1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %% %%
3: %% %%
4: %% Plain TeX file for the paper: %%
5: %% %%
6: %% "Supersymmetric Quantum Mechanics with a Point Interaction" %%
7: %% %%
8: %% by %%
9: %% T. Uchino and I. Tsutsui %%
10: %% %%
11: %% %%
12: %%%%%%%%%%%%%%%%%%%%%%%%%% October 2002 %%%%%%%%%%%%%%%%%%%%%%%%%%
13:
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15: %%%%%%%% Collection of macro's for use %%%%%%%%%
16: %%%%%%%% with plain tex %%%%%%%%%
17: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
18:
19: %%% First some fonts %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20: \font\bigbold=cmbx12
21: \font\ninerm=cmr9 \font\eightrm=cmr8 \font\sixrm=cmr6
22: \font\fiverm=cmr5
23: \font\ninebf=cmbx9 \font\eightbf=cmbx8 \font\sixbf=cmbx6
24: \font\fivebf=cmbx5
25: \font\ninei=cmmi9 \skewchar\ninei='177 \font\eighti=cmmi8
26: \skewchar\eighti='177 \font\sixi=cmmi6 \skewchar\sixi='177
27: \font\fivei=cmmi5
28: \font\ninesy=cmsy9 \skewchar\ninesy='60 \font\eightsy=cmsy8
29: \skewchar\eightsy='60 \font\sixsy=cmsy6 \skewchar\sixsy='60
30: \font\fivesy=cmsy5 \font\nineit=cmti9 \font\eightit=cmti8
31: \font\ninesl=cmsl9 \font\eightsl=cmsl8
32: \font\ninett=cmtt9 \font\eighttt=cmtt8
33: \font\tenfrak=eufm10 \font\ninefrak=eufm9 \font\eightfrak=eufm8
34: \font\sevenfrak=eufm7 \font\fivefrak=eufm5
35: \font\tenbb=msbm10 \font\ninebb=msbm9 \font\eightbb=msbm8
36: \font\sevenbb=msbm7 \font\fivebb=msbm5
37: \font\tenssf=cmss10 \font\ninessf=cmss9 \font\eightssf=cmss8
38: \font\tensmc=cmcsc10
39:
40: %%% Some Families %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
41: \newfam\bbfam \textfont\bbfam=\tenbb \scriptfont\bbfam=\sevenbb
42: \scriptscriptfont\bbfam=\fivebb \def\Bbb{\fam\bbfam}
43: \newfam\frakfam \textfont\frakfam=\tenfrak \scriptfont\frakfam=%
44: \sevenfrak \scriptscriptfont\frakfam=\fivefrak \def\frak{\fam\frakfam}
45: \newfam\ssffam \textfont\ssffam=\tenssf \scriptfont\ssffam=\ninessf
46: \scriptscriptfont\ssffam=\eightssf \def\ssf{\fam\ssffam}
47: \def\smc{\tensmc}
48:
49: %%% Definition of 8 point %%%%%%%%%%%%%%%%%%%%%%%%%%%%
50: \def\eightpoint{\textfont0=\eightrm \scriptfont0=\sixrm
51: \scriptscriptfont0=\fiverm \def\rm{\fam0\eightrm}%
52: \textfont1=\eighti \scriptfont1=\sixi \scriptscriptfont1=\fivei
53: \def\oldstyle{\fam1\eighti}\textfont2=\eightsy
54: \scriptfont2=\sixsy \scriptscriptfont2=\fivesy
55: \textfont\itfam=\eightit \def\it{\fam\itfam\eightit}%
56: \textfont\slfam=\eightsl \def\sl{\fam\slfam\eightsl}%
57: \textfont\ttfam=\eighttt \def\tt{\fam\ttfam\eighttt}%
58: \textfont\frakfam=\eightfrak \def\frak{\fam\frakfam\eightfrak}%
59: \textfont\bbfam=\eightbb \def\Bbb{\fam\bbfam\eightbb}%
60: \textfont\bffam=\eightbf \scriptfont\bffam=\sixbf
61: \scriptscriptfont\bffam=\fivebf \def\bf{\fam\bffam\eightbf}%
62: \abovedisplayskip=9pt plus 2pt minus 6pt \belowdisplayskip=%
63: \abovedisplayskip \abovedisplayshortskip=0pt plus 2pt
64: \belowdisplayshortskip=5pt plus2pt minus 3pt \smallskipamount=%
65: 2pt plus 1pt minus 1pt \medskipamount=4pt plus 2pt minus 2pt
66: \bigskipamount=9pt plus4pt minus 4pt \setbox\strutbox=%
67: \hbox{\vrule height 7pt depth 2pt width 0pt}%
68: \normalbaselineskip=9pt \normalbaselines \rm}
69:
70: %%% Definition of 9 point %%%%%%%%%%%%%%%%%%%%%%%%%%%%
71: \def\ninepoint{\textfont0=\ninerm \scriptfont0=\sixrm
72: \scriptscriptfont0=\fiverm \def\rm{\fam0\ninerm}\textfont1=\ninei
73: \scriptfont1=\sixi \scriptscriptfont1=\fivei \def\oldstyle%
74: {\fam1\ninei}\textfont2=\ninesy \scriptfont2=\sixsy
75: \scriptscriptfont2=\fivesy
76: \textfont\itfam=\nineit \def\it{\fam\itfam\nineit}%
77: \textfont\slfam=\ninesl \def\sl{\fam\slfam\ninesl}%
78: \textfont\ttfam=\ninett \def\tt{\fam\ttfam\ninett}%
79: \textfont\frakfam=\ninefrak \def\frak{\fam\frakfam\ninefrak}%
80: \textfont\bbfam=\ninebb \def\Bbb{\fam\bbfam\ninebb}%
81: \textfont\bffam=\ninebf \scriptfont\bffam=\sixbf
82: \scriptscriptfont\bffam=\fivebf \def\bf{\fam\bffam\ninebf}%
83: \abovedisplayskip=10pt plus 2pt minus 6pt \belowdisplayskip=%
84: \abovedisplayskip \abovedisplayshortskip=0pt plus 2pt
85: \belowdisplayshortskip=5pt plus2pt minus 3pt \smallskipamount=%
86: 2pt plus 1pt minus 1pt \medskipamount=4pt plus 2pt minus 2pt
87: \bigskipamount=10pt plus4pt minus 4pt \setbox\strutbox=%
88: \hbox{\vrule height 7pt depth 2pt width 0pt}%
89: \normalbaselineskip=10pt \normalbaselines \rm}
90:
91: %%% Macro to generate the equation #'s automatically.
92: %%% To use start each new section (eg 3) with the commands
93: %%% \secno=3 \meqno=1 :this will start the equations with (3.1)
94: %%% Then in place of \eqno(3.1) type \eqn\descriptivename . To refer
95: %%% back to the equation simply type (\descritivename)
96: %%% For the appendixset \secno=0, \appno=1\meqno=1 etc
97: %%%
98: \global\newcount\secno \global\secno=0 \global\newcount\meqno
99: \global\meqno=1 \global\newcount\appno \global\appno=0
100: \newwrite\eqmac \def\romappno{\ifcase\appno\or A\or B\or C\or D\or
101: E\or F\or G\or H\or I\or J\or K\or L\or M\or N\or O\or P\or Q\or
102: R\or S\or T\or U\or V\or W\or X\or Y\or Z\fi}
103: \def\eqn#1{ \ifnum\secno>0 \eqno(\the\secno.\the\meqno)
104: \xdef#1{\the\secno.\the\meqno} \else\ifnum\appno>0
105: \eqno({\rm\romappno}.\the\meqno)\xdef#1{{\rm\romappno}.\the\meqno}
106: \else \eqno(\the\meqno)\xdef#1{\the\meqno} \fi \fi
107: \global\advance\meqno by1 }
108:
109: %%% Macro to do the refs %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
110: \global\newcount\refno \global\refno=1 \newwrite\reffile
111: \newwrite\refmac \newlinechar=`\^^J \def\ref#1#2%
112: {\the\refno\nref#1{#2}} \def\nref#1#2{\xdef#1{\the\refno}
113: \ifnum\refno=1\immediate\openout\reffile=refs.tmp\fi
114: \immediate\write\reffile{\noexpand\item{[\noexpand#1]\ }#2\noexpand%
115: \nobreak.} \immediate\write\refmac{\def\noexpand#1{\the\refno}}
116: \global\advance\refno by1} \def\semi{;\hfil\noexpand\break ^^J}
117: \def\nl{\hfil\noexpand\break ^^J} \def\refn#1#2{\nref#1{#2}}
118: \def\listrefs{\vfill\eject\immediate\closeout\reffile%\parindent=20pt
119: \centerline{{\bf References}}\bigskip\frenchspacing%
120: \input refs.tmp\vfill\eject\nonfrenchspacing}
121:
122: \def\ann#1#2#3{{\it Ann.\ Phys.}\ {\bf {#1}} ({#2}) #3}
123: \def\cmp#1#2#3{{\it Commun.\ Math.\ Phys.}\ {\bf {#1}} ({#2}) #3}
124: \def\jgp#1#2#3{{\it J.\ Geom.\ Phys.}\ {\bf {#1}} ({#2}) #3}
125: \def\jmp#1#2#3{{\it J.\ Math.\ Phys.}\ {\bf {#1}} ({#2}) #3}
126: \def\jpA#1#2#3{{\it J.\ Phys.}\ {\bf A{#1}} ({#2}) #3}
127: \def\ijmp#1#2#3{{\it Int.\ J.\ Mod.\ Phys.}\ {\bf A{#1}} ({#2}) #3}
128: \def\ijtp#1#2#3{{\it Int.\ J.\ Theor.\ Phys.}\ {\bf {#1}} ({#2}) #3}
129: \def\mplA#1#2#3{{\it Mod.\ Phys.\ Lett.}\ {\bf A{#1}} ({#2}) #3}
130: \def\nc#1#2#3{{\it Nuovo Cimento} {\bf {#1}D} ({#2}) #3}
131: \def\np#1#2#3{{\it Nucl.\ Phys.}\ {\bf B{#1}} ({#2}) #3}
132: \def\pl#1#2#3{{\it Phys.\ Lett.}\ {\bf {#1}B} ({#2}) #3}
133: \def\plA#1#2#3{{\it Phys.\ Lett.}\ {\bf {#1}A} ({#2}) #3}
134: \def\pr#1#2#3{{\it Phys.\ Rev.}\ {\bf {#1}} ({#2}) #3}
135: \def\prA#1#2#3{{\it Phys.\ Rev.}\ {\bf A{#1}} ({#2}) #3}
136: \def\Prb#1#2#3{{\it Phys.\ Rev.}\ {\bf B{#1}} ({#2}) #3}
137: \def\prD#1#2#3{{\it Phys.\ Rev.}\ {\bf D{#1}} ({#2}) #3}
138: \def\prl#1#2#3{{\it Phys.\ Rev.\ Lett.}\ {\bf #1} ({#2}) #3}
139: \def\ptp#1#2#3{{\it Prog.\ Theor.\ Phys.}\ {\bf {#1}} ({#2}) #3}
140: \def\rmp#1#2#3{{\it Rev.\ Mod.\ Phys.}\ {\bf {#1}} ({#2}) #3}
141: \def\prp#1#2#3{{\it Phys.\ Rep.}\ {\bf {#1}C} ({#2}) #3}
142: \def\zpc#1#2#3{{\it Z.\ Phys.}\ {\bf C{#1}} ({#2}) #3}
143:
144: %%% Numbering does not start on title page %%%%%%%%%%%
145: \newif\iftitlepage \titlepagetrue \newtoks\titlepagefoot
146: \titlepagefoot={\hfil} \newtoks\otherpagesfoot \otherpagesfoot=%
147: {\hfil\tenrm\folio\hfil} \footline={\iftitlepage\the\titlepagefoot%
148: \global\titlepagefalse \else\the\otherpagesfoot\fi}
149:
150: %%% Abstract %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
151: \def\abstract#1{{\parindent=30pt\narrower\noindent\ninepoint\openup
152: 2pt #1\par}}
153:
154: %%% A nicer footnote (\note) %%%%%%%%%%%%%%%%%%%%%%%%%
155: \newcount\notenumber\notenumber=1 \def\note#1
156: {\unskip\footnote{$^{\the\notenumber}$} {\eightpoint\openup 1pt #1}
157: \global\advance\notenumber by 1}
158:
159: %%% Date %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
160: \def\today{\ifcase\month\or January\or February\or March\or
161: April\or May\or June\or July\or August\or September\or October\or
162: November\or December\fi \space\number\day, \number\year}
163:
164: %%% More general stuff %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
165: \def\pagewidth#1{\hsize= #1} \def\pageheight#1{\vsize= #1}
166: \def\hcorrection#1{\advance\hoffset by #1}
167: \def\vcorrection#1{\advance\voffset by #1}
168:
169: %%% Output layout of the text %%%%%%%%%%%%%%%%%%%%%%%%
170: \pageheight{23cm}
171: \pagewidth{15.7cm}
172: \hcorrection{-1mm}
173: \magnification= \magstep1
174: \parskip=5pt plus 1pt minus 1pt
175: \tolerance 8000
176: \def\bsk{\baselineskip= 14.5pt plus 1pt minus 1pt}
177: \bsk
178:
179: %%% Definition of extra symbols R, C, Z, N %%%%%%%%%%%
180: \font\extra=cmss10 scaled \magstep0 \setbox1 = \hbox{{{\extra R}}}
181: \setbox2 = \hbox{{{\extra I}}} \setbox3 = \hbox{{{\extra C}}}
182: \setbox4 = \hbox{{{\extra Z}}} \setbox5 = \hbox{{{\extra N}}}
183: \def\RRR{{{\extra R}}\hskip-\wd1\hskip2.0true pt{{\extra I}}
184: \hskip-\wd2 \hskip-2.0 true pt\hskip\wd1}
185: \def\Real{\hbox{{\extra\RRR}}} % Actual special symbol: R
186: \def\CCC{{{\extra C}}\hskip-\wd3\hskip 2.5 true pt{{\extra I}}
187: \hskip-\wd2\hskip-2.5 true pt\hskip\wd3}
188: \def\Complex{\hbox{{\extra\CCC}}\!\!} % Actual special symbol: C
189: \def\ZZZ{{{\extra Z}}\hskip-\wd4\hskip 2.5 true pt{{\extra Z}}}
190: \def\Zed{\hbox{{\extra\ZZZ}}} % Actual special symbol: Z
191: \def\NNN{{{\extra N}}\hskip-\wd5\hskip -1.5 true pt{{\extra I}}}
192: \def\En{\hbox{{\extra\NNN}}} % Actual special symbol: N
193:
194: %%% Some useful macros %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
195: \def\frac#1#2{{#1\over#2}}
196: \def\dfrac#1#2{{\displaystyle{#1\over#2}}}
197: \def\tfrac#1#2{{\textstyle{#1\over#2}}}
198: \def\ket#1{|#1\rangle}
199: \def\bra#1{\langle#1|}
200: \def\pmb#1{\setbox0=\hbox{$#1$} \kern-.025em\copy0\kern-\wd0
201: \kern.05em\copy0\kern-\wd0 \kern-.025em\raise.0433em\box0 }
202: \def\noblackboxes{\overfullrule 0pt}
203: \def\ve{\vfill\eject}
204: \def\pa{\partial}
205: \def\tr{\hbox{tr}}
206: \def\({\left(}
207: \def\){\right)}
208: \def\la{\langle}
209: \def\ra{\rangle}
210:
211: \def\R{{\Bbb R}}
212: \def\X{{\Bbb X}}
213: \def\Z{{\Bbb Z}}
214: \def\C{{\Bbb C}}
215:
216: \def\figone{line1.eps}
217: \def\figtwo{circleline.eps}
218: \def\figthree{spectline.eps}
219:
220: \input epsf
221:
222: %\let\omitpictures=Y
223: \let\omitpictures=N
224:
225: %%% References for the paper:
226:
227: {
228:
229: \refn\RS
230: {M. Reed, B. Simon,
231: \lq\lq Methods of Modern Mathematical Physics\rq\rq ,
232: {\sl Vol.II}, Academic Press, New York, 1980}
233:
234: \refn\AGHH
235: {S. Albeverio, F. Gesztesy, R. H{\o}egh-Krohn and H. Holden,
236: \lq\lq Solvable Models in Quantum Mechanics\rq\rq,
237: Springer, New York, 1988}
238:
239: \refn\FCT
240: {T. F\"{u}l\"{o}p, T. Cheon and I. Tsutsui,
241: {\it Classical Aspects of Quantum Walls in One Dimension}, KEK preprint
242: 2001-134, quant-ph/0111057, to appear in {\it Phys.\ Rev.}\ {\bf {A}}}
243:
244: \refn\CFT
245: {T. Cheon, T. F\"{u}l\"{o}p and I. Tsutsui,
246: \ann{294}{2001}{1-23}}
247:
248: \refn\TFC
249: {I. Tsutsui, T. F\"{u}l\"{o}p and T. Cheon,
250: {\it J. Phys. Soc. Jpn.} {\bf 69} (2000) 3473}
251:
252: \refn\Witt
253: {E. Witten,
254: {\it Nucl.Phys.} {\bf B 185} (1981) 513}
255:
256: \refn\Junker
257: {G. Junker,
258: \lq\lq Supersymmetric Methods in Quantum
259: and Statistical
260: Physics\rq\rq,
261: Springer,
262: Berlin, 1996}
263:
264: \refn\AG
265: {N.I. Akhiezer and I.M. Glazman,
266: \lq\lq Theory of Linear Operators in Hilbert Space\rq\rq,
267: {\sl Vol.II},
268: Pitman Advanced Publishing Program, Boston, 1981}
269:
270: \refn\FT
271: {T. F\"{u}l\"{o}p and I. Tsutsui,
272: \plA{264}{2000}{366}}
273:
274: \refn\TFCtwo
275: {I. Tsutsui, T. F\"{u}l\"{o}p and T. Cheon,
276: {\it Journ. Math. Phys.} {\bf 42} (2001) 5687}
277:
278: }
279:
280:
281: %%% frontpage
282:
283: %%% Output of frontpage %%%%%%%%%%%%%%%%%%%%
284:
285: \pageheight{23cm}
286: \pagewidth{15.7cm}
287: \hcorrection{0mm}
288: \magnification= \magstep1
289: \def\bsk{%
290: \baselineskip= 16.8pt plus 1pt minus 1pt}
291: \parskip=5pt plus 1pt minus 1pt
292: \tolerance 6000
293:
294: %\vsize 19.2cm \voffset -1.2cm %% For temporary purposes !!!
295:
296: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
297:
298: %%% %%% %%%
299: % FrontPage
300: %%% %%% %%%
301:
302: \hfill
303: %\phantom
304: {KEK Preprint 2002-107}
305: \vskip -4pt
306: \hfill \phantom{quant-ph/0207xxx}
307:
308: \vskip 42pt
309:
310: %%% Setting of the baselineskip for frontpage
311: {\baselineskip=18pt
312: %%%
313:
314: \centerline{\bigbold
315: Supersymmetric Quantum Mechanics}
316: \centerline{\bigbold
317: with a Point Singularity}
318:
319: \vskip 30pt
320:
321: \centerline{\smc
322: Takashi Uchino\footnote{${}^*$}
323: {\eightpoint email:\quad uchino@post.kek.jp}
324: \quad
325: {\rm and}
326: \quad
327: Izumi Tsutsui\footnote{${}^\dagger$}
328: {\eightpoint email:\quad izumi.tsutsui@kek.jp}
329: }
330:
331:
332:
333: \vskip 7pt
334:
335: {
336: \baselineskip=13pt
337: \centerline{\it
338: Institute of Particle and Nuclear Studies}
339: \centerline{\it
340: High Energy Accelerator Research Organization (KEK)}
341: \centerline{\it Tsukuba 305-0801}
342: \centerline{\it Japan}
343: }
344:
345:
346:
347: \vskip 100pt
348:
349: \abstract{%
350: {\bf Abstract.}\quad
351: We study the possibility of supersymmetry (SUSY)
352: in quantum mechanics in one
353: dimension under the presence of a point singularity.
354: The system considered is the
355: free particle on a line
356: $\R$ or on the interval $[-l, l]$ where the point singularity lies at
357: $x = 0$. In one dimension, the singularity is known to admit a $U(2)$
358: family of different connection conditions which include as a special case the
359: familiar one that arises under the Dirac delta
360: $\delta(x)$-potential. Similarly, each of the walls at $x = \pm l$ admits
361: a $U(1)$ family of boundary conditions including the Dirichlet and the
362: Neumann boundary conditions. Under these general connection/boundary
363: conditions, the system is shown to possess an $N = 1$ or $N = 2$ SUSY
364: for various choices of the singularity and the walls, and the SUSY
365: is found to be \lq good\rq{} or
366: \lq broken\rq{} depending on the choices made.
367: We use the supercharge
368: which allows for a constant shift in the energy, and argue that if the system
369: is supersymmetric then the supercharge is
370: self-adjoint on states that respect the
371: connection/boundary conditions specified by the singularity.}
372:
373: \vskip 10pt
374: %{\baselineskip=10pt
375: %{\ninepoint
376: %\indent{PACS codes: 3.65.-w, 2.20.-a, 73.20.Dx\hfill\break}
377: %\indent{Keywords: Singular interaction, Inequivalent Quantizations,
378: % Caustics}
379: %}
380: %}
381: %
382: %\vskip 10pt
383: %
384: %\noindent
385: %Number of manuscript pages: xx, \quad figures: x, \quad tables: x
386: %
387: %\centerline{
388: %(Running title: {\sl M{\" o}bius Structure of the Spectral Space
389: %under Point Interaction})
390: %}
391: %
392: %%% End setting of the baselineskip for frontpage
393: }
394: %%%
395: %\bigskip
396: \ve
397:
398: %%% Output layout of the text %%%%%%%%%%%%%%
399:
400: \pageheight{23cm}
401: \pagewidth{15.7cm}
402: \hcorrection{-1mm}
403: \magnification= \magstep1
404: \def\bsk{%
405: \baselineskip= 15.2pt plus 1pt minus 1pt}
406: \parskip=5pt plus 1pt minus 1pt
407: \tolerance 8000
408: \bsk
409:
410: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
411:
412:
413: \secno=1 \meqno=1
414:
415: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
416: \bigskip
417: \noindent{\bf 1. Introduction}
418: \medskip
419: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
420:
421: It has been known for some time that, in one dimension, quantum mechanics admits
422: various different singular point interactions
423: parametrized by the group $U(2)$
424: [\RS, \AGHH]. These include the familiar singularity of
425: the Dirac
426: $\delta(x)$-potential of arbitrary strength which gives rise to discontinuity
427: in the derivative of the wave function, but the generic connection
428: condition in the $U(2)$ family develops discontinuity in both the wave
429: function and its derivative. In mathematical terms, this is equivalent
430: to the fact that the free Hamiltonian operator, defined on the line
431: $\R$ with the singular point removed, admits a $U(2)$ family of
432: self-adjoint extensions.
433: If one considers an interval $[-l, l]$ with point singularity,
434: then in addition to the $U(2)$ family of the singularity, the system is characterized
435: further by the property of the endpoints $x = \pm l$ each of which has a
436: $U(1)$ family of possible boundary conditions (see, {\it e.g.}, [\RS, \FCT]).
437: These varieties in the connection/boundary conditions have been shown to
438: accommodate interesting physical phenomena, such as duality, anholonomy (Berry
439: phase) and scale anomaly, which are normally found in more complicated systems
440: or in quantum field theory [\CFT].
441:
442: The varieties are also expected to furnish a room
443: to realize novel quantum systems with supersymmetry (SUSY). In fact, in our
444: previous work [\CFT, \TFC], we found that there occurs a double degeneracy in the
445: energy level for some specific choices of the conditions, and an attempt was
446: made to reformulate the system into a SUSY quantum mechanics. There, we
447: encountered the problem of
448: how to ensure the self-adjointness of the supercharge
449: under the given connection/boundary conditions. Another problem that needs
450: to be addressed is how to preserve the conditions under the transformations
451: generated by the supercharge. These properties are crucial
452: for the very benefit of SUSY and should be maintained, since otherwise the
453: generic degeneracy in the level and/or the positive semi-definiteness of the energy
454: will not be guaranteed.
455:
456: In this paper, we provide a full analysis on the possibility of SUSY
457: quantum mechanics for these systems, {\it i.e.}, a free particle on the line $\R$
458: or on the interval $[-l, l]$ with a point singularity at $x = 0$. We find that,
459: for a large variety of the connection/boundary conditions, these systems indeed
460: possess an $N = 1$ or
461: $N = 2$ SUSY (the latter case being the Witten model [\Witt]). The
462: supercharge we use is a slightly extended version of the conventional one
463: and allows for a constant shift in the energy. With this
464: supercharge, the two
465: properties mentioned above are shown to be maintained fully,
466: if one takes the energy shift into account.
467: The examples presented
468: include cases where the SUSY is
469: \lq good\rq{} or \lq broken\rq{} [\Junker], showing that these systems, though being
470: simple, embody the essential features of SUSY quantum mechanics observed in
471: other models so far.
472:
473: \secno=2 \meqno=1
474:
475: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
476: \bigskip
477: \noindent{\bf 2. Supersymmetry on a line with point singularity}
478: \medskip
479: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
480:
481: Let us first explore the possibility of SUSY on
482: a line
483: $\R$ in the presence of a point singularity at $x = 0$.
484: The system is defined by
485: the free Hamiltonian,
486: $
487: H = -{\hbar^2\over {2m}}\frac{d^2}{dx^2},
488: $
489: on $\R$ with the point $x = 0$ removed, and the singularity can be
490: characterized by a set of connection conditions at $x = 0$ for the wave
491: function
492: $\psi(x)$ belonging to
493: the Hilbert space ${\cal H} = L^2(\R\backslash\{0\})$.
494: The system may equally be formulated by cutting the space in half (see Fig.1) and
495: identifying
496: ${\cal H}$ with $L^2(\R^+) \otimes \C^2$, where
497: instead of $\psi(x)$ one considers
498: the vector-valued wave function,
499: $$
500: \Psi(x) = \left( {\matrix{{\psi_+(x)}\cr
501: {\psi_-(x)}\cr}
502: }
503: \right),
504: \qquad x \in \R^+,
505: \eqn\no
506: $$
507: defined
508: from
509: $\psi(x)$ by $\psi_+(x) = \psi(x)$ for $x > 0$ and
510: $\psi_-(-x) = \psi(x)$ for $x < 0$.
511: This way we introduce a
512: $\C^2$-graded structure into the system allowing for accommodating SUSY,
513: where now the Hamiltonian takes the form
514: $$
515: H = -{\hbar^2\over {2m}}\frac{d^2}{dx^2}\otimes I,
516: \eqn\ha
517: $$
518: with $I$ being the identity matrix acting on the $\C^2$ vector.
519:
520: Before proceeding further, let us recall that the set of connection conditions
521: which ensures the
522: self-adjointness of the Hamiltonian $H$ in (\ha) is
523: provided by [\AG, \FT, \CFT]
524: $$
525: (U-I)\Psi(+0)+iL_0(U+I)\Psi^\prime(+0)=0.
526: \eqn\nyoro
527: $$
528: Here $U$, called the \lq characteristic matrix\rq{}, is an arbitrary
529: $U(2)$ matrix characterizing uniquely the self-adjoint domain of $H$, which we
530: denote by
531: ${\cal D}_U(H)$, and we used
532: $\Psi'(x) = \frac{d}{dx}\Psi(x)$.\note{%
533: We note that $\Psi'$ used in [\CFT] has an extra minus sign in the second
534: component, but this sign factor
535: is unnecessary here due to the mapping of the negative coordinate to the positive
536: one.
537: }
538: The conditions (\nyoro) may also be written as
539: $$
540: U\Psi^{(+)}(+0) = \Psi^{(-)}(+0), \qquad \hbox{with} \quad
541: \Psi^{(\pm)} = \Psi\pm iL_0\Psi^\prime.
542: \eqn\nyoronyoro
543: $$
544:
545: A system is said to be \lq supersymmetric\rq{} if it has
546: self-adjoint operators $Q_i$, $i = 1, \,2,\ldots$, called supercharges, such
547: that
548: $\{Q_i, Q_j\} = H \,\delta_{ij}$
549: (see, {\it e.g.}, [\Junker]).
550: {}For the free Hamiltonian, the standard form of the
551: supercharges is
552: $Q_i = -i\lambda\frac{d}{dx} \otimes \sigma_i$ where $\lambda =
553: \hbar/\sqrt{2m}$ and $\sigma_i$ are the Pauli matrices. Formally, these
554: supercharges satisfy the relation with the Hamiltonian $H$ in (\ha). However,
555: this is not quite sufficient to prove that the system has a SUSY, since
556: operators are defined not just by the differential operations but also by the domains
557: on which they operate. In fact, the supercharges $Q_i$ may not preserve the
558: self-adjoint domain ${\cal D}_U(H)$ of the Hamiltonian for a given $U$. This
559: can be seen, for example, by considering the domain
560: ${\cal D}_U(H)$ for $U = \sigma_3$, for which the connection conditions read
561: $\psi_+^\prime(+0) = \psi_-(+0) = 0$. The state
562: $\Psi(x) = (0, xe^{-x})^T$ belongs to
563: ${\cal D}_U(H)$ but the transformed state
564: $Q_1\Psi(x) = (i\lambda(x-1)e^{-x}, 0)^T$
565: do not fulfill the connection conditions and
566: hence $Q_1\Psi(x) \not\in {\cal D}_U(H)$.
567: This problem is generic for any domain
568: ${\cal D}_U(H)$, because the supercharges $Q_i$ involve a derivative and,
569: accordingly, they generate a state given basically by the derivative of the
570: original state. Obviously, there is no reason to expect the generated state to
571: remain in the same domain as the original. Under these
572: circumstances, all we can hope for is perhaps to demand that the supercharges map
573: any eigenstate of the Hamiltonian $H$ to some (but not necessarily the same)
574: eigenstate of $H$ specified by the same
575: $U$. This is reasonable because the benefit of SUSY in quantum mechanics
576: is that it may lead to the degeneracy of energy levels by generating an
577: eigenstate from a known eigenstate with the same energy by the operation of $Q_i$. If
578: this is the property we want for SUSY in the system, then we do not need to require
579: $Q_i$ to preserve the domain ${\cal D}_U(H)$, as long as the above demand for
580: eigenstates is fulfilled.
581:
582:
583: \topinsert
584: \epsfxsize 7.0cm
585: \ifx\omitpictures N \centerline{\epsfbox {\figone}} \fi
586: \abstract{{\bf Figure 1.} The system on a line $\R$ with a
587: point singularity at $x = 0$
588: may be identified with the system of two half lines $\R^+$ with the
589: probability flow between the two ends $x = 0$ allowed. }
590: \bigskip
591: \endinsert
592:
593: To seek for supercharges fulfilling this demand, let us consider a slightly more
594: general form than the standard one,
595: $$
596: Q =
597: -i\lambda\frac{d}{dx}\otimes\sigma_{\vec{a}}+ {\bf 1} \otimes\sigma_{\vec{b}} ,
598: \eqn\spp
599: $$
600: where ${\bf 1}$ is the identity operator in $L^2(\R^+)$ and
601: $$
602: \sigma_{\vec{a}} = \sum_{i = 1}^3{a_i \sigma_i},
603: \qquad
604: \sigma_{\vec{b}} = \sum_{i = 1}^3{b_i \sigma_i},
605: \qquad
606: \vert \vec{a}\vert = 1 ,
607: \qquad
608: \vec{a}\cdot \vec{b} = 0,
609: \qquad
610: a_i, \,\, b_i \in \R.
611: \eqn\choice
612: $$
613: The standard form is obtained if we choose $\vec{a} = (1, 0, 0)$ or $(0, 1,
614: 0)$ and $\vec{b} = 0$.
615: The properties (\choice) for the vectors $\vec{a}$ and
616: $\vec{b}$ lead to the relation
617: $2 Q^2 = H + \vert {\vec b} \vert^2$. Now, given a domain ${\cal D}_U(H)$ of the
618: Hamiltonian $H$ specified by the conditions (\nyoro) with some $U$,
619: our demand for $Q$ to be a supercharge is that
620: $$
621: Q\,\Psi(x) \in {\cal D}_U(H)
622: \eqn\no
623: $$
624: for any $\Psi(x)$ which is an eigenstate,
625: $H\,\Psi(x) = E\,\Psi(x)$,
626: of the Hamiltonian $H$.
627: If there are two (or more) such independent
628: operators, then one may choose an appropriate basis $Q_i$ whereby one
629: has
630: $$
631: \{Q_i, Q_j\} = (H+\vert {\vec b} \vert^2)\, \delta_{ij}.
632: \eqn\acc
633: $$
634: We note that our generalized supercharge (\spp)
635: allows the constant shift $\vert {\vec b} \vert^2$ in
636: (\acc) for
637: $ {\vec b} \ne 0$, which is
638: not harmful to SUSY since the shift can always be absorbed into the Hamiltonian
639: by the corresponding energy shift. It is, however, important to notice that
640: due to the constant in the SUSY algebra (\acc) the Hamiltonian $H$ in (\ha) may
641: no longer be positive semi-definite even in the presence of SUSY.
642: Accordingly, the question of the system being good SUSY or broken
643: SUSY [\Junker] will be examined by taking this energy shift into account;
644: they are seen by the property of whether the supercharges annihilate the
645: ground state,
646: $Q_i\Psi^{{\rm grd}} = 0$, rather than whether a zero energy state exists or not.
647:
648:
649: To examine if a given $U$ admits such supercharges, we first note that
650: any $U \in U(2)$ can be decomposed as $U = V^{-1}DV$ with some matrix $V \in SU(2)$
651: and a diagonal matrix,
652: $$
653: D =
654: \left( \matrix{ e^{i\theta_+} & 0 \cr
655: 0 & e^{i\theta_-} }\right),
656: \qquad \theta_\pm \in [0, 2\pi).
657: \eqn\diago
658: $$
659: Observe that, in view of the conditions (\nyoro) which specify the
660: self-adjoint domains corresponding to $U$, if
661: $\Psi(x)
662: \in {\cal D}_U(H)$ then $W \Psi(x) \in {\cal
663: D}_{WUW^{-1}}(H)$ for any $W \in U(2)$. This implies that,
664: if there exists a pair $(U, Q)$ satisfying the above demand, so does the
665: pair $(WUW^{-1}, WQW^{-1})$ where $WQW^{-1}$ is again written in the form (\spp).
666: Choosing in particular
667: $W = V$, we find that the pair $(D, VQV^{-1})$ also satisfies the demand. For this
668: reason, with no loss of generality, we restrict ourselves below
669: to the case where $U$ is
670: diagonal. In other words, once a solution $(D, Q)$ is found for some $D$ and
671: $Q$, then
672: $(U, V^{-1}QV)$ gives the desired solution.
673:
674: To find the solutions $(D, Q)$, we observe that the charge $Q$ in (\spp) induces the
675: transformation on an eigenstate (with energy $E$) and its derivative as
676: $$
677: \eqalign{
678: \Psi(x) &\mapsto -i\lambda\sigma_{\vec{a}}\,\Psi^\prime(x)+\sigma_{\vec{b}}\,\Psi(x),
679: \cr
680: \Psi^\prime(x) &\mapsto
681: -i\lambda\sigma_{\vec{a}}\,\Psi^{\prime\prime}(x)+\sigma_{\vec{b}}\,
682: \Psi^\prime(x) =
683: i\lambda^{-1}E\sigma_{\vec{a}}\,\Psi(x)+\sigma_{\vec{b}}\,\Psi^\prime(x) .
684: }
685: \eqn\trr
686: $$
687: Our demand that the transformed state satisfy the same connection conditions
688: (\nyoro), or (\nyoronyoro), then implies
689: $$
690: \eqalign{
691: \Bigl[
692: \{ -\lambda(D-I) \sigma_{\vec{a}} + 2 L_0 D \sigma_{\vec{b}} \}
693: &+ \{\lambda(D-I)
694: \sigma_{\vec{a}} -2 L_0 \sigma_{\vec{b}} \} D \cr
695: &- L_0^2 \lambda^{-1}E (D+I) \sigma_{\vec{a}} (D + I)
696: \Bigr] \Psi^{(+)}(+0) = 0.
697: }
698: \eqn\senyoro
699: $$
700: An important point to be noted is that the original conditions (\nyoronyoro) provide
701: relations among the components between the two vectors $\Psi^{(+)}(+0)$ and
702: $\Psi^{(-)}(+0)$, but not among those within each of the vectors.
703: This means that the equality (\senyoro) holds without the vector
704: $\Psi^{(+)}(+0)$. We then see that,
705: since the energy $E$ varies with the eigenstate considered, for the
706: conditions (\senyoro) to be identical to the original ones (\nyoro) which are
707: independent of $E$, the last term in the square bracket in (\senyoro) must vanish
708: separately from the rest.
709: The conditions for SUSY therefore become
710: $$
711: (D+I) \sigma_{\vec{a}} (D+I) = 0,
712: \eqn\musa
713: $$
714: and
715: $$
716: \lambda(D-I) \sigma_{\vec{a}} (D-I)
717: + 2 L_0 [D, \sigma_{\vec{b}}] = 0.
718: \eqn\musi
719: $$
720:
721: {}From (\musa) one obtains
722: $\det(D+I) = 0$, and from (\musi) one finds $D \neq -I$.
723: This shows that one of the eigenvalues of $D$ must be $-1$ while the other cannot be
724: $-1$. Since the two eigenvalues in
725: (\diago) can be interchanged by the conjugation
726: $D \rightarrow WDW^{-1}$
727: with
728: $W = i\sigma_1$, the diagonal matrix
729: $D$ can always be taken to be
730: $$
731: D =
732: \left(
733: \matrix{ e^{i\theta} & 0 \cr
734: 0 & -1 }
735: \right), \qquad
736: \theta \ne \pi.
737: \eqn\dia
738: $$
739: {}For this $D$ the condition (\musa) is fulfilled if
740: $$
741: \sigma_{\vec{a}} = \cos\alpha\, \sigma_1 + \sin\alpha\, \sigma_2
742: = e^{-i{\alpha\over 2}\sigma_3}\sigma_1 e^{i{\alpha\over 2}\sigma_3},
743: \eqn\ndia
744: $$
745: where $\alpha \in [0, 2\pi)$ is an arbitrary angle parameter.
746: If we consider in (\spp)
747: the simple supercharge
748: $Q$ with $ \vec b = 0$, then from (\musi)
749: we have $\theta = 0$, {\it i.e.}, $D =
750: \sigma_3$ and the supercharge
751: $Q$ specified by the $\sigma_{\vec{a}}$ in (\ndia).
752: {}For $Q$
753: with
754: $\vec
755: b\ne 0$, we combine (\musa) and (\musi) to find
756: $$
757: [\sigma_3, \sigma_{\vec{b}}] = {{2i\lambda}\over{L(\theta)}}\sigma_{\vec{a}},
758: \eqn\newcondtn
759: $$
760: where we have defined
761: $$
762: L(\theta) = L_0 \cot{{\theta}\over 2},
763: \eqn\no
764: $$
765: which provides a physical length scale to the system [\TFCtwo].
766: The condition (\newcondtn)
767: can then be solved by
768: $$
769: \sigma_{\vec{b}} = {\lambda\over{L(\theta)}}
770: \left\{
771: \sin\alpha\, \sigma_1
772: - \cos\alpha\, \sigma_2
773: \right\} + c\, \sigma_3
774: = -{\lambda\over{L(\theta)}} e^{-i{\alpha\over 2}\sigma_3}\sigma_2
775: e^{i{\alpha\over 2}\sigma_3} + c\, \sigma_3,
776: \eqn\ndib
777: $$
778: with $c \in \R$ being arbitrary. Collecting all, we find that the supercharge
779: takes the form $Q = q(\alpha, c; \theta)$ with
780: $$
781: q(\alpha, c; \theta) =
782: -i\lambda\frac{d}{dx}\otimes
783: e^{-i{\alpha\over 2}\sigma_3}\sigma_1 e^{i{\alpha\over 2}\sigma_3} +{\bf 1}
784: \otimes
785: \left[-{\lambda\over{L(\theta)}}
786: e^{-i{\alpha\over 2}\sigma_3}\sigma_2
787: e^{i{\alpha\over 2}\sigma_3}
788: + c\,
789: \sigma_3\right] .
790: \eqn\spform
791: $$
792: {}For the general $U = V^{-1}DV$, the supercharge is given by
793: $$
794: Q = V^{-1}q(\alpha, c; \theta)V,
795: \eqn\gespform
796: $$
797: as noted above.
798:
799: We therefore have learned that, if the characteristic matrix $U$ has
800: eigenvalues $-1$ and $e^{i\theta} \ne -1$, the system admits two independent
801: supercharges, {\it i.e.}, an
802: $N = 2$ SUSY. For other
803: $U$, no SUSY is allowed under the $Q$ in (\spp).
804: Since the conjugation by $V$ in (\gespform) merely rotates the
805: vectors in the basis $\sigma_i$, for any
806: $U$ that enjoys the $N = 2$ SUSY one may use the concise basis set of
807: supercharges,
808: $$
809: Q_1 = -i \lambda\frac{d}{dx} \otimes \sigma_1
810: - {\lambda\over L(\theta)} \otimes\sigma_2,
811: \qquad
812: Q_2 = -i \lambda\frac{d}{dx} \otimes \sigma_2
813: + {\lambda\over L(\theta)} \otimes\sigma_1,
814: \eqn\mogura
815: $$
816: by setting $\alpha =
817: 0$ or $\alpha = \pi/2$ with $c = 0$.
818: These supercharges
819: satisfy (\acc) with
820: $\vert {\vec b} \vert^2 = [\lambda/L(\theta)]^2$.
821: We remark that, if we restrict ourselves to the
822: simple
823: $Q$ with
824: $\vec b = 0$, then from (\musa) and (\musi) we find that an
825: $N = 2$ SUSY arises only if the eigenvalues of $U$ are $\pm 1$.
826:
827:
828: \secno=3 \meqno=1
829:
830: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
831: \bigskip
832: \noindent{\bf 3. Supersymmetry on an interval with point singularity}
833: \medskip
834: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
835:
836: Next we investigate the possibility of SUSY in the system of an
837: interval with a point singularity ({\it i.e.}, a quantum well with a point
838: interaction). Our assumption for the supercharges remains to be of the form (\spp),
839: but now we need to take into account the boundary effect at both ends of the
840: interval. Let the interval be $[-l, l]$ with the point singularity placed
841: at $x = 0$. As before, we remove the point $x = 0$ from the interval and identify
842: the Hilbert space ${\cal H} = L^2([-l, l]\backslash\{0\})$ with $L^2((0, l])
843: \otimes \C$. The Hamiltonian (\ha) then possesses self-adjoint domains
844: ${\cal D}_{\tilde U}(H)$, where now the characteristic matrix $\tilde U$
845: belongs to
846: $U(2)
847: \times U(1)
848: \times U(1)$ because the point singularity at $x = 0$ furnishes a
849: $U(2)$ arbitrariness while each $U(1)$ corresponds to the arbitrariness
850: provided by the two
851: ends at $x = \pm l$. We write ${\tilde U} = U \times D_l$ where
852: $U \in U(2)$ is the characteristic matrix associated with $x = 0$ and
853: $D_l\in U(1)\times U(1)$ is the one associated with
854: the two ends
855: $x =\pm l$. If we regard the two ends as a special case of point singularity
856: where no
857: probability flow is allowed (see Fig.2), then in terms of the boundary vectors
858: $\Psi(l)$ and
859: $\Psi^\prime(l)$, the matrix $D_l$ may be taken to be a diagonal $U(1)\times
860: U(1)$ matrix embedded in $U(2)$. With these, the self-adjoint domain ${\cal
861: D}_{\tilde U}(H)$ can be specified by the connection conditions at
862: $x = 0$,
863: $$
864: (U - I)\Psi(+0) + iL_0(U +I)\Psi^\prime(+0) = 0,
865: \eqn\ltem
866: $$
867: together with the boundary conditions
868: %\note{%
869: %If we followed the formulation presented in [\FT], the second component of the
870: %boundary vectors
871: %$\Psi(l)$ and
872: %$\Psi^\prime(l)$ at the ends would have a minus sign. The minus sign, however,
873: %can always be absorbed into the matrix $D_l$ to validate (\lten).}
874: at $x = \pm l$,
875: $$
876: (D_l - I)\Psi(l) + iL_0(D_l +I)\Psi^\prime(l) = 0.
877: \eqn\lten
878: $$
879:
880: \topinsert
881: \epsfxsize 7.0cm
882: \ifx\omitpictures N \centerline{\epsfbox {\figtwo}} \fi
883: \abstract{{\bf Figure 2.} The system on the interval $[l,-l]$ with
884: a point singularity at $x = 0$ may be identified with the system of two half
885: intervals $(0, l]$ with the probability flow between the two ends $x = 0$ allowed.
886: The flow is not allowed at the other two ends $x = l$.}
887: \bigskip
888: \endinsert
889:
890:
891: Our task is again to find a pair $({\tilde U} = U \times D_l, Q)$
892: such that the supercharge
893: $Q$ preserve the connection/boundary conditions (\ltem) and (\lten).
894: Recall that a supercharge that preserves the
895: conditions (\ltem) and (\lten) specified by
896: $U$ and $D_l$, respectively, exists only if $U$ and $D_l$ are of the form (\dia)
897: for $\theta \ne \pi$, up to the conjugation by some $V \in SU(2)$
898: for $U$. We therefore have
899: $$
900: U = V^{-1} D V, \qquad
901: D =
902: \left( \matrix{ e^{i\theta} & 0 \cr
903: 0 & -1 } \right) ,
904: \qquad
905: D_l =
906: \left( \matrix{ e^{i\theta_l} & 0 \cr
907: 0 & -1 } \right) ,
908: \eqn\saru
909: $$
910: for $\theta, \, \theta_l \ne \pi$, and
911: our question is whether there is a supercharge $Q$ compatible to both of the
912: conditions (\ltem) and (\lten). In terms of $q(\alpha, c;
913: \theta)$ in (\spform), the supercharge corresponding to $x = 0$ is
914: $V^{-1} q(\alpha, c; \theta) V$ whereas
915: the one corresponding to $x = \pm l$ is
916: $q(\alpha_l, c_l; \theta_l)$, and hence our requirement of compatibility reads
917: $$
918: Q = V^{-1} q(\alpha, c; \theta) V = q(\alpha_l, c_l; \theta_l).
919: \eqn\supcond
920: $$
921: Comparing the terms involving the
922: derivative in (\supcond), one finds
923: $$
924: \tr\,\sigma_i \left[
925: V^{-1}
926: e^{-i{\alpha\over 2}\sigma_3}\sigma_1 e^{i{\alpha\over 2}\sigma_3}
927: V
928: \right] =\tr\, \sigma_i \left[
929: e^{-i{\alpha_l\over 2}\sigma_3}\sigma_1 e^{i{\alpha_l\over 2}\sigma_3}
930: \right],
931: \eqn\no
932: $$
933: for $i = 1$, 2, 3, which can be made more explicit by using the parametrization
934: $$
935: V = e^{i{\mu\over 2}\sigma_2}e^{i{\nu\over 2}\sigma_3},
936: \qquad
937: \mu \in [0, \pi], \quad \nu \in [0, 2\pi),
938: \eqn\zerpo
939: $$
940: as
941: $$
942: \eqalign{
943: \cos\alpha \cos\mu \cos\nu
944: - \sin\alpha\sin\nu &= \cos \alpha_l, \cr
945: \cos\alpha \cos\mu \sin\nu
946: + \sin\alpha\cos\nu &= \sin \alpha_l, \cr
947: \cos\alpha \sin\mu &= 0.
948: }
949: \eqn\no
950: $$
951: These are satisfied if
952: $$
953: \mu = 0,
954: \qquad
955: \alpha_l = \nu + \alpha,
956: \qquad
957: \hbox{or}
958: \qquad
959: \mu = \pi,
960: \qquad
961: \alpha_l = \nu - \alpha \pm \pi,
962: \eqn\scdc
963: $$
964: or otherwise if
965: $$
966: \alpha = {\pi \over 2},
967: \qquad
968: \alpha_l = \nu + {\pi \over 2},
969: \qquad
970: \hbox{or}
971: \qquad
972: \alpha = {{3\pi} \over 2},
973: \qquad
974: \alpha_l = \nu - {\pi \over 2}.
975: \eqn\fstc
976: $$
977:
978: The remaining conditions which arise from
979: the non-derivative term in (\supcond) are
980: $$
981: \eqalign{
982: &\tr\, \sigma_i
983: \left[
984: -{\lambda\over{L(\theta)}}
985: V^{-1}e^{-i{\alpha\over 2}\sigma_3}\sigma_2
986: e^{i{\alpha\over 2}\sigma_3} V
987: + c\,
988: V^{-1}\sigma_3 V
989: \right] \cr
990: &\qquad\qquad =
991: \tr\, \sigma_i
992: \left[
993: -{\lambda\over{L(\theta_l)}}
994: e^{-i{\alpha_l\over 2}\sigma_3}\sigma_2
995: e^{i{\alpha_l\over 2}\sigma_3}
996: + c_l\,
997: \sigma_3
998: \right].
999: }
1000: \eqn\nondercond
1001: $$
1002: Among these conditions the one corresponding to the third component $i = 3$ can
1003: always be fulfilled by adjusting the free constants $c_l$ and $c$.
1004: Also, only
1005: one of the remaining two components, say $i = 2$, is important because, once
1006: its corresponding condition is met, the construction of the
1007: supercharge ensures that the other
1008: component $i = 1$ fulfills its condition, too. The
1009: $i = 2$ component of (\nondercond) is
1010: $$
1011: -{\lambda\over{L(\theta)}}
1012: \left(\cos\alpha \cos\nu
1013: - \sin\alpha \cos\mu \sin\nu
1014: \right) + c \sin\mu \sin\nu =
1015: -{\lambda\over{L(\theta_l)}} \cos \alpha_l.
1016: \eqn\scscdcond
1017: $$
1018:
1019: Now, for (\scdc) we observe that (\scscdcond)
1020: simplifies into $L(\theta) = \pm L(\theta_l)$ or $\theta = \pm \theta_l$.
1021: We thus realize that, if the point singularity at $x = 0$
1022: and the endpoints at $x = \pm l$ are characterized by
1023: $$
1024: U =
1025: \left( \matrix{ e^{i\theta} & 0 \cr
1026: 0 & -1 } \right) ,
1027: \qquad
1028: D_l =
1029: \left( \matrix{ e^{i\theta} & 0 \cr
1030: 0 & -1 } \right) ,
1031: \qquad
1032: \theta \ne \pi,
1033: \eqn\csone
1034: $$
1035: or
1036: $$
1037: U =
1038: \left(\matrix{ -1 & 0 \cr
1039: 0 & e^{-i\theta}} \right) ,
1040: \qquad
1041: D_l =
1042: \left( \matrix{ e^{i\theta} & 0 \cr
1043: 0 & -1 } \right) ,
1044: \qquad
1045: \theta \ne \pi,
1046: \eqn\cstwo
1047: $$
1048: the system has the supercharge
1049: preserving the connection/boundary conditions simultaneously. The supercharge
1050: is
1051: $$
1052: Q = q(\alpha, c; \theta) ,
1053: \eqn\fstspc
1054: $$
1055: where $q(\alpha, c; \theta)$ is given in (\spform). Since the angle parameter
1056: $\alpha$ in the derivative term in (\fstspc)
1057: remains arbitrary, we see that the system admits an
1058: $N = 2$ SUSY.
1059:
1060: On the other hand, for (\fstc) we observe that
1061: (\scscdcond) reduces to $\nu = 0$, $\pi$, or
1062: $$
1063: \pm {\lambda\over{L(\theta)}}
1064: \cos\mu
1065: + c \sin\mu =
1066: \pm {\lambda\over{L(\theta_l)}}.
1067: \eqn\no
1068: $$
1069: This relation may be used to determine the constant $c$ in favor of
1070: $\mu$, $\theta$ and $\theta_l$.
1071: The relevant supercharge then reads
1072: $$
1073: Q = q(\nu \pm \pi/2, c; \theta) ,
1074: \eqn\scdspc
1075: $$
1076: where $\nu$ is one of the
1077: parameters in (\zerpo). In contrast to (\fstspc),
1078: the angle parameter in (\scdspc) is determined by $\nu$ in $U$, and hence
1079: the system admits only an
1080: $N = 1$ SUSY.
1081:
1082: To see in more detail the content of the SUSY systems we have
1083: found, we consider, for instance, the case (\csone) whose
1084: connection/boundary conditions are
1085: $$
1086: \eqalign{
1087: \psi_+(+0) +L(\theta)\, \psi_+^\prime(+0) &= 0,
1088: \qquad \psi_-(+0) = 0, \cr
1089: \psi_+(l) +L(\theta)\, \psi_+^\prime(l) &= 0,
1090: \qquad \psi_-(l) = 0.
1091: }
1092: \eqn\cbone
1093: $$
1094: The eigenstates of the Hamiltonian $H$ with positive energy
1095: $E_n = \hbar^2k_n^2/(2m) > 0$ satisfying (\cbone) are doubly degenerate,
1096: $$
1097: \Psi^{(n)}_+(x) =
1098: N_+^{(n)}\left( \matrix{ -L(\theta)\,k_n \cos k_n x + \sin k_n x \cr
1099: 0 } \right),
1100: \qquad
1101: \Psi^{(n)}_-(x) =
1102: N_-^{(n)}\left( \matrix{ 0 \cr
1103: \sin k_n x } \right),
1104: \eqn\no
1105: $$
1106: where $k_n = n\pi/l$ with $n = 1,2,3,\ldots$, and
1107: $N_+^{(n)}$ and $N_-^{(n)}$
1108: %$N_+^{(n)} = [l\{(L(\theta)k_n)^2 + 1\}/2]^{-1/2}$ and
1109: %$N_-^{(n)} = (l/2)^{-1/2}$
1110: are normalization constants.
1111: Under the SUSY transformations generated by (\mogura), these two states
1112: interchange each other,
1113: $\Psi^{(n)}_\pm \mapsto Q_i\Psi^{(n)}_\pm \propto \Psi^{(n)}_\mp$
1114: for $i = 1, 2$.
1115: The ground state is given by
1116: $$
1117: \Psi^{{\rm grd}}_+(x) =
1118: N_+^{{\rm grd}} \left( \matrix{ e^{-x/L(\theta)} \cr
1119: 0 } \right),
1120: \eqn\ora
1121: $$
1122: with energy
1123: $E^{\rm grd} = - \hbar^2/\{2m(L(\theta))^2\}
1124: = -[\lambda/ L(\theta)]^2 < 0$,
1125: % and $N^{{\rm grd}} = [L(\theta)(1 - e^{-2l/L(\theta)})/2]^{-1/2}$,
1126: which
1127: exists for any
1128: $\theta \ne \pi$
1129: except when $\theta = 0$ which yields the Neumann condition,
1130: $\psi_+'(+0) = \psi_+'(l) = 0$. The ground state is unique and annihilated by the
1131: supercharge
1132: $Q_i\Psi_+^{{\rm grd}} = 0$, which shows that the system possesses a good SUSY.
1133:
1134:
1135: {}For the case (\cstwo), on the other hand, the
1136: connection/boundary conditions read
1137: $$
1138: \eqalign{
1139: \psi_+(+0) &= 0,
1140: \qquad \psi_-(+0) -L(\theta)\, \psi_-^\prime(+0) = 0,
1141: \cr
1142: \psi_+(l) +L(\theta)\, \psi_+^\prime(l) &= 0,
1143: \qquad \psi_-(l) = 0.
1144: }
1145: \eqn\cbtwo
1146: $$
1147: The eigenstates with positive energy are then
1148: $$
1149: \Psi^{(n)}_+(x) =
1150: N_+^{(n)} \left( \matrix{ \sin k_n x \cr
1151: 0 } \right) ,
1152: \qquad
1153: \Psi^{(n)}_-(x) =
1154: N_-^{(n)}\left( \matrix{ 0 \cr
1155: \sin k_n(x-l) } \right) ,
1156: \eqn\no
1157: $$
1158: where the discrete $k_n > 0$ are determined as solutions of
1159: $L(\theta) k_n + \tan(k_n l) = 0$. Similarly,
1160: there arise the ground states,
1161: $$
1162: \Psi^{{\rm grd}}_+(x) =
1163: N_+^{{\rm grd}} \left( \matrix{ \sinh \kappa x \cr
1164: 0 } \right) ,
1165: \qquad
1166: \Psi^{{\rm grd}}_-(x) =
1167: N_-^{{\rm grd}}\left( \matrix{ 0 \cr
1168: \sinh \kappa (x-l) } \right) ,
1169: \eqn\no
1170: $$
1171: with $E^{{\rm grd}} = -\hbar^2\kappa^2/(2m) < 0$,
1172: where $\kappa > 0$ satisfies
1173: $L(\theta) \kappa + \tanh \kappa l = 0$ which has a solution for
1174: $-l < L(\theta) < 0$. If $L(\theta) = -l$, we have in addition the
1175: zero energy states,
1176: $$
1177: \Psi^{{\rm zero}}_+(x) =
1178: N_+^{{\rm zero}} \left( \matrix{ x \cr
1179: 0 } \right) ,
1180: \qquad
1181: \Psi^{{\rm zero}}_-(x) =
1182: N_-^{{\rm zero}}\left( \matrix{ 0 \cr
1183: x-l } \right) .
1184: \eqn\no
1185: $$
1186: Irrespective of the energy (positive, negative or zero), all states are doubly
1187: degenerate and the degenerate pair of states are related by the SUSY
1188: transformations generated by $Q_i$. The degeneracy of the ground states implies
1189: that, in contrast to the previous case, the SUSY is broken here. Note that the
1190: energy of the ground states does not attain the lower bound,
1191: $E^{{\rm grd}} > - [\lambda/ L(\theta)]^2$.
1192:
1193:
1194: In particular, if we choose $\theta = 0$, then the conditions (\cbone) become
1195: the Neumann type $\psi_+^\prime(+0) = 0 = \psi_+^\prime(l)$ and the
1196: Dirichlet type $\psi_-(+0) = 0 = \psi_-(l)$, whereas the
1197: conditions (\cbtwo) become their combinations, $\psi_+(+0) =
1198: 0 =\psi_+^\prime(l)$ and $\psi_-^\prime(+0) = 0 = \psi_-(l)$.
1199: These are the models known earlier [\Junker] whose supercharges are
1200: given without the constant term in (\spp).
1201: Taking into account the $N = 1$ systems realized
1202: on the interval under (\scdspc), we
1203: see that, under the supercharge (\spp) with a constant term, the general point
1204: singularity leads to
1205: a much richer variety of SUSY systems than known before.
1206:
1207: If we restrict ourselves to the
1208: simple $Q$ with $\vec b = 0$, then from the result in section 2, we know
1209: that both of
1210: $U$ and $D_l$ must have the eigenvalues $+1$ and $-1$, that is,
1211: $U = V^{-1}\sigma_3 V$ and $D_l = \sigma_3$ (up to the exchange of the
1212: eigenvalues). This implies the connection/boundary conditions,
1213: $$
1214: \eqalign{
1215: e^{i\nu} \psi_+(+0) - \cot {\mu\over 2}\, \psi_-(+0) &= 0,
1216: \qquad e^{i\nu} \psi_+^\prime(+0) + \tan {\mu\over 2}\,
1217: \psi_-^\prime(+0) = 0,
1218: \cr
1219: \psi^\prime_+(l) &= 0,
1220: \qquad
1221: \psi_-(l) = 0,
1222: }
1223: \eqn\cbtwo
1224: $$
1225: under which we have the eigenstates,
1226: $$
1227: \Psi^{(n)}(x) = N^{(n)}
1228: \left( \matrix{ - e^{-i\nu} \cos k_n (x-l) \cr
1229: \sin k_n (x-l) } \right) ,
1230: \qquad
1231: k_n = {{n\pi + \mu/2}\over{l}},
1232: \eqn\no
1233: $$
1234: for $n \in \Z$. Each eigenstate is invariant
1235: under the SUSY transformation by $Q$, and as shown in Fig.3,
1236: the energy levels $E^{(n)} = \hbar^2 k_n^2/(2m)$
1237: are not degenerate unless $\mu = 0$ or $\pi$.
1238:
1239: \topinsert
1240: \epsfxsize 5.0cm
1241: \ifx\omitpictures N \centerline{ \epsfbox {\figthree} } \fi
1242: \abstract{{\bf Figure 3.}~Energy levels of the $N = 2$ SUSY system
1243: possessing the simple supercharge $Q$ with $\vec b = 0$. The levels
1244: are not degenerate unless $\mu = 0$ or $\pi$.}
1245: \bigskip
1246: \endinsert
1247:
1248: {}For the interval we have yet another extension based on the \lq half
1249: parity\rq{} transformation,
1250: $$
1251: {\cal X}: \Psi(x) \mapsto
1252: ({\cal X}\Psi)(x) = \left( {\matrix{{\psi_+(l - x)}\cr
1253: {\psi_-(x)}\cr}
1254: }
1255: \right),
1256: \eqn\hprty
1257: $$
1258: which is well-defined
1259: in ${\cal H}$ and fulfills ${\cal X}^2 = \hbox{id}$. For the
1260: system characterized by $\tilde U = U \times D_l$ with $U$ and
1261: $D_l$ given either by (\csone) or (\cstwo), this half parity
1262: induces a change in the
1263: characteristic matrix $\tilde U \mapsto \tilde U^{\cal X} = U^{\cal X} \times
1264: D_l^{\cal X}$, namely, if $\Psi \in {\cal D}_{\tilde U}(H)$ then ${\cal
1265: X}\Psi \in {\cal D}_{\tilde U^{\cal X}}(H)$, where $U^{\cal X}$ and
1266: $D_l^{\cal X}$ are given by interchanging the upper left components of
1267: $U$ and $D_l$ with the extra sign $\theta \rightarrow -\theta$ .
1268: Clearly, we have
1269: $Q ({\cal X}\Psi) \in {\cal D}_{\tilde U^{\cal
1270: X}}(H)$ for the supercharge
1271: $Q = {\cal X} q(\alpha, c; \theta) {\cal X}^{-1}$.
1272: The half parity transformation ${\cal X}$ leaves the good SUSY system
1273: (\csone) unaltered, but it turns the broken SUSY system (\cstwo) into
1274: one with
1275: $$
1276: U^{\cal X} =
1277: \left(\matrix{ e^{-i\theta} & 0 \cr
1278: 0 & e^{-i\theta}} \right) ,
1279: \qquad
1280: D^{\cal X}_l =
1281: \left( \matrix{ - 1 & 0 \cr
1282: 0 & -1 } \right) ,
1283: \qquad
1284: \theta \ne \pi.
1285: \eqn\csthree
1286: $$
1287: Under these the connection/boundary conditions read
1288: $$
1289: \eqalign{
1290: \psi_+(+0) -L(\theta)\, \psi_+^\prime(+0) &= 0,
1291: \qquad \psi_-(+0) -L(\theta)\, \psi_-^\prime(+0) = 0,
1292: \cr
1293: \psi_+(l) &= 0,
1294: \qquad \psi_-(l) = 0.
1295: }
1296: \eqn\exbb
1297: $$
1298: Evidently, this system has a spectrum identical to the case (\cstwo) and
1299: the $N = 2$ SUSY is broken. The present case corresponds to
1300: the \lq self-dual\rq{} subfamily mentioned earlier in Ref.[\TFC], which
1301: pointed out that the system becomes a Witten model at $\theta = \pi$ but
1302: fell short of obtaining the full realization of SUSY quantum mechanics
1303: for other $\theta$ due to the
1304: question of the self-adjointness of the supercharge, which we now address
1305: next.
1306:
1307:
1308: \secno=4 \meqno=1
1309:
1310: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1311: \bigskip
1312: \noindent{\bf 4. Self-Adjointness of the supercharge}
1313: \medskip
1314: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1315:
1316: An important question which remains to be answered is
1317: whether our supercharge
1318: $Q$ in (\gespform) can be defined as a self-adjoint operator, and if so,
1319: whether its self-adjoint domain
1320: is compatible with the self-adjoint domain of the
1321: Hamiltonian.
1322:
1323: To answer the former half of the question, we first observe that the problem of
1324: the self-adjointness can be examined by the form (\spp) because, analogously
1325: to the case of the Hamiltonian, a self-adjoint domain for (\gespform), if
1326: any, can be obtained from a domain for (\spp) by the conjugation of the matrix
1327: characterizing the domain of the supercharge. Now, for the
1328: interval $[-l, l]$, for example, the supercharge
1329: (\spp) being self-adjoint implies
1330: $$
1331: \int_{+0}^l\Psi^\dagger(x) (Q\Psi)(x) dx
1332: - \int_{+0}^l(Q\Psi)^\dagger(x) \Psi(x)
1333: dx = 0.
1334: \eqn\marimo
1335: $$
1336: Clearly, the constant term in $Q$ drops out from this
1337: condition and hence does not affect the domain determined from
1338: (\marimo).
1339: Taking
1340: the freedom of conjugation (including the half parity
1341: ${\cal X}$, if necessary) into account, with no loss of generality we
1342: can restrict ourselves to the simple form
1343: $Q = -i\lambda\frac{d}{dx}\otimes\sigma_2$. Then, the
1344: condition (\marimo) reduces to
1345: $$
1346: \Psi^\dagger(l)\sigma_2\Psi(l) -
1347: \Psi^\dagger(+0)\sigma_2\Psi(+0)= 0.
1348: \eqn\yago
1349: $$
1350: This can be fulfilled if
1351: $\Psi(0) = M\Psi(l)$ with some $U(1, 1)$ matrix $M$, which satisfies
1352: $M^\dagger \sigma_2 M = \sigma_2$. Other solutions are given by
1353: $\psi_+(+0) + u\,\psi_-(+0) = 0$ and
1354: $\psi_+(l) + u_l\,\psi_-(l) = 0$, where
1355: $u, u_l \in \R \cup \{\infty\} \simeq U(1)$.
1356: The entire family of the solutions is given by the sum of these, and
1357: they actually form a $U(2)$ group. That this is the case may be argued
1358: by invoking the theory of self-adjoint extensions [\RS] as follows.
1359:
1360: Let ${\cal D}(Q)$ be a symmetric domain of
1361: $Q$ defined by
1362: $$
1363: {\cal D}(Q)= \left\{\Psi\, \big|\, \Psi\in AC([0,l])\otimes\C^2, \,
1364: Q\Psi\in L^2((0,l])\otimes\C^2,\,
1365: \Psi(+0) = \Psi(l) = 0 \right\},
1366: \eqn\no
1367: $$
1368: where $AC([0,l])$ denotes the space of absolutely continuous functions on
1369: the interval $[0,l]$.
1370: The domain of the adjoint $Q^\dagger$ is then given by
1371: $$
1372: {\cal D}(Q^\dagger)= \left\{\Psi\, \big|\,
1373: \Psi\in AC([0,l])\otimes\C^2, \,
1374: Q\Psi\in L^2((0,l])\otimes\C^2 \right\}.
1375: \eqn\no
1376: $$
1377: The eigenvalue equation,
1378: $Q\Psi_{\pm i} = (-i\lambda\frac{d}{dx}\otimes\sigma_2)\Psi_{\pm i} = \pm
1379: i g\, \Psi_{\pm i}$ for any $g > 0$, has the solutions,
1380: $$
1381: \Psi^{(1)}_{\pm i} =
1382: \left( \matrix{ \pm i \cr 1 } \right) e^{g x/\lambda} ,
1383: \qquad
1384: \Psi^{(2)}_{\pm i} =
1385: \left( \matrix{ \pm i \cr -1 } \right) e^{- g x/\lambda} .
1386: \eqn\eifun
1387: $$
1388: Thus the deficiency indices of the operator $Q$ are $(2, 2)$ showing that
1389: the supercharge $Q$ admits a $U(2)$ family
1390: of self-adjoint extensions for the interval.
1391: These extensions are characterized by the aforementioned boundary conditions.
1392:
1393: {}For the line $\R$, the above discussion applies more or less unchanged, except
1394: that the contribution from the endpoint
1395: $x = l$ is now absent. The deficiency indices of the operator $Q$ become $(1,
1396: 1)$ since $\Psi^{(1)}_{\pm i}$ are non-normalizable there. The resultant
1397: $U(1)$ family of self-adjoint domains is realized by the
1398: boundary condition,
1399: $\psi_+(+0) + u\,\psi_-(+0) = 0$.
1400:
1401: We next turn to the latter half of the question, namely, the compatibility of
1402: the two self-adjoint domains of $Q$ and $H$. This, again, can be answered
1403: affirmatively. In fact, we show that any self-adjoint domain ${\cal D}_U(H)$
1404: is a subset of the self-adjoint domain of the supercharge $Q$ associated with
1405: the Hamiltonian $H$. To see this,
1406: let $f_k$, $k = 1, 2$ be the eigenvectors of the characteristic matrix $U$
1407: specifying the Hamiltonian. We then have
1408: $Uf_k =e^{i\theta_i}f_k$, where one of the eigenvalues, say $e^{i\theta_2}$, is
1409: $-1$ while the other $e^{i\theta_1}$ is not $-1$. We now decompose the
1410: boundary vectors
1411: $\Psi(+0)$ and
1412: $\Psi^\prime(+0)$ into the eigenvectors of the
1413: characteristic matrix $U$ of the Hamiltonian,
1414: $$
1415: \Psi(+0)= \sum_{k = 1}^{2}{\langle f_k, \Psi(+0)\rangle \, f_k}, \qquad
1416: \Psi^\prime(+0) = \sum_{k = 1}^{2}{\langle f_k, \Psi^\prime(+0)\rangle
1417: \,f_k},
1418: \eqn\no
1419: $$
1420: where $\langle \cdot, \cdot \rangle$ is the innerproduct for $\C^2$-vectors.
1421: In terms of these, the connection conditions (\nyoro) become
1422: $$
1423: \sum_{k=1}^{2}{\bigl[ (e^{i\theta_i}-1)\langle f_k, \Psi(+0)\rangle \,
1424: + iL_0(e^{i\theta_i}+1) \langle f_k, \Psi^\prime(+0)\rangle \bigr] f_k } = 0.
1425: \eqn\mizusumasi
1426: $$
1427: From the independence of the eigenvectors $f_k$ and $e^{i\theta_2} = -1$, we
1428: find
1429: $$
1430: \langle f_2, \Psi(+0)\rangle = 0.
1431: \eqn\yamori
1432: $$
1433: On the other hand, we recall that
1434: the existence of the supercharge requires (\musa). For $\sigma_{\vec{n}}$
1435: now taken by $\sigma_2$, and for general $U$, (\musa) becomes
1436: $(U+I) \sigma_2 (U+I) = 0$.
1437: Multiplying this by $f_1$ from the right and by
1438: $f_1^\dagger U^\dagger$ from the left, we obtain
1439: $$
1440: f_1^\dagger\,\sigma_2\, f_1 = 0.
1441: \eqn\tanisi
1442: $$
1443: From (\yamori) and (\tanisi), we see that
1444: $$
1445: \Psi^\dagger(+0)\sigma_2\Psi(+0) =
1446: \sum_{j,k = 1}^{2}{\langle f_k, \Psi(+0)\rangle
1447: \langle \Psi(+0), f_j\rangle
1448: \,f_j^\dagger\sigma_2 f_k} = 0.
1449: \eqn\yadokari
1450: $$
1451: So far we have considered only for $x = +0$, but
1452: the contribution from $x = l$ can be evaluated analogously to show that
1453: $\Psi^\dagger(l)\sigma_2\Psi(l) = 0$. We therefore realize that the
1454: requirement for the self-adjointness (\yago) of $Q$ is ensured
1455: for any states belonging to the domain of $H$ for which an associated
1456: supercharge
1457: $Q$ exists, and that this is true for both the line and the
1458: interval system.
1459:
1460: An important consequence of this is that all eigenvalues of $Q$ are real
1461: and, hence, the operator $Q^2$ is positive semi-definite. From this
1462: we find the lower bound of the spectrum,
1463: $H = 2Q^2 - [\lambda/L(\theta)]^2\otimes {\bf 1} \ge -
1464: [\lambda/L(\theta)]^2\otimes {\bf 1}$, which is attained by the ground state
1465: (\ora) in the good SUSY case.
1466:
1467:
1468: \secno=5 \meqno=1
1469:
1470: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1471: \bigskip
1472: \noindent{\bf 5. Conclusion and discussions}
1473: \medskip
1474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1475:
1476: We have seen that a rich variety of $N = 1$ and $N = 2$ SUSY systems appear
1477: on the line and the interval with a singular point. The key element for this
1478: is that we consider the entire family of quantum singularities, and that we
1479: extend the supercharge by introducing a constant term allowing for a shift in
1480: the energy. The resultant $N = 2$ SUSY systems are the Witten models
1481: and exhibit different features depending on the choice of the characteristic
1482: matrix that specifies the singularity.
1483:
1484: The self-adjoint domain of the
1485: supercharge $Q$ is seen to contain the self-adjoint domain of the Hamiltonian
1486: $H$. As a result, the supercharge $Q$ may be expected to ensure the double
1487: degeneracy of energy levels by its operation on the eigenstates. Indeed, this
1488: has been seen in the first two (and the fourth) examples discussed
1489: in the interval case in section 3, but
1490: not in the third one where no degeneracy arises in general. The standard tool to
1491: establish the degeneracy is the Witten parity operator $W$, which is
1492: self-adjoint and satisfies
1493: $W^2= {\bf 1}$, $[W, H]=0$ and $\{W, Q\}=0$.
1494: Obviously, for these conditions we need to examine if the domains of the
1495: operators invloved --- in particular the
1496: domain ${\cal D}_{\tilde U}(H)$ of the Hamiltonian --- change
1497: under the operation of
1498: $W$, and this is a highly nontrivial matter. However, if we assume that
1499: $W$ is given by a $2\times 2$ Hermitian matrix acting on the
1500: graded Hilbert space, then on account of the boundary conditions (\nyoro) we see
1501: that
1502: ${\cal D}_{\tilde U}(H)$ is preserved under $W$ if $[W, U] = [W, D_l] = 0$.
1503: The first two (and the fourth) examples have their ${\tilde U}$
1504: that fulfills this demand with
1505: $W = \sigma_3$, and consequently allow the degeneracy to occur in the
1506: eigenspaces of $\sigma_3$, {\it i.e.}, the upper and the lower components of
1507: the vector states $\Psi$. In contrast, in the
1508: third example, $U$ and $D_l$ do not in general allow a common operator commuting
1509: the both simultaneously, implying that such $W$ cannot exist. In fact, this
1510: seems to be the case for a generic pair of $U$ and $D_l$ unless there underlies
1511: some mechanism to ensure the degeneracy.
1512:
1513: {}Finally, we mention that it is straightforward to extend our analysis to more
1514: complicated systems with point singularities, including a circle with singular
1515: points or a quantum circuit whose vertices may be regarded as point
1516: singularities. For these systems, all we need is to put an appropriate
1517: $U(2)$ matrix to each of the point singularity and seek the supercharge that
1518: preserves the domain of the Hamiltonian.
1519: The extension may also include systems with a
1520: potential $V(x)$ which develops a singularity at its divergent point, such as the
1521: Coulomb potential. Singular potentials that arise in integrable models,
1522: such as the Calogero-Moser models, may also be of interest for the possibility
1523: of accommodating SUSY under the general singularity.
1524:
1525:
1526: \bigskip
1527: \noindent
1528: {\bf Acknowledgement:}
1529: I.T.~thanks T. Cheon and T. F\"{u}l\"{o}p for useful discussions.
1530: This work has been supported in part by
1531: the Grant-in-Aid for Scientific
1532: Research on Priority Areas (No.~13135206) by
1533: the Japanese
1534: Ministry of
1535: Education, Science, Sports and Culture.
1536:
1537:
1538: \baselineskip= 15.5pt plus 1pt minus 1pt
1539: \parskip=5pt plus 1pt minus 1pt
1540: \tolerance 8000
1541: \listrefs
1542:
1543: \bye
1544:
1545:
1546:
1547:
1548:
1549: