gr-qc0401082/MGX.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %% ws-procs975x65.tex   :   10 October 2003
3: %% Text file to use with ws-procs975x65.cls written in Latex2E.   
4: %% The content, structure, format and layout of this style file is the 
5: %% property of World Scientific Publishing Co. Pte. Ltd. 
6: %% Copyright 1995, 2002 by World Scientific Publishing Co. 
7: %% All rights are reserved.
8: %%
9: %% Proceedings Trim Size: 9.75in x 6.5in
10: %% Text Area: 8in (include runningheads) x 5in
11: %% Main Text is 10/13pt					  
12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
13: %%
14: 
15: %\documentclass[draft]{ws-procs975x65}
16: \documentclass{ws-procs975x65}
17: 
18: \begin{document}
19: 
20: \title{Phenomenology of Space Time Fluctuations}
21: 
22: \author{R. Aloisio\footnote{invited speaker}} 
23: 
24: \address{INFN - Laboratori Nazionali Gran Sasso \\
25: SS 17 bis, Assergi (AQ) Italy, \\ 
26: E-mail: roberto.aloisio@lngs.infn.it}
27: 
28: \author{P. Blasi}
29: 
30: \address{INAF - Osservatorio Astrofisico Arcetri \\
31: Largo E. Fermi 5, 50125 Firenze Italy\\ 
32: E-mail: blasi@arcetri.astro.it}
33: 
34: \author{A. Galante}
35: 
36: \address{Dipartimento di Fisica, Universit\`a di L'Aquila\\
37: Via Vetoio, 67100 Coppito (AQ) Italy\\ 
38: E-mail: angelo.galante@lngs.infn.it}
39: 
40: \author{A.F. Grillo}
41: 
42: \address{INFN - Laboratori Nazionali Gran Sasso \\
43: SS 17 bis, Assergi (AQ) Italy, \\ 
44: E-mail: aurelio.grillo@lngs.infn.it}
45: 
46: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
47: % You may repeat \author \address as often as necessary      %
48: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49: 
50: \maketitle
51: 
52: \abstracts{
53: Quantum gravitational effects may induce stochastic fluctuations in the
54: structure of space-time, to produce a characteristic foamy structure. 
55: It has been known for some time now that these fluctuations may have
56: observable consequences for the propagation of cosmic ray particles
57: over cosmological distances. While invoked as a possible explanation for 
58: the detection of the puzzling cosmic rays with energies in excess of the 
59: threshold for photopion production (the so-called super-GZK particles), we
60: demonstrate here that lower energy observations may provide 
61: strong constraints on the role of a fluctuating space-time structure.
62: We note also that the same fluctuations, if they exist, imply that 
63: some decay reactions normally forbidden by elementary conservation laws, 
64: become kinematically allowed, inducing the decay of particles that are seen 
65: to be stable in our universe. Due to the strength of the prediction, we are 
66: led to consider this finding as the most severe constraint on the classes of 
67: models that may describe the effects of gravity on the structure of space-time.
68: We also propose and discuss several potential loopholes of our approach, 
69: that may affect our conclusions. In particular, we try to identify the
70: situations in which despite a fluctuating energy-momentum of the
71: particles, the reactions mentioned above may not take place.
72: }
73: 
74: \section{Introduction}
75: 
76: In the last few years the hunt for possible minuscule violations of the
77: fundamental Lorentz invariance (LI) has been object of renewed interest,
78: in particular because it has been understood that cosmic ray physics has
79: an unprecedented potential for investigation in this field
80: \cite{kir,lgm,cam,colgla,noi,spain}. Some authors \cite{cam,colgla,berto}
81: have even invoked possible violations of LI as a plausible explanation to
82: some puzzling observations related to the detection of ultra high energy
83: cosmic rays (UHECRs) with energy above the so-called GZK feature \cite{gzk}, 
84: and to the unexpected shape of the spectrum of photons with super-TeV 
85: energy from sources at cosmological distances.
86:  
87: Both types of observations have in fact 
88: many uncertainties, either coming from limited statistics of very rare events,
89: or from accuracy issues in the energy determination of the detected 
90: particles, and most likely the solution to the alleged puzzles will come from 
91: more accurate observations rather than by a violation of fundamental 
92: symmetries.
93:  
94: For this reason, from the very beginning we proposed \cite{noi} that 
95: cosmic ray observations should be used as an ideal tool to constrain 
96: the minuscule violations of LI, rather than as evidence for the need 
97: to violate LI. The reason why the cases of UHECRs and TeV gamma rays 
98: represent such good test sites for LI is that both are related to physical
99: processes with a kinematical energy threshold, which is in turn very sensitive
100: to the smallest violations of LI. UHECRs are expected to suffer severe 
101: energy losses due to photopion production off the photons of the cosmic 
102: microwave background (CMB), and this should suppress the flux of particles 
103: at the Earth at energies above $\sim 10^{20}$ eV, the so called GZK feature. 
104: 
105: Present operating experiments are AGASA \cite{AGASA}
106: and HiRes \cite{Hires}, and they do not provide strong evidence either 
107: in favor or 
108: against the detection of the GZK feature \cite{demarco}. A substantial 
109: increase in the statistics of events, as expected with the Auger project
110: \cite{Auger}
111: and with EUSO \cite{EUSO}, 
112: should dramatically change the situation and allow to detect
113: the presence or lack of the GZK feature in the spectrum of UHECRs. These
114: are the observations that will provide the right ground for imposing a 
115: strong limit on violations of LI. For the case of TeV sources, the process
116: involved is pair production \cite{gamgam} of high energy gamma rays on
117: the photons of the infrared background. In both cases, a small violation
118: of LI can move the threshold to energies which are smaller than the 
119: classical ones, or move them to infinity, making the reactions impossible.
120: The detection of the GZK suppression or the cutoff in the gamma ray 
121: spectra of gamma ray sources at cosmological distances will prove that 
122: LI is preserved to correspondingly high accuracy \cite{noi}.
123: 
124: The recipes for the violations of LI generally consist of requiring an 
125: {\it explicit} modification of the dispersion relation of high energy 
126: particles, due to their propagation in the ``vacuum'', now affected by 
127: quantum gravity (QG). This effect is generally parametrized by introducing 
128: a typical mass, expected to be of the order of the Planck mass ($M_P$), 
129: that sets the scale for QG to become effective. 
130: 
131: However, explicit modifications of the dispersion relation are not really
132: necessary in order to produce detectable effects, as was recently
133: pointed out in Refs. \cite{ford,ng1,ng2,lieu} for the case of propagation 
134: of UHECRs. It is in fact generally believed that coordinate measurements 
135: cannot be performed with precision better than the Planck distance (time) 
136: $\delta x \geq l_P$, namely the distance where the metric of space-time 
137: must feature quantum fluctuations.
138: A similar line of thought implies that an uncertainty in the measurement
139: of energy and momentum of particles can be expected, according with the  
140: relation $\delta p \simeq \delta E \simeq p^2/M_P$. As discussed also in 
141: Refs. \cite{ng1,ng2} the apparent problem of super-GZK particles
142: may find a solution also in the context of this uncertainty approach.
143: 
144: We discuss here this appealing approach more in detail, by taking into 
145: account the effects of the propagation of CRs in the QG vacuum in the 
146: presence of the universal microwave background radiation. A fluctuating
147: metric implies that different measurements of the particle energy or 
148: momentum may result in different outcomes. Therefore it becomes important to
149: define the probability that the {\it measured} energy (momentum) of a particle 
150: is above some fixed value. Note that averaging
151: over a large number of measurements would yield the {\it classical} values
152: for the energy and momentum. The process of measurement mentioned above, 
153: during the propagation of particles over cosmological distances occurs at
154: each single interaction of the particle with the environment. At each
155: interaction vertex, the fluctuating energy/momentum of the particle is 
156: compared with the kinematic threshold for the occurrence of some physical 
157: process (in our case the photopion production).
158: A clear consequence of this approach is that particles with classical energy
159: below the standard Lorentz invariant threshold have a certain probability of 
160: interacting. In the same way, particles above the classical threshold have a 
161: finite probability of evading interaction. We show here that the most striking
162: consequences of the approach described above derive from low energy particles 
163: rather than from particles otherwise above the threshold for photopion 
164: production. 
165: 
166: However, the possibility of a fluctuating energy and momentum is mainly 
167: constrained by other processes that could arise. The fluctuations 
168: of energy and momentum are responsible, infact, for decaying processes 
169: otherwise impossible, typically prevented by energy and momentum conservation.
170: These decaying processes represent the most stringent test of the proposed 
171: model. In the present paper we will discuss these decaying processes, showing 
172: how they could arise. From a general point of view a particle 
173: propagating in a fluctuating vacuum acquires an energy dependent fluctuating 
174: effective mass (the fluctuating dispersion relations introduced in 
175: \cite{noi2}) which may be responsible for kinematically forbidden decay 
176: reactions to become kinematically allowed. 
177: 
178: If this happens, particles that are known to be stable would decay, provided 
179: no other fundamental conservation law is violated (e.g.: baryon number 
180: conservation, charge conservation). A representative example is that of the 
181: reaction $p\to p+\pi^0$, that is prevented from taking place only due to 
182: energy conservation. With a fluctuating metric, we find that if the initial 
183: proton has energy above a few $10^{15}$ eV, the reaction above can take place 
184: with a cross section typical of hadronic interactions, so that the proton 
185: would rapidly lose its energy. Similar conclusions hold for the 
186: electromagnetic process $p\to p+\gamma$.
187: 
188: The fact that particles that would be otherwise stable could
189: decay has been known for some time now \cite{gmestres,liberati} and
190: in fact it rules out a class of non-fluctuating modifications of the 
191: dispersion relations for some choices of the sign of the modification: 
192: the new point here is that it does not appear to be possible to fix the 
193: sign of the fluctuations, so that the conclusions illustrated above 
194: seem unavoidable. This result represents the most striking test of the 
195: fluctuating picture discussed in this paper and could in principle invalidate 
196: the basis of the proposed model itself.
197: 
198: The plan of the paper is the following: in \S 2 we discuss the effect of 
199: fluctuations on the propagation of high energy particles, setting also the
200: computational framework of the paper.
201: In \S 3, we discuss, mainly from the astrophysical point of view, the 
202: possibility of putting under experimental scrutiny some of the conclusions
203: reached in \S 2. In section \S 4 we will discuss the decays of stable 
204: particles induced by fluctuations. Finally in section \S 5 we argue that the 
205: comparison of our predictions with experimental 
206: data indicates a strong inconsistency, implying that the framework of 
207: quantum fluctuations currently discussed in most literature is in fact 
208: ruled out. The strength of this conclusion leads us to try to identify 
209: possible loopholes in our working assumptions. The ways to avoid the
210: dramatic effects of the fluctuating energy-momentum of a particle should
211: be mainly searched in the dynamics of Quantum Gravity.
212: These effects, in which our knowledge is poor to say the least, might 
213: forbid processes even when these processes are kinematically allowed
214: due to the fluctuations in the energy and momentum. 
215: 
216: \section{The effect of Space-Time fluctuations on the propagation
217: of high energy particles.}
218: 
219: While electroweak and strong interactions propagate through space-time,
220: gravity turns out to be a property of the space-time itself. This simple 
221: statement has profound implications in the quantization of gravity. Our 
222: belief that gravity can be turned into a quantum theory immediately implies 
223: that the structure of space-time has quantum fluctuations itself. 
224: Another way of rephrasing this concept is that space-time is expected to have 
225: a granular (or foamy) structure, where however the size of space-time cells 
226: fluctuates stochastically, thereby causing an intrinsic uncertainty in the 
227: measurements of space-time lengths, and indirectly of energy and momentum of 
228: a particle moving through space-time. The uncertainty appears on scales 
229: comparable with the Planck scale (the quantization scale of gravity).
230: 
231: It is generally argued that measurements of distances (times) smaller than 
232: the Planck length (time) are conceptually unfeasible, since the process of 
233: measurement collects in a Planck size cell an energy in excess of the Planck
234: mass, hence forming a black hole, in which information is lost.
235: This can be translated in different ways 
236: into an uncertainty on energy-momentum measurements \cite{ng1,ng2}. The
237: Planck length is a good estimate of the uncertainty in the De Broglie 
238: wave-length $\lambda$ of a particle with momentum $p$. Therefore 
239: $\delta \lambda \approx l_P$, and  $\delta p = \delta (1/\lambda) \approx 
240: (p^2 l_P)=(p^2/M_P)$.
241: 
242: Speculating on the exact characteristics of the fluctuations induced by
243: QG is beyond the scope of the present paper, and it would probably be
244: useless anyway, since the current status of QG approaches does not allow
245: such a kind of knowledge. We decided then to adopt a purely phenomenological
246: approach, in which some reasonable assumptions are made concerning the 
247: fluctuations in the fabric of space-time, and their consequences for the 
248: propagation of high energy particles are inferred. Comparison with 
249: experimental data then possibly constrains QG models.
250: 
251: Following \cite{ng1}, we assume that in each measurement:
252: 
253: \begin{itemize}
254: 
255: \item{the values of energy (momentum) fluctuate around their average values
256: (assumed to be the result theoretically recoverable for an infinite number
257: of measurements of the same observable): 
258: \begin{equation}
259:  E \approx  {\bar E} + \alpha \frac{\bar{E}^2}{M_P} 
260: \label{eq:Ebar}
261: \end{equation}
262: \begin{equation}
263:  p \approx  {\bar p} + \beta \frac{\bar{p}^2}{M_P} 
264: \label{eq:pbar}
265: \end{equation}
266: with $\alpha, \beta$ normally distributed variables and $p$ the modulus 
267: of the 3-momentum (for simplicity we assume rotationally invariant 
268: fluctuations);}
269: 
270: \item{the dispersion relation fluctuates as follows: 
271: \begin{equation}
272: P_\mu g^{\mu\nu} P_\nu = E^2-p^2 + \gamma \frac{p^3}{M_P}=m^2
273: \label{eq:PmuPmu}
274: \end{equation}
275: and $\gamma$ is again a normally distributed variable.}
276: 
277: \end{itemize}
278: 
279: Ideally, QG should predict the type of fluctuations introduced above, but,
280: as already stressed, this is currently out of reach, therefore we assume here 
281: that the fluctuations
282: are gaussian. Our conclusions are however not sensitive to this assumption:
283: essentially any symmetrical 
284: distribution with variance $\approx 1$, within a large factor, would
285: give essentially the same results.
