cond-mat0603341/imp.tex
1: % Thesis chapter on impurities in LL
2: 
3: 
4: \chapter{Impurities in Luttinger liquids}
5: \label{sec:imp}
6: \vspace{2cm}
7: \textit{The exactly soluble Luttinger model provides 
8:   a generic scenario 
9:   for one-dimensional Fermi systems with repulsive interactions,
10:   denoted as ``Luttinger liquid''. 
11:   The low-energy physics is completely determined by a
12:   few interaction-dependent characteristic parameters 
13:   describing the  power-law exponents of the correlation
14:   functions.
15:   Already a single static impurity leads to 
16:   peculiar modifications of the electronic properties of Luttinger 
17:   liquids.
18:   Even for a weak impurity potential, physical observables behave as 
19:   if the system is split into two parts in the low-energy limit.
20:   The local density of states near the
21:   impurity and the conductance through the impurity vanish as power 
22:   laws.}
23: \vspace{0.75cm}
24: 
25: 
26: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27: 
28: 
29: \section{Luttinger liquids}
30: \label{sec:imp:luttinger}
31: 
32: 
33: In one-dimensional interacting Fermi systems Fermi-liquid theory is not 
34: valid. 
35: The breakdown of Fermi-liquid theory is indicated already in second 
36: order perturbation theory, where the reduction of the quasi-particle 
37: weight at the Fermi surface due to interactions
38: diverges logarithmically. 
39: These divergencies can be treated by a 
40: weak-coupling renormalization-group method applied to an effective
41: low-energy theory known as $g$-ology model \cite{Sol79}. 
42: Depending on the values of the bare couplings, the renormalized 
43: couplings flow either to strong coupling, and hence out of the 
44: perturbatively controlled regime, or to a fixed-point Hamiltonian, 
45: the exactly soluble Luttinger model \cite{Tom50,Lut63,ML65}.
46: The term ``Luttinger liquid'' has been introduced for the latter 
47: systems, in analogy with the mapping of low-energy states of 
48: interacting electron systems onto the Fermi gas in higher dimensions 
49: for Fermi liquids \cite{Hal81a}.
50: The normal gapless metallic phase is characterized by 
51: {\it i)} a continuous momentum distribution with a power-law singularity at
52: the Fermi surface, described by a nonuniversal exponent $\alpha$;
53: {\it ii)} a single-particle density of states which vanishes as 
54: $|\omega|^{\alpha}$ near the Fermi energy, implying the absence of 
55: fermionic quasi-particles;
56: {\it iii)} finite charge and spin-density responses for long wavelengths 
57: and the existence of collective bosonic charge and spin-density modes;
58: {\it iv)} power-law singularities in various correlation functions with 
59: interaction-dependent exponents;
60: {\it v)} separation of spin and charge degrees of freedom.
61: There are several good reviews on one-dimensional Fermi systems, 
62: recent reviews are presented in Refs.~\cite{Voi95,Gia03}.
63: In the following we will summarize the most important results.
64: 
65: 
66: Theoretical work on interacting fermions in one dimension has 
67: progressed along different lines. Besides the perturbative 
68: investigation of the weak-coupling limit \cite{Sol79}, 
69: Luttinger-liquid theory is usually formulated using the bosonization 
70: technique \cite{ML65,Hal80,Hal81b,Hal81a,LP74,Mat74}.
71: A different approach is based on the Bethe-ansatz method for special 
72: integrable models \cite{Gia03}. The computation of correlation 
73: functions is however very difficult from the complicated
74: expressions for the eigenfunctions.
75: 
76: 
77: As proposed in a seminal work by Haldane \cite{Hal81a},
78: the low-energy physics of the Luttinger model is 
79: generic for interacting fermions in one dimension
80: with repulsive interactions. In the language of
81: the renormalization group the Luttinger model Hamiltonian is the
82: fixed-point Hamiltonian for a large class of one-dimensional 
83: fermions with repulsive interactions. 
84: The Luttinger model can be solved exactly at any interaction; 
85: it is characterized by a linear dispersion 
86: relation, and the electron-electron interaction is limited to forward 
87: scattering only \cite{Gia03}.
88: Umklapp and backscattering processes, as well as additional terms for 
89: more general models arising from band curvature
90: are irrelevant and vanish in the low-energy limit \cite{Hal81a}.
91: As in the Landau Fermi liquid a few parameters
92: completely determine the low-energy physics.
93: The charge degrees of freedom of Luttinger liquids are described by a
94: sound velocity $v_{\rho}$ and the dimensionless parameter $K_{\rho}$,
95: and the spin degrees of freedom are characterized by a spin-wave 
96: velocity $v_{\sigma}$ and $K_{\sigma}$.
97: All correlation functions are uniquely parametrized by $K_{\nu}$ and
98: the velocities of the collective modes 
99: $v_{\nu}$, with $\nu = \rho,\sigma$, in the low-energy limit;
100: the corresponding exponents are determined by $K_{\nu}$.
101: For noninteracting particles $K_{\rho} = K_{\sigma} = 1$. 
102: In the absence of a magnetic field, the ground state is 
103: spin-rotationally invariant and $K_{\sigma} = 1$, while
104: $K_{\rho} <1$ ($>1$) for repulsive (attractive) forces.
105: 
106: 
107: For $K_{\sigma} = 1$ the momentum distribution function 
108: exhibits a power-law singularity at the Fermi level with exponent
109: $\alpha = (K_{\rho}^{\phantom{-1}}\!\!\! + K_{\rho}^{-1} - 2)/4$
110: for any nonvanishing interaction \cite{Gia03}. 
111: For $\alpha <1$ the momentum distribution function 
112: near $k_F$ obeys a power law
113: %
114: \begin{equation}
115:   |n(k)-n(k_F)| \sim |k-k_F|^{\alpha} \; .
116: \end{equation}
117: %
118: The spectral function has the form
119: %
120: \begin{equation}
121:   N(\om)\sim|\om|^{\alpha}
122: \end{equation} 
123: %
124: in the low-energy limit.
125: Landau quasi-particle excitations are absent in the Luttinger liquid.
126: The power laws hold also for nonsoluble generalizations of the model 
127: with a nonlinear dispersion \cite{Hal81a}.
128: 
129: 
130: In the Luttinger model the charge and spin density modes are
131: exact undamped eigenstates, and any excited state of the 
132: model is a superposition of these elementary excitations.
133: This becomes particularly explicit in the bosonized
134: form of the Luttinger model \cite{ML65}. 
135: The Luttinger model Hamiltonian conserves charge and the
136: $z$ component of spin separately on each Fermi point.
137: Charge and spin excitations are completely independent, as
138: the respective terms in the Hamiltonian commute. This phenomenon 
139: is called ``spin-charge separation'' \cite{Gia03}, 
140: charge and spin propagate with different velocities.
141: 
142: 
143: Concerning the leading low-energy long-wavelength response 
144: functions there is no difference between Fermi and Luttinger liquids
145: \cite{Gia03}.
146: Thermodynamic properties as the compressibility and the 
147: susceptibility do not differ from the Fermi-liquid description 
148: and the modification due to the interaction leads to renormalized 
149: coefficients depending on $K_{\nu}$ and $v_{\nu}$.
150: Differences between Fermi and Luttinger-liquid behavior arise only
151: from the enhanced phase space for forward scattering in one
152: dimension.
153: Marked differences appear
154: in the single-particle propagator, which determines the momentum 
155: distribution function and the spectral density for single-particle
156: excitations. In a Fermi liquid residual interactions modify the
157: propagator only on a subleading level, leading for example to a small 
158: quasi-particle decay rate, while in a Luttinger liquid forward
159: scattering affects the leading low-energy behavior.
160: Another distinctive feature of Luttinger liquids is the singular
161: behavior of density correlations with momenta near $2k_F$ 
162: \cite{Gia03}.
163: 
164: 
165: Conservation laws play a crucial role in one-dimensional 
166: Fermi systems \cite{Metzner:9701012}.
167: In addition to the usual charge and spin conservation
168: the discrete structure of the Fermi surface in one dimension leads to 
169: an additional conservation law: separate charge 
170: conservation in low-energy scattering processes for particles near the
171: left and right Fermi points, respectively. 
172: Separate spin conservation is spoiled by the 
173: backscattering process generally present in  
174: models of spin-$\frac{1}{2}$ fermions. In most cases of interest,
175: in particular for the models considered in the present work, 
176: the backscattering amplitude scales to 
177: zero at low energies, and the separate spin 
178: conservation is restored asymptotically.
179: The velocities associated with the corresponding conserved currents 
180: provide a complete parametrization of the low-energy physics 
181: \cite{Hal81a,MC93}. 
182: 
183: 
184: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
185: 
186:    
187: \section{Impurity effects}
188: \label{sec:imp:impeff}
189: 
190: 
191: An important aspect of Luttinger-liquid behavior concerns the peculiar 
192: modification of the electronic properties in the presence of impurities.
193: For Luttinger liquids with repulsive interactions ($K_{\rho}<1$) 
194: already a single static impurity has a strong effect at low energy
195: scales, even if the impurity potential is relatively 
196: weak \cite{LP74,Mat74,AR82,GS88,KF92b,KF92a,FN93,FN93s,YGM94}.
197: In general the interplay of disorder and interactions is still a 
198: challenging issue, 
199: although the properties of noninteracting disordered
200: electronic systems are rather well understood.
201: In one-dimensional noninteracting systems disorder leads to localization 
202: of all electrons; the localization length
203: characterizing the exponential decay of the wave function is of the 
204: same order as the mean free path \cite{Gia03}.
205: On the other hand interactions strongly affect the properties of the 
206: pure system, leading to Luttinger-liquid behavior. Thus in one dimension
207: a particularly strong mutual influence of disorder and interactions 
208: is expected.
209: 
210: 
211: Relevant parameters in the description of disorder are
212: the strength of the individual impurity $V$ and the impurity density 
213: $n_{\rm imp}$. The variation of these two parameters leads to different 
214: physical effects. 
215: In the limiting case of very weak individual impurities with a dense 
216: distribution the effect of a single impurity is negligible and collective 
217: effects dominate;
218: the corresponding relevant length scale is $\sim 1/n_{\rm imp}$.
219: As a consequence of the central-limit theorem, 
220: for continuous systems the disorder can be 
221: described by a Gaussian distribution in the limit 
222: $n_{\rm imp} \to \infty$ and $V \to 0$ for constant $n_{\rm imp} V^2$ 
223: measuring the disorder strength \cite{Gia03}.
224: The main results for Gaussian disorder from a perturbative treatment 
225: can be summarized as follows \cite{Gia03}. 
226: Interactions are effectively renormalized by disorder, which is reversely 
227: affected by interactions.
228: Repulsive interactions generally enhance localization 
229: whereas attractive ones reduce this effect.
230: For spinless fermions attractive interactions 
231: enhance superconducting fluctuations, 
232: leading to an effective screening of the disorder. 
233: For spin-$\frac{1}{2}$ fermions
234: a competing effect arises. The tendency towards a uniform charge 
235: distribution inhibits the coupling to disorder, leading to
236: an increase in the localization length for strong interactions
237: in the pure Hubbard model.
238: For the extended Hubbard model with a local as well as nearest-neighbor 
239: interaction this effect is reduced.
240: The opposite limit examined in the present work corresponds to  
241: strong and dilute impurities. In this case
242: collective effects do not play any role and the problem essentially 
243: reduces to a single isolated impurity.
244: An interesting unsolved problem concerns the combination of single 
245: impurity and collective effects
246: at intermediate scales: depending on whether 
247: collective effects become important before the individual impurities 
248: renormalize to high barriers, a different characteristic behavior is 
249: expected.
250: 
251: 
252: In the following we consider the case of a single or double impurity,
253: where the effects of coherent scattering 
254: from many impurities are absent.
255: The asymptotic low-energy properties of Luttinger liquids with 
256: a single impurity have been investigated by 
257: mapping the problem onto an effective field theory,
258: where terms which are expected to be
259: irrelevant in the low-energy limit are neglected. 
260: For attractive interactions the impurity is
261: irrelevant in the renormalization-group sense and scales to zero
262: at low energies.  
263: For repulsive electron systems with $K_{\rho}<1$ the essential 
264: properties from the perturbative bosonic
265: renormalization-group calculation and the boundary
266: conformal field-theory analysis can be summarized as follows.
267: The backscattering amplitude generated by a weak impurity is a 
268: relevant perturbation which grows as $\Lam^{(K_{\rho}-1)/z}$,
269: for a decreasing energy scale $\Lambda$, where
270: $z$ is the number of spin components.
271: This behavior can be traced back
272: to the power-law singularity of the $2k_F$ density response
273: function in a Luttinger liquid. 
274: On the other hand, the tunneling amplitude through a weak link 
275: between two otherwise separate wires is irrelevant and scales to zero as 
276: $\Lambda^{\alpha_B}$, with the boundary exponent 
277: %
278: \begin{equation}
279:   \alpha_B = \frac{1}{z}(K_{\rho}^{-1}-1)
280: \end{equation}
281: %
282: depending only on the interaction strength and band
283: filling, but not on the impurity parameters.
284: At low energy scales any impurity thus effectively ``cuts'' 
285: the system into two parts with open boundary conditions at 
286: the end points, and physical observables are controlled by the 
287: open chain fixed point. 
288: 
289: 
290: In particular,
291: the local density of states near the impurity is suppressed as 
292: %
293: \begin{equation}
294:   \rho (\omega) \sim |\omega|^{\alpha_B}
295: \end{equation}
296: %
297: for $|\omega| \to 0$.
298: 
299: 
300: Long-range Friedel oscillations in the density profile induced by 
301: boundaries or impurities decay 
302: with a power law at long distances \cite{EG95} as
303: %
304: \begin{equation}
305:   \label{eq:dens}
306:   n(x) \sim x^{-K_{\rho}} 
307: \end{equation}
308: %
309: for spinless fermions, 
310: where $x$ measures the distance from the impurity or boundary. For
311: spin-$\frac{1}{2}$ fermions $K_{\rho}$ is replaced by $(K_{\rho}+1)/2$ 
312: in Eq.~(\ref{eq:dens}).
313: 
314: 
315: The conductance through an infinite Luttinger liquid with a single impurity
316: vanishes at low temperatures as
317: %
318: \begin{equation}
319:   G(T) \sim T^{2\alpha_B} \; .
320: \end{equation}
321: % 
322: The conductance through a single impurity of variable strength
323: can be collapsed onto a single curve by a one-parameter
324: scaling ansatz.  
325: For resonant scattering at double barriers 
326: the distance between the two barriers and the detuning from resonance
327: introduce additional scales and a more complex behavior is observed.
328: The Lorentzian resonance line shape for
329: noninteracting electrons is modified by the interaction, and
330: for appropriate parameters the 
331: conductance exhibits distinctive power-law scaling 
332: as a function of temperature \cite{KF92c,FN93,Fur98,NG03,PG03,YGM94}. 
333: 
334: 
335: Note that the above power laws are strictly valid only in the
336: absence of two-particle backscattering. For spin-$\frac{1}{2}$ 
337: fermions they are in general modified by logarithmic corrections. 
338: The asymptotic behavior is universal in the sense that the exponents
339: depend only on the properties of the bulk system, via $K_{\rho}$,
340: while they do not depend on the impurity strength or shape, except
341: in special cases such as resonant scattering at double barriers, 
342: which require fine-tuning of parameters.
343:  
344: 
345: The asymptotic low-energy properties of Luttinger liquids with a
346: single impurity are rather well understood. Universal power laws
347: and scaling functions have been obtained by bosonization, conformal
348: field theory, and exact solutions for the low-energy asymptotics in 
349: special integrable cases \cite{Gia03}.
350: Numerical results from exact diagonalization and
351: DMRG applied to the lattice model of spinless fermions
352: with nearest-neighbor interaction confirmed 
353: the field theoretical scenario and the validity
354: of the underlying assumptions \cite{EA92,Med98}.
355: These methods are however limited to lattice systems with about 
356: $1000$ sites and do not allow for
357: a systematic analysis of the crossover between the weak
358: and strong-impurity limit.
359: Moreover, only a restricted set of observables can be evaluated with 
360: affordable computational effort.
361: In this context the fRG provides a complementary technique 
362: for microscopic models of interacting fermions with impurities,
363: which does not only capture correctly the universal low-energy
364: asymptotics, but allows one to compute observables on all energy
365: scales, providing thus also nonuniversal properties, and in
366: particular an answer to the important question at what scale the
367: ultimate asymptotics sets in. That scale can indeed be surprisingly
368: low, and the properties above it very different from the asymptotic
369: behavior.
370: 
371: 
372: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
373: 
374: 
375: \section{Experimental realization}
376: \label{sec:imp:exp}
377: 
378: 
379: The progress in the fabrication of artificial low-dimensional structures
380: led to advanced experimental verification of the theoretical 
381: predictions. 
382: %Following the recent overview on the experimental verification 
383: %of Luttinger-liquid behavior in Ref.~\cite{Sch05}, 
384: We present a short list of the most
385: promising systems and of the employed experimental techniques.
386: For a detailed discussion and references to the most recent publications
387: and review articles on the subject we refer to Ref.~\cite{Sch05}.
388: 
389: 
390: Strictly one-dimensional systems are a theoretical idealization, the 
391: coupling to an experimental probe as well as the coupling between 
392: several Luttinger liquids is not completely understood \cite{Gia03}. 
393: The coupling between the chains in a strongly
394: anisotropic three-dimensional compound leads to the development of 
395: long-range order at very low temperatures in the phase for which the 
396: algebraic decay of the corresponding
397: correlation function of the single-chain Luttinger liquid is the slowest.
398: In appropriate temperature and energy regimes Luttinger-liquid behavior 
399: can be expected in 
400: several systems with a predominantly one-dimensional character, as 
401: highly anisotropic quasi one-dimensional conductors, organic conductors 
402: like the Bechgaard salts, as well as inorganic materials, artificial 
403: quantum wires in semiconductor
404: heterostructures or on surface substrates, carbon nanotubes, and 
405: fractional quantum Hall fluids \cite{Sch05}.
406: In particular, single-wall carbon nanotubes
407: are expected to show Luttinger-liquid behavior
408: with $K_{\rho} \sim 0.2 -0.3 $ down to very
409: low temperatures, despite the presence of two
410: low-energy channels \cite{EG97,KBF97}.
411: 
412: 
413: Experimental techniques used to verify Luttinger-liquid behavior involve 
414: mainly high resolution photoemission and transport measurements, in 
415: addition to optical properties \cite{Sch05}.
416: A careful analysis of experimental data 
417: indicating power-law behavior and signatures of spin-charge separation
418: reveals partly inconsistent interpretations.
419: The discussion on the modification of the 
420: quantized value $e^2/h$ for noninteracting electrons
421: in a single channel by the interaction to $K_{\rho} \,(e^2/h)$
422: indicates a sensitive dependence on the schematization of the contacts, 
423: a challenging theoretical as well as experimental problem \cite{Sch05}. 
424: Experimental results for cleaved-edge overgrowth quantum wires and
425: carbon nanotubes indicate power laws
426: of the conductance consistent with Luttinger-liquid behavior. 
427: In the last few years, ultracold gases in optical lattices 
428: have opened up an entirely new area of physics,
429: where strong correlations can be studied
430: with unprecedented flexibility and control of the parameters. 
431: Further work is necessary
432: for clear experimental evidence of Luttinger-liquid behavior.
433: 
434: 
435: 
436: 
437: %%% Local Variables: 
438: %%% mode: latex
439: %%% TeX-master: "thesis"
440: %%% End: 
441: