hep-ph0404015/2.1.tex
1: \documentclass[preprint,aps,showpacs]{revtex4}
2: \usepackage{epsfig}
3: %\usefont{U}{dsrom}{m}{n}
4: \oddsidemargin=0.0in
5: \evensidemargin=0.0in
6: \topmargin=0.25in
7: \textwidth=6.45in
8: \textheight=8.95in
9: \headheight=15pt
10: \newcommand{\Be}{\begin{equation}}
11: \newcommand{\Ee}{\end{equation}}
12: \newcommand{\Bea}{\begin{eqnarray}}
13: \newcommand{\Eea}{\end{eqnarray}}
14: \newcommand{\nonum}{\nonumber} 
15: \newcommand{\NL}{\nonumber \\}                    %%
16: \newcommand{\mtxfnt}{\usefont{U}{dsss}{m}{n} \selectfont}  %useSpecialFont 
17: \setcounter{page}{0}
18: 
19: \begin{document}
20: 
21: \setcounter{page}{0}
22: 
23: \title{Large Mixing from Small: \\
24: Pseudo-Dirac Neutrinos and the Singular Seesaw}
25: \author{G.J. Stephenson, Jr.}\email{ GJS@baryon.phys.unm.edu}
26: \affiliation{Department of Physics and Astronomy, 
27: University of New Mexico, Albuquerque, NM 87131}
28: \author{T. Goldman}\email{ tgoldman@lanl.gov}
29: \affiliation{Theoretical Division, MS-B283, 
30: Los Alamos National Laboratory, Los Alamos, NM 87545}
31: \author{B.H.J. McKellar}\email
32: { b.mckellar@physics.unimelb.edu.au} 
33: \author{M. Garbutt}\email{ mgarbutt@treasury.gov.au}
34: \affiliation{University of Melbourne 
35: Parkville, Victoria 3052, Australia}
36: 
37: \begin{flushright}
38: \vspace{-1.5in}
39: {LA-UR-04-1736}\\
40: \vspace{-0.1in}
41: {hep-ph/0404015}\\
42: \vspace*{0.2in}
43: \end{flushright}
44: 
45: \begin{abstract}
46: 
47: If the sterile neutrino mass matrix in an otherwise conventional 
48: seesaw model has a rank less than the number of flavors, it is 
49: possible to produce pseudo-Dirac neutrinos. In a two-flavor, 
50: sterile rank 1 case, we demonstrate analytic conditions for large 
51: active mixing induced by the existence of (and coupling to) the 
52: sterile neutrino components. For the three-flavor, rank 1 case, 
53: ``3+2'' scenarios with large mixing also devolve naturally as we 
54: show by numerical examples.  We observe that, in this approach, 
55: small mass differences can develop naturally without any requirement 
56: that masses themselves are small.  Additionally, we show that 
57: significant three channel mixing and limited experimental resolution 
58: can combine to produce extracted two channel mixing parameters at 
59: variance with the actual values. 
60: 
61: \end{abstract}
62: 
63: \pacs{14.60.Pq, 14.60.St, 14.60.Lm, 23.40.Bw}
64: \maketitle
65: 
66: \section{Introduction}
67: 
68: Conventional wisdom holds that neutrinos ought to be Majorana 
69: particles with very small masses, due to the action of a 
70: ``seesaw'' mechanism\cite{see-saw}, which is built on the 
71: concept of quark-lepton symmetry\cite{GSMcK}.  Alternatively, 
72: there have been theoretical suggestions regarding pseudo-Dirac 
73: neutrinos in the past\cite{wolfm2,oldref}, and again more 
74: recently\cite{GSMcK,CHM,othr}, i.e., that neutrinos may well be 
75: Majorana particles occuring in nearly degenerate pairs. These 
76: can occur within the framework of the so-called ``singular'' 
77: see-saw where the rank of the mass matrix describing the 
78: (presumed to be) heavy neutrinos, which have no interactions 
79: (often referred to as ``sterile'' neutrinos) in the standard 
80: model~(SM), is less than maximal.
81: 
82: Recent results from Kamiokande\cite{superK} on atmospheric 
83: neutrinos, from Sudbury\cite{SNO} on solar neutrinos, and from 
84: KamLand\cite{KamL} on long baseline reactor neutrinos, appear 
85: to require oscillations between nearly maximally mixed (active 
86: neutrino) mass eigenstates.  Each of these analyses, however, 
87: argues that this mixing cannot be dominantly to sterile states 
88: such as are found in pseudo-Dirac pairs. On the other hand, the 
89: concatenation of the data from these experiments with that from 
90: LSND\cite{LSND} and other short baseline data does not appear 
91: to fit into a theoretical structure which only includes mixing 
92: among three active Majorana neutrinos.  Many have therefore 
93: been motivated to consider the effects of additional (sterile, 
94: Majorana) neutrino states, the existence of which is accepted 
95: in the conventional ``see-saw'' extension of the SM, although 
96: there the actual states are generally precluded from appearing 
97: directly in experiments by an assumption that the masses of the 
98: sterile states are very large. 
99: 
100: We investigate here how small flavor mixing effects in the 
101: sterile sector can lead to large mixing among active neutrinos 
102: in the presence of a singular see-saw. (In Ref.\cite{CHM}, 
103: large mixing  was achieved by means of a mass hierarchy in 
104: the Dirac mass sector.) Paralleling a convention in the quark 
105: sector, we assume the mass and flavor bases for the charged 
106: leptons are simultaneously diagonal, so that all flavor violations 
107: and oscillation phenomena are described as arising from the 
108: neutrino mixing angles alone. 
109: 
110: It should be noted that there is no accepted principle 
111: that specifies the flavor space structure of the mass 
112: matrix assumed for the sterile sector. Some early 
113: discussions\cite{see-saw,wolfm2} implicitly assume that 
114: a mass term in the sterile sector should be proportional 
115: to the unit matrix.  This has the pleasant prospect, 
116: in terms of the initial argument for the see-saw, that 
117: all active neutrino flavors have small masses on the 
118: scale of other fermions.  However, since there is no 
119: obvious requirement that Dirac masses in the neutral 
120: lepton sector are the same as Dirac masses in any 
121: other fermionic sector, this result is not compelling.  
122: Indeed, Goldhaber has argued for a view of family 
123: structure and self-energy based masses that naturally 
124: produces small neutrino masses\cite{Maurice}. We 
125: discuss here a more conventional possibility which 
126: arises from a minimal modification of the standard see-saw, 
127: namely that the rank of the mass matrix for the sterile 
128: sector is less than the number of flavors. Note that 
129: this does not conflict with quark-lepton symmetry which 
130: applies only to the number and character of states. 
131: 
132: In this paper, which is an extension of reference\cite{hep}, 
133: we shall concentrate on the case of a rank $1$ sterile matrix, 
134: relegating the rank $2$ case to some remarks at the end. (The 
135: analysis of short baseline data by Sorel, Conrad and 
136: Shaevitz\cite{Sorel} suggests that the rank $2$ case may not 
137: actually occur in Nature.) We note in passing that some Grand 
138: Unified Theories include more than $3$ fermions that are 
139: neutral under all of the interactions in the SM; a $4\times4$ 
140: or larger, rank $1$ sterile mass matrix could lead to $3$ 
141: pseudo-Dirac pairs of neutrinos involving all of the active 
142: neutrinos of the SM. 
143: 
144: Concentrating on a $3$-dimensional sterile space, we consider 
145: rank $1$ to be a natural case because whatever spontaneous 
146: symmetry breaking produces mass in that flavor space necessarily 
147: defines a specific direction. Before including the effects of the 
148: sterile mass, we assume three non-degenerate Dirac neutrinos, with 
149: Dirac masses, $m_1~<~m_2~<~m_3$, (although this is not essential,) 
150: which are each constructed from one Weyl spinor which is active 
151: under the $SU(2)_W$ of the SM and one Weyl spinor which is sterile 
152: under that interaction. (Being neutrinos, both Weyl fields have no 
153: interactions under the $SU(3)_C$ or the $U(1)$ of the SM.) There 
154: is then an MNS matrix\cite{MNS} which relates these Dirac mass 
155: eigenstates to the flavor eigenstates in a manner completely 
156: parallel to that of the CKM matrix\cite{ckm} for quarks.  Note, 
157: however, that these matrix elements are not the ones extracted 
158: directly from experiment, as the mass matrix in the sterile sector 
159: induces additional mixing.
160: 
161: We next use the Dirac mass ($m_{D}$) eigenstates to define 
162: bases in both the $3$-dimensional active flavor space and the 
163: $3$-dimensional sterile flavor space\cite{fn1}. Following the 
164: spirit of the original see-saw, we exclude any initial Majorana 
165: mass term in the active space. If the Majorana mass matrix in 
166: the sterile space were to vanish also, the three flavors of 
167: Dirac neutrinos would be a mixture of (Dirac) mass eigenstates 
168: in a structure entirely parallel to that of the quarks. 
169: 
170: A rank $1$ sterile mass matrix may be represented as a vector 
171: of length $M$ oriented in some direction in the $3$-dimensional 
172: sterile space. If that vector lies along one of the axes, then 
173: the Dirac neutrino that would have been formed from it and its 
174: active neutrino partner partake of the usual see-saw 
175: structure (one nearly sterile Majorana neutrino with mass 
176: approximately $M$ and one nearly purely active neutrino with 
177: mass approximately $m_D^2/M$) and the other two mass eigenstates 
178: remain Dirac neutrinos. If that vector lies in a plane 
179: perpendicular to one axis, the eigenstate associated with that 
180: axis will remain a pure Dirac neutrino, and the other two pairs 
181: of states form one pseudo-Dirac pair and a pair displaying the 
182: usual see-saw structure. Both of these pairs are mixtures of the 
183: $4$ Weyl fields associated with the two mixing Dirac neutrinos. 
184: In general, the structure consists of $2$ pseudo-Dirac pairs and 
185: one see-saw pair, all mixed.
186: 
187: As we implied above, the very large mixing required by the 
188: atmospheric neutrino measurements could have been taken to be 
189: evidence for a scheme involving pseudo-Dirac neutrinos. (This, 
190: after all, follows Pontecorvo's initial suggestion\cite{BP}.) 
191: However, pure mixing into the sterile sector is now strongly 
192: disfavored\cite{nsm}.  It is evident from the discussion above 
193: that there is a region of parameter space (directions of the 
194: vector) in which the two pseudo-Dirac pairs are very nearly
195: degenerate, giving rise to the possibility of strong mixing in 
196: the active sector coupled with strong mixing into the sterile 
197: sector.  We explore this point here. 
198: 
199: The organization of the remainder of the paper is as follows: 
200: In the next section we discuss a two flavor, $4\times 4$ neutrino 
201: mass matrix analytically. In Sec.\ref{genmass}, we present 
202: the general $6\times 6$ mass matrix and discuss the parameterization 
203: of the sterile mass matrix and various limiting cases.  We show 
204: the spectrum for a general case.  In Sec.\ref{2pds}, we apply 
205: our analysis to the case where the plane in question is 
206: perpendicular to the axis for the middle value ($m_2$) Dirac 
207: mass eigenstate, raising the possibility of near degeneracy  
208: between pseudo-Dirac pairs. Moving away from that plane produces 
209: large  mixing amongst the members of those pseudo-Dirac pairs. 
210: In Sec.\ref{example}, we show an example of the oscillation 
211: patterns that are produced and how limited experimental resolution 
212: can lead to errors in the extraction of physical parameters if 
213: the data analysis assumes only two channel mixing.  Finally, we 
214: remark on the structures expected for a rank $2$ sterile matrix 
215: and then reiterate our conclusions.
216: 
217: \section{Two flavor case}\label{2flavor}
218: 
219: In our examination of the consequences of assuming a rank $1$ 
220: mass matrix in the sterile subspace, we will show below that 
221: there are certain parameter ranges for which there is very 
222: large mixing induced in the active subspace, even though there 
223: is no explicit mixing among the original Dirac bispinors.  To 
224: see how this arises, it is useful to look at the two flavor 
225: model for which we can obtain an analytic description of the 
226: mass eigenvalues as a power series in $\frac{1}{M}$.  We then 
227: can find the eigenfunctions, again as a power series in
228: $\frac{1}{M}$, and look at the ratio of the coefficients for 
229: the two active components.  We examine the conditions which 
230: allow for large active mixing when there is no mixing in the 
231: original Dirac space.
232: 
233: In the next section we shall discuss the case where two 
234: pseudo-Dirac pairs are nearly degenerate and follow 
235: the mixing patterns as we move away from that region of 
236: parameter space.  To facilitate that discussion, we explore 
237: this subsystem where analytic approximations are available, 
238: i.e., the limit where one Dirac mass eigenstate remains 
239: uncoupled from all of the other states.  Anticipating the 
240: following section, we decouple what is there $m_2$.  That 
241: is, we examine a two flavor system in which the Dirac mass 
242: eigenvalues are $m_1$ and $m_3$. 
243: 
244: It is useful to define:
245: \Bea
246: m_0^2  & = & m_1^2 \cos^2 \theta + m_3^2 \sin^2 \theta \label{1} \\
247: a & = &\frac{ \left(m_1^2 - m_3^2\right)\sin\theta \cos
248: \theta}{m_0\sqrt{2}} \label{2} \\
249: b & = & \frac{m_1m_3}{m_0} \label{3}
250: \Eea
251: and $c = \cos\theta$, $s = \sin\theta$.  Note the additional 
252: $1/\sqrt{2}$ factor in $a$.  These refer to the mass matrix
253: \Be
254: {\cal{M}}_{1} =
255: \left(\begin{array}{cccc}
256: 0 & 0 & m_{1} & 0 \\
257: 0 & 0 & 0 & m_{3} \\
258: m_{1} & 0 & Ms^{2} & Mcs \\
259: 0 & m_{3} & Mcs & Mc^{2}
260: \end{array}
261: \right) 	\label{mass4}
262: \Ee
263: where $m_1, m_3$ are Dirac masses for the two neutrino flavors 
264: and $M$ is the single nonzero mass eigenvalue in the sterile 
265: sector. The angle $\theta$ defines the deviation of the direction 
266: in the sterile subspace of the eigenvector for this nonzero mass 
267: from one of the flavor axes defined by the Dirac mass eigenstates. 
268: Note that the structure in Eq.(\ref{mass4}) is equivalent to the 
269: assumption that the MNS\cite{MNS} analog of the CKM\cite{ckm} 
270: matrix for the quarks is the unit matrix. 