286: Furthermore we {\it assume} that  $\alpha$, $\beta$ and $\gamma$ 
287: are uncorrelated random variables; again, this assumption reflects our 
288: ignorance in the dynamics of QG 
289: 
290: The fluctuations described above will in general derive from metric 
291: fluctuations of magnitude $\delta g^{\mu\nu} \sim h^{\mu\nu} \frac{l_P}{l}$ 
292: \cite{cam,ng2}. Our assumption reflects the fact that, while the magnitude of
293: the fluctuation can be guessed, we do not make any assumption on its 
294: tensorial structure $h^{\mu\nu}$.
295: 
296: Our interest will be now concentrated upon processes of the type
297: 
298: $$a+b\to c+d$$
299: 
300: where we assume that a kinematic threshold is present; in the realm of 
301: UHECR physics (a,b) is either ($\gamma, \gamma_{3K}$)
302: or ($p,\gamma_{3K}$) and (c,d) is ($e^+,e^-$) or ($N,\pi$).
303: 
304: To find the value of initial momenta for which the reaction occurs we write
305: down energy-momentum conservation equations and solve them with the
306: help of the dispersion relations, as discussed in detail in \cite{noi}.
307: 
308: The energy momentum conservation relations are (in the laboratory frame,
309: and specializing to the case in which the target (b) is a low energy
310: background photon for which fluctuations can be entirely neglected)
311: \begin{equation}
312: E_a+ \alpha_a \frac{E^2_a}{M_P} + \omega = 
313: E_c+ \alpha_c \frac{E^2_c}{M_P}+ E_d + \alpha_d \frac{E^2_d}{M_P}
314: \label{eq:disp1}
315: \end{equation}
316: \begin{equation}
317: p_a+ \beta_a \frac{p^2_a}{M_P} - \omega = 
318: p_c+ \beta_c \frac{p^2_c}{M_P}+ p_d + \beta_d \frac{p^2_d}{M_P}.
319: \label{eq:disp2}
320: \end{equation}
321: These equations refer to head-on collisions and collinear
322: reaction products, which is appropriate for threshold computations. Together
323: with the modified dispersion relations, these equations, after 
324: some manipulations, lead to a cubic equation for the initial momentum 
325: as a function of the momentum of one of products, and, after minimization, 
326: they define the threshold for the process considered. In figure 1 we 
327: report the distribution of thresholds in the $\approx 70 \%$ of cases
328: in which the solution is physical; in the other cases the kinematics 
329: does not allow the reaction.
330: 
331: This threshold distribution can be interpreted in the 
332: following way: a particle with energy above $\sim 10^{15}$ eV has essentially 
333: $70 \%$ probability of being above threshold, and therefore to be absorbed. 
334: In the other $30 \%$ of the cases the protons do not interact.
335: 
336: In (\ref{eq:disp1},\ref{eq:disp2}) the fluctuations are taken independently 
337: for each particle, which is 
338: justified as long as the energies are appreciably smaller than the Planck
339: energy. At that point it becomes plausible that different particles 
340: experience the same fluctuations, or more precisely fluctuations of the
341: same region of space-time. It is instructive to consider this case in some 
342: more detail: we introduce then the four-momenta (and dispersion 
343: relations) of {\it all} particles fluctuating in the same way. 
344: Specializing to proton interaction on CMBR, the equation which defines
345: the threshold $p_{th}$ is \cite{noi}:
346: \begin{equation}
347: \eta \frac{2 p_0^3}{(m_{\pi}^2+2m_{\pi}m_p)  M_P} 
348: \frac{m_{\pi}m_p}{(m_{\pi}+m_p)^2} 
349: \left ( \frac{p_{th}}{p_0} \right ) ^3   
350: + \left ( \frac{p_{th}}{p_0} \right )   -1 = 0
351: \end{equation}
352: where $\eta$ is a gaussian variable with zero average and variance
353: of the order of (but not exactly equal to) one, and $p_0$ is
354: the L.I. threshold (GZK). The threshold is the positive solution of
355: this equation.
356: 
357: \begin{figure}[ht]
358: %\epsfxsize=10cm   %width of figure - will enlarge/reduce the figures
359: %\epsfbox{fig3.eps}
360: %\figurebox{2cm}{3cm}{} %to have a box alone 
361: \centerline{\epsfxsize=3.8in\epsfbox{fluct.eps}}   
362: \caption{Threshold distribution for $p \gamma_{3^oK} \to 
363: N \pi$. In the $30\%$ of cases the reaction is not allowed.}
364: \end{figure}
365: \eject
366: 
367: The coefficient of the cubic term is very large, of the 
368: order of $10^{13}$ in this case, so that unless $\eta$ 
369: is $O(10^{-13})$,
370: we can write, neglecting pion mass
371: \begin{equation}
372: p_{th}\approx p_0 \left ( \frac{m_p^2 M_P}{\eta p_0^3} 
373: \right )^{\frac{1}{3}}.
374: \label{eq:thre}
375: \end{equation}
376: When $\eta$ becomes negative, the above equation has no 
377: positive root; this happens essentially in $50 \%$ of the cases.
378: Since the gaussian distribution is  flat in a small interval 
379: around zero, the distribution of thresholds for positive $\eta$ 
380: peaks around the 
381: value for $\eta \approx 1$, meaning that the threshold moves almost 
382: always down to a value of $\approx 10^{15}$ eV \cite{noi}; 
383: essentially the same result holds for fluctuations affecting only  the 
384: incident (highest energy) particle.
385: For {\it independent} fluctuations of final momenta,
386: the asymmetry in the probability 
387: distribution of allowed thresholds arises from the fact that 
388: even
389: exceedingly small negative values of the fluctuations lead 
390: to unphysical solutions. 
391: 
392: Building upon our findings, we now apply the same calculations to 
393: the case of UHECR protons propagating on cosmological distances.
394: An additional ingredient is needed to complete the dynamics of the
395: process of photopion production, namely the cross section. The rather
396: strong assumption adopted here is that the cross section remains the
397: same as the Lorentz invariant one, provided the reaction is 
398: {\it kinematically} allowed. This implies that the interaction lengths
399: remain unchanged.
400: 
401: In order to assess the situation of UHECRs, we first consider the 
402: case of particles above the threshold for photopion production in 
403: a Lorentz invariant world. According with eqs. (\ref{eq:disp1},
404: \ref{eq:disp2}),
405: in this case particles have a probability of $\approx 30 \%$ of being
406: not kinematically allowed to interact inelastically with a photon in the 
407: CMBR. Therefore, if our assumption on the invariance of the interaction 
408: length is correct, then each proton is still expected to make photopion
409: production, although with a slightly larger pathlength. 
410: 
411: The situation is however even more interesting for particles that are
412: below the Lorentz invariant threshold for the process of photopion 
413: production. If the energy
414: is below a few $10^{18}$ eV, a galactic origin seems to be in good 
415: agreement with measurements of the anisotropy of cosmic ray arrival 
416: directions \cite{agasa_a,fly_a}. We will not consider these energies 
417: any longer. On the other hand, at energies in excess of $10^{19}$ eV,
418: cosmic rays are believed to be extragalactic protons, mainly on the ground of
419: the comparison of the size of the magnetized region of our Galaxy and
420: the Larmor radius of these particles. 
421: We take these pieces of information as the basis 
422: for our line of thought. If the cosmic rays observed in the energy range 
423: $E> 10^{19} \rm{eV}$ are extragalactic protons,
424: then our previous calculations apply and we may expect that these particles 
425: have a $\sim 70\%$ probability of suffering photopion 
426: production at each interaction with the CMB photons, even if their energy 
427: is below the classical threshold for this process. Note that the pathlength 
428: associated with the process is of the order of the typical pathlength for 
429: photopion production (a few  tens of Mpc), therefore we are here discussing 
430: a dramatic process in which the absorption length of particles drops from 
431: Gpc, which would be pertinent to particles with energy below 
432: $\sim 10^{20}$ eV in a Lorentz invariant world, to several Mpc, with a 
433: corresponding suppression of the flux. 
434: What are the consequences for the observed fluxes of cosmic rays? 
435: The above result implies that {\it all} protons with $E>10^{15}$ eV
436: are produced within a radius of 
437: several tens of Mpc, and above this energy there is no dramatic change 
438: of pathlength with energy. 
439: There is no longer anything
440: special about  $E \sim 10^{20}$ eV, and 
441: any mechanism invoked to explain the flux of super-GZK particles must
442: be at work also at lower energies. 
443: 
444: The basic situation remains the same in the case of pair production as the
445: physical process under consideration. For a source at cosmological distance,
446: a cutoff is expected due to pair production off the far infrared background
447: (FIR) or the microwave background. Using the results in \cite{noi} we expect
448: that the modified thresholds are a factor $0.06$ ($0.73$) lower than the 
449: Lorentz invariant ones for the case of interaction on the CMBR (FIR).
450: There is also a small increase in the pathlengths above the threshold,
451: which would appear exponentially in the expression for the flux. Therefore
452: there are two effects that go in opposite directions: the first moves the
453: threshold to even lower energies, and the second increases the flux of
454: radiation at Earth because of the increase of the pathlength. 
455: It seems that geometry fluctuations do not provide an immediate explanation
456: of the possible detection of particles in excess of the expected ones from
457: distance sources in the TeV region. In any case the experimental evidence 
458: for such an excess seems at present all but established.
459: 
460: \section{Astrophysical observations}
461: 
462: As discussed in the previous section fluctuations in the space-time metric 
463: may induce a violation of Lorentz invariance that changes the thresholds for 
464: the photopion production of a very high energy proton 
465: off the photons of the CMBR, or for the pair production of a high energy
466: gamma ray in the bath of the FIR or CMBR photons.
467: 
468: For the case of UHECRs interacting with the CMBR, we obtained a picture that 
469: changes radically our view of the effect of QG on this phenomenon, as 
470: introduced in previous papers: not only particles with energy above 
471: $\sim 10^{20}$ eV are affected by the fluctuations in space-time, but also 
472: particles with lower energy, down to $\sim 10^{15}$ eV seem to be affected
473: by such fluctuations. In fact the latter, as a result of a fluctuating 
474: space-time, may end up being above the threshold for photopion production, 
475: so that particles may suffer significant absorption. Our conclusion is that 
476: all particles with energy in excess of $\sim 10^{15}$ eV eventually detected 
477: at Earth would be generated at distances comparable with the pathlength 
478: for photopion production ($\sim 100$ Mpc). 
479: A consequence of this is that there is no longer anything special 
480: characterizing the energy $\sim 10^{20}$ eV. 
481: 
482: Since the conclusion reached in the previous section is quite strong, it is 
483: important to summarize in detail some tests that may allow to understand 
484: whether the current or future astrophysical observations are compatible with 
485: the scenario discussed in this paper.
486: 
487: a)  Future experiments \cite{Auger,EUSO} 
488: dedicated to the detection of UHECRs will provide a
489: substantial increase in the statistics, so that the spectral features of 
490: the UHECRs in the energy region $E>10^{19}$ eV can be resolved, and
491:  further indications on the nature of primaries and their 
492: possible extragalactic origin will be obtained. In particular
493: the present possible disagreement between  
494: AGASA \cite{agasa_f} and HiRes \cite{Hires_a} will be clarified.
495: 
496: One should also keep in mind that an evaluation of the expected flux in 
497: terms of sources distributed as normal galaxies is
498: in contradiction with AGASA data by an amount ranging from $2$ to $6 \sigma$
499: depending on the assumed source spectrum \cite{blanton}. 
500: Since the nature of the sources is not known, it is not
501: clear if their  abundance within the absorption pathlength
502: is sufficient to explain the observed flux in presence of space-time 
503: fluctuations, nor if they can induce observable anisotropies.
504: 
505: In any case, in a Lorentz invariant framework a suppression in the flux 
506: at $\sim 10^{20}$ eV is expected. If such a feature is 
507: unambiguously detected in the UHECR spectrum, no much room would be left for
508: the fluctuations of space-time discussed in this paper, since in this scenario
509: nothing special happens around $10^{20}$ eV.
510: In quantitative terms \cite{noi} 
511: this would imply  a phenomenological bound on $l_P$ 
512: now interpreted as a parameter: 
513:  $l_P < 10^{-46}$ cm instead of 
514: $l_P \approx 10^{-33}$ cm; in other words, only fluctuations with 
515: variance $\approx 10^{-13}$, instead of $1$, would be allowed
516: \footnote{ Alternatively, 
517: one can assume a more general form of fluctuations, i.e. $\delta E \approx 
518: E (E/M_P)^{\alpha}$ and similar for momentum and dispersion relations 
519: \cite{tom}. In this case the basic conclusions reached here remain unchanged.}
520: 
521: b) According with our findings, all particles with energy in excess of 
522: $\sim 10^{15}$ eV lose their energy by photopion production on cosmological
523: spatial scales, as a result of the metric fluctuations. This energy ends up 
524: mainly in gamma rays, neutrinos and protons. The protons pile up in the
525: energy region right below $\sim 10^{15}$ eV. The gamma ray component
526: actually generates an electromagnetic cascade that ends up contributing 
527: low energy gamma rays, in the energy band accessible to instruments like
528: EGRET \cite{EGRET} and GLAST \cite{GLAST}. 
529: This cascade flux cannot be larger than the measured 
530: electromagnetic energy density in the same band
531: $\omega_{cas}^{exp}=10^{-6}~eV/cm^{3}$ \cite{EGRET}. The cascade flux in our 
532: scenario can be estimated as follows. Let $\Phi(E)=\Phi_0 (E/E_0)^{-\gamma}$
533: be the emissivity in UHECRs ($\rm{particles}/cm^3/s/GeV$). Let us choose the
534: energy $E_0=10^{10}$ GeV and let us normalize the flux to the observations
535: at the energy $E_0$. The total energy going into the cascade can be shown 
536: to be $$\omega_{cas} \approx \frac{5\times 10^{-4}}{\gamma-2}~~ 
537: x_{min}^{2-\gamma} ~~\xi ~~eV ~cm^{-3},$$ 
538: where $\xi$ is the fraction of energy going into gamma rays in each 
539: photopion production, and $x_{min}=(E_{th}/E_0)=10^{-4}$ for $E_{th}=
540: 10^{15}$ eV. It is easy to see that, for $\gamma=2.7$, the cascade bound 
541: is violated unless $\xi \ll 10^{-3}$.
542: 
543: One note of warning has to be sent concerning the development of the 
544: electromagnetic cascade: the same violations of LI discussed here affect 
545: other processes, as stressed in the paper. For instance pair production 
546: and pion decay are also affected by violations of LI \cite{amepai}. 
547: Therefore the possibility that the cascade limit is exceeded concerns 
548: only those scenarios of violations of LI that do not inhibit appreciably 
549: pair production and the decay of neutral pions.
550: 
551: The protons piled up at energies right below $10^{15}$ eV, would be a 
552: nice signature of this scenario, but it seems difficult to envision a 
553: way of detecting these remnants. In fact, even a tiny magnetic field
554: on cosmological scales would make the arrival time of these particles to
555: Earth larger than the age of the universe. Moreover, even assuming an exactly
556: zero extragalactic magnetic field, these particles need to penetrate the 
557: magnetic field of our own Galaxy and mix with the galactic cosmic rays, 
558: making their detection extremely problematic if not impossible.
559: 
560: Clearly a more detailed flux computation, taking into account propagation of 
561: primaries as well as generation and propagation of the secondaries is needed 
562: in order to assess in a more quantitative way observable effects of possible 
563: metric fluctuations on UHECRs.
564: 
565: Let us conclude this section sending a note of warning concerning 
566: Eq. (\ref{eq:thre}), in this expression the dependence on the CMB photon
567: energy is washed out by the approximation done (we have neglected the pion 
568: mass). From the physical point of view this corresponds to the appearence 
569: of an effective mass (momentum dependent) of the proton due to the effect 
570: of fluctuations. The effective mass of the proton may be responsible 
571: for the decay of this particle. As we will discuss in the next section 
572: the possibility of a decaying proton is a very stringent test for the 
573: fluctuations picture much powerful than the astrophysical observations 
574: discussed in the present section. 
575: 
576: \section{Decay of stable particles}
577: 
578: Let us discuss in this section the most striking test of the models 
579: that predict energy and momentum fluctuations. We will discuss here 
580: the possibility that these fluctuations may induce particles decays otherwise 
581: impossible. This possibility, already discussed in the framework of 
582: non-fluctuating modifications of the dispersion relation 
583: \cite{gmestres,liberati}, could in principle rule out the models with 
584: fluctuations. In this section we will discuss the basic features of the
585: decays, leaving a detailed discussion of the implications and possible way
586: out to the next section.
587: 
588: We will consider three specific decay channels, that illustrate
589: well, in our opinion, the consequences of the quantum fluctuations introduced
590: above. We start with the reaction 
591: $$p\to p + \pi^0$$
592: and we denote with $p$ ($p'$) the momentum of the initial (final) proton, and
593: with $k$ the momentum of the pion. Clearly this reaction cannot take place 
594: in the reality as we know it, due to energy conservation. However, since 
595: fluctuations have the effect of emulating an effective mass of the particles, 
596: it may happen that for some realizations, the effective mass induced to the 
597: final proton is smaller than the mass of the proton in the initial 
598: state, therefore allowing the decay from the kinematical point of view. Since
599: no conservation law or discrete symmetry is violated in this reaction, it 
600: may potentially take place. For the sake of clarity, it may be useful to 
601: invoke as an example the decay of the $\Delta^+$ resonance, which is 
602: structurally identical to a proton, but may decay to a proton and a pion
603: according to the reaction $\Delta^+\to p+\pi^0$, since its mass is larger
604: than that of a proton. From the physical point of view, the effect of the
605: quantum fluctuations may be imagined as that of {\it exciting} the proton,
606: inducing a mass slightly larger than its own (average) physical mass.
607: 
608: Following the discussion of the previous sections we expect to find that for 
609: momenta above a given threshold, depending on the value of the random 
610: variables, the decay may become kinematically allowed. In general, the 
611: probability for this to happen has to be calculated  numerically from the 
612: conservation equations supplemented by the dispersion relations \cite{noi3}. 
613: 
614: Although a full calculation is possible, it is probably more instructive
615: to proceed in a simplified way, in which only the fluctuations in
616: the dispersion relation of the particle in the initial state are taken
617: into account. Neglecting the corresponding fluctuations in the final state
618: should not affect the conclusions in any appreciable way, unless the
619: fluctuations in the initial and final states are correlated (we will return 
620: to this possibility at the end of section \S 5).
621: 
622: In this approximation, the threshold for the process of proton decay to a 
623: proton and a neutral pion can be written as follows (neglecting corrections 
624: to order higher than $p/M_P$):
625: 
626: \begin{equation}
627: \gamma \frac{2 p_{th}^3}{M_P} 
628: -2 m_{\pi} m_p - m_{\pi}^2 =0,
629: \end{equation}
630: with solution
631: \begin{equation}
632: p_{th}= \left (\frac{(2m_p m_{\pi} +  m_{\pi}^2 )M_P}{2 \gamma } 
633: \right )^{\frac{1}{3}}.
634: \end{equation}
635: For negative values of $\gamma$, the above equation has no positive
636: root; this happens in $50\%$ of the cases. Since the gaussian distribution 
637: is essentially flat in a small interval around zero, the distribution of 
638: thresholds for positive $\gamma$ ({\it i.e. } in the remaining 50 $\%$ of the
639: cases) peaks around the value for $\gamma \approx 1$, meaning that the 
640: threshold moves almost always down to a value of $\approx 10^{15}$ eV 
641: \cite{noi,noi2}; essentially the same result holds for generic fluctuations 
642: ({\it i.e.} not confined to the dispersion relations) affecting only the 
643: incident particle, namely the one with the highest energy \cite{noi3}.
644: 
645: The reason why the effects of fluctuations are expected to occur at such 
646: low energies is that starting from that energy region the fluctuation term
647: becomes comparable with the rest mass of the particle. In fact the same 
648: concept of rest mass of a particle may lose its traditional meaning at 
649: sufficiently high energies \cite{lieu}.
650: 
651: It can be numerically confirmed that {\it independent} fluctuations of  
652: momenta (and/or of the dispersion relations) of the decay products are more 
653: likely to make the decay easier rather than more difficult, due to the non 
654: linear dependence of the threshold on the strength of fluctuations:
655: the probability that the decay does not take place is in fact $\approx 30 \%$. 
656: In the remaining cases, the decay will occur if the momentum of the initial 
657: proton is larger than $p_{th}$\cite{noi3}. The distribution of $p_{th}$ 
658: is essentially identical to the one reported in \S 2 for the photopion 
659: production.
660: 
661: All the discussion reported so far remains basically unchanged if similar
662: reactions are considered. For instance the reaction $p \to \pi^+ n$ is
663: kinematically identical to the one discussed above. For all these reactions,
664: we expect that once they become kinematically allowed, the energy loss
665: of the parent baryon is fast. For the case of nuclei, all the decays that do
666: not change the nature of the nucleon leave (A,Z) unchanged, so we do
667: not expect any substantial blocking effect in nuclei.
668: 
669: Another reaction that may be instructive to investigate is the spontaneous 
670: pair production from a single photon, namely \cite{noi3}
671: $$\gamma \to e^+ e^-.$$
672: In this case, following the calculations described above, we obtain the 
673: following expression for the threshold:
674: \begin{equation}
675: p'_{th}= \left (\frac{4 m_e^2 M_P}{2 \gamma' }, 
676: \right )^{\frac{1}{3}}
677: \end{equation}
678: and $p'_{th}$ is of the order of $10^{13}$ eV.
679: Again, if the reaction becomes kinematically allowed, there does not seem 
680: to be any reason why the reaction should not take place with a rate 
681: dictated by the typical cross section of electromagnetic interactions. 
682: 
683: Finally, we propose a third reaction that in its simplicity may represent
684: the clearest example of reactions that should occur in a world in which 
685: quantum fluctuations behave in the way described above. Let us consider
686: a proton that moves in the vacuum with constant velocity, and let us consider
687: the elementary reaction of spontaneous photon emission. In the Lorentz 
688: invariant world the process of photon emission
689: is known to happen only in the presence of an external field that 
690: may provide the conditions for energy and momentum conservation. However, 
691: in the presence of quantum fluctuations, one can think of the gravitational
692: fluctuating field as such an external field, so that the particle can in fact 
693: radiate a photon without being in the presence of a nucleus or some other 
694: external recognizable field. The threshold for this process, calculated
695: following the above procedure, is 
696: \begin{equation}
697: p_{th}'' \approx \left (\frac{m^2 M_P \omega}{\gamma''}
698: \right )^{\frac{1}{4}},
699: \end{equation} 
700: where $\omega$ is the energy of the photon. This threshold approaches zero 
701: when $\omega \to 0$: for instance, if $\omega=1$ eV, then $p_{th} \approx 300$ 
702: GeV for protons and $p_{th} \approx 45$ GeV for electrons. In other words there
703: should be a sizable energy loss of a particle in terms of soft photons. 
704: This process can be viewed as a sort of bremsstrahlung emission of a charged
705: particle in the presence of the (fluctuating) vacuum gravitational potential.
706: 
707: Based on the arguments provided in this section, it appears that all particles
708: that we do know are stable in our world, should instead be unstable at 
709: sufficiently high energy, due to the quantum fluctuations described above. 
710: In the next section we will take a closer look at the implications of the 
711: existence of these quantum fluctuations, and possibly propose some plausible 
712: avenues to avoid these dramatic conclusions.
713: 
714: \section{Discussion and Outlook}
715: 
716: If the decays discussed in the previous section could take place, our
717: universe, at energies above a few PeV or even at much lower energies 
718: might be unstable, nothing like what we actually see.
719: The decays $$nucleon\to nucleon+\pi$$ would start to be kinematically allowed 
720: at energies that are of typical concern for cosmic ray physics, while 
721: the spontaneous emission of photons in vacuum might even start playing
722: a role at much lower energies, testable in laboratory experiments. 
723: Without detailed calculations of energy loss rates it is difficult to 
724: assess the experimental consequences of this process. 
725: 
726: For the nucleon decay, the situation is slightly simpler if we assume
727: that the quantum fluctuations affect only the kinematics but not the
728: dynamics, an assumption also used in in the photopion production study 
729: \cite{noi2}. In this case one would expect 
730: the proton to suffer the decay to a proton and a pion on a time scale of the
731: same order of magnitude of typical decays mediated by strong interactions.
732: This would basically cause no cosmic ray with energy above $\sim 10^{15}$
733: eV to be around, something that appears to be in evident contradiction 
734: with observations 
735: \footnote{From a phenomenological point of view, consistency 
736: with experiments would require either that the variance of the fluctuations 
737: considered above is ridiculously small ($<10^{-24}$) or, allowing more generic
738: fluctuations $\Delta l \propto l_P (l_P/l)^{\alpha}$, that a fairly large value
739: for $\alpha$ should be adopted \cite{noi2}.}. 
740: 
741: In the following we will try to provide a plausible answer to these three very 
742: delicate questions:
743: \begin{enumerate}
744: \item If the particles were kinematically allowed to decay, and there
745: were no fundamental symmetries able to prevent the decay, would it 
746: take place? 
747: 
748: \item Is the form adopted for the quantum fluctuations correct and if so,
749: how general is it?
750: 
751: \item If in fact the form adopted for the fluctuations is correct, how
752: general and unavoidable is the consequence that (experimentally)
753: unobserved decays should take place?
754: \end{enumerate}
755: 
756: Although the result that particles are kinematically allowed to
757: decay is fairly general, the (approximate) lack of relativistic
758: invariance forbids the computation of life-times \footnote{In fact
759: life-times can be in principle estimated in approaches in which it is
760: possible to make transformations between frames \cite{lieu,dsr,jap},
761: despite the lack of LI.}.
762: Two comments are in order: first, the phase space for the decays described
763: above, as calculated in the laboratory frame, is non zero and in fact 
764: it increases with the momentum of the parent particle. The effect of 
765: fluctuations can be seen as the generation of an effective (mass)$^2 
766: \propto p^3/M_P$. A similar effect, although in a slightly different 
767: context, was noted in \cite{colgla}.
768: 
769: Second, we do not expect dynamics to forbid the reactions:
770: one must keep in mind that we are considering very small effects, at
771: momenta much smaller than the Planck scale. For instance the gravitational
772: potential of the vacuum fluctuations is expected to move quarks in a
773: proton to excited levels, not to change its content, nor the properties 
774: of strong interactions. 
775: 
776: There is a subtler possibility, which must be taken very seriously in our
777: opinion, since it might invalidate completely the line of thought illustrated
778: above, namely that the quantum fluctuations of the momenta of the particles
779: involved in a reaction occur on time scales that are enormously smaller than 
780: the typical interaction/decay times. This situation might resemble the
781: so called Quantum Zeno paradox, where continuously checking for the
782: decay of an unstable particle effectively impedes its decay. 
783: This possibility is certainly worth a detailed study, that would however
784: force one to handle the intricacies of matter in a Quantum Gravity regime. 
785: We regard this possibility as the most serious threat to the validity of 
786: the arguments in favor of quantum fluctuations discussed in this paper 
787: and in many others before it. 
788: 
789: Let us turn out attention toward the question about the correctness and
790: generality of the form adopted for the momentum fluctuations. It is 
791: generally accepted that the geometry of space-time suffers profound 
792: modifications at length (time) scales of the order of the Planck
793: length (time), and that this leads to the emergence of a minimum
794: measurable length. This may be reflected in a non commutativity of 
795: space-time and in a generalized form of the uncertainty principle.
796: 
797: The transition from uncertainty in the length or time scales to uncertainty
798: in momenta of particles is undoubtedly more contrived and deserves some
799: attention. The expressions in Eqs. (\ref{eq:Ebar},\ref{eq:pbar}) and 
800: (\ref{eq:PmuPmu}) have been 
801: motivated in various ways \cite{ford,ng1,noi2,lieu,camacho} in previous
802: papers. For instance, the condition $\Delta l \ge l_P$ seems to imply the 
803: following
804: constraint on wavelengths $\Delta \lambda \ge l_P$, otherwise it would be 
805: possible to design an experimental set-up capable of measuring distances with 
806: precision higher than $l_P$. Therefore $\Delta p \propto \Delta (\lambda^{-1})
807: \propto  l_P p^2$. Similar arguments have been proposed, all based to some
808: extent on the de Broglie relation $p \propto \lambda^{-1}$. 
809: 
810: There is certainly no guarantee that the de Broglie relation continues
811: to keep its meaning in the extreme conditions we are discussing, in 
812: particular in models in which the coordinates and coordinate-momentum 
813: commutators are modified with respect to standard quantum mechanics and 
814: the representation of momentum in terms of coordinate derivatives generally 
815: fails. For instance in a specific (although non-relativistic) example 
816: \cite{kempf} the existence of a minimum length is shown to imply that
817: \begin{equation}
818: p = \frac{2}{\pi l_P} \tan \left ( \frac{\pi l_P}{2 \lambda} 
819: \right ).
820: \end{equation}
821: In other words, the de Broglie relation may be modified in such a way that
822: a minimum wavelength corresponds to an unbound momentum.
823: Notice, however, that we are considering here the effects of these
824: modifications at length scales much larger than the Planck scale, where the
825: correction is likely to be negligible. In general, if $p \propto
826: \lambda^{-1}g(l_P/ \lambda)$ then $\Delta p \propto l_P  p^2 +
827: p ~O(l_P^2 p^2)$. Hence, we do not expect that the result shown in the
828: previous Section is appreciably modified.
829: 
830: Last but not least we notice that the fluctuations in the dispersion 
831: relations can be easily derived from fluctuations of the (vacuum) metric in 
832: the form given in \cite{camacho}:
833: \begin{equation}
834: ds^2 = (1+\phi)dt^2-(1+\psi)d{\bf r}^2
835: \end{equation}
836: where $ \phi, ~\psi $ are functions of the position in space-time. 
837: 
838: The fluctuations of the dispersion relation, Eq. (\ref{eq:PmuPmu}), follow 
839: if $\phi \ne \psi$ ({\it i.e.} non conformal fluctuations), assuming at 
840: least approximate validity of the de Broglie relation; if $\phi=\psi$ a 
841: much milder modification (O($p m^2/M_P$)) follows.
842: 
843: Having given plausibility arguments in favor of the form adopted for the
844: fluctuations, at least for the case of non conformal fluctuations, we are 
845: left with the goal of proving an answer to the last question listed above, 
846: namely does a decay actually occur once it is kinematically allowed?
847: Certainly the answer is positive if one continues to assume momentum and
848: energy conservation, and modifications of these conservation laws with 
849: random terms of order $O(p^2/M_P)$ do not change this conclusion.
850: The question then is whether we are justified in assuming energy and momentum
851: conservation in the form used above. For instance, in the so-called Doubly
852: Special Relativity (DSR) \cite{dsr}, theories and in general in models with
853: deformed Poincare' invariance, the conservation relations may be
854: modified in a non trivial, non additive and non abelian way. For instance, 
855: in the case of proton decay considered above, momentum conservation
856: may read as \cite{dsr,jap}
857: \begin{equation}
858: {\bf p}_p \approx {\bf p}'_{p}+(1+l_P E'_p) {\bf p}_{\pi}\quad\quad
859: {\rm or}
860: \quad\quad 
861: {\bf p}_p \approx {\bf p}'_{\pi}+(1+l_P E'_{\pi}) {\bf p}_p.
862: \end{equation}
863: This certainly makes the probability of being above threshold smaller.
864: However in order to qualitatively modify our results this probability
865: should be in fact vanishingly small. For the case of {\it low} energy 
866: cosmic rays, this probability should be of the order of a typical decay
867: time divided by the residence time of cosmic rays (mostly galactic at
868: these energies) in our Galaxy. 
869: 
870: We are led to conclude that allowing for modifications of the conservation 
871: relations does not appear to improve the situation to the point that  
872: the strong conclusions derived in the previous section can be avoided.
873: In the same perspective, cancellation between fixed modifications 
874: of the dispersion relation and fluctuations (of the same order of 
875: magnitude) does not seem a viable way to proceed.
876: 
877: It is important however to notice that we have considered the above 
878: fluctuations as independent. In a full theory one should take into account
879: possible correlations between fluctuations. The effect of correlations is
880: very important because it pushes to higher energies the fluctuation scale
881: of the particle momentum (energy). Let us discuss in more detail this point.
882: Quantum fluctuations of the momenta of the particles involved in a reaction
883: occour on time scales that are much smaller than the typical interaction 
884: time. Particles during the interaction time experience a large number of 
885: fluctuations, typically
886: 
887: $$ N=\frac{\tau}{\tau_P}=\frac{1}{p\tau_P}=\frac{M_P}{p}~, $$
888: where we have used $\tau\sim 1/p$ for the interaction time scale and 
889: $\tau_P\sim 1/M_P$ for the fluctuation time scale. Assuming independent
890: fluctuations of energy and momentum the fluctuation variance $\sigma$ will be
891: 
892: $$ \sigma^2 = \frac{p^3}{M_P\sqrt{N}}=\frac{p^3}{M_P} 
893: \left (\frac{p}{M_P} \right )^{1/2}~, $$
894: and the fluctuation variance becomes of the order of the proton mass 
895: $\sigma\simeq m_p$ already at momentum $p\simeq 10^{17}$ eV. 
896: In this case the situation resembles as discussed above and, for instance, 
897: the decaying of the proton arises already at lower energies. Let us consider 
898: now the case in which there is some degree of correlation in the momentum 
899: (energy) fluctuations. In this case the fluctuation variance $\sigma$ will be
900: 
901: $$ \sigma^2 = \frac{p^3}{M_P N^{\alpha}}=\frac{p^3}{M_P} 
902: \left (\frac{p}{M_P} \right )^{\alpha}~, $$
903: where we have introduced the exponent $\alpha>1/2$ that parametrizes the 
904: effect of correlations. In this case the fluctuation variance becomes of the 
905: order of the proton mass at larger energies, namely $\sigma\simeq m_p$
906: at momentum of the order of
907: 
908: $$ p\simeq M_P \left (\frac{m_p}{M_P} \right )^{\frac{2}{3+\alpha}}~. $$
909: 
910: A detailed analysis of possible correlations
911: between fluctuations, namely an analytic determination of $\alpha$, is 
912: impossible at this stage because it implies a better knowledge of the theory,
913: and in particular of the dynamics of the QG regime.
914: 
915: Finally, a separate discussion is needed for those theories that 
916: include the relativity principle (exemplified by DSR models). The DSR 
917: theories are characterized by an extended Lorentz invariance \cite{dsr}
918: with two separate invariant scales: the light velocity and the Planck length.
919: Moreover, in the low energy limit of DSR, or for distances much larger 
920: than the Planck length, the usual Lorentz invariance is recovered. 
921: 
922: Using these two characteristics of the DSR theories it is easy to proove that 
923: particle kinematics in DSR is the same as in the usual Lorentz invariant 
924: theories. This result holds in the case in which there are no fluctuations
925: of energy and momentum. In the most general case in which fluctuations of 
926: energy and momentum are taken into account it is difficult to prove that the 
927: situation remains unchanged. Nevertheless, if in DSR the relativity principle 
928: remains at work also in the fluctuating case the DSR approach seems the most 
929: promising in order to escape the particles decays discussed in this paper 
930: that seems to invalidate all the other models.
931: 
932: \begin{thebibliography}{0}
933: 
934: \bibitem{kir}
935: D.A. Kirzhnits and V.A. Chechin, Sov. Jour. Nucl. Phys. {\bf 15}, 585 (1971).
936: 
937: \bibitem{lgm}
938: L. Gonzalez-Mestres, Proc. 26th ICRC (Salt Lake City, USA), {\bf 1}, 179 
939: (1999).
940: 
941: \bibitem{cam}
942: G. Amelino Camelia, J. Ellis, N.E. Mavromatos and S. Sarkar
943: Nature {\bf 393} (1998) 763.
944: 
945: \bibitem{colgla}
946: S. Coleman and S.L. Glashow, Phys. Rev. {\bf D59}, 116008 (1999).
947: 
948: \bibitem{noi}
949: R. Aloisio, P. Blasi, P.L. Ghia and A.F. Grillo,
950: Phys. Rev. {\bf D62}, 053010 (2000).
951: 
952: \bibitem{spain}
953: J.M. Carmona, J.L. Cortes, J. Gamboa and  F. Mendez, 
954: {\it preprint} hep-th/0301248. 
955: 
956: \bibitem{berto}
957: O. Bertolami and C.S. Carvalho, Phys. Rev. {\bf D61}, 103002 (2000);
958: O. Bertolami, {\it preprint} astro-ph/0012462.
959:  
960: \bibitem{gzk}
961: K. Greisen, Phys. Rev. Lett. {\bf 16}, 748 (1966);
962: G.T. Zatsepin and V.A. Kuzmin, Pis'ma Zh. Ekps. Teor. Fiz. {\bf 4}, 114  
963: (1966) [JETP Lett. {\bf 4}, 78 (1966)].
964: 
965: \bibitem{AGASA}
966: N. Hayashida {\it et al.} [AGASA collaboration], Phys. Rev. Lett. {\bf 73},
967: 3491 (1994).
968: 
969: \bibitem{Hires} 
970: T. Abu-Zayyad {\it et al.} [Hires collaboration], astro-ph/0208243.
971: 
972: \bibitem{demarco}
973: D. De Marco, P. Blasi and A.V. Olinto,  
974: Astrop. Phys. {\bf 20}, 53 (2003).
975:  
976: \bibitem{Auger}
977: J. Bl\"umer {\it et al.} [Auger Collaboration], J. Phys. {\bf G29} 867 (2003).
978: 
979: \bibitem{EUSO}
980: EUSO space Observatory, http://www.euso-mission.org.
981: 
982: \bibitem{gamgam}
983: A.I. Nikishov, Sov. Phys. - JETP {\bf 14}, 393 (1962);
984: P. Goldreich and P. Morrison, Sov. Phys. - JETP {\bf 18}, 239 (1964);
985: R.J. Gould and G.P. Schreder, Phys. Rev. Lett. {\bf 16}, 252 (1966).
986: 
987: \bibitem{ford}
988: L.H. Ford Int. J. Theor. Phys. {\bf 38} 2941 (1999).
989: 
990: \bibitem{ng1}
991: Y.J. Ng, D.S. Lee, M.C. Oh and H. van Dam, 
992: Phys. Lett. {\bf B507} 236 (2001).
993: 
994: \bibitem{ng2}
995: Y.J. Ng, Int. J. Mod. Phys. {\bf D11} 1585 (2002) 
996: ({\it preprint} gr-qc/0201022).
997: 
998: \bibitem{lieu}
999: R. Lieu, Astrophys. J. {\bf 568} L67 (2002).
1000: 
1001: \bibitem{noi2}
1002: R. Aloisio, P. Blasi, A. Galante, P.L. Ghia and A.F. Grillo,
1003: Astrop. Phys. {\bf 19}, 127 (2003).
1004: 
1005: \bibitem{gmestres}
1006: L. Gonzalez-Mestres, {\it preprint} hep-ph/9905430.
1007: 
1008: \bibitem{liberati}
1009: T. Jacobson, S. Liberati and D. Mattingly, 
1010: Phys. Rev. {\bf D66}, 081302 (2002). 
1011: 
1012: \bibitem{agasa_a}
1013: N.Hagashida {\it et al.}, Astrop. Phys. {\bf 10} 303 (1999).
1014: 
1015: \bibitem{fly_a}
1016: D.J. Bird {\it et al.}, Ap. J. {\bf 511} 739 (1999).
1017: 
1018: \bibitem{agasa_f}
1019: N. Sakay et al. Proceedings of 2001 ICRC.
1020: 
1021: \bibitem{Hires_a}
1022: C.C.H. Jui {\it et al.}, Proceedings of 2001 ICRC.
1023: 
1024: \bibitem{blanton}
1025: M. Blanton, P. Blasi and A.V. Olinto, Astrop. Phys. {\bf 15}, 275 (2001).
1026: 
1027: \bibitem{tom}
1028: M. Jankiewicz, T.W. Kephart and T.J. Weiler, {\it preprint} hep-ph/0312221 
1029: 
1030: \bibitem{EGRET}
1031: P. Sreekumar {\it et al.} [EGRET collaboration], 
1032: Astroph. J. {\bf 494} (1998) 523.
1033: 
1034: \bibitem{GLAST}
1035: GLAST Tlescope, http://www-glast.stanford.edu.
1036: 
1037: \bibitem{amepai}
1038: G. Amelino-Camelia, Phys. Lett. {\bf B528} (2002) 181.
1039: 
1040: \bibitem{noi3}
1041: R. Aloisio, P. Blasi, A. Galante and A.F. Grillo, 
1042: Astrop. Phys. {\bf 20}, 369 (2003). 
1043: 
1044: \bibitem{dsr}
1045: N.R. Bruno, G. Amelino-Camelia and J. Kowalski-Glikman, 
1046: Phys. Lett. {\bf B522} 133 (2001). G. Amelino-Camelia, 
1047: Int. J. Mod. Phys. {\bf D11} 1643 (2002). G. Amelino-Camelia
1048: Nature {\bf 418} 34 (2002).
1049: 
1050: \bibitem{jap}
1051: T. Tamaki, T. Harada, U. Miyamoto and T. Torii, 
1052: Phys. Rev. {\bf D65} 083003 (2002).
1053: 
1054: \bibitem{camacho}
1055: A. Camacho, Gen. Rel. Grav. {\bf 35} 319 (2003). 
1056: 
1057: \bibitem{kempf}
1058: A. Kempf, G. Mangano and R.B. Mann, Phys. Rev. {\bf D52} 1108 (1995).
1059: 
1060: 
1061: \end{thebibliography}
1062: 
1063: \end{document}
1064: 
1065: 
1066: