1: %%
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 2005 December 5
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17: %%
18: \documentclass[12pt,preprint]{aastex}
19:
20: %\usepackage{natbib}
21:
22: %% manuscript produces a one-column, double-spaced document:
23:
24: %\documentclass[manuscript]{aastex}
25:
26: %% preprint2 produces a double-column, single-spaced document:
27:
28: %\documentclass[preprint2]{aastex}
29:
30: %% Sometimes a paper's abstract is too long to fit on the
31: %% title page in preprint2 mode. When that is the case,
32: %% use the longabstract style option.
33:
34: %% \documentclass[preprint2,longabstract]{aastex}
35:
36: %% If you want to create your own macros, you can do so
37: %% using \newcommand. Your macros should appear before
38: %% the \begin{document} command.
39: %%
40: %% If you are submitting to a journal that translates manuscripts
41: %% into SGML, you need to follow certain guidelines when preparing
42: %% your macros. See the AASTeX v5.x Author Guide%% for information.
43:
44: \newcommand{\vdag}{(v)^\dagger}
45: \newcommand{\myemail}{jojo@donegal.uchicago.edu}
46:
47: %% You can insert a short comment on the title page using the command below.
48:
49: \slugcomment{Astrophysical Journal}
50:
51: %% If you wish, you may supply running head information, although
52: %% this information may be modified by the editorial offices.
53: %% The left head contains a list of authors,
54: %% usually a maximum of three (otherwise use et al.). The right
55: %% head is a modified title of up to roughly 44 characters.
56: %% Running heads will not print in the manuscript style.
57:
58: \shorttitle{Composition of Primary Cosmic-Ray Nuclei at High Energies}
59: \shortauthors{Ave et al.}
60:
61: %% This is the end of the preamble. Indicate the beginning of the
62: %% paper itself with \begin{document}.
63:
64: \begin{document}
65:
66: %% LaTeX will automatically break titles if they run longer than
67: %% one line. However, you may use \\ to force a line break if
68: %% you desire.
69:
70: \title{Composition of Primary Cosmic-Ray Nuclei at High Energies}
71:
72: %% Use \author, \affil, and the \and command to format
73: %% author and affiliation information.
74: %% Note that \email has replaced the old \authoremail command
75: %% from AASTeX v4.0. You can use \email to mark an email address
76: %% anywhere in the paper, not just in the front matter.
77: %% As in the title, use \\ to force line breaks.
78:
79: \author{M. Ave, P.J. Boyle, F. Gahbauer\altaffilmark{a},
80: C. H\"{o}ppner\altaffilmark{b}, J.R. H\"{o}randel\altaffilmark{c},
81: M. Ichimura\altaffilmark{d}, D. M\"{u}ller, A. Romero
82: Wolf\altaffilmark{e}}
83:
84: \affil{Enrico Fermi Institute, The University of Chicago, \\933 E 56th
85: Street, Chicago, IL 60637, USA}
86:
87: %% Notice that each of these authors has alternate affiliations, which
88: %% are identified by the \altaffilmark after each name. Specify alternate
89: %% affiliation information with \altaffiltext, with one command per each
90: %% affiliation.
91:
92: \altaffiltext{a}{Currently at: University of Latvia, Latvia.}
93: \altaffiltext{b}{Currently at: Technische Universit\"{a}t M\"{u}nchen, Germany.}
94: \altaffiltext{c}{Currently at: Radboud University Nijmegen, The Netherlands.}
95: \altaffiltext{d}{Currently at: Hirosaki University, Japan.}
96: \altaffiltext{e}{Currently at: University of Hawaii, USA.}
97:
98: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
99: %% style, abstract will output a Received/Accepted line after the
100: %% title and affiliation information. No date will appear since the author
101: %% does not have this information. The dates will be filled in by the
102: %% editorial office after submission.
103:
104: \begin{abstract}
105: The TRACER instrument (``Transition Radiation Array for Cosmic
106: Energetic Radiation'') has been developed for direct measurements of
107: the heavier primary cosmic-ray nuclei at high energies. The instrument
108: had a successful long-duration balloon flight in Antarctica in
109: 2003. The detector system and measurement process are described,
110: details of the data analysis are discussed, and the individual energy
111: spectra of the elements O, Ne, Mg, Si, S, Ar, Ca, and Fe (nuclear
112: charge Z=8 to 26) are presented. The large geometric factor of TRACER
113: and the use of a transition radiation detector make it possible to
114: determine the spectra up to energies in excess of 10$^{14}$ eV per
115: particle. A power-law fit to the individual energy spectra above 20
116: GeV per amu exhibits nearly the same spectral index ($\sim$ 2.65 $\pm$
117: 0.05) for all elements, without noticeable dependence on the elemental
118: charge Z.
119:
120: \end{abstract}
121: %% Keywords should appear after the \end{abstract} command. The uncommented
122: %% example has been keyed in ApJ style. See the instructions to authors
123: %% for the journal to which you are submitting your paper to determine
124: %% what keyword punctuation is appropriate.
125:
126: \keywords{General : Acceleration of particles and Cosmic rays ---
127: methods: data analysis --- ISM: abundances}
128:
129: %% From the front matter, we move on to the body of the paper.
130: %% In the first two sections, notice the use of the natbib \citep
131: %% and \citet commands to identify citations. The citations are
132: %% tied to the reference list via symbolic KEYs. The KEY corresponds
133: %% to the KEY in the \bibitem in the reference list below. We have
134: %% chosen the first three characters of the first author's name plus
135: %% the last two numeral of the year of publication as our KEY for
136: %% each reference.
137:
138:
139: %% Authors who wish to have the most important objects in their paper
140: %% linked in the electronic edition to a data center may do so by tagging
141: %% their objects with \objectname{} or \object{}. Each macro takes the
142: %% object name as its required argument. The optional, square-bracket
143: %% argument should be used in cases where the data center identification
144: %% differs from what is to be printed in the paper. The text appearing
145: %% in curly braces is what will appear in print in the published paper.
146: %% If the object name is recognized by the data centers, it will be linked
147: %% in the electronic edition to the object data available at the data centers
148: %%
149: %% Note that for sources with brackets in their names, e.g. [WEG2004] 14h-090,
150: %% the brackets must be escaped with backslashes when used in the first
151: %% square-bracket argument, for instance, \object[\[WEG2004\] 14h-090]{90}).
152: %% Otherwise, LaTeX will issue an error.
153:
154: \section{Introduction}
155: \label{sec:intro}
156:
157: The energies of cosmic rays observed near earth extend over a very
158: wide range, from below 10$^8$ eV to more than 10$^{20}$ eV per
159: particle. Up to about 10$^{10}$ eV per amu, the particle energies and
160: intensities are significantly affected by solar modulation, but for
161: the nearly ten remaining decades, the arriving particles are believed
162: to represent the ambient cosmic-ray population in the local galaxy.
163: Over this range, the total cosmic-ray intensity (the ``all particle''
164: differential energy spectrum, see, for instance, the compilation of
165: \citet{cronin97}) decreases monotonically by more than 30 orders of
166: magnitude, initially as a power law $\propto$ E$^{-2.7}$, steepening
167: slightly to E$^{-3.0}$ at the ``knee'' above 10$^{15}$ eV per
168: particle, and exhibiting small changes in slope again (a ``second
169: knee'' and an ``ankle'') in the 10$^{18}$ -- 10$^{19}$ eV region.
170: Besides these relatively minor changes, the overall spectrum is
171: remarkably featureless, even though a variety of processes may
172: contribute to the cosmic-ray flux. The general consensus is that the
173: particles below the knee are generated inside our Galaxy and are
174: contained by the galactic magnetic fields for millions of years, while
175: at the highest energies ($\geq$ 10$^{18}$ eV) the gyration radii
176: become so large that galactic containment is no longer effective, and
177: that the cosmic rays may then be of extragalactic origin.
178:
179: The all-particle energy spectrum refers to the overall spectrum of
180: cosmic rays without differentiating the individual components. To
181: obtain deeper insight, details of the composition of the cosmic-ray
182: particles must be studied. For instance, very accurate measurements
183: of not only the elemental, but also the isotopic composition are now
184: available at low energies ($\leq$ 10$^{9}$ -- 10$^{10}$ eV per
185: particle). These have determined such important quantities as the
186: average containment time of cosmic rays in the galaxy from
187: observations of radioactive clock-nuclei ($\tau \approx 1.5 \times
188: 10^{7}$ y at these energies (first reported by \citet{gmunoz75}; more
189: recent work by \citet{yanasak01}), and they have excluded fresh
190: supernova ejecta as source material of cosmic rays from the abundances
191: of Co- and Ni-isotopes \citep{wiedenbeck99}. At relativistic energies
192: only the abundances of the elemental species, without isotopic detail,
193: have been accessible to measurements (e.g. \citet{engelmann90}), but
194: even these encounter increasingly severe systematic and statistical
195: uncertainties as the energy increases. Very few direct observations
196: (e.g. \citet{muller91}) have provided composition detail with
197: single-element resolution above ~10$^{13}$ eV per particle. The work
198: described in this paper represents an attempt to improve this
199: situation, and to determine the energy spectra of individual elemental
200: species up to much higher energies.
201:
202: The current paradigm of the origin of galactic cosmic rays postulates
203: first-order Fermi acceleration in interstellar shock fronts from
204: supernova (SN) explosions as the most potent contributor to the
205: cosmic-ray flux below the knee. This mechanism, first proposed by
206: \citet{bell78}, must be very efficient indeed, since as much as $\sim$10\%
207: of the total kinetic energy released in supernova explosions is
208: required to sustain the cosmic-ray flux. The SN shock acceleration
209: mechanism predicts a source energy spectrum in the form of a power law
210: $\propto$ $E^{-\Gamma}$, with a spectral index $\Gamma \approx 2.0$
211: for strong shocks. Such a spectrum is much harder than the $E^{-2.7}$
212: spectrum that seems to be typical for most primary cosmic-ray nuclei
213: below the knee. This difference in spectral slope can be explained if
214: the propagation and containment of cosmic rays in the galaxy are
215: energy-dependent. Such a behavior had indeed been inferred, even
216: before the SN shock acceleration model was formulated, on the basis of
217: measurements of the ``L/M ratio'', i.e., of the abundance of
218: secondary, spallation-produced cosmic rays (such as the light elements
219: Li, Be, B) relative to their primary parents (such as C and O),
220: \citep{juliusson72}. One concludes from such measurements that the
221: average amount of interstellar matter encountered by cosmic-ray nuclei
222: during their propagation through the galaxy decreases with energy. The
223: effect is often parameterized by a propagation pathlength $\Lambda$,
224: which depends on energy as $\Lambda \propto E^{-0.6}$ for relativistic
225: nuclei \citep{engelmann90}. The primary cosmic-ray spectrum at the
226: source would then be approximately $\propto E^{-2.1}$ or $E^{-2.2}$,
227: close to what the shock acceleration model predicts. This fact
228: provides strong but indirect evidence for the validity of the shock
229: acceleration model.
230:
231: However, it must be kept in mind that current measurements of the L/M
232: ratio do not extend beyond 10$^{11}$ eV per amu (or about 10$^{12}$ eV
233: per particle). Hence, there is an extended region of energies
234: (10$^{12}$ -- 10$^{15}$ eV per particle) where the SN shock
235: acceleration model is assumed to be valid, albeit without the benefit
236: of much observational support. The SN shock acceleration process is
237: expected to become inefficient at energies around $Z \times$ 10$^{14}$
238: eV (where $Z$ is the atomic number of a cosmic-ray nucleus)
239: \citep{lagage83}, but this limit is not observationally confirmed.
240: One might refer to the steepening of the all-particle spectrum at the
241: "knee" as evidence, but even then the origin of cosmic rays with
242: energies far beyond the knee remains a mystery.
243:
244: The SN-shock acceleration model has recently found strong
245: observational support through the detection of TeV gamma rays from
246: shell-type SN-remnants \citep{aharonian04}, although there remains
247: a debate whether the parent particles of these gamma rays are electrons
248: or, indeed, protons and nuclei \citep{berezhko06}.
249:
250: Clearly, the present observational evidence about high-energy cosmic
251: rays is inadequate to provide answers to some of the most fundamental
252: questions about their origin and galactic propagation, and there
253: exists an undeniable need for improved measurements at higher
254: energies. The TRACER program has been developed to provide direct
255: measurements of the energy spectra of the heavier cosmic-ray nuclei
256: through balloon flights above the atmosphere, up to energies between
257: 10$^{14}$ and 10$^{15}$ eV per particle [The acronym TRACER stands for
258: ``Transition Radiation Array for Cosmic Energetic Radiation'']. Higher
259: energies are currently studied through indirect observations with
260: ground-based air-shower installations. However, attempts to obtain
261: composition details from air-shower observations are affected by a
262: number of systematic uncertainties. One of the goals of TRACER is to
263: approach the energy region of air shower measurements, with the hope
264: of providing some cross-calibration with the indirect measurements.
265:
266:
267:
268:
269: \section{Observational Technique}
270: \label{sec:technical}
271:
272: \subsection{General Principles}
273:
274:
275: A successful cosmic-ray measurement above the atmosphere must
276: determine at least two quantities for each particle: the charge $Z$ and
277: the energy $E$. Measurement of $Z$ is commonly accomplished by utilizing
278: the fact that all electro-magnetic interactions scale with $Z^{2}$.
279: Hence, devices that measure the ionization energy loss (dE/dx) such as
280: plastic scintillators or gas proportional counters, or that measure
281: the intensity of Cherenkov light, are well suited and are used in
282: TRACER.
283:
284: The energy determination is a greater challenge, mainly because of the
285: large detector areas that are needed at high energy. Extrapolations
286: from lower energies indicate that a successful measurement of nuclei
287: heavier than helium requires detectors with exposure factors of at
288: least several 100 m$^{2}$ sr days in order to approach the 10$^{15}$
289: eV per particle region. Such exposures can be accomplished with
290: long-duration (i.e., several weeks) balloon flights if the detectors
291: have sensitive areas of a few square meters. A classic calorimeter
292: that measures the total energy absorbed in matter after a particle
293: interacts would become much too massive if it were to have an area of
294: this magnitude. Hence, TRACER uses electromagnetic interactions not
295: just to determine $Z$, but also to measure the energy (or more
296: exactly, the Lorentz-factor $\gamma \approx E/mc^{2}$): measurement
297: of Cherenkov light produced in an acrylic Cherenkov counter at low
298: energies (up to a few 10$^{9}$ eV per amu), measurement of the
299: specific ionization in gases at intermediate energies (10$^{10}$ to
300: 10$^{12}$ eV per amu), and measurement of transition radiation at the
301: highest energies ($>$ 5 $\times$ 10$^{11}$ eV per amu).
302:
303: Figure \ref{resp} shows the typical energy responses for the detectors
304: used in TRACER. The light intensity of an acrylic Cherenkov counter
305: (refractive index = 1.49) increases quickly with energy above a
306: threshold of 0.325 GeV per amu ($\gamma \approx$ 1.35) until it
307: approaches saturation above 20 GeV per amu ($\gamma \approx$ 22). The
308: ionization loss in gas decreases with energy, reaches a minimum at
309: $\gamma$ = 3.96, and then increases logarithmically with $E$.
310: Transition radiation (TR) x-rays are generated in a multi-surface
311: radiator (mats of plastic fibers in TRACER) and are detected in a gas
312: proportional counter. Hence, the TR detector measures an ionization
313: loss signal, and at sufficiently high energies ($>$ 5 $\times$
314: 10$^{11}$ eV per amu), a superimposed signal due to the absorption of
315: TR photons. The TR signal rises rapidly with the Lorentz factor
316: $\gamma$ and reaches saturation in the $\gamma$ = 10$^{4}$ -- 10$^{5}$
317: region. The use of TR to measure the highest cosmic-ray energies was
318: first implemented for the CRN detector in 1985 \citep{lheureux90}, but
319: remains unconventional. It permits the construction of detectors of
320: very large area. As the TR intensity increases with $Z^{2}$, this
321: technique is well suited for measurements of the heavier cosmic rays
322: ($Z >$ 3), but not applicable to protons or He-nuclei where the
323: signals are affected by large statistical fluctuations.
324:
325: Note that the response curves shown in figure \ref{resp} for the
326: detectors of specific ionization and TR are double-valued: large
327: signals can either be produced by high-energy particles, or by
328: particles below minimum ionization. The TRACER instrument must be
329: launched at latitudes with low geomagnetic cutoff energies where a
330: significant cosmic-ray intensity arrives with sub-relativistic
331: energies. The inclusion of a plastic Cherenkov counter which
332: identifies the low-energy particles is then essential in order to
333: remove the degeneracy of ionization and TR response. Overall, the
334: counter combination chosen for TRACER measures the energies of
335: cosmic-ray nuclei over a very wide range, in excess of four orders of
336: magnitude.
337:
338:
339:
340:
341: \subsection{The TRACER Instrument}
342:
343: A schematic cross-section of the TRACER instrument is shown in figure
344: \ref{tracer}. The detector elements are rather large in area (206 cm
345: $\times$ 206 cm), and are vertically separated by 120 cm. This leads
346: to an overall geometric factor of about 5 m$^{2}$sr. The instrument
347: has been previously described \citep{gahbauer04}, and a detailed
348: technical description of the detector elements is currently being
349: prepared for a separate publication. Here, we will present a brief
350: summary of the relevant properties of the instrument.
351:
352: As figure \ref{tracer} shows, the individual components of TRACER are,
353: from top to bottom:
354:
355: 1. A plastic scintillator sheet (BICRON-408 material), with active
356: area 200 cm $\times$ 200 cm, and 0.5 cm thick\footnote[1]{Due to
357: the interspersed wavelength shifter-bars, the geometric area of
358: these counters is slightly larger than the active area.}. The
359: scintillator is viewed by 24 photomultiplier tubes (PMT) via
360: wavelength shifter bars.
361:
362: 2. Four double layers of single-wire proportional tubes, oriented in
363: two orthogonal directions. Each tube is 200 cm long and 2 cm in
364: diameter. The tube walls are wound from aluminized mylar and are
365: about 127 $\mu$m thick.
366:
367: 3. Another four double layers of proportional tubes of identical
368: design, but each double layer located below blankets of plastic
369: fiber material, which form a radiator to generate transition
370: radiation.
371:
372: All tubes are filled with a mixture of xenon and methane (at equal
373: parts by volume) and are operated in flight at a pressure of 0.5
374: atmospheres.
375:
376: 4. A second scintillator of identical design, but read out by just 12
377: PMTs, and located below the lowest layer of proportional tubes.
378:
379: 5. A Cherenkov counter, with active area 200 cm $\times$ 200 cm and
380: 1.27 cm thick, at the bottom of the detector stack. The Cherenkov
381: material consists of an acrylic plastic (refractive index = 1.49),
382: doped with wavelength shifter material. Signals are read out via
383: wavelength-shifter bars connected to 24 PMTs (as for the scintillator
384: on top).
385:
386: The two scintillators serve as triggers for the instrument and,
387: together with the Cherenkov counter, determine the elemental charge
388: $Z$ of individual cosmic-ray nuclei traversing the instrument. The
389: proportional counter arrays measure the energy (or Lorentz-factor $
390: \approx E/mc^{2}$) of the cosmic-ray particles at high energies, and
391: they determine the trajectory of each particle traversing the
392: instrument in two projections. The signals in the top four double
393: layers simply are a measure for the ionization energy loss ($dE/dx$)
394: of the particle. The small logarithmic increase above the
395: minimum-ionization level provides an estimate of the particle energy
396: up to Lorentz factors of a few hundred. The lower four double layers
397: of tubes form a transition radiation detector (TRD). As the response
398: curves in figure \ref{resp} show, the TRD signals are identical to the
399: dE/dx signals of the tubes above, except at the highest energies
400: ($\gamma$ $>$ 400), where the rapid increase of the signal with
401: increasing $\gamma$ permits an energy measurement. The signals of the
402: proportional tubes (a total of about 1600) are processed through VLSI
403: chips (``Amplex'', \citet{beuville90}). The finite dynamic range of
404: these devices restricts the range of elemental species for which data
405: could be recorded in this flight, from oxygen ($Z$ = 8) to iron ($Z$ =
406: 26).
407:
408: The TRACER instrument derives its heritage from the CRN detector that
409: was developed by our laboratory in the 1980s for space flight
410: \citep{lheureux90}. In fact, the radiator material used for the
411: TRACER-TRD is identical to that of CRN. However, CRN employed
412: thin-window multiwire proportional chambers to detect transition
413: radiation, and hence, needed a heavy pressurized container. The
414: proportional tubes used in TRACER instead can easily withstand
415: external low pressure; and thus, a pressurized gondola is not required
416: and not used for TRACER.
417:
418: \section{Balloon Flight and Instrument Performance}
419: \label{sec:inst}
420: TRACER had first a one-day balloon flight in Fort Sumner, New Mexico,
421: in 1999. The results from this flight have been published
422: \citep{gahbauer04}. A circumpolar long-duration balloon (LDB) flight
423: was originally planned for a flight in the northern hemisphere with a
424: launch from Alaska, but could not be conducted because of the lack of
425: relevant international agreements. Therefore, an LDB flight had to be
426: planned in Antarctica, but had to wait until 2003, when a new launch
427: vehicle commensurate with the weight of TRACER (6,000 lbs., including
428: ballast, telemetry, and balloon-related instrumentation) became
429: operational. TRACER was launched near McMurdo, Antarctica, on December
430: 12, 2003. During its two-week flight (see figure~\ref{map}) at an
431: average height of 37.75 km (3.9 g/cm$^2$) with excursions not
432: exceeding $\pm$ 1.75 km, it collected data from a total of 5 $\times$
433: 10$^{7}$ cosmic-ray events. The data were transmitted at 500 Mbit/sec
434: to the ground (when the instrument was within telemetry range) and
435: stored on six disks on-board and recovered after termination of the
436: flight.
437:
438: Overall, the instrument performed very well during this flight. In
439: particular, the entirely passive thermal insulation kept the
440: temperatures of all detector components close to room temperature,
441: with diurnal variations of at most a few degrees C. There was no
442: indication of a deterioration of the performance of the proportional
443: tubes due to gas poisoning for the duration of the flight. This result
444: would be of great practical importance if flights of such devices for
445: much longer duration (for instance in space) are anticipated: there
446: seems to be no need for gas-purging at the time scales of at least a
447: few weeks, except for the provision of make-up gas to correct for
448: minor gas leaks. The data acquisition system operated very
449: efficiently, with a dead time of just 6\%, but the total time during
450: which data could be recorded was limited to ten days, due to a
451: catastrophic failure of the rechargeable lithium batteries (which were
452: used to buffer the solar-power system) towards the end of the first
453: orbit of the balloon around the South Pole.
454:
455:
456:
457: \section{Data Analysis}
458: \label{sec:analysis}
459:
460: \subsection{Trajectory Reconstruction}
461: \label{sec:trajectory}
462: The first step in the analysis of the data is the accurate
463: reconstruction of the trajectory of each cosmic-ray particle through
464: the instrument. The entire array of proportional tubes is used for
465: this purpose. It consists of four double layers of tubes in each the
466: x- and y- direction. Hence, ideally a trajectory is characterized by
467: eight tube ``hits'' in each projection. In practice, sometimes the
468: particle may pass between adjacent tubes of a given layer and not
469: generate a hit in this layer. In addition, spurious signals in tubes
470: outside the particle trajectory may be recorded.
471:
472: As a first estimate, the path is obtained using the center of each of
473: the tubes hit in an event. All possible combinations of these tubes
474: are fit with straight lines in the $X$ and $Y$ projection, and the
475: combination in each projection with minimum $\chi^2$ per degree of
476: freedom and maximum number of tubes used is kept. This method allows
477: an accuracy of 5 mm in track position. The procedure is quite
478: efficient: less than 5~\% of cosmic-ray nuclei are missed in the
479: process.
480:
481: A preliminary estimate of the charge of each particle is then
482: performed using the technique described in the following section. A
483: subset of the data from which we expect the same average signal in any
484: tube (high energy $O$ or $Ne$ nuclei) is then used to correct for
485: deviations of the tube positions from their ``ideal'' positions, and
486: to match signal gains within 5~\% or less. Corrections due to
487: variations in the tube gain with time either due to temperature
488: changes or small gas leakages are also calculated and applied. In
489: general, gain variations in a given tube are negligible \emph{along}
490: that tube. Finally, the tube signals are corrected to account for
491: minor non-linearities in the front end electronics at large
492: signals. All these corrections are necessary to reduce the systematic
493: uncertainties in the tube signals below the level of unavoidable
494: physical fluctuations (``Landau fluctuations'') in the energy
495: deposited in the gas.
496:
497: As a final step, the trajectory is refined by using the fact that the
498: energy deposit in each tube is proportional to the track length in
499: that tube. Taking this fact into account, an accuracy of 2 mm in the
500: lateral track position is achieved, which corresponds to a 3~\%
501: uncertainty in the total pathlength through all the tubes. The
502: accuracy and efficiency of this method are verified with a GEANT4
503: simulation \citep{agostinelli03} of the instrument.
504:
505: The accurate knowledge of the particle trajectories is essential to
506: the analysis for two reasons: first, it permits corrections of the
507: scintillator and Cherenkov signals due to spatial non-uniformities and
508: zenith-angle variations (see section \ref{sec:charge}). Second, it
509: makes possible to determine accurately the average energy deposit of a
510: cosmic-ray particle when it traverses the tubes. Specifically, the
511: average signals are as follows:
512:
513: In the top four double layers, the average specific ionization is defined as
514:
515: \vspace{0.1cm}
516: \begin{center}
517: \begin{equation}
518: \label{dedxeq}
519: \langle \frac{dE}{dx} \rangle = \frac{\sum_{i=1}^{i=8} \Delta{E}_i}{\sum_{i=1}^{i=8} \Delta{x}_i}
520: %\frac{dE}{dx}+TRD = \frac{\sum_{i=9}^{i=16} dE_i}{\sum_{i=0}^{i=8} dx_i}
521: \end{equation}
522: \end{center}
523: %\vspace{0.2cm}
524:
525: and in the bottom four double layers (which form the TRD), the average
526: signal is
527:
528: \begin{center}
529: \begin{equation}
530: \label{dedxtreq}
531: %\frac{dE}{dx} = \frac{\sum_{i=1}^{i=8} dE_i}{\sum_{i=0}^{i=8} dx_i}
532: \langle \frac{dE}{dx}+TR \rangle = \frac{\sum_{i=9}^{i=16} \Delta{E}_i}{\sum_{i=9}^{i=16} \Delta{x}_i}
533: \end{equation}
534: \end{center}
535: \vspace{0.1cm}
536:
537: Here, $\Delta{E}_{i}$ and $\Delta{x}_{i}$ refer to energy deposit and
538: pathlength, respectively, in tube number $i$. The two quantities defined
539: in (\ref{dedxeq}) and (\ref{dedxtreq}) are, of course, identical
540: within fluctuations at energies where no transition radiation is
541: detectable ($TR = 0$). In order to exclude large fluctuations
542: associated with short pathlengths, we remove tubes with $\Delta{x}_{i}
543: <$ 1 cm from the summation.
544:
545: \subsection{Charge Analysis}
546: \label{sec:charge}
547:
548: As a first step to determine the charge of each particle, the signal
549: recorded by each PMT is corrected to account for the spatial
550: non-uniformities in the counter responses. These corrections are
551: derived using the tracking information together with response maps
552: recorded with muons at ground and verified with the flight data
553: themselves in an iterative procedure. Subsequently, for each event the
554: average signal of all PMTs in each counter is calculated, and then
555: normalized to vertical incidence. Figure \ref{CerScint} shows a
556: scatter plot of the $top$ scintillator signal vs the Cherenkov
557: signal. It is apparent that particles with the same charge are
558: clustered along lines. The position along the line depends on the
559: primary energy: a concentration of events around maximum Cherenkov and
560: minimum scintillation signals corresponds to energies of minimum
561: ionizing particles and above, and the energies decrease for events
562: towards the left of this concentration.
563:
564: The scintillator signals depend on particle charge as $Z^{1.65}$, due
565: to non-linear effects in the light yield, while the Cherenkov signals
566: follow the expected $Z^2$ dependence. The residual scintillation in
567: the Cherenkov counter is visible in figure \ref{CerScint}, and is
568: subtracted out in the analysis. Figure \ref{Zhisto} shows a charge
569: histogram for all charges obtained from summing along lines of
570: constant charge in figure \ref{CerScint}. The charge resolution
571: evolves from 0.25 charge units for $Z$=8 ($O$ nuclei) to 0.5 charge
572: units for $Z$=26 ($Fe$ nuclei) at low energies ($<$ 10 GeV per amu),
573: and increases slightly to 0.3 and 0.6 charge units, respectively, at
574: the highest energies.
575:
576: Note that figure \ref{Zhisto} is generated just on the basis of the
577: signals of the top scintillator, and thus depicts a histogram of
578: $Z_{top}$. One can, in the same fashion, construct a histogram of
579: $Z_{bot}$, using the signals of the bottom scintillator. For the
580: further analysis, we require that $Z_{top}$ and $Z_{bot}$ are
581: consistent within fluctuations for each accepted event. In practice,
582: this means that particles with $Z_{bot} < Z_{top}$ are rejected. In
583: almost all cases these are particles that have undergone a nuclear
584: spallation while traversing the detector material between top and
585: bottom of the instrument (for numerical detail, see section 5 and
586: figure \ref{interactions}).
587:
588: The procedure to select a sample of a given element for further
589: analysis is as follows: first, with a very tight cut on $Z_{bot}$,
590: clean, penetrating nuclei are selected, and their distribution in
591: $Z_{top}$ is investigated. This yields the values for charge
592: resolution mentioned above, and it indicates the efficiency if a
593: charge cut on $Z_{top}$ is made. If the cut on $Z_{top}$ is of the
594: order of $\Delta{Z}_{top} \ge \pm 0.5$ charge units, the efficiency is
595: quite high. For certain elements, a tighter cut on $Z_{top}$ must be
596: made, and the efficiency is then reduced. Numerical details are given
597: in section 5 and table \ref{tab:eff}. It should be noted that the
598: tight cut on $Z_{bot}$ is only applied in order to determine the form
599: of the charge distribution on $Z_{top}$. For the final analysis, only
600: a very loose cut on $Z_{bot}$ is used to remove interacting
601: particles. This cut leads to a negligible reduction in overall charge
602: selection efficiency.
603:
604: Finally, it should be noted that ``edge'' effects can compromise the
605: charge resolution. To avoid this problem, tracks with an impact point
606: within 1 cm of the wavelength-shifter bars are excluded. This
607: creates a dead area in the detector of about 4 \% of the total area.
608:
609: \subsection{Energy Analysis}
610: \label{sec:energy}
611:
612: \subsubsection{Overview}
613:
614: The energy of each cosmic-ray nucleus is obtained from the combined
615: signals of the Cherenkov counter and of the proportional tubes. The
616: top four double layers of tubes measure the specific ionization of
617: each particle, $\langle dE/dx \rangle$, while the bottom four double
618: layers detect transition radiation, measuring $\langle dE/dx + TR
619: \rangle$. While the main objective of the Cherenkov counter is to
620: identify low energy particles, below the level of minimum ionization,
621: the rapid increase of the signal with energy provides a good energy
622: measurement up to about 3 GeV per amu.
623:
624: At energies between minimum ionization and the onset of TR (3 to 400
625: GeV per amu) the signals in the dE/dx tubes and the TR tubes are the
626: same and increase logarithmically with energy. This increase is slow
627: and can be used for an energy measurement unless it is obscured by
628: statistical signal fluctuations. The relative level of the
629: fluctuations decreases with the nuclear charge as $1/Z$. Therefore,
630: the energy measurement improves with increasing $Z$. Note that each
631: individual proportional tube provides an estimate for the specific
632: ionization by measuring the ratio
633: $\Delta{E}_{i}/\Delta{x}_{i}$. Hence, for any particle up to 16 tubes
634: provide independent measurements, which are required to be consistent
635: within fluctuations. For the final analysis, the average $\langle
636: dE/dx \rangle$, as defined in equation \ref{dedxeq} and
637: \ref{dedxtreq}, is used.
638:
639: At the highest energies ($>$ 400 GeV per amu), the signals from the
640: dE/dx tubes and from the TR tubes diverge, and the rapid increase of
641: the TR signals with particle energy provides an excellent energy
642: measurement. However, particles in this energy region are extremely
643: rare, less abundant than particles in the minimum-ionization region by
644: more than four orders of magnitude. To uniquely identify these
645: particles, it is necessary but not sufficient to just require that
646: low-energy nuclei are rejected on the basis of their Cherenkov
647: signals. In addition, we require that the measurement of their
648: ionization energy loss $\langle dE/dx \rangle$ (performed with four
649: double layers of dE/dx tubes) places them at an energy level well
650: above minimum ionization (for details, see section~\ref{sec:hep} and
651: table~\ref{tab:flux1}). Thus, the combination of dE/dx tubes with the
652: TRD tubes is crucial for the success of the TRACER measurement at the
653: highest energies.