271: 
272: It is useful to transform ${\cal{M}}_{1}$ into the form
273: \Be
274: \cal{M} =
275: \left(\begin{array}{cccc}
276: m_0 & 0 & 0 & a \\
277: 0 & -m_0 & 0 & -a \\
278: 0 & 0 & 0 & b \\
279: a & -a & b & M
280: \end{array}
281: \right)
282: \Ee
283: in order to see that, to lowest order, the three small eigenvalues 
284: are $\pm m_0, 0$.  (Note the minus sign on the $a$ in the (2,4) and
285: (4,2) positions.) The matrix effecting the transformation ${\cal{M}} 
286: = \Omega^{\dag} {\cal{M}}_{1} \Omega$ is
287: \Be
288: \Omega =
289: m_{0}^{-1}\left(\begin{array}{cccc}
290: m_{1}s/\sqrt{2} & -m_{1}s/\sqrt{2} & m_{3}c & 0 \\
291: -m_{3}c/\sqrt{2} & m_{3}c/\sqrt{2} & m_{1}s & 0 \\
292: m_{0}s/\sqrt{2} & m_{0}s/\sqrt{2} & 0 & m_{0}c \\
293: -m_{0}c/\sqrt{2} & - m_{0}c/\sqrt{2} & 0 & m_{0}c
294: \end{array} \right)
295: \Ee
296: 
297: This suggests writing the characteristic equation as:
298: \Be
299: \mu \left(m_{0}^{2} - \mu^{2}\right)\mu(M - \mu) =
300:   2 \mu^{2} a^{2} - \left(m_{0}^{2} - \mu^{2}\right) b^{2}
301: \Ee
302: which is convenient for iterative solution in a series in $M^{-1}$.
303: The usual equation obtained directly from $\left| {\cal{M}}_{1} - 
304: \mu \right. ${\mtxfnt 1}$\left. \right| = 0$,
305: \Be
306: \mu^{4} - \mu^{3}M - \mu^{2}\left(m_{1}^{2} + m_{3}^{2}\right) + \mu
307: m_{0}^{2} M + m_{1}^{2} m_{3}^{2} = 0,
308: \Ee
309: is just the same equation.
310: 
311: The solutions to order $M^{-2}$ are 
312: \Bea
313: \mu_1 & = & m_0 - \frac{a^{2}}{M} -
314: \frac{a^{2}}{m_{0}M^{2}}\left(m_{0}^{2} - \frac{a^{2}}{2} -
315: b^{2}\right) \label{4solns1}  \\
316: \mu_2 & = & -m_0 - \frac{a^{2}}{M} +
317: \frac{a^{2}}{m_{0}M^{2}}\left(m_{0}^{2} - \frac{a^{2}}{2} -
318: b^{2}\right) \label{4solns2}  \\
319: \mu_3 & = & - \frac{b^2}{M} + {\cal O}(M^{-3}) \label{4solns3}  \\
320: \mu_4 & = & M +\frac{b^2}{M} +2 \frac{a^2}{M} +{\cal O}(M^{-3}) 
321: \Eea 
322: 
323: Notice that the eigenvalues sum to $M$ as they must and that 
324: the $\pm m_{0}$ eigenvalues are shifted in opposite directions 
325: at $O(M^{-2})$ but in the same direction at $O(M^{-1})$, which 
326: is a small amount for sufficiently large $M$. The latter shift 
327: is why these form a pseudo-Dirac pair rather than simply a 
328: Dirac bispinor.  Note also that $\mu_{3}$ and $\mu_{4}$, do 
329: not acquire $O(M^{-2})$ corrections; their next correction 
330: is at the next higher order.
331: 
332: Having obtained the eigenvalues, we now solve for the eigenvectors.
333: Since our interest is in the mixing in the active sector, it is useful
334: to carry this exercise out in the original representation, that of
335: ${\cal{M}}_{1}$.  In this representation, we define the $i^{th}$ 
336: eigenvector as
337: \Be
338: \phi_i = \left( \begin{array}{c} \alpha_i \\ \beta_i  \\ \gamma_i 
339: \\ \delta_i \end{array} \right),
340: \Ee
341: where $\alpha_i$ and $\beta_i$ are the two active components and
342: $\gamma_i$ and $\delta_i$ are the two sterile components.
343: 
344: Picking three equations, we find
345: \Bea
346: -\mu_i \alpha_i + m_1 \gamma_i & = & 0  \nonumber \\
347: -\mu_i \beta_i + m_3 \delta_i & = & 0   \nonumber  \\
348: m_3 \beta_i  + M s c \gamma_i +(M c^2 - \mu_i) \delta_i & = & 0 \label{3eqs}
349: \Eea
350: A number of points are immediately clear from Eqs.(\ref{3eqs}): 
351: Since $\mu_{4} \sim M$, $\beta_{4}$ and $\alpha_{4}$ are small 
352: (${\cal O}(m_{D}/M)$) so the fourth eigenstate is almost entirely 
353: decoupled from the active sector.  Conversely, since $\mu_{3} 
354: \sim {\cal O}(m_{D}^2/M)$, $\gamma_{3}$ and $\delta_{3}$ are 
355: small (${\cal O}(m_{D}/M)$) so the third eigenstate resides almost 
356: entirely in the active sector. Finally, since $\mu_{1}$ and $\mu_{2}$ 
357: are of order ${\cal O}(m_{D})$, $\gamma_{1,2}$ and $\delta_{1,2}$ 
358: are comparable with $\beta_{1,2}$ and $\alpha_{1,2}$ so these two 
359: eigenstates are generally strongly mixed between the active and 
360: sterile sectors, i.e., they form a pseudo-Dirac pair. 
361: 
362: Substituting for $\gamma_i$ and $\delta_i$ gives an equation for the 
363: ratio
364: \Bea
365: \frac{\beta_i}{\alpha_i} & = & -\frac{M \mu_i s c}{[\mu_i (M c^2 -
366: \mu_i) + m_3^2]}  \nonumber \\  & = & -\frac{s c}{[c^2 -
367: \frac{\mu_i}{M} + \frac{m_3^2}{M \mu_i}]}. \label{ratio}
368: \Eea
369: Note that if either $s = 0$ or $c = 0$, one pair of states forms 
370: a purely Dirac bispinor and the other becomes the usual see-saw 
371: pair of Majorana states.
372: 
373: For the light mass eigenstates, the ratio $\beta_{i}/\alpha_{i}$ 
374: is a measure of mixing in the active sector. Solving Eq.(\ref{ratio}), 
375: we find that
376: \Be
377: | \frac{\beta_{1}}{\alpha_{1}} | \,\,\, =  
378: | \frac{\beta_{2}}{\alpha_{2}} | \,\,\, =
379: | \frac{\alpha_{3}}{\beta_{3}} | \,\,\, = 
380: \frac{m_{3}}{m_{1}} {\rm tan}(\theta), \label{lrgmx}
381: \Ee
382: where the last equality is correct to ${\cal O}(m_{D}^2/M^2)$ 
383: and the first two are correct to ${\cal O}(m_{D}/M)$. It is 
384: apparent that, in all three states, the mixing of the active 
385: components can be large simultaneously. 
386: 
387: Turning back to the amplitudes of the sterile components, 
388: we see from the first two lines of Eq.(\ref{3eqs}) that 
389: \Bea
390: \frac{\gamma_{i}}{\alpha_{i}} & = & \frac{\mu_{i}}{m_{1}} \NL
391: \frac{\delta_{i}}{\beta_{i}} & = & \frac{\mu_{i}}{m_{3}} .
392: \Eea	 
393: Hence, for a large range of values of ($m_{1}$, $m_{3}$, 
394: $\theta$), these ratios are ${\cal O}(1)$ for $i=1$ and 
395: $2$, which is, of course, characteristic of a pseudo-Dirac 
396: pair. As long as $M$ is large, $\gamma$, $\delta$ are small 
397: for $i=3$ since $\mu_{3} \sim 0$, and huge for $i=4$ since 
398: $\mu_{4} \sim M$. This reiterates the fact that the massive 
399: sterile state is quite decoupled, while the light sterile 
400: can be strongly coupled into the active states and the 
401: pseudo-Dirac states significantly mixed across all four 
402: components. 
403: 
404: %\subsection{Two Flavor Conclusion}
405: 
406: Thus, in this simple, two flavor model, we have demonstrated 
407: that a misalignment of the direction vector for the heavy 
408: sterile mass with the axes determined by the Dirac mass 
409: eigenstates necessarily induces mixing in the active sector 
410: for all of the light Majorana mass eigenstates, even with a
411: unit MNS matrix for the Dirac mass matrix. This point has 
412: been raised previously in Ref.\cite{rabi} in a different context. 
413: 
414: Moreover, large mixing of active states results over a region 
415: of the $(m_1, m_3, \theta)$ parameter space where the mass ratio 
416: and the deviation angle of the sterile components from flavor 
417: alignment approximately compensate, i.e., near the line determined 
418: by setting the rightmost quantity in Eq.(\ref{lrgmx}) to one. 
419: Mixing in the Dirac sector by an MNS matrix should not alter the 
420: general feature of achieving large mixing "naturally". 
421: 
422: Finally, we note explicitly the difference in oscillation structure 
423: between this $4\times4$ neutrino mass matrix and the $2\times2$ 
424: Majorana (or Dirac) mixing usually applied to interpret experiments. 
425: As shown by Eqs.(\ref{3}) or Eqs.(\ref{4solns1}, \ref{4solns2}, 
426: \ref{4solns3}), instead of one mass difference, here 
427: 
428: \pagebreak 
429: 
430: \noindent there are at least two 
431: independent mass differences, even in the limit of large sterile 
432: mass ($M$). Thus, simple two channel analyses are not guaranteed 
433: to extract the true physical oscillation parameters from experimental 
434: results. This problem is exacerbated in the $6\times6$ case that 
435: we discuss in the next Section, in which at least four independent 
436: mass differences appear where it has been conventionally assumed that 
437: there can only be two. 
438: 
439: \section{General mass matrix}\label{genmass}
440: 
441: The flavor basis for the active neutrinos and the pairing to 
442: sterile components defined by the (generally not diagonal) 
443: Dirac mass matrix could be used to specify the basis for the 
444: sterile neutrino mass matrix, $M_S$. Instead we take the basis 
445: in the $3\times 3$ sterile subspace to allow the convention 
446: described below. This implies a corresponding transformation 
447: of the Dirac mass matrix, which is irrelevant at present 
448: since the entries in that matrix are totally unknown. 
449: 
450: We define our convention for the choice of axes in the $3\times 3$ 
451: sterile subspace as follows. Denote the nonzero mass eigenvalue 
452: of the rank $1$ by $M$ and choose its eigenvector initially 
453: in the third direction. Then rotate this vector, first by 
454: an angle of $\theta$ in the $1-3$ plane and then by $\phi$ 
455: in the $1-2$ plane. The rotation is chosen so that the 
456: Dirac mass matrix which couples the active and sterile 
457: neutrinos becomes diagonal, i.e., the basis is defined by 
458: Dirac eigenstates. This produces a $3 \times 3$ mass matrix 
459: in the sterile sector denoted by
460: \Be
461: M_S = M \left[ 
462: \begin{array}{ccc}
463: \cos^2 \phi \sin^2 \theta  &  \cos \phi \sin \phi \sin^2 \theta  &
464: \cos \phi \sin \theta \cos \theta  \\  \cos \phi \sin \phi \sin^2 \theta & 
465: \sin^2 \phi \sin^2 \theta  &  \sin \phi \sin \theta \cos \theta  \\ 
466: \cos \phi \sin \theta \cos \theta & \sin \phi \sin \theta \cos \theta & 
467: \cos^2 \theta   \end{array}  \right]. 
468: \Ee
469: 
470: In this representation, the Dirac mass matrix is diagonal 
471: by construction
472: \Be
473: m_D = \left[ \begin{array}{ccc}
474: m_1   &  0  &  0  \\  0  &  m_2  &  0  \\  0  &  0  &  m_3
475: \end{array} \right].  
476: \Ee
477: 
478: Note that there are special cases.  For $\theta = 0$ and any 
479: value for $\phi$,
480: \Be
481: M_S = \left[ \begin{array}{ccc}
482: 0 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & M \end{array} \right]  .
483: \Ee
484: For $\theta = \pi / 2$ and $\phi = 0$, 
485: \Be
486: M_S = \left[ \begin{array}{ccc} 
487: M & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{array} \right], 
488: \Ee
489: and, for $\theta = \pi / 2$ and $\phi = \pi / 2$, 
490: \Be
491: M_S = \left[ \begin{array}{ccc} 0 & 0 & 0 \\ 0 & M & 0 \\ 0 & 0 & 0
492: \end{array} \right]. 
493: \Ee
494: These are equivalent under interchanges of the definition 
495: of the third axis. 
496: 
497: The $6 \times 6$ submatrix\cite{fn2} of the full $12 \times 
498: 12$ is, in block form, 
499: \Be
500: {\cal M} = \left[ \begin{array}{cc} 0 & m_D \\ m_D & 
501: M_S \end{array} \right].   
502: \Ee
503: Since we are ignoring CP violation here, no adjoints or 
504: complex conjugations of the mass matrices appear. 
505: 
506: Note that, in the chiral representation, the full $12 \times 12$ matrix is
507: \Be
508: \left[ \begin{array}{cc} 0 & {\cal M} \\ 
509: {\cal M} & 0 \end{array} \right]. \nonumber
510: \Ee
511: Thus the full set of eigenvalues will be $\pm$ the eigenvalues of
512: ${\cal M}$.  Where it matters for some analysis we keep track of 
513: the signs of the eigenvalues, however for most results we present 
514: positive mass eigenvalues.
515: 
516: After some algebra, we obtain the secular equation 
517: \Bea
518: 0 & = & \lambda^6 -M \lambda^5 -(m_1^2 + m_2^2 + m_3^2) \lambda^4 \nonumber \\
519: & & + M [m_3^2 \sin^2 \theta + m_2^2 (\sin^2 \theta \cos^2 \phi + \cos^2 \theta)] 
520: \lambda^3 \nonumber \\ & & + (m_1^2 m_2^2 + m_2^2 m_3^2 + m_3^2 m_1^2) 
521: \lambda^2 \\
522: & & - M (m_1^2 m_2^2 \cos^2 \theta + m_2^2 m_3^2 \cos^2 \phi \sin^2 \theta 
523: \nonumber \\
524:   &  & + m_3^2 m_1^2 \sin^2 \phi \sin^2 \theta) \lambda \nonumber \\
525:   &  & - m_1^2 m_2^2 m_3^2. \nonumber
526: \Eea  
527: 
528: \pagebreak
529: 
530: This may be rewritten as 
531: \Bea
532:  0 & = & (\lambda^2 - m_1^2) (\lambda^2 - m_2^2) 
533: (\lambda^2 - m_3^2) \nonumber \\ &   &    
534: - \lambda M \left( \lambda^4 -\left[ m_3^2 \sin^2 \theta 
535: + m_2^2 (\sin^2 \theta \cos^2 \phi + \cos^2 \theta ) \right. \right. \nonumber \\
536: &  & \left. \left. + m_1^2 ( \sin^2 \theta \sin^2 \phi + \cos^2 \theta) 
537: \right] \lambda^2 \right. \\
538: &  & \left. +m_1^2 m_2^2 \cos^2 \theta + m_2^2 m_3^2 \sin^2 \theta 
539: cos^2 \phi  \right. \nonumber \\
540: &  & \left. + m_3^2 m_1^2 \sin^2 \theta \sin^2 \phi \right). \nonumber
541: \Eea
542: 
543: The special cases follow directly.  For $\theta = 0$, we find
544: \Be
545: (\lambda^2 - m_1^2) (\lambda^2 - m_2^2)
546: (\lambda^2 - M \lambda - m_3^2) = 0, 
547: \Ee
548: for $\theta = \pi / 2$ and $ \phi = 0$
549: \Be
550: (\lambda^2 - m_2^2) (\lambda^2 - m_3^2)
551: (\lambda^2 - M \lambda - m_1^2) = 0, 
552: \Ee
553: and for $\theta = \pi / 2$ and $\phi = \pi / 2$
554: \Be
555: (\lambda^2 - m_3^2) (\lambda^2 - m_1^2)
556: (\lambda^2 - M \lambda - m_2^2) = 0. 
557: \Ee
558: 
559: If $m_1^2 = m_2^2 = m_3^2 = m^2$, then we find
560: \Be
561: (\lambda^2 - m^2)^2 (\lambda^2 - M \lambda - m^2) = 0. 
562: \Ee
563: 
564: Due to the wide range of possibilities inherent in the system, 
565: it is useful to examine specific numerical examples.  For the
566: immediate exercise, we have picked the following parameters:
567: $m_1 = 1$, $m_2 = 2$, $m_3 = 3$ and $M = 1000$. The relatively 
568: small value of $M$ is chosen so that the splittings are not so 
569: tiny as to be difficult to discern. 
570: 
571: For this choice, the eigenvalues have a definite pattern 
572: for all values of $\theta$ and $\phi$.  There are two very 
573: close pairs, with mass eigenvalues between $1$ and $3$.  
574: There is one very small eigenvalue, of order $10^{-3}$ 
575: reflecting the ratio of $m_D$ to $M$, and one large eigenvalue 
576: of order $10^{3}$ (i.e., of order $M$).  Treating the last 
577: two as a pair despite their disparity in mass allows us to 
578: present results in tabular form, one for each pair, for sets 
579: of angles $\theta , \phi = \pi / 8, \pi / 4, 3 \pi / 8 $.
580: 
581: First, for the lower mass close pair, we have
582: \Be
583: \begin{array}{lccc}
584: \theta \backslash \phi & \pi /8  &  \pi /4  & 3 \pi / 8  \\
585:              &         &          &            \\
586: \pi / 8      & 1.398125 & 1.230175 & 1.068477   \\
587:              & 1.394934 & 1.228025 & 1.067688    \\
588:              &          &          &            \\
589: \pi / 4      & 1.809478 & 1.478863 & 1.151936   \\
590:              & 1.808183 & 1.477134 & 1.150941   \\
591:              &          &          &             \\
592: 3 \pi / 8    & 1.877166 & 1.562977 & 1.18999    \\
593:              & 1.876742 & 1.561911 & 1.189146  \end{array}
594: \Ee
595: 
596: Then, for the next mass pair with close eigenvalues, 
597: we find
598: \Be
599: \begin{array}{lccc}
600: \theta \backslash \phi & \pi / 8  &  \pi / 4 & 3 \pi / 8   \\
601:              &          &          &             \\
602: \pi / 8      & 2.038992 & 2.107688 & 2.158044    \\
603:              & 2.038729 & 2.107156 & 2.157407    \\
604:              &          &          &              \\
605: \pi / 4      & 2.347974 & 2.46348  & 2.529128    \\
606:              & 2.346047 & 2.462176 & 2.52809     \\
607:              &          &          &              \\
608: 3 \pi / 8    & 2.816525 & 2.847539 & 2.868607    \\
609:              & 2.815691 & 2.846972 & 2.868186   
610: \end{array} 
611: \Ee
612: 
613: \pagebreak 
614: 
615: Finally, even though it does not directly impact the 
616: argument, we display the remaining pair in order to 
617: present a complete set.  
618: \Be
619: \begin{array}{lccc}
620: \theta \backslash \phi & \pi / 8  &  \pi / 4 & 3 \pi / 8   \\
621:              &          &          &             \\
622: \pi / 8      & 1000.008 & 1000.008 & 1000.008    \\
623:              & 0.00444  & 0.005366 & 0.006778    \\
624:              &          &          &              \\
625: \pi / 4      & 1000.005 & 1000.006 & 1000.006    \\
626:              & 0.001997 & 0.002717 & 0.004248    \\
627:              &          &          &              \\
628: 3 \pi / 8    & 1000.003 & 1000.003 & 1000.004    \\
629:              & 0.001289 & 0.001819 & 0.003092   
630: \end{array} 
631: \Ee
632: 
633: \section{Two nearly degenerate pseudo-Dirac pairs}\label{2pds}
634: 
635: Applying the techniques of the last section, we find the angle 
636: $\theta$ such that $m_2$ and the eigenvalue for the pseudo-Dirac 
637: pair above, $m_0$, are approximately degenerate.  We then vary  
638: $\phi$ away from $0$ and display the eigenfunctions.  To 
639: illustrate the general nature of the result, we have changed 
640: the Dirac masses from the even spacing used above.
641: 
642: In Table I, the Dirac masses are taken to be $m_1 =  1$, $m_2 
643: =  1.1$, and  $m_3 =  3$.  The value for $m_2$ has been changed 
644: from above so that we can demonstrate that small angles in the 
645: sterile sector can lead to large mixing in the active sector.  
646: Again, in order to display the structure of the spectrum, we have 
647: chosen $M = 1000$, rather than a larger value, expected to be more 
648: realistic, but which would suppress the difference scale between 
649: the pairs. The angles are given in degrees.
650: 
651: Table I represents only a small part of the available parameter 
652: space; the values of the angles are chosen to display some 
653: interesting possible features. First, $\theta$ has been chosen so 
654: that, at $\phi = 0$, the Dirac pair at $m_2$ is bracketed by the 
655: pseudo-Dirac pair.  Such a value of $\theta$ exists for any pattern 
656: of the Dirac masses.  Then, for small values of $\phi$, there are 
657: always two nearly degenerate pseudo-Dirac pairs.
658: 
659: Note that, for $\phi = 0$, there is no mixing between the 
660: field labelled by $2$ and the remaining fields, while for 
661: the next entry at $\phi = 2.25$ degrees there is considerable 
662: mixing.  That mixing increases with $\phi$ as the difference 
663: bewteen the eigenvalues increases.  The pattern described by 
664: the centroids of the pseudo-Dirac pairs is fixed by the angles 
665: $\theta$ and $\phi$.  If $M$ is increased, that pattern hardly  
666: changes.  The primary effect of increasing $M$, consistent 
667: with the analysis in Sec.\ref{2flavor}, is to decrease the 
668: separation of the two members of each pseudo-Dirac pair while 
669: producing the usual see-saw behavior for the remaining pair. 
670: Thus, tiny differences in mass between masses that are not 
671: especially small themselves, are, in the usual sense of the 
672: term, natural in this approach. 
673: 
674: The implication for oscillation phenomena is clear.  A given 
675: weak interaction produces an active flavor eigenstate which 
676: is some linear combination of the three active components 
677: listed in Table I.  That then translates into a linear 
678: combination of the six mass eigenstates.  From Table I, it 
679: is clear that the involvement of the heavy Majorana see-saw 
680: state is minimal, so the system effectively consists of the 
681: light Majorana see-saw state and the four Majorana states 
682: arising from the two pseudo-Dirac pairs. These five states 
683: include all three active neutrinos, generating a natural 
684: 3+2 scenario. 
685: 
686: Since these five mass eigenstates have both active and 
687: sterile components, the subsequent time evolution will 
688: involve both flavor changing oscillations and oscillation 
689: into (and back out of) the sterile sector. This can lead 
690: to very complex oscillation patterns, as there are $10$ 
691: mass differences, $4$ of which are independent. A specific 
692: example is discussed in the next section. 
693: 
694: Finally, inspection of the column labelled ``1active'' for 
695: $\phi = 2.25$ or $\phi = 4.5$, for example, shows that 
696: the presence of a rank $1$ sterile mass matrix can 
697: seriously change any mixing pattern of the MNS type\cite{MNS}, 
698: from that which would have obtained with purely Dirac neutrinos.
699: 
700: \section{example}\label{example}
701: 
702: In Figs.1 to 3, we plot the oscillation patterns that appear for 
703: the parameters set by the second entry in Table I.  Fig.1 gives 
704: an overview of the case where an active neutrino (labelled $1$) 
705: is produced initially. The plot is given versus $L/E$, where $L$ 
706: is the distance from the neutrino source and $E$ is the energy 
707: (bin) of the neutrino observed. (As shown in the Appendix, $L/p$,  
708: where $p$ is the momentum of the neutrino, might well be the 
709: more correct variable to use, but the difference is certainly 
710: irrelevant in all conceivable neutrino experiments.) 
711: 
712: The (compressed scale) Fig.1 shows rapid oscillations between 
713: active neutrinos $1$ and $3$ with a later appearance of the 
714: active neutrino $2$. Note the large mixing among all three 
715: channels of active neutrinos. The mixing to sterile neutrinos 
716: is large also, but occurs on a much larger $L/E$ scale, 
717: corresponding to the much smaller mass difference (approximate 
718: degeneracy) of the pseudo-Dirac pairs. 
719: 
720: Fig.2 extracts from Fig.1 the appearance of active neutrino 
721: $2$ at small $L/E$ (short baseline experiments). Clearly, 
722: attempting to fit this highly nonsinusoidal behavior with two 
723: channel sinusoidal mixing will generally not yield physical 
724: mixing parameters in good agreement with the actual three 
725: channel case. 
726: 
727: Fig.3 emphasizes how such an error may be magnified by limited 
728: resolution in an experiment. The heavy black curve is the average 
729: over 100 $L/E$ units of the probability for finding the initial 
730: neutrino flavor. It approximates the shape of the {\bf envelope} 
731: of the high frequency oscillations. A two channel analysis would 
732: clearly find a small difference between the squared masses for 
733: the mixing from active neutrino $1$ to active neutrino $3$ 
734: despite the obviously larger value demonstrated by the rapid 
735: oscillation cycles. A similar conclusion follows from the 
736: cycle-averaged curve for appearance of active neutrino $3$. 
737: Finally, one would be tempted to conclude that the mixing to 
738: active neutrino $2$ is small or negligible, when in fact it 
739: is about as large as any other mixing in the full case. 
740: 
741: Labelling active neutrino $1$ as the muon neutrino, active neutrino 
742: $3$ as the tau neutrino and active neutrino $2$ as the electron 
743: neutrino illustrates our concerns about the strong conclusions 
744: drawn from atmospheric and accelerator neutrino experiments by 
745: means of two channel mixing analyses. Similar concerns\cite{tw} 
746: have also been raised in the literature previously. 
747: 
748: \section{Rank $2$}
749: 
750: We have not discussed the case of rank $2$ matrices explicitly,
751: although the pattern is obvious.  In such a case, there would 
752: be two see-saw pairs and one pseudo-Dirac pair, leading to 
753: three active and one sterile light neutrino.  While this pattern 
754: has been analyzed in the literature, we do not find any compelling 
755: pattern for it in the sterile sector.  Furthermore, the current 
756: consistency of all neutrino oscillation data can be accomodated 
757: much more easily (and perhaps only, as indicated by Ref.\cite{Sorel},) 
758: in the rank $1$ case discussed in this paper. Therefore we do  
759: not discuss rank $2$ at this time.
760: 
761: \section{Conclusions}
762: 
763: We have considered here the effects on neutrinos in the SM of the 
764: recurrently successful and conventional constraint of quark-lepton 
765: symmetry, namely, the existence of six independent Weyl spinor 
766: fields of neutrinos, three corresponding to active and three 
767: corresponding to sterile neutrinos. In the now venerable see-saw 
768: approach, the latter three effectively disappear from the excitation 
769: spectrum, leaving small Majorana masses for the active states as a 
770: residuum. We have examined the effect on this system of a rank less 
771: than three character of the $3 \times 3$ mass matrix in the sterile 
772: sector and studied the rank $1$ case, in particular. 
773: 
774: In the rank $1$ case on which we have focused, we find that 
775: the neutrino fields naturally form into two pseudo-Dirac 
776: pairs, leaving only one almost pure Majorana active neutrino 
777: and one conventionally very heavy sterile Majorana neutrino. 
778: More importantly, we also find a naturally strong mixing 
779: between the active and sterile parts of the two pseudo-Dirac 
780: pairs.  Further, we find that this can easily affect the 
781: mixing between active neutrinos even if it is otherwise small. 
782: That is, even if the Dirac mass matrix induced mixing analogous 
783: to what is known to occur in the quark sector is small or absent, 
784: mixing between active neutrinos can develop with large values. 
785: In a two flavor case, we demonstrated analytically that this 
786: strong mixing can develop over a wide range of parameters.
787: 
788: We have chosen a limited relative value of the sterile neutrino 
789: mass scale, $M$, that allows for easy discernment of the nature 
790: of the effects. It should be noted, however, that the primary 
791: effect of increasing $M$ is to decrease the separation of the 
792: two members of each pseudo-Dirac pair while producing the usual 
793: see-saw behavior for the remaining pair. Thus, tiny differences 
794: in mass between masses that are not especially small themselves 
795: are, in the usual sense of the term, natural in this approach. 
796: This is contrary to the general expectation that the small mass 
797: differences responsible for the observed neutrino oscillation 
798: phenomena presage small absolute masses for all of the neutrinos. 
799: Furthermore, increasing the value of $M$ without altering the 
800: Dirac masses retains the features and scales of the oscillations 
801: essentially unchanged; the only significant change is that the 
802: appearance of the sterile components is delayed to even greater 
803: values of $L/E$. 
804: 
805: The features described above are most easily discerned in the 
806: case when the Dirac mass terms for the neutrinos are well separated 
807: in value. It remains conceivable that, if their differences are 
808: small for some other reason, then the splitting between the pseudo-Dirac 
809: pairs may be larger than that between flavors. In this case, it 
810: is still true that large flavor mixing is naturally induced.
811: 
812: Finally, we presented a specific model which raises concerns 
813: about the conclusions drawn from analyses of neutrino oscillation 
814: experiments in terms of two channel mixing: Such analyses may 
815: be misleading regarding the true values of physical parameters. 
816: After the completion of this work, we learned of papers\cite{ernst} 
817: which have independently suggested that such a concern may well be 
818: justified. 
819: 
820: \vspace*{-0.1in}
821: \section{Acknowledgments}
822: \vspace*{-0.1in}
823: This research is supported in part by the Department of Energy under
824: contract W-7405-ENG-36, in part by the National Science Foundation 
825: under NSF Grant \# PHY0099385 and in part by the Australian Research 
826: Council.
827: 
828: \newpage
829: \begin{thebibliography}{99}
830: \bibitem{see-saw} M.\ Gell-Mann, P.\ Ramond and R.\ Slansky, 
831: 	in {\it Supergravity}, Proceedings of the Workshop, 
832: 	Stony Brook, New York, 1979, ed. by P.\ van 
833: 	Nieuwenhuizen and D.\ Freedman (North-Holland,
834: 	Amsterdam, 1979), p. 315; T.\ Yanagida, in {\it 
835: 	Proceedings of the Workshop on the Unified Theories 
836: 	and Baryon Number in the Universe}, Tsukuba, Japan, 
837: 	1979, edited by O.\ Sawada and A.\ Sugamoto (KEK 
838: 	Report No. 79-18, Tsukuba, 1979), p.95; R.N.\ 
839: 	Mohapatra and G.\ Senjanovic, Phys. Rev. Lett.  
840: 	{\bf 44} (1980) 912; S.L.\ Glashow, in {\it Quarks 
841: 	and Leptons}, Cargese (July 9-29, 1979), eds. M.\ 
842: 	Levy {\it et al.} (Plenum, 1980, New York), p. 707.
843: \bibitem{GSMcK} T.\ Goldman, G.J.\ Stephenson Jr.\ and 
844: 	B.H.J.\ McKellar, Mod. Phys. Lett. A{\bf 15} (2000) 
845: 	439; nucl-th/0002053.
846: \bibitem{wolfm2} L. Wolfenstein, Phys. Lett. {\bf B107} 
847: 	(1981) 77; Nucl. Phys. {\bf B186} (1981) 147.
848: \bibitem{oldref} M.\ Kobayashi, C.S.\ Lim and M.M.\ Nojiri, 
849: 	Phys. Rev. Lett. {\bf 67} (1991) 1685.
850: \bibitem{CHM} Y.\ Chikira, N.\ Haba and Y.\ Mimura, Eur. 
851: 	Phys. J. C {\bf 16} (2000) 701; hep-ph/9808254.  
852: \bibitem{othr} E.J.\ Chun, C.W.\ Kim and U.W.\ Lee, Phys. 
853: 	Rev. D~{\bf 58} (1998) 093003; hep-ph/9802209; 
854: 	Kevin Cahill, hep-ph/9912416; hep-ph/9912508. 
855: \bibitem{superK} Y.\ Fukuda {\it et al.}, Phys. Rev. Lett.  
856: 	{\bf 81} (1998) 1562; {\bf 82} (1999) 2644; Phys. 
857: 	Lett. B {\bf 476} (1999) 185; W.W.\ Allison {\it et al.}, 
858: 	Phys. Lett. B {\bf 449} (1999) 137. 
859: \bibitem{SNO} Q.R.\ Ahmad {\it et al.}, Phys. Rev. Lett. 
860: 	{\bf 89} (2002) 011301; {\bf 87} (2001) 071301; S.\ 
861: 	Fukuda {\it et al.} Phys. Lett. B {\bf 539} (2002) 
862: 	179.
863: \bibitem{KamL} K.\ Eguchi {\it et al.}, Phys. Rev. Lett. 
864: 	{\bf 90} (2003) 021802. 
865: \bibitem{LSND} C.\ Athanassopoulos {\it et al.}, Phys. Rev. 
866: 	Lett. {\bf 77} (1996) 3082; A. Aguilar {\it et al.}, 
867: 	Phys. Rev. D{\bf 64} (2001) 112007. 
868: \bibitem{Maurice} M.\ Goldhaber, Proc. Nat. Acad. Sci., USA 
869: 	{\bf 99} (2002) 33; hep-ph/0201208.
870: \bibitem{hep} B.H.J.\ McKellar, G.J.\ Stephenson, Jr., T.\ 
871: 	Goldman and M.\ Garbutt, contributed paper at the 
872: 	XXth Int. Symp. on Lepton and Photon Interactions at 
873: 	High Energies, Rome, July 2001, and the Int. Europhys. 
874: 	Conf. on High Energy Physics, Budapest, July 2001; 
875: 	hep-ph/0106121. For an earlier version of this paper, 
876: 	see also, G.J.\ Stephenson, Jr., T.\ Goldman, B.H.J.\ 
877: 	McKellar and M.\ Garbutt, ``3+2 neutrinos in a see-saw 
878: 	variation'', hep-ph/0307245. 
879: \bibitem{Sorel} M.\ Sorel, J.\ Conrad and M.\ Shaevitz, ``A 
880: 	combined analysis of short-baseline neutrino experiments 
881: 	in the (3+1) and (3+2) sterile neutrino oscillation 
882: 	hypotheses'', hep-ph/0305255.
883: \bibitem{MNS} Z.\ Maki, M.\ Nakagawa and S.\ Sakata, Prog. 
884: 	Theor. Phys. {\bf 28} (1962) 870. 
885: \bibitem{ckm} N. Cabibbo, Phys. Rev. Lett. {\bf 10} (1963) 531; 
886: 	M. Kobayashi and T. Maskawa, Prog. Theor. Phys. {\bf 49} 
887: 	(1973) 652.
888: \bibitem{fn1} {In much of the literature, the sterile space is 
889: 	referred to as ``Right-handed'' (or just R) and the 
890: 	active space as ``Left-handed'' (or just L), which 
891: 	follows from the behavior of the components of a Dirac 
892: 	neutrino where the neutrino is defined as that neutral 
893: 	lepton emitted along with a positively charged lepton. 
894: 	Since we are dealing with mass matrices, which necessarily 
895: 	all couple left-handed representations to right-handed 
896: 	representations, we choose to refer to the Weyl neutrino 
897: 	representations as active and sterile. ``Sterile'' refers 
898: 	only to the SM; these neutrino states may have non-SM 
899: 	interactions.}
900: \bibitem{BP} B.\ Pontecorvo, Zh. Eksp. Teor. Fiz. {\bf 33} 
901: 	(1957) 549; {\bf 34} (1958) 247.
902: \bibitem{nsm} S.\ Fukuda {\it et al.} Phys. Rev. Lett. 
903: 	{\bf 85} (2000) 3999.
904: \bibitem{fn2}{We use the states rather than the field operators 
905: 	to define the mass matrix; see, for example, the 
906: 	discussion in the review article on double beta decay 
907: 	by W.C. Haxton and G.J. Stephenson, Jr., Prog. Part. 
908: 	Nucl. Phys. (Sir Denys Wilkinson, ed.) {\bf 12}, p. 409, 
909: 	Permagon Press, New York, 1984.}
910: \bibitem{rabi} K.S. Babu, B. Dutta and R.N. Mohapatra, Phys. Rev. D 
911: 	{\bf 67} (2003) 076006; hep-ph/0211068.
912: \bibitem{tw} H. P\"{a}s, L. Song and T.J. Weiler, Phys. Rev. D {\bf 67} 
913: 	(2003) 073019; hep-ph/0209373.
914: \bibitem{ernst}{D.C. Lattimer and D.J. Ernst, ``Three-neutrino model 
915: 	analysis of the world's oscillation data'', nucl-th/0310083; 
916: 	See also, D.C. Lattimer and D.J. Ernst, ``Bounds on the angles 
917: 	for the parameterization of three neutrino mixing'', 
918: 	nucl-th/0402019.}
919: \end{thebibliography}
920: 
921: \newpage
922: 
923: \section{APPENDIX: Two Mass Eigenstate Oscillations}
924: 
925: There has been some confusion and discussion in the literature 
926: regarding the space and time dependence of neutrino oscillations. 
927: We briefly present here an argument in the rest frame of a state 
928: of a given flavor that demonstrates the time dependence unequivocally. 
929: By boosting the observer instead of the state, we demonstrate the 
930: equivalence of the usual $L/E_{\nu} \approx L/p_\nu$ dependence, 
931: derived in several ways, to the time dependence in the rest frame. 
932: Hence the figures in the main sections of this paper can be viewed 
933: as either variation with $L/E_{\nu}$, $L/p_\nu$ or time. 
934: 
935: We begin with a flavor eigenstate composed of two different mass 
936: eigenstate contributions, in their common rest frame. Let
937: \Bea
938: c \equiv cos(\theta) & ; & s \equiv sin(\theta) \NL
939: |\nu_{f}> & = & \;\;\, c \, |\nu_1> + s \, |\nu_2> \NL
940: |\nu_{g}> & = & -s \, |\nu_1> + c \, |\nu_2> 
941: \Eea 
942: to fix conventions for neutrino flavors $f$ and $g$ composed 
943: in two channel mixing from neutrino mass eigenstates $\nu_1$ 
944: with mass $m_1$ and $\nu_2$ with mass $m_2$. The time evolution 
945: of the state initially in flavor $f$ in this common rest frame 
946: is given by
947: \Bea
948: & & |\nu(t)> = e^{-\imath {\bf H} t} \, |\nu(0) \equiv \nu_{f}> \NL
949: & = & c \, e^{-\imath m_1 t} |\nu_1> + s \, e^{-\imath m_2 t} |\nu_2> 
950: \Eea 
951: The probability of appearance of ${\nu}_{g}$ from the ${\nu}_{f}$ 
952: source is given by
953: \Bea
954: & & |<\nu_{g}|\nu(t)>|^2  = |e^{-\imath\frac{(m_1+m_2)t}{2}}|^2 \NL 
955: & \times & c^{2}s^{2} \, | -e^{\imath\frac{(m_2-m_1)t}{2}} + 
956: e^{-\imath\frac{(m_2-m_1)t}{2}}|^2 \NL
957: & = & sin^{2}(2\theta) \, sin^{2}(\frac{\Delta m\, t}{2}) 
958: \Eea 
959: Viewed from a frame moving past these states with velocity $\beta$, 
960: the relation between position $L$ in the moving frame and the 
961: time is affected both by the velocity and by time dilatation, i.e., 
962: \Be
963: t = L/[{\beta \gamma}] \sim L m_{{av}}/p_{{\nu}}  , 
964: \Ee
965: where we choose the motion of the frame to be consistent with 
966: the ratio of the average neutrino mass and mean neutrino momentum. 
967: Hence
968: \Be
969: |<\nu_{g}|\nu(t)>|^2 \sim sin(2\theta) \, sin^2({\Delta m^2} 
970: L/[4 p_{{\nu}}])
971: \Ee
972: As usual, the units are determined by the relation 
973: \Be
974: {1.27} \frac{\Delta m^2}{eV^2} \frac{L_{osc}}{km} 
975: \frac{GeV/c}{p_{\nu}} = \pi
976: \Ee
977: consistent with all conventional analyses and expectations. 
978: 
979: It should be noted that different energy (mass) eigenstates do 
980: not interefere with each other. The effect derives entirely 
981: from the independent phase advance of the individual states and 
982: their translation relative to the laboratory rest frame. 
983: 
984: \newpage
985: 
986: \noindent TABLE I: Eigenmasses for various values of $\theta$ and 
987: $\phi$ for cases of two approximately degenerate pseudo-Dirac pairs.
988: \begin{verbatim}
989: __________________________________________________________________________
990: \end{verbatim}
991: $\theta =  9.324078$,   $\phi =  0$
992: 
993: \begin{verbatim}
994:       mass     1active    2active   3active  1sterile  2sterile  3sterile
995: 
996:    1.099328   0.635032   0.000000 -0.310533  0.698108  0.000000 -0.113793
997:    1.100680  -0.633620   0.000000  0.314383  0.697413  0.000000 -0.115345
998:    1.100000   0.000000   0.707107  0.000000  0.000000  0.707107  0.000000
999:    1.100000   0.000000  -0.707107  0.000000  0.000000  0.707107  0.000000
1000:    0.007438   0.441883   0.000000  0.897064 -0.003287  0.000000 -0.002224
1001: 1000.008789   0.000162   0.000000  0.002960  0.162017  0.000000  0.986784
1002: \end{verbatim}
1003: 
1004: $\theta =  9.324078$,   $\phi =  2.25$
1005: 
1006: \begin{verbatim}
1007:       mass     1active    2active    3active  1sterile  2sterile  3sterile
1008: 
1009:    1.095953   0.479130  -0.468214  -0.225940  0.525106 -0.466489 -0.082539
1010:    1.096608   0.437964  -0.514829  -0.208027 -0.480274  0.513243  0.076041
1011:    1.103359   0.416946   0.529767  -0.212981  0.460041  0.531383 -0.078333
1012:    1.104056  -0.458049  -0.484588   0.235669  0.505710  0.486376 -0.086730
1013:    0.007438   0.441553   0.015769   0.897088 -0.003285 -0.000109 -0.002224
1014: 1000.008789   0.000162   0.000007   0.002960  0.161892  0.006361  0.986784
1015: \end{verbatim}
1016: 
1017: $\theta =  9.324078$,   $\phi =  4.5$
1018: 
1019: \begin{verbatim}
1020:       mass     1active    2active    3active  1sterile  2sterile  3sterile
1021: 
1022:    1.092254   0.479875  -0.471453  -0.217491  0.524155 -0.468127 -0.079183
1023:    1.092888  -0.458763   0.495815   0.209390  0.501371 -0.492614 -0.076279
1024:    1.107010   0.416602   0.526536  -0.221472  0.461189  0.529886 -0.081725
1025:    1.107726   0.437718   0.503654  -0.234323 -0.484866 -0.507196  0.086521
1026:    0.007439   0.440571   0.031517   0.897156 -0.003273 -0.000217 -0.002226
1027: 1000.008789   0.000162   0.000014   0.002960  0.161517  0.012712  0.986784
1028: \end{verbatim}
1029: 
1030: $\theta =  9.324078$   $\phi =  22.5$
1031: 
1032: \begin{verbatim}
1033:       mass     1active    2active    3active  1sterile  2sterile  3sterile
1034: 
1035:    1.062925   0.550356  -0.405921  -0.179528  0.584987 -0.392239 -0.063608
1036:    1.063381   0.546548  -0.411257  -0.179609 -0.581185  0.397574  0.063663
1037:    1.134871   0.337840   0.568726  -0.249457  0.383405  0.586755 -0.094367
1038:    1.135731   0.341702   0.564710  -0.254038 -0.388074 -0.583058  0.096172
1039:    0.007475   0.409265   0.154109   0.899298 -0.003058 -0.001048 -0.002241
1040: 1000.008789   0.000150   0.000068   0.002960  0.149684  0.062001  0.986784
1041: \end{verbatim}
1042: 
1043: $\theta =  9.324078$   $\phi =  45$
1044: 
1045: \begin{verbatim}
1046:       mass     1active    2active    3active  1sterile  2sterile  3sterile
1047: 
1048:    1.030458   0.632073  -0.290233  -0.127244  0.651329 -0.271878 -0.043708
1049:    1.030692  -0.630801   0.292859   0.127989  0.650162 -0.274406 -0.043972
1050:    1.163620   0.226485   0.612428  -0.270973  0.263544  0.647849 -0.105102
1051:    1.164612   0.227955   0.610618  -0.274587 -0.265472 -0.646488  0.106595
1052:    0.007566   0.315141   0.286490   0.904762 -0.002384 -0.001969 -0.002282
1053: 1000.008789   0.000115   0.000126   0.002960  0.114563  0.114563  0.986784
1054: ________________________________________________________________________
1055: \end{verbatim}
1056: 
1057: \begin{figure*}[h] 
1058: \includegraphics[height=4.0in]{2K.eps} 
1059: \caption{ Oscillations for all six channels commencing 
1060: from one active flavor with appearance probabilities for 
1061: the two other actives and the steriles using the second 
1062: set of angle parameters in Table I.} 
1063: \label{2K} 
1064: \end{figure*} 
1065: 
1066: \begin{figure*}[h] 
1067: \includegraphics[height=4.0in]{appear.eps} 
1068: \caption{ Appearance probability for one neutrino 
1069: flavor using the second set of angle parameters 
1070: in Table I. The arbitrary units of the abscissa may 
1071: be viewed as time in the rest frame or $L/E$ in the 
1072: laboratory. See Appendix.} 
1073: \label{appear} 
1074: \end{figure*} 
1075: 
1076: \begin{figure*}[h] 
1077: \includegraphics[height=4.0in]{runavs.eps} 
1078: \caption{ Effects of limited resolution in $L/E$. 
1079: Running averages over 100 units have been taken for 
1080: each of the active curves. (The curves for the sterile 
1081: neutrinos remain as in Fig.1.)} 
1082: \label{runavs} 
1083: \end{figure*} 
1084: 
1085: \end{document}
1086: 
1087: