing that, the matrices are assumed to be in the CKM basis. When rotating them to the ew basis, NaN's were showing up. It seems that there is a different convention for the Yukawas in the function Switch_from_superCKM which was taken over from SPheno.

It should be sufficient to change in Boundaries_MSSM.f90

Code: Select all

!---------------------------------------------------------
  ! CKM matrix at Q, shifting phases according to PDG form
  !---------------------------------------------------------
  CKM_Q =  Matmul(uU_L, Transpose(Conjg(ud_L)) )

to

Code: Select all

  CKM_Q =  Matmul(uU_R, Transpose(Conjg(ud_R)) )

Cheers,
Florian

PS: The routine Switch_to_superCKM looks fine.

[BjHerrmann]

Hi,
and thanks for the fix! I have implemented this and it seems to run fine.

Just one thing about the routine that you mentioned. The part of the code that you mentioned above now looks like this:

Code: Select all

  
  !---------------------------------------------------------
  ! CKM matrix at Q, shifting phases according to PDG form
  !---------------------------------------------------------
  ! CKM_Q =  Matmul(uU_L, Transpose(Conjg(ud_L)) )
  ! Correction as indicated by F. Staub in SARAH Forum 
  ! (see here: http://stauby.de/sarah_userforum/viewtopic.php?f=4&p=1640#p1640)
  CKM_Q =  Matmul(uU_R, Transpose(Conjg(ud_R)) )
  uD_L(1,:) = uD_L(1,:) / Conjg(CKM_Q(1,1)) * Abs(CKM_Q(1,1))
  uD_L(2,:) = uD_L(2,:) / Conjg(CKM_Q(1,2)) * Abs(CKM_Q(1,2))
  uU_L(2,:) = uU_L(2,:) / CKM_Q(2,3) * Abs(CKM_Q(2,3))
  uU_L(3,:) = uU_L(3,:) / CKM_Q(3,3) * Abs(CKM_Q(3,3))
  !-------------------------------------------------------------------
  ! also the right quark must be multiplied with the conjugate phase
  ! as otherwise the masses get complex
  !-------------------------------------------------------------------
  uD_R(1,:) = uD_R(1,:) / CKM_Q(1,1) * Abs(CKM_Q(1,1))
  uD_R(2,:) = uD_R(2,:) / CKM_Q(1,2) * Abs(CKM_Q(1,2))
  uU_R(2,:) = uU_R(2,:) / Conjg(CKM_Q(2,3)) * Abs(CKM_Q(2,3))
  uU_R(3,:) = uU_R(3,:) / Conjg(CKM_Q(3,3)) * Abs(CKM_Q(3,3))
  CKM_Q =  Matmul(uU_L, Transpose(Conjg(ud_L)) )

So in the upper part there is the multiplication that concerns the right sector, and below the one corresponding to the left sector. However, the comment in the middle seems to suggest the opposite... maybe this is the difference in convention between SPheno and SPhenoMSSM that you mentioned, but I ask to be sure

Cheers,
Björn

[FStaub]

Yes, the Yukawas in SPheno are transposed.