Jpn. J. nppl. Phys. Vol. 39 (2000) Pa. :2 t?—122o
`Part I. No. 3A, March Ztlflt)
`©2t]lltl Publication Board, .lt'1].‘.‘tlIll.‘.5U Joumal offitpplictl Physics
`
`Control of the Anchoring Energy of Rubbed Polyimide Layers
`
`by lrradiation with Depolarized UV-Light
`
`Denis ANoaii;N1<o'-3, Yuriy Kutuoz‘ . Michinori t~Itstttt~:.»tw;t3*-‘, Yuriy Rnzultcovl-3 and John L. Wt=.sT""
`'.’n.trn'orc’ ofPl‘I',v.s‘i'tLs‘. 4-I5 l“'m.rpe:'t Mm-lrl. K_t't'v.
`.7'5.3.'t‘i5-‘J. Ul'rrr.riot=
`1H. H. Will.t Pla_}.\;1'('.t Lohoiwnry. [.r':ri1:m:-;r'r_t'c.fBJ'lZrrm‘. Royal Fort. Il*na’:ill' .-lvennre. Bristol 8.5%‘ ITL, UK
`3LEqtr1'z.t’['r_ie'trrl i'li.t'n'lI.i.re. it’:-or Stare U.I1ioer,vr'.'_tt Kcm‘. (Jilrin. 44342 USA
`“T.nrkt.-lid R(\\’ll'£il‘(.’.li Lrrl'tm'm‘or}-'. JSR C'nr_:mr:rt‘imr. 25 iifiwrkigtroktr. T'.w.lml'm, lbomki 3[l_'i—tJ.‘t'tH', Joplin:
`[Received Atigust G, 1999; accepted for publication December I5. I999)
`
`Exposure of rubbing polyirnide (Pi) film to depolarized ultraviolet (UV) light suppressed the offer.-tive anchoring energy of
`liquid crystal {LC} with aligning surface. Polarized light changed the orientations] distribution of PI molecules obtained by
`rubbing by changing both the anchoring energy and easy axis direction. These results Show that ultraviolet exposure can be
`effectively used to control anchoring parameters.
`KEYWORDS:
`liquid crystals. photoaligrtment. anchoring energy, polyimide. UV light
`
`1.
`
`Introduction
`
`Thin polyirnide (Fl) films are the most commonly em-
`ployed liquid crystal (LC) alignment layers. They have good
`thermal stability and provide stable planar and tilted LC align-
`ment. Two techniques are utilized to produce LC alignment
`on P1 films. Standard method of rubbing PI films"-"l changes
`the topography of the layer and induces an anisotropic orien-
`tation of polymer chains along the rubbing direction.“ Con-
`ventional wisdotn postulates that orientation of the polymer
`chains results in epitaxial alignment of the liquid crysta|.‘”l
`More recently l-lascgawa er of.“ Showed that irradiation of
`polyimidcs with polarized UV light produces planar align-
`mcnt of LC with the orientation of the easy axis. 8, per-
`pendicular to the polarization vector, E. They assumed that
`irreversible chemical reaction caused by linearly polarized
`light results in an anisotropy of PI surface and in align-
`ment of LC parallel to the easy axis. A theoretical model
`of LC photoalignment on P1 films was proposed by Johnson
`er of." In this model the alignment arises from the angular
`dependence of'tl1e probability of photoreuctiort. The rcsu[t—
`ing anisotropy of the photosensitive bond distribution results
`in an ani.~;on'opic interaction with LC. which induces an easy
`anus.
`
`Kim er oi.“-°l believed that the alignment in polyimid-cs
`is produced by photo-reaction of the C(O)=N bonds of the
`irnidc rings. UV irradiation selectively dissociates thcsc pho-
`tosensitive mocitics.
`In the case of the initially isotropic
`PI films, polarized exposure produces an anisotropic angular
`distribution of disassociated and ttnrcactcd imidc fragments.
`This theory is supported by the production of both UV-visible
`and infrared dichroisrn in the polarized UV exposed Iilrns. We
`postulate that interaction between the LC molecules and the
`unrcactcd moieties is stronger explaining why the director of
`the LC aligns perpendicular to the polarization of UV light.
`Kim at of.“ demonstrated that irradiation ofrubbed Pl sur-
`faces with polarized UV light can be used to fine-tune the
`LC alignment by adjusting the angular distribution of the Pl
`aligning fragments. Moreover they found that irradiation of
`the 1''] surface with ttnpolarized UV light destroys the rubbed-
`induccd alignmcrtt.
`In this paper we show that irradiation of rubbed PI layers
`I-I-mail address:
`
`‘Author to whom correspondence should be addressed.
`johnwcsl@;scorpio.kcn1.ctiu
`
`with depolarized UV light can be used to control the anchor-
`ing energy of LC on the surface. Using both polarized and
`unpolarizcd UV exposure we can prcciscly control the align-
`ment direction and anchoring energy and determine the dis-
`tribution of the reacted polymer.
`
`2. Light—induced Change in Anchoring Energy of
`Rubbed Pl Surface
`
`We have modeled the alignment of liquid crystal on a
`photo—e)tposcd.
`rubbed Pl surface. We assume the sur-
`facc consists of long polymer chains. which contain axially
`aligned blocks with photosensitive groups. These blocks
`align the LC rnolcculcs parallel to their anisotropy axis. UV
`light irradiation leads to decomposition of the photosensitive
`fragments in the blocks. As a result, the LC aligns with the
`unrcactcd blocks and therefore perpendicular to the irradia-
`tion polarization direction.
`The initial angular distribution of the pliotoscrtsitivc blocks
`is dcterniitlcd by the Pl treatment. In the case of a layer made
`by spin-coating the precursor polyamit: acid followed by ther-
`mal imidization. the angular distribution of the blocks is as-
`sumed to be isotropic i11 the azimuthal plane. Rubbing of the
`PI films results in anisotropic distribution of pltotoscnsilivc
`blocks, which can be described as a Gaussian distribution?’
`-
`2
`
`Noi¢n=Nucxp[-.];(¢ m¢")l
`
`(I)
`
`Where No is the concentration oi‘P[ blocks. go, is the direction
`of robbing coinciding with the direction of the easy axis on
`the 1'‘! surface. and u.-' is the width ofthc blocks distribution.
`
`We assume that the PI polymer clients do not change ori-
`entation aft-er light-induced dccomposition.'"’ Tltcrcfore, ac-
`cording to the model proposed by Johnson er off‘ the angular
`distribution N (git. I} of undamaged blocks can be found from
`the equation
`
`HNI90. 1}
`T
`
`=-flU;yE;E:Niqj, fl
`
`where n',-__.- = o'J_§,-J,- + n(,l,-Q is a tensor of light absorption. 1'
`is a unit vector along the particular irnidc bond. 13, are com-
`ponents of the light vector E . or is the rate of photochemical
`dissociation. Solving eqn. ('2) for the case of arbitrary polar-
`ized light, we obtain the angular distribution of the undam-
`aged blocks in the form
`
`Page 1 of 4
`
`Tianma Exhibit 1022
`
`l2|'i‘
`
`Page 1 of 4
`
`Tianma Exhibit 1022
`
`

`
`12155
`
`Jpn. J. Appl. Phys. Vol. 39 t_2fitlo‘.t Pt. I. No. JA
`
`D. mtoIuF.r~'t<o oi ml.
`
`Nlril. I) = Ngttbl expl—e'Ir|:
`= No(¢>) exp { —ori't I:
`
`_)
`,)
`
`0'" +01
`
`J.
`
`CI“-I-O‘
`
`-1- gll-",:|SlI12(iJ + £3 cos2¢)]}
`+ ‘%(l —e2_l'” cos2(¢2—t{.-}:”
`
`{3}
`
`is the intensity of UV
`1’
`Here it}, are the Stokes parameters,
`light. rm, is the exposure time, ti! is the angle that the major
`axis of the polarization ellipse makes with the .r axis in the
`plane of the Pl surfaces, e is the ellipticity of the incident
`light.
`Analyzing expression {3} we see. that elliplically or lin-
`early polarized light changes the distribution function N (to. r)
`in a cornplex way. This may result in changes of both the
`easy axis direction and anchoring energy on the aligning sur-
`face.”-"’ In contrast, circularly polarized light (.5. = E; =
`0. E3 2:
`|)or completely depolarized light (t3 = E; = .53 = 0)
`does not change the angular distribution function of polymer
`fragtnents. Rather,
`this type of exposure only changes the
`number of undamaged blocks aligned in a given direction.
`Hence.
`the orientation of the distribution function is fixed
`while the concentration of the undamaged blocks changes.
`One can see from eq. (3) that it decreases exponentially with
`the exposure dose:
`
`Nlqfutl = N9[¢i exp (-org“ ‘:cu ft)
`
`(4)
`
`To find out how the change in the microscopic distribution
`of the polymer chains influences the macroscopic properties
`ofthc interface (the value of anchoring energy and direction
`of easy orientation axis) we assume that the interaction poten-
`tial bcrwcen the LC and polymer blocks is
`
`Ultpu. ti’) = C sinzttpu — 95),
`
`(5)
`
`where (flu is the director angle on the polymer surface with
`respect to the x axis and angle to defines the axis of anisotropy
`of a polymer block to the same axis. In this case the surface
`free energy density of LC can be written as fol lows
`
`,:{:ittrfi§9D»fl Z I Ulrpnu ¢iNl¢»u’ld¢'-
`-rt
`
`(Gil
`
`integrating (6) we itnlnediately obtain the corresponding
`surface torque
`.
`1
`.
`(7)
`not = 3_fs..rr;'<'lion = EWU. ml stnfilfitr. ml — trail
`Where WU. to) is the surface anchoring energy,
`,b’(r. wl
`defines the direction of the easy orientation axis.
`For depolarized or circularly polarized light the easy axis
`direction remains lixcd and coincides with the rubbing direc-
`tion. The anchoring energy decrcascs exponentially with the
`exposure doze
`
`Wit‘. to) = Wutntl exp
`
`{8)
`
`the anchoring energy of unirradiated Pl,
`is
`Here I41;
`1' = 2/o:H_t:r" + try) is the characteristic time ofthc dissocia-
`tion of imide bonds.
`
`Thus. treatrncttt of the Pl surface with depolarized light
`changes the strength of the easy orientation axis but does not
`affect its direction. 'l‘herefore_. one can control al'1Cl'lGI'll1g pa-
`
`Page 2 of 4
`
`the easy axis direction can be ad-
`ramcters independently:
`justed during the rubbing process. Then the fine-tuning of the
`anchoring energy can be achieved by exposure with depolar-
`ized light.
`
`3. Experiment
`
`We tested the above theory using a polyimide with the
`chemical formula shown in Figure 1. The chemical struc-
`ture is consistent with the theory developed above. The poly-
`mer consists of axially aligned blocks of relatively photo-
`stablc benzene and the photosensitive cyclobutane tetracar-
`hoxylic dianhydridc groups. As was shown in our previ-
`ous paper.'3'”'
`these groups are easily decomposed under
`UV—light exposure by cyclobutane ring cleavage. We found
`that polarized irradiation produced planar alignment of the
`LC perpettdicnlar to the polarization of UV light.” As pos-
`tulated above, the LC aligns along the axis of rernainirtg Pl
`clients.
`
`The precursor polyamic acid films were synthesized from
`the reaction between tetracarboxylic dianhydride and di-
`arninc.
`‘l'hc polyamic films were deposited by first spin-
`coating dilute solutions of the polyatnic acid on glass sub-
`strates covered with ITO electrodes. The resulting polyomic
`films were imidized by curing at 250°C for I h. The thickness
`ofthe resulting PI film was about 5 mm. Pl films were rttbbed
`in one direction. The rubbed substrates were exposed with
`UV light produced by a 450W Xenon lamp (Oriel, model
`6266}. The light was incident normal to the surface. The
`UV light was polarized with a surface film polarizer (Oriel,
`model 27320) whose effective range was between 230 nm and
`770 nm. The power of UV light alter passing through the po-
`larizer was about how at 254 11111.. The intensity ol‘UV light
`in the plane oftlte film was in the range l—ltl(J mWx‘cm3_
`The anchoring energy on the inadiatcd PI surface was stod-
`icd in a twist cell filled with a 4’-n-pentyl-4-cyanobiphcnyl
`(K15, Merle). Thickness of the cell, it = 20,ttrn, was con-
`trolled by cylindrical spacers. The reference substrate used
`an unirradiated rubbed Pl layer producing strong planar an-
`choring of the LC. The tested substrate was covered with
`P1. that have t.lifl°ercnt strip—li!<c regions irradiated for differ-
`ent times. The angle between the easy axis on the re fereitce
`substrate, em; {given by the rubbing direction) and the light-
`induced easy axis on the tested substrate, em. (given by po-
`larization ofthe light) was go, 2 57°. The cell was filled with
`LC in the isotropic state [T = 100°C) and slowly cooled to
`room temperature to avoid possible flow alignment.
`
`Fig.
`
`I. Chemicals’[ructttrcoi'polyitniL|c mrttertul.
`
`Page 2 of 4
`
`

`
`Jpn. J. Appl. Phys. Vol. 39 (2000) PL l, No_LlA
`
`D. Annninwxo emf.
`
`121a
`
`
`
`3-B:1:or
`E3t-
`
`Reference — rubbed PI
`Tested — rubbed F'l + UV
`
`depolarized light enen‘:1
`
`Fig. 2. LC cell in crossed polarizers. The directions ofrubbing on the ref-
`erence and tested substrates are depicted with arrows. One can see the
`stripes. which brightness decrease with time that corresponds to the de-
`crease of the twist angle.
`
`0
`
`1 D
`
`20
`
`30
`
`40
`
`50
`
`Exposure time, min
`
`Fig. 3
`
`T\.visl angle vs exposure time. ])epotarized light.
`
`The birefringence of I-115 is rather high, An 2 0.18. Cor-
`respontlingly, for the chosen experimental conditions, the di-
`mensionless ratio 2nAnL {A >> 1 and light beam propagates
`in the cell in the adiabatic (Mauguin) regime, i.e. the polar-
`ization follows the deviation of the director in the cell”) This
`allows the orientation of the director on the tested surface to
`
`be determined using polarizing microscope.
`The rubbing direction was set parallel to the polarizer axis.
`The analyzer was rotated to obtain the rninimutn output in
`light intensity. In this position the angle between analyzer and
`polarizer axes corresponds to the twist angle, gig, between the
`director on the reference and tested surfaces.
`
`We observed that the angle (pg. was equal to Q5, 2 57° in the
`unexposed regions of the tested surface and monotonically
`decreased with increase of the exposure time (Fig- 2 and 3).
`Uniform. homogeneous alignment was observed in the cell in
`the whole range of the exposure times. Finally, a good planar
`structure was achieved in the eell.”’
`The monotonic decrease of the director twist angle demon-
`strates a decrease in the anchoring energy with an increase in
`the exposure time. The value of the angle gag can be found
`from the balance of surface and bulk elastic torques, which,
`according to eq. (7) has the form
`I
`ta: + is sin Ztivo - ta) = 0.
`whereé is the anchoring parameter, E = WL,Hx’3;, K3; is the
`twist elastic constant.
`
`('9)
`
`The numerical solution to this equation for the experi-
`mental data gon and parameters K3; = 3.6 X lD"7erg;’cm,
`L = 20 pm is shown in Fig. 4. The exponential fitting curve
`is presented in the same figure with a dashed line. It is seen
`that the decrease in the anchoring energy is well described
`by the exponential function, demonstrating the validity of the
`
`20
`
`3D
`
`40
`
`50
`
`Exposure time. l‘. min
`
`light.
`Fig. 4. Anchoring parameter vs the exposure time; clcpolariactl
`tixperirriental data was calculated using eq. (9). Dotted line presents
`the exponential
`fit with K33 = 3.5 x In lergrcm, L = Ztliim.
`M. ‘v 6 X 10” erg/cn11,zmd r = 6.]; lrrtil1_
`
`Johnson model for our PI. The fitting of the dependence of
`W(tc,.,,) with exposure time allows evaluation of the anchor-
`ing energy W9 ~ 6 x l(}‘3 ergfcml and the characteristic rate
`r: = ?.fcrI(o'" + 01) = (6:|: 1) min of P1 decomposition. The
`knowledge of the values Wu and r allows measurement of the
`width to ofthe distribution profile N(w. t).
`The angular distribution function, N (to. I), for the ease of
`linearly polarized light is given by the formula
`
`NW}. fl :2 Nuttfilcxp [——r.r.’r,,q,l:
`
`“It +01
`
`2
`
`+ ?cos2tt35~ III)“
`
`('10)
`
`Page 3 of 4
`
`11C
`
`3to1:.
`an
`
`E'
`
`5J:
`DE
`
`5.9
`
`Eto
`
`E9
`
`Page 3 of 4
`
`

`
`1220
`
`JpiI.J. Appl. Phys. Vol. 39 (zoom Pl. 1. No. 3n
`
`D. Aunniemto oi oi.
`
`Both the concentration of undamaged bonds and the maxi-
`mum of their angular distribution changes during irradiation.
`As a result, the anchoring energy decreases. The easy orien-
`tation axis rotates perpendicular to the UV light polarization
`direction.
`The orientation of the director in a cell can be found from
`
`the balance of the surface and bulk torques
`
`K.7.‘Pilr”- — l-'5urfl(90- 3-71’) E 0
`
`(11)
`
`This equation can be solved numerically using the param-
`eicrs IV“. and r. determirlcd above and the directions :11. and
`ti; which are given by the experimental geometry. The ex-
`perimental data for the angle gag. which we observed in the
`combined cell with initial planar alignment (oi, = (J) and the
`angle between the direction of rubbing and UV light polar-
`ization 1}: = 45“ are presented in Fig. 5. The dashed line
`represents the fit of the data to the cq. (1 1] with w = 0.5, an-
`choring energy W" = 1.8 x l{i'3 erglcmz and characteristic
`time 1 = 6 min.
`This is a sensitive method for the detennination of the dis-
`
`tribution width. The dotted lines in Fig. 5 are plotted for
`to = 0.4 and 0.6. The value of the anchoring energy strongly
`
`depends on the value of the parameter in.
`
`4. Conclusions
`
`layer with
`We have shown that irradiation of rubbed PI
`depolarized UV light can be used to control the anchoring
`energy of LC‘ on rubbed Pl surface. Unpolarizcd exposure
`does not change the angular distribution function of the PI
`lragments.
`It does change the anchoring energy of the PI
`layer without reorientation of the easy axis on the Pl surface.
`The combination of irradiation with depolarized and polar-
`ized light can be used for measuring the angular distribution
`of PI fragments.
`
`Acknowledgments
`
`This work was supported by the National Science Foun-
`dation Science and Technology Center for Advanced Liq-
`uid Crystalline Optical Materials (ALCOM), grant DMR
`89-20147. the Ukrainian State Fund for Fundamental Slud-
`ies Support, grant No. #358-97-13, grant No. l329:’l3, grant
`UB1-315 of CRDF Cooperative Grants Program. DA. ac»
`knowledges support through grant PSU()82002 oi" [menia-
`Ijonal Sores Science Education Program and Overseas Re-
`search Students Award No. 0RSi’990070 I 5.
`
`1) J. Cognttrtl: Aflgrinit-iii ofNcmm|'t.' i'..‘qrm'r.f Cr_v.ri'r.ri‘.~' win’ T'iit'ir ii-i'i'.\'ri.W.\‘.
`Moi. Crysl. :3: Liq. Cryst. Suppl. (1982).
`2) M. Mahajan and C. Rosenblatt: J. Appl. Phys. 83 t_l‘}‘Jii) 75-19.
`3} M.Geary, J. W.Cioot|by, A. R. Kmetz and]. S. I-‘alt.-l.' J. Appl. Phys. 62
`(198?) -1 I00.
`41 A. J. Pidtiuck. G. P. !.lry:m—1Srnwn. S. D. lriiual-um and it. Buiiitisier: Liq.
`Crysl. 21 (I996) 759.
`5] A. J. Piddut:k.(i. P.Bryan—Brown_. S. D. Haslam. R. Bannister. I. Kilcly.
`'1'. J. McMa-sler and L. Iioogaau-tl:
`J. Vac. Sci.
`l‘ci:l1no1. A H {I995}
`1T3].
`6} M. liasegawa and Y. "|‘iaia: J. Phuuipolyln. Sci. 'l‘echnol. 3 (1995) ‘£03.
`1'}
`J. Chen. D. I..Johnson. P. J. Bos. X. Wang and J. l.. West: Phys. Rev. [-1
`54[lS'9(I] 1599.
`8} Y. Wang. A. Kanezawa. T. Shiono. T. iltctla. Y. Mutsuki and Y. 'l‘akeuelii:
`Appl. Phys. Lctt. T2 (199?) 345.
`9) J.-H. Kim. S. Kumar and S.-D. Lee: Phys. Rev. E 51 (I998) 5644.
`10) This assumption is valid only for relatively small UV domes. During
`exposure polymer chains grow shorter which could result in increasing
`their Inability and reerieruatioii. broadening their angular dislribiilion.
`II} M. Nisliikawa. B. Talicri and J. L. West: Anni. I‘hys. Letl.
`'.-'2 (19934:
`2403.
`I2) M. Nishlkawa M, J. L. West and Yu. Reznilcov: Liq. Cryst. 261.199‘-J;
`5'35.
`I3) P. G. De Genomes and J. Frost.‘ iilie i"hy.n'c.9 njLiquio'C't3'.uoJs(19931.
`Id] The authors ufthe rel‘.
`l I observed an appearance ofa domain slilirenv
`texture on the Pl surfaces irradiated with unpolarized light.
`
`3|]
`
`25
`
`It.‘5'
`
`
`
`GReorientationangle.dog 3
`
`5
`
`.-
`
`_
`
`Combined planar coll
`Tested - rubbed Pl 4- UV
`
`E_. :10. w: C|.5.1:=6.B min
`
`D
`
`‘ID
`
`20
`
`SD
`
`40
`
`50
`
`Exposure time, min
`
`Fig. 3. Twist angle vs exposure time, geometry with polarized UV light.
`Different curves are plotted for dilTerenl widths of lhe Pi blocks distribu-
`liuns.
`
`Page 4 of 4
`
`Page 4 of 4