1: % Talk at Radhep-2000 Conference
2: % UCLA, Nov. 16-18, 2000
3: % Draft as of April 14, 2003
4: % One reference updated
5: \documentstyle[epsfig,12pt]{aipproc}
6: \def \beq{\begin{equation}}
7: \def \eeq{\end{equation}}
8: \def \efi{Enrico Fermi Institute Report No.\ EFI}
9: \topmargin 0.1in
10: \begin{document}
11: \title{Extensive Air Shower Radio Detection:\\
12: Recent Results and Outlook
13: \footnote{Invited talk presented by J. Rosner at RADHEP-2000 Conference, UCLA,
14: Nov.\ 16--18, 2000, proceedings published by AIP. \efi~2000-57,
15: astro-ph/0101089.}}
16: \author{Jonathan L. Rosner and Denis A. Suprun}
17: \address{Enrico Fermi Institute and Department of Physics \\
18: University of Chicago, Chicago, IL 60637 USA}
19: \maketitle
20: \begin{abstract}
21: A prototype system for detecting radio pulses associated with extensive
22: cosmic ray air showers is described. Sensitivity is compared
23: with that in previous experiments, and lessons are noted for future studies.
24: \end{abstract}
25:
26: \section{Introduction}
27:
28: The observation of the radio-frequency (RF) pulse associated with extensive air
29: showers of cosmic rays has had a long
30: and checkered history. In the present report we describe an attempt to
31: observe such a pulse in conjunction with the Chicago Air Shower Array (CASA)
32: and Michigan Muon Array (MIA) at Dugway, Utah. Only upper limits
33: on a signal have been obtained at present, though we are still processing
34: data and establishing calibrations.
35:
36: In Section~2 we review the motivation and history of RF pulse detection.
37: Section~3 is devoted to the work at CASA/MIA, while Section~4 deals with
38: some future possibilities, including ones associated with the planned
39: Pierre Auger observatory. We conclude in Section~5. This report is an
40: abbreviated version of a longer one \cite{nim}.
41:
42: \section{Motivation and history}
43:
44: \subsection{Auxiliary information on shower}
45:
46: Present methods for the detection of an extensive air shower of cosmic rays
47: leave gaps in our information. The height above ground at which showers
48: develop cannot be provided by ground arrays, though stereo detection by air
49: fluorescence detectors is useful. Composition of the primary particles,
50: another unknown, is correlated with shower height, with heavy
51: primaries leading to showers which begin higher in the atmosphere.
52:
53: Radio detection can help fill such gaps. The electric field associated with a
54: charge $|e|$ undergoing an apparent angular acceleration $\ddot{\theta}$ is
55: $|{\cal E}| = 1.5 \times 10^{-26} \ddot{\theta}$ V/m, where time is measured in
56: seconds \cite{Allan,Feyn}. For typical showers the charges of radiating
57: particles are expected to be able to act coherently
58: to give a pulse with maximum frequency component $\nu_{\rm max} ({\rm MHz})
59: \simeq 10^6/R^2 ({\rm m})$, where $R$ is the distance of closest approach of
60: the shower axis to the antenna. Showers originating higher in the atmosphere
61: are expected to have higher-frequency components. Thus RF detection may be
62: able to add information on shower height and primary composition, and to
63: provide a low-cost auxiliary system in projects such as the
64: Pierre Auger array \cite{Auger}.
65:
66: \subsection{Pulse generation mechanisms}
67:
68: Several possibilities have been discussed for generation of a pulse by
69: air showers. Cosmic rays could induce the atmosphere to act as a giant spark
70: chamber, triggering discharges of the ambient field gradient \cite{RRW}.
71: Compton scattering and knock-on electrons can give rise to a negative charge
72: excess of some 10 to 25\% at shower maximum \cite{Ask}. Separation of positive
73: and negative charges can occur in the Earth's magnetic field as a result of a
74: $q {\bf v} \times {\bf B}$ force \cite{KL}. This last mechanism
75: is thought to be the dominant one accounting for atmospheric
76: pulses with frequencies in the 30--100~MHz range
77: \cite{Allan}, and will be taken as the model for the signal for which the
78: search was undertaken. The charge-excess mechanism is probably the major
79: source of an RF signal in a dense material such as polar ice \cite {ZHS},
80: but is expected to be less important in the atmosphere.
81:
82: \subsection{Early measurements}
83:
84: The first claim for detection of the charge-separation mechanism utilized
85: narrow-band techniques at 44 and 70~MHz \cite{Jelley,Weekes}.
86: A Soviet group reported signals at 30~MHz \cite{Sov}, while a University of
87: Michigan group at the BASJE Cosmic Ray Station on Mt.~Chacaltaya,
88: Bolivia \cite{Chac} studied pulses in the 40--90~MHz range.
89:
90: The collaboration of H. R. Allan {\it et al.}\ \cite{Allan} at Haverah Park in
91: England studied the dependence of signals on primary energy
92: $E_p$, perpendicular distance $R$ of closest approach of the
93: shower core, zenith angle $\theta$, and angle $\alpha$ between
94: the shower axis and the magnetic field vector. Their results
95: indicated that the electric field strength per unit of frequency,
96: ${\cal E}_\nu$, could be expressed as
97: \beq \label{eqn:E}
98: {\cal E}_\nu = s \frac{E_p}{10^{17} {\rm~eV}} \sin \alpha \cos \theta
99: \exp \left( - \frac{R}{R_0(\nu, \theta)} \right)~~~\mu{\rm
100: V}~m^{-1}~ {\rm MHz}^{-1}~~~,
101: \eeq
102: where $R_0$ is an increasing function of $\theta$, equal (for example) to $(110
103: \pm 10)$ m for $\nu = 55$~MHz and $\theta < 35^\circ$. The constant $s$ was
104: originally claimed to be 20. The Haverah Park observations were
105: recalibrated to yield $s = 1.6$ (0.6 $\mu$V m$^{-1}$~MHz$^{-1}$ for a
106: $10^{17}$~eV shower at $R = 100$ m) while observations in the U.S.S.R. gave
107: $s = 9.2$ (3.4~$\mu$V~m$^{-1}$~MHz$^{-1}$ at $R = 100$ m) \cite{Atrash}.
108: To estimate the corresponding signal
109: strength in \cite{Jelley,Weekes}, we note that the signal power for
110: showers of average primary energy $E = 5 \times 10^{16}$~eV was measured to be
111: about 4 times that of galactic noise, for which \cite{Allan} ${\cal E}_\nu^{\rm
112: Gal} \simeq 1$--$2~\mu$V m$^{-1}$~MHz$^{-1}$. Thus, for such showers, one
113: expects ${\cal E}_\nu \simeq 2$--$4~\mu$V m$^{-1}$~MHz$^{-1}$.
114: Similarly, the estimate \cite{Weekes} of an average pulse power $V_{\rm peak}^2
115: /2 R = 10^{-12}$ W gives $V_{\rm peak} = 10~\mu$V for $R = 50~\Omega$. Using
116: the relation between pulse voltage and ${\cal E}_\nu$ \cite{Allan}
117: \beq \label{eqn:VE}
118: V = 30 G^{1/2} (\delta \nu/\nu) {\cal E}_\nu~~~,
119: \eeq
120: where
121: $G$ is the antenna gain, and $\delta \nu$ is the bandwidth centered at
122: frequency $\nu$, we find for an assumed $G = 5$ (7~dB) (it is not quoted in
123: Ref.\ \cite{Jelley}) and $\delta \nu/ \nu = 2.75/44$ \cite{Jelley}, a value of
124: ${\cal E}_\nu \simeq 2.4~\mu$V m$^{-1}$~MHz$^{-1}$ at a primary energy of
125: $5 \times 10^{16}$~eV, or about $5~\mu$V m$^{-1}$~MHz$^{-1}$ at $10^{17}$~eV if
126: ${\cal E}_\nu$ scales linearly with primary energy \cite{Allan}. For $G=5$ the
127: data of Refs.\ \cite{Jelley,Weekes} thus would favor the higher field-strength
128: claims of the U.S.S.R. group cited in Ref.\ \cite{Atrash}.
129:
130: More recent claims include pulses with
131: components at or below several MHz \cite{Agasa,Yak,GS,Gau}, and
132: at VHF frequencies \cite{GS,Gau}. The Gauhati University group has reviewed
133: evidence for pulses at a wide range of frequencies \cite{Gau}.
134:
135: \subsection{Pulse characteristics}
136:
137: The Haverah Park observations are consistent with a model
138: in which the pulse's onset is generated
139: by the start of the shower at an elevation of about 10 km above
140: sea level, while its end is associated with the greater total
141: path length (shower $+$ signal propagation distance) associated
142: with the shower's absorption about 5~km above sea level. If a vertical
143: shower is observed at a distance of 100~m from its core, the pulse should
144: rise and fall back to zero within about 10~ns, with a subsequent longer-lasting
145: negative component. High frequencies should be less visible far from
146: the shower axis. Heavy primaries should lead to showers originating higher in
147: the atmosphere, with consequent higher-frequency RF components as a result
148: of the geometric aspect ratio with which they are viewed by the antenna, and
149: possibly a greater ${\cal E}_\nu$ for a given primary energy \cite{Allan}.
150: The polarization of the pulse should be dictated by the mechanism of pulse
151: generation: e.g., perpendicular to the line of sight with component
152: along ${\bf v} \times {\bf B}$ for the charge-separation mechanism.
153:
154: \subsection{RF backgrounds}
155:
156: Discharges of atmospheric electricity will be detected at random
157: intervals at a rate depending on local weather
158: conditions and ionospheric reflections. Man-made RF
159: sources include television and radio stations, police and other
160: communications services, broad-band sources (such as ignition
161: noise), and sources within the experiment itself. The
162: propagation of distant noise sources to the receiver is a strong
163: function of frequency and of solar activity.
164:
165: Galactic noise can be the dominant signal in exceptionally
166: radio-quiet environments for frequencies in the low VHF (30--100~MHz)
167: range \cite{Allan}. For higher frequencies in such
168: environments, thermal receiver noise becomes the dominant effect.
169:
170: \section{The installation at CASA}
171:
172: The CASA/MIA detector is located about 100~km southwest of Salt Lake City,
173: Utah, at the Dugway Proving Ground
174: \cite{CASAnim}. The Chicago Air Shower Array (CASA) is
175: a rectangular grid of $33 \times 33$ stations on on the desert's surface.
176: The inter-station spacing is 15~m. A station has four 61
177: cm $\times$ 61~cm $\times$ 1.27~cm sheets of plastic scintillator
178: each viewed by its own photomultiplier tube (PMT). When a signal
179: appears on 3 of 4 PMTs in a station, a ``trigger request pulse''
180: of 5~mA with 5~$\mu$s duration is sent to a central trailer, where
181: a decision is made on whether to interrogate all stations for a
182: possible event. Details of this trigger are described in Ref.~\cite{CASAnim}.
183: When this experiment was begun the CASA array had been reconfigured to remove
184: the 4 westernmost ``ribs'' of the array. For
185: runs performed in 1998, the easternmost rib had also been removed.
186:
187: The University of Michigan designed and built a muon detection
188: array (MIA) to operate in conjunction with CASA. It consists of
189: sixteen ``patches,'' each having 64 muon counters, buried 3~m
190: below ground at various locations in the CASA array.
191: Each counter has lateral dimensions 1.9 m $\times$ 1.3
192: m. Four of the patches (numbered 1 through 4), each about 45~m
193: from the center of the array, lie on the corners of a skewed
194: rectangle; four (numbered 5 through 8), each about 110 m from the
195: center of the array, lie on a rectangle with slightly different
196: skewed orientation, and eight (numbered 9 through 16) lie on the
197: sides and corners of a rectangle with sides $x \simeq \pm 180$~m
198: and $y \simeq \pm 185$~m, where $x$ and $y$ denote East and North
199: coordinates.
200:
201: \subsection{Expected integral rates}
202:
203: The CASA trigger threshold is a few $\times 10^{14}$~eV and corresponds to
204: a trigger rate of 10--20 Hz. The expected rates above ($10^{15},~10^{16},~
205: 10^{17},~10^{18}$)~eV are about (1~Hz, 1 per 2 min, 1 per 4~hr, and 2 per mo),
206: respectively, over the 1/4 km$^2$ area of the array.
207: A primary energy of at least $10^{17}$~eV seems to be
208: needed if radio signals are to exceed the galactic noise level of
209: 1--2 $\mu$V m$^{-1}$~MHz$^{-1}$. Such pulses could be generated by a
210: $10^{17}$~eV vertical shower with axis 100~m from the antenna under the most
211: optimistic estimates. Since the whole array should see such
212: showers only every few hours, and most have axes farther from the
213: antenna than 100 m, the possibility of accidental noise pulses during such
214: long time intervals reduces the expected sensitivity considerably.
215:
216: \subsection{The ``radio shack'' at CASA}
217:
218: A survey of the CASA/MIA site determined that within the array, broad-band
219: noise associated with computers, switching power supplies, and other
220: electronics was so intense that no RF searches could be undertaken. The same
221: was true at any position within the perimeter of the array.
222: Consequently, an antenna was mounted on top of a mobile searchlight tower at
223: a height of 10~m about 24~m east of the eastern edge of the array,
224: corresponding to $x = 263.8$~m,
225: $y=0$~m. The antenna, a 9-element portable log-periodic antenna manufactured
226: by Dorne and Margolin, was acquired from FairRadio Co.~in Lima, Ohio, for about
227: \$60. Its nominal bandwidth is 26--76~MHz but it was measured to have usable
228: properties up to 170~MHz.
229:
230: The signal was fed through 60~ft.\ of RG-58U cable,
231: filtered by a high-pass filter admitting frequencies
232: above 23~MHz, preamplified using a Minicircuits ZFL-500LN preamplifier with
233: 26~dB of gain, low-pass-filtered to admit frequencies below 250~MHz, and fed
234: to the oscilloscope at a sensitivity of 5~mV/division. This constituted the
235: ``wide-band'' configuration used for most data acquisition runs. The filters
236: were Minicircuits BNC coaxial models. A ``narrow-band'' configuration with
237: response between 23 and 37~MHz and two preamplifiers had substantially poorer
238: signal-to-noise ratio in distinguishing transient signals from background.
239:
240: A trigger based on the coincidence of seven of the eight outer muon ``patches''
241: was set to select large showers. Each muon patch was set to produce a
242: trigger pulse when at least 5 of its 64 counters registered a minimum-ionizing
243: pulse within $5.2$~$\mu$s of one another. The pulses were then combined
244: to produce a summed pulse, fed to a discriminator, whose output was amplified
245: and sent over a cable (with measured delay time 2.15~$\mu$s) to the RF trailer.
246: No evidence for pickup of the trigger pulse from the antenna was found. The
247: trigger corresponded to a minimum shower energy somewhat below $10^{16}$~eV,
248: based on the integral rate \cite{REF} at $10^{18}$~eV of 0.17/km$^2$/day/sr.
249:
250: A Tektronix TDS-540B digitizing oscilloscope registered filtered
251: and preamplified RF data on a rolling basis. These data were then captured
252: upon receipt of a large-event trigger and stored on hard disk using a National
253: Instruments GPIB interface. Data were
254: taken using separate computers (at different times), allowing
255: for analysis both at the University of Washington and at
256: Chicago. The Washington system used a Macintosh Quadra 950
257: running Labview, while the Chicago system used either a Dell XPS200s Pentium
258: desktop computer or a Dell Latitude LM laptop computer
259: running a C program adapted from those provided by National Instruments.
260: Each trigger caused 50~$\mu$s of RF data, centered around the
261: trigger and acquired at 1~GSa/s, to be saved.
262:
263: The total trigger rate ranged between about 20 and 50 events per hour,
264: depending on intermittent noise sources in the trigger system.
265: Concurrently, the CASA on-line data acquisition system was
266: instructed to write files of
267: events in which at least 7 out of the 8 outermost muon patches
268: produced a patch pulse. These files
269: typically overlapped with the records taken at the RF trailer to
270: a good but not perfect extent as a result of occasional noise on
271: the trigger line.
272:
273: \subsection{Raw data and RF backgrounds}
274:
275: To remove strong Fourier components associated with narrow-band RF
276: signals which were approximately constant over the duration of
277: each data record, a MATLAB routine performed
278: the fast Fourier transform of the signal and renormalized the
279: large Fourier components to a given maximum intensity. Fig.~1
280: shows the fast Fourier transform of a typical RF signal before
281: and after this procedure was applied. In each case the data were
282: acquired using the ``wide-band'' filter configuration mentioned above, whose
283: response cuts off sharply below 23~MHz.
284:
285: % This is Figure 1 (Figure 2 of NIM article)
286: \begin{figure}
287: \centerline{\epsfysize = 4.7 in \epsffile{nim2.eps}} \caption{Top
288: panel: Fourier spectrum (in arbitrary units) of RF signals
289: at Dugway site. Prominent features include video and audio carriers
290: for TV Channels 2, 4, 5, 7, and 11,
291: and the FM broadcast band between 88 and 108~MHz. Bottom panel:
292: Fourier spectrum (same vertical scale) after renormalization of
293: large Fourier components to an arbitrary maximum magnitude. The
294: continuum between 23 and 88~MHz was not detectable in Chicago;
295: TV and FM signals were found to be almost 40~dB stronger there,
296: so gain was reduced correspondingly.}
297: \end{figure}
298:
299: % This is Figure 2 (Figure 3 of NIM article)
300: \begin{figure}
301: \centerline{\epsfysize = 7 in \epsffile{nim3.eps}}
302: \caption{Effect of Fourier coefficient shrinkage on detectability
303: of a transient. Top panel: raw RF record (in arbitrary units)
304: with simulated signal superposed. Middle panel: record (same
305: scale) after Fourier coefficient shrinkage. Here a maximum
306: Fourier coefficient magnitude of $10^3$ (in the units of Fig.~1)
307: has been imposed. Bottom panel: the same record after denoising
308: with a wavelet routine.}
309: \end{figure}
310:
311: The effect of digital filtering on detectability of a transient is
312: illustrated in Fig.~2. The top panel shows the RF record whose
313: Fourier transform was given in Fig.~1, on which has been
314: superposed a simulated transient of peak amplitude 14.5
315: digitization units. (The data acquisition scale ranges from $-128$ to $+127$
316: digitization units; one scale division on the oscilloscope corresponds to 25
317: units.) The transient is invisible beneath the large
318: amplitude associated with television and FM radio signals. The
319: middle panel shows the result after application of the Fourier
320: coefficient shrinkage algorithm. The bottom panel shows the same record
321: after denoising with a wavelet routine \cite{DFS}.
322: The records in Figs.~1 and 2 were obtained for 32,768 data points
323: obtained at a 1~ns sampling interval, with the trigger at the 20,000th point.
324:
325: \subsection{Signal simulation}
326:
327: To quantify signal processing efficiency, we generated the expected signal,
328: fed it through the same preamplifier and filter configurations used for
329: data acquisition, and superposed it on records otherwise free of
330: transients. We successively reduced the amplitude of the
331: superposed test signal until it was no longer detectable, thereby
332: obtaining an estimate of sensitivity.
333:
334: % This is Figure 3 (Figure 4 of NIM article)
335: \begin{figure}
336: \centerline{\epsfysize = 4.7 in \epsffile{nim4.eps}}
337: \caption{Analytic depiction of typical pulse presented to
338: filter-preamplifier configuration. Top panel: time dependence
339: of pulse $f(t) = \theta(t) t^2[e^{-0.4t} - e^{-0.02t}/ 8000] (t$
340: in ns); bottom panel: Fourier spectrum of pulse for $-20~\mu$s
341: $\le t \le 12.768~\mu$s (calculated analytically). In the top panel, the short
342: bar above the pulse denotes $\delta$, the time difference between onset and
343: maximum, while the longer bar below the pulse denotes
344: the duration of the positive component.}
345: \end{figure}
346:
347: A Hewlett-Packard Arbitrary Waveform Generator was used to
348: generate signals whose typical characteristics are illustrated in Fig.~3.
349: These signals were taken to have the form $f(t) = \theta(t)
350: At^2(e^{-Bt} - C e^{-Dt})$ with the coefficient $C$ chosen so
351: that $f(t)$ has no DC component, and $D$ corresponding to a long
352: duration of the negative-amplitude component. For all pulses we
353: chose $D = B/20$, so that $C = (8000)^{-1}$ cancels the DC
354: component. The Fourier components of the test pulse fall off
355: smoothly with frequency. The initial $t^2$ behavior was chosen
356: so that both the test pulse and its first derivative vanish at
357: $t=0$, as might be expected for a pulse from a developing shower. We
358: chose $B = 0.8,~0.4,~0.2,~0.1$ corresponding to a time difference between
359: pulse onset and maximum of $\delta = 2.5~,5,~10,~20$~ns
360: and simulated both narrow-band (23--37~MHz) and broad-band (23--250~MHz)
361: configurations.
362:
363: The shape of the pulse of Fig.~3 is affected by preamplification and filtration
364: as shown in Fig.~4 for the broad-band example. (The narrow-band configuration
365: leads to a longer ringing time.) The noise is associated with the system used
366: to generate the test pulse, and the fact that the Fourier transform
367: is taken over a much longer time than the duration of the pulse. The sharp
368: feature at 125~MHz is a local artifact.
369:
370: Systematic studies of signal-to-noise ratios have been performed
371: so far only for the simulated pulses with $\delta=5$~ns applied
372: to the broad-band front end. A typical pulse of this type
373: gave a front end output of 21~mV peak-to-peak, acquired at an
374: oscilloscope sensitivity of 5~mV per division. Since each division
375: corresponds to 25 digitization units, the peak-to-peak range
376: is about 104 digitization units, or slightly less than half the
377: dynamic range (255 units, or 8 bits). Positive and negative
378: peaks are thus about 52 digitization units each.
379:
380: The stored test signal is then multiplied by a scale factor and
381: added algebraically to a collection of RF records in which, in
382: general, randomly occurring transients will be present. One then
383: inspects these records to see if the transient can be
384: distinguished from random noise.
385:
386: For the broad-band data we estimated that pulses with input
387: voltages corresponding to about 1/5 the original test pulse can
388: be distinguished from average noise (not from noise spikes!).
389: Since the original test pulse had a peak value of 1.3~mV, this
390: corresponds to sensitivity to an antenna output of about $V_{\rm
391: pk} \simeq 260~\mu$V. The ability to detect such a pulse with an
392: effective bandwidth of about 30~MHz corresponds to a threshold
393: sensitivity at the level of order $3~\mu$V/m/MHz \cite{nim}.
394:
395: % This is Figure 4 (Figure 5 of NIM article)
396: \begin{figure}
397: \centerline{\epsfysize = 4.7 in \epsffile{nim5.eps}}
398: \caption{Test pulse of Fig.~3 after broad-band filtration (23--250~MHz) and
399: preamplification. Top panel: time dependence of pulse; bottom
400: panel: Fourier spectrum of recorded pulse for $-20~\mu {\rm s}
401: \le t \le 12.768~\mu$s.}
402: \end{figure}
403:
404: Preliminary studies of simulated pulses applied to the
405: narrow-band front end suggest a considerably poorer achievable
406: signal-to-noise ratio, despite the expectation that the signal
407: should have a large portion of its energy between 23 and 37~MHz.
408: It appears difficult to detect a pulse from the antenna below
409: about 0.7~mV, which for a bandwidth of 14~MHz corresponds to a
410: threshold sensitivity of $7~\mu$V/m/MHz \cite{nim}.
411: Studies of possible improvements of the analysis
412: algorithm for the narrow-band data are continuing.
413:
414: \subsection{Transients detected under various conditions}
415:
416: % This is Figure 5 (Figure 7 of NIM article)
417: \begin{figure}
418: \centerline{\epsfysize = 5 in \epsffile{nim7.eps}} \caption{Top
419: panel: time-vs.-amplitude plot for maxima of 880 events recorded
420: in January 1998 (RF run 23498 only) with CASA HV supplied to all stations. All
421: events recorded with East-West antenna polarization. Bottom
422: panel: time distribution of transients.}
423: \end{figure}
424:
425: Several means were used to characterize transients. One method
426: with good time resolution involved the shrinkage of large Fourier coefficients.
427: One can then search for peaks of each data record, plotting their
428: amplitude against time relative to the trigger. One such
429: plot is shown in Fig.~5 for a data run in which CASA HV was
430: delivered to all boxes. A strong accumulation of transients,
431: mostly with amplitude just above the arbitrarily chosen threshold
432: (mean + 3 $\sigma$), is visible at times $-5$ to $-7~\mu$s relative to
433: the trigger. In a comparable plot for a run in which CASA HV was
434: completely disabled (Fig.~6), only a small accumulation at times $-6$ to
435: $-7~\mu$s is present. This excess appears due to transients with
436: predominantly high-frequency components (over 100~MHz). Since signal
437: pulses are expected to have more power below 100~MHz (see Fig.~4, bottom)
438: we believe that this accumulation is not due to shower radiation, but
439: most likely arises from the muon patches, one of which is within
440: 75~m of the antenna.
441:
442: A typical transient occurring in a run with CASA HV on is shown in Fig.~7. The
443: transients are highly suppressed (though not in all runs) when
444: CASA boxes within 100 m of the antenna are disabled.
445:
446: The time distribution of event maxima above an arbitrary
447: threshold for 880 events taken with CASA HV on (one run
448: from January 1998) is shown in the bottom panel of Fig.~5. The
449: mean arrival time is about $6~\mu$s before the trigger, with a
450: distribution which is slightly broader for pulses arriving
451: earlier than the mean. This broadening may correspond to some
452: jitter in forming the trigger pulse from the sum of muon patch
453: pulses.
454:
455: % This is Figure 6 (Figure 8 of NIM article)
456: \begin{figure}
457: \centerline{\epsfysize = 5 in \epsffile{nim8.eps}} \caption{Top
458: panel: time-vs.-amplitude plot for maxima of 824 events recorded
459: in January 1998 with CASA HV disabled. All events recorded with
460: East-West antenna polarization. Bottom panel: time distribution
461: of transients.}
462: \end{figure}
463:
464: As mentioned earlier, the time for the trigger pulse to propagate
465: from the central station to the RF trailer was measured to be
466: $2.15~\mu$s. One expects a similar travel time for pulses to
467: arrive from muon patches to the central station.
468: Moreover, the muon patch signals are subjected to delays so that
469: they all arrive at the central station at the same time for a
470: vertically incident shower. Thus, the peak in Fig.~5 is
471: consistent with being associated with the initial detection of a
472: shower by CASA boxes. This circumstance was checked by recording
473: CASA trigger request signals simultaneously with other data; they
474: coincide with transients such as those illustrated in Fig.~7
475: within better than $1/2~\mu$s.
476:
477: % This is Figure 7 (Figure 9 of NIM article)
478: \begin{figure}
479: \centerline{\epsfysize = 4.9 in \epsffile{nim9.eps}}
480: \caption{Signal of a typical transient associated with CASA
481: operation. Top panel: before denoising; bottom panel: after
482: denoising.}
483: \end{figure}
484:
485: The RF signals from the shower are expected to arrive around the
486: same time as, or at most several hundred nanoseconds before, the
487: transients associated with CASA operation. They would propagate
488: directly from the shower to the antenna, whereas transients from
489: CASA stations are associated with a slightly longer total path
490: length from the shower via the CASA station to the antenna. There
491: will also be some small delay at a CASA station in forming the
492: trigger request pulse. Thus, we expect a genuine signal also to
493: show up around 6--7 $\mu$s before the trigger. However, for data
494: recorded with CASA boxes disabled, no significant peak with the
495: expected frequency spectrum is visible
496: in this time window. The upper limit on the rate of events giving
497: rise to such a peak can be used to set a limit on RF
498: pulses associated with air showers.
499:
500: In Table \ref{tab:triggs} we summarize the triggers taken under optimum
501: conditions, corresponding to broad-band data acquisition under conditions
502: of minimum ambient noise. These triggers represent about 50 hours of data.
503: Data were taken with both East-West and North-South antenna polarizations.
504: Since noise from CASA boxes was found to be a significant source of RF
505: transients, data were taken with some or all CASA boxes disabled
506: by turning off high voltage (HV) supply to the photomultipliers.
507:
508: % This is Table 1
509: \begin{table}
510: \caption{Broad-band data recorded under lowest-noise conditions.
511: \label{tab:triggs}}
512: \begin{center}
513: \begin{tabular}{l c c c c} \hline
514: Antenna & CASA & CASA & Partial & Total \\
515: Polarization & HV on & HV off & CASA HV & events \\ \hline
516: East-West & 4503 & 857 & 1834 & 7194 \\
517: North-South & 677 & 641 & 366 & 1821 \\ \hline Total
518: events & 5180 & 1498 & 2200 & 9015 \\ \hline
519: \end{tabular}
520: \end{center}
521: \end{table}
522:
523: \subsection{Sensitivity estimate}
524:
525: The main difficulty associated with pulse detection is that signal pulses are
526: not easily distinguishable from large spurious pulses originating from
527: atmospheric discharges. Both air shower pulses and these background noise
528: pulses can considerably exceed the average noise level. We estimate that all
529: signal pulses should arrive between 7 and 6 $\mu$s before the trigger, while
530: the time distribution of noise pulses is assumed to be uniform. Hence, a
531: sufficiently large relative accumulation of pulse maxima in that time bin was
532: adopted as a key criterion in search of signal pulses. We also employed several
533: criteria to increase the fraction of signal pulses: The pulse should be larger
534: than some specified magnitude threshold, bandwidth limited and approximately
535: uniform within its limited bandwidth. To date, no significant accumulation has
536: been detected.
537:
538: At present we are only able, after accounting for the rate of accidental
539: noise pulses, using Monte-Carlo simulation, and taking an antenna gain of
540: $G = 2.5$ and a cable attenuation factor of 1.4~dB, to set an upper limit of
541: $s = 54$ in Eq.~(\ref{eqn:E}). This is to be compared with the original
542: Haverah Park result $s=20$ \cite{Allan}, the recalibrated result $s=1.6$
543: \cite{Atrash}, and the Soviet group's result $s=9.2$ \cite{Atrash}.
544: The noise level at Dugway is too high and the acquired sample is too limited
545: in statistics and dynamic range to allow us to place upper limits strict enough
546: to check the claims of the two groups. We hope that further
547: improvement of the data processing technique will reduce the
548: noise contribution. The magnetic dip angle $\gamma$ is much smaller at the
549: Auger site in Argentina ($34^{\circ}$ versus $68^{\circ}$ at Utah), leading
550: one to expect bigger electric fields for vertical showers
551: and facilitating the detection of shower radiation. We also expect
552: that the Argentina site will be quieter than Dugway and some
553: clarity as regards the calibrating factor will be established.
554:
555: \section{Outlook}
556:
557: \subsection{Further possibilities for processing present data}
558:
559: We are still hoping to improve our sensitivity to the point that we can
560: see a true RF signal from a shower even when CASA HV is not disabled. Such a
561: signal should precede CASA-related transients by
562: at least the delay of formation of a phototube signal.
563: An important calibration will be achieved if we can determine whether we are
564: sensitive to galactic noise, which may be responsible for the continuum
565: between 23 and 88~MHz visible in Fig.~1.
566:
567: We are still exploring improved methods for removing constant RF signals
568: from our records. In this respect we are limited by the 8-bit dynamic range
569: of the TDS-540B oscilloscope. Data taken with various configurations of
570: boxes near the antenna disabled may help us to better characterize the
571: CASA-related transients.
572:
573: The triggered data may be useful in a rather different context. It has
574: been proposed that radar methods be used to detect ion trails associated
575: with extensive air showers \cite{PG}. In our case we may be able to
576: investigate sudden enhancements of the signals of distant television
577: signals (on Channels 3, 6, 8, and 12) correlated with receipt of a large-event
578: trigger.
579:
580: \subsection{Considerations for Auger site}
581:
582: A number of questions have been suggested by the present investigation if
583: RF pulse detection is to be considered as an adjunct to the Auger array.
584:
585: \begin{itemize}
586:
587: \item How far from the shower axis can antennas detect pulses from showers with
588: energies above $10^{18}$ or $10^{19}$~eV? The answer determines whether a
589: sparse array (e.g., one with the same density as Auger stations)
590: would be sensitive to the RF pulse.
591:
592: \item What dynamic range for a data-acquisition system is needed so that a
593: transient signal survives digital filtering? Apparently 8-bit range is not
594: enough.
595:
596: \item What RF interference exists at each site? Surveys would be desirable.
597: They could pinpoint not only narrow-band sources due to broadcasting stations
598: but potentially dangerous broadband sources from switching power supplies,
599: computers, etc. It would be best to undertake such studies after prototype
600: systems are in place.
601:
602: \item What power budget would an RF detection system need? Each Auger
603: solar-powered station is limited to a total budget of 10 watts. Presumably
604: an RF system would have to use auxiliary power, particularly for its
605: fast-digitization and memory components.
606:
607: \item What is the envisioned minimum cost per RF station? It would presumably
608: be dominated by the data-acquisition system; the antennae and preamplifiers
609: would probably be cheap by comparison. A preliminary estimate is less than
610: \$3K per station \cite{nim}.
611:
612: \end{itemize}
613:
614: The Southern Hemisphere Auger site is progressing well toward an engineering
615: array of 40 stations, as we have heard at this conference \cite{Zas}. It is
616: hoped that in a couple of years an investigation at that site of the
617: feasibility of RF pulse detection can be undertaken.
618:
619: \section{Conclusions}
620:
621: No ``golden signal'' has been seen for an RF transient associated with
622: extensive air showers of cosmic rays at the CASA site. With further
623: processing, the data may permit the setting of useful upper limits on
624: signals relevant to at least some of the previous claims. A
625: number of useful lessons have been learned if a similar technique is to
626: be tried in conjunction with the Auger project.
627:
628: \section{Acknowledgements}
629:
630: It is a pleasure to thank Mike Cassidy, Jim Cronin, Brian Fick, Lucy Fortson,
631: Joe Fowler, Rachel Gall, Kevin Green, Brian Newport, Rene Ong, Scott Oser,
632: Daniel F. Sullivan, Fritz Toevs, Kort Travis, Augustine Urbas, and John
633: Wilkerson for collaboration and support on various aspects of this experiment.
634: Thanks are also due to Bruce Allen, Dave Besson, Maurice Givens, Peter Gorham,
635: Kenny Gross,
636: Dick Gustafson, Gerard Jendraszkiewicz, Larry Jones, Dave Peterson, John
637: Ralston, Leslie Rosenberg, David Saltzberg,
638: Dave Smith, M. Teshima, and Stephan Wegerich for
639: useful discussions. This work was supported in part by the Enrico Fermi
640: Institute, the Louis Block Fund, and the Physics Department of the University
641: of Chicago and in part by the U. S. Department of Energy under
642: Grant No.~DE FG02 90ER40560.
643:
644: % Journal and other miscellaneous abbreviations for references
645: \def \ajp#1#2#3{Am.\ J. Phys.\ {\bf#1}, #2 (#3)}
646: \def \apny#1#2#3{Ann.\ Phys.\ (N.Y.) {\bf#1}, #2 (#3)}
647: \def \app#1#2#3{Acta Phys.\ Polonica {\bf#1}, #2 (#3)}
648: \def \arnps#1#2#3{Ann.\ Rev.\ Nucl.\ Part.\ Sci.\ {\bf#1}, #2 (#3)}
649: \def \cmts#1#2#3{Comments on Nucl.\ Part.\ Phys.\ {\bf#1}, #2 (#3)}
650: \def \cn{Collaboration}
651: \def \cp89{{\it CP Violation,} edited by C. Jarlskog (World Scientific,
652: Singapore, 1989)}
653: \def \epjc#1#2#3{Eur.\ Phys.\ J. C {\bf#1}, #2 (#3)}
654: \def \f79{{\it Proceedings of the 1979 International Symposium on Lepton and
655: Photon Interactions at High Energies,} Fermilab, August 23-29, 1979, ed. by
656: T. B. W. Kirk and H. D. I. Abarbanel (Fermi National Accelerator Laboratory,
657: Batavia, IL, 1979}
658: \def \hb87{{\it Proceeding of the 1987 International Symposium on Lepton and
659: Photon Interactions at High Energies,} Hamburg, 1987, ed. by W. Bartel
660: and R. R\"uckl (Nucl.\ Phys.\ B, Proc.\ Suppl.\, vol. 3) (North-Holland,
661: Amsterdam, 1988)}
662: \def \ib{{\it ibid.}~}
663: \def \ibj#1#2#3{~{\bf#1}, #2 (#3)}
664: \def \ichep72{{\it Proceedings of the XVI International Conference on High
665: Energy Physics}, Chicago and Batavia, Illinois, Sept. 6 -- 13, 1972,
666: edited by J. D. Jackson, A. Roberts, and R. Donaldson (Fermilab, Batavia,
667: IL, 1972)}
668: \def \ijmpa#1#2#3{Int.\ J.\ Mod.\ Phys.\ A {\bf#1}, #2 (#3)}
669: \def \ite{{\it et al.}}
670: \def \jhep#1#2#3{JHEP {\bf#1}, #2 (#3)}
671: \def \jpb#1#2#3{J.\ Phys.\ B {\bf#1}, #2 (#3)}
672: \def \lg{{\it Proceedings of the XIXth International Symposium on
673: Lepton and Photon Interactions,} Stanford, California, August 9--14 1999,
674: edited by J. Jaros and M. Peskin (World Scientific, Singapore, 2000)}
675: \def \lkl87{{\it Selected Topics in Electroweak Interactions} (Proceedings of
676: the Second Lake Louise Institute on New Frontiers in Particle Physics, 15 --
677: 21 February, 1987), edited by J. M. Cameron \ite~(World Scientific, Singapore,
678: 1987)}
679: \def \kdvs#1#2#3{{Kong.~Danske Vid.~Selsk., Matt-fys.~Medd.} {\bf #1}, No.~#2
680: (#3)}
681: \def \ky85{{\it Proceedings of the International Symposium on Lepton and
682: Photon Interactions at High Energy,} Kyoto, Aug.~19-24, 1985, edited by M.
683: Konuma and K. Takahashi (Kyoto Univ., Kyoto, 1985)}
684: \def \mpla#1#2#3{Mod.\ Phys.\ Lett.\ A {\bf#1}, #2 (#3)}
685: \def \nat#1#2#3{Nature {\bf#1}, #2 (#3)}
686: \def \nc#1#2#3{Nuovo Cim.\ {\bf#1}, #2 (#3)}
687: \def \nima#1#2#3{Nucl.\ Instr.\ Meth.\ A {\bf#1}, #2 (#3)}
688: \def \np#1#2#3{Nucl.\ Phys.\ {\bf#1}, #2 (#3)}
689: \def \npbps#1#2#3{Nucl.\ Phys.\ B Proc.\ Suppl.\ {\bf#1}, #2 (#3)}
690: \def \PDG{Particle Data Group, L. Montanet \ite, \prd{50}{1174}{1994}}
691: \def \pisma#1#2#3#4{Pis'ma Zh.\ Eksp.\ Teor.\ Fiz.\ {\bf#1}, #2 (#3) [JETP
692: Lett.\ {\bf#1}, #4 (#3)]}
693: \def \pl#1#2#3{Phys.\ Lett.\ {\bf#1}, #2 (#3)}
694: \def \pla#1#2#3{Phys.\ Lett.\ A {\bf#1}, #2 (#3)}
695: \def \plb#1#2#3{Phys.\ Lett.\ B {\bf#1}, #2 (#3)}
696: \def \pr#1#2#3{Phys.\ Rev.\ {\bf#1}, #2 (#3)}
697: \def \prc#1#2#3{Phys.\ Rev.\ C {\bf#1}, #2 (#3)}
698: \def \prd#1#2#3{Phys.\ Rev.\ D {\bf#1}, #2 (#3)}
699: \def \prl#1#2#3{Phys.\ Rev.\ Lett.\ {\bf#1}, #2 (#3)}
700: \def \prp#1#2#3{Phys.\ Rep.\ {\bf#1}, #2 (#3)}
701: \def \ptp#1#2#3{Prog.\ Theor.\ Phys.\ {\bf#1}, #2 (#3)}
702: \def \rmp#1#2#3{Rev.\ Mod.\ Phys.\ {\bf#1}, #2 (#3)}
703: \def \rp#1{~~~~~\ldots\ldots{\rm rp~}{#1}~~~~~}
704: \def \si90{25th International Conference on High Energy Physics, Singapore,
705: Aug. 2-8, 1990}
706: \def \slc87{{\it Proceedings of the Salt Lake City Meeting} (Division of
707: Particles and Fields, American Physical Society, Salt Lake City, Utah, 1987),
708: ed. by C. DeTar and J. S. Ball (World Scientific, Singapore, 1987)}
709: \def \slac89{{\it Proceedings of the XIVth International Symposium on
710: Lepton and Photon Interactions,} Stanford, California, 1989, edited by M.
711: Riordan (World Scientific, Singapore, 1990)}
712: \def \smass82{{\it Proceedings of the 1982 DPF Summer Study on Elementary
713: Particle Physics and Future Facilities}, Snowmass, Colorado, edited by R.
714: Donaldson, R. Gustafson, and F. Paige (World Scientific, Singapore, 1982)}
715: \def \smass90{{\it Research Directions for the Decade} (Proceedings of the
716: 1990 Summer Study on High Energy Physics, June 25--July 13, Snowmass, Colorado),
717: edited by E. L. Berger (World Scientific, Singapore, 1992)}
718: \def \tasi{{\it Testing the Standard Model} (Proceedings of the 1990
719: Theoretical Advanced Study Institute in Elementary Particle Physics, Boulder,
720: Colorado, 3--27 June, 1990), edited by M. Cveti\v{c} and P. Langacker
721: (World Scientific, Singapore, 1991)}
722: \def \yaf#1#2#3#4{Yad.\ Fiz.\ {\bf#1}, #2 (#3) [Sov.\ J. Nucl.\ Phys.\
723: {\bf #1}, #4 (#3)]}
724: \def \zhetf#1#2#3#4#5#6{Zh.\ Eksp.\ Teor.\ Fiz.\ {\bf #1}, #2 (#3) [Sov.\
725: Phys.\ - JETP {\bf #4}, #5 (#6)]}
726: \def \zpc#1#2#3{Zeit.\ Phys.\ C {\bf#1}, #2 (#3)}
727: \def \zpd#1#2#3{Zeit.\ Phys.\ D {\bf#1}, #2 (#3)}
728:
729: \begin{thebibliography}{99}
730:
731: \bibitem{nim} K. Green, J. L. Rosner, D. A. Suprun, and J. F. Wilkerson,
732: \nima{498}{256}{2003}.
733: %%CITATION = ASTRO-PH 0205046;%%
734:
735: \bibitem{Allan} H. R. Allan, in {\it Progress in Elementary Particles and
736: Cosmic Ray Physics}, v. 10, edited by J. G. Wilson and S. G.
737: Wouthuysen (North-Holland, Amsterdam, 1971), p. 171, and
738: references therein.
739:
740: \bibitem{Feyn} R. P. Feynman, R. B. Leighton, and M. Sands, {\it The Feynman
741: Lectures in Physics,} Addison-Wesley, Reading, Mass., 1963, Sec.\ I-28.
742:
743: \bibitem{Auger} J. W. Cronin, \rmp{71}{S165}{1998}; \npbps{80}{33}{2000};
744: D. Zavrtanik, \npbps{85}{324}
745: {2000}. For the Pierre Auger Project Design Report see
746: {\tt http://www.ses-ng.si/public/pao/design.html}.
747:
748: \bibitem{RRW} R. R. Wilson, Phys.~Rev. {\bf 108}, 155 (1967).
749:
750: \bibitem{Ask} G. A. Askar'yan, Zh.~Eksp.~Teor.~Fiz.~{\bf 41}, 616 (1961)
751: [Sov.~Phys.--JETP {\bf 14}, 441 (1962)];
752: Zh.~Eksp.~Teor.~Fiz.~{\bf 48}, 988 (1965) [Sov.~Phys.--JETP {\bf
753: 21}, 658 (1965)].
754:
755: \bibitem{KL} F. D. Kahn and I. Lerche, Proc.~Roy.~Soc.~{\bf A 289}, 206
756: (1966).
757:
758: \bibitem{ZHS} E. Zas, F. Halzen, and T. Stanev, Phys.~Rev.~D {\bf 45}, 362
759: (1992).
760:
761: \bibitem{Jelley} J. V. Jelley {\it et al.}, Nature {\bf 205}, 327 (1965);
762: Nuovo Cimento {\bf A46}, 649 (1966); N. A. Porter {\it et al.},
763: Phys.~Lett.~{\bf 19}, 415 (1965).
764:
765: \bibitem{Weekes} T. Weekes, this conference.
766:
767: \bibitem{Sov} S. N. Vernov {\it et al.}, Pis'ma v ZhETF {\bf 5}, 157 (1967)
768: [Sov.~Phys.--JETP Letters {\bf 5}, 126 (1967)];
769: Can.~J.~Phys.~{\bf 46}, S241 (1968).
770:
771: \bibitem{Chac} P. R. Barker, W. E. Hazen, and A. Z. Hendel, Phys.~Rev.~Lett.
772: {\bf 18}, 51 (1967); W. E. Hazen, {\it et al.}, {\it ibid.} {\bf
773: 22}, 35 (1969); {\bf 24}, 476 (1970).
774:
775: \bibitem{Atrash} V. B. Atrashkevich et al., Yad.~Fiz.~{\bf 28}, 366 (1978).
776:
777: \bibitem{CASAnim} A. Borione {\it et al.}, Nucl.~Instrum.~Meth.~A {\bf 346},
778: 329 (1994).
779:
780: \bibitem{Agasa} K. Kadota {\it et al.}, Proc.~23rd International Conference on
781: Cosmic Rays (ICRC-23), Calgary, 1993, v.~4, p.~262; Tokyo
782: Workshop on Techniques for the Study of Extremely High Energy
783: Cosmic Rays, Tanashi, Tokyo, 27 -- 30 Sept. 1993.
784:
785: \bibitem{Yak} P. I. Golubnichii, A. D. Filonenko, and V. I. Yakovlev,
786: Izv. Akad. Nauk {\bf 58}, 45 (1994).
787:
788: \bibitem{GS} C. Castagnoli {\it et al.}, Proc.~ICRC-23, Calgary, 1993,
789: v.~4, p.~258.
790:
791: \bibitem{Gau} R. Baishya {\it et al.}, Proc.~ICRC-23, Calgary, 1993, V.~4,
792: p.~266; Gauhati University Collaboration, paper submitted to this conference.
793:
794: \bibitem{REF} M.~A.~Lawrence, R. J. O. Reid, and A. A. Watson,
795: J.~Phys.~G {\bf 17}, 733 (1991).
796: \bibitem{mutrig} R. Gall and K. D. Green, UMC-CASA note, Aug.~23, 1996
797: (unpublished).
798:
799: \bibitem{DFS} D. F. Sullivan, Master's Thesis, University of Chicago, 1999
800: (unpublished).
801:
802: \bibitem{PG} P. W. Gorham, ``On the possibility of radar echo detection of
803: ultra-high energy cosmic ray- and neutrino-induced extensive air
804: showers,'' hep-ex/0001041, January, 2000 (unpublished).
805:
806: \bibitem{Zas} E. Zas, this conference.
807:
808: \end{thebibliography}
809: \end{document}
810: #!/bin/csh -f
811: # this uuencoded Z-compressed .tar file created by csh script uufiles
812: # for more information, see e.g. http://xxx.lanl.gov/faq/uufaq.html
813: # if you are on a unix machine this file will unpack itself:
814: # strip off any mail header and call resulting file, e.g., rhf.uu
815: # (uudecode ignores these header lines and starts at begin line below)
816: # then say csh rhf.uu
817: # or explicitly execute the commands (generally more secure):
818: # uudecode rhf.uu ; uncompress rhf.tar.Z ;
819: # tar -xvf rhf.tar
820: # on some non-unix (e.g. VAX/VMS), first use an editor to change the
821: # filename in "begin" line below to rhf.tar_Z , then execute
822: # uudecode rhf.uu
823: # compress -d rhf.tar_Z
824: # tar -xvf rhf.tar
825: #
826: uudecode $0
827: chmod 644 rhf.tar.Z
828: zcat rhf.tar.Z | tar -xvf -
829: rm $0 rhf.tar.Z
830: exit
831:
832: