cond-mat0204534/223.tex
1: \documentclass[10pt,onecolumn]{myart}  
2: \usepackage{graphics}
3: \usepackage{amssymb}
4: \newcommand{\vek}[1]{\mbox{\bf #1}}
5: \newcommand{\bmatrix}[1]{\underline{\mathbf{#1}}}
6: \newcommand{\ubm}[1]{\underline{\mbox{\boldmath$#1$\unboldmath}}}
7: %\renewcommand{\thefootnote}{\arabic{footnote}}
8: \begin{document}
9: \begin{frontmatter}
10: 
11: \title{Hamiltonians and Green's functions which interpolate
12: between two and three dimensions}
13: 
14: %\author[\label1]{Rainer Dick\thanksref{rainer}}
15: \author{Rainer Dick\thanksref{rainer}}
16: 
17: %\address[\label1]{Department of Physics and Engineering Physics,
18: \address{Department of Physics and Engineering Physics,
19:         University of Saskatchewan,\\ 116 Science Place, Saskatoon, 
20:         SK S7N 5E2, Canada}
21: 
22: \thanks[rainer]{rainer@sask.usask.ca}
23: 
24: \begin{abstract}
25: I propose to use Hamiltonians which contain two-dimensional
26: and three-dimensional terms for the description of 
27: two-dimensional systems in physics.
28: 
29: As a model system the evolution of three-dimensional
30: wavefunctions in the presence of an infinitely thin layer
31: is studied.
32: The model predicts distance laws for correlation functions
33: which interpolate between two-dimensional and three-dimensional
34: behavior. It also predicts
35: that in certain cases transmission probabilities
36: through thin layers should depend not only on the transverse,
37: but also on the longitudinal
38: momentum of the infalling particles.
39: 
40: The model also yields a static potential which interpolates between
41: the two-dimensional logarithmic potential at small distances
42: and the three-dimensional $1/r$-potential at large distances.
43: \end{abstract}
44: %\begin{keyword}
45: % keyword1 \sep keyword2
46: % keywords here, in the form: keyword \sep keyword
47: % PACS codes here, in the form: \PACS code \sep code
48: % \PACS % 04.50.+h \sep 98.80.Cq \sep 98.80.Hw
49: %\end{keyword}
50: \end{frontmatter}
51: 
52: \section{Introduction}\label{intro}
53: 
54: Two dimensions played a prominent role 
55: in the development of physics in the last twenty years.
56: 
57: On the experimental side this was driven e.g.\ by the needs
58: of very large scale integration, by
59: applications of semiconducting layer structures, 
60: by exploitations of surface catalytic effects,
61: and by the
62: development of atomic-scale surface analysis and manufacturing tools
63: like scanning tunneling microscopy and atomic force microscopy,
64: to mention only a few developments in this area.
65: 
66: On the theoretical side interest in two-dimensional field theories 
67: was largely driven
68: by string theory \cite{GSW}, where fundamental excitations
69: are described by covariant two-dimensional field theories,
70: and by the realization that in two-dimensional critical systems
71: with a rotational symmetry scaling symmetry may be elevated
72: to full conformal invariance \cite{BPZ}.
73: 
74: The purpose of the present paper is to point out that recent developments
75: in the mathematical formulation of brane world models may also
76: inspire new developments in the physics
77: of low-dimensional systems, and help us to acquire a better understanding
78: of the transition between three-dimensional and two-dimensional behavior
79: in these systems. 
80: 
81: The present work was specifically motivated by the
82: brane world model of Dvali, Ga\-ba\-da\-dze and Porrati, who
83: recently proposed and analyzed a model which combined gravity on
84: a $(3+1)$-dimensional manifold (a "3-brane") with gravity in an ambient
85: $(4+1)$-dimensional bulk \cite{DGP}. 
86: They observed that the combination
87: of gravity in different dimensions yields a gravitational potential
88: which interpolates continuously between the three-dimensional $-1/r$ potential
89: at small distances and the four-dimensional $-1/r^2$ potential
90: at large distances, with a transition scale $\ell_{DGP}\simeq m_3^2/m_4^3$
91: set by the ratios of the reduced Planck masses on the brane and in the bulk.
92: Since we are not concerned with gravity in the present
93: paper I will not explicitly write down the model in terms of intrinsic
94: and extrinsic curvature terms, see \cite{rd1}, but instead refer to
95: the related model of Dvali, Gabadadze and Shifman \cite{DGS}, which combines 
96: a Maxwell term in Minkowski space (coordinates $x^\mu=\{t,\vek{r}\}$) 
97: with a Maxwell term in an ambient
98: $(4+1)$-dimensional bulk (coordinates $x^M=\{t,\vek{r},x^\perp\}$):
99: \begin{equation}\label{eq:actionDGS}
100: S=-\left. \frac{1}{4q_3^2}
101: \int\! dt\int\! d^{3}\vek{r}\,F_{\mu\nu}F^{\mu\nu}\right|_{x^\perp=0}
102: -\frac{1}{4q_4^2}\int\! dt\int\! d^{3}\vek{r}\int\! dx^\perp\,
103: F_{MN}F^{MN}.
104: \end{equation}
105: The resulting Coulomb potential on the $(3+1)$-dimensional
106: Minkowski space\footnote{See Ref.\ \cite{rd2} for a discussion
107: of dynamical potentials on the brane and in the bulk.} is \cite{DGS}
108: \begin{equation}\label{eq:Vdgs}
109: A^0(\vek{r})=\frac{q_3}{4\pi r}\left[
110: \cos\!\left(\frac{2q_3^2}{q_4^2}r\right)
111: -
112: \frac{2}{\pi}\cos\!\left(\frac{2q_3^2}{q_4^2}r\right)
113: \mbox{Si}\!\left(\frac{2q_3^2}{q_4^2}r\right)
114: +\frac{2}{\pi}\sin\!\left(\frac{2q_3^2}{q_4^2}r\right)
115: \mbox{ci}\!\left(\frac{2q_3^2}{q_4^2}r\right)
116: \right],
117: \end{equation}
118: with the sine and cosine integrals
119: \[
120: \mbox{Si}(x)=\int_0^x\! d\xi\,\frac{\sin\xi}{\xi},
121: \qquad
122: \mbox{ci}(x)=-\int_x^\infty\! d\xi\,\frac{\cos\xi}{\xi}.
123: \]
124: The ratio of gauge couplings defines a length scale
125: \begin{equation}\label{eq:ldgs}
126: \ell=\frac{q_4^2}{2q_3^2},
127: \end{equation}
128: and $A^0$ interpolates between a three-dimensional
129: distance law at short distances and
130: a four-dimensional distance law at large distances:
131: \[
132: r\ll\ell:
133: \quad
134: A^0(\vek{r})=\frac{q_3}{4\pi r}
135: \left[1+
136: \frac{2r}{\pi\ell}\!\left(\gamma-1+\ln\!\left(\frac{r}{\ell}\right)\right)
137: +\mathcal{O}\!\left(\frac{r^2}{\ell^2}\right)\right],
138: \]
139: \[
140: r\gg\ell:
141: \quad
142: A^0(\vek{r})=\frac{q_3\ell}{2\pi^2 r^2}
143: \left[1-2\frac{\ell^2}{r^2}
144: + \mathcal{O}\!\left(\frac{\ell^4}{r^4}\right)\right].
145: \]
146: 
147: Action principles of the kind (\ref{eq:actionDGS}) were denoted as 
148: dimensionally hybrid action principles in \cite{rd2}.
149: 
150: Of course, this does not simply carry over to 
151: low-dimensional systems in condensed matter or
152: statistical physics:
153: Dimensionally hybrid action principles would not be a suitable 
154: tool for model building in theoretical investigations of these
155: systems because time derivatives generically will appear
156: only as bulk terms in the Lagrangian of a particle interacting
157: with a low-dimensional structure.
158: 
159: Therefore the main proposal of the present work is to use
160: {\it dimensionally hybrid Hamiltonians} in model building
161: for low-dimensional systems: A combination of two-dimensional
162: and three-dimensional terms and the ensuing interpolating 
163: correlation functions may help to narrow the gap between the
164: powerful methods of two-dimensional field theory and realistic
165: thin layers or surface structures in physics and technology.
166: 
167: In the sequel I will use this idea
168: to discuss non-relativistic particles interacting with
169: a thin layer. The system becomes a dimensionally hybrid system
170: with a specifically two-dimensional component
171: through the assumption that particles in the layer 
172: have a kinetic energy different from particles outside of the layer,
173: e.g.\ as a consequence of mass renormalization $M\to m$
174: due to the interaction of the particles with the
175: components of the layer. 
176: This yields some straightforward but interesting results.
177: 
178: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
179: In the next section I will show in a simple model that transmission
180: probabilities through thin layers in this class of models depend
181: also on the momentum parallel to the layer if $\mu=m(L)/L|_{L\to 0}$
182: remains finite. The discussion
183: of Green's functions in these models
184: will be the subject of Secs.\ \ref{sec:greens} and \ref{sec:greensK}.
185: Sec.\ \ref{sec:greens} contains in particular
186: a potential which interpolates between two-dimensional and
187: three-dimensional distance laws.
188: 
189: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
190: \section{Dimensionally hybrid Hamiltonians
191: and Green's functions}\label{sec:hybrid}
192: 
193: To investigate implications
194: of dimensionally hybrid Hamiltonians
195: for the description of the interaction of particles
196: with a thin layer we assume in the present section
197: that the layer is planar and homogeneous
198: and therefore
199: generates a potential $U(z)$, where $z$ is transverse to the layer.
200: 
201: In realistic two-dimensional
202: systems particles are not strictly bound to the layer,
203: and the effective particle mass in the layer may
204: be changed due to interactions.
205: This motivates the following Hamiltonian for particles
206: of mass $M$ in the presence of the layer 
207: \begin{eqnarray}\label{eq:HQFT}
208: H&=&
209: \left.
210: \frac{\hbar^2}{2\mu}\int\! d^2\vek{x}
211: \nabla\psi^+\cdot\nabla\psi\right|_{z=0}\\
212: \nonumber
213: &&
214: +\int\! d^2\vek{x}\int\! dz\left(\frac{\hbar^2}{2M}
215: \left(\nabla\psi^+\cdot\nabla\psi+\partial_z\psi^+
216: \cdot\partial_z\psi\right)+\psi^+U\psi\right).
217: \end{eqnarray}
218: Here and in the sequel all vectors are 2-dimensional
219: vectors in the layer.
220: 
221: The assumption behind the Hamiltonian (\ref{eq:HQFT})
222:  is that the same field $\psi$ may
223: describe e.g.\ free electrons in the bulk and large polarons\footnote{See e.g.
224: \cite{kittel,madelung,devreese} for introductions to polarons in solids.}
225: or other collective excitations involving conduction
226: electrons in the layer. 
227: The parameter $\mu$ has dimensions of mass per length, and in a limiting
228: procedure from layers of finite thickness $L$ would correspond to
229: \begin{eqnarray}\label{eq:mu}
230: \mu=\lim_{L\to 0}\frac{m(L)}{L},
231: \end{eqnarray}
232: where $m(L)$ would be the mass of the modes in the layer.
233: 
234: The corresponding 
235: equation of motion for stationary single-(quasi)particle
236: wavefunctions is
237: \begin{equation}\label{eq:schrodinger}
238: E\psi(\vek{x},z)=
239: -\delta(z)\frac{\hbar^2}{2\mu}\Delta\psi(\vek{x},0)
240: -\frac{\hbar^2}{2M}\left(\Delta+\partial_z^2\right)
241: \psi(\vek{x},z)+U(z)\psi(\vek{x},z).
242: \end{equation}
243: 
244: The Fourier {\it ansatz}
245: \begin{equation}\label{eq:fourier1}
246: \psi(\vek{x},z)=\frac{1}{2\pi}\int\!d^2\vek{k}\,\psi(\vek{k},z)
247: \exp(\mathrm{i}\vek{k}\cdot\vek{x})
248: \end{equation}
249: yields the separated equation
250: \begin{equation}\label{eq:sep1}
251: \left(E-\frac{\hbar^2k^2}{2M}\right)\psi(\vek{k},z)
252: =-\frac{\hbar^2}{2M}\partial_z^2\psi(\vek{k},z)
253: +U(z)\psi(\vek{k},z)+\delta(z)\frac{\hbar^2k^2}{2\mu}
254: \psi(\vek{k},0).
255: \end{equation}
256: Every solvable model of one-dimensional quantum mechanics
257: gives a solution to this class of layer models, with the kinetic
258: term of the layer modes only generating a cusp
259: proportional to $(M/\mu)k^2$ in $\ln\psi(\vek{k},z)$.
260: 
261: Obviously, the large longitudinal momentum modes are strongly
262: affected by the existence of layer modes, but we will see in a moment
263: that the two-dimensional kinetic term can also have
264: a strong impact on modes with small longitudinal momentum.\\
265: 
266: \begin{center}
267: \scalebox{0.8}{\includegraphics{planmod.eps}}
268: \end{center}
269: 
270: \noindent
271: Fig. 1: Contribution of a virtual planar mode to a particle
272: penetrating a thin homogeneous layer.\\[0.1ex]
273: 
274: It is a trivial exercise to adapt solutions of one-dimensional
275: quantum mechanics to the cusp imposed by the layer modes,
276: but it may be worthwhile to write down the modifications 
277: of the transmission coefficient due to the
278: layer modes when the layer potential represents
279: a work function: $U(z)=-w\delta(z)$.
280: The transmission coefficient for an infalling particle
281: of momentum $\{\vek{k},k_\perp\}$ is
282: \begin{equation}\label{eq:T}
283: T(\vek{k},k_\perp)=
284: \left[1+\frac{M^2}{k^2_\perp}\left(
285: \frac{w}{\hbar^2}-\frac{k^2}{2\mu}\right)^2
286: \right]^{-1},
287: \end{equation}
288: i.e.\ the layer modes increase the transmission probability
289: for low longitudinal momentum modes $0<\hbar^2k^2<4\mu w$,
290: and decrease the transmission probabilities for the modes of higher
291: longitudinal momentum.
292: 
293: This model also predicts a resonance in transmission for a
294: certain value
295: $\hbar^2k^2=2\mu w$ of the {\it longitudinal} momentum.
296: This is as a genuine consequence of the two-dimensional kinetic term
297: in (\ref{eq:HQFT}) and may be the simplest way to test
298: the viability of the idea of dimensionally hybrid Hamiltonians
299: in low-dimensional systems.
300: 
301: Obviously, the prediction of dependence of transmission probabilities
302: on longitudinal momenta requires finiteness of the parameter 
303: $\mu$ (\ref{eq:mu}), i.e.\ a derivation of the "phenomenological"
304: Hamiltonian (\ref{eq:HQFT}) from a limiting procedure of purely
305: three-dimensional models will require a thorough study of finite
306: size effects on mass renormalization in solids.
307: 
308: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
309: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
310: \section{Correlations on a layer}\label{sec:greens}
311: 
312: A model similar to (\ref{eq:HQFT}) allows for a neat discussion of the
313: impact of combinations of kinetic terms from different dimensions
314: on the "two-dimensional" correlation functions on the layer.
315: 
316:  For this we assume that the layer is not necessarily homogeneous,
317: but generates a strongly localized potential
318: \[
319: V(\vek{x},z)=u(\vek{x})\delta(z)
320: \]
321: along the layer:
322: \begin{eqnarray}\label{eq:HQFT2}
323: H&=&
324: \frac{\hbar^2}{2M}\int\! d^2\vek{x}\int\! dz
325: \left(\nabla\psi^+\cdot\nabla\psi+\partial_z\psi^+
326: \cdot\partial_z\psi\right)\\
327: \nonumber
328: &&
329: +\left.
330: \int\! d^2\vek{x}\left(\frac{\hbar^2}{2\mu}
331: \nabla\psi^+\cdot\nabla\psi
332: +\psi^+u\psi\right)\right|_{z=0}.
333: \end{eqnarray}
334: 
335: The generating functional for correlation functions on the layer
336: is\footnote{As usual
337: $\delta/\delta j$ acts from the right if $\psi$ is fermionic.}
338: \begin{equation}\label{eq:Z}
339: Z[j,j^+]=\int\! d\psi d\psi^+\exp\!\left(
340: -\beta H[\psi,\psi^+]
341: -\int\! d^2\vek{x}[\psi^+(\vek{x},0) j(\vek{x})+j^+(\vek{x})\psi(\vek{x},0)]
342: \right)
343: \end{equation}
344: \[
345: =
346: \exp\!\left(-\beta\int\! d^2\vek{x}
347: \frac{\delta}{\delta j(\vek{x})}u(\vek{x})
348: \frac{\delta}{\delta j^+(\vek{x})}\right)
349: Z_0[j,j^+]
350: \]
351: with
352: \begin{equation}\label{eq:Z0}
353: Z_0[j,j^+]\propto
354: \exp\!\left(\frac{2M}{\hbar^2\beta}
355: \int\! d^2\vek{x}\int\! d^2\vek{x}'\,j^+(\vek{x})G(\vek{x}-\vek{x}',0)
356: j(\vek{x}')\right).
357: \end{equation}
358: Since 
359: (\ref{eq:HQFT2}) is a free theory from a field theory point of view,
360: the "two-dimensional" correlations in it
361: can be calculated from tree-level diagrams, which involve
362: only the restriction
363: of the free Green's function $G(\vek{x}-\vek{x}',z)$ to the layer
364: and insertions of the layer potential.
365: 
366: The free Green's function used in (\ref{eq:Z0}) satisfies
367: \begin{equation}\label{eq:greeneq0}
368: \left(\Delta+\partial_z^2\right)G(\vek{x},z)
369: +2\ell\delta(z)\Delta G(\vek{x},0)=-\delta(\vek{x})\delta(z)
370: \end{equation}
371: where
372: \begin{equation}\label{eq:ell}
373: 2\ell=\frac{M}{\mu}.
374: \end{equation}
375: 
376: The {\it ansatz}
377: \begin{equation}\label{eq:g0ans}
378: G(\vek{x},z)=\frac{1}{(2\pi)^3}\int\! d^2\vek{k}\int\! dk_\perp\,
379: G(\vek{k},k_\perp)\exp[\mathrm{i}(\vek{k}\cdot\vek{x}
380: +k_\perp z)]
381: \end{equation}
382: yields
383: \begin{equation}\label{eq:greeneqk0}
384: \left(k^2+k_\perp^2\right)G(\vek{k},k_\perp)
385: +\frac{\ell}{\pi}k^2 \int\! dk'_\perp\, G(\vek{k},k'_\perp)=1.
386: \end{equation}
387: This determines the $k_\perp$-dependence of the propagator
388: \begin{equation}\label{eq:kperpdep1}
389: G(\vek{k},k_\perp)=\frac{f(k)}{k^2+k_\perp^2}.
390: \end{equation}
391: 
392: With
393: \[
394: \frac{1}{\pi}\int_{-\infty}^\infty\! dk'_\perp\,
395: \frac{1}{k^2+k'^2_\perp}
396: =\frac{1}{k}
397: \]
398: we find the Green's function
399: \begin{equation}\label{eq:statickp}
400: G(\vek{k},k_\perp)
401: =\frac{1}{(1+k\ell)(k^2+k_\perp^2)}
402: \end{equation}
403: or
404: \begin{equation}\label{eq:staticz}
405: G(\vek{k},z)
406: =\frac{1}{2k(1+\ell k)}\exp(-k|z|).
407: \end{equation}
408: 
409: The solution\footnote{I follow the conventions of \cite{AS}
410: for Bessel and Struve functions.}
411: \begin{equation}\label{eq:pot}
412: G(\vek{x},z)
413: =\frac{1}{8\pi^2}\int_0^\infty\! dk\int_0^{2\pi}\!
414: d\varphi\,
415: \frac{\exp[k(\mathrm{i}r\cos\varphi-|z|)]}{1+k\ell}
416: =\frac{1}{4\pi}\int_0^\infty\! dk\,
417: \frac{\exp(-k|z|)}{1+k\ell}J_0(kr)
418: \end{equation}
419: of (\ref{eq:greeneq0})
420: can be thought of as the electrostatic potential of a unit charge
421: on the layer, if the fields which are continuous across the layer
422: make a special contribution to the Hamiltonian
423: of the electromagnetic field\footnote{In the spirit of the philosophy
424: advocated in this paper the electromagnetic fields in
425: (\ref{eq:Helm}) would simultaneously describe photons
426: in the bulk and polaritons in the layer.}:
427: \begin{equation}\label{eq:Helm}
428: H[F]=
429: \ell\int\! d^2\vek{x}\left(\vek{E}^2+B_\perp^2\right)
430: +\frac{1}{2}\int\! d^2\vek{x}\int\! dz
431: \left(\vek{E}^2+E_\perp^2+\vek{B}^2+B_\perp^2\right),
432: \end{equation}
433: e.g. as a consequence of a non-vanishing limit of
434: \[
435: 2\ell=\lim_{L\to 0}(\epsilon_r L).
436: \]
437: Here $\epsilon_r$ is the relative permittivity of the layer
438: and $L$ its transverse extension.
439: 
440:  The perturbation
441: series (\ref{eq:Z},\ref{eq:Z0}) requires the Green's function
442: on the layer, which can be expressed as a linear combination
443: of a Struve function and a Bessel function of the second kind:
444: \begin{equation}\label{eq:Phi}
445: \Phi(\vek{x})=G(\vek{x},z)\Big|_{z=0}
446: =\frac{1}{8\ell}
447: \left[\mathbf{H}_0\!\left(\frac{r}{\ell}\right)
448: -Y_0\!\left(\frac{r}{\ell}\right)\right].
449: \end{equation}
450: This interpolates between two-dimensional and three-dimensional
451: distance laws
452: \[
453: r\ll\ell:\quad
454: \Phi(\vek{x})=\frac{1}{4\pi\ell}\left[
455: -\gamma-\ln\!\left(\frac{r}{2\ell}\right)
456: +\frac{r}{\ell}
457: +\mathcal{O}\!\left(\frac{r^2}{\ell^2}\right)\right],
458: \]
459: \[
460: r\gg\ell:\quad
461: \Phi(\vek{x})=\frac{1}{4\pi r}\left[
462: 1-\frac{\ell^2}{r^2}+\mathcal{O}\!\left(\frac{\ell^4}{r^4}\right)
463: \right].
464: \]
465: 
466: $\Phi(\vek{x})$ along with the limiting cases is plotted in Fig.\ 2.\\[2mm]
467: 
468: \begin{minipage}[t]{70mm}
469: \noindent
470: Fig. 2: The solid line is the Green's function (\ref{eq:Phi})
471: on the layer
472: as a function of $x=r/\ell$, in units of $\ell^{-1}$. 
473: The upper dashed line is the 
474: three-dimensional $1/4\pi r$ potential, and the lower dashed line
475: is the two-dimensional logarithmic potential.
476: \end{minipage}
477: \hspace*{15mm}
478: \begin{minipage}[t]{6cm}
479: \vspace*{20mm}
480: \scalebox{0.45}{\includegraphics{223pot.eps}}
481: \end{minipage}
482: 
483: Polaron masses in semiconductors are of the order of effective
484: band masses. Due to strong band curvature the
485: ratios of polaron masses to the mass of free electrons 
486: can be of order $m/M\simeq 10^{-2}$, so one may hope that
487: in thin semiconducting layers of thickness $L$ the parameter
488: $\ell$ is of order $\ell\simeq 10^2L$.
489: Two-dimensional
490: distance laws for correlation functions might then be realized up to distances
491: of order $10L$, and intermittent behavior for distances between $10L$
492: and $200L$. For a possible realization of (\ref{eq:Phi}) as an electromagnetic
493: potential in thin layers, semiconducting compounds involving Pb might be the
494: best bet due to their high relative permittivities of order $\epsilon_r\sim 10^2
495: - 10^3$ \cite{LB}. The corresponding field strength on the layer is
496: \begin{equation}\label{eq:field}
497: -\partial_r\Phi(\vek{x})
498: =\frac{1}{8\ell^2}
499: \left[\mathbf{H}_1\!\left(\frac{r}{\ell}\right)
500: -Y_1\!\left(\frac{r}{\ell}\right)-\frac{2}{\pi}\right].
501: \end{equation}
502: 
503: \vspace*{4mm}
504: 
505: \begin{minipage}[t]{70mm}
506: \noindent
507: Fig. 3: The solid line is the field strength 
508: per charge (\ref{eq:field})
509: on the layer
510: as a function of $x=r/\ell$, in units of $\ell^{-2}$. 
511: The dashed line approaching the solid line for $x<1$ is the 
512: two-dimensional $1/4\pi\ell r$ field, and the other dashed line
513: is the three-dimensional $1/4\pi r^2$ field.
514: \end{minipage}
515: \hspace*{15mm}
516: \begin{minipage}[t]{6cm}
517: \vspace*{20mm}
518: \scalebox{0.45}{\includegraphics{223force.eps}}
519: \end{minipage}
520: 
521: This reduction of the force between charges
522: in a thin dielectric layer with finite
523: $2\ell=\lim_{L\to 0}(\epsilon_r L)$ can pictorially be understood as
524: a consequence of the fact that field lines are refracted away
525: from the layer when they leave the layer. This reduces the field lines e.g.\ 
526: between two opposite charges at short distances, since the field
527: lines cannot re-enter the
528: layer on short scales, whereas for large separation of the
529: two charges along the layer re-entry renders the refraction effect
530: negligible.
531: 
532: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
533: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
534: \section{The Green's function for scattering
535: from two-dimensional potentials on the layer}\label{sec:greensK}
536: 
537: The stationary wave equation from (\ref{eq:HQFT2}) is
538: \begin{equation}\label{eq:schrodinger2}
539: E\psi(\vek{x},z)=
540: \delta(z)\left(-\frac{\hbar^2}{2\mu}\Delta+u(\vek{x})\right)\psi(\vek{x},0)
541: -\frac{\hbar^2}{2M}\left(\Delta+\partial_z^2\right)
542: \psi(\vek{x},z).
543: \end{equation}
544: 
545: We have already noticed that this can be solved exactly for
546: $u(\vek{x})$=constant.
547: 
548: In discussing scattering of bulk particles from the layer in the
549: model (\ref{eq:HQFT2}) we could proceed using the ordinary three-dimensional
550: Green's function for scattering of waves of energy $E=\hbar^2K^2/2M$
551: and treat the full two-dimensional contribution to Eq.\
552: (\ref{eq:schrodinger2}) as a perturbation. However, 
553: here I rather would like
554: to treat only the layer potential $u(\vek{x})$
555: as a perturbation. This has the virtue of reducing
556: the perturbation for large longitudinal momenta.
557: 
558: The relevant unperturbed wave is then
559: \begin{eqnarray}\label{eq:psiin}\nonumber
560: \psi_0(\vek{x},z)&=&\frac{1}{\sqrt{2\pi}^3}
561: \exp\!\left(\mathrm{i}\vek{K}_\|\cdot\vek{x}\right)
562: \left[\Theta(-z)\left(\exp\!\left(\mathrm{i}K_\perp z\right)
563: +\frac{K_\|^2\ell}{\mathrm{i}K_\perp-K_\|^2\ell}
564: \exp\!\left(-\mathrm{i}K_\perp z\right)\right)\right.\\
565: &&\left.
566: +\Theta(z)\frac{K_\perp}{K_\perp+\mathrm{i}K_\|^2\ell}
567: \exp\!\left(\mathrm{i}K_\perp z\right)
568: \right]
569: \end{eqnarray}
570: where again the definition (\ref{eq:ell}) was used.
571: 
572: The relevant Green's function $G_K$
573: for propagation of bulk plane waves of energy
574: \[
575: E=\frac{\hbar^2 K^2}{2M}=\frac{\hbar^2}{2M}\left(\vek{K}_\|^2+K_\perp^2\right)
576: \]
577:  has to satisfy
578: \begin{equation}\label{eq:greeneq}
579: \left(\Delta+\partial_z^2+ K^2\right)G_K(\vek{x},z)
580: +2\ell\delta(z)\Delta G_K(\vek{x},0)=-\delta(\vek{x})\delta(z),
581: \end{equation}
582: and the solution proceeds similarly to the solution of (\ref{eq:greeneq0}).
583: The Fourier {\it ansatz} (\ref{eq:g0ans}) yields
584: \begin{equation}\label{eq:greeneqk}
585: \left(k^2+k_\perp^2- K^2\right)G_K(\vek{k},k_\perp)
586: +\frac{\ell}{\pi}k^2 \int\! dk'_\perp\, G_K(\vek{k},k'_\perp)=1,
587: \end{equation}
588: which determines the $k_\perp$-dependence of the propagator
589: \begin{equation}\label{eq:kperpdepk1}
590: G_K(\vek{k},k_\perp)=\frac{f(k)}{k^2+k_\perp^2- K^2}.
591: \end{equation}
592: At this stage the possibility of poles complicates the calculation slightly:
593: In evaluating the integral in (\ref{eq:greeneqk})
594: with (\ref{eq:kperpdepk1}) for $k<K$
595: we have to make a judicious choice on how to shift the poles
596: or the integration path
597: at $k_\perp=\pm\sqrt{K^2-k^2}$, corresponding to 
598:  correct physical boundary conditions on $G_K(\vek{k},z)$.
599: The correct choice turns out to be
600: $k_\perp=\pm(\sqrt{K^2-k^2}+\mathrm{i}\epsilon)$
601: since $G_K(\vek{k},z)$ is supposed to describe outgoing scattered waves
602: from the layer if $k<K$, i.e.\ we have
603: \begin{eqnarray}\label{eq:kperpdep2}
604: G_K(\vek{k},k_\perp)&=&\frac{f(k)}{k^2+k_\perp^2-K^2-\mathrm{i}\epsilon}\\
605: \nonumber
606: &=&f(k)\left(\mathcal{P}\frac{1}{k^2+k_\perp^2-K^2}
607: +\mathrm{i}\pi\delta(k^2+k_\perp^2-K^2)\right).
608: \end{eqnarray}
609: With
610: \[
611: \frac{1}{\pi}\int_{-\infty}^\infty\! dk'_\perp\,
612: \frac{1}{k^2+k'^2_\perp-K^2-\mathrm{i}\epsilon}
613: =\frac{\Theta(k^2-K^2)}{\sqrt{k^2-K^2}}
614: +\mathrm{i}\frac{\Theta(K^2-k^2)}{\sqrt{K^2-k^2}}
615: \]
616: we find the Green's function
617: at large longitudinal wavelength $k<K$ 
618: \begin{equation}\label{eq:greenlowk}
619: G_K(\vek{k},k_\perp)
620: =\frac{\sqrt{K^2-k^2}}{(\sqrt{K^2-k^2}+\mathrm{i}k^2\ell)
621: (k^2+k_\perp^2-K^2-\mathrm{i}\epsilon)},
622: \end{equation}
623: \begin{equation}\label{eq:greenlowk2}
624: G_K(\vek{k},z)
625: =\frac{1}{2k^2\ell-2\mathrm{i}\sqrt{K^2-k^2}}
626: \exp\!\left(\mathrm{i}\sqrt{K^2-k^2}|z|\right),
627: \end{equation}
628: while the short longitudinal wavelength part is 
629: \begin{equation}\label{eq:greenlargek}
630: G_K(\vek{k},k_\perp)
631: =\frac{\sqrt{k^2-K^2}}{(\sqrt{k^2-K^2}+k^2\ell)
632: (k^2+k_\perp^2-K^2-\mathrm{i}\epsilon)},
633: \end{equation}
634: \begin{equation}\label{eq:greenlargek2}
635: G_K(\vek{k},z)
636: =\frac{1}{2\sqrt{k^2-K^2}+2k^2\ell}\exp\!\left(-\sqrt{k^2-K^2}|z|\right).
637: \end{equation}
638: Of course, the Green's function again reduces to the usual
639: three-dimensional
640: result for $\ell\to 0$, i.e.\ if the modes in the layer become so heavy
641: that they decouple. 
642: 
643: With an incoming plane wave the
644: integral equation following from (\ref{eq:schrodinger2},\ref{eq:greeneq})
645: is
646: \begin{equation}\label{eq:inteq}
647: \psi(\vek{x},z)
648: =\psi_0(\vek{x},z)
649: -\frac{2M}{\hbar^2}
650: \int\! d^2\vek{x}'\,G_K(\vek{x}-\vek{x}',z)u(\vek{x}')\psi(\vek{x}',0).
651: \end{equation}
652: In a Born approximation this yields with (\ref{eq:psiin})
653: \begin{equation}\label{eq:psiborn}
654: \psi(\vek{x},z)
655: =\psi_0(\vek{x},z)
656: -\frac{2M}{\sqrt{2\pi}^7\hbar^2}\frac{K_\perp}{K_\perp+\mathrm{i}K_\|^2\ell}
657: \int\! d^2\vek{k}\,\exp(\mathrm{i}\vek{k}\cdot\vek{x})
658: G_K(\vek{k},z)u(\vek{k}-\vek{K}_\|),
659: \end{equation}
660: where the normalization of the Fourier transformed layer potential is
661: \[
662: u(\vek{q})=\int\! d^2\vek{x}\,\exp(-\mathrm{i}\vek{q}\cdot\vek{x})u(\vek{x}).
663: \]
664: 
665: Eq.\ (\ref{eq:psiborn}) together with (\ref{eq:greenlargek2}) implies that
666: no particular scattering wave is generated by the short wavelength components
667:  at $|\vek{q}|<K_\|+\sqrt{K_\|^2+K_\perp^2}$
668: of the layer potential, which is simply a statement of the limited resolving
669: power of the external wave. However, beyond that the $\ell$-dependence of
670: (\ref{eq:greenlowk2}) and (\ref{eq:psiborn}) implies that the two-dimensional
671: kinetic term in (\ref{eq:HQFT2}) reduces potential scattering at large $K_\|$.
672: 
673: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
674: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
675: 
676: \section{Summary}\label{sec:conc}
677: 
678: Two-dimensional field theory is a very appealing subject with many powerful
679: results. One reason for this is because every tensor and spinor field on a 
680: 2-manifold decomposes into
681: covariant primary fields which provide 1-dimensional representations
682: of the corresponding symmetry groups\footnote{This is true
683: beyond the realm of conformal transformations if the Beltrami
684: parameters on the 2-manifold are used to decompose tensors
685: and spinors into 
686: {\it covariant
687: primary fields}, see Secs.\ 1 and 2 in \cite{rdfp1} for tensors
688: and \cite{hermann} for spinors.}.
689: 
690: However, the assumption of strict two-dimensionality seems too restrictive
691: when it comes to comparisons with actual layers or surface structures in
692: physics.
693: Conservative theoretical approaches to low-dimensional structures in 
694: physics and technology therefore rely on 
695: genuine three-dimensional Hamiltonians and only restrict the 
696: locations and momenta of particles to a surface or a layer 
697: (see e.g.\ Sec.\ 9.2 in \cite{madelung}). In these approaches
698: two-dimensionality is only taken into account at a kinematical level,
699: at the expense of sacrificing the powerful methods and results of
700: two-dimensional field theory.
701: 
702: On the other hand, recent results in brane theory taught us that
703: straightforward combinations of four-dimensional terms and five-dimensional
704: terms in action principles yield interpolating Green's functions
705: on the brane,
706: and it is apparent from the functional integral representation that
707: this property must also hold for higher order correlation functions on the
708: brane.
709: 
710: As mentioned above this cannot carry over directly to low-dimensional systems
711: in condensed matter and statistical
712: physics, but it initiated the present proposal to use dimensionally
713: hybrid Hamiltonians $H=\ell h_2+H_3$
714: for theoretical investigations of low-dimensional
715: structures in physics.
716: 
717: Such an approach has the prospect to provide more realistic results
718:  than strictly two-dimensional field theory, while at the same time
719: utilizing the power of two-dimensional field theory for the
720: determination e.g.\ of equilibrium correlation
721: functions in the limiting cases $\ell\to\infty$ or $k\ell\gg 1$.
722: 
723:  Further virtues of this approach are predictions on the
724: transition behavior between two-dimensional and three-dimensional
725: distance laws in layers and surface structures, and a better understanding
726: on how two-dimensional structures might be approached in the more
727: conservative purely three-dimensional framework, through studies
728: of finite size effects on effective masses and permittivities
729: in three-dimensional models.
730: 
731: To illustrate the use and some straightforward consequences
732: of dimensionally hybrid
733: Hamiltonians a homogeneous layer
734: and layers with strongly localized potentials 
735: were studied. In these settings the
736:  field $\psi$ describes simultaneously bulk particles of mass $M$
737: and excitations of mass $\mu$ per transverse length in the layers.
738: These models 
739: might serve as an approximation to thin layers of semiconductors
740: or polar solids, where the planar modes would correspond to large
741: polarons.
742: The model predicts a strong dependence of transmission probabilities
743: on longitudinal momentum.
744: 
745: The Green's function in the model also yields
746: a static potential which interpolates between the logarithmic
747: two-dimensional distance law at distances $\ll\ell$ and the three-dimensional
748: Coulomb law at distances $\gg\ell$.\\[1ex]
749: 
750: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
751: 
752: \noindent
753: {\bf Acknowledgement:} This work was supported by NSERC.
754: 
755: \begin{thebibliography}{18}
756:  \bibitem{GSW}M.B. Green, J.H. Schwarz, E. Witten, Superstring Theory
757:   (Cambridge University Press, Cambridge, 1987).
758:  \bibitem{BPZ}A.A. Belavin, A.M. Polyakov, A.B. Zamolodchikov,
759:   {\it Infinite conformal symmetry in two-dimensional quantum field theory},
760:   Nucl. Phys. B 241 (1984) 333--380.
761:  \bibitem{DGP}G. Dvali, G. Gabadadze, M. Porrati,
762:   {\it 4-D gravity on a brane in 5-D Minkowski space},
763:   Phys. Lett. B 485 (2000) 208--214.
764:   %%CITATION = HEP-TH 0005016;%%
765:  \bibitem{rd1}R. Dick, {\it Brane worlds},
766:   Class. Quantum Grav. 18 (2001) R1--R23.
767:   %%CITATION = HEP-TH 0105320;%%
768:  \bibitem{DGS}G. Dvali, G. Gabadadze, M. Shifman, {\it (Quasi)localized gauge
769:   field on brane: dissipating cosmic radiation to extra dimensions?},
770:   Phys. Lett. B 497 (2001) 271--280.
771:  \bibitem{rd2}R. Dick, D.M.E. McArthur, {\it A photon mass on the brane},
772:   hep-th/0203271, Phys. Lett. B (in press).
773:   %%CITATION = HEP-TH 0203271;%%
774:  \bibitem{kittel}C. Kittel, Quantum Theory of Solids, 2nd ed.
775:   (Wiley, New York, 1987).
776:  \bibitem{madelung}O. Madelung, Introduction to Solid-State Theory,
777:   3rd pr. (Springer, Berlin, 1996).
778:  \bibitem{devreese}J.T. Devreese, {\it Polarons},
779:   in: Encyclopedia of Applied Physics, Vol. 14, eds. G.L. Trigg
780:   and E.H. Immergut
781:   (Wiley-VCH, New York, 1995) pp. 383-409.
782:  \bibitem{AS}M. Abramowitz, I.A. Stegun, Handbook of Mathematical Functions
783:   (United States National Bureau of Standards, Washington, 1972).
784:  \bibitem{LB}Landolt--B\"ornstein Numerical Data and Functional Relationships
785:   in Science and Technology Vol. III/17f, ed. O. Madelung
786:   (Springer, Berlin, 1983).
787:  \bibitem{rdfp1}R. Dick, {\it Topological aspects 
788:   of chiral fields in two dimensions and superstring vertices},
789:   Fortschr. Phys. 40 (1992) 519--591.
790:  \bibitem{hermann}H. Nicolai, {\it New linear systems for 2D Poincar\'{e}
791:   supergravities}, Nucl. Phys. B 414 (1994) 299--328.
792: \end{thebibliography}
793: 
794: \end{document}
795: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
796: 
797: 
798:  \bibitem{rd3}R. Dick, {\it Half--differentials and fermion propagators},
799:   Rev. Math. Phys. 7 (1995) 689--708.
800:   %%CITATION = HEP-TH 9410099;%%
801:  \bibitem{rdlmp}R. Dick, {\it Conformal gauge fixing in 
802:   Minkowski space}, Lett. Math. Phys. 18 (1989) 67--76.
803: \end{thebibliography}
804: 
805: \end{document}
806: 
807: 
808: 
809: