Muon Work at CalDet

Calibration

Introduction

• Selected beam muons have been used to obtain a set of values to calibrate the detector
• Muon energy deposits are histogrammed and fitted, per strip end.
• The peaks, obtained from the fits, can then be used to characterize the response of the detector to the passage of a MIP

• Select tracks that are at least 8 planes long in both views, that are within a 10 degree cone (wrt beamline) of the most upstream track hit in the detector
• Ensure that no planes have more than 2 hits in 5 consecutive strips (shower veto)
• Use only hits on a track
• Correct ADC values using LI drift point information to some fixed point in time - two types of correction attempted:

    quadratic parameterisation:
           quad_corrected_adc = QuadCorrected(adc_t2)*(QuadCorrected(drift_adc)/drift_pin)_t1/(QuadCorrected(drift_adc)/drift_pin)_t2

   Quadratic terms for each strip-end come from Phil.

• Histogram ADC values per strip end and fit
• Fitting function is an approximation to a Landau convolved with a Poisson and returns a value for the peak of the underlying Landau distribution

Work done

I want to compare MIP values on a strip-end by strip-end basis for a range of different runs to ensure that the above procedure is valid.
I've been looking at:

• Runs of the same energy separated in time
• Runs taken before and after HV change
• Long runs that span a whole day(night), to investigate temperature effects

Runs before and after the HV change

Looking at two 3 GeV runs taken 2 weeks apart and straddling the HV change, I examined the ability of the drift calibration to correct for the ~20% gain change. Using the driftpoints, I corrected the ADC value back to a time before the HV change and then obtained the MIP values in terms of photoelectrons using Ryan's old gains.

The plots below show the fractional difference between the sets of MIP values. The upper plot is without the stability correction and lower plot is with:

The stability correction narrows the distribution significantly as we would hope. However, although the mean value changes, it is still not consistent with zero after the correction i.e. the MIP values have changed by ~1.2% between the two runs.

Encouraging Aside:

I also looked at two 2 GeV runs separated by ~24h both of which were before the HV change. The plot below is the fractional difference in MIP value between the earlier and later runs:
 

The mean of the distribution is consistent with zero and the statistical variation on the fit results is ~ 2%
On short timescales therefore, we can see there are no significant shifts in the MIP values.

Notice that there are a surplus of entries in the wings of the Gaussian; these can be understood in a 2D histogram of strip vs plane weighted by fractional difference. Only the central strips contain enough beam muon energy deposits to be fitted, however at the boundary between the fitted and not fitted strips there are a number of poor fits, and it is these entries that can be seen in the gaussian wings in the plot above.
 

Changes in Strip Response

The sets of data used in the plot above were the first ~2 hours of two long runs that started in the evening and ran overnight until morning. The MIP values were found for each ~2 hour period to see whether there was a trend that would point to a temperature dependance. The scintillator light output is expected to vary with temperature; approximately 0.3% per degree.

The graphs below show the variation in the average MIP value in terms of NPEs vs time (click on the plot for an eps version):

(The blue dots are the run before the HV change and red after)

(Fractional Difference between blue and red runs on a strip-end by strip-end basis vs time)

• There is a ~1% drop in the average values between the red and blue, as we have already seen.
• The second blue point appears to be anomolous, however, the file representing the 3rd point was unusable (empty). (Why is 2nd point anomolous? - scatter plot of strip vs plane weighted by frac diff between 1st and 2nd blue points )
• A time dependance may be there, but most of the points are within each others error bars.

(I have left out the 2nd blue point which shows a ~ -1% shift from the first point so that the rest could be seen more clearly)

In this plot, significant changes in the MIP values can be seen for the blue run, especially in the morning.

(The blue run was dated 23rd Sept, and the red 7th Oct.)

I have also looked at all other +/- 3 GeV runs and plotted the % shifts vs time from the first run in the series:

Work using Phil's Quadratic Corrections

Only some channels are available at present and only for runs after the HV change.

Comparing runs 11285 and11285++, (using the same sample of channels):

Using the simple correction:

Using the quadratic:

The width is ~10% narrower using the quadratic corrections. The mean is further from zero, (in the same direction).
 

Looking at Cosmics

Started looking for cosmics out to 70degrees wrt beamline.
I histogram the path length corrected ADCs and fit as usual.
 
 

Remarks and Conclusions

• Over short periods, the stability of the MIP values has been shown and the fit results have a statistical error of ~2%
• With a longer period/extreme gain change, systematic shifts of ~1-2% have been observed.
• Smaller shifts are also observed on short timescales perhaps due to changing environmental conditions, however, these are not sufficient to explain the 1% shift alone.
• A more sophisticated method to correct for the gain variations has been attempted using a quadratic parameterisation from Phil.

Properties of Muons

The beam makes it possible to investigate certain properties of muons at CalDet, including:

• Range as a function of energy
• dE/dx as a function of energy
• Differences between +ves and -ves

Looking at some last plane hit distributions:

2GeV

1.4GeV

Taking muons that stop in the fitted peaks a clean sample of beam energy muons can be obtained.
Measurements of the range and dE/dx have been made at several beam energies, the relevant plots are below (eps versions):

 Range plot
 green dE/dx plot
 clear dE/dx plot
 

Another measurement that has been made is the change in dedx as a function of plane for 2 GeV muons. To do this I look at the average MIP value per plane for 2 GeV muons ( dedx for 2GeV muons ), and then remove scintillator variations by dividing by the 3 GeV numbers ( dedx for 3GeV muons ):
 

The increase in dedx for the stopping muons can clearly be seen, (measured stopping plane for 2 GeV muons is ~ plane 55).

The plot below is the same thing, except that now, I have demanded that the muon stops in planes 54, 55 or 56. The red line is MC generated by Leo : (n.b. the black line is wrong, all entries should be shifted one bin to the right)

The main difference before plane 56 is that the MC rises more quickly. This is because I am cutting on stopping plane whereas Leo is not.