654:
655:
656: \subsubsection{Identification of Sub-Relativistic Particles ($<$ 3 GeV/amu)}
657: %\subsection{Identification of Low Energy Particles }
658: \label{sec:lep}
659:
660: To identify low energy particles, the signals from the Cherenkov
661: counter are compared with those of the dE/dx tubes. The dE/dx
662: response, shown in figure \ref{resp}, is well described by the
663: Bethe-Bloch formula. The Cherenkov counter response (figure
664: \ref{resp}) must be slightly modified to account for effects of delta
665: rays generated by the primary particle in the instrument, while the
666: signals from the proportional tube array remain unaffected by delta
667: rays. These effects are well understood and have been previously
668: studied and reported \citep{gahbauer03b,romero05b}. The scatter plot
669: in figure \ref{CerdEdx} shows the correlation of the ionization and
670: Cherenkov signals measured for iron nuclei. The solid line represents
671: the average response expected from simulations. The Cherenkov signal
672: increases with energy until it reaches saturation, while the
673: ionization signal clearly exhibits the level of minimum ionization
674: ($\sim$ 3 GeV per amu). Note that the gray scale of the scatter plot
675: is logarithmic. Plots like this one for each charge exhibit the level
676: of fluctuations in the Cherenkov signal: it is about 8 \% for O nuclei
677: and 2.5 \% for Fe nuclei. A cut in the Cherenkov signal is placed at a
678: level that corresponds to the minimum in the ionization signal. Events
679: with Cherenkov signals below this cut are used to generate their
680: energy spectrum from 0.8 to 2.3 GeV per amu. Events with Cherenkov
681: signals above the cut are used to construct the energy spectra above
682: 12 GeV per amu as described in the following section.
683:
684:
685:
686: \subsubsection{High Energy Particles}
687:
688: %\subsubsection{Identifying the Highest Energy Particles ($>$ 3 GeV/amu)}
689: \label{sec:hep}
690:
691: After removing all particles below about 3 GeV per amu, the rare high
692: energy particles are identified from the combination of the signals
693: from the dE/dx tubes and the TRD. These detectors have been calibrated
694: at accelerators using singly charged particles. Figure \ref{TRcurve}
695: shows the measured response for the TRD of the CRN instrument
696: \citep{lheureux90}. The radiator combination and gas mixtures of
697: TRACER are identical to those of the CRN instrument. However, CRN used
698: plane multiwire proportional chambers rather than the layers of
699: cylindrical proportional tubes of TRACER. We have ascertained in
700: Monte-Carlo simulations that the accelerator calibrations for CRN remain
701: valid for the detector geometry of TRACER. Most of the data shown in
702: figure \ref{TRcurve} have been published previously
703: \citep{lheureux90}.
704:
705: Below 400 GeV/amu we expect no observable transition radiation. Both
706: signals, $\langle dE/dx \rangle$ and $\langle dE/dx + TR \rangle$,
707: increase logarithmically with energy and will cluster along the
708: diagonal in a correlation plot of TRD response vs ionization response
709: as illustrated in figure \ref{TRDvsDEDX}. Above 400 GeV per amu, TR becomes
710: observable, and the signal from the TRD will increase above the
711: ionization signal. This manifests itself as a deviation from the
712: diagonal in the correlation plot.
713:
714: %% FIGURES 7 and 8
715:
716: As an example, figure \ref{TRdEdx} shows the observed cross
717: correlation between the TRD and the ionization signals for neon nuclei
718: ($Z$ = 10) above minimum ionization. The small black points represent
719: the numerous events with energies below the onset of TR. The rare high
720: energy particles with clear TR signals are highlighted. As expected,
721: the data follow the response illustrated in figure
722: \ref{TRDvsDEDX}. Note that the highest energy events (the $TR$
723: events) stand out without any background in other regions of the
724: scatter plot. The most energetic neon nucleus in this sample of data
725: has an energy of $6 \times 10^{14}$eV.
726:
727: We must emphasize again that the off-diagonal position of these higher
728: energy events defines them uniquely as TR-events for the selected
729: charge (Z=10 in figure \ref{TRdEdx}). Any spillover due to
730: misidentified charges, with Z either larger or smaller than the
731: selected value, would lead to signals along the diagonal, but would
732: not contaminate the well defined ``TR-tails''.
733:
734:
735: \section{Absolute Flux Measurements}
736: \label{sec:flux}
737:
738: %The overall geometric factor of TRACER is 5.04 m$^2$ sr, and the total
739: %exposure time is 787200 seconds (9 days 3 hours).
740:
741: Each event that passes the data analysis cuts is classified in energy
742: as either as a \emph{Cherenkov Event} ($<$ 3 GeV per amu),
743: \emph{dE/dx Event} (10 - 400 GeV per amu) or \emph{TR Event} ($>$ 400
744: GeV per amu). Events are sorted into energy bins of width
745: $\Delta{E}_i$ and a differential energy spectrum is constructed for
746: each elemental species. We present the spectra in terms of an absolute
747: flux $dN/dE$ at the top of the atmosphere. To convert from the number of
748: events $\Delta{N}_i$ in a particular energy bin $\Delta{E}_i$ to an
749: absolute differential flux $\mathrm{d}N/\mathrm{d}E (i)$ one must
750: compute the exposure factor, effective aperture, efficiency of the
751: cuts, and unfold the instrument response:
752:
753: \begin{equation}\label{dNdE}
754: \frac{\mathrm{d}N}{\mathrm{d}E}(i)= \frac{\Delta N_i}{\Delta E_i} \cdot \frac{1}{T_l} \cdot \frac{1}{\varepsilon_i} \cdot \frac{1}{A_i} \cdot C_i
755: %\frac{\mathrm{d}N}{\mathrm{d}E}(i)= \frac{\Delta N_i C_i}{\Delta E_i T_l \varepsilon_i A_i}
756: %\frac{overlaps_i \cdot N_i}{Exposure \cdot Aperture}\cdot \Delta{E^{-1}}
757: \end{equation}
758:
759: with $\Delta{N}_i$ the number of events, $\Delta{E}_i$ the energy
760: range, $T_l$ the live-time, $\varepsilon_i$ the efficiency of analysis
761: cuts, $A_i$ the effective aperture and $C_i$
762: the ``overlap correction'' due to misidentified events from
763: neighbouring energy bins. For Cherenkov events one also must take into
764: account that $\Delta{E}_i$, the energy interval on top of the
765: atmosphere, is not equal to the measured energy interval because of
766: energy losses in the atmosphere and in the detector.
767:
768: \emph{Livetime} --- The live time is determined by comparing the raw
769: trigger rate with the number of events stored on disk. With a raw
770: trigger rate of 60 Hz, the dead time was negligible throughout the
771: flight, and the total live time is 787,200 seconds.
772:
773: \emph{Tracking Efficiency} --- The efficiency of the trackfit is 95 \%
774: for all charges, and is independent of energy. Included is a consistency check in the signal of the
775: tubes contributing to the $\langle dE/dx \rangle$ and $\langle dE/dx
776: + TR \rangle$ measurements.
777:
778: \emph{Charge Efficiency} --- For the more abundant elements (O, Ne,
779: Mg, Si) a charge cut of $Z_{top} \pm$ 0.5 is made. The efficiencies
780: range from 85 \% to 72 \% (O, Si) and are listed in table
781: \ref{tab:eff}. For the less abundant elements (S, Ar, Ca) a more
782: conservative cut of $Z_{top} \pm$ 0.1 is imposed to avoid
783: contamination from adjacent charges, which yields a reduced efficiency of
784: 17 \%. For iron, a cut of $Z_{top} \pm$ 1.0 is used at an efficiency
785: of 90 \%. The consistency cut on the charge at the bottom of the
786: instrument $Z_{bot}$ can be sufficiently loose to not affect the
787: overall efficiency.
788:
789: \emph{Effective Aperture} --- The overall geometric factor of TRACER
790: is 5.04 m$^2$ sr. For the flux computation an effective aperture $A_i$
791: is calculated, which includes the geometric factor, losses due to
792: nuclear spallations, and ``dead'' regions in the detector.
793:
794: \begin{equation}\label{geometry}
795: A_i = A \cdot 2 \pi \int^{\pi /2}_{\theta=0}P_I(\theta)P_D(\theta)\cos \theta \ \mathrm{d}(\cos \theta)
796: \end{equation}
797:
798: with $A$ the area of the detector (206 cm $\times$ 206 cm),
799: $P_I(\theta)$ the probability of survival against spallation in the
800: atmosphere and in the instrument as a function of zenith angle
801: $\theta$ (see figure \ref{interactions}), and $P_D(\theta)$ the
802: probability that a particle passing through the top scintillator will
803: traverse the entire instrument without encountering any insensitive
804: detector regions. The quantity P$_{D}(\theta)$ is determined from a
805: Monte Carlo simulation. Insensitive detector regions lead to a
806: reduction in the overall geometric factor by about 10~\%.
807:
808: Losses due to nuclear interactions are computed using a combination of
809: measured charge-changing cross sections and parameterizations obtained
810: by \citet{heckmann78} and \citet{westfall79}. The interaction losses
811: are taken to be independent of energy but do depend on the particle
812: mass. Interactions in the atmosphere are determined by measuring the
813: residual atmosphere above the payload, which for this flight was
814: monitored by pressure sensors and GPS modules attached to the
815: instrument. The average residual atmosphere was 3.91
816: g/cm$^2$. Interaction probabilities in the instrument are obtained by
817: numerical computation using a complete list of materials in
818: TRACER. The fraction of interacting nuclei in the instrument can also
819: be obtained experimentally from a comparison of the reconstructed
820: nuclear charges for each element in the top and bottom
821: scintillators. Both methods lead to consistent results. As an example,
822: an oxygen or iron nucleus at an incident angle of 30$^\circ$ has a
823: probability of survival in the atmosphere of 83~\% or 73\% and
824: probability of survival in the instrument of 70~\% or 49~\%,
825: respectively (see figure \ref{interactions} for all elements).
826:
827: \emph{Energy Overlap Corrections} --- As TRACER has three detectors, each
828: with a different energy response and energy resolution (see
829: figure~\ref{resp} and figure~\ref{eres}), the widths of the energy
830: intervals into which the data are sorted must be appropriately
831: chosen. The intrinsic relative signal fluctuations decrease with
832: increasing $Z$, essentially as $1/Z$ (figure~\ref{eres}). Therefore,
833: the number of acceptable energy bins increases for elements with
834: higher charge number.
835:
836: As the shape of the energy spectra is not expected to vary rapidly
837: with energy, choosing too fine a subdivision of the energy scale might
838: just lead to increased statistical uncertainty in each energy bin, and
839: hence, not improve the quality of the results. More importantly, the
840: widths of the bins must be commensurate with the energy resolution of
841: the detector. Otherwise, overlap corrections for events being
842: misinterpreted from a neighbouring energy bin would become substantial
843: and could possibly bias the result for a steeply falling spectrum.
844:
845: In the current analysis of the TRACER data, the bin widths are chosen
846: fairly wide, such that the overlap correction factor can be determined
847: with the help of an iterative Monte-Carlo simulation. The corrections
848: are computed for each energy bin and element. They are fairly small,
849: in general $\le 20~\%$ (i.e. $0.8 \leq C_i \leq 1.2$).
850:
851: % ------------------------------------------------------------------
852: \section{Measured Energy Spectra}
853: \label{sec:spectra}
854:
855: % this goes in here
856:
857: %The data points are plotted according to a method discussed in detail
858: %in \cite{lafferty95}, except for the highest energy bin where, to
859: %accumulate statistics, the flux stated is an integral flux.
860:
861: The analysis yields, for each species, numbers of events $\Delta{N}_i$
862: sorted into energy bins $\Delta{E}_i = E_{i+1} - E_{i}$. The ratio
863: $\Delta{N}_{i}/\Delta{E}_{i}$ defines the differential intensity $dN/dE$
864: for this interval:
865:
866: \begin{equation}\label{dndei}
867: \frac{dN}{dE}(i) = \frac{\Delta{N}_{i}}{\Delta{E}_{i}}
868: \end{equation}
869:
870: We assume that the energy spectrum is represented by a power law
871: $CE^{-\alpha}$ over this interval, such that:
872:
873: \begin{equation}\label{dn}
874: \Delta{N}_{i} = \int^{E_{i+1}}_{E_{i}} C E^{-\alpha} dE
875: \end{equation}
876:
877: We then plot the spectral value $\frac{dN}{dE}(i)$ at an energy level
878: $\hat{E}$ defined such that:
879:
880:
881: \begin{equation}\label{dnde}
882: \frac{dN}{dE} = C \cdot \hat{E}^{-\alpha},
883: \end{equation}
884:
885:
886: \begin{equation}\label{ehat}
887: \hat{E} = \frac{1}{(E_{i+1} - E_{i})} \quad \frac{1}{(1-\alpha)} \quad (E_{i+1}^{1-\alpha} - E_{i}^{1-\alpha})^{-1/\alpha}
888: \end{equation}
889:
890: Since $\hat{E}$ depends on the value of $\alpha$, an iterative
891: procedure has been used. It is found that the dependence of $\hat{E}$
892: on $\alpha$ is not very strong.
893:
894: For the spectra presented in this work, the highest energy intervals
895: are integral bins, with $E_{i+1} \rightarrow \infty$. We then plot the
896: data point at the median energy for this bin, $\hat{E}_{int}$:
897:
898: %\begin{equation}\label{eint}
899: %\hat{E}_{int} = [(E^{1-\alpha}_{i} + E^{1-\alpha}_{i+1})/2]^{1/1-\alpha}
900: %\end{equation}
901:
902: \begin{equation}\label{eint}
903: \hat{E}_{int} = [E^{1-\alpha}_{i}/2]^{1/1-\alpha}
904: \end{equation}
905:
906: The value of the flux at this energy is:
907:
908: %\begin{equation}\label{fint}
909: %\left| \frac{dN}{dE} \right| _{int} = \frac{N(\alpha-1)}{ (E^{1-\alpha}_{i} - E^{1-\alpha}_{i+1})} \cdot \hat{E}_{int}^{-\alpha}
910: %\end{equation}
911:
912: \begin{equation}\label{fint}
913: \left| \frac{dN}{dE} \right| _{int} = \frac{\Delta{N}_{i}(\alpha-1)}{ E^{1-\alpha}_{i}} \cdot \hat{E}_{int}^{-\alpha}
914: \end{equation}
915:
916: The energy spectra, in terms of absolute intensities, for the elements
917: O, Ne, Mg, Si, S, Ar, Ca, and Fe are presented in tables
918: \ref{tab:flux1} and \ref{tab:flux2} and are plotted in figure
919: \ref{tracerspec}. The low energy data points ($\le 2.3$ GeV per amu)
920: come from an analysis of the Cherenkov counter, using a restricted
921: angular acceptance. Note that the energy in the table is given in
922: units of kinetic energy per amu, whereas in figure \ref{tracerspec}
923: the units are given in terms of kinetic energy per particle. Data from
924: the TRACER 2003 flight are indicated by the solid squares. For clarity
925: the intensity of each element is scaled by a factor shown on the
926: left. Existing data from measurements in space with HEAO-3
927: \citep{engelmann90} and CRN \citep{muller91} are shown for
928: comparison. In the energy regions where overlap exists, the agreement
929: of the results appears to be quite good.
930:
931: We note the large range in intensity (ten decades) and particle energy
932: (four decades) covered by TRACER. As can be seen, the energy spectra
933: extend up to and beyond 10$^{14}$ eV. The eight elements for which
934: spectra are shown in figure \ref{tracerspec} represent the major
935: primary species from oxygen to iron; contributions to these species
936: from spallation produced secondary elements are expected to be
937: small. We may fit the high-energy part ($>$ 20 GeV per amu) of each
938: spectrum to a power law in energy $E^{-\alpha}$. The resulting
939: spectral indices are shown in figure \ref{powerlaw}. Remarkably, there
940: is no significant trend of the spectral indices with charge $Z$, and
941: all indices fit well to an average of $\alpha$ = 2.65 $\pm$ 0.05. It
942: must be emphasized that this result is not an artifact of the method
943: of defining the differential intensities described in the previous
944: sections. The similarity of the spectral shapes must indicate that
945: acceleration and propagation mechanisms are essentially the same for
946: all species. However, the statistical errors do not exclude the
947: possibility of deviations from a pure power-law behavior for
948: individual spectra.
949:
950: %\section{Sources of Uncertainty}
951: %\label{sec:systematics}
952:
953: \section{Comparison with other data}
954: \label{sec:otherdata}
955:
956: In figure \ref{tracerspec}, the energy spectra obtained with TRACER
957: are shown together with measurements in space with the HEAO-3
958: satellite up to $ \sim $ 40 GeV per amu, and with the CRN detector on
959: the Space Shuttle up to about 1000 GeV per amu. However, a
960: multi-decade plot such as that of figure \ref{tracerspec} may obscure
961: the finer details in a steeply falling spectrum. To illustrate this,
962: we show the energy spectra for O, Ne, and Fe, in figure \ref{fe-ne-o}
963: multiplied with $E^{2.5}$. It is apparent that the differences between
964: individual data sets seem to be mostly statistical.
965:
966: We note that figure \ref{fe-ne-o} also includes a recent determination
967: of the high-energy iron spectrum that was performed with the HESS
968: Imaging Air Cherenkov Telescope on the ground. Here it was possible
969: for the first time to observe directly the Cherenkov light of
970: iron-group nuclei in the atmosphere before they interact
971: \citep{aharonian07a}. Again, the agreement with the TRACER data is
972: good, although uncertainty about the choice of a nuclear interaction
973: model affects the interpretation of the HESS measurement.
974:
975: There are also several recent or current balloon-borne detectors that
976: are designed to obtain measurements of the high-energy cosmic-ray
977: composition, for instance RUNJOB \citep{derbina05}, ATIC
978: \citep{panov06}, and CREAM \citep{seo06}; all of these employ mostly
979: calorimetric techniques. The RUNJOB experiment has published energy
980: spectra for mostly elemental groups. The iron spectrum from this
981: experiment is also included in figure \ref{fe-ne-o}. Again we see that
982: there is good agreement with the TRACER data. Spectra for a few
983: individual heavy elements, more limited in energy coverage and
984: statistics than TRACER, have been reported for ATIC and CREAM at
985: conferences \citep{panov06,ahn07, zei07}. These results are not
986: included here as they are still labeled preliminary. However, the
987: agreement in the regions of overlap appears to be quite good.
988:
989: Data for oxygen and iron from TRACER are compared in figure
990: \ref{airshower} with spectra derived from indirect air-shower
991: observations by the EAS-TOP collaboration \citep{navarra03}, and by
992: the KASCADE group for two different nucleus-nucleus interaction models
993: \citep{antoni05}. These groups do not report results for individual
994: elements: the fluxes for the ``CNO group'' most likely have about twice
995: the intensity than oxygen alone, while the ``iron group'' probably is
996: dominated by iron. Our results do not yet overlap with the energy
997: region of the air shower data, but the gap is becoming smaller, in
998: particular for oxygen. Additional measurements should indeed lead to
999: significant constraints on the air shower interpretations.
1000:
1001: \section{Conclusions}
1002: \label{sec:conclusions}
1003:
1004: The energy spectra of the cosmic-ray nuclei determined in this
1005: investigation extend to energies around $10^{14}$ eV per particle, or
1006: significantly higher for the more abundant nuclei. We believe that
1007: this data set represents the most comprehensive information about the
1008: spectra of the heavier primary nuclei currently available.
1009:
1010: We have studied in some detail how these results provide constraints
1011: on our current understanding of the origin of galactic cosmic rays and
1012: of models of acceleration and propagation. We realize that the upper
1013: energy limit of the measurement is still below the region of the
1014: cosmic-ray ``knee'', so perhaps it is not surprising that no
1015: significant changes in cosmic-ray composition are yet
1016: observed. However, the great similarity between the individual energy
1017: spectra is remarkable. The result of our studies of these features is
1018: currently being prepared for a separate publication.
1019:
1020: It is clear that an extension of the measurements to still higher
1021: energy is desirable. The dynamic range of the TRD used here is not
1022: exhausted with the current measurement. Hence, in order to reach
1023: higher energy, the TRACER detector would either have to be increased
1024: in size, which seems impractical as the instrument is already the
1025: largest balloon-borne cosmic-ray detector in existence, or would have
1026: to be subjected to additional long-duration balloon flights. Ideally,
1027: an instrument like TRACER should be exposed in space for several
1028: years.
1029:
1030: An important objective for measurements beyond the current TRACER
1031: results would be an extension of the dynamic range of the detector,
1032: such that the light secondary elements (Li, Be, B) as well as carbon
1033: and nitrogen can be covered together with the heavier nuclei. If
1034: successful to sufficiently high energies ($\sim 10^3$ GeV per amu),
1035: this measurement would lead to a determination of the L/M abundance
1036: ratio, i.e., to a determination of the propagation pathlength at high
1037: energy.
1038:
1039: This would be an essential step towards a better understanding of the
1040: galactic propagation of cosmic rays. To make this measurement
1041: possible, the electronics of TRACER has recently been completely
1042: refurbished, and another successful long-duration balloon flight was
1043: performed in 2006. The analysis of these new data is currently in
1044: progress and results will be reported in due course.
1045:
1046: \acknowledgements
1047:
1048: We gratefully acknowledge the contributions of Drs. German Hermann
1049: and Scott Wakely to the design and construction of TRACER. We
1050: appreciate the services of the University of Chicago Engineering
1051: Center, especially Gary Kelderhouse, David Pernic and Gene Drag. We
1052: thank the staff of the Columbia Scientific Balloon Facility, especially
1053: David Sullivan, and the National Science Foundation and United States
1054: Antarctic Program for support in conducting this flight. We are
1055: indebted to Scott Cannon for his contribution in the development of
1056: the thermal protection for the instrument. This work was supported by
1057: NASA grants NAG5-5305, NN04WC08G and NNG06WC05G. MI acknowledges the
1058: Grant-in-Aid for Scientific Research of the Japan Society for the
1059: Promotion of Science (JSPS), No. 17540226. Numerous students have
1060: participated in the construction of the instrument under support from
1061: the Illinois Space Grant Consortium.
1062:
1063: %% Tables may also be prepared as separate files. See the accompanying
1064: %% sample file table.tex for an example of an external table file.
1065: %% To include an external file in your main document, use the \input
1066: %% command. Uncomment the line below to include table.tex in this
1067: %% sample file. (Note that you will need to comment out the \documentclass,
1068: %% \begin{document}, and \end{document} commands from table.tex if you want
1069: %% to include it in this document.)
1070:
1071: %% \input{table}
1072:
1073:
1074:
1075: \begin{thebibliography}{}
1076: \bibitem[Ahn et al.(2007)]{ahn07} Ahn, H.S. et al., 2007, Proc. of 30th International Cosmic Ray Conference (ICRC)
1077: \bibitem[Agostinelli et al.(2003)]{agostinelli03} Agostintelli, S. et al. 2003, NIM A, 1, 250
1078: \bibitem[Aharonian et al.(2007a)]{aharonian07a} Aharonian, F. et al. 2007a, \prd, 75, 4
1079: \bibitem[Aharonian et al.(2004)]{aharonian04} Aharonian, F. et al. 2004, Nature, 432, 75
1080: \bibitem[Antoni et al.(2005)]{antoni05} Antoni, T. et al. 2005, APh., 24, 1
1081: \bibitem[Bell(1978)]{bell78} Bell, A.R 1978, \mnras, 182, 147
1082: \bibitem[Berezhko \& V\"{o}lk(2006)]{berezhko06} Berezhko, E.G. \& V\"{o}lk, H.J. 2006, \aap, 451, 981
1083: \bibitem[Beuville et al.(1990)]{beuville90} Beuville, E. et al. 1990, NIM A, 288, 157
1084: \bibitem[Cronin et al.(1997)]{cronin97} Cronin, J., Gaisser, T.K., \& Swordy, S.P. 1977, Scientific American, 276, 44
1085: \bibitem[LaGage \& Cesarsky(1983)]{lagage83} Lagage, P.O. \& Cesarsky, C.J. 1983, \aap, 125, 249L
1086: \bibitem[Derbina et al.(2005)]{derbina05} Derbina, V. et al. 2005, \apjl, 628, L41
1087: \bibitem[Engelmann et al.(1990)]{engelmann90} Engelmann, J. et al. 1990, \aap, 233, 96
1088: \bibitem[Gahbauer et al.(2004)]{gahbauer04} Gahbauer, F., Hermann, G., H\"{o}randel, J., M\"{u}ller D., \& Radu, A.A. 2004, \apj, 607, 333
1089: \bibitem[Gahbauer et al.(2003)]{gahbauer03b} Gahbauer, F., M\"{u}ller D., Hermann, G., H\"{o}randel, J., \& Radu, A. 2003, Proc. 27th ICRC, 1, 2245
1090: \bibitem[Garcia Mu\~{n}oz et al.(1975)]{gmunoz75} Garcia Mu\~{n}oz, M, Mason, G.M, \& Simpson, J.A, 1975, \apjl, 201, L141
1091: \bibitem[Heckmann et al.(1978)]{heckmann78} Heckmann, H.H., Greiner, D.E, Lindstrom, P.J., \& Shwe, H. 1978, \prc, 17, 1735
1092: \bibitem[Juliusson et al.(1972)]{juliusson72} Juliusson, E., Meyer, P., \& M\"{u}ller D., 1972, \prl, 29, 445
1093: %\bibitem[Kieda et al.(2001)]{kieda01} Kieda, D., Swordy, S.P and Wakely, S.P. 2001, Proc 27th ICRC, 4, 1533
1094: %\bibitem[Lafferty \& Wyatt(1995)]{lafferty95} Lafferty, G.D. and Wyatt, T.T. 1995, NIM A, 355, 541
1095: \bibitem[L'Heureux et al.(1990)]{lheureux90} L'Heureux, J., Grunsfeld, J.M., Meyer, P., M\"{u}ller, D., \& Swordy, S.P. 1990, NIM A, 295, 246
1096: \bibitem[Mewaldt et al.(2001)]{mewaldt01} Mewaldt, R.A. et al. 2001, Space Science Reviews, 99, 27
1097: \bibitem[M\"{u}ller et al.(1991)]{muller91} M\"{u}ller, D., Swordy, S.P., Meyer, P., L'Heureux, J., \& Grunsfeld, J. 1991, \apj, 374, 356
1098: %\bibitem[M\"{u}ller et al.(2007b)]{muller07b} M\"{u}ller, D. et al. 2007, NIMA, in preparation
1099: \bibitem[Navarra et al.(2003)]{navarra03} Navarra, G. et al. 2003, Proc. 28th International Cosmic Ray Conference, 1, 147
1100: \bibitem[Panov et al.(2006)]{panov06} Panov, A. et al. 2006, ArXiv e-print, astro-ph/0612377
1101: \bibitem[Romero-Wolf(2005)]{romero05b} Romero-Wolf, A. 2005, MSc. Thesis, The University of Chicago, Unpublished
1102: %\bibitem[Swordy et al.(1990)]{swordy90} Swordy, S.P., Grunsfeld, J., L'Heureux, J., Meyer, P. and M\"{u}ller, D. 1990, \apj, 349, 625
1103: \bibitem[Seo et al.(2006)]{seo06} Seo, E.S. et al. 2006, Advances in Space Sciences, in press
1104: %\bibitem[Wakely(2002)]{wakely02} Wakely, S.P. 2002, APh., 18, 67
1105: \bibitem[Westfall et al.(1979)]{westfall79} Westfall, G. D., Wilson, L. W., Lindstrom, P. J., Crawford, H. J., Greiner, D. E., \& Heckman, H. H., \prc, 19, 1309
1106: %\bibitem[Wiedenbeck et al.(1998)]{wiedenbeck98} Wiedenbeck, M.E et al. 1998, AAS, 194, 2804H
1107: \bibitem[Wiedenbeck et al.(1999)]{wiedenbeck99} Wiedenbeck, M.E et al. 1999, \apjl, 523, L61
1108: \bibitem[Yanasak et al.(2001)]{yanasak01} Yanasak, N. E. et al., 1975, \apj, 563, 768
1109: \bibitem[Zei et al.(2007)]{zei07} Zei, R. et al., 2007, Proc. of 30th ICRC
1110: \end{thebibliography}
1111:
1112:
1113:
1114:
1115:
1116:
1117:
1118: \clearpage
1119:
1120: \begin{table}
1121: \caption{\label{tab:eff} Charge selection efficiencies}
1122: \begin{center}
1123: \begin{tabular}{lcccccccc}
1124: Element & {\bf 8} & {\bf 10}& {\bf 12} & {\bf 14} & {\bf 16} & {\bf 18}& {\bf 20} & {\bf 26}\\
1125: Cut (Charge Units) & $\pm$0.5 & $\pm$0.5 & $\pm$0.5 & $\pm$0.5 & $\pm$0.1 & $\pm$0.1 & $\pm$0.1 & $\pm$1.0 \\
1126: Efficiency & 0.89 & 0.83 & 0.75 & 0.72 & 0.17 & 0.18 & 0.18 & 0.90 \\
1127: \end{tabular}
1128: \end{center}
1129: \end{table}
1130:
1131:
1132: \clearpage
1133:
1134: %% TRACER TABLE FLUX
1135:
1136:
1137: \newcommand{\ibin}{b}
1138:
1139: \begin{table}
1140: \begin{minipage}[h]{20cm}
1141:
1142: \caption{\label{tab:flux1} {\bf Measured Intensities (Oxygen - Silicon)}}
1143:
1144: \footnotesize
1145:
1146: \begin{tabular}{lr@{ - }lrrr@{}l@{$\times$}l} \hline
1147: {\bf Element} & \multicolumn{2}{c} {\bf Energy Range} & \multicolumn{1}{c}{\bf Kinetic Energy\tablenotemark{1}} & \multicolumn{1}{c}{\bf Number of} & \multicolumn{3}{c} {\bf $\quad$ Differential Intensity}\\
1148: & \multicolumn{2}{c} {\bf (GeV/amu) } & \multicolumn{1}{c}{\bf \^{E} (GeV/amu)} & \multicolumn{1}{c}{\bf Events } & \multicolumn{3}{c} {\bf $\quad$ (m$^2$ s sr GeV/amu)$^{-1}$}\\ \hline
1149:
1150:
1151: \vspace{0.05cm} \\
1152: \bf{O (Z = 8)} & 0.8 & 1.0 & 0.9$\qquad \quad$ & 25437 $\quad$ & $\qquad$(1.71 $\;$&$\pm$ 0.01)&$\;10^0$ \\
1153: & 1.0 & 1.3 & 1.1$\qquad \quad$ & 24237 $\quad$ & $\qquad$(1.38 $\;$ &$\pm$ 0.01)&$\;10^0$ \\
1154: & 1.3 & 1.9 & 1.5$\qquad \quad$ & 38226 $\quad$ & $\qquad$(8.95 $\;$ &$\pm$ 0.05)&$\;10^{-1}$ \\
1155: & 12 & 292 & 48$\qquad \quad$ & 456918 $\quad$ & $\qquad$(7.27 $\;$ &$\pm$ 0.01)&$\;10^{-4}$ \\
1156: & 1315 & 2627 & 1834$\qquad \quad$ & 77 $\quad$ & $\qquad$(4.4 $\;$ &$\pm$ 0.5)&$\;10^{-8}$ \\
1157: & 2627 & 6371 & 4001$\qquad \quad$ & 29 $\quad$ & $\qquad$(5.8 $\;$ &$\pm$ 1.1)&$\;10^{-9}$ \\
1158: & 6371 & 9834 & 7872$\qquad \quad$ & 5 $\quad$ & $\qquad$(9.1 $\;$ & $^{+ 6.0} _{- 3.9}$)&$\;10^{-10}$ \\
1159: & 9834 & & 15611$\qquad \quad$ & 6 $\quad$ & $\qquad$(2.0 $\;$ &$^{+ 1.2} _{- 0.8}$)&$\;10^{-10}$ \\
1160: \vspace{0.05cm} \\
1161: \bf{Ne (Z = 10)} & $\quad$ 0.8 & 1.0 & 0.9 $\qquad \quad$ & 4031 $\quad$ & $\qquad$(3.08 $\;$ &$\pm$ 0.05)&$\;10^{-1}$ \\
1162: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 3790 $\quad$ & $\qquad$(2.42 $\;$ &$\pm$ 0.04)&$\;10^{-1}$ \\
1163: & $\quad$ 1.3 & 1.9 & 1.5 $\qquad \quad$ & 5639 $\quad$ & $\qquad$(1.48 $\;$ &$\pm$ 0.02)&$\;10^{-1}$ \\
1164: & $\quad$ 12 & 292 & 48 $\qquad \quad$ & 57991 $\quad$ & $\qquad$(1.21 $\;$ &$\pm$ 0.01)&$\;10^{-4}$ \\
1165: & $\quad$ 895 & 1316 & 1080 $\qquad \quad$ & 16 $\quad$ & $\qquad$(3.2 $\;$ &$\pm$ 0.8)&$\;10^{-8}$ \\
1166: & $\quad$ 1316 & 2627 & 1834 $\qquad \quad$ & 13 $\quad$ & $\qquad$(8.2 $\;$ &$\pm$ 2.3)&$\;10^{-9}$ \\
1167: & $\quad$ 2627 & 6371 & 4001 $\qquad \quad$ & 4 $\quad$ & $\qquad$(8.9 $\;$ & $^{+ 6.9} _{- 4.2}$)&$\;10^{-10}$ \\
1168: & $\quad$ 6371 & 9834 & 7872 $\qquad \quad$ & 1 $\quad$ & $\qquad$(2.2 $\;$ & $^{+ 5.0} _{- 1.8}$)&$\;10^{-10}$ \\
1169: & $\quad$ 9834 & & 15611 $\qquad \quad$ & 2 $\quad$ & $\qquad$(7.8 $\;$ & $^{+ 10.1} _{- 5.0}$)&$\;10^{-11}$ \\
1170: \vspace{0.05cm} \\
1171: \bf{Mg (Z = 12)} &$\quad$ 0.8 & 1.0 & 0.9 $\qquad \quad$ & 4603 $\quad$ & $\qquad$(4.02 $\;$ &$\pm$ 0.06)&$\;10^{-1}$ \\
1172: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 4184 $\quad$ & $\qquad$(2.96 $\;$ &$\pm$ 0.05)&$\;10^{-1}$ \\
1173: & $\quad$ 1.3 & 1.9 & 1.5 $\qquad \quad$ & 6339 $\quad$ & $\qquad$(1.89 $\;$ &$\pm$ 0.02)&$\;10^{-1}$ \\
1174: & $\quad$ 12 & 292 & 48 $\qquad \quad$ & 59374 $\quad$ & $\qquad$(1.60 $\;$ &$\pm$ 0.01)&$\;10^{-4}$ \\
1175: & $\quad$ 696 & 1316 & 946 $\qquad \quad$ & 23 $\quad$ & $\qquad$(3.6 $\;$ &$\pm$ 0.7)&$\;10^{-8}$ \\
1176: & $\quad$ 1316 & 2627 & 1834 $\qquad \quad$ & 9 $\quad$ & $\qquad$(6.3 $\;$ & $^{+ 2.9} _{- 2.1}$)&$\;10^{-9}$ \\
1177: & $\quad$ 2627 & 6371 & 4001 $\qquad \quad$ & 1 $\quad$ & $\qquad$(2.4 $\;$ & $^{+ 5.6} _{- 2.0}$)&$\;10^{-10}$ \\
1178: & $\quad$ 6371 & & 10113 $\qquad \quad$ & 1 $\quad$ & $\qquad$(6.7 $\;$ & $^{+ 15.4} _{- 5.5}$)&$\;10^{-11}$ \\
1179: \vspace{0.05cm} \\
1180: \bf{Si (Z = 14)} &$\quad$ 0.8 & 1.0 & 0.9 $\qquad \quad$ & 3127 $\quad$ & $\qquad$(3.07 $\;$ &$\pm$ 0.06)&$\;10^{-1}$ \\
1181: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 2812 $\quad$ & $\qquad$(2.27 $\;$ &$\pm$ 0.04)&$\;10^{-1}$ \\
1182: & $\quad$ 1.3 & 1.9 & 1.5 $\qquad \quad$ & 4292 $\quad$ & $\qquad$(1.39 $\;$ &$\pm$ 0.02)&$\;10^{-1}$ \\
1183: & $\quad$ 12 & 73 & 28 $\qquad \quad$ & 36369 $\quad$ & $\qquad$(5.23 $\;$ &$\pm$ 0.03)&$\;10^{-4}$ \\
1184: & $\quad$ 73 & 292 & 138 $\qquad \quad$ & 2278 $\quad$ & $\qquad$(7.2 $\;$ &$\pm$ 0.2)&$\;10^{-6}$ \\
1185: & $\quad$ 567 & 696 & 627 $\qquad \quad$ & 18 $\quad$ & $\qquad$(1.3 $\;$ &$\pm$ 0.3)&$\;10^{-7}$ \\
1186: & $\quad$ 696 & 1316 & 946 $\qquad \quad$ & 24 $\quad$ & $\qquad$(4.0 $\;$ &$\pm$ 0.8)&$\;10^{-8}$ \\
1187: & $\quad$ 1316 & & 2088 $\qquad \quad$ & 6 $\quad$ & $\qquad$(2.2 $\;$ & $^{+ 1.3} _{- 0.9}$)&$\;10^{-9}$ \\
1188:
1189:
1190: \end{tabular}
1191: \tablenotetext{1}{see text for definition of $\hat{E}$}
1192: \end{minipage}
1193: \end{table}
1194:
1195: \normalsize
1196:
1197: \clearpage
1198:
1199: %% TRACER TABLE FLUX
1200:
1201:
1202:
1203: \begin{table}
1204: \begin{minipage}[h]{20cm}
1205:
1206: \caption{\label{tab:flux2} {\bf Measured Intensities (Sulphur - Iron)}}
1207:
1208: \footnotesize
1209:
1210: \begin{tabular}{lr@{ - }lrrr@{}l@{$\times$}l} \hline
1211:
1212: {\bf Element} & \multicolumn{2}{c} {\bf Energy Range} & \multicolumn{1}{c}{\bf Kinetic Energy\tablenotemark{1}} & \multicolumn{1}{c}{\bf Number of} & \multicolumn{3}{c} {\bf $\quad$ Differential Intensity}\\
1213:
1214: & \multicolumn{2}{c} {\bf (GeV/amu) } & \multicolumn{1}{c}{\bf \^{E} (GeV/amu)} & \multicolumn{1}{c}{\bf Events } & \multicolumn{3}{c} {\bf $\quad$ (m$^2$ s sr GeV/amu)$^{-1}$}\\ \hline
1215:
1216: \\
1217: \bf{S (Z = 16)} & $\quad$ 0.8 & 1.0 & 0.9 $\qquad \quad$ & 156 $\quad$ & $\qquad$(7.5$\;$ &$\pm$ 0.6)&$\;10^{-2}$ \\
1218: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 158 $\quad$ & $\qquad$(6.3$\;$ &$\pm$ 0.5)&$\;10^{-2}$ \\
1219: & $\quad$ 1.3 & 1.8 & 1.5 $\qquad \quad$ & 170 $\quad$ & $\qquad$(3.3$\;$ &$\pm$ 0.3)&$\;10^{-2}$ \\
1220: & $\quad$ 1.8 & 2.3 & 2.0 $\qquad \quad$ & 106 $\quad$ & $\qquad$(2.0$\;$ &$\pm$ 0.2)&$\;10^{-2}$ \\
1221: & $\quad$ 12 & 41 & 21 $\qquad \quad$ & 1212 $\quad$ & $\qquad$(2.1$\;$ &$\pm$ 0.1)&$\;10^{-4}$ \\
1222: & $\quad$ 41 & 130 & 70 $\qquad \quad$ & 212 $\quad$ & $\qquad$(1.1$\;$ &$\pm$ 0.1)&$\;10^{-5}$ \\
1223: & $\quad$ 130 & 412 & 223 $\qquad \quad$ & 32 $\quad$ & $\qquad$(5.2$\;$ &$\pm$ 0.9)&$\;10^{-7}$ \\
1224: & $\quad$ 529 & 792 & 644 $\qquad \quad$ & 4 $\quad$ & $\qquad$(2.1$\;$ & $^{+ 1.6} _{- 1.0}$)&$\;10^{-8}$ \\
1225: & $\quad$ 792 & & 1257 $\qquad \quad$ & 3 $\quad$ & $\qquad$(2.6$\;$ & $^{+ 2.5} _{- 1.4}$)&$\;10^{-9}$ \\
1226: \vspace{0.05cm} \\
1227: \bf{Ar (Z = 18)} & $\quad$ 0.8 & 1.0 & 0.9 $\qquad \quad$ & 70 $\quad$ & $\qquad$(4.0$\;$ &$\pm$ 0.5)&$\;10^{-2}$ \\
1228: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 61 $\quad$ & $\qquad$(2.9$\;$ &$\pm$ 0.4)&$\;10^{-2}$ \\
1229: & $\quad$ 1.3 & 1.8 & 1.5 $\qquad \quad$ & 66 $\quad$ & $\qquad$(1.5$\;$ &$\pm$ 0.2)&$\;10^{-2}$ \\
1230: & $\quad$ 1.8 & 2.3 & 2.0 $\qquad \quad$ & 39 $\quad$ & $\qquad$(8.9$\;$ &$\pm$ 1.4)&$\;10^{-3}$ \\
1231: & $\quad$ 12 & 41 & 21 $\qquad \quad$ & 454 $\quad$ & $\qquad$(7.1$\;$ &$\pm$ 0.3)&$\;10^{-5}$ \\
1232: & $\quad$ 41 & 130 & 70 $\qquad \quad$ & 85 $\quad$ & $\qquad$(3.8$\;$ &$\pm$ 0.4)&$\;10^{-6}$ \\
1233: & $\quad$ 130 & 412 & 223 $\qquad \quad$ & 11 $\quad$ & $\qquad$(1.5$\;$ &$\pm$ 0.5)&$\;10^{-7}$ \\
1234: & $\quad$ 452 & 696 & 558 $\qquad \quad$ & 2 $\quad$ & $\qquad$(1.9$\;$ & $^{+ 2.5} _{- 1.3}$)&$\;10^{-8}$ \\
1235: & $\quad$ 696 & & 1105 $\qquad \quad$ & 1 $\quad$ & $\qquad$(3.2$\;$ & $^{+ 7.3} _{- 2.6}$)&$\;10^{-9}$ \\
1236: \vspace{0.05cm} \\
1237: \bf{Ca (Z = 20)} & $\quad$ 0.8 & 1.0 & 0.9 $\qquad \quad$ & 99 $\quad$ & $\qquad$(5.5 $\;$ &$\pm$ 0.6)&$\;10^{-2}$ \\
1238: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 79 $\quad$ & $\qquad$(3.7 $\;$ &$\pm$ 0.4)&$\;10^{-2}$ \\
1239: & $\quad$ 1.3 & 1.8 & 1.5 $\qquad \quad$ & 99 $\quad$ & $\qquad$(2.2 $\;$ &$\pm$ 0.2)&$\;10^{-2}$ \\
1240: & $\quad$ 1.8 & 2.3 & 2.0 $\qquad \quad$ & 61 $\quad$ & $\qquad$(1.3 $\;$ &$\pm$ 0.2)&$\;10^{-2}$ \\
1241: & $\quad$ 12 & 41 & 21 $\qquad \quad$ & 610 $\quad$ & $\qquad$(9.9 $\;$ &$\pm$ 0.4)&$\;10^{-5}$ \\
1242: & $\quad$ 41 & 130 & 70 $\qquad \quad$ & 88 $\quad$ & $\qquad$(4.1 $\;$ &$\pm$ 0.4)&$\;10^{-6}$ \\
1243: & $\quad$ 130 & 412 & 223 $\qquad \quad$ & 19 $\quad$ & $\qquad$(2.8 $\;$ &$\pm$ 0.6)&$\;10^{-7}$ \\
1244: & $\quad$ 452 & 696 & 558 $\qquad \quad$ & 3 $\quad$ & $\qquad$(1.8 $\;$ & $^{+ 1.7} _{- 1.0}$)&$\;10^{-8}$ \\
1245: & $\quad$ 696 & & 1105 $\qquad \quad$ & 1 $\quad$ & $\qquad$(1.9 $\;$ & $^{+ 4.3} _{- 1.6}$)&$\;10^{-9}$ \\
1246: \vspace{0.05cm} \\
1247: \bf{Fe (Z = 26)} & 0.8 & 1.0 & 0.9 $\qquad \quad$ & 1821 $\quad$ & $\qquad$(2.21 $\;$ &$\pm$ 0.05)&$\;10^{-1}$ \\
1248: & $\quad$ 1.0 & 1.3 & 1.1 $\qquad \quad$ & 1669 $\quad$ & $\qquad$(1.65 $\;$ &$\pm$ 0.04)&$\;10^{-1}$ \\
1249: & $\quad$ 1.3 & 1.8 & 1.5 $\qquad \quad$ & 2166 $\quad$ & $\qquad$(1.02 $\;$ &$\pm$ 0.02)&$\;10^{-1}$ \\
1250: & $\quad$ 1.8 & 2.3 & 2.0 $\qquad \quad$ & 1388 $\quad$ & $\qquad$(6.2 $\;$ &$\pm$ 0.2)&$\;10^{-2}$ \\
1251: & $\quad$ 12 & 23 & 16 $\qquad \quad$ & 10589 $\quad$ & $\qquad$(1.22 $\;$ &$\pm$ 0.02)&$\;10^{-3}$ \\
1252: & $\quad$ 23 & 41 & 30 $\qquad \quad$ & 4468 $\quad$ & $\qquad$(2.80 $\;$ &$\pm$ 0.04)&$\;10^{-4}$ \\
1253: & $\quad$ 41 & 92 & 60 $\qquad \quad$ & 2132 $\quad$ & $\qquad$(4.7 $\;$ &$\pm$ 0.1)&$\;10^{-5}$ \\
1254: & $\quad$ 92 & 412 & 183 $\qquad \quad$ & 543 $\quad$ & $\qquad$(1.9 $\;$ &$\pm$ 0.1)&$\;10^{-6}$ \\
1255: & $\quad$ 529 & 606 & 566 $\qquad \quad$ & 9 $\quad$ & $\qquad$(1.3 $\;$ & $^{+ 0.6} _{- 0.4}$)&$\;10^{-7}$ \\
1256: & $\quad$ 606 & 895 & 733 $\qquad \quad$ & 12 $\quad$ & $\qquad$(5.2 $\;$ &$\pm$ 1.5)&$\;10^{-8}$ \\
1257: & $\quad$ 895 & & 1421 $\qquad \quad$ & 11 $\quad$ & $\qquad$(7.4 $\;$ &$\pm$ 2.2)&$\;10^{-9}$ \\
1258:
1259:
1260: \end{tabular}
1261: \tablenotetext{1}{see text for definition of $\hat{E}$}
1262: \end{minipage}
1263: \end{table}
1264:
1265: \normalsize
1266:
1267: \clearpage
1268:
1269: %% FIGURES
1270:
1271:
1272: %% TRACER RESPONSE CURVES
1273: \begin{figure}[t]
1274: \begin{center}
1275: %\includegraphics[width=0.98\textwidth]{energy-signal1.eps}
1276: \includegraphics[width=0.98\textwidth]{f1.eps}
1277: \end{center}
1278: \caption{Energy response of detection techniques used in TRACER:
1279: Cherenkov counter (CER), energy loss in gases (DEDX), and Transition
1280: Radiation (TRD).}
1281: \label{resp}
1282: \end{figure}
1283:
1284:
1285: \begin{figure}[t]
1286: \begin{center}
1287: %\includegraphics[width=.78\textwidth,bb=0 0 3in 300]{JoJo1.JPG}
1288: \includegraphics[width=.9\textwidth]{f2.eps}
1289: \end{center}
1290: \caption{Schematic drawing of TRACER}
1291: \label{tracer}
1292: \end{figure}
1293:
1294:
1295: %% TRACER TRAJECTORY
1296: \begin{figure}
1297: \begin{center}
1298: \includegraphics[width=.9\textwidth]{f3.eps}
1299: \end{center}
1300: \caption{Trajectory of 2003 TRACER Antarctic Long Duration Balloon Flight.}
1301: \label{map}
1302: \end{figure}
1303:
1304: %% TRACER CHARGE ANALYSIS
1305: \begin{figure}[ht]
1306: \begin{center}
1307: \includegraphics[width=.99\textwidth]{f4.eps}
1308: \end{center}
1309: \caption{Scatter plot of top scintillator vs. Cherenkov signals in arbitrary units.}
1310: \label{CerScint}
1311: \end{figure}
1312:
1313: %% TRACER CHARGE ANALYSIS
1314: \begin{figure}[ht]
1315: \begin{center}
1316: \includegraphics[width=.99\textwidth]{f5.eps}
1317: \end{center}
1318: \caption{Charge histogram for all events measured in flight.}
1319: \label{Zhisto}
1320: \end{figure}
1321:
1322:
1323: %% TRACER LOW ENERGY PARTICLES
1324: \begin{figure}[h]
1325: \begin{center}
1326: \includegraphics[width=0.99\textwidth]{f6.eps}
1327: \end{center}
1328: \caption{Scatter plot of dE/dx vs. Cherenkov signals for iron
1329: nuclei. The black line is the average response obtained from
1330: simulations. The gray scale is logarithmic.}
1331: \label{CerdEdx}
1332: \end{figure}
1333:
1334: %% TRACER HIGH ENERGY PARTICLES
1335:
1336: \begin{figure}[h]
1337: \begin{center}
1338: \includegraphics[width=.99\textwidth]{f7.eps}
1339: \end{center}
1340: \caption{Energy response of the Transition Radiation Detector as
1341: measured at accelerators. Except for two highest-energy data points,
1342: these measurements have been published previously
1343: \citep{lheureux90}. The open symbols and dashed lines refer to
1344: measurements of dE/dx only.}
1345: \label{TRcurve}
1346: \end{figure}
1347:
1348: \begin{figure}[h]
1349: \begin{center}
1350: \includegraphics[width=.99\textwidth]{f8.eps}
1351: \end{center}
1352: \caption{Correlation of average responses (square root) of Transition
1353: Radiation and specific ionization detectors for relativistic
1354: nuclei. Four elements are displayed to illustrate the charge
1355: dependence of the responses.}
1356: \label{TRDvsDEDX}
1357: \end{figure}
1358:
1359:
1360: %% NEON SCATTER PLOT
1361: \begin{figure}[h]
1362: \begin{center}
1363: \includegraphics[width=0.90\textwidth]{f9.eps}
1364: \end{center}
1365: \caption{Scatter plot of TR vs. dE/dx signal for neon nuclei. The
1366: highlighted points represent the highest energy events measured with
1367: the TRD. As expected the transition radiation events have signals in
1368: the dE/dx detector which are well above the minimum ionization
1369: level. Units are arbitrary.}
1370: \label{TRdEdx}
1371: \end{figure}
1372:
1373: %% INTERACTIONS
1374: \begin{figure}[h]
1375: \begin{center}
1376: \includegraphics[width=0.9\textwidth]{f10.eps}
1377: \end{center}
1378: \caption{Survival probability all nuclei incident at a zenith angle of
1379: 30$^\circ$ in a residual atmosphere of 3.91 g/cm$^2$ (dash-dot),
1380: within the TRACER instrument (dash), and total survival probability
1381: (solid).}
1382: \label{interactions}
1383: \end{figure}
1384:
1385: %% INTERACTIONS
1386: \begin{figure}[h]
1387: \begin{center}
1388: \includegraphics[width=0.9\textwidth]{f11.eps}
1389: \end{center}
1390: \caption{Energy resolution (1$\sigma$) of individual detector
1391: subsystems vs charge $Z$, and for typical energies: Cherenkov:
1392: $\gamma$ = 2 (dashed line), dE/dx: $\gamma$ = 100 (dash-dot) and TRD:
1393: $\gamma$ = 1000 (solid line)}
1394: \label{eres}
1395: \end{figure}
1396:
1397: %% TRACER ENERGY SPECTRA
1398: \begin{figure}[t]
1399: \begin{center}
1400: \includegraphics[width=.90\textwidth]{f12.eps}
1401: \end{center}
1402: \caption{Differential energy spectra vs energy per particle of the
1403: cosmic ray nuclei : O, Ne, Mg, Si, S, Ar, Ca and Fe. Results from the
1404: TRACER 2003 flight are indicated by the solid squares. Existing data
1405: from the HEAO-3 experiment (open diamonds : \citet{engelmann90}) and
1406: the CRN experiment (open crosses : \citet{muller91}) are shown for
1407: comparison. The dashed line represents an independent power-law fit to
1408: each spectrum above 20 GeV per amu.}
1409: \label{tracerspec}
1410: \end{figure}
1411:
1412:
1413: %% TRACER POWER LAWS
1414: \begin{figure}[ht]
1415: \begin{center}
1416: \includegraphics [width=0.90\textwidth]{f13.eps}
1417: \end{center}
1418: \caption{Spectral indices of a best power-law fit to the combined
1419: TRACER and CRN data above 20 GeV per amu. The line indicates the an
1420: average spectral fit of E$^{-2.65}$.}
1421: \label{powerlaw}
1422: \end{figure}
1423:
1424: %% TRACER IRON SPECTRUM
1425: \begin{figure}[h]
1426: \begin{center}
1427: \includegraphics [width=0.90\textwidth]{f14.eps}
1428: \end{center}
1429: \caption{Energy spectra multiplied with $E^{2.5}$ for O, Ne, and Fe,
1430: for different observations: TRACER (solid squares), HEAO-3 (diamonds:
1431: \citet{engelmann90}), CRN (crosses: \citet{muller91}), RUNJOB (filled
1432: circles: \citet{derbina05}) and HESS (open stars: QGSJET model, filled
1433: stars: SYBILL model; \citet{aharonian07a}). The error bars shown are
1434: statistical.}
1435: \label{fe-ne-o}
1436: \end{figure}
1437:
1438:
1439:
1440: %% TRACER AIRSHOWERS
1441: \begin{figure}[h]
1442: \begin{center}
1443: \includegraphics [width=0.90\textwidth]{f15.eps}
1444: \end{center}
1445: \caption{Energy Spectra from TRACER (solid squares) compared with the
1446: interpretation of air shower data of KASCADE (solid triangles, for two
1447: different interaction models : \citet{antoni05}) and of EAS-TOP (open
1448: crosses, two data points for each energy represent upper and lower
1449: limits: \citet{navarra03}). The spectra are for oxygen and for iron
1450: for TRACER, but for the ``CNO-group'' and the ``Fe-group'' for the
1451: other observations.}
1452: \label{airshower}
1453: \end{figure}
1454:
1455: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1456:
1457:
1458: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1459:
1460:
1461: %% The following command ends your manuscript. LaTeX will ignore any text
1462: %% that appears after it.
1463:
1464: \end{document}
1465:
1466: %%
1467: %% End of file `sample.tex'.
1468: