cond-mat0509642/fluc.tex
1: \documentclass[twocolumn,pre,twoside,showpacs,showkeys,floatfix, superscriptaddress]{revtex4} 
2: 
3: \usepackage{amsmath}
4: \usepackage{amssymb}
5: \usepackage{graphicx}
6: \usepackage{dcolumn}
7: \usepackage{bm}
8: 
9: \def\d{\mathrm{d}}
10: \def\vec#1{\mathbf{#1}}
11: 
12: 
13: \begin{document}
14: 
15: \title{Non-monotonic fluctuation spectra of membranes pinned \\ or tethered discretely to a substrate}
16: 
17: \author{Rolf-J\"urgen Merath}
18: \affiliation{
19:     II.\ Institut f\"ur Theoretische Physik, 
20:     Universit\"at Stuttgart, 
21:     Pfaffenwaldring 57, 
22:     70550 Stuttgart, 
23:     Germany  
24: }
25: \affiliation{
26:    Max-Planck-Institut f\"ur Metallforschung, 
27:    Heisenbergstra\ss e 3, 
28:    70569 Stuttgart, 
29:    Germany
30: }
31: 
32: \author{Udo Seifert}
33: \affiliation{
34:     II.\ Institut f\"ur Theoretische Physik, 
35:     Universit\"at Stuttgart, 
36:     Pfaffenwaldring 57, 
37:     70550 Stuttgart, 
38:     Germany  
39: }
40: 
41: 
42: %% \date{September 21, 2005}
43: \date{\today}
44: 
45: 
46: 
47: \begin{abstract}
48: The thermal fluctuation spectrum of a fluid membrane coupled harmonically to a solid support by an array of tethers is calculated. 
49: For strong tethers, this spectrum exhibits non-monotonic, anisotropic behavior 
50: with a relative maximum at a wavelength about twice the tether distance. 
51: The root mean square displacement is evaluated to estimate typical membrane displacements. 
52: Possible applications cover pillar-supported or polymer-tethered membranes. 
53: \end{abstract}
54:    
55: 
56: \pacs{82.70.-y, 87.16.Dg}
57: \keywords{supported membrane, fluctuation spectrum}
58: 
59: \maketitle
60: 
61: 
62: 
63: 
64: %-------------------------------------------------------------------------------
65: {\sl Introduction.} --- 
66: Thermal shape fluctuations of fluid membranes in the vicinity of a substrate depend both on the elasticity of the membrane 
67: and the specific type of interaction with the substrate~\cite{US_AdvPhys_1997}. 
68: For laterally homogeneous substrates the combination of steric, van der Waals and electrostatic interactions 
69: determines the strength of these fluctuations. 
70: Such spectra, as measured experimentally using video microscopy \cite{Zilker_1987, Raedler_1995}, 
71: decrease with increasing wave-vector 
72: since shorter wave-length fluctuations cost more bending energy. 
73: A qualitatively different type of interaction arises from tethering or pinning a membrane at discrete points to a substrate. 
74: By adjusting the length of the tethers the mean distance between substrate and membrane can be controlled. 
75: As tethering molecules mostly thio\-lipids are used which consist of a lipid tail, a hydrophilic spacer 
76: (e.\,g.\ peptides \cite{Naumann_Knoll_Langmuir_2003, Bunjes_Knoll_Langmuir_1997} 
77: or polymers \cite{Tanaka_2003_Wiley, Tanaka_2004}) and a sulfur-based linker to the substrate.   
78: Likewise, end-functionalized membrane proteins can serve as tethers~\cite{Giess_Knoll_BiophysJ_2004}. 
79: 
80: Micro- or nano-patterned substrates involving equi\-distant silicon pillars or gold dots 
81: have so far mainly been used to study the interaction with cell membranes \cite{Spatz_2004} 
82: or an actin cortex \cite{Spatz_2003} 
83: but similar experiments with vesicles should become possible as well. 
84: If the membrane binds specifically and firmly to these structures, it is effectively pinned at these points. 
85: For future biotechnological applications of such systems an understanding of how the thermal shape fluctuations are affected 
86: by tethering or pinning is of paramount interest. 
87: 
88: In this paper, we determine the spectrum of shape fluctuations of such membranes. 
89: Surprisingly, we find a large range of parameters for which the fluctuations exhibit a non-monotonic behaviour 
90: with a maximum at a wavelength of the order of the tethering or pinning distance. 
91: Previous theoretical work on membrane conformations on structured
92: substrates focussed on groove-like or rough substrates \cite{Andelman_Langmuir_1999, Andelman_PRE_2001} 
93: or vesicles adhering strongly to chemically patterned substrates~\cite{Lipowsky_JPhysCondMat_2005}. 
94: In a formally related theoretical development, 
95: motivated by the plasma membrane of red blood cells, fluctuations of a compound system coupling a fluid membrane to the cytoskeleton 
96: were investigated~\cite{Gov_2003, Fournier_2004, Gov_2004, Lin_Brown}. 
97: 
98: 
99: 
100: 
101: %-------------------------------------------------------------------------------
102: \vspace{0.2cm}
103: {\sl The model.} --- 
104: The membrane is linked at discrete points $\vec{r}_{\alpha}$ to a substrate by $N$ springs 
105: as sketched in Fig.~\ref{fig:Modell-Arten}. 
106: \begin{figure}[!b]  
107: \hfil \includegraphics[scale=0.63, clip]{Fig1.eps}   \hfil 
108: \vspace{-0.2cm} 
109:    \caption{\label{fig:Modell-Arten}Sketch of a membrane tethered to a substrate at \mbox{discrete} sites. 
110:                        A membrane conformation is characterized by the height profile $h(\vec{r})$, 
111:                        which denotes a displacement relative to the flat configuration at a height 
112:                        given by the rest length $l^{~}_0$ of the tethers. 
113:            }       
114: \vspace{-0.12cm}
115: \end{figure}
116: The membrane surface is parameterized by a height profile $h(\vec{r}) \equiv h(x, y)$  
117: over a rectangular substrate of extensions $L_x$ and $L_y$. 
118: The height $h( \vec{r})$ refers to the displacement with respect to the rest length $l^{~}_0$ of the tethers. 
119: The total energy functional 
120: \begin{equation}  \label{eqn:E_Helfrich+Federn}
121:     E  \; \equiv \;  \frac{\kappa}{2} \int\limits_{0}^{L_x} \!\!   \d x  \!  \int\limits_{0}^{L_y} \!\!  \d y     \,
122:                      \left[\nabla^2 h(\vec{r}) \right]^2    
123:                      \; + \;  \sum\limits_{\alpha = 1}^{N} \frac{K_{\alpha}}{2} \;  h^2(\vec{r_{\alpha}})   
124: \end{equation}   
125: comprises membrane elasticity with bending rigidity $\kappa$
126: and harmonic tethers with strength $K_\alpha$.
127: Assuming periodic boundary conditions in the lateral direc\-tion, a Fourier expansion of the height profile reads 
128: \begin{equation}  \label{eqn:Fourierreihe} 
129:    h(\vec{r})  \: \equiv \:   \sum_\vec{k} \,h^{~}_{\vec{k}}\, e^{i\, \vec{k} \cdot \vec{r}}  
130: \end{equation}
131: with $\vec{k}  \equiv  (k_x, k_y)$ and 
132: $k_{x,y}  = (0, \pm 1, \pm 2, \ldots) \cdot (2 \pi / L_{x,y})$\,. 
133: Since $h(\vec{r})$ is real, the complex Fourier coefficients obey $h^{~}_{-\vec{k}} = h_{\vec{k}}^{*}$. 
134: The spatial average of the height profile, $h^{~}_\vec{0}$, is real. 
135: The energy can be written as      
136: \mbox{$E  \:  =  \:  \rule[-0.18cm]{0cm}{0.52cm}{}    
137:                  \frac{1}{2} \,\sum_{\vec{k},\vec{k}^\prime} \,  h^{*}_{\vec{k}} \, D^{~}_{\vec{k}, \vec{k}^{\prime}} \hspace{0.03cm} 
138:                                                                  h^{~}_{\vec{k}^{\prime}}  $}
139: with a non-diagonal coupling matrix $D^{~}_{\vec{k}, \vec{k^{\prime}}}$. 
140: Following a scheme introduced by Lin and Brown~\cite{Lin_Brown}, a transformation to \mbox{\emph{independent}} variables leads to 
141: \begin{equation}  \label{eqn:cMc}
142:     E ~ = ~ \vec{c} \cdot \vec{M} \, \vec{c}  
143:     ~ \equiv ~ \sum\limits_{r, r^\prime} c^{~}_r \:  M^{~}_{r, r^\prime}  \: c^{~}_{r^\prime}  \ \ .
144: \end{equation}
145: The components $c^{~}_r$ (with $r = 0, 1, 2, 3, \ldots$) of the vector
146: $\vec{c}$ can be grouped into three sectors,        
147: \mbox{$\vec{c} \equiv ( \frac{1}{2} h^{~}_{\vec{0}}, \;   \{ \mathrm{Re}\,h_{\vec{q}} \}  , \; \{ \mathrm{Im}\,h_{\vec{q}} \}  )$}, 
148: where $\vec{q}$ runs through all independent, non-vanishing wave-vectors. 
149: With the definitions 
150: \begin{equation}
151:    E^{~}_r \: \equiv \: \left\{ \begin{array}{ll}    0              &  \quad \mathrm{for~}r=0  \\ 
152:                         \kappa \, L_x L_y \, |\vec{q}(r)|^4   &   \quad \mathrm{for~}r>0   \end{array} \right.   
153: \end{equation}
154: and  
155: \begin{equation} 
156:    m_r^{\alpha} \: \equiv \: \left\{ \begin{array}{ll} 
157:         \!\!\!        ~~\:   \sqrt{2 K_{\alpha}}      
158:                                                &  ~ \mathrm{for~} r=0 ~\mathrm{(1^{st}~sector)}  \hspace*{-4cm}   \\ 
159:         \!\!\!    ~~\:    \sqrt{2 K_{\alpha}} \, \cos(\vec{q}(r) \cdot \vec{r}_{\alpha})  &  
160:                                        ~ \mathrm{for~the~2^{nd}~sector}      \hspace{-4cm}  \\     
161:         \!\!\! -\, \sqrt{2 K_{\alpha}} \, \sin(\vec{q}(r) \cdot \vec{r}_{\alpha})  &  ~ \mathrm{for~the~3^{rd}~sector} 
162:          \hspace{-4cm}   \end{array}     \right.   
163: \end{equation}
164: the matrix $\vec{M}$ 
165: can be simplified to take the form 
166: \vspace{-0.15cm}
167: \begin{equation}  \label{Def.MatrixM}
168:     M^{~}_{r,r^{\prime}} \, = \, E^{~}_r \cdot \delta^{~}_{r,r^{\prime}} 
169:                         \:+\: \sum_{\alpha = 1}^{N} \, m_r^{\alpha} \cdot m_{r^{\prime}}^{\alpha}   \ \ . 
170: \end{equation}
171: The fluctuation spectrum $ \rule[-0.15cm]{0cm}{0.14cm}{}   \langle \, |h^{~}_{\vec{k}}|^2 \, \rangle $ 
172: can then be extracted from the inverse matrix $\vec{M}^{-1}$, 
173: where \mbox{$M^{-1}_{r, r^{\prime}}   =   \frac{2}{k_B T} \, \langle \, c^{~}_r \, c^{~}_{r^\prime}  \rangle$~\cite{Merath}}. 
174: 
175: 
176: Below we will compare the spectrum of the discretely tethered membrane with a simplified model called ``continuous springs'' 
177: where a membrane fluctuates in a laterally homogeneous 
178: harmonic potential of strength $\gamma$ with energy functional  
179: \begin{equation}  \label{eqn:E_with_gamma}
180: E^{\scriptscriptstyle (\gamma)}_{~}  \; \equiv \;  \frac{1}{2} 
181:     \int\limits_{0}^{L_x} \!\!   \d x  \!  \int\limits_{0}^{L_y} \!\!  \d y     \,
182:     \Big\{  \kappa  \left[\nabla^2 h(\vec{r}) \right]^2    
183:     \, + \,  \gamma \, h^2(\vec{r})
184:  \Big\}     \ \ .
185: \end{equation}   
186: The fluctuation spectrum of this system reads 
187: \begin{equation}   \label{eq:spectrum_gamma}
188:     { \langle \, |h^{~}_\vec{k}|^2 \, \rangle }^{\scriptscriptstyle (\gamma)}  
189:     \; = \; \frac{k_B T}{L_x L_y \, (\kappa \, |\vec{k}|^4 + \gamma)}  \ \ .
190: \end{equation}
191: The overall strength of the discrete and the continuous springs are comparable for 
192: $\gamma =  \frac{1}{L_x \, L_y}  \,\sum_{\alpha = 0}^{N} \, K_{\alpha}$, 
193: since these para\-meters lead to the 
194: same effective spring \mbox{constant} of the spring brush as a whole.  
195: 
196: 
197: 
198: 
199: 
200: %-------------------------------------------------------------------------------
201: \vspace{0.2cm}   
202: {\sl Fluctuation spectra.} --- 
203: We focus on a membrane attached to a \emph{quadratic} array 
204: \footnote{A \emph{hexagonal} array leads to qualitatively similar results --- up to the different symmetry~\cite{Merath}.} 
205: of equally strong ($K_\alpha \equiv K$), equidistant tethers, with $\Delta$ as tether lattice constant, 
206: in the limit $L_{x, y} \to \infty$.
207: First, the case of \emph{pinning}, i.\,e., infinitely strong springs ($K \to \infty$), is analyzed. 
208: Dimensional analysis yields the spectrum in the form 
209: \begin{equation}  \label{eq:Def.g}
210:    \left.{ \langle \, |h^{~}_{\vec{k}}|^2  \, \rangle }\right|_{\mathrm{pinned}}  
211:        \; = \;  \left.{ \langle\, |h^{~}_{\vec{0}}|^2 \,\rangle }\right|_{\mathrm{pinned}}     \cdot  g\left(k_x\Delta, \, k_y\Delta \right) 
212: \end{equation}
213: with the amplitude at the origin ($\vec{k} = \vec{0}$) given by
214: \begin{equation}  \label{eq:Def.s}
215:    \left.{ \langle \, |h^{~}_{\vec{0}}|^2  \, \rangle }\right|_{\mathrm{pinned}} 
216:            \; = \; s \cdot \frac{k_B T}{\kappa \, N} \Delta^2
217: \end{equation}
218: and a numerically determined prefactor $s \simeq 3.9\cdot 10^{-3}\,$. 
219: The scaling function $g(k_x\Delta, k_y\Delta)$ is shown in Fig.~\ref{fig:quadrat._1}. 
220: \begin{figure}[!tb]   
221: \hfil \hspace{-0.30cm} \includegraphics[scale=0.67, clip]{Fig2.eps} \hfil  
222: \vspace{-0.2cm}     
223:      \caption{\label{fig:quadrat._1}The scaling function $g(k_x\Delta, k_y\Delta)$ of the fluctuation spectrum 
224:                                     of a fluid membrane pinned by a quadratic array of infinitely strong springs. 
225:              }
226: \end{figure}
227: The fourfold symmetry of the spectrum reflects the symmetry of the tether array. 
228: This spectrum is non-monotonic with four relative maxima and four saddle points. 
229: The wavelength at the maxima is 
230: $\lambda_{\mathrm{\scriptscriptstyle max}} \simeq 1.11 \cdot 2\, \Delta\,$. 
231: Since undulation modes with a wavelength twice the distance $\Delta$ fulfill a pinning condition 
232: (i.\,e.\ no displacement at the site of the tethers), 
233: a naive guess would yield $\lambda_{\mathrm{\scriptscriptstyle max}} =  2\, \Delta\,$. 
234: For a quasi-one-dimensional system (with $k_y \equiv 0$) this would indeed be true~\cite{Merath}. 
235: However, for the full two-dimensional case, the interplay of all fluctuation modes 
236: apparent in the coupling matrix $\vec{M}$ (\ref{Def.MatrixM}) 
237: leads to deviations from this naive expectation towards a slightly larger wavelength. 
238: 
239: 
240: 
241: From this spectrum, one can derive various one-dimensional spectra shown in 
242: Fig.~\ref{fig:cut_hard}. 
243: Cuts through the origin and a maximum, and through the origin and a saddle
244: point, respectively, show the non-monotonicity. 
245: The \mbox{saddle} point appears at a (by a factor of approximately $\sqrt{2}$\,) larger wavelength than the maxima, 
246: corresponding to the larger tether distance in this direction. 
247: Furthermore, the average of the 2-d spectrum with respect to the azimuth angle is displayed. 
248: In the averaged spectrum, the wavelength of the maximum is 
249: $\lambda^{\mathrm{\scriptscriptstyle azi}}_{\mathrm{\scriptscriptstyle max}} \simeq 1.2 \cdot 2\, \Delta\,$.  
250: It is larger than the wavelength corresponding to the relative maxi\-mum in the \mbox{2-d} spectrum 
251: but smaller than the wavelength corresponding to a saddle point. 
252: 
253: 
254: \begin{figure}[!tb]   
255: \hfil \hspace{-0.4cm} \includegraphics[scale=0.70, clip]{Fig3.eps} \hfil 
256: \vspace{-0.32cm}    
257:      \caption{\label{fig:cut_hard}One-dimensional plots of the universal 2-d fluctuation spectrum $g(k_x\Delta, k_y\Delta)$ 
258:              of a membrane pinned to a quadratic array of pinning sites (cf.\ Fig.~\ref{fig:quadrat._1}). 
259:              Cuts in two main directions, through a maximum (curve 1) and a saddle point (curve 2), 
260:              of the 2-d spectrum are shown and compared to an azimuthal average (curve 3). 
261:              Curve 4 shows an azimuthal average of a spectrum for a \emph{hexagonal} arrangement of pinning sites. 
262:              The closer packed sites induce a six-fold symmetric 2-d spectrum situated below the spectrum of the quadratic system. 
263:              The constant $s$ for the hexagonal pattern, 
264:              $s^{\mathrm{(hex.)}} \simeq  2.5\cdot10^{-3}$, 
265:              is different than in the quadratic case (\ref{eq:Def.s}).  
266:              For comparability, the hexagonal spectrum is divided by the prefactor 
267:              $\langle\, |h_{\vec{0}}|^2 \,\rangle^{}_{\mathrm{pinned}}$ 
268:              of the quadratic case (\ref{eq:Def.g}). 
269:              }
270: \end{figure}
271: 
272: 
273: 
274: 
275: For \emph{finite spring constants $K$}, the fluctuation spectrum of a membrane tethered by a quadratic array 
276: can be expressed as 
277: \begin{equation}  \label{eq:Def.f}
278:    \langle \, |h^{~}_{\vec{k}}|^2  \, \rangle \; = \;   
279:          s \,  \frac{k_B T}{\kappa \, N} \Delta^2    \cdot f\left(\frac{K \, \Delta^2}{\kappa}, \, k_x\Delta, \, k_y\Delta\right) 
280: \end{equation}
281: with a scaling function $f$ that incorporates the spring elasticity. 
282: Since $f$ is continuous and $f(\infty, k_x\Delta, k_y\Delta) = g(k_x\Delta, k_y\Delta)$, 
283: for sufficiently large spring constants $K$ the non-monotonicity of the spectrum persists. 
284: In Fig.~\ref{fig:Vgl.div.K}, the wavelength-dependence of the azimuthally averaged scaling function $f$ 
285: is shown for different values of $K\,\Delta^2 / \kappa$  
286: and compared to the spectrum (\ref{eq:spectrum_gamma}) of 
287: a continuous confining harmonic potential. 
288: For weak springs, a tethered membrane behaves 
289: like a membrane with continuous confinement.        
290: For moderately strong spring constants, deviations from the continuous spectrum occur, 
291: with the spectrum of the discrete tethering
292:  being systematically larger than the continuous one. 
293: The positions of the relative maxima in the 2-d spectra depend on $K \Delta^2 / \kappa$. 
294: For infinitely strong springs the maximum of the spectrum reaches its lowest possible wavelength. 
295: For decreasing spring constants the wavelength corresponding to the relative maxima increases. 
296: Finally, below a critical value of the spring constant, 
297: $\left. {K \Delta^2 / \kappa  }\right|_\mathrm{crit} \, \simeq \, 100 $, 
298: the relative maxima disappear and the spectrum decays monotonically. 
299: 
300: 
301: 
302: \begin{figure}[!tb]
303: \hfil \hspace{-0.2cm} \includegraphics[scale=0.76, clip]{Fig4.eps} \hfil 
304: \vspace{-0.32cm}    
305:      \caption{\label{fig:Vgl.div.K}The scaling function $f(K \, \Delta^2 / \kappa, \, k_x\Delta, \, k_y\Delta)$ 
306:               for the fluctuation spectrum of a quadratically tethered
307:               membrane               for different spring stiffnesses
308:               (\ref{eq:Def.f}).                                                                                                                
309:               Azimuthal averages of the 2-d spectra are compared to results from the continuous spring model (\ref{eq:spectrum_gamma}). 
310:               For comparision, the spectra for the continuous model are scaled by the prefactor 
311:               $\langle\, |h_{\vec{0}}|^2 \,\rangle^{}_{\mathrm{pinned}}$ 
312:               of the discrete case.     
313:               Curve 1 shows the azimuthal average of the \mbox{2-d} spectrum for pinning, i.\,e.\ 
314:               infinitely strong springs.  
315:               In the comparable continuous case ($\gamma \to \infty$), the spectrum vanishes.   
316:               Curve 2 depicts the spectrum for strong continuous springs: 
317:               While the ``discrete spectrum'' is still undistinguishable from curve 1, 
318:               the ``continuous spectrum'' is located far underneath. 
319:               Curves 3 and 4 give the corresponding discrete and
320:               continuous spectrum for moderately strong springs, respectively. 
321:               For weaker springs the two cases yield nearly the same data shown in curve 5. 
322:               The spectrum of a free membrane without tethers is shown as curve 6. 
323:              }
324: \end{figure}
325: 
326: 
327: 
328: 
329: 
330: %-------------------------------------------------------------------------------
331: \vspace{0.2cm}  
332: {\sl Real-space fluctuations.} --- 
333: We now determine the average width of the membrane in real space. 
334: The height profile can be written as 
335: \begin{equation}
336:    h(\vec{r}) = 2 \, \vec{w}(\vec{r}) \cdot \vec{c} 
337: \end{equation}
338: with~\cite{Lin_Brown}
339: \begin{equation}
340:    \vec{w}(\vec{r}) \equiv ( 1, \{ \cos(\vec{q} \cdot \vec{r}) \},  \{ -\sin(\vec{q} \cdot \vec{r}) \}   ) \ \ . 
341: \end{equation}
342: Given the inverse matrix $\vec{M}^{-1}$, the mean square displacement reads 
343: \begin{eqnarray} \label{eq:MSD}
344:   \langle \, h^2(\vec{r}) \, \rangle & = & 
345:              4 \, \sum\limits_{r, r^\prime} w^{~}_r(\vec{r}) \: \langle \, c^{~}_r \: c^{~}_{r^\prime}  \rangle \: 
346:                                             w^{~}_{r^\prime}(\vec{r}) \nonumber  \\ 
347:              & = & 2 \; k_B T \; \sum\limits_{r, r^\prime} w^{~}_r(\vec{r}) \:  M^{-1}_{r, r^\prime}  \: w^{~}_{r^\prime}(\vec{r})   \ \ .
348: \end{eqnarray}
349: Typical membrane elongations are represented by the root mean square displacement (RMSD) $\sqrt{\langle  h^2(\vec{r}) \rangle}$. 
350: Only if the maximum of this RMSD is significantly smaller than the rest length of the springs, 
351: this model can be trusted quantitatively since it does not yet include the steric hindrance by the substrate. 
352: The maximum of the RMSD is a measure to estimate how far above a substrate the membrane has to be placed 
353: in order to avoid undesirable contact or adhesion of the bilayer 
354: to the substrate. 
355: 
356: 
357: 
358: For \emph{pinning} ($K \to \infty$),  we find 
359: \begin{equation}   \label{eq:Def.u}
360:    \left. { \sqrt{\langle \, h^2(\vec{r}) \, \rangle\,} }\right|_{\mathrm{pinned}} \;  
361:            \equiv \; \sqrt{\frac{k_B T}{\kappa}\,} \Delta     \cdot  u\left(\displaystyle \frac{x}{\Delta}, \frac{y}{\Delta} \right)
362: \end{equation} 
363: with a scaling function $u(x / \Delta, y / \Delta)$     
364: shown in Fig.~\ref{fig:Matratzen-Plot}. 
365: The maximum of the RMSD in the pinned case occurs in the center of the array and is given by 
366: \begin{equation}  \label{eq:max_pinned}
367:    \left. \mathrm{max}\{ \sqrt{\langle  h^2(\vec{r}) \rangle \,} \, \}\right|_{\mathrm{pinned}} \; 
368:    \equiv \; p \cdot \sqrt{\frac{k_B T}{\kappa}\,} \Delta
369: \end{equation}
370: with a constant $p \simeq 0.124\,$. 
371: 
372: 
373: For \emph{finite} spring constants $K$, the maximum of the RMSD takes the form 
374: \begin{equation}   \label{eq:Def.w}
375:    \mathrm{max}\{ \sqrt{\langle  h^2(\vec{r}) \rangle \, } \, \} \; \equiv \; p \: \sqrt{\frac{k_B T}{\kappa}\,} \Delta 
376:                              \cdot  w\left( \frac{K \Delta^2}{\kappa} \right) 
377: \end{equation} 
378: with a scaling function $w$.     
379: For weak springs, $w(K \Delta^2 / \kappa)$ approaches 
380: $\frac{1}{p \, \sqrt{8\,}}  \left( \frac{\kappa}{K \, \Delta^2}  \right)^{1/4}$, 
381: which follows from the known result 
382: \begin{equation}
383:      { \langle  h^2(\vec{r}) \rangle }^{\scriptscriptstyle (\gamma)} 
384:                =   \,  \frac{k_B T}{8 \sqrt{\kappa \, \gamma\,}}
385: \end{equation}
386: for a membrane bound in a harmonic potential~\cite{Safran_1994}. 
387: For strong springs, $w$ \mbox{approaches} 1.  
388: The cross-over between these two scaling limits occurs around $K\Delta^2 / \kappa = (p \, \sqrt{8\,})^{-4} \simeq 66.1$, 
389: where the two asymptotes intersect. 
390: 
391: 
392: 
393: \begin{figure}[!tb]  
394: \vspace{0.2cm}
395: \hfil \hspace{-0.1cm} \includegraphics[scale=0.64, clip]{Fig5.eps} \hfil 
396: \vspace{-0.1cm}   
397:     \caption{\label{fig:Matratzen-Plot}Scaling function $u(x/\Delta, y/\Delta)$ for the RMSD, $\sqrt{\langle \, h^2(\vec{r}) \, \rangle\,} $, 
398:                        of a membrane pinned at a quadratic array of sites. 
399:             } 
400: \end{figure}
401: 
402: 
403: 
404: 
405: 
406: %-------------------------------------------------------------------------------
407: \vspace{0.4cm}
408: {\sl Summarizing perspective.} --- 
409: Discrete tethering of a membrane to a substrate has a profound implication on the fluctuation spectrum. 
410: For strong enough tethers, this spectrum becomes non-monotonic with a maximum determined by the spacing of the tethers. 
411: An interpretation of such a spectrum in terms of a continuous model would 
412: require a term $- |\sigma| (\nabla h)^2$ in (\ref{eqn:E_with_gamma}) implying 
413: a negative ``surface tension''.     
414: For sufficiently weak tethers, however, the discrete tethering can indeed be replaced by a continuous harmonic confining potential. 
415: The explicit introduction of an additional regular (positive) surface tension, 
416: e.\,g.\ caused by the area constraint of a vesicle whose bound part is considered~\cite{US_PRL_1995}, 
417: into the energy (\ref{eqn:E_Helfrich+Federn}) of our model poses no problems. 
418: 
419: Our quantitative data on the mean displacements caused by these fluctuations should provide valuable hints for the experimentalists, 
420: how large the rest length $l^{~}_0$ of the pillars and tethers has to be in
421: order to avoid contact with the substrate. 
422: For a smaller $l_0$, one should 
423: include a direct interaction with the substrate 
424: which could be attractive due to van-der-Waals interaction and/or
425:  repulsive due to steric interactions. 
426: In \mbox{either} case one could include an effective potential $V(h)$ to the energy functional. 
427: Its mini\-mization will then lead to a laterally inhomogeneous ``ground state'' profile $l^{~}_0(\vec{r})$ 
428: as in studies of membrane adhesion to structured substrates~\cite{Andelman_PRE_2001}. 
429: In another extension of our model one can study the fluctuation of two almost parallel membranes connected by polymeric linkers. 
430: The fluctuation of the relative distance between the membranes, i.\,e., the peristaltic mode, 
431: is then governed by an effective Hamiltonian (\ref{eqn:E_Helfrich+Federn}) 
432: and will show the non-monotonic behavior for strong enough linkers as well. 
433: 
434: 
435: 
436: 
437: %-------------------------------------------------------------------------------
438: \begin{thebibliography}{00}
439:    \bibitem{US_AdvPhys_1997} For a review, see U.\ Seifert, Adv.\ Phys.\ {\bf 46}, 13 (1997).  
440:    \bibitem{Zilker_1987} A.\ Zilker, H.\ Engelhardt, and E.\ Sackmann, J.\ Phys.\ (Paris) {\bf 48}, 2139 (1987). 
441:    \bibitem{Raedler_1995} J.\,O.\ R\"adler, T.\,J.\ Feder, H.\,H.\ Strey, and E.\ Sackmann, Phys.\ Rev.\ E {\bf 51}, 4526 (1995).  
442:    \bibitem{Naumann_Knoll_Langmuir_2003} R.\ Naumann, S.\,M.\ Schiller, F.\ Giess, B.\ Grohe, K.\,B.\ Hartmann, I.\ K\"archer, I.\ K\"oper,
443:                                          J.\ L\"ubben, K.\ Vasilev, and W.\ Knoll, Langmuir {\bf 19}, 5435 (2003). 
444:    \bibitem{Bunjes_Knoll_Langmuir_1997} N.\ Bunjes, E.\,K.\ Schmidt, A.\ Jonczyk, F.\ Rippmann, D.\ Beyer, H.\ Ringsdorf, P.\ Graber, W.\
445:                                         Knoll, and R.\ Naumann, Langmuir {\bf 13}, 6188 (1997). 
446:    \bibitem{Tanaka_2003_Wiley} A.\ F\"ortig, R.\ Jordan, K.\ Graf, G.\ Schiavon, O.\ Purrucker, and M.\ Tanaka, 
447:                                Macromolecular \mbox{Symposia} {\bf 210}(1), 329 (2003).
448:    \bibitem{Tanaka_2004} O.\ Purrucker, A.\ F\"ortig, R.\ Jordan, and M.\ Tanaka, Chem.\ Phys.\ Chem.\ {\bf 5}, 327 (2004).
449:    \bibitem{Giess_Knoll_BiophysJ_2004} F.\ Giess, M.\,G.\ Friedrich, J.\,H.\ Heberle, R.\,L.\ Naumann, and W.\ Knoll, 
450:                                        Biophys.\ J.\ {\bf 87}, 3213 (2004). 
451:    \bibitem{Spatz_2004} J.\,P.\ Spatz, Chem.\ Phys.\ Chem.\ {\bf 5}, 383 (2004).
452:    \bibitem{Spatz_2003} W.\,H.\ Roos, A.\ Roth, J.\ Konle, H.\ Presting, E.\ \mbox{Sackmann}, and J.\,P.\ Spatz, 
453:                         Chem.\ Phys.\ Chem.\ {\bf 4}, 872 (2003). 
454:    \bibitem{Andelman_Langmuir_1999} P.\,S.\ Swain and D.\ Andelman, Langmuir {\bf 15}, 8902 (1999). 
455:    \bibitem{Andelman_PRE_2001} P.\,S.\ Swain and D.\ Andelman, Phys.\ Rev.\ E {\bf 63}, 051911 (2001). 
456:    \bibitem{Lipowsky_JPhysCondMat_2005} R.\ Lipowsky, M.\ Brinkmann, R.\ Dimova, T.\ Franke, J.\ Kierfeld, and X.\ Zhang, 
457:                                         J.\ Phys. Cond.\ Mat.\ {\bf 17} S537 (2005). 
458:    \bibitem{Gov_2003} N.\ Gov, A.\,G.\ Zilman, and S.\ Safran, Phys.\ Rev.\ Lett.\ {\bf 90}, 228101 (2003).
459:    \bibitem{Fournier_2004} J.-B.\ Fournier, D.\ Lacoste, and E.\ Rapha\"el, 
460:                            Phys.\ Rev.\ Lett.\ {\bf 92}, 018102 (2004). 
461:    \bibitem{Gov_2004} N.\ Gov and S.\,A.\ Safran, Phys.\ Rev.\ E {\bf 69}, 011101 (2004).
462:    \bibitem{Lin_Brown} L.\,C.-L.\ Lin and F.\,L.\,H.\ Brown, Biophys.\ J.\ {\bf 86} (2), 764 (2004).
463:    \bibitem{Merath} R.-J.\  Merath,
464:                      thesis, Universit\"at Stuttgart (2004). 
465:    \bibitem{Safran_1994} S.\ A.\ Safran, {\it Statistical Thermodynamics of Surfaces, Interfaces and Membranes}, 
466:                          Frontiers in Physics Vol.\ 90 (Addison-Wesley, Reading, MA, 1994). 
467:    \bibitem{US_PRL_1995} U.\ Seifert, Phys.\ Rev.\ Lett.\ {\bf 74}, 5060 (1995). 
468: \end{thebibliography}
469: 
470: \end{document}           
471:    
472: