Nuclear Instruments and Methods in Physics Research A 433 (1999) 513}517
Single-electron pulse-height spectra in thin-gap parallel-plate
chambers
P. Fonte  *, Rui Ferreira-Marques  , V. Peskov , A. Policarpo 
Instituto Superior de Engenharia de Coimbra, Coimbra, Portugal
Departamento de Fisica da Universidade de Coimbra, Coimbra, Portugal
LIP - Laboratorio de Instrumentacao e Fisica Experimental de Particulas, Portugal
NASA/Marshal Space Flight Center, Huntsville, AL, USA
Abstract
Single-electron pulse-height spectra were measured in 0.6 and 1.2 mm parallel-plate chambers developed for the TOF
system of the ALICE/LHC-HI experiment. Mixtures of Ar with ethane, isobutane, and SF

were studied. The observed
spectrum shows a clear peak for all gases, suggesting e$cient single-electron detection in thin parallel-plate structures.
The pulse-height spectrum can be described by the weighted sum of an exponential and a Polya distribution, the Polya
contribution becoming more important at higher gains. Additionally, it was found that the maximum gain, above 10, is
limited by the appearance of streamers and depends weakly on the gas composition. The suitability of each mixture for
single-electron detection is also quantitatively assessed. 1999 Elsevier Science B.V. All rights reserved.
Keywords: Single-electron pulse-height spectra; Parallel-plate chambers
1. Introduction
The advent of heavy-ion collision physics has
renewed the interest in Time-of-Flight (TOF) techniques
for particle identi"cation, since the momentum
spectrum of many of the large number of
particles emerging from the interaction is within the
range covered by TOF.
The ALICE [1] experiment at the CERN's Large
Hadron Collider heavy-ion program will include
a TOF barrel system covering the $1 rapidity
*Corresponding author. European Laboratory for Particle
Physics (CERN), CH-1211, Geneve 23, Switzerland.
Present address: Royal Institute of Technology,
Frescativagen 24, Stockholm, Sweden, S-10405.
E-mail address: fonte@lipc."s.uc.pt (P. Fonte)
range. The technology now chosen to implement
this detector is based on parallel-plate chambers,
similar in construction to the ones tested in this
work.
We, hereby, investigate the characteristics of the
Pulse Height Spectra (PHS) of avalanches initiated
by single electrons extracted from the chamber
cathode by UV photons. This experimental situation
is relatively similar to the one found in the
TOF application because when an ionizing particle
crosses an amplifying parallel-plate gap only those
electrons created close to the cathode by ionizing
collisions will be subject to the full gap ampli"cation.
For a sub-millimeter gap only a few electrons
at most will be in such conditions.
The measured PHS constitutes also a basic input
information to any Monte Carlo model of such
detectors.
0168-9002/99/$- see front matter 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 9 0 0 2 ( 9 9 ) 0 0 3 0 9 - 5 SECTION IX
Fig. 1. (a) Structure of the detector cell and (b) experimental
setup.
2. Experimental setup
The detector cell is constituted by a gas gap
con"ned by two identical metallic electrodes. These
electrodes were deposited on pro"led ceramic
plates [2] and the electrical contact made through
a small perforation in the plate, placed outside the
active gap volume. The gap width is de"ned by a set
of four spacers placed in the corners of the squareshaped
ceramic plates and also lying outside the
active gap volume. A cross section of the detector
cell can be seen in Fig. 1a.
The experimental setup is schematically represented
in Fig. 1b. The cell is illuminated from the
side by light emitted from a D

continuous discharge
lamp, attenuated by an adjustable colorneutral
"lter.
The detector is read out from the cathode by
a spectroscopy pre-ampli"er and ampli"er (CANBERRA
2001) chain and the resulting signals analyzed
by a pulse height analyzer. The spectra are
calibrated in charge by injection of a known voltage
pulse to the pre-ampli"er test input. The 50 s
integration time of the amplifying chain allows the
full signal charge to be integrated. The counting
rate is kept below a few hundred Hz in order to
ensure that even after integration most pulses correspond
to a single avalanche. Since the shape of
the measured pulse height spectra was independent
of the light source intensity, we conclude that most
of the avalanches were initiated by single photo-
electrons.
On an independent measurement made with
a chamber having one glass and one metallic electrode,
but otherwise similar, it was established that
for a given light source intensity and gas mixture
the counting rate strongly depended on whether the
cathode was the metallic surface or the glass plate,
being higher for the former case. This indicates that
most of the electrons are released from the metallic
surface and not from the gas, in which case there
would be no di!erence between both situations.
The anode is readout by a fast current ampli"er
[3,4] with a rise time of 2 ns. This ampli"er was
used for the time-resolved measurements presented
below.
The chamber was kept in a gas-tight box where
an externally prepared gas mixture would continuously
#ow. No dependence of the measurements on
the gas #ow was found.
3. Results
For all mixtures studied the observed PHS can
be "tted (with 

(2) by the weighted sum of an
exponential and a gamma distribution, with
weighting factors a and b, and parameters

,

,
and k:
y"a

e\HV#b
xI\ I

e\HV
(k)
. (1)
The mean value of this distribution, equal to the
gas gain when x corresponds to the number of
collected electrons, is given by
G"
a

#
bk

being the variance given by
"
a


#
kb


.
Also known as the Polya distribution.
514 P. Fonte et al. / Nuclear Instruments and Methods in Physics Research A 433 (1999) 513}517
Fig. 2. Typical observed pulse height spectra for low gain
(4;10) and high gain (3;10), with the adjusted analytical
distributions (see text) superimposed. The gas mixture was
Ar#10% isobutane.
An example of the adjustment of the distribution
(Eq. (1)) to the data is presented in Fig. 2.
A "gure of merit, 04p41, which measures the
departure of the observed distribution from a pure
exponential distribution corresponding to the same
gas gain, is plotted in Fig. 3a as a function of the gas
gain for several gas mixtures and gaps of 1.2 and
0.6 mm.
p"



y!
e\V%
G dx. (2)
For each curve the point at larger gain corresponds
to the maximum achievable gain, de"ned as
the gain at which a breakdown rate of a fraction of
Hz is observed. Such information is summarized in
Fig. 4.
From Fig. 3a it can be inferred that the departure
from an exponential PHS occurs mainly (p'0.1)
at the larger gains, above 10 for argon}hydrocarbon
mixtures and above 10 for mixtures containing
SF

. The e!ect is larger for the 1.2 mm gap than
for the 0.6 mm gap and it is much smaller for
mixtures containing SF

(a strongly electronegative
gas). At high gains, above 10, three families of lines
with distinct magnitudes of p can be recognized.
Since the PHS shape changes with the gas composition
(most notoriously with SF

which yields
the lowest values of p) it is unlikely that the observed
shape will be due to an irregular deposition
of primary charge or chamber gain inhomogeneities
(on the edges for instance).
Another useful "gure of merit can be de"ned as
the e$ciency for single-electron detection at a given
detection threshold. This detection e$ciency, taken
as the ratio between the number of counts above
the detection threshold and the total number of
counts, was calculated from the measured pulse
height spectra. Notice that the pedestal was about
2000 electrons wide, corresponding to less than one
bin of the multichannel analyzer. The results, for
a detection threshold of 10 collected electrons, are
presented in Fig. 3b.
The three families of lines already identi"ed in Fig.
3a at gains above 10, corresponding to di!erent
magnitudes of p, can also be identi"ed in Fig. 3b. For
those mixtures where the PHS is more peaked (higher
p) a given e$ciency is achieved at a lower gain, but
since the mixtures having a smaller value of p seem
to achieve larger gains, the "nal e$ciency is similar.
To check whether the observed PHS shape is due
to some form of feedback (photon and ion feedback
being the most common forms), we looked for the
presence of aftercurrents following the main avalanche
but no statistically signi"cant evidence of
such processes was found.
In Fig. 5, we show a typical current signal in the
vicinity of a breakdown event. The characteristic
precursor}streamer structure of the Raether breakdown
mechanism [5}8] is clearly visible, in sharp
contrast with the slow current growth typical of the
feedback-mediated Townsend breakdown mechanism
[5}8], thereby con"rming that feedback does
not play a major role in the chamber behavior.
The results shown in Fig. 4 indicate that, although
some optimization is possible, the maximum
achievable gain for all gases studied lies
within the same order of magnitude. The maximum
avalanche charge is more than one order of magnitude
smaller than the well-known Raether criterion
of streamer breakdown (10 electrons). This may be
related to the fact that the streamer is triggered by
P. Fonte et al. / Nuclear Instruments and Methods in Physics Research A 433 (1999) 513}517 515
SECTION IX
Fig. 3. (a) Departure of the observed distribution from a pure exponential distribution, as de"ned by Eq. (2); (b) Detection e$ciency for
single photoelectrons, considering a detection threshold of 10 collected electrons.
a space charge e!ect that depends on the charge
density inside the avalanche, that is, on the inverse
of the cube of the avalanche linear dimensions.
Since the gaps are very narrow and the width of the
primary charge deposition is null (1 electron) necessarily
the avalanche will be smaller than in the
(more common) case of a gap with a few mm width.
4. Conclusions
We studied the PHS of avalanches initiated by
single electrons in thin-gap PPCs suited for TOF
detectors. The measured PHS constitutes a basic
input information to any Monte Carlo model of
such detectors.
516 P. Fonte et al. / Nuclear Instruments and Methods in Physics Research A 433 (1999) 513}517
Fig. 4. Maximum achievable gain for several di!erent gas and
gap length conditions. Some estimated representative error bars
are also displayed.
Fig. 5. Current signal for a breakdown event recorded on
a digital oscilloscope via the fast ampli"er chain described
above. The horizontal scale corresponds to 50 ns/division and
the vertical scale is in arbitrary current units. The gas was
Ar#10% ethane on a 0.6 mm gap.
At large gains the PHS shape departs from the
exponential distribution observed at lower gains,
forming a peak. The magnitude of such departure
( p) depends upon the gas gain, the gap width and
the nature of the gas mixture.
No relevant photon or ion feedback was observed,
ruling out such origin for the peaked PHS
shape. The dependency of the spectrum shape on
the gas mixture, for a given detector, seems to
exclude that the e!ect would be caused by gain
inhomogeneities within the chamber or on its
edges.
The maximum achievable gain lies between 10
and 8;10 for all mixtures studied, limited by the
appearance of streamers.
For some gas mixtures single photoelectron detection
e$ciencies in excess of 98% were measured
for a detection threshold of 10 collected electrons.
For such mixtures an inverse relation seems to exist
between the magnitude of p and the maximum
achievable gain, resulting in identical maximal
detection e$ciencies.
Acknowledgements
Some of the necessary hardware was provided by
ITEP, Moscow, but this work would never have
been started without the friendly encouragement of
Dr. Francois Piuz.
References
[1] ALICE - Technical Proposal for A Large Ion Collider
Experiment at the CERN LHC, CERN/LHC/95-71,
December 1995.
[2] V. Akimov et al., Nucl. Instr. and Meth. A 344 (1994)
120.
[3] R. D'Alessandro et al., preprint ITEP 51}95.
[4] A. Martemianov, Proceedings of the First Workshop on
Electronics for LHC Experiments, Lisbon, 11}15 September,
1995.
[5] H. Raether, Electron Avalanches and Breakdown in Gases,
Butterworths, London, 1964.
[6] J.M. Meek, in: J.A. Rees (Ed.), Electrical Breakdown of
Gases, MacMillan, London, 1973.
[7] P. Fonte et al., Nucl. Instr. and Meth. A 310 (1991)
140.
[8] P. Fonte, IEEE Trans. Nucl. Sci. NS-43 (1996) 2135.
P. Fonte et al. / Nuclear Instruments and Methods in Physics Research A 433 (1999) 513}517 517
SECTION IX