ebook img

Point-defect haloing in curved nematic films PDF

0.25 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Point-defect haloing in curved nematic films

Point-defect haloing in curved nematic films ∗ Isaku Hasegawa Department of Applied Physics, Graduate School of Engineering, Hokkaido University, Sapporo 060-8628, Japan Hiroyuki Shima Department of Applied Physics, Graduate School of Engineering, Hokkaido University, Sapporo 060-8628, Japan and Department of Applied Mathematics 3, LaCa`N, Universitat Polite`cnica de Catalunya (UPC), Barcelona 08034, Spain (Dated: January 6, 2010) We investigate the correlation between the point disclination energies and the surface curvature 0 modulation of nematic liquid crystal membranes with a Gaussian bump geometry. Due to the 1 correlation, disclinations feel an attractive force that confines them to an annulus region, resulting 0 in a halo distribution around the top of the bump. The halo formation is a direct consequence of 2 the nonzero Gaussian curvatureof the bump that affects preferable configurations of liquid crystal n molecules around thedisclination core. a J PACSnumbers: 61.30.Jf,61.30.Hn,81.40.Lm,02.40.-k 6 I. INTRODUCTION tures. In actual systems, however, a slight physical (and ] t chemical)disturbancecausesexcessdisclinations[19–22]. f o These disclinations can travel over a surface via thermal It is generally assumed that topological defects in liq- s excitation,and they flow almostfreely if the disclination . uid crystal membranes are excluded in the ground state t densityissufficientlylowtoignoretheirinteraction. The a becausetheyincreasetheelasticfreeenergyofthesystem m to more than that in the defect-free equilibrium state. only possible force that restricts the free motion of the excess disclinations arises from a coupling between the - The situation is quite different when the liquid crystal d surface curvature and the disclination energy. The sub- phases are embedded in a curved surface. In a curved n stratecurvaturebreaksthetranslationalsymmetryofthe membranewithnonzeroGaussiancurvature,theappear- o molecularconfigurationinthevicinityofthedisclination ance of topological defects can be favorable or even in- c core, which may cause a shift in the disclination energy. [ evitable because of geometrical and/or topological con- As a result,the energeticallypreferentialpositions ofex- straints. A typical example is a disclination-containing 3 cess disclinations are believedto be tunable by imposing equilibrium state in a nematic liquid crystal that is con- v an appropriate curvature modulation. strained in a spherical surface [1]. The defect structure 8 In this article, we study the energies of disclinations 2 insuchanematicsphericalshellhasrecentlybeenrecon- in a nematic thin film assigned on a Gaussian bump. It 2 sideredboththeoretically[2–5]andexperimentally[6,7], 4 and these investigations revealed richer physical proper- is found that disclinations with a +1 charge prefer posi- 2. ties thanthose thathaveinitially been suggested. Inad- tioningatinflectionpointsofthebump,implyingthefor- mation of a “halo,”i.e.,an annulus distribution of excess 1 dition to the nematic cases, the smectic phases of liquid disclinations around the top of the bump. In addition, 9 crystalfilms[8,9]andcolumnarphasesofablockcopoly- 0 mer assembly [10] have also exhibited an interplay be- disclinationswitha 1chargearefoundtoexhibitdouble − : concentric halo structures, provided the bump height is v tweendisclinationstructuresandthe geometry/topology relatively large. These halo formations are consequences i of the constraint surface. In particular, smectic ordering X ofthe non-trivialcouplingbetweenthe disclinationener- on a sinusoidal substrate has indicated a direct coupling r between the substrate curvature and the preferential lo- gies and the surface curvature modulation. a cationsofdisclinationsinequilibriumstates[9]. Thecou- pling is purely a geometric curvature effect in the sense that the system involves no topological requirement for II. MODEL the existence of disclination. The geometric curvature alsostronglyinfluencestheorientationalorderofthespin lattice models defined on curved surfaces [11–18]. We consider nematic liquid crystals confined in a thin curved membrane with thickness h, in which the Many intriguing results have been obtained with re- moleculesalignparalleltothe tangentialdirectionsinde- gardtothegloballowest-energystateonacurvedsurface pendently of the depth in the membrane. In line with with a minimum number of disclinations, which should the continuum theory [19, 23], the elastic free energy of be relevantto liquid crystalphases at very low tempera- themolecularensemblewithinthevolumeAhisgivenby F =h dA K1(D n)2+ K3(D n)2 , (1) ∗Electronicaddress: [email protected] d Z (cid:26) 2 · 2 × (cid:27) A 2 where the unit vector n (called the director field) de- shape when it is placed on a curved surface, whereas it scribes the local molecular alignment. D is a vector op- is an exact circle when placed on a flat plane. eratorwhosecomponentsD arethecovariantderivative Forlateruse,weintroduceanewpolarcoordinatesys- µ [24]onthecurvedsurface. K1andK3aretheelasticcon- tem(l,θ)ontheflatplanesuchthatthecoordinateorigin stants associated with splay and bend distortions of the (i.e., l =0) locates at the point of interaction of the flat molecular configuration, respectively [25]. plane and the vertical line penetrating the disclination Supposethatthedirectorlieswithinathincurvedshell core (see Fig.1). Then, the basis vectors a1 ∂R/∂l ≡ having a Gaussian bump geometry. Points on the bump and a2 ∂R/∂θ span the curved surface and allow us are represented by a three-dimensional vector R(r,φ) to write≡n = niai (i = 1,2), where ni and ai depend given by on l and θ (See Appendix B). The director orientation around the core is given by the angle ψ between n and r2 a1. It is natural to assume that R(r,φ)=rcosφi+rsinφj+vexp k, (2) (cid:18)−2r2(cid:19) 0 ψ = (s 1)θ+ψ0, (3) − where i, j, and k are basis vectors of the Cartesian co- where s denotes the disclination charge that quantifies ordinate system, and r and φ are the polar coordinates how many times n rotates about the core. When s = assignedontheunderlyingflatplane(seeFig.1). Thege- +1, for instance, all n contained within the disclination ometric deviation of a bump from a flat plane is charac- terized by a dimensionless parameter α v/r0 that rep- area A make the same angle ψ0 with respect to a1 (see ≡ Fig.3(a)); on the other hand, when s = 1, the angle resents the ratio of the height to the width of the bump. − increases proportionally to θ (see Fig.4(a)). Note that the constant r0 in the last term in Eq.(2) de- In the calculation, we consider disclinations with two termines the points of inflection of the Gaussian bump, i.e., across which the sign of (∂2R/∂r2) k changes. We charges, s = 1. The disclination area A and the mem- · brane thickn±ess h are set to be A = 400π µm2 and demonstrate below that r0 plays a prominent role in de- h=4µmbyreferringtothematerialconstantsofactual scribing the correlation between the disclination energy liquid crystal membranes [26]. The values of the elastic and the substrate curvature. Let us locate a disclination core on a point R = constantsK1andK3 arethoseof4-methoxybenzylidene- ′ 4-butylaniline (MBBA), p-azoxyanisole (PAA), and 4- R(r ,φ ), where φ is anarbitraryconstantwithout loss c c c ′ pentyl-4-cyanobiphenyl(5CB) presented in Table I [25]. of generality. The disclination size is sufficiently smaller than the bump width, and the perimeter of the discli- Theconstantψ0 inEq.(3)isfixedtobezerowithoutloss of generality,bearing in mind the fact that other choices nation is defined by the locus of points whose distances fromthe coreare equalin the sense ofgeodesic distance; ofψ0 leadtonoessentialchangeinthe conclusionofthis work. therefore, the perimeter slightly deviates from a circular III. RESULT Figure 2 shows plots of the disclination energy F of d MBBAasafunctionofthecorepositionr . Thevertical c axis shows the difference in F between the bump and d the flat plane, i.e., F (α = 0) F (α = 0). The discli- d d 6 − nation charge s is s = +1 in (a) and 1 in (b), and the − parameter α ranges from α = 0 to 1.2 in both (a) and (b). Fors=+1,Fd takesthe minimumatrc =r0 forall values ofα, indicating thata disclinationis stable atthe inflection point of the bump r =r0 independently of the bump height. On the other hand, the plot for s = 1 − exhibits a transitionfroma single downwardpeak struc- ture(α<0.7)intoadouble-wellstructure(α>0.7)such TABLE I: Elastic constants of MBBA, PAA, and 5CB re- ported in Ref.[25] Elastic constants MBBA PAA 5CB FIG.1: SchematicofaGaussianbumpandmathematicalno- ◦ ◦ ◦ (pN) (25 C) (122 C) (26 C) tationsusedinthepaper. Thethickarcdepictedonthebump K1 6 6.9 6.2 isaprojectionofthethicklinesegmentontheunderlyingflat K3 7.5 11.9 8.2 plane. 3 FIG. 2: Disclination energy in deformed MBBA films as a function of the core position rc. The parameter α characterizing thebumpheight ranges from 0to1.2 (asindicated). (a) Energy of apositively charged disclination (s=+1). Themagnitude of a downward peak at rc = r0 monotonically increases with α. (b) Energy of a negatively charged disclination (s = −1). Transition from a single-well to a double-well structureis observed at α=0.7. that r0 becomes unstable. The contrasting behaviors of Figures 3(a) and 3(b) illustrate how the configuration F fors=+1and 1areattributedtosurfacecurvature varies with a shift in the core position r from r = 0 d c c − effects on the director configuration close to the core, as (a) to rc = r0 (b). In Fig.3(a), n aligns radially around will be showninSec.IV. For PAAand5CB,we haveob- thetopofthebump,exhibitingrotationalsymmetrysuch tained similar profiles of Fd exhibiting a downwardpeak thatn a1andn a2 =0ateverypoint. Thissymmetry k · at rc = r0 for s = +1 and a double-well structure for breaks down for rc > 0. By increasing rc, the perimeter s = 1, whereas the peak heights differ slightly from ofthedisclinationdistortsslightlyandbecomeselliptical, − those shown in Fig.2 because of quantitative differences andthen, the relationshipn a2 =0nolongerholds(see · in K1 and K3. We have also confirmed that the peak Fig.3(b)), except along the four specific line segments positions (both downward and upward) are independent (θ = 0, π, π, 3π). The most important observation in 2 2 of our choice of ψ0 defined by Eq.(3). Fig.3(b) is a suppression of the splay deformation in the We emphasize that the energy scales of the potential shadowed region. The area of the corresponding region wellsshowninFigs.3(a)and3(b)exceedthe thermalen- ergy scale at ambient temperature. For instance, the depth of the single well for s =+1 and α=1.0 given in Fig.3(a) is approximately 5.7 10−18J; this is of the or- der of 103 k T at T =25◦C. ×This fact suggeststhat the B potential well at r = r0 overwhelms the thermal excita- tionthatcausesarandommovementofdisclinationsover thesurface. Asaresult,anensembleoftrappeddisclina- tions forms a halo with radius r around the top of the c Gaussianbump, providedthatthedisclinationdensity is sufficiently low to ignore their repulsive interaction. The same scenario applies to disclinations with a charge of 1, implying the formation of doubly consentric halos − observed for α>1.0. IV. DISCUSSIONS FIG. 3: Configurations of the director n around the core of a disclination with a charge of +1. With a shift in the core The physical origin of a downwardpeak in the plot of position from rc =0 (a) to rc =r0 (b), the radial symmetry F for s=+1 (see Fig.2(a)) is a curvature-induced shift ofthedirectorconfiguration breaksdownsuchthatthesplay d deformation in the shadowed region in (b) is suppressed. inthedirectorconfigurationaroundthedisclinationcore. 4 FIG. 4: Configurations of n in a disclination with a charge of −1. The core positions rc are rc =0 in (a), 0<rc <r0 in (b), and rc =r0 in (c). In thelight- and dark-shadowed region, thesplay deformation is suppressed and enhanced,respectively. ismaximizedatrc =r0,asaresultofwhichadownward ceeds the enhancement, and it increases in the opposite peak Fd arises at rc =r0 at the expense of a slightsplay situation. The degree of splay enhancement in the dark enhancement within a limited region around θ = π2. regions is maximized at rc = r0; this explains why r0 is ± Parallelargumentstotheprecedingparagraphaccount the most unstable point for s= 1 disclinations. − for the double-well structure of F for s = 1 that is d − We have numerically confirmed that the splay term shown in Fig.2(b). Figures 4(a)–(c) show the variation of the n configuration under s = 1 and α > 0.7, each K1(D n)2 in Eq.(1)playsa dominantrolein describing · − the r - (and α-) dependences of F . Figures 5(a)–(b) figure corresponding to the core position: r = 0 (a), c d 0 < rc < r0 (b), and rc = r0 (c). The dircector con- show contour plots of the splay term K1(D·n)2 within thedisclinationareaofs= 1. Weobservethatwithan figurationin(a)exhibitsafour-foldrotationalsymmetry − increase in r , the local splay energy decreases at points thatbreaksdownasr increases. Symmetrybreakingin- c c near θ =π/2 or 3π/2 but increases near θ =0 or π; this ducesasuppressionofthesplaydeformationinthelight- is in agreement with our previous discussions. A more shadowedregionsshowninFig.4(b)and4(c)inamanner detailed version of the study that covers the variations similar to the case of s = +1. In contrast, the splay de- formationinthedark-shadowedregionsisenhancedwith of the disclination size A and the initial angle ψ0 will be presented elsewhere. an increase in r . These two competing effects result in c the double-well structure of F for s = 1 cases. That d − is, F decreases with r when the suppression effect ex- d c V. SUMMARY In conclusion, we investigated the elastic energy of disclinations that arise in a curved nematic membrane with a Gaussian bump geometry. When α < 0.7, the disclination energy F exhibits a downward peak at r = d r0 independently of the sign of the disclination charge, suggestinga curvature-drivenhalo formationof disclina- tions around the top of the bump. Furthermore, when α > 0.7, F for s = 1 exhibits a double-well structure d − FIG.5: Contourplotsof thesplay termK1(D·n)2 of discli- in which the point r = r0 becomes unstable in contrast nations with s = −1 placed at rc = 0 (a) and at rc = r0 to the case of s=+1; this implies doubly concentric ha- (b). Light (dark) gray indicates a small (large) value of the los peculiar to negatively charged disclinations. In both term at the position. With an increase in rc, the splay term cases of s = 1, the well depth overcomes the thermal ± decreases around the line segments of θ =π/2 and 3π/2 but energy scale at room temperature, suggesting the exper- increases around those of θ=0 and π. imental feasibility of our theoretical predictions. 5 Acknowledgements have 1 0 haWraefowrotuhlde flriukeitftuol dthisacnukssfioonrsK. .IHYaiksutbhoanaknfdulHfo.rOthrie- gµν =(cid:18)0 l2 (cid:19), (A4) financial support from the 21st Century COE Program “Topological Science and Technology”and support from as expected. theJapanSocietyforthePromotionofScienceforYoung scientists. HSacknowledgesM.Arroyoforassistingwith the work carriedout of UPC. This work is supported by Appendix B: Contravariant components ni the Kazima foundation and a Grant-in-Aid for Scientific ResearchfromtheJapanMinistryofEducation,Science, We presentthe derivationof the contravariantcompo- Sports and Culture. nent ni of the vector n=nia in terms of a . First, the i i relative angle between n and a1 is denoted as ψ. Sim- Appendix A: Metric tensor gµν =aµ·aν ilarly, the angle between a1 and a2 is denoted as Ψ. It follows from Eq.(A3) that In this Appendix, we derive the metric tensor gµν = a1 a2 g12 aµ·aν intermsofthecovariantbasisvectorsa1 =∂R/∂l cosΨ = a1·a2 = √g11√g22, (B1) anda2 =∂R/∂θintroducedinSec. II. Westartwiththe | || | basis vectors e1 = ∂R(r,φ)/∂r and e2 = ∂R(r,φ)/∂φ sinΨ = |a1×a2| = √g , (B2) associatedwiththe polarcoordinatesystem(r,φ)onthe a1 a2 √g11√g22 | || | flatplane. Intermsofe1 ande2,themetrictensorg˜µν = eµ eν becomes where g =det[gµν]. We thus have · f(r) 0 α2r2 r2 1 g˜µν =(cid:18) 0 r2 (cid:19), f(r)=1+ r02 exp(cid:18)−r02(cid:19).(A1) cos(Ψ−ψ) = √g11g22(g12cosψ+√gsinψ),(B3) We use a coordinate transformation from (r,φ) to (l,θ) whichleadsustotheexplicitformsofthecovariantcom- by shifting the coordinate origin from r = 0 to r = rc. ponents ni =n ai such as This is accomplishedby multiplying both sidesofg˜ by · µν the Jacobian β defined by n1 = n a1 cosψ =√g11cosψ, (B4) | || | ∂r/∂l ∂φ/∂l β = , (A2) n2 = n a2 cos(Ψ ψ) (cid:18)∂r/∂θ ∂φ/∂θ (cid:19) | 1|| | − = (g12cosψ+√gsinψ). (B5) √g11 whose elements can be computed by using the follow- ing relationships: r = r2+2r lcosθ+l2 and φ = −1 c c Finally, the formulas nj = n gij with gij = [g ]−1 is tan lsinθ/(rc+lcosθ)p. Finally we obtain i ij { } employed to obtain g = a a =βg˜ (β)−1 µν µ ν µν · 1 1 g12 n = (cosψ sinψ), (B6) = r12 (cid:18)lpfq((r1)p2+f(qr2)) l2lp(fq((1r)−q2f+(rp)2))(cid:19),(A3) √g11 − √g − 2 g11 n = sinψ. (B7) r g where p = r cosθ + l and q = r sinθ. In particular, c c when α = 0 (i.e., a flat plane), f(r) 1, and thus, we ≡ [1] T. C. Lubensky and J. Prost, J. Phys. II France 2, 371 M. M´arquez, D. R. Nelson and D. A. Weitz, Phys. Rev. (1992). Lett. 99, 157801 (2007). [2] D.R. Nelson, Nano Lett. 2,1125 (2002). [7] T. Lopez-Leon and A. Fernandez-Nieves, Phys. Rev. E [3] V. Vitelli and D. R. Nelson, Phys. Rev. E 74, 021711 79, 021707 (2009). (2006). [8] X. Xing, Phys.Rev. Lett.101, 147801 (2008). [4] G.SkaˇcejandC.Zannoni,Phys.Rev.Lett.100, 197802 [9] L. R. G´omez and D. A. Vega, Phys. Rev. E 79, 031701 (2008). (2009). [5] H. Shin, M. J. Bowick and X. Xing, Phys. Rev. Lett. [10] C. D. Santangelo, V. Vitelli, R. D. Kamien and D. R. 101, 037802 (2008). Nelson, Phys.Rev.Lett. 99, 017801 (2007). [6] A.Ferna´ndez-Nieves,V.Vitelli,A.S.Utada,D.R.Link, [11] H. Shima and Y. Sakaniwa, J. Phys. A: Mech. Gen. 39, 6 4921 (2006). [20] N. D. Mermin, Rev.Mod. Phys.51, 591 (1979). [12] H.ShimaandY.Sakaniwa,J.Stat.Mech.(2006)P08017. [21] A. E. Romanov, Eur. J. Mech. A-Solid 22, 727 (2003). [13] R. Dandoloff and A. Saxena, Phys. Lett. A 358, 421 [22] M. Kleman and J. Freidel, Rev. Mod. Phys. 80, 61 (2006). (2008). [14] I. Hasegawa, Y. Sakaniwa and H. Shima, Surf. Sci. 601, [23] C. Y. Young, R. Pindak, N. A. Clark and R. B. Meyer, 5232 (2007). Phys. Rev.Lett. 40, 773 (1978). [15] S.K.Baek,P.MinnhagenandB.J.Kim,Europhys.Lett. [24] H. Shima and T. Nakayama, Higher Mathematics for 79, 26002 (2007). Physics and Engineering (Springer-Verlag, 2010), in [16] V.L.Carvalho-Santos,A.R.Moura,W.A.Moura-Melo press. and A. R.Pereira, Phys. Rev.B 77, 134450 (2008). [25] I.W.Stewart,TheStaticandDynamicContinuumThe- [17] S. K. Baek, P. Minnhagen and B. J. Kim, Phys. Rev. E ory of Liquid Crystals (Taylor & Francis, London and 79, 011124 (2009). New York,2004). [18] Y. Sakaniwa and H. Shima, Phys. Rev. E 80, 021103 [26] R. R. Shah and N. L. Abbott, J. Phys. Chem. B 105, (2009). 4936 (2001). [19] P.G.deGennes,ThePhysicsofLiquidCrystals(Claren- don Press, Oxford, 1974).

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.