betaFeedback

WorksPhysics and Astronomy

The Position Resolution of a Charged Particle Hodoscope Consisting of Triangular Scintillator Counters

Abstract

Hodoscopes consisting of triangular counters with embedded wavelength shifting fibers are becoming more popular as they are relatively inexpensive, are easily fabricated, need no cooling systems, employ low-cost photodetectors, require only modest voltages, and perhaps most importantly, have good position resolution. Through interpolation of the light yields between adjacent counters, position resolutions that are typically an order of magnitude better than the counter separation can be obtained. Here, after a brief explanation of how the interpolation is done to improve the position resolution, we derive an approximate expressions for the uncertainty in the interpolated xx and yy positions as a function of the impact position of the charged particle.

Introduction

Hodoscopes consisting of arrays of triangular scintillator counters are becoming increasingly popular due to their modest cost, ease of fabrication, ease of operation, and their ability to operate in harsh environments. More importantly, position resolutions better than one-tenth of the distance between the triangle centers have been achieved (see Refs. [1, 2, 3, 4, 5, 6, 7], for example). We first derive the simple expressions that give the xx and yy coordinates of a charged particle going through a plane of triangular scintillator counters by interpolation of the charged particle path lengths through the adjacent counters using the relative light yields in each counter. We then derive an approximate expression of the uncertainties in the positions using Poisson estimates for the light yield distributions.

Determining the xx and yy Coordinates

We derive below the xx and yy positions found by interpolation of the light yields in the adjacent counters shown in Figure 1 and Figure 2 below. The results for the interpolated xx- and yy-coordinates are:

The counter geometry showing a charged track traversing the inverted counter to the left of the center counter

Figure 1. The counter geometry showing a charged track (in blue) traversing the inverted counter to the left of the center (upright) counter.

The counter geometry showing a charged track in blue traversing the inverted counter to the right of the center upright counter

Figure 2. The counter geometry showing a charged track (in blue) traversing the inverted counter to the right of the center (upright) counter.

x±=±w4(E1−E2E1+E2),(1)x_{\pm} = \pm\frac{w}{4}\left(\frac{E_1-E_2}{E_1+E_2}\right), \tag*{(1)}
y=h2(E1−E2E1+E2),(2)y = \frac{h}{2}\left(\frac{E_1-E_2}{E_1+E_2}\right), \tag*{(2)}

where the + sign in Eq. (1) refers to the case shown in Figure 1 where the leftmost triangle is inverted and the − sign in Eq. (1) refers to the case shown in Figure 2 where the leftmost triangle is upright. In both of the above equations ww is the triangle base width, hh is the triangle height, and E1E_1 and E1E_1 are the charged particle path lengths shown in Figure 1 and Figure 2.

To experimentally determine the xx and yy positions, we replace the path lengths E1E_1 and E2E_2 with their photoelectron yield proxies, N1N_1 and N2N_2, to get:

x±≈±w4(N1−N2N1+N2),(3)x_{\pm} \approx\pm\frac{w}{4}\left(\frac{N_1-N_2}{N_1+N_2}\right), \tag*{(3)}
y≈h2(N1−N2N1+N2),(4)y \approx\frac{h}{2}\left(\frac{N_1-N_2}{N_1+N_2}\right), \tag*{(4)}

where the approximate sign indicates the fact that N1N_1 and N2N_2 are imperfect measures of path lengths E1E_1 and E2E_2.

Case 1: Charged Particle Traverses the Junction Between the Left and Center Counters

We derive Eq. (1) and Eq. (2) referring to the geometry shown in Figure 1 below. Each counter is an isosceles triangle of width ww and height hh. We assume that the charged particle’s incident angle is such that it only traverses no more than two counters, and we assume that the two counters have a perfect isosceles triangular shape. The origin of the coordinate system is at the midpoint of the interface between the two counters: M+M_+ in Figure 1. The coordinates of the point at which a charged particle passes the boundary between two counters is P+P_+ in Figure 1. The charged particle’s path (shown in blue in Figure 1), has a length E2E_2 in the left (inverted) counter and E1E_1 in the center (upright) counter.

We wish to find the best estimate of the coordinates of the point P+(x,y)P_+(x,y), relative to the origin of the coordinate system M+M_+ in Figure 1. We have from the geometry of similar triangles:

y1y2=E1E2ory2=E2E1y1.\frac{y_1}{y_2} = \frac{E_1}{E_2} \quad\text{or}\quad y_2 = \frac{E_2}{E_1}y_1.

In addition, y1+y2y_1+y_2 equals the height of the counter, or:

y2=h−y1.y_2 = h-y_1.

Setting these two above expressions for y2y_2 equal to one another, we have:

y1=hE1E1+E2.y_1 = h\frac{E_1}{E_1+E_2}.

We can now solve for yy:

y=y1−h2=hE1E1+E2−h2,y = y_1 - \frac{h}{2} = h\frac{E_1}{E_1+E_2} - \frac{h}{2},
y=h2(E1−E2E1+E2).(5)y = \frac{h}{2}\left(\frac{E_1-E_2}{E_1+E_2}\right). \tag*{(5)}

Does this equation make sense? If the path lengths E1E_1 and E2E_2 are equal, then y=0y=0, which is correct. If E1=0E_1=0 then y=−h/2y=-h/2, and if E2=0E_2=0, then y=+h/2y=+h/2, both of which are correct.

Once the yy-coordinate is known (y1y_1), the xx-coordinate can be found using the slope:

S=ΔyΔx=hw/2=2hw.S = \frac{\Delta y}{\Delta x} = \frac{h}{w/2} = \frac{2h}{w}.

The xx-coordinate is then:

x=yS=w2hh2(E1−E2E1+E2),x = \frac{y}{S} = \frac{w}{2h}\frac{h}{2}\left(\frac{E_1-E_2}{E_1+E_2}\right),
x=w4(E1−E2E1+E2).(6)x = \frac{w}{4}\left(\frac{E_1-E_2}{E_1+E_2}\right). \tag*{(6)}

Does this equation make sense? If the path lengths E1E_1 and E2E_2 are equal, then x=0x=0, which is correct. If E1=0E_1=0 then x=−w/4x=-w/4, and if E2=0E_2=0, then x=+w/4x=+w/4, both of which are correct.

Case 2: The Charged Particle Traverses the Junction Between the Center and Right Counters

This derivation proceeds exactly as that shown above for Case 1, and hence will not be repeated. The only difference is that the in going from the yy coordinate to the xx coordinate a minus sign appears as the slope of the junction between the two counters, as shown in Figure 2 is negative.

We have for the yy position in Figure 2:

y=h2(E1−E2E1+E2),(7)y = \frac{h}{2}\left(\frac{E_1-E_2}{E_1+E_2}\right), \tag*{(7)}

exactly the same equation as found in Case 1 above.

Does this equation make sense? If the path lengths E1E_1 and E2E_2 are equal, then y=0y=0, which is correct. If E1=0E_1=0 then y=−h/2y=-h/2, and if E2=0E_2=0, then y=+h/2y=+h/2, both of which are correct.

The xx-coordinate is:

x=−w4(E1−E2E1+E2).(8)x=-\frac{w}{4}\left(\frac{E_1-E_2}{E_1+E_2}\right). \tag*{(8)}

The minus sign in front of Eq. (8) is the only difference between this value for xx and that from the previous section.

Does this equation make sense? If the path lengths E1E_1 and E2E_2 are equal, then x=0x=0, which is correct. If E1=0E_1=0, then x=+w/4x=+w/4, and if E2=0E_2=0, then x=−w/4x=-w/4, both of which are correct.

Angular Determination

We emphasize that the derivations of Eq. 1 and Eq. 2 are independent of the charged particle incident angle, as long as it does not traverse more than two counters. And, indeed, the incident angle cannot be determined by this method. Figure 3 shows two charged-particle tracks, T1T_1 and T2T_2 going through the left and right counter junctions at exactly the same xx and yy points, but at widely different angles. The point of this is to show that one cannot tell the angle from only one plane of triangular counters. Another plane is needed.

Two pairs of charged particles going through adjacent triangular counters

Figure 3. Two pairs of charged particles going through adjacent triangular counters. Each track of each pair has the same xx and yy coordinates, although their incident angles are quite different.

The Interpolated Position Uncertainty

We replace the distances traversed by the charged particles in the counters, E1E_1 and E2E_2 by their photoelectron yield proxies, N1N_1 and N2N_2, respectively in Eq. (1). We have then

x=±w4(N1−N2N1+N2).x = \pm\frac{w}{4}\left(\frac{N_1-N_2}{N_1+N_2}\right).

The uncertainty in xx as a function of the uncertainty in N1N_1 and N2N_2:

∂x∂N1=±w4(N1+N2)−(N1−N2)(N1+N2)2=±w2N2(N1+N2)2,(9)\frac{\partial x}{\partial N_1} = \pm\frac{w}{4}\frac{(N_1+N_2)-(N_1-N_2)}{(N_1+N_2)^2} = \pm\frac{w}{2}\frac{N_2}{(N_1+N_2)^2}, \tag*{(9)}
∂x∂N2=±w4−(N1+N2)−(N1−N2)(N1+N2)2=±w2−N1(N1+N2)2.(10)\frac{\partial x}{\partial N_2} = \pm\frac{w}{4}\frac{-(N_1+N_2)-(N_1-N_2)}{(N_1+N_2)^2} = \pm\frac{w}{2}\frac{-N_1}{(N_1+N_2)^2}. \tag*{(10)}

We add the two uncertainties in quadrature to get the total uncertainty in xx:

(Δx)2=(∂x∂N1)2(ΔN1)2+(∂x∂N2)2(ΔN2)2.(11)(\Delta x)^2 = \left(\frac{\partial x}{\partial N_1}\right)^2(\Delta N_1)^2+\left(\frac{\partial x}{\partial N_2}\right)^2(\Delta N_2)^2. \tag*{(11)}

We wish to estimate the uncertainty in parameters N1N_1 and N2N_2. For charged particles going though a thin counter, this is given by a Landau distribution [8]. However, to simplify our derivation we assume the light yield follows a Poisson distribution in order to allow a simple estimation of their uncertainties:

(ΔN1)2=σ12=N1and(ΔN2)2=σ22=N2.(\Delta N_1)^2 = \sigma_1^2 = N_1 \quad\text{and} \quad(\Delta N_2)^2 = \sigma_2^2 = N_2.

Using these two values for σ1\sigma_1 and σ2\sigma_2 we find, using Eq. (11) in conjunction with Eqs. (9) and (10):

σx2=(w2)2[N22(N1+N2)4N1+N12(N1+N2)4N2],=(w2)2N1N2(N1+N2)4(N1+N2),=(w2)2N1N2(N1+N2)3.\begin{aligned} \sigma_x^2 &= \left(\frac{w}{2}\right)^2\left[\frac{N_2^2}{(N_1+N_2)^4}N_1+\frac{N_1^2}{(N_1+N_2)^4}N_2\right],\\ &= \left(\frac{w}{2}\right)^2\frac{N_1N_2}{(N_1+N_2)^4}(N_1+N_2),\\ &= \left(\frac{w}{2}\right)^2\frac{N_1N_2}{(N_1+N_2)^3}. \end{aligned}

Finally,

σx=(w2)N1N2(N1+N2)3.(12)\sigma_x = \left(\frac{w}{2}\right) \sqrt{\frac{N_1N_2}{(N_1+N_2)^3}}. \tag*{(12)}

The yy position uncertainty, σy\sigma_y, is easily determined to be:

σy=hN1N2(N1+N2)3.(13)\sigma_y = h \sqrt{\frac{N_1N_2}{(N_1+N_2)^3}}. \tag*{(13)}

Note that Eq. (12) and Eq. (13) are true for both Case 1 and Case 2 described above. Do these equations make sense? If N1N_1 or N2N_2 is 0 then the charged particle is going through the apex of one of the counters and σx\sigma_x or σy\sigma_y is zero.

At what value of xx is the uncertainty maximal? To find out we set the derivative of Eq. (12) equal to zero. We have, setting N2=NT−N1N_2 = N_T - N_1, where NT=N1+N2N_T = N_1 + N_2:

∂σx∂N1=(w2)∂∂N1(N1(NT−N1)NT3/2),=(w2)12NT3/2(NT−N1)−N1N1(NT−N1),∂σx∂N1=w4NT3/2N2−N1N1N2.(14)\begin{aligned} \frac{\partial\sigma_x}{\partial N_1} &= \left(\frac{w}{2}\right) \frac{\partial}{\partial N_1}\left(\frac{\sqrt{N_1(N_T-N_1)}}{N_T^{3/2}}\right), \\ &= \left(\frac{w}{2}\right) \frac{1}{2N_T^{3/2}} \frac{(N_T-N_1)-N_1}{\sqrt{N_1(N_T-N_1)}}, \\ \frac{\partial\sigma_x}{\partial N_1} &= \frac{w}{4N_T^{3/2}} \frac{N_2-N_1}{\sqrt{N_1N_2}}. \tag*{(14)} \end{aligned}

We also have a similar expression for the derivative of the yy position uncertainty:

∂σy∂N1=h2NT3/2N2−N1N1N2.(15)\frac{\partial\sigma_y}{\partial N_1} = \frac{h}{2N_T^{3/2}} \frac{N_2-N_1}{\sqrt{N_1N_2}}. \tag*{(15)}

Both equations are zero when N1=N2N_1 = N_2, which happens when the charged particle crosses midway up the junction between the two counters where the path lengths through the adjacent counters are equal. If we let N1=N2=12NTN_1 = N_2 = \frac{1}{2}N_T, then Eq. (12) reads:

σx max=(w2)14NTNT(NT)3=(w4)1NT.(16)\sigma_{x\ \mathrm{max}} = \left(\frac{w}{2}\right) \sqrt{\frac{\frac{1}{4}N_TN_T}{(N_T)^3}} = \left(\frac{w}{4}\right) \sqrt{\frac{1}{N_T}}. \tag*{(16)}

In Figure 4 the xx-position uncertainty is plotted as a function of the xx position in units of the base width, ww, assuming a photoelectron yield of 100 PE/hh (a light yield achieved by Ref. [9]). Figure 5 shows the resolution for a counter of 40 ×\times 20 mm2^2 (w×hw \times h) and the same PE yield. The greatest uncertainty occurs when the charged particle crosses halfway up the junction between adjacent counters, and has a value of σx=0.025⋅w\sigma_x = 0.025 \cdot w for a photoelectron yield of 5 PE/mm. The position resolution averaged over all incident xx values is σx av=0.019⋅w\sigma_{x\ \mathrm{av}} = 0.019 \cdot w.

Figure 4. The xx position uncertainty for normally incident particles versus the xx position along the base of the center triangular counter, from Eq. (12). Units are counter base width (ww) and a light yield of 100 PE/hh is assumed. The resolution is best at the corners of the center triangular counter and worst midway up its two sides.

Figure 5. The xx position uncertainty for normally incident particles versus the xx position along the base of the center triangular counter, from Eq. (12). The counter is 40 ×\times 20 mm2^{2} (w×hw \times h) and a light yield of 5 PE/mm is assumed. The resolution is best at the corners of the center triangular counter and worst midway up its two sides.

For a triangular counter of width 40 mm, such as the counter described in Ref. [9], that resolution is σx av=0.76\sigma_{x\ \mathrm{av}} = 0.76 mm, using their measured photoelectron yield of 100 PE/hh. The worse resolution for such a counter is 1.0 mm.

Our result is a lower limit on the position uncertainty one expects from a real hodoscope. Real position resolutions are worse for several reasons, such as imperfections in the counter

geometry, electronics noise, and the fact that assuming a Poisson distribution for the photoelectron yield underestimates the tails in the real light yield distributions. A recent study of the position resolution of a hodoscope consisting of 40×20 mm240 \times20\ \mathrm{mm}^{2} (w×hw \times h) counters with a light yield of 100 PE/h100\ \mathrm{PE}/h (and using Eq. (1) to determine the xx position) shows a similar behavior in the position resolution as a function of the charged particle impact position as shown in Fig. 5 [9]. They find a maximum xx-position uncertainty of roughly 2 mm, about a factor of two worse than what we estimate here. However, with a toy Monte Carlo that uses a Landau+Gaussian fit to the measured light yield distribution, they find excellent agreement with the data.

References

  1. [1]M. Adams et al., “A new detector technique using triangular scintillating strips to measure the position of minimum ionizing particles,” Nucl. Instrum. Meth. A 378 (1996) 131, and P. Baringer et al., “Cosmic-ray tests of the DØ preshower detector,” Nucl. Instrum. Meth. A 469 (2001) 295.DOI
  2. [2]G. Ambrosi et al., “The MU-RAY project: Volcano radiography with cosmic-ray muons,” Nucl. Instrum. Meth. A 628 (2011) 120. A. Anastasio et al., “The MU-RAY experiment. An application of SiPM technology to the understanding of volcanic phenomena,” Nucl. Instrum. Meth. A 718 (2013) 134. A. Anastasio et al., “The MU-RAY detector for muon radiography of volcanoes,” Nucl. Instrum. Meth. A 732 (2013) 423,DOI
  3. [3]L. Aliaga et al., “Design, calibration, and performance of the MINERvA detector,” Nucl. Instrum. Meth. A 743 (2014) 130.DOI
  4. [4]S. Cecchini et al., “Test of scintillator bars coupled to Silicon Photomultipliers for a charged particle tracking device,” Nucl. Instrum. Meth. A 846 (2017) 1,DOI
  5. [5]G. Baccani et al., “The MIMA project. Design, construction and performances of a compact hodoscope for muon radiography applications in the context of archaeology and geophysical prospections,” JINST 13 (2018) P11001.DOI
  6. [6]P. Bernardini et al., “On the performances of a particle tracking detector based on triangular scintillator bars read out by silicon photomultipliers,” Nucl. Instrum. Meth. A 967 (2020) 163882,DOI
  7. [7]X. Luo et al., “Development and commissioning of a compact Cosmic Ray Muon imaging prototype,” Nucl. Instrum. Meth. A 1033 (2022) 166720.DOI
  8. [8]P. V. Vavilov, Sov. Phys. JETP 5, 749 (1957), (Zh. Eksp. Teor. Fiz. 32, 920 (1957)], L.D. Landau, J. Exp. Phys. (USSR) 8, 201 (1944).
  9. [9]“Position Resolution of a Scintillator Hodoscope Employing Triangular Counters with Embedded Wavelength Shifting Fibers,” E. Craig Dukes, Ralf Ehrlich, Daniel Lee, Sydney Roberts, Luke Watson, Alan Bross, Sten Hansen, and Paul Rubinov, arXiv:2608.10282 (2026).arxiv.org/abs/2608.10282

Paper details

vlnc:W_mnyzzfrq
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 2

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